Continental Lithosphere - Deep Seismic Reflections PDF

Continental Lithosphere:
Deep Seismic
Geodynamics
Geodynamics Series
I. Dynamics of Plate Interiors
A W. Bally, P. L Bender, T. R. McGetchin, and R. I. Walcott (Editors)
2. Paleoreconstruction of the Continents
M. W. McEIhinny and D. A. Valencio (Editors)
3. Zagros, Hindu Kush, Himalaya, Geodynamic Evolution
FL K. Gupta and F. M. Delany [Editors]
4, Anelasticity in the Earth
F. D. Stacey, M. & Patterson, and A. Nicholas [Editors]
5. Evolution of the Earth
R. J. O'Connell and W. & Fyfe [Editors]
(x Dynamics of Passive Margins
R. A. Scrutton [Editor]
7. Alpine-Mediterranean Geodynamics
FL Berckhemer and K. Hs( [Editors]
8. Continental and Oceanic Riffs
G. P(lmason, P. Mohr, K. Burke, R. W. Girdler, R. J. Bridwell, and G. E Sigvaldason [Editors]
9. Geodynamics of the Eastern Pacific Region, Caribbean and Scotia Arcs
Rmon Cabr, & J. [Editor]
0. Profiles of Orogenic Belts
N. Rast and F. M. Delany [Editors]
11. Geodynamics of the Western Pacific-lndonesian Region
Thomas W. C. Hilde and Seiya Uyeda [Editors]
2. Plate Reconstruction From Paleozoic Paleomagnetism
R. Van der Voo, C R. Scotese, and N. Bonhommet [Editors]
13. Reflection Seismology. A Global Perspective
Muawia Barazangi and Larry Brown [Editors]
'14, Reflection Seismology. The Continental Crust
Muawia Barazangi and Larry Brown [Editors]
'15. Mesozoic and Cenozoic Oceans
Kenneth J. Hs( [Editor]
'16. Composition, Structure and Dynamics of the Lithosphere-Asthenosphere System
K. Fuchs and C. Froidevaux [Editors]
'17. Proterozoic Lithospheric Evolution
A. KrOner [Editor]
'18. Circum-Pacific Orogenic Belts and Evolution of the Pacific Ocean Basin
J. W. H. Monger and J. Francheteau [Editors]
'19. Terrane Accretion and Orogenic Belts
Evan C Leitch and Erwin Scheibner [Editors]
20. Recent Plate Movements and Deformation
K. Kasahara [Editor]
2'1. Geology of the USSR: A Plate-Tectonic Syn
L. P. Zonenshain, M. I. Kuzmin, and L. M, Natapov B.M. Page
Continental Lithosphere:
Deep Seismic Reflections
Rolf Meissner
Larry Brown
Hans-JOrgen DOrbaum
Wolfgang Franke
Karl Fuchs
Friedrich Seifert
Editors
Geodynamics Series
Volume 22
Publication No. 183 of the International Lithosphere Program
American Geophysical Union
Washington,
Published under the aegis of the AGU Books Board.
Library of Congress Cataloging-in-Publication Data
Continental lithosphere : deep seismic reflections / Rolf
Meissner...[et al.], editors.
p. cm. -- (Geodynamics series: v. 22)
"The 4th International Symposium on Deep Reflection
Profiling of the Continental Lithosphere, Bayreuth, Germany, 4-
7 September, 1990"-- Intro.
ISBN 0-87590-522-6
1. Earth--Crust--Congresses. 2. Seismic reflection method-
-Congresses. I. Meissner, Rolf. II. International Symposium on
Deep Reflection Profiling of the Continental Lithosphere (4th:
1990: Bayreuth, Germany) III. Series.
QE511.C67 1991
551.1 '3--dc20 91-34553
CIP
ISBN 0-87590-522-6
Copyright 1991 by the American Geophysical Union, 2000 Florida Avenue, NW,
Washington, DC 20009
Figures, tables, and short excerpts may be reprinted in scientific books and
journals if the source is properly cited.
Authorization to photocopy items for internal or personal use, or the
internal or personal use of specific clients, is granted by the American
Geophysical Union for libraries and other users registered with the Copyright
Clearance Center (CCC) Transactional Reporting Service, provided that the
base fee of $1.00 per copy plus $0.10 per page is paid directly to CCC, 21
Congress Street, Salem, MA 01970. 0277-6669/87/$01. + .10.
This consent does not extend to other kinds of copying, such as copying
for creating new collective works or for resale. The reproduction of multiple
copies and the use of full articles or the use of extracts, including figures and
tables, for commercial purposes requires permission from AGU.
Printed in the United States of
CONTENTS
Introduction L. Brown and R. Meissner xi
NORTH AMERICA
COCORP Seismic Investigations of the Surrency Bright Spot beneath the Southern Appalachians,
USA T. L. Pratt, E. Hauser, T. Hearn, and T. Reston 1
Deep Seismic Reflection Constraints on Paleozoic Crustal Structure and Moho from the Buried
Southern Appalachian Orogerk,..J.H. McBride 9
High-Resolution P- and S-Wave Deep Crustal Imaging across the Edge of the Colorado Plateau,
USA: Increased Reflectivity caused by Initiating Extension J.M. Howie, T. Parsons, and
G. A. Thompson 21
Crustal Anisotropy and the Structure of the Mohorovicic Discontinuity in Western Nevada
of the Basin and Range Province R. Carbonell and S. B. Smithson 31
Seismic Reflectors in High-Grade Metamorphic Rocks of the Kapuskasing Uplift: Results of
Preliminary Drill Site Surveys B. Milkereit, J. A. Percival, D. White, A. G. Green, and
M. H. Salisbury 39
Seismic Reflectivity Patterns of the Kapuskasing Structural Zone J. Wu and R. F. Mereu 47
The Structure of the Archean Crust in SW Greenland from Seismic Wide-Angle Data: A
Preliminary Analysis K. Gohl, R. B. Hawman, S. B. Smithson, and Y. Kristoffersen 53
EUROPE
Late- amd Post-Orogenic Evolution of the Crust Studied from ECORS Deep Seismic Profiles
C. Bois and ECORS Scientific Party 59
The DEKORP Surveys: Major Achievements for Tectonical and Reflective Styles
R. Meissner and the DEKORP Research Group 69
Reflectivity of a Proterozoic Shield: Examples from BABEL Seismic Profiles across Fennoscandia
BABEL Working Group 77
Reflectivity Variations of Variscan Terranes in Germany R. Bittner and T. Wever 87
Structure of the Saar-Nahe-Basin (SW-Germany) from DEKORP Profiles 1-C and 9-N
A. Henk 91
Seismic Anisotropy of the Earth's Crust of the Urals and its Possible Relation to Oriented
Cracking and to Stress State S. Kashubin 97
INTEGRATED SEISMIC EXPERIMENT OBERPFALZ 1989
3-D Vertical Incidence Seismic Reflection Survey at the KTB Location, Oberpfalz M. Stiller
CONTENTS
Investigation of Structures and Anisotropy by a 3D-Expanding-Spread Experiment H. Wiederhold 115
Integrated P- and S-Wave Borehole Experiments at the KTB-Deep Drilling Site E. Liischen,
W. S611ner, A. Hohrath, and W. Rabbel 121
Processing and Analysis of MSP-Experiments within the KTB-Project M. Kemper
and H. P. Harjes 135
3-D Wide-Angle Investigations in the KTB Surroundings as part of the "Integrated Seismics
Oberpfalz 1989 (ISO89)", First Results H. Gebrande, M. Bopp, M. Meichelb6ck, and
P. Neurieder 147
THE ALPS
Crustal Indentation in the Alps--An Overview of Reflection Seismic Profiling in Switzerland
P. Heitzmann, W. Frei, and P. Lehner 161
Crustal Reflections beneath the Alps and their Foreland; Geodynamic Implications
J. L. Mugnier and J. M. Marthelot 177
Crustal Reflections from the Alpine Orogen: Results from Deep Seismic Profiling
O. A. Pfiffner, L. Levato, and P. Valasek 185
Integrated Analysis of Normal Incidence and Wide-Angle Reflection Measurements across the
Eastern Swiss Alps ETH Working Group on Deep Seismic Profiling 195
Seismic Data from the Alps and their Bearing on the Formation of Granulite and Eclogite Nappes
H. Laubscher 207
SOUTH AFRICA
A Deep Seismic Reflection Profile across the Archean-Proterozoic Witwatersrand Basin, South
Africa R. J. Durrheim, L. O. Nicolaysen, and B. Corner 213
CONTINENTAL MARGINS
Deep Penetration Seismic Reflection Profiling across the Southeastern United States Continental
Margin J. Oh, J. D. Phillips, J. A. Austin, Jr., and P. L. Stoffa 225
Images of the Plate Boundary beneath Southern Alaska T.M. Brocher, M. J. Moses,
M. A. Fisher, and C. D. Stephens 241
LITHOPROBE Reflection Transect of Southwestern Canada: Mesozoic Thrust and Fold Belt
to Mid-Ocean Ridge F. A. Cook, J. L. Varsek, and R. M. Clowes 247
Seismic Structure of the Northern Cascadia Accretionary Prism: Evidence from New Multichannel
Seismic Reflection Data G. D. Spence, R. D. Hyndman, E. E. Davies, and C. J. Yorath 257
First Deep Seismic Reflection Transect from the Gulf of Lions to Sardinia (ECORS-CROP Profiles
in Western Mediterranean) B. De Voogd and ECORS Group 265
Structure and Evolution of two Adjoining Segments of the West African Margin from Deep
Seismic Profiling J. Wannesson, J. C. lcart, and J. Ravat
CONTENTS
COMPARATIVE AND GENERAL STUDIES
Reflectivity Caused by Texture-Induced Anisotropy in Mylonites S. Siegesmund,
M. Fritzsche, and G. Braun 291
Correlating of Crustal Reflections with Geologic Outcrops: Seismic Modeling Results from the
Southwestern USA and the Swiss Alps R. K. Litak, R. H. Marchant, L. D. Brown, O. A. Pfiffner,
and E. C. Hauser 299
Reflections from a Relic Moho in Scotland D.B. Snyder 307
The Reflection Moho along the COCORP Northwest USA Transect E.I. Prussen 315
Upper Crustal Poisson's Ratios in the Colorado Plateau, Northern Arizona, from Multi-Component
Wide-Angle Seismic Recording R. A. Johnson and K. A. Hartman 323
Velocity-Resistivity Correlations in the Deep Crust G. Marquis and R. D. Hyndman 329
Determination of the fine Structure of Crustal Boundaries from their Reflected Phases R. Carbonell
and S. B. Smithson 335
Lower Crustal Lithology from Shear Wave Seismic Reflection Data G. Ward, M. Warner, and
the BIRP Syndicate 343
Layers Thicknesses in the Lower Crust: Modelling and Spectral Analysis of BIRPS Data A. Paul
and R. Hobbs 351
Constructive Interference-Geophysical Mythology Re-Examined B. $paargaren and M. Warner 359
Comparative Investigations of Reflectivity Based on Deep Seismic Reflection Data from Europe
and the United States of America P. Sadowiak, R. Meissner, and L. D. Brown 363
Seismic Patterns across Crustal Suture Zones Th. Wever and P. Sadowiak 371
Deep Seismic Profiling and Continental Evolution K.D. Nelson 377
METHODICAL AND TECHNICAL DEVELOPMENTS
Multiple Suppression in Deep Water R. J. J. Hardy and R. W. Hobbs 383
A Complete Waveform Inversion and its Application to ECORS Data S.C. Singh, M. Dietrich,
and F. Chapel 391
Geometrical Migration of Line-Drawings: A Simplified Method applied to ECORS Data G. Snchal
and F. Thouvenot 401
Energy- and Power Sections in Seismic Interpretation R. Bittner and W. Rabbel 409
Importance of Selecting the Most Effective Processing Sequences: Reprocessing of Kapuskasing
Crustal Reflection Data Z. Hajnal, D. Scott, B. I. Pandit, and R. Reilkoff 417
Comparison of Coincident High-Resolution Wide-Aperture and CDP Profiling along the Southwest
Coast of Norway S. J. Deemer and C. A. Hurich 435
Source-Generated Noise in Marine Seismic Profiles: The Limits of Reflection Detectability in the
Upper Crust C.A. Hurich
Acknowledgments
Thanks are due to all reviewers who helped to improve the manuscripts and to make this publication a success.
E. Banda, Barcelona, Spain
H.-J. Behr, G6ttingen, FRG
D. J. Blundell, Egham, U.K.
C. Bois, Rueil-Malmaison, France
R. Clowes, Vancouver, Canada
G. Dohr, Hannover, FRG
H. Dfirschner, Hannover, FRG
H. A. K. Edelmann, Hannover, FRG
J. Fertig, Hannover, FRG
D. Finlayson, Canberra, Australia
P. Firbas, Brno, CSFR
E. Fifth, Kiel, FRG
H. Gebrande, Mnchen, FRG
P. Giese, Berlin, FRG
A. Green, Ottawa, Canada
H.-P. Harjes, Bochum, FRG
K. Hinz, Hannover, FRG
R. v. Huene, Kiel, FRG
C. Juhlin, Perth, Australia
F. Keller, Clausthal-Zellerfeld, FRG
H. Kern, Kiel, FRG
S. Klemperer, Stanford, U.S.A.
P. Lehner, Richterswil, Switzerland
E. Lschen, Karlsruhe, FRG
R. Marschall, Hannover, FRG
J.-M. Marthelot, Strasbourg, France
P. Matte, Montpellier, France
B. Milkereit, Ottawa, Canada
H. Miller, Bremerhaven, FRG
W. Mooney, Menlo Park, U.S.A.
G. Mller, Frankfort, FRG
K. D. Nelson, Ithaca, U.S.A.
J. Oliver, Ithaca, U.S.A.
R. A. Price, Kingston, Canada
C. Prodehl, Karlsruhe, FRG
W. Rabbel, Kiel, FRG
C. Reichert, Hannover, FRG
K.-J. Sandmeier, Karlsruhe, FRG
S. B. Smithson, Laramie, U.S.A.
D. B. Snyder, Cambridge, U.K.
K. Weber, G6ttingen, FRG
F. Wenzel, North Ryde,
THE 4TH INTERNATIONAL SYMPOSIUM ON DEEP
REFLECTION PROFILING OF THE CONTINENTAL
LITHOSPHERE
Bayreuth, Germany, 4-7 September 1990
An Introduction by L. Brown and R. Meissner
With the spirit of the Nibelungen lingering from The first morning was devoted to marine
the annual Richard Wagner Festival which had just investigations along shelf areas and across the
ended, 200 scientists from 23 countries met at continent-ocean transition at passive margins. New
Bayreuth University, only a mile away from this images of seaward dipping layers at a passive
historic and picturesque city, to renew their margin, this time off the coast of Argentina (K.
attack on old problems with Gala. Hintz), and rifted crust off Nova Scotia (Ch.
Keen/LITHOPROBE) and in the Gulf of Lions (B.
The 4th International Symposium on Deep Seismic DeVoogd/ECORS-CROP) continue to refine our under-
Reflection Profiling of the Continental Lithos- standing of the initial stages of continental
phere, like its predecessors at Cornell University breakup and ocean formation. New results across
in the U.S. (1984), Cambridge University in Great active compressional margins of western North
Britain (1986) and Canberra, Australia (1988), America were shown (M. Talwani/EDGE, Spence), but
provided a rich review of new tectonic discoveries perhaps some of the most intriguing new sections
from, and technical developments of, seismic re- were from major transform boundaries. Reflection
flection profiling of the continental lithosphere. transects straddling a transform fault off the
Effective national programs in various countries, African coast (J. Wannesson) revealed dramatic
as well as individual initiatives are responsible differences in crustal extension, while EDGE
for the success of this series of meetings and have profiles off the California Coast (A. Levander)
developed a truly international basis for coopera- demonstrated the interaction between near-vertical
tion. The application of reflection techniques to strike slip and low angle detachments.
basic geological problems has provided a wealth of
new data on crustal structure and evolution.
One of the highlights of the conference was the
The meeting in Bayreuth consisted of 90 oral presentation of results from the BABEL transects in
presentations, numerous poster displays, an excur- the Baltic Sea and Gulf of Bothnia. This coopera-
sion to the German Continental Deep Drilling site tire venture between BIRPS, Geophysics Kiel, and
and various pre- and post- conference excursions. the Scandinavian countries delineated an early
It was preceded by a three-day workshop on Proterozoic suture in the north, upper crustal dyke
"Paleozoic Orogens in Central Europe" in the swarms, complex variations in lower crustal and
framework of IGCP project 233 in Gttingen and by Moho reflectivity and some mantle reflections (R.
another workshop on progress on "Controlled Source Meissner, D. Snyder). It showed the Tornquist Zone
Seismology" in Kiel. It was followed by an opening in the south to be part of an inverted basin with
ceremony for the deep hole of the German Conti- a significant bulge of the lower crust. By using
nental Deep Drilling Program (KTB) near Bayreuth huge tuned airgun and geophone arrays, seismic
and an international seminar on Superdeep Drilling energy was recorded out to 700 km offset, providing
in Regensburg. Except for one day with many wide angle-data for velocity estimation at great
technical contributions, no parallel session took depths.
place. The presentations were arranged according to
regional and tectonic aspects, giving the national
working groups a chance to display their results in Another highlight was the comprehensive review
a homogeneous and concise way. of Swiss (NFP-20), French (ECORS) and
xii INTRODUCTION
(CROP) results across the Alps. A whole afternoon shear (T. Reston). Following the technical sessions
was devoted to reviewing the reflection signature and poster (TERRA Abstracts 2, 1990; Deep
and tectonic fine structure of the Alps (P.Lehner, Reflection Profiling, 151-207), attenders moved to
J. Mugnier, O. Pfiffner, A.Steck). Specialized pro buses for a tour of the KTB drillsite followed by
cessing (L. Levato), lab measurements of rock a dinner at a nearby restaurant whose revolving
properties (S. Sellami), seismic modeling (R. floor helped everybody to have a view on the KTB
Marchant, R. Litak) and integrated analysis of rig once a whil, also kept the scientific
near-vertical and wide-angle results (J.Ansorge) discussions moving.
constituted an especially broad approach to the
interpretation of the Alpine transects. The seismic On Thursday, the 3 day of the symposium, an
results have clearly stimulated renewed discussion overview of reflection work in the Soviet Union (N.
of Alpine lithospheric evolution, with emphasis on Pavlenkova) started the presentations on reflection
the nature of crustal shortening and possible studies in Asia and Eastern Europe. Wide-angle work
interfingering of lower crust and upper mantle (P. seems still to form the backbone of crustal studies
Heitzmann, H. Laubscher). in many eastern countries (K. Kaila) with the
notable exception of Czechoslovakia where an
A public lecture in the evening on the German extensive near-vertical program has demonstrated
Continental Deep Drilling Project (KTB) by the fine structure of the Carpathians and the
Emmermann provided a natural prelude to the more Bohemian Massif (C. Tomek), and Hungary, where
than 3 hours of the second day that were devoted to ample reflection studies have revealed a shallow
the European Variscides. Beginning with an intro- crust with some strong reflections from the mantle
duction to the new DEKORP Atlas (R. Meissner), this (K. Posgay). The Bohemian Massif around the KTB-
session provided an overview of the great achieve- drilling site was also studied by ample wide-angle
ments of the first 5 years of DEKORP's operation. and near vertical reflection work in a kind of site
Seismic imaging of Variscan sutures (Weber, Franke) selection (DEKORP, Gebrande). Reflection work for
and the Rhenish Massif (Ch. Reichert, O. Oncken) the study of the Urals (V. Sokolov, S. Kashubin)
indicated that compressional structures are still will undoubtedly become more prominent in the
preserved, at least in the rigid upper crust, while future.
the lower crust is often affected by post-Variscan
extension with the formation of a laminated lower In a parallel session, a group of presentations
crust. A review of COCORP's results in the on technical aspects of reflection profiling re-
contemporaneous Appalachians (J. McBride) added a ported attempts to map reflectivity patterns (P.
comparative perspective for the central European Sadowiak), to identify the origin of specific
efforts. reflections (J.Schmoll), to better display
structural variations using displays of energy and
The rest of the second morning was devoted to related seismic attributes (R. Bittner), to
presentations by the Canadian Lithoprobe group. The discriminate source-generated noise in marine
range of topics covered the excellent data about surveys (C. Hurich), to measure reflection polarity
the Mid-Continental Rift (obtained by marine (Ch. Rathleff, T. Pratt) and velocity (C.
techniques in Lake Michigan), new interpretations Boennemann), and to apply full waveform inversion
of the Kapuskasing uplift (now apparently a shallow techniques (S.Singh). These presentations confirm
angle outcrop), combined geophysical surveys in the that reflection data contain an abundance of
east and new terrane studies in the west of Canada. information which has yet to be fully exploited.
The interdisciplinary nature of LITHOPROBE studies
was emphasized (R. Clowes); for example, the
Vancouver Island traverse integrates The afternoon sessions were introduced by a re-
electromagnetic and seismic studies to relate view of the Late and Post-orogenic evolution of the
reflection patterns with conductivity anomalies. It European crust as revealed by ECORS and other deep
was further argued (R.Hyndman) that the apparent seismic results (C.Bois), which emphasized the
correlation between lower crustal reflections and problem of what happened to the mountain roots and
high conductivity is underappreciated by the the asymmetry of Tertiary rifting. The new ECORS-
seismic community, an assertion which led to DEKORP results reveal just such an asymmetry in
substantial debate. The presentations were rounded both the northern and southern Rhinegraben (M.-
up by a discussion on the role of deep crustal and A.Gutscher), although of differing polarity. The
mantle reflections in accommodating lithospheric discussion of basin structure and
L. BROWN AND R. MEISSNER xiii
continued with reports on preliminary results from variously as a Caledonian suture (D. Snyder), and
a new ECORS survey in the Aquitaine Basin of ore relics of delamination (M. Warner). Offsetting
southern France (M.Daignieres), on a survey across relations among BIRPS deep crustal reflections were
the Archean-Proterozoic Witwatersrand Basin in used to construct the sequential history of struc-
South Africa (N. Nicolaysen), and on a multi- tural development in the lower crust (S.Klemperer).
disciplinary analysis of Porcupine Basin west of From Norway a study of coincident near-vertical and
Ireland (M.Tate).
Later in the afternoon attention turned to
North-America, beginning with a status report on
COCORP's latest efforts (L. Brown), including a
lithosperic transect of the intracratonic Williston
Basin and a combined P and S vibroseis study of a
midcrustal "bright spot" in the southern Appa-
lachians. Shear waves were also an important part
wide angle results (S. Deemer) again demonstrated
the relative and complementary merits of these two
modes of recording. The session ended with a pre-
sentation of a comprehensive model for craton
evolution, that attempts to relate seismic reflec-
tion patterns with models of lithospheric under-
plating and delamination (D. Nelson).
The afternoon session returned to North America
of studies across the Basin and Range-Colorado with a Wagnerian review of seismic reflection and
Transition zone (R. Johnson, J. Howie), and, as refraction results in the U.S. (W. Mooney) with the
reported in other sessions, were the basis of Leitmotiv of emphasizing application of various
experiments in the Basin and Range (R. Carbonell), geophysical techniques to crustal research.
Britain (G. Ward), and Germany (E. Lueschen). This Terranes in Alaska and the Sierra Nevada seem to
session also included a discussion of the tectonic have lost parts of their lower crust, while the
implications of deep seismic results from the Colorado Plateau transition indicates a low V, in
McKenzie delta in northern Canada (R. Stephenson), the lower crust with extremely poor reflectivity
OBS results off Lofoten, Norway, and expanding (J. McCarthy). The value of seismic modeling was
spread studies from the eastern Mediterranean (C. illustrated with examples from the southwestern
Truffert). U.S. and the Alps (R. Litak). Two more critical
In a parallel technically oriented session
tuning effects (B. Spaargaren), migration (G.
Senechal), attribute analysis (J. Marthelot) and
overall processing strategies (Z. Hajnal) were
comparisons of wide-angle and near vertical results
(S. Mangino, C. Gohl) completed the first part of
the afternoon sessions.
The final session of the meeting was devoted to
discussed. A presentation on the relative merits of DEKORP's detailed investigations around the KTB-
vibroseis vs dynamite (B. Damotte) provoked con- borehole, beginning with an introduction to the
siderable discussion, with a growing consensus that planning, acquisition and first results of the
explosives provide better penetration for lower major 3-D seismic survey (M. Stiller), the first
crustal while vibroseis provides greater resolution such experiment to be carried out in outcropping
of upper crustal structure. The implication of crystalline. Strong anisotropies and weak reflec-
borehole results (C. Weber, C. Juhlin), spectral tivity in the uppermost crust reduce the possi-
modeling (A. Paul) and anisotropy (S. Siegesmund) bility for a powerful stacking, but several new
for the interpretation of crustal reflections were attempts like energy displays and azimuth stacking
explored, and gravity results were used to argue are promising. Four more papers described special
for a new deep seismic profile in Algeria (M. anisotropy experiments (H. Wiederhold), P- and S-
Djeddi). wave borehole studies (E. Lueschen), the outcome of
The last day of the meeting began with a concise
review of BIRPS' activities, including an announce-
ment that their seismic atlas will soon be availa-
ble and plans for future surveys in the Banda Sea
and Mediterranean (R. Hobbs). Among the recent
moving source profiling (M. Kemper) and the
technical status of borehole-geophone chains (J.
Mylius).
In conclusion, the Bayreuth reflection symposium
scientific achievements are shear-wave studies of demonstrated the impressive and continuing progress
deep crustal reflectivity by using explosives on in crustal studies using seismic techniques, both
land (G. Ward), an attempt to use a two-ship ex- in terms of tectonic resolution and technical
periment for amplitude-offset studies (C. Peddy), refinement. Some of the thematic highlights are
areal "mapping" of mantle reflections, interpreted summarized as
xiv INTRODUCTION
(1) Better correlations between geological
outcrops and deep-penetrating reflections,
especially faults
(7) Use of Vibrators and explosives for irapro
ring resolution in upper and lower crust.
(2) More attempts to determine Pisson' s
constant from simultaneous P- and S-wave
studies, using 3-component geophones and
correlations of Vp/V I to crustal petrology.
(3) Greater use of integrated near-vertical
(8) Use of huge, tuned airgun arrays for in-
creasing seismic signals for near vertical
and wide angle/refraction studies at sea,
replacing explosives even in studies of
the deeper lithosphere.
and wide angle recording, combining These and other highlights of the conference found
crustal fine structure and velocity. their entry also into the proceedings of the
symposium. About 60 papers were received until
(4) Attempts to map "reflectivity patterns" December 1990, and 48 papers were selected for
and their correlation to certain tectonic publication. Some dublication in densely populated
units or temperature regimes. and geologically exciting areas could not be
avoided. The division of papers into various
(5) Strong indication that laminated lower chapters follows closely that of the conference.
crust is younger than upper crust, at The editors believe that the selection represents
least in the Variscan-Caledonian areas. the current knowledge about the architecture of the
continental crust and the technique of obtaining
(6) Use of 3-D or network studies for impro- this information. The editors would like to thank
ring determination of fine structure, all contributors and reviewers for their effort and
including anisotropy.
COCORP SEISMIC INVESTIGATIONS OF THE SURRENCY BRIGHT SPOT
BENEATH THE SORN APPALACHIANS, USA
Thomas Pratt, Ernest Hauser, Thomas Hearn 1 , Timothy Reston 2
Institute for the Study of the Continents, Cornell University, Ithaca, New York, USA
Abstract. An unusually strong midcrustal reflector, the Surrency Bright
Spot (SBS), was found at a depth of approximately 16 km during earlier
COCORP profiling in the southeastern U.S. In addition to being very
bright, the SBS is notable for being unusually flat and horizontal for
about half of its 4 km length. As these characteristics are similar to those
of fluid-caused reflections at shallow depths, it has been suggested that the
SBS may be caused by in situ midcrustal fluids. If caused by fluid
enclosed in fracture porosity in solid rock, the reflection would be
expected to exhibit a negative polarity from the top of the porous zone as
well as showing anomalous amplitude-versus-offset reflectivity. A series
of seismic reflection experiments have therefore been undertaken by
COCORP to measure the reflection properties of the SBS. A small
reflection seismic experiment with an explosive source was conducted to
determine the polarity of the SBS. The experiment consisted of four
dynamite shots, each recorded at three receiver stations by Seismic Group
Recorders (SGR) borrowed from Amoco Production Company.
Comparison of the dynamite records with geophone polarity tests indicate
that the SBS is characterized by a positive reflection coefficient at its top.
This result itself does not negate the fluid hypothesis - a fluid-fluid
interface could cause the positive reflection as well as the 'flat-spot' nature
of the reflector. However, modeling indicates that the SBS waveform is
most simply interpreted as originating from a circa 120 m thick, high-
impedance layer with wavelet tuning contributing to the unusually large
amplitude. In this light, a fluid model becomes much more ad hoc.
Though the fluid hypothesis is not precluded by this experiment, we feel
that the SBS is more likely caused by a mafic or ultramafic layer, either
tectonically or magmatically emplaced. A much more extensive
experiment was undertaken in the summer of 1990 to obtain 3-component
CDP and expanding spread (ESP) profiles over the SBS, as well as
crossline and three dimensional control. Initial results show that the SBS
is well imaged on all of the P-wave data to 36 km offsets and is visible
on some of the S-wave data as well. In the third dimension the SBS is
subhorizontal and of comparable lateral extent as on the main profile.
INTRODUCTION
One of the most striking features yet imaged on COCORP reflection
seismic profiles is an extremely bright reflector found beneath the
southern Appalachians, termed the Surrency Bright Spot (SBS) (Figs. 1
and 2). The SBS is a relatively small (4 kin) reflector lying at 5.8 to 6
1 Now at Dept. of Physics, New Mexico State University, Las
Cruces, New Mexico 88003-0001
2 Now at GEOMAR, Wischhofstrasse 1-3, 2300 Kiel 14, F. R. G.
Continental Lithosphere: Deep Seismic Reflections
Geodynamics 22
o 1991 American Geophysical Union
sec traveltime (about 16 km depth) in the midcrust. The amplitude of the
SBS is about 10 dB greater than the surrounding reflectors; it is certainly
one of the brightest reflections imaged by COCORP to date. In addition
to its unusually large amplitude the SBS is curious because it is flat and
level for half of its length (Fig. 2).
Given these characteristics for the SBS, it has been suggested that the
extreme reflectivity is caused by a fluid-filled porous zone (Wille, 1987;
Brown and others, 1987). This hypothesis is by analogy with 'bright
spots' found on shallow seismic sections; such bright spots are
sometimes diagnostic of fluid- or gas-filled porous zones which can also
have 'flat-spots' due to the gravitational control on a fluid-fluid interface
within the porous zone (e.g., Backus and Chen, 1975; Ensley, 1984).
Though the nature of a fluid(s) at the depth of the SBS is unknown, free'
deep-crustal fluids have been recognized from deep drilling and
hypothesized from magnetotelluric studies (Kozlovsky, 1984; Gough,
1986; MacDonald, 1988; Bailey and others, 1989; Hyndman and Shearer,
1989). Presumably any such fluids causing the SBS would be trapped
within fracture porosity. A similar mechanism has been suggested to
cause a bright reflection beneath the Black Forest in Germany (Luschen
and others, 1987). The SBS lies within an hypothesized Late Paleozoic
suture zone delineated on the basis of drillhole, magnetic, and seismic
reflection data (Nelson and others, 1985). A possible source for the
fluids could thus be formational and/or metamorphic waters derived from
underthrust sedimentary rocks (e.g., Fyfe, 1986).
To determine the cause of its reflectivity, COCORP has recently
reexamined the SBS using a variety of seismic experiments. The first of
these was a small dynamite experiment carried out in late 1989, the goal
of which was to determine the reflection polarity of the SBS. In addition,
seismic wavelet modeling of the dynamite data was carried out to further
constrain the SBS geometry.. COCORP returned to the SBS in the
summer of 1990 for our most elaborate field effort to date. Shear-wave
(both transverse and radial) and P-wave expanding spread profiles, crossing
S-wave and P-wave common-depth point (CDP) profiles, and a P-wave 3-
D experiment were acquired using a 650-channel telemetered seismic
system with 8 large vibrators. The goals of these experiments were to
further constrain the physical characteristics of the SBS, measure the bulk
seismic properties of the crust, and test the applicability of shear-wave
vibrators for deep crustal work.
Analysis of the dynamite experiment is currently being completed, and
processing of the data from the larger experiment is in its initial stages.
This paper describes the results obtained to date.
POLARITY TEST
A key test of the fluid hypothesis is the polarity of the SBS reflection.
The top of a porous zone would presumably mark a downward decrease in
velocity and density (Koefoed, 1955; Gardner and others, 1974; Gregory,
1976; Carlson and Herrick, 1990), and thus have negative reflection
coefficien Other hypothesized causes of the SBS reflectivity (e.g., mafic
sill or ultramafic body), however, should produce a positive reflection
coefficient at the top of the SBS. Unfortunately, seismic reflection data
obtained with vibrators, such as the earlier COCORP survey, are plagued
2 SURRENCY BRIGHT SPOT
vibroseis data, true amplitude stacked section amplitude decay plot
0.0
?
GEORGIA '
, 8rigm Sot ,
I COCORP GA- 1
,
o krn oo -
lndex mad
lO km
0.0
5.0
10.0
12.0
log of ampiRude;
two dynamite traces, averaged
2O dB
Fig. 1. The Surrency Bright Spot (SBS) as imaged on a true-amplitude
version of the original COCORP stacked vibroseis section and on an
amplitude decay plot of the dynamite data obtained over the feature. No
scaling has been applied to the stacked section; only the shallow Atlantic
Coastal Plain strata and the exceptionally strong SBS are visible. The
amplitude decay plot shows that the SBS has an amplitude approximately
10 dB above background levels. Section is plotted at a vertical:horizontal
ratio of about 1:1 at 6 km/s.
with uncertainty regarding the reflection polarity because of the
complexity of the vibrator source and its interaction with the ground.
The polarity was therefore ambiguous on the previously acquired data.
To determine the reflection polarity of the SBS a small seismic
reflection experiment was conducted using four dynamite shots as a
source. Dynamite can be assumed to have an initial outward motion upon
detonation, and thus a compressional wave should be the first arrival at
any given subsurface reflector. Charges ranging in size from 11.4 to 20.5
kg (25 to 45 lbs) were placed at the bottom of four 19m holes (Fig. 3)
and tamped with coarse sand. Upon detonation, the resulting reflections
from the SBS were recorded on three Seismic Group Recorders (SGRs)
borrowed from Amoco Production Company. Receiver arrays were 91.5
m (300ft) in length and consisted of 12, 10Hz P-wave geophones. The
largest source-receiver offset was 3.5 Pan; the reflected energy therefore had
effectively vertical travelpaths and receiver arrays should not effect the
reflected waveforms. A total of 12 traces (4 shots into three receivers)
were recorded; static (including uphole) and normal moveout corrections
were applied to the traces and they were arranged in CDP order (Fig. 4,
center).
The onset of the SBS reflection package is apparent on the recorded data
as a sharp negative (leftward on the plot in Fig. 4, center) deflection
occurring at about 5.8 sec two-way traveltime (dashed line on Fig. 4).
Comparison with tap tests conducted on the geophones and with first
(refracted) arrivals in the data show that this negative deflection is in
response to an upward geophone motion at the onset of the SBS
reflection. The initial outward motion of the explosive source and the
upward geophone motion indicate that the upper surface of the SBS has a
positive reflection coefficient; this implies that the SBS is composed of a
relatively high-impedance (high velocity and/or density) material.
WAVELET MODELING
Though the results of the polarity test imply that the source of the SBS
has higher impedance than the overlying material, this does not actually
eliminate the fluid hypothesis. It is conceivable that the top of the
porous zone is non-reflective, perhaps because of a gradual onset of
porosity, and the inital reflection is actually coming from a fluid-fluid
boundary within the porous zone. Though this requires two fluids to be
present, it is the implication of the 'flat-spot' model proposed for the
SBS. Such a fluid model, however, would be characterized by another
positive reflection at the base of the porous zone, in contrast to a thin,
high-impedance body which would have a negative reflection coefficient at
its base.
To distinguish between these two possiblities, wavelet modeling of the
traces from the dynamite experiment was undertaken. The modeling
procedure consisted of convolving a source wavelet with a reflectivity
function composed of 'spikes' representing each of the interfaces. The
synthetic models were then compared with the observed field data.
The primary difficulty in this modeling procedure is obtaining a source
wavelet. In similar studies, the wavelet has commonly been taken from
the first arrivals recorded in the field (e.g., Goodwin and others, 1989). At
Surrency, however, there is considerable attenuation of the higher-
PRATT ET AL. 3
Surrency Bright Spot; vibroseis stacked section (scaled)
1 km
Fig. 2. Detail of the Surrency Bright Spot (SBS) on the original
vibroseis stacked section after the application of automatic gain control
(AGC). Note the nearly 2.0 km long level portion at the northern half of
the SBS and the apparent diffraction tail to the south. The reflection
lying approximately 0.5 see below the SBS is believed to be caused by
multiply-reflected energy. Section is plotted at a vertical:horizontal ratio
of about 1:1 at 6 km/s.
frequency seismic energy as it propagates through the shallow sedimentary
section; hence, the frequency specmun of the first arrivals is far wider than
that of the SBS (Fig. 5). To circumvent this problem, we extracted a
minimum-phase wavelet directly from the televent portion of the seismic
trace using a spectral factorizafion technique (e.g., Claerbout, 1976;
Yilmaz, 1987, appendix B-4). This process produces a seismic wavelet
with an amplitude speclmm identical to the input trace, but whose phase
spectrum has been rearranged to front-load the energy within the trace
(Robinson and Treitel, 1980). Examples of such wavelets extracted from
different portions of a seismic trace from the Surrency dynamite
experiment are shown in Fig. 5A.
Modeling using the wavelet extracted by the above process is valid
only if two important assumptions are accepted. First, it is assumed that
the dynamite produced a minimum-phase wavelet upon detonation. This
assumption is difficult to evaluate, but it is widely believed that explosive
sources produce a minimum-phase wavelet (e.g., White and O'Brien,
1974) and it can be shown that any ghosts or reverberations are also
minimum phase (Robinson and Treitel, 1980, chapter 11 and appendix
13-1). The second assumption is that the large-amplitude reflection from
the SBS is not altering the frequency spectrum of the trace, and thus the
wavelet extracted therefrom. In other words, the reflectivity series is
random and the amplitude spectrum of the trace is a scaled version of the
amplitude spectrum of the source wavelet. This assumption is the basis
for all predictive aleconvolution and seems justified here given that the
amplitude spectrum and wavelet extracted from windows both above (2.0-
4.0 sec) and including the SBS (5.0-7.0) are nearly identical (Fig. 5A).
Before modeling, the phase spectrum of the source wavelet must be
modified to account for the phase shift introduced by the geophones. The
geophones used in the experiment have a damping of 67% of critical, and
their phase response can be computed using the equation for a damped
harmonic oscillator:
Phi = 180 - tan'l[2Dfnf/(fn2-f2)]
where D=-damping (0.67), f= frequency, and fn = natural frequency of the
geophone (10 Hz). The minimum-phase wavelets extracted from the data
and the same wavelets after applying the geophone 'filter' are shown in
Fig. 6 after smoothing with a zero-phase, 4-40 Hz filter to remove high-
frequency components introduced by windowing in the wavelet extraction
process. The digital recorders and receiver arrays should have a negligible
effect on the phase of the recorded signal given the frequency range and
incidence angle of the SBS reflection.
Four types of models (Fig. 7, bottom) were used to create synthetic
traces for comparison with the observed field records. Based on the
polarity test, each of the models has a positive reflection coefficient at the
top of the SBS reflector. The first model is a simple boundary (single
interface model). Because the amplitude is arbitrary, there is only one
permutation of this model. The second model is of a thin, high-
impedance layer with a positive reflection coefficient at its upper boundary
and an equal but negative reflection at its base. Because the body can have
any thickness, a range of thicknesses from 10 msec to 50 msec traveltime
were tried; these traveltimes would represent layers between approximately
30 and 160 m in thickness (assuming a velocity of 6.5 km/sec within the
layer). Two variations on a fluid model were also tested. The first of
these ('fluid model 1') consisted of a positive reflector at the top of the
SBS, as determined from the polarity test, and a second positive reflector
of equal amplitude at a later traveltime representing the base of the porous
zone. The second fluid model is the same except that the lower reflector
(the base of the porous zone) has only half the amplitude of the upper.
As with the high-impedance model, the thickness of the hypothesized
body was examined by varying the traveltime difference between the
reflectors from 10 msec to 50 msec.
The results of the modeling exercise for 4 representative CDPs are
shown in Fig. 7, and all of the synthetic traces from the best-fitting high-
impedance and fluid models are shown next to the field data in Fig. 4.
The single-interface model produces a wavelet package which is
considerably shorter in duration than the observed SBS reflection sequence
(Fig. 7). We thus feel that multiple interfaces are involved in producing
the SBS reflection. This was expected because the limited lateral extent
of the SBS makes a half-space model unlikely.
The high-impedance model produces a reflection sequence whose
duration varies with the layer thickness, as one would expect (Fig. 7).
The models with 30 to 40 msec layer thickness produce a reflection
sequence very similar in duration, lobe spacing, and lobe amplitude to that
of the observed data except that the initial lobe is smaller in amplitude
than on the field records. Increasing the thickness of the model body
introduces added complexities to the reflection package in the later lobes.
Our best approximation to the SBS body was produced with a 36 msec
(120 m thick) high-impedance model with a lower reflection amplitude
0.7 times that of the upper (Fig. 4). This model has the added complexity
of unequal reflection coefficients on each boundary of the body, but this
N
station 220
.... I , ,
Surrency Brlgnt Spot
230 power 240 250
line
/, kl ,'l"l I I , , I , , I 'Il
approximate
scale
Fig. 3. Diagram of the reflection point coverage of the dynamite
experiment along the main north-south profile over the SBS. Receivers
locations are shown as triangles, shot locations are shown as asterisks,
and the raypaths are shown as straight lines. The source and receiver
locations were chosen to maximize the subsurface coverage of the SBS
while avoiding the noise associated with the powerline and road overpass.
4 SURRENCY BRIGHT SPOT
high-impedance model
observed data fluid model
o o o o o ...... rr o o o o cc o o o o o ......
Fig. 4. Single-fold field records from the dynamite experiment,
arranged by common-depth-point (CDP) with the same CDP number as
the original COCORP profile (Fig. 1). The dashed line at approximately
5.86 sex is the interpreted onset of the SBS reflection package. Note the
leftward deflection of the dynamite traces at this time. The 'high-
impedance' and 'fluid' model sections are discussed in the text. Data are
scaled to equalize the largest values in each trace.
A wavemete
00
Fig. 5. Seismic wavelets and corresponding amplitude spectra from
selected time windows from one trace of the dynamite experiment Note
that the pulse begins as a sharp wavelet with a broad frequency spectrum
(0.0-2.0 sec window) but quickly broadens as the higher frequencies are
attenuated. Most of the attenuation appears to be occurring in the upper
2.0 sex of traveltime. The wavelets were extracted from the seismic traces
using the Kolmogoroff method of spectral factorizafion (Claerbout, 1976),
could be caused by the presence of several thin layers or a shear zone
overlying or underlying the body, thus changing the reflection amplitude
on one boundary.
The fluid model containing two equal positive reflection coefficients
(fluid model 1) produces a reflection package whose duration is larger than
the observed reflection (Fig. 7). Fluid model 2, with a smaller reflection
coefficient on the second reflector, produces a reflection which is in good
agreement with the observed data when the reflection spacing is about 50
msec. Our best match of any of the fluid models was with a 54 msec
/L
Q
u
0000 ...... CUCU
0.0
0.0
0.2
Fig. 6. The minimum-phase wavelets extracted from the dynamite data
and the same wavelets after accounting for geophone phase characteristics.
The wavelets correspond to the recorded traces (observed data) shown in
Figure 4. Data are scaled to equalhe the largest values in each trace.
FRATr ET AL. 5
" synthetic traces synthetic traces
=1-I
1'1 I
&)l I hll fluM I fluid . hl-lmpl fld
CDP 500
11o
60
6.0 '
fluid
megel
u
0 0
synthetic traces synthetic traces
u e J "t:3 - high-impedance fluid
. fluid
high-Impedance fluid I fluid model I
- 18 layer model 2
CDP 51:3 ,:> layer m0dell I m0de, 2 CDP
o o o o o o o o o o
o o o o o o o o o olo o o o o
.0 (' ' ' / / ' i
single hlgh-lmpeclance fluid fluid
reflectivity series Interface layer model 1 model 2
used to compute the :
synthetic t. races
Fig. 7. Results for four CDPs of convolving the gcophonc-filtcred
source wavelets with single-interface, high-impedance, and two fluid
models. The traveltime differences of the reflectors in the latter three
models, which represent thin beds, were varied in 10 msec increments to
simulate different layer thicknesses. The model reflectivity functions are
shown in the box at the bottom. Data are scaled to equalize the largest
values on each trace.
reflection spacing (about 150 m layer thickness) and a lower reflection
coefficient is 0.5 times that of the upper (Fig. 4). The drawback to the
fluid hypothesis is that the polarity test requires that two or more fluids
be present, and that the top of the porous zone be relatively non-reflective.
Though not impossible, these constraints make the fluid model complex
in comparison with the relatively simple high-impedance model.
The modeling therefore demonstrates that several interfaces are n to
adequately model the SBS reflection. A reasonable approximation to the
SBS reflection package can be obtained using a simple double-interface,
120 m thick high-impedance model, although the amplitude characteristics
of the bounding surfaces must be slightly different. The fluid model
requires that the upper surface of the porous zone be essentially non-
reflective, that the porous layer be approximately 150 m thick, and that at
least two fluids are contained in the porous zone.
Because of its simplicity, we suggest that the SBS is most likely
caused by a thin (circa 120 m), high-impedance body lying within the
midcrust. Two likely alternatives are a mafic sill, perhaps related to the
Mesozoic rifting and associated basaltic rocks in the area, and an
ultramarie sliver which was tectonically emplaced during the Late
Paleozoic collision between North America and Africa. Either of these
could produce reflection coefficients (milo of reflected to incident wave
amplitudes) approaching 0.1 across a single boundary; given the tuning
effects of a 30 to 40 msec pair of interfaces the net reflection amplitude
could be significantly larger. It must also be kept in mind that the
{5 SUCY BRIGHT SPOT
flatness and horizontal attitude of the SBS also contribute to providing
excellent reflecting surfaces.
SHEAR AND P-WAVE EXPANDING-SPREAD PROFILES
During the summer of 1990, a much more extensive S- and P-wave,
expanding spread profiling (ESP) investigation of the SBS was undertaken
by COCORP. ESPs define the reflectivity characteristics of the reflector
under a range of illumination angles and, espially when shear-waves are
also used, are a sensitive indicator of the material properties of the
reflector (e.g., Koefoed, 1955; Ostrander, 1984; Ensley, 1984; Robertson
and Pritchett, 1985). ESP and shear-wave analyses are especially
sensitive to porous zones and have been used to identify fluids within the
shallow crust. Some workers have claimed to have been able to
distinguish between differing types of pore fluids (Yu, 1987).
Expanding-spread profiles were collected in all three wave modes (P,
SV, SH) out to offsets of 36 kin, resulting in incidence angles of nearly
50 at the SBS. Both ESP and normal-incidence CDP profiles were
collected in each wave mode by using separate 3-component receiver
arrays positioned near the vibrators as well as at large offsets. Four shear-
wave (with rotatable pads) and four compressional-wave vibrators were
used as sources. A crossline CDP profile in each wave mode and a P-
wave 3D experiment were collected to examine the geometry and attitude
of the SBS in more detail.
Initial results show that the SBS is extremely well imaged on all of the
P-wave data, including the ESP data at all offsets. The 3D and crossline
experiments show that in the third dimension the SBS is nearly horizontal
and has an extent that is about equal to that on the original profile (about
4 inn). The shear-wave reflectivity is more problematic, but a distinct
reflection is observed at the predicted traveltime for the SBS on the brute
stacks of the shear-wave data. This is among the deepest reflections yet
recorded from a shear-wave vibrator source. These preliminary results
indicate that in addition to constraining the SBS reflectivity, the
COCORP experiment will provide accurate measurements of the Vp/Vs
ratio within the upper and middle crust by observing traveltime differences
between the different wave modes.
CONCLUSIONS
An analysis of explosion data acquired over the Surncy Bright Spot
(SBS) shows that it has a positive reflection coefficient at or near its
upper surface. Waveform modeling demonstrates that two or more thinly
spaced reflectors contribute to the large reflection amplitude. A simple,
120 m thick high impedance layer, with a positive reflection at its upper
boundary and a nearly equal-sized negative reflection coefficient at its base,
provides a reasonable match to the observed data when modeled. Such a
model could represent a mafic intrusion or thin ultramafic sliver, either of
which could produce appropriate reflection coefficients. Fluid models
based on analogies with "flat spots" seen on shallow seismic sections can
also provide a reasonable match but require a fluid-fluid interface within a
-150 m thick porous zone which has a relatively non-reflective upper
surface. We currently favor a mafic intrusion or ultramafic sliver in the
midcrust as the most reasonable explanation for the SBS because of its
simplicity; additional data collected this past summer should resolve the
issue.
Acknowledgements. Funding for the data acquisition was provided
primarily by the National Science Foundation with additional support
from BP Exploration. Amoco Production Company generously donated
equipment and substantial personnel time to the study, and in particular
we thank John Myers, Dan Johnson, and Richard Heiser for their efforts.
Data were acquired by personnel from Grant-Norpac using equipment
provided by Amoco Production Company. The experiment took place on
the property of the Union Camp Corporation who kindly allowed us to
freely roam their property. Data processing was done primarily on the
Cornell Theory Center's IBM 3090-600E computers at the Cornell
National Supercomputer Facility with additional processing on the
COCORP MEGASEIS and Sierra Geophysics processing systems. Greg
Steiner, Bob Smalley, and Jer-Ming Chiu of the University of Tennessee,
Memphis, provided some of the field monitoring equipment as well as
instruction on its usage. The paper has benefited from discussions with
Art Barnes, Larry Brown, Sial Kaufman, Doug Nelson and Jack Oliver at
Cornell as well as Dave Okaya, Ron Clowes, and Tom Brocher. The
Consortium for Continental Reflection Profiling (COCORP) is supported
by National Science Foundation grant EAR-8916129. Institute for the
Study of the Continents (INSTOC) contribution number 155.
REFERENCES
Ake, J.P., and Sanford, A.R., 1988, New evidence for the existence and
internal structure of a thin layer of magma at mid-crustal depths near
Socorro, New Mexico, Bulletin of the Seismological Society of
America, v. 78, p. 1335-1359.
Akima, H., 1970, A new method of interpolation and smooth curve
fitting based on local procedures, Journal of the ACM, v. 17, p. 589-
602.
Almoghrabi, H., and Lange, J., 1986, Layers and bright spots,
Geophysics, v. 51, p. 699-709.
Backus, and Chen, 1975, Flat-spot exploration, Geophysical Prospecting,
v. 23, p. 533-577.
Bailey, R.C., Craven, J.A., Macnae, J.C., and Polzer, B.D., 1989,
Imaging of fluids in Arcbean crust, Nature, v. 340, p. 136-138.
Brown, L.D., Wille, D., Zheng, L., DeVoogd, B., Mayer, J., Hearn, T.,
Sanford, W., Caruso, C., Zhu, T.-F., Nelson, K.D., Potter, C.,
Hauser, E., Klemperer, S., Kaufman, S., and Oliver, J., 1987,
COCORP: new perspectives on the deep crust, Geophysical Journal,
v.89, p. 47-54.
Brown, L.D., Chapin, C.E., Sanford, A.R., Kaufman, S., and Oliver,
J.E., 1980 Deep structure of the Rio Grande Rift from seismic
reflection profiling, Journal of Geophysical Research, v. 85, p. 4773-
4800.
Carlson, R.L., and Herrick, C.N., 1990, Densities and porosities in the
oceanic crust and their variations with depth and age, Journal of
Geophysical Research, v. 95, p. 9153-9170.
Claerbout, J.F., 1976, Fundamentals of Geophysical Data Processing,
McGraw-Hill, New York, New York, 274 pp.
deVoogd, B., Serpa, L., and Brown, L.D., 1988, Crustal extension and
magmatic processes: COCORP profiles from Death Valley and the Rio
Grande rift, Geological Society of America Bulletin, v. 100, p. 1550-
1567.
Domenico, Slq., 1974, Effect of water saturation on seismic reflectivity
of sand reservoirs encased in shale, Geophysics, v. 39, p. 759-769.
Ensley, R.A., 1984, Comparison of P- and S-wave seismic data: A new
method for detecting gas reservoirs, Geophysics, v. 49, p. 1420-1431.
Frost, B.R., Fyfe, W.S., Tazaki, K., and Chan, T., 1989 Grain-boundary
graphite in rocks and implication for high electrical conductivity in the
lower crust, Nature, v. 340, p. 134-136.
Fyfe, W.S., 1986, Fluids in deep continental crust, in Barazangi, M., and
Brown, L., eds., Reflection Seismology: The Continental Crust,
American Geophysical Union Geodynamics Series 14, p. 33-40.
Gardner, G.H.F, Gardner, L.W., and Gregory, A.R., 1974, Formation
velocity and density - the diagnostic basics for stratigraphic traps,
Geophysics, v. 39. p. 770-780
Goodwin, E.B., Thompsen, G.A., and Okaya, D.A., 1989, Seismic
identification of basement reflectors: the Bagdad reflection sequence in
the Basin and Range province-Colorado Plateau transition zone,
Arizona, Tectonics, v. 8, p. 821-831.
Gregory, A.R., 1976, Fluid saturation effects on dynamic elastic
properties of sedimentary rocks, Geophysics, v. 41, p. 895-921.
Gough, D.I., 1986, Seismic reflectors, conductivity, water and stress in
the continental crust, Nature, v. 323, p. 143-144.
Hauser, E.C., Gephart, J., Latham, T., Oliver, J.E., Kaufman, S.,
Brown, L.D., and Lucchitta, I., 1987, COCORP Arizona transect:
Strong crustal reflections and offset Moho beneath the transition zone,
Geology, v. 15, p. 1103-1106.
Hufich, C.A., and Smithson, S.B., 1987, Compositional variation and
the origin of deep crustal reflections, Earth and Planetary Science
Letters, v. 85, p. 416-426.
Hyndman, R.D., and Hyndman, D.W., 1968, Water saturation and high
electrical conductivity in the lower continental crust, Earth and
Planetary Science Letters, v. 4, p. 427-432.
PRATF ET AL. 7
Hyndman, R.D., and Shearer, P.M., 1989, Water in the lower continental
crust: modelling magnetotelluric and seismic reflection results,
Geophysical Journal International, v. 98, p. 343-365.
Juhlin, C., 1988, Interpretation of the seismic reflectors in the Gravberg-1
well, in Boden, A., and Eriksson, K.G., eds, Deep drilling in crystalline
bedrock, Springer-Verlag, Berlin, p. 113-121.
Koefoed, O., 1955, On the effect of Poisson's ratio of rock strata on the
reflection coefficients of plane waves, Geophysical Prospecting, v. 3, p.
381-387.
Koslovsky, Y.A., 1984, The world's deepest well, Scientific American, v.
251, n. 6, p. 98-104.
Luschen, E., Wenzel, F., Sandmeier, K.-J., Menges, D., Ruhl, Th.,
Stiller, M., Janoth, W., Keller, F., Soilnet, W., Thomas, R.,
Krohe,A., Stenger, R., Fuchs, K., Wilhelm, H., and Eisbacher, G.,
1987, Near-vertical and wide-angle seismic surveys in the Black Forest,
SW Germany, Journal of Geophysics, v. 62, p. 1-30.
MacDonald, G.J., 1988, Major questions about deep continental
structures, in Boden, A., and Eriksson, K.G., eds, Deep drilling in
crystalline bedrock, Springer-Verlag, Berlin, p. 28-48.
Matthews, D.H., and Cheadle, M.J., 1986, Deep reflections from the
Caledonides and Variscides west of Britain and comparison with the
Himalayas, in Barazangi, M., and Brown, L.D., eds., Reflection
Seisinology: A Global Perspective, American Geophysical Union
Geodynamics Series, v. 13, p. 5-19.
McBride, J.H., Nelson, K.D., and Brown, L.D., 1989, Evidence and
implications of an extensive early Mesozoic rift basin and basalt/diabase
sequence beneath the southeast coastal plain, Geological Society of
America Bulletin, v. 101, p. 512-520.
Nelson, K.D., Arnow, J.A., McBride, J.H., Willemin, J.H., Huang, J.,
Zheng, L., Oliver, J.E., Brown, L.D., and Kaufman, S., 1985, New
COCORP profiling in the southeastern United States. Part I: Late
Paleozoic suture and Mesozoic rift basin, Geology, v. 13, p. 714-718.
Ostrander, W.J., 1984, Plane-wave reflection coefficients for gas sands at
nonnormal angles of incidence, Geophysics, v. 49, p. 1637-1648.
Ragland, P.C., Hatcher, R.D., Jr., and Whittington, D., 1983, Juxtaposed
Mesozoic diabase dike sets from the Carolinas: a preliminary
assessment, Geology, v. 11, p. 394-399.
Ricker, N., 1953, The form and laws of propagation of seismic wavelets,
Geophysics, v. 18, p. 10-40.
Robertson, J.D., and Pritchett, W.C., 1985, Direct hydrocarbon detection
using comparative P-wave and S-wave seismic sections, Geophysics, v.
50, p. 383-393.
Robinson, E.A., and Treitel, S., 1980, Geophysical Signal Analysis,
Prentice-Hall, Englewood Cliffs, New Jersey, 466 pp.
Shankland, T.I., 1989, A case of two conductors, Nature, v. 340, p. 102.
White, R.E., and O'Brien, P.N.S., 1974, Estimation of the primary
seismic pulse, Geophysical Prospecting, v. 22, p. 627-651.
Widess, M.B., 1973, How thin is a thin bed?, Geophysics, v. 38, p.
1176-1180.
Wille, D.M., 1987, The COCORP Surrency bright spot: fluid in the deep
crust?, M.S. Thesis, Cornell University, 46 pp.
Yardley, B.W.D., 1986, Is there water in the deep continental crust?,
Nature, v. 323, p. 111.
Yilmaz, O., 1987, Seismic Data Processing, Society of Explorations
Geophysicists, Tulsa, Oklahoma, 526 pp.
Yu, G., 1985, Offset-amplitude variation and controlled-amplitude
processing, Geophysics, v. 50, p. 2697-2708.
DEEP SEISMIC REFLECTION CONSTRAINTS ON PALAEOZOIC CRUSTAL
STRUCTURE AND DEFINITION OF THE MOHO IN THE BURIED SOUTHERN
APPAI.CHIAN OROGEN
John H. McBride 1 and K. Douglas Nelson
Institute for the Study of the Continents,
Snee Hall, Cornell University, Ithaca, New York 14853-1504, USA
Abstract. Experimental reprocessing of COCORP deep
seismic reflection data collected over the buried Southern
Appalachian orogen (southeastern USA) provides a more
accurate characterisation of middle/lower crustal structure and
Moho, thus allowing a quantitative basis for geologic
interpretation. Reprocessing and migration of data over the
late Palaeozoic suture (Alleghanian=Hercynian) between North
America and relict west Africa reveal varying north-vergent
thrust geometries including ramp-and-flat, antiformal, and
planar structures. Interpretation of a strike-parallel line over
the suture zone implies that the suture is dominated by discrete
bands of south-dipping thrusts which are expressed, in strike
view, as a "layered" fabric of sub-horizontal reflections. The
internal structure of the suture zone varies dramatically along
strike becoming broader and less steep eastward across the
Atlantic Coastal Plain as the amount of crustal "overlap"
between relict west AfriCan and North American terranes
progressively increases. Suture-zone reflections are abruptly
truncated in the lower crust by a ubiquitous sub-horizontal
reflection Moho which is anomalously shallow generally at 33-
36 km (relative to a much deeper Moho beneath the Blue Ridge
and Inner Piedmont to the west). The character of the
reflection Moho varies throughout the area and, after
deconvolution and frequency filtering, appears either as a
simple doublet/triplet or more complex multicycle event. This
multicyclic character may be suggestive of magmatic
underplating by mafic igneous sills intruded during a regional
thermal event brought on by early Mesozoic extension and
1 Now at Bullard Laboratories, Department of Earth
Sciences, Madingley Rise, University of Cambridge,
Cambridge CB3 0EZ, England.
Geodynamics 22
crustal thinning related to the initial rifting and opening of the
North Atlantic. The broad correspondence between a highly
reflective Moho underlying thinned crust and the region of
early Mesozoic rifting, together with its horizontal truncation
of Palaeozoic suture-zone structure, suggests that the Moho is
a dynamic boundary that is early Mesozoic in age and was
produced by rifting processes.
Introduction and Regional Setting
In the southeastern United States, the Appalachian orogen
disappears abruptly beneath the Atlantic Coastal Plain
obscuring the suture between North American and relict west
African crusts (Figure 1). The structure of the buried portion
of the orogen has been the subject of intense interest and
speculation for geologists. The purpose of this study is to
utilise experimental processing and testing of deep seismic
reflection data collected over the Georgia Coastal Plain by
COCORP (Consortium for Continental Reflection Profiling) in
order to provide a more quantitative basis for the geologic
interpretation of Palaeozoic compressional structure and the
reflection Mohorovicic discontinuity, and then to constrain
crustal cross-sections through the orogen.
Information on the pre-Mesozoic basement of the Southern
Appalachians is sharply divided across the Fall Line (inland
limit of Cretaceous-Tertiary onlap) (Figure 1). Mapping of the
diverse metamorphic and sedimentary rocks exposed north of
the Fall Line indicates that this region is composed of two
northeast-trending composite terranes (Inner Piedmont terrane;
Carolina or Avalon terrane), that were juxtaposed in Devonian-
Carboniferous time (now separated by the "Central Piedmont
suture"--Taconic or Acadian(?)) [Hatcher, 1987; Horton et
al., 1989] (see map, Plate 1). Subsequently, in
Carboniferous-Early Permian time, these terranes were
transported, en masse, westward over the North American
continental margin during terminal continent/continent
convergence between North America and west Africa
10 DEEP STRUCTURE OF BURLED SOUTHERN APPALACHIANS
'"?! GRENVILLE
..:...:L BASEMENT
%'""5,.
'"i:..,..;.:<.,
21
19
30
Fig. 1. Location map showing major crustal terranes and COCORP survey lines. Grid= magnetic intensity
high; stipple= magnetic intensity low; hachured= Bouguer gravity gradient.
(Alleghanian orogeny). South of the Fall Line, the subsurface
geology below the Coastal Plain is known only from scattered
drilling but indicates that lower Palaeozoic platform
sedimentary strata and Precambrian/lower Palaeozoic felsic
volcanic rocks with west African affinities ("Suwannee
terrane") underlie northern Florida and southernmost Georgia
[Chowns and Williams, 1983; Dallmeyer et al., 1989]
(Figures 1 and 2). Triassic-Jurassic continental clastic
sediments ("South Georgia basin") [Behrendt, 1986; McBride
et al., 1989] overlap the northern limit of the Suwannee terrane
(map, Plate 1), and thus obscure the Alleghanian (Middle
Carboniferous to Permian, =Hercynian) suture between North
America and a fragment of west Africa accreted during the late
Palaeozoic assembly of Pangea [Secor et al., 1986].
Previous interpretation of COCORP data in eastern
Georgia postulates that the Blue Ridge and Inner Piedmont,
and probably the Eastern Piedmont, were transported along a
low-angle d6collement ("Southern Appalachian d6collement")
over autochthonous Precambrian Grenville basement which
itself extends at least as far to the southeast as the seaward
edge of the Inner Piedmont [Cook and Oliver, 1981; Cook et
al., 1983]. It is thus implied that all the terranes of the
Piedmont provinces have been displaced northwestward by the
last major episode of thrusting during the Alleghanian orogeny
[Secor et al., 1986]. However, the relationship between this
deformation and where (and how) the Alleghanian sole
d6collement roots into the lower crust and upper mantle
beneath the Coastal Plain has remained unclear. Also unclear
have been the nature and extent of the early Mesozoic
"overprint" of rifting and magmatism associated with the early
opening of the North Atlantic Ocean [Nelson et al., 1986; de
Boer et al., 1988].
Results and Interpretation
Dipping Reflection Structure Wittu'n the Late Palaeozoic Suture
Zone
On all three crossings of the Brunswick magnetic anomaly
(BMA) (Figure 1), the middle and lower crust is marked by a
complex, steeply dipping zone of reflections and diffractions
which, together with the coincident B MA, separates North
American Appalachian terranes from relict west African crust,
and thus has been interpreted as marking the late Palaeozoic
"whole-crustal" suture [Nelson et al., 1985; Nelson et al.,
1987; Tauvers and Muehlberger, 1987; McBride and Nelson,
1988]. However, the character of the reflective zone varies
along strike, changing from a narrow, steeply dipping zone in
western Georgia to a shallow dipping, more broadly
distributed zone in eastern Georgia (Figure 2). This study
examines the internal fine-structure of this zone and poses a
possible tectonic explanation for its variation.
The western Georgia transect over the BMA (lines 13 &
14) shows the greatest amount of, and variation in, internal
structural detail, and thus is useful for finer-scale processing
z
z
n-
ILl
!--
Z
o
E
u
11
12 DEEP STRUCTURE OF BURIED SOUTHERN APPALACHIANS
N , BMA S
VP 400 200 H 200 300 400 800 I000 1200
0 I I II , I i I It I , I ,
BRIFT BASIN .... '... '' T ........... '
- o -- '' - ". - "- magnetized
' -%- :. .:!-3 ' '".. - ' "..
S : . .j_:;; ,-, -t?;:'.; :.;:::. ' ..... ' .......... zone :-
.oo ,i.7o ,7o, i,,o 7o, 9or 7o,
o,, ,,,,, , ,, I
ii
S _:' '"'PJJcC;. .-. ': ...... ".RIFT "BAS.I. magnetized
E 8 A. :.:;.,::,.;Lt. ' ..... ' .... .... zone
et... .t-t ' - --:. ........................
% I = I11- II -I1 I
0 50 /00
I I
k[/omefers
Fig. 2. Summary interpretive line drawings for the main COCORP deep reflection (unmigrated) transects
across the Georgia Coastal Plain. Data are displayed from 0 to the full 16 s (two-way time) correlated record
length. Dotted outline shows highly magnetised zone deduced from magnetic modeling constrained by the
reflection data [McBride and Nelson, 1988]. S.A.D.: Southern Appalachian d&ollement; PMB: Pine
Mountain belt [Nelson et al., 1987a]; BMA: Brunswick magnetic anomaly minimum. Vertical exaggeration =
1:1 @ 6.0 km/s.
and testing. Seismic time migrations were applied to merged
CDP-stacked sections together with a coherency enhancement
filter developed at Cornell University [Zheng and Brown,
1986]. This processor involves slant stacking within a sliding
window and then, in order to reduce artefacts, modifying the
processed stack according to the intertrace coherence with the
original section. The coherency-filtered unmigrated stack
clearly shows a dominant pattern of south-dipping reflections
(arrows, Figure 3), with three north-dipping reflections on line
13 between 2 and 5 s (-6 & 15 km, @ 6 krn/s). Individual
reflector segments are typically no longer than 6-8 km, and
dips typically range from -30 in the upper crust to -20 in the
lower crust, apparent unmigrated dip (@ 6 km/s). An unusual
concentration of diffractions occurs on line 14 between 6 and 9
s (-18 and 27 km, @ 6 km/s) indicating the presence of
several "sharp-edged" point sources (small intrusions or steep
structure?). Such a diffraction pattern is not a characteristic
feature of the suture zone on the three transects (e.g., Figure
2). South of the dipping reflective zone, the crust is
unrefiecfive excepting three north-dipping events. This change
in reflectivity, which appears to be distributed across a south-
dipping line, probably not a single, planar fault (Figure 3), is
interpreted to mark a terrane boundary between a relatively
unreflective west African "upper plate" and a more reflective
thrust root zone "lower plate" [see also Nelson et al., 1985].
Migration was used primarily to focus on specific
problems involving steep dips. A reduced velocity of 3.5
km/s is used to partially migrate a portion of the section in
Figure 4 while minimising distortion (with increasing travel
time) and over-migration [Warner, 1987]. The principal effect
of the migration is to concentrate the dipping reflective zone
into a more compact and well-defined package of events
consisting of several separated and distinct dipping reflectors.
An explanation for this pattern is that the reflector segments
represent thrusted pods or lozenges bound by shear zones that
may "ramp" upward in the direction of vergence (Figure 4).
Because of the smaller Fresnel zone diameter in the upper
crust, imaging the finer structure of such thrust complexes
may be easier at shallower levels. One possible thrust
complex is imaged below about VP 200 between 2 and 5 s,
line 14 (Figures 4 and 5). Figure 5 shows a small portion of
this zone migrated with a velocity of 6 km/s, which is
probably closer to the true interval velocity and can produce a
more stable migration due to the shorter travel times (_ 5 s)
and seismic wavelengths involved. The migrated reflector
structure is reminiscent of a partial thrust duplex or a single
MCBRIDE AND NELSON 13
300
NVp I
..,
200 100
I I
-- - . . . .,,,.. .,:....,;.-
300 200
I I I
lOO s
o
CZ-MZ
- 4
- g
- 8
-]12
0 km 10 GA-14 GA-13 unmigrated
I I
Fig. 3. Merged unmigrated section of deconvolved lines 13 and 14 with coherency enhancement filter. Major
dipping reflections (small dots), interpreted diffractors (dots, diamonds), and the reflection Moho (large dots,
M) are noted. Diffractors are located only roughly; source may lie out of the plane. Dashed line marks
interpreted terrane boundary (not necessarily as a fault) separating lower and upper plates in the suture zone.
Vertical exaggeration = 1:1 @ 6.0 krn/s.
thrusted anticline verging northward [e.g., Suppe, 1985].
Such structures would be typical of other major thrust root
zones in the Appalachian fold belt [Ando et al., 1983; Secor et
al., 1986; q2oruh et al., 1987; Hatcher, 1987; Pratt et al.,
1988]. The interpretation of the prominent north-dipping
reflectors is more problematic, but they may represent back
thrusts in the upper plate of the suture.
The three-dimensional relationship of the dipping reflection
structure of the suture zone can be best analysed where the line
11-13-14 and line 19-21 transects are tied by cross-line 20
vP 13oo .200 i100 1300 200
.......... -.`..`..``.`...`..-....-;-`....`.`-....`.:`:.``...`:i! ?-2 ' ':.:i e.:2:-;: '.::4. '
36 0 20 GA-14 GA-13 migrated
I I
km
Fig. 4. Partially migrated (@ 3.5 krn/s) CDP section and interpretation of merged lines ]3 and ]4. Vertical
exaggeration = 1:] @ 6.0 km/s. Solid lines mark interpreted thrusts and dashed lines indicate possible "fiat"
thrust surfaces. Distortion is caused in the middle of the section due to partially zeroed CDP traces at the
juncture of lines.
VP 260 220 180 140
6.0 km/s conslont velocily migrofion
30km
i !
Form ine Structure
INTERPRETATION COCORP 1990
Fig. 5. Detail of migrated (@ 6 km/s) line 14
[modified from unpublished COCORP work].
(Figure 6). The deep structure below line 20 is characterised
by bands of sub-horizontal reflections between 2 and 12 s that
are separated by relatively reflection-free zones. This
"layered" fabric is interpreted to represent a strike view of the
corresponding set of south-dipping surfaces as seen on the
intersecting dip profiles. Furthermore, the separation of
reflections into distinct zones may be related to north-vergent
thrusted sheets within the suture exposed in strike view.
Although this area has been widely affected by early Mesozoic
rifting and magmatism [McBride et al., 1989], the observed
layered character is not interpreted as an "extensional fabric"
associated with rifting as has been postulated for major
extensional provinces in general [cf. Matthews and Cheadle,
1986; Warner, 1990].
The line 16-17 transect, at about 200 km east of lines 13
and 14, exhibits a relatively simple pattern of gently south-
dipping events below the BMA (Figures 2 and 7). The most
coherent reflections on the record are a line of 1 to 3-km long
events between 4 and 6 s that rises in the section to the north
below a relatively blank record. Before migration, these
events show an arcuate, upward concave character and may
include multiples [Wille, 1987] (Figure 2). After applying a
-5.5 krn/s migration [Wille, 1987] (Figure 7), the events
collapse into a pattern of bent and flattened segments with an
antiformal shape somewhat similar to that imaged on line 14
(Figure 5). Taken as a whole, this reflector pattern appears to
represent the upper surface (i.e., separating upper and lower
plates) of a broad thrust complex that climbs and verges
northward. Although detailed interpretation here is not
unambiguous, the alternation between dipping and flat events
may result from a ramp-and-flat surface typical of thrust belts
in general [e.g., Suppe, 1985, p. 343]. The increased
broadness of the dipping reflective zone on the line 16-17
transect indicates an eastward widening of the Alleghanian
suture zone, as well as an overall change in structural style.
The Reflection Moho in the Buried Southern Appalachians
Sub-horizontal reflections appear in discontinuous strands
at between 11.0 and 12.0 or 12.5 s (33-36 or 37.5 km @ 6
km/s) on the three COCORP transects across the Georgia
Coastal Plain. The reflection strands, which are typically 5-10
km long and are equally well-developed beneath the early
Mesozoic basins and the suture zone (Figure 2), approximately
match the depth of the Moho as determined from regional
refraction data and thus have been interpreted as the "reflection
Moho" [Nelson et al., 1985; Behrendt, 1986; Costain et al.,
1989]. Seismic amplitude studies [Arnow, 1987] indicate that
the characteristically blank section below the Moho represents
an actual transparent zone in the upper mantle and not an
artefact of lost signal penetration.
Raw, unprocessed shot records (Figure 8) typically show
the Moho as a relatively strong, simple, 2-3 cycle event clearly
visible in the near-source offsets. Because this event may
contain reverberations, possibly generated in the flat-lying
Atlantic Coastal Plain sequence (Cretaceous-Tertiary) in the
upper -1 s, deconvolution testing was necessary to derive a
more accurate image of the reflector and to address the
significance of its apparent layered character. This study
considers two areas where the Moho is particularly well-
expressed, beneath one of the main Triassic-Jurassic basins
and beneath the dipping reflection zone along the B MA.
Beneath the main basin depocentre at the northern end of line
11, the Moho appears as isolated packages of multicyclic
events (e.g., Figure 9). In order to attack the problem of
short-path multiples, several deconvolution parameter trials
were designed from autocorrelations by varying the lag time
(travel time to second zero-crossing in autocorrelation
function) between 48 and 100 ms. Both pre- and post-stack
deconvolution applications were tested, although due possibly
to noise and the complexity of the multiples, deconvolution
applied post-stack often produced superior results over pre-
stack. Deconvolutions having longer time lags (100 ms)
produced a better defined image. The main effect of the
deconvolution has been to transform a reverberative
multicyclic section into one with 2 or 3 events dominated by a
single prominent reflection (Figure 9). Although the events
appear laterally discontinuous, because the Fresnel zone width
,oo
BMA ..a...,/?' f --. '" N,,.. :too I ,_,:,.
-- . '-. - ,.a I-IlII
: ', - c,, .., , o
/ : Alleghanian :'
. ' . G-14 GA-21 GA:19
vp !00 300 500 700
8 J .... UNE NE 832
' -- S
6A-20 I [,6
50 km
Fig. 6. Three-dimensional perspective diagram of reflection line drawings (unmigrated) of two dip sections
through the Alleghanian suture and overlying rift basins, with intersecting strike section (line GA-20, vertical
exaggeration = 1' 1 @ 6.0 km/s). Note sub-horizontal fabric associated with the dipping reflection zones.
O-
N vP I00 200 240 300 400 S
I i ! i i
:.--.,. - .-..,.-. --..._':,.__=_.-,..-.-%--_ ___ _._"- ::. _'-,_ , --.-. _.... _.. _ .._ -._ _.
:::?:'i?; ::::-:::. ::::.:'; :::': :.:' ..:: ::,'. ?: 0 P ; :: ::6 L-.:'i'-:':.".!: ::::::::::::::::::::::::::::: '::'' '"" ' "- "-'":-- - - - -'- -- '
:,_:: .:-::h_..:::::?:: ,.'..?:-::.:::? .":-: .:':':: e_:':: :,G:-: : ::: ..... '" ' '-- '' *c -" "-- ':::: :-:--.'::: ....... " '--'-- :'. - .-- --':i'.:: ::
. ,......, ....... ,_:....,
i:'?-.:_,.'i..2LU;::'?;;.:i.;:::i:V:..% 2d:. T.-:':',:?:'.:: :':-'-.,V' '.-:..':::?.-L;'i:;--.::r:%B:.':' i?;::5:::.,:';:
IOkm
GA-16 MIGRATED
COCORP
CZ-MZ
..... 15 km
-'-- 30 km
Fig. 7. Migrated (@ -5.5 km/s) CDP section and interpretation of line 16. Note the gradual northward rise of
dipping and flat reflection segments (A, B, C). M: Moho. Data processing from Wille [1987]. Vertical
exaggeration = 1:1 @ 6.0 km/s.
11
12
N
COMMON SOURCE GATHER
SOUTH END GA-11
I TRACES 96
VP $TA. SPACING = 330 FT (100.6 m)
Fig. 8. Small portion of unprocessed vibrator point record
from the southern end of line 11 showing well-developed
Moho reflection.
at Moho traveltimes is probably at least 4.0 km (for a spherical
wave; frequency < 25 Hz; velocity > 6.0 km/s; Sheriff and
Geldart, 1982), most apparent fine-scale lateral variation is
probably artificial. On the other hand, an idealised vertical
resolution of 60-133 m is possible for average velocities
between 6 and 8 km/s and for frequencies between 15 and 25
Hz using the "quarter wavelength" (or Rayleigh) criterion (and
wavelength = velocity/frequency) [Sheriff and Geldart, 1982].
These values correspond to a reflection time "resolution" of
--30 ms which, however idealised, is well below the observed
vertical separation between Moho reflections (Figure 9).
Thus, the vertical "structure" of the reflection Moho on the
COCORP transects may represent actual layering and not
necessarily reverberation or interference.
Reprocessing of a portion of the lower crustal reflection
data has been undertaken for line 13 in order to better
understand the relationship between dipping Palaeozoic
compressional structure and the reflection Moho. The Moho
beneath the BMA dipping reflection sequence, after applying
deconvolution, appears as a simple, two-cycle reflection at the
base of a discontinuous band of subhorizontal reflections
(Figure 10). A similar relationship is seen along lines 13, 14,
15, 19, 20, and 21 below the BMA dipping sequence (Figures
2 and 6). As argued above, this band may represent an actual
thin "layered" sequence at the base of the crust just above the
seismic Moho. On lines 16 and 17, where the South Georgia
basin is much less developed or absent, the Moho appears as a
sharp, flat reflection following pre-stack deconvolution [Wille,
1987] (northern end of line 16, Figure 7). A striking feature
of the Moho is its truncation of planar dipping reflections that
project downward into the lower crust. Also notable is that no
dipping reflections have been observed to cross-cut or
continue below the reflection Moho anywhere along the
ransects (Figures 2 and 6). This behaviour is suggested in
Figure 10 where planar dipping events continue down to, but
do not cut, the reflective band at the base of the crust. Dipping
events just above the Moho remain planar and show no listric
or asymptotic character as they deepen as has been interpreted
from deep reflection data sets elsewhere in the Appalachians
[e.g., Phinney, 1986; Hall et al., 1990].
The above analysis suggests that the multicyclic nature of
the reflection Moho may represent a thin "layered" sequence at
the base of the crust. The thickness of this zone ranges from
1.3 + 0.2 km below the South Georgia basin on line 11 to 3.2
+ 0.5 km below the BMA dipping reflection zone on line 13
(@ 6-8 km/s). Based only on observed wavelet "thicknesses"
and separations, individual layers 100 to 300 m thick might be
expected. This observation is consistent with models of wide-
angle seismic data for the lower crust and Moho in general that
suggest a laminated zone involving individual layers of about
100 m thickness [Wenzel et al., 1987; Paul and Nicollin,
1989].
Synthesis
The analysis of the reflection data provides for more
detailed and accurate "images" of the reflection Moho and
Palaeozoic crustal structure. Constrained crustal cross-
sections (Plate 1) through the Southern Appalachian orogen
demonstrate both the along-strike correlation of Palaeozoic
compressional structure as well as significant along-strike
variability, and the relation of this structure to the Moho.
GA-11 STACK vP 450
11
12
N NON-DECONVOLVED S
11
s
12
DECONVOLVED
km
0 2
!
Fig. 9. Portion of CDP stack (unmigrated) of Moho on line
11 below the South Georgia basin. Note effects of pulse
deconvolution (gap= 100 ms). Vertical exaggeration = 1:1 @
6.0 krn/s.
N
10 j
.
12
200 100
VP 300
--::-' ,'+'::.':'2,,.. ' ....::'_-:.-:.:'.:-:-.-::.i'.: ::-:::;.:::.:"..'....:..'...?:..'t.::.?,:.'.,...'.'5. :'.-'z ::.?_._. .;:. :_r,....-.?.,. ;.::,-.:. ,::...'F'. ?-
k.-'...'--'-",-'.'.'-'.-.',.--." ..... .' ' -,.' '--'..'."-: '- 1,' --?' .",,.': -._..-.-.-t....,, .--:..--., -.Z-,/.,'. :,,t:o..:,.'.-.v: ...,,':t.._*.'....'. :..-.-.-
. :_.-._-:'- :: z:.: :'.,.- :- :._-_..-.:2--.:.-.: --, C. ..... .' .':' :."--,- ': .'. :--:',;,...'":- .."-. 2 :a;'2' :-' ..- -"":...- -':.'--' ::..'.'-: "' .:-?--: Z':':- ;L:.:'.: : :.X ...-....-::
["--' .- ':-.--. -..,.:-. ,.'-' _L-':. '-?y_-' :,,_:-.-: ..; .-::-.'-_ :-.'. -': -&: --'.._.-._,. : :.: :; .-.: Z!-:: -7 '... - - : .:-.-'., _.-.:,'. :..-.-. :.': :: '--'.--- '_' -.:':! &% .._z,;-.' .:. :...-,-'-'.....%- '-. 5-'. ::: .? --_-&'[:_: -.'- :
o lo km GA-13
I I
s
Fig. 10. Portion of line 13 below the Brunswick magnetic anomaly dipping reflection sequence showing the
more complex character of the Moho (M). Vertical exaggeration = 1' 1 @ 6.0 km/s.
Palaeozoic Compressional Structure
In western Georgia, regional geology [Hooper and
Hatcher, 1988] indicates that the Inner Piedmont--Eastern
Piedmont (Carolina-Avalon) terrane boundary ("Central
Piedmont suture") projects to immediately south of the Pine
Mountain Belt and forms the lower plate of the BMA dipping
reflection zone on lines 13 and 14 (section Y-Y', Plate 1). The
timing of this suturing is uncertain but may be associated with
the mid-Palaeozoic closure of the Iapetus Ocean between the
Avalonian arc and the Inner Piedmont during the Acadian or
Taconic(?) orogeny [Hatcher, 1987]. Thus, at least in
westernmost Georgia, both the Inner and Eastern Piedmont
have been transported over autochthonous Grenville basement
out of a relatively narrow root zone below lines 13 and 14
(Figure 3). This implies that the composite Inner and Eastern
Piedmont with its embedded Central Piedmont suture was
transported northwestward during the Alleghanian orogeny,
and that at depth the Central Piedmont suture merges with (and
is overprinted by) the Alleghanian suture represented by the
zone of dipping reflections on lines 13 and 14. The North
American-west African suture cannot be located as accurately
but would be very near to the upper plate of the BMA dipping
reflection zone (section Y-Y', Plate 1). Constrained magnetic
intensity modeling of the BMA [McBride and Nelson, 1988]
suggests that the outboard portion of the suture was intruded
(reactivated?) by early Mesozoic-age, highly magnetised
(Figure 2) mafic/ultramafic material (gabbroic composition?)
as part of a zone of rift-stage crust [cf. Hutchinson et al.,
1983] that extended onland from the present-day offshore
along the East Coast magnetic anomaly. The implication of all
the above is that in western Georgia, mid- and late Palaeozoic
sutures have occupied the same, or nearly the same, thrust root
zone which was reactivated yet again by a concentration of
early Mesozoic intrusion (Plate 1, section Y-Y').
On the eastern Georgia transect (section X-X', Plate 1),
the BMA dipping reflection zone is shallower-dipping and
broader--here regional geology [Chowns and Williams, 1983;
Secor et al., 1986] suggests that the northern limit of west
African crust extends further northward and produces a
broadened zone of "crustal overlap" of west African and North
American Piedmont terranes [Tauvers and Muehlberger, 1987]
as shown in Plate 1. This crustal overlap appears to have been
accommodated by a more distributed deformation within the
suture. Likewise, the Eastern Piedmont terrane, between the
west African and Central Piedmont sutures, is even more
expanded as an allochthonous sheet transported over thinned
Grenville or "transitional" crust [Cook et al., 1983; Phinney
and Roy-Chowdhury, 1989]. Along the eastern Georgia
section, the crust is marked by a wider distribution of
deformation including a possible additional root zone below
the Augusta fault. Phinney and Roy-Chowdhury [1989]
interpret this feature as a root zone for the Taconic thrusts
further north. The location of the southernmost edge of
Grenville basement here is unknown but analysis of the
reflectivity of the dipping reflection zone down-dip of the
Augusta fault (Plate 1) suggests a limit as shown in section X-
X' [cf. Phinney and Roy-Chowdhury, 1989] although a root
zone much further outboard (e.g., offshore) is also
conceivable [Seeber and Armbruster, 1981].
In summary, Alleghanian convergence has transported the
entire Inner and Eastern Piedmont northwestward over an
eastward-widening autochthon between western and eastern
Georgia. The widening of the Alleghanian allochthon is
shown diagrammatically in Plate 1 as the eastward opening of
the zone defined by arrows (between , and '). Thrusting
associated with Alleghanian convergence roots into the lower
crust and upper mantle at the North American/west African
suture (,, Plate 1) and rises into the upper crust or surface at
the Central Piedmont suture (', Plate 1) and beyond in the
Valley and Ridge thrust belt.
An Early Mesozoic Age for the Moho?
The enhanced reflectivity and relative shallowness of the
reflection Moho below the Piedmont and the Atlantic Coastal
Plain lead to a suggestion of an early Mesozoic age for the
Moho below the Appalachian orogen [see also Nelson et al.,
1986; Nelson et al., 1987b; McBride et al., 1988]. The
seismic expression of the reflection Moho on the COCORP
data consists of a relatively thin package of discontinuous
events. Much thicker and well-developed "layered" sequences
in the lower crust occur elsewhere and have been associated
with regional extension [e.g., U.S. Basin and Range,
Klemperer et al., 1986; North Sea, Matthews and Cheadle,
1986; western Europe, Bois et al., 1988] or a crust-mantle
de3collement [Matte and Him, 1988]. No clear explanation for
these sequences has emerged, but possible interpretations have
included high-strain shear zones IReston, 1988], or basaltic
sills [Nelson, 1991; Warner, 1990]. The presence of a layered
reflection sequence at the Moho may be suggestive of material
added to the base of the crust [e.g., Latham et al., 1988]. The
mechanism of magmatic "underplating" of the lower
continental crust [Furlong and Fountain, 1986] by, for
example, ponding of basaltic magma from the mantle
[Meissner et al., 1983] could explain the observed highly
reflective and layered expression of the Moho. For the
Southern Appalachians, an interpretation of basaltic sill
intrusion at the base of the crust is consistent with a regional
geologic history of extension and graben formation (e.g., in
the South Georgia basin) together with the widespread
occurrence of early Mesozoic tholeiitic basalt/diabase dikes,
sills, and/or flows in the upper crust [Nelson et al., 1986; de
Boer et al., 1988; McBride et al., 1989]. Furthermore, the
nuncation of planar thrust surfaces within the suture zone by
the reflection Moho carries a strong implication that the Moho
is a younger feature that cross-cuts, and thus is unrelated to,
older compressional snucture. Because the youngest tectonic
event in the Appalachians is manifested as faulting and
magmatism associated with early Mesozoic crustal thinning
and heating, and the initial opening of the North Atlantic
Ocean, a probable candidate age for the Moho is Triassic-
Jurassic.
Cook and Oliver [ 1981] have previously interpreted the
presence of thinned crust in the Appalachians as a relict late
Precambrian/early Palaeozoic rifted margin (i.e., seaward limit
of Precambrian Grenville basement) delineated by the
Appalachian gravity gradient (Figure 1), although as pointed
out by Hutchinson et al. [1983], the observed crustal thickness
differential seems too small to represent a true continent-ocean
crust transition (i.e., only ~10 instead of 30 km). The relict
margin hypothesis can be rephrased by asking, where is the
easternmost extent of Precambrian Grenville basement? In the
New England Appalachians, the restored position of exposed
Grenville basement is far east of the gravity gradient and
presumed edge of crustal thinning [Stanley and Ratcliffe,
1985]. As stressed above, the southern limit of Grenville
basement in the Southern Appalachians is unknown but
probably extends beyond the gravity gradient (Figure 1 and
Plate 1) [Phinney and Roy-Chowdhury, 1989] and thus would
be unrelated to crustal thinning, whatever the cause. As
argued in the present study, an explanation for crustal thinning
and Moho development involving early Mesozoic rifting
magmatic processes is more attractive in being consistent with
the post-Palaeozoic geologic history of the Appalachians.
Acknowledgments. Data processing and reprocessing
were performed at the Institute for the Study of the Continents
at Cornell University. This research was supported largely by
United States National Science Foundation grants to the
COCORP project, most recently EAR-89161129. E.I.
Prossen suggested some of the steps used in data processing.
AGU reviews by P. Matte and an anonymous referee
improved the paper. Institute for the Study of the Continents
Contribution No. 166.
References
Ando, C. J., B. L. Czuchra, S. L. Klemperer, L. D. Brown,
M. J. Cheadle, F. A. Cook, J. E. Oliver, S. Kaufman, T.
Walsh, J. B. Thompson, Jr., J. B. Lyons, and J. L.
Rosenfeld, Crustal profile of mountain belt: COCORP deep
seismic reflection profiling in New England Appalachians
and implication for architecture of convergent mountain
chains, Am. Assoc. Petrol. Geol. Bull., 68, 819-837,
1984.
Arnow, J. A., Deep structure from continental collision:
COCORP deep seismic reflection profiles across the
southern Appalachians [M.S. thesis]: Cornell University,
Ithaca, New York, 82 pp., 1987.
Behrendt, J. C., Structural interpretation of multichannel
seismic reflection profries crossing the southeastern United
States and the adjacent continental margin - decollements,
faults, Triassic (?) basins and Moho reflections, in
Reflection seisrnology: The Continental Crust, Geodyn.
Ser., vol. 14, edited by M. Barazangi and L. D. Brown,
pp. 201-214, AGU, Washington, D.C., 1986.
Bois, C., M. Cazes, A. Hirn, A. Mascle, P. Matte, L.
Montadert, and B. Pinet, Contribution of deep seismic
reflection profiling to the knowledge of the lower crust in
France and neighbouring areas, Tectonophys. 145, 253-
275, 1988.
Chowns, T. M., and C. T. Williams, Pre-Cretaceous rocks
beneath the Georgia Coastal Plain - regional implications,
in Studies Related to the Charleston, South Carolina,
Earthquake of 1886Tectonics and Seismicity, edited by
G. S. Gohn, U.S. Geol. Surv. Prof. Pap. 1313, L1-L42,
1983.
Cook, F. A., L. D. Brown, S. Kaufman, and J. E. Oliver,
The COCORP seismic reflection traverse across the
southern Appalachians, AAPG Stud. Geol., 14, 61 pp.,
1983.
Cook, F. A., and J. E. Oliver, The late Precambrian-early
Paleozoic continental edge in the Appalachian orogen, Am.
J. $ci., 281, 993-1008, 1981.
(oruh, C., J. K. Costain, R. D. Hatcher, Jr., T. L. Pratt, R.
T. Williams, and R. A. Phinney, Results from regional
vibroseis profiling: Appalachian ultra-deep core hole site
study, Geophys. J. R. astr. Soc., 89, 147-156, 1987.
Costain, J. K., R. D. Hatcher, Jr., C. (oruh, T. L. Pratt, S.
R. Taylor, J. J. Litehiser, and I. Zietz, Geophysical
characteristics of the Appalachian crust, The Geology of
North America, Volume F-2, in The Appalachian-Ouachita
Orogen in the United States, edited by R. D. Hatcher, Jr.,
W. A. Thomas, and G. W. Viele, pp. 385-416,
Geological Society of America, Boulder, Colo., 1989.
Dallmeyer, R. D., M. Caen-Vachette, and M. Vi.lleneuve,
Emplacement age of the post-tectonic granites in southern
Guinea (West Africa) and the peninsular Florida subsurface:
Implications for origins of southern Appalachian exotic
terranes, Geol. Soc. Am. Bull., 99, 87-93, 1987.
de Boer, J. Z., J. G. McHone, J. H. Puffer, P. C. Ragland,
and D. Whittington, Mesozoic and Cenozoic magmatism,
in The Geology of North America, Volume 1-2, The
Atlantic Continental Margin, U.S., edited by R. E.
Sheridan and J. A. Grow, pp. 217-241, Geological
Society of America, Boulder, Colo., 1988.
Furlong, K. P., and D. M. Fountain, Continental crustal
underplating: Thermal considerations and seismic-petrologic
consequences, J. Geophys. Res., 91, 8285-8294, 1986.
Hall, J., G. Quinlan, F. Marillier, and C. Keen, Dipping shear
zones and the base of the crust in the Appalachians,
offshore Canada, Tectonophys., 173, 581-593, 1990.
Hatcher, Jr., R. D., Tectonics of the southern and central
Appalachian internides, Ann. Rev. Earth Planet. Sci., 15,
337-362, 1987.
Hooper, R. J., and R. D. Hatcher, Jr., Pine Mountain terrane,
a complex window in the Georgia and Alabama Piedmont;
evidence from the eastern termination, Geology, 16, 307-
310, 1988.
Horton, J. W., Jr., A. A. Drake, Jr., and D. W. Rankin,
Tectonostratigraphic termnes and their Paleozoic
boundaries in the central and southern Appalachians, in
Terranes in the Circum-Atlantic Paleozoic Orogens, edited
by R. D. Dallmeyer, Spec. Pap. Geol. Soc. Am. 230,
213-246, 1989.
Hutchinson, D. R., J. A. Grow, K. D. Klitgord, and B. A.
Swift, Deep structure and evolution of the Carolina trough,
in Studies in Continental Margin Geology, edited by J. S.
Watkins and C. L Drake, Am. Assoc. Petrol. Geol. Mem.,
34, pp. 129-154, 1983.
Hutchinson, D. R., J. A. Grow, and K. D. Klitgord, Crustal
structure beneath the southern Appalachians:
Nonuniqueness of gravity modeling, Geology, 11,611-
615, 1983.
Klemperer, S. L., T. A. Hauge, E. C. Hauser, J. E. Oliver,
and C. J. Potter, The Moho in the northern Basin and
Range province, Nevada, along the COCORP 40N seismic
reflection transect, Geol. Soc. Am. Bull., 97, 603-618,
1986.
Matte, P., and A. Him, Seismic signature and tectonic cross
section of the Variscan crust in western France, Tectonics,
7, 141-155, 1988.
Matthews, D. H., and M. J. Cheadle, Deep reflections from
the Caledonides and Variscides west of Britain and
comparison with the Himalayas, in Reflection seismology:
The Continental Crust, Geodyn. Ser., vol. 14, edited by M.
Barazangi and L. D. Brown, pp. 5-19, AGU, Washington,
D.C., 1986.
McBride, J. H., and K. D. Nelson, Integration of COCORP
deep reflection and magnetic anomaly analysis in the
southeastern United States: Implications for the origin of the
Brunswick and East Coast magnetic anomalies, Geol. Soc.
Am. Bull., 100, 436-445, 1988.
McBride, J. H., K. D. Nelson, and L. D. Brown, Regional
Bouguer gravity and deep seismic reflection patterns in the
Appalachian-Ouachita orogen [abstract], EOS Trans. AGU,
1988.
McBride, J. H., K. D. Nelson, and L. D. Brown, Evidence
and implications of an extensive early Mesozoic rift basin
and basalt/diabase sequence beneath the Southeast Coastal
Plain, Geol. Soc. Am. Bull., 101, 512-520, 1989.
Nelson, K. D., A unified view of craton evolution motivated
by recent deep seismic reflection and refraction results,
Geophys. J. Int., 1991.
Nelson, K. D., J. A. Arnow, J. H. McBride, J. H. Willemin,
J. Huang, L. Zheng, J. E. Oliver, L. D. Brown, and S.
Kaufman, New COCORP profiling in the southeastern
United States. Part I: Late Paleozoic suture and Mesozoic
rift basin, Geology, 13,714-718, 1985.
Nelson, K. D., J. A. Arnow, M. Giguere, and S. Schamel,
Normal-fault boundary of an Appalachian basement
massif?.: Results of COCORP profiling across the Pine
Mountain belt in western Georgia, Geology, 15, 832-836,
1987a.
Nelson, K. D., J. H. McBride, and J. A. Amow, Deep
reflection character, gravity gradient, and crustal thickness
variation in the Appalachian orogen: relation to Mesozoic
extension and igneous activity [abstract], Geol. Soc. Am.
Abstr. Prog., 18, 704-705, 1986.
Nelson, K. D., J. H. McBride, J. A. Amow, D. M. Wille, L.
D. Brown, J. E. Oliver, and S. Kaufman, Results of recent
COCORP profiling in the southeastern United States,
Geophys. J. R. astr. $oc., 89, 141-146, 1987b.
Paul, A., and F. Nicollin, Thin crustal layering in Northern
France: observations and modelling of the PMP spectral
content, Geophys. J. Int., 99, 229-246, 1989.
Phinney, R. A., A seismic cross section of the Appalachians;
The orogen exposed, in Reflection seismology: The
Continental Crust, Geodyn. $er., vol. 14, edited by M.
Barazangi and L. D. Brown, pp. 157-172, AGU,
Washington, D.C., 1986.
Phinney, R. A., and K. Roy-Chowdhury, Reflection seismic
studies of crustal structure in the eastern United States, in
Geophysical framework of the continental United States,
edited by L. C. Pakiser, and W. D. Mooney, Geol. Soc.
Am. Mem., 172, 613-653, 1989.
Pratt, T. L., I. Coruh, J. K. Costain, and L. Glover, III, A
geophysical study of the Earth's crust in central Virginia:
Implications for Appalachian crustal structure, J. Geophys.
Res., 93, 6649-6667, 1988.
Reston, T. J., Evidence for shear zones in the lower crust
offshore Britain, Tectonics, 7, 929-945, 1988.
Secor, Jr., D. T., A. W. Snoke, and R. D. Dallmeyer,
Character of the Alleghanian orogeny in the southern
Appalachians: Part III. Regional tectonic relations, Geol.
Soc. Am. Bull., 97, 1345-1353, 1986.
Seeber, L., and J. G. Armbruster, The 1886 Charleston,
South Carolina earthquake and the Appalachian
Detachment, J. Geophys. Res., 86, 7874-7894, 1981.
Sheriff, R. E., and L. P. Geldart, Exploration Seismology
Volume I, 253 pp., Cambridge University Press, 1982.
Stanley, R. S., and N.M. Ratcliffe, Tectonic synthesis of the
Taconian orogeny in western New England, Geol. Soc.
Am. Bull., 96, 1227-1250, 1985.
Suppe, J., Principles of structural geology, 537 pp., Prentice-
Hall, Inc., Englewood Cliffs, New Jersey, 1985.
Tauvers, P. R., and W. R. Muehlberger, Is the Brunswick
magnetic anomaly really the Alleghanian suture?, Tectonics,
6, 331-342, 1987.
Warner, M., Migration-why doesn't it work for deep
continental data?, Geophys. J. R. astr. Soc., 89, 21-26,
1987.
Warner, M., Basalts, water, or shear zones in the lower
continental crust?, Tectonophys., 173, 163-174, 1990.
Wenzel, F., K. Sandmeier, and W. Wilde, Properties of the
lower crust from modeling refraction and reflection data, J.
Geophys. Res., 92, 11575-11583, 1987.
Wille, D. M., The COCORP Surrency bright spot: possible
evidence for fluid in the deep crust [M. S. thesis]: Cornell
University, Ithaca, New York, 46pp., 1987.
Zheng, L., and L. D. Brown, Application of coherency
enhancement to deep reflection data, EOS Trans., 67,
AGU, 303, 1986.
HIGH-RESOLUTION P- AND S-WAVE DEEP CRUSTAL IMAGING ACROSS THE EDGE
OF THE COLORADO PLATEAU, USA :INCREASED REFLECTIVITY CAUSED BY
INITIATING EXTENSION
John M. Howie, Tom Parsons, and George A. Thompson,
Department of Geophysics, Stanford University, Stanford, California 94305
Abstract. The Transition Zone bordering the southwestern edge of
the Colorado Plateau is a region that is undergoing the early stages
of extension, as the Basin and Range province encroaches into non-
extended Colorado Plateau crust. A high-resolution seismic reflection
profile imaging the southwestern edge of the Colorado Plateau was
acquired in 1989 by Stanford University in cooperation with the U.S.
Geological Survey 1989 Colorado Plateau seismic refraction program.
Three-component recording allowed us to compile a single fold S-wave
common midpoint (CMP) section to accompany a P-wave CMP sec-
tion. The P- and S-wave sections reveal a "transparent" upper crust,
underlain by middle crust with a laminated sub-horizontal reflective
texture. Beginning about 9 s two way travel-time (twtt) (P-wave),
a gradual die-out of reflectivity occurs with the Moho located be-
tween 12 and 13 s twtt (~38 to 40 km depth). An increase in middle
and lower crustal reflectivity is observed beneath Chino Valley, at the
northeastern edge of the Transition Zone, as compared to the north-
ern end of the profile beneath the Colorado Plateau. This increased
reflectivity in combination with local seismicity and high heat flow
may be an indication of continental crust in the initial stages of ex-
tension. The vertical distributiort of reflectivity both on the P- and
S-wave sections also suggests changes with depth in the way the crust
is responding to tectonic extension. If the magnitude of crustal reflec-
tivity in extending regions is an indication of amounts of ductile shear
or magmatic intrusion, then we are observing a zone of mid-crustal
deformation located 10-20 km above the Moho.
Introduction
The Colorado Plateau of the southwestern USA has survived as
a coherent, relatively undeformed block of continental crust bordered
by Cenozoic extension in the Basin and Range province to the west,
and the Rio Grande rift to the east. Along the southwestern margin
of the Colorado Plateau, the crust thins by 10 km or more across
the Transition Zone to the Basin and Range province [Warren, 1969;
Hauser et al., 1987; Hauser and Lundy, 1989; McCarthy, et al., in
press]. However the shape of the structural transition in the lower
crust and processes related to crustal thinning at the edge of the
plateau remain poorly understood. This paper presents results from a
high-resolution deep seismic reflection experiment carried out by the
Stanford Crustal Geophysics Project across the physiographic edge
Geodynamics 22
21
of the Colorado Plateau. The Stanford Colorado Plateau reflection
experiment used a high-resolution array to record the large explo-
sive sources of the U.S. Geological Survey PACE (Pacific to Arizona
Crustal Experiment) 1989 Colorado Plateau refraction program (Fig-
ure 1). Our main objective was to image the structure and reflective
character of the middle and lower crust between the Transition Zone
and the Colorado Plateau to study the processes that are active at
this tectonic boundary.
We employed an integrated or "full-wavefield" seismic approach to
gain as much information as possible regarding rock properties and
structure of the crust across this major tectonic boundary. These
data indicate a transition from highly reflective crust to essentially
non-reflective crust between the extended Transition Zone ad the
unextended Colorado Plateau. The bulk of P- and S-wave reflectivity
originates from a laminated 8 km thick mid-crustal zone (~20 km
depth). Distinct P- and S-wave reflections have been correlated al-
lowing us to calculate interval Poisson's ratio values. The character
and distribution of P- and S-wave reflectivity, combined with surface
geology, heat-flow, and seismicity data, allow us to constrain possible
causes of the mid-crustal reflectivity.
Full-Wavefield Seismic Profiling Methods
The analysis of coincident wide-angle and vertical incidence seis-
mic data has long been recognized as a powerful method with which
to study crustal structure and evolution [Mooney and Brocher, 1987].
The importance of collecting coincident P- and S-wave information
for constraining crustal composition is also well known [Christensen
and Fountain, 1975; Kern, 1982]. However the spatial sampling and
data processing of these coincident data types can be quite different.
Working cooperatively over the past several years, Stanford and the
U.S. Geological Survey have attempted to break down the operational
distinctions between traditional refraction and reflection methods by
recording the "full-wavefield." We have done this by recording large
explosions at a continuous range of incident angles (vertical incidence
to wide-angle) on high density three-component arrays. The essential
components of the full-wavefield method as we have employed it are:
High-Energl Ezplosive Sources. Large explosive sources (450-2700
kg) similar to those commonly used for refraction profiles can po-
tentially provide excellent signal-to-noise and broad frequency band-
width for imaging the entire crust at both vertical-incidence and wide
angles [J archow et al., 1990]. Comparative studies of large explosive
sources and vibroseis have demonstrated that for many applications
explosives may be superior for imaging the deep crust [Brocher and
Hart, 1988; Jarchow et al., 1990]. The high signal-to-noise ratio of
the large explosions also eliminates the need to stack. The low-fold
22 P- AND S-WAVE REFLECTIVITY, COLOKADO PLATEAU
b 5o 8o
112' 110'
Ash rk .....
20 km
km
Tertiary and Quaternary Alluvium
Tertiary Volcanic Rocks
Paleozoic Sedimentary Rocks
.'/.
Precambrian Granitic Rocks
Shot Point
Mid-point locations
50 m spacing 1 oomponent
........ 25 m spacing 1 component
3 oomponenta-'
j Recording Spread Geometry
Paulden
Fig. 1. Geologic map showing the Stanford Colorado Plateau reflection experiment located at the northwestern edge of
the Transition Zone between the extended Basin & Range province and the non-extended Colorado Plateau. Upper inset
map shows the position of the Stanford profile relative to the USGS Pacific to Arizona Crustal Experiment (PACE), and
the lower inset shows the actual recording geometry used for the Stanford profile.
stacked sections that result (usually 100% sections) actually have im-
proved resolution over conventional high-fold stacked records because
less lateral averaging takes place [Klemperer et al., 1986]. It is impor-
tant to note however, that by focusing on the middle and lower crust
the upper 2-3 s of the crust is poorly imaged because of the coarse
shot spacing usually employed (5 km or more).
The large explosions also often generate strong S-wave energy
[Kullinger and Lund, 1986; Goodwin and McCarthy, 1990], especially
when the shotholes are located in bedrock. The S-wave energy is
probably generated directly by the shot [White and Sengbush, 1964]
or by a surface P-S conversion [Fertig, 1984].
High Density Recording Arrays. Close station spacing (<75 m)
allows improved correlation and spatial resolution of phases that are
laterally discontinuous or weak in amplitude. The close station spac-
ing is also essential for resolving S-wave phases that are often weaker
and less coherent.
Vertical-Incidence to Wide-Angle Recording. By recording energy
at a broad range of incident angles from vertical to wide-angle we
obtain the velocity resolution of wide-angle data and the spatial res-
olution of vertical-incidence data. As a result, knowledge about large
scale or gross properties of the crust (i.e. velocity within lower crustal
layers) can be combined with high-resolution vertical-incidence infor-
mation (i.e. subhorizontal reflective texture that might indicate fine
scale layering) to distinguish changes in reflectivity due to composi-
tion from those due to theology [Holbrook, 1988]. The continuous
range of incident angles also aids the identification of weak reflected
phases at vertical-incidence because they can be correlated continu-
ously from post-critical angles (high amplitude) to pre-critical angles
(low amplitude).
P- and S-Wave Recording. P- and S-wave velocities derived from
3-component recording can be used to determine average Poisson's
ratio ((r) for the entire crust and for intervals within the crust. Al-
though the relationship between (r and composition is non-unique,
it is an important parameter for constraining lithologic composition
and is particularly sensitive to quartz content [Tarkov and Vavakin,
1982]. A comparison of P- and S- reflectivity can also be an impor-
tant aid in lithologic determinations [Goodwin and McCarthy, 1990;
Holbrook et al., 1988], especially in situations where reflectivity is
caused by compositional layering.
Recording Geometry
The Stanford Colorado Plateau reflection experiment was centered
in Chino Valley of north-central Arizona (Figure 1). A stationary
recording array ~ 45 km long was deployed utilizing two large indus-
HOWIE ET AL. 25
try recording systems. A total of 1700 channels were recorded at 924
stations with an average station spacing of 50 m. Areas with sensitive
upper crustal targets had station spacings of 25 m. Several short (1-
5 km) 3-component arrays were positioned strategically along the 45
km reflection spread. Intermittent, short, closely spaced 3-component
arrays were chosen over widely spaced 3-component stations covering
the entire line so that P- and S-wave data could be compared at
the same, close station spacing. Twenty-four large explosions (450-
2700 kg) were shot into the spreaA at offsets ranging from 0 to 250+
km. The Stanford reflection sprea constitutes a 45-km-long high-
resolution "window" embedded in the larger regional U.S. Geological
Survey refraction array (200 km array length with 300 m average
station spacing).
Single Fold Common Midpoint Sections
P-wave
The P-wave section was constructed using shot-receiver offsets be-
tween 5 and 35 km. The near offsets were not included because of
strong broadband groundroll contamination. The majority of the off-
sets used are from 5-15 km with the longer offsets coming from the
off-end shots (km 0.10 km and km 42-50 on Figure 2). Processing in-
cluded static corrections (shot, receiver, datum), spherical divergence
gain correction (t2), NMO, and bandpass filtering (16-36 Hz). The
NMO velocity used assumed an average crustal P-wave velocity of 6.3
km/s based on early results from wide-angle inversion [McCarthy and
Kohler, 1990]. Lateral amplitude balancing was applied for display
purposes to reduce variation in amplitude between shots of different
sizes.
Upper Crust - (0-18 km depth). The upper crust (0-6 s twtt, 0-
18 km depth) reveals high amplitude laterally continuous reflections
of the Bagdad Reflection sequence set against an otherwise trans-
parent background. The Bagdad Reflection sequence was identified
on COCORP Arizona Transect seismic profiles [Hauser, et al., 1987]
and was interpreted as either a complex of subhorizontal intrusive
sills or shear zones. Goodwin et al. [1989] used waveform modeling
to determine that individual reflections within the sequence resulted
from thin (~25 m) high velocity layers, probably subhorizontal mafic
intrusions.
Middle Crust - (18-8 km depth). For purposes of this study,
the middle crust is distinguished from the upper crust by the highly
reflective texture seen at vertical incidence. Beginning at about 6
s twtt (~18 km depth) the crust becomes more reflective with an
increase of relatively short discontinuous reflections. The zone of
highest reflectivity (between 7 and 9 s) is bounded by semi-continuous
bands of reflections at 7 s (PA) and 9 s (PB) (Figure 3). These
reflective bands are correlated across 2/3 or more of our profile (km
0-35, Figure 2).
At the north end of our profile, the strength of the reflective band-
ing and the lower crustal reflectivity decreases significantly. The lat-
eral change in reflectivity cannot be directly related to either a change
in surface conditions, or poor signal penetration. Thus we are observ-
ing a fundamental change in the reflective character of the crust as
we cross from the Transition Zone to the Colorado Plateau. The high
amplitude sub-horizontal reflectivity that is strongest beneath Chino
Valley (km 20-30, Figure 2) is replaced by relatively non-reflective
crust with a few distinct reflections (i.e. km 35-40, 6-7 s twtt, Figure
2).
Lower Crust - (8-40 km depth). Below 9 s twtt, the crust be-
comes graclationally less reflective approaching the Moho (km 5-10
and 20-30, Figure 2; Figure 3). We estimate the Moho between 12
and 13 s twtt (36-40 km depth) where a subtle drop in reflectivity
occurs. More detailed estimates of the crustal thickness and shape of
the Moho must be deferred to the ongoing interpretation of the U.S.
Geological Survey wide-angle data.
C- wa v
A 50 km S-wave CMP profile was also constructed, although mid-
point coverage was not as complete as on the P-wave section. Process-
ing for the S-waves included rotation to radial and transverse com-
ponents (transverse component shown in Figure 2), spherical diver-
gence correction, NMO (assuming a--0.25 and average crustal Vp-6.3
km]s), and bandpass filtering (2-12 ttz). Lateral trace balancing was
applied for display purposes. The upper 5 s twtt of the section was
muted to eliminate strong P-wave and surface wave energy recorded
on horizontal geophones. Source-receiver offsets included for the S-
wave section are 0-35 km.
The general pattern of S-wave reflectivity shows a strong corre-
spondence to the P-wave reflectivity. The upper crust is transparent
(5-12 s twtt), the middle crust is highly reflective (12-18 s twtt) and
the lower crust becomes gradationally less reflective towards the Moho
(18-22 s twtt) (Figures 2,3,4). The transparent upper crust is best
seen at kilometers 8-10 and 25-35 from 6-12 s twtt (Figure 2). Dif-
fuse lower crustal reflectivity is clearly seen at kilometers 8-10 from
12-18 s twtt (Figure 2). A short data segment (~ km 26, Figures
2,3,4) also exhibits distinct S-reflections at 12 s ($A) and 16 s ($B)
twtt that correlate closely with the 7 and 9 s P-wave reflections (PA
and PB, Figures 3,4). In several locations, the Moho on the S-wave
section is easier to identify than on the P-wave section and is defined
by a die-out of reflectivity. For example, at kilometers 8-10 (Figure
2) fairly distinct S reflectivity near 22 s twtt (~40 km depth) drops
off sharply.
Poisson's Ratio
Windows of high quality S-wave data can be used with the cor-
responding P-wave data to constrain Poisson's ratio within the crust
(Figures 3,4). Poisson's ratio can be calculated directly from P- and
S-wave travel-times using the simple expression:
0.5
a-'
1.0- [s ]2
A value of a for the interval from the surface to ~28 km depth (9 s
twtt, P) was calculated by correlating a distinct pair of high amplitude
P and S reflections (PB and SB, Figures 3 and 4). The average
value of a for the upper and middle crust is 0.23-0.24. The primary
assumption in the calculation is that the events are reflecting off the
same horizon at depth. By correlating absolute amplitudes (Figure 4),
we eliminate possible picking errors related to the unknown polarity
of the shear waves. Also, since we are correlating the shape of the
amplitude curves and not just a single time pick for both P and S the
error in the calculation is relatively small (4-0.005) (Figure 4).
Poisson's ratio combined with velocity information, usually Vp,
can be a useful indicator of crustal composition. For crustal rocks, a
averages ~0.25, with quartz-r;ch rocks slightly lower and mafic rocks
slightly higher. Poisson's ratio of 0.24 with average crustal velocities
of ~6.3 km/s indicates probable granitic to quartz dioritic average
compositions in the upper and middle crust [Simmons and Brace,
1965; Brooks, 1985; Simmons, 1964].
In addition to the vertical incidence estimates of a, correspond-
ing wide-angle phases such as Pg, Sg, PrnP and SrnS can provide
regional estimates of a for the upper crust and the whole crust (Fig-
ure 5). The results of Goodwin and McCarthy [1991], located 60 km
southwest of our study, give consistent estimates of auppercrust=0.24-
0.25 and ami d _ lowercrust=0.27 and suggest an intermediate com-
position for the middle and lower crust.
Lower Crustal Reflectivity
Two types of reflectors can be identified in the middle crust of
the Transition Zone: 1) high amplitude laterally continuous events
26 P- AND S-WAVE REFLECTIVITY, COLOHADO PLATEAU
VERTICAL
DISTANCE (km)
26.0 27.0
2 2
4 4
12 [2
14 14
16 6
SP 32
TRANSVERSE
DISTANCE (km)
26.0 27.0
12
SB
24 24
Fig. 3. Windows of P- and S-wave data near km 26 with a CMP spacing of 12.5 m. The data has been displayed
with spherical divergence correction and bandpass filtering, but no NMO. The vertical component section shows the high
amplitude laterally continuous reflections PA and PB set within a more discontinuous laminated reflective texture in
the lower crust. By contrast the transverse component shows two distinct reflections SA and SB set within a relatively
non-reflective background. Poisson's ratio for various phases can be calculated directly from travel-times determined on
this plot. The strong correlation between the A and B reflections is seen. P and S are particularly similar in shape
and amplitude. The bar beneath each plot shows the group of traces summed to create the amplitude decay curves
shown in Figure 4.
that are reflective to both P- and S-wave energy, and 2) shorter dis-
continuous events that comprise a laminated texture on the P-wave
section, but are only slightly reflective to non-reflective to S-waves
(Figure 3). For a given frequency, resolution of P- and S-waves differ.
However in this case, S-wave energy with a dominant frequency near
10 Hz and P-wave energy with a dominant frequency near 20 Hz have
approximately the same wavelength and thus about the same vertical
resolution.
Possible causes of the laterally continuous reflections include com-
positional layers, shear zones, mafic intrusions and partial melts.
Shear zones or mafic intrusions are the two most popular causes,
but quantitatively distinguishing fluid filled shear zones (low velocity
layers)[O'Connell and Budiansky, 1974] from intrusions (high velocity
layers) [Simmons, 1964] at these depths is particularly difficult.
The laminated reflectivity is similar to that observed beneath tec-
tonically extended areas [BlttPS and ECORS, 1986; McCarthy and
Thompson, 1988]. It probably results from subhorizontal alignment
and segregation of minerals during ductile flow or shearing that occurs
during extension [Wever and Meissner, 1987; McCarthy and Thomp-
son, 1988; Reston, 1990]. Laminated reflectivity is most often asso-
ciated with zones that are theologically weak [Meissner and Kusznir,
lg87].
It is particularly striking that the reflective zone beneath the Tran-
sition Zone is located 10-20 km above the base of the crust. If crustal
reflectivity is associated with crustal theology in extending regions
then this observation suggests a theologically weak mid-crust over a
theologically strong lower crust [Meissner and Kusznir, 1987]. Early
results from inversion of the coincident wide-angle data indicate that
the velocities at the base of the crust can be no faster than about
6.6 km/s [McCarthy and Kohler, 1990]. In addition, results from the
central and southwestern Transition Zone [McCarthy et al., in press;
Goodwin and McCarthy, 1990] show a similar reflective mid-crust
HOWIE ET AL. 27
9
.c 10
SP-32 Amplitude Decay db
-20 -10 0
S P
o'= 0.24
0.23
10
S-wave window P-wave window
10
I km
Fig. 4. Clce-up of the P- and S-wave reflection pair used to calculate an average Poisson's ratio for the upper and
middle crust beneath Chino Valley (km 26, Figure 2,3). In this display S-wave travel-time was converted to P-wave
travel-time using a Poisson's ratio of 0.235. We have assumed that this reflection pair is from the same depth interface
because of the striking similarity of relative amplitude and reflection chaxacter. The compaxison of amplitude decay
curves in the left panel displays the effects of wrying Poisson's ratio from 0.23 to 0.24 on the correlation of the P- and
S-wave amplitude peaks.
but do not indicate a major velocity boundary that corresponds with
the change in reflectivity between the middle and lower crust. Thus
we have no evidence for a high velocity mafic layer that might be
rheologically strong. However, these observatons could be explained
by a change in composition from quartz-dominated to plagioclase-
dominated between the middle and lower crust. This compositional
change would result in significant rheological differences [Meissner and
Kusznir, 1987] with little change in velocity [Babuka, 1981]. The low
of 0.24 to the base of the middle crust supports this interpretation.
Therefore the most probable causes for the mid-crustal reflectivity are
either a theologically weak zone that has undergone ductile shear, or
a localized zone of magmatic intrusion. Ongoing analysis of S-wave
polarization and detailed modelling of reflectivity, combined with re-
sults from the wide-angle inversion may help us better constrain why
reflectivity is limited to the mid-crust.
The Transition Zone offers a unique opportunity to image the
crust in the early stages of extension, and contrast that image with
the adjacent non-extended Colorado Plateau crust to the northeast
Basin and Range Transition Zone Colorado Plateau
Highly Extended Initiating Extension No Extension
0
3O
45
0 distance km 100
2OO
:'.";'sson's R;'I':;, Stanford 1987 I:XOfile Stanford 1989 profile
Heat Flow
125 I,
wE100 -
75
50
0 distance km 100 200
Fig. 5. Composite schematic plot of reflectivity inferred from the Stanford 1989 and 1987 reflection experiments
in the Transition Zone, USGS PACE refraction data, and the COCORP Arizona profiles [Hauser et al., 1987; Goodwin
and McCarthy, 1990]. There is a mrked increase in mid-crustal reflectivity in the Transition Zone compared with the
extended but inactive southern Basin and Range, and the non-extended Colorado Plateau. The ellipses show Poisson's
ratio values calculated from long offset and near-vertical incident phases [Goodwin and McCarthy, 1990]. Heat flow is
plotted below, as derived from the contour map of J. Sass and A. Lschenbruch [unpublished data, 1991]. There is a
correlation between high heat flow, active seismicity, and middle crustal reflectivity in the Transition Zone.
28 P- AND S-WAVE REFLECTIVITY, COLOl:LkDO PLATEAU
and highly extended Basin and Range crust to the southwest. A
schematic synthesis of crustal reflectivity combining our results with
earlier work in the Basin and Range province and the Transition Zone
[Hauser et al., 1987; Goodwin and McCarthy, 1990](Figure 5) high-
lights the relationship between extension and crustal reflectivity that
is discussed in detail by McCarthy et al. [1990]. The activity of local
extension in the Transition Zone is evidenced by high heat flow (80-
100 mW/m 2) [J. Sass and A. Lachenbruch, unpublished data, 1991]
and the 1976 magnitude 4.9 earthquake on the Big Chino normal fault
[Eberhart-Phillips et al., 1981] (Figure 1). North of the Big Chino
fault (which marks the physiographic edge of the Colorado Plateau),
there is no further surface evidence of extension, and the surface heat
flow drops to the Colorado Plateau average of 60 mW/m 2 [Krieger,
1965; J. Sass and A. Lachenbruch, unpublished data, 1991]. We ob-
serve a correlative lateral die-out of mid-crustal reflectivity in the
non-extended Colorado Plateau crust to the north compared with
the extended Transition Zone crust. The north end of our profile
reveals a relatively non-reflective crustal column very similar to CO-
CORP profiles in the interior of the Colorado Plateau [Hauser and
Lundy, 1989]. In contrast, the highly extended, but currently inactive,
Basin and Range province to the southwest does not have the distinct
mid-crustal reflectivity. Here the crust is reflective throughout with
a relatively sharp Moho reflection [Hauser et at., 1987]. These ob-
servations suggest that the Transition Zone reflectivity is the result
of localized extension related tectonism in the mid-crust. The recent
seismic activity on the Big Chino normal fault, and the localized ele-
vated heat flow suggest that this reflective character is probably the
result of current and active extension.
Conclusions
Initial results from the Stanford Colorado Plateau reflection ex-
periment indicate that distinct differences between the crust of the
Transition Zone and the Colorado Plateau can be seen within 10-20
km of the physiographic boundary between these tectonic provinces.
1. The crust at the northeastern edge of the Transition Zone is
36-40 km thick. Moho is defined by a die-out of reflectivity
that is slightly more pronounced on the S-wave section than
the P-wave section.
2. The upper and middle crust beneath Chino Valley have an aver-
age r=0.23-0.24 and Vp=6.3 km/s indicating granitic to quartz
dioritic average compositions in the upper and middle crust.
3. A striking change in reflective character is observed from a
reflective middle crust beneath the Transition Zone to a non-
reflective middle crust beneath the Colorado Plateau. A highly
reflective zone in the middle crust (~6-9 s twtt, P) seen on the
southwestern 2/3 of our profile dies out within ~20 km of the
physiographic edge of the Colorado Plateau. The crust at the
northeastern end of our profile (northeast of the physiographic
edge of the Colorado Plateau) is relatively non-reflective and
similar to deep reflection profiles from the interior of the Col-
orado Plateau [Hauser and Lundy, 1989].
4. If the magnitude of crustal reflectivity in extending regions is an
indication of amounts of ductile shear or magmatic intrusion,
then the highly reflective middle crust within the Transition
Zone suggests the localization of extension related deformation
in this part of the crust. Ongoing analysis is focussed on the
seismic properties of the reflective zone in an effort to better
understand why the reflectivity and underlying tectonic activity
is located in the middle crust and not in the lower crust, where
it is most often observed.
5. The increase in mid-crustal reflectivity in the Transition Zone
relative to the Colorado Plateau and the extended but inactive
southern Basin and Range province corresponds with an area
of high heat flow, and active extensional seismicity. We suggest
that the Transition Zone reflective texture is the result of recent
and ongoing tectonic extension.
Acknowledgments. This research results from the cooperation and
generosity of a large number of individuals and organizations. The
success was possible only with the encouragement and close coordi-
nation of Jill McCarthy and the U.S. Geological Survey. We thank
Amoco Production Co. (Fred Metrailer and John Myers), ARCO
Oil and Gas Co. (Gary Crews, Bob Dart and George Clark), and
Grant-Norpac Co. (Hugh Fraser and Ken Freeman) for making per-
sonnel and equipment available during the experiment. We also thank
the Gas Research Institute for their financial support. For their assis-
tance in the field, we particularly thank Ken and Sally Krauter, Bruce
Beaudoin and Stun Ruppert. We are indebted to John Sass and Art
Lachenbruch for allowing us to include the heat flow transect across
our study area from their unpublished heat flow map for the Mohave,
southern Basin and Range, and Colorado Plateau region. All data
processing was done on Jon Claerbout's Stanford Exploration Project
CONVEX computer. This work was supported by NSF grants EAR
89-03541 and EAR 90-17667.
References
Babuka, V., Anisotropy of Vp and Vs in Rock-Forming Minerals: J.
Geophys., 50, 1-6, 1981.
BIRPS and ECORS, Deep seismic reflection profiling between
England, France and Ireland, J. Geol. Soc. London, 143, 45-52,
1986.
Brocher, T. M., and P. E. Hart, Direct Comparison of Dynamite and
Vibroseis Methods for Deep Crustal Seismic Reflection Studies in
the Southwestern Basin and Range, Nevada, USA, EOS Trans.
A GU, 69, 1313, 1988.
Brooks, S. G., Seismic velocities from crustal sections in northern
Scandinavia, PhD Thesis, Univ. of East Anglia, England, 1985.
Christensen, N. I., and D. M. Fountain, Constitution of the lower
continental crust based on experimental studies of seismic velocities
in granulite, Geol. Soc. Am. Bull., 86, 227-236, 1975.
Eberhart-Phillips, D., R.M. Richardson, M.L. Sbar, and R.B. Her-
rman, Analysis of the 4 February 1976 Chino Valley, Arizona,
Earthquake, Bull. Seismol. Soc. Am., 71, 787-801, 1981.
Fertig, J., Shear waves by an explosive point-source: The earth as a
generator of converted P-S waves, Geophys. Prospecting, $, 1-17,
1984.
Goodwin, E., and J. McCarthy, The lower crust of the Transition
Zone, Arizona, J. Geophys. Res., 95, 20097-20109, 1990.
Goodwin, E., G.A. Thompson, and D.A. Okaya, Seismic Identifica-
tion of Basement Reflectors: The Bagdad Reflection Sequence in
the Basin and Range Province-Colorado Plateau Transition Zone,
Arizona, Tectonics, 8, 821-832, 1989.
Hauser, E. C., J. Gephart, T. Latham, L. Brown, S. Kaufman, J.
Oliver, and I. Lucchitta, COCORP Arizona Transect: Strong crustal
reflections and offset Moho beneath the transition zone, Geology,
15, 1103-1106, 1987.
Hauser, E. C., and J. Lundy, COCORP Deep Reflections; Moho at
50 km (16s) beneath the Colorado Plateau, J. Geophys. Res.,
7071-7081, 1989.
Holbrook, W. S. Wide-angle seismic studies of crustal structure and
composition in Nevada, California and southwest Germany, PhD
Thesis, Stanford University, Stanford, CA, p. 214, 1988.
Holbrook, W. S., D. Gajewski, A. Krarnrner, and C. Prodehl, An in-
terpretation of wide-angle compressional and shear waves in south-
west Germany: Poisson's ratio and petrological implications, J.
Geophys. Res., 93, 12081-12016, 1988.
HOWIE ET AL. 29
Jarchow, C. M., E. B. Goodwin, and R. D. Catchings, Are large ex-
plosive sources applicable to resource exploration?, Leading Edge,
9, 12-17, 1990.
Kern, H., P- and S-wave velocities in crustal and mantle rocks under
the simultaneous action of high confining pressure and high tem-
perature and the effect of rock microstructure, in High Pressure
Researches in Geoscience, edited by W. Schreyer, 15-45, Schweizer-
bert, Stuttgart, Federal Republic of Germany, 1982.
Klemperer, S., T.A. Hauge, E.C. Hauser, J.E. Oliver, and C.J. Potter,
The Moho in the northern Basin and Pnge province, Nevada,
along the COCORP 40 N seismic-reflection transect, Geol. Soc.
Am. Bull., 97, 603-618, 1986.
Krieger, M. H., Geology of the Prescott and Paulden Quadrangles,
Arizona, U.S. Geol. Surv. Prof. Pap., 467, United States Gov-
ernment Printing Office, Washington, D.C., 1965.
Kullinger, B., and C. E. Lund, A preliminary interpretstion of S-
wave traveltimes from Fennolora data, Tectonophysics, 16, 365-
374, 1987.
Meissner, R., and N.J. Kusznir, Crustal viscosity and the reflectivity
of the lower crust, Annales Geophysicae, 5B, 365-374, 1987.
McCarthy, J., and W. Kohler, Preliminary two-dimensional mod-
eling of the crustal structure across the Colorado Plateau-
Basin and Range transition, Arizona, LOS Trans. A GU, 71,
1564, 1990.
McCarthy, J., S.P. Larkin, G.S. Fuis, R.W. Simpson, and K.A. Howard,
Deep Anatomy of a Metamorphic Core Complex: Seismic Re-
fraction and Wide-Angle Reflection Profiling in southeastern
California and western Arizona, J. Geophys. Res., in press.
McCarthy, J., D.A. Okaya, E.B. Goodwin, Geophysical Character
of the Middle and Lower Crust in the Southwestern U.S.:
Results from Seismic Profiling, Abstracts wiih Programs, 86th
Annual Meeting Cordilleran Section Geol. Soc. Amer., , 65-66,
1990.
McCarthy, J., and G.A. Thompson, The seismically reflective crust
beneath highly extended terranes: Evidence for its origin in exten-
sion, Geol. Soc. Am. Bull., 100, 1361-1374, 1988.
Mooney, W.D., and T.M. Brocher, Coincident seismic reflection /
refraction studies of the continental lithosphere: A obal review,
Reviews of Geophysics, 5, 723-742.
O'Connell, R. J., and B. Budiansky, Seismic velocities in dry and
saturated cracked solids, J. Geophys. Res., 79, 5412-5426, 1974.
Reston, T.J., The Lower Crust and the Extension of the Continental
Lithosphere: Kinematic Analysis of BIRPS Deep Seismic Data,
Tectonics, 9, 1235-1248, 1990.
Simmons, G., Velocity of shear waves in rocks to 10 kilobars, J. Geo-
phys. Res., 69, 1123-1130, 1964.
Simmons, G., and W.F. Brace, Comparison of static and dynamic
measurements of compressibility of rocks, J. Geophys. Res., 70,
5649-5656, 1965.
Tarkov, A. P., and V.V. Vavakin, Poisson's ratio behavior in various
crystalline rocks: applications to the study of the Earth's interior,
Phys. Earth Planet. Interiors, 9, 24-29,1982.
Warren, D. H., A Seismic-Refraction Survey of Crustal Structure in
Central Arizona, Geol. Soc. Am. Bull., 80, 257-282, 1969.
Wever, Th., and R. Meissner, About the nature ofrefiections from the
lower continental crust, Annales Geophysicae, 5B, 349-352, 1987.
White, J. E., and R. L. Sengbush, Shear waves from explosive sources,
Geophysics, 8, 1001-1019, 1963.
CRUSTAL ANISOTROPY AND THE STRUCTURE OF THE MOHOROVICIC DISCONTINUITY
IN WESTERN NEVADA OF THE BASIN AND RANGE PROVINCE
Ramon Carbonell and Scott B. Smithson
Program for Crustal Studies, Department of Geology and Geophysics
University of Wyoming, P.O. Box 3006 Laramie, Wyoming, 82071.
Abstract. Near the intersection of perpendicular Introduction
lines, the 1986 Nevada PASSCAL wide-angle seismic
reflection data show strikingly different
responses of the deep crust and Moho in each The Basin and Range is characterized by
direction. The seismic signature of the crust and Cenozoic extension varying in ages and amounts,
Moho in the dip line (east-west) is characterized which may reach areally up to 200% [Miller,
at normal incidence by strong, subparallel, 1988]. The geology, which is characterized by
laterally discontinuous, multicyclic events listric and normal faults, and the extensional
lacking observable diffractions. In the strike tectonic environment can be found elsewhere
line (north-south), the crust is highly [Miller et al., 1983; Hamilton, 1987;
transparent with short events and a high- Allmendinger et al., 1987; Hauser et al., ]987;
amplitude, laterally continuous Moho reflection. Speed et al., 1988; Snoke and Miller 1988; and
The PmP phase presents an abrupt high frequency Miller et al., 1988].
attenuation from subcritical to postcritical The 1986 PASSCAL experiment in western Nevada
offsets characteristic of tunneling waves. PmS [Nevada PASSCAL 1986 Lithospheric seismic
and/or SmP and SmS phases are also identified as experiment, Working group, 1988], included two
lower frequency dispersed wave trains. Forward perpendicular reflection lines, one roughly
modeling of amplitudes by the reflectivity north-south (parallel to strike of major geologic
algorithm, reflection coefficients estimates, and features, faults, strike line), and a second line
heat flow data suggest such high velocity parallel to the extension direction, roughly
contrasts for the Moho structure that east-west (parallel to the dip of the major
interlayering of partial melts with mantel (high geologic features, dip direction) (Figure 1).
velocity) material at the base of the crust is a Shots recorded by the reflection spreads provide
strong possibility. Finite difference forward a wide-angle data set. The 3-D control due to the
modeling indicates that mid- and/or lower crustal perpendicular orientation of the recording
anisotropy achieved by placing a stack of spreads provides new insight into the crustal and
tectonic lenses (boudins) strongly elongated Moho structure of the Basin and Range province.
parallel to the stretching direction (dip Processing included: 1) Trace equalization
direction), disrupts a strong laterally (i.e., trace balancing). Each trace was scaled by
continuous event to a series of short multicyclic the RMS (root mean squared) of the amplitudes of
reflection segments. The synthetic seismic record the last second of the seismogram. The main
sections parallel and perpendicular to the longer assumption being that near the end of the record
axes of the heterogeneities (i.e., lenses) little or no signal is recorded thus it is an
closely model the seismic signature observed in estimate of the background noise, which is
the E-W and N-S shot gathers, respectively. Thus assumed to be c>nstant along the line. 2) High
the deep crust may consists of a series of pass filtering (5.0 to 8.0 Hz). Since the natural
lenses, highly elongated in the E-W (stretching) frequency of the geophones used varied from 8-10
direction, overlying a Moho marked by ponded Hz, this step removed the frequency peaks
lenses of about 5% partial melt. recorded below the natural frequency of the
sensors, suppressing surface related noise (i.e.,
ground roll) and strongly improving the quality
of the data. 3) Spherical divergence correction.
A velocity function derived by Hawman et al.
[1990] was used in a spherical divergence routine
Continental Liosphere: Dp Seismic Refltions from a standard process i ng package (DISCO) to
Geodynamics22 obtain true relative amplitude plots. 4) Spectral
1991 Amecan Geophysic Union analysis. A window 0.9 s long was extracted
31
32 BASIN & RANGE CRUSTAL ANISOTROPY & MOHO STRUCTURE
around the PmP phase; linear tapers 0.250 s long
were used at either end to avoid side lobes.
Description of the Data
Strikingly different seismic signatures for
the crust and Moho are observed at the
intersection of the orthogonal lines (Figure 2a
and 2b). The broad reflecting bands at 4, and 6,
s and the broad diffuse Moho are all
characterized by short multicyclic reflection
segments which lack observable diffractions in
the east-west (i.e., stretching) direction. These
contrast with the transparent crust and laterally
continuous, high amplitude Moho event,
characteristic of the north-south line. In
general the seismic signature can be
3km characterized by far-offset events that interfere
- (i.e., criss-cross) with other mid- and lower
Fig. 1. Locations of shot points (numbered crustal phases, and by short randomly distributed
stars) and receivers for the 1986 Nevada PASSCAL reflection segments at subcritical distances. PmP
Basin and Range seismic experiment. Ranges are arrives with approximately 0.3-0.5 s delay in the
stippled, Basins unshaded. Reflection spreads north-south with respect to the east-west line
extended 20 Km west and north of shot point 4 and (Figure 2a and 2b). One-dimensional inversion
20 Km east of shot point 4-C. (-.-): refraction schemes [Hawman et a]. 1990] revealed total
lines; ( ): COCORP lines (Klemperer et al.,, crustal thicknesses of 31 and 30 km for the
1986); (===): overlapping COCORP and refraction north-south and east-west lines respectively. The
lines ( After 1986 PASSCAL Working group, 1988). observed time delay and/or depth difference in
0.0
Offset (km)
- _
:.
.
_
Fig. 2. a: Reflection recording for shot point 4 Km or larger offsets. A 4.0 s long AGC window has
along the N-S profile. A: short and scattered been applied for display purposes to these shot
mid-crustal events. B: coherent long events at 20 gathers.
CARBONELL AND SMITHSON 33
Offset
Fig. 2. b: Reflection recordings for shot point at 4, 6 s TWTT. A 4.0 s long AGC window has been
4-B along the E-W profile. 50-100 m receiver applied for display purposes to these shot
spacing. Heavy arrows indicate reflecting bands gathers.
the perpendicular lines possibly indicates a techniques which involve the analysis of two
dependence of the velocity with orientation on a consecutive arrivals provide reflection
crustal scale (i.e., anisotropy). coefficient estimates that reach 0.240.02 for
Although the PmP is the most prominent phase the Moho structure [Carbonell and Smithson,
in all of the shots, converted phases such as PmS 1990].
and SmP, and the SmS are also recognizable The average amplitude spectrum of the mid- and
(Figure 2b and 2c) at some offsets. The lower crustal events and of PmP at subcritical
triangular shape of these converted (PmS/SmP and offsets (Figure 4) has a characteristic peak at
SmS), for instance the SmS arrival in Figure 22-24 Hz. The peak frequencies of the converted
(2c), suggests a frequency dependence of velocity and SmS phases are shifted towards lower values
(i.e., dispersion). True relative amplitude plots (Figure 4). The PmP phase presents an anomalous
(Figure 3a) show the high amplitude obtained for abrupt high frequency attenuation beyond critical
the Moho which is of the same magnitude as the offsets (Figure 5) characteristic of tunneling
basement critical reflection. Spectral ratio waves [Fuchs and Schulz, 1976].
o
.. - <: . .: . _''-; .- .
x2.0 ..*:.=._q.. .. .,. ?, -:.:. , .. - .
24.0
26.0 :.:....::_:..:=::=._ - .:. .. _.:::
-.-- -- :-? - .... ._-:-.-,-. _. .:_-_. -.. _- ?--
-= x_,:-: ---: ,. c-_-_ --- ....- . . : - -
30.0
';:: . _ .
_ _ .
Fig. 2. c: Reflection recording for shot point 2
along the E-W profile. A 4.0 s long AGC window
has been applied for display purposes to these
shot gathers.
Forward modeling
Elastic two- and three-dimensional finite
field propagating within the model, deforming the
wave front, and redistributing the energy, thus
generating maxima and minima (Figure 6). The
synthetic shot gather recorded parallel to the
short axis of three dimensional elongated boudins
(Figure 7) shows short scattered events at mid-
crustal depth, and the seismic section recorded
along the log axes of these lenses shows strong
laterally continuous multicyclic arrivals (Figure
7b).
The two-dimensional finite difference models
demonstrate that the seismic signature of a
highly reflective horizontally layered structure
(Moho) changes from a laterally continuous event
(for small heterogeneities with a horizontal
length of less than 25 % of a Fresnel zone,
small-lens case) to a diffuse band of short
laterally discontinuous reflections (for large
heterogeneities with a horizontal length of 1-].5
Fresnel zones, large-lens case). The synthetic
seismograms generated by the two-dimensional
finite difference algorithm for the small-lens
case reveal a 0.3s time delay for the Moho
arrival with respect to the large-lens case
(Figure 8).
1 oo A
E
< 0.50
z0 O0
< 0 2.5 5.0 7.5 10.0' 12'.5
TiBe (s)

' 1.00. , .., I ........ i "
0 0.1 0.2 0.3 0.
Tie (s)
lb)
i9. 3. a: Observed true absolute ampliude-vs-
ime plo for an offse of 5,6 m for e
difference modeling of the wave field propagating gather recorded along the N-S line for shot point
through a model consisting of a flat highly 4. This plot was obtained by adding 10 traces
reflective structure overlayin by lens-shaped centered at this offset after the application of
heterogeneities, demonstrates that its seismic a spherical divergence correction using a
signature is strongly dependent on the lateral velocity function derived from Hawman et al.,
dimensions of the lenses [Carbonell and Smithson, (1990). A: Basement reflection; B: PmP phase. b:
1991]. The model (Figure 6a) consists of a 2 by 2 True absolute amplitude-vs-time plot for an
km (2 by 2 by 1 km for the three-dimensional offset of 5.6 Km for a synthetic finite
model Figure 7a) block, although this is a scaled difference modeled time section with a vertical
down model it illustrates the major features of P-velocity model shown at the top right corner.
wave-field propagation in the presence of A: reflection from a reference interface with a
heterogeneities. The scattering effects due to reflection coefficient of 0.2. B: Reflection from
the elongated inhomogeneities disrupt the wave the layered structure representing the Moho.
lures 5 and 6 were originally in color but mistakenly printed in
black and white. This corrected page represents the color as it
should be.
so
0 12 2 216 ,8 60
Frequency
Fig. 4. Mean amplitude spectrum of the three
reflecting bands present in shot gather from shot
point 4-B recorded along the E-W profile (Figure
2 b.) and mean amplitude spectrum for the SmS
phase (bottom).
The seismic sig%atures along, and 15
perpendicular to the long axes of the boudins in 16
the three dimensional mode], the characteristics l?
of the Moho reflection and its differences in 18
two-way travel time for the small and the large 19
heterogeneities of the two-dimensional models 20
closely resemble our observations. In Europe Zl
different seismic signatures depending on azimuth
have frequently been reported [Meissner, 1986;
Wever et al., 1987]. Reston [1987] suggested
similar boudin structures for the interpretation
of the seismic signatures observed in the SWAT-1
and SWAT-2 CMP profiles.
Discussion
The high reflection coefficient, estimated for
the Moho reflection, the reflectivity [Carbonell
and Smithson, 1990] and finite difference
amplitude modeling of the subcritical reflections
(Figure 3b) support a high impedance layered
structure for the Moho, achieved by alternating
high and low velocity layers. High velocities
the crust mantel boundary in comparable tectonic
environments [Smithson, 1989; Va]asek et al.,
1989]. Such a structure could reasonably account
for the dispersion observed in the converted and
SmS arrivals due to melt squirt phenomena and for
the selective high frequency attenuation observed
at postcritical offsets and presumably due to the
high impedance contrasts. However, such features
can be also accounted for by considering other
wave kinematic processes.
Highly extended terraes may be characterized
by many scales of anisotropy. These scales
include: small-scale anisotropy involving
differences in velocity due to the fabric of
individual layers which contribute to
reflectivity [Fountain et al., 1984; Valasek et
al, 1987; Christensen, 1989[, centimeters-thick
mylonitic zones, lower crustal-scale plastic
flow, and hundreds of kilometers thick
anisotropic layers [Savage et al., 1990]. Far-
offset diffractions and randomly distributed
events at subcritical offsets in the wide-angle
shot gathers [Gibson and Levander, 1988], and the
finite difference two- and three dimensional
modeling results indicate that mid- and/or lower
crustal heterogeneities may be responsible for
Offset (krn) m
(a)
=20 '
30
&o
Prop
(b)
!
1oo
25
.. 0
Vp
(c)
within the range of 7.90.3 km/s and low Fig. 5. a: Subcritical to postcritical PmP phase
velocities below 6.5 km/s. While the high from shot gather recorded from shot point 8-2.
velocity may be indicative of layers of mantle b: Amplitude spectrum of the previous phase
material, the low velocities coupled with the showing the abrupt attenuation of the high
high temperatures (800-900 C) inferred for this frequencies that takes place beyond the critical
area [Lachenbruch and Sass, 1978] strongly angle (heavy arrow). c: Synthetic modeling of the
suggests the possibility of at least partially selective postcritical attenuation Arrow heads
melted rocks of crustal or mantel origin. Similar indicate the attenuated phase (after Fuchs and
stratified media have been purposed as models for Schulz, 1976).
(a)
(b)
the strikingly different seismic signatures in
the east-west and north-south directions
[Carbonell and Smithson, 1991]. The lack of
diffractions suggests that differences in the
velocities characterizing these heterogeneities
are small, or if they are large they are tapered
[Goodwin and Thompson, 1988].
These results coupled with the constraints
placed on the thickness of the structures at mid-
and lower crustal and Moho depths by the
amplitude and phase spectra [Khudzinskii, 1962;
Berzon, 1965] of the arrivals and by the
velocities obtained by Hawman et al., [1990],
favor a geological model consisting of three
dimensional tectonic lenses (i.e., boudins)
strongly elongated parallel to the stretching
direction (with a length of about one Fresnel
zone or more in the east-west direction. This is
approximately 1-1.5 km for a depth of 12 km and a
25 hz signal, a width of about 25 % of a Fresnel
zone in the north-south direction, which is
approximately 0.4 km for a depth of 12 km and 25
Hz signal and, a thickness ranging from 80 - 250
m). Ray tracing forward modeling by Ojo and
Mereu, [1986] demonstrated that the presence of
heterogeneities and/or the existence of
topography in a reflecting interface increases
the raypath due to the multiple refractions, thus
increasing the TWTT. This can very well bias
velocity estimates towards lower velocities.
These east-west strongly elongated boudins can
account for: 1) the differences on the crustal
Fig. 6. Elastic finite-difference modeling. a: seismic signature, 2) the time delay for the PmP
Sketch of the scaled crustal model consisting of phase and, 3) the difference in the seismic
mid-crustal heterogeneities and a layered highly signatures of the Moho, observed near the
reflective structure. b: Snapshot of the wave intersection point of the two perpendicular
field propagating through the model showing the lines. The COCORP seismic reflection lines
redistribution of the energy (maxima in red) and [Klemperer et al., 1986, 1987; Allmendinger,
the distortion of the wave front. 1987] are parallel to the extension direction
(a) (b) Offset (m)
(c) Offset (m)
-fillJill 1111!iJ r ill HIIIi i tlllIllilllliIljiJllllllllllliIIIIl)lJlll)l IllIIlllIIllll liJlliilllllllllllllliiiiiiiliiiJilJJillllllllllllJFlJ')JJTJtlJiJllJllllllllj
Fig. 7. Synthetic record section obtained by
three-dimensional, elastic, finite-difference crustal lenses. b: seismograms in the Y
algorithm. a: the model with elongated mid- directions. c: seismograms in the X direction.
(a)
Offset
'lllflJI;11}, ,
thus sampling the most favorable reflectivity.
Such a model indicates large scale anisotropy for
the crust in this area and provides a geodynamic
model to accommodate extension [Hamilton, 1987].
Geologically this can be described as a network
of anastomosing shear zones delineating boudins
[Reston, 1987] or lenses (i.e., pods) that
"float'" in a matrix that have undergone greater
plastic deformation (Figure, 9).
Conclusions
Offset
Finite difference modeling of the strikingly
different seismic signature displayed by
perpendicular shots of the 1986 Nevada PASSCAL
seismic reflection experiment which roughly
sample the same crustal column suggest a
geological model for the crust in this area
characterized by large scale anisotropy. Three-
dimensional tectonic lenses (boudins) elongated
parallel to the extension direction floating in a
matrix that underwent greater plastic deformation
constitute a model with a seismic signature that
closely agrees with the observed features and
provides the necessary geodynamics to accommodate
Fig. 8. Synthetic seismo9Lms obtained by a two- extension. The high reflection coefficients
dimensional, elastic, finite-difference algorithm required to mode] the dynamic characteristics of
for the model in Figure (6a). a: Synthetic the PmP phase, and the high heat flow data,
seismograms (shot gather) for the small-lens indicate that interlayering of partially melted
model. b: Synthetic seismograms for the large- rocks (probably basalts based on the tectonic
lens case. A: Crustal events; B: PmP phase. setting) with high velocity mantle material is a
possible fine structure for the Mohorovicic
discontinuity in this area.
Acknowledqments. This research was partially
supported by NSF grant EAR-8519153. A thoughtful
review by Dr. R. Meissner greatly improved the
final version of this article.
References
Fig. 9. Structurally simplified interpretative
crustal block not drawn to scale, representing
the essential elements of a 30-31 Km thick
Allmendinger, R.W., Hauge, T.A., Hauser, E.C.,
Potter, C.J., and Oliver, J., Tectonic
heredity and the layered lower crust in the
Basin and Range province, western United
States, Continental Extension Tectonics, Geol.
Soc. Sp. Pub. 28, 223-246, 1987b.
Berzon, I.S., The determination of a thinly
layered medium by the simultaneous use of
amplitude and phase spectrum characteristics
of the layer, Bulletin (Izv.) of the Academy
of Science USSR, 5, 363-367, 1965.
crustal column for the Basin and Range. Large Carbonell, R., and Smithson, S.B., On the
scale anisotropy achieved by the strongly determination of the fine structure of crustal
elongated boudins (unshaded) at mid-crustal boundaries from the dynamic characteristics of
depths, they are delineated by anastomosing shear their reflected phases, (abstr.) th.
zones (stippled). The horizontal dimensions are: International Symposium on Deep Seismic
100 % and 25 % of a Fresnel zone for the E-W and Reflection Profilin 9 o__f the Continental
N-S line respectively (approximately 1.5 Km, and Lithosphere, 1990.
0.4 Km, for a depth of 12 Km respectively). The Carbonell, R., and Smithson, S.B., Large scale
layered Moho which consists of interlayering of anisotropy within the crust in the Basin and
melts (dark layers) or partially melted rocks Range Province, Geology, (In press), 1991.
(i.e., magma) with high velocity mantel rocks Christensen, N.I., Reflectivity and seismic
(dashed layers) is accentuated by shear properties of the deep continental crust, J.
accompanying extension and/or intrusions. Geophys. Res. 94, 17793-17804, 1989.
Fountain, D.M., Hurich, D.A., and Smithson, S.B., Miller, E.L., Gans, P.B., Right, J.E., and
Seismic reflectivity of mylonitic zones in the Sutter, J.F., Metamorphic history of the East-
crust: Geology, 12, 195-198, 1984. Central Basin and Range province: Tectonic
Fuchs, K., and Schulz, K., Tunneling low setting and relationship to magmatism, in
frequency waves through the subcrustal Ernst, W.G., ed., Metamorphism and crustal
lithosphere: J. Geophs., 42, 175-190, 1976. evolution of the western United States: Rube
Goodwin, E.B., and Thompson, G.A., The Volume VII, 649-682, 1988.
seismically reflective crust beneath highly Nevada PASSCAL 1986 Basin and Range Lithospheric
extended terranes: Evidence for it origin in Seismic Experiment Working Group, The 1986
extension, Geol. Soc. of Am. Bull., 100, 1616- PASSCAL Basin and Range Lithospheric Seismic
1626, 1988. Experiment, EOS (Trans. AGU), 69, 593-598,
Gibson, B.S., and Levander, A.R., Lower crustal 1988.
reflectivity patterns in wide-angle seismic Ojo, S.B., and Mereu, R.F., The effect of random
recordings, Geophs. Res. Let. 15, 617-620,1988. velocity functions on the travel times and
Hamilton, W., Crustal extension in the Basin and amplitude of seismic waves, Geophs. J. R.
Range province, southwestern United States. astr. Soc. 84, 607-618, 1986.
M.P. Coward, J.F. Dewey and P.L. Handcock ed., Reston T.J., Spatial Interference, reflection
Continental Extensional Tectonics: Geol. Soc. character and the structure of the lower crust
SD. Pub. 28, 155-176, 1987. under extension. Results from 2-D seismic
Hauser, E., Potter, C.J., Hauge, T.A., Burgess, modeling, Annales Geophsicae, , 339-348,
S., Burtch, S., Mutschler, J., Allmendinger, 1987.
R.W, Brown, L., Kaufman S., and Oliver, J., Savage, M.K., Silver, P.G., and Meyer, R.P.,
Crustal structure of the eastern Nevada from Observations of teleseismic shear-wave
COCORP deep seismic reflection data, Geol. splitting in the Basin and Range from portable
Soc. of Am. Bull. 99, 833-844, 1987 and permanent stations, Geophs Res. Let.,
Hawman, R.B., Colburn, R.H., Walker D.A., and 17, 21-24, 1990.
Smithson, S.B., Processing and interpretation Smithson, S.B., Contrasting types of lower crust:
of refraction and wide-angle reflection data in Mereu, R.F., Muller, S., Fountain, D.M.,
from the 1986, Nevada PASSCAL Experiment, J. Properties and Processes of the Earth's Lower
Geophys. Res. 95, 4657-4691, 1990. Crust, Geophysical Monographs Series, American
Khudzinskii, L.L., On the determination of Geophysical Union, Washington, D.C., Vol., 51,
parameters of layers of intermediate thickness 53-63, 1989.
from the spectra of reflected waves, Bulletin Snoke, A.W., and Miller, D.M., Metamorphic and
(Izv.) of the Academy of Science USSR, 5, 439- tectonic history of the north eastern Great
444, 1962. Basin, in Ernst, W.G., ed., Metamorphism and
Klemperer, S.L., Hauge, T.A., Oliver, J., and crustal evolution of the western United
Potter, C.J., The Moho in northern Basin and States: Rube Volume VII, 606-648, 1988.
Range province, Nevada, along the COCORP 40 N Speed, R.C., Elison, M.W., and Heck, F.R.,
seismic-reflection transect: Geol. Soc. of Am. Phanerozoic tectonic evolution of the Great
Bull., 97, 603-813, 1986. Basin, in Ernst, W.G., ed., Metamorphism and
Klemperer, S.L., A relation between continental crustal evolution of the western United
heat flow and the seismic reflectivity of the States: Rube Volume VII, 572-605, 1988.
lower crust: J. Geophys. Res., 61, 1-11, 1987. Valasek, P.A., Hawman, R.B., Johnson, R.A., and
Lachenbruch, A.H., and Sass, J.H., Models of an Smithson, S.B., Nature of the lower crust and
extending lithosphere and heat flow in the Moho in Eastern Nevada from "Wide-Angle"
Basin and Range Province, in Smith, R.B., and reflection measurements, Geophvs. Res. Let.
Eaton, G.P., eds., Cenozoic Tectonics and 1--4, 1111-1114, 1987
Regional Geophysics of the Western Cordillera: Valasek, P.A, Snoke, A.W., Hurich, C.A., and
Geol. Soc. of America Memoir, 152,209-250,1978. Smithson, S.B., Nature and origin of seismic
Meissner, R., The continental crust: 6 reflection fabric, Ruby-East Humboldt,
geophysical approach, Academic Press. London, metamorphic core complex, Nevada, Tectonics,
New York, 1986. , 391-415, 1989
Miller, E.L., Gans, P.B., and Garing, J., The Wever, Th., Trappe, H., Meissner, R., Possible
Snake Range decollement-An exhumed mid- relations between crustal reflectivity,
tertiary ductile-brittle transition, crustal age, heat flow, and viscosity of the
Tectonics, , 239-263, 1983. continents, Annales Geophsicae, 5B, 255-266.
SEISMIC REFLECTORS IN HIGH-GRADE METAMORPHIC ROCKS OF THE
KAPUSKASING UPLIFT: RESULTS OF PRELIMINARY DRILL SITE SURVEYS
Bernd Milkereit, John A. Percival, Don White, Alan G. Green and Matt H. Salisbury
Geological Survey of Canada, Ottawa, K1A 0Y3, Canada
Abstract. The nature and origin of deep-crustal reflections can be
studied in the Kapuskasing uplift of the central Superior Province,
where middle to lower crustal rocks breach the surface. High seismic
velocities and reflection zones can be traced up dip from near to their
normal depths at > 20 km to within a few hundred metres of the
surface. The gently dipping reflection zones, defined by
LITHOPROBE Vibroseis sections, project to a high-grade metamorph-
ic sequence of interlayered mafic and felsic gneiss and tonalitc.
Physical property studies of representative rock samples from the study
area confirm the presence of high densities and high compressional
wave velocities at shallow depths. Additional dynamite high resolution
seismic reflection profiling was carried out in 1989 as a preliminary site
survey for the Canadian Continental Drilling Program (CCDP). The
main goal of the survey was to provide lateral control of the laminated
reflections observed at shallow depths on the Vibroseis profiles. The
strong reflections seen on the Vibroseis profiles can be traced with
confidence into the dynamite profiles. Amplitude-versus-offset (avo)
studies of these shallow reflections show an amplitude decay with offset
that is compatible with a simple model of interlayered mafic, felsic and
tonalitic rocks.
Introduction and Motivation
During the past decade, the application of deep seismic reflection
profiling has given rise to considerable speculation and controversy
about the reflectivity of the lower continental crust, the variation of
reflectivity patterns, and the origin of deep seismic reflections (e.g.
Klemperer et al., [1987]). For example, the presence of basaltic sills
[Warner, 1990],free fluids [Matthews and Cheadle, 1986;Hyndman and
Shearer, 1989], depositional layering [Cannon et al., 1989], highly
strained contacts between compositionally layered gneissic rock
[Smithson et al., 1977, 1986; Green et al., 1990] and shear zones
[Reston, 1990a] have been proposed to explain some of the observed
lower crustal reflectivity patterns.
Deciphering Lower Crustal Reflectivity From Surface Observations
Approaches to resolving the causes of reflectivity in the crystalline
crust fall into three principal categories. The first involves placing
reasonable bounds on the physical parameters of crustal rocks by
conducting either detailed analysis and modelling of recorded wave field
data (e.g. Sandmeier et al. [1987];Hobbs, [ 1990]) or specially designed
Geodynamics 22
geophysical experiments such as combined seismic reflection and
refraction surveys (e.g. Gajewski et al., [1987];Holbrook et al. [1991]),
shear wave studies (e.g. Liischen et al. [1990])or electromagnetic
soundings (e.g. Jones [1987]).
The second approach is to drill seismic reflectors and study their
physical properties in situ. Scientific drilling in basement rocks has
provided a wide range of explanations: open fracture zones are the
source of strong reflections in a Canadian granitic body [Green and
Mair, 1983]; reflections from the Siljan Ring in Sweden appear to be
the result of basaltic sills or their metamorphic equivalents [Juhlin,
1990];in the Kola Superdeep hole, seismic reflections may be correlated
with high porosity zones [Mints et al., 1987]; and reflections from
mylonitic gneisses and fractures have been identified in a 5.5 km deep
well in a highly extended terrane of the Cordillera in southeastern
Arizona [Goodwin and Thompson, 1988]. However, drilling technology
does not yet permit direct sampling of reflectors in the lower crust.
The third approach is to study lower crustal, high-grade rocks now
exposed at the earth's surface. For example, Green et al. [1990]
presented seismic reflection data from high-grade metamorphic terranes
in Canada characterized by prominent reflection zones that can be
correlated with surface features such as highly strained contacts
between compositionally layered gneissic rock. Compositional layering
has also been identified as the main source of reflectivity in a physical
rock property study by Christensen [1989]; synthetic seismograms
generated from shallow drill core data recovered from a high-grade
metamorphic terrane located in the Inner Piedmont showed that seismic
reflections may originate from subhorizontal metamorphic layering.
Exposed sheets of high-grade rocks in tilted cross-sections of the
continental crust offer the opportunity to study the origin of lower
crustal reflectivity. Based on petrological and physical rock property
studies, nearly continuous cross-sections have been identified in the
Ivrea zone (e.g. Burke and Fountain [1990]), Calabria [Kern and
Schenk, 1988] and the Kapuskasing uplift [Percival and Card, 1983;
Fountain et al., 1990]. In the Kapuskasing uplift, seismic reflections can
be traced from lower and middle crustal levels to the surface [Percival
et al., 1989; Geis et al., 1990].
Geological Setting
The Kapuskasing uplift (KU) is a northeast-trending, fault-bounded
zone of medium- to high-grade metamorphic rocks in the Superior
Province, the largest Archcan craton (Fig. 1). The sequence exposed
in the KU has been interpreted as an oblique cross section of Archcan
crust, thrust to the surface along the west-dipping Ivanhoe Lake fault
zone (ILFZ). Brittle deformation associated with the early Proterozoic
uplift occurred mainly along the ILFZ [Bursnail, 1990]. Lithologically,
the cross section reveals a transition from shallow-level metavolcanic
rocks and tonalitic gneiss (greenschist-amphibolite facies) in the Wawa
39
40 SEISMIC REFLECTORS IN HIGH-GRADE METAMORPHIC ROCKS
Fig. 1. Geology of the central Superior Province showing location of
the Kapuskasing structure, the exposed midcrustal discontinuity (C-C)
and cross section A-B of Figure 2. Area of site surveys for the proposed
drill site is indicated by ,. Inset: location of the Superior Province
(diagonal ruling and inside box).
gneiss terrane to a deeper-level, layered heterogeneous sequence (upper
amphibolite and granulite facies) in the Kapuskasing zone. The
high-grade rocks were thrust onto low-grade metavolcanic rocks of the
Abitibi greenstone belt to the east. Based on a wide range of geologi-
cal and geophysical studies a crustal-scale cross section (Fig. 2) through
the KU was constructed by Percival et al. [1989], showing horizontal
shortening by brittle upper-crustal thrusting and ductile lower-crustal
thickening (see also Handy [1990]). Key evidence for this model of
uplifted middle to lower crust in the KU includes: a systematic eastward
increase in palcopressure and metamorphic grade [Percival, 1983];
eastward increases in the density and velocity of surface rocks [Fountain
et al., 1990] and the identification of a possible Conrad discontinuity
[Percival, 1986]; a paired gravity anomaly (60 mGal peak to trough)
modelled as a west-dipping slab of high density material [Percival and
WAW A Kaouskasing A B IT I B I
,. SUBPROVINCE C Zone. /,7zU B PR OVl IE ip BO
...... ,.. .
,1 2 3 4
Fig. 2. Cross section A-B across the Kapuskasing uplift (no horizontal
exaggeration) summarizing seismic reflection and refraction data
(modified after Percival et al., 1989). The reflective sequence projects
to the surface about 10 km west of the ILFZ (Ivanhoe Lake fault zone).
The location of high resolution seismic surveys is indicated by a, and
data example is shown in Figure 3. C = midcrustal discontinuity; 1 =
metavoleanic rocks; 2 = tonalitic rocks; 3 = intefiayered marie gneiss,
paragneiss and tonalitic gneiss; 4 = anorthosite.
McGrath, 1986]; a uniform westward decrease of isotopic ages,
indicative of top-down cooling of the former crustal column [Percival
and Krogh, 1983]; a zone of high crustal velocities 7.0 km/s in
places) that can be traced from the surface in the KU to mid-crustal
depths [Boland and Ellis, 1989]; and a 5 to 10 km increase in crustal
thickness beneath the KU [Boland and Ellis, 1989].
Seismic Reflection Data
Under the LITHOPROBE program, regional and high-resolution
seismic surveys were conducted across the KU in 1987/88. The
reflection data image the Kapuskasing structure as a relatively thin
thrust sheet [Percival et al., 1989], bounded by low angle faults that
merge into a detachment with ramp anf flat geometry [Geis et al.,
1990]. These faults, which have a fiat zone at 10-12 km depth, are
responsible for the emplacement of a seismically reflective sequence of
amphibolite and granulite facies rocks that have been interpreted to be
analogues of seismically reflective 'in situ' middle to lower crust. The
shallow reflective sequence is the subject of this pre-drilling site study.
Location of the study area with respect to the KU and the proposed
cross section is indicated by a star (*) in Figures 1 and 2.
A key portion of the seismic reflection data that contributed to the
cross section of the KU is shown in Figure 3. The migrated
high-resolution section images west-dipping reflections projecting to the
surface in the ILFZ, which is 1-2 km wide at this location. At the
western end of the profile, an approximately 1.5 km thick lamella of
high-amplitude reflections projects toward the surface to a metamorphic
sequence of interlayered marie and felsic gneiss and tonalitc. The
prominent west-dipping reflections and features associated with the
ILFZ are underlain by subhorizontal and east-dipping events of
unknown origin. For a more detailed discussion of the high-resolution
V,,. ANORTHOSITE AND
ANORTHOSITIC GABBRO
-k
0 .
4
_L .,,,,.,ss .D _J 'v:" I
CATA -
CLASTIC
-I-FOLIATED TONAL'TE-I ZON
CHAPLEAU BLOCK (GRANULITE FACIES) 0 km 5
LINE 2 MIGRATED I I I
,
Fig. 3. F-K migrated seismic section from central Kapuskasing uplift.
Shallow layered, west-dipping reflections can be projected to near the
surface in the Kapuskasing uplift. Vibroseis data were recorded with the
following parameters: 2 vibrators, 20-130Hz sweeps, 120-channels, 2 ms
sample interval, 20-m station and source interval, 60 fold coverage.
MILKEREIT ET AL. 41
Vibroseis data from the KU see Green et al. [1990] and Geis et al.
[ 1990].
In this paper we report on the findings of preliminary site surveys
conducted in preparation for a proposal to drill the lamella reflections,
which cannot be correlated directly with surface exposures. In
particular, new rock property and reflection seismic studies have been
conducted. Results of detailed geological mapping and exposure
enhancement in the central KU area are summarized by Percival et al.
[ 1990]. The main objective of the rock property study was to fine tune
the existing regional data base and assess the role of anisotropy in
generating reflections; high-resolution seismic reflection studies were
carried out to establish true dip and lateral continuity of the shallow,
apparently gently dipping sequence of layered reflections. The mapped
image of these reflections will subsequently be used to (1) define a
drilling target and (2) extrapolate results from core studies and
borehole geophysics away from the hole, thereby overcoming some of
the constraints imposed by single site studies.
Physical Rock Property Studies
A model of the uppermost lower continental crust has been derived
from compressional wave velocities (Vp) measured at confining
pressures of up to 600 MPa for representative rock samples from
throughout the Kapuskasing uplift [Fountain et al., 1990]. The model
includes a layered, seismically heterogeneous zone with an average
velocity of 6.84km/s. The laboratory data support refraction results
that show a high velocity zone beneath a large area centred on the KU
[Boland and Ellis, 1989].
Representative rock samples from some of the major lithologic
units (anorthosite, mafic gneiss and tonalite) in the vicinityof the study
area were collected and densities and compressional wave velocities
were determined. Compressional wave velocities at various confining
pressures up to 600 MPa for cores were cut normal and perpendicular
to foliation are summarized in Figure 4.
The tonalite has the lowest velocities (< 6.8 km/s at 600 MPa) and
is seismically anisotropic. In contrast, the anorthosite has relatively high
velocities (< 7.4 km/s) and is only weakly anisotropic. The mafic
gneiss has the highest velocities (< 7.6km/s) and significant anisotropy.
Seismic velocities are higher parallel to foliation than normal to it.
Seismic anisotropy in mafic gneiss has been related to the preferred
orientation of hornblende [Percival and Fountain, 1989]. Bulk densities
of 2.78 g/cm 3 for anorthosite, 3.03 g/cm 3 for mafic gneiss and 2.81
g/cm 3 for tonalite were determined.
These new physical property studies of representative rock samples
from the study area confirm the presence of high densities and high
compressional wave velocities. Average velocity and density contrasts
between typical KU rock units in the study area are as high as 0.7km/s
and 0.25 g/cm respectively, with reflection coefficients for layered
sequences as high as 0.1. These observations confirm earlier findings
based on regional sampling of the KU reported by Percival and
Fountain [1989] and Fountain et al. [1990]. The regional Vp values
were used by Reston [1990b] to constuct a synthetic seismogram. The
scale of layering (metres to kilometres) and high reflection coefficients
predict significant reflections from at least the layered mafic/tonalitic
gneiss sequence.
More Seismic Data
A dynamite high-resolution seismic reflection survey was carded out
in 1989 as a preliminary site survey for the Canadian Continental
Drilling Program (CCDP). The main goal of the survey was to provide
true dips and lateral control of the laminated reflections observed at
shallow depths on the high-resolution (sweep frequencies 20-130 Hz)
LITHOPROBE Vibroseis profile (Fig. 3). To accomplish this,
".' TONALITE I[ for
, TONALITE I-fol ,.
' MAFIC GNEISS I[ fol
'. MAgiC GNEISS ii._!__ol,
--ANORTttOSITE
' --sorosir

0 00 200 300 400 500 600
Pressure (MPo)
7.4
5.8
Fig. 4. Compressional wave velocity data for a tonalite, anorthosite and
mafic gneiss from the survey area as a function of pressure. Above
about 100 MPa confining pressure, velocities greater than 7.0km/s are
observed for anorthosite and mafic gneiss. Velocity measurements
parallel and perpendicular to foliation demonstrate velocity anisotropy.
dynamite profiles were run nearly perpendicular to the Vibroseis
profile. Basic data acquisition parameters for the Vibroseis and
dynamite surveys are documented in Milkereit et al. [1990] and are
summarized in the captions to Figures 3 and 5, respectively.
Important data processing steps included (1) interactive first break
analysis and computation of static corrections to compensate for
significant variations in overburden thickness along the profiles,
(2) deconvolution and filter tests of the dynamite data, which exhibited
high freqencies between 60 and 100 Hz, and (3)spectral balancing to
allow comparison between the dynamite and Vibroseis sections.
Static Corrections
Although high frequency seismic data may improve both lateral and
vertical resolution, such data require the computation of accurate static
corrections in order to compensate for near-surface inhomogeneities.
All first break travel times have been picked and refraction static
corrections have been computed. The weathering models showed a
large variation in overburden thickness (glacial drift) along the dynamite
profile, from 0 m (basement outcrop) to more than 20 metres over only
a few surface stations.
Figure 5 illustrates how important accui'ate static corrections are for
processing high frequency seismic reflection data from crystalline crust.
The application of simple elevation corrections (step 1) or long
wavelength static corrections (step 2) does not yield a high quality
seismic image (24-fold dynamite data example). The seismic image
improves only after applying the short and long wavelength static
corrections in step 3. This example emphasizes that the computation
of static corrections is the singlemost important step for processing high
resolution seismic data. All subsequent processing and data analysis
(e.g. velocity analysis, trim statics, etc.) are based on the static corrected
500m 500m
i i
STEP I STEP 2
500m 500m
o ................................ : ........................................ : ........................... ;- ......... ';:: ................................. 0
g.: ::. .... ,,,..=:.. ,::,:,..:.........;,. ...: :::,:!i..';:'...i. ?-
7-:,:-:,: r,. :. :::::,:' :,:;, -"::: . :::'., . ': ,,..'7.. ,.', "7. -:" ' ' ': :' ,'l..'...--.' ','. .....
STEP 3 STEP 4
Fig. 5. Example of the effect of static corrections for high resolution
seismic data. The application of simple elevation corrections (step 1) or
long wavelength static corrections (step 2) does not provide an
acceptable seismic image. Most processilg steps (e.g. velocity analysis)
for high frequency data require the use of total (short and long
wavelength) static corrected data (step 3). The final stack (step 4) with
updated stacking velocities and correlation statics applied, is based on
processing step 3. Dynamite data were recorded with the following
parameters: 0.25 kg charges in shallow drill holes, 96-channels, 2 ms
sample interval, 12.5-m station and 25-m source interval, 24-fold
coverage.
data of step 3. The final stack, with updated mutes, stacking velocities,
trim statics, bandpass filter and coherency enhancement is shown in
step 4.
The Target
Spectral enhancement in the 50 to 100 Hz range has been applied
to the Vibroseis CMP stacked section to match the high frequency
content of the dynamite survey. At the intersection of the dynamite and
Vibroseis profiles, both data sets show good signal-to-noise ratios
(Fig. 6). The strong reflections observed on the east-west Vibroseis
profile (Fig. 3) can be correlated confidently with similar events on the
the north-south dynamite profile. The fence diagram shows the
coincidence of reflector positions at the point of intersection. The
continuous reflections (at about 300 to 350 ms on the east-west profile
and at 250 to 300 ms on the north-south profile, indicated by an arrow)
provide good lateral control of the subsurface reflectivity. In addition,
they define a true northwesterly dip of 15 degrees for these events.
The zone of high reflectivity imaged in Figure 6 defines a target to
study the nature and origin of reflectors in high-grade metamorphic
rock. In particular, the target is accessible since prominent reflectors
occur at relatively shallow depths; the reflectors can be traced from
near the surface to mid-crustal levels (permitted by regional
LITHOPROBE data); the lateral continuity ofrefiectors overhundreds
of meters and gentle dips in the study area will allow modelling and
extrapolation of geophysical logging and coring data in the vicinity of
the drill site.
Origin of Reflectors in High-Grade Rocks
Amplitude-versus-offset (avo) data recorded over shallow reflectors
in the crystalline crust may provide clues to what causes crustal
reflections . Figure 7 shows a true amplitude shot gather (dynamite
source) from the proposed drill site area. Data are highpass filtered
75 Hz) in order to suppress shear waves, first breaks (refracted
phases) are muted, noisy traces are edited and geometrical spreading
MILKEREIT ET AL. 43
Fig. 6. Fence diagram of high-resolution Vibroseis and dynamite
profiles. Position of shallow seismic reflector is indicated by arrow.
Spectral balancing (60 - 130 Hz) has been applied to the Vibroseis
reflection data to match the high frequency content of the dynamite
survey. Note the continuous reflections at 350 to 300 ms on the E-W
Vibroseis profile and at 300 to 250 ms on the N-S dynamite profile
(indicated by an arrow). These reflections show an overall N-W dip in
the survey area. The location of the proposed drill site is indicated by
ar.
corrections proportional to time are applied. The grey-scale plot is
coded according to the trace envelope of the true amplitude data.
Amplitude variations of more than 40 dB are observed. An important
observation is that the amplitude associated with a shallow reflection
(indicated by an arrow in Fig. 7) decays with offset (-10 dB over 1 km).
Maximum source-receiver offset is 1.2 km, reflector depth is
approxilnately 850 metres, and reflector dip is less than 15 degrees.
These field parameters limit the maximum angle of incidence at the
reflector to _< 40 degrees. The observed negative ave trend probably
indicates seismic energy partitioning as a function of angle of incidence
(uniformly strong first breaks demonstrate that the decay is not due to
changing near surface conditions). Energy partitioning requires a
significant change in elastic properties (compressional and shear wave
velocities, densities) at depth. In light of the observed seismic
anisotropy of rock samples from the study area (Fig. 4) we do not
attempt to invert the observed ave trend for changes in elastic
properties. Nevertheless, the negative ave trend shown in Figure 7
places centtaints on the variation of Poisson's ratio with depth [Louie
and Clayton, 1987]; the most likely explanation being a decrease of
Poisson's ratio at approximately 850 m. The observed amplitude decay
does not preclude a thin layer model. Comparison of the Kapuskasing
ave data with those reported for the Siljan Ring reflections [Juhlin,
1990] show the amplitude data to be consistent with a thin high-velocity
layer model. However, a significant shift of reflection peak frequency
with offset is a key indicator for thin layering [Louie and Clayton, 1987].
For example, frequency effects would be expected for a dynamite source
with a dominant frequency of about 90 Hz, a thin high velocity layer of
7200 m/s and a layer thickness of about 20 metres. Such a significant
shift of reflection peak frequency with offset is not clearly observed in
the KU data, but it may be difficult to measure (see examples given by
Juhlin [ 1990]).
Discussion and Summary
In an attempt to determine the source of middle and lower crustal
reflectivity a number of regional and high resolution seismic reflection
profiles have been recorded across the Kapuskasing uplift. Crossing
Vibroseis and dynamite profiles reveal a prominent package of gently
dipping reflections at shallow depth. Reflections may arise from
interlayered high-grade mafic and felsic gneisses and tonalitc, like those
observed and sampled at the surface; there is now no doubt that lower
crustal, upper ampibolite to granulite facies rock assemblages do reflect.
(km) 1.0
o!
-40-_
dB
0.5 0
'T" .... :: "':'h"':[:!';'" ',:i.:".i[i: " '
Fig. 7. A true amplitude shot gather from the southern end of the
dynamite reflection profile. Data are bandpass filtered (75 to 130 Hz),
first breaks are muted and geometrical spreading corrections are
applied (proportional to time). The grey-scale plot is coded according
to the trace envelope of the true amplitude data. Amplitude variations
of more than 40 dB are observed. Note the amplitude decay with offset
(-10 dB) that is associated with a shallow reflection (arrow). This may
indicate seismic energy partitioning as a function of angle of incidence.
Energy partitioning requires a significant change of elastic properties
(compressional and shear wave velocities, densities) at approximately
850 m depth (see also stacked section, Figure 6).
Measured densities and velocities in the layered sequence indicate that
it is a reasonable analogue for reflections observed in the middle to
deep crust.
These conclusions lend support to the widely held view that deep
crustal reflectivity results from interlayering of marie and felsic rocks.
However, unlike many models that relate the marie layers to late
basaltic sills (e.g., Warner [ 1990];Juhlin [ 1990]; Holbrook et al. [ 1991 ];
Nelson [in press]), the Kapuskasing uplift reveals an older marie
component that may comprise much of the lower crust [Boland and
Ellis, 1989; Fountain et al., 1990], intruded by tonalitic sheets. The
reflective package occurs within an interlayered transition zone between
the predominently tonalitic upper crust and the more marie lower crust.
This observation may have widespread implications for the
interpretation of reflectivity in terranes without evidence of extension
or underplating.
Drilling is still required to correlate the reflections with specific
lithologieal or structural interfaces. It is unclear as to whether the
reflections originate from simple intrusive contacts between sheet-like
bodies of marie, tonalitic and anorthositic rock, from coeval and later
intense compressional stress zones that are related to displacement of
rock units under high pressure, or from late tectonic extensional
lozenges within the same rock sequence. Structurally imposed seismic
anisotropy could be an important factor in enhancing reflectivity in the
latter two interpretations (but see Christensen [1989]). The planned
drilling project will identify the reflective zone by direct penetration,
followed by geophysical logging, core studies and vertical seismic
profiling.
The successful high-frequency seismic studies in the Kapuskasing
uplift, together with reports of reflected energy with frequencies as high
as 80 or 100 Hz from dynamite and high frequency Vibroseis surveys
across crystalline basement [Roy-Chowdury and Phinney, 1987; Goleby
et al., 1987], raises an important question in reflection seismology-
whether many of the earlier conclusions on reflectivity patterns based
on relatively low frequency signals, are valid or not?
Acknowledgments. We thank C. Spencer, A. Jones and S. Smithson
for helpful reviews of the manuscript. Vibroseis reflection data were
acquired and initially processed by Veritas Geophysical and Veritas
Seismic under contract to LITHOPROBE; the 1989 dynamite data were
acquired by GPR International and initially processed by Western
Geophysical under contract to the Canadian Continental Drilling
Program (CCDP). Funding for LITHOPROBE is provided by NSERC
and the Geological Survey of Canada; funding for CCDP's pre-site
seismic survey was provided by the Ontario Geological Survey.
Geological Survey of Canada contribution 41990 and LITHOPROBE
publication 217.
References
Boland, A.V. and R.M. Ellis, Velocity structure of the Kapuskasing
uplift, northern Ontario, from seismic refraction studies, J. Geophys.
Res., 94, 7189- 7204, 1989.
Burke, M.M. and D.M Fountain, Seismic properties of rocks from an
extended continental crust - new laboratory measurements from the
Ivrea Zone, Tectonophysics, 182, 119 - 146, 1990.
Bursnail, J.T., Deformation sequence in the southeastern Kapuskasing
structural zone, Ivanhoe Lake, Ontario, Canada, in: M.H. Salisbury
and D.M. Fountain (eds.), Exposed Cross-Sections of the Continental
Crust, Kluwer Academic Publishers, Dortrecht, p. 469 - 484, 1990.
Cannon, W.F.,A.G. Green, D.R. Hutchinson, M.W. Lee, B. Milkereit,
J.C. Behrendt, H.C. Halls, J.C. Green, A.B. Dickas, G.B. Morey, R.
Sutcliffe and C. Spencer, The North American Midcontinent Rift
beneath Lake Superior from GLIMPCE seismic reflection profiling,
Tectonics, 8, 305- 332, 1989.
Christensen, N.I.,Refiectivity and seismic properties of the deep
continental crust, J. Geophys. Res., 94, 17793 - 17804, 1989.
Fountain, D.M.,M.H. Salisbury and J. Percival, Seismic structure of the
continental crust based on rock velocity measurements from the
Kapuskasing uplift, J. Geophys. Res. 95, 1167 - 1186, 1990.
Gajewski, D., W.S. Holbrook, and C. Prodehl, Combined seismic
reflection and refraction profiling in southwestern Germany -
detailed velocity mapping by the refraction survey, Geophys. J.R. astr.
Soc., 89,333- 338, 1987.
Geis, W.T.,F.A. Cook, A.G. Green, B. Milkereit, J.A. Percival, and
G.F. West, Thin thrust sheet formation of the Kapuskasing Structure
Zone revealed by Lithoprobe seismic reflection data, Geology, 18,513
- 516, 1990.
Goodwin, E.B. and G.A. Thompson, The seismically reflective crust
beneath highly extended terranes; evidence for its origin in extension,
Geol. Soc. Am. Bull., 100, 1616- 1626, 1988.
Goleby, B.R., C. Wright, B.L.N. Kennett, Preliminary deep reflection
studies in the Arunta Block, Central Australia, Geophys. J.R. astr.
Soc., 89,437 - 442, 1987.
Green, A.G. and J.A. Mair, Subhorizontal fractures in a granitic pluton:
their detection and implication for radioactive waste disposals,
Geophysics, 48, 1428- 1449, 1983.
Green, A.G.,B. Milkereit, J. Percival, A. Davidson, R. Parrish, F. Cook,
W. Geiss, W. Cannon, D. Hutchinson, G. West, R. Clowes, Origin
of deep crustal reflections: results from seismic reflection profiling
across high grade metamorphic terranes in Canada, Tectonophysics,
173,627 - 638, 1990.
Handy, M.R.,The exumation of cross sections of the continental crust:
structure, kinematics and rheology, in: M.H. Salisbury and D.M.
Fountain (eds.), Exposed Cross-Sections of the Continental Crust,
Kluwer Academic Publishers, Dortrecht, p. 485 - 507, 1990.
Hobbs, R.W., Effective Q determination using frequency methods on
BIRPS data, Tectonophysics, 173, 25- 30, 1990.
Holbrook, W.S.,R.D. Catchings and C.M. Jarchow, Origin of deep
crustal reflections: implications of coincident seismic refraction and
reflection data in Nevada, Geology, 19, 175 - 179, 1991.
Hyndman, R.D. and P.M. Shearer, Water in the lower continental crust:
modelling magnetotelluric and seismic reflection results, Geophys. J.
Int., 98,343 - 365, 1989.
Jones, A.G.,MT and reflection: an essential combination, Geophys. ZR.
astr. Soc., 89, 7 - 18, 1987.
Juhlin, C., Interpretation of the reflections in the Siljan Ring area based
on results from Gravberg-1 borehole, Tectonopltysics, 173,345- 360,
1990.
Kern, H. and V. Schenk, A model of velocity structure beneath
Calabria, southern Italy, based on laboratory data, Earth and
Planetary Science Letters, 87,325- 337, 1988.
Klemperer, S.L. and the BIRPS group, Reflectivity of the crystalline
crust: hypotheses and tests, Geophys. J.R. astr. Soc., 89, 217- 222,1987.
Louie, J.N. and R.W. Clayton, The nature of deep crustal structures in
the Mojave Desert, California, Geophys. J.R. astr. Soc., 89,125 - 132,
1987.
Lfischen, E., B. Nolte and K. Fuchs, Shear-wave evidence for an
anisotropic lower crust beneath the Black Forest, southwest
Germany, Tectonophysics, 173,483 - 493, 1990.
Matthews, D.H. and M.J. Cheadle, Deep reflections from the
Caledonides and Variscides west of Britain and comparison with the
Himalayas, in: M. Barazangi and L.D. Brown (eds.), Reflection
Seismology: A Global Perspective, Amer. Geophys. Un. Geodyn. ser.,
13, p. 5-19, 1986.
Milkereit, B., D.J. White, J.A. Percival, K. Vasudevan and P.C.
Thurston, High resolution seismic reflection profiles across the
Kapuskasing structure, Ontario Geological Survey Open File 5781,
1991.
Mints, M.V.,N.I. Kolpakov, V.S. Lanev and M.S. Rusanov, The
character of the subhorizontal seismic boundaries within the upper
part of the earth's crust (according to data from the Kola ultradeep
well), Geotectonics, 21,444- 451, 1987.
Nelson, K.D., A unified view of Craton evolution, Tectonophysics, in
press.
MILKEREIT ET AL. 45
Percival, J.A.,High-grade metamorphism in the Chapleau-Foleyet area,
Ontario, Amer. Mineral., 68,667 - 686, 1983.
Percival, J.A.,A possible exposed Conrad discontinuity in the
Kapuskasing uplift, Ontario, in: M. Barazangi and L.D.Brown (eds.),
Reflection Seismology: The Continental Crust, Amer. Geophys. Un.
Geodyn. ser., 14, p. 135- 141, 1986.
Percival, J.A. and K.D. Card, Archean crust as revealed in the
Kapuskasing uplift, Superior province, Canada, Geology, 11,323-
326, 1983.
Percival, J.A. and D.M. Fountain, Metamorphism and melting at an
exposed example of the Conrad discontinuity, Kapuskasing uplift,
Canada, in: D. Bridgwater (ed.), Fluid movements - Element transport
and the composition of the deep crust, Kluwer Academic Publishers,
51 - 60, 1989.
Percival, J.A. and T.E. Krogh, U-Pb zircon geochronology of the
Kapuskasing structural zone and vicinity in the Chapleau-Foleyet
area, Ontario, Can. J. Earth Sci., 20, 830- 843, 1983.
Percival, J.A. and P.H. McGrath, Deep structure and tectonic history
of the northern Kapuskasing uplift of Ontario: an integrated
petrological-geophysical study, Tectonics, 5,553 - 572, 1986.
Percival, J.A.,A.G. Green, B. Milkereit, F.A. Cook, W. Geis and G.
West, Seismic reflection profiles across deep continental crust
exposed in the Kapuskasing uplift structure, Nature, 342, 416 - 420,
1989.
Percival, J.A.,J.T. Bursnail, D.E. Moser and D.M. Shaw, Site survey for
the Canadian Continental Drilling Program's pilot project in the
Kapuskasing uplift, Ontario Geological Survey Open File (in press),
1991.
Reston, T.J., Shear in the lower crust during extension: not so pure and
simple: Tectonophysics, 173, 175- 183, 1990a.
Reston, T.J, The structure of the crust and uppermost mantle offshore
Britain, deep seismic Reflection profiling and crustal cross-sections,
in: M.H. Salisbury and D.M. Fountain (eds.), Exposed Cross-Sections
of the Continental Crust, Kluwer Academic Publishers, Dortrecht, p.
603 - 621, 1990b.
Roy-Chowdhury, K. and R.A. Phinney, Improved resolution of
reflections from the crystalline upper crust, Geophys. J.R. astr. Soc.,
89, 35 - 40, 1987.
Sandmeier, K.J., W. W.lde and F. Wenzel, Physical properties and
structure of the lower crust revealed by one- and two-dimensional
modelling, Geophys. J.R. astr. Soc., 89,339 - 344, 1987.
Smithson, S.B. ,P.N. Shive and S.K. Brown, Seismic velocity,reflections,
and structure of the crystalline crust, in: Amer. Geophys. Un.
Geophysical Monograph 20, The Earth's Crust, 254- 270, 1977.
Smithson, S.B.,R.A. Johnson and C.A. Hurich, Crustal reflections and
crustal structure, in: M. Barazangi and L.D. Brown (eds.), Reflection
Seisinology: The Continental crust, Amer. Geophys. Un. Geodyn. set.,
14, p. 21 - 32, 1986.
Warner, M., Basalts, water or shear zones in the lower continental
crust?, Tectonophysics, 173,163 - 174, 1990.
SEISMIC REFLECTIVITY PATrERNS OF THE KAPUSKASING STRUCTURAL ZONE
Jianjun Wu and Robert F. Mereu
Department of Geophysics, University of Western Ontario
London, Ontario, Canada. N6A 5B?
Abstract. In 1987-88 a deep seismic reflection 1987; Mooney and Brocher, 1987; Smithson et al., 1987;
survey was conducted over the Kapuskasing structural Wever et at, 1987; and Meissner et at., 1990).
zone (KSZ) in the Canadian shield as a part of the A consensus which has emerged from these studies is
LITHOPROBE Project. Approximately 360 km of vibroseis that a transparent upper crust overlying a reflective
data were acquired along 9 profiles which crossed the lower crust with the Moho clearly defined by layered
boundaries between the KSZ and Val Rita Block to the reflections is a characteristic of young extensional
west and between the KSZ and the Abitibi belt to the terranes, while a reflective upper crust underlain by
east. a transparent lower crust with the Moho poorly defined
A statistical analysis of the data reveals that may be more characteristic of old shield areas. In
there are large differences in the reflectivity pat- this paper it is shown that contrasting reflectivity
terns of the various blocks. The pattern of the KSZ is patterns can also be seen between adjacent geological
typical for old shield areas, i.e. a relatively re- blocks in the Archean Canadian shield and that there
flective upper crust underlain by a largely non- may not always be a simple relation between crustal
reflective lower crust with no indication of the Moho. reflectivity and crustal age. The data used here was
In contrast, the whole crust in the adjacent Val Rita obtained from the 1987-88 Lithoprobe seismic reflec-
Block and Abitibi region is highly reflective. Acom- tion survey conducted over the Kapuskasing structural
parison of the reflection and wide-angle refraction zone (KSZ) in the Superior province of the Canadian
data from 1984 Kapuskasing seismic refraction survey shield (Fig. 1) (see also Mitkereit et al., this
further confirms these crustal reflectivity differ- volume). The data set consists of 360 km of 60-fold
ences. Both data sets also suggest that the Moho is vibroseis profiles (Percival et al., 1989). These data
well defined under the Val Rita Block. Our results were acquired with the same parameters and processed
show that the contrasts may have resulted from the ef- with the same procedure.
fects of the thrust like tectonic processes which oc-
curred during the development of the KSZ. These pro- Methodology
cesses could have caused brittle faulting in the upper
crust and ductile flow in the lower crust. The results The statistical method used in this study is a mo-
indicate that after tectonic processes have created a dified version of that first introduced by Dohr (1957)
reflectivity pattern in the crust, subsequent tectonic and later adopted by Wever et al. (1987). First a co-
periods will not be able to destroy it. herency filter with a 100 ms time window and 32 ms in-
crement was applied to all stacked data to automarl-
Introduction rally extract coherent events, here treated as reflec-
tions (Fig. 2a). Then the number of reflections picked
Over the past 20 years there has been a dramatic within each time interval for each common-mid-point
increase in the number of deep seismic reflection sur- (CMP) trace was added up along each segment or the
veys conducted over the world (see for example, whole section according to regional geology, normaliz-
Barazangi and Brown, 1986; and Mereu et al., 1990). ed by the total number of CMP traces and plotted
These studies have shown that the nature of the crust against the travel time to form a histogram as shown
is exceedingly complex and that large differences in in Fig. 2b. This kind of plot presents the distribu-
reflection pattern exist when different regions are tion of reflections with depth. Theoretically it would
compared with each other. An examination of these pat- be better to have used migrated data, however these
terns and how they relate to tectonic processes such were not used here as it was felt that migrating arti-
extension and convergence has_been the subject of nu- facts could contaminate the results.
merous papers (Matthews and Cheadle, 1986; Brown,
Results
Geodynamics 22 The method described above was applied in the same
o 1991 American Geophysical Union manner to all the stacked data. The main results are
47
48 SEISMIC REFLECTIVITY PATTERNS OF THE KSZ
' ...--4-,'q I
,' x,,, ';a-. c,...,
.' nl:! ,':'-
: ,,
,- ' :::.:- 8 ' '" " 4a
:, ;:. ,,--- ,...
WAWA ,' BELT .:: ABITIBI :BELT "-.
....' ,,' -'... - ....... :I- -
" 4 ' '/
""'" .,...' "" ....': EAPUSEASIN BELT
:","' ..'"':' round iver Block I
-- ,.' Shawmere Complex
,'" ', ..... Ivanhoe Lake Fault Zone[
" .... 1984 Refraction Line 7
," 1987-88 Reflection Line
. K 0 M Superior I I I I
6 s TWT. The lower crust and upper mantle are for the
most part transparent with no indication of the Moho.
One of the most striking aspects of the data is
that across the ILFZ into the AB to the east, the
crustal reflectivity shows a very significant increase
as seen in Fig. 2. Similar observations are also shown
along lines 6 and 8. It is apparent that the ILFZ not
only acts as a geological boundary separating the KSZ
from the AB but also as a reflectivity boundary iso-
lating the poor reflection zone (the KSZ) from the
high reflection zone (the AB).
In contrast to the CB, the entire crust in the sur-
vey region of the VRB is highly reflective (Fig. 4).
Line 10 shows exactly the same pattern as line 9. Near
horizontal reflectors are predominant across the seis-
mic sections. Moreover they are strong and continuous.
The rapid disappearance of the reflectors seen at 14 s
TWT marks the reflection Moho (Brewer et al., 1983;
and Matthews and Cheadle, 1986), essentially a boun-
dary contrasting the reflective lower crust with the
seismically transparent upper mantle. This reflection
Moho is in good agreement with the refraction Moho
(see below).
Comparison of the Reflection and Refraction Data
A large-scale seismic refraction survey was carried
out across the Kapuskasing belt in 1984 {Fig. 1). The
refraction data has been interpreted by Wu (1987),
Fig. 1. The 1984 seismic refraction and 1987-88 seis- Boland and Ellis (1989), and Wu and Mereu (1990). Com-
mic reflection surveys. Note that refraction line CJ parison of the near vertical reflection and long range
partially overlaps with reflection profiles 1 and 4, refraction data displays some interesting features
so does refraction line EH with reflection profiles 2, common to both methods. First of all, both data sets
3 and 4. show the KSZ has a complex geological structure. Two
refraction sections shot from two ends of line CJ
along the axis of the KSZ are presented in Figs 5a and
5b. Except for first arrivals, no other major events
presented in Figs 2-4. The data can be divided into such as the PmP, wide-angle reflection from the Moho,
three groups according to the survey line locations: can be identified on the section. Pronounced coda en-
(1) lines 2, 3 ,4 and part of lines 1, 5 and 6 are in ergy is seen everywhere following the first arrival,
the Chapleau Block (CB); (2) south-east portions of implying that rocks in the KSZ are highly heterogene-
lines 1, 5, 6 and 8 are in the Abitibi belt (AB); and ous. Absence of the PmP reflections indicates that the
(3) lines 9 and 10 are in the Val Rita Block (VRB) of Moho under the KSZ is a wide transition zone. These
Wawa belt (Fig. 1). Seismic reflections in the CB show observations are consistent with those seen from the
a very complex pattern in which most reflectors are reflection data in the CB as described above. The Pn
seen in the upper crust (Fig. 2). The major thrust arrivals, refraction from the uppermost mantle, are
fault in the region, Ivanhoe Lake fault zone (ILFZ), clearly seen on shot C but not on reverse shot J. With
is clearly imaged and can be traced to within 1 km of limited Pn arrivals and with no PmP, the depth to the
the surface on the reprocessed seismic section from Moho cannot be constrained with much confidence. 2-D
line 5. In general, shallow reflections can be related seismic ray-tracing modeling (Fig. 5c) indicates that
to surface rock outcrops (Percival et al., 1989). As the Moho under the KSZ is a wide transition zone
the earth's crust is both laterally and vertically he- (about 10 km thick).
terogeneous, it is doubtful whether a histogram from Refraction line AE cuts through the VRB and the
one survey line would be representative of the entire Groundhog River block in a direction almost perpen-
survey region. Fig. 3 gives the overall reflection dicular to the KSZ (Fig. 1). A refraction section shot
distribution in the CB, which is constructed by using at A along the line is presented in Fig. 6a. Large am-
all the data within the block (more than 6000 CMP plitude PmP reflections are prominent on the section
traces). Statistically this is believed to represent between 120-180 km, implying that the Moho under sam-
the reflection signature of the CB, and hence the re- pling points (middle point between the shot and sta-
fleetion signature of the Kapuskasing belt as the CB tions) is a relatively sharp boundary. 2-D ray-tracing
is its most important part. The reflectivity changes modeling (Fig. 6b) shows that the Moho under line AE
resulting from the effects of source coupling and pro- varies in depth from 40-45 km (Wu, 1987). This appears
cessing would be averaged out. Overall in the CB most in good agreement with the reflection Moho beneath the
of the reflections (67.) are concentrated in the first VRB, defined as the base of the lower crustal reflect-
WU AND MEREU 49
N
2
12-
14-
16
LINE DRAWING OF LINE 1
(UNMIGRATED)
o KM
,
THE KSZ I= THE AB
THE ILFZ
REFLECTION DENSITY
0 4 8 0 4 8 12 16 20
0 , I , I , I , I ,
2
- _
'
16 THE KSZ THE ABITIBI BELT
Fig. 2. Line drawing of line 1 and segment reflectivity histograms.
2
4
6
8
-10
-12
1o-
c 1-
14-
REFLECTION DENSITY
0 4 8 12
0 , I , I , I
Fig. 3. The average reflection distribution for the
Chapleau Block.
LINE DRAWING OF LINE 9 REFLECTION DENSITY
(UNMIGRATED) 0 4 12 16 20 24
O N S O I , I , I , I , I , I
2 .-"'" '-- ' 2
14 Moho
18 18
-Val Rita t]1oek-
o 5 KM
Fig. 4. Line drawing of line 9 and reflectivity histogram;
a
0 30 60 90
b
194 KAPUSKASING EXPERIMENT LINE CJ SHOT C
2o 5o 8o 21o 24o 27o ,3o 33o
DISTANCE (KM)
194 KAPUSKASING EXPERIMENT LINE CJ SHOT J
0 30 60 20 150 10 21 240 270 300 0 330
DISTANCE (KM)
0 $0 / 100 150 200 250 300 30 600
0 , , , 1 , , L_, I .... I .... I .... I .... I .... L .... I i ,
2o 6.6
6.8
7.O
8.1
Fig. 5. a) Seismic refraction section shot at C along line CJ. b) Seismic
refraction section shot at J along line CJ. c) 2-D P velocity model along line CJ.
ing zone at about 14 s TWT (Fig. 4). This is equiva- (1983) show that the KSZ is distinguished from the ad-
lent to 42-45 km for an assumed average P velocity of jacent Wawa, Abitibi, Quetico and Opatia belts by an
6.0-6.5 km/s. The agreement is consistent with the abrupt increase in metamorphic grade and by broader
study by Mooney and Brocher (1987), who compare coin- positive gravity and aeromagnetic anomalies. The KSZ
cident reflection and refraction surveys around the transects the east-west trends of the adjacent ter-
world and find that in many cases the reflection and ranes, indicating that it formed at a later stage in
refraction Mohos are the same. The marked contrast be- the shield development. Using evidence from surface
tween refraction sections shot along lines CJ and AE geology, geobarometric data and gravity modeling,
is evident and further confirms the reflectivity dif- Percival and Card (1983) interpreted the KSZ as an up-
ferences between the KSZ and VRB. thrust cross section of middle to lower Archean crust.
Recent Analysis of high resolution seismic data resur-
Discussion and Conclusions veyed along line 2 by Gels et al. (1990) confirmed the
interpretation except that the KSZ is a thin thrust
The observed reflectivity patterns must owe their sheet traced to a depth of 12 km. All these indicate
origins to the tectonic processes which occurred in that the Wawa and Abitibi belts are parts of a former-
the historical development of the present day shield. ly continuous belt, now interrupted by the KSZ.
Geological studies such as those by Percival and Card The reflectivity of these belts could originate
WU AND MEREU 51
a 1964 KAPUSKASING EXPERIMENT LINE AE SHOT A
?o
o 30 60 90 ao 15o 8o ao 240 a?o 3oo
DISTANCE (KM)
b ^ c E
0 So m 1oo I$O 200 25ol 3oo :350
I1 .... f , .L, , I .... .... I .... I1.,, , I .... f .... 1 , , ,
...... ...............................
20 l'::;;:;;'_'_'_;:- ;: ........................ 65 ....
-,.- ........... 7'.-1'
_ 0 7.2
'" 8.2
6O
Fig. 6. a) Seismic refraction section shot along line AE. Note large
amplitude PmP arrivals between 120-180 km. b)2-D P velocity model
along line AE.
from either layering due to ancient igneous intrusion thanks is given to K. Vasudevan at the Lithoprobe
or metasedimentary or metavolcanic layering. It is seismic processing facility for his technical
speculated that since the Superior craton was estab- assistance with the DISCO processing package. The
lished, both the Wawa and Abitibi belts have not been Kapuskasing data was collected by Veritas Geophysical
involved in any major tectonic process. Therefore, and processed by Veritas Seismic. This study was
this layering including the Moho has been well pre- funded by research grants from NSERC (Grant A1793),
served. The lack of reflections in the lower crust LITHOPROBE, and EMR Canada and CSEG scholarships.
under the KSZ is interesting. It is possible that the
large scale crustal compression from which the KSZ was References
formed, caused uplifting and near surface faulting in
the upper crust. The upper crustal reflectivity could Barazangi, M., and Brown, L. (Eds.), 1986. Reflection
then easily result from sheared mylonite zones Seismology: A Global Perspective. AGU Geodyn. Set.,
(Smithson et al., 1986; and Green et al., 1990). The 15, 311p.
same compression force could have produced ductile Boland, A.V., and Ellis, R.M., 1989. Velocity Struc-
flow in the lower crust. This flow in turn might have ture of the Kapuskasing Uplift, North Ontario, From
resulted in small scale fold or disrupted structures Seismic Refraction Experiment Studies. J. Geophys.
(Percival et al., 1989), which had the effect of scat- Res., 94, 7189-7204.
tering seismic energy and erasing both the near ver- Brewer J.A., Matthews, D.H., Warner M.R., Hall, J.,
tical and wide angle PmP reflections. This process Smythe,D.K., and Whittington, R.J., 1983. BIRPS deep
destroyed the original geological layering and seismic reflection studies of British Caledonides.
increased the thickness of the Moho transition zone. Nature, 305, 206-210
In conclusion, we have demonstrated that the Brown, L.D., 1987. Lower continental crust: variations
crustal reflectivity and the nature of the Moho in the mapped by COCORP deep seismic profiling. Ann.
Canadian shield are highly variable from one tectonic Geophys., 5b, 325-330.
block to another. Our study may indicate that crustal Dohr, G., 1957. Zur reflexionsseismischen Erfrassung
age does not always have much control over crustal re- sehr tiefer Unstetigkeitsflichen. Erdi5 und kohe,
flection behavior. The pattern once established tends 10: 278-281.
to be well preserved and may well show a state that Geis, W.T., Cook, F.A., Green, A.G., Milkereit, B.,
the crust was left in after its last major tectonic Percival, J.A., and West, G.F., 1990. Thin thrust
event. sheet formation of the Kapuskasing structural zone
revealed by Lithoprobe seismic reflection data.
Acknowledgments. The authors would like to acknow- Geology, 18: 513-516.
ledge S.B. Smithson for his critical review. Special Green, A., Milkereit, B., Percival, J., Davidson, A.,
Parrish, R., Cook, F., Geis, W., Cannon, W., Percival, J.A., Green, A.G., Milkereit, B., Cook,
Hutchinson, D., West, G., and Clowes, R., 1990. F.A., Geis, W.T., and West, G.F., 1989. Lithoprobe
Origin of deep crustal reflections: results from reflection profiles cross exposed deep crust: Kapu-
seismic profiling across high-grade metamorphic skasing uplift. llature, $45: 416-420.
terranes in Canada. Tectonophysics, 172:627-638 Smithson, S.B., Johnson, R.A., and Hurich, C.A., 1986.
Matthews, D.H., and Cheadle, M.J., 1986. Deep Crustal reflections and crustal structure. In: Re-
reflections from the Caledonides and Vaiscides west flection Seisinology: The Continental Crust.
of Britain and comparison with the Himalayas. Geodyn. Set., 14, 21-32.
In: Reflection Seismology: A Global Perspective. AGU Smithson, S.B., Johnson, R.A., Hurich, C.A., Valasek.
Geodyn. Set., 12, 5-20. P.A., and Branch, C., 1987. Deep crustal structure
Meissner, R., Wever, Th., and Sadowiak, P., 1990. and genesis from contrasting reflection patterns: an
Reflectivity patterns in the Variscan mountain belts integrated approach. Geophysical Journal of R. astr.
and adjacent ares: an attempt for a pattern Soc., 89: 67-72.
recognition and correlation to tectonic units. Wever, Th., Trappe, H., and Meissner, R., 1987.
Tectonophysics, 172: 361-378. Possible relations between crustal reflectivity,
Mereu, R.F., Mueller, St., and Fountain, D. (Eds.), crustal age, heat flow, and the viscosity of the
1990. The Lower Crust: Properties and Processes. AGU continents. Ann. Geophys., 5B: 255-266.
Monograph, 51, 338p. Wu, J., 1987. The analysis of the data of the 1984
Mooney, W.D., and Brocher, T.M., 1987. Coincident Kapuskasing seismic experiment. M.Sc thesis, Univer-
Seismic Reflection/Refraction Studies of the sity of Western Ontario.
Continental Lithosphere: A Global Review. Reviews of Wu, J., and Mereu, R.F., 1990. The nature of the
Geophysics, 25: 723-742. Kapuskasing structural zone: results from the 1984
Percival, J.A., and Card, K.D., 1983. Arcbean crust as seismic refraction experiment. Proceedings of IVATO
revealed in the Kapuskasing uplift, Superior provin- Advanced Institute, M.H. Salisbury (Ed.), Kluwer Do-
ce, Canada. Geology, 11: 323-326. ordrecht, 563-586.
THE STRUCTURE OF THE ARCHEAN CRUST IN SW GREENLAND FROM SEISMIC WIDE-ANGLE DATA:
A PRELIMINARY ANALYSIS
Karsten Gohl 1 Scott B Smithson 1 and Yngve Kristoffersen 2
Abstract. An extensive seismic wide-angle and conducted along the coast and two fjords of SW
vertical-incidence experiment was conducted over Greenland (Fig. 1) to investigate the structure of
Archean crust in SW Greenland. 15 PASSCAL REFTEK the ancient Archean crust. Participants were the
recorders were placed on 35 locations along the University of Wyoming, the University of Bergen,
coast and fjords recording 6000 cu.in. air-gun IRIS, the Geodetic Institute of Denmark, and Victor
shots from offsets as far as 350 km. 100-150 m shot McGregor from Greenland. Some of the oldest rocks
spacing allows one to identify several unaliased in the world [McGregor, 1973] are exposed in the
pre- and post-critical coherent phases from Archean block in SW Greenland. One goal of the
throughout the crust and uppermost mantle. The most experiment was to trace Archean sutures from
prominent P-wave phases consist of Pg- and PmP- surface observations down into the crust and to
arrivals, mid- and lower crustal (PIP) reflections, obtain constraints on a possible assembly of
and Pn-arrivals. Excellent quality PiP and PmP- Archean terranes [Nutman et al., 1989]. A more
phases can be traced from near-vertical to post- general purpose was to invert and model seismic
critical distances. S-wave arrivals include Sg and normal-incidence and wide-angle data for velocity
SmS (or S1S)-phases, and a strong Sn-phase at structure of oldest Archean crust that possibly
offsets between 200 and 280 km. To obtain local underwent extension during rifting of the Labrador
one-dimensional estimates of the velocity Sea and Davis Strait in the Cretaceous and Tertiary
structure, an extremal inversion method was [Srivastava, 1978; Hinz et al., 1979]. This study
applied. All receiver gathers were slant-stacked focuses on a preliminary analysis and a one-
and semblance-filtered. Pre- and post-critical dimensional inversion of densely-spaced wide-angle
arrivals were picked in the tau-p domain and recordings.
inverted. Depth bounds show that velocities exceed
6.0 km/s at depths below 4 to 6 km, followed by a Geology and Previous Geophysical Studies
low gradient to the bottom of the crust where
velocities exceed 7.0 km/s. The Moho depth is The Archean block in southern Greenland consists
estimated to be 30 - 40 km with increasing depth of 3.8 BA old rocks affected by younger events and
northward. Depth bounds derived for two of the contains several Archean sutures [McGregor, 1973,
stations show a major discontinuity appearing 1979; Bridgwater et al., 1976; Brown et al., 1981;
between 6 and 8 km above the Moho. Mafic material, McGregor et al., 1986; Robertson, 1986]. Nutman et
possibly accreted during the opening of the al. [1989] describe four terranes with ages of 2750
Labrador Sea, might be responsible for the high to 3820 MA that assembled between 2750 and 2550 Ma.
velocities between both reflectors. All terranes underwent amphibolite to granulite
facies metamorphism and were intensively folded.
Introduction The Archean block in Greenland was separated
from its counterpart in Canada, the Nain Province,
In September 1989, an xtensive seismic by rifting between Greenland and North America.
vertical-incidence and wide-angle experiment was Rifting started in the southern Labrador Sea in the
1) Program for Crustal Studies, Department of
Geology and Geophysics University of Wyoming,
Laramie, WY 82071
2) Seismological Observatory, University of
Bergen, N-5007 Bergen
Geodynamics 22
Late Cretaceous, continued in the northern Labrador
Sea in the Early Paleocene, and ended in the Davis
Strait and Baffin Bay in the Early Oligocene
[Srivastava, 1978; Srivastava and Tapscott, 1986].
Previous geophysical investigations in SW
Greenland include gravity and magnetic studies in
the Godthaabfjord region [Thorning, 1986;
Geological Survey of Canada, 1988a,b; Woodside and
Verhoef, 1989; Forsberg, unpublished data; Speece,
unpublished data]. Gravity maps show a decrease in
the long-wavelength Bouguer anomaly field northward
of the western Godthaabfjord region, indicating a
53
54 GREENLAND CRUSTAL STRUCTURE FROM SEISMIC DATA
Data Acquisition
north dipping Moho. A steep gradient in the Bouguer
anomaly follows the continental slope along the
passive margin.
15 REFTEK recorders were placed on 35 locations
along the coast, fjords, and farther inland. An
array of 6000 cu. in. air-guns was fired along a
Wide-Angle Data
north-south striking offshore line (WA) covering
offsets up to 350 km, and along the Godthaab- (GOD)
and Ameralik- (AM) fjords to obtain data farther
inland (Fig. 1). For the first time, these
broadband instruments, designed and built for the
PASSCAL project as part of IRIS, were used for a
large-scale controlled source experiment. To avoid
the danger of losing data due to instrument failure
in their first deployment, two recorders were
placed on key locations. The shot interval from
line WA was 1 minute, for the eMF-lines 30 seconds.
A portion of the CMP-shooting was recorded by land-
stations as well. We chose a recording length of 30
seconds to record up to six days without servicing.
The positions of the land-stations were located
using the Global Positioning System (GPS).
Positions off-shore were obtained by an integrated
naVigation system.
Receiver gathers show excellent quality arrivals
at offsets to 280 km (Fig. 2 and 3). All receivers
were corrected for elevation statics with sea-level
as datum. Also, high and low frequency noise was
eliminated by a band-pass filter of 6 to 30 Hz.
Dominant frequencies for most phases lie between 8
and 10 Hz. All phases show strong reverberations
from the bubble effect and sea-bottom multiples.
Most of the reverberations were eliminated by
applying a predictive deconvolution operator
designed for the dominant arrival (Fig. 2).
Major vertical component phases recorded along
the coast show, even after statics correction, a
first break (Pg) with highly varying apparent
velocities between 5.4 and 7.1 km/s due to drastic
changes' in sea-bottom topography and lateral
velocity variations as well as dip effects within
the uppermost crust. One of the most prominent
features of the vertical component of stations IK
and KA is a strong, continuous lower crustal pre-
and post-critical reflection (PlP) crossed by an
equally strong Moho reflection (PmP) (Fig. 2). The
P1P- and PmP events can be correlated through
interfering S-wave arrivals to near-vertical
offsets with few interruptions in their continuity.
Mid-crustal phases were identified in all
recordings of profile WA. Their coherency is very
weak at pre-critical distances and increases with
offset. A refracted upper mantle phase (Pn) with an
apparent velocity of 8.1 to 8.3 km/s appears at far
offsets at station IK (Fig. 2).
Other prominent features include strong shear-
wave arrivals, originated from mode conversions at
the water-seafloor contact. A strong, multi-cyclic
direct S-wave (Sg) phase appears in all recordings.
Stations IK and KA show a band of coherent S-wave
energy reflected from the lower crust (SlS) or Moho
(SmS). Unfortunately, this phase can not be traced
farther due to a limited recording window. As a
result of wrap-around recordings of previous shots,
energy could be recorded at travel times
corresponding to 60 to 90 seconds at offsets as far
as 280 km. Within that window, all stations, except
- PROfiLES
CUP PROFILI!S
- .. ARCHEAN SUTURES
SEISMIC RECORDERS
A GODTHM8
100 km
I II " I
o
I I I I
\ .
D Proterozok: .MI....
"'ehun Granitic intrusians
ARCHEAN TERRANES
Tr. Brodt. 'e".an. (2750 2100 MA)
.. raaiuHnueq terrana C2100 3000 MAl
A.u. le,,&ft. (2140 3010 MAl
lliiliill FurinOehavn terrane 3100 MA)
Fig. 1. Location map of shots and receiver
stations of the seismic experiment in SW Greenland.
Solid lines: positions of source array for wide-
angle recordings consisting of 5 airguns, shot
interval 60 s; thick-dotted lines: positions of CMF
lines, source array consisted of 3 air-guns, shot
interval 30 s, 48-channel streamer with 50 m
spacing; land-station (REFTEK instruments): UK
Uman, KA Kangeq, IK Ikaarisat, GR Graedefjord, MA
Manitsoq (2 stations), GO Godthaabs-fjord (11
stations), AM Ameralikfjord (8 stations), BA Base;
star: Godthaab; thin-dotted lines: boundaries of
Archean terranes [from Nutman et al., 1989];
striped area in inlet map marks Archean Block.
GOHL ET AL. 55
170 100 o o
N offset (Rs) B
Fig. 2. Receiver gather from station IK (shot profile WA), vertical component, travel-
time reduced with 8 km/s, predictive deconvolution applied.
UMAN, show coherent energy within a multi-cyclic phases show strong amplitudes when they reach post-
band, identified as a shear-wave refraction from critical angles. The tau-p gathers for stations KA
the upper mantle (Sn) with an apparent velocity of and IK show clear images of pre- and post-critical
4.1 to 4.3 km/s (Fig. 3). After rotation has been reflections from the lower crust and Moho.
applied, it appears that the transverse component
contains a more coherent Sn-phase than the radial Extremal Inversion
component, an unexpected result.
Picks in the tau-p domain were made at the onset
Processing of each arrival. Uncertainties assigned to the
intercept-time picks were based on the reciprocal
As a preliminary step in our interpretation of bandwidths and S/N ratio of the arrivals. They did
densely spaced wide-angle data, we performed an not incorporate scatter due to local dip effects.
inversion of individual receiver gathers to These uncertainties ranged between 0.08 and 0.2 s.
estimate depth bounds assuming local 1-D velocity The model parameterization consisted of a stack
structures. Although the bounds are probably biased of layers of constant slownesses with slownesses
to some extent by local dip effects, the inversion decreasing (velocities increasing) with depth. The
results provide useful starting points for deriving
more realistic models. Each receiver gather was
split into overlapping windows of 200 traces and km
then transformed into the tau-p domain. A ray- 222 249
parameter increment of 0.002 s/km was used, -.z.:....' ] ...
starting from ray-parameter zero and ending at 0.35
s/km, so that the ray-parameter range for S-wave a .:.".' . .' ,
,'
arrivals were included. After multiplying by a ._.,.:, ...... .... ..
coherency filter derived from the smoothed
semblance (Stoffa et al. 1981) the individual
, , -7..- . y.. ' 'J' -.
panels were summed to form a composite tau-p
s e c t i on. 68 :.:...:: .,:, -, -,.:+:-.[..,
The envelope of the composite slant stack of
station KA (Fig. 5) shows a strong P-wave branch
and a weaker S-wave branch. Several arrivals can be
matched with the respective time-offset gather. The 72'=
Pg-phase consists of several high-amplitude
clusters between ray-parameters 0.185 and 0.125 Fig. 3. Rotated horizontal components (transverse
s/km indicating variations in the first break and radial) of far offset recordings from station
alignment of the offset gather. Several mid-crustal KA (shot profile WA), containing multi-cyclic Sn.
radial
km
222 249
'.,-' ,. :., .,` .- .....
,.! ...:.. .:.. ,
:', ,. -', , .:..:.
.,- :l.V.'.-. ,S" ... .f .
[:-: -.,. ,q .--H:'(,'% ..'
: ;-,. -:;:, . ,. ,., , .: .:'e-'-- -- ":' ,';,''',,.
...
.. :
56 GREENLAND CRUSTAL STRUCTURE FROM SEISMIC DATA
Fig. 4. Receiver gather from station IK (shot Fig. 5. Envelope plot of coherency-filtered slant
profile AM), vertical component, travel-time stack from the vertical component of station KA-
reduced with 8 km/s. Note the strong dip of PmP. North (shot profile WA), containing strong P-wave
and weaker S-wave branches. Threshold for the
semblance-filter was set to level where correlation
slowness for the uppermost layer (0.19 s/km) was of P-wave arrivals display best. Notice the high
chosen from the near-offset first break velocities. amplitude PiP and PmP arrivals.
A minimum slowness of 0.125 s/km was chosen to
allow for the possibility of high velocities near
the base of the crust. The model is parameterized
such that the number of layers is larger than the Therefore, the results of the inversion may be
number of data picks to more closely approximate biased by a slope of the Moho perpendicular to the
the real earth. Input data include refractions and profile.
post-critical reflections. Pre-critical reflections The observation of a distinct high-velocity zone
were also included wherever they could be traced at the bottom of the transitional crust suggests
back from the critical point; this helped to that mantle-derived material might have intruded
contrain depth bounds for specific target and accreted to the Archean crust in SW Greenland
slownesses [Hawman and Phinney, submitted paper]. during the beginning of the Labrador Sea rifting
The extremal method produces a 1-D inversion process. This interpretation implies that the
rifting of the region was more volcanic than
(Fig. 6) of data acquired over a large offset range
on a crust that is probably extremely previously thought (White and McKenzie, 1989).
heterogeneous. Localized depth bounds show a
northward increase in crustal thickness from 30 to
40 km, in agreement with gravity data [Speece,
unpublished data]. Several mid-crustal
O
discontinuities appear at varying depths between 5
and 25 km, indicating a complex structure with
lateral heterogeneities. A first order
discontinuity, 6 to 8 km above the Moho, can be 20
derived from the strong, coherent PIP-phase of
stations KA and IK. A high velocity gradient for
the uppermost crust is observed. Velocities exceed km40
6.0 km/s below depths of 4 to 6 km for most
recordings, but, thereafter, the gradient remains
6 8
T-r 1''
6 8
' I '"
6 8 6 8
[,-r -r l''
km/s
,, , r, I r'
L
GRI-1N KAI-1N IK3-1N UMI-1S UMI-1N
low for the larger part of the crust. The bottom of Fig. 6. Results from extremal inversion using tau-
the crust appears to consist of a high-velocity p picks. Profiles represent minimum and maximum
layer with a large gradient from 7.0 to 7.6 km/s. depth bounds for each slowness, here displayed as
The WA profile was located parallel to the layer velocity. Depth bounds are narrow for the
continental margin. A first study of data from the upper and lowermost crust, because more coherent
Ameralikfjord indicates a Moho with local dip phases exist. Note the increase in crustal
angles of 20 to 30 degrees eastward (Fig. 4). thickness toward northern recordings.
GOHL ET AL. 57
Conclusions Geological Survey of Canada, Gravity anomaly map of
the continental margin of Eastern Canada' Geol.
1. Densely spaced 3-component seismic wide-angle Surv. of Canada, Map 1708A, scale 1'500 000,1988a.
data were acquired from the Archean crust in SW Geological Survey of Canada, Magnetic anomaly map
Greenland covering offsets of up to 350 km. of the continental margin of Eastern Canada'
Prominent P-wave phases observed in the receiver Geol. Surv. of Canada, Map 1709A, scale
gathers include strong Pg, PmP, PiP (lower crustal 1'500 000, 1988b.
reflection), and several pre- and post-critical Hinz, K., H.-U. Schlueter, A. C. Grant, S. P.
mid-crustal arrivals as well as a Pn-phase. Strong Srivastava, D. Umpleby, and J. Woodside,
coherent S-wave arrivals include Sg, SmS (or SiS) Geophysical transects of the Labrador Sea'
phases, and a surprisingly coherent Sn phase at 200 Labrador to southwest Greenland, Tectonophysics,
to 280 km offset. 59, 151-183, 1979.
2. Extremal inversion of composite tau-p data of McGregor, V. R., The early Precambrian gneisses of
the coastline data assuming local 1-D structures the Godthaab district, West Greenland, Philos.
show an increase in crustal thickness from 30 to 40 Trans. R. Soc. London, Ser. A, 273, 343-358,
km northward that is consistent with results from 1973.
gravity surveys. The results are strongly affected McGregor, V. R., Archaean gray gneisses and the
by the complicated 3-D structure of an Archean origin of the continental crust Evidence from
crust in proximity of a passive continental margin. the Godthaab region, West Greenland, in
Depth bounds of mid-crustal discontinuities are Trondhjemites, Dacites and Related Rocks, edited
wide and vary among the station recordings, by F. Barker, pp. 169-205, Elsevier, Amsterdam,
indicating intense lateral heterogeneities in the 1979.
mid- and upper crust. The velocity profiles show a McGregor, V. R., A. P. Nutman, and C. R. L. Friend,
The Archean geology of the Godthaabfjord region,
high velocity gradient for the uppermost crust
where velocities exceed 6 km/s below depths of 4 to southern West Greenland, Tech. Rep.. 86-04, pp.
6 km and a rather low gradient to lower crustal 113-169, Lunar and Planet. Inst., Houston, 1986.
depths. Several profiles show a distinct high- Nutman, A. P., C. R. L. Friend, H. Baadsgaard, and
velocity layer (7.0 to 7.6 km/s) at the bottom of V.R. McGregor, Evolution and assembly of Archean
the crust. gneiss terranes in the Godthaabfjord region,
3. The observation of a 3 to 8 km thick high- southern West Greenland' structural, meta-
velocity zone at the bottom of the Archean morphic, and isotopic evidence, Tectonics, ,
transitional crust implies that the SW Greenland 573-589, 1989.
margin was affected by more active volcanic rifting Robertson, S., Evolution of the late Archaean lower
than previously thought. Mafic and ultra-mafic continental crust in southern West Greenland, in
material might have been underplated as part of the
rifting process.
Acknowledgments. The authors are grateful to
the crew of the R.V. "Hakon Mosby" and to the
members of the land acquisition crew, Chris
The Nature of the Lower Continental Crust, edited
by J. B. Dawson, D. A. Carswell, J. Hall, and K.
H. Wedepohl, Geol. Soc. Special Publ. No. 24,
pp. 251-260, 1986.
Srivastava, S. P., Evolution of the Labrador Sea
and its bearing on the early evolution of the
North Atlantic, Geophys. Jour. R. Astr. Soc., 52,
313-357, 1978.
Humphreys, Allan Tanner, Reid Fletcher, Rick Srivastava, S. P., and C. R. Tapscott, Plate
Blenkner, Mark Skelton, Peter Skjellerup, and kinematics of the North Atlantic, Chapter 23 in
Marvin Speece. We also want to thank Rick Williams The Western North Atlantic Region, edited by P.
and Victor McGregor for their support in the R. Vogt and B. E. Tucholke, The Geology of North
experiment. Peter Skjellerup participated in the Amerika, v. M, pp. 379-404, 1986.
processing of the data. Thanks to Rob Hawman for Stoffa, P. L., P. Buhl, J. B. Diebold, and F.
providing the inversion program and reviewing the Wenzel, Direct mapping of seismic data to the
manuscript. This project was supported by NSF grant domain of intercept time and ray parameter' A
DPP-8821974. plane wave decomposition, Geophysics, 46, 255-
267, 1981.
Thorning, L., A decade of geophysical surveying in
References Greenland, in Developments in Greenland Geology,
edited by F. Kalsbeek and W. S. Watt, Groenlands
Bridgwater, D., L. Keto, V. R. McGregor, and J.S. Geol. Undersol., Rapp. 128, pp. 12-133, 1986.
Myers, Archean gneiss complex of Greenland, in White, R. S., and D. P. McKenzie, Magmatism at rift
Geology of Greenland, edited by A. Escher and W. zones' The generation of volcanic continental
S. Watt, Groenlands Geol. Undersog., pp. 18-74, margins and flood basalts, J. Geophys. Res., 94,
Copenhagen, 1976. 7685-7730, 1989.
Brown, M., C. R. L. Friend, V. R. McGregor, and W. Woodside, J. M., and J. Verhoef, Geological and
T. Perkins, The late Archaean Qorqut granite tectonic framework of eastern Canada as
complex of southern West Greenland, J. Geophys. interpreted from potential field imagery, Geol.
Res., 86, 10617-10632, 1981. Survey of Canada, Paper 88-26 1989.
POST-OROGENIC EVOLUTION OF THE EUROPEAN CRUST
STUDIED FROM ECORS DEEP SEISMIC PROFILES
C. Bois* and ECORS Scientific Parties
Abstract. Recent deep seismic profiles acquired by the
ECORS project have been compared to former surveys. The
Gulf of Lions basement is a segment of the Pyr6n6es-
Provence belt beneath which the crustal root was removed in
the late Eocene. Metamorphism and/or magmatic intrusion
played a part in crustal removal as in the Variscan crust. The
evolution of the crust beneath the Tertiary rifts of the Upper
Rhinegraben, Bresse and the Gulf of Lions is characterized by
a small extension of the basement, great crustal thinning, high
P-wave velocities in the deep crust and asthenospheric
upwelling. Interpretation suggests that interactions between
asthenosphere, lithosphere and crust should have occurred in
relation to Alpine belt development. The Cretaceous and
Triassic-Jurassic basins also show a discrepancy between
basement extension and crustal thinning, probably also related
to interactions between mantle and crust. In contrast, the
layered lower crust, unthinned beneath the Triassic-Jurassic
basins, suggests that underplating occurred during or after the
crustal thinning which controlled basin subsidence. The
present review shows that the Moho and, to a certain extent,
the layered lower crust can be adjusted through time. Their
changes were governed not only by mechanical deformation
but also by transfer of material and modification of the
physical properties near the crust-mantle boundary.
Introduction
In recent years, the ECORS deep seismic project has
mainly investigated Tertiary structures, either compressional
or extensional, which occur in the southern and eastern parts
of France (Fig. 1). The Alps profile has been continued across
the Jura mountains and the Bresse rift over about 100 km,
resulting in a 365 km traverse from the Po Plain to the French
Massif Central. The Tertiary rift system that occurs in eastern
France has also been investigated by two profiles across the
Rhinegraben (80 and 120 km) in co-operation with DEKORP
and by a 600 km survey in the Gulf of Lions complemented
by a 200 km CROP profile on the Sardinia margin. The aim
of this paper is to review these profiles and to draw some
interpretational conclusions and speculations from their
*Institut Franqais du P6trole B.P. 311, 92506 Rueil-
Malmaison Cedex, France
Geodynamics 22
1991 American Geophysical Union
59
comparison with former investigations (Bois and ECORS,
1990). The main topics that have been addressed are the
disappearance of the mountain roots and the subsidence in
platform areas.
The disappearance of mountain roots
Orogenic belts of different ages show a prominent
difference in the geometry of the layered lower crust and the
Moho (Bois et al., 1989; Bois and ECORS, 1990). These
features are almost flat beneath the Variscan belt, whereas they
are folded and faulted beneath the Tertiary belts where a crust-
mantle imbrication can be observed (Fig. 2). Since all these
belts were formed by similar collisional processes, deep
crustal roots should have existed at the end of the Variscan
orogeny. They were subsequently removed when the crust-
mantle transition was re-equilibrated, a new Moho was formed
and layering was created in the lower part of the crust. In
several places, for instance in the English Channel, the
dipping reflections interpreted as Variscan thrusts associated
with the Lizard suture are intersected at an angle by the lower
crustal layering and the Moho (Bois et al., 1990a) (Fig. 2).
This relationship on a regional scale suggests that
metamorphism and/or magmatic intrusion played an important
role in the removal of the crustal root by obliterating a
previously deformed and thickened crust.
No similar process has been observed in the Alps and
Pyr6n6es belts. A wide-angle seismic reflection survey has
traced the Moho from 37 to 55 km beneath the Alps,
indicating the presence of a major crustal root (ECORS-
CROP, 1989). In the Pyr6n6es, despite the low-amplitude of
deep crustal reflections, the crustal root can unambiguously be
observed in the seismic reflection profile (Choukroune et al.,
1989) (Fig. 2). Although the orogenic deformation was
completed 30 Ma ago in the Pyr6n6es, there is no evidence of
incipient crustal root removal or the formation of any new
lower crustal layering.
The Gulf of Lions basement is a segment of the
1000 km Pyr6n6es-Provence fold belt formed in the late
Cretaceous-Eocene by the collision between the European
plate and the Iberian-Sardinian-Corsican block (Arthaud and
S6guret, 1981) (Fig. 1). Paleozoic rocks that have been
deformed, gently metamorphosed and intruded by granite are
found in the Gulf of Lions beneath the Tertiary cover (Cravatte
et al., 1974). These rocks, which have been thrust over
Permian and Mesozoic formations, indicate that the area
underwent a major shortening, probably having led to a crustal
thickening like in the Pyr6n6es. The boundary between the
Iberian-Sardinian-Corsican and European plates is the North
60 POST-OROGENIC EVOLUTION OF THE EUROPEAN CRUST
x\
CELTIC
\
ATLANTIC OCEAN
4
IRELAND: :'"
ENGLAND
BAY
OF
BISCAY
lice 9
!
!
SPAIN
NORTH
SEA
FRANCE
*ARDENNE
.,,,
100 200 300 Km .,--'"- '
I I I
WES TERN
BLACK
FOREST
PO PLAIN
CORSICA
Fig. 1. Main structural features of France and adjacent areas and the location of the deep
seismic profiles. 1 = Paleozoic basement; 2 = Pyrenean-Provenqal belt, 3 = Alpine belt;
4 = major thrusts and faults; 5 = orientation of a subduction; 6 = ECORS near-vertical
seismic reflection profile; 7 = ECORS refraction and wide-angle seismic reflection
profile; 8 = profile acquired by a foreign partner; 9 = other deep seismic profile; 10 =
international boundary; CF = Cevennes fault; NPF = North Pyrenean fault.
Pyrenean fault, which outcrops in the Pyr6n6es and should
extend eastwards and approximately follow the edge of the
continental shelf (Fig. 1). Farther south, the Pyr6n6es-
Provence belt was considerably affected by Oligocene rifting
and the Miocene opening of the western Mediterranean basin
(Biju-Duval et al., 1977).
The seismic profiles in the Gulf of Lions (Fig. 3) show
a number of south-dipping reflections in the crust, some of
which extend with high amplitude into the layered lower crust
and even the upper mantle. By comparison with the Pyr6n6es
profile (Fig. 2), they have been interpreted as major
Pyrenean-Provenqal crustal thrusts (Burrus et al., 1990).
However, the northward subduction of the southern plate with
mantle imbrication cannot be seen as in the Pyr6n6es profile.
On the contrary, the Moho, the location of which has been
confumed by Expanding Spread Profiles (ESPs) (Le Douaran
et al., 1984), forms a fairly continuous trend of reflections
with subhorizontal to gently northerly dips. It is lined by
subparallel shorter and weaker reflections which form the
layered lower crust (Fig. 3). The Gulf of Lions sections
show a strong similarity with the English Channel section
(Fig. 2), and the geometric relationship between the dipping
thrusts, the Moho and the lower crustal layering suggests that
metamorphism and/or magmatic intrusion also played a part in
the removal of the crustal root. This deep process was
certainly related to the leveling of the basement, which shows,
between the rifts, a smooth surface sealed by Oligocene and
early Miocene deposits (Fig. 3). The early Tertiary re-
BOIS AND ECORS 61
BUDE SSE N LIZARD THRUST EN G LIS H C H ANN EL S
BASIN ,ooo .... , DO 3400 3000 zooo ooo
, :.., ...... .__:._:.:.,_.,....v. ,, .,._ 0 O. , ---*-4-- , _ --.!, . ,-a.-.,. ..... _: 0
L- ,..7 7':v",_-*,:,.? s 1. L...."'">. ...- ..... . ': _ ' ........ _2 s
_,... . 7. ,,, , _. --_---:.-_::-;.--- _ - - 7-__-_ ..... -__--:-. .... ,,
S N
EB PFT TB NT AZ NPF AM PFT AB
: .............. =:.,:..,,.,: .... --/, ... - --- __......... .-.
.. ,, -_ _ ====================== -..- _ . -
' . ' - '_ - ...... ' I
25 PYReNeeS
Fig. 2. Above = Interpreted line drawing of portions of the SWAT 3 and 8 profiles
(unmigrated) after Bois et al., 1990a. See location in Fig. 1. 1 = layered lower crust;
2 = sedimentary basin; 3 = shallow layering; 4 = main seismic feature; C = Caledonian;
V = Variscan. Below = Interpreted line drawing of the ECORS Pyrenees profile. See
location in Fig. 1. 1 = layered lower crust; 2 = upper pan of the Paleozoic basement;
3 = late Cretaceous and Tertiary; 4 = main fault, thrust or detachment; 5 = borehole;
AB = Aquitaine basin; AM = Arize North Pyrenean massif; AZ = axial zone; EB =
Ebro basin; NPF = North Pyrenean fault; NT = Nogueras thrusts; PFT = Pyrenean
frontal thrust; TB = Tremp detached basin.
equilibration of the crust to a thickness of 30 km or less
appears to be a small-size analogue of the Permian-early
Mesozoic large-scale re-equilibration of the Variscan crust.
Deep seismic imaging in rifts and platform basins
Gulf of Lion
After the completion of the Pyr6n6es-Provence fold belt,
several rifts were formed in the shelf of the Gulf of Lions as
the result of a N 140 extension (Arthaud and S6guret, 1981).
These rifts are parallel to major late-Variscan faults such as the
Cevennes fault (Fig. 1). The Oligocene-Aquitanian synrift
series has a maximum thickness of 2 km and is overlain by
the Miocene formations related to the opening of the western
Mediterranean basin and the late Miocene to Recent wedge of
the Rh6ne delta (Burrus et al., 1991). The thickness of the
whole sedimentary pile amounts to 5-7 km, added to 2-3 km
of water. Actually, the lithosphere was stretched over a larger
area and the Tertiary basin extends farther north, up to the
Cevennes fault. The northern shoulder of the stretched crustal
segment belongs to the Massif Central that was mainly uplifted
in the late Miocene and Pliocene. On the other side, Sardinia
and Corsica were uplifted in the same period together with the
Alps and Apennines. Alkaline volcanism was active in the
Oligocene and Miocene in the Gulf of Lions and fanher north.
Calc-alkaline volcanism is known mainly south of the western
Mediterranean basin (Sardinia) but also in Provence and the
southern Alps, suggesting that the Apulian plate was
subducted northwestwards in this period beneath a Sardinian-
Corsican active margin (Biju-Duval et al., 1977).
In the seismic profiles (Fig. 3), the rifts are bounded by
steep faults and show only a gentle rotation of the layers.
There is no major crustal detachment and the rate of basement
extension is estimated to be 1.2 (Bessis and Bums, 1986).
Beneath the Gulf of Lions margin, the Moho is uplifted and
the crust thinned by a factor greater than 5 on the edge of the
continental shelf. The layered lower crust is about 2 s TWT
thick in the north and less than 0.5 s TWT in the south. The
corresponding stretching rate of a 25 km crust would amount
to 1.8-2.0. ESPs carded out along Profile 1 (Fig. 3) have
found anomalous P-wave velocities (7.0-7.7 km/s) at the
bottom of the crust. The asthenosphere was uplifted beneath
the western Mediterranean basin and may have reached the sea
bottom between the Oligocene and the early Miocene (23-
19 Ma) (Biju-Duval et al., 1977).
A SIROCCO ESP,201 ESP 202 ESP 203
o _. - , ........ ]o
.
8
. - - ..m-.;. ............ .. . - ....
' 0 IOkm --
- GULF OF LIONS 1
14-
1 ESP 201 GLP - ESP 20;5
B ",, ,o,o , , , ? _ ,,o , / ,,oo i, 20?0
t ............... '"
1 ..... - - - ... .,; :::::.....::...., ,..,::::..,..; .._.....,...,__:., ..:: -_- ...... . ..... _ ... . .... ,-- ___
. .................. :: ........................ ..,, ... _ .......... _--- -...._., -.:.,.. .. ,:-: _ ...........
/ r .............
1- -'-- ...... --"'--'''-'-- - - '---'- -'.Z:"-"-;-:'i:Z"__....-.-:'- '"
I 0 t ........ ::::::: ........ ' .......... :f:'f-':. " ............. ' ............. . ......
I t .... - _
4 ? - o GULl: Ol: LIONS 2
$
-4900
..
Fig. 3. Interpreted line drawing of the ECORS Gulf of Lions profile (unmigrated). See
location in Fig. 1. Dotted line = diffraction; 1 = late Miocene-Recent; 2 = Miocene
(B = Burdigalian); 3 = Oligocene-Aquitanian; 4 = layered lower crust; 5 = main
dipping reflections interpreted as Pyrenean-Provenqal thrusts; 6 = borehole.
The Bresse rift
This north-trending rift is an elongated trough
(180 x 40 km) resulting from an E-W extension (Fig. 1).
The rift infill of about 1000 m along the seismic profile
consists essentially of Oligocene-Aquitanian rocks with a thin
late Miocene-Pliocene veneer (Fig. 4). This late subsidence
probably resulted from the overthrusting of the eastern
shoulder of the rift in the late Miocene by the Jura nappe
(Bergerat et al., 1990). The Maconnais hills in the west may
be interpreted as the flexural bulge related to the emplacement
of the Jura nappe. The lithosphere was actually stretched over
a wider area, and other asymmetric rifts with up to 2500 m of
Oligocene sediments can be observed in Limagne beyond the
end of the seismic profile (Figs. 1 and 4). The regional
basement was affected by a large wavelength uplift with a
maximum elevation of 1500 m located south of the Limagne
rift (Lucazeau and Bayer, 1982). This uplift occurred after the
late Miocene and mainly since the late Pliocene in relation to
the Plio-Pleistocene volcanic climax.
Beneath the Bresse rift infill, the pre-rift series shows a
complicated geometry because of a mechanical recoupling in
the Triassic salt and the gradual slippage of the overlying
Jurassic and Cretaceous formations during the development of
a horst-and-graben structure in the underlying Paleozoic rocks
(Bergerat et al., 1990). Basement extension in the Bresse area
appears to be extremely small. Though detailed data are not
available in Limagne, the extension in the whole region is
estimated to be no greater than 5-10 km.
The region is underlain by a broad asymmetric Moho
uplift (Perrier and Ruegg, 1973) (Fig. 4). The seismic
reflection section shows only the eastern limb of this feature,
which passes in the east into the flexure of the crust below the
Alpine front. The layered lower crust shows a constant
thickness of 3.5 s TWT, and the mode of crustal stretching
cannot be unambiguously determined from the data at hand.
But the crustal thinning observed would require that a 30 km
crust was stretched by 25-30 km. Seismic refraction has
found a P-wave velocity of 7.3 krn/s at the bottom of the crust
in Limagne (Perrier and Ruegg, 1973). Geophysical and
petrographic evidence indicates the presence in this area of an
asthenospheric diapir whose uplift in the last 5 Ma is related
to the evolution of volcanism and lava composition (Coisy and
Nicolas, 1978).
This north-trending rift (300 x 40 km) was initiated by
an E-W extension in the Oligocene-Aquitanian (Fig. 1). Its
infilling is moderate in the south (1500 m) where subsidence
stopped in the Miocene. The subsidence that resumed in the
northern part of the graben led to a thicker Tertiary infill
(3000 m) (Villemin et al., 1986). The graben's shoulders
were uplifted in the south (Vosges and Black Forest, Fig. 1)
to a maximum elevation of 1500 m. This uplift started as early
as the Oligocene, but it was largely completed after the late
Miocene together with a regional erosion encroaching upon the
graben infill itself. Alkaline volcanism was active from the late
Cretaceous till the Miocene, but very few extrusions occurred
in the Oligocene.
The southern seismic profile (Fig. 5) shows an
asymmetric rift bounded in the west by an east-dipping fault
BOIS AND ECORS 63
NW
MASSIF BRESSE
CENTRAL
..... ' "i" '"'"' ";"-'' ' ' ' ...... ' ....... ' '' ' ' ..... ' ..... '::::' '"' - ' ":';' '" ....
...... '" .... '"'"':':':':::':'"'" .... Lt ............ ' ::.">': ':.-..- '., '.:<... :';-<-;.:':-:;
i .... ................ -" .... :--'"' ""
'f[l111
- ]!1111 ,.11h,111. ,, ..,.,,.,, ,,,
,,mm '".. III "!!111,iiii!ill, [., IIIIII!! IIII .,, I ; ,,, '"I, '
4
.%.''"'"";':':.::' _ _
...:- ...-;-..V ........ .. -. -e.%-4.;.;
;- --+:-:-' '-'-'--.-.-.-.-.-..-.-:.:.:-.-.-.-:-;<<-' __ '-x-:<-x-,-.
.:,:::3<.,-I:::::;-:_ ....... t_ _ ;:::::::::.:.:.. ===================== ...... :.:.:::::::;;'.. -- ........ :::::::(:
:<<<< -<_ :::::::::::"::''::::::' . - .-:::
12 0 IOkm
I I
NW MASSIF CENTRAL SE
LIMAGNE MAC6NNAIS BRESSE JURA MOLASSE
I :'.= ..... ,'""": I' I" '" ........ '::::" ....... :':"';;'"' "i"
. I I I IIII '"' ........ ;'
OR
o
0 2.5 50 km
t I i
Fig. 4. Above = Interpreted line drawing of the ECORS Jura-Bresse profile. See
location in Fig. 1; Below = Regional section across Jura, Bresse and Limagne based on
the deep seismic profile and seismic refraction and wide-angle reflection data (Perrier and
Ruegg, 1973). 1 =Tertiary; 2 = Mesozoic; 3 = upper part of the Paleozoic basement;
4 = main fault or detachment; 5 = shallow crustal layering; 6 = layered lower crust.
and formed by blocks generally tilted westwards (Brunet al.,
1991). In the northern profile, the asymmetry is reversed,
with the main border fault and the deepest part of the rift being
clearly in the east whereas the blocks are little or not tilted
(Wenzel et al., 1991). The basement extension amounts to 4-
5 km in both sections without any relation to the change in the
infill thickness (Villemin et al., 1986). Moho steps seem to be
present beneath the eastern side of the rift in the southern
profile and beneath its western side in the northern profile
(Fig. 5). Associated with east- and west-dipping reflections
respectively, they suggest that simple shear may have
controlled the crustal stretching, perhaps down to the Moho.
RHINE
2000 W DONNERSBERG ODENWALD !:
3340 3000 2500 ', I 15100 I020
0 I I : ,iilli[illi[[iilll v v ,: ' ........... . I __........... i . ' i." . .-i: 0
_ -
g ..... ,.-....---..-..-.; ........ .... ? ..... .....:.. . .
I ............... ........ ..... ......
........................ ..::. . ................ ; ..... _ ;;:.;;?: ...... .... _ ...
..... <.::<.:< ::..::" ...--'-:.:._::::-:- --_
.................................... .... ..... ............. :::::::..:::::::,,,::.:::.:::: .................... ........ :.:. ...................... . ....
.:-:::'-'' .... .-. .- j::..-;.::-:-' - ..
I 0 2km
RHIgRABEN NORTH
W E
LOrrAINE VOSGE8 RHINE BLAOK FOREST
2500 000 500 000 500 300
I i I
2- / /
::::::::::::::::::::::::::::::: ...... ......... ..... ..:........:.....::. -.::::::::::::::::-.-.-::-..:.:.:........:.:.:.:.::
::..:----.-...: ..:::.:::.:. x ...... ...:.:.: -...: ............ ..:
.... ----.-. .:... ===============================================-......:...:.- ':::.::::.:.....:... ....................... :.- ......
.... ::::::: ..... ....._...._....::.:::.:. ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
_ .-.:::::::.--. ................ ... .......... ......:. ............. ,..-..,:.:,:::: ................. .: .......................................... ...<::::::::::::::::::::.:::::.:::.:::::::::;::::..J............":::.::::::::::.:.:.. .... - ...... :::?:?:?:?:::?:?:?"-'::.::-- 8
_:::---- ................ ..... - . ..............................
/ ' ':?:'"":???:-:'.;:'"''"'"' ............
_ - '"'"'""--------'-'-' -
RHINEGRABEN SOUTH
-10
10-
Ivvvvvl
Fig. 5. Interpreted line drawing of the ECORS-DEKORP Upper Rhinegraben profiles.
See location in Fig. 1. Dotted line = diffraction; 1 = Tertiary; 2 = Mesozoic; 3 =
Carboniferous and Permian (P = Permian, S = Stephanian, W = Westphalian)' 4 =
layered lower crust; 5 = Permian volcanics; 6 = main fault; 7 = borehole.
The layered lower crust is clearly visible beneath both sides of
the rift, but it can be observed beneath the graben only in the
southern profile where it is rather thinned. The rift is underlain
by a broad Moho bulge that is 150 km wide (Fig. 5). This
feature, discovered by seismic refraction has its apex near the
southern profile (Edel et al., 1975). It roughly mimics the
shoulder surface and fades northwards, though subsidence
has been greater in that direction. A 30 km crust could have
been stretched by 13 km to complete the observed crustal
thinning. Beneath the rift, seismic refraction has indicated
anomalously high P-wave velocities (7.2-7.8 km/s) at the
bottom of the crust (Edel et al., 1975). Gravity and heat flow
modelling indicate the presence of an asthenospheric
upwelling beneath the rift (Werner and Kahle, 1980).
Discussion
The evolution of the crust beneath and around the
Tertiary rifts reviewed above is mainly characterized by (1) a
moderate extension of the basement, (2) broad Moho uplifts
BOIS AND ECORS 65
NNW
KrnO
20
30
ssw
SWAT 5
S LH PARENTIS N
0 -'-
,. 5
1o
vr - 'Moho
15
5 -I
15
Fig. 6. Above = Section across the North Celtic Sea basin along the SWAT 5 profile
after Bois et al. (1990b). See location in Fig. 1. 1 = Tertiary; 2 = late Cretaceous and
Albian; 3 = early Cretaceous; 4 = late Jurassic; 5 = middle and early Jurassic; 6 =
Permian (?)-Triassic; 7 = layered lower crust. Below = Interpreted line drawing of a
portion of the ECORS Bay of Biscay profile (after Pinet et al., 1987a). 1 = Albian, Late
Cretaceous and Tertiary; 2 = Early Cretaceous; 3 = Triassic-Jurassic; 4 = layered
lower crust; 5 = main fault or bottom of the Mesozoic cover; LH = Landes high.
corresponding to great crustal thinning (3) a comparatively late
uplift of the shoulders, (4) a thinned lower crust with
anomalously high P-wave velocities and (5) asthenospheric
upwellings. The relationship between stretching and thinning
of the lithosphere has been extensively debated, leading to the
concepts of active rifting where the driving force is an
asthenospheric upwelling and passive rifting where this is
lithospheric stretching. The pre-rift doming of the surface
associated with active rifting was created by a deep heat
source, so it may have been of small amplitude, spread over a
broad area and overlooked (Moretti and Froidevaux, 1986).
The presence of such surface doming cannot be
unambiguously demonstrated in the Rhinegraben and Limagne
areas. However, the pre-Oligocene leveling of the basement in
the Gulf of Lions suggests that rifting was actively initiated
there.
In the first stage of passive rifting, stretching and
thinning of the crust should match, which is contrary to what
is observed in the seismic sections. The European Tertiary
rifts were initiated 35 Ma ago and, assuming they were
passively initiated, they may have become more and more
similar to active rifts during their later evolution because of
asthenospheric upwelling: The crust-mantle interaction
suggested by the high seisrmc velocities in the lower crust and
volcanic activity may have led to the great crustal thinning
observed in the seismic profiles. However, the round shape
and large size of the Moho uplifts related to the Rhinegraben,
Limagne and Bresse rifts and their geometric relation to the
overlying rifts suggest that they were completed by
asthenospheric upwellings of greater amplitude than the initial
ones, having probably occurred later than the rifting.
The European Tertiary rift system developed during the
Alpine orogeny to which it was closely related. The early
Alpine N-trending stress may have determined the orientation
of many rifts through basement rupturing or the reactivation of
pre-existing faults. The impingement of the Apulian Spur
against the European plate, which created major folds and
nappes in the Alpine belt, may have produced the Oligocene
extension in the foreland (Tapponnier, 1977). The Moho
bulges observed at present beneath the Rhinegraben and
Limagne areas may have been enhanced by the Mio-Pliocene
Alpine events that triggered a redistribution of the hot
asthenospheric material associated with a volcanic climax and
a surface uplift.
There are, however, differences between the Gulf of
Lions and the other Tertiary rifts. In the Gulf of Lions, the
crust was late in being equilibrated. There was a plate
boundary in the east (Sardinia), and the Apulian plate was
subducted beneath the whole area in the Oligocene and
Miocene. These particular geodynamic features may explain
that, in the Gulf of Lions, rifting resulted in oceanic opening.
The Tertiary rifts can be compared with the Mesozoic
basins of the European platform. The SWAT 5 profile shot
across the North Celtic Sea basin (Fig. 1) shows an example
of a Triassic-Jurassic basin that is 80 km wide and 10 km
deep and displays a roughly symmetrical saucer shape in
cross-section (Fig. 6). It was essentially infilled by Triassic-
Jurassic deposits whereas the Cretaceous series is not very
thick. Despite a few subordinate horsts and anticlines, the
basin's bottom remains comparatively smooth. The basement
extension ratio amounts to 1.2 whereas the crustal thinning
corresponds to an average stretching of 1.6 (Dyment, 1990).
In contrast to the Tertiary rifts, the layered lower crust does
not show any change in thickness beneath the basin, with the
whole crustal attenuation being concentrated in the upper crust
(BIRPS and ECORS, 1986; Dyment, 1990; Bois et al.,
1989). The other Triassic-Jurassic basins crossed by deep
seismic profiles show similar characteristics, suggesting that
crustal stretching did not play any significant part in the
initiation of subsidence (Bois et al., 1990b)
The early Cretaceous subsidence was associated with
the rifting of the North Atlantic margin. The Parentis basin
(Fig. 1) is about 100 km wide and was superimposed on a
wider Triassic-Jurassic basin (Pinet et al., 1987ab) (Fig. 6).
Fault-controlled rifting occurred in the early Cretaceous,
associated with fast subsidence. The Cretaceous series
amounts to 7 km whereas the whole sedimentary pile may be
thicker than 15 km. Underneath, the Moho has been uplifted
by 12 km, the crust has been thinned within a large area to a
minimum of 5 km and the layered lower crust has almost
disappeared. The Moho uplift is supported by ESPs, which
have also found high-velocity rocks in the remaining crust,
suggesting the presence of mantle intrusions (Marillier et al.,
1988). Despite uncertainty as to the real configuration of the
basin at the end of the rifting stage due to inversions and
possible strike-slip, the basement extension should not have
been greater than 1.2 whereas the crustal thinning corresponds
to an average stretching value of 1.6 with a maximum of 6.
Creep of the deep crust has been considered as a possible
cause for such great thinning, because the basin is near to the
North Atlantic continental margin (Pinet et al., 1987ab).
However, the deep crust has never been found to be denuded
farther west, and it is more probable that mantle uplift was
related to asthenosphere diapirism and the formation of a new
and shallower Moho.
One conclusion of the deep seismic surveys carried out
in France and adjacent areas is that a discrepancy between
stretching and thinning of the crust has been observed in all
platform basins and rifts, whatever the age of subsidence may
be. Several arguments for an underestimation of the basement
extension have been proposed (e.g. Coward and Trudgill,
1989). Inversions may have changed the original
displacements along the normal faults, and erosion may have
decreased their heave. However, inversions were minor in
most of the profiles studied. The seismic sections may not
have detected very small faults, which could account for larger
stretching than what was measured. But, assuming that
unrecognized faults have dips similar to the visible ones, they
cannot imply more than 0.1-0.2 of additional extension. Part
of the extension might have been oblique or perpendicular to
the available sections. But the main extension should have
occurred perpendicular to the basin's axis, hence parallel to the
seismic sections. Vertical and/or lateral heterogeneity in
lithospheric stretching might also account for the small
apparent basement extension. Although such a mechanism can
be envisaged in some cases, it cannot explain all the features
observed, especially the great Moho uplifts beneath the
Tertiary rifts. We believe that lithospheric stretching alone
cannot entirely account for the crustal thinning observed.
Interaction between asthenosphere, lithosphere and crust has
to be envisaged (Logatchev, 1987; Olsen et al., 1987). The
Moho rise may result from metamorphism and/or mantle
intrusion, as suggested by the high seismic velocities
frequently found at the bottom of the crust.
However, prominent morphological differences occur
between the wide Triassic-Jurassic basins and the narrow
Tertiary rifts. These differences are reflected in the shape of
the Moho and the thickness of the layered lower crust. They
have been related to the geodynamic context of the platform in
which subsidence occurred (Morgan and Ramberg, 1987).
The Triassic-Jurassic basins were formed in a period of major
continental splitting, whereas the Tertiary rifts developed in an
area of continental collision. However, the Cretaceous basins
are generally smaller than the Triassic-Jurassic ones, though
they were also associated with continental splitting and are
even nearer to the continental margin. The age of the basins
might also be a critical factor. Beneath the Triassic-Jurassic
basins, the gentle basement sagging indicates that thermal
subsidence overbalanced fault-controlled rifting. Lithosphere
cooling started more than 200 Ma ago in these basins and
should have been completed since the late Jurassic. On the
other hand, the unthinned layered lower crust beneath these
basins implies that crustal underplating took place during or
after basin subsidence. The Tertiary rifts show a younger
stage of thermal evolution. The lack of broad basins above
most of these rifts except in the Gulf of Lions may result from
high heat flow and the presence of hot and light material at
depth. The Cretaceous basins may represent an intermediary
stage. They show some features similar to the Tertiary rifts,
such as Moho uplifts and a thinned layered lower crust, but
they are overlain by real basins like the Triassic-Jurassic
basins yet of smaller size.
Conclusion
Deep seismic profiles have provided critical data on the
complex evolution of the European crust since the end of the
Vailman orogeny:
BOlS AND ECORS 67
(1) In the Permian-early Mesozoic the roots of the
Variscan belt were removed, the crust was re-equilibrated to a
thickness of about 30 km, a new Moho was formed and
layering was created in the lower crust.
(2) The initiation and subsequent evolution of the
Triassic-Jurassic basins seem to have resulted from crust-
mantle interaction controlling first crustal thinning then
underplating.
(3) Beneath the Cretaceous basins, crustal thinning is
greater than basement extension and also suggests crust-
mantle interaction.
(4) In the Eocene the early Alpine compression resulted
in the formation of the Pyr6n6es-Provence belt, the early
deformation in the Alps and the formation of N-S fractures in
the foreland, which guided the development of later rifts.
Whereas the crustal root has been preserved in the Pyr6n6es,
the crust was re-equilibrated in the early Tertiary in the Gulf
of Lions.
(5) The European Tertiary rifts resulted from the
impingement of the Apulian Spur against the European plate,
while the underlying Moho uplifts suggest the influence of
later asthenosphere upwellings related to the Mio-Pliocene
Alpine events. The subduction of the Apulian plate may have
played a part in the opening of the Gulf of Lions. In all these
rifts, great crustal thinning appears together with moderate
basement extension.
The present review shows that the Moho and, to a
certain extent, the layered lower crust can be adjusted through
time. Deep seismic data suggest that magmatic inmasion and/or
metamorphism played a major part in the re-equilibrafion of a
crust deformed and thickened by an orogeny. Beneath the
basins, the Moho's uplifts should result not only from crustal
stretching, which is generally small, but also from intrusion
and/or metamorphism in the deep crust. On the other hand,
crustal underplating at a later date may have brought the Moho
down beneath the Triassic-Jurassic basins. Therefore, the age
of a basin might be a critical factor in its present morphology
together with the geodynamic context in which it has
developed.
Acknowledements. The paper has benefited greatly
from critical reviews by J. Burrus, B. Colletta and I. Moretti
and three anonymous reviews. The ECORS project is being
carried out by the Insfitut Franqais du P{trole, Insfitut National
des Sciences de l'Univers (CNRS), Elf Aquitaine and the
Institut Franqais de Recherches pour l'Exploitation de la Met.
The Spanish ECORS, ENIEPSA petroleum company (now
REPSOL), Italian CROP and German DEKORP projects
participated in some of the profiles discussed above.
References
Arthaud F. and S6guret M., Les structures pyr6n6ennes du
Languedoc et du Golfe du Lion (Sud de la France). Bull,
. G6ol. France, 7.23, 1, 51-63 (in French), 1981.
Bergerat F., Mugnier J.L., Guellec S., Truffert C., Cazes M.
and Damotte B., Extensional tectonics and subsidence of
the Bresse basin: an interpretation from ECORS data. In:
Deep Structure of the Alt>s, edited by Roure F. et al.,
Mem. Soc. Geol. France, 56, Mem, $oc, Geol. Suisse,
1, Stec. Vol. Soc. Geol. Italiana, 1, 145-156, 1990.
Bessis i:. and Burrus J., Etude de la subsidence de la marge
du Golfe du Lion. Bull. Centre Recherches SNPA. Pau,
10, 123-141 (in French), 1986.
Biju-Duval B., Dercourt J. and Le Pichon X., From the
Tethys ocean to the Mediterranean sea: a plate tectonic
model of the evolution of the western Alpine system. In:
Structural history_ of the Mediterranean basins, edited by
Biju-Duval B. and Montadert L., Proceed. 25th CIESM
meeting, Split 1976, Editions Technip, Paris, 143-164,
1977.
BIRPS and ECORS, Deep seismic reflection profiling
between England, France and Ireland, J. Geol. Soc.
London. 143, 45-52, 1986.
Bois C., Pinet B. and Route F., Dating lower crustal features
in France and adjacent areas from deep seismic profiles. In:
Properties and Processes of Earth's Lower Crust, edited by
Mereu R. et al., AGU-IUGG Geophys. Monogr. 51, 6,
17-31, 1989.
Bois C. and ECORS Scientific Party, Major geodynamic
processes studied from the ECORS deep seismic profiles in
France and adjacent areas. Tectonophysics, 173, 397-410,
1990.
Bois C., Lefort J.P., Le Gall B., Sibuet J.C., Gariel O., Pinet
B. and Cazes M., Superimposed Variscan, Caledonian and
Proterozoic features inferred from deep seismic profiles
recorded between southern Ireland, southwestern Britain
and western France. Tectonophysics, 177, 15-37, 1990a.
Bois C., Gariel O. and Pinet B., Ecors Deep Seismic Surveys
across Paleozoic and Mesozoic Basins in France and
Adjacent Areas: a Review. In: The Potential of Deep
.Seismic Profiling for Hydrocarbon Exploration, edited by
Pinet B. ad Bois C., Editions Technip, Paris, 381-401,
1990b.
Brun J.P., Wenzel F. and the ECORS-DEKORP team, 1990,
Crustal Scale Structure of the Southern Rhinegraben from
ECORS-DEKORP seismic reflection data. Geology, (in
press), 1991.
Burrus J., Voogd B. de, Olivet J.L., Ravat J., and ECORS
team, ECORS deep seismic reflection study in the Gulf of
Lions (NW Mediterranean): first results. Tect0ni (in
press), 1991.
Choukroune P. and ECORS team, The ECORS Pyrenees deep
seismic profile: reflection data and the overall smacture of
an orogenic belt. Tectonics, 8,1, 23-39, 1989.
Coisy P. and Nicolas A., Regional smacture and geodynamics
of the upper mantle beneath the Massif Central. Nature,
274, 429-432, 1978.
Coward M.P. and Trudgill B., Basin development and basin
structure in the Celtic Sea basins (SW Britain). Bull. SO..
Geol. France, 8, 5, 3, 423-436, 1989.
Cravatte J., Dufaure P., Prim M. and Rouaix S., Les
sondages du Golfe du Lion; Stratigraphic et
s&limentologie. Notes et Mmoires, Compagnie Franqaise
des P6trole, Paris, 2, 209-274 (in French), 1974.
Dyment J., Some complementary approaches to improve the
deep seismic reflections studies in sedimentary basin
environment; the Celtic Sea basin. In: The Potential of
Deep Seismic Profilin for Hydrocarbon Exploration, _ _
edit.d by Pinet B. and Bois C., Editions Technip, Pads,
403-423, 1990.
Edel J.B., Fuchs K., Gelbke C. and Prodehl C., Deep
structure of the southern Rhinegraben area from seismic
refraction investigations. J, Geophys, 41, 333-356, 1975.
ECORS-CROP Deep Seismic Sounding Group, A new picture
of the Moho under the Western Alps. Nature, 337, 249-
251, 1989.
Le Douaran S., Burrus J. and Avedik F., Deep structure of
the north-western Mediterranean basin: results of a two-
ship seismic survey. Marine Geol., 55,325-345 (1984).
Logatchev N.A. and Zorin Yu.A., Evidence and causes of the
68 POST-OROGENIC EVOLUON OF THE EUROPEAN CRUST
two-stage development of the Baikal rift. Tectonophysics,
143, 225-234, 1987.
Lucazeau F. and Bayer R., Evolution geothermique et
g6odynamique du Massif Central franqais depuis
l'Oligocne. Ann. Geophys., 38, 3, 405-429, 1982.
Marillier F., Tomassino A., Patriat P. and Pinet B., Deep
structure of the Aquitaine shelf: constraints from expanding
spread profiles in the ECORS Bay of Biscay transect.
Marine Petr. Geol., 5, l, 65-74, 1988.
Moretti I. and Froidevaux, Physical models of extensional
tectonics. In: Thermal modelline in sedimentary basins,
edited by Burrus J., Editions echnip, Paris,-107-123,
1986.
Morgan P. and Ramberg I.B., Physical changes in the
lithosphere associated with thermal relaxation after rifting.
Tectonophysics, 14), 1-11, 1987.
Nicolas A., Him A., Nicolich R. and Polino R., Lithosphere
wedging in the Western Alps inferred from the ECORS-
CROP traverse. Geology, 18, 7, 587-590, 1990.
Olsen K.H., Baldridge W.S. and Calender J.F., Rio Grande
Rift: and overview. Tectonophysics, 143, 119-139, 1987.
Perrier G. and Ruegg, Structure profonde du Massif Central
franais. Ann, O0phys.. 29, 435-502 (in French), 1973.
Pinet B., Montadert L. and ECORS Scientific Party, Deep
seismic reflection and refraction profiling along the
Aquitaine shelf (Bay of Biscay). Geophys. J.R. Astr.
Soc., 8__9, 305-312, 1987a.
Pinet B., Montadert L., Mascle A., Cazes M. and Bois C.,
New insights on the structure and formation of sedimentary
basins from deep seismic profiling in Western Europe. In:
Petroleum Geology of North West Europe, edited by
Brooks J. and Glennie K.W., Graham and Trotman,
London, 1, 11-31, 1987b.
Tapponnier P., Evolution tectonique du systme alpin en
Mediterran6e' poinqonnement et 6crasement rigide-
plastique. Bull. Soc. G6ol. France, 7, 19, 3, 437-460 (in
French), 1977.
Tucker R.M. and Arter G., The tectonic evolution of the
North Celtic Sea and Cardigan Bay basins with special
reference to basin inversion. Tectonophysics, 137, 291-
308, 1987.
Villemin T., Alvarez F. and Angelier J., The Rhinegraben:
extension, subsidence and shoulder uplift. Tectonophysics,
128, 47-59, 1986.
Wenzel F., Brun J.P. and the ECORS-DEKORP working
group, A Deep Reflection Seismic Line across the Northern
Rhine Graben. Earth Planet. Sci. Lett. (in press), 1991.
Werner D. and Kahle H.G., A geophysical study of the Rhine
Graben. I-Kinematics and geothermics. Geophys. J.R.
Astr. Soc., 62, 617-629, 1980.
Ziegler P.A., Geological Atlas of Western and Central
Europe. Shell Internationale Maatshapij, Elsevier,
Amsterdam, 130 p., 1982.
THE DEKORP SURVEYS: MAJOR RESULTS IN TECTONIC AND REFLECTIVE STYLES
R. Meissner and the DEKORP Research Group
Institute for Geophysics, University Kiel, Olshausenstr. 40, 2300 Kiel, Germany
Abstract. Looking back on more than five minations on zircon minerals and their coincidence
years of DEKORP's activity the great increase of with pan-African collision events support the idea
our knowledge on the deep structure and evolution by Gebauer et al. [1989] that the various Variscan
of the Variscan mountain belts in Central Europe units ere continental terranes, split apart from
is reviewed. Collisions and sutures have been Gondwana, collided partly with Laurussia, partly
imaged, and especially in the Rhenohercynian a with each other.
complex interwedging and interfingering of crustal The strong modification, shortening and thick-
units are observed. Similar processes can still be ening of the Variscan crust during its orogenic
verified in the Moldanubian/Saxothuringian area of cycles makes it nearly impossible to see anything
the Oberpfalz on the flanks of the large Bohemian from the pre-Variscan stage in the reflection
Massif. In contrast, the center part of the Mol- records. In addition, the desintegration of the
danubian and Saxothuringian shows numerous signs Variscan units was extremely intense and modified
of post-Variscan crustal modifications, like flat- the crust considerably, the most important devel-
tening and equilibration of the Moho, lower crust opment possibly being the removal (or "creeping
lamination, thin mafic lower crusts and many away") of the crustal mountain roots, together
plutons, possibly responsible for the rather with the erosion and the general thinning of the
transparent upper crust. These and other types of extended crust [Meissner et al., 1987]. Post-oro-
reflectivity and their correlation to tectonic genic volcanism and a secondary indenter tectonics
terranes and their boundaries can be explained by [Bachtadse and Van der Voo 1986] supported the
different thermal histories and different heat disruption of Central and Western Europe.
supplies with their strong impact on crustal
rheology. Reflectivity of Variscan Collisions
Introduction Although Variscan collisions started in the
south around 380 my [Ziegler, 1984] and progressed
It is the intention of this paper to review to the north until around 300 my we start our des-
the major achievements of DEKORP's first 5 years cription with the externides at the North Variscan
and integrate the results in the framework of tec- Deformation Front (NVDF) in the north. The line
tonic evolution and reflectivity patterns as they BELCORP/DEKORP 1-A and DEKORP 1-B could show that
are known today (DEKORP = D__eutsches K__ontinentales the thin-skinned tectonics of the Aachen Thrust/
eflexionsseismisches rogramm). DEKORP and some Faille du Midi as part of the NVDF [Meissner et
pre-DEKORP lines cover the Variscan internides and al. 1981] has its root 100 km to the south at a
externides, and they constitute by far the most depth of 15 km, forming a ramp and flat structure
extensive high-quality seismic reflection data [DEKORP Res. Group, 1991]. This means that it
within this late Paleozoic mountain belt (see extends well into the crystalline basement, simi-
Figure 1). They are always accompanied by wide- lar to the situation of the profile ECORS "Nord de
angle studies. The geological, tectonic evolution la France" further west [Bois et al., 1986]. A
of the belt is described in various other papers compilation of 12 seismic reflection profiles all
of this volume. Contribution to its genesis are along the NVDF up to western Ireland shows a
from Matte [1986], Ziegler [1983, 1984], Bachtadse similar thin-skinned pattern over more than 2000
and van der Voo [1986], Behr et al. [1984], Franke km [Meissner et al., 1990].
et al. [1990] and others, who all see the origin No more thin-skinned, but a deep reaching
of the belt by plate tectonic models. Age deter- compressional heritage is found east of the Rhine
River on profile DEKORP 2-N [Franke et al., 1990]
(Fig. 1). This profile shows a complex interfin-
ContinenmlLiosphere: Dp Seisn Refltions gering and indentation which is stronger than that
Geodynanfis22 on BELCORP/DEKORP 1-A (D 1A) and reaches deeper.
o 1991AmeHcanGeophysil Union Figure 2 shows the northern part of the NVDF (D
69
70 MAJOR RESULTS IN TECTONIC AND REFLECTIVE STYLES
5/..
52
50
6 8
:E231 PC
':'/ 0,5,010,0 km 5
48 o.
10 12 14
BASIN
The southern part of the Rhenish Massif shows
the strongest single reflections of all DEKORP
lines. It is speculated that during Variscan com-
pression slices from the "GieBen" ocean, a part of
the Rheic Suture, were incorporated into the shor-
tening foreland and thrust belts of the Rhenish
54 Massif. Figure 4 shows a section of this part of
DEKORP 2-N together with data of two single cover-
age (!) records from a pre-DEKORP survey in 1968
in the same area [Glocke and Meissner, 1976]. The
speculated incorporation of oceanic (or other
mafic) material into the southern Rhenish Massif
52 is supported by a strong increase of velocities
and a doubling of the lower crust compared to that
of the Saxothuringian zone in the south [Aichroth
et al., 1991] as will be discussed in the next
section. "Decollement" nappes like those of the
NVDF with its thin-skinned tectonics are not seen
in the internal collision zones of the Variscan
50 belts, but signs of enhanced compressions in form
of some remnants of former crustal roots, a buck-
ling of the lower crust or deep reaching NNW
vergent ramps are observed along the internal
collisions [Wever et al., 1990]. It has to be
mentioned that these relatively small signs of
former sutures are not seen in any of the refrac-
480 tion or wide-angle profiles.
Reflectivity of Post-Variscan Events
6 8 10 12 14
While thrust, nappes and crustal shortening
Figure 1. Simplified geological/tectonic map of have to be considered as signs of Variscan com-
Germany with DEKORP (D) and KTB (K) network 1984- pression, the great abundance of granitoids be-
1990. A: Ardennes, BM: Bohemian Massif, H: Harz tween 370 and 280 ma has to be related to syn- and
Mtns., LBM: London-Brabant Massif, NVDF: North post-orogenic extensional regimes. Their ages and
Variscan Deformation Front, O: Odenwald, R: Ries their abundance decrease from south to north
crater, RS: Rheinisches Schiefergebirge, S: [Matte, 1986]. As the outer Carboniferous basins
Schwarzwald, SNB: Saar-Nahe Basin, SP: Spessart are devoid of granites their great production in
Mtns., V: Vosges Mtns., (1)-(4): basement of the the Variscan belts must be related to Variscan
Variscan Foredeep, Rhenohercynian, Saxothuringian thickening (and subsequent stretching). High ini-
and Moldanubian, (5) Saar-Nahe Basin. tial S7Sr ratios and the absence of mafic intru-
sions indicate their origin from the middle and
lower crust. The post-orogenic granites must cover
1A) which is compared with the northern rim of the the same timespan as the maximum of erosion and
Rhenohercynian along DEKORP 2-N. Note the similar the disappearance of the mountain roots [Meissner
abrupt ending of the tectonized crust, the similar et al., 1987]. It is important to note that grani-
poor reflectivity of the northern foreland and the tic plutons so far cannot be observed by seismic
difference in reflectivity on both parts of the methods directly because their Vp values .are
Rhenish Massif. The Variscan foreland in general identical with their surroundings in the upper
shows no clear and continous reflections in the crust. They rose in times of maximum heating and
upper or lower crust. Especially the London- extension and seem to have produced a rather
Brabant Massif, which was strongly deformed during transparent upper crust, while lamination started
the Caledonian orogeny, has a chaotic and disrup- in the hot and low viscous lower crust. As the
rive reflectivity, as demonstrated by Bouckaert et lamination often crosses the collisional belts,
al. [1988]. At present we can only speculate that their lower base marking the new Moho, it is
a multiple and complex compression in a cold en- definitely post-Variscan. Even if we consider that
vironment may have created such a disruptive ap- the elevation of the collisional belts was only 2
pearance. km high, the crustal root according to isostatic
Figure 3 shows a detailed location map of the balance and present experience [Woollard, 1969]
seismic lines DEKORP 2 (D 2N and D 2S) and the (Meissner and Mooney, submitted, 1991) had at
southern part of the European Geotraverse (EGT) least an excess thickness of 7 x 2 km = 14 km. But
crossing the Variscan units Moldanubian, Saxothu- DEKORP lines crossing the belts show only about 1
ringian and Rhenohercynian. This figure covers the km excess thickness or even less. The whole lower
area discussed together with the next two figures. crust has rejuvenated. This disappearance of crus-
o
5
lO
..-..
MEISSNER AND DEKORP RES. GROUP 71
vv . /SVA ) Eifel Nord-Sod Zone O Upper
Efehan
Erosion
0 20 40 50 80 [ (k o ........
BELCORP/DEKOIP 1-A
NNW SSE
So tonerr o Eb be Siege n
Thrust Anticl. inorium Thrust
I 150km 100km I
c I I I I I I I I i I I I 0
"L::-:.,_ :__-_-. -:.. , '::,..:,?--z::, :,,_t,, .--- ':-.--,,-z-.-.,--. _,--:- -/; -'-:-;:'-- - , ----.: -: -I
---i- _.-r:r'. ._.--,-. .... ."..-. _--.-. -.,-., .- :-. -:. ,;--_-. -...,..-..-- , _--.- ..... -,_
.}'- - .--: . .z -_ -.,,' .- - .=-. -- "- - .-- -. - .% - .. e - .-. ---' ' - -" .- -:-- -- "C.-.-" -" '; -'..-- - . ' .e'_ ,,. - '-- -
.... - _ .-.- .{_ .....:._..,. :-..,,, ..... - _ . -. .:.: _ __-_: -, ......... ,,,:., _.=_ -. - ..... .
- -" - '-.'- "'- --'",x ,-,-....-.,,.* _ _ '.-- .=. _ - F..:..-,"'.. - -.: .._ ...' - -,-'x. .C'"- '' .---. .'.x. ..,.
]: . . :- ,-.,,.:,,. -,- ,,.__.... _-.-,.-, -.,:.-::,.: :-:..- -__.-.: .... _., .:.-...._ -.: -- L._<.:-.- -,,,_-_,
' '" " -'_"2':" _ _-,'."..:..-,.:.-, 77..c -'-_ _- - ----- --.-,..,. -- -?'... -=--- ._-._-
/ -: -- . - _-.'-.,- e.--- :: :-;. - ..x.'..'E-%-t:_' -: -. --, ,-- -,--: =-_-. -- - -- .- _----. :; =-:.. __ --. = -- - : - '1
,- / -:' -=.-.: -- ':_---_ .-t- "--- '-=--T,.:-'- :; - C ' _:-:'..z-: - :._.z._-:_=,.:- ,-'-'- ---- -" - /-
t) - - :-, ..... . - ,,;,, _-,..--e_---.:. :.:..._ '--,,,=3 - : __..: :L':-. = t-=-'-,,, .-- .... "'- -:=.g-W_-,__.-=- _ _ ::)
' -'- -': -: '- - - -= ...... __--:_. :'- ' = '"-- -.--=-- -'---" - ----'---'-: r -'- ---- ---_1
--- I : '- _-. - - :::-_-- --_ --: --__---: - :- ? '--_ _-----'-&--______-:--_-_-?-----':d '-'
be / =- --- :' -- :--"--- '-- :-: -:---: - - .... --_:c'------' - ' .... .-I
z I --- -"- ' - '. ' - ' - --"-'--'"-:'----"" -- .-"'---I
-/: '-- .- - '- .......... " .?''---__---.?-_:.:
/-- '- -- , - - , -. -- _:,-*-,: '.?:-- - -------
0'1-- -_.._:- . ___ -_..:-_.:-__--__-:: :-::...-_ .::_:-.::-_- ---:__.?- F-u
/ b
Figure 2. The northern part of the Rhenish Massif, (a) thin-skinned tectonics in the west
(BELCORP/DEKORP I-A), versus (b) more complex indentations in the east (DEKORP 2-N).
tal mountain roots in post-orogenic extensional weil-known superposition system of sediments, by
regimes must have been accompanied by huge ver- which deep layers get increasingly older can defi-
tical and horizontal movements in the lower crust nitely not be applied to the lower continental
and uppermost mantle, possibly supported by mag- crust.
matic intrusions. As a consequence of these pro- The widespread appearance of lower crust la-
cesses the lower crust, the uppermost mantle and mellae in most - but not in all - areas of Varis-
the Moho became younger than the upper crust. The can and Caledonian Europe has been reported fre-
8E 10E
52N o 52N
48ON : , I j I , I Z, 8N
8E 10E
Figure 3. Location map of the southern part of the
EGT profile and of profiles DEKORP 2-N and 2-S
crossing the Moldanubian, Saxothuringian and Rhe-
nohercynian units of the Variscides. 6, 8: shot-
point number 6 and 8 from a single-coverage profile
[Glocke and Meissner, 1976].
flectivity in the whole area and can not be over-
looked. Similar triangle-shaped bodies are inclu-
ded in other terranes as well from the Variscan,
e.g. the Long Island Platform [Phinney and Roy-
Chowdhuri, 1989] to the Tertiary, e.g. parts of
the Himalaya [Mattauer, 1986]. We believe that
some movements (possibly inverse ?) along the
triangle-shaped B-body continued in post-Variscan
time when desintegration, disruption and equili-
bration, intrusions (diffraction) and Moho flat-
tening took place, similar to processes observed
along parts of the American Appalachians [Mat-
tauer, 1986].
Profile DEKORP 2-S and the EGT line provide
another chance to compare near-vertical and wide-
angle seismic data. The lower crust in the Mol-
danubian and the $axothuringian, marked by Vp-
velocities of 6.4 to 7.0 km/s is only 6 to 8 km
thick. It begins below 20 km in the south, and
some delamination, stretching and/or flattening
SE
o km
o
quently [Matthews, 1986; BIRPS and ECORS, 1986;
Lschen et al., 1987; Meissner, 1989; Sadowiak et
al., 1991]. In the following, diffraction clus-
ters, apparently emerging from a laminated lower
crust, will be discussed. Their largest concen-
tration is on the line DEKORP 2-$ in the $axothu-
ringian. It has been suggested [DEKORP Res. Group,
1985] that the diffractions might constitute mafic
intrusions from the mantle, which got stuck in the
lower crust because of decreasing buoyancy and
increasing viscosity. A comparison of the location
of the diffraction cluster with velocity data from
the European Geotraverse (EGT) in the Moldanubian
[Aichroth et al., 1991] suggests a correlation
with strong anomalies of upper mantle velocities
(Figures 5a and b). High Vp-values, up to 8.4
km/s, are observed in the upper mantle beneath the
diffraction cluster and might be related to a
strongly depleted mantle, which possibly has lost
some of its temporarily fluid constituents in form
of magmatic blobs to the lower crust of the Molda-
nubian. It seems also possible that the crust in
the Saxothuringian terrane was especially weak.
The triangle-shaped body B in Figure 5b in the
upper crust seems to represent a "sub-terrane"
r"] = 5P. 6: D strong energy
0 = SP. 8 : ( strong energy
within the Saxothuringian which was squeezed be- Figure 4. Seismogram (unmigrated) of the southern
tween adjacent terranes during the Variscan com- part of profile DEKORP 2-N in the Rhenish Massif
pressional phase. During the completion of this near GieBen with strong reflectivity [after Franke
paper the new Variscan terrane map of Matte et al. et al., 1990]. Reflections from a single-coverage
[1990] came out and showed a terrane boundary profile from 1968 [Glocke and Meissner, 1976] are
exactly at its seismically detected location. Its marked. Strong reflectivity might indicate the
steep NW dipping SE part shows the strongest re- involvement of slices of oceanic material. M: Moho.
z. oo
o o
__--__-
lO -- .....
20 ...... 6
- - -66 - - -
30
o l o
500 600 ( k m )
! : :s..:: - -:: -- 6 0 __j0
vL< v 6; .... [o
....... 6 2 .............. [20
6 2 km/s
CONRAD --_ --
"_-_' ]---- 6 8 ....... _-_-----_ IVJOHO ......... 6{i
NW
RHENOHERCYNIAN
De,on. I VocanicslFysch NappeslDevonien [ Tertiory
I DILL I GIESSEN I ..... /
lSYNCLINEI NAPPE I'u"u3/WETTERAU
SE
SAXOTHURINGIAN [ MOLDANUBIAN
I c,,t.Ro,l Buntar I Huschelkalk
SWAB
I
50 200 150 100 50 (kin) 0
I
0 ' ' ' '0
12 12
D 2-N I D 2-5 J D i f f r o c t i o n s J
Figure 5. Comparison of the ..GT profile and D..KORP 2-$/2-N in the area of int. esection
of both lines (see Figure 3), (a) EGT line with V-isolines, Conrad and Moho disconti-
nuities [after Aichroth et al., 1991], (b) automatic line drawing (migrated) of DEKORP 2-
S/2-N [Meissner and Bortfeld, 1990], Conrad (C) and Moho (M) included from (a),
B: possible boundaries of a mini-terrane, "diffractions": area of strong diffraction
clusters in unmigrated section. Note the coincidence of "wide-angle" and "reflection" Moho
NW of km 150 and the begin of lamination above the "wide-angle" Conrad discontinuity.
has definitely to be invoked (Fig. 5a). Comparing postulated that creep in low -zones generated the
the thickness of the laminated (reflective) lower subhorizontal smoothing of any inhomogeneities and
crust in Fig. 5b with that of the (wide-angle) hence the laminated appearance of reflectivity.
lower crust (Fig. 5a) below the "Conrad", as The same viscosity () level might be found in
marked by the 6.4 km/s isoline, the observation of sialic material at the bottom of the upper crust
Wever [1989] is confirmed, that the lamination and in the more mafic lower crust. Figure 6a shows
generally starts at much shallower depth than the an -model of Variscan crust, on present tempera-
Conrad Discontinuity. Whereas the "reflection" ture estimates. The Saxothuringian and the Mol-
Moho either coincides with or is slightly deeper danubian terranes show low V-values and probably
than the "wide-angle" Moho, the "reflective" Con- weak, sialic material down to at least 20 km. From
rad, i.e. the begin of lamination, starts well in the rheological point it seems probable that hot
the low-velocity zone (LVZ) of the upper crust mafic material, after entering the weak middle
which is definitely still sialic and above the crust and cooling, became rather rigid and formed
"real" Conrad discontinuity. The lamellae, which some hard plumes in the weak matrix of the sialic
are so widespread in Western Europe [Sadowiak et middle and lower crust (Fig. 6b), giving rise to
al., 1991], do not seem to be related to velocity diffractions as seen in Figs. 5a, b.
and hence petrology, as suggested already much The situation in the Rhenohercynian along
earlier. Also the crustal LVZ in the $axothurin- lines BELCORP/DEKORP 1-A and DEKORP 2-N and at the
gian and Moldanubian is especially pronounced and flanks of the Bohemian Massif (DEKORP 4) is dif-
may be considered another indication for sialic ferent. Here, apparently the whole temperature (T)
and weak terranes in general. level never reached the high values of the centers
of the Moldanubian and Saxothuringian. During and
after the orogeny the crust behaved much more
Rheological Consequences brittle. This conclusion is based on three obser-
vations:
It was stated in various studies [Strehlau and (1) No lower crust lamination is observed
Meissner, 1987; Smith and Bruhn, 1984] that the along these profiles (attributed to an absence of
widespread appearance of seismic lamination in the smoothing effects due to only medium T and medium
lower crust do not correlate with any special ).
velocity value V but cover the whole range of (2) No transparent upper crust is seen (at-
crustal V-values. This seems to agree with cer- tributed to missing granitic plutons due to the
rain crustal viscosity () levels, and it was same medium -T levels)
o b
>ln'r I
. . y+ Transparent /
20 _ 20-
o- m
c
6 7 8
' ' ' Vp
(kmls)
I
W Br,ttle
z (kin) z (km)
Figure 6. Concept of rheology (a), reflectivity granitic plutons, non-reflective, (3) late mafic
and structure (b), and velocity-depth function (c) intrusions, diffractire, (4) granitic-gneissic
in Saxothuringian and Moldanubian terranes; material (upper crust), (5) mafic material (lower
(a) viscosity () model of Strehlau and Meissner crust), (6) peridotitic (upper mantle);
[1987], C: Conrad discontinuity, M: Moho discon- (c) Vp-Z model derived from the EGT [modified after
tinuity; (b) (1) reflectivity (schematic), (2) Aichroth et al., 1991].
(3) Single reflections, ramp and flat struc- Inside the $axothuringian there might be an ad-
tures and dipping bands of reflectors cover the ditional small terrane as observed by a suspect
whole crust (attributed to a dominance of inden- triangle-shaped body in the upper and middle crust
tations and interfingering of rather hard layers along line DEKORP 2-S in accordence with Matte et
under the same medium -T levels). al. [1990]. In contrast to the Rhenohercynian, the
The latter statement is supported by a compa- $axothuringian and the Moldanubian were strongly
rison with Vp-values from wide-angle measurements affected by post-Variscan extension and heating.
along the reflection lines. For BELCORP/DEKORP 1-A This is seen by (1) a flat and shallow Moho, (2) a
the ramp and flat structure of the Aachen over- thin lower crust (based on wide-angle velocity
thrust seems to be an especially hard splinter studies) (3) the presence of lamellae in the lower
with an increased V in the south i.e. in the crust. An additional observation comes from the
crystalline basement (while in the sediments near central part of DEKORP 2-5: In addition to the
Aachen a weak gouge material is found) [DEKORP lamellae many strong diffractions are observed in
Res. Group, 1991]. Also for line DEKORP 2-N the the lower crust, tentatively interpreted as mafic
accumulation of reflections at various depth in- magmas which got stuck in the lower crust after
tervals correlate with high V layers, [Giese et cooling. They might have left a depleted upper
al., 1990] supposed to have also a stiffer rheo- mantle, as suggested by the high Vp-velocities
logy [Meissner and Tanner, 1990]. (8.4 km/s) observed here. It should also be
stressed that the "reflection" Conrad, i.e. the
Discussion and Conclusions upper boundary of lower crust lamination, is con-
siderably higher than the "wide-angle" Conrad, as
One of the main results of the DEKORP research found by identically arranged wide-angle surveys,
is the observation that the various Variscan ter- while the "reflection" and the "wide-angle" Moho
ranes have different styles of reflectivity, de- are mostly identical except for the SE part of
pending on their tectonic and thermal development. line DEKORP 2-5 where the profiles diverge from
The Rhenohercynian presents a huge fold and thrust each other.
belt with an abrupt termination to the north. West Especially strong lamination is seen on the
of the Rhine River the North Variscan Deformation KTB Schwarzwald profiles. It might have been cre-
Front extends to more than 2000 km up to Ireland, ated or enhanced by thermal and stress patterns of
mainly in form of thin-skinned tectonics around the Tertiary Rhinegraben. In contrast, the Ober-
the London Brabant Massif. The DEKORP 1 line, west pfalz area on the flanks of the large Bohemian
of the Rhine River, in addition, shows clear signs Massif [DEKORP 4 and the KTB lines) is not so
of Tertiary magmatism in the whole crust and post- strongly affected by post-Variscan thermal events,
Varican desintegration and extension in the $aar- similar to part of the Rhenohercynian. It does not
Nahe Trough. DEKORP 2-N, east of the Rhine River, exhibit much lower-crust lamellae but shows a
seems to show the original thrusting and inter- complex interwedging of the $axothuringian below
fingering much better although no thin-skinned (and into) the Moldanubian. Towards the northeast,
termination to the north. Anomalously strong re- however, reflectivity in the lower crust in-
flections in the middle crust suggest the presence creases, and a thermal peak from the Tertiary Eger
of slices of oceanic lithosphere, especially in graben might be responsible for the beginning
the southern part of line DEKORP 2-N. lamination [Trappe and Wever, 1990].
The boundary between the $axothuringian and In these and other areas the presence and de-
the Rhenohercynian is clearly visible in the re- velopment of reflectivity can not be understood
flection records, especially in the upper crust. without taking into account thermal and rheologi-
cal processes. The exponential dependence of vis- b.y. old detrital zircons, Tectonophysics, 157,
cosity on temperature makes crustal theology very 81-96, 1988.
sensitive to thermal effects. Crustal collisions Giese, P., Ibeken, S., Baler, B., and K. Schulze-
and sutures are primarily cold processes, but Frerichs, Accompanying seismic refraction inves-
subsequent collapse and extension provide heat. tigations along the profile DEKORP 2-North, Geol.
Basic modifications of deep crustal structure, Rundschau, 7--9, 567-579, 1990.
like Moho-flattening, lamination and intrusions, Glocke, A., and R. Meissner, Near-vertical reflec-
have taken place in post-orogenic environments. tions recorded at the wide-angle profile in the
Rhenish Massif, in Explosion Seismoloqy in Cen-
Acknowledqments. We acknowledge with thanks tral Europe, edited by P. Giese, C. Prodehl and
the continuous support of DEKORP by the Ministery A. Stein, Springer, Berlin, 252- 256, 1976.
of Research and Technology (BMFT) and by the Get- Lschen, E., et al., Seismic reconnaissance studies
man Research Association (DFG) in Bonn. Special of the deep drilling location KTB- Black Forest,
thanks are due to C. Bols for critical comments SW Germany, J. Geophys., 6--2, 1-30, 1987.
and reviewing, P. Sadowiak for editing and H. Mattauer, M., Intracontinental subduction, crust-
Peters for preparing the figures. Contribution mantle decollement and crustal stacking wedge in
number 417 of Inst. fr Geophysik, Kiel. the Himalayas and other collisional belts, in
Collisional Tectonics, edited by M. Coward and A.
References Ries, Spec. Publ. Geol. Soc. London, Vol. 19,
Aichroth, B., Prodehl, C., and H. Thybo, Crustal Blackwell, Oxford, 73-50, 1986.
structure along the central segment of the EGT, Matte, P., Maluski, H., Rajlich, P., and W. Franke,
Tectonophysics, in press, 1991. Terrane boundaries in the Bohemian Massif= Result
Bachtadse, V., and R. Van der Voo, Paleomagnetic, of large-scale Variscan shearing, Tectonophysics,
evidence for crustal and thin-skinned rotations 177, 151-170, 1990.
in the European Hercynides, Geophys. Res. Lett., Matthews, D.H., Seismic reflections from the lower
1--3, 161-164, 1986. crust around Britain, in The Nature of the Lower
Behr, H., Engel, W., Franke, W., Giese, P., and K. Continental Crust, edited by J. Dawson, D. Cars-
Weber, The Variscan belt in Central Europe= main well, J. Hall and K. Wedepohl, Geol. Soc. London
structures, geodynamic implications, open ques- Spec. Publ., Vol. 24, Blackwell, Oxford, 11-21,
tions, Tectonophysics, 109, 15-40, 1984. 1986.
BIRPS and ECORS, Deep seismic reflection profiling Matte, Ph., Tectonics and plate tectonic model for
between England, France and Ireland, J. Geol. the Variscan belt in Europe, Tectonophysics, 126,
Soc. Lond., 143, 45-52, 1986. 329-374, 1986.
Bois, C., Damotte, B., Mascle, A., Torreilles, G., Meissner, R., Rupture, creep lamellae and croco-
Galdeano, A., Him, A., Matte, P., and J. Rao- diles= happenings in the continental crust, Terra
ult, Deep seismic reflection profiling of the Nova, , 17-28, 1989.
crust in northern France= Programme ECORS, in Meissner, R., and B. Tanner, Structure and tec-
Reflection Seismoloqy: A Global Perspective, tonics of the Oberpfalz area and its surroun-
edited by M. Barazangi and L.D. Brown, AGU, Geo- dings, in Super Deep Drill Continental Drillinq
dyn. Set., vol. 14, Washington, 43-55, 1986. and Deep Geophysical Soundinq, edited by K.
Bouckaert, J., Fock, W. and N. Vandenberghe, First Fuchs et al., Springer Verlag, Stuttgart, 120-
results of the Belgian geotraverse 1986 133, 1990.
(BELCORP), Ann. Geol. Belqique, 111, 279-290, Meissner, R., Bartelsen, H., Murawski, H., Thin-
1988. skinned tectonics in the northern Rhenish Massif,
DEKORP Research Group, First results and prelimi- Germany, Nature, 290, 399-401, 1981.
nary interpretation of deep reflection seismic Meissner, R., Wever, Th., and E.R. Flh, The Moho
recordings along profile DEKORP 2-S, J. Geophs.,
57, 137-163, 1985.
DEKORP Research Group, Results of the DEKORP 1
(BELCORP-DEKORP) deep seismic reflection studies
in the western part of the Rhenish Massif, Geo-
phys. J. Int., in press, 1991.
Franke, W., Bortfeld, R.K., Brix, M., Drozdzewski,
G., Drbaum, H.J., Giese, P., Janoth, W., J-
in Europe - implications for crustal development,
Ann. Geophysicae, 5__B, 357-364, 1987.
Meissner, R., Wever, Th., Sadowiak, P., and the
DEKORP Res. Group, Reflectivity patterns in the
Variscan mountain belts and adjacent areas: an
attempt for a pattern recognition and correlation
to tectonic units, Tectonophysics, 173, 361-378,
1990.
dicke, H., Reicheft, Chr., Scherp, A., Schmoll, Phinney, R.A., and K. Roy-Chowdhury, Reflection
J., Thomas, R., Thinker, M., Weber, K., Wiesner, seismic studies of crustal structure in the east-
M.G., and H.K. Wong, Crustal structure of the ern United States, in Geophysical Framework of
Rhenish Massif: results of deep seismic reflec- the Continental United States, edited by L.C.
tion lines DEKORP 2-North and 2-North-Q, Geolo- Pakiser and W.D. Mooney, Geol. Soc. Am., Memoir
qische Rundschau, 79, 523-566, 1990. 172, 613-653, 1989.
Gebauer, D., Williams, I., Compston, W. and M. Gr- $adowiak, P., Wever, Th., and R. Meissner, Deep
nenfelder, The development of Central European seismic reflectivity patterns in specific tec-
crust since the Early Archean based on convent- tonic units of Western and Central Europe, Geo-
ional and ion-microprobe dating of up to 3.84 phys. J. Int., 105, in press, 1991.
Smith, R., and R. Bruhn, Intraplate extensional along the central part of the European Geotra-
tectonics of the eastern Basin-Range: inferences verse in Germany: a review, Tectonophysics, 176,
from structural style from seismic reflection 87-101, 1990.
data, regional tectonics, and thermal-mechanical Wever, Th., Conrad and top of the reflective lower
models of brittle-ductile deformation, J. Geo- crust - do they coincide? Tectonophysics, 157,
phys. Res., 89, 5733-5762, 1984. 39-58, 1989.
Strehlau, J., and R. Meissner, Estimation of Woollard, P., Regional variations in gravity, in
crustal viscosities and their role in geodynamic The Earth's Crust and Upper Mantle, edited by
processes, in The Composition, Structure and P.J. Hart, AGU, Geophys. Monoqr., vol. 13, 320-
Dynamics of the Lithosphere Asthenosphere System, 341, 1969.
edited by K. Fuchs and C. Froidevaux, Geodyn. Ziegler, P., Geoloqical Atlas of Western and
Ser, vol. 16, AGU, Washington, 69-87, 1987. Central Europe, Elsevier, Amsterdam, 130 pp,
Trappe, H., and Th. Wever, Seismic evidence of in 1982.
creased tectonothermal activity near the Ober- Ziegler, P., Caledonian and Hercynian crustal
pfalz deep continental drilling location (SE Ger- consolidation of Western and Central Europe - a
many), Geol. Rundschau, 7--9, 649-658, 1990. working hypothesis, Geol. Mijnbouw, 6--3, 93-108,
Wever, Th., Meissner, R., Sadowiak, P., and the 1984.
DEKOPRP Res. Group, Deep reflection seismic data
REFLECTIVITY OF A PROTEROZOIC SHIELD: EXAMPLES FROM BABEL SEISMIC PROFILES ACROSS FENNOSCANDIA
BABEL Working Group*
Abstract. We present examples of recently in the past mainly in Phanerozoic terranes. In
acquired marine seismic reflection profiles across other parts of the BABEL sections we see an appa-
the Baltic Shield. Our regional transect across a rently transparent crust, a diffractire crust, or
distance of 1350 km shows a higly variable range of regions without a clear reflectivity transition
reflectivity patterns in the crust. Some of the from crust to mantle, features that have commonly
data exhibit a clear reflection Moho and bright been recognized in Precambrian areas in other
lower-crustal reflections that have been recognized surveys. Though this paper is based only on brute
stacks, the excellent quality of the BABEL data
clearly shows that it is inappropiate to distin-
4embers of the BABEL Working Group: R. Meissner, guish generically between Phanerozoic and Precam-
P. Sadowiak , S. Thomas, T. Wever:, Institute for brian reflectivity patterns. Rather, the variabili-
Geophysics, University Kiel, Olshausenstr. 40-60, ty in reflectivity along the BABEL profiles seems
D-2300 Kiel 1, Germany; T. Dickmann, E.R. Flh, more probably due to differences in structures and
GEOMAR, Wischhofstrasse 1-3, D-2300 Kiel 14, Ger- lithology than to differences in crustal age.
many; T. Dahl-Jensen 3. R.W. Hobbs, S.L. Klempe-
rer '4, D.H. Matthews, D.B. Snyder, BIRPS, Bullard Introduction
Laboratories, Cambridge University, Madingley
Rise, Madingley Road, Cambridge CB3 0EZ, UK; R. As deep seismic reflection profiling has pro-
Long, T. Matthews, Department of Geological Sci- gressed during the last 15 years, various theo-
ences, Durham University, South Road, Durham DH1 ries about the reflectivity of the continental
3LE, UK; D.J. Blundell, Department of Geology, crust have evolved. In the past five years corre-
RHB New College, Egham TW20 0EX, UK; A. Berthel- lations between crustal reflectivity and tectonic
sen, H. Thybo, Institut for Almen Geologi, ster environments have been suggested. The most popu-
Voldgade 10, DK-1350 Copenhagen K, Denmark; N. lar hypotheses for crustal reflections are mafic
Balling, E. Nrmark, Department of Earth Scien- intrusions, compositional layering, ductile shear
ces, Aarhus University, Finlandsgade 6, DK-8200 zones in high-grade rocks, and fluids trapped in
Aarhus N, Denmark; C.-E. Lund, H. Palm, L.B. Pe- the lower crust (for a summary see Matthews
dersen, R.G. Roberts, Department of Solid Earth [1986]).
Geophysics, Uppsala University, Box 556, S-75122 As more deep seismic reflection profiles in
Uppsala, Sweden; S.-A. Elming, Department of Geo- different geological and tectonic areas become
physics, Lule& Technical University, S-95187 Lu- available, covering perhaps about 50,000 km
le&, Sweden; P. Heikkinen, H. Korhonen, U. Luos- worldwide, lower-crustal reflectivity becomes
to, Institute of Seismology, Helsinki Universi- less of a common global feature. Moreover, rari-
ty, Et. Hesperiankatu 4, SF-00100 Helsinki, Fin- ous reflectivity patterns that are independent of
land; S.-E. Hjelt, K. Komminaho, Department of the seismic source (Vibroseis - explosives - air
Geophysics, Oulu University, SF-90570 Oulu, Fin- guns) can be observed and correlated with speci-
land; J. Yliniemi, Geophysical Observatory, Oulu fic tectonic units [Sadowiak et al., 1991]. It
University, SF-90570 Oulu, Finland. turned out that lamellae and bands of reflections
authors for correspondence, :now at Forschungsan- are predominant in extensional areas of post-
stalt der Bw. fr Wasserschall- und Geophysik orogenic collapsed regions whereas many Precam-
(FWG), Klausdorfer Weg 2-24, D-2300 Kiel 14, Ger- brian crustal units surveyed until now exhibit a
many, 3now at Greenland Geological Survey, ster "shield-type" reflectivity [Wever et al., 1987]
Voldgade 10, DK-1350 Copenhagen K, Denmark, 4now characterized by diffractions, by reflectivity
at Department of Geophysics, Mitchell 366, Stan- decreasing with depth, and by the paucity of Moho
ford University, Stanford, CA 94305-2215, U.S.A. reflections (for details see Meissner [1986]).
These conclusions are based largely on data from
the North American [Brown et al., 1983; Gibbs,
Continenmlosphere: Dp Seisnfic Refltions 1986] and Australian cratons [Finlayson, 1982].
Geodynan 22 It is not clear whether the general conclusion
o 1991AmcdcanGeophysical Union can be drawn that stable cratonic areas have fun-
77
78 REFLECTIVITY OF THE BALTIC SHIELD
damentally different reflectivity patterns from Sweden, across the Gulf of Bothnia, the Aland
Phanerozoic areas [Meissner, 1986]. islands, to southern Finland. In the same area,
The new marine deep seismic reflection pro- approximately 1.3 Ga old Jotnian red sandstones
files of the BABEL survey (Baltic nd othnian are locally present, including the occurrences in
choes from the ithosphere), a joint venture of the Gulf of Bothnia lAxberg, 1981]. They are in-
British, Danish, Finnish, German, and Swedish truded by c. 1.25 Ga old dolerite sheets of re-
scientists carried out in September/October 1989
in the Baltic Sea and Gulf of Bothnia, consist of
nine deep seismic profiles totalling 2,268 km of
high-quality data forming a 1,350 km traverse
from the edge of the Arcbean nucleus to Phanero-
zoic Western Europe (Figure 1). The BABEL survey
acquired both deep seismic reflection data and
coincident refraction/wide-angle reflection data
providing a large dataset to study the crustal
structure of the Baltic Shield and its transition
to the Phanerozoic crust of Western Europe across
the Tornquist Zone. In this paper we describe the
seismic near-vertical reflection data based on
high-quality brute stacks. We emphasize the re-
flection character and compare this to reflecti-
vity patterns from other areas. Technical details
of the experiment are summarized by BABEL Workinq
Group [1990]. Other regional geological results
from the reflection lines, e.g. from the Torn-
quist Zone, and results from the wide-angle data
are published elsewhere (see e.g., BABEL Working
Group [1991a, 1991b]). In this paper some typical
characteristics as well as some of the more stri-
king features will be illustrated and discussed.
Geological Evolution of the Baltic Shield
The following summary of the evolution of the
Baltic Shield is based to a large extent on Ga&l
and Gorbatschev [1987] and Gorbatschev et al.
[1979].
The formation of the Precambrian shield star-
ted in SE-Karelia 3.5-3.2 Ga ago. After large-
scale accretion an Arcbean proto-shield or nucle-
us was formed in the northeastern half of the
present shield. During the Early Proterozoic the
consolidated Arcbean craton was intruded by dyke
swarms and divided by rifts. In the northern part
of the craton a NW-SE trending spreading ocean
developed [Marker, 1989]. The subsequent Sveco-
karelian/Svecofennian orogeny (2.0-1.8 Ga) re-
joined the separated Arcbean terranes. Closing of
the ocean in the north resulted in the formation
of a collisional suture about 1.9 Ga ago [Marker,
1989].
More or less coeval with the evolution in the
nucleus, the crust of the Svecofennian domain was
accreted at the southwestern margin of the former
KARELIDES I
:%:::::
*, SVECO-
, FENNIDES
[, '" ,nland
SVECO FENNIDES
' . B
sou.:t
ESTE R N"t.',.
OROGENS%.
) '"//' CA":EDONIDE" :::
5' ; ermany
Arcbean craton. Late ascent of crustal-derived Fig. 1. Location map of BABEL reflection profiles
granitoids contributed to this until 1.78 Ga. (thick solid lines marked by numerals and letters)
Between 1.8 and 1.55 Ga ago when the GothJan and seismometer arrays or stations (triangles)
orogeny evolved further to the SW the granites together with some tectonic-stratigraphic ter-
and porphyrites of the NS-trending Trans-Scandi- ranes. Circled numerals along the profiles in-
navian Igneous Belt were intruded in adjacent dicate figure number illustrating those segments
parts of the Svecofennian domain, and more dis- within this paper. Stippled pattern indicates
tally, 1.7-1.55 Ga old post-orogenic Rapakivi tectonic boundaries. A1 = land islands, Bo =
granites and dyke swarms were emplaced within a Bornholm island, 6 = 61and island, TZ = Tornquist
broad E-W trending zone extending from northern Zone.
BABEL WORKING GROUP 79
gional extent and a cumulate thickness of a few (1) Complex, hiqhly variable reflectivity pat-
hundred meters [Gorbatschev et al., 1979]. Most terns with diffractions as well as bands and lay-
of the Gothian orogen in the SW was reworked by erinq in the lower crust
the Grenvillian/Sveconorwegian orogeny 1.2-0.9 Ga
ago. Figure 2 shows the northern part of BABEL line
With the close of the Caledonian orogeny du- 6 in the Bothnian Sea. This section exhibits very
ring the Devonian the western border of the pre- rapid lateral variation in reflection character.
sent shield had finally been outlined. The northern 15 km are marked by continous hori-
The Tornquist Zone is a still younger, NW-SE zontal reflections in the lower crust, i.e. bands
running tectonic lineament in Europe that cuts at 7.4-8.4 s, 9.5-10.5 s and 12.8-13.6 s TWT. The
through shield-type Precambrian basement across Moho, assumed to occur near 15 s TWT, is charac-
the southern Baltic Shield, whereas it marks in terized by numerous diffractions and no continous
Poland the border between the Precambrian Plat- reflections. Further south, diffractions dominate
form and Paleozoic Central Europe [BABEL Working the middle and lower crust (6-15 s TWT). "Reflec-
Group, 1991b]. tions" in the upper crust are caused by reverbe-
rations within the water between surface and a
Reflectivity Patterns of the Baltic Shield hard limestone seafloor.
A reflective lower crust of subhorizontal re-
The BABEL transect (Figure 1) starts in the flections beginning at 10 s TWT and ending with
south within the North German Basin, underlain by the reflection Moho at 15 s TWT is seen at the
Caledonian basement, and runs across the Torn- eastern end of BABEL line 7 in the southeastern
quist Zone near Bornholm island to the platform Bothnian Sea (Figure 3). In contrast, the mid-
region adjacent to the Baltic Shield border. Tra- and upper crust show isolated dipping reflections
versing 1350 km, it passes the land islands and or diffractions. This reflection character resem-
the Skellefte district, ending in the northern bles that seen throughout much of Phanerozoic
Bothnian Bay. Most of the data come from the Pro- Western Europe (e.g., BIRPS and ECORS [1986], L-
terozoic Svecofennides, yet they exhibit consi- schen et al. [1986]). However, this is the only
derable diversity in reflectivity. Hence, it is example of such "layering" on BABEL profiles.
not possible to characterize the data by only one
type of reflectivity. Nearly all sections exhibit {2) Low reflectivity and diffractions
probable diffractions, either dominating or occu-
ring in addition to other features. In some parts Not all parts of the BABEL lines show Moho
we see prominent reflections in the lower crust, reflections or bands of reflections in the lower
similar to the bands of reflections of profiles crust. Many sections exhibit low reflectivity
in post-orogenic areas of Phanerozoic Europe. A throughout the whole crust. Diffractions are also
sharp reflection Moho, marked by a single dis- present in these parts and Moho reflections are
tinctive reflection rather than by the downward questionable or absent, e.g. on Figure 4 which
termination of lower-crustal reflections, is shows a part of line B off the island of 61and in
imaged in some areas at depths ranging from 13 s the Baltic Sea. Fewer reflections occur in this
to 16 s two-way traveltime (TWT). Other regions, section, particularly toward the base of the
however, are marked by a transparent crust or by crust at about 14 s TWT, although a weak north-
a diffractive crust, and do not show any clear ward-dipping reflective zone cuts across the
transition from crust to mantle. In the follow- crust from about 8 s in the south to 12 s TWT in
ing, we will describe the BABEL data by showing the north.
both typical as well as striking reflectivity
patterns. In particular, we show examples demon- (3) A briqht, dippinq packaqe of reflections
strating
(1) complex, highly variable reflectivity pat- Figure 5 shows the southern part of BABEL pro-
terns with diffractions as well as bands and lay- file 2 (migrated) in the Gulf of Bothnia. It is
ering in the lower crust, characterized by a package of strong complex re-
(2) low reflectivity and diffractions, flections dipping to the north between 15-19 s
(3) a bright dipping package of reflections, TWT, the base of which is coincident with the
(4) strong subhorizontal reflections in the Moho determined from wide-angle observations
upper crust, and [BABEL Working Group, 1990]. Dipping reflections
(5) variability of reflection character. occur throughout most of the crust. The Moho is
A detailed interpretation of the data will be observed at about 13.5 s TWT just north of this
the subject of subsequent papers. Since we cannot section. Nearly the same pattern is imaged on the
show all the data, the parts presented are chosen parallel BABEL line 4 to the east. Here also di-
to give an insight into the variability. Each vergent crustal reflections and a Moho offset are
figure shows a segment of about 50 km of migrated found [BABEL Working Group, 1990]. This pattern
or unmigrated data. All seismic sections are dis- looks similar to that observed in Phanerozoic
played without vertical exaggeration, for a velo- orogens, e.g. the Pyrenees [Choukroune et al.,
city of 6.5 kms ', and have had gain balancing 1989] where the divergent reflections within the
applied. crust seem to mark tectonically imbricated crust
$0 REFLECTIVITY OF THE BALTIC SHIELD
s N
SP 2200 BABEL Line 6 1410
o
lO
20
lO KM
Fig. 2. Seismic reflection profile (brute stack) from BABEL
line 6 (SP 2200-1410). This section illustrates rapid lateral
variation in reflection and diffraction patterns within the
mid- to lower crust (6-15 s TWT). The Moho, assumed to occur
near 15 s TWT, is characterized by numerous diffractions and
no continous reflections. Reflections at upper left are caused
by reverberations within the water layer and a hard limestone
seafloor layer.
on opposite sides of a suture. The offset Moho clearest seismic evidence for Proterozoic plate
and crustal root zone have been interpreted as a tectonics to date, and support a uniformitarian
former subduction zone. The similarity to reflec- view of tectonic processes.
tivity patterns of Phanerozoic collisional belts,
together with conductivity anomalies, leads us to (4) Stronq subhorizontal reflections in the upper
suggest that the reflectivity pattern imaged on crust
BABEL lines 2 and 4 represents an Early Protero-
zoic suture zone in the Baltic Shield [BABEL Wor- Figure 6 illustrates the northern part of
king Group, 1990]. Our new data provide the BABEL line 1 in the northern Bothnian Sea and
E
SP 1700
BABEL Line 7
w
2350
lO
..
. .
.

.... .
20
lO KM
Fig. 3. Seismic reflection profile (brute stack) from BABEL
line 7 (SP 1700-2350) in the southeastern Bothnian Sea. Here
the lower crust (10-15 s TWT) exhibits numerous subhorizontal
reflections whereas the mid- and upper crust have isolated
dipping reflections or diffractions resembling the reflection
character seen throughout much of Western Europe.
shows strong subhorizontal reflections at 3-4.3 s data [Juhlin and Pedersen, 1987]. Drilling and
TWT and at 12-15 s TWT. The midcrust lacks promi- vertical seismic profiling found that the reflec-
nent reflections and the Moho is poorly defined. tors in the Siljan area are associated with dole-
We tentatively attribute the upper crustal re- rite sills intruded into granite. Their thicknes-
flections to dolerite sills, because of nearby ses range from a few meters up to 60 m and they
outcrop of 1.2-1.25 Ga sills in Sweden and Fin- extend over an area of at least 800 km 2 [Juhlin,
land [Gorbatschev et al., 1979], and because they 1990].
look similar to reflection data in the Siljan No upper crustal reflections can be observed
area (Sweden) about 350 km southwest of the BABEL on BABEL line 6 which lies parallel to the west
S N
SP 6720 BABEL Line B 6220
lO
20
lO KM
Fig. 4. Seismic reflection profile from BABEL on this section, although one northward-dipping
line B (SP 6720-6220) off the island of 6tand in reflective zone cuts across the crust from 8 s TWT
the Baltic Sea proper (f-k migrated at 2.0 kms ' to in the southwest to 12 s in the NE.
remove water-born energy). Fewer reflections occur
of BABEL line 1. Thus, the dolerite sheets do not (5) Variability of reflection character
extend across the whole Bothnian Sea as had been
suggested previously [Gorbatschev et at., 1979]. In order to present a larger view of the va-
Similar upper crustal reflections were imaged riabte reflection character, Figure 7 shows a
by COCORP in Arizona [Hauser et at., 1987] and line drawing of profile BABEL A. The line traver-
also observed in the Amadeus Basin in Australia ses from the North German Basin in the SW over
[Wright et al., 1990]. the Ringkbing-Fyn-High, crosses the Tornquist
s
SP 8900
BABEL Line 2
N
6100
lO
lO KM
Fig. 5. Seismic reflection profile from BABEL line 2 in
the Gulf of Bothnia (f-k migrated at 6.5 km '). A bright
packet of reflections is illustrated dipping to the north
at 15-19 s TWT coincident with the Moho determined from
refraction data and interpreted as representing a Prote-
rozoic suture [BABEL Working Group, 1990].
Zone near Bornholm island and continues to the beneath the platform margin in the north. The
Fennoscandian Platform south of 61and island. same variation from platform to basin reflec-
The line exhibits various distinctive reflectivity patterns is seen in the southern North Sea
tivity patterns with a striking similarity to the [Blundell et al., 1991]. The deeper crust within
variation observed on lines BIRPS MOBIL 6A and 7 the Tornquist Zone is not clearly imaged, pos-
in the southern North Sea between England and sibly due to near-surface complexity. For details
France [Blundell et al., 1991]: On this profile a see [BABEL Working Group, 1991b].
highly reflective lower crust, ending at a re-
flection Moho at about 11 s TWT, and a transpa- Discussion and Conclusions
rent upper crust characterize the section beneath
the basin area in the south. A markedly diffrac- The BABEL data exhibit highly variable reflec-
tion-dominated region with no clear Moho occurs tivity patterns. Low reflectivity with no sharp
S N
SP 4270 BABEL Line 1 58o
lO
2o
lO KM
Fig. 6. Seismic reflection profile from BABEL line 1
(SP 4270-5810) in the northern Bothnian Sea (f-k mi-
grated at 2.0 kms 4 to remove water-born energy). This
section shows strong subhorizontal reflections at 3-4 s
TWT and at 12-15 s TWT; the shallow ones are tentively
attributed to dolerite sills.
reflection Moho contrasts with reflective lower tions in the deeper crust [Lee and Hutchinson,
crust with bands of reflections and prominent 1990] and weak Moho reflections are also reported
Moho reflections as well as with diffraction- from eratons in Australia [Finlayson, 1982] and
dominated segments. South Africa [Durrheim, 1987]. The sections, how-
Nearly all profiles in cratonic areas, acqui- ever, show more reflectivity in the upper than in
red mostly on land so far, are characterized by the lower crust.
decreasing reflectivity and diffractions [e.g., Assuming the difference in reflectivity be-
Brown et al., 1983; Gibbs, 1986]. LITHOPROBE re- tween the Baltic Shield and other cratonic areas
flection lines on the Canadian Shield, paris real, i.e. not just caused by the acquisition
ticularly those recorded in the Great Lakes parameters, and assuming that the reflectors are
[Green et al., 1989], occasionally show reflec- old, it follows that the cratonic areas of Fen-
SW NE
I RingkCbing- [ i Tornqulst Zone I
Fyn-High
Fennoscandian Platform'
300 km 200 km 100 km 0 km
o
I-':.%'---=_--_-'r-_z-
5 -, - - ' - --'-" - "--"' ' ...... -':<:':- ' -' - '--'---"- - -'?:'--'"'--:;':;-'
I - - -- ' -- - :- '-- . ..... -: - - ........ :-;';'r---'- .... -'=;--. --.2: --'-
...... =- ........... :- ..... -?_-'-'?4=- .g.-,---..- .-.. --- ."
/ , . '-- - . --, .... .---.:.
/ _
' '- - ? --- .-'--- - - "" - - "" - -.. ':'-- ' "x ". ....
- 15 ,.
' BABEL Line A , ' ' .... '
B0'rnhlm
Fig. 7. Line draing of BABDT. :profiLe A shoing a larger section of variable reflecton
character. The Line extende from the CaLedonian German-Danish Basin in the SW to the
ennoecandan Platform in the bid and roeeee the Tornquet ?,one (T,) near ]Bornholm
ieland. ote the bande of reflection in the loer cruet eouth of the TZ arid the df-
fraction-dominated cruet north of the T,.
noscandia and North America must have formed Stockholm Contributions in Geoloqy, 3--6, 151-213,
under different conditions. Various cratons may 1981.
have undergone a very different evolution and an BABEL Working Group, Evidence for Early Protero-
identical structure or reflectivity should not zoic plate tectonics from seismic reflection
necessarily be expected. Another possibility is profiles in the Baltic Shield, Nature, 348, 34-
that there might be a selective process to 38, 1990.
destroy reflections. One possible process is an- BABEL Working Group, Recording marine airgun shots
orogenic magmatism, which is particularly exten- at offsets between 300 and 700 km, Geophys. Res.
sive in North America [Emslie, 1978; Lindh and Lett., 1--8, in press, 1991a.
Gorbatschev, 1984], but is also widespread in the BABEL Working Group, Deep seismic survey images
Baltic Shield [Gorbatschev et al., 1987]. Pre- crustal structure of Tornquist zone beneath
existing reflections might be destroyed by re- southern Baltic Sea, Geophs. Res. Lett., 1--8, in
peated dike injection, resulting in diffractions press, 1991b.
[Nelson, 1991]. BIRPS and ECORS, Deep seismic reflection profiling
If the reflectors have formed in Phanerozoic between England, France and Ireland, J. qeol.
times, it is not yet clear why the Baltic Shield $oc. Lond., 143, 45-52, 1986.
and the North American craton show differences in Blundell, D.J., Hobbs, R.W., Klemperer, $.L.,
reflectivity. Scott-Robinson, R., Long, R.E., West, T.E., and
As we have seen from the BABEL data the con- E. Duin, Crustal structure beneath the central
ventional concept of reflectivity in Precambrian and southern North Sea from BIRPS deep seismic
shield and platform areas may need to be revised. reflection profiling, J. qeol. Soc. London,
Additional high-quality data from other Precam- 148,445-458, 1991.
brian cratonic areas are necessary to contribute Brown, L.D., $erpa, L., $etzer, T., Oliver, J.,
to the discussion about reflectivity and its ori- Kaufman, S., Lillie, R., Steiner, D., and D.W.
gin. Steeples, Intracrustal complexity in the United
States mid-continent: preliminary results from
Acknowledqments. Principal funding for Pro- COCORP surveys in northeastern Kansas, Geoloq,
ject BABEL was from NERC (U.K.) and from the Com- 1__1, 25-30, 1983.
mission of the EC within the framework of the Re- Choukroune, P., and ECOR$ Team, The ECOR$ Pyrenean
search and Development Project JOULE. Additional deep seismic profile reflection data and the
funds came from BIRP$ Industrial Associates; the overall structure of an orogenic belt, Tecto-
Swedish Natural Sciences Res. Council; the Aca- nics, , 23-39, 1989.
demy of Finland; the DFG (Germany); and the Da- Durrheim, R.J., Seismic reflection and refraction
nish Natural Science Res. Council. The marine studies of the deep structure of the Agulhas
seismic data were recorded and are being pro- Bank, Geophs. J. R. astr. Soc., 8--9, 395-398,
cessed by Prakla-Seismos A.G. and will become 1987.
available for the cost of reproduction from the Emslie, R.F., Anorthosite massifs, Rapakivi gra-
British Geological Survey, Edinburgh EH9 3LA, nites, and Late Proterozoic rifting of North
U.K. America, Precambrian Res., !, 61-98, 1978.
Finlayson, D., Seismic crustal structure of the
References Proterozoic North Australian Craton beneath
Tennant Creek and Mount Isa, J. Geophs. Res.,
Axberg, S., Seismic stratigraphy and bedrock geo- 8--7, 10569-10578, 1982.
logy of the Bothnian Sea, northern Baltic, Gal, G., and R. Gorbatschev, An outline of the
Precambrian evolution of the Baltic Shield, processing techniques for marine, crustal-re-
Precambrian Res., 3--5, 15-52, 1987. flection seismic data, U.S. Geol. Surv. Bull.
Gibbs, A.K., Seismic reflection profiles of Pre- 1987, 22, 1990.
cambrian crust: a qualitative assessment, in Lindh, A., and R. Gorbatschev, Chemical variation
Reflection Seismoloq: The Continental Crust,
edited by M. Barazangi and L. Brown, Geodyn.
S9r., vol. 14, American Geophysical Union, 95-
106, 1986.
Gorbatschev, R., Solyom, Z., and I. Johansson, The
central Scandinavian dolerite group in J'mtland,
central Sweden, GPP, 101, 177-190, 1979.
in a Proterozoic suite of granitoids extending
across a mobile belt - craton boundary, Geol.
Rundschau, 73, 881-893, 1984.
Lschen, E. et al., Seismic reconnaissance studies
of the continental deep drilling KTB-location
Black Forest, SW Germany, J. Geophs., 61, 1-10,
1987.
Gorbatschev, R., Lindh, A., Solyom, Z., Laitakari, Marker, M., Tectonic interpretation and new crus-
I., Aro, K., Lobach-Zhuchenko, $.B., Markov,
M.S., Ivliev, A.I., and I. Bryhni, Mafic dyke
swarms of the Baltic Shield, in Mafic Dyke
Swarms, edited by H.C. Halls and W.F. Fahrig,
Geol. Assoc. Canada, Spec. Paper 34, 361-372,
1987.
Green, A.G., Cannon, W.F., Milkereit, B., Hut-
chinson, D.R., Davidson, A., Behrendt, J.C.,
Spencer, C., Lee, M.W., Morel--l'Huissier, and
W.F. Agena, A "GLIMPCE" of the deep crust be-
tal modelling along the Polar Profile, Northern
Baltic Shield, in Proceedinqs of the sixth work-
shop on the European Geotraverse (EGT) Project,
edited by R. Freeman and St. Mueller, 9-21,
European Science Foundation, Strasbourg, 1989.
Matthews, D.H., Seismic reflections from the lower
crust around Great Britain, in The Nature of the
Lower Continental Crust, edited by J.B. Dawson,
D.A. Carswell, J. Hall and K.H. Wedepohl, Black-
well, Oxford, 11-21, 1986.
neath the Great Lakes, in Properties and Pro- Meissner, R., The continental crust - A qeophysi-
cesses of Earth's Lower Crust, edited by R.F. cal approach, Academic Press, London, 426 pp,
Mereu, St. Mueller and D.M. Fountain, Geophs. 1986.
Monograph 51, American Geophysical Union, 65-80, Nelson, K.D., A unified view of craton evolution
1989. motivated by recent deep seismic reflection and
Hauser, E.C., Gephart, J., Latham, T., Oliver, J., refraction results, Geophys. J. Int., 105, in
Kaufman, S., Brown, L., and I. Lucchitta, COCORP press, 1991.
Arizona transect: strong crustal reflections and Sadowiak, P., Brown, L.D., and R. Meissner, Seis-
offset Moho beneath the transition zone, Geolo- mic reflectivity patterns in Europe and North
LY, 1__5, 1103-1106, 1987. America, in this volume, 1991.
Juhlin, C., Interpretation on the reflections in Wever, Th., Trappe, H., and R. Meissner, Possible
the Siljan area based on the results from the
Gravberg-1 borehole, Tectonophysics, 173, 345-
360, 1990.
Juhlin, C. and L.B. Pedersen, Reflection seismic
investigations of the Siljan impact structure,
Sweden, J. Geophys. Res., 92, 14113-14122, 1987.
Lee, M.W., and D.R. Hutchinson, True-amplitude
relations between crustal reflectivity, crustal
age, heat flow and viscosity, Annales Geophysi-
cae, 5B, 255-266, 1987.
Wright, C., Goleby, B.R., Collins, C.D.N., Korsch,
R.J., Barton, T., Greenhalgh, , $.A., and $.
Sugiharto, Deep seismic profiling in central Au-
stralia, Tectonophsics, 173, 247-256, 1990.
REFLECTIVITY VARIATIONS OF VARISCAN TERRANES IN GERMANY
R. Bittner and Th.' Wever
Institut fur Geophysik, Universitt Kiel, Leibnizstr. 15, D-2300 Kiel.
Abstract. Interpretations of the Variscan Later this zonation has been extended to western
orogenic domain involve a plate tectonic evolu- Europe.
tion. They require inter-province oceans, but no Besides geological mapping, geophysical inves-
ophiolites or other relics have been found so tigations in Germany confirmed different charac-
far. The widths of the oceans are also debated, teristics for the various subprovinces. From
and paleomagnetic data do not support any parti- ample refraction/wide-angle profiling in Germany
cular interpretation. crustal thickness is well constrained. The aver-
It is one of the main objectives of the deep age thickness is about 28-30 km with only minor
reflection seismic projects in Germany to study variations. The only exceptions are the Rhinegra-
crustal fine structure and decipher the evolution ben and its immediate neighbourhood ( 25 km)
of the Variscides. Profiles now total more than and the western flank of the Bohemian Massif
2,400 km in length. Line DEKORP 2 (470 km) tra- ( > 35 km).
verses all Variscan provinces, and exhibits dis- The Conrad discontinuity has been found almost
tinctive terrane-related properties, both in everywhere in Germany. But in contrast to the
terms of structure and distribution of reflec- Moho, a considerable variability both in refrac-
tions. tion velocity and depth had been already noted by
We apply the method of automatic energy histo- the 1960s [German Res. Group for Explosion Seis-
grams (based on final stacks in digital format on mology, 1964]. The refraction velocity was found
magnetic tape) to more clearly extract informa- to vary between 6.4 and 7.0 km/s, the depth bet-
tion about reflectivity from the data in an un- ween 10 and 21 km. More recent data show more
biased manner. The results yield convincing evi- details, but in general, the heterogeneity was
dence for the existence of different terranes already manifest. For the Rhenohercynian and
welded together during the various tectonic com- Saxothuringian, Giese [1983] reports strong in-
pressions. tracrustal discontinuities which are underlain by
Introduction
Already by 1927 Kossmat had reported a tecto-
nic zonation of the Variscides of Central Europe.
He defined the provinces on the basis of similar
evolution of magmatism, tectonics, and facies. He
mapped different characteristics within each of
the provinces. Starting in the south, where the
low-velocity zones suggesting a tectonic inter-
pretation of thrusts.
Potential field data indicate a relatively
simple (homogeneous) crust in the Moldanubian,
whereas the Saxothuringian is characterized by a
great number of anomalies [Behr et al., 1984].
Also electrical anomalies are observed in the
Saxothuringian and Rhenohercynian.
In summary, not only geological and tectonic
data indicate fundamental differences between the
earliest deformation occurred, the Variscan pro- provinces, but their physical properties change
vinces are the Moldanubian, the Saxothuringian, so as to indicate a different construction in the
the Rhenohercynian, and the Subvariscan Foredeep. various subprovinces. The aim of this paper is to
examine the near-vertical deep reflection seismic
data for evidence to support the conclusions
derived earlier.
now at: Forschungsanstalt der Bw. fur Wasser-
schall- und Geophysik, Klausdorfer Weg 2-24,
D-2300 Kiel 14.
Energy-histograms as a tool for mapping
reflectivity
Geodynamics 22
The two major recent geoscientific projects
DEKORP (Deutsches K__ontinentales eflexionsseis-
misches rogramm) and KTB (ontinentales ief-
bohrprogramm der undesrepublik Deutschland) have
87
88 REFLECTIVITY VARIAnONS IN GERMANY
contributed much to our understanding of the
variscan orogen. However, also new questions have
arisen. Comparing the different seismic signatu-
res directly from the seismograms is a problem
which was encountered even in the earliest deep
seismic recordings in the 1950s. Dohr (1957)
introduced a statistical technique for evaluating
deep seismic reflections (reflectivity-his-
tograms). This work was extended by Liebscher
(1962) and Wever (1984]. In a first attempt Wever
et al. (1987) classified types of crustal reflec-
tivity distribution and prepared a preliminary
map for West Germany.
The old statistical method was very time-con-
suming and has frequently been criticized as
being subjective. In order to overcome both prob-
lems we introduced a simple method of producing
'energy-histograms' directly from the final stack
tapes (for details see Bittner and Wever, 1991].
This method sums the sauares of amplitudes of
seismic traces for defined intervals along a
profile with a window sliding along the time
axis. Tests of several window lengths have been
performed. The best for general interpretations
proved to be a 1 second window, with more de-
tailed investigations requiring a shorter window
length. Detailed examples are given by Bittner
(1989). For the profile segments here we chose a
length of 10 km. The resulting energy-histograms
are compatible with the reflectivity-histograms.
Reflectivity of the Variscan terranes
In this paper we apply the method outlined
above to the profile DEKORP 2, the only line
crossing all Variscan zones (Fig. 1). All other
profiles are considered in a forthcoming paper
(Wever and Bittner, in prep.). In order to com-
pare the energy-histograms with the corresponding
profile segments we combine the automatic line
drawings (ALD) of profiles DEKORP 2-S and DEKORP
2-N (Fig. 2).
DEKORP 2-S, the southern branch of DEKORP 2
(Fig. 1), was shot with explosives as the first
long-range deep reflection line in Germany in
1984 (DEKORP Res . Group, 1985J. DEKORP 2-N, the
northern branch, was observed in 1986 (Franke et
al., 1990J.
After tests had proved that the vibroseis
technique is useful for detecting lower crustal
reflections, this technique was used for DEKORP
2-N. However, a higher coverage was required to
yield the same data quality. Therefore, these
segments have different acquisition characteris-
tics. While DEKORP 2-S has a 25 fold coverage,
the southern/northern parts of DEKORP 2-N have an
average coverage of 137 and 64, respectively. The
eMP spacing of DEKORP 2-S is 40 m, 20 m for
DEKORP 2-N (for more details see Meissner and
Bortfeld, 1990J.
Data quality is generally very good on both
lines and many new structures could be observed.
For extracting the energy-histograms we used the
parameters mentioned above.
so 100 km
Fig. 1. Map showing the location of DEKORP 2
crossing the Variscan provinces in Germany.
1 = Basement of the Rhenohercynian, 2 = Basement
of the Saxothuringian, 3 = Basement of the Mol-
danubian, F = Frankfurt/Main, HH = Hamburg,
K =Koln, M - Munchen, R - Nordlinger Ries impact
structure, NVDF = North Variscan Deformation
Front.
The Moldanubian
On older seismic data (Liebscher, 1962; Angen-
heister and Pohl, 1976] the reflectivity of the
Moldanubian is clearly characterized by an almost
non-reflective upper crust and finger-type re-
flections, i.e., two or more strong peaks [Wever
et al., 1987 J in the lower crust. The same pat-
tern is displayed along DEKORP 2-S up to km 90.
The only exception is below the Ries impact cra-
ter area (km 15-32), where scattering of energy
in the upper few kilometers is too strong.
The Saxothuringian
The Saxothuringian reflectivity is in clear
contrast to that of the Moldanubian. An almost
unstructured crust is seen in the energy-histo-
grams. The upper crust shows many inclined re-
flections of unknown origin. From km 180 to the
northwest, reflectivity in the lower crust again
becomes dominant, partly due to lower crustal
diffractions. Further northwest, the picture be-
comes less clear again.
BITrNER AND WEVER 89
T
W
T
[]
10
D E K 0 R P 2 S DISTANCE [kml
240 220 200 180 160 140 120 100 80 60 40 Res 0
. i . .. i . ! . , .
.-,.. . .-" ;:";,' .' '.','' -,, ?.'.': ::.:]: - :. , -L" ..?- -] ,.:'c.:-.--"-"-.:::.::/,'-. ''""-"!- ',,-.:':' - ' :: '. ' ' -'"'i:'..-?'',-"i""','.::':'-' ,- t 0
:f?-::? ?-'.,-::::' .'-..-, ;*-.'.:-;-:s..:-%?-,,,?.-..;.,:.::=c--:-.T.:'-'..:,L'...':::_!?:-=.../.z'..-:-?.?.' - -.,. r?..'.-* .'- -: - .-:.:'.-..'-;K'-.";r. ' '
[S]o v
-] 5
1 10
DEKORP 2N DISTANCE
220 200 180 160 140 120 100 80 60 40 20 0 0
' - ' -'"' '- - -* - ' - ' ' :' *' ' .--*,_*'r..J"-.*..' -..'*-T-- ............... *-,'", - --**-- *---' .?,*,--'-- -' - ' .... -" -
............ , . .... . = . _-,.,...., ., .. .... * .. .....,_-.,-.,,.._ .... *?.,,, ..........
,:.*. .. ;.;-.. '- .,.-; 'L'.',:,. ;*.'.'., *.-::;.c?.: ::,:-;,;*...; .;.<.' .*:,:': - ' :' . -.:-,'",;;; ';-*..:L' . ': -- ].;. .;.:,*,.. i .-.,..>'.:-. -..,.:*.?,7:::.,,,,:,i'?;.,z...c::..,.*:_=:=; *;:].';:\*;--* .,.]'..'.
I .... ;?: .- '.::i.:".':'.:.'.:,i'-:.::;:'':.*;,%.'::'"?!'';.:.. ' :" .:"?;:!?'-:'::*;'.-'::'::--* _ . ' .
T
W
T
Is]
T
W
T
Is]
Fig. 2. Automatic line drawings of DEKORP 2 and energy-histograms.
1
The Rhenohercynian energy-maxima within the upper 4 s TWT. The crys-
talline crust is almost non-reflective. The ex-
The Rhenish Shield, comprising most of the planation may be of technical (poor energy pene-
Rhenohercynian investigated with the deep reflec- tration due to strong absorption within the sedi-
tion seismic method in Germany, is rather well ments) or geological nature (lack of impedance
known for its 'strong and distinct reflectivity. contrasts). The high variability of lower crustal
Already in 1968 single-coverage data [Glocke and reflectivity in the Northwest German Basin has
Meissner, 1976] revealed very strong reflections already been pointed out by Trappe [1989].
below its southern part. DEKORP 2-N shows a simi-
lar character: the lower crust exhibits very Summary and Conclusions
strong and distinct reflections (resulting in
seven peaks of reflected energy below 5 s TWT in Based on seismograms and energy-histograms it
the southernmost histogram). By comparison with seems justified to conclude that the near-verti-
the seismogram, it becomes clear that up to km cal seismic reflection method can distinguish be-
140 most upper crustal peaks in the histogram tween different terranes: all Variscan subpro-
represent bands, of dipping reflections which vinces show characteristics which differ strong-
correlate well with thrusts known from outcrops. ly. The energy-histograms reveal a clear internal
Lower crustal reflectivity is also concentrated division of the Moldanubian crust, which stands
in bands resulting in peaks of the histogram. The in strong contrast to the $axothuringian with its
'fingers' in the histogram differ from those of higher heterogeneity, which is also evident in
the Moldanubian in lateral continuity and number. gravity and magnetic fields. The Rhenohercynian,
with its remarkable number of thrusts and narrow
bands of lower crustal reflections, results in an
The $ub-Variscan Foredeep even different kind of energy-histograms.
It seems unlikely that the contrasting tecto-
The sub-Variscan Foredeep is characterized by nic styles observed on the DEKORP data could have
its thick sedimentary strata which produce strong been produced in close proximity. We thus can add
90 REFLECTIVITY VARIATIONS IN GERMANY
more arguments for former separation and indepen- Franke, W., and 15 others, Crustal structure of
dent evolution of the Variscan terranes. the Rhenish Massif: results of deep seismic
reflection lines DEKORP 2-North and 2-North-Q,
Geol Rundschau, 79, 523-566, 1990.
Acknowledqements. We acknowledge the conti- German Res. Group for Explosion $eismology,
nuous support of the project DEKORP by the Bun- Crustal structure in Western Germany, Z. Geo-
desministerium fr Forschung und Technologie phys., 3--0, 209-234, 1964.
(BMFT) under grant RG 8703 6. Discussions with Giese, P., The evolution of the Hercynian crust
many colleagues, esp. with Dr. Henning Trappe
(BEB, Hannover) helped to develop our ideas.
Publication No. 420 from Institut fr Geophysik,
Universitt Kiel.
References
Angenheister, G. and J. Pohl, Results of seismic
investigations in the Ries Crater area (South
Germany), in Explosion Seismoloqy in Central
- some implications to the uplift problem of
the Rhenish Massif, in Plateau Uplift, edited
by K. Fuchs, K. v. Gehlen, H. Mlzer, H.
Murawski, and A. Semmel, pp. 303-314, Sprin-
ger, Heidelberg, 1983.
Glocke, A. and R. Meissner, Near-vertical
reflections recorded at the wide-angle profile
in the Rhenish Massif, in Explosion Seismoloqy
in Central Europe, edited by P. Giese, C.
Prodehl, and A. Stein, pp. 252-256, Springer,
Heidelberg, 1976.
Europe, edited by P. Giese, C. Prodehl and A. Kossmat H. (1927). Gliederung des varistischen
Stein, pp. 290-302, Springer, Heidelberg, Gebirgsbaues, Abh. Schs. Geol. Landesamtes,
1976. , 39 pp.
Behr, H., W. Engel. W. Franke, P. Giese, and K. Liebscher, H.-J., Reflexionshorizonte der tiefe-
Weber, The Variscan belt in Central Europe: ren Erdkruste im bayerischen Alpenvorland, ab-
main structures, geodynamic implications, open geleitet aus Ergebnissen der Reflexionsseis-
questions, Tectonophysics, 109, 15-40, 1984. mik, Z. Geophys., 2--8, 162-184, 1962.
Bittner, R. $ignal/Noise-Verbesserung eng vermes- Meissner, R. and R. Bortfeld, (Eds.), The DEKORP
sener seismischer Daten unter Verwendung eines Atlas, 90 pp. and 80 seismic sections, $prin-
Energie-Mehrspurfilters, Doctor Thesis, 137 ger, Heidelberg, 1990.
pp., University of Kiel, 1989. Trappe, H., Deep seismic profiling in the North
Bittner R., and Th. Wever, Energy-histograms for German Basin, First Break, !, 173-184, 1989.
the characterization of deep reflection seis- Wever, Th., Hufigkeitsstatistik yon $teilwin-
mic profiles, Geophys. J. Int., 105, (1991, in kelreflexionen aus Gebieten der kontinentalen
press). Kruste und ihre Anwendung auf die Analyse yon
DEKORP Research Group, First results and prelimi- Krustenstrukturen, Diploma Thesis, 163 pp.,
nary interpretation of deep reflection seismic University Kiel, 1984.
recordings along profile DEKORP 2-$, J. Geo- Wever, Th., H. Trappe, and R. Meissner, Possible
phys., 5--7, 137-163, 1985. relations between crustal reflectivity, crus-
Dohr, G., Ein Beitrag der Reflexionsseismik zur tal age, heat flow and the viscosity of the
Erforschung des tieferen Untergrundes, Geol. continents, Ann. Geophysicae, 5__B, 255-266,
Rundschau, 4--6, 17-26, 1957. 1987.
STRUCTURE OF THE SAAR-NAHE-BASIN (SW-GERMANY)
FROM DEKORP PROFILES 1-C AND 9-N
Andreas Henk
Institut fuer Geologie, Pleicherwall 1, 8700 Wuerzburg, Germany
Abstract. Two deep seismic reflection profiles have
been acquired by the DEKORP program across the
northeastern part of the Late Palaeozoic Saar-Nahe-
Basin (SW-Germany). The geological interpretation
reveals a halfgraben structure bounded to the NW by a
fault zone with complex movement history consisting of a
subvertical fault and a southeast-dipping normal fault.
The latter is a detachment fault which delimits the
Permo-Carboniferous basin fill. On the basis of
DEKORP profiles 1-C and 9-N a model for the evolution
of the Saar-Nahe-Basin is developed.
Introduction
The Saar-Nahe-Basin in SW-Germany is a NE-SW-
trending Permo-Carboniferous basin within the Variscan
fold belt. Situated just southeast of the Hunsriick Slate
Belt, it forms part of a much larger structure extending
over 300 km from Lorraine in NE-France to the Wetterau
region north of Frankfurt/M.. The Saar-Nahe-Basin was
initiated most likely in Early Westphalian times, and
deposition of entirely continental sediments partly inter-
bedded with volcanics continued to the end of the Rotlie-
gend. Pre-rift sediments of Middle Devonian through
Lower Carboniferous age are underlain by crystalline
basement. The present outcrop of the basin are mainly
synrift strata. Only in the SE are sediments preserved,
which are thought to have been deposited during the
thermal subsidence phase. The basin is bounded to the
NW by the Hunsrack Boundary Fault which also separa-
tes two of the tectonostratigraphic units of the Variscan
fold belt in Central Europe, the Rhenohercynian and the
Saxothuringian zones [Franke, 1989]. This fault zone had
already been investigated by a short deep reflection profi-
le [Meissner et al., 1980]. Other major elements of the
basin structure paralleling the Variscan strike are the
Nahe Syncline, the Pfalz Anticline and the Pfalz Syncline
(Fig. 1).
Geodynamics 22
91
In 1988 DEKORP acquired two deep seismic reflection
profiles with a record length of 16 s Two Way Travel
Time (TWT) covering the northeastern part of the Saar-
Nahe-Basin. The complete data sets are available in the
DEKORP atlas [Meissner and Bortfeld, 1990]. DEKORP
1-C crosses the Hunsriick Slate Belt and its Southern
Boundary Fault as well as part of the Saar-Nahe-Basin
perpendicular to their Variscan strike and ends just south
of the Pfalz Anticline [DEKORP Research Group, 1991].
DEKORP 9-N crosses the northern Rhinegraben and
covers the eastern parts of the Saar-Nahe-Basin [Wenzel
et al., 1991].
The sections were processed using a FD-(finite diffe-
rence) migration. Additional application of a FK-migrati-
on (frequency-wavenumber) for the Nahe-Syncline in the
location range of 1460 to 1760 of DEKORP 1-C improved
the quality in this part of the section (Fig 2.) [DEKORP
Research Group, 1991].
Parts of the DEKORP lines 1-C and 9-N, which cover
the Saar-Nahe-Basin, are considered here for geological
interpretation. As the paper concentrates on the structu-
ral style of the basin and the stratigraphic subdivision of
the basin fill only the upper 7 s TWT are considered. The
deep crustal structure deduced from DEKORP 1-C and
9-N and its implication for the evolution of the Variscan
orogen are discussed in more detail in DEKORP Rese-
arch Group [1991] and Wenzel et al. [1991].
Geological interpretation based on DEKORP 1-C and 9-N
The geological interpretation based on a manual line
drawing (Fig. 3.) of time sections of DEKORP 1-C and 9-
N is shown in Fig. 4. The crust beneath the Saar-Nahe-
Basin is characterized by a segmentation into three dis-
tinctively different reflectivity patterns. The uppermost
strongly reflective part representing the Palaeozoic basin
fill is undedain by a nearly transparent wedge. The base
of the wedge is characterized by several arched reflector
bands above an up to 2 s TWT thick, highly reflective
zone. At the intersection of 1-C and 9-N its true strike
and SE dip can be determined. In general, the reflections
of the 1-C part of the section are stronger and more
continious compared to DEKORP 9-N.
92 STRUCTURE OF THE SAAR-NAHE-BASIN (SW-GERMANY)
Fig. 1. Map of the northeastern Saar-Nahe-Basin showing the main geological features and the
location of seismic lines DEKORP 1-C and 9-N.
NW DEK 1-C SE w DEK 9N E
Nahe Syncline Pfalz Anticline Pfalz Syncline
Monzingen I 0dernheim 1
(6 km N) (10 km
Locations lO00 1400 . 1200 3000 2800 2600 Locations
, : ...... .... ......... . ........ . t., ..... ...;:....._....,....,:...... ..... ,... L
Fig. 2. Migmt sfions of DEKO 1-C d 9-N.
The Saar 1 borehole located about 60 km SW of the
DEKORP profiles can be used as reference for the deep
structure of the Saar-Nahe-Basin. The borehole penetra-
tes about 4600 m of syn-rift and 1000 m of pre-rift sedi-
ments. The crystalline basement is reached at a depth of
5662 m [Kneuper, 1976]. Projection of the Saar 1 boreho-
le suggests that the transparent wedge in the seismic
section is the pre-Variscan basement of the Saar-Nahe-
NW
Locations
TWT (s)
HENK
DEK 1-C s. w DEK 9N
Nahe Syncline Pfalz Anticline Pfalz Syncline
Monzingen I Odernheim 1
(a k,, st) (Io k,,
1600 140 1200 3000 2800 2600
i i i
- :-, _- - . ._
Locations
TWT (s)
1
2
Fig. 3. Manual line drawing for DEKORP 1-C and 9-N.
93
NW
Locations
TWT (s)
1
DEK 1-C
Nahe Syncline
Monzingen I Odernheim 1
(6 km NE) (10 km N)
ss DEK 9N
Pfalz Anticline Pfalz Syncline
A 1700 14yO 1200 3000 2800 2600
_ _ _
. . D'nne'sbeJ
I I I I I I I I I I I t 211.[11kam ,, ' .,re,i.., i: :.,:: , - _
+ + + +
+ + + + + + +
+ + + + + + +
+ + + + +
Fault
0 10 km
i
Locations
TWT (s)
Tertiary Stephanian Devonian
':'.:'":...': Upper Rotliegend Westphalian crystalline basement
Lower Rotliegend M'Devonian - L'Carboniferous rhyolite
Fig. 4. Geological interpretation for DEKORP 1-C and 9-N.
Basin thus forming part of the Mid-German Crystalline
Rise at the northern rim of the Saxothuringian zone. Its
poorly reflective appearance could be due to abundant
granitic intrusions as well as to an extensive reequilibra-
tion of early Variscan lithologic and structural fabric. The
top of the basement is marked by the change from the
upper strongly reflective part to the transparent wedge. It
is located at 3.3 s TWT (6.7 km) close to the Hunsflick
Boundary Fault and rises to about 1.2 s TWT (1.9 km) in
the eastern part of the basin. Again, by comparison with
94 STRUCTURE OF THE SAAR-NAHE-BASIN (SW-GERMANY)
the $aar 1 borehole the crystalline basement can be
expected to be overlain by a unit of maximal 0.4 s TWT
(1 km) thickness consisting of non-metamorphic Middle
Devonian to Lower Carbonifeous rocks, mainly carbona-
tes [Kneuper, 1976].
The base of the Permo-Carboniferous fill of the $aar-
Nahe-Basin is interpreted as the lowest reflector of basin-
wide extent unconformably overlying the pre-rift-
sediments and crystalline basement respectively. As the
basin was initiated possibly in Late Namurian, more likely
Early Westphalian times, a time gap of 10 to 20 Ma exists
to the underlying strata. The frequent strong reflections in
the oldest part of the basin filling are related to conglo-
meratic beds and coal seams. The preserved Westphalian
strata show a progressive thickness reduction from
0.9 s TWT (2.2 km) at the Hunsrtick Boundary Fault to
zero in the Pfalz Syncline.
The thickness variation is at least in part due to erosion
prior to the $tephanian. Its base, the Holz Conglomerate,
corresponds to a pronounced reflector band located at 2.3
s TWT (4 km) in the Nahe Syncline and at 1 s TWT (1.5
km) in the Pfalz Anticline. The angular unconformity of
the Holz Conglomerate to the underlying strata is clearly
visible particularly in the anticlinal parts of the profile.
Towards the E, the Stephanian truncates progressively
older rocks until at 9-N location 2850 Stephanian directly
rests on crystalline basement. In the Nahe Syncline the
Westphalian-Stephanian-transition seems to be confor-
mable, but a hiatus could still exist.
The $tephanian and the overlying Lower and Upper
Rotliegend have a more uniform appearance in the seis-
mic section compared to the Westphalian. Stronger re-
flections may be due to conglomeratic horizons and basal-
tic sills. The boreholes Monzingen 1 and Odernheim 1
[Habicht, 1966], located 6 km and 10 km respectively NE
of DEKORP I-C, as well as surface outcrop were used to
correlate the upper part of the intra-basin reflections with
stratigraphic boundaries.
Between locations 2700 and 2900, DEKORP 9-N runs
in between two rhyolite outcrops (Fig. 1). The seismic
section shows in this part between 0.5 and 2.5 s TWT a
large number of short reflectors bounded by a transparent
zone. This is interpreted as an indication of an huge
laccolith which probably connects the surface exposures
at depth. Apparently, the rhyolite forms no compact body
as some stratigraphic reflectors are still visible (see also
Haneke, 1987).
The total thickness of the Permo-Carboniferous fill of
the Saar-Nahe-Basin varies between 1 and 3 TWT. A
depth-conversion using sonic logs and published seismic
interval velocities reveals a maximum thickness of the
preserved basin fill of 6.5 km.
The strongly reflective part beneath the crystalline
wedge can be interpreted in various ways. This zone may
represent a southeast-dipping Variscan thrust with an
anastomosing pattern of shear zones and associated
duplex structures. Alternatively, it can be interpreted as
relicts of the sedimentary cover of the former "Giessen
Ocean" sheared off from the oceanic crust and stacked
between the continental crust of the Saxothuringian and
Rhenohercynian zones [DEKORP Research Group,
1991]. An interpretation as magmatic sills of Permian age
may also be posssible.
Northwest of the Hunsrtick Boundary fault, which is
intersected by DEKORP 1-C at Loc. 1730, the crustal
reflectivity is distinctly different from the parts of the
sections described so far. The upper crust down to 5 s
TWT shows only a weak reflectivity, which is likely to be
a result of the very low to low grade Palaeozoic rocks as
well as the steeply dipping structural grain of the Huns-
rack Slate Belt. In contrast, the lower crust is again more
reflective.
The Hunsrack Boundary Fault correlates in part with a
subvertical zone separating the two regions of different
crustal reflectivity. It can be traced to depths of at least 8
s TWT truncating the strong reflections beneath the
crystalline wedge. For the formation of the $aar-Nahe-
Basin this steep fault was only of very minor importance,
because the reflections of the basin fill are delimited by a
major normal fault. This detachment is located at 3.3 s
beneath the Nahe Syncline and soles out in a midcrustal
level at about 6 s TWT. Its shape is listric in the time
section. Between locations 1600 and 1700 of DEKORP l-
C, the exact localisation of the detachment, particularly in
the final stack (see Meissner and Bortfeld, 1990), is
complicated by a number of reflectors, which are related
to the dip of the strata in the Nahe Syncline. Apparently,
the listric fault coincides only in its upper steep part with
the subvertical branch of the Hunsrack Boundary Fault
known from surface exposure. Below 2 s TWT none of
the intra-basin reflectors can be traced to the subvertical
fault branch as one would expect for a fault being active
as a normal or strike-slip fault during Westphalian,
Stephanian and Lower Permian times. The strong gene-
tic relation of the Saar-Nahe-Basin to the listric fault can
also be seen on unpublished industrial seismic lines
running parallel to the DEKORP 1-C profile.
Discussion and Conclusions
The interpretation of DEKORP 1-C and 9-N shows a
classical halfgraben structure for the Saar-Nahe-Basin.
The halfgraben is bounded to the NW by two major fault
zones which coincide only in their upper part. The subver-
tical branch of the Hunsrack Boundary Fault extends into
the lower crust and possibly even down to the Moho. The
second fault zone has a listtic shape in the time section
and detaches at midcrustal levels of about 16 km. All
intra-basin reflectors are delimited by this fault thus
indicating the prevailing structural control of the Saar-
Nahe-Basin by a detachment and not a subvertical fault.
The thickness distribution of the Permo-Carboniferous
strata proves the synsedimentary activity of the Hunsrack
Boundary Fault as well as the synsedimentary formation
of the major synclinal and anticlinal structures.
Most likely, the Saar-Nahe-Basin formed within the
collapsed hanging wall of a major reactivated Variscan
thrust. Along this low angle fault the Mid-German Cry-
stalline Rise was overthrusted on the Hunsrack Slate Belt
during the Vistan and Namurian [Franke, 1989]. Such a
reactivation of thrusts as low angle normal faults during
the extensional collapse of the Variscan orogen had
already been suggested by Behr et al. [1984] and Eis-
bacher et al. [ 1989].
Prior to the basin formation, most likely during a short
time span between about 325 and 315 Ma, dextral strike
slip movement occured on the Hunsrack Boundary Fault.
In this tectonic phase the subvertical fault branch was
active truncating the strong reflectors beneath the crystal-
line wedge. Then the Saar-Nahe-Basin formed by oblique
extension on the detachment fault. The synsedimentary
fault pattern and the orientation of intrusive rocks indica-
te a minimum principal stress oriented WSW-ENE. This
is in close agreement with the regional stress-field dedu-
ced by Arthaud and Matte [1977] from the Late Palaeozo-
ic dextral mega shear zone between the Appalachians and
the ldral.
The Hunsrack Boundary Fault was reactivated in Late
Mesozoic-Cenozoic times in conjunction with the Alpine
collision and the Rhine graben rifting [Schwab, 1987].
Past models which interpreted the Saar-Nahe-Basin as a
symmetrical graben or being underlain by a Permian
batholith can be ruled out. Additionally, no evidence for
the Saar-Nahe-Basin being a typical pull-apart structure
can be derived from DEKORP seismic lines 1-C and 9-N.
Acknowledgements. The author gratefully acknowledges
the DEKORP Research Group, in particular R. Meissner
and C. Reichert, for providing the seismic data. Reviews
by V. Lorenz, O. Oncken, R. Meissner and an anonymous
reviewer helped to improve the manuscript.
References
Arthaud, F. and Matte, P., Late Palaeozoic strike-slip
faulting in southern Europe and North Africa: Results of
HENK 95
a right-lateral shear zone between the Appalachians and
the Urals, Geol. Soc. Am. Bull., 88, 1305-1320, 1977.
DEKORP Research Group, Results of DEKORP 1 /
BELCORP deep seismic reflection studies in the west-
ern part of the Rhenish Massif along profile DEKORP
1, Geophys. J_ Int., in press, 1991.
Franke, W., Tectonostratigraphic units in the Variscan
belt of Central Europe, Geol. Soc. Amer..Spec. P, 230,
67-90, 1989.
Habicht, H., Die permo-karbonischen Aufschlullbohrun-
gen der Nahe-Senke, des Mainzer Beckens und der
Zweibracker Mulde, Z_. deutsch. geol. Ges., 115, 631-
649, 1966.
Haneke, J., Zur Genese und stratigraphisch-tektonischen
Stellung eines permokarbonen Rhyolith-Domes im Saar-
Nahe-Gebiet (SW-Deutschland). - Pollichia-Buch, 10,
147 pp., 1987.
Kneuper, G., Regionalgeologische Folgerungen aus der
Bohrung Saar 1, Geol. Jb__, A_. 27, 499-510, 1976.
Meissner, R., Bartelsen, H. and Murawski, H., Seismic
reflection and refraction studies for investigating fault
zones along the Geotraverse Rhenohercynicum, Tec-
tonophysics, 64, 59-84, 1980.
Meissner, R. and Bortfeld, R.K., DEKORP-Atlas, Sprin
ger, 19 pp. + 80 seismic sections, 1990.
Schwab, K., Compression and right-lateral strike-slip
movement at the Southern Hunsrack Border fault
(Southwest Germany), Tectonophysics, 137, 115-126,
1987.
Wenzel, F., Brun, J.-P. and the ECORS-DEKORP
Research Group, A deep reflection seismic line across
the Northern Rhine Graben, Earth Planet. Sci. Lett., in
press, 1991.
SEISMIC ANISOTROPY OF THE EARTHVS CRUST OF THE URALS
AND ITS POSSIBLE RELATION TO ORIENTED CRACKING AND TO STRESS STATE
Sergey Kashubin
Bazhenov Geophysical Expedition of PGO vURALGEOLOGIYA , Scheelite, Sverdlovsk
District 624051, USSR
Extended Abstract. This paper reports on a (4) Deviation of compressional and shear wave
field experiment and first interpretations of particle motion is found from rays and
azimuthal anisotropy of seismic waves in the Ural wavefronts, as well as from the source-receiver
Mountains. Field observations were conducted by direction;
the Geological organization "Uralgeologiya". (5) The ratio of velocities and amplitudes of
Figure 1 shows shotpoints and multichannel compressional and shear waves depends on the
array stations. We used 25, 48-channel sets of azimuth.
geophones, located on an area of 5000 square km The azimuth of the general axes of the
with a spacing of 10 km, on four crossing lines. anisotropy ellipsoid and the coefficient of
Every set has 24 vertical geophones, with 100 m anisotropy were derived independently from
spacing on two orthogonal lines of 1100 m length, compressional and shear waves. The results do not
as well as three-component and twelve-component contradict each other. Preliminary velocity
inclined azimuth-oriented geophones [Galperin, models are shown in figure 5. They suggest that
1984]. Shots were located so that the the earth's crust is essentially anisotropic in
investigated area could be examined by seismic the depth interval from 3 to 7 km. In this depth
rays at all azimuths and up to distances of 60 range, maximum and minimum of P- and S-wave
km. velocities are presented with respect to the
Computer processing of multichannel seismic overburden, depending on the profile azimuth.
records included standard processing techniques, There may be different interpretations of this
narrow-band low-pass filtering, and controlled effect. It is customary to explain elastic
direction polarization filtering. This made it anisotropy by the following factors [Crampin et
possible to reliably distinguish first arrivals al., 1984]:
of compressional waves as well as shear waves of a) aligned crystal mineral orientation:
different polarization. b) aligned crack orientation;
Thus, we obtained travel times, apparent c) stress-induced state.
velocities and particle-motion directions for
both compressional and shear waves as a function
of the azimuth. We initially tried to make an
isotropic 3-D velocity model for the P- and S-
waves to a depth of about 8 km. However, many
peculiarities of the wave field could not be
satisfactorily explained with the isotropic
model.
The following observations were made (see
figures 2,3,4):
(1) The traveltimes and velocities of P- and S-
waves depend on the direction of propagation;
(2) Shear-wave splitting is seen in the particle
motions and in the time delays between the split
S-waves;
(3) The particle motion of the faster S-wave
depends on azimuth;
Geodynamics 22
0 10 20 30KM
I i I I
Fig. 1. Configuration of shotpoints (o) and
receiver arrays (V) of the Muslumov site in the
Urals. The frame serves as a location reference
for figures 2 and 5.
97
98 SEISMIC ANISOTROPY OF THE URALS
t - x/,.4, s
10-
8 -
6 -
4 -
2 -
0 -
N
\
, I , , X,km
10 20 30 40 50 60
Fig. 2. Traveltime curves for profiles with
different azimuths. Curves labelled with Po and
S o correspond to direct P and S-waves. The dashed
lines show travel times perpendicular to Po(S9)
direction. Reduction velocity for time scale xs
6.4 km/s.
A priori data on the geological structure of
the area make a dominant influence of the first
factor unlikely, because the primary north-south
orientation of the Urals differs from the Nl10-
120E orientation of maximum seismic velocities.
On the other hand, the composition and genesis of
the rocks are typical of isotropic rocks
[Kashubin et al., 1984]. A dominance of the other
two factors is more probable, especially if one
Vs
-5
-4
-3
0 0 10 20 360
AZIMUTH OF OBSERVED PROFILE
--// .ill
/ A I [ B
So
\ \ \ \ I
\ I !
I A
_
/ / /
/ / / / /
/ / / c
\ \
I 5 IOKM
i ! i
Fig. 4. Azimuth representation (from N clockwise)
of the general axis of the velocity anisotropy
ellipsoid for P-waves (top) and for S-waves
(bottom). Letters mark the following depth
intervals:
A - from 0 to 2 km, B - from 2 to 3 km,
C - from 3 to 4 km, D - from 4 to 5 km,
E - from 5 to 6 km, F - from 6 to 7 km,
Fig. 3. P and S-wave velocity dependence in the Anisotropy coefficients K = [(Vmax-Vmin) / V] *
center of the array. The form of the curve can be 100 % given by numbers'
approximated by the formula' 1 - from 0% to 2%, 2 - from 3% to 5%,
V(w) = A*sin(2w)+B*cos(2w)+C w = azimuth 3 - from 6% to 10%, 4 - over 10%.
KASHUBIN 99
N
A
vz,ocT Vr, M/s
2
8 -
$,o 6,0 7,0
voc rs, M/s
2,5 3,0 3,5
0 I I
2
e6 /
8
B
4,0 C
takes into account the results of mine
measurements which indicate that in the Urals the
east-west component of the lateral compressive
stress is essentially larger than the vertical
and north-south components.
Without additional information it is
impossible to separate the individual effects of
any of these factors on elastic anisotropy.
However, in our case when the composition of the
rocks and their intrinsic properties are known
from combined geological and geophysical
research, such an attempt can be made.
According to the theory of Blot [1965] elastic
wave velocities increase in the direction of
maximum compressire stress and do not change in
perpendicular direction. In the case of oriented
cracks velocities decrease in the direction
perpendicular to the cracks and do not change
parallel to the cracks [Crampin et al., 1980]. In
the depth interval from 3 to 7 km (Figs. 4,5) P
and S-wave velocities observed at azimuths of
NllO-120E are equal to velocities in unchanged
rocks. Velocities observed in N200-210E are
lower in comparison with unchanged rocks. This
effect suggests an influence of oriented cracks.
This interpretation does not contradict the
presence of a NllO-120E oriented compressire
stresses in this region [Kashubin et al., 1984].
References
Blot, M., Mechanics of incremental deformations,
New York, 502 p., 1965.
Crampin, S., McGonigle, R., Bamford, D.,
Estimating crack parameters from observations
of P-wave velocity anisotropy, Geophysics, 45,
N3, 345-360, 1980.
Crampin, S., Chesnokov, E. M., Hipkin, R. A.,
Seismic anisotropy - the state of the art,
First Break, , N3, 9-18, 1984.
Gal'perin, E. I., The polarization method of
seismic exploration, D.Reidel Publishing
Company, Dordrecht, 267 p., 1984.
Fig. 5. Tentative, first velocity-depth functions
derived from observations from four different
azimuths. A: orientation and symbols (see fig. Kashubin, S.N., Pavlenkova, N.I., Yegorkin, A.V.,
1); B: P-waves, C: S-waves. Crustal heterogeneity and velocity anisotropy
Maximum anisotropy apparently between 3 and 7 km from seismic studies in the USSR, Geophys.
depth. astr. Soc., 76, 221-226, 1984.
3-D VERTICAL INCIDENCE SEISMIC REFLECTION SURVEY
AT THE KTB LOCATION, OBERPFALZ
Manfred Stiller
DEKORP Processing Center, Institute for Geophysics of the
Technical University Clausthai, Arnold-Sommerfeld-Str.1
D-3392 Clausthal-Zellerfeld, FR Germany
Abstract. Within the Integrated Seis- with the KTB drilling site in the center.
mics Oberpfalz (ISO 89) the 3-D reflec- Geologically it can be divided into two
tion survey attains a key role. The subareas which are separated by the Fran-
DEKORP Processing Center was charged with conian Line- in the northeast granitic,
the scientific planning and the complete gneissic and amphibolitic outcrops domi-
3-D processing. The survey area is loca- hate, in the southwest Mesozoic and
ted in the FR of Germany, northeastern Paleozoic sediments [DEKORP Research
Bavaria, its center is the KTB borehole Group, 1987, 1988]. For processing the
(KTB = Kontinentales Tiefbohr-Programm survey area had to be divided into 4
der Bundesrepublik Deutschland; German subareas A, B, C, D of about equal size
Continental Deep Drilling Program). due to disc capacity. This will also
In this paper the following topics allow to obtain some results about the
will be discussed' planning (aims, theor- area around the KTB borehole at an ear-
etical acquisition configuration, field lier stage (see chapter 2).
inspection, exact positioning of all Figure 2 shows the two alternating
locations, full 3-D pre-binning), acqui- single array set-ups as they were used
sition (transposition of all locations for the seismic survey. Each one consists
into the survey area, summary of all of 10 geophone lines each with 48 geo-
actual field parameters), and first phone groups. The distance between the
results (actual 3-D bin-coverage, concept lines was 400 m, the distance between the
of 3-D processing, azimuthal velocity group centers was 100 m. The shot point
determination, preliminary 3-D stacks). line was perpendicular to the geophone
The field data comprise 3327 records lines, 8 km long with 40 shots on it,
each with 478 channels with 12 s record- spaced in a recurring sequence. Although
ing-time at 4 ms sampling rate, which the word 'shot' is mainly used in this
amounts to a total data volume of about paper (as it is usual in seismics) it is
18 GigaBytes covering a 3-D block of always meant 'vibration' because it was a
about 20 km side length. The processing vibroseis survey with the following par-
is not yet finalized, but a few insights ameters- 5 vibrators, 8-fold vertical
into seismic fine structures are already stacking, up-sweep 12 - 48 Hz, sweep
possible in order to facilitate geo- length 20 s, recording time after corre-
tectonic interpretation of the KTB data. lation 12 s with 4 ms sampling rate.
Figure 3 shows an example of a raw
The Field Work field file recorded with such a single
array set-up, the true position of the
shot is marked in Fig. 2 with an X. Its
Figure i shows a map of the survey answer to the ten geophone lines can
area which is the subject of this paper. clearly be seen. At about 4 to 5 s TWT
It is approximately 20 km x 20 km large there are the 'Erbendorf' reflections,
well known from 2-D profiles recorded in
Contenmlosphere:DpSscRefltions 1985 [Bortfeld and Meissner, 1990]. Those
Geodacs22 events are one of the main targets of the
1991AmeficanGysilUnion KTB.
101
102 3-D SEISMIC REFLECTION SURVEY AT THE KTB
Gneissic and
amphibolitic
outcrops

}ksseureuth
Mesozoic
sedimentary
covel'
?alkene
x x Granitic
xx outcrops
bendori C
x iudisch:
xx ch Sch6nticht
D
Pfaffeuth'
. -
' etadt
, e Parkstein
6000 m
en
I I
Fig. 1. Location map of the 3-D survey in northeastern Bavaria (FR Ger-
many) with geological annotations and processing subareas A, B, C, D.
The KTB borehole site is located in subarea B.
The longest offsets (more than 6 km) very few theoretical shot locations (Fig.
occur nearly in direction of the shot 6a) could be reached with vibrator
point lines, in geophone line direction trucks, so appropriate substitute sites
they are never longer than 2.5 km. The had to be found. Therefore, a few months
Franconian Line was not to be crossed before the survey started the survey area
with large offsets to avoid stacking was visited for several weeks to locate
problems due to high velocity contrasts, all roads and tracks that were prac-
so the shotpoint lines had to be oriented ticable with vibrators and mark them on
parallel to it. topographic maps.
When the 40 shots of one shot point These 'vibrator' maps were digitized
line were done, the whole single array in the DEKORP Processing Center to get
set-up moved by 800 m in geophone line the true coordinates and to test (with a
direction, and the process was repeated self-prepared interactive computer pro-
until 21 shot point lines were completed. gram) the suitability of all vibrator-
Then the whole set-up moved by 4 km in reachable sites as substitute locations.
shot point line direction, and another Finally all 2100 shot locations were
840 shots were made in the next swath. specified, a full 3-D pre-binning was
Figure 4 illustrates this proceeding. performed to guarantee the required sub-
60 % of the shot locations are used twice surface coverage in each bin. All desired
because the shot point lines overlap by shot points were plotted on topographic
20 points in each direction. Therefore, maps and given to the contractor before
with 4 swaths only 2100 shot locations the start of the field work.
were needed for the theoretical 3360 During the field work it was necessary
shots during the whole survey. to check that there was no difference
The theoretical subsurface coverage of between the planned locations on the maps
this acquisition configuration is 15- and the actual, surveyed locations in the
fold in the center area (Fig. 5a) by field. If there was a deviation the
using a bin size of 50 m x 50 m. In respective point had to be newly marked
practice the problem was that only a in the field, resurveyed and rechecked
STILLER 103
+
4
4
+

4
.+
4.
.+

4.
4
4.
4.
+
4.
4.

4
4. 4.
+

+ 4
4 +
4
+
+ 4
4
.+ .
+

+

+
+

4

+
+
+
+
+

+
+
+
+
+
+
+

+
+
+
+
+
+
+ +
+
+
+
+ +
+
+
+
+
+ +
+ +
.+
+
+ +
+ +
+ + +
+ + .+
+
+
+ +
+ .+
+ +
+ +
+ +
+
4 4 4

+
4 4
+ ,
+ +
+ +
+
+ +
+ 4 4
, . +
CH478
Fig. 3. Monitor disDlay of a raw field file down to 5 s TWT recorded
with a single (10 lines) array set-uD (shot position see Fiq. 2).
2
Vibrator lines
Interval 8 0 0 m .,,
Single array .
s e t - u p ////_///
Fig. 4. Scheme for carrying out the survey (see text for exDlanations).
STILLER 105
fold- 1 3, 4- 7', 8 11, 12 1S, >1S
b
Fig. 5. Subsurface coverage' (a) theoretical, (b) actual distribution.
Fig. 6. Source (*) and receiver (+) locations- (a) theoretical, (b) ac-
tual positions.
until all positions were properly sited. the theoretical one, although the loca-
All this was done before any seismic tion map looks very different (Fig. 6b).
records were made. The required coverage of 12-15 fold is
With this unusual procedure it was achieved in nearly the whole area; there
possible to achieve an actual subsurface are even a few parts with more than 15-
coverage (Fig. 5b) which is (regarding fold. The loss of traces (dead geophone
the bin size) absolutely comparable to locations, inaccessable shot locations)
STILLER 107
TABLE 1. Summary of all important survey parameters
method:
eguipment:
sampling rate'
instrument filter'
source:
vertical stacking rate-
sweep frequency:
sweep length:
reflection time'
recording configuration-
datum level:
number of swaths:
separation between swath centers:
geophone lines per swath-
total number of geophone lines:
separation between geophone lines'
Vibroseis
SERCEL SN 568 (telemetric)
4 ms
8 Hz 18 dB/o. 88.8 Hz 72 dB/o.
5 vibrators VVEA
5 - 8 fold
12 - 48 Hz (up-sweep)
20 s
12 s (after correlation)
cross array
500 m above msl
4
4 km
10
40
400 m
geophone groups per geophone line and swath- 208
geophone groups per shot: 478 (theor. 480)
geophone groups per geophone line and shot' 48 (first line 46)
separation between geophone groups-
geophones per geophone group:
total number of geophone locations'
shot point lines per swath'
separation between shot point lines:
shot points per shot point line and swath-
total number of shot points per shot line:
separation between shot points:
total number of shots:
total number of shot locations:
bin size'
area covered by a single array set-up-
one-fold covered area: 17.9 x 19.1 km
15-fold covered area: 15.2 x 14.4 km
100 m
18 (3-vane fan)
8320
21
800 m
40
100
100 - 300 m
3327 (theor. 3360)
2084 (theor. 2100)
50 x 50 m
2.4 x 5.8 km
(theor. 17.8 x 18.4 km)
total number of bins: 357 x 382 = 136.374 (theor. 356 x 368 = 131.008)
total number of traces: 1.562.678 (theor. 1.612.800)
source-receiver offset: 2 - 6580 m (theor. 72-6212 m)
maximum fold: 19 (theor. 15)
average fold: 11.5 (theor. 12.3)
lies only at about 3 . Altogether 18 parts of the 3-D volume have been per-
GigaBytes of data (250 magnetic tapes) formed.
were recorded within 4 months (August to Figure 7 shows the processing flow
November 1989). Table 1 summarizes all chart for this survey and the relational
important survey parameters. More details structures between seismic and control-
of the field work have been described in ling data. The circular symbols at the
a special issue about ISO 89 [Rehling and top of the figure characterize magnetic
Stiller, 1990]. tapes supplied by the contractor. The
thin-framed rectangles represent mathema-
tical procedures to apply, and the broad
The Processing arrows at the right indicate the flow of
all seismic data from the field tapes to
an interpretable 3-D data block.
The processing of this 3-D data, carried The thick-framed rectangles symbolize
out by DEKORP Processing center in Claus- the basic data files stored on discs
thal, had already started during field (seismic data and controlling databases)
work with extractions of single-fold and the broad bidirectional arrows indi-
covered sections and with geometry cate the relational structure between
coding. In the meantime brute stacks of them. The trapezoids in the center of the
I 1
Pre-stack database
(Iocat'lons, coordinates, elevations,
dead traces, field record and spread
Information, bin assignments,
azimuths. static corrections, offsets,
disk pointers, etc.)
Edit
CMP Sort
Single-fold
cross and
In I-In
!
Edit of data subsets,
data validation,
first arrival check
Header
application
Pre-stack subdatabases
Edit of
analysis ranges
Post-stack databases
(stacking velocities, dips and dip
directions, mute times. elevations.
disk pointers, etc.)
Pre-stack data
(total of 1.562.678 traces)
, [
^nalytlcal scaling
Static corrections
3 D - Interpolated
muting
Random scaling
Bandpass filtering
^utomatlc scaling
t
I 3D-azimuth dependent
dynamic corrections
3D-residual static
corrections
Stack (15 fold)
Post-stack data
(total of 357382 bins)
Location map, CDP fold plot, CDP bin
and CMP scatter plot, angular
distribution plot. offset distribution
plot, azimuth-offset spider plots,
eloclty ellipses, contour and relief
plots of eleYatlons, static corrections
and horizons, etc.
__ Coherency filtering. J/ automatic line-drawing I
I Cross and In-line sections,
Two-pass 2D-
FD -mlgatlon
,,,ystem z
Fig. 7. Flow chart of the used 3-D processing sequence (see text for
explanations) .
STILLER 109
ltlO
N
Strike, Vmi n I t (8 OO Dip direction,
,joo v.,.,,
sin'.: i, (Vm'."
interpretation system to facilitate at a
very early stage geo-tectonic predictions
during the drilling of the main KTB
borehole.
Details of 2-D processing steps in
deep seismics have already been de-
scribed [Stiller and Thomas, 1989]. The
main problem in 3-D seismic processing
is the definition of optimum stacking
velocities. Some general remarks to this
are following: if a seismic profile
crosses an inclined horizon the stacking
velocity is not the same for each survey
direction. The highest stacking velocity
occurs, if the source-receiver line is in
dip direction, the lowest in direction of
strike. Each azimuth 8 needs another
Fig. 8. Dependence of stacking velocity stacking velocity for the same horizon.
from azimuth e for a dipping layer (8 is The velocity function of 8 is an ellipse
the horizontal angle between source-re- with an excentricity depending on the
ceiver direction and dip direction; is amount of dip (Fig. 8).
the angle of dip) and formula for result- This is well known from 2-D seismics,
ing excentricity. but there it is not critical because all
sources and receivers are arranged on
one line (usually perpendicular to the
strike), so it has always the same
figure represent the analyses necessary
for the determination of optimum process- angle 8 against the dip direction of a
ing parameters, and the paper symbols at certain layer. In 3-D seismics, however,
the bottom characterize the plots avail- information is compiled from all azimuth
able for evaluation. Unmigrated data directions. This can be seen by regarding
volumina are fed to an interactive a so-called 'spider' diagram for one
specific bin (Fig. 9). Sources and re-
H
Z45.7 6
255,
2,1.'0.7
U
.;
75.7':'
, .......... '- -.:/-'120 7 '
.,; . i
iR'..:.": ..-: .::.:.;'"
':.: :.- '.::.,,;'""'
. . ' ..;.:,;...>-
65.7
ceivers had been positioned all around to
give the 16-fold coverage for the reflec-
tion points within this bin.
Therefore, velocity analyses are to be
performed in different directions to
determine (and later apply) the optimum
stacking velocities for each individual
azimuth. For this purpose a semicircle is
divided into 4 ranges separated by 45
and laid over the spider (see Fig. 9).
Separately 4 velocity analyses are to be
evaluated, each one with all the traces
of just one sector. But with 16-fold
coverage there are seldom more than 4
traces in each 45 range, which is not
enough for a stable normal moveout deter-
mination. Therefore, macrobins were
introduced consisting of 5 adjacent
single bins to stabilize the S/N-ratio
for stacking.
Of course, there is only optimum reso-
lution if the orientation of the macro-
bin is parallel to the strike of a
horizon. Therefore, different profile
directions (with different macrobin
Fig. 9. All source-receiver (S-R) line orientations) are to be analysed, each
constellations for one bin with their az- one separately for each azimuth
imuths and offsets. Each line belongs to range. For each horizon the optimum cor-
one of four 45 azimuth ranges 8 , 8z, 83 rection velocities are then used to con-
and 04 (y, v, x, u: bin directions). struct an individual velocity ellipse
Brutestack
!aline 311
T
Brutestack
7-fold trace mixing
Brutestack
7-fold trace mixing
coherency filtering
S/It - Azimuth 3
O. OOO
1.0OO
i, 2.0OO
Brutestack
!nllne 491
.000
4.000
T
W
T
Brutestack
Z-fold t,'ace mixi.g
Brutestack
7-fold trace mixing
coherehey filtering
S/It - Azimuth 3
4.000
Fig. 10. Improvement of S/N-ratio in an azimuth-restricted stack (Oe)
by summing up 7 adjacent inline sections (SW - NE) and additional coher-
ency filtering (subarea B): (a) for a steeply inclined horizon 2500 m SE
of the KTB, (b) for the Erbendorf reflections 6600 m NW of the KTB.
(least squares method). Dip direction and determined very carefully, the informa-
angle of dip can be calculated from its tion from each azimuth range is not
orientation and excentricity. sufficiently the same, so the stack of
Nevertheless, for any selected profile all traces within the bins is poorer than
direction the seismic imaGinG should the stack of each individual azimuth
theoretically be the same after applica- range, although the nominal coverage is
tion of all corrections, independent from higher.
azimuth. But exactly this is the problem So far, there is no satisfactory
with the survey around the KTB borehole. explanation for this phenomenon, maybe
Even if the correction parameters are the correct velocity curves are no more
STILLER 111
T
w
T
Fig. 11. Azimuth-restricted (el) cross-
line stack after summing up 7 adjacent
crossline sections (NW - SE) and additio-
nal coherency filtering, showing struc-
tures of the Erbendorf complex between 3
and 5 s TWT (2550 m NE of the KTB) in
subarea B.
ellipses but disturbed or distorted when
high and irregular anisotropy and dip
dependence are superimposed. Dip moveout
and refraction effects occur because of
steep dips (60 and more) of some hori-
zons.
In this paper only brute stacks with
restricted azimuth ranges are presented,
all located in subarea B (see Fig. 1 for
identification), which is only processed
so far. This subarea is about 6 km x 13
km large and contains the KTB location in
the center. Only 7 s reflection time have
been processed. There are two main
reflection ranges present in the whole
area: a steeply NE-dipping horizon bet-
ween 0.7 s and 2.1 s (Fig. 10a) and the
Erbendorf reflection series between 3.5 s
and 4 s (Fig. 10b).
To enhance the S/N-ratio 7 adjacent
inline sections (inline means geophone
line direction) are stacked together and
0 0 0 0 0 0 0 0 0
a coherency filter in inline direction is Fig. 12. Series of 5 time slices from
applied. The improvements by these two 3888 ms to 3920 ms with a spacing of 8 ms
procedures can clearly be seen. After showing the strike of an Erbendorf event
migration the dipping element will lie in for bins in subarea B.
a depth of about 6.5 km and the Erbendorf
complex at a depth of somewhat more than
10 km, the target depth of the KTB. It
has been proved that the steeply inclined
structure i,s no reflected refraction
event, maybe it is the Franconian Line
itself. Figure 11 shows again the promi-
nent Erbendorf reflections but in an
azimuth-restricted crossline section
(crossline means direction of shot point
lines), 7 adjacent crossline sections are
stacked, and coherency filtering in
crossline direction is applied.
Another possibility to see amplitude
characteristics and spatial position of
reflections is the display of time
slices. Figure 12 shows the appearance of
the Erbendorf event for constant times
over the entire subarea B. Its strike is
visible in each time slice and its dip by
comparing successive slices. The slices
are from 3888 ms to 3920 ms with a
spacing of 8 ms.
Finally, with the aid of an inter-
active seismic interpretation system, two
views into the stacked data volume are
given. In Figure 13a two inline sections
are combined with a time slice (a so-
called 'skeleton') to give a spatial
imagination of the steeply dipping ele-
ment in the upper time range mentioned
before. In Figure 13b the data cube is
cut open at the time range of a prominent
Erbendorf reflection. In this so-called
'chair' display its strike in a certain
depth can be seen on the horizontal
slice, while its dip is visible on the
adjoining vertical slices.
Conclusions
As a rsum it can be stated that high
resolution 3-D seismics in a crystalline
area is helpful but extremely hard to
handle in comparison to 2-D seismics in
the same area and to conventional 3-D
seismics in sedimentary areas. A general
problem in deep seismics is the low
reflectivity of most of the crustal
regions. Special problems with the ISO 89
3-D survey are the very steeply inclined
(60 and more) near-surface structures,
the probably high and irregular aniso-
tropy, and the often differing apparent
reflectivity (position and strength of
events)' in different offset and azimuth
Fig. 13. Display of the spatial position ranges. Additional difficulties result
of reflections using an interactive seis- from insufficient azimuthal coverage per
mic interpretation system' (a) 3-D 'skel- bin in connection with the rather low-
eton' showing a steeply inclined horizon, energetic vibroseis method.
(b) 3-D 'chair' showing Erbendorf reflec- But the processing is going on and
tions. probably will require new processing
STILLER 113
ways (e.G. 3-D dip moveout, anisotropy-
corrected velocities, envelope stacks,
spatial coherency enhancements, diffrac-
tion models) .
References
DEKORP Research Group, Results of Deep
Reflection Seismic Profiling in the
Oberpfalz (Bavaria). Geophys. J. R.
astr. Soc., 89, 353-360, 1987.
DEKORP Research Group, Results of the
DEKORP 4 / KTB Oberpfalz Deep Seismic
Reflection InvestiGations. J. Geophys.,
62, 69-101, 1988.
Meissner, R., and Bortfeld, R. K., The
DEKORP Atlas, Results of Deutsches Kon-
tinentales Reflexionsseismisches Pro-
York,
Stiller,
o
D
I
(
D
1
1
Sti
SprinGer, Berlin Heidelberg New
19 pp, 80 figures, 1990.
M., and Thomas, R., Processing
f Reflection-seismic Data in the
EKORP Processing Center, Clausthai.
n: Emmermann, R., and WohlenberG, J.,
Eds.), The German Continental Deep
rillinG ProGram (KTB). SprinGer, Ber-
in Heidelberg New York, pp 177-232,
989.
11er, M., and RehlinG, J., 3-D Reflec-
tion Seismic Survey of the Area around
the KTB Drilling Site. In: Drbaum, H.-
J., et al., (Eds.), KTB Report 90-6b
(DEKORP Report, InteGrated Seismics
Oberpfalz 1989). Niederschsisches Lan-
desamt fr BodenforschunG, Hannover,
pp 9-53, 1990.
INVESTIGATION OF STRUCTURES AND ANISOTROPY BY A 3D EXPANDING-SPREAD EXPERIMENT
H. Wiederhold
Niedersichs. Landesamt f. Bodenforschung, Stilleweg 2, 3000 Hannover 51, Fed. Rep. Germany
Abstract. In the scope of the Integrated is constant, that means at the centre of the sur-
Seismics Oberpfalz 1989 (IS089) at the Oberpfalz vey at the KTB. The aim was to get a spatial ve-
continental deep drilling (KTB) location, a 3D locity distribution for this area down to Moho
expanding-spread experiment was carried out by depth, so called 'seismic Mohole'.
the DEKORP group in 1989. The vibrator signals of In the following, an overview of the field
the 3D seismic survey were recorded with addi- configuration, the collected data and problems
tional geophone spreads in such a way that a c. with which one has to deal is given.
40-fold 2.6 * 3.8 km 2 CMP-area centered at the
KTB location was obtained. The source-receiver
offsets are in the range of 0-24 km and azimuth Field Investigations and Data Material
angles of 0-180 o are achieved. The data set con-
sists of 189,460 seismic traces. The configura- The field investigations took place in August/
tion and data description from this experiment September 1989. As seismic sources, the Vibroseis
are presented, and problems in the velocity anal- signals of the 3D seismic survey were used. The
ysis caused by geology, dipping reflections and planning, preparation and realization of the 3D
anisotropy are discussed. seismic survey is described in detail by Rehling
and Stiller [1990] and Stiller [1991, this vol-
Introduction ume]. Figure 1 gives a sketch of the actual field
geometry of the 3D expanding-spread experiment.
Since the beginning of the continental deep The area of the 3D seismic survey is shown just
drilling programme (KTB) and the continental as the vibrator traverses and, perpendicular to
seismic reflection programme (DEKORP) of the them, the geophone lines and the KTB borehole at
Federal Republic of Germany, the Oberpfalz area the centre. For recording of the symmetrical
has been the target of intensive geological and spread, symmetrical with respect to KTB, a DFS-V
geophysical investigations [DEKORP Research recording unit with 120 channels owned and opera-
Group, 1988; Schmoll et al., 1989]. With start of ted by the DEKORP group was used. In order to
drilling, the need for a 3D seismic survey to integrate this experiment into the 3D seismic
connect the 2D seismic reflection profiles and survey, the geophone locations were chosen to be
the drillhole was expressed. In 1989, with the the same as those used for the 3D seismics; thus,
completion of the pilot hole, this could be re- no additional locations had to be surveyed. The
alized, and many piggyback seismic experiments spread for one shot included two parallel lines
were carried out with the 3D survey. The whole with 60 channels each, a geophone group spacing
survey is known as Integrated Seismics Oberpfalz of 100 m and a distance of 400 m between both
1989 (IS089) and was coordinated by the DEKORP lines (same as with the 3D seismic survey).
group. An overview of the experiments is given in While, for example, the vibrators were on swath
Drbaum et al. [1990]. 4, the 'expanding-spread' was on geophone lines 5
This paper deals with a 3D expanding-spread and 6 on swath 1. Correspondingly, while the
experiment that was directly connected with the vibrators were on swath 3, the 'expanding-spread'
3D seismic survey. The basic idea, first sugges- was on lines 15 and 16 of swath 2. Now the mid-
ted by members of the Geophysical Institute of points of the geophone spread and the vibrator
Technical University Clausthai, was to record all traverse are symmetrical with respect to the KTB.
Vibroseis source signals of the 3D seismic survey The spread has been kept fixed for one vibrator
in such a way, that the source-receiver midpoint traverse (=40 source points). With the change of
the vibrator traverse (=800 m distance) the geo-
phone spread moved by 8 geophones (=800 m) in the
ContinenlLiosphere: Dp Seismic Refitions opposite direction; thus, the covered area always
Oeodynamics 22 remains the same. By the extension of the vibra-
o 1991 American GeophysicM Union tor traverse and the geophone spread, a subsur-
115
116 3D EXPANDING- SPREAD-EXPERIMENT
area of 3D seismic survey
vibrator traverse .. '
. "*'/ '5 j geophone spread
. KTB ..-
CMP-area
face area of about 4050 m times 4200 m is set-up.
The field parameters are listed in Table 1.
Originally, the expanding-spread recording
should take place during the whole 3D survey, but
due to budget restrictions only two swaths in-
stead of 4 were observed in this way.
Figure 2 gives an overview of the surveyed
area. In the middle, the main geological units
are shown: the Variscan granites of Steinwald,
Falkenberg and Friedenfels in the northeast, the
zone of Erbendorf-Vohenstraug (ZEV) with meta-
basitic gneisses and amphibolites in the middle
of the survey area, where the KTB is also loca-
ted. To the southwest follows the Franconian
Line, a steep fault zone, striking NW/SE, that
separates the crystalline area from the sedimen-
Fig. 1. Sketch of the geometry of the 3D expan- tary basin west of it. In the upper part, the
ding-spread experiment. The area shown is 20 km * topography of the area is shown. The Franconian
20 km.
TABLE 1. Field parameters for the 3D expanding-
spread experiment of the Integrated
Seismics Oberpfalz 1989.
Method Vibroseis
Equipment
DFS-V, 120 channels
sampling rate 4 ms
recording time 33.7 s
reflection time 13.7 s
source signal upsweep 12-48 Hz, 20 s
vertical stack 5 - 8 fold
data format SEGB, multiplexed
in a plane, 2 receiver
lines with 60 geophones,
400 m line interval,
100 m receiver intervail,
centre of receiver spread
and centre of vibrator
traverse are symmetrical
related to KTB.
Receiver geometry
layout
Source parameters same as 3D seismics
vibrators 5 VVEA (16 to)
pattern 48 m
interval alternating:
100-200-300-200-100 m,
40 source points per
traverse und receiver spread,
in total 21 vibrator-
traverses, subdivided into
2 swaths.
4570000 East 4520000
4500000 45 ? 0000 4520000
Fig. 2. Surveyed area (20 km * 20 km) of the
Integrated Seismics Oberpfalz 1989 with the KTB
Coverage 42 fold (theoretical) drillhole in its centre. For location see Stiller
[1991], this issue. Top: Topography (vertical
CMP interval 50 m exaggeration: 20); centre: Main geological
units: I ..... wq Steinwald granite massif,
CMP area (1 fold) 3550 m x 4150 m (theoretical) iv vv I Falkenberg/Friedenfels granite massif,
4050 m x 4200 m (actual) gneisses and amphibolites,
I',' ,q sediments. FL= Franconian Line. Bottom:
CMP area 2350 m x 3750 m (theoretical) Location map of actual vibrator points (scat-
(ca 40 fold) 2650 m x 3800 m (actual) tered) and geophone lines (5,6,15,16). The
covered subsurface area is marked.
i i i
4500000 45 ? 0000 4520000
WIEDERHOLD 117
0.00 4.00 8.00 12.00 16.00 20.00 24.00
VP - PG - OFFSET IN KM
30 - 60
< -I 0.00 4.00 8.00 12.00 16.00 20.00 24.00
_ VP - PG - OFFSET IN KM
o
n o o
.., o - 30
z_
0'.00 ' 4'.00 ......... 2.00 ' 16.00 20.00 8.00 12.00
VP - PG - OFFSET IN KM
km offsets are obtained, and for the greater azi-
muth ranges up to 18 km are obtained. Compared
with the 3D seismic survey itself, with maximum
offsets of 6.5 km, this experiment provides reli-
able conditions for velocity analyses. The range
generally covered best is 12-16 km and 12-20 km
respectively, with about 240 traces per offset in
60 - 90 100 m intervals.
A look at the subsurface coverage for these
selected azimuth ranges shows a concentration of
the better covered bins (7-16 fold) in a smaller
area, that is not the same for the different
ranges. This concentration will be advantegeous
when it is necessary to enlarge the bin size to
get a better coverage.
Figure 4 gives as an example of the data a
common shot gather for two parallel geophone
spreads of 120 traces. Channels 1-60 belong to
one geophone line and channels 61-120 to the
other one parallel at 400 m distance. Beside the
first arrivals in the range 1.8 to 2.8 s, strong
reflections are observed in the range 4-5 s, cor-
responding to vertical reflection times of ap-
proximately 3.5-4 s, a depth interval that is
known to be strongly reflective. Arrivals from
crystalline reflectors are in general not very
pronounced. However, it can be expected that
application of static corrections will improve
the correlation of the traces. Final static
corrections have only recently become available.
Fig. 3. Offset distribution for 30 o azimuth
ranges. The range 90 o - 180 o is covered sym-
metrically. The azimuth of 0 corresponds to the
direction of the geophone lines and means 233.2 o
against north (compare with Figure 1 and 2).
Discussion and Outlook
The main aim of this expanding-spread experi-
ment is the determination of seismic velocities
and thus the detection of anisotropy. Preliminary
values of the velocities can be obtained by nor-
Line in the foreground as well as the Steinwald mal moveout corrections, constant velocity
granite massif with up to 800 m height above sea stacks, x2-t 2 diagrams or by -p inversion The
level are clearly displayed The lowest part of error arising by applying hyperbolic normal move-
Figure 2 shows the actual vibrator points and out functions at the large offsets is unimportant
geophone locations of the 3D expanding-spread ex- compared to influences from geology, dipping in-
periment. A total of 189,460 seismic traces has terfaces and anisotropy. As all these factors
been recorded The area covered in the subsurface affect azimuth-dependent velocities, the velocity
is marked analysis should be restricted to single azimuth
The characteristics of this covered area are ranges
as follows: the surface geometry gives a CMP or Figure 5 shows the influence of the large,
bin spacing of 50 m. So the area is divided into seismically different geologic units with a sim-
S0 m times 50 m bins There are 81 bins in the plified geologic model The seismic velocities
direction of the geophone lines and 84 bins in for these geologic units are based on refraction
the direction of the vibrator traverses An area seismic measurements for static corrections: 3500
of 2650 m times 3800 m is covered with about 37 m/s for the sediments west of the Franconian
to 42 fold Line, 5500 m/s for the gneisses and amphibolites
The traces that fall into one single bin in- and 5000 m/s for the granite massifs Of course
clude different source-receiver azimuths and off- these are not the velocities for greater depths
sets All together azimuths of 0 to 180 o are a- and the boundaries are not so simple, but the
chieved. Figure 3 shows the offset distribution contrast between these units will exist. Calcu-
of the traces, already split into 30 o steps for lating traveltimes along cross sections A and B
different azimuth ranges. The distributions in (Figure 5) results in different average veloci-
the range 90 o to 180 are symmetrical to 90 o ties depending on the path in the different geo-
(Figure. 3). Maximum offsets of 24 km occur for logic units. For example integrating along the
O '
the 30 u to 60 range. Thzs azzmuth range is cov- raypaths one gets 4750 m/s for section A and
ered best. With smaller azimuth ranges, up to 20 5300 m/s for section B; a difference of about
118 3D EXPANDING-SPREAD-EXPERIMENT
WIEDERHOLD 119
I A
=4750 m/s
Y v=5300 m/s
10%. The velocity seems to be azimuth dependent.
This influence decreases with decreasing offset.
Seismic anisotropy in this area is known from
core samples of the KTB-borehole. Zang et al.
[1989] show velocities determined on a foliated
gneiss sample by horizontal transmission of P-
waves for all azimuth ranges (0-360 o ) and under
different hydrostatic pressures up to 360 MPa.
There is a decrease of anisotropy with increasing
hydrostatic pressure - that can be explained by
closing of microcracks. At the pressure corre-
sponding to the depth of the core sample (60 MPa
and about 2 km) there is still about 10% aniso-
tropy. Also at 360 MPa pressure there is a con-
siderable amount of anisotropy.
The influence of dipping reflectors and aniso-
tropy on velocity is shown in Figure 6. The left
part demonstrates the well-known influence of a
dipping reflector on the estimated seismic velo-
cities. The velocity depends on the angle be-
tween profile direction and strike of the reflec-
tor as well as on the dip angle , resulting in
an elliptical velocity distribution. 0nly a quar-
ter of the ellipse is shown. On the right side,
the influence of seismic anisotropy (transverse
isotropy) is shown. As a first approximation,
elliptical anisotropy was used for estimating
maximum velocity differences. A combination of
both effects yields a velocity distribution of
quasi-elliptical shape. For anisotropy axes in
the strike and dip direction, there is a super-
Fig. 5. Simplified geological model with seismic position. For anisotropy axes different from the
velocities from refraction measurements, and strike and dip directions, the ellipse is turned.
average velocities v integrated along the ray- Trying to recalculate the dip from these ellipses
paths in the cross sections A and B. without knowledge of anisotropy or, vice versa,
azimuth dependent NMO-velocities
dipping reflector anisotropy
v vx
V -- V
' NMO
NMO 1 - sin2 cos2t9
v=5000 m/s
=40
vy
dip
cos219 + (vx/vy) 2 sin2t9 '
vy
vx=5500 m/s
vy=5000 m/s
combination of both vy
_6)
vx=5000 m/s
vx dip dip
anisotropy axes equivalent to strike/dip anisotropy axes different to strike/dip (by z =45 )
Fig. 6. Influence of dip and anisotropy on normal moveout velocity V . = dip
angle, = angle between profile direction and dip of reflector or v respectively, v
and v = possible velocity components in the case of anisotropy.
y
120 3D EXPANDING-SPREAD-EXPERIMENT
deducing anisotropy without considering the dip DHrbaum, H.J., C. Reichert and K. Bram (eds),
will give wrong values. Integrated Seismics Oberpfalz 1989, KTB Report,
These small and simple examples just show the 90-6b/DEKORP Report, NLfB Hannover 1990.
difficulties and problems with which one has to Rehling, J.G. and M. Stiller, 3D reflection
deal when interpreting velocity data from this survey of the area around the Continental Deep
expanding-spread experiment while determining and Drilling Project site, KTB Report, 90-6b/DEKORP
separating the influences of geology, dipping Report, NLfB Hannover 1990.
reflectors and anisotropy. The most promising way Schmoll, J., R. Bittner, H.-J. DHrbaum, T. Hein-
to obtain true interval velocities, anisotropy richs, R. Meigner, C. Reichert, T. RHhl and H.
values and the layering of reflectors seems to be Wiederhold, Oberpfalz deep seismic reflection
with raytrace modelling methods. survey and velocity studies, Emmermann, R. and
J. Wohlenberg (eds.), The German continental
Acknowledgments. The Bundesministerium fHr deep drilling program (KTB), Springer-Verlag
Forschung und Technologie (BMFT) is gratefully Berlin Heidelberg, 1989.
acknowledged for the financial support of this Stiller, M., 3D vertical incidence seismic re-
project. flection survey at the KTB location, Oberpfalz,
References 1991 (this volume).
Zang, A., K. Wolter and H. Berckhemer, Strain
DEKORP Research Group, Results of the DEKORP 4/ recovery, microcracks and elastic anisotropy of
KTB Oberpfalz deep seismic reflection investi- drill cores from KTB deep well, Scientific
gations, J. Geophs., 62, 69-101, 1988. Drilling, , 115-126, 1989.
INTEGRATED P- AND S-WAVE BOREHOLE EXPERIMENTS AT THE KTB-DEEP DRILLING SITE IN
THE OBERPFALZ AREA (SE GERMANY)
Ewald Lschen, Walter S611ner, Achim Hohrath
Geophysikalisches Institut, UniversitRt Karlsruhe, Hertzstrasse 16,
D- 7500 Karlsruhe 21, F.R.G.
Wolfgang Rabbel
Institut ffir Geophysik, UniversitRt Kiel, Olshausenstrasse 40-60,
D-2300 Kiel 1, F.R.G.
Geodynamics 22
Abstract. Data are presented from integrated Introduction
special experiments which have been performed in
conjunction with a seismic 3-D reflection survey The interpretation of standard 2-D reflection
at the KTB (Continental Deep Drilling Program) profiling data collected in crystalline areas
site in the Oberpfalz area. often suffers from the lack of direct calibra-
The main objective of our investigations was tion. Deeper crustal reflections [e.g. DEKORP
to study the nature of P- and S-wave reflections Research Group, 1988; Lschen et al., 1987]
and velocities, the Poisson's ratio and the seis- differ significantly from those known from
mic anisotropy in a medium of crystalline rocks industrial sedimentary basin surveys. Although
around the KTB site, where the pilot-borehole has they are often very bright, reflections from
reached a depth of 4000 m. crystalline rocks are discontinuous and irregular
The program described here consists of: or, in the lower crust, laminated in appearance.
1) shearwave 2-D reflection profiling (SCMP) with The German Continental Deep Drilling Program
two 10-12 km long lines crossing the KTB-site, (KTB) offers a unique opportunity to study seis-
source: 2 horizontal vibrators with different mic reflections of the crystalline basement and
orientations, 3-component recording; their nature in more detail. An integrated seis-
2) shearwave moving source profiling (S-MSP); mic survey has been designed under the auspices
downhole recording of the source points of the of DEKORP (DEutsches KOntinentales Reflexions-
SCMP; seismik Programm) to provide an extensive data
3) vertical seismic profiling (VSP) down to 3660 base which is comparable to the techniques used
m with different source azimuths and offsets in hydrocarbon exploration, including 3-D reflec-
(zero-offset, 4 km, 8 km) and P- and S-wave tion surveys and vertical seismic profiling
sources: (VSP). This program includes many non-standard
4) multiple azimuth shearwave experiment (MASE) experiments, particularly using wide-angle, down-
with 4 km and 8 km offset and horizontal vibrator hole and shearwave techniques.
sources (radial and transversal orientation). Shearwave observations, by means of their
The VSP surveys display steeply dipping polarization and by determination of the seismic
reflections, increasing in number below 3000 m anisotropy [Helbig and Mesdag, 1982, Crampin,
depth. Horizontal structures, preferentially seen 1987a,b], in combination with standard P-wave
in the surface profiling, are the exception. measurements, possess the potential to obtain
Seismic anisotropy revealed by polarization additional information on lithology and struc-
analysis and traveltimes of split shearwaves can ture, crack and microcrack distribution. Experi-
be explained by the predominant azimuth and dip mental studies by Nur and Simmons [1969] have
of rock foliation. shown that velocity anisotropy can be induced by
the stress field. Laboratory measurements on
crystalline samples [Kern, 1982] demonstrated
that the V /V ratio can be a very useful litho-
R s
logical indicator. The KTB offers a direct link
between experiments on core samples, well-log-
ging, and seismic illumination by downhole sur-
veys and surface seismic measurements. Thus,
121
122 INTEGRATED P- AND S-WAVE BOREHOLE EXPERIMENTS
seismic images of the crystalline environment can stacking in this case is expected to enhance the
be calibrated by findings in the borehole and, shearwave signal to noise ratio and to allow for
furthermore, the regional significance of velocity analyses.
lithological and structural features in the The layout of the three component receivers
borehole can be tested. with 80 m spacing was asymmetric to the source
This article describes a package of four field points (see figure 2) in order to enable velocity
experiments, completed within the Integrated analysis for greater moveouts and to provide a
Seismic Program (IS089) between July and November better balance between the energy of first
of 1989. They focus on application of shearwave arrivals and late reflections, which is of
and downhole techniques in order to derive crucial importance when using the Vibroseis
information which complements standard P-wave technique. The deployment of three-component
techniques. We used a series of standard and non- receiver stations, particularly of the horizontal
standard field configurations as well as geophone strings, needs special care concerning
supplementary experiments with different source ground coupling and orientation. Therefore we
techniques. The basic principles of these decided to deploy them only once, and rechecked
techniques, which came into use in industrial the layout frequently, keeping it in a stationary
hydrocarbon exploration in the eighties, are position, while the source moved along the line
provided by several excellent review and in 80 m intervals. The maximum coverage is
introductory articles, e.g. Hardage [1985] and therefore produced on a sector with the KTB-hole
Danbom and Domenico [1986]. in its centre and decreases towards the ends of
Our intention was to illuminate the target the lines.
area around and below the pilot borehole (maximum Two horizontal vibrators (type VVCS, Prakla-
depth 4000 m) with seismic waves of different Seismos AG) were used as shearwave sources with
resolution characteristics for different angles an upsweep of 9 to 43 Hz and 30 s length. The
and raypaths, thereby addressing the problem of orientation of the horizontally vibrating
distinguishing lateral inhomogeneities from baseplates was transverse (Y) to the line on all
seismic anisotropy. The experiments described in points. Additionally, on every third point we
this article should not be regarded as also used Y-45 and Y+45 orientations.
independent experiments. Their full resolving In addition to the Vibroseis survey,
power is expected during integration of all explosives were used on one shotpoint on each
experiments, particularly in conjunction with the line, to produce a single-fold shotgather with
contemporarily completed 3-D seismic survey. the site of the KTB-hole in the centre of its
The processing is not yet completed. coverage (compare figure 2, shots are labeled
Particularly, polarization studies require a with 5-shot). These shotpoints consisted
systematic approach in the future. Therefore, actually of 5 distinct holes according to the
this paper describes the present stage of the three-hole technique (or Camouflet method,
data base after completion of most of the stan- Edelmann, 1985), which were fired, the centre
dard processing sequence. First noteworthy shot first, and recorded separately. Detailed
results are also outlined. Figure 1 shows the field parameters as well as a comparison between
area of main interest indicated by a frame of 15 Vibroseis and explosive results are presented by
km side length and 15 km depth around the pro- Lschen et al. [1990b].
jected KTB-hole. Reflections of almost discon- Major problems result from the topography and
tinuous character were recognized below 1 s TWT from highly variable near-surface velocities and
by the presite survey [DEKORP Research Group, weathering thickness. Since static corrections
1988]. The most prominent group of reflections is are not yet optimal, we present an example of
found between 3 and 4 s TWT, which define the so- single 3-component vibrator-point gathers re-
called Erbendorf body [Franke, 1989]. garded as typical from both lines (figure 4).
Figure 2 represents a location map of all the Processing only consisted of demultiplexing, cor-
experiments to be described in this chapter. It relation, vertical stacking of 2 field records
covers nearly the same area which has been (each one 5-fold vertically stacked during re-
surveyed by the 3-D reflection technique. Figure cording). No further processing has been applied,
3 shows the basic principles applied to each the data show true amplitudes.
corresponding experiment. P-wave onsets of the direct wave can be seen
on the vertical component, while the horizontal
Shearwave 2-D reflection profiling (SCMP)
components reveal stronger S-wave first arrivals.
The SCMP-experiment has been designed anal- Although the source orientation was transversal
ogous to study the near-vertical shearwave (Y), the energy of the recorded S-waves is shared
response of crustal structures (down to the by all three components. This indicates strong
crust-mantle boundary) known from the previous P- heterogeneities in the subsurface which may cause
wave profiling. Since it could not be expected to complex changes in the polarization of S-waves.
exceed the lateral und vertical resolution of P- The direct waves of lines SCMP1 (SW-NE) and SCMP2
waves, this experiment was designed to produce (NW-SE) are characterized by different phase
multifold CMP-coverage for two short perpendicu- velocities. This is a first indication of a
lar profiles, each approx. 5 km long. CMP- pronounced seismic anisotropy.
LOSCHEN ET AL. 123
L
50 km
[ Saxothuringian Paleozoic rocks
autochthonous (" Thuringian")
allochthonous ("Bavarian")
[ Tepla-Barrandian (ZEV) & M6nchberg
gneisses & amphibolite
ii greenschist & serpentinite
:,;....'....11... ' Moldanubtan S sir
(Drosendorf)
post-tectonic granites
Tertiary Eger Graben
Moho
prominent seismic reflections
interpreted as:
leclonic
..... stratigraphic
%,[ strongly reflective,
high- veloclly zone
(only in seclion)
Fig. 1. Study area marked by a frame of approximately 15 km side length and 15
km depth. Interpretive 3-D representation of the crustal structure around the
KTB drill site derived from surface mapping and 2-D reflection profiling during
pre-site investigations (from Franke, 1989). The surface of the marked target
area corresponds also to location map of figure 2.
Travelling in NW direction, both wave types the fracture orientation must be considered.
show increased velocities and an increased Vp/V s Alternative explanations in terms of lateral het-
ratio of 1.8 (Poisson's ratio = 0.28; crustal erogeneities, e.g. the granite-gneiss contact,
average of the Poisson's ratio is 0.25, Vp/7 = are less likely. The raypaths considered here
1.73). These values are representative a pass the ZEV body (Zone Erbendorf Vohenstrauss)
depth range of approximately 1 - 2 km around the consisting of gneisses and amphibolites (see
KTB site. The difference in Vp is 16%, in V s 11%o figure 1).
This is confirmed independently by a velocity The record of figure 4 also reveals distinct
evaluation of the S-MSP data (compare next chap- seismic energy on the horizontal components,
ter). A possible relationship of this azimuthal particularly at 7 s TWT, presumably S-wave
anisotropy with the recent stress field and / or reflections. There is no energy seen on the
2 km
i
MASE
815
Fig. 2. Location map of the study area, reduced from scale 1:25000. SCMP:
Shearwave 2-D reflection profiling, dots are vibrator points, solid lines are
stationary 3-component receiver spreads. S-MSP: Shearwave moving source profile,
simultaneous recording of all SCMP source points with 3-component downhole
geophone units. VSP: Vertical seismic profiling, 4 source points: 101 (zero
offset), 401 and 402 (4 km offset) and 802 (8 km offset), SP: explosive source,
VP (P): vertical vibrator, VP (S): horizontal vibrator, MP (L) Marthor source
(hammer, left), MP (R) Marthor, right, CIPHER: circular polarized horizontal
excitation. MASE: Multiple azimuth shearwave experiment, vibrator source points
401-420 (4 km offset) and 801-820 (8 km offset).
vertical component. The subsurface covered by This must be substantiated by further polariza-
reflections is nearly identical in the center of tion analysis of the complete dataset.
the spreads, but different in azimuth. The
reflected energy is characterized by two distinct Shearwave moving source profiling (S-MSP)
phases, one on each component. Obviously, these
elements are different in polarization by 90 A moving-source-profile (or walkaway survey,
degrees. We argue that these phases might be compare figure 3) was gathered at the same time
split shearwaves, propagating through an aniso- as the SCMP survey. This was achieved by re-
tropic medium above the reflecting structure. cording all SCMP sources simultaneously by a
LOSCHEN ET AL. 125
SCMP
VSP
S -MSP
downhole tool. A digital 5-unit geophone chain
(SEKAN 5, Prakla-Seismos AG) was placed in a sta-
tionary position between 3590 m and 3690 m depth,
the spacing being 25 m, each unit consisting of 3
components. Each unit was equipped with a mag-
netic compass for orientation.
Figure 5 shows an example of the first line
representing a horizontal component (HI) of the
second unit (of 5 units), located at 3665 m depth
with an azimuth of N 2 (H2 at N 92). The traces
correspond to the vibrator location moving from
NE over the KTB-site to the SW.
The direct shearwave arrivals seen between 1
and 2 s traveltime, form a typical hyperbola-like
curve. Based on the CMP-concept, this allows for
a velocity analysis similar to the move-out
analysis in surface reflection studies. Such an
analysis gives a shearwave velocity of 3130 m/s
Fig. 3. Schematic diagram with the basic for line SCMP1 (figure 5) and 3240 m/s for line
principles of the SCMP, S-MSP, VSP and MASE SCMP2. These velocities must be regarded as aver-
experiments. For abbreviations see figure 2. age velocities for the medium around the borehole
[ SCMP [OFS48+DFS120) LINE $CMP2 VP 197 RZ Y
X Y Z
Fig. 4. Vibrator gather of VP 197 of line SCMP 2. 56 stationary channels for
each component. Offset range is from 3 km to 7.4 km. Orientation of the vibrator
baseplate is transverse (Y). Data were demultiplexed, correlated (extended
correlation), 10-fold vertical stack (2 field records). No further processing.
True amplitudes. Note highly correlated P-wave first arrivals on the vertical
component and direct S-wave onsets on the horizontal components. S-wave
reflections of different polarization at about 7 s TWT.
KT8-ISO: S-4FVE flSP S-UNIT-TOOL LINE SCMP 1
SW
NE
profiling and the lithology in the borehole. Its
application in hydrocarbon exploration has been
widely established during the eighties [Hardage,
1985].
P-wave sources as well as S-wave sources were
applied. In order to cover a maximum range of
frequencies and corresponding resolution
characteristics, different source types for each
wave type were also used: dynamite, vertical
vibrator, horizontal vibrator, horizontal hammer
(Marthor, trademark of IFP).
Four different source locations were used
(compare figure 2). The dynamite shotpoint SP 101
and the S-vibrator VP 101 S were located at the
same place at 60 m offset, while the other zero-
offset sources VP 101 P, MP 101 were located at
160 m and 200 m offset. The other three source
points 401, 402 and 802 were located at 4 km and
8 km offset, respectively, and in different azi-
muths (see figure 2). They were considered to
provide a link to the MASE-Experiment and to
illuminate the target area with a wider range of
different raypaths, which is important for in-
version techniques and for anisotropy studies.
The downhole tool (SEKAN 5) and the recording
equipment were the same as described above.
Besides routine work (demultiplexing, editing,
sorting, etc.) the basic processing was as fol-
lows: in the case of Vibroseis sources, the data
were vertically stacked (2 records) and then
correlated in two versions: zero-phase
correlation, minimum-phase correlation. After
these steps the components were rotated into
Fig. 5. Moving source (walkaway) profile with appropriate coordinate systems [e.g. Toksbz and
the shearwave sources of the SCMP survey. The Stewart, eds., 1984], e.g. into the geographic
traces are recorded with a horizontal component system using the compass readings or into the
(HI) located in 3665 m depth (selected from 5 direction of the raypath using automatic
units, located between 3590 and 3690 m, with 3 procedures, maximizing the energy or using the
components each). Note hyperbolic-like direct S- covariance matrix [DiSiena et al., 1984;
wave arrivals and later arrivales marked in the Kanasewich, 1981; Benhama et alo, 1988].
figure. True relative amplitudes. Optionally, a deconvolution was then applied.
This is of particular importance for all zero-
offset VSPs. The resolution of the signals is
between surface and downhole receiver. The enhanced and multiples are suppressed.
discrepancy is consistent with the azimuth- Thus, four different zero-offset VSPs are now
dependent velocities of the SCMP direct waves available, showing P- and S-waves of different
mentioned above. After normal move-out correction energy and signal frequencies. Additionally,
of the arrival times, short-period deviations can several P to S conversions are visible starting
be used advantageously for calculating static at the P-wave branch of the direct arrivals. The
corrections for the SCMP survey. Also, later S- direct S-wave, which originates at the surface,
wave arrivals are seen on the traces of the is propagating in different phases of their own
southwestern vibrator locations behind the direct velocity, clearly seen in the vibrator profiles
arrivals (marked in figure 5). Arriving nearly at VP 101 P (figure 6) and VP 101 S. This is a first
the same time from about 25 vibrator positions direct manifestation of shearwave splitting
(arrivals are aligned horizontally), they may be according to seismic anisotropy. The slowest
interpreted as reflections from a steep (40-60 ) phases also could be interpreted as Stoneley
structure dipping to the east. This structure waves (borehole guided wave). But this
might be intersected by the borehole below 4000m. interpretation is very unlikely, because the
observed amplitude decay is typical for body
Vertical seismic profiling (VSP) waves and because there are no such waves
reflected from the bottom of the borehole.
Vertical seismic profiling provides the Although the vibrator and the Marthor source were
closest link in interpretation between reflec- oriented transversly with respect to the source-
tions observed in 2-D and 3-D surface reflection borehole azimuth, the recorded S-wave energy is
LOSCHEN ET AL. 127
2
Fig. 6. Original section of VSP VP 101 P (vertical vibrator), N-component, AGC
with 300 ms window length. Two split shearwaves are marked.
not restricted to the HT-component. This indi- S-wave generation is considered to be caused by
cates that polarization of the signal is changing conversion of P to S energy at heterogeneities
along its raypath due to the complexities and the near the source. The difference in amplitudes
anisotropy of the medium. In the case of dynamite between the two splitting S-waves indicates that
sources, the S-wave generation is most likely due the polarization of the converted energy is
to conversion of the P-wave at the surface or at nearly parallel to the slower direction of the
lateral inhomogeneities near the source [Fertig, medium.
1984]. For the vibrator and the Marthor, the In case of zero-offset VSP wavefield separa-
shearwaves can be regarded as originating tion techniques can be applied to suppress the
directly at the source. downgoing wavefield (direct waves) and to enhance
In figure 6 the two continuously splitting the upgoing reflections [e.g. Hardage, 1985].
shearwaves can clearly be recognized. Hodograms Usually, f-k filter or median filters are applied
of the corresponding signals, presented in figure for this purpose. The traces are then time
7, clearly show, that the faster and the slower shifted according to the two-way-traveltime cor-
shearwaves are different in polarization by about rection. After performing a corridor stack, the
90 . This means that the S-wave with particle VSP now can be compared directly with a surface
motion in NW direction is faster than the other. reflection profile. On the other hand, these
E o
o
o N-i i::
o o
o
( o
rO ,,.-I .,.-I
c O
O(4-1
o
o
o
o
J.N / .
LOSCHEN ET AL. 129
seismic measurements now can be calibrated by the zero-offset profile SP 101 representing the
borehole measurements. reflected P-waves.
Figure 8 corresponds to the upgoing (re- In the VSPs of the crystalline environment
flected) wavefield after f-k filtering of the fundamental differences are discernible compared
DEPTH [M]
o
2
TWT
3
Fig. 8. Upgoing wavefield of VSP SP 101 after f-k filtering, radial component,
in two-way-traveltime representation. Trace interval 25 m. Bandpass 40-110 Hz,
AGC with 300 ms window for plotting. Muting after f-k filtering along first
arrival traveltime curve.
to data from sedimentary strata. Instead of long is responsible for the dipping reflector. The
continuous reflected branches, numerous shorter question about the nature of the horizontal
segments are visible. Only a few of them form reflector is still open.
long consistent branches, although often inter- Consistent reflections are seldom seen above
rupted. We attribute this to a generally lower 3000 m, although shorter segments correlate with
energy of reflections, hence a lower signal to variations in the sonic log. Below 3000 m depth,
noise ratio, which is caused by energy losses due the intensity of the reflection pattern is
to numerous small scale impedance contrasts. The generally increased. Inclined reflections showing
wavefield therefore must be considered as an the characteristic dip and a slight bending are
interference pattern or back-scattered energy dominating at all depth ranges. All possible dips
from numerous single reflections. are present, the steepest reflectors are visible
Reflections from horizontal structures would in unfiltered versions, represented by downgoing
show up parallel to the depth axis. Only one waves behind the direct wave. With this
horizontal reflection is visible in figure 8 information from the VSPs, processing of 2-D and
originating at 3200 m depth. The sonic log shows 3-D surface measurements could focus on dipping
a pronounced low-velocity peak at this depth, reflections using modified stacking velocities or
where a fracture zone is known from other dip move-out (DMO) stacking techniques.
borehole measurements and from coring, but it is
characterized by a strong dip. Curiously, at the Multiple azimuth shearwave experiment (MASE)
same depth in the VSP another reflection
originates which is oblique to the depth axis, Shearwave sources were distributed over semi-
indicating a pronounced dip. It also reveals a circles around the KTB-hole, with special empha-
curvature typical for reflections from dipping sis on the analysis of azimuth-dependent aniso-
structures. It must be assumed that the fracture tropy. Two circles of 4 km and 8 km radius,
N E N E N E N E
P
Is]
Fig. 9. Data from MASE. All traces in this example are from Z component of the
second downhole unit in 3375 m depth. From left to right: 4 km offset with
transverse source orientation, 4 km offset with radial orientation, 8 km offset
with transverse orientation, 8 km offset with radial orientation. In each panel
the traces start on the left with the source location in the NW. Static
corrections include correction for distance, but no near-surface effects at the
source. True relative amplitudes.
LOSCHEN ET AL. 131
N
respectively, were selected, starting NW over E
to SE, each one with 20 source points (see figure
2). The far-offset source points of the VSP were
also located on these half-circles.
One vibrator (VVCS) was used for radius 4 km
and 2 vibrators for radius 8 km. At each point
the transverse (Y, with respect to source-
borehole azimuth) orientation of the baseplate
movement was recorded first und then the radial
(X) orientation.
The downhole tool of 5 geophone units (SEKAN
5) was located at 3300-3400 m depth, with 25 m
spacing, during the whole experiment. At each
source point an extensive shallow shearwave
refraction survey was conducted. The goal was to
correct static effects on time delays and
anomalies in polarization caused by the near-
surface weathering layer.
Figure 9 displays the traces recorded with the
vertical component of unit 2 in 3375 m depth. It
is selected as an example out of 5 units located
in 3300 to 3400 m depth, each one with 3 compo-
nents. The traces represent the vibrator posi-
tions starting in the NW moving around the KTB-
site on two different radii. The first group of
traces (labelled 4 T, lefthand) are from trans-
verse orientation of the S-wave vibrator on the
4-km-circle, the second group (4 R) from radial
orientation, the same with the 8-km-circle.
Static corrections have been applied which ac-
count for slight variations in the source-re-
ceiver distance. Variations in travelpath length
(calculated with 3-D coordinates of receiver and
source) are reduced with an average S-wave vel-
ocity of 3070 m/s.
Shearwaves are the dominant phases on figure
9. The first S-wave clearly displays a traveltime
anomaly consistently on both radii and both
source orientations. As expected from azimuthal
anisotropy, this anomaly has a sinusoidal
behaviour as a function of the azimuth.
' Conclusions in terms of anisotropy can be drawn
after application of additional static correc-
tions for near-surface effects and polarization
analysis.
Systematic analysis of the polarization using
variable source and three receiver components is
- ' expected to provide a quantitative image of
seismic anisotropy, which may be caused by the
,-, ..... recent stress pattern [Nur and Simmons, 1969], or
Fig. 10. Schematic diagram showing the suggested ,-,,m -' effect of the rock foliation on shearwave
splitting. Lower part shows an example of the
, effect seen in particle motion diagrams of
shearwave direct arrivals of VSP VP 101 S (S-wave
vibrator) at 3200 m depth. Motion is first in NW-
TillelM /1SLrC T -
SE, then in SW-NE direction with a delay of about
" '" " 100 ms.
by orientated fracture systems [Crampin, 1987b], the seismic wavefield, we expect more quanti-
or by texture and foliation of the rock. The MASE tatire results on seismic anisotropy, its nature
provides observations variable in azimuth, but and its distinction from lateral heterogeneity.
constant in offset. The S-MSP on the contrary Further evaluation will be accomplished with
allows for observations constant in two azimuths, reference to the other results of ISO89, in par-
but variable in offset. The VSPs complete the ticular to the 3-D survey.
observation scheme with data variable in depth. The initial results presented in this paper
The initial results presented in this paper indicate that crack-related and, in particluar,
show a remarkable coincidence with the orienta- texture-related seismic anisotropy plays a
tion of the most compressire stress in the KTB significant role for in-situ observations in most
deduced from televiewer observations [Mastin et crystalline rocks. In this respect they confirm
al., 1991]. The orientation of this stress field findings by Lschen et al. [1990a] from modelling
is almost N 160 . The azimuthal difference in P- integrated near-vertical and wide-angle P and S-
and S-wave velocities observed in the SCMP wave reflections from the lower crust in SW
survey, the fastest waves being observed in NW Germany. Seismic anisotropy must be taken into
direction, may be attributed to the stress field account before interpreting seismic velocities
and to the possible preferred orientation of purely in terms of lithology.
cracks. This is very likely for the uppermost 1
or 2 km. The velocities deduced from the first Acknowledgements. Financial support from the
arrivals of the SCMP survey also correspond to BMFT (Federal Ministry of Science and Technology,
this depth range. Bonn; grant RG 8801) is gratefully acknowledged.
Below the upper depth range, particularly The DEKORP management at the NLfB (Geological
between 1500 and 3000 m, we suggest that shear- Survey of Lower Saxony, Hannover) is thanked for
wave splitting observed in figures 6 and 7 is administrative survices. The crew Schwanitz of
mainly caused by rock foliation, which is known PRAKLA-SEISMOS AG, Hannover, is thanked for
to almost NW in strike direction and subvertical technical management of the field work. Proces-
in dip [Emmermann, 1989]. The downgoing split sing is being done at Karlsruhe University using
shearwaves diverge continuously in their travel- a CONVEX C1 computer and DISCO software.
time reaching a difference of about 100 ms at Contribution No. 444 of the Geophysical
3000 m depth (figure 6). This corresponds to a Institute, Karlsruhe.
velocity anisotropy of the order of 10 %. This
value can be directly compared with experimental
laboratory data on representative samples by Kern References
and Schmidt [1990]. All the KTB samples show
significant velocity anisotropy due to oriented Benhama, A., C. Cliet, M. Dubesset, Study and
microcracks (with major contributions in the low- applications of spatial directional filtering
pressure range, in particular for the upper 2 in three-component recordings, Geophysical
kilometers) and to crystallographic preferred Prospecting, 36, 591-613, 1988.
orientation. Sillimanite and biotite bearing Crampin, St., Geological and industrial
metapelitic gneisses dominate in the depth range implications of extensive-dilatancy
of 1500-3500 m. They show a considerable intrin- anisotropy, Nature, 328, 491-496, 1987a.
sic (texture and composition-related) velocity Crampin, St., Crack porosity and alignment from
anisotropy of more than 10 % for P and S-waves. shear-wave VSPs, in Shear-wave Exploration,
Maximum S-wave splitting is observed for raypaths eds. S.H. Danbom and S.N. Domenico,
parallel to the foliation plane, the fastest S- Geophysical Developments, SEG Special Publ.,
wave being polarized parallel to the foliation 1, 227-251, 1987b.
plane, the slowest normal to the foliation [Kern DEKORP Research Group, Results of the DEKORP4/KTB
and Schmidt, 1990]. Oberpfalz deep seismic reflection
In figure 10 the suggested relation between investigations, Geophys., 62, 69-101, 1988.
foliation and the in-situ observed shearwave Danbom, S.Ho and S. N. Domenico, (Editors),
splitting is shown. Cracks and microcracks with Shear-Wave Exploration, Geophysical
the same orientation can produce the same effects Development Series, Vol.l, Society of
[Crampin, 1987b] and supplement the texture-re- Exploration Geophysicists, Tulsa, Okla., 274
lated effect in the upper depth range. It is im- p., 1986.
portant to note that this behaviour is observed DiSiena, J.P., J.E. Gaiser, and D. Corrigan,
for near-vertical raypaths only (zero-offset Horizontal components and shear wave analysis
VSP). When oblique raypaths are considered (S- of three-component VSP data, in Vertical
MSP, far-offset VSP, blASE) the polarization of S- Seismic Profiling, Part B: Advanced Concepts,
waves gets much more complex and difficult to in- eds M.N. Toks6z, R.R. Stewart, Geophysical
terpret. Press, London, p. 177-188, 1984.
From a more systematic analysis of the Edelmann, H.A.K., Shear-wave energy sources, in
polarization in the future, taking into account Seismic Shear Waves, Part B, Applications, ed.
the variation of source orientations, offsets and G. Dohr, Geophysical Press, London, p. 134-
azimuths at the surface and depth variations of 177, 1985.
L0SCHEN ET AL. 133
Emmermann, R., The KTB pilot hole' tectonic
setting, technical data and first results, in
The German Continental Deep Drilling Program
(KTB), eds. R. Emmermann, J. Wohlenberg,
Springer Verlag, Berlin, p. 527-553, 1989.
Fertig, J., Shear waves by an explosive point-
source' the earth surface as a generator of
converted P-S waves, Geophysical Prospecting,
32, 1-17, 1984.
Franke, W., The geological framework of the KTB
drill site, Oberpfalz, in The German
Continental Deep Drilling Program (KTB), eds.
R. Emmermann, J. Wohlenberg, Springer Verlag,
Berlin, p. 37-54, 1989.
Hardage, B.A., Vertical seismic profiling, Part
A' principles, Handbook of Geophysical
Exploration, Vol. 14A, Geophysical Press,
London, 509 p., 1985.
Helbig, K. and C.S. Mesdag, The potential of
shearwave observations. Geophysical
Prospecting, 30, 413-431, 1982.
Kanasewich, E.R., Time Sequence Analysis In
Geophysics, The Univ. of Alberta Press,
Winnipeg, 480 p., 1981.
Kern, H.,P-and S-wave velocities in crustal and
mantle rocks under the simultaneous action of
high confining pressure and high temperature
and the effect of the rock microstructure, in
High-Pressure Researches In Geoscience, ed. W.
Schreyer, Schweizerbartsche
Verlagsbuchhandlung, Stuttgart, 15-45, 1982.
Kern, H. and R. Schmidt, Petrophysical
investigations on KTB core samples at
simulated in situ conditions, Scientific
Drilling, 1990,
Lscheh, E., F. Wenzel, K.-J. Sandmeier, D.
Menges, T. Rhl, M. Stiller, W. Janoth, F.
Keller, W. S611ner, R. Thomas, A. Krohe, R.
Stenger, K. Fuchs, H. Wilhelm, and G.
Eisbacher, Near-vertical and wide-angle
seismic surveys in the Black Forest, SW
Germany, J. Geophys., 62, 1-30, 1987.
Lschen, E., B. Nolte, and K. Fuchs, Shear-wave
evidence for an anisotropic lower crust
beneath the Black Forest, southwest Germany,
Tectonophysics, 173, 483-493, 1990a.
Lschen, E., W. S611ner, A. Hohrath, W. Rabbel,
Integrated P- and S-wave borehole experiments
at the KTB-deep drilling site, in KTB Report
90-6b - Dekorp Report, eds. H.-J. Drbaum, Ch.
Reichert, K. Bram, NiedersRchsisches Landesamt
fur Bodenforschung, Hannover, 85-134, 1990b.
Mastin, L., B. Heinemann, A. Krammer, K. Fuchs,
M.D. Zoback, Stress orientation in KTB Pilot
hole determined from wellbore breakouts,
Scientific Drilling, in press, 1991.
Nur, A. and G. Simmons, Stress-induced velocity
anisotropy in rock: an experimental study, J.
Geophys. Res., 74, 6667-6674, 1969.
Toks6z, M.N. and R.R. Stewart, (eds), Vertical
Seismic Profiling, Part B: Advanced Concepts,
Handbook of Geophysical Exploration, Vol 14B.,
Geophysical Press, London, 419 p., 1984.
PROCESSING AND ANALYSIS OF MSP-EXPERIMENTS
WITHIN THE KTB-PROJECT
M. Kemper and H.-P. Harjes
Institute of Geophysics, Ruhr-University Bochum,
P.O. Box 102148, 4630 Bochum, FRG
Abstract. At the KTB drill-site several MSP-surveys
were conducted. Their aim was the prediction of discon-
tinuities that should be hit by the main drill hole. By itera-
tive forward and inverse modeling a subsurface image was
obtained. In addition to a steeply dipping reflector element
at approximately 4500 m depth and a few near-horizontal
elements in the depth interval from 6000 m to 8000 m, the
analysis suggests a highly shattered, inhomogeneous rock
unit down to at least 3.5 s TWT, without discontinuities of
considerable spatial extension.
Introduction
To improve the subsurface images obtained from conven-
LXUIICLi. :)LLAXI&,, IJXLJXIX;,, :)LJIJXIXLII.LLI. 2)1111/ J.I. L41, E:lL.IUXXLXLJXX
techniques using geophones in downhole arrays are avail-
able. These seismic measurements are generally considered
as Vertical Seismic Profiles (VSP). A VSP survey is nor-
mally executed with a single source position near the well
head, and therefore provides subsurface information only
within the Fresnel zone surrounding the well.
Beside this, extended acquisition schemes like the Mov-
ing Source Profile (MSP) allow an increased resolution of
the subsurface. An MSP is derived from a conventional
VSP geometry and differs by the use of many additional
source positions on a profile crossing the well site (Fig-
ure 1). This configuration allows not only the illumination
of .the borehole environment but also that of the subsurface
away from the well. The use of several source positions
further permits detection of dipping elements and allows a
Geodynamics 22
135
multifold coverage of the subsurface. Analysis of direct
waves can be used to deduce the seismic velocity of the
overburden (Harjes et al., 1990).
With this in mind, we recently conducted three MSP sur-
veys within the framework of the Integrated Seismic Survey
Oberpfalz (ISO-89) at the pilot-hole of the German deep
drilling project (KTB) located in the Oberpfalz, FRG
(Stiller, 1991). These experiments amend seismic surface
surveys and serve as a calibration of reflection elements
seen on surface profiles to lithological information from
borehole data. The most challenging part of the MSP sur-
vey is the prediction of discontinuities that should be hit in
the main hole below the geophone depths used for the MSP
surveys.
Whereas the MSP-technique is well established in sedi-
mentary structures, the MSP-experiments hereafter
described were performed in a crystalline environment. Due
to the different characteristics of such a location, e.g. the
short lateral continuity of reflection elements, no specific
target horizon was detected by seismic pre-surveys in the
first 4 s two-way time (TWT) which could be used for tun-
ing the MSP survey.
MSP data acquisition
A first MSP-experiment (MSP 1) was conducted in
December 1988 to gain experience with the site-specific
problems of data acquisition and processing in the KTB-
area, i.e. to get a first insight into the geological structure at
the location of the pilot hole and its extension to the neigh-
bouring site of the main deep drilling project planned at
about 200 m distance (Figure 2). Additionally, two seismic
surface profiles were recorded simultaneously using a 120
trace geophone spread along both vibrator profiles (Harjes
et al., 1990).
136 PROCESSING AND ANALYSIS OF MSP-EXPERIMENTS
1.0
2.0
3.0
4.0
Moving Source Profil / Welk-ewey VSP
OFFSET (krn)
Fig. 1. Moving Source Profile (MSP) survey, schematic.
Within the ISO-89 experiment, a full MSP-profile (MSP
2) was recorded. A receiver chain of five oriented, three-
component geophones was used, separated by a distance of
25 m. By altering the chain depth four times, a depth range
from 3210 m to 3685 m was covered. This gives a theoreti-
cal 20-fold coverage of the subsurface. The vibrator line
26000
24000
22000
20000
8000
16000
14000
, I , i , i i i , i ,
NS/WE-PROFILE
.T. %. MSP-2, VP
KTB'
KTB-VB
" ,,,,,, :...
, , ,
000 6000 'soob 0000 2000 aooo 6000
R45
Fig 2. Location map of the study area. Shown are the
vibrator positions (VP) of MSP2 and the seismic surface
profiles (NS/WE-profiles), acquired while MSP1, also the
profiles KTB8502 and DEKORP 4N. In the lower left the
Franconian line (F.L.) is indicated, one of the regional geo-
logical structures.
for this multifold MSP-profile was oriented in NE-SW
direction to fit into the grid of the surface seismic data of
the ISO-89 3D-campaign (Stiller, 1991). Along the vibrator
lines, combisweep signals (20 - 80 Hz) of 20 sec duration
were produced at 50 m interval. 5 -10 sweeps were stacked
and correlated with an effective recording time of 12.5
seconds at a sampling rate of 2 msec.
MSP data
Typical examples of raw data are plotted in Figure 3 and
Figure 4. Figure 3 shows the Z-component from depth
position 3235 m. The clearest coherent events can be asso-
ciated with the direct P-wave and the direct S-wave. The
direct S-wave seems to be generated by mode-conversion at
the near-surface weathering layer and can be recognized by
its larger moveout and also by its different spectral content.
The incident energy shows a long reverberant tail, and the
source wavelet itself appears very broadened. Between the
P- and S-wave some distinct reflection elements can be
recognized, primarily at the eastern part. Obviously the long
coda of the S-wave masks later reflection events; neverthe-
less, after closer inspection reflection elements with short
lateral continuity can be found. Figure 4 shows the
corresponding crossline component. Again the dominant S-
wave is recognized. The previously mentioned reflection
elements between the P- and S-arrival are clearly distin-
guishable.
MSP data processing
Initially, a conventional VSP/MSP data processing was
applied, e.g. as described by Ahmed et al. (1986). But it is
by no means evident that these procedures, developed for
sedimentary structures, are applicable to crystalline struc-
tures. Therefore a modified MSP-processing scheme for a
SUN-workstation was developed as outlined in Figure 5.
After data editing, static corrections were applied to
correct for near-surface effects at the source position. Stat-
ics were obtained from the ISO-89 3D-campaign.
Acquisition of the total seismic wavefield with the three-
component borehole geophones allows separation of the
different wavetypes (P, S, etc.) encountered in the data set
by single axis rotation. This separation requires estimation
of the direction of particle motion. In a first step, the two
horizontal components were oriented towards the inline
component Hp and the crossline component T (Figure 6).
Further rotation towards the incident P-energy yields the
KEMPER AND HARJES 137
SW
NE
Fig. 3. Raw data, z-component, depth position 3235 m.
1.00
1
2.00
2.50
3.00
3.50
data in three components P, V - orthogonal to P- and T.
Wavefield separation was achieved by FK-filtering of sin-
gle, fixed-offset source VSP's or Common-Shot-Gathers.
VSP's were constructed by sorting traces from all MSP's
with a common shot point but varying recording depth.
Prior to the FK-filtering, amplitude balancing was applied
to each VSP. Finally, the data set was resorted again to
reach MSP or Common-Receiver-Gather.
SW
NE
2.00
Fig 4. Raw data, crossline component, depth position 3235 m.
2.50
3.00
3.50
Results of processing are shown in Figure 7 and Fig-
ure 8. Here, the upgoing P and the upgoing T component
of the same depth-position as before are displayed. The P-
and S-arrivals are clearly weakened though still visible.
This phenomena has to be attributed to the complicated
structure of the dominant wavelet and insufficient
consideration at least of processes like wavelet shaping and
deconvolution. The upgoing T-component shows a distinct
UPGOING
WAVEFIELD
I MSP-RAW U OATA I
DATA H EDITING I
I
APPLY STATIC l
I
ESTIMATE
FIRST ARRIVAL
PARTICAL MOTION
I
ROTATION (PVT)
I
SORT TO OFFSET VSP'S
COMMON SHOT GATHER
I
AMPLITUDE BALANCING
I
WAVEFIELD SEPARATION I
I
DOWNGING
WAVEFIELD
I
SORT TO MSP'S
COMMON RECEIVER GATHER
Fig 5. Flow chart of three-component MSP processing
sequence.
band of reflection elements in the NE-part (R1, see figure
8), and a slight improvement in the lower part of the sec-
tion. Nevertheless, the overall transparency still remains.
The single offset VSP's (Figure 9), which build up an
MS P, show several remarkable events.
MS P data analysis
Due to the lack of clear reflections and insufficient spa-
tial aperture of the MSP-geometry, migrated images
obtained by conventional migration algorithms are dom-
ROTATION
-Z
H1 T
HORIZONTAL
PLANE
Fig 6. Principle of axis rotation. For explanation see text.
inated by artifacts ('smiles'). However, an image of the
subsurface was obtained by simple kinematic migration and
by incorporating surface seismic data.
The seismic surface data acquired during MSP 1 were
used to locate the dominant reflection R1. The stacked and
unmigrated surface section along a W-E profile is shown in
Figure 10. A prominent reflector, steeply dipping from west
to east, can easily be recognized. Illuminating this reflector
by an MSP-geometry and comparing calculated ray-
theoretical traveltimes (Janik, 1990) with observed travel-
times from the MSP data (Figure 11) suggests that both
events are identical. At the least, the reflector can be
located at a depth of approximately 3590 m to 4545 m, dip-
ping approximately 30 to southeast.
For migration of less dominant reflection elements, espe-
cially for those which occur after the direct S-arrival, the
VSP-CDP-mapping introduced by Dillon and Thomson
(1984) was used. For plane layering, this mapping
transforms each time sample of a trace to its depth position.
The assumption of plane layering is fundamental for the
mapping, which constitutes a correct kinematic migration.
This method seems to be inadequate for application to
the actual data set because any restriction of layer dip in a
crystalline environment has to be questioned. To avoid this
problem all results obtained by CDP-mapping have been
checked for correctness by a subsequent ray-theoretical for-
ward modeling, eliminating all artifacts and misleading
results, e.g. inclined structures. Only near-horizontal
migrated elements are accepted for further interpretation.
In Figure 12 all these elements are summarized.
Discussion
The most challenging part of the MSP-experiment was the
mapping and prediction of discontinuities, especially those
that should be hit when the drilling operation will reach
depths below the pilot hole. Results obtained hitherto are
summarized in Figure 12. Incorporating surface data, the
clearest reflection that could be located by the MSP-
interpretation, is predicted to be crossed by KTB at a depth
of 4600 m by the main hole. Later reflections are mapped
at depths from 6000 m to 8000 m.
The results fit into the general idea of the subsurface at
the KTB, with steeply inclined elements in the upper part
and near horizontal elements in the deeper part (DEKORP
Research Group, 1988). The almost transparent character of
the sections indicates no substantial change in the overall
material properties, and suggest a rather highly shattered,
SW
NE
Fig 7. MSP upgoing wavefield, P-component, depth position 3235 m.
inhomogeneous zone without discontinuities of considerable
spatial extension. Consequentely, the limited lateral exten-
sion of the reflection element R1 permits no correlation
with the Franconian line, the major fault zone (see again
Figure 2) nearby the KTB location (Franke, 1989).
The results presented here were achieved by iterative for-
ward and inverse modeling. Conventional processing was
not applicable to the MSP-data, which were mainly
SW
NE
Fig 8. MSP upgoing wavefield, T-component, depth position 3235 m.
influenced by the heterogeneous crystalline environment.
The lack of target horizons and disappointing migration and
stacking results were direct consequences. Thus, for further
analysis a target area should be defined using the data from
the ISO-89 surface (3D) campaign, which are not yet avail-
able. At least one additional MSP-experiment at greater
depth (e.g. 6000 m) might be necessary to update the loca-
tion of the deeper discontinuities.
u
o
[ - a. ' 8 a ' ' 8 - a I
o .: ,
i ' - a 8 8 a
I : 8 a 8 : I
I I I I
w
KEMPER AND HARJES
E
Fig. 10. Seismic surface WE-profil, unmigrated stack. The dipping reflection
element can be related to the reflection R1 seen on the T-component (Figure
8).
o
-lOOO
-2000
-3000
-5000
-6000
-7000
-8000
-9000
SW NE
Dip
-3000 -2000 -1000 0 1000 2000 3000 4000 5000 6000 7000
offset Ira]
: :
:
: : : :

.

-o.6 ... ........ ! ........ i ..... i ..... i ................ i ....................
:

-o.8 .4 .................... ! ...................... ! .......
-1.4 Traveltimes from model
* Traveltimes from data
-1.6
-3000-2600-100 ; ld00 2d00 3(30 4000 500 6000
offset [m]
Fig. 11. Results of kinematic forward modeling for reflection element R1. For explanation see text.
SW NE
F.L.
OFFSET 8000
DEPTH
3000 [m]
I,
!
m - 454m i
3590 I
Dip 30' I
6195m_+ 125m
I 7100m+ 150m
7600 m + 200 m
lOOO
2000
3000
4000
6000
7000
- 8O00
Fig. 12. Schematic of mapped elements at KTB. 'F.L.' indicates the Franconian line, a much debated and major
fault zone (Franke, i989).
References (Eds.), The German Continental Deep Drilling Program
Ahmed H., P.B. Dillon, S.E. Johnstad and C.D. Johnston, (KTB), Springer Verlag, Berlin, 37-54, 1989.
Northem Viking Graben multilevel three-component Harjes, H.-P., M. Janik and M. Kemper, Moving So.urce
walkaway V SPs - a case history. First Break 4, NO. 10,
9-27, 1986.
DEKORP Research Group, Results of the DEKORP4/
KTB Oberpfalz deep seismic reflection investigations.
J. Geophys. 62,69-101, 1988.
Dillon, P.B. and R.C. Thomson, Offset source VSP surveys
and their image reconstruction. Geophysical Prospecting
32,790-811, 1984.
Profiling - A link between KTB borehole data and
seismic surface measurements, KTB Ret>ort 90-6b , _
Niedersichsisches Landesamt ftir Bodenforschung,
Hannover, 137-155, 1990.
Janik, M., Die strahlenseismische Wellenftrontenmigration
und ihre Anwendbarkeit auf MSP-Sektionen aus kristal-
linen Regionen, Diplomarbeit, Institut ftir Geophysik,
Ruhr-Universitit Bochum, 1990.
Franke, W., The geological framework of the KTB drill Stiller, M., 3-D Vertical Incidence Seismic Reflection
site, Oberpfalz. In: R. Emmennann, J. Wohlenberg Survey at the KTB Location, Oberpfalz. This volume.
3-D WIDE-ANGLE INVESTIGATIONS IN THE KTB SURROUNDINGS AS PART OF THE
"INTEGRATED SEISMICS OBERPFALZ 1989 (IS089)", FIRST RESULTS
Helmut Gebrande, Michael Bopp, Michael Meichelb6ck, Peter Neurieder
Institut f6r Allgemeine und Angewandte Geophysik, Theresienstr. 41/IV, 8000 M6nchen 2, FRG
Abstract. The high reflectivity of the upper and pects of wide-angle reflectivity, seismic wave ve-
middle crust of the Oberpfalz, both for near verti- locities and some fundamental observations of seis-
cal and wide-angle reflections, was one reason for mic anisotropy in the KTB surroundings. Except for
selecting this area as the site for the German Con- some more general informations on crustal structure
tinental Deep Drilling Program KTB. of the Oberpfalz derived from large-scale seismic
A peculiarity in the middle crust beneath the refraction measurements in the 1960's and early
KTB site is the so-called Erbendorf Body (EB) 1970's [Giese 1976; Peters 1974] most knowledge
giving rise to extremely strong wide-angle reflec- available before IS089 (Integrated Seismics Ober-
tions. Its position right on the border between two pfalz 1989) originated from KTB pre-site surveys
tectonometamorphic units of the Variscan fold belt, [DEKORP Research Group 1988, Gebrande et al. 1989,
unusual high P-wave velocities (over 7.0 km/s) in Schmoll et al. 1989]. The most outstanding feature
its lower part at 11 to 14 km depth, and associated in the wide-angle domain is the so-called Erbendorf
dipping reflectors suggest a possible lower crustal Body (EB), a high-velocity body at 11 to 14 km
origin of the EB. A specially designed and so far depth producing extremely strong reflections in the
unique wide-angle 3D-survey was carried out as part 40 to 60 km distance range. Detailed information on
of the program "Integrated Seismics Oberpfalz 89" its P-wave velocities, shape and NW-SE extent were
for investigating the spatial extent, the velocity provided by a systematic, multiply covered wide-
distribution and the internal structure of the EB. angle survey along the DEKORP4 line in 1985.
The first results show clearly that the EB is not a The latest result of this survey is a wide-angle
local phenomenon beneath the DEKORP4 line, but that image of the earth crust in a NW-SE section through
it exists, yet with remarkable complexities, be- the KTB site (Fig. 1). It was obtained by applica-
neath the whole covered area between the Franconian tion of the 2D isochron-migration process on wide-
Line and the Falkenberg granite complex. angle data from the 41 to 90 km offset range
The wide-angle shots were also recorded by four [Schmidt 1990]. 95 shot recordings of the contrac-
3-component geophones in the KTB pilot hole at 3195 tor's 200 channel spread operated from 41 to 58 km
to 3295 m depth and provided convincing evidence offset and of 24 3-channel MARS stations operated
for S-wave splitting in the upper crust SE of the by universities at 60 to 90 km offset were jointly
KTB location. This anisotropy effect seems to be and consistently processed to provide Fig. 1. In
related to the overall strike and dip of rock fo-
many respects (Moho, lower crust, SE dipping re-
liation in the zone of Erbendorf-VohenstrauB (ZEV). flectors) the picture is conformable to the steep-
angle migration result [Schmoll et al. 1989]. In
1. Introduction spite of the general weaker structural resolving
power of wide-angle reflections, some special
It is one of the major goals of the Continental structures are imaged even more emphatically due to
Deep Drilling Program KTB to contribute to a better their incidence dependent reflection coefficients.
understanding of the nature of seismic reflectors This is true, e.g., for the EB, which is the con-
in the subsedimentary continental crust. The high spicuous wedge-shaped body at 3.5 to 5 s two-way
reflectivity of the upper and middle crust at the time (TWT) beneath the KTB site (arrow in Fig. 1).
KTB site, both in the vertical incidence as well as Different methods of velocity analysis and model
in the wide-angle domain [DEKORP Research Group, calculations have consistently revealed P-wave ve-
1988], offers promising preconditions for this ob- locities up to or even over 7.0 km/s in the lower
jective. The present report deals mainly with as- part of the EB [DEKORP Research Group 1988,
Gebrande et al. 1989].
These unusual properties of the EB, together
ContinenmlLiosphere: Deep Seismic Refltions with its likewise pronounced steep-angle reflec-
Geodynamics22 tivity, have given rise to several speculations on
o 1991AmeficanOeophysic Union its lithological composition and its geodynamic
147
148 3-D WIDE-ANGLE INVESTIGATIONS IN THE KTB SURROUNDINGS
NW KTB8501 KTB 8502 i
HP-KH '70 80' 90 KTB 100
T
KTB8503 SE
I
110 120 130 14.0
i
I, I I I o
DEKORP4: ISOCHRON !'IIORRTION
OFFSET RRNOE: 41 - 90 Krl
DEPTH
20
30
km
Fig. 1. Seismic image o the earth crust along a NW/SE-section through the KTB site obtained by 2D iso-
chron-migration of DEKORP4 wide-angle data from the 41 to 90 km offset range [after SCHMIDT, 1990]. The
Erbendorf Body (EB) appears as a wedge-shaDed structure in the middle crust beneath the KTB location.
significance. in general the EB is expected to be a mer DEKORP4 line (Fig. 3) and once in SW-NE direc-
two-dimensional SW-NE striking structure marking tion with shot-points 201 to 221 near to the pre-
the Saxothuringian/Moldanubian suture at depth site reflection line KTB8504 [DEKORP Research Group
[Franke 1989]. On the other side, there are many 1988]. For the latter setting shot and receiver
indications, e.g., from older refraction lines lines have been rotated by 90 compared to Fig. 2.
[Gebrande et al. 1989] and from the KTB reflection Both configurations cover the same CMP rectangle
lines [DEKORP Research Group 1988], that lateral with the KTB drill site in its centre.
heterogeneities perpendicular to the DEKORP4 line
may be of comparable importance to in-line hetero-
,,
geneities. In principle it could even not be ruled
out, from the solely one-dimensional measurements
along DEKORP4, that the observed strong wide-angle
reflection amplitudes of the EB may be partly due
to focusing by off-line structures, and that the
wide-angle image of Fig.1 may contain significant
side effects. Therefore, a consequent 3D wide-angle
survey was planned and realized as part of the pro-
gram "Integrated Seismics Oberpfalz 1989" (IS089)
devoted to a comprehensive seismic investigation of
the KTB surroundings.
2. Layout and Realization of the wide-angle
3D-survey
The basic principle of the realized 3D wide-
angle survey is shown in Fig.2. The scheme consists
of a linear sequence of 20 shot-points with one km
spacing and a transversely oriented receiver
spread. It provides single coverage in a rectangle
midway between shots and receivers, the edges of
which are half as long as the shot and receiver
spreads, respectively. It is advisable to choose
the mean shot-to-receiver offset equal to the
ot line
*I MP-area
transverse spread
[] in-line spread (PCM)
[]
critical distance of the reflector to be mapped, Fig. 2. Basic scheme for 3D wide-angle measure-
which is about 40 km for the EB. A supplementary ments' recording a linear shot series by a trans-
in-line receiver spread has been added for improved verse oriented receiver spread of equal length at
velocity resolution. about the critical distance provides single CMP
The basic scheme of Fig.2 has been applied twice coverage in a square area in between. The supple-
for the IS089 wide-angle survey, once in SE-NW di- mentary in-line spread serves for improved velocity
rection with the shot-points 101 to 120 on the for- resolution.
GEBRANDE ET AL. 149
; Bo
- ,.,.,
Hof '"'"r- .. J/'" o
o ;' '\ "' Kraslice N
, .| /.--'
,,_, Ell'- ....
k, ,., CSFR
FRG '", i A5
/ B2
; A4
--,,
o
I . "'.. Cheb . A2
,\: '- B3 I
'-- A, 1
" 224/6
201 . KTB
;223
i 120 222 ', '
I . '% I lO km
% Weid:h' %)) "'""i
[=221, .. ...
% lol
.
In addition, all shots were also recorded in the
KTB pilot hole by 3-component geophones. A geophone
chain with 4 active sensors was used during shot
series 101 to 120, and a single 3-component sensor
during the rest of the wide-angle experiment. De-
tails on position and orientation are given in Ta-
ble 1.
3. Data Processing and First Results
3.1Preorocessing
Due to the different recording instruments used
in the wide-angle survey, extensive preprocessing
was required to generate a homogeneous data set. We
had decided to process the whole data volume with
the DISCO software package [C.S.D. 1988], a standard
processing system in exploration geophysics, par-
tially modified and extended for wide-angle appli-
cations. Therefore some extra effort was neccessary
for the conversion of the non-standard PCM data
into the internal DISCO format. As a first step to
this end, PCM data were already transformed during
the experiment into a slightly modified ESSTF-for-
mat (European Standard Seismic Tape Format) and
stored on computer-readable 9-track magnetic tape
by means of the DEKORP PCM-playback computer, which
was installed in the field headquarter; simultane-
ously generated analogue outputs were used for
quality control.
Fig. 3. Location map of IS089 wide-angle and re- For the straightforward input of the ESSTF data
fraction measurements' 101-120, 201-227 shotpoints into the DISCO system an input-module "PCM" has
in the FRG; BO shot in the GDR; B1-B5, A1-A5 shot- been developed, which demultiplexes the block-mul-
points in the CSFR; unnumbered symbols are loca- tiplexed ESSTF data, converts them to DISCO format,
tions of quarry blasts. Lines Reflexl and Reflex2 cuts out the desired time windows from 0 to 32 s
are 312-channel receiver spreads formed by 5 uni- traveltime, and adds 62 shot and receiver specific
versity and DEKORP reflection units; perpendicular information headers to each seismic trace. Finally,
spreads PCM1 and PCM2 consist of 20 PCM stations data were resampled to 4 ms to become compatible
with a total of 120 channels. with multichannel reflection data.
The receiver spreads normal to the shot lines,
termed Reflexl and Reflex2 in the location map
(Fig. 3):., were observed by means of 5 university-
and DEKORP-owned reflection instruments with a to-
tal of 312 channels and 80 m receiver group spac-
ing. The perpendicular spreads PCM1 and PCM2 were
formed by 20 PCM-stations with a total of 120 chan-
nels, half of which were used for three-component
recordings.
Field work was carried out from 9th to 21th
Editing was rather time consuming and included
the following steps:
- correction of polarity reversals due to non-
standard polarities of some borrowed geophones,
- despiking, especially for data from the first
rainy week,
- elimination of dead or strongly disturbed
traces,
- zeroing of strong transient disturbances, e.g.
by cars crossing cables,
- mute until about 300 ms before first arrivals,
Furthermore the data were corrected for the
October, 1989. It was integrated in the ISO89 3D different receiver responses due to the use of 1.0,
steep-angle survey and closely coordinated and syn- 2.0 and 4.5 Hz geophones and different recording
chronized with a simultaneous refraction survey by systems. An uniform receiver response (natural fre-
GEOFYZIKA BRNO and ZIPE, Potsdam, in the adjacent quency 4.5 Hz, damping 0.62, generator constant 100
territories of the CSFR and the former GDR. In that Vs/m) was simulated by recursive simulation filter-
way, shots fired in the Oberpfalz could also be ing [Seidl 1980]. Finally, a 6 to 40 Hz band-pass
recorded in northwestern Bohemia, southern Saxony filter was applied to all data. The efficiency of
and vice-versa. Altogether 64 borehole shots (47 in these processing steps is exemplified by Fig. 4,
the FRG, 17 in the CSFR) and 10 quarry blasts (9 in top and centre.
the FRG and 1 in the former GDR) were observed. The Fig. 4, centre, also reveals, and this is con-
contractor's recording unit for the steep-angle 3D- firmed by comparative inspections of other shot re-
survey recorded the shots as well and served as cords, that static corrections are a matter not to
master unit for radio controlled shot release and be ignored in wide-angle seismics. Because of miss-
remote start of the other reflection units. ing nearby shots, however, they cannot be evaluated
150 3-D WIDE-ANGLE INVESTIGATIONS IN THE KTB SUPOUNDINGS
;0 40 Offset (km)
cance of the final results. That's why we feel it
worthwhile to do all the tedious preprocessing
briefly described in this subsection.
As a typical shot record along spread Reflex2,
Fig. 4 already anticipates qualitatively some main
results of the present investigation:
1. The strong wide-angle reflections with 1 to 2 s
reduced traveltime, which were known so far
only from the NW-SE-striking DEKORP4 line and
by which the Erbendorf Body (EB) was originally
defined, appear also for SW-NE propagated
waves over the whole length of spread
Reflex2 (Fig.3).
2. The EB wide-angle reflections show up again
with large amplitudes and apparent velocities
over 7.0 km/s.
This proves that the EB is not a narrow local
feature beneath the DEKORP4 line and that its large
reflection amplitudes and apparent velocities are
not just the results of focusing by off-line het-
erogeneities. These conclusions are corroborated by
all other IS089 data processed and analysed so far.
3.2 Simulated Constant-Offset (COF) and Zero-Offset
ZOF) Sections
For a first qualitative assessment of horizontal
structural' variations we use simulated constant-
offset (COF) sections. They are obtained from sin-
gle-shot sections or CMP gathers by partial dynamic
corrections to a specified target offset (e.g. 40km
in Fig. 5); traveltime differences due to variable
Fig. 4. Example for signal improvement by consecu- shot-to-receiver offsets are thereby eliminated. We
tire preprocessing steps. Top: raw data of shot 203 have implemented the COF simulation process as a
recorded by spread Reflex2. Centre: same data after NMO correction with a subsequent inverse NMO cot-
editing, mute, restitution and 6-40 Hz band-pass rection to the target offset.
filtering. Bottom: Same data as above with static In principle, any target offset would serve the
corrections applied. purpose, but a reference value near to the mean
shot-to-receiver offset has the advantage of least
signal stretching and of smallest distortions in
by standard refraction interpretation of first at- case of inaccurate NMO velocities. For that reason
rivals. Therefore we designed a two-step correla- 40 km. has been used as the COF target offset in
tion method for the determination of short-wave- Fig. 5. Examples are shown for three shots (205,
length statics and applied it, for the sake of 213,220) recorded on spread Reflex2, which are rep-
higher resolution, to resampled data (from 4 to 1 resentative for the whole series of shots 201 to
ms). Based on 500 ms time windows around the first 221, by which the KTB surroundings were scanned in
arrivals, cross-correlation functions of adjacent SW-NE direction. For a first overview COF simula-
traces were stacked over 5 neighbouring shotpoints tions were performed with a constant velocity of
and a delay-time curve was defined along the spread 5.95 km/s, which - according to the DEKORP4 wide-
by adding the individual delays of the correlation angle results - is the best estimate of the average
maxima picked by the interpreter. In this way 4 v value down to the EB.
delay-time curves were obtained for each spread. In Fig. 5 the wide-angle reflections from the EB
After removal of a moving average over 5 km from appear around 8 s modified time with great ampli-
each curve, the remaining delay-time curves were in tudes and widespread continuity. This is very sur-
almost perfect agreement and their means were ap- prising in view of the generally short reflection
plied as first-order static corrections. After elements observed in the crystalline crust by
that, the procedure was repeated with a narrower steep-angle seismics, e.g. on profile KTB8502
time window of 100 ms and the resulting second-or- [DEKORP Research Group 1988] which partially coin-
der delays were added to the first-order static cides with our spread Reflex2. The EB reflections
corrections. The effect of applying these correc- form a band which broadens from NW to SE, i.e. from
tions is illustrated by comparison of Fig. 4, cen- shotpoint 205 towards 220. In the SE-part (shot
ter and bottom. The general improvement of the sig- 220) the EB seems to split up in an upper, possibly
nal-to-noise ratio from top to bottom in Fig. 4 is overthrusted and a lower segment, which descends to
remarkable and will certainly increase the signill- the SE.
GEBRANDE ET AL. 151
sw DEKORP/.
CMP4 ;)387 ;)362 2337 23t2 ;)287 :3262 ;)237 22t2 287 2162 2137
7 : ,,.: SHOT 220
!':- , ...... ::'.','i:Ji:,.-.-.'.-.'7';;;;,:',',:'".,'.;;'.'.',',;;,,;;',,..,, .................. ,, , , .......... ,,
.K!i:.. , ' ,,,,, .'.T.:::...:"._ .... .-.-',::.';; ..... : ..... ::::.".,'.:,'T',;:;:;'F' ,, ,. ' ' ,, .
I '.'--.:'-':=,,,',7-.:::'-L.:;.. '" '. ...... '.''" ,, ...;.::::L;'.".::;'7,: ::::': .... " '""':'::':'" "' r,..'.'..;:,L;'.. ::: :'., ' ...... : ....... ::: ...... '::::: '" '
?.:.:! ;.:,...;..,-_:,.:.,.:.. - ---..-..._.::!. .-- ........ . ............. ,-.,: .... ,,
., ?---:.:.'..:-_" .,:-,:.
2 8 "-"" ' ', ,' ,' ....... ,
-
I- . '.'-' .- ........... .:... ,....,,,,.,:.,,. ,.., ,,,-.,..,.
,,,
'. ,, ..,: 't,,,-.,,:,,,,." i;';,;;,.:,;...-... ...... --.;;.
' :,..' :..',-'L; ..... .,.-t,,..' .... ':;.;;::;',';,',:.::'. :,'i.:---"::,'..-.,,".,
0 : ..... ,"".' .... :'.';; ' ,,,,,,a,,' 7' 'P .... ' ..,,, .
...*,'"."-'"*"'I;"""-'".1"= ' ,N,, ,, , , , ,,
Fig. 5. Three shot records along spread Re1ex2 transformed into constant-offset sections (COF) by partial
NMO corrections' reference offset is 40 km. The typical wide-angle reflections from the Erbendorf Body
appear around 8 s modified time. An increase in traveltime from NW to SE and a splitting up of reflections
in the SE-part (shot 220) is observed.
Fig. 6 displays the same data as simulated zero- rect comparability with steep-angle seismic results
offset (ZOF) sections. A ZOF section may be consid- and of simple depth interpretation (with the risk
ered as a special COF section with total NMO cor- of oversimplifications, however). If applied to
rections applied, i.e. with target offset zero. far-offset wide-angle data, serious signal distor-
Simulated ZOF sections provide the benefit of di- tion arises, because the NMO corrections revoke the
sw DEKORP/+
NE
2262 223'/ 22t2 2187 2162 2137 2112
SHOT 205
3-
, ,,,,,:,, , , ...... ,, ' ,, ,, .. ,_. ................ , SHOT 220
......... ::::1 ' Ii, l` . ., .... ' 'I , ' ........... I ............. I .............. ' . I , I
,i "'1 i Ii1",, I ! ' ' ,I , ' ..... II'I'IH'" h,, i, I , liH, ll j ' ' ')' III i i , , i Illl` t , , ii ii ,,i ' i ,
,I""q"lMl,l'lll. II i , - . "' 114'I I,l`#111.11- '1'111"1411'1"I'. .......... i,,11tllJ I1, l`ll I "1 II , i . HI I h, . , , i,,I, i,.,i I ,
,,, , .. .................. ,,,'.,",' ,;',,',,,, ..... .:, .......... '", ,,,..uq,%,"..:"-' ..... -: .......... ;-::.,k._-'._ ', , , ..... , ....... 1 ..... %,.k",,.', ..... ', ,, ..... , ', ........
#., ,.,.,i,,, i,, .,... ,,' * .. ' i
:.,,: ..... ,"':. ............. ,' '",,,,' ,,, ,,::'. .... ' "'.:,,,':,.. ' .... ..,,... . ,d,---=":--
,. :"..'.,,,,,, ..... ,.,.,,,;,,, ,,,, .... .... : ........... , ..... .. ..... ,,,,,, ..... .,,,,,::,,', .,.,,,,,..., ,, .... Z_,.r,.-,...,.,,.,,,, ",?.",,.,,,,.,.,..
..... :-',:', .... .,,,':..tt,,t,,..,,';,.:!'. ................ :',. ,:,'.;,,; '"
I
Fig. 6. Simulated zero-offset sections for the same shot records as in Fig. 5 obtained by NMO corrections
with an average velocity of 5.95 km/s' Pg-onsets, not being roerly transformed by NMO corrections, have
been suppressed by a ram mute.
GEBRANDE ET AL. 153
time-convergency of adjacent reflections with in- covered, from the Franconian Line to the Falkenberg
creasing offset and thereby yield considerable sig- granite complex, with the same characteristics as
nal stretching at zero offset. The Pg-onsets which earlier along the DEKORP4 line.
are most affected by stretching, and which moreover IS089 has also provided evidence that the EB is
are kinematically not properly transformed by stan- not a singularity in the middle crust of the imme-
dard NMO corrections, have therefore been sup- diate KTB surroundings. Similar bodies may exist in
pressed in Fig. 6 by gradual muting. Structures other places of the Saxothuringian/Moldanubian
with less than about 3.5 s zero-offset two-way time boundary zone. Fig. 7 shows a simulated ZOF-section
can not be resolved by our wide-angle observations of two stacked shots from shotpoints B2/CSFR re-
starting at 25 km minimum offset. corded along spread Reflexl (Fig. 3) with a very
According to the common-midpoint concept, how- sharp wide-angle reflection of about the same TWT
ever, the displayed reflections can be assigned to as the EB but with otherwise different signature.
!2.5 km long, SW-NE trending midpoint lines begin- The reflector seems to dip from the SE margin of
ning at about the Franconian Line and crossing the the Fichtelgebirge towards the Tertiary Cheb Basin.
former DEKORP4 line 3.5 km NW, 1.0 km SE and 4.2 km Parallel to this a marked reflection from the top
SE, respectively, from the KTB location. Comparing of the lower crust appears. This might indicate a
the three ZOF simulations, again obtained with Tertiary rejuvenation of the Variscan crust at the
VNMO=5.95 km/s, the dipping of the base of the EB transition to the Cheb Rift.
towards SE and its bisection in the SE are clearly Simulated COF and ZOF sections are only very
discernible. These observations are corroborated by rough first approximations for imaging crustal
the whole sequence of ZOF sections for shotpoints structures by wide-angle reflections and should be
201 to 221, interpreted cautiously. For a detailed and reliable
The main point to be stressed by Figs. 5 and 6 imaging of strong horizontal heterogeneities the
is that the EB reflections have been observed per- migration process is even more important in wide-
pendicular to the DEKORP4 line in the whole area angle seismics than it is in steep-angle seismics
SW CHEB BASIN NE
'Offset (m) ............ ' '
2.00
3.00
4.00
5.00
6.0B
7.00
8.00
2.00
31:',i'
3.00
4.00
6.00
7.99
8.00
Fig, 7. Zero-offset-corrected stacked section of under Reflexl and seems to dip towards the Tertiary
two shots from location B2 (CSFR) recorded along Cheb basin' a change of crustal style in that di-
spread Reflexl. A very sharp wide-angle reflector rection is indicated by the appearance of strong
with TWT comparable to the Erbendorf Body appears reflections from the top of the lower crust.
[Gebrande et al. 1989]. A suitable 3D-version of the up by the most coherent signals and thereby repre-
isochron wide-angle migration method - Fig.1 is the sents the dominant velocity-depth distribution
result of a 2D-version - is presently being devel- half-way between shotpoint and receivers. Horizon-
oped. For its hopefully successful application, a tal dark stripes marked by arrows in Fig. 8 may be
comprehensive 3D velocity model will be a prerequi- due to subcritical reflections and indicate possi-
site. ble depths of first-order discontinuities.
The velocity-depth function picked by the geo-
3.3 Velocity Analyses physicist from Fig. 8 and controlled by forward
modeling is shown, together with the result from
The independent control or improvement, respec- the XTV-inversion of the same data, and with tau-p-
tively, of the 2D velocity model derived from inversions for shots 225/227 in Fig. 9. A gradient
DEKORP4 data [DEKORP-Research Group 1988; Gebrande zone with velocities from 6.3 to 7 km/sec around
et al. 1989] and its extension to a 3D-model con- 2.5 km depth (marked by the asterisk in Fig. 8)
stitute the major targets of the ISO89 wide-angle could not be verified by forward modeling, but
survey. Seismic wave velocities are required for turned out to be an artifact of a near-surface
reliable structural imaging by steep- or wide-angle heterogeneity; it was therefore not included in the
migration techniques as well as for the petrophysi- velocity-depth function in Fig. 9. The tau-p- and
cal, lithological and finally geodynamic interpre- the XTV-method provide very similar results down to
tations. a depth of about 3 km, where a low-velocity layer
We started velocity analysis of ISO89 data with is expected according to both solutions. The V(z)
spread Reflex2 because in this case the wide-angle function derived from shots 224/226 is expected to
data set includes observations down to zero offset, be valid in the area of the Falkenberg granite com-
which are important for the velocity resolution in plex about 10 km NE from the KTB site.
the upper crust. We used two independent methods:
- a special variant of the tau-p-inversion
technique [Meichelbck 1988] and
- the so-called XTV-inversion [Gebrande et al.
1985], which is essentially a traveltime based
stripping technique.
Fig. 8 shows an example for the tau-p method
applied to stacked shot records 224 and 226 (Fig.
3). By means of an automatic picking algorithm,
manifold paths are traced through the slantstack
transformed wavefield and converted to the V-z-do-
main by the tau-sum method [Diebold et al. 1981].
The stacked amplitudes are transferred from the
tau-p into the V-z-domain and plotted as a filled-
area contour-map, in which the dark crest is built
5.0 5.5 6.0 6.5 velocity
(km/sec)
1
4
5 6 7 velocity
.. I I I
.... . (km/sec)
224/226 ( tau-p-inv. )
224/226 ( XTV-inv. )
225/227-SW ( tau-p-inv. )
......... 225/227-NE ( tau-p-inv. )
depth ! \
,(km )
Fig. 9. Velocity-depth distributions derived by
the tau-p and the XTV-method for the SW and the NE
parts of line Reflex2.
depth
(km)
Fig. 8. Example for automatic velocity-depth de-
The shots 225/227 were situated in a split-
spread position in the centre of spread Reflex2
near to a boundary between granites in the SW and
mica schists in the NE and have provided signifi-
cantly different V(z) functions for both directions
(Fig. 9). The depth of penetration is only 2 km for
the split-spread data.
With regard to the EB, the tau-p inversion of
termination by the modified tau-p wave-field trans- the shot 211, recorded on spread Reflex2 is of spe-
formation after Meichelbfick [1988]. The dark crest cial interest because the corresponding midpoints
is built up by the most coherent signal amplitudes to which the results of one-dimensional velocity
of the shot records 224 and 226 and represents the determinations should be attributed cross the KTB
dominant velocity-depth distribution for the area site (Fig. 10). The results are limited by missing
of the Falkenberg granite about 10 km NE of the KTB observations in the 0 to 25 km offset range but,
site. Arrows mark depths of probable undercritical nevertheless, the V(z) function seems to be well
reflections. The asterisk indicates an artifact constrained at depths greater 5 km by clear first
produced by a local lateral inhomogeneity. and later arrivals recorded on spread Reflex2. A1-
GEBRANDE ET AL. 155
4 5 6 7 velocity
that time. Details of geophone positions and orien-
tation of components are given in Table 1.
As a first step we tested our two-dimensional
isotropic velocity model derived from DEKORP4 wide-
angle data against the ISO89 down-hole observa-
tions. Fig. 11 shows the P-wave section (a) in re-
duced time for the shot series 101 to 120 recorded
by the geophone at 3295 m depth in comparison with
traveltimes (b) predicted from the model (c). The
observed arrival times are about 0.1 s earlier than
predicted, but the average apparent velocities co-
incide rather well. The more or less horizontal
incidence of the seismic rays predicted by the
model calculations fits with the results of polari-
zation analyses of the 3-component recordings,
which were performed by calculating the direction
of the principal axis with the largest eigenvalue
of the covariance matrix [Kanasewich, 1973].
Surprisingly, the azimuth of the P-wave polari-
Fig. 10. Comparison of velocity-depth functions
for the KTB vicinity, derived by the tau-p method zation proved to be in close agreement with the
from the single shot 211 recorded on line Reflex2 shot-to-receiver azimuth (Table 1). This means that
(SW-NE azimuth) and from CMP-sorted seismograms ob- either there is no significant P-wave anisotropy or
served 1985 along DEKORP4 (NW-SE azimuth). The ex- the azimuth of wave propagation, i.e. the SE-NW
istence of P-wave velocities around 7.0 km/s at 12 direction, is close to a specific direction of pure
to 14 km depth is corroborated by the ISO89 data. longitudinal polarization in an anisotropic upper
Velocity resolution in the upper 5 km is poor due crust. Larger differences between P-wave polariza-
to the rather large offset minimum of 25 km. tion and propagation directions for other shot-
points, e.g. for quarries NNE from the KTB site,
are in fayour of the second explanation, and this
becomes indeed evident from the investigation of
shear-waves. In Fig. 12 typical S-wave signals of a
shot (no. 113) recorded in three different depth
TABLE 1' Receiver locations of the 3-component
borehole geophone chain and orientation of its
horizontal components according to compass
readouts and P-wave polarization
lowance for sediments beneath the sHotpoint was
made by a static correction of 300 ms prior to the
slantstack. P-wave velocities around 7.0 km/s in
the 12 to 14 km depth range corroborate former re-
sults froin the DEKORP4 line and confirm that they
are not an artifact by some strange side-effect.
The unusual high wave velocities as well as the
high wide-angle reflectivity of the Erbendorf Body
seem to be regional phenomena of the Earth's crust
in the surroundings of KTB. The pronounced low-re-
locity zone obtained for the shot 211 is a surprise depth component horizontal receiver-azimuth [NE]
and needs further verification. [m] from compass from polarization
readout analysis
3.4 Polarization Anal.vsis and Shear-Wave-SDlittin
The operation of 3-component geophones recording
nearly all seismic sources of the IS089 project in Z
the KTB pilot hole offered the unique possibility 3195 HI
to analyse traveltime and wave-fields of the wide- H2
angle shots at depth. In view of convincing evi-
dence for ultrasonic P- and S-wave anisotropy in z
almost all core samples from the KTB pilot hole 3220 HI
H2
[Lippmann et al. 1989; Zang et al. 1989] the ques-
tion of large scale seismic in-situ anisotropy in Z
the KTB surroundings calls for an answer. It's a
3245 Hi
question of fundamental interest well matching to H2
the KTB project, but it is also a question of great
practical importance because significant anisotropy z
would require serious modifications in seismic 3270 HI
processing. H2
The down-hole recordings from the first week of
the ISO89 wide-angle measurements with shotpoints z
101 to 120 are especially appropriate for an explo- 3295 HI
ration of anisotropy effects because 4 (of 5) H2
3-component geophones were in perfect operation at
52 59 +/- 5
142 149 +/- 5
H2 defectiv
19 26 +/- 4
109 116 +/- 4
24 24 +/- 2
114 114 +/- 2
248 249 +/- 4
338 339 +/- 4
P-ra
incid,
DEPTH
195
1245
127(
1295
[rn]
up
90
down
81
86
84
82
87
80
86
78
RADIAL HORIZONTAL COMPONENT (depth 3295 m, Vred=6km/s)
/ // / / 1 -r 1 1' I t I I I I 1 f
! i 9 _ ' SHOT
i I I
10 20 30
HORIZONTAL OFFSET [km]
. 4 4 i- - + :- + + + + + + T + 4- + t 4
110 120
-.'-':.:.::::; .-..-.'z,-;;."..-.,..;,., ..- . .. -.- - .... -'v;::.':?':,"- ;' '
I
lOO 11o
PROFILE-COORDINATE rkm]
v = 6.1
v = 5.8 km/s
120
Fig. 11. Comparison of observed (a,d) and predicted (b,c) traveltimes and angles of incidence for 3-compo-
nent down-hole measurements at 3295 m depth.
positions are plotted in the vertical plane normal and this is also true for most of the other shot
to wave propagation and P-wave polarization. The and receiver positions. This delay cannot be the
components as well as the hodographs are displayed result of a very local anisotropy, e.g. merely at
providing clear evidence for birefringence (shear- the source or receiver. With a reasonable aniso-
wave splitting)'The S-wave trains start consis- tropy coefficient of 5 to 10%, the waves must
tently with linear polarization dipping steeply (65 propagate through at least several km of aniso-
degree) to the SW nd continue with more elliptical tropic rocks to bring forth the observed delay.
particle motion, which seems to be due to a super- Fig. 14 shows record sections of the decomposed
imposed orthogonally polarized shear-wave with fast and slow shear-waves for the total shot series
slower propagation velocity. All records exhibit recorded at 3295 m depth. Shear-wave splitting is
this general behaviour. evident, but there is no linear increase of delay-
Decomposition of the 3zcomponent seismograms times with shot-to-receiver distances, as would be
into mutually orthogonal components parallel and expected for a medium with constant anisotropy.
normal to the P-wave polarization direction pro Fluctuations seem to be of a scale of about 10 km;
vides an almost perfect separation of P- and S- at the scale of seismic field experiments, however,
waves as demonstrated by Fig. 13a. By an appropri- they are not simply averaged out.
ate additional rotation of the transverse compo- From comparison with laboratory investigations
nents in the normal plane the fast and the slow on KTB core samples [Lippmann et al. 1989; Zang et
shear-waves can also be separated. This is shown in al. 1989] and from theoretical considerations it is
Fig. 13b for the four records of shotpoint no. 113. expected that rock foliation is the main cause for
It becomes evident that the delay of the slow the observed anisotropy and that the polarization
shear-wave is more than one period (about 100 ms), direction of the fast S-wave is in the plane of
GEBRANDE ET AL. 157
VER'nCAL (Z) vs. TRANSVERSE HORIZONTAL (X) COMPONENT
(shot 113, depth 3245 m, bandpass 6-15 I-Iz)
SW / NE
! x
Z-KOHP
X-KOHP
VER'i3CAL (Z) vs. TRANSVERSE HORIZONTAL (X) COMPONENT
(shot 113, depth 3270 m, bandpass 6-15 Hz)
SW / NE
Z-KOHP ----- ---'- ,/ V / . l
X-KOHP .---"-
5 e.m 5..sa 51.17 37 5.33 TIME {MSEC)
a
b
VERTICAL (Z) vs. TRANSVERSE HORIZONTAL (X) COMPONENT
(shot 113, depth 3295 m, bandpass 6-15 Hz)
SW / NE
X-KOH '
-
TIME (NSEC) C
Fig. 12. S-wave components and S-wave hodographs S-wave predominate on the Z-component due to their
showing S-wave splitting. The arrivals of the fast steeply dipping polarization.
foliation. This would imply that the dipping po- tarization analysis to resolve textural lithologic
larization of the fast shear-wave reflects the properties. With consequent methodical development
dominant inclination of foliation of gneisses in it might become a diagnostic tool for the seismolo-
the zone of Erbendorf-VohenstrauB SE of the KTB gist as useful as the polarization microscope is
location. This demonstrates the potential of po- for the petrologist.
158
3-D WIDE-ANGLE INVESTIGATIONS IN THE KTB SURROUNDINGS
ooo
[s]
I
o
o
o
o
o .
o o
o
.r4
o
o
o D.

GEBRANDE ET AL. 159
SHOT 12 119118117 116 115 114 11311 111 118 le9 le8 le7 186 15 14 183 le 181
I 0 20 0
.ORIZL OS
FAST SHEAR WAVE (depth 3295 m, bandpass 6-15Hz, Vred--3.5krn/s)
a -
SHOT IPO 119118117 116 IlS 114 113112 111 110 189 188 107 106 ItS 184 103 102 101
I ' I I
10 20 30
HORIZONTAL OFFSET [km]
SLOW SHEAR WAVE (depth 3295 m, bandpass 6-15Hz, Vred=3.5krn/s)
Fig. 14. Comparison of the separated fast (a) and
slow (b) shear-waves for all shotpoints recorded in
3295 m depth' shear-wave splitting is beyond all
doubt.
4. Conclusions
3D isochron migration method beeing presently de-
veloped.
From the 3-component recordings of the wide-angle
shots in the KTB pilot hole it became evident that
seismic anisotropy is an intrinsic property of the
upper Earth's crust in the KTB surroundings.
Anisotropy is not simply averaged out at the scale
of seismic wavelengths or field experiments and
will have to be considered for a more advanced
processing.
Acknowledgements. The funding of the investigations
by the Federal Ministry of Science and Technology
(BMFT), Bonn, is gratefully acknowledged. We appre-
ciate the continuous support by the DEKORP steering
committee and by H. Soffel, director of the In-
stitut fQr Allgemeine und Angewandte Geophysik,
MQnchen. Thanks for excellent cooperation during
field work are due to the teams from the Geophysi-
cal Institutes of Aachen, Clausthai, Karlsruhe,
M6nster, Z6rich and M6nchen, from the Geological
Survey of Lower Saxony, the Prakla-Seismos crew and
to the collegues from Geofyzika Brno.
References
C.S.D., CogniSeis Development Inc.: VAX-DISCO 7.3
User's Manual l&2, Houston, Texas, 1988
DEKORP Research Group, Results of the DEKORP4-KTB
Oberpfalz deep seismic reflection investigations.
J.Geophws.,62, 69-101, 1988.
Diebold, J.B., and P.L. Stoffa, The traveltime
equation, tau-p mapping and inversion of common
midpoint data. Geophysics, 46, 238-254, 1981
Franke, W. , The Geological Framework of the KTB
Drill Site Oberpfalz. In: Exploration of the
Continental Crust through Drilling, IV,
Emmetmann, R., Wohlenberg, J., eds., 37-54,
Springer-Verlag, Berlin, 1989
Gebrande, H., and H. Miller, Refraktionsseismik.
In: Angewandte Geowissenschaften - Band II,
Bender, F., ed., 226-260. Stuttgart, F. Enke
Verlag, 1985
Gebrande, H., Bopp, M., Neurieder, P. and Schmidt,
T., Crustal Structure in the Surroundings of the
KTB Drill Site as derived from Refraction and
Wide-Angle Seismic Observations. In: The German
Continental Deep Drilling Program (KTB),
Emmermann, R., Wohlenberg, J., eds., 151-176,
Giese, P., Results of data generalization. In:
Explosion seismology in Central Europe, Giese,
P., Prodehl, C. Stein, A., eds., 339-346,
ISO89 3D-wide-angle observations have confirmed the
existence of the Erbendorf Body (EB) as a strongly
reflecting high velocity structure at 11 to 14 km Kanasewich, E.R., Time Seouence Analysis in
depth beneath the KTB site. Furthermore they pro- Geophysics. University of Alberta Press, 1973
vided evidence that the EB is not a local feature Lippmann, E., B6cker, C., Huenges, E., Rauen, A.,
along the DEKORP4 line but that it extends in SW-NE Wolter, K., Soffel, H.C., Rock physical
direction, i.e. parallel to the suspected Saxo- properties: first results of the KTB-
thuringian/Moldanubian suture, at least from the field-labratory. Scientific Drilling, 1, 143-149,
Franconian Line to beneath the Falkenberg granite 1989
pluton. The complex internal structure of the EB Meichelb6ck, M., Tau-P-Inversion - mit Anwendungen
and its possible geodynamic significance will pre- auf DEKORP4-Weitwinkel- und Refraktionsmessungen.
sumably become clearer after the application of the Dipl.- Arbeit, Inst. f. Allg. u. Angew. Geophysik
der Ludwig-Maximilians-Universitfit Mfinchen,
119 pp., 1988
Peters, K., Ergebnisse der Gravimetrie im Bereich
der Mfnchberger Gneismasse und der Refyk-
tionsseismik lngs eines Profils fiber die
Gneismasse. Diss. Ludwig-Maximilians-Universitt
Mfnchen, 131 pp., 1974
Schmidt, T., Seismisches Abbilden durch Isochro-
Migration yon Weitwinkeldaten am BeisDiel des
DEKORP4-Proils. Diss. Ludwig-Maximilians-
Universitt, in preparation, 1990
Schmoll, J., Bittner, R., Dfirbaum, J., Heinrichs,
T., MeiBner, R., Reichert, C., Rfihl, T.,
Wiederhold, Oberpfalz Deep Seismic Reflection
Survey and Velocity Studies. In: Exploration of
the Continental Crust through Drilling, IV,
Emmermann, R., Wohlenberg, J., eds., 99-149,
Seidl, D., The Simulation Problem for Broad-Band
Seismograms. J. GeoDhws., 48, 84-93, 1980
Zang, A., Wolter, K., Berckhemer, H., Strain
recovery, microcracks and elastic anisotropy of
drill cores from KTB deep well. Scientific
Drilling, l, 115-126, 1989
CRUSTAL INDENTATION IN THE ALPS -
AN OVERVIEW OF REFLECTION SEISMIC PROFILING IN SWITZERLAND
Peter Heitzmann l, Walter Frei 2, Peter Lehner 3, & Paul Valasek 2
Abstract. Reflection seismic profiling across the
Swiss Alps displays the structural relationship of the
junction between the European and the African plates
with a wedge of African crust indented into the Euro-
pean one during Neo-Alpine deformation. This structural
configuration clearly postdates the Eo-/Meso-Alpine
thrusting of the African plate over the European one and
the imbrication of the thinned south European margin.
Introduction
In the framework of NFP-20 (National Research
Program on the Deep Structure of Switzerland - a multi-
disciplinary research program of the Swiss National
Science Foundation) geophysical and geological inve-
stigations are concentrated on three traverses across
the Swiss Alps (Pfiffner et al. 1990 a, b, Schmid et al.
1990, Bernoulli et al. 1990, Bertotti 1990). The main part
of the program consists of three reflection seismic pro-
files recorded in 1986 - 1990 connected with two refrac-
tion lines (Fig. 1). A third refraction line on the eastern
traverse was recorded within the European Geotraverse
(EGT) program.
Geological setting
The Alps mark the suture of the Late Creta-
ceous/Tertiary collision between the European and Afri-
can plates. From N to S the following main tectonic units
can be discerned (Fig. 1):
- The EUROPEAN FORELAND with the southward
thickening MOLASSE BASIN and the peripheral
JURA MOUNTAINS.
- The EXTERNAL MASSIFS (Aiguilles Rouges, Aar),
as crystalline complexes with their proper Mesozoic
cover are tectonically overthrusted by the sedimen-
tary HELVETIC NAPPES. Both units formed part of
the ancient southern European continental margin.
- The INTERNAL CENTRAL ALPS. Three domains
compose this part of the Alps:
(i) the PENNINIC domain as the former Euro-
pean southern continental margin was par-
tially extremly thinned during the Mesozoic
extension phase. During Alpine orogenesis
it was deformed to a more than 15 km thick
nappe complex, characterized by basement
nappes;
(ii) the OPHIOLITE suite marks, as a dismem-
bered complex, the old oceanic domain
between the two continental parts of the
plates;
(iii) the AUSTROALPINE domain is a part of the
overthrusted African plate.
- The SOUTHERN ALPS - also of African origin - are
separated from the Central Alps by the INSUBRIC
LINE. In this region a Variscan basement is overlain
by Mesozoic/Tertiary sediments.
The main goal of the seismic investigation was to
extend the structural information from the surface out-
crops to greater depth, in order to obtain a two- to
three-dimensional image of the deep structure of the
Alps and -in combination with the results of other geo-
logical and geophysical research - a better understan-
ding of the tectonic evolution of the Alpine chain during
its oceanic and orogenic phases. Special attention was
centered on the structural relations between the two
former plates.
1 Swiss National Geological Survey, CH-3003 Bern,
Switzerland
2 Institut of Geophysics, ETH-H0nggerberg, CH-8093
Z0rich, Switzerland
3 Geological Institute, ETH-Zentrum, CH-8092 Z0rich,
Switzerland
Geodynamics 22
Data acquisition and seismic processing
Field work was contracted to Prakla-Seismos AG,
Hannover, Compagnie GSnSrale de G8ophysique
(CGG), Paris and in a smaller amount to Swiss enterpri-
ses. In addition the Institute of Geophysics, ETH Z0rich,
supported the recording campaigns. To get information
from both the near surface structures and the deeper
crustal features, two acoustic sources, Vibroseis and
dynamite shots, were applied. Generally the following
field parameters were used, but often special configura-
tions adapted to the local goals were necessary:
161
162 CRUSTAL INDENTATION IN THE ALPS
o .
w
ILl
c
c
c
E
HEITZMANN ET AL. 163
SANTISALP
NFP-20 RFLECTION
SEISMIC PROFILES SIMMIBACH
ALTENDOR '
LOCATIONS OF _ I ", '",
DYNAMITE SHOTS SEVELEN
AZMOOS 4 .! .''
TH ZORICH DFS-V SARGANS 0/':'' ....... '' ................
SEISMIC 'PROFlEES .....
HArtz: JENINS
VALEN' i PeAN j
UNTERVAZ ...........
KUNKELS 1 KLOSTERSO_
I oc..
BREIL/BRIGELS
THUSIS CANOVA
OVALS l ZILLIS
[AUSSERFERRERA
HINTERRHEIN _ INNERFERRERA
TORRE ' j' ...... '"',,,,./' . CR(T
.."
VALBEL,LA SOVRANA
BIASCA [ STA' DO'ENIA '"".. .' ......
/ o \
LOORI"O / CHIAVENN ............ "";
SEMENTINAU BELLINZ)NA DELEBIO (EG'r)
iSON 7 'i" .....
LUGAGGIA
) .{" SAN PIETRO
LUGANO o :
." eLE BOLETTE
.......... i . -. ONI
'" I;"" SALORINO 50 km
SESEGLIO
Fig. 2. Eastern and southern Swiss Alps, with the location of the seismic profiles of the NFP-20
eastern and southern traverses.
Table 1 NFP-20 General recording parameters
Sources
- Dynamite 50 - 1200 kg
- Vibroseis 70 t
VP distance 40 m
sweep 60 sec
frequency 10 - 48 Hz
Data processing is executed in the computing cen-
ters at the ETH, ZOrich, with SSL's Phoenix software on
VAX-780 (Valasek at al. 1990) and at the University of
Lausanne with Geovecteur from CGG on a CRAY II (du
Bois et al. 1990).
Main results
Eastern traverse
The eastern traverse (Fig. 2) begins in the Helvetic
nappes, crosses the tectonic window of the eastward
plunging Aar Massiv and continues through the Penninic
Mesozoic sediments ("BOnderschiefer") before ending
up in the upper Penninic nappe complex of eastern
Switzerland, which is built up of crystalline nappes sepa-
rated from each other by thin zones of Mesozoic meta-
sediments (Fig. 1). The eastern traverse reaches neither
the Austroalpine nor the South Alpine domains and
therefore does not enter into the ancient African domain
(or Adriatic plate).
z
o
E
, , 'L,: " "x" :
, ,,' .,, ..,
,: i; i 'i ' ' :, ,
, ,:: ,,,
i.
,, , . /
:. '
. ,. . , I! ,.
,," i i,:; '
1
.,' , I . .
./ \ '" I
:1 , ,:, ,,,..:
. , /
. :
/ !1' '.1' ' ,,'// ' '" j ';.
,. ,i / '"' 25
": ,"1 \;
-
" '; \
i . . \ I
', .
'.. , / 'g..
, r e.' ! .//
L *
SONO03S NI 31_1_NO1_1_039-.13
N S
=======================================================
, ::iiiii::-: -. -o 60
o 50 KM 100
Fig. 4. Synthetic geological cross section along the NFP-20 eastern traverse
COL DE JAUN
LAUBEGG
ANKENBURG
TTEN
, IFFIGENALP
k RAWIL
. / .CORDONA
G BOUZERU e ISSOIE
SION )
MISSN
ZINAL
NFP-20 REFLECTION SEISMIC PROFILES
LOCATIONS OF DYNAMITE SHOTS
ETH ZORICH DFS-V SEISMIC PROFILES
BALTSCHIEDER
AROLLA I ZERMATTi MATTMARK
..'
ZMUTT '..J FINDELEN .............': FURl
,ooO ""'o.,.
........... OERViNiA '". ...........
oO., ........
VISP 0
VISPERTERMINEN
LPETRAN
/ ...........
20 km
SAN GIACOMO
CROP
Fig. 5. Western Swiss Alps, location of the seismic profiles of the western traverse.
N
s i mmen
013 1S6 2q6 2S6 306
I I I I i i
10-
20-
30-
40-
,50-
60-
"- . .3 S W E
Granges Uol d'lqnnlvlers Sialden Zermott Furl (14-E)
406 506 556 606 156 206 256 306 356 106 156 206 256 306 356 406 4.56 506 106 156 206
I I i I i i i i i i I i i I i i i i I I i
::. '7..c ,4.'2; ..... '
b.:-c:...5:.:%_:.:?;;..-":.".: :37'b '.-...-.4.--,,,.:'.-'.:::.: :;:';.-:..,. T..'-.::.::::...-7.E;'7'-:.:7 ':'-E'-:."-'.:? '::'. :.::-,.'.-:...,:-"...,.- z. 12
J" ;. ':'" "':'. ':'":'4.--:?'7:'Z ""c" '": '-' '. '"? IJJ
;. :.;:_?.,....---<. ..:...::.;:...:,... :. -;::.7%:.,...:;;...,..'_,_ -. -5. oo
--:z:...:..::'." :.-'--:%-.,::-. ':" '"' ":'"' .... :"-" ..... :=':;:"'::-%'":' :"'--'::...'---.":;:7<'.:'i- 6, Z
..' '."-' :- ".z_' - . . :-; .4','; -L'.,:.:.'"-.,-;, -';,,.. -'..=,'...'..;".-.':...
"Z '-":L?_" :'.:.':,-2."' :,.'.,.::.'%.,.--..,.-;.F ~. -.':E '.-,-:': ::. = .',.- - 7. --
'77"':-"':'"-";7:'?"::'+'""'""'";"&?;3::7i'" ;?:;.:.....:.`...%::?`::,..;.::`::.-:..-..::.:. ;:''::'"' :':' ..... '*:"':" e. LLI
- :'.:; ...:% :- '.,!:.Z_ .(,?'--;-:%,.7%. '--..-.. '-% :-: ?- . :.-:.-"-'j "-'_:: ::::.:'.--,.. . ':-Z .: :',.. ".-: .... ::- -C :.'":,':: '.,. ;.;.:4
.':.:..::'-: --.::'L_-;.';",: .... c:;:.-..:-:.':." :.":,,',:;';!.;-.::..-'-',;'.'.,L.;-:-;.'C'. '; 9.
;::.%.?:,.:';.:;:''-.::_,-C%.::.:.4T-'"--' %::''":::";.71.: -'::-..:':'-,'-'.L:'. ...... .:.?:..;?;-.:.;:..-:....:;,:,...:.:,,.::.;.,:.:..:z. 10. Z
-:-7' '.: ....... "':%:..-"'.':::-%"--:.':.'-.-:'%;":::': ;"-;:..-.. -'1-.-".:;:;-::.'3 :'..: '::...;E!?.r.:.:,'. 18. I'-
':::'"' -. -:' '- ;'".:":'"" 2;.-- ''-.: -?:;".:-.;'..!:',:??,:%.4::' , 3. 0
;'':-".;'_"-' :>-.'3..<:. '=.-- :"%',.,.:.:' '...?'..-'.::.._"--',..-:.;;.:-c:..,:;;:"'.-. .... .*!.::Z~:"--L,.;.,. ;,f': :-' .;' ?.-';.':' .2 ILl 14. _1
i?': E X P L O S I O N SE I SlY11 C "-..]!?' .'i71.:'.'.':",i;; :';:' ;:' ;:!?;...::;"!; - ":': ..... ':' '"" ":' :::'"" ::;x,, ::CC-?;-..':-_'",_.' :'.., .;.":.':'..../-.':';::--7.?..'.'?.:: 1,7.
Fig. 6. Explosion seismic profiles of the NFP-20 western traverse sections Wl, W2, W3 and W4.
In a general profile the following features are dis-
played from N to S (Fig. 3 a + b, 4):
- In the northernmost part of the profile various dis-
crete subhorizontal reflections down to 3 sec TWT
mark the Helvetic nappes, with internal thrusts, and
the sole thrust, which steepens up towards the
South where all the features crop out.
- The crust/mantle boundary, the Moho, forms a very
pronounced reflection band in the North at depth of
12 sec TWT and dips gently towards the South to
reach a depth of 16 sec TWT in the central part of
the profile.
The crustal structure of the northern part compares
well with similar profiles from ECORS (Cazes et al.
1990) or DEKORP (Behr & Heinrichs 1987, DEKORP
Research Group 1990) in Central Europe. The up-
per crust with a thickness of 5 sec TWT (,-15 km) is
rather transparent with only a very few reflections,
whereas the lower crust (4 sec TWT,,-12 km) is
almost over its whole extension highly reflective with
a pronounced "Moho-band". Towards the South the
thickness of the upper crust is nearly doubled below
the Aar massif, either by tectonic overthrusting or by
upfolding. This structural configuration implies a
midcrustal detachment to allow independent defor-
mation of the two crustal complexes.
- In the southern part of the profile subhorizontal
reflections to depths of approx. 5 see mark the
zones of Mesozoic metasediments which separate
the crystalline cores of the Penninic nappes. The
tectonic cross section derived from reflection seis-
mic recording corresponds exactly to that con-
structed with the method of projection parallel to
axial plunge applied for a long time in the Swiss Alps
(see e.g. Argand 1911).
- At depth of about 14 sec TWT reflections can be
interpreted as the (South Alpine) Moho. The gap in
Moho in the central part of the profile was first inter-
preted as a disappearing of this feature under the
central part of the Alps, but refraction seismic
(Burness & Giese 1990, Ye & Ansorge 1990), ray-
tracing and the results of the southern traverse (see
4.3) show clearly that reflections from the southward
dipping Moho were not registered because the
recording disposition did not reach far enough to
the South.
An additional W-E trending line from the central part
of the eastern traverse (Thusis) over the Austroalpine
nappes to the Penninic window of the lower Engadine
displays the eastward plunge of the whole nappe pile
and the structural relationship between the two Penninic
domains.
ZERMATT SAN GIACOMO
I I
501 551 601 651 701 7S1 801 8S1 901 951 100110S1 1101 1151120112511301 13S1
i ! ! ! i ! ! ! ! ! ! ! ! ! ! ! ! ! 0
.,':.' : .,.? .. . ...... ' ;'.. .
._.: ;..,.:,-.: . . ..-
...: .; ..... . ........ ?;..-:-.' :-.
:'.*:;::.: .. *. . .-.:-.
_.;::-:? :';'..C ...... ;.;: '.':-,';'.--:5:: .... .' 2,
.. : .,-- ::- ..... .'; ..... .':-:.;?.:-:. .;.,;',.. .: - . -.- ' :.- , . ;..
: :-... ;..-., , . ,:- ??': ......::.. _::c - ',-: .:.:' :- .... ';_..;.- ::- ..,":?:; ..... '_.
. ..-,_ c.: .: -,.. -.. :.. ,..; .... : ..... ...-., :. ?: .- .. - :" . -
:;- '. ::.' --:" -': ';-a 7:. '-::'- .'. .: :.'- ;''." -...'- :'-' : "-'-''? '-':"" .
-..:... - .. -,-' . .- . . -.'J .... ...... : -.. - - :. .'.?:_.':.:. :::: :.;.....:..' :..-.::.
' ',- ... -:'-'-<: '* '-' ..,...-_ ::':.. -:c :;: :- ..- -/.. ;: :.. ;?.: '. -' ..'. :..:-- ,' .'5.
- - : {7'. "':- . .;:. ;:": ":';:::;:: ..'%'::?';'??:-::. ': ;;';'.-.-.: .... :'::': --' :'.:;..'....:'
.: .;_. .-.. :?.: -; -. -..._.: 2,-.: .... .:: .... ::.: ...:..;..'.? '...: :- :;.... :. '.. .... :.:
.. ,,.:;/_... , ':. .?<- - .... ?.%:-:-E'.. . ;:;( -:;'... :?.' a;:;- :? :=?.'-' .';:. ::.:';';..-' -..':,:
<.<:? .' .-, ....... .:... -...->.: ;-:. :.--. ....:.. ...... :....-::: :';?:.?. '.....;-' -.'..:a;:. '::: :':., ._ -.? .-...:.
. (- 'c:...:,:.- :..-. ;.- ';:. , . : .;;,;::-' :.. .. .... .-..-.:c-.._.:; .... ...:.. --' :: :-'-':,. ..'.::., .
'-. -....:.:--- '?.:2: . ::. ..' :'.. : ;--V.::.:*: ._ -C__ . ..-. _." ...... '._. .,:-=.,. ,. ...::: .:'. :..?-..
: '?"' .:" .' '" - '. '/-,:"-;<-':.": '. .... :.-'...'':' -":.- :" :- ' :'-..' .... : .... ':::':;;::;: .-':7,:'":-'':;"';
'- .... ...:.?:_-.:': :;' '. :', .5-":::.;.-.. '. ::: -::,' ;'".. ..'.' :'.'.%.':. ' -' '"::.-.-':.:'.'<';::.:."'.'..L'
..._;.: '--. .. ,, ,-./.' .7 .-....',_..; ..-7.:;: :: ..... -' -; "-'.': ..... , -" '. ' . -- '.::2:' .... :: ;' T::": ' '. .... : ....
..... - . :: ..-: .' :..::<.:. ..:?:-.. .-,..--:.. . ::.,,' .. . .- .. .:;;... -* a.:c ...: ...:-,.
.. ;. . ,.. _5...- ..... .?: ,:. , .,. :' . ,,. '.-,.' .': ..'.6:-.-'_:: '.. "... ';'
..:..- ..- ..:.:. . ..;,:. = --.,: ;:'..:.:: ...... ;..,.')..:;.'.:.:. ;,-...- -,: :.:.. .. :. :-5 : , :. .:.:c .'. j
: .". ................ = - ,--:-- - .... '-.*. ---7 . . .... ". -:?. :?- '.' ?< .... ;:::." e'
, ... ?,'..:.,..-::. .. '. : ,... .. :;:. _.: ..-.-,.:.-Z... _ .... ::- , :,..:. ..... . . ... :.:..,.
Fig. 7. Explosion seismic profile of the NFP-20 western traverse section W3 and prolongation to Val
d'Aosta.
Western Traverse
The western traverse (Fig. 5) lies west of the Central
Alpine updoming of the Aar massif and the lower Penni-
nic nappe complex in a similar tectonic position as the
eastern one. It is composed of four branches. The first
one begins in the external Alps, crosses the Rawil
depression between the Aar and Aiguilles Rouges mas-
sifs to arrive in the Rhone valley at the Penninic front.
The second profile is nearly the continuation of the first
one and runs along the Val d'Annviers through the
upper Penninic nappes. The third (N-S) and forth (E-W)
sections in the Mattertal are displaced from the others to
E, but structurally they can be correlated by axial pro-
jection.
The results from the western traverse confirm those
from the eastern one (Fig. 6 - 10). Here also the Moho is
plunging from a depth of less than 12 sec TWT (-,35 km)
z
N S W E
__.
.._-_. ---. .... ..-. :-. '-'<. ;..-'.':.
_:: .... .-.....r.'--. '"t
---_ ........_.7- ..... :- ....
r.- ..r : -' :- ,.. -.--
,_.
..
' ...3%..:. '""'- ./
:.,. . ._ ." .
": "':" :" ""- " '' ' "'*'" ""' "" "'L -
;. - ,.. ,.:... -:'j: '' ... , ..;.
!.". '~ '.',:-: . .:..--' ....
4o-.-".;";::.".::":'"i.?.:L:":' ' ' .... --,' ..... ' .... .... :-'-f,i_'.,.:;:.._.
,:;, ::...'.:. ,..,
.<.. _... m...x..: ,..., ,'...,..
-., '-' .,-":': "'/ ' ..... :'- '.-":;E,-'..:..-;-. ...... .., ..
':' '" '"'/140H -' '":'':":
::. .,. .._.., '-. .-_ ":.,... _..,,,..., ..... .._ _> --
-' . .... -.),..ji. :L:' . ..:
50
...... .: ;... ;.-.,
'- ...?
"' ;" ':'' ;:c.. ',:-
"" .- '"'"'C..:.
., :.:.,5" .',,
/- ..

... :.' .:,:.
---.. ...:;' ..-
z
o
o
z
..,:_'"-.-'-_,_.'.-
z
":":' "-" "-- 0
,., . .-..._. ,-',?
.... '"; '/"2 '"1 I-'
'"
,.:.';:. ..,,,_.- _
..: .:,, ... .d
..-..714 LL!
,-,,' ,o
..(/ . -

.. .,,
:y ..:"
' ....... 16
,.
'-.,.
..,.
,./
--18 .._
:.,, .,.....
Fig. 8. Vibroseis profiles of the NFP-20 western traverse sections W3 and W4.
to more than 17 sec TWT (~51 km) under the region of
Zermatt. In the upper part of the profile the base of the
Helvetic nappes in the northern part and the internal
structure of the Penninic nappe pile in the South are
displayed. A new feature, not detected on the eastern
traverse, is the reflective zone in the Val d'Anniviers sec-
tion W2 dipping to the north from 5 see TWT under Zinal
to depth of 8 see TWT under Sierre.
New lines were recorded in 1990 further west (Val
de Bagnes - Martigny - St-Maurice) to have a modest
three dimensional control in the Valais Alps and a pos-
sible link to the French-Italian ECORS-CROP profile
(Nicolas et al. 1990).
Southern traverse
The southern traverse (Fig. 2, 11) completes the
reflection seismic profiling across the Swiss Alps
towards the South. Seven sections give an entire cross
section through the Central and the Southern Alps. It
begins in the lower Penninic nappe complex, crosses
the Central Alpine "root zone" (the southern steep belt of
ZWEISIMMEN RAWIL SIERRE ZlNAL ZERMATT SAN GIACOMO
I I I I !
W2 W3 ---
Wl .,/ HELVETIC NAPPES .__
-50 /-/o
LINE DRAWING
Fig. 9. Line drawing of the NFP-20 western traverse.
o
-14 J-
-16 =j
-18 jJ'j
-2o
SIERRE ZERIVlATT
20--
40-.-,
60- I I
o 5o K M 100
Fig. 10. Synthetic geological cross section along the NFP-20 western traverse.
LODRINO
ORDUN
-SEMENTINA
HINTERRHEIN
SANTA DOMENIC
SAN VlTTORE
-- P LA.'I'TA
BREGAGLIA
LUGAGGIA
SAN PIET
LE BOLLETTE
MONTONI
O 10 5Okm
Fig. 11. Southern Swiss Alps, locations of the sections Sl, S2, S3, S4, S5, S8 and S7 of the NFP-20
southern traverse recorded in 1988.
the Central Alps) and the Insubric Line (Heitzmann
1987, Schmid et al. 1989), runs through the South
Alpine basement complex and ends at the southern-
most point of Switzerland in the Mesozoic/Tertiary
sedimentary cover.
The main results of these investigations are - from N
to S - the following:
- The profile S2 (Fig. 12) over the frontal part of the
Lucomagno-Leventina unit (the lowermost tectonic
unit which crops out in the Ticino region) displays
discrete reflective zones at depths of 1.5 sec TWT,
between 2 and 3 sec TWT and under 4 sec TWT
very comparable to the reflective zones separating
the crystalline nappes in the higher Penninic com-
plex on the eastern and western traverses. The
same features are displayed also in the uppermost
part of the Val Calanca profile (Fig. 13, 14). There-
fore the Lucomagno-Leventina unit does not repre-
sent the lowermost tectonic unit, but it could be itself
a nappe separated from lower nappes and the rela-
tive autochthon by Mesozoic metasediments. In
consequence the overall shortening during Alpine
overthrusting and nappe formation may be several
tens to possibly more than 100 km longer than cal-
culated until now.
SHOT
TORR
140 160 1BO 200 220 241 260 200
s
o
STATION
O. 000
1. 000
2. 000
3. 000
4. 000
S. 000
6. 000
Fig. 12. Vibroseis profile of the NFP-20 southern traverse section S2.
VALS
VAL C A L A rqc A LUGAN0 CHIASS0
I I I I o
- ..... " ' ' ':7-,-' ' :";';-'" ' ' ' ..... ":' '" ' '' ""' 'i
.i:=:(:.':;":: ',......;..'.-. _ .i" ';'::' -'" ........ ":-"
,. .-'.., .': '':': ::' :-''.-7 :-:-: '...-:.':'
. : ..-..... .'.-': .:... ..... ?...;.--...
?, .. :.. ....:.,. .. "' .: .::c,;..::L.. ::. .
, ":':' .' L - :' .' :'; ;:,'...-..:.- %:_.-';-".-i-::( -
'"'x.,..,. -.:.....' .'.; :.-L'-'--,;"...--..=-t -
-',..: '.i. . :'4::.;%,&:-.::::.;. ,., ::L-:.:.--"...
(:; :L?.:,--.4 -:- .-;..?.::;::?,.:7,.:;'..,..:L -,'..:-.,:,"z.'-'.'::..,::,:? .':z:'.-"','.,'.--...-:..
S 95 KM S3 .S4 S5
EXPLOSION SEISMIC
Fig. 13. Explosion seismic profiles of the NFP-20 southern traverse sections Sl (Calanca), S3
(Lugano), S4 and S5 (Chiasso).
20-
NORTH IEI, :
% VALS VAL CALANCA
30-
40-
50-
KM 110
SOUTH
'14OH O
LUGANO CHIASSO
-- MOHO
lO z
12 ...I
'"'?'' 14
-10
1 =
Fig. 14. Line drawings of the NFP-20 southern traverse sootions 8,1 (Oalanca), $ (Lugano), 8,4
and $5 (hiasso).
30
50
km
CENTRAL ALPS
Penninic
European Foreland "Massifs" nappe complex
LU
'" SOUTHERN ALPS
z
N / basement sediments s
"" N ..... .,.<,.:::...::.:...:............ --.-.-.:............-......-......,-,.;.,,.., ............
d_ x .... i:::.-:':..:i:.-'::,
-_ .:,,,..',,::-.,.:...':..'.:/''"'':":'"ii......:ii ..... .`.!!iiii..::....:ii :':"-"
..........
/OWer crust -"- ':*iiii'""'"""'"'" '"'":'"'"""' "'"" "
100 km I NFP 20
PNR 20
SYNTHETIC CROSS SECTION
Fig. 15. The NFP-20 synthetic crustal cross section through the Swiss Alps displays the indentation
of the European and the African plates.
o
30
50
km
- At depth of 16.5 to 18 sec TWT strong reflections
can be correlated with the reflections of the Moho
on the eastern traverse. Therefore a continuing
southward dipping of the European Moho under the
Southern Alps can be assumed; this is in good
agreement with the EGT refraction seismic survey
across the Alps (Buness & Giese 1990, Ye &
Ansorge 1990). To confirm this interpretation a new
Alpine traverse was recorded in 1990 extending
from Meiringen to Southern Switzerland.
- The continuation towards the surface of the steeply
N dipping reflection zone in the Val Calanca section
(Fig. 13, 14) coincides with the outcropping rnylonite
belt of the Insubric Line. This feature is therefore
interpreted as the continuation of the Insubric line to
depth. This interpretation of a north dipping Insubric
Line was also proposed by Heitzmann (1987) based
on structural and gravimetric arguments.
- South of the Insubric Line, respectively underlying
this northward dipping feature, one finds a zone with
low reflectivity to depth of about 8 sec TWT and a
zone with strong reflectivity at a depth between 8
and 11 sec TWT. The upper zone represents the
outcropping upper crust of the Southern Alps
whereas the lower zone would represent its lower
crust. The whole South Alpine crust is here thrusted
at a depth of about 11 sec TWT (~33 krn) over the
underlying European elements with the southward
dipping European Moho at the base.
- In the South Alpine crustal complex two very
distinctive northward dipping reflection horizons can
be correlated with upper crustal thrust faults (Fig.
13, 14; Schumacher 1990). This deformation of the
South Alpine upper crust explains very well the
doublin of its thickness in the region north of
Lugano compared with the reduced thickness in the
profile further south at Chiasso (Fig. 13, 14).
- The crustal section in the Mendrisiotto region
(Chiasso in Fig. 13, 14) displays a vague Moho
reflection at about 12 sec TWT -in agreement with
results from refraction seismic - and the base-
ment/sediment contact is found above 5 see TWT.
This seismically not very well structured South
Alpine crust with a thickness of not more than 20 km
can be taken as a model of South Alpine (African)
crust tectonically thinned during the Mesozoic
extension phase.
- The thickness of the Southalpine sedimentary com-
plex in the Mendrislotto region of more than 4 sec
TWT (Fig. 13, 14), which exeeds the normal strati-
graphical thickness, can be explained by thrust faul-
ting in the sedimentary cover as proposed by Laub-
scher (1985).
Fig. 16. Schematic block diagramm of the crustal relationship between the European and the African plates in
the Swiss Alps.
Conclusions
Reflection seismic profiling through the Swiss Alps
reveals an asymmetric deep structure (Fig. 15, 16) with
the European crust plunging to depths of more than 60
km under the overthrusted African plate which on the
other side indented the European thrust complex.
There is good evidence to assume that during the
Eo-Alpine and Meso-Alpine phases the African (Adriatic)
plate was thrusted over the European one. The Eo-
Alpine event can be related to HP-LT metamorphism in
the European/Adriatic suture zone. During the Meso-
Alpine deformation, imbrication of the southern margin
of the European crust led to the formation of the Central
Alpine nappe complex. The Eo- and Meso-Alpine
deformation was generally north-vergent and resulted in
a asymmetrical mountain belt.
During the Neo-Alpine phase, a wedge of the Afri-
can plate was indented into the European crust. From
structural analysis and petrological and geochronologi-
cal studies we conclude that this deep crustal Neo-
Alpine process of indentation is related to decreasing
metamorphism in the Central Alps and the formation of
the Insubric Line. In contrast to the earlier generally
north-vergent thrust system, the Neo-Alpine orogeny
can be interpreted as a rootless symmetrical thrust
system toward the northern as well as the southern
foreland.
Acknowledments. The NFP-20 programsdirection is
grateful to the Swiss National Science Foundation
(grants 4020-10822, 4020-10902, 4020-10903., 4020-
10955) and the NFP-20 group of experts for financial
and scientific support. The reviews are appreciated.
Institute of Geophysics, ETH Z[3rich, Contribution No.
670.
References
Argand, E., Ees nappes de recouvrement des Alpes
Pennines et leurs prolongements structuraux, Beitr.
Geol. Karte Schweiz N.F. 31, 1-26, 1911.
Behr, H.J. & T. Heinrichs, Geological interpretation of
DEKORP 2-S: A deep seismic reflection profile
across the Saxothuringian and possible implications
for the Late Variscan structural evolution of Central
Europe, Tectonophysics 142, 173-202, 1987.
Bernoulli, D., P. Heitzmann & A. Zingg, Central and
southern Alps in southern Switzerland. Tectonic
evolution and first results of reflection seismis, in
Deep structure of the Alps, edited by F. Roure, P.
Heitzmann & R. Polino, Mm. Soc. gol. France
156, M6m. Soc. G6ol. Suisse 1, Vol. Spec. Soc.
Geol. Ital. 1,289-302, 1990.
Bertotti, G., The deep structure of the Monte Generoso
basin. An extensional basin in the South Alpine
Mesozoic passive continental margin, in Deep
structure of the Alps, edited by F. Roure, P. Heitz-
mann & R. Polino, M6m. Soc. g6ol. France 156,
Mm. Soc. Gol. Suisse 1, Vol. Spec. Soc. Geol.
Ital. 1., 303-308, 1990.
Buness, H. & P. Giese, A crustal section through the
northwestern Adriatic plate, in The European Geo-
traverse Project' Integrative Studies, Results form
the 5th earthscience study center in Rauischholz-
hausen, 26.3.- 7.4.1990, edited by R. Freeman, P.
Giese & St. Mueller, ESF, Strasbourg, 1990.
Cazes, M., A. Mascle, J.-F. Raoult, Interpretation de la
vibrosismique rflexion, in Etude de la croOte ter-
restre par sismique profonde. Profil Nord de la
France, edited by M. Cazes & G. Torreilles, Editions
Technip, Paris, 53-77, 1988.
DEKORP Research Group (Franke, W., R.K. Bortfeld, M.
Brix, G. Drozdzewski, H.J. DOrbaum, P. Giese, W.
Janoth, H. J0dicke, Chr. Reichert, A. Scherp, J.
Schmoll, R. Thomas, M. ThOnker, K. Weber, M.G.
Wiesner & H.K. Wong), Crustal structure of the
Rhenish Massif: results of deep seismic reflection
lines DEKORP-2 North and 2-North-Q, Geol. Rund-
schau 79, 523-566, 1990.
Du Bois, I., L. Levato, J. Besnard, R. Marchant, R. Oli-
vier, M. Ouwehand, S. Sellami & J.-J. Wagner,
Aspects particuliers du traitement sismique des pro-
fils alpins suisses du PNR-20, in Deep structure of
the Alps, edited by F. Roure, P. Heitzmann & R.
Polino, M6m. Soc. g6ol. France 156, M6m. Soc.
Gol. Suisse 1, Vol. Spec. Soc. Geol. Ital. 1, 47-54,
1990.
Heitzmann, P., Evidence of late Oligocene/early Mio-
cene backthrusting in the Central Alpine "root zone",
Geodin. Acta 1,183-192, 1987.
Laubscher, H.P., Large-scale, thin-skinned thrusting in
the Southern Alps: Kinematic models, Geol. Soc.
Amer. Bull. 96, 710-718, 1985.
Nicolas, A., R. Polino, A. Hirn, R. Nicolich and Ecors-
Crop working group, Ecors-Crop traverse and deep
structure of the western Alps. A synthesis, in Deep
structure of the Alps, edited by F. Roure, P. Heitz-
mann & R. Polino, Mm. Soc. gol. France 156,
Mm. Soc. Gbol. Suisse 1, V01. Spec. S0c. Geol.
Ital. 1, 15-28, 1990.
Pfiffner, A.O., E.M. Klaper, A.-M. Mayerat & P. Heitz-
mann, Structure of the basement-cover contact in
the Swiss Alps, in Deep structure of the Alps, edited
by F. Roure, P. Heitzmann & R. Polino, Mb' _m.. Soc.
gol. France 156, Mm. Soc. Gol. Suisse 1, V01.
Spec. Soc. Geol. Ital. 1,247-262, 1990 a.
Pfiffner, A.O., W. Frei, P. Valasek, M. Stauble, L. Levato,
L. Du Bois, S.M. Schmid & S.B. Smithson, Crustal
shortening in the Alpine orogen: Results from depp
seismic reflection profiling in the Eastern Swiss Alps,
line NFP-20-East, Tectonics 6., 1327-1355, 1190 b.
Schmid, S.M., H.R. Aebli, F. Heller & A. Zingg, The role
of the Periadratic Line in the tectonic evolution of the
Alps, Geol. Soc. London, Spec. Publ. 45, 153-171,
1989.
Schmid, S.M., Ph. ROck & G. Schreurs, The significance
of the Schams nappe for the reconstruction of the
paleotectonic and orogenic evolution of the Penni-
nic zone along the NFP-20 East traverse (Grisons,
eastern Switzerland), in Deep structure of the Alps,
edited by F. Roure, P. Heitzmann & R. Polino, _M_.rn.
Soc. gol. France 156, Mm. Soc. Gol. Suisse 1,
Vol. Spec. Soc. Geol. Ital. 1,263-289, 1990.
Schumacher, M.E., Alpine basement thrusts in the
eastern Seengebirge, Southern Alps (Italy/Switzer-
land), Eclogae geol. Helv. 83, 645-663, 1990.
Valasek, P., W. Frei, M. Stauble & K. Holliger, Proces-
sing of the NFP-20 seismic reflection traverses
across the Swiss Alps by the ETH-Zurich data pro-
cessing center, in Deep structure of the Alps, edited
by F. Roure, P. Heitzmann & R. Polino, M6_m.. Soc.
gol. France 156, Mm. Soc. Gol. Suisse 1,
Spec. Soc. Geol. Ital. 1, 55-64, 1990.
Ye, S. & J. Ansorge, A crustal section through the Alps
derived from the EGT seismic refraction data, in The
European Geotraverse Proiect: Integrative Studies,
Results form the 5th earth science study center in
Rauischholzhausen, 26.3. - 7.4.1990, edited by R.
Freeman, P. Giese & St. Mueller, ESF, Strasbourg,
1990.
CRUSTAL REFLECTIONS
BENEATH THE ALPS AND THE ALPINE FORELAND:
GEODYNAMIC IMPLICATIONS
Jean-Louis. Mugnier* and Jean-Michel Marthelot**
Abstract The ECORS-CROP deep seismic profile provides a 320
km long traverse across the western part of the Alpine collision belt.
The most characteristic reflections are recorded: 1) in the lower crust,
which displays the strong reflectivity typical of the European
Hercynian crust beneath the northwestern foreland and the Belledonne
Massif; 2) at the boundary between the external and the internal zones,
where two southeast dipping reflective zones cross the crust from less
than 1 s to 5 s and flatten at this level beneath the metamorphic
Pennine zone. The lateral variations in the crustal reflectivity of the
lower crust cannot be correlated to the Alpine shortening. The foreland
lithosphere seems to be unaffected by the Alpine deformation and
supports a flexure that increases the depth of the Moho from 9 s
beneath the Massif Central to 14 s beneath the Belledonne Massif.
Thickening of the external part of the Alpine crust is induced by a thin-
skinned thrust belt, a foreland basin and an upper crustal imbrication.
The southeast dipping reflective zones are linked to major geological
features, i.e. the bottom of the sediments resting on the back of the
Belledonne Massif and a major Alpine suture including some basic
and ultrabasic rocks of the narrow Valais domain. A dense and high-
velocity lithospheric heterogeneity, inherited from the floor of this
domain, acts as an indenter beneath the metamorphic belt. It forces the
European upper crust upward mainly by thrusting with a west vergence,
while the lower crust is buried beneath this heterogeneity.
Introduction
ECORS-CROP shot a 320 km long seismic profile (Fig. 1)
(Damotte et al. 1990) to improve the interpretation of deep Alpine
structures. Results from other geophysical experiments, such as
gravimetry (Bayer and Lanza 1989), wide angle seismic reflection
(ECORS-CROP Deep Seismic Sounding Group 1989), and magnetism
(Rey et al. 1990), have been integrated in the interpretation of the
seismic data (Nicolas et al. 1990, a and b).
This paper describes the study of deep seismic reflection profiles
forming a complete traverse across the Alpine collision belt.
Particular attention has been given to the following aspects:
1) the lateral variations in the crustal reflectivity along the profile to
*ECORS Program and URA CNRS 69, rue Maurice Gignoux,
38031, Grenoble cedex, France
**ECORS Program and EOPG, URA CNRS 323, 5 rue Descartes,
Strasbourg, 67084, France
Geodynamies 22
monitor the effect of crustal thickening on the reflective lower crust;
2) the geological meaning of reflective zones that cross the entire
upper crust and terminate close to the surface.
Implementation And Processing
Of The Deep Seismic Profile
The implementation details are described by Damotte et al. 1990.
The sources were generally vibrators except in the Jura and part of the
Vanoise where 20 to 30 kg explosive shots were used. The French part
of the profile was shot and processed by CGG (Compagnie Gtntrale de
G6ophysique), the Italian part by OGS (Osservatorio Geofisico
Sptrimentale). No aleconvolution was applied except for the Jura-
Bresse profile.
In addition to industrial processing, several non-standard
techniques developed at EOPG (Ecole et Observatoire de Physique du
Globe) were applied in order to enhance the presentation of the stacked
sections. The lateral coherence of the events (Hatton et al. 1986) is
computed in a moving window of 15 traces for a range of dips less
than 0.25 s/km. For each sample, the maximum value of coherence
(varying between 0 and 1) and the associated value of the local dip are
retained. A plot of the coherence allows a condensed representation of
the seismic section where the lateral variations in the continuity of the
reflections are preserved. Alternatively, a plot of the samples having a
coherence greater than a given threshold enables an objective selection
of the main events (Fig. 2).
As shown by Milkereit (1987), the coherence values of the
reflections can also be used to weight the contribution of each sample
in a Kirchoff migration of noisy seismic sections. A similar technique
has been developed at EOPG (Bano 1989). Fig. 3 shows the result of the
coherency-weighted migration of the section from the Bornes massif to
the internal zones of the Alps at a constant propagation velocity of 6
km/s. Note that because the dip discrimination of the coherence
measured in a narrow window is insufficient compared with the dip
sampling interval used in the migration, artefacts such as migration
smiles are not totally eliminated, especially for long travel times.
Lateral Variations In Deep Crustal Reflectivity
Of The External Zones
The bottom of the reflective crust is very regular (Fig. 2). The few
abrupt changes in two way travel time (TWT) to the Moho are linked
to shallow velocity changes. For example, below the Saleve structure
(km 127 on Fig. 2), all the reflections in the upper, middle and lower
crust, have a 10 dip eastwards beneath the East flank of the cover fold.
This fold is located at the hangingwall of a ramp and increases the
thickness of high velocity layers of limestones and evaporites compared
to the adjacent Molassic sandstones. The lower crust displays the strong
reflectivity typical of the laminated European Hercynian crust
177
178 CRUSTAL REFLECTIONS AND GEODYNAMICS OF THE ALPS
I IOOkm :: Molasse and
SEISMIC I "
TRANSECT Bresse basin
LOC^.,o. I
:_..-.
... .:-.T':, '" .....
............ . ..... :_.:...:...:,.;::.....
Fig. 1. Main tectonic units of the Western Alps and location of the
ECORS-CROP profiles. (Bresse, Jura, Alp2 andAlpl ECORS-CROP).
The locations of the Swiss NFP 20 West profiles are also shown.
(DEKORP Research Group, 1990). The lateral variations of its seismic
pattern are partly linked to the variations in source efficiency
(Damotte et al. 1990) and to signal absorption in the upper crust. For
example, the lower crust appears very bright below the crystalline
outcrops (Massif Central and Belledonne Massif), and very faint below
the recent fluvial formations of the western part of the Bresse graben.
The reflective lower crust has a maximum thickness (4 s TWT)
below the Jura and the Belledonne massif and is slightly thinner (3 s
TWT) below the western part of the Molasse basin. The top of the
Hercynian crust , overlaid by Triassic evaporites and Jurassic
limestones is very bright beneath the Molasse basin (a on Fig. 7). The
entire Hercynian crust, defined from the Triassic unconformity to the
bottom of the layered crust, has a fairly regular thickness of 9-10 s
TWT, except below the Bresse basin (8 s TWT), and below the
Belledonne Massif (13 s TWT). Below the Bresse, the thinning of the
Hercynian crust is compensated by the basin infill, and no local uplift of
the bottom of the lower crust is evidenced. This upper crustal thinning
of the Bresse is linked to an upper crustal stretching which is transferred
westwards along a detachment in the lower crust and/or in the
lithosphere beneath the Massif Central (Bergerat et al. 1990). Below
the Belledonne Massif, the relatively transparent upper crust is 9 s TWT
thick, while the thickness of the laminated lower crust ranges is about 4
s TWT, a classical value for a European Hercynian crust . This
thickening of the upper crust is induced by stacking of upper crustal
thrust sheets (M6nard and Thouvenot 1984), displaced westwards along
a decoupling level in the middle crust. The barely visible reflections
seen in the coherency plot at 4 s TWT below the Belledonne Massif
could be linked to the Alpine sole thrust (b on Fig. 7). Summarizing, the
bottom of the lower crust deepens continuously from 9 s TWT
below the Massif Central to 14 s TWT below the Belledonne Massif.
This deepening is probably induced by the flexure (Beaumont 1981) of
an elastic lithosphere (Karner and Watts 1983) loaded by a tectonic
wedge. A thin skinned thrust belt (Jura mountains) (Guellec et al. 1990),
a foreland basin displaced along the Triassic evaporite decollement
(Molasse basin), and an upper crustal imbrication are incorporated
(Mugnier and Vialon 1986) (Fig. 7) in this wedge.
The Main Reflective Zones At The Boundary
Between The External And Internal Zones
Structures close to the surface
The axial zone of the Alpine arc is classically divided into an
external and an internal part, following the paleogeographic origins of
the rocks. The Pennine thrust is defined as the footwall thrust of the
western internal nappe, but the structural boundary between the two
domains is in fact a complex zone (Valaisan zone, see Fig;4), several
kilometres wide, in which numerous sheets of various paleogeographic
origins and of various metamorphic grades are incorporated (Tardy et
al. 1990) in a zone resembling an accretionary wedge (Polino et al.
1990). Basement slices, mafic pillow lava and even ultramafic rocks
("basic rocks of the Versoyen") (Loubat 1975) are also incorporated in
this complex thrust zone along the ECORS profile. Some outcrops of
serpentinites are attributed to ultramafic mantle rocks (Schurch 1987).
Locally the "Versoyen" displays blueschist and eclogite facies
metamorphism (Schurch 1987). These rocks may have been part of a
narrow oceanic domain ("the Valaisan domain") north of and parallel to
the main oceanic Piedmont basin (Lemoine and Trtimpy 1987). The
profile crossed this zone along a sinuous trajectory and the common
mid-points (CMP) have a swath pattern (Fig. 4). In a crucial zone for
interpretation, the profile ran parallel to the Pennine front and
crossed a deep lateral ramp (Butler 1985) between two external
basement thrust sheets (Mont-Blanc and Belledonne). Therefore, the
deep seismic data, which show two SE dipping reflective zones in this
NW BRESSE JURA MOLASSE BASIN BORNES BELLEDONNE VANOISE G. PARAOIS SESIA PO PLAIN SE
3000 2000 1000 960 460 2000 1500 1000 500 70 340 840 1340 1840
O .................. - : , . .....,,,. ..... - - -, ........ , .... - 0
I- - -' -,.....,- .-., - ...... .... -;, g..x------ '-?"' ;"+_-': --' ' "'-:' ? .:.' ....... >-.:'-"' .... '':" -: ............
1 ':" "' -' '-'" - :'" ':" -' ::'' '::>':""":: ":" ' ' '"'??':'::'::'::: .... ':''':-' ...... ' >" ""
:. . :? ::. :::.:-'.. .
- ::::::::::::::::::::: :. .... ::?: ..... ......
......... .,. ............. , .............. , .:. ,:,. ........... . . ':":<-':". ' '.':' ':':-"::':?.:'::"' -'":(.::-' .::::'-' :.'::.":' ::::::::::::::::::::::::::::
2 / ..... ,. :'-':-.:--':... ,:..:,:: .::.: .':: .... :'--. ..... . .... : :..-'..:.,:. ' . ,: .'-'? :.:: ,". '.::.'?'::' ":. '::: :: ::::..::.:. '. :::.:,":,:.:::;: n2
4 . "-- ". ..... : ......... :.: '": :" ..... . ::-,::: "::'- '- ' :."::...::.'.:,.':::-.: ':: ..:::..:->.:=.': n
, , :..:--' .. , :'-:.--. . ' . .... . -:.. ...:... :....; ..::..--:.., :..,.:': ', .,.: ::-':
8 - ' ' ': -:"" _ ' ' ' ' - . ' .:. ::. -:::.:. :':,.::-' '' ... "?:'?:::a
2o . ' ': .. .', ' ' . .' . : ":: '" ":' ' '':' "":' '
- . .: :..... . :: :-. .. -.- . :.> ..:, ....:,.,:- zo
22 . '. ,:. "'' -' ..:':' . "' " ' - ". ' ' ,' . - '..? - .'.: .... :'::. : - .-. ' ":::::
s ' -: , -'.:-'.,, ' ,, - ,:':? ,' ' , " , ',, ' : '' ' , ,,, ,.. .
0 25 50 75 100 125 150 175 20 25 20 275 30 ' S KM
Fig. 2. Events having a lateral coherence greater than 0.5 for any dip
value less than 0.25 s/km on the stacked section. The numbers located
above the plot refer to the CMP of the different profiles composing the
traverse.
MUGNIER AND MARTHELOT 179
D CB A
NW BORNES BELLEEONE VANOISE
ooo oooV v o
i--_: .... .... ,.. - _:... -..:.-... _-- .
-.. --- . -- -<: . _< .... .. ...... -.: . ..... ..__- .... :-. .. &..:o --..-.----. ;;..:.a=r 1 -- .
............................. ...... ...... .... ..... ........ ...... ..... ' ' '" - '': :-'--;"':;-" < =- - --- ' "'-'--;"-'"'' ':':
_ _ _:-...
'":'-'- "'';--' - - ' ...... '";'-"--.,:*'-' *"" ....... ' '--',':'*' ', * ---:-'"".,-'- :--- -".'-"?::"'"- ..... - ' *":,:.',,4._--.. I 4
G. PARADIS SE
12
60 70 80 90 100 KM
Fig. 3. Coherency-weighted migration of the profile at the boundary
between the external and internal zones of the Alps. A constant
velocity of 6 km/s is used in the migration. A, B, C and D refer to the
nodes of the CMP swath on Fig. 4.
The correspondence of one of the brightest reflective zones on the
profile with one of the serpentinites outcrops in the Alps suggests a
major tectonic contact, linked to the closure of the Valais domain, as
already proposed by Lemoine 1984. In the following text, this major
contact is called the Pennine suture. Late reactivation in an out-of-
sequence thrusting (Butler 1984, Roure et al. 1989), postdates the
closure of the Valais domain.
The root of the Pennine suture
The upper SE dipping reflection zone is clear to 5 s TWT (Fig. 2).
At deeper levels, a coherency-weighted migration (Fig. 3) simplifies the
complex reflection pattern of the stacked section. These results
suggest that (Fig. 6): a) the Pennine suture has a staircase trajectory
and flattens out at a depth of 5-6 sTWT beneath the internal nappes; b) a
duplex structure is located at the hanging-wall of the suture, and
induces a succession of synformal-antiformal structures.
Comparison between deep seismic profiling
and o[her eorhvsical data _ _ _
Fig. 4. Simplified geological map and CMP swath at the boundary
between the external and internal Alps. A,B,C and D refer to the
nodes of the CMP swath.
area (Fig. 2 and 3), cannot be directly correlated to the surface
geology along the ECORS-CROP profile.
Instead, the comparison of a NW-SE geological section (modified
from Tardy et al. t1986) perpendicular to the strike of the structures,
with a migrated seismic section (Fig. 5) where the average CMP line
displays the same trend, suggests that the lower (westward) reflective
zone corresponds to the top of the external basement massifs (Mont-
Blanc), and that the upper (eastward) one is the continuation at depth of
the tectonic boundary at the top of the basic complex (Versoyen).
The two strong reflective zones are linked in both cases to
lithological contrasts. However, shear fabrics and/or fluids along the
tectonic boundaries could also induce some reflections.
Other geophysical experiments have been performed together with
the deep seismic profiling. They have led to the development of a
synthetic model (Nicolas et al . 1990 a). The critical results were
acquired by wide angle reflection (W) experiments 0ECORS-CROP
Deep Seismic Sounding Group 1989), and are as follows:
1) there is a highly reflective horizon 22-30 km below the Pennine
zone, that corresponds to a velocity contrast of more than 1 km/s
(Thouvenot et al. 1990). As the WAR event is located outside the
plane, projection and velocity assumptions affect the comparison of the
two data sets. The shallow WAR event could be linked to the same
horizon as the highly coherent event at 6- 7 s TWT on the profile ((c)
on Fig. 6 and 7), or to the bottom of the reflective zone at 1 Os TWT on
the profile ((d) on Fig. 6 and 7). From projection studies (Thouvenot
et al. 1990), there is no actual constraint on the western limit of the
reflector, and the west dipping reflections ((e) on Fig.6 and 7) could be
located over the leading edge of the structure outlined by the WAR
experiment;
EXTERNAL ZONES INTERNAL ZONES
- VALAISAN -- VANOISE
METAMORPHIC
BASIC
SERIES
CRYSTALLINE (VERSOYEN)
MASSIF
', - "?Z' &_---,'_,-.., x .,-,,
'x ,... ' -x x .,,.."-.' ---.--. ' X - ,, 0--'/ . ,,,, o. , vv,
..... . .... - X\\x' \ o //_' .... . .... ,..,.,.
Fig. 5. Comparison between a NW-SE cross-section through the
Pennine front (broken line on Fig. 4 is cross-section location) and a
C.G.G. migration (wave equation). A,B and C refer to the nodes of the
CMP swath.
ANOISE G. PARADIS D CB A v
.-..-:..,..-._-.. -.-,, .....,,_..... .5_:-_ -.___.?.::--.: .... . ........
-...' "at ".' r?'!.' .'" '"' '::.- .'.""' ...._'-,,r,__.%:...-..-,.:. ' . . ,., . '.
h:'- ' .... .' "', "S .- 7'_- " '"'--" '.?":" :-,17.:.:. -:.-','-... ~ '--:.--,:. ,,::'
I' ..... "."; e- -. , .'. , ', ., ':=-.:'.. ..' .-:- .'. :-,- "..':
i. . .. . . ' ' ' .: ' ..';. -,:.-- . -..c . ' .... .
I:.... .......... -./.....'.::. :.:...:-. ... :..':.:..... .. ... ..,
..:,..:.,,.:.-- :..:.. -. -.,. .... -:, . .... .--=...,:. .,.- . .-..-.:.:i, '"_-_. ,
'. '-' .- la ..... :.'':-.' .... -;.:-':i .,':' ..: -." ""
.; ..... '-- : -..?r.-.-':.-::..=:--i?: ..... ?':-:-' "7, .... ' '"". ":'." -. .....
I ' '-"' ...... 1_.i."'-..' :' -.-,t,"' ...... ,.--:. .... : ..':-_..':-.::.-'..: ' -" ":':.-:
/' .-:::'??:'.":."... :- .="-'-:'-:'E.?:!'-!:":-'-',-=: -'-:-../'C'::'--': i' .... '"'
/." :..-' "'!':'::-::''i::.:L.::: .:i "'".' '-'i':'-'.L .-..:".:".:'."..'" 77..';? ": . .!:.. :E:.-' .......
:' :?: :. :i '.
,i '".- -. :7.""' ..' .... ":-.'-'.>-'::-'":..' .-.-".-.' ...,.
Fig. 6. Interpretation of a windows extracted from the coherency
weighted migration (see Fig. 3). (c), (d), and (e) are the locations of
the reflections cited in the text.
2) the Moho reflection is recorded from the fordand to the Vanoise
where it reaches a depth of around 55 km (18 s TWT on the profile)
((0 on Fig. 7). Note that, as no reflection was recorded deeper than
10 s on the profile, the WAR experiment appears more efficient for
studying the deeper part of the collision belt.
Gravimetric modeling confirms that the effects of a shallow, thin
and high-density slice beneath the WAR interface at a depth of 22-30
km is compatible with the observed gravimetric anomaly in the Alps
(Bayer and Lanza 1989). This thin dense slice is located to the west,
and at a deeper level than the already known "Ivrea" body which induces
the main positive Bouguer anomaly in the Alps (Berckhemer 1968).
The top of the "Ivrea" body is seen on the ECORS-CROP profile ((g)
on Fig. 7).
Discussion
Many Alpine collision models that fit the geometrical constraints
provided by the deep seismic profiling in the Alps can be imagined
(Nicolas et al. 1990a). However, a simple sketch can be suggested for
the continent-continent collision that takes into account three of the
major features of the Alps, i.e.:
- the tectonic culmination of the External Crystalline Massif at the
hangingwall of the Alpine sole thrust (Mtnard and Touvenot 1984 ,
Butler 1985 );
- the Pennine suture (Lemoine et Trtimpy 1987, Vialon 1990) ;
-the dense and high velocity slice beneath the Internal Alps (ECORS-
CROP Deep Seismic Sounding Group 1989, Bayer and Lanza 1989).
In this sketch, it is suggested that the dense body is located over
the Alpine sole thrust. Its displacement, which exceeds 80 km, is
balanced in the upper crust both by 35-50 km in sequence thrusting
beneath the External Crystalline Massifs (Guellec et al. 1990,
Mugnier et al. 1990) and by out-of-sequence thrusting (Roure et al.
1990) along the Pennine suture. Back-thrusts, located somewhere over
the top of the dense body may link the external thrust system to the
internal out-of sequence system ( Route et al. 1990) . The crude
restoration (Fig. 8) suggests that the present dense slice may be
inherited from the floor structure of the Valaisan domain and could
therefore be an oceanic crust, or more probably a mantle body uplifted
during Tethyan crustal extension. From these geometrical and
,w eaEss Ju.^ UOL^SS,= .^s,. .o,.,=s ..LO.. V^.O,S,= . ,^.^D,S SS,^ ,0 ,L^,, SE
3000 2000 1000 960 46.0 20,00 1.570 10pO 5010 Zp 30 840 1340 184%
_ / -:'.,,,RI.FI:,.. . .... ' .. ,':- , ]...z..',....... *.,.-.;..;, ii/l__uT,a,l. . _. :'.':..,'..'-,.'.'-;MET,aMORPHICB..ELT_?;>: . .: .,-..-',..':. . :-.:, -
z 1 -:'.-;-.-:;.-..:.-::7.. ;- . . .- .. "- :: .. ' -.':.-''.': -,,':-:':.-'-?..:.'.T.':'. "; .... ....-7=N .... ::.-'..'.-.'?..',;/f.:",,"''?.'.-':::-" '--';'.::.-:'.:-'h ''.'-':-i:,-:..,vJ---.,.'t 4
4/' ' ' ' ' '" " -' .... -.g.-.-: o- .' - - "': "- ' ' ' ':-' .-] -" ,:c--E-C.::- ' '. .'-.- -'...-'',.r:''-,''_,,-;:'.,u,--T. ',:4 .,'.,.-..-'--:'i;!.'..".
I..: ' '. ': '.',. ':" . '-..?'.'.'.-' '..: ','+:.: .: 7. ':: . ' "'" .'.: ' -'/..'. :'.:.? ,:.;.': ,,,-.:.Y...
'1'. . ' ...: , . . ..,.,,:..,.. _-.: .c,.,-:..,'.,,:,,: .., .. :. .:. , ,.' . -, .: .-, .:. -.::.:::...' ::.., k_,j-'-_.-:::':.:.:i-(,g;F,c. ; ,,.:.:...:,..'.,:.y::.:,, . .:?..-!,_-.=:<.'.:CRUST I
-r:-- "'--. ':='- -' : ' ' - ' ' :'?--_'='-=.:----.',:,;::/:._--'_.-:'' : :.: . -:: ?::-- ;" -'-"-,'-, :'i'..',.T."--"' - ..,---x:.. .'.5.,:::..-'-', r-:-::'? "-.- ..:..:;:e::'.'- :--'".- . ',:,,'rLE. "'% :: "'-"':'"?i .... ; 8
,j',-_'..: ' .-'--.,..:?..'.-? . -,.,._:-,.,'..::-.:;'_-::, .... :._?...THE .LOWRCi,,./'. ,.:-- ,.'_ .' ' '.'.-:..-.,.'-:?'"DENSEI ...:'...". .. ;". ' c.:':: .: "-' '- ' ':":"...":::'".':':::'1
'"l: ' '; ' ' : :"'? ::."..?,'-.-'.:. '--"::-',. -''. ;-..-"':'-.'---' "-Jl:: ...... : .... . ,..: - - ..7c, -.. . -...-:- ,. - ...?. : , .-.._..,.: ;. -.:? :. , .,:. :,...':=: - ::.. :..,..:..;....: . ? ;
2t..'" ',- . .: ': '. -' ':2 :.:' ".:" ":,", %':: '7" . .' ' ': ' :':'-- ' '-= :" :J.:'L .. ' "' -. - :-. : : : . '?. :. - ,-'.'=: .... :'??. -?::.;..:" ::...:" ':':'-.-?'/'
.... . . ..-:.-::..,. . .- :.::::.. .. . .- .. .::_ . : ...... ....
18 ' '. ' ' : ." .:.: - ' ' : ' ' - ", . - .':= , -":' ,'-. '-' :.'." '-":' - '-. :':."
1- . ..... ...: '.' -. ' : ,- .. .. ,- .: -, ',,,,.: ...... .:.. '..-- ..,:
2o . .. - ..... -,- .- .'-': . '.
J. . .. . ,.-:: .,. .., .. ...:. . . .... -. , . : ...... ,, . ,;; :!:-'- --... ?-: :-. . . . . :; .. . . , :- -.. , - ' . .,-.Ip '?. ::- .. - . ., ' 7 . - ' '--""
., .-.'.:' ' .." ' -:' .- : .,:-'" '.-: ' .... ' . ---" . ........ : ', ':-::, : ': , s '2' '''o " '7' 'do' o 260 2 2o oo K
Fig. 7. Interpretation of the entire stacked section of the profile. Same
caption as Fig. 2. (a), (b), (c), (d), (e), (f) and (g) are the locations of
the reflections cited in the text.
A MAJOR UTHOSPHEEIC INDENTATION *
o,w^,D m,u, oT o.ou,c mu, .+r-if]
[:;:;:,:.:..........-.---;.:.:.--___k... --'."
-.
:.-.:... ':. :...':. I
STACKING OF THE ME"TAMOIi!I UNi
AIAISA# FIOOI..
European crust
Externc c11ne
rr13sstfs
Internal zone
Apulian crust
C MAIN $UBDUC1tON
OF 'IHE PIEDMONT
Fig. 8. Sketch of alpine evolution. A, B, C, refer to three stages of
the evolution, respectively at 20 Ma, 80 Ma and 130 Ma. The ages of
the transitions between the three stages are here not precised.
kinematic considerations it is inferred that the dense slice acts as an
indenter bounded by two tectonic contacts, and embedded in the
European crust (Mugnier et al. 1989, Frei et al; 1989).
In fact there is now strong evidence that deep indentation, or
"crocodile tectonics" as defined by Meissner (1989), is a classical
feature in the orogenic belts (DEKORP Research Group 1990). The
areas of detachment (the sole and roof thrusts of the indenter and their
foreward continuations) are areas of weakness, i.e. of low viscosity
(Meissner and Kusznir 1987). In the case of the Alps, the contrasting
strength between felsic and mafic/ultramafic rocks in the 10-30 km
depth range (Molnar 1988) could then have a major effect: during the
compression stage, the hard heterogeneity, i.e. the Valaisan floor,
would not rupture but would transmit stresses to surrounding weaker
rocks (the overriding internal nappes and the foreland crust) and a
detachment would propagate through the middle crust of the
European foreland.
However, the high-density body is not necessarily made up of
mantle rocks (Nicolas et al. 1990 a), as mixed acid-mafic rocks affected
by eclogite facies metamorphism have the same range of density and
velocity values as the mantle rocks (Ringwood 1975). Surface
occurrences of eclogite facies metamorphism in the Alps (Gillet et al.
1985) and records by the Swiss geotraverse (Rybach et al. 1980) of a
shallow, high-velocity zone (6.7 km/s) interpreted as granulite rocks
(Pfiffner et al. 1988) along the East NFP 20 profile, point to the
relevance of this question. However, the tectonic sketch proposed
here accounts for a dense body of mafic or ultramafic composition.
Conclusion
The view of the deep structures of the Alps and their foreland
provided by the ECORS-CROP reflection profiling emphasizes the
contrast between the behaviour of the foreland lithosphere and the
behaviour of the Alpine lithosphere.
The foreland lithosphere seems unaffected by the Alpine
deformation and supports an elastic flexure that smoothly increases the
depth of the Moho without affecting the reflective lower crust.
In the axial belt, the entire lithosphere is deformed. A major pre-
existing lithospheric heterogeneity, inherited from the structures of the
Tethyan margin, acts as an indenter. It forces the European upper
crust upward along major thrusts and the lower crust downward.
Aknowledements. ECORS is a joint programm of IFP, INSU-
CNRS, SNEA(P) and IFREMER. This work has been supported by
"action accompagnement ECORS" of INSU-CNRS. We thank R.
Meissner, P. Vialon and two anonymous reviewers for their critical
comments and suggestions, which led to great improvements in the
original manuscript. G. Sumners is gratefully aknowledged for its
contribution to the quality of the English.
References
Bano, M. (1989).- Extraction automatique des rflexions,
modlisation des diffractions et migration des donnes
de sismique profonde ECORS Ph. Thesis, EOPG,
Strasbourg, 155 pp.
Bayer R. and Lanza R. (1989).- Gravity modelling along
the ECORS-CROP profile through the Western Alps.-
Tectonophysics, 162, 203-218.
Beaumont C. (1981).- Foreland basins. Geophys. J.R.
astro. Soc., 65, 291-329.
Bergerat F., Mugnier J.L., Guellec S., Truffert C., Cazes
M., Damotte B., Roure F.- (1990). Extension tectonics
and subsidence of the Bress basin a view from Ecors
data. Mern. $oc. gdoL France, Paris, 145-156.
Berckhemer, H. (1968).- Topographie des "Ivrea-Korpers"
abgeleitet aus seismichen und gravimetrischen Daten
Schweizerische Mineralogische und Petrographische
Mitteilungen, 48, 235-246.
Butler R.W.H. (1984).- Balanced cross-section and their
implications for the deep structure of the northwest
Alps: discussion.-J. Struct. Geol., 6, 603-606.
Butler R.W.H. (1985).- Thrust tectonics, deep structure
and crustal subduction in the Alps and Himalayas.-J.
Geol. Soc. London, 85, 1437-873.
Damotte B. et Nicolich R., Cazes M., Guellec S. (1990).-
Mise en oeuvre, traitement et presentation du profil
plaine du Pt. Massif Central.- Mdrn. Soc. gdol. France,
H.S., 156, 65-71.
DEKORP Research Group (1990).- Reflectivity patterns
in Variscan Mountain belts.- Tectonophysics, 173, 361-
378.
ECORS-CROP Deep Seismic Sounding Group (1989).- A
new picture of the Moho under the western Alps.-
Nature, 337, 249-251.
Frei W., Heitzmann P., Lehner P., Muller St., Olivier R.,
Pfiffer A., Steck A., Valasek P. (1989).- Geotraverse
across the Swiss Alps.- Nature, 6234, 544-548.
Gillet P., Davy P., Ballewe M., Choukroune P. (1985).-
Thermomecanical evolution of a collision zone the
exemple of the Western Alps. Terra Cognita, 5, 399-404.
Guellec S., Mugnier J.L., Roure F. et Tardy M. (1990).-
Neogene evolution of the alpine foreland: a view from
ecors data and balanced cross-section. Soc. gdol. France
156, 165-184.
Hatton, L., Worthington M.H. and Makin J. (1986).-
Seismic data processing: theory and pratice. Blackwell,
Oxford, 860
Karner, G.D., Pad Watts, A.B., 1983, Gravity anomalies
and flexure of the lithosphere, J.Geoph. Res., 88, B12,
445-477.
Lemoine, M. (1984).- La marge occidentale de la Ttthys
ligure. In G. Boillot, Ed. Les marges continentales
actuelles et fossiles autour de la France.-Masson, Paris
342 pp.
Lemoine,M., Trumpy, R. (1987).- Pre-oceanic rifting in
the Alps.- Tectonophysics, 133, 305-320.
Lemoine M., Tricart P. Boillot G. (1987).- Ultramafic and
gabbroic ocean floor of the LigUrian Tethys: in search
of a genetic model.- Geology, 15, 622-625.
Loubat H. (1975).- La zone du versoyen. Ttmoin possible
d'une intersection entre dorsale volcanique octanique
et marge continentale.- Arch. Sci. (Genve), 28, 101-116.
Mnard, G. Thouvenot, F. (1984).- Ecaillage de la
lithosphere europtenne sous les Alpes occidentales
arguments gravimttriques et sismiques lits t l'anomalie
d'Ivrea.- Bull Soc. Gdol Fr., 26, 875-884.
Meissner, R. (1989).- Rupture, creep, lamellae and
crocodiles Happenings in the continental crust, Terra
Nova, 1, 17-28.
Meissner, R., Kuznir N. (1987).- Crustal viscosity and the
reflectivity of the lower crust, Am. Geophys., 5(B), 365-
374.
Milkereit B. (1987).- Migration of noisy crustal seismic
data. J. Geoph. Res. 92, 7916-7930.
Molnar P. (1988).- Continental tectonics in the after math
of plate tectonics.- Nature, 6,186, 131-137.
Mugnier J.L. et Vialon P. (1986).- Deformation and
displacement of the Jura cover on its basement.-J.
Struct. GeoL, 8, 373-387.
Mugnier J.L., Guellec S., Menard G. et Roure F. (1989).-
G6omtrie et structures du socle des Alpes externes
dtduites des proills ECORS-CROP Alpes ! et ECORS
Alpes 2.- C.tAcad. Sci., Paris, 309, (II), 733- 739.
Mugnier J.L., Guellec S., Menard G., Roure F., Tardy M.
et Vialon P. (1990).- Crustal balanced cross-sections
through the external Alps deduced from ECORS
profile.-Mern. $oc. gdoL France, 15( 203-216.
Nicolas A., Polino R., Hirn A., Nicolich R. and ECORS-
CROP Working-Group (1990).- ECORS-CROP
traverse and deep structure of the Western Alps.-Mdm.
$oc. gdoL France, Paris, 156, 15-27.
Nicolas A., Hirn A., Nicolich R., Polino R., ECORS-
CROP Working Group (1990).- Lithospheric wedding
in the western Alps inferred from the ECORS-CROP
Traverse.- Geology, 18, 587-590.
Pfiffner O. et Frei W., Finckh P. and Valasek P. (1988).-
Deep seismic reflection profiling in the Swiss Alps
explosion seismology results for line NFP 20 East.-
Geology, 16, 987-990.
Polino R., Dal Piaz G.V. and Gosso (1990).- The Alpine
Cretaceous orogeny : an accretionary wedge model
based on integrated stratigraphic, petrologic and
radiometric data.- Mdrn. $oc. gdoLFr., Paris, 156, 345-
367.
Rey D., Quartza T., Mouge P., Miletto M., Lanza R.,
Galdtano A., Carozzo M.T., Armando E., Bayer R.
(1990).- Gravity and acromagnetic maps on the western
Alps contribution to the knowledge on the deep
structures along the Ecors-Crop seismic profile.- Mdrn.
Soc. Gdol. Fr., Paris, 156, 107-122.
Roure F., Polino R. and Nicolich R. (1989).-
Poinonnement, rttrocharrriages et chevauchements
post-basculement dans les Alpes occidentales:
6volution intracontinentale d'une chaine de collision.-
C.R. Acad. Sci., Paris, 309, (II), 283-290.
Ringwood A.E. (1975).- Composition and petrology of the
Earth's mantle ;New York; Mc Graw-Hill, 618
Rybach L., Mueller S., Milnes A., Ansorge J., PBPrnoulli
D., Frey M. (1980).- The Swiss geotraverse Basel-
Chiaso- a review.-Eclogae geoL Helv., 73, 437-462.
Schiirch M.F. (1987).- Les ophiolites de la zone du
Versoyen : t6moin d'un bassin t 6volution
m6tamorphique complexe.- Thse Universitd de Gendve,
153 pp.
Tardy M., Fudral S., Deville E. (1986).- Coupe des Alpes
fran;aises, in rapport d'implantation du proill "structure
profonde des Alpes". lnstitut Franqais du Pdtrole Paris,
73 pp.
Tardy M., Deville E., Fudral S., Guellec S., M6nard G.,
Thouvenot F. et Vialon P. (1990).- Interpr6tation
structurale des donn6es du profil de sismique r6flexion
profonde ECORS-CROP Alpes entre le front Pennique
et la ligne du Canavese.- Mm. Soc. gol Fr., Paris, 156,
217-227.
Thouvenot F., Senechal G., Hirn A., Nicolich R., Hirn A.
(1990).-Ecors-Crop wide-angle reflection seismics :
contraints on deep interfaces beneath the Alpes.
Mdm. Soc. gol Fr., 156, 97-107.
Vialon P., (1990).- Structures profondes des Alpes et
6volution g6odynamique une introduction. - Mdm. Soc.
gol. Fr., Paris, 156, 7-15.
CRUSTAL REFLECTIONS FROM THE ALPINE OROGEN:
RESULTS FROM DEEP SEISMIC PROFILING
O. Adrian Pfiffner 1 Lucia Levato 2 and Paul Valasek 3'4
Abstract. Near-vertical deep seismic reflection Introduction
profiling through the eastern Swiss Alps offers
the opportunity to study exhumed crust and to
calibrate reflections by downplunge projections. Deep seismic reflection profiling is confronted
Lithologic layering, including intrusive contacts with the task of explaining the source of the re-
and relatively thin layers of metamorphosed Meso- flectivity of parts of crustal sections without the
zoic sedimentary rocks seem to be the primary help of calibration by well data as in the case for
reason for the observed reflective nature of the example of petroleum exploration. Reflectivity has
upper crust. Reflectivity is strongly controlled by been attributed to metamorphic or magmatic laye-
the orientation of this layering. Shear zones seem ring, shear zones or the presence of fluids. Only
to be responsible for reflections only ff they are rarely can down-plunge projections be used to
part of thick mylonite belts. identify the sources of reflectivity and ff so, only
Combined partial stacks of varied offsets used the shallow parts of the crust can be treated this
in conjunction with complete stacks proved to be way. Due to the particular geologic structures in
a valuable tool to overcome the difficulties asso- the Swiss Alps such down-plunge reflections are
ciated with 3-D effects of Alpine structure like reliable to a depth of some 10 kin. The aim of this
hummocky surfaces with pronounced cross dip. paper is to discuss possible origins of crustal re-
By combining refraction and near-vertical reflec- flections using this technique and to discuss
tion data it was possible to trace the European crustal reflections in the framework of collision
crust-mantle boundary into the Adriatic mantle. A tectonics.
wedge-shaped indenter of Adriatic lower crust The Swiss reflection profile NFP 20-EAST fol-
was apparently forced into the European crust, lows the trace of the European Geotraverse EGT
splitting it apart and delaminating it. Defocussing and crosses imbricated slivers of continental crust
effects due to internal deformation of the indenter from the core of the Alpine collision zone. Flakes
prevented the penetration of seismic energy in of upper crustal crystalline rocks were thrust
near-vertical reflection profiling. The stacking of northward on top of each other in the course of
delaminated slivers of upper crust during the col- the Alpine collision. In the later stages of the
lision resulted in a nappe pile. More or less co- Alpine orogeny this pile of basement flakes was
eval north- and south-directed thrusting lead to a subjected to erosion. Consequently, some of the
remarkably symmetrical thrust belt overlying a rocks observed at the surface today represent
southvergent subduction zone. former intermediate crust with greenschist and
amphibolite facies Alpine metamorphism.
The northern portion of the seismic line NFP
i Geologisches Institut, Universitiit Bern, Balt- 20-EAST is a transect through the Helvetic zone,
zerstr. 1, CH-3012 Bern, Switzerland the southern part is a transect through the Pen-
2 GRANSIR, Institut de Giophysique, Universiti ninic nappes of the Alps. The various tectonic
'de Lausanne, BSFH2, Dorigny, CH-1015 Lau- units are characterized by a pronounced axial
sanne, Switzerland plunge of up to 35 [Pffffner et al. 1990a]. These
3 Institut fiir Geophysik, ETH-HSnggerberg, structural features are best illustrated by the
CH-8093 Ziirich, Switzerland structural contour map shown in fig. 1. Here the
4 now at Dept. of Geology & Geophysics, Uni- top of the individual basement nappes, i.e. the
versity of Wyoming, Laramie, WY 82071, U.S.A. contact between crystalline basement rocks and
Triassic cover rocks, are contoured. As discussed
in Pfiffner et al. [1990b] and Sellami et al. [1990]
this interface is likely to be responsible for some
Continental Lithosphere: Deep Seismic Reflections of the seismic reflections. fig. 1 shows that the
Geodynamics 22 potential reflections have a relatively complex,
1991 American Geophysical Union hummocky shape and that CMP points are likely to
185
18(5 CRUSTAL REFLECTIONS FROM THE ALPINE OROGEN
southward subsurface continuation corresponds to
the Aar and Gotthard basement massifs. These
massifs are dominated by steeply dipping Variscan
structures, such as a prominent foliation, mag-
matic and other lithological contacts. It is there-
fore not a surprise that these segments of upper
crust, forming the southern margin of the Euro-
pean plate, appear seismically transparent.
The prominent reflections in the southern part
of the NFP 20-EAST transect can be attributed to
the Penninic nappes and will now be treated in
more detail. This nappe pile was strongly affected
by the Alpine deformations. Variscan structures
were in many instances completely overprinted.
As a consequence many of the lithologic contacts
are now at a shallow angle and often even parallel
to a penetrative foliation (either due to shearing
related to thrusting or isoclinal folding). From
this it follows that the internal structure of the
Penninic basement nappes seems more favorable
for producing seismic reflections from magmatic
and metamorphic layering as compared to the
steeply dipping structures discussed above. The
origin of the reflections from this southern part
of NFP 20-EAST will be illustrated by two repre-
sentative examples of typical Alpine geology.
The first example, shown in fig. 3, covers the
Spliigen zone, where a thin layer of Mesozoic car-
bonates, is pinched between two blocks of cry-
stalline basement rocks, the Tambo and Suretta
nappes. Comparing the two seismic sections of
fig. 3 it follows that the partial stack images the
Spliigen zone as a much more continuous reflec-
tion band, a feature which fits the geologic
structure of a continuous band of quartzites and
Fig. 1. Structure contour map of the basement- carbonates much better. The geologic profile con-
cover contact of the major tectonic units in the structed from field observations by Zurfliih [1961]
Alps of eastern Switzerland (altidues are in km). and Blanc 1965 shows that the Spliigen zone in its
The seismic line NFP 20-EAST follows the dyna- lower part consists of quartzites and carbonates,
mite shotpoints which extend along km 750. Inset: which are considered to be remnants of the au-
Location of the transect (arrow) in the framework tochthonous cover of the underlying Tambo base-
of the Alpine chains in Europe. ment. The upper part in the north is made up of
scatter off the line of recording. Nevertheless allochthonous carbonates detached along Triassic
these structure contour maps are a reliable base evaporitic layers (cargneu]e and anhydrite). The
for the identification of reflection events. overlying Suretta basement nappe consists of in-
The near-vertical reflection seismic data of terlayered orthogneisses and paragneisses.
NFP 20-EAST are discussed in Pfiffner et al. For a suite of samples of the various rock
[1988] (explosion data) and [1990b] (vibroseis types, seismic velocities and densities were deter-
data) in detail. Fig. 2 is a line drawing which mined in the laboratory [Sel]ami et al. 1990]. Most
was prepared from an unmigrated Vibroseis sec- of the norma] incidence reflection coefficients
tion. It shows a highly reflective uppermost crust computed for possible lithologic contacts are
around or above 5%. They are particularly high
except for the segment between shot points Va-
lens and Tamins, which corresponds to an uplif-
ted area where crystalline basement rocks outcrop
at the surface (Aar massif). The lower crust
lacks significant reflectivity, but a 1-2 s thick
band of reflections marks the crust-mantle boun-
dary. This will be discussed in more detail below.
Origin of upper crustal reflections
for the contacts between carbonates and crystal-
line basement rocks which are thus thought to be
the origin of the observed reflections.
The continuous band of reflections marked A
in fig. 3b is easily identifiable as the top of the
Spliigen zone according to down-plunge projec-
tions [see also Pfiffner et al. 1990b]. Also, field
observations suggest that the Spliigen zone
should thicken towards the north. Considering
the high reflection coefficients between carbonates
The segment of the NFP 20-EAST transect be- and crystalline basement rocks the reflection B in
tween shot points Valens and Tamins and its fig. 3B can be correlated to the base of the Splii-
PFIFFNER ET AL. 187
s
Simmi- Alten- Seve-Azmoos Ragaz Valens Kunkels Tamins Rothen- Canova Thuss Zdlis Ausser - Inner- Cr6t Burg Sovrana
bach dorf len
I I I I I I I
__.=- --_._ _ ...?c.=;_-/.... _
lO km
brunnen ferrera
I I I I oo, ol I I J
--- '"' ' ....... - - " /
/.%.
. .....%
/
Fig. 2. Line drawing of the Vibroseis dataset of NFP 20-EAST. Unmigrated time section
(8-48 Hz, 240 channels, 80 m group spacing, 40 m vibration point spacing, CMP traces
spaced at 2040 m, datum plane 700 m, coverage 120 fold). Localities given are shot-
points of near-vertical profiling by explosion seismology.
gen zone. The reflection C, on the other hand, the basement at the macroscopic scale (fig. 1)
occurs where the intrusive contact between the and the mesocopic scale (fig. 3c)it is conceivable
Truzzo granite gneiss and the paragneisses of the that scattering and defocussing effects are inevi-
Tambo nappe is expected. Given the reflectivity table even for near-vertical incidence and more
coefficient of 5% calculated for samples PF 9 and important when stacking the data from the entire
PF 16 [see also Sellami et al. 1990], this magmatic spread.
contact is the likely source for this reflection. The second example, involves a stack of thrust
Although some weak reflections from the inte- sheets of variable composition. Fig. 4a shows a
riot of the Suretta nappe seem to exist (e.g. D complete stack with discontinuous reflections
in Fig. 3B), the prominent layering of the Rofna (marked A), and a dipping event (marked B). In
Porphyry and the paragneisses of the Timun com- the partial stack shown in fig. 4b both, reflec-
plex does not seem to express itself seismically. tions A and B appear much more continuous.
One explanation might be that this layering is be- Moreover, additional reflections emerge (C in fig.
yond the resolution of the seismic experiment. 4b). The reflections B can be attributed to Meso-
The same might be true for a layer of mylonitized zoic carbonates overlying crystalline basement
Rofna Porphyry at the base of the Suretta nappe, rocks of the Gotthard massif [Pfiffner et al.
which has a pronounced velocity anisotropy with 1990b]. They will not be explored any further
the slow velocity (5.6 km/s, as compared to 6.4- here. Reflections A and C, on the other hand,
6.5 km]s) perpendicular to the gently east dip- are related and associated to the Aul and neigh-
ping foliation. In summary, the origin of reflec- boring nappes,
tions in the example of the Spliigen zone are best Fig. 4c is a geological profile drawn after data
explained by impedance contrasts related to litho- published by Nabholz [1945] projected down
logic and magmatic layering. plunge and covering the upper part of the seismic
The difference between the .stacks of fig. 3a section. The profile shows the complex geometry
and b might be caused by 3-D effects. DuBois et of these nappes which is characterised by inter-
al. [1990] in a study on special processing ap- layering of carbonates, lenses of ophiolites,
plied to Alpine seismic sections ascribe offset-de- crystalline basement and thick sequence of Juras-
pendent features to effects of geometry on the sic calcareous mica-schists (Biindnerschiefer).
ray path. Considering the hummocky geometry of The carbonates and ophiolites possess high ave-
188 CRUSTAL REFLECTIONS FROM THE ALPINE OROGEN
Crbt Burg
a!
I I
b!
2200 2300
Sovrana
0.5
-1.5
0.5
Cro! Burg Sovrana
i I 1
[_.+_ .... -- -___ -- Suretta nappe " I
--15 -- / S TWT WT
PF 4 6 29
(PF 8) 7 01
PF 10 08 PF 16 01
PF9 619
1 km
sample p ,
2 69 Marble
2 85 Lrlqz_L Banded marble and dolomite
Ivy,,; ,,,, Evaporites (cargneule and anhydrite)
ou,,z,,e
2 69 [-- --1-[ Orthogneiss 265
2 78 [---- ---'1 Paragnelss, micaschest, amphibolte
H&V
Fig. 3. Comparison of Vibroseis section and geologic profile from the Spliigen zone.
Localities are shotpoints (see fig. 2). A: Migrated stack of the entire spread. B: Mi-
grated stack of near traces spaced 981-1760 m from vibration point. C' Detailed geologic
profile showing a layer of sedimentary rocks (Spliigen zone) sandwiched between the
Tambo and Suretta basement nappes. Laboratory data of densities (p) and mean com-
pressional velocities (Vp) are from Sellami et al. (1990). Samples in brackets are from
equivalent rocks outcropping outside the area displayed.
PFIFFNER ET AL. 189
Canova
1400
a!
1500
Thusis
.
I km
c!
Canova
Canova
1400
Thusis
1500
1 km
s
Thusis
I
Grav a
Grava
-2 s
TWT
Aul -2 s
TWT
.._._ Vals Imbricates
Zervrella
1 km
H&V
sample 9p
' 4 5 2 67 [ Shaly-sandy calcareous (mlca)schlsl
PF 14 6 8 t 2 96 OphlOllles Jurassic
PF 13 6 99 2 70 [--[--[--[ (Calclle)marble, carbonate brecclas, quartzlie
(PF 8) 70t 2 85 Dolomite (marble), cargneule, quarlzlle Triassic
(PF 3) 60t 2 e2 [--i- -'-t-[ Augen gneiss (orthogneiss) / pre-Trlasslc
PF 2 6 05 2 64 [ --[ Paragnelss and micaschist
Fig. 4. Comparison of Vibroseis section and geologic profile from the Aul and neigh-
boring nappes. Canova and Thusis are shot points (see fig. 2). A: Migrated stack of
entire spread. B: Migrated stack of intermediate traces spaced 5661-6440 m from vibra-
tion point. C' Detailed geologic profile obtained by down-plunge projection. The Aul
nappe forms an approximately I km thick layer of carbonates interlayered with ophiolite
slivers. The Zervreila imbricate forms the top of the Adula nappe (figs. 2, 5). Labora-
tory data of densities (p) and mean compressional velocities (Vp) are from Sellami et
al. (1990). Samples in brackets are from equivalent rocks outcropping above the area
displayed. *: Vpare from first arrivals of dynamite shots, p from Schwendener 1984.
rage seismic velocities varying between 6.8 and only isolated, short and laterally discontinuous
7.2 km]s and high densities in the range between reflection events, for which a fluid related origin
2.5 to 3.0 g]cm3. These values, given in fig. 4c, has been evoked by Pfiffner et al. [1988]. The
were measured in the laboratory [Sellami et al. dipping events at 10-15s TWT beneath shot points
1990] and are corroborated by refraction seismic Burg and Sovrana migrate out of the section
experiments, which lead to the postulation of a [Holliger 1990] and arise within the Southern
high velocity layer of 6.5 km/s at a depth equi- Alps. As' discussed by Pfiffner et al. [1990b] the
valent to 2s TWT [Miiller et al. 1980; Ye and An- reflections might be caused by shear zones due to
sorge 1990]. either the Mesozoic stretching, Cenozoic collision
Considering reflection coefficients computed or 1ithologic banding. They are not explored fur-
from the seismic velocities and densities given in ther in this paper.
fig. 4c, reflections from several interfaces of the The north dipping reflections at about 8s TWT
Aul and associated nappes can be expected. The beneath shot points Burg and Sovrana also mi-
highest reflection coefficients, Rc=14-25%, are grate out of the section [Holliger 1990]. They can
between the shaly-sandy, calcareous micaschists be attributed to the mylonite belt of the Insubric
(the Biindnerschiefer) and the carbonates. The Line [Pfiffner et al. 1990b]. The Insubric Line
ophiolites show high Rc values when in contact separates the high grade Penninic nappes from
with marble (3.3%) but a low value (0.5%) for the the Southern Alps (fig. 5). This major fault ac-
observed contact with dolomite. The crystalline comodated Cenozoic dextral strike slip and south-
basement rocks are expected to produce reflec- directed reverse faulting [Heitzmann, 1987,
tions when in contact with micaschists and dolo- Schmid et al. 1989]. According to seismic data of
mites (Rc=17% and 11%) and possibly marble line NFP 20-S1 acquired 30 km west of the sou-
(Rc=3%). Considering that some of the thinner them end of the transect studied here, the In-
lenses of e.g. dolomites and ophiolites are less subtic line seems to flatten out at depth [Bet-
than 30 m thick (i.e. less than 1/4 or even 118 of noulli et al. 1990].
the wavelength), it follows that the fine structure Both, the Vibroseis (fig. 2) and the explosion
is probably beyond the resolution of the seismic seisinology dataset [Pfiffner et al. 1988] indicate a
experiment. Constructive interference is likely to prominent band of reflections marking the crust-
be hampered by the complex shape of indivudual mantle transition. In the European plate it ex-
layers (pinch-and-swell structures, folds at va- tends from 11-12s TWT beneath shot point Simmi-
rious scales and lateral discontinuities). bach to 16-17s TWT beneath shot point Canova,
The reflections visible in fig. 4b are more li- where it disappears with a southerly dip. Seismic
kely to be associated with the more regular larger refraction data [Miiller et al. 1980; Ye and An-
scale structures. The subhorizontal event A may sorge 1990] place the Moho at the base of this
well stem from the carbonate-ophiolite layer for- reflection band, with a jump in seismic velocities
ming the top of the Grava and the bottom of the from 6.2-6.5 to 8.2 km/s.
Tomiil thrust sheets. According to field observa- At the southern end of the transect horizontal
tion [Nabholz 1945] this layer thins out towards reflections at 13-14s TWT beneath shot point Soy-
the north, which might explain the disappearance rana are interpreted as the crust-mantle boundary
of the reflection north of station 1500. The re- of the Adriatic plate, which is also compatible
flection C might be associated with the contact with refraction data (op. cit. ). fig. 5 summarizes
between the micaschists (Biindnerschiefer) of the the crustal structure of the Alpine orogen, ex-
Grava and the carbonate-ophiolite layer of the Aul tending the seismic line NFP 20-EAST to include
thrust sheet. To the north this reflection disap- the Southern Alps and the South- and North-al-
pears. Considering the geologic structure (fig. pine foreland basins. One of the outstanding fea-
4c) this might be due to alefocussing effects cau- tures to emerge is the fairly symmetric structure
sed by the complex folds of carbonate lenses displayed by the imbricate thrust sheets involving
south of shotpoint Canova at a depth equivalent upper crustal rocks. The geometry of the crust-
to about 2s TWT. The seismic section does not mantle boundary shown in fig. 5, although clearly
give any indication of the frontal antiform of the visible on parts of the near-vertical reflection
Aul thrust sheet. data, is based on a new interpretation of the re-
The observed reflections in the seismic section fraction data [Ye and Ansorge 1990]. The refrac-
of fig. 4 match the geological structure rather tion data were formerly interpreted with the Moho
well, although down-plunge projection is over a forming a continuous 52 km deep trough beneath
large distance of 30 kin. Similar to the Spliigen the central portion of the Alps. The near-vertical
zone, reflections seem also to be related to litho- reflection data, however, suggested that this
logic layering and - as to be expected - depend "trough" was highly asymmetric, with the Moho of
on the geometry of the interfaces. the European plate dropping to successively grea-
ter depths towards the south, whereas the Moho
beneath the Southern Alps seems to be situated at
Reflections from the lower crust a significantly higher level. The new model pre-
sented here, compatible with both the reflection
and refraction datasets, suggests a gently dip-
With the exception of the southernmost portion ping slab of European lower crust extending be-
of the seismic line, the lower crust (fig. 2) shows neath the upper mantle of the Adriatic plate's
PFIFFNER ET AL. 191
NFP 2O - EAST
S.ntis Chur
Insubric Line Lecco
10 A a r Adula
Simano
Gotthard
20 upper crust -. __ ...

,,
3O
40-
50-
60-
70-
80-
90-
lOO
upper crust
lower crust
,.. crust
..,ro p '
mant ,",' Cenozoic intrusives *. ..... Fragments of oceanic crust & sed.
Mesoz.-Cenoz. of Penninic nappes
.... "'": Mesozoic of Austroalpine nappes '.- .' '. Cenozoic } Helvetic nappes, Pre-Triassic basement /Mesozoic North- & Southalpine foreland
i i
20 150 100
mantle
lO
2o
3o
4o
50
6o
70
- 80
90
, 100
50 [km]
Fig. 5. Geologic profile and model of the crustal structure based on surface data, deep
seismic near-vertical reflection and refraction profiling. A wedge of the Adriatic plate is
forced into the European plate splitting it apart. The upper crust is stacked up sym-
metrically, the lower crust got subducted.
northern edge. Subduction of European lower of the segment of lower crust is at least of the
crust has been proposed earlier [e.g. Laubscher order of 40-50 km judging from the dimension of
and Bernoulli 1982] and is required as a conse- the lowermost imbricate of upper crust implied by
quence of crustal balancing. The stack of upper this late phase shortening (fig. 5). It is proposed
crustal rocks recognised within the Helvetic and that this lower crust was forced as wedge into
Penninic zones requires a considerable volume of the European crust splitting it apart and thereby
lower crust, from which it was detached. This thickening it. The wedge was likely to have un-
lower crust was thinned during the Mesozoic rif- dergone extensive folding and shearing. The re-
ting associated with the opening of Tethys. The sulting complex internal structure may prevent
exact amount of this extension is difficult to the penetration of seismic energy and would ex-
assess. Nevertheless the estimated length of plain the lack of reflections from near-vertical re-
stacked upper crustal material within the Penninic flection profiling. The top of the wedge cortes-
zone suggests that a substantial amount of lower ponds to a zone of high velocities as determined
crust must have disappeared into the mantle du- from refraction seismology. These could stem from
ring Cenozoic subduction. In the model presented fragments of oceanic crustal material of Tethys.
in fig. 5, the south-dipping European Moho is The oceanic suture zone between the Austroalpine
taken to indicate the trace of this subduction and Penninic nappes marking the former plate
zone. It points to the area beneath the Po Plain boundary between the European and Adriatic
where several geophysical parameters suggest the plates is represented by nappes containing mafic
existence of a lithospheric root [Milllet 1989]. and ultramafic rocks (indicated by dots in fig.
Similar to the Penninic zone, the Southern 5). This suture zone must have extended down
Alps also require a certain volume of lower crust into the mantle, but is truncated in the south by
to satisfy crustal mass balance constraints. In the the Insubric Line. The continuation of the Euro-
external (southern) part of the Southern Alps im- pean-Adriatic suture zone at depth is shown to
portant crustal shortening seems to have occurred straddle the Insubric Line down to about 25 km
very late in the collision. This is indicated by the where it is disrupted by the indenting wedge.
Pliocene basin fill of the Apennine foredeep, The site of this crustal indentation is also
which lies discordantly above the folded and im- marked by the generation of granitic melt which
bricated Miocene sediments of the Southalpine intruded some of the Penninic nappes. In fig. 5
fold-and- thrust belt buried beneath the Po Plain this magmatism is shown to be spatially and tem-
[Roeder 1989 and references therein]. The length potally related to the mylonitization along the In-
subtic Line, which might have played a role in crust-mantle boundary beneath this wedge ob-
channelling magma ascent. A prominent subhori- served in the near-vertical reflection experiments,
zontal reflection at 7 s TWT beneath shotpoint while the refraction experiments (undershooting
Sovrana in fig. 2 is tentatively interpreted as the the wedge) indicate a continuation of the Euro-
top of an intrusive body. pean lower crust into the mantle beneath the
Adriatic plate. The reflection events near the
southern end of the seismic line, which are
Discussion and conclusions thought to originate from the stretched crust of
the Adriatic plate could be compared to "lamellae"
[Meissner et al. 1990].
In the transect through the eastern Swiss Alps Integrating also the refraction seisinology data
the pronounced reflectivity of the predominantly it can be shown clearly that the Cenozoic collision
crystalline basement rocks of the upper crust is stage of the Alps involved southdipping subduc-
mainly due to gently dipping lithologic and mag- tion of lower crust. The upper crust was peeled
matic layering. Particularly strong reflections are off and stacked up to a pile of thrust sheets. Im-
associated with contacts between carbonates and brication was to the north in the northern part of
gneissic rocks, carbonates/ophiolites and mica- the orogen, i.e. the Penninic and Helvetic zones,
schists, and between granitic and gneissic rocks. and coeval but south-directed in the Southern
To the north of the transect, where magmatic and Alps. The latest stage of the collision is marked
metamorphic layering dip steeply, the upper crust by an indenting wedge of Adriatic crust stuck in
appears seismically transparent. the delaminated European crust.
The prominent foliation present in the high
grade rocks, which parallels the lithologic laye-
ring and nappe boundaries, can in general not be Acknowledgments. NFP 20-EAST is financed by
related to high reflectivity. Either the velocity the Nationales Forschungsprogramm 20 of the
anisotropy associated with such a foliation is by Schweizerischer Nationalfonds. We thank I. Blaser
itself not sufficient to produce reflections, or the and W. Schaad for technical assistance. The pa-
thickness of the ductile shear zones parallel to per benefitted from thoughtful comments by two
the basal thrust faults of basement nappes is be- anonymous reviewers.
yond the resolution of the seismic experiment.
Reflections seemingly correlated to the mylonite
belt of the Insubric Line suggest that ductile de- References
formation zones of adequate thickness for seismic
resolution can produce reflections. This conclu-
sion is also corroborated by results from the Bernoulli, D., Heitzmann, P. and Zingg, A.,
seismic line NFP 20-W3 [S.A.R. 1988] on which Central and Southern Alps in southern Switzer-
the mylonite belt of the Simplon Line in western land- tectonic evolution and first results of
Switzerland is responsible for a prominent reflec- reflection seismics. In: Route, F., Heitzmann,
tion event. P. and Polino, R. (eds): Deep structure of the
The use of partial stacks of varied offsets in Alps. Mm. Soc. gol. France, 156, 289-302,
conjunction with complete stacks turned out to be 1990.
a fruitful approach in unravelling the upper crus- Blanc, B.L., Zur Geologie zwischen Madesimo
tal structure because partial stacks very often und Chiavenna (Prov. Sondrio, Italia), Mitt.
produce more coherent reflection events than en- Geol. Inst. ETH und Univ. Ziirich, 37, 134
tire stacks. DuBois et al. [1990] relate this effect pp., 1965.
to the influence of the subsurface structure and DuBois, L., L. Levato, J. Besnard, R. Marchant,
geometry on the ray paths of the far-offset tra- R. Olivier, M. Ouwehand, S. Sellami, and J.J.
ces. This study identifies scattering and defocus- Wagner, Aspects particuliers du traitement sis-
sing of seismic rays as being mainly due to irre- mique des profils alpins suisse du PNR 20. In:
gularly shaped surfaces and lensoid, folded and Route, F., Heitzmann, P. and Polino, R.
laterally discontinuous layers. Partial stacks (eds): Deep structure of the Alps. Mm. Soc.
should not be used independantly and particular Gol. France, 156, 47-54, 1990.
attention has to be paid to 3-D effects on 2-D mi- Heitzmann, P., Evidence of late Oligocene/early
gration. Miocene backthrusting in the central alpine
The pattern of reflectivity in the Alps lacks "root zone", Geodyn. Acta, 1, 183-192, 1987.
features like crocodiles which are recognised in Holliger, K., A composite depth-migrated seismic
the Variscan chain and attributed to interwedging reflection section along the Alpine segment of
of layers [Meissner et al. 1990]. The postulated EGT derived from NFP20 eastern and southern
indenter of Adriatic crust into European crust traverses. In- R. Freeman, P. Giese and St.
corresponds kinematically to a crocodile but is a Milllet (eds): The European Geotraverse, inte-
larger scale feature. It lacks internal reflections, grative studies. Eurolean Science Foundation,
which could be explained by complex deformation Strasbourg, France, 245-253, 1990.
within the indenting wedge. This would also ex- Laubscher, H., and D. Bernoulli, History and
plain the disappearance of reflections from the deformation of the Alps, in Mountain Building
PFIFFNER ET AL. 193
Processes, edited by K. J. Hsii, Academic S.A.R., Schweizerische Arbeitsgruppe fiir Refle-
Press, 169-180, 1982. xionsseismik, Vorliiufige Ergebnisse der Alpen-
Meissner, R., Wever, Th. and Sadowiak, P., Re- traversen des NFP 20 "Geologische Tiefenstruk-
flectivity patterns in the Variscan mountain tur der Schweiz", Bull. Vet. schweiz. Petro-
belts and adjacent areas' an attempt for a pat- leum-Geol. u.-Ing., 54/126, 1-30, 1988.
tern recognition and correlation to tectonic Schmid, S.M., Aebli, H.R., Heller, F. and
units. Tectonophysics, 173, 361-378, 1990.
Miiller, St., J. Ansorge, R. Egloff and E. Kiss-
ling, A crustal cross section along the Swiss
Geotraverse from the Rhinegraben to the Po
Plain, Eclogae Geol. Helv., 7_3, 463-483, 1980.
Milllet, St., Deep-reaching geody-namic processes
in the Alps. In' Coward, M.P., Dietrich, D.
and Park, R.G. (eds.) Alpine Tectonics, Geol.
Soc. Spec. Publ. 4__5, 303-328, 1989.
Nabholz, W.K., Geologie der Biindnerschieferge-
Zingg, A., The role of the Periadriatic Line in
the tectonic evolution of the Alps. In- Coward,
M.P., Dietrich, D. and Park, R.G. (eds.)
Alpine Tectonics, Geol. Soc. Spec. Publ. 4_5,
153-171, 1989.
Schwendener, H., Ein gravimetrisches Krusten-
Mantel-Modell fiir ein Profil vom nSrdlichen A1-
penvorland bis an die Ligurische Kiiste, Geodii-
tisch-geophysikalische Arbeiten in der Schweiz,
3__6, Schweiz. Geodiitische Kommission, 1984.
birge zwischen Rheinwald, Valser- und Safien- Sellami, S., F. Barblan, A.-M., Mayerat, O.A.
tal, Eclogae Geol. Helv., 38, 1-119, 1945.
Pfiffner, O.A., W. Frei, P. Finckh, and P. Va-
lasek, Deep seismic reflection profiling in the
Swiss Alps' Explosion seisinology results for
line NFP 20-EAST, Geology, 16, 987-990, 1988.
Pfiffner, O.A., E.M. Klaper, A.-M. Mayerat,
and P. Heitzmann, Structure of the basement-
cover contact in the Swiss Alps. In' Route, F.,
Heitzmann, P. and Polino, R. (eds)' Deep
Pfiffner, K. Risnes, and J.J. Wagner, Com-
pressional wave velocities of samples from NFP
20-EAST profile. In- Route, F., Heitzmann, P.
and Polino, R. (eds): Deep structure of the
Alps. Mm. Soc. Gol. France, 156, 77-84, 1990.
Triimpy, R., Penninic-Austroalpine boundary in
the Swiss Alps' a presumed former continental
margin and its problems, Am. J. Sci., 275-A,
209-238, 1975.
structure of the Alps. Mm. Soc. Gol. France, Weisseft, H.J. and Bernoulli, D., A transform
156, 247-262, 1990a. margin in the Mesozoic Tethys' evidence from
Pfiffner, O.A., Frei, W., Valasek, P., Stiiuble, the Swiss Alps, Geol Rdsch. 74/3, 665-679,
M., Levato, L., DuBois, L., Schmid, S.M. and 1985.
Smithson, S.B., Crustal shortening in the A1- Ye, S. and Ansorge, J., A crustal section
pine orogen- results from deep seismic reflec- through the Alps derived from the EGT seismic
tion profiling in the Eastern Swiss Alps, Line reflection data. In- R. Freeman, P. Giese and
NFP 20-EAST, Tectonics, 9/6, 1327-1355, St. Mailer (eds)- The European Geotraverse,
1990b. integrative studies. European Science Founda-
Roeder, D., South Alpine thrusting and trans-A1- tion, Strasbourg, France, 221-236, 1990.
pine convergence. In' Coward, M.P., Dietrich, Zurfliih, E., Zur Geologie des Monte Spluga,
D. and Park, R.G. (eds.) Alpine Tectonics, Mitt. Geol. Inst. ETH und Univ. Ziirich, C]83,
Geol. Soc. Spec. Publ. 4__5, 211-227, 1989. 117 pp., 1961.
INTEGRATED ANALYSIS OF SEISMIC NORMAL INCIDENCE AND WIDE-ANGLE
REFI.JCTION MEASUREMENTS ACROSS THE EASTERN SWISS ALPS
ETH Working Group on Deep Seismic Profiling*
*J. Ansorgel, K. Holliger 1, P. Valasekl, 2, S. Ye 1,
P. Finckh 1, R. Freeman 1, W. Frei 1, E. Kissling 1, P. Lehner 3,
H. Maurer 1, St. Mueller 1, S.B. Smithson 2, and M. Stiiuble 1
Abstract. Deep crustal seismic data extending along a N-S
transect through the eastern Swiss Alps have been compiled to
form a detailed image of the crustal structure across a portion
of this continental collision zone. The data sample the entire
Alpine edifice beginning in the Molasse Basin in the north and
ending in the Po Plain. The image incorporates both large-
scale refraction and wide-angle reflection measurements as
well as near-vertical deep-crustal reflection profiles. The
reflection surveys are part of the Swiss National Science
Foundation Program NFP-20 and are almost coincident with
the Alpine segment of the European Geotraverse (EGT).
Examples of the two types of data are displayed separately
and then combined as a single composite section. The EGT
wide-angle reflection data are shown also as NMO-corrected
profiles. The standard, reduced traveltime sections were used
to select seismic events to base forward modeling by 2D
raytracing. The NMO-corrected wide-angle data reproduce the
sections in a normal-incidence format which can be compa,'ed
directly to the deep-crustal reflection profiles. The reflection
data from several of the NFP-20 traverses are compiled after
migration in a line-drawing format onto a single transect.
These migrated reflection data are superimposed on the crustal
velocity structure obtained from refraction modeling which
includes generalized velocity information from profiles
crossing the EGT. Most of the main structural seismic features
correlate well between the reflection and refraction data sets. In
particular, a remarkable agreement exists between the
reflection and refraction Mohos. From the refraction data the
base of the European crust is shown to dip from 35 km depth
lInst. of Geophysics, ETH-Htnggerberg, CH-8093 Ziirich,
Switzerland;
2Dept. of Geology and Geophysics, University of
Wyoming, Laramie, Wyoming 82071, USA;
3National Research Program 20 Coordination, Speerstrasse
39, CH-8805, Richterswil, Switzerland)
Geodynamics 22
in the north to 56 km beneath the Penninic nappes. A 18 km
thick slab of lower crust is subducted further south under the
Southern Alps. On the reflection data, this interface correlates
well with the base of a 3-6 km thick package of deep
reflections which suggests a transitional velocity-depth
distribution. Steeply dipping structures such as the Insubric
line are less well imaged on the reflection profile which
crosses it, but are clearly visible on the more distant high-
resolution reflection spreads. Overall, the advantage of this
integration is a unified data base whereby the large-scale
velocity distribution derived from the refraction measurements
is resolved in fine stmctmal details by the reflection data.
Introduction
The Alps are the surface feature of the recent continental
collision zone between the European and African lithospheric
plates, which has been investigated by various geological and
geophysical methods over many years. The most detailed
seismic crustal cross section available until recently was
compiled by Mueller et al. [1980] which was based on
structural interpretations of a reasonably large amount of
seismic wide-angle data projected onto a NW-SE oriented
geotraverse through Switzerland between the towns of Basel
and Chiasso. Recently two large-size longterm research
programs have added a wealth of new data from which a fully
revised structural picture of the Earth' s crust under the Alps
could be developed. This report summarizes the present status
of studies on the gross crustal structure under the central and
southern Alps of Switzerland using active seismic methods.
One data set was acquired as part of the seismic refraction
and wide-angle reflection survey along the north-south
oriented European Geotraverse (EGT) (Fig. 1) [Galson and
Mueller, 1986]. In 1986 a large-scale seismic refraction
experiment was carded out along the Central Segment
(EUGEMD of the EGT between Kiel in northern Gy and
Genova at the Ligurian coast. The EUGEMI Working Group
[1990] has published a detailed description of this experiment.
All data and technical parameters are contained in the
compilations by Aichroth and Prodehl [1990], Aichroth et al.
[1990], Maistrello et al. [1991], and Ye and Ansorge
[1990a,b].
195
196 INTEGRATED ANALYSIS ACROSS THE EASTERN SWISS ALPS
N 6 8 10 12 14 16
50
48
46
44
2 40
i
Cenfra['
Marseille
4 6
Rhenish Massi
Luxembourg
.Bern
I
8 10 o
EGT
Bohemian
.M(Jnchen
Aps
100 km
50
48
46
44
12 14 16 E
Fig. 1. Location of the European Geotraverse (EGT) (--) and the deep reflection traverses of the Swiss
National Science Project NFP-20 (**) in central Europe with major tectonic units.
The complementary second set of data was provided by the
implementation of the special research program NFP-20 of the
Swiss National Science Foundation [Frei et al., 1989]. The
near-vertical deep seismic reflection surveys of this program
comprise a western, an eastern and a southern traverse across
the western and eastern Alps of Switzerland which were
carried out in 1987, 1986, and 1988, respectively, at the
locations indicated in Figs. 1 and 2. The eastern and southern
traverses are situated mostly parallel and close to the Alpine
section of the EGT [Pfiffner et al., 1988, 1990; Bernoulli et
al., 1990; Valasek et al., 1991]. In order to probe the entire
crust from the near surface sediments to the Moho Vibroseis
and explosions were used as sources along all three sections.
The extensive processing procedures are described in detail by
Valasek et al. [ 1990].
All the aforementioned data were obtained along profiles
which are oriented more or less perpendicularly to the tectonic
strike of the Alps (Fig. 2), which in such a structurally
complex area may lead to severe spatial aliasing problems in
the case of the seismic wide-angle data. Therefore, the reliable
interpretation of these data depends heavily on the velocity and
structural information obtained from the seismic refraction
profiles parallel or subparallel to the tectonic strike. This vital
information was taken from the third set of data which were
collected over many years and interpreted by several authors
[Ottinger, 1976; Egloff, 1979; Deichmann et al., 1986; Maurer
and Ansorge, 1991; Yan and Mechie, 1989]. Figure 2 gives an
impression of the density of these seismic refraction
experiments which go back as far as 1958 [cf. Mueller et al.
[1980] and references therein].
This survey of the crustal structure covers geographically the
north-south section between the Molasse Basin and the Po
Plain across the Central Alps of Switzerland (Fig. 1). It
traverses the tectonic units of the Molasse Basin, the Helvetic
zone, the Aar Massif, the Penninic nappe zone, the Southern
Alps and the Po Plain. Figure 2 shows in more detail the
locations of the seismic refraction profiles and reflection
traverses of the three data sets described above.
Seismic Wide-angle Data
Two examples of the seismic refraction data are shown here
which have their common shotpoint ED (Fig. 2) at the
northern boundary of the Helvetic zone. The reversed profile
in Figure 3 runs parallel to the northern edge of the Helvetic
nappes between shotpoints ED and J (Fig. 2) [Maurer and
Ansorge, 1991]. It is one of the more recent profiles parallel to
the tectonic strike which cross or are closely connected with
the north-south oriented EGT traverse, whereas the section in
Figure 4 was obtained on the EGT proErie southward from
shotpoint ED. The correlations of the most prominent phases
which were interpreted as wide-angle reflections or as
refracted arrivals are indicated.
The forward modeling was carded out by iterative raytracing
with the program RAY87 developed by Luetgert [Hill et al.,
1985]. The resulting ray diagrams of crustal phases for the
ETH WORKING GROUP 197
.0 7.0 e.o '9.o EGT 10.0E
' ............ .......:.:.:.:.:.:.:.:.:.:.:.
..::. s i
..5':':'::'-'.
IL
MC EB i
7.0 ,1o i
Fig. 2. Location map of the EGT profile across the Alps, refraction seismic surveys parallel or subparallel
to the tectonic strike, and NFP-20 reflection traverses. 1: Tertiary sediments, Molasse; 2: Tabular Jura
mountains; 3: Folded Jura mountains; 4: Crystalline massifs and basement; 5: Helvetic autochthonous
sediments, Helvetic and Ultrahelvetic nappes; 6: Penninic nappes; 7: Austroalpine nappes; 8: Southern
Alps; 9: Tertiary intmsives and volcanics.
EE,ED,EC,EB: Shotpoints of the EGT;
ET eastern, ST southern, WT western seismic reflection traverses (***) of the Swiss National Science
Project NFP-20 with shotpoints M1,2.
Supplementary and older shotpoints and profiles: A,B,C,D: Alpine Longitudinal Profile (ALP75); AT:
Albiano/Trento; ES: Eschenlohe; GG: Gornergrat; LB: Lago Bianco; LL: Lago Lagorai; LT:
Lentate/Varese, MC: Mt. Cenis; OG: Orgelet; SB:Steinbrunn; SU: Sulz; TO: Osttirol; TS: Tournus; J:
Jaunpass; SA: Saviese; BR: Brigels; S4,S5: Black Forest; ZI: Hohenzollerngraben,
record sections ED-southwest and ED-south are shown in the
lower parts of Figures 3 and 4 together with a comparison of
observed and calculated traveltimes of the main phases. As
was to be expected, the Helvetic crustal cross section differs
significanfiy from that along the north-south line by its
continuity along the profile if one disregards the shallow
sedimentary structure. The recordings of section ED-south are
strongly influenced by the Helvetic nappes overlying parts of
the Molasse Basin and by the adjacent Aar Massif.
The velocity-depth distributions derived from this and older
profiles crossing the EGT serve as the basic outline for the
initial model of the crustal cross section along the EGT. This is
then modified iteratively by raytracing until a satisfying fit
between the correlated and calculated traveltimes is achieved
(Fig. 4, lower part). The earlier derived structure of the initial
model at the points of intersection have to be kept unchanged
as much as possible. The system of reversed and overlapping
profiles from the EGT shotpoints ED (northern boundary of
the Helvetic zone), EC (Insubric line), EE (northern boundary
of the Molasse Basin), and EB (southern boundary of Po
Plain) (partially situated slightly outside the frame of Figure 2)
together with the structural constraints provided by the older
profiles has led to the velocity-depth distribution shown in
Figure 7a. Additional record sections were obtained on the
EGT profile from explosions M1 and M2 (Figs. 2 and 7a) of
the eastern NFP-20 reflection traverse [Pfiffner et al., 1988,
1990] (Fig. 7a). The average station spacing of 3 km on the
north-south profiles limits the resolution of shallow structures.
Therefore, the derived seismic near-surface structure
represents the geological situation only schematically.
Normal Incidence Seismic Reflection Data
As an example of the NFP-20 near-vertical reflection data
Figure 5 shows an assemblage of single fold dynamite
recordings of four subsegments from the southern traverse
(Fig. 2). First processing steps include filtering, as well as
static and dynamic corrections. The following weighted
coherency ffitering and plotting of only the strongest positive
amplitudes leads to the computer-generated unmigrated line
drawing form. The complete processing sequence is described
by Valasek et al. [ 1990]. Despite the poorer spatial resolution
limited by the average distance of 5 km between the shots
r--] 6
LJ 5

x !
0
t-- -1
5
200 190 180 170 160 150 ltl0 130 120 110 100 90 80 70 60 50 tl0 30 20 10 0
ED
,
correlated
calculated
-lO
150
ED
6.15
6.35
6.25
6.60
6.40
8.10
I ' ' ' 510 .... I .... I '
0 100 150
DISTRNCE [KM]
Fig. 3. Record section of the profile from shotpoint ED to J along the northern
margin of the Helvetic zone (Fig. 2) with reduced traveltime (reduction velocity 6.
km/s) together with raytracing diagram, correlated, and calculated phases. PIP,
P2P, P3P are wide-angle reflections from crustal interfaces.
considerable significant shallow and deep reflectivity appears
on all sections. On section $1 a prominent north-dipping
reflector can be observed between 4 s in the south and 6.5 s
two-way traveltime (TWT) in the north below the shallow
reflections from the Penninic units (P). A band of strong
reflections (L2) appears in the lower crust dipping slightly to
the north from 7.5 s to 9 s TWT. Less pronounced reflections
(M2) are visible near the northern and southern ends of
sections S1 and S3, respectively, between 12 and 13 s TWT.
Considerably clearer is a broad band of reflected energy (M 1)
on section $1 dipping southward between 16 s and 18 s TWT.
Another way to bring out the main structural features of the
Alpine crust is shown in Figure 6 [Valasek, 1990]. The EGT
wide angle reflections from the four shotpoints EE, ED, EC,
and EB (Fig. 2) were processed similarly to the near-vertical
reflection data with special emphasis on the deep crustal
reflections. The individual seismograms were shifted to a
common datum of 700 m and a "NMO correction" was applied
with an average crustal velocity of 6.1 km/s with respect to a
dip of the European Moho of 15 degrees. After muting of
shallow reflections and refracted arrivals the signals, which
were recorded beyond the critical distances of lower crustal
events, were then plotted at their midpoint locations. Figure 6a
shows a composite section of the EGT wide-angle reflection
recordings and the corresponding unmigrated NFP-20
reflection data from the various subsegments of the eastern and
southern traverse. In Figure 6b the main crustal reflection
events are correlated and labeled as described in the previous
paragraph and by Valasek et al. [1991], i.e. H (Helvetic zone),
P (Penninic zone), L1,L2 (lower crustal reflections), and
M1,M2 (reflections from the base of the crust). Figure 6 thus
displays the TWT topography of lower-crust and Moho
reflections along the Alpine traverse.
The reliable integrated analysis of the reflection and
refraction data along a single continuous north-south traverse
requires two more steps. The reflection data of the various
segments especially on the southern segment have to be
projected onto a single continous profile which allows to
extend the eastern traverse further south. Subsequently the
composite reflection section has to be migrated properly using
a representative smoothed crustal velocity structure.
The first condition was fulfilled by projecting the time
sections of the southern reflection segments onto the
southward extended eastern traverse roughly east-west along
the trend of the Bouguer gravity anomaly after the gravity
contribution of the Ivrea body was subtracted [Kissling, 1984;
Kissling et al., 1983]. Most likely this corrected Bouguer
anomaly represents the gross structure of the Alpine crust best
over a few tens of kilometers.
The second step depends heavily on the proper choice of
migration velocities which are the true average layer veocifies
of the considered crustal depth range [Valasek and Holliger,
1990; Holliger, 1990 and 1991; Holliger and Kissling, 1991].
For this purpose the third data set, i.e. all significant crustal
refraction profiles running parallel to the strike of the Alps and
lying close to the investigated traverse were reanalyzed for the
major crustal velocities and interfaces. This has led to a
schematic four-layer model with 5.0 km/s for sediments, 6.1
0 1o 20 30 qo 50 60 '70 80 90
> SOUTH oSR.CE
AHPLITUDE TRACE NORHALIZED:
100 110
(KH)
120 lqO 130 150 160 170 180 190
BANDPASS 1.5 - 15 CHZ3
6
.
lO
20
-30
qO
50
0
[ I I I , I , I
+x +
[ I , I , I , [ I
x correltted
+ X calculated
XX X
X
xx x .+xl- xd. x+ x x x l-
X X XX xxX X X X X X X X
+ +x++
+'1' X X
Xx x
I ' I ' I ' I ' I ' I ' I ' I ' i
20 q0 60 80 100 120 lq0 160 180
EC
I [ I I i I [ I
' I ' I ' ' I ' I ' I ' I ' I
20 q0 60 80 100 120 lq0 160 180
DISTANCE (KH)
Fig. 4. Record section of the profile from shotpoint ED to the south on the EGT (Fig. 2) perpendicular to
the tectonic strike of the Alps with reduced traveltime together with raytracing diagram, correlated, and
calculated phases. The correlated phases are explained in the text. For seismic velocities see Fig. 7a.
-3
-2
-1
lO
20
3O
q0
50
N
STAT I ON
0.000
1.000
4. 000
?.000
11.000
12.000
17.000
xwx ()
NF-88-S1 NF-88-S3
NF-88-S4 NF-88-SS
STAT I ON
:_?:..:::3:;,",.'.!-2.. ,-'-...-: -- C'.::m. ,,.c.. :c..:..., '
. ......... ....... ::. , :[m ';..,:::,*'( ..... ..,:..:-
' . ........ :. ,- -:-.' _ -.-- ::::'""::'-:"" ':".: ....... 3:,,;..''::,?':;:: 2. o
.... ' -: "-:.7'-. ',:...::-...:.' ' ' ." 'c':': .;::. - -.;:;:-: "f:'["::"" ;::? -'::'c'
.... !-?:.:. '. ,
:.: - '.':;. :.: ....:.. ' : . ';.'-. :::... ',,'::.c...'
L2 .'::7-'-. -- :: .,.:.::;: :':':. ,:::'.:.'":f. '- '
.... --::'::'..-':..;, :a-'f'::5- ,,z. ::: ..... -(- ' - ':.".::'<':'::"."-.:"'' ;.:?':.::.:.'.:-'::Y' .:-;J.':;":;3. "::;:.% '"?': ' O
::::':.':;:-c :.-.:::..-":.'..:. f..-;.:'c.......'?T...:-.:-.;2'.:<' -.' -'...':;: -.-."5
l O
1S O
:'?',':'..::::..:': -..._:'...:"--.c..:.. " :;:::.'.;?':.:;'::'-:.::'::4::'..;f:/:':?':' "';;.?. ':"?'?'";.-" "'.Y.'/C:":::
l OOO
-? ..-C.'..:;:..,:;]:?'- :,:..:. :;;..:?. ;.;..:..:. :.Z.:: :-;-.cm -
O
o TWT (s)
Fig. 5. Line drawing of the southern reflection traverse (ST in Fig. 2) recorded from dynamite sources
only. Single-fold subsection S 1, S3, S4, S5 are shown. The labeled reflections are discussed in the text.
km/s for the upper crust, 6.5 km/s for the lower crust, and 8.1
km/s for the sub-Moho velocity. The previously composed
line drawings and the reflection points of the average velocity
model were then depth migrated simultaneously by an
advanced ray-theoretical method developed by Holliger [ 1990,
1991] the result of which is shown in Figure 7b.
Crustal cross section
In a final step the crustal models along the Eastern traverse
through the Swiss Alps derived from refraction, wide-angle
reflection and near-vertical reflection surveys have to be
merged and compared. The two independently derived crustal
cross sections displayed in Figures 7a and 7b are combined in
the overlay of Figure 7c. They show a very good agreement in
crustal structure.
The careful forward modeling for the reversed and
overlapping EGT wide-angle profiles has led to the velocity-
depth distribution in Figure 7a for the section between the
Molasse Basin and the Po Plain. In the north the model joins
the crustal structure derived by Aichroth & Prodehl [1990] for
the central segment of the EGT and in the south two rather
similar extensions of the crustal structure to the Ligurian sea
were proposed by Ye and Ansorge [1990b] and by Buness
and Giese [ 1990].
The least differentiated crustal structure underneath the
sediments is found under the southern Molasse Basin and the
northern part of Helvetic zone. A comparison with the
migrated reflection data of Fig. 7b is somewhat scanty because
the survey does not continue into the Molasse Basin and the
migrated deep crustal reflections do not reach far enough to the
north.
The section between shotpoint ED at the northern margin of
the Alps and shotpoint EC at the Insubric line shows
pronounced horizontal structural variations. Major elements in
the upper crust are the relatively low velocity in the Aar Massif
compared to the well documented higher velocities in the
metamorphic Penninic nappes derived from the observations
of shots at M1, M2 and EC. The upper and middle crust
beneath the Aar Massif seems to be devoid of significant
reflections. It separates the highly differentiated uppermost
crust consisting of the Helvetic nappes with well proven
structural details in the north [Pfiffner et al., 1990] from the
Penninic units in the south. Stguble et al. [1991] have
determined the structure between the Molasse Basin and the
North Shot E D 5hot E C South
DISTANCE (km ) So
o
lO
TWT
Isl
20
30
(a) ,,o.m km
,oo, ET so ;S1 , S3 =S5 3oo
13 93 km 147 km
182 - 290 km 230 - 72 km
o
TWT
Isl
20
30
North ShoL E D Shot E C South
DISTANCE (km ) So
0 oo ET Sl S3 sS5 3oo
l i)' 'l ' ' j , ,. 1'' ...;_7:':..> ...... ';'-' ...:,..... ....:_..__ .:- ....,.._ ':*':1 '-':':' 'i, t, , ,,1 t, , 0
' [ ....... :':- .-:;"'::" : .... '-'-:"- '- ';:':'::":: ".:'' "-:: Z-..TR.:"- '-' -" , -" ,-: , i ' i ' '1
,o ,,/,','l "' :'"'""'" ' ' ' 'i"'
TWT TWT
Isl > . 4I
,,li 20
Fig. 6. a) Composite section of the EGT refraction/wide-angle reflection data across the Alps and Po Plain
and the NFP-20 reflection data. The offsets from the four EGT refraction shots are shown below the
section. The distance scale above the Profile is measured from shotpoint E. b) Correlation of the main
crustal boundaries. The white circles represent the superposition of the base of the European (M1) and
Adriatic (M2) crust obtained from the deeper portions of the NFP-20 reflection data. (H: Helvetic domain,
P: Penninic domain, LI: European lower crust, L2: Adriatic lower crust).
Aar Massif with the same NFP-20 reflection data in more
detail than can be presented here. The interpretation of the
shallow wide-angle arrivals PbP observed further to the south
(Fig. 4) suggests a reflecting zone or horizon dipping
southward at the northern base of the Aar Massif which was
not found by Stuble and Pfiffner [1991] in the normal
incidence reflection data. This problem has to be studied more
carefully. The high velocity layer (6.5 kin/s) under shotpoints
M1 and M2 can be correlated with an interface within the pile
of Penninic nappes by comparison with the migrated reflection
traverse [Pfiffner, 1990].
Another high-velocity layer with a velocity of 6.6 km/s
derived from an early reflected phase PiP (Fig. 4) is found at
about 20 km depth under the Penninic zone which coincides
with the north dipping broad band of high reflectivity at the
same depth range and position on Figure 7b and the overlay in
Figure 7c. This may be explained as the top and front of the
Adriatic lower crust which is pushed northward as a wedge
between the European upper and lower crust. Crustal wedging
in this area has been proposed previously by Mueller et al.
[ 1980] from the along-strike seismic wide-angle data and by
Kahle et al. [1980] from gravity data. This mechanism of
middle crustal shortening is compatible with mass balance
considerations [e.g. Buffer, 1990] and the concentration of late
orogenic vertical movement along the Insubric Line and in the
Lepontine area [Schmid et al., 1989].
N Molasse Helvetic Aar Penninic Insubric Southern
Basin N appes M ass if N appes Line Alp s Po-Plain
El:) M,2 EC
. I .... !, t !
- -- 4.8 5''"'".9 "'---:-?-'------' -------' --6 i .... i--t. ,T 5.9 --
.......... . ...... .... :' - .... 6.0 .................... --
.8 6.0
O--
i 40
50
60
6.2 ......... 6.5 6.6
EGT 8.2
(a>
70--- , I .... I .... I ' ' ' ' I ' '
o 50 1oo 150
$AEiJAU LB.ES ALP?$ LBiPU LBiLL I,B-_VE $UDALP?7
5 0 -10 0 10 20 80 qO 50 60 70 80 90 100 110 120
.... , .... i .... [ .... .-..i .... ..1:... .... ,....1..1 .... i .... i .... i .... i .... I .... i .... i .... i .... i .... i .... s...i..l ,.H...J.i .... [ .... i ........ i ........ , ..... o
t - - - ":'=- ' ': .':.
, .,;=,-, _ .
60 60
o ..................................................................................................................................................................................
ET S3 S5
S1
Peinic Iubc Souem
Napes Line Alps Po-Pln
N Molasse Helvetic Aar
Basin Nappes Massif
o
20
30
40
50
7O
I I I
L i , i i i i _
o--j - - - - _- _ .. _- _--___i,__: ....... .......... =.-. .... --]-- o
,, .- _........ .. . -- . -' ,....._ . . . .
........ --- _ ' :.-=:'3,:,"-' :- z,'.
'-' : ' '; ' ' .... - I
--] x "' " ,. -'" - '- " .- ..... ". _-_--3--- - _ '%_ [--
,.,,,,_ _ . o ,.o .. - _ ..o. ... . _
,-, 2o---! .. -',- '"-- - "- ...... _ _ -_::,, ;,:=: - - -. 2-- I
- -- ' = =-' ,.. - -. -- ' ..v=- ,. _'-,. ' "----. ----: '_.' ',,. . _
o , .... :.----. ' - -- ----- - ' '': '_ / o
Z ' - - . ' - --.: _ -..A'_z -- =, _.,-.:f:_.--'-'V
c '= ='- -]-":'-- --.--..'" ' - r"-- -' '.- ---,'"' I--
- .... .... -- ,,.,:-_'t,.-----, - _/"- ';' , I
- ., , -- . _--- _.-'..:. ,--'- ....
. --._%.% +: =. ,, =
50 x- . -- 50
. .. '.,
6O 6O
()
7O .... I ' ' 7O
0 50 100 150
DISTANCF. (n)
Beginning at about 10 km south of shotpoint ED the lower
crust is differentiated by a transitional layer of about 10 km
thickness with an average velocity of 6.5 km/s on top of the
south-dipping Moho. This layer is characterized by an
increasing reflectivity with depth which is highest in the 3 to 6
km above the Moho. A similar reflective transition zone of 10
km thickness in the lower crust was also found by Maurer
[1989] and Maurer and Ansorge [1991] along the northern
margin of the Alps (Fig. 3). This lower crustal transition zone
is interrupted by a pronounced gap in the reflectivity which has
no counterpart in the refraction model. Holliger [ 1991] and
Holliger and Kissling [ 1991] suggest that this feature is caused
by scattering in the complex upper crustal geology of the
Penninie domain. The north Alpine reflection character appears
again in the deepest part of the crustal root zone north of the
Insubrie line. This implies that the European lower crust is a
pre-Alpine - possibly Varisean - feature and extends relatively
undisturbed from the northern Molasse Basin to about 120 km
south of shotpoint ED where it is subdueted under the Adrafie
plate.
Shotpoint EC is located almost exactly on the Insubrie line.
Therefore, nothing can be said about the seismic nature of this
fault from the refraction data. Contrary to that this geological
feature at the surface is perfectly imaged by the strong north-
dipping reflections on section S1 after migration and can
excellently be traced to a depth of 17 km (Fig. 7b). Further
north its counterpart the Penninie front is equally well
displayed which is outcropping at the southern margin of the
Aar Massif.
The next interval ranges from the Insubrie line well into the
Po Plain. The low-velocity layer in the upper crust under the
Southern Alps (5.8 km/s) is not an outcome from the EGT
data alone. It was derived by Deichmann et al. [1986] from the
SUDALP77 profile (Fig. 2) and introduced here again. The
differentafion in the upper crust is supported by a band of high
reflectivity in the same depth range under segment S5 of the
southern traverse (Fig. 7b). Perhaps it represents the stack of
southward oriented thrust complex suggested by Roeder
[ 1989]. The transition to the lower crust occurs at about 19 km
under the center of the Po Plain dipping northward to reach 24
km depth under the Southern Alps. The velocity increases at
this boundary to 6.6 km/s.
The Moho dips down gently to the south from 27 km under
the northern end of the Molasse Basin [Ye and Ansorge,
1990b] to 35 km at the northern end of the Helvefie overthrust
Fig. 7. a) Crustal cross section between the Molasse Basin in
the north and the Po Plain in the south. The upper crustal
structure under the Southern Alps was taken from Deiehmann
et al. [1986]. (---) change of velocity gradient, (---)
approximate layer boundaries. EC, ED and M1,2 are EGT and
NFP-20 shotpoints, respectively (Fig. 2). b) Depth-migrated
reflection data from the composite section of the eastern and
southern NFP-20 traverses indicated by wavy lines on the
bottom. Zero distance corresponds to the northernmost CMP.
Clear wide-angle reflections which were correlated on the
profiles crossing the north south traverse, labelled at the top of
the diagram, are represented by stars, doubtful ones by
crosses. These have been migrated as well. e) Overlay of the
velocity-depth model derived from the EGT refraction and
wide-angle reflection data and the depth migrated line
drawings of the NFP-20 eastern and southern traverses.
under shotpoint ED. Starting at about 20 km north of ED the
slope increases to reach the largest clearly determined depth of
56 km at 95 km south of ED, i.e. under shotpoint M2. After a
careful modeling of all Pn and PMP arrivals including the ones
from shotpoints north and south of the Alps a very steep offset
in the Moho has to be assumed at about 10 km north of EC.
The vertical throw of at least 12 km between the downdipping
Moho from the north and the much steeper ulxtip to the south
depends of course on the detailed velocity-depth structure of
the overlying crust. In previous cross sections the Moho was
assumed to be a continuous interface because of the lack of
enough data [Mueller et al., 1980]. Later on a horizontal gap in
the Moho in the area of the greatest crustal thickness was
discussed based on the unmigrated near-vertical eastern NFP-
20 reflection traverse. Holliger [ 1991] has removed this
discrepancy by the proper depth migration of the reflection
data as shown in Figures 7b and 7c [Holliger and Kissling,
1991]. The Moho lies at 33 and 29 km depth under the margin
of the Southern Alps and the center of the Po Plain,
respectively. The northern European Moho is clearly defined
by the sudden temafion of the high reflectivity, refracted and
wide-angle reflected observations. Contrary to that the base of
the Adriatic crust is much less well outlined on the reflection
data, whereas the refraction and wide-angle reflection data are
equally good as further north. The general increase of the
noise level in that area of the southern reflection traverse is
most likely an explanation for this. The EGT refraction profile
crosses the southern Alps in the less populated area east of
Lake Como (Fig. 2).
There is general agreement that the European plate is
subdueted under the Adriatic lithosphere where the Alpine
crust is thickest [e.g. Valasek et al., 1991]. However, it is
unclear how deep the European Moho can be followed as a
guiding interface in the lower lithosphere and how well the
corresponding velocity-depth distribution is controlled by the
presently available data. Buness and Giese [1990] have
derived from later phases observed at distances up to 370 km
on the recordings of shots ED to the south and EA from the
Ligurian coast to the north a velocity range between 6.2 and
7.4 km/s for the subdueted lower European crust. Ye and
Ansorge [1990b] have shown that the same phases can also be
interpreted differently which leaves the question about the
detailed sub-moho velocity structure open.
Pffffner [1990] and Valasek et al. [1991] have combined
geological evidence and models from reflection and refraction
interpretations to draw up preliminary crustal cross sections
which are partly based on the models presented here. It is the
purpose of this paper to present a first consistent cross section
of the Alpine ernst along a north-south traverse derived from
migrated near-vertical, refraction, and wide-angle data. The
reader is cautioned against the assumption that Figure 7e is
already the final model mainly for two reasons. Firsfly, the
projection of the data which reflect the shallow structure is
certainly not correct in all details. The significant in parts east-
west oriented structural variations especially in the Helvefie
units, in the southern Penninie realm and in the Southern Alps
explain some of the discrepancies. Secondly, not all the details
of the along-strike wide-angle profiles have been included in
the basic starting model which sets the boundary conditions
for the interpretation of the north-south oriented wide-angle
EGT profiles by raytracing. Therefore some modifications
may still arise from such a more rigorous design of the input
model. Nevertheless, for the first time a continuous two-
dimensional velocity-depth distribution derived independently
from the EGT data could be combined with the properly
migrated NFP-20 reflection data where the migration process
itself is based on an independent determination of velocities
from older along-strike refraction profiles. These velocities are
average values over larger distances and a two-layered crust
(see stars and crosses in Fig. 7b,c). It will be possible to
assign true layer velocities to the gross features of the
reflection survey in a tectonically highly complicated area of
the Alps.
Acknowledgements. Funding-for the reflection survey was
provided by the Swiss National Science Foundation through
the special support of the National Research Program 20
"Deep Geological Structure of Switzerland". The large scale
seismic refraction experiments are part of the European
Geotraverse Project which is administered by the European
Science Foundation and again supported by the Swiss
National Science Foundation. We would also like to thank all
our colleagues who participated in the field work and data
processing of both projects.
This paper is Contribution No. 672 of the Institute of
Geophysics, ETH Hoenggerberg, CH-8093 Ziirich,
Switzerland.
References
Aichroth, B., and Prodehl, C., EGT Central segment
refraction seismics, In: R. Freeman, St. Mueller (Eds),
Sixth EGT Workshop: Data Compilations and Synoptic
Interpretation, pp. 105-127, European Science Foundation,
Strasbourg, France, 1990.
Aichroth, B., Ye, S., Feddersen, J., Maistrello, M., and
Pedone, R., A compilation of data from the 1986 European
Gotraverse Experiment (Main Line) from Genova to Kiel.
Geophysical Institute, University of Karlsruhe (Federal
Republic of Germany), Open File Report 90-1, Part I, 115
pp., Part II, 89 plates, 1990.
Bernoulli, D., Heitzmann, P., and Zingg, A., Central and
Southern Switzerland: Tectonic evolution and first results of
reflection seismics, In' F., Route, P., Heitzmann, R,.
Polino, (Eds.), Deep structure of the Alps, pp. 289-302,
Mem. Soc. geol. Fr., Paris, 156; Mem. Soc. geol. suisse,
Z'tirich, 1; Vol. spec. Soc. Geol. It., Roma, 1, 1990.
Buness, H., and Giese, P., A crustal section through the
northwestern Adriatic plate, In: R. Freeman, P. Giese, St.
Mueller (Eds.), The European Geotraverse: Integrafive
Studies, pp. 297-304, European Science Foundation,
Strasbourg, France, 1990.
Buffer, R.W.H., Notes on crustal balancing along the Alpine
segment of the European Geotraverse, In: R. Freeman, P.
Giese, St. Mueller (Eds.), The European Geotraverse:
Integrafive Studies, pp. 263-276, European Science
Foundation, Strasbourg, France, 1990.
Deichmann, N., Ansorge, J., and Mueller, St., Crustal
structure of the Southern Alps beneath the intersection with
the European Geotraverse, Tectonophysics, 126, 57-83,
1986.
Egloff, R., Sprengseismische Untersuchungen der Erdlcruste
in der Schweiz, PhD Thesis, ETH Ziirich, 167 pp., 1979.
EUGEMI Working Group, The European Geotraverse seismic
refraction experiment of 1986 from Genova, Italy, to Kiel,
Germany, Ttonophy$ics, 176, 43-57, 1990.
Frei, W., Heitzmann, P., Lehner, P., Mueller, S., Olivier,
R., Pfiffner, O.A., Steck, A., and Valasek, P.,
Geotraverses across the Swiss Alps, Nature, 340, 544-548,
1989.
Galson, D.A., and Mueller, St., An introduction to the
European Geotraverse Project: First results and present
plans, Tectonophysics, 126, 1-30, 1986.
Hill, D.P., Kissling, E., Luetgert, J.H., and Kradolfer,U.,
Constraints on the upper crustal structure of the Long Valley
- Mono Craters volcanic complex, Eastern California, from
seismic refraction measurements, J. Geophys. Res., 90,
11135-11150, 1985.
Holliger, K., A composite, depth-migrated deep seismic
reflection section along the Alpine segment of the EGT
derived from the NFP20 eastern and southern traverses, In:
R. Freeman, P. Giese, St. Mueller.(Eds.), The European
Geotraverse: Integrarive Studies, pp. 245-254, European
Science Foundation, Strasbourg, France, 1990.
Holliger, K., Ray-based image reconstruction in controlled-
source seismology with an application to seismic reflection
and refraction data in the central Swiss Alps, PhD Thesis,
ETH Z'tirich, 156 pp., 1991.
Holliger, K., and Kissling, E., Ray-theoretical depth
migration: Methodology and application to deep seismic
reflection data across the eastern and southern Swiss Alps,
Eclogae Geol. Helvet., 84, in press, 1991.
Kahle, H.-G., Mueller, St., Klingele, E., Egloff, R., Recent
dynamics, crustal structure and gravity in the Alps, In: N.A.
Moerner (Ed.), Earth's Rheology, Isostasy, and Eustasy,
pp. 377-388, John Wiley and Sons, New York, 1980.
Kissling, E., Three-dimensional gravity model of the northern
Ivrea-Verbano zone, In: J.J. Wagner, St. Mueller (Eds.),
Materiaux pour la Geologic de la Suisse, Geophysique, 21,
53-61, 1984.
Kissling , E., Mueller, St., and Werner, D., Gravity
anomalies, seismic structure and geothermal history of the
Central Alps, Ann. Geophys., 1(1), 37-46, 1983.
Maistrello, M., Scarascia, S., Ye, S., and Him, A., -EGT
1986 Central Segment- Compilation of seismic Data
(additional Profiles and Fans) in Northern Apennines, Po
Plain, Western and Southern Alps, Istituto per la Geofisica
della Litosfera C.N.R., Milano (Italy), Open File Report, 24
pp., 56 plates, 1991.
Maurer, H., Die Struktur unter dem Schweizerischen
Alpenvorland, Diploma Thesis, ETH Ziirich, 109 pp.,
1989.
Maurer, H., and Ansorge, J., The structure of the Earth's
crust under the northern margin of the Swiss Alps,
Tectonophysics, in press, 1991.
Mueller, St., Ansorge, J., Egloff, R., and Kissling, E., A
crustal cross section along the Swiss Geotraverse from the
Rhinegraben to the Po Plain, Eclogae geol. Helv., 73 (2),
463-483, 1980.
Ottinger, Th., Der Aufbau der Erdkruste unter dem
Schweizerischen Teil des refraktionsseismischen
Alpenlingsprofils yon 1975, Diploma Thesis, ETH Z'tirich,
166 pp., 1976.
Pfiffner, O.A.,-Crustal shortening of the Alps along the EGT
profile, In: R. Freeman, P. Giese, St. Mueller.(Eds.), The
European Ge, otravers: Integrative Studies, pp. 255-262,
European Science Foundation, Strasbourg, France, 1990.
Pfiffner, O.A., Frei, W., Finckh, P., and Valasek, P., Deep
seismic reflection profiling in the Swiss Alps: Explosion
seismology results for line NFP20-East, Geology. 16, 987-
990, 1988.
Pfiffner, O.A., Frei, W., Valasek, P., S tuble, M., Levato,
L., DuBois, L., Schmid, S.M. and Smithson, S.B., Crustal
shortening in the Alpine orogen: Results from deep seismic
reflection profding in the eastern Swiss Alps, Line NFP 20-
East, Tect0nics, 9, 1327-1355, 1990.
Roeder, D., South-Alpine thrusting and trans-Alpine
convergence, In: M.P. Coward, D. Dietrich, R.G. Park
(Eds.), Alpine Tect0nic$, pp.211-227, Geological Society
Special Publication, 45, 1989.
Schmid, S.M., Aebli, H.R., Heller, F., and Zingg, A., The
role of the Periadriatic Line in the tectonic evolution of the
Alps, In: M.P. Coward, D. Dietrich, R.G. Park (Eds.),
Alpine Tectonics, pp. 153-171, Geol. Soc. London Spec.
Publ., 45, 1989.
Stiiuble, M., and Pfiffrier, O.A., Evaluation of the seismic
response of basement thrust and fold geometry in the
Central Alps based on 2-D raytracing, Annales Tektonicae,
in press, 1991.
Stiiuble, M., Smithson, S.B., and Pfiffner, O.A.,
Interpretation of reflection profries in the Eastern Swiss Alps
based on 3-D raytracing modeling, Tectonics, in press,
1991.
Valasek, P., Integrated processing and presentation of seismic
reflection and refraction data recorded along the Alpine
segment of the European Geotraverse (EGT), In: R.
Freeman, P. Giese, St. Mueller.(Eds.), The European
Geotraverse: Integrative Studies, pp. 237-243, European
Science Foundation, Strasbourg, France, 1990.
Valasek, P., and Holliger, K., Approaches towards an
integrated interpretation of the NFP20 Deep Crustal
Reflection Profiles along the Alpine Segment of the EGT,
In: R. Freeman, St. Mueller (Eds.), Sixth EGT Workshop;
Data Compilations and Synoptic Interpretation, pp. 137-
147, European Science loundafion, Strasbourg, France,
1990.
Valasek, P., Frei, W., Stiuble, M., and Holliger, K.,
Processing of the NFP-20 seismic reflection traverses
across the Swiss Alps by the ETH-Z'tirich processing center,
In: F. Roure, P. Heitzmann, R. Polino (Eds.), Deer _
structure of the Alps, pp. 55-64, Mem. Soc. geol. Fr.,
Paris, 156; Mem. Soc. geol. suisse, Ziirich, 1; Vol. spec.
Soc. Geol. It., Roma, 1, 1990.
Valasek, P., Mueller, St., Frei, W., and Holliger, K., Results
of NFP20 seismic reflection profiling along the Alpine
section of the European Geotraverse (EGT), Geophysical
Journal Int., 105, 85-102, 1991.
Yan, Q.Z., and Mechie, J., A fine structural section through
the crust and lower lithosphere along the axial region of the
Alps, Geothvsical Journal, 98, 465-488, 1989.
Ye, S., and-Ahsorge, J., The compilation of deep seismic
refraction data along the EGT main profile between the
Danube River and the Gulf of Genova, In: R. Freeman, St.
Mueller (Eds.), Sixth EGT Workshop: Data Compilations
and Synoptic Interpretation, pp. 129-136, European Science
Foundation, Strasbourg, France, 1990a.
Ye, S., and Ansorge, J., A crustal section through the Alps
derived from seismic refraction data, In: R. Freeman, P.,
Giese, St., Mueller (Eds.), The Eurovean Geotraverse:
Integrative Studies, pp. 221-236, ]uropean Science
Foundation, Strasbourg, France, 1990b.
SEISMIC DATA FROM THE ALPS AND THEIR BEARING ON THE FORMATION OF GRANULITE AND ECLOGITE NAPPES.
by H. Laubscher
Geological Institute of the University, Bemoullistrasse 32, CH4056 Basel, Switzerland
Abstract. Several models for the uplift to the surface of deep-seated
masses have been proposed in the past, largely on surface evidence.
Geophysical data from the deeper parts of young collision zones had been
scant, but recent seismic surveys through Alpine ranges provided
important information. Although these data emphasize Neogene tectonics
which severely disrupted older structures that contain eclogitic to
granulific bodies, they suggest various ways for the submergence, deep-
seated metamorphism, and re-emergence of originally surficial rocks. In a
section through central Switzerland, three modes of the localized uplift of
deep-seated masses may be surmised: (1) Delamination of lower crust and
its wedging into the middle crust of the opposite plate; (2) partition of
transpression into dip-slip- and strike-slip-dominated structures, with
locally continued underthrusting along the latter; (3) crustal
brachyanficlines ("flower structures") associated with transpression and
uplift of slivers along steep shear zones. The transpressive scenarios (2)
and (3) are probably important for eclogitic rocks from the early
Cretaceous when according to plate tectonics models strike-slip
predominated. Transpression-induced localized depression and uplift of
both oceanic and continental crustal slivers may explain a number of
otherwise puzzling features such as the juxtaposition as exotic terranes of
eclogific slivers of Austroalpine, Piemontese, and European affinities.
Introduction
How do granulitic and eclogitic rocks, formed at depths exceeding 25-30
kin, make their way to the surface? Several scenarios are imaginable and
have been offered for consideration: Backflow of eclogifized material along
the subduction zone, e.g. with the help of enhanced fluid pressure (Ernst,
1970, 1973, 1988); repeated in-sequence thrusting of thick crustal slices
at the leading edge of the upper plate, combined with tectonic denudation
(Platt, 1986; Andersen and Jamtveit, 1990); wedging of the lower crust
of one plate into the middle crust of the opposite plate with continued
underplating (Laubscher, 1970, 1977: "wedge-into-split-apart"; Oxburgh,
1972: "flake tectonics", with similar implications). Surface geology in
the Alps had established that basement nappes as a rule are very thin (Fig.
1), consisting only of the uppermost few km of the Mesozoic continental
crust. This makes it difficult to bring to the surface deep-seated masses as
the basal part of thick crustal slivers, although occasionally this seems to
have happened, as in the case of the Ivrea body; the present position of
this body, however, is the result of a long and complex history (Zingg,
Geodynamics 22
1983). Direct geophysical evidence from recently active collision zones
was scant or lacking when these models were proposed, although for
Laubscher (1970) the "bird's head" shape of the Ivrea body postulated by
geophysicists (e.g. Kaminski and Menzel, 1968) was suggestive.
New data from the deep reflection surveys across Alpine ranges.
The first deep reflection surveys through Alpine ranges active in the
Tertiary (ECORS Pyrenees team, 1988, Bayer et al., 1987) showed
wedges of presumable lower crust and upper manfie somehow pushed into
higher parts of the crust (Fig. 2, 3). They seemed to confirm that the
wedge-into-split-apart or flake tectonics model is a viable one (Laubscher,
1988; Route et al., 1989, 1990; Nicolas et al, 1990; compare Meissner,
1989: "crocodiles"). Subsequent deep reflection lines through the Swiss
Alps (e.g. Frei et al., 1989) added further evidence. They confirmed the
presence of a high-velocity body in the middle crust of the Central Alps
as postulated on refraction data by Mtiller et al. (1980). Its layering (Frei
et al., 1989; Valasek and Holliger, 1990) suggests an origin in the lower
crust, and its position below the projection of the important middle
Miocene Lombardic thrust (Laubscher 1990 a, b) would appear to imply
that it is Adriatic lower crust delaminated from the upper manfie (Fig. 4).
These wedges are still far below the surface and will need some assistance
from underplating and erosion, and particularly tectonic denudation, in
order to reach the surface. Such tectonic denudation has been active
repeatedly in the Alps, e.g. in the Oligocene (Laubscher, 1983; Nievergelt
et al., in press) and in the early Miocene (Merle et al., 1989) when it was
assisting the rise of the Lepontine area as a pull-apart within the dextrally
transpressive Insubric system (Laubscher, in press).
The wedge model corresponds to scenario 2 (middle Miocene) of Fig. 5.
Scenario 1 and 3 are associated with dextral transpression in the late
Oligocene-early Miocene and the late Miocene (-Recent?), respectively.
They are parts of crustal en 6chelon folds, probably flower structures
(Harding and Lowell, 1979), in which underthrusting at presumably steep
shear zones resulted in local uplifts of 10-20 kin. They were insufficient
to raise lower crust material to the surface but demonstrate their
efficiency in bringing up deep-seated rocks. Some of these rocks had
originally been deposited on the surface - e.g. the pre-Mesozoic
stronalites and marbles of the Ivrea zone (Zingg 1983), the coesite bearing
pyrope quartzites of the Dora Maim (Chopin 1984, 1987), the pelitic
granulites of Calabria (Schenk, 1984; Kern, 1990) or the eclogitic pillow
lavas of Zermatt (Bearth, 1959). The transpressive scenario may also
account for this deep burial: Material balance demands that the masses
squeezed up be somehow compensated by masses squeezed down. How
exactly this functions at strike-slip dominated transpression zones such as
the San Andreas fault system and even more so in the deep parts of
207
208 SEISMIC DATA FROM THE ALPS
N EM IL S
0 200 km
Fig. 1 Thin basement nappes (shaded) as observed and projected from the surface, in a cross-section through eastern Switzerland
(compare TrQmpy 1980; for western Switzerland compare Escher et al., 1987). Early refraction work revealed essentially the
generalized Moho 0VI) and indications of the lower crust (dark shading). In between, structure remained enigmatic.
S _.."-" \N P F N
0
- ,, -- /
...................... ==================================================
50
km
Fig.2 Interpretation of the ECORS reflection line across the Pyrenees
after Roure et al. (1989). Light shading: upper (brittle) crust; ruled: lower
crust (dark shading) and upper manfie (light shading), which seem to dip
together into the subduction zone (compare Fig. 4). There, thin slivers of
lower crust (and occasionally upper manfie, see Fig. 3) are detached and
wedged into middle crust or even thrust onto upper crust of the foreland.
NPF: North Pyrenean fault, a steep fault zone that seems to accommodate
large mounts of strike-slip.
intracontinental tmnspression is controversial; however, ,arguably some
sort of inverted flower structure would be more efficient than the familiar
inclined subcluction zones in moving material rapidly to great depths.
This is particularly important for the transfer- (transform-) dominated
early Cretaceous stages of the Alpine orogeny (Dewey et al., 1973)
which, according to some data (e.g. Dal Piaz et al., 1972; Monit, 1985)
produced a number of the more important eclogite bodies.
A scenario for early Cretaceous eclogites in the Alps.
The tectonic situation of pre-Neogene rock masses is not recorded on
the seismic lines because Neogene transpression on the order of 200 to
300 km dissected and obliterated earlier relationships (Laubscher, 1988).
Direct observational data from the Alps appear to be insufficient for
inferring the Cretaceous scenario with any confidence. Viewed as part of
simple subduction models they even seem to be contradictory (Winkler,
1988; Trommsdorff et al., 1990; Faupl and Tollmann, 1979; Th0ni,
1981). However, plate tectonics models that sketch the development of
the western Tethys as part of the opening history of the Atlantic may be
helpful (Dewey et al., 1973; Dercourt et al., 1986; Ziegler, 1988)
although in detail they do not appear to be too well constrained (Dewey et
al., 1989).
Fig 6 depicts the essentials of an early Cretaceous scenario in the Alps
according to Dewey et al. (1973), omitting secondary features such as the
Valais-Northpenninic trough. While within a large-scale sinistral strike-
slip situation the Piemontese ocean basin developed as a pull-apart, the
internal Dinarides were severely compressed as a sort of transverse range
with obduction of an already existing oceanic crust. A North Piemontese
transfer zone (NPT; compare Weisseft and Bernoulli, 1985) may be
postulated to have joined the Piemontese ocean and the internal Dinarides.
The Adria subplate or promontory of Africa occupied the site between the
Dinarides and the Piemontese ocean, with the Austroalpine domain (AA)
in the N and the Southern Alps (SA) in the S of the continental margin.
The crucial question for the problem of Early Cretaceous subduction and
olxluction of both oceanic (ophiolites, e.g. Dal Piaz and Ernst, 1978) and
continental crust (e.g. Dal Piaz et al., 1972) in the Alps concerns the
exact nature of the North Piemontese transfer zone. Particularly where it
joins the internal Dinarides it may have been dragged to the south (Fig. 6)
which would have caused transpression. Considering the tremendous
amount of sinistral strike-slip faulting, even temporary transpression
would have a decisive influence on crustal structure, as suggested by
scenario 1 and 3 in Fig. 5.
A mountain range in some ways similar to the early Cretaceous Alps
are the Pyrenees (partly intracontinental, considerable strike-slip, modest
compression, compare Roure et al. 1989). Their familiar fan-shape
associated with the steep Noah Pyrenean fault is reminiscent of a crustal-
scale flower structure. The recent reflection survey and subsequent
interpretations (e.g. Route et al., 1989) bear out the fact that slivers of
lower crust (and possibly upper mantle) were raised tectonically by many
kilometers (Fig. 2). The interpretation by Route et al. (1989) envisages a
kinematics of left-lateral pull-apart in the middle Cretaceous which,
coupled with tectonic denudation and isostatic uplift, raised portions of
lower crust to comparatively shallow levels before compression set in. In
contrast, the early Tertiary phase is considered purely compressive. In
view of the large dextral strike-slip component of the oblique Africa-
WNW ESE
PF
2wt B :: G P CL
' Po plain
_ _ ....:.....:....,..:.:.: ..... .......: :.:.::.:......._ ................
10 i 30
20 60
km
Fig.3 Interpretation of the ECORS-CROP reflection line through the
western Alps, according to The ECORS-CROP deep seismic sounding
group (1989). Ornamentation as in Fig. 2. B: Belledonne external massif;
PF: Penninic front; GP: Gran Paradiso; CL: Canavese line. Interpretatio
show thin wedges of lower crust and/or upper mantle.
H. LAUBSCHER 209
40
IE.xtertlal S
N Helvetic nappes MassTs Penninic nappes
L Southern Alps
.......
60
Fig.4 Intepretation of refraction (Mueller et al., 1980) and reflection (Frei et al., 1989) data from the central Alps according to
Laubscher (1990 b). The heavy lines are migrated reflections (v=6 km/s) believed representative of entire reflection bands.
Ornamentation as in Fig. 2 except light shading=reflection bands with possible lower crust signature; dark shading = high-v body
fxom refraction data. The northern half of the reflection data is projected from NFP 20 East and should be viewed with some
caution because of pronounced non-cylindricity of the central Alps. The possible candidates for lower crustal wedges (reflection)
cannot be combined easily, and the refraction high-v body apppears largely in the gap between them. IL = Insubric line.
EUROPA
D A D R I
50 km
Fig.5 Three scenarios (numbered boxes), due to different tectonical phases, of localized upldt of deep-seated masses may be
inferred from a combined geological-geophysical section across the central Alps (compare Laubscher 1990 c). Dash-dot lines
indicate the main shear zones (thrusting) of the three Neogene (Neoalpine) phases, from older to younger: (1) lightest shading,
Insubric-Helvetic phase; (2) intermediate shading, Lombardic phase; (3) darkest shading: Windows phase. The upper (brittle) crust
is indicated by light shading. The exact configuration of the southern end of Europa (plate name as distinct from Europe, the
present continent) subduction is controversial.
Europa convergence from the late Cretaceous on (compare Fig. 8) one
may wonder whether this is a realistic view.
For comparison, consider the Alps. As stated above, an early
Cretaceous sinistral pull-apart in the Piemontese ocean was coupled with
transverse compression in the Dinarides and probable sinistral ,
transpression in their Alpine tail. In the area of cross section P (Fig. 6)
this transpression was essentially intracontinental (including the
continental margin) not unlike that in Fig. 2 and Fig. 7. In contrast to
the Pyrenees, however, the Alps were subsequently subjected to dextral
transpression of many hundreds of kilometers (Fig. 8), giving rise to the
pile of thin nappes shown in Fig. 1. The early Cretaceous structures were
disrupted, severely deformed and incorporated in these nappes in a way
difficult to unravel.
Fig. 7, from the link between the Andes of Venezuela and Colombia
after Meier et al. (1987), is a rather suggestive example for the efficiency
of transpressional structures to raise deep-seated masses as narrow slivers.
/
/
1
/
i
I
I
, \
', \
,
', \
,'
Fig.6 Early Cretaceous (148-80 Ma) plate tectonics scenario, based on Dewey et al. (1973). Horizontal ruling: Piemontese
oceanic pull-apart. Shading: Dinaric subduction zone. NPT = North-Penninic transfer zone. AA = Austroalpine domain, SA =
Southalpine domain. P = transpressive cross-section (compare Fig. 4). PZ = Pyrenean zone.
Fig.7 The La Mulera dextmlly transpressive structure (Tfichira, W
Venzuelan Andes) with the isolated Cerro Rangel basement sliver (Meier
et al., 1988), an example of localized uplift along a transpression zone.
In this structure, a narrow outcrop of basement appears in the midst of
Cretaceous to Miocene sediments, surrounded by a complex system of
thrusts and strike-slip faults. Extrapolation of the surface observations is
evidently difficult, but a system of thrusts and strike-slip faults after the
fashion of a "flower structure" (Harding and Lowell, 1979; compare
particularly their seismic section Fig.6) appears unavoidable. The strike-
slip component of this transpressive feature may be estimated at about 10
km. For strike-slip on the order of 1000 km, the depressing and uplifting
of limited slivers of both oceanic and continental crust on the order of 100
km (Chopin, 1984, 1987; Schreyer, 1988) does not seem to be too
fantastic a figure. The juxtaposition of high-pressure crustal slivers of
Austroalpine (Sesia), Piemontese (ophiolites) and European (Monte Rosa
to Dora Maira) affinities, otherwise rather puzzling, is consistent with
exotic terranes in a strike-slip dominated model.
Fig. 8 depicts a late Cretaceous-Paleocene scenario (omitting again
branches such as the Valais-Northpenninic zone) based on the plate
tectonics of Dewey et al. (1973), when opening of the North Atlantic
resulted in dextral transpression in the Alps with large-scale nappe
development in the Austroalpine and Piemontese domains. Simple shear
at the base of these nappes resulted in the observed dismembering and
scattering of ophiolites, including their high-pressure bodies. Continued
NW motion of the Adriatic promontory resulted in their arcuate
arrangement as suggested in Fig. 8.
Conclusions.
The geophysical data from the Alps, and particularly the recent
seismic reflection surveys, suggest various ways for the submergence,
deep-seated metamorphism, and re-emergence of originally surficial rocks.
The data emphasize Neogene tectonics; they are ambiguous and need
interpretation in 3d kinematic schemes. In a section through central
Switzerland, three modes of the localized uplift of deep-seated masses are
suggested: (1) delamination of lower crust and its wedging into the middle
crust of the opposite plate; (2) partition of transpression into dip-slip- and
strike-slip-dominated structures, with locally continued underthrusting
along the latter; (3) crustal brachyanticlines ("flower structures")
associated with tmnspression. The transpressive modes (2) and (3) are
probably important for eclogitic rocks from the early Cretaceous when
according to plate tectonics models strike-slip predominated.
Transpression-induced localized depression and uplift of both oceanic and
H. LAUBSCHER 211
Fig.8 Late Cretaceous-Paleocene plate tectonics scenario, based on Dewey et al. (1973) according to Laubscher ("The Arc of the
Western Alps today"; paper read at the annual convention of the Swiss Geological Society 1990 and unpublished manuscript,
1991). The apex A of the Adriatic indenter moves to A', obliquely to boundaries inherited from the early Cretaceous phase (Fig.
3). The shallow portions of the Austroalpine and Piemontese domains are sheared off and thrust as a pile of nappes onto the
foreland. The Austroalpine transfer zone (AAT) is simultaneously active as a dextral strike-slip fault, causing a clockwise rotation
of the nappes in front of it. The Piemontese ocean is nearly eliminated at the end of this time interval. The first thrust zone
(shading) is followed in sequence by other thrusts (dashes).
continental crustal slivers may explain a number of otherwise puzzling
features such as the juxtaposition as exotic terranes of eclogitic slivers of
Austroalpine, Piemontese, and European affinities. The geophysical data
do not, unfortunately, shed any light on this problem as Neogene
motions have severely disrupted and even obliterated older structures.
However, a comparison with the Pyrenees where the deep structure was
observed by the recent reflection survey has added a wealth of data, and
with transpressive structures studied by surface geologists or known from
industrial seismic sections, helps in devising preliminary models.
References
Andersen, T. B., and Jamtveit, B., Uplift of deep crust during orogenic
extensional collapse: a model based on field studies in the Sogn-
Sunnfjord region of western Norway. Tectonics 9, 1097-1111, 1990.
Bayer, R., Cazes, M., Dal Piaz, G.V., Damotte, B., Elter, G., Gosso,
G., Hirn, A., Lanza, R., Lombardo, B., Mugnier, J.-L., Nicolas, A.,
Nicolich, R., Polino, R., Roure, F., Sacchi, R., Scarascia, S.,
Tabacco, I., Tapponnier, P., Tardy, M., Taylor, M., Thouvenot, F.,
Torreilles, G. and Villien, A., Premiers r6sultats de la travers6e des
Alpes occidentales par sismique r6flexion verticale (Programme
ECORS-CORP). CJ. Acad. Sci. Paris, t. 305, SErie H, 1461-1470,
1987.
Bearth, P., Ober Eklogite, Glaukophanschiefer und metamorphe
Pillowlaven. Schweizerische Mineralogische und Petrographische
Mitteilungen, 39, 267-286, 1959.
Chopin, C., Coesite and pure pyrope in high-grade blueschists of the
western Alps: A first record and some consequences. Contributions to
Mineralogy and Petrology, 86, 107-118,1984.
Chopin, Ch., Very-high-pressure metamorphism in the western Alps:
New petrologic and field data. Terra Cognita, 7, 94, 1987.
Dal Piaz, G. V., and Ernst, W. G., Areal geology and petrology of
eclogites and associated metabasites of Piemonte ophiolite nappe, Breuil
St. Jacques area, Italian western Alps. Teconophysics, 51, 99-126,
1978.
Dal Piaz, G. V., Hunziker, J.C., and Martinotti, La zona Sesia-Lanzo e 1'
evoluzione tettonico-metamorfica delle Alpi nordoccidentali interne.
Memorie della Societd Geologica Italiana,11, 433 - 466, 1972.
Dercourt, J., and 18 coauthors, Geologic evolution of the Tethys belt
from the Atlantic to the Tamirs since the Lias. Tectonophysics 123,
241-315, 1986.
Dewey, J. F., Helman, M. L., Turco, E., Hutton, D. H. W., and Knott,
S. D., Kinematics of the western Mediterranean. Geological Society
Special Publications No. 45, pp. 265-283, 1989.
Dewey, J.F., Pitman, W. C., Ryan, W. B. F., Bonin, J., Plate
tectonics and the evolution of the Alpine system. Geol. Soc. America
Bull., 84, 3137 - 3184, 1973.
ECORS Pyrenees Team, The Ecors deep reflexion seismic survey across
the Pyrenees. Nature, 331,508-510, 1988.
Ernst, W. G., Interpretative synthesis of metamorphism in the Alps.
Geological Society of America Bulletin, 84, 2053-2078, 1973.
- Tectonic contact between the Franciscan m61ange and the Great Valley
Sequence- crustal expression of a late Mesozoic Benioff Zone. Journal
of Geophysical Research, 75, 886-901, 1970.
- Tectonic history of subduction zones inferred from retrograde blueschist
P-T paths. Geology, 16, 1081-1084, 1988.
Escher, A., Masson, H., and S teck, A., Coupes g6ologiques des Alpes
occidentales suisses. Rapp. geol. Serv. hydr. geol. Suisse, 2, 1987.
Faupl, P., and Tollmann, A., Die Rossfeldschichten: Ein Beispiel f'tir
Sedimentation im Bereich einer tektonisch aktiven Tiefseerinne aus der
kalkalpinen Unterkreide. Geologische Rundschau, 68, 93-120, 1979.
Frei, W., Heitzmann, P., Lehner, P., Muller, St., Olivier, R., Pfiffner,
A., Steck, A., Valasek, P., Geotraverses across the Swiss Alps. Nature,
340, 544- 548, 1989.
Harding, T.P., and Lowell, J.D., Structure styles, their plate-tectonics
habitats, and hydrocarbon traps in petroleum provinces. The American
Association of Petroleuzm Geologists Bulletin, 63, 1016-1058, 1979.
Kaminski, W., and Menzel, H., Zur Deutung der Schwereanomalie des
Ivrea-KOrpers. Schweizerische Mineralogische und Petrographische
Mitteilungen, 48, 255-260, 1968.
Kern, H., Laboratory seismic measurements: an aid in the interpretation
of seismic field dam. Terra Nova, 2, 617-628, 1990.
Laubscher, H.P., Bewegung und Warme in der alpinen Orogenese.
Schweiz. mineral., petrogr. Mitt., 50, 565-596, 1970.
- The Tectonics of Subduction in the Alpine System. Memorie della
Societd Geologica Iraliana, 13, 275-283, 1977.
- The late Alpine (Periadriatic) intrusions and the Insubric Line. Memorie
della Societd Geologica Iraliana, 26, 21-30, 1983.
- Material balance in Alpine orogeny. Geol. Soc. of America Bulletin,
100, 1313-1328, 1988.
- The problem of the deep structure of the Southern Alps: 3-D material
balance considerations and regional consequences. Tectonophysics, 176,
103-121, 1990a.
- Seismic data and the deep structure of the central Alps. In R. Freeman
and St. Mueller, (ed.), Proceedings of the Sixth Workshop on the
European Geotraverse (EGT) Project: Data compilations and synoptic
interpretation. European Science Foundation, Strasbourg. pp. 149-156,
1990b.
- Deep seismic data from the central Alps: mass distributions and their
kinematics. Mdm. Soc. gdol. suisse, 1,335-343, 1990c.
- The Alps- a transpressive pile of peels. In K. McClay, (ed), Thrust
tectonics. Unwin Hyman, London, in press.
Meier, B., Schwander, M., and Laubscher, H., The tectonics of Tachira -
a key to North-Andean tectonics. In J.-P. Schaer and J. Rodgers (eds.),
The anatomy of mountain ranges, Princeton Univ. Press, Princeton, N.
J., pp. 229 - 237, 1987.
Meissner, R., Rupture, creep, lamellae and crocodiles: happenings in the
continental crust. Terra Nova, 1, 17-28, 1989.
Merle, O., Cobbold, P. R., and Schmid, S., Tertiary kinematics in the
Lepontine dome. Geological Society Special Publications, 45, 113-134,
1989.
Moni6, P., La m6thode 39Ar-40Ar appliqu6e au m6tamorphisme alpin
dans le massif du Mont-Rose (Alpes Occidentales). Chronologie
d6tail16e depuis 110 Ma. Eclogae Geologicae Helvetiae, 78, 487-516,
1985.
Mueller, St., Ansorge, J., Egloff, R., and Kissling, E., A crustal cross
section'along the Swiss Geotraverse from the Rhinegraben to the Po
plain. Eclogae Geologicae Helvetiae, 73,463-483, 1980.
Nicolas, A., Him, A., Nicolich, R., Polino, R., and ECORS-CROP
working group, Lithospheric wedging in the Western Alps inferred from
the ECORS-CROP traverse. Geology, 18, 587-590, 1990.
Nievergelt, P., Liniger, M., Froitzheim, N., and Ferreiro-
Maehlmann,R.,The Turba mylonite zone: an Oligocene extensional
fault at the Pennine-Austroalpine boundary in eastern Switzerland. Terra
Nova, in press.
Oxburgh, R., Flake tectonics and continental collision. Nature, 239, 202-
204, 1972.
Platt, J.P., Dynamics of orogenic wedges and the uplift of high-pressure
metamorphic rocks. Geological Society of America Bulletin, 97, 1037-
1053, 1986.
Roure, F., Choucroune, P., Berastegui, X., Mufioz, J. A., Viilien, A.,
Matheron, P., Bareyt, M., Seguret, M., Camara, P., and Deramond, J.,
ECORS deep seismic data and balanced cross sections: Geometric
constraints on the evolution of the Pyrenees. Tectonics, 8, 41 - 50,
1989.
Roure, F., Polino, R., and Nicolich, R., Poinqonnement, r6trocharriages
et chevauchements post-basculement dans les Alpes occidentales:
6volution intracontinentale d'une chane de collision. C.R. Acad. Sci.
Paris, t. 309, Sdrie II, 283 - 290, 1989.
Schenk, V., Petrology of felsic granulites, metapelites, metabasics,
ultramafics, and metacorbonates from southern Calabria (Italy): prograde
metamorphism, uplift and cooling of a former lower crust. Journal of
Petrology, 25, 255 - 298.
Schreyer, W., Subduction of continental crust to mantle depths:
petrological evidence., Episodes, 11, 97 - 104, 1988.
Th6ni, M., Degree and evolution of the Alpine metamorphism in the
Austro-Alpine unit W of the Hohe Tauern in the light of K/Ar and
Rb/Sr age determinations on micas. Jahrbuch der Geologischen
Bundesanstalt, 124, 111-174, 1981.
Trommsdorff, V., Dietrich, V., Flisch, M., Stille, P., and Ulmer, P.,
Mid-Cretaceous, primitive alkaline magmatism in the Northern
Calcareous Alps: Significance for Austroalpine geodynamics.
Geologische Rundschau, 79, 85 - 98, 1990.
Trfimpy, R., An outline of the geology of Switzerland. Geology of
Switzerland- a guide-book, Schweiz. Geol. Kommission, Wepf & Co.
Publ., Basel/New York. pp. 5 - 104, 1980.
Valasek, P., and Holliger, K., Approaches towards an integrated
interpretation of the NFP20 deep crustal reflection profiles along the
Alpine segment of the EGT. In R. Freeman and St. Mueller, (ed.),
Proceedings of the Sixth Workshop on the European Geotraverse
(EGT) Project: Data compilations and synoptic interpretation. European
Science Foundation, Strasbourg, pp. 137-148, 1990.
Van den Driessche, J., and Brun, J.P., Shear partitioning in oblique
collisions. Terra Cognita 7, 120, 1987.
Weissen, H., and Bernoulli, D., A transform margin in the Mesozoic
Tethys: evidence from the Swiss Alps. Geologische Rundschau, 74,
665-679, 1985.
Winkler, W., Mid- to early Late Cretaceous flysch and m61ange
formations in the western part of the eastern Alps. Palaeotectonic
implications. Jahrbuch der Geologischen Bundesanstalt, 131, 341-389,
1988.
Ziegler, P. A., Evolution of the Arctic-North Atlantic and the western
Tethys. The American Association of Petroleum Geologists, Memoir
43, 1988.
Zingg, A., The Ivrea and Strona-Ceneri Zones (Southern Alps, Ticino and
N-Italy) - A Review. Schweiz. mineral. petrogr. Mitt. 63, 361, 1983.
A DEEP SEISMIC REFLECTION PROFILE ACROSS THE ARCHEAN-PROTEROZOIC WITWATERSRAND
BASIN, SOUTH AFRICA
R. J. Durrheim 1, L. O. Nicolaysen 2 and B. Corner 1
Abstract. The 112 km, 16 s two-way time
trans-Witwatersrand basin seismic reflection
profile stretches from the Ventersdorp dome,
across the Potchefstroom syncline and terminates
in the center of te Vredefort dome. The
Potchefstroom syncline rests on Archean basement
and contains Archean (Dominion Group and
Witwatersrand Supergroup) and early Proterozoic
(Ventersdorp Supergroup and Transvaal Sequence)
strata. The disposition of the strata is clearly
imaged on the seismic section - in particular a
series of half-grabens on the basin margin, the
unconformity between the Witwatersrand and
Ventersdorp Supergroups, and the overturned strata
forming the collar of the Vredefort dome.
The crystalline basement is divided into three
domains on the basis of the seismic fabric: the
domain stretching from the Ventersdorp dome to
Potchefstroom characterized by several distinct
zones of strong, sub-horizontal reflections; the
relatively transparent domain between
Potchefstroom and the collar of the Vredefort
dome; and the domain of occasional reflections,
within the core of the Vredefort dome. The
reflectivity observed within the crystalline
basement is probably due to sills intruded during
Precambrian extension, or ductile banding. The
western portion of the profile shows a change in
reflective character at approx. 12 s TWT, which
probably marks the crust/mantle transition. The
absence of a distinct "reflection Moho" suggests a
smooth transition from crust to mantle over a
distance of a few kilometers.
The fundamental basis for the three domains
which display contrasted seismic fabrics in the
1Department of Geophysics, University of the
Witwatersrand, Private Bag 3, WITS 2050, South Africa.
2Bernard Price Institute of Geophysical Research,
University of the Witwatersrand, Private Bag 3, WITS
2050, South Africa
Oeodynamics 22
basement is not understood. Systematic mapping of
such basement domains by reflection profiling may
provide insights regarding processes responsible
for localising stratified basins.
Introduction
One of the principal objectives of deep seismic
profiling is to study reflections at near-normal
incidence from crystalline rocks in the crust and
upper mantle, using the techniques developed by
the oil industry for mapping structures in
sedimentary basins. During the past decade deep
seismic reflection profiling has produced many new
insights into continental structure and evolution
[e.g. Barazangi and Brown, 1986a and 1986b;
Matthews and Smith, 1987; Leven et al., 1990].
However, in general the crustal processes
operating during the Archean remain poorly
illuminated by deep seismic profiling. To date,
relatively few deep seismic reflection profiles
have been acquired in Archean provinces, mostly in
North America. The Archean provinces that have
been surveyed include the Kapuskasing structural
zone [Green et al., 1990], Abitibi greenstone belt
[Jackson et al., 1990] and the COCORP Minnesota
profile [Gibbs et al., 1984] in the Superior
Province; the COCORP Wind River [Smithson et al.,
1980] and Montana plains [Latham et al., 1988]
profiles in the Wyoming Province; the Yilgarn
Block of Australia [Mathur et al., 1977; Moss and
Dooley, 1988]; and the Limpopo Belt of southern
Africa [De Beer et al., 1988].
The profile described in this paper images a
section through the Kaapvaal craton of South
Africa. This continental shield stabilised during
the Archean, and formed the' basement for
substantial deposition of "cover" strata during
the Archean and lower Proterozoic. The Archean
crust remains preserved in a relatively undeformed
state. The 112 km, 16 s two-way-time seismic
reflection profile stretches from an arc of
basement granite, across th Potchefstroom
syncline of the Witwatersrand basin in which
virtually the full succession of Archean and early
Proterozoic strata are preserved, and terminates
within a dome of basement granite (the Vredefort
dome) situated in the centre of the basin (Figure
213
214 WITWATERSRAND BASIN SEISMIC REFLECTION STUDY
1). Reflection seismics has been used for the past
decade to explore for the gold- and
uranium-bearing strata contained within the
Witwatersrand basin [Campbell and Peace, 1984;
Pretorius, Jamison and Irons, 1989] . The existence
of a zone of strong reflectivity at approximately
2.3 s two-way time within the Archean basement was
recognized in an earlier study, and interpreted to
correspond with the transition from upper to
middle crust [Durrheim, 1986] .
One of the primary objectives of the trans-
Witwatersrand bas in deep seismic reflection .
profile was to throw light on the mechanism of
formation of a thick basin in the late Archean and
early Proterozoic. It was hoped that the profile
would illustrate the regional framework within
which the basins are set, and whether a sub-
basinal geologic feature existed which led to the
lengthy period of episodic sedimentation and the
preservation of the sediments. The fundamental
question to be answered is whether subsidence was
controlled by a deep-seated array of faults
(penetrating through the crust or soling in a
decoupling horizon), or if quasi-plastic necking
took place in the deep crust. Another goal was to
learn about the constitution of the middle and
deep crust. Finally, it was anticipated that the
profile would provide information concerning the
present day disposition of the supracrustal strata
preserved in the Potchefstroom' syncline, es-
pecially the geometry of the overturned Vredefort
collar rocks.
26 27
28o-l-
26
27
28
Johannesb, urg""
I L_,/ ,--',,: L
W. itwatersrand/.*. ---"' 7
\ Witwatersrand/
29
26
. .
27
29
km
I00 km
[] TRANSVAAL SEQUENCE
[] VENTERSOORP SUPERGROUP
2 8o I--] CENTRAL RAND GRCiUP
1'721 WEST RAND GROUP
I DOMINION GROUP
[ BASEMENT GNEISS
Fig. 1. Location of the seismic reflection profile and the geology of the
Witwatersrand basin with the Paleozoic cover removed [after Borchers, 1964 and
Jones, 1988]. Approximated localities of goldfields and gaps between them are: EV -
Evander; ER - East Rand; WR - West Rand; WW - Far West Rand; P - Potchefstroom gap;
KE- Klerksdorp east; KW - Klerksdorp west; KS - Klerksdorp south; B - Bothaville;
N - Free State north; FC - Free State central; and FS - Free State south. Granite
domes are: H - Hartebeestfontein; VE - Ventersdorp; and VR - Vredefort. The seismic
profile is the line VE-P-VR marked with arrowheads. The Machavie line is a major
lineament recognised by gravity and magnetic studies [Corner and Wilsher, 1989].
DURRHEIM, NICOLAYSEN AND CORNER 215
Geology
Tectonic history and stratigraphy of the central
Kaapvaal craton.
The Kaapvaal craton is one of the ancient
nuclei around which the African continent was
built. The crystalline rocks constituting the bulk
of the craton are generally referred to as
basement granites". In fact this crystalline
foundation encompasses a range of compositions
from granites to tonalites, and fabrics which vary
from those of igneous origin to gneissic or
migmatitic. Within these crystalline rocks are
arcuate and canoe-shaped inliers, the
metamorphosed and structurally deformed greenstone
belts, constituted of Archean sedimentary and
volcanic strata. The entire suite of Archean rocks
which underpins the craton is also known
informally as the Basement Complex, of which the
greenstone belts comprise less than 10% of the
exposure. The Basement Complex crops out in a 200
km broad strip along the northeastern margin of
the craton, but over most of the craton it is
concealed by younger cover rocks except for a few
isolated inliers such as the Johannesburg,
Vredefort and Ventersdorp domes (Figure 1). A
profound unconformity separates the Basement
Complex from the relatively undisturbed cover
rocks.
Since 3.0 Ga the Kaapvaal craton has been
relatively stable, with tectonic movements
providing the appropriate environments for forming
a succession of Archean and Proterozoic basins
exhibiting fluvial, deltaic and continental-shelf
types of deposition. The last period of pervasive
regional metamorphism and deformation experienced
by the craton took place more than 2.6 Ga ago
[Walraven et al., 1990].
The succession of Archean and early Proterozoic
rocks document a lengthy and complex tectonic
history involving several periods of extension and
contraction. The rocks of the Dominion Group and
Witwatersrand Supergroup are believed to have been
deposited in the Witwatersrand basin between 3.1
and 2.7 Ga [Walraven et al., 1990]. The Dominion
Group is a predominantly volcanic succession lying
unconformably upon the Basement.. Complex, and is
interpreted to represent the protobasinal phase of
the greater Witwatersrand basin. Although largely
concealed beneath younger strata, the Dominion
Group is known to reach a thickness of 2700 m in
the vicinity of the Klerksdorp goldfield [Tankard
et al., 1982]. The Witwatersrand Supergroup is
divided into the (lower) West Rand Group and the
(upper) Central Rand Group. The predominantly
argillaceous sediments of the West Rand Group
conformably overlie the Dominion Group, or overlap
the Basement Complex. The West Rand Group averages
4650 m in thickness, and exceeds 7000 m in the
northwest [Tankard et al., 1982]. The Central Rand
Group consists predominantly of coarse-grained
subgreywacke, now quartzite, with less than 10%
conglomerate. The maximum thickness of the Central
Rand Group is almost 3000 m in the centre of the
basin. Deposits of gold and uranium occur in the
conglomerates, which represent alluvial fans
formed on the perimeter of the basin. The
Witwatersrand Supergroup is covered unconformably
by the predominantly volcanic Ventersdorp
Supergroup (ca. 2.7 Ga); and the Transvaal
Sequence (2.5-2.3 Ga) consisting of dolomitic
rocks, shales, lavas and quartzites. Finally,
Phanerozoic Karoo Sequence sediments, generally
less than 1 km in thickness, cover much of the
Witwatersrand basin.
Vredefort structure
The Vredefort dome, situated in the centre of
the Witwatersrand basin, is one of the Earth's
largest cryptoexplosion structures. The core of
basement granite is about 40 km in diameter, and
is surrounded by a polygonal collar of steeply
dipping and overturned supracrustal strata, which
were deposited on an eroded basement surface. A
suite of alkalic eruptives, surrounded by an
aureole of coarsely recrystallised metamorphic
hornfelses, forms a prominent Yeature of the
collar strata in the northwestern sector. The
basement granite and collar rocks display
well-developed shock deformation features such as
the presence of coesite and stishovite,
emplacement of pseudotachylite veins, lamellar
deformations in quartz, and the widespread
occurrence of shatter cones [Nicolaysen, 1990].
Strong geochemical gradients within the granitic
core have been interpreted as evidence that the
Archean crust has been turned non edge" during the
2 Ga updoming event [Hart et al., 1981]. A
traverse from the collar to the center of the
structure is thought to represent an upturned
section through the Archean crust. The granitic
core is divided into two roughly concentric zones.
The Vredefort upper crust consists of massive
granites or granite-gneisses, knownas the Outer
granite-gneiss (OGG). In the center of the
structure (10-15 km below the original cratonic
surface) rocks of granitic composition, in granu-
lite facies, occur; these rocks are termedthe
Inlandsee leucogranulite (ILG). The OGG and ILG
are separated by a narrow (ca. 10 m) zone of
pseudotachylite and brecciated dolerite termed the
Vredefort discontinuity, with which a 2-3 kmbroad
magnetite-rich region of charnockitization is
associated. In the centre of the Vredefort dome a
serpentinised ultramafic body has been intersected
by a drillhole, and is interpreted to be a sample
of the mantle [Hart et al., 1990].
Reflection seismic experiment
Acauis ition
_
Field trials were carried out during April
1988 in the Kroonstad district over a profile 6 km
in length. The acquisition parameters are listed
in Table 1. The processing sequence followed that
conventionally applied to Witwatersrand basin
data. We were given access to the record section
from 2- 24 s TWT. Many strong reflections are
observed from 2-6 s. These are attributed to the
supracrustal strata. From 6-10 s there are
numerous short reflecting segments. A band of
reflecting segments is recognised between
11.8-12.3 s, and interpreted to be the reflection
Moho. From 12.3-24 s there are virtually no
coherent reflections. It was found that the low
frequency end (8-12 Hz) of the vibrator sweep
caused considerable damage to the gravel roads on
which the vibrators operated, which were water-
logged at the time due to exceptionally high
rainfall. Bandpass filtered sections showed that
the data was not noticeably degraded by elimina-
ting the frequencies below 15 Hz.
The acquisition parameters for the
Witwatersrand basin profile were determined after
consideration of the field trial, and a cautious
attempt to "trade-off" between the-length of the
line and the acquisition parame.t.ers which affect
record quality and the depth of investigation
(i.e. number of repeat sweeps, s.weep length,
listening time, source station interval). The
geophone group interval was limited to 50 m, as
this was the maximum cable length that could be
supplied by the only contractor operating in
southern Africa at that time. The dominant
groundroll wavelength on outcropping Precambrian
strata in the region of the Witwatersrand basin is
in excess of 200 m. Consequently vibrator and
geophone arrays cannot be effectively used to
reduce the groundroll [Weder, 1990 ] . The
TABLE 1. Acquisition parameters for the deep
seismic reflection field trials conducted in the
Kroonstad district.
SOURCE
acquisition parameters selected for the profile
are listed in Table 2.
Data acquisition proceeded without. incident. It
took 18 field days for the contractor to complete
the profile. The survey commenced at the
northwestern end of the profile, near the town of
Ventersdorp. Some degradation of data quality was
experienced when traversing the outcropping
dolomitic rocks of the Transvaal Sequence, which
have an irregular bedrock surface, and are some-
times cavernous. This results in strong near-
surface scattering of the energy. Undershooting
was used to pass through the towns of Potchefstroom
and Vredefort, resulting in a reduction of CDP
coverage. The crooked roads passing through the
mountainous collar of the Vrede{ort dome also
resulted in a reduction of CDP coverage.
A brute stack of the first 30 km of the profile
was examined while the survey was in progress. The
data appeared to be of good quality, with many
strong, continuous reflectors arising from the
crystalline basement. However, it was decided not
to reduce the fold of the stack in order to
increase the length of the profile, but to adhere
to the original acquisition parameters. This
decision was subsequently vindicated, as the final
TABLE 2. Acquisition parameters for the
trans-Witwatersrand basin deep seismic
reflection profile.
SOURCE
Type
Sweep length
Sweep range
Repeat sweeps
Spacing
5 in-line Failing Y2700 vibrators
10 m pad separation
28 s
12-61.5 Hz
6, 4 m move-up between sweeps,
pattern centered between
stations
100 m
Type 5 in-line Failing Y2700 vibrators
Sweep length 30 s
Sweep range 8-60 Hz
Repeat sweeps 8
Spacing 100 m
SPREAD
Geophone
groups 120
Spacing 25 m
Geometry end on
Spread length 3 km
CDP coverage 0-30-0 fold
RECORDING
Sampling 4 ms
Listening time 54 s
Record length 24 s
SPREAD
Geophones SM4/UB, 10 Hz
Geophone groups 120
Geophone pattern 24 geophones/array
Spacing 50 m
CDP interval 25 m
Geometry split straddle
3225-(50)-275-0-275-(50)-3225
Spread length 6.45 km
CDP coverage 0-30-0 fold
RECORDING
Sampling 4 ms
Listening time 44 s
Record length 16 s
Low-cut filter out
High-cut filter 62.5 Hz, 72 db/oct
Notch out
Tape format SEG-D, 9 track
section shows that the northwestern segment
contains the greatest number of deep crustal
reflectors. A relaxation in the acquisition
parameters would have made it difficult to assess
whether changes in the seismic character are due to
variations in the basement geology, or a result of
a change in the survey specifications.
Data processing
Reflection seismic data have been gathered as a
routine in the Witwatersrand basin for gold and
uranium exploration since 1982. The geological
environment is very different to that commonly
encountered in hydrocarbon prospects, and several
factors have been identified which require a
modification of the processing parameters [March,
(a) The indurated Precambrian strata have high
seismic velocities (generally > 6 km/s) and
densities. Low reflection coefficients (of the
order of 5%) are encountered. Fortunately
attenuation is also very low, and so reflections
are reliably obtained.
(b) Due to the very high formation velocities,
the moveout on the CDP gather is very small when
compared to that typically encountered in oil-field
prospecting, restricting the sensitivity of
conventional velocity analysis. For the trans-
Witwatersrand basin profile we were limited to a
spread of about 6 km (120 channels, geophone group
interval of 50 m). A split-spread configuration was
selected as we anticipated steep and variable dips
would be encountered as the profile traversed the
Potchefstroom syncline. As the maximum offset is
only about 3225 m, the determination of interval
velocity is probably only reliable in the first 1
S.
(c) Very little multiple activity appears to
be present in the data, in spite of clear primary
reflections. This is probably due to the low
reflection coefficients, and the fact that there is
usually no significant decrease in reflection
coefficient or general increase in seismic velocity
with depth.
(d) The migration of data with high seismic
velocity and no regular increase in velocity with
depth is one of the most difficult parts of the
processing. In practice it is usual to regard the
velocity of the whole section as virtually
homogeneous. The structure of the Witwatersrand
basin is also known to be complex, and out-of-plane
events may appear on the section. Although gross
structural information can be obtained from the
migrated section, it is necessary to treat the data
with caution until the interpretation can be
verified by loop closure or a full
three-dimensional survey.
(e) Reflected refractions give rise to
arrivals on the CDP gather which are difficult to
distinguish from the primary reflections, and at
present no automated process exist to remove these
arrivals from a record section. Near-vertical
structures such as faults and dykes are known to be
common in the Witwatersrand basin, and consequently
reflected refraction events are likely to be
present in the seismic section.
Data processing was executed taking the above
considerations into account. The sequence is des-
cribed in Table 3. The unmigrated stacked section
was released on Open File by the South African
Geological Survey in January 1989. Finite-
difference migration was carried out at the
University of the Witwatersrand. Only the upper 8 s
of the section was migrated, as the difficulty of
using wave equation methods to migrate deep
continental data is well known [Warner, 1987]. It
was found that a migration velocity of 6000 m/s was
satisfactory, causing many diffractions to collapse
to point foci.
Interpretation
An interpretation of the seismic reflection
profile is shown in Figure 2. This figure can be
used to establish the position of the reproduced
seismic panels (Figures 3 to 5) on the profile.
Supracrustal strata
The geologically mapped contacts between
formations were projected downdip to identify the
TABLE 3. Data processing sequence applied to the
trans- Witwatersrand basin deep seismic reflection
profile.
1.Demultiplex
2.Edit
3.Statics - field statics applied to floating
datum
4.F-K filter - full on at 3000 m/s
5.Scale - whole trace equalisation followed by
spherical divergence correction
6.Sort
7.Deconvolution - minimum phase predictive
deconvolution
8.Scale - automatic gain control using 100 ms
windows
9.NMO - revised velocities derived
10.Statics - data corrected from floating datum
to regional datum of 1300 m
11.Autostatics - surface consistent residual
statics
12.NMO - revised velocities applied
13.Mute
14. Statics - data corrected to regional datum
15.Autostatics
16.Autostatics
17.Stack
18.Scale - whole trace equalisation
19.Filter - time variant bandpass filter
6-12-62-74 Hz 0-8 s
6-12-46-58 Hz 5-16 s
20.Coherency
NW
0 lo0
12
16
SE
5 10 15 20 25 km Potchefstroom Vredefort
I
600 1100 1600 2100 2600 3100 3600 41,00 CDP
i i i ' i I i I
w- v , / \
V C W _ X
- - ,
__' / h,,,,.'
X
SEISMIC FABRIC ZONES
Fig. 2. Interpretation of the trans-Witwatersrand
basin deep seismic reflection profile. Geology: X
- Basement Complex, W - Dominion and West Rand
Groups, C - Central Rand Group, V - Ventersdo.p
Supergroup, M - Black Reef Formation and Malmani
Subgroup, P - Pretoria Group, R - Roodekraal
Complex.
reflections marking these interfaces, and to
determine the characteristic "seismic fabric" of
each unit. Where possible, the interpretation has
been constrained by drillhole information supplied
by mining companies. The following seismic fabrics
and reflectors are recognised, working upwards
through the stratigraphic succession (note that
formations may have been eliminated by erosion or
tectonism in some areas). The basement granite is
generally seismically transparent, although in some
areas continuous zones of strong reflections occur.
The alternating shales, lavas and quartzites of the
overlying Dominion and West Rand Group rocks give
rise to strong continuous reflections. The quart-
zites of the Central Rand Group and the basaltic
lavas of the Ventersdorp Supergroup are essentially
transparent, although the contact between these
units is marked by a strong reflection. The
dolomitic rocks at the base of the Transvaal
Sequence give rise to low amplitude reflections
with poor continuity, but the contact with the
underlying Ventersdorp lavas is marked by a strong
reflection. The upper formations of the Transvaal
Sequence are shales, characterised by strong
continuous reflections, divided by a seismically
transparent andesitic lava formation.
The northwestern segment of the profile shows
Dominion and West Rand Group strata preserved in
several half-grabens (Figure 3). The strata dip to
the southeast at ca. 30 . The faults forming the
southeastern margins are interpreted to be reverse
faults of Central Rand age, reactivated as normal
faults during the Ventersdorp rift episode. The
faults cannot be traced deep into the basement, as
similar rock types are juxtaposed. The faults do
not appear to displace the reflective bands in the
crystalline basement. Transvaal strata are pre-
served in a minor syncline in the vicinity of CDP
1600, and in the major Potchefstroom syncline. In
the center of the syncline (CDP 2400) the contact
between the dolomitic rocks at the base of the
Transvaal Sequence and the basaltic lavas of the
Ventersdorp is at 1 s TWT. Although a virtually
complete succession crops out along the collar of
the Vredefort dome, the steep dips and intense
faulting result in sparse reflections. In one
possible interpretation, overturned Dominion and
West Rand Group rocks are represented by a set of
reflections which dip to the southeast in the upper
1 s TWT, and then to the northwest from 1-2 s TWT
(Figure 4).
Basement complex
The crystalline basement may be divided into
three zones on the basis of the seismic fabric. The
first zone extends from the outcropping Basement
Complex rocks (the Ventersdorp 'dome) on the
northwestern end of the profile as far as
Potchefstroom. It is characterized by numerous
subhorizontal reflections, both isolated and in
complex packages. The upper 4 s is shown in Figure
5. Virtually all the events show some curvature on
the unmigrated section - those above 1.5 s TWT tend
to collapse to a focus after migration. Zones of
strong reflections which are continuous for more
than 20 km occur from 2.4- 2.8 s TWT and 3.4-4.4 s
TWT. These reflective bands appear to parallel the
base of the supracrustal strata. The origin of
these reflections is discussed in the following
section. From 6-12 s TWT there are numerous sub-
horizontal reflections, whose horizontal length
6OO
!;,
4
11oo
CDP
1600 2100
I
Fig. 3. Migrated panel from the
trans-Witwatersrand basin deep seismic reflection
profile. X - reflectors within the Basement
Complex, W - base of Dominion or West Rand Group,
V - base of Ventersdorp Supergroup, M - base of
Black Reef Formation and Malmani Subgroup, P -
base of Pretoria Group. Four adjacent traces are
summed giving an equivalent trace spacing of 100
m.
W X,
interpretation .
Fig 4. Panel (CDP 3540-4000) from the trans-Witwatersrand basin deep seismic
reflection profile showing the collar of the Vredefort dome.
generally do not exceed a Fresnel zone (for a
frequency of 20 Hz and a velocity of 6 k m/s, the
first Fresnel zone radius at a depth of 20 km is 3
km). These reflections probably arise from undu-
lating interfaces or an ensemble of reflective
lenses. The gradual decrease of reflectivity at 12
s TWT is thought to mark the crust-mantle tran-
sition. The absence of a "reflection Moho" suggests
a smooth transition from crustal to mantle.
The second zone extends from Potchefstroom to
the collar of the Vredefort structure. It is
characterised by the complete absence of reflected
energy between 6 s TWT and the base of the section
at 16 s TWT. The boundary between this zone and the
160 200 CDP 240 280
Fig 5. Migrated panel from the trans-Witwatersrand basin
deep seismic reflection profile. This panel shows the
reflective character of the Ventersdorp dome.
first zone is vertical and abrupt. A major
east-west striking lineament (the Machavie line,
see Figure 1), which has been recognised from
gravity and magnetic studies, intersects the
seismic profile at this point [Corner and Wilsher,
1989]. The possibility that the absence of deep
reflections is due to either near-surface
scattering or entrapment of energy within the
supracrustal strata was carefully considered. The
absence of outcropping dolomitic rocks and the
sharp definition and continuity of the reflectors
in the upper 2 s of the section discount the loss
of energy due to near-surface scattering. Although
a large number of reflectors occur in the upper 3 s
of the section, none are abnormally strong. When
the generally low values obtained for the
reflection coefficients in the Witwatersrand basin
are considered, the entrapment of most of the
energy within the supracrustal strata is considered
to be unlikely. Consequently we are of the opinion
that the absence of reflections is genuine, due
either to an absence of layers with a contrast in
acoustic impedance, or extreme complexity of
structure. No vestige of a reflection Moho is
detected.
The third zone lies within the core of the
Vredefort dome. It is characterised by occasional
inclined (curving) events, generally not more than
2 km in length. No crust/mantle transition is
recognised.
Discussion
Origin of reflectors within the crystalline crust
Numerous conditions capable of producing seismic
reflections can exist within the crystalline crust.
Examination of outcrops and drill core have
suggested several causes for the reflectivity
observed in the upper crystalline crust: aniso-
tropic fabrics, trapped fluids, major lithological
variations (gneissic banding on a very large
scale), and fault zones manifesting compositional
change (which may be primary or due to metasomatic
activity). Hypotheses regarding the cause of
reflectivity in the lower crust may be divided into
several classes: primary igneous layering; suites
of sub-parallel mafic sills; reflective ductile
strain banding, amplifying the original litho-
logical contrasts; and contrasts in the physical
properties of the banded rocks, including the fluid
content [Klemperer et al., 1987]. When interpreting
the reflectivity observed in the Archean crust, it
must be remembered that the features causing
reflections represent the current state of the
crust. The effects of the most recent tectonic
event will usually be dominant, and earlier
structures may be reworked.
The results of numerous seismic profiles of
the continental crust over a wide range of tectonic
and age provinces in Europe and north America
indicate that the upper crystalline crust is
largely devoid of reflections, even thoughwell-
logs display velocity perturbations of the proper
wavelength and magnitude to produce large amplitude
reflections [Mooney and Brocher, 1987]. This
paradox may be explained by the suggestion that the
velocity perturbations in the upper crust are
probably neither well laminated nor of long scale
length because this portion of the crust behaves
nonductilely - extension and compression is
achieved brittlely by cataclasis and folding,
producing short features with steep dips which are
not well imaged by conventional seismic profiling
methods [Mooney and Brocher, 1987].
The observation of strong, continuous
reflectors in the upper crystalline crust of the
Ventersdorp dome is thus unexpected, especially
when the lengthy tectonic history of the area is
considered. The Ventersdorp dome represents a major
zone of basement uplift, but is very poorly
exposed. The occasional fresh outcrops in the
center of the dome reveal a homogeneous, massive,
medium- to coarse-grained granodiorite. Very
coarse-grained pegmatites occur throughout [Robb
and Meyer, 1985]. The strongly reflective bands at
2.4-2.8 s and 3.4-4.4 s are a prominent feature of
the northwestern section of the profile. It has
been shown that similar seismic fabrics are
produced by zones consisting of many reflective
lenses packed together [Reston, 1987], and by
undulating reflective surfaces [Blundell and
Raynaud, 1986]. There are a wide ange of geologic
structures which could produce these reflector
geometries. These include: igneous intrusions or
diapiric upwelling; strongly folded or corrugated
surfaces of significant impedance contrast with the
radius of curvature less than 10 k m; and extensive
faulting juxtaposing rocks of differing impedance.
A striking feature of the reflective zones
within the crystalline basement is the parallelism
with the base of the supracrustal strata (Figure
3). A zone of strong reflectivity some 2.3 s into
the basement of the Witwatersrand basin was first
described by Durrheim [1986]. It was suggested that
this zone marked the transition from the upper
granite-gneiss crust to the middle crust composed
of granulites of dioritic bulk composition, but
probably bimodal in actual composition (i.e. a
mixture of felsic and mafic granulite). This
reflective zone has also been correlated with a
semi-continuous magnetite-rich zone which crops out
in the core of the Vredefort structure and at the
craton boundary [Corner et al., 1990]. A reflective
zone parallel to the basement contact is also found
in the eastern part of the Witwatersrand basin. A
30 km profile described by Weder [1990] shows the
zone lying about 1 s below the basement contact.
The evidence now suggests that several sub-
horizontal reflective zones, with lateral extent of
the order of 10's of kilometers, exist within the
basement to the Witwatersrand basin. The thickness
and depth of the zones vary, and in some areas
several reflective zones are stacked.
The large lateral extent of the reflective
zones suggests that they signify reconstitution of
the crust during basin formation, either through
metamorphism or through ductile deformation
accompanied by the intrusion of sills. The
parallelism of the reflectors with the base of the
Dominion Group strata suggest that the eruption of
the Dominion lavas is the event most likely to be
responsible for the reflectors. This event (3.06
Ga) produced pervasive metamorphism of the basement
rocks [Walraven et al., 1990]. Other less plausible
candidate events are the Ventersdorp Supergroup
volcanism (2.7 Ga) and the intrusion of the
Bushveld Complex (2.06 Ga). In the model of
Clendenin et al. [1988] for the development of the
Ventersdorp Supergroup, the Ventersdorp dome is
interpreted to lie within the zone of maximum
crustal extension. Massive eruptions of basaltic
lava accompanied this rifting. Numerous sills of
Bushveld Complex age intrude the supracrustal rocks
in the vicinity of the Vredefort dome.
Similar reflections within granitic basement
have been recorded elsewhere. Strongly laminated
reflection fabrics have been recorded in the upper
crystalline crust at several localities in North
America: from 2-4 s in Precambrian granites and
gneisses near the Picacho Mountain core complex in
south-central Arizona (the reflections were
originally interpreted as overthrust sediments, and
were drilled to a depth of 5492 m); and from 4-10 s
in the Snake River metamorphic core complex in
eastern Nevada. Both areas lie in the Basin and
Range Province, and experienced large amounts of
extension in the Tertiary. The laminated reflection
fabric is attributed to ductile stretching and
contemperaneous magmatic intrusion during this
episode [McCarthy, 1986]. High-amplitude horizontal
reflectors in granitic crust in the vicinity of the
Siljan structure in Sweden have been proven by
drilling to be due to dolerite sills [Juhlin,
1990].
The seismically transparent domain beneath the
Potchefstroom syncline is probably due to
structural and petrological complexity. The
Johannesburg dome, some 100 km northeast of the
Potchefstroom syncline, has been mapped in detail.
It is comprised of two main varieties of gneiss
(mesocratic biotite tonalitic gneiss and leuco-
cratic biotite tonalire and trondhjemite gneiss)
which display polyphase folding, as well as green-
stone remnants [Tankard et al. , 1982] . The
formation of the gneisses has been attributed to a
complex interplay of processes involving stoping
and assimilation of supracrustal greenstones by
granitic fluids followed by potash metasomatism.
The deformation in the gneisses is attributed to
partial mobiliztion which resulted in "flowage
folding" [Anhaeusser, 1973] . Basement of this type
would probably be essentially seismically
transparent.
The core of the Vredefort structure has been
interpreted as representing a section of the crust
turned "on edge". Within the Vredefort middle
crust, pods of mafic granulites are found within
the leucogranulite [Hart et al. , 1981] . The
reflection seismic signature, characterized by
occasional inclined or curved events, is in accord
with that likely to be produced by steeply dipping
rocks containing occasional lenses. of reflective
material.
Crust/mantle transition
No "reflection Moho" is observed on the trans-
Witwatersrand basin seismic profile, although a
decrease in reflectivity on the northwestern part
of the profile at ca. 12 s is thought to indicate
the base of the crust. Models of the Moho boundary
consisting of a thick zone with a rather smooth
transition from crust to mantle velocities, or more
generally models that lack a stack of thin layers
with strong velocity contrasts between them, are
unlikely to reflect large amounts of energy at
normal incidence and high frequencies. However, at
wide-angle and low frequency he reflection
coefficient is similar to that for models with
first order discontinuities or velocity laminations
at the base of the crust [Braile and Chiang, 1986].
Seismic refraction experiments on the central
Kaapvaal craton have shown that the crust is about
35 km in thickness [Willmore et al., 1952, Gane et
al., 1956, Hales and Sacks, 1959]. The records
obtained during a recent seismic refraction
experiment were interpreted to indicate a velocity
gradient zone 1-3 km thick at the base of the
crust. This zone gives rise to high-amplitude,
low-frequency (< 5 Hz) dispersed PmP and SmS
arrivals at offsets greater than 150 km. Occasional
SmP or PmS reflections with a dominant frequency of
2-3 Hz are also observed [Durrheim, 1989]. As the
vibrator sweep commenced at 12 Hz, the seismic
wavelengths are too short to be reflected by the
gradient zone. Similar results were obtained on the
ECORS "north of France" section over the Pre-
cambrian Brabant massif. Reflection profiling
showed only a transparent lower crust, and
virtually no indication of a crust-mantle boundary;
while a refraction/wide-angle reflection survey
using an explosive source produced clear low-
frequency arrivals (< 10 Hz)/ from the Moho
reflector [Hirn et al., 1987]. This was interpreted
as indicating a smooth crust-mantle velocity
transition.
Conclusion
The seismic reflection profile across the
Witwatersrand basin shows the structure of basins
formed during the late Archean and early
Proterozoic on the Kaapvaal craton of southern
Africa. The region has had a lengthy tectonic
history. Consequently features that would be
diagnostic of the formation mechanism of the early
basins have largely been overprinted. A striking
feature of the profile are the reflective zones
within the crystalline basement. These zones dip
beneath the basin, approximately parallel to the
base of the supracrustal strata. The most probable
origin of these reflectors are intrusive sills and
ductile strain banding associated with Dominion
Group extension and volcanism at 3.06 Ga. The
absence of a distinct "reflection Moho" suggests a
smooth transition from crust to mantle.
Acknowledgements. This project was funded by
the National Geophysics Programme, Geological
Survey of S A, University of the Witwatersrand,
Johannesburg Consolidated Investments Co. Ltd.,
and Gold Fields of South Africa Ltd. The field
trials were sponsored by Gencor Ltd. Rockplan
Ltd. were the contractors responsible for the
acquisition and initial processing. Annette Maher
interpreted the seismic image of the collar of the
Vredefort Dome.
References
Anhaeusser, C. R., The geology and geochemistry of
the Archean granites and gneisses of the
Johannesburg-Pretoria dome, Geol. Soc. S. Afr.
Spec. Publ. 3, 361-385, 1973.
Barazangi, M. and L. Brown (Eds.), Reflection
Seismology' a global perspective, Geodynamics
series, 13, AGU, Washington, 1986a.
Barazangi, M. and L. Brown (Eds), Reflection
Seismology'the continental crust Geodynamics
series, 1--4, AGU, Washington, 1986b.
Blundell, D. J. and B. Raynaud, Modeling lower
crust reflections observed on BIRPS profiles,
in Reflection Seismology' a global perspective,
edited by M. Barazangi and L. Brown, pp.
287-296, Geodynamics Series, 13, AGU,
Washington, 1986.
Borchers, R., Exploration of the Witwatersrand
System and its extensions, in The geology of
some ore deposits in southern Africa vol 1,
edited by S. H. Haughton, pp. 1-23, Geol. Soc.
S. Afr., Johannesburg, 1964.
Braile, L. W. and C. S. Chiang, The continental
Mohorovicic discontinuity: results from
near-vertical and wide-angle seismic reflection
studies, in Reflection Seismology: a global
perspective, edited by M. Barazangi and L.
Brown, pp. 257-272, Geodynamics Series, 13,
AGU, Washington, 1986.
Campbell, G. and D. G. Peace, Seismic reflection
experiments for gold exploration, Wits. basin,
Republic of South Africa, paper presented at
the 46th Meeting of the European Association of
Exploration Geophysicists, London, 19-22 June,
1984.
Clendenin, C.W., E.G. Charlesworth, S. Maske and
A. A. de Gasparis, Normal simple shear model
for the evolution of the early Proterozoic
Ventersdorp Supergroup, South Africa, Inform.
Circ. Econ. Geol. Res. Unit 201,.20 pp., Univ.
Witwatersrand, Johannesburg, 1988.
Corner, B., R. J. Durrheim and L. O. Nicolaysen,
Relationships between the Vredefort structure
and the Witwatersrand basin as interpreted from
regional gravity and aeromagnetic data,
Tectonophysics, 171, 49-62, 1990.
Corner, B. and W. A. Wilsher, Structure of the
Witwatersrand basin derived from interpretation
of aeromagnetic and gravity data, in
Proceedings of Exploration '87: Third Decennial
International Conference on Geophysical and
Geochemical Exploration for Minerals and
Groundwater, edited by G. D. Garland, pp.
532-546, Ontario Geological Survey, Special
Volume 3, 1989.
De Beer, J. H., K. Geerthsen, E. H. Stettler, S.
Aldrich, A. du Plessis and P. Smit, Geophysics
and the deep structure of the Limpopo belt and
northern part of the Kaapvaal Craton, Abstracts
Geocongress '88, pp. 145-148, Geological
Society of South Africa, 1988.
Durrheim, R. J., Recent reflection seismic
developments' in the Witwatersrand basin, in
Reflection Seismology: a global perspective,
edited by M. Barazangi and L. Brown, pp. 77-83,
Geodynamics Series, 13, AGU, Washington, 1986.
Durrheim, R. J., A seismic investigation of the
Kaapvaal craton, Ph. D. thesis, 171 pp.,
Witwatersrand University, 1989.
Gane, P. G., A. R. Arkins, J.P. F. Sellschop and
P. Seligman, Crustal structure in the
Transvaal, Bull. Seism. Soc. Am. 46, 293-316,
1956.
Gibbs, A.K., B. Payne, T. Setzer, L.D. Brown, J.D.
Oliver and S.. Kaufman, A Seismic reflection
study of the Precambrian crust of central
Minnesota, Bull. Geol. Soc. Am., 95, 280-294,
1984.
Green, A., B. Milkereit, J. Percival, A. Davidson,
R. Parrish, F. Cook, W. Gels, W. Cannon, D.
Hutchinson, G. West and R. Clowes, Origin of
deep crustal reflections: seismic profiling
across high-grade metamorphic terranes in
Canada, Tectonophysics, 173, 627-638, 1990.
Hales, A. L. and I. S. Sacks, Evidence for an
intermediate layer from crustal structure
studies in the eastern Transvaal, Geophysical
Journal , 15-33, 1959.
Hart, R. J., M. A. G. Andreoil, C. B. Smith, M. L.
Otter and R. J. Durrheim, Ultramafic rocks in
the centre of the Vredefort structure (South
Africa): possible exposure of the upper
mantle?, Chemical Geology, 83, 233-248, 1990.
Hart, R. J., L. O. Nicolaysen and N.H. Gale,
Radioelement concentrations in the deep profile
through the Precambrian basement of the
Vredefort structure, J. Geophys. Res. 86,
10639-10652, 1981.
Hirn, A., B. Damotte, G. Torreilles and the ECORS
Scientific Party, Crustal reflection seismics:
the contributions of oblique, low frequency
and shear wave illuminations, Geophys. J. R.
astr. Soc., 89, 287-296, 1987.
Jackson, S. L., R. H. Sutcliff, J. N. Ludden, C.
Hubert, A. G. Green, B. Milkereit, L. Mayrand,
G. F. West and P. Verpaelst, Southern Abitibi
greenstone belt: Archean crustal structure from
seismic-reflection profiles, Geology, 18,
1086-1090, 1990.
Jones, M. Q. W., Heat flow in the Witwatersrand
basin and environs and its significance for the
South African shield geotherm and lithosphere
thickness, J. Geophys. Res. 93, 3243-3260,
1988.
Juhlin, C., Interpretation of the reflections in
the Siljan Ring area based on the results from
the Gravberg-1 borehole, Tectonophysics, 173,
345-360, 1990.
Klemperer, S. L. and the BIRPS group, Reflectivity
of the crystalline crust: hypotheses and tests,
Geophys. J. R. astr. Soc., 89, 217-222, 1987.
Latham, T. S., J. Best, T. Chaimov, J. Oliver, L.
Brown and S. Kaufman, COCORP profiles from the
Montana plains: the Archean cratonic crust
and a lower crustal anomaly beneath the
Williston basin, Geology, 16, 1073-1076, 1988.
Leven, J. H., D. M. Finlayson, C. Wright, J. C.
Dooley and B. L. N. Kennett (Eds.), Seismic
probing of the continents and their margins,
Tectonophysics, 173, 1990.
March, D. W., Seismic acquisition and processing
in old geological formations, S A Geophys. Ass.
Yearbook, 39-41, 1988.
Mathur, S. P., F. J. Moss and J. C. Branson,
Seismic and gravity investigations along the
geotraverse, Western Australia, 1969, Bur.
Miner. Resour. Aust. Bull., 191, 1977.
Matthews, D. and C. Smith (Eds.), Deep seismic
reflection profiling of the continental
lithosphere, Geophys. J. R. astr. Soc., 89,
1987.
McCarthy, J., Reflection profiles from the Snake
Range metamorphic core complex: a window into
the mid-crust, in Reflection Seismology: the
continental crust, edited by M. Barazangi and
L. Brown, pp. 281-292, Geodynamics Series, 14,
AGU, Washington, 1986.
Mooney, W. D. and T. M. Brocher, Coincident
seismic reflection/refraction studies of the
continental lithosphere: a global review, Rev.
of Geophys., 25, 723-742, 1987.
Moss, F. J. and J. C. Dooley, Deep crustal lower crustal structure revealed by COCORP
reflection recordings in Australia 1957-1973 - crustal reflection profiling in the Wind River
I. Data acquisition and presentation, Range, Wyoming, Earth Planet. Sci. Lett., 46,
Geophysical Journal, 93, 229-237, 1988. 295-305, 1980.
Nicolaysen, L. O., The Vredefort structure: an Tankard, A. J., M.P. A. Jackson, K. A. Eriksson,
introduction and a guide to recent literature, D. K. Hobday, D. R. Hunter and W. E. L. Minter,
Tectonophysics, 171, 1-6, 1990. Crustal Evolution of southern Africa. 3.8
Pretorius, C. C., A. A. Jamison and C. Irons, Billion years of Earth history, 523 pp.,
Seismic exploration in the Witwatersrand basin, Springer-Verlag, New York, 1982.
Republic of South Africa, in Proceedings of Walraven, F., R. A. Armstrong and F. J. Kruger, A
Exploration '87: Third Decennial International chronostratigraphic framework for the
Conference on Geophysical and Geochemical north-central Kaapvaal craton, the Bushyeld
Exploration for Minerals and Groundwater, Complex and the Vredefort structure,
edited by G. D. Garland, pp. 241-253, Ontario Tectonophysics, 171, 23-48, 1990.
Geological Survey, Special Volume 3, 1989. Warner, M., Migration - why doesn't it work for
Reston, T. J., Spatial interference, reflection deep continental data ? Geophys. J. R. astr.
character and the structure of the lower crust Soc., 89, 21-26, 1987.
under extension. Results from 2-D seismic Weder, E. E. W., The significance of gravity and
modelling, Annales Geophysicae, 5B,339-348,1987. seismic reflection techniques in deriving a new
Robb, L. J. and M. Meyer, The nature on the structural model for the areas south of the
Witwatersrand hinterland: conjectures of the central Rand gold fields, Ph.D. thesis, 173
source-area problem, Inform. Circ. Econ. Geol. pp., Pretoria University, 1990.
Res. Unit, 178, 25 pp., Univ. Witwatersrand, Willmore, P. L., A. L. Hales and P. G. Gane, A
Johannesburg, 1985. seismic investigation of crustal structure in
Smithson, S. B., J.A. Brewer, S. Kaufman, J.E. the western Transvaal, Bull. Seism. Soc. Am.
Oliver and R. L. Zawislak, Complex Archean 42, 53-80, 1952.
DEEP-PENETRATION SEISMIC REFLECTION IMAGES ACROSS THE SOUTHEASTERN
UNITED STATES CONTINENTAL MARGIN
Jinyong Oh 1 , Joseph D. Phillips, James A. Austin, Jr., and Paul L. Stoffa 1
The University of Texas Institute for Geophysics, 8701 Mopac Blvd., Austin, Texas 78759
Abstract. Approximately 1,200 km of deep-penetration
reflection (16 s, 240 channels) seismic data acquired over the
Southeastern Georgia Embayment along six profiles using a
large (10,800 in3), tuned airgun source array and a 6,000 m
streamer receiver provide detailed images of basement
reflections associated with Mesozoic rifting and deeper pre-
Mesozoic crustal structures. The new, high-quality seismic
profiles that result from such seismic data processing
techniques as f-k filtering and inside muting to suppress
multiples and dip moveout to enhance images indicate: 1)
Wedge structures consisting of several seaward-dipping,
subaerially-emplaced Jurassic basalt flows characterize the
basement hinge zone. 2) The inferred Mesozoic basins
which have been defined from previous magnetic and seismic
studies below the Brunswick magnetic anomaly (BMA) are
not real structures. They result from diffraction artifacts. 3)
A crustal-scale ramp structure interpreted as the suture zone,
bounded on its south side by south-dipping intracrustal
reflections, is observed at varying distances north of the
BMA, suggesting that the BMA is not directly related to the
suture.
Introduction
The United States Atlantic passive continental margin
originated from the separation of Africa and North America
in the Early Mesozoic [Pitman and Talwani, 1972]. It has
been considered a good example of a non-volcanic margin
covered with thick (8-13 kin)postrift or drift sediments
[Bally, 1981; Klitgord et al., 1988]. This thick sedimentary
overburden usually prevents the penetration of seismic
energy into basement structures. Consequently, the imaging
of the margin's deep crustal structure, which is critical to
understanding its evolution, has proved difficult.
lAlso at Department of Geological Sciences, The
University of Texas at Austin.
Geodynamics 22
Many geological and geophysical studies (including
commercial seismic exploration and scientific deep-ocean
drilling) have been done over the southeastern U.S., both
onshore and offshore [Chowns and Williams, 1983; Sheridan
et al., 1988]. Magnetic data have been used to infer
locations of boundaries between major crustal types and to
estimate the depth to basement underlying thick sedimentary
basins [e.g., Klitgord and Behrendt, 1979]. Gravity data
have also been used to model the depth to the Mohorovicic
discontinuity (Moho) across these basins [e.g., Hutchinson
et al., 1983]. Geopotential anomalies have been an
important source of information concerning this passive
margin's deep crustal structure and evolution, even though
their solutions are non-unique.
Most of our current knowledge of the postrift stratigraphic
history of the continental margin along the U.S. east coast
is based on commercial contractor-acquired, multichannel
seismic reflection (MCS) profiling obtained between 1973
and 1979 by the United States Geological Survey (USGS)
[Folger et al., 1979]. Most of the USGS MCS profiling used
relatively small energy sources (generally, 1,000-2,000 in3
at 2,000 psi) with 2,400- to 3,600-m-long hydrophone
receivers [Sheridan et al., 1988]. Except for a few lines
[e.g., Behrendt et al., 1983], no coherent reflections from
deep crustal depths were observed. Therefore, sub-basement
events have not been regionally correlated.
In contrast, the Consortium for Continental Reflection
Profiling (COCORP) acquired onshore seismic lines across
Georgia that imaged a large complex of southward-dipping
reflections extending downward through the entire crust
[Nelson et al., 1985a, b; 1987]. These were interpreted as a
Late Paleozoic "Alleghanian" suture between Africa and
North America.
More recently, an ocean bottom seismometer (OBS)
refraction program conducted across the Carolina trough
using explosives [Tr6hu et al., 1989] suggested that a "7.2
km/s" deep crustal layer, similar to that found beneath the
Baltimore Canyon trough off New lersey by the LASE
experiment [LASE Study Group, 1986], extends southward to
the Carolina trough and the adjacent oceanic basin. Both
layers have been interpreted as anomalously thick plutonic
materials that result from crustal underplating during late-
225
226 DEEP-PENETRATION SEISMIC REFLECTION IMAGES,
stage continental rifting and early sea-floor spreading [White
and McKenzie, 1989]. These interpretations suggest that the
origin of the U.S. Atlantic margin may involve more
extensive magmatic activity than previously thought.
In June 1988, The University of Texas Institute for
Geophysics collected approximately 1,200 km of deep-
penetration MCS reflection data along six profiles off the
southeastern U.S. margin (Figure 1). These data, together
with previous MCS records and geopotential field data, have
provided important new information about the early
evolutionary phase of the margin during Mesozoic
continental separation [e.g., Austin et al., 1990]. Based on
these new seismic images, we are able to develop more
detailed deep crustal cross-sections, thereby allowing us to
constrain the mode of continental crust deformation during
extension.
Tectonic Setting
The superposition of compressional fabrics caused by
collision of the North American and African plates in the
Late Paleozoie and extensional structures resulting from
rifting in the Early Mesozoic is the major tectonic
characteristic of the eastern U.S. continental margin
[Klitgord et al., 1983]. Along its southern part, in the
Southeast Georgia Embayment, the offshore portion of the
margin is underlain by two basins, the Carolina trough and
Blake Plateau basin, separated by the Blake Spur Fracture
Zone (BSFZ) (Figure 1). This area exhibits major geological
and geophysical features which undoubtedly reflect the
compressional and subsequent extensional history seen
onshore. These features include prominent magnetic
anomaly patterns, extensive distribution of Mesozoic
volcanic rocks, and a large contrast between the two basins
in their respective widths, thickness of sediment, salt
diapirism, and postrift depositional style [Klitgord and
Behrendt, 1979; Hutchinson et al., 1983; Dillon et al.,
1983; Dillon and Popenoe, 1988].
In this area, the Brunswick (BMA) and the East Coast
(ECMA) magnetic anomalies are interpreted as important
markers of major crustal boundaries [Klitgord et al., 1988]
(Figure 1). The ECMA, a large positive magnetic anomaly
trend extending from Canada to South Carolina, has been
interpreted as the continent-ocean boundary [Klitgord and
Behrendt, 1979], although details of the origin of this
anomaly remain controversial [e.g., Emery et al., 1970;
Hutchinson et al., 1983; Alsop and Talwani, 1984; McBride
and Nelson, 1988; Austin et al., 1990; Hall, 1990]. The
BMA, a negative anomaly trend located about 100 km
landward of and nearly parallel with the ECMA south of Cape
Hatteras, curves onshore off Georgia [Klitgord et al., 1988].
Offshore, the BMA occurs along the landward edge of the
marginal sedimentary basins previously described [Klitgord
and Behrendt, 1979], whereas onshore it has been interpreted
to mark a Late Paleozoie suture zone boundary between
African and North American crust [Daniels et al., 1983;
Nelson et al., 1985a, 1987]. Consequently, contrasting
magnetic models have been proposed to account for the
BMA: a dipping, highly magnetized single slab with
changing azimuth, emplaced in the upper plate of the Late
Paleozoic suture [McBride and Nelson, 1988], a Mesozoic
rift basin filled with non-magnetic material [Hutchinson et
al., 1983] and a Mesozoic volcanic wedge extruded during
continental break-up [Austin et al., 1990].
Among the upper crustal structures imaged by pre-existing
MCS lines, the major geological feature representing vertical
changes is the basement hinge zone. It is interpreted to
mark the seaward edge of undeformed or slightly extended
continental crust [Klitgord et al., 1988]. Across this zone,
acoustic basement rapidly deepens. Farther seaward, the
basement seismic reflection image is often lost as a result of
thick overlying sediments [Hutchinson et al., 1983; Tr6hu et
al., 1989]. Rift-stage crust is found between the ECMA and
the hinge zone. This implies that the dominant seismic
expression of the crust landward of the hinge zone should
relate to pre-Mesozoic tectonics, while Mesozoic extension
should dominate seaward. However, formation of the hinge
zone has not been understood in detail, although it appears
common to all the marginal basins off eastern North
America [Klitgord et al., 1988] and to many other passive
margins around the world [e.g., Austin and Uchupi, 1982].
In summary, most of the conflicting interpretations (e.g.,
the origin of the BMA) and the unanswered questions (e.g.,
crustal characteristics below the basement hinge zone)
regarding the evolution of passive margins require
knowledge of the basement and deeper crustal structure.
Data Acquisition
Based on the examination of existing MCS data, six
survey lines were selected in the Southeast Georgia
Embayment to image basement structures and deeper crustal
features associated with the major magnetic anomalies,
especially the BMA (Figure 1). Since all profiles crossed
the BMA at a high angle, except BA-2, we refer to them as
Brunswick Anomaly (BA) lines. The BA lines were acquired
using the GECO My, a commercial vessel equipped with a
10,800 in 3 (at 2,000 psi) tuned airgun source array fired
every 50 m and a 6,000-m-long streamer with 240 channels
spaced every 25 m. Figure 2a shows this tuned airgun array
configuration. Data were recorded for 16 s at a 4 ms
sampling interval. The average autocorrelation function
from a typical record and its frequency spectrum are shown in
Figures 2b and 2e. It is clear that uniform, low frequency
energy suitable for deep crustal profiling was generated.
Data Processing
All six B A lines have been processed through stack using
Compagnie G6n6rale de G6ophysique Geovecteur software
installed on the Cray X-MP/24 supercomputer of The
University of Texas System Center for High Performance
Computing. The MCS field data, recorded in SEG-D format,
were sorted into common depth point (CDP) gathers every
OH ET AL. 227
z
+
o+
(D
o
o

o
o
z z
228 DEEP-PENETRATION SEISMIC REFLECTION IMAGES
(a)
14.2m
GECO WIDE CROSSLINE ARRAY
145m _ I_ 275m
Stbd 3 660 415 280 1195 145 105
Stbd 2
Stbd 1
Port1
Port _
SAT. NAV. ANTENNA PARAVANE
Port 3
' I
I
I 18.5m
I
_1
62m
STREAMER
6000m
CENTER OF FIRST GROUP
CENTER OF SOURCE (
SAT. NAV. ANTENNA
DISTANCE SAT. NAV. ANTENNA TO CENTER OF SOURCE 145m
DISTANCE CENTRE OF SOURCE TO CENTER OF FIRST GROUP 275m
TOTAL VOLUME 10800 cu. ln.
TOTAL No. OF ARRAYS = 6
(b)
0 100 200 300
zOO
500 (ms)
(c)
-12
-24
25 50 75 100 125 (Hz)
OH ET AE. 229
12.5 m, resulting in 60-fold data with 100 m trace spacing.
The initial (brute) stacks for each line were processed using
standard techniques, such as spike deconvolution, bandpass
filtering, normal moveout (NMO) correction, outside muting.
Four adjacent stacked CDPs were then combined to increase
the fold to 240, improving the signal-to-noise ratio (S/N)
on the stacked section and resulting in a 50 m CDP spacing.
These brute stack sections were used to interpret the major
reflection events and for stacking velocity analyses based on
velocity spectra every 2.5 kin. Even in the brute stacks,
good images of the Moho and deep crustal structures were
observed. However, the combination of thick carbonate
layers, which occurred in the study area, and the large energy
source used in the data acquisition contribute to generating
persistent multiples which often obscured upper crustal
events, including the top of basement. Processing solutions
to suppress these multiples included frequency-wavenumber
(f-k) filtering together with NMO overcorrection and inside
muting. Also, the dip moveout (DMO) processing was
employed to enhance dipping events.
F-k filtering of NMO overcorrected CDP gathers for
multiple suppression is not a new technique [Ryu, 1980].
However, it is not effective in deep penetration, sub-
basement studies unless the MCS data are acquired with a
large receiving array because of the small NMO differences
between primaries and multiples. First, to alleviate the
spatial aliasing problem, four neighboring CDP gathers were
grouped into one 240-trace gather with a 25 m trace-
separation before f-k filtering. NMO overcorrections using
the "abnormal" velocities (intermediate between the higher
primary and the lower multiple stacking velocities) were
then applied, producing overcorrected primary events and
undercorrected multiples. After rejecting the low velocity
arrivals, which contained nearly all the multiple energy, the
NMO overcorrection was removed using the same "abnormal"
velocities. Figure 3 illustrates this technique for CDP
gathers from line B A-5.
Even after f-k filtering, strong multiple noise often
remained in the near offset traces (e.g., arrows with "M" of
Figure 3d). The best way to remove this residual multiple
energy was to eliminate these near traces using an "inside"
(near offset) mute before stack. For our final stacks, up to
one-third of the nearest offset traces (<2250 m offset) were
muted. After both f-k filtering and inside muting, the S/N
improvement is remarkable on all BA lines in comparison
with their initial stacked section. For example, in Figure
4b, from line BA-4 between shotpoint (SP) 2000 and 2100,
the continuous, high-amplitude, and two-cycle reflector at
3.3-3.5 s (named reflector "J") and several dipping layers
below "J" are imaged clearly. These events are totally
blurred on the brute stack section of Figure 4a where
multiples are dominant.
Generally, structures with divergent dips cannot be imaged
properly using conventional processing CDP methods,
because the dips of the reflectors are not considered in
deriving the stacking velocities [Yilmaz, 1987, p. 334].
DMO or partial pre-stack migration, is one method that
attempts to resolve the conflicting dip problem. In this
study, 6 common offset sections, after 10:1 trace
composition of CDP gathers using partial offset stacks, were
used for DMO processing. Figure 5a shows the result of
DMO processing from part of line BA-6 (SP 2680-3320),
where the basement hinge zone lies.
Post-stack migration using finite difference time
[Claerbout and Doherty, 1972] and split-step Fourier depth
[Stoffa et al., 1990] were also applied to image upper crustal
structures. The velocities used for the depth migration were
based on interval velocities initially derived from stacking
velocities and then refined by the analysis of the CDP data
in the intercept time-ray parameter (x-p) domain [Stoffa,
1985]. Figure 5b shows a split-step Fourier depth migration
of the DMO processed data after being converted back to two
way travel time for comparison purposes. In the resulting
time section, it is clear that the landward-dipping events are
collapsed by migration and imaged as high-angle faults.
Results
We report here on two significant crustal images which
have emerged from the deep-penetration MCS reflection
profiling; wedge-shape structures seaward of the basement
hinge zone and prominent southeastward-dipping (seaward)
reflection zones in the deep crust. Because these two images
must represent different spatial structures as well as structures
formed by important geological events, we will present them
separately along with an interpretation of their geologic
significance.
Seaward-dipping Wedge Structures
In many passive continental margins around the world,
seaward-dipping layers or reflectors have been observed
[Hinz, 1981], for example, the V6ring Plateau west of
Norway [Eldholm et al., 1989]. Through deep sea drilling
cores, the nature of this section is confirmed as a thick
sequence of subaerially extruded tholeiitic basalts. Although
the existence of such thick volcanic flows along the U.S.
Fig. 2. (a) The tuned airgun source array of the GECO My used for the BA lines consisted of 36-airguns
(equivalent to 10,800 in 3) fired at ~2,000 psi every 50 m. A hydrophone streamer with a 25 m group interval and
240 channels (6,000 m length) was used for the receiving array. (b) The average normalized autocorrelogram
from a typical common depth point record of BA lines shows a high (7:1) primary peak-to-bubble ratio. The
bubble pulse period of ~145 ms was used for the design of the deconvolution operator. (c) The frequency
spectrum of (b). The vertical axis represents the normalized amplitude in dB. It is clear that this source array
produces fairly uniform energy with good low frequency-content (5-60 Hz) suitable for deep crustal prof'fiing.
0 0 0 0 0
0 0 0 0 0
(03S)
OH ET AL. 231
2100
2000 2100 2000
I I o o I
;;1t1111i lll,,';.&Ii [ ! I,,1, ,.: ....
i ',-.;: '" ? "' :v. r4 ..... ,'" .'{ ", ,' ...... IR ::.,tv, ,
3.0 l?.i%iii;:;iiillm' ..... ':,.!Jill}jlli',,l,':,,:t::.!))]:; :.
: ,: Jilt
(a) (b)
Fig. 4. The effect of f-k filtering and inside muting for the uppermost 6 s on part of the BA-4 stack section
between SP 2000 and 2100 (equal to 5 km). Figure 4a is the initial stack section where multiples dominate the
record. Nora the continuous, high-amplitude dipping reflections below 3 s in Figure 4b after both f-k filmring and
inside muting. The event at 3.3-3.5 s can be traced into land where it is named reflector "J". These dipping
reflections including "J", which are obscured by multiples originating from high reflectivity interfaces in postrift
sediments in Figure 4a, are inmrpreted to be Mesozoic subaerial basalt flows.
east coast margin has been suggested by several studies
[e.g., Hinz, 1981; Benson and Doyle, 1988], there has been
no clear seismic evidence for them due to thick overlying
postrift sediments. Mechanisms for the generation of the
seaward-dipping reflections have been suggested [e.g.,
Mutter et 1., 1982], although there is no universally
accepted model. Nonetheless, whether the U.S. east coast
continental margin originated from volcanic or non-volcanic
processes will significantly affect both its early tectonic
evolution and subsequent subsidence history.
BA-6 Profile. This 190-km-long profile, which crosses
the BMA (SP 3550), the southern Carolina trough, and the
NW $E
(a) 3300 3200 3100 3000 2900 2800 2700
3.0 ....... ..- --.- I I I t I i
p%,i"q,,,l ' .' { Itnql ""--, '; .... " W't''LI h" ...... "' Up,;3;'"';'. " ....... ; ,',t".' ....... ",;;1,; ,,,,,, ',,.,' ........ h,.. ' t: ,.,. , ,' .,,,, .,,,:.-,, ........ :: .....
Fig. 5. (a) The result of dip moveout (DMO) and (b) split-step Fourier migration from part of BA-6 between $P
2680 and 3320 at 3-9 s (see Figure I for location). The continuous, high-amplitude dipping reflector between 3
and 6 s is "J". The basement hinge zone is located at SP 3100. We observed two different type of dipping events
below "J" on the DMO section (Figure 5a): weak landward-dipping events between shot points 3100 to 3300 and
strong seaward-dipping reflections forming a wedge-shape structure between shot points 2700 to 3100. The
landward dipping events are also present on USGS line 32 where they were interpreted as basin-fill reflectors.
However, in the migrated time section (Figure 5b), it is clear that these landward-dipping events are imaged as
high-angle faults. Thus, the original landward dipping events are interpreted as diffraction tails. The seaward-
dipping reflectors, Austin et al. [1990] interpreted as Mesozoic subaerial basalt flows, producing a volcanic wedge
which may be the cause of the BMA. Note that a bright continuous dipping reflection as part of this seaward-
dipping wedge penetrates from 5 s to >9 s (i.e., to >20 km depth) in Figure 5b.
OH ET AL. 233
ECMA (SP 1800), is parallel to USGS line 32 (Figure 1).
Figure 5a shows a part of BA-6 where a distinct, two-cycle
high-amplitude, smooth reflector at ~4-6 s is observed. This
reflector is traceable from the northwestern end of the line at
~1.4 s, across the basement hinge zone (SP 3100) to the
base of the Carolina trough at ~6.5 s. We interpret this
reflection as the previously identified "J" reflector, a
Mesozoic basalt layer drilled onshore near Charleston, South
Carolina (' in Figure 1) which is dated 184+3.3 Ma [Dillon
et al., 1983; Hamilton et al., 1983; Lanphere, 1983; Austin
et al., 1990]. As in previous refraction data onshore [e.g.,
Ackermann, 1983], we measured a head wave velocity of ~6
km/s for this volcanic layer in our shot records. Above "J"
near the basement hinge zone, onlapping reflectors are
interpreted as postrift sediments. Beneath "J" at the hinge
zone, seaward fanning, high-amplitude "bright" reflectors
form a wedge-shape structure. The lowest one can be traced
to >9 s (>20 km depth in Figure 5b). Similar reflections can
be observed on USGS line 32, although they were not
interpreted by Hutchinson et al. [1983]. On the other hand,
beneath "J" near SP 3000 ("seaward" of the BMA; Figure
5a), weak, northwest-dipping diffractions at ~4-6 s are
observed. These may correspond to the "landward-dipping
reflections" of Hutchinson et al. [1983] on the USGS line
32.
BA-5 Profile. This long (252 kin) line extends into the
northern Blake Plateau basin from offshore Charleston and
crosses the B MA and southernmost portion of the ECMA
trend (Figure 1). The middle of this line, between the hinge
zone and the BMA (SP 3700), crosses the northwestward
projection of the BSFZ at a low angle. As in BA-6, "J" is
easily traced. The most conspicuous image in this profile is
a--50 km wide, asymmetric basin at ~3-6 s between SP 4200
and 5200, just seaward of the BMA (Figure 6). Its northern
edge corresponds to the basement hinge zone. The topmost
reflector of this structure, which has previously been
interpreted as an angular unconformity separating Triassic or
Paleozoic, low-grade metamorphic rocks below from clastic
Mesozoic sediments above [Dillon et al., 1983, 1985], is
now revealed as "J". The deepest reflection within this
basin reaches 6 s (--12 km depth) at SP 4850. This basin,
which may be bounded along its southeast edge by a north-
dipping normal fault, exhibits progressive shallowing in the
dip of basin-fill reflectors including "J" along its northwest
edge. This implies that this basin is a synrift half graben
presumably related to Mesozoic extension before or during
"J" emplacement. On the landward side of this synrift basin,
between the BMA and SP 4200, we observe several
southward dipping-reflections beneath "J" showing a dip of
~15 (at 5 km/s) and the same high-amplitude, two-cycle
reflection signature pattern of "J".
BA-4 Profile. This 220-km-long line extends
northwestward from the northern maxgin of the Blake Plateau
basin to offshore Charleston (Figure 1). Again, the key
observation is a wedge-shape structure just seaward of the
basement hinge zone (SP 2400 at ~3 s) (Figure 7). The
geometry of this wedge shows distinctive seaward fanning of
high-amplitude reflectors very similar to that of B A-5 and -6
(Figures 5a and 6), as well as USGS line 32. The top layer
is again the "J" reflector, which is easily traced from the
landward end of the profile. We have estimated the
thickness of these layers within the wedge using our high-
resolution BA images. For example, the "J" reflector at 3.7
s on BA-4 is >300 rn in thickness (0.125 s interval at 5
km/s). Also, the lower reflector's depth of-8 s (-20 km)
matches that of BA-6 (see Figure 5a).
Discussion. A fundamental observation on the BA lines is
the ubiquity of the "J" reflector, which probably corresponds
to the breakup unconformity on USGS line 32 [Hutchinson
et al., 1983]. It is believed to cover an area of more than
100,000 km 2 in the southeastern U.S. onshore and offshore
[Nelson et al., 1987]. At the Clubhouse Crossroad in South
Carolina near the landward extensions of lines BA -3, -4,
and-5 (' in Figure 1), one of three wells encountered a 256-
m-thick subaerial basalt flow unit at 775 meters depth which
has been dated as early Middle Jurassic on the basis of
4Ar/39Ar [Lanphere, 1983]. "J" offshore must correlate with
this flow.
Accordingly, we interpret the seaward-dipping wedges
beneath the Southeast Georgia Embayment to be nearly
contemporaneous features consisting of a series of Mesozoic
subaerial volcanic flows intercalated with sediments which
give rise to high acoustic impedance contrasts. The strong
support for this interpretation is the similar geometry of
wedge structures along the three BA lines and USGS line 32
and the configuration of onlapping postrift sediments, as
well as the similarity among prominent reflectors such as
their consistent high-amplitude, two-cycle signature
character and similar frequency content.
We interpret the asymmetric basin in BA-5 (Figure 6) to
be a synrift half graben filled with mainly volcanic
materials. It may be related to the positive magnetic
anomaly seen at the center of this basin which is the
southwestward continuation of the ECMA (see Figure 1).
This synrift basin may give insight about the formation of
the basement hinge zone. The depth of the hinge zone
deepens rapidly from -3 s (SP 4200) to 4.2 s (SP 5190).
The sharp drop of basement is likely caused by synrift
faulting along the southeast edge of this basin. An
important observation is that this basin image can be
matched very well with a proposed Mesozoic rift basin filled
with at least 6 km of sediments on COCORP GA-19 beneath
the BMA on the Georgia Coastal Plain [Nelson et al., 1987;
McBride et al., 1989].
Landward of the wedge structures a seismic pattern similar
to the "landward-dipping reflections" of USGS 32 is seen in
Figure 5a. Hutchinson et al. [1983] interpreted them as non-
magnetic sediment filling a synrift basin, and named this
feature the Brunswick graben. They proposed that the
magnetic basement beneath/flanking the graben is the cause
of the BMA. However, as explained earlier, we have
interpreted these events as diffraction tails which are
collapsed by migration (Figure 5b), not true dippin
sedimentary layers. Consequently, there is little evidence
LLI
o o
o
I
(j.u)
OH ET AL. 235
LLI
Z
o o o
o o
c. xl 1.
i i
(j.u)
0 0 0 0 0 0 0 0

(03) grill I::i^VI:IJ. AVM-OM/
E
for a graben as depicted by Hutchinson et al. [1983] on
USGS Line 32. These diffractions probably originate from
small (<0.05 s, i.e., 125 m at 5 km/s), normal fault steps of
"J" [Austin et al., 1990] or fractures within the basement
related to the formation of the basement hinge zone.
It is more likely that the wedge structures axe related to the
BMA along BA-4, -5 and -6 lines. A strong magnetic
anomaly could originate from a wedge of strongly
magnetized volcanic rocks surrounded by relatively non-
magnetic basement and overlying sediments [Phillips et al.,
1990]. In fact, Austin et al. [1990] generated a magnetic
model for the wedge structure along BA-6 which
demonstrated that the seawaxd-dipping wedge structure on
BA-6 could be responsible for the BMA. They further
suggested that the lateral extension and thickening of the
wedge seawaxd into the Carolina trough could also cause the
ECMA. Therefore, the origin of the two major magnetic
anomalies along the southeast U.S. margin might both be
related to rifting during the Mesozoic separation of Africa
and America.
The irregular distribution of low amplitude magnetic
anomalies within the Blake Plateau basin have suggested to
some workers that no oceanic crust was formed here by sea-
floor spreading [Klitgord and Behrendt, 1979; Dillon and
Popenoe, 1988]. However, the seaward-dipping reflectors of
the BA-4 and -5 lines suggest that considerable oceanic-type
volcanic material together with associated intrusive rocks
form the basement foundation of the northern Blake Plateau
basin.
Southward-dipping, lntracr ustal Reflections
The most prominent deep crustal seismic reflection
package on all the BA lines is interpreted as Moho with a
-11-12 s (-35 km) appaxent depth. Moho shows a
distinctive reverberant arrival pattern and can be traced along
most BA profiles (labeled "M" in Figure 8), despite its
occasionally broken and sporadic appearance. Moho has
previously been recognized in reflection data at only a few
locations along the U.S. Atlantic continental maxgin [e.g.,
Behrendt et al., 1983; Hutchinson et al., 1986]. Although
determination of the Moho configuration and its regional
implications are essential in understanding the overall
structure of the margin, detailed presentation of our Moho
observations axe beyond the scope of this paper. However,
caxeful inspection of deep portions (>10 s) of the BA stack
sections included here demonstrate that a seismic data
acquisition scheme using a laxge tuned airgun source and
long receiver array is one of the best ways to image deep
crustal structures including Moho, even beneath thick
carbonate sediments.
Another pronounced deep crustal image ig the southwaxd-
dipping, intracrustal reflections seen along all profiles.
They axe similax to the intracrustal reflectors seen on the
COCORP Georgia transects, which have been interpreted to
be a Late Paleozoic "Alleghanian" suture [Nelson et al.,
1985a]. It is cleax that the precise location and chaxacter of
this suture is very important for understanding the tectonic
evolution of the southeastern U.S. maxgin. In fact, after
Wilson's [1966] suggestion that most of northern Florida
may be a portion of Africa left behind after the opening of
the modern Atlantic, many studies have attempted to
determine the original, Pangea super-continent configuration
prior to its Mesozoic break-up and to locate this suture [e.g.,
Chowns and Williams, 1983; Higgins and Zietz, 1983;
Klitgord et al., 1983; Horton et al., 1984; Tauvers and
Muelberger, 1987]. Because of their proximity, the deep
crustal seismic images of the COCORP Georgia profiles can
be compaxed with their marine counterpaxt, the southwaxd-
dipping, intracrustal structures imaged on the BA profiles.
Figure 8 is composed of portions of unmigrated lines BA-
6 through BA-1 between 4 and 12 s (see Figure 1 for
locations). They axe aligned at the trough of the BMA. The
shape of the southward dipping reflection zone is easily
recognized because of its prominent reflection character. In
the case of BA-3 (Figure 8d), however, strong diffractions
probably due to the disrupted shallow-basement surface
structure obscure the dipping reflector image.
Many prominent, southward-dipping intracrustal
reflections and diffractions can be traced to extend down to
deep crustal levels, sometimes greater than the Moho axrival
times (-12 s), for example on BA-5 line (Figure 8b).
Although the diffraction events collapse and appeax above
the Moho after migration, they still remain as major deep
crustal images. These dipping reflections and diffractions
appear to be the southern limit of several, crustal-scale ramp
structures which have generally two aspects on the BA lines;
namely, imbricated or simple shapes. The two sets of south-
dipping reflection shapes axe overlapped on BA-4 line
(Figure 8-c)i Whereas a single dipping reflection package is
found on other lines. The depth to the top of these ramps
generally appeax to be -5-6 s. The width of the south-
dipping zone vaxies from 30 to 40 km, with average dip of
30 (at 6 km/s). The BMA is generally located above or
seaward of this southeastward-dipping intracrustal reflection
zone. The vaxiation of distance between the top of the ramp
and the trough of the BMA is from -2-3 km on BA-1 line to
38~39 km on BA-4 and-5 lines.
We can correlate the BA southward-dipping reflective
zones with the COCORP Georgia profiles (GA-13, -14, -16A,
and-19) in terms of their general shape and pattern of
reflections. Most of all, they form a prominent crustal-scale
ramp(s) from deep to middle crustal levels. An important
observation is their spatial relationship to the BMA. Both
onshore and offshore, they axe located either directly beneath
or landwaxd of the BMA, although there axe laxge vaxiations
in distance between the trough of the BMA and the top of
this zone from line to line. Reflection images within the
crust also change to the north and south of this zone. To
the north, the deep crust is dominated by diffractions and
vaxious oriented reflections. To the south, the Mesozoic
extension structures dominate the upper crust and few
prominent reflections axe seen beneath them.
OH ET AL. 237
300
i" . ..>. / -
.....
-300
4.0
4.0 (b)
12.0
8.0' ,
o (d)
8 8.o
: - .
20 km
Fig. 8. Sections of BA profiles showing the intracrustal seismic images between 4 and 12 s from BA-6 (top;
Figure 8a) to BA-1 (bottom; Figure 8e). All lines, which are aligned along the BMA minima, are unmigrated.
The most prominent images within the crust are the southward-dipping reflection zones with 30-40 km width,
extending to lower crustal levels with average dips of 30 . In Figure 8d from BA-3, strong diffraction energy
probably due to the disrupted shallow-basement structure partially obscures the southward dipping reflectors
image. These dipping reflections are the southern limit of the crustal-scale ramp structures which are correlated to
the COCORP intracrustal reflectors of Nelson et al. [1985] who considered them to be a possible Late Paleozoic
suture between North America and Africa. Note the significant distance variations between the BMA and this
suture zone from line to line, which confirm that the cause of the BMA is not this suture. "M" at 11 - 12 s indicats
Moho reflections.
On the basis of crustal fabric, these intracrustal reflection
images can be compared with "Zone 3" proposed by
Allmendinger et al. [1987] to represent a complex basement
thrust ramp and duplex zone (i.e., compressional fabric) with
a pronounced post-thrusting extensional history. In fact,
Nelson et al. [1985] have interpreted the southward dipping
reflections of the COCORP Georgia profiles as a Late
Paleozoic suture based on their uniqueness as the only major
crustal-penetrating reflection features between the proposed
African basement and known North American basement and
their appropriate position as the potential root zone for Late
Paleozoic, inner Piedmont nappes in the southern
Appalachians. They further suggest that the BMA marks
this Late Paleozoic suture zone and that the ECMA simply
represents its northward continuation. This implies that the
central-north Atlantic opened along the Paleozoic
Alleghanian suture. Subsequently, McBride and Nelson
[1988] further proposed that a single induced magnetized
slab emplaced "immediately" above the southern flank of the
suture is the cause of both the BMA and the ECMA.
However, whatever the origin of the dipping slab is, the
geometry of the McBride and Nelson [1988] magnetic model
is not compatible with the B A seismic images. The major
problem is the location of the northern edge of their
proposed dipping slab relative to the south-dipping
reflections of the suture zone. As stressed earlier, there is
significant variation in the distance between the southward-
dipping reflective zones (i.e., the proposed suture) and the
trough of the BMA. In the Georgia COCORP profiles, the
BMA is located just above the top of the suture zone, which
would be compatible with a magnetized body lying
immediately south of the suture. However, along lines BA-5
and BA-4 (Figure 8b and 8c), the top of the southward-
dipping complex is -40 km northward of the BMA.
Therefore, the BMA cannot be caused by a magnetic body
emplaced immediately adjacent to the suture zone.
Another conflicting observation of the magnetic models
of the McBride and Nelson [1988] is the depth and induced
magnetization strength of the dipping slab. In their
magnetic modeling, the depth to the top of the dipping slab
is less than 7 km along USGS line 32 in the Carolina
trough, and reaches 25 km at the bottom. In line BA-6
(Figure 5b), we image the subaerial volcanic layer extending
from 5 km to more than 20 km depth as part of the seaward-
dipping wedge between the BMA and ECMA. To apply their
intrusive or tectonic slab model to our crustal images, nearly
all of the slab would have to be situated deeper than 20 km
or well beneath the seaward dipping volcanic wedge. This
would require much stronger induced magnetization than is
reasonable for rocks at such great depth. A similar
geometric arrangement prevails along lines B A-4 and -5 (see
Figures 6 and 7).
Along line BA-3, two wells have reached Paleozoic
section beneath post-Mesozoic drift sediment: the Transco
1005-1 well near SP 4100 and the COST GE-1 well near SP
4900 (Figure 1). The latter reached the Upper Devonian
basement showing relatively low grade of metamorphic, and
the former, which is located above the seaward-dipping
reflective zone, penetrated undeformed Silurian and Lower
Ordovician sedimentary and meta-sedimentary rocks [Dillon
and Popenoe, 1988]. The lithology and mineralogy of
Transco 879-m-thick strata shows strong similarity to
undisturbed Paleozoic deposits from onshore drill holes
southwest of line BA-1 [Poppe and Dillon, 1989]. This
suggests that, if the southward-dipping intracrustal structures
represent a zone of thrusting due to continental collision,
the emplacement of the proposed suture along the B A lines
is probably an Early Paleozoic or pre-Cambrian event, not a
Late Paleozoic "Alleghanian" event, since it did not disturb
the overlying Ordovician-Silurian sedimentary strata.
In summary, we can correlate the southward-dipping
intracrustal reflectors on the B A lines to the COCORP
intracrustal reflectors and consider them to be a possible
suture zone. However, the cause of the BMA is not this
proposed suture.
Conclusions
The interpretation of the images from deep-penetration
MCS profiling can be summarized as follows:
1. The high-amplitude, continuous, two-cycle "J"
reflector, which correlates with a previously identified early
Middle Jurassic subaerially volcanic flow drilled onshore,
can be traced from land to the basement hinge zone, where
postrift sediments onlap it.
2. At the basement hinge zone along profiles BA-4, -5
and -6, seaward-dipping wedge-shape structures consisting of
a series of high amplitude reflectors, including "J" at the
top, are imaged. These wedge-shaped structures lie seaward
of the BMA, and have a similar spatial geometry along each
line. The reflectors within these wedges are interpreted to be
the Mesozoic, rift-related subaerial volcanic rocks that may
be responsible for the BMA along BA-4, -5 and -6.
3. The previously postulated Mesozoic rift graben on just
the "landward side" of the BMA, which were proposed from
the previous aeromagnetic and MCS studies, are not imaged
on any BA profiles. Particularly, the "Brunswick graben",
which was suggested as the cause of the BMA on USGS line
32, was most likely diffraction tails from faults.
4. The southward-dipping intracrustal reflection zone,
which has been interpreted as a suture, is observed on all BA
profiles but with significant distance variations from the
axis of the BMA. Accordingly, the BMA cannot be related
directly to a suture in our data. The highly magnetized
simple slab model suggested by McBride and Nelson [1988]
to account for the B MA is not be supported by the new deep-
crustal seismic images. The age of this crustal-scale ramp
structures along B A lines is probably as old as the Early
Paleozoic.
Acknowledgements: This work was supported by the
National Science Foundation Grant OCE-8711300. D. Dean
helped to process the seismic data, and W. Lloyd drafted the
illustrations. We would like to acknowledge the support of
OH ET AL. 239
The University of Texas Systems Center for High
Performance Computing.
The University of Texas Institute for Geophysics
contribution number 865.
References
Alsop, L. E., and M. Talwani, The east coast magnetic
anomaly, Science, 226, 1189-1191, 1984.
Ackermann' H. D., Seismic-refraction study in the area of the
Charleston, South Carolina, 1886 Earthquake, in Studies
Related to the Charleston, South Carolina, Earthquake of
1886; Tectonics and Seismicity, edited by G. S. Gotm,
U.S. Geol. Surv. Prof. Pap., 1313, F1-F20, 1983.
Allmendinger, R. W., K. D. Nelson, C. J. Potter, M.
Baraangi, L. D. Brown, J. E. Oliver, Deep seismic
reflection characteristics of the continental crust,
Geology, 15, 304-310, 1987.
Austin, J. A., Jr., and E. Uchupi, Continental-oceanic crustal
transition off Southwest Africa, Am. Assoc. Pet. Geol.
Bull., 64, 501-526, 1982.
Austin, J. A. Jr., P. L. Stoffa, J. D. Phillips, J. Oh, D. S.
Sawyer, O. M. Purdy, E. Reiter, and J. Markis, Crustal
structure of the Southeast Georgia embayment-Carolina
trough: Preliminary results of a composite seismic image
of a continental suture(?) and a volcanic passive margin,
Geology, 18, 1023-1027, 1990.
Bally, A. W., Atlantic-type margins, Geology of passive
continental margins' History, Structure and
Sedimentologic Record (With Special Emphasis on the
Atlantic Margin), Educ. Course Note Ser. 19, pp. 11-148,
Am. Assoc. Pet. Geol., Tulsa, Okla., 1981.
Behrendt, J. C., R. M. Hamilton, H. D. Ackermann' V. J.
Henry, and K. C. Bayer, Marine multichannel seismic-
reflection evidence for Cenozoic faulting and deep crustal
structure near Charleston, S.C., in Studies Related to the
Charleston, South Carolina, Earthquake of 1886; Tectonics
and Seismicity, edited by G. S. Gohn, U.S. Geol. Surv.
Prof. Pap., 1313, J1-J29, 1983.
Behrendt, J. C., and M. S. Grim, Structural elements of the
U.S. Atlantic continental margin delineated from second
vertical derivative of aeromagnetic data, U.S. Geological
Survey Map GP-956, scale 1:5,000,000, 1983.
Benson, R. N., and R. G. Doyle, Early Mesozoic rift basins
and the development of the United States middle Atlantic
continental margin, in Triassic-Jurassic Rifting,
Continental Breakup and the Origin of the Atlantic Ocean
and Passive Margins, Part A, edited by W. Manspeizer,
pp. 99-127, Elsevier, Amsterdam, 1988.
Claerbout, J. F., and S. M. Doherty, Downward continuation
of moveout-corrected seismogram, Geophysics, 37, 741-
768, 1972.
Chowns, T. M. and C. T. Williams, Pre-Cretaceous rocks
beneath the Georgia Coastal Plain, in Studies Related to
the Charleston, South Carolina, Earthquake of 1886;
Tectonics and Seismicity, edited by G. S. Gohn, U.S.
Geol. Surv. Prof. Pap., 1313, L1-L41, 1983.
Daniels, D. L., I. Zietz, and P. Popenoe, Distribution of
subsurface lower Mesozoic rocks in the southeastern
United States as interpreted from regional aeromagnetic
and gravity maps, in Studies Related to the Charleston,
South Carolina, Earthquake of 1886; Tectonics and
Seismicity, edited by G. S. Gotre, U.S. Geol. Surv. Prof.
Pap., 1313, K1-K24, 1983.
Dillon, W. P., K. D. Klitgord, C. K. Paull, Mesozoic
development and structure of the continental margin off
South Carolina, in Studies Related to the Charleston,
South Carolina, Earthquake of 1886; Tectonics and
Seismicity, edited by G. S. Gotre, U.S. Geol. Surv. Prof.
Pap., 1313, N1-N16, 1983.
Dillon, W. P., C. K. Paull, and L. E. Gilbert, History of the
Atlantic continental margin off Florida; The Blake Plateau
basin, Geologic evolution of the United States Atlantic
margin, edited by C. W. Poag, pp. 189-215, Van Nostrand
Reinhold, New York, 1985.
Dillon, W. P., and P. Popenoe, The Blake Plateau Basin and
Carolina Trough, in The Geology of North America, Vol.
1-2, The Atlantic Continental Margin U.S., edited by R. E.
Sheridan and J. A. Grow, pp. 291-328, Geological Society
of America, Boulder, Colo., 1988.
Eldholm, O., J. Thiede, and E. Taylor, The Norwegian
continental margin: Tectonics, volcanic and
paleoenvironmental framework, in Proceedings of the
Ocean Drilling Program, Vol. 104, Scientific Results, pp.
5-26, College Station, Tex., 1989.
Emery, K. O., E. Uchupi, J. D. Phillips, C. O. Bowin, E. T.
Bunce, and S. T. Knott, Continental rise off Eastern North
America, Am. Assoc. Pet. Geol. Bull., 54, 44-108, 1970.
Folger, D. A., Evolution of the Atlantic continental margin
of the United States, in Deep drilling results in the
Atlantic Ocean: Continental margins and
paleoenvironment: American Geophysical Union, Maurice
Ewing Ser. Vol. 3, edited by M. Talwani, W. Hay, and W.
B. F. Ryan, pp. 87-108, AGU, Washington, D.C., 1979.
Hall, D. J., 1990, Gulf Coast-East Coast magnetic anomaly
I: Root of the main crustal decollement for the
Appalachian-Ouachita orogen, Geology, 18, 862-865,
1990.
Hamilton, R. M., J. C. Behrendt, and H. D. Ackermann,
Land multichannel seismic-reflection evidence for tectonic
features near Charleston, South Carolina, in Studies
Related to the Charleston, South Carolina, Earthquake of
1886; Tectonics and Seismicity, edited by G. S. Gohn,
U.S. Geol. Surv. Prof. Pap., 1313, 11-118, 1983.
Higgins, M., and I. Zietz, Geologic interpretation of
geophysical maps of the pre-Cretaceous "basement"
beneath the Coastal Plain of the Southeastern United
States, in Contributions to the Tectonics and Geophysics
of Mountain Chains, edited by R. D., Jr. Hatcher, H.
Williams, and I. Zietz, Mere. Geol. Soc. Am., 158, 125-
130, 1983.
Hinz, K., A hypothesis on terrestrial catastrophs; wedges of Nelson, K. D., J. A. Amow, J. H. McBride, J. H. Willemin,
very thick, oceanward-dipping layers beneath passive J. Huang, L. Zheng, J. E. Oliver, L. D. Brown, and S.
continental margins, Geol. Jahrb., Reihe E., Geophysik, Kaufman, New COCORP profiling in the southeastern
22, 3-28,1981. United States. Part I, Late Paleozoic suture and Mesozoic
Horton, J. W., Jr., I. Zietz, and T. L. Neathery, Truncation rift basin, Geology, 13, 714-718, 1985a.
of the Appalachian Piedmont beneath the coastal plain of Nelson, K. D., J. H. McBride, J. A. Amow, J. E. Oliver, L.
Alabama: Evidence from new magnetic data, Geology, 12, D. Brown, and S. Kaufman, New COCORP profiling in the
51-55, 1984. southeastern United States. Part II, Brunswick and East
Hutchinson, D. R., J. A. Grow, K. D. Klitgord, and B.A. Coast Magnetic anomalies; Opening of the north-central
Swift, Deep structure and evolution of the Carolina Atlantic Ocean, Geology, 13, 718-721, 1985b.
Trough, in Studies in Continental Margin Geology, edited Nelson, K. D., J. H. McBride, J. A. Amow, J. A. Wille, J.
by Watkins, J. S., and Drake, C. L., Mere. Am. Assoc. E. Oliver, L. D. Brown, and S. Kaufman, Result of recent
Pet. Geol., 34, 129-152, 1983. COCORP profiling in the southeastern United States,
Hutchinson, D. R., K. D. Klitgord, and R. S. Detrick, 1986, Geophys. J. R. Astron. Soc., 89, 141-146, 1987.
Rift basins of the Long Island platform, Geol. Soc. Am. Phillips, J. D., L.M. Gahagan, D. M. Miller, J.A. Austin,
Bull., 97, 688-702, 1986. Jr., J. Oh, P. L. Stoffa, and D. Sawyer, Synthetic
Klitgord, K. D., and Behrendt, J. C., Basin structure of the Modeling of the East Coast and Brunswick Magnetic
U.S. Atlantic margin, in Geological and Geophysical Anomaly Trends, Southeastern U.S.: Geologic
Investigations of Continental Margins, edited by J.S. Interpretation, Eos-Trans., 71, 1616, 1990.
Watkins, L. Montadert, and P. W. Dickerson, Mere. Am. Pitman, W. C., and M. Talwani, Sea floor spreading in the
Assoc. Pet. Geol., 29, 85-112, 1979. North Atlantic, Geol. Soc. Am. Bull., 83, 619-649, 1972.
Klitgord, K. D., W. P. Dillon, and P. Popenoe, Mesozoic Poppe, L. J., and W. P. Dillon, Petrology of Some Lower
tectonics of the southeastern United States Coastal Plain Ordovician-Silurian Sedimentary Strata From the Southeast
and Continental Margin, in Studies Related to the Georgia Embayment, U.S. Outer Continental Shelf,
Charleston, South Carolina, Earthquake of 1886; Tectonics Southwestern Geology, 29, 169-190, 1989.
and Seismicity, edited by G. S. Gohn, U.S. Geol. Surv. Ryu, J. V., Decomposition of seismic gathers into velocity
Prof. Pap., 1313, P1-P15, 1983. components by a space-time filter, Presented at the 50th
Klitgord, K. D., D. R. Hutchinson, and H. Schouten, U.S. Ann. Int. SEG Meeting, 1980.
Atlantic continental margin: Structural and tectonic Sheridan, R. E., J. A. Grow, and K. D. Klitgord, Geophysical
framework, in The Geology of North America, Vol. 1-2, data, in The Geology of North America, Vol. 1-2, The
The Atlantic Continental Margin U.S., edited by R.E. Atlantic Continental Margin U.S., edited by R. E.
Sheridan and J. A. Grow, pp. 19-55, Geological Society of Sheridan and J. A. Grow, pp. 177-196, Geological Society
America, Boulder, Colo., 1988. of America, Boulder, Colo., 1988.
Lanphere, M. A., 4Ar/39Ar ages of basalt from Clubhouse Stoffa, P. L., Analysis and processing of wide-angle
Crossroads Test Hole # 2, near Charleston, South reflection and refraction seismic data in the tau-p domain,
Carolina, in Studies Related to the Charleston, South Adv. Geoph. Data Proc., 2, 81-117, 1985.
Carolina, Earthquake of 1886; Tectonics and Seismicity, Stoffa, P. L., J. T. Fokkema, R. M. de Luna Freire, and W.
edited by G. S. Gohn, U.S. Geol. Surv. Prof. Pap., 1313, P. Kessinger, Split-step Fourier migration, Geophysics,
B1-B8, 1983. 55, 410-421, 1990.
LASE Study Group, Deep structure of the U.S. East Coast Tauvers, P. R., and W. R. Muehlberger, Is the Brunswick
passive margin from large aperture seismic experiments magnetic anomaly really the Alleghanian suture?,
(LASE), Mar. Pet. Geol., 3 , 234-242, 1986. Tectonics, 6, 331-342, 1987.
McBride, J. H., and K. D. Nelson, Integration of COCORP Trhu, A.M., A. Ballard, L. M. Dotman, J. F. Getmast, K.
deep reflection and magnetic anomaly analysis in the
southeastern United States: Implications for origin of the
Brunswick and East Coast magnetic anomalies, Geol. Soc.
Am. Bull., 100, 436-445, 1988.
McBride, J. H., K. D. Nelson, and L. D. Brown, Evidence
and implications of an extensive early Mesozoic rift basin
D. Klitgord, and A. Schreiner, Smacture of the lower crust
beneath the Carolina Trough, U.S. Atlantic Continental
Margin, J. Geophy. Res., 94, 10585-10600, 1989.
White, R. S., and D. P. McKenzie, Magmatism at rift zones:
the generation of volcanic continental margins and flood
basalts, J. Geophy. Res, 94, 7685-7729, 1989.
and basalt/diabase sequence beneath the southeast Coastal Wilson, J. T., Did the Atlantic close and then reopen?,
Plain, Geol. $oc. Am. Bull., 101, 512-520, 1989. Nature, 211. 676-688, 1966.
Mutter, J. C., M. Talwani, and P. L. Stoffa, Origin of Yilmaz, O, Seismic data processing, pp. 526, Society of
seaward-dipping reflectors in oceanic crust off the Exploration Geophysicists, Tulsa, Okla., 1987.
Norwegian margin by "subaerial sea-floor spreading",
Geology, 10, 353-357, 1982.
IMAGES OF THE PLATE BOUNDARY BENEATH SOUTHERN ALASKA
Thomas M. Brocher, Michael J. Moses, Michael A. Fisher, Christopher D. Stephens,
and Eric L. Geist
U.S. Geological Survey, M/S 977, 345 Middlefield Road, Menlo Park,
California 94025
Abstract. The geometry of the plate boundary beneath
Prince William Sound in Southern Alaska has been deter-
mined from a combined seismic reflection and wide-angle
reflection/refraction experiment. Strong reflections dipping
landward from 6 to 8 s originate from a mid-crustal velo-
city discontinuity at a depth of 15 to 24 km. Wide-angle
diving waves indicate that the reflections originate at the
top of a 6.7 km/s layer, which underlies inferred accreted
sedimentary rocks having velocities less than 6 krn/s. This
reflective, high-velocity, mid-crustal layer is also highly
magnetic; and thus may represent basalts of the underthrust
Yakutat terrane. The seismic character of the reflector, its
subparallel relationship to the subducting oceanic plate, its
location near the top of the Wadati-Benioff seismic zone,
and its correspondence in dip and possibly depth to the
inferred slip plane for the great 1964 earthquake suggests
that this prominent mid-crustal reflector represents the inter-
plate decollement between North America and the lower
crust of the subducting Yakutat terrane sutured to the top of
the Pacific plate.
Introduction
Modern marine seismic reflection profiles recorded both
on- and offshore by fixed receivers provide coincident
seismic reflection and refraction images of continents and
their margins. Such onshore-offshore studies in the North
and Baltic Seas [Jokat and Flueh, 1987; BABEL Working
Group, in press], the Great Lakes [Mereu et al., 1990], the
central coast of California [Levander and Putzig, in press;
Trehu, in press], have allowed the determination of crustal
structure and velocities to help interpret the coincident
reflection profiles. Here we present a synthesis of seismic
reflection and refraction data acquired during a onshore-
offshore experiment in the Gulf of Alaska and Prince Wil-
liam Sound [Brocher and Moses, 1990] with aeromagnetic
data and hypocentral locations of microearthquakes to
derive a tectonic model for the Gulf of Alaska.
Geologic Setting
Southern Alaska is composed of a series of accreted,
far-traveled terranes [Jones et al., 1987]. In the study area
Geodynamics 22
(Figure 1), the Aleutian Trench forms the southeastern
boundary of the Prince William terrane, a belt of Paleogene
rocks that extends southwestward from Prince William
Sound to Kodiak Island and beyond. The Prince William
terrane consists primarily of turbidites, intercalated pillow
basalt flows, and granitic plutons, formed in place at the
Aleutian Trench, and is bounded to the northwest by the
Chugach termne, which is composed of similar lithologies
of Cretaceous age [Plafker, 1987]. The Contact Fault
forms the boundary between the two terranes. The Prince
William terrane is bounded on the east by the Yakutat ter-
rane, which is a composite terrane consisting of Lower
Eocene and possibly Paleocene oceanic crust, upper Meso-
zoic flysch, and Eocene granitoid intrusions [Plafker, 1987].
The Yakutat terrane is underthrust beneath the Prince Wil-
liarn terrane along a northeasterly trending suture through
the vicinity of Kayak Island [Bruns, 1985], although the
lower crust of the Yakutat terrane may be decoupled from
the upper crust at a location to the southeast of the Kayak
Island suture [Plafker, 1987]. Structural relations observed
in seismic reflection profiling of the Alaskan margin sug-
gest that the Yakutat terrane has been moving with the
Pacific plate for the last 5 Ma [Bruns, 1985; Plafker et al.,
1989].
The Seismic Experiment
In 1988, the U.S. Geological Survey (USGS) acquired
1107 km of multichannel seismic reflection profiles along 6
lines in the Gulf of Alaska and Prince William Sound,
Alaska [Fisher et al., 1989b]. These data were acquired in
support of the Trans-Alaska Crustal Transect (TACT), a
comprehensive geological and geophysical study of the
crust along a corridor parallel to the Trans-Alaska oil pipe-
line [Page et al., 1986]. The location of one line, known
as the TACT line, is highlighted in Figure 1. The 40 gun,
tuned source array totalling 126 liters (7700 cu. in.), was
fired at 50 m intervals (approximately 20 s) along the
reflection lines. At 16 sites onshore, the USGS temporarily
deployed three-component analog seismic recorders
[Brocher and Moses, 1990]. The wide-angle data were
highpass filtered above 3 Hz, deconvolved, and plotted as
seismic record sections with and without Automatic Gain
Control. Three adjacent traces were stacked to enhance
signal levels.
The 40-fold seismic reflection data were processed at the
USGS in Menlo Park and at the Houston Area Research
241
242 IMAGING THE PLATE BOUNDARY BENEATH S. ALASKA
153ow 149W
Alaska
BRF
terrane
145W
141W
61 N
Kodiak
I.
2OO m
Montague Middleton
I. I
4000 m
Yakutat
terrane
Pacific plate
j63 mm/yr
Gulf of
Alaska
1 O0 km
59ON
57N
Fig. 1. Terrane map of southern Alaska centered on the
northern Gulf of Alaska and Prince William Sound (PWS)
showing the location of TACT multichannel seismic
reflection line and locations of receivers (filled circles) used
Consortium. The relative-amplitude processing of these data
included multiple suppression, deconvolution, and f-k filter-
ing to reduce strong diffractions generated by the rough
seafloor topography. The data shown in Figure 2 were used
to constrain the interpretation of the wide-angle data and to
map the results from the TACT line to other reflection lines
not as well instrumented with wide-angle recorders.
Results
In the vertical-incidence data, a prominent mid- to
lower-crustal reflection is seen between 6 and 8 s two-way
traveltime (TWTT) and can be traced nearly continuously
from the NW end of the TACT line seaward to Middleton
Island, near the shelf break (Figure 2). In the wide-angle
data, this reflection can be observed throughout Prince Wil-
liam Sound and has a low-frequency reverberant character
that is 1 to 2 s long (Figure 3). This reflection has an
apparent dip of 2 landward, and can be traced from a
depth of about 17 km beneath Montague Island, to about 24
km beneath the Contact Fault (Figure 2).
The wide-angle profiles record useful arrivals produced
by the airgun army source along the entire TACT line, to
maximum offsets of 290 km. Forward modeling of the
travel times of refractions and wide-angle reflections was
used to obtain a preliminary velocity-depth model along the
to obtain record sections shown in Figures 3 and 4. BRF
and CF are the Border Ranges and Contact faults, respec-
tively. Knight Island (K.I.) is the location for the seismi-
city data shown in Figure 5.
TACT line (Figure 4). Receiver density was highest within
Prince William Sound, although recorders were also posi-
tioned on Middleton Island and at the seaward end of the
TACT line [Brocher and Moses, 1990]. The forward
modeling indicates that the units above the prominent mid-
crustal reflector have seismic velocities less than 6 km/s, as
one would expect for the accreted sedimentary rocks and
granitic plutons of the Prince William terrane. Wide-angle
diving waves define a 3- to 4-km-thick unit immediately
below the reflector with velocities of around 6.7 kns; the
large velocity contrast between rocks above and below the
prominent reflection observed from this mid-crustal horizon,
of course, accounts for the strength of the reflection.
Refraction data indicate a landward-dip (northwesterly)
on Moho of 9 . Discontinuous, weak reflections may define
the top of the subducting oceanic plate to as deep as 10 s
TWTT. While the velocity model shown in Figure 4 is
consistent with our interpretation of the reflection Moho
beneath the shelf, neither the available reflection data nor
wide-angle refraction data constrain the depth of the Moho
on the northwesternmost end of the TACT line (Figure 4).
Other Geophysical Constraints
Relocation of relatively well-constrained earthquake
hypocenters using a one-dimensional approximation to our
velocity model, and considerations of regional stress orien-
BROCHER et al. 243
Slope
Magnetic
Anomaly
Nw C:.nntnt f.t ,P. ro. je. ct!o,n of, Montague I. I ;,m,,t,, I Aleutian Trench SE
............... r,,nght Isana Distance km / .............. /
350 ,I, 300 ,I, ,I, 250 ,I, 200 ' * ,I, 50 00 * 50
... i:!:i:i:!:i::!:i:.::...::i:i:i:!:.:::!:!:i:i::i:!:i:.::k.:i83..::!:i:i:i:::::.:::. . ........ .........:.:.'.-:.:.:-.-.:.:.:.:-.'.:.:.:.:.:.:.:.:.:.:.
tltit tet ..... :'" ....
10-
. _
VE 1.6:1
14
Fig. 2. Interpreted line drawing of the unmigrated TACT line showing prominent mid-crustal reflection
between 6 and 8 s TWTT. The seismic velocities of the rocks above this reflection, inferred to represent
the plate boundary, are consistent with their interpretation as accreted sedimentary rocks and granitic plu-
tons of the Prince William terrane of North America.
Distance, km
10 20 30 40
Nw .............. z ...................... z .................. ................. z ..................... ...................... s. .................. z ................... s. ............. SE
8
Fig. 2 Inset. Illustration of a portion of the unmigrated TACT line from which the line drawing shown in Fig. 2
was made.
tations inferred from focal mechanisms [Page et al., 1988,
1989], suggest that the mid-crustal reflector is at or near the
top of the Wadati-Benioff seismic zone (Figure 5). These
earthquakes were recorded by both regional and temporary
seismographic stations deployed in the vicinity of Knight
Island (Figure 1). For earthquakes beneath Knight Island
having well-constrained focal mechanisms, the principal
axes of least compressire stress are subhorizontal and gen-
244 IMAGING THE PLATE BOUNDARY BENEATH S. ALASKA
PWS Line Gather
sw NE
25 3 25 km
...................... ,.. , . .,. , , , .,, , , , .,.. , , ,,, , , , ,,, ,
6 :
8 , . -...-,- -% - Mid-crustal
-----'-' ':-' ';4 -- reflector
12 _, _ '--' ' ".: : M o h o
Fig. 3. Portion of a record section obtained from the Prince William Sound, showing prominent mid-
crustal reflections (between 6 and 7 s). The section has been plotted as a reflection gather (unreduced)
using Automatic Gain Control. Each trace represents the sum of three adjacent airgun shots.
NW
4
Contact
fault
I
, ,...,;.-.~..- .,,;:. ,: .. ,%..,4... ,.: o, ...c:,z.'-',.......,
; ."i.7.",:.%'.?.,,.-'": ,,.,,-.......-.,-..'.. ,. > . ,:.-;.,, .,-.:,..,.:
'', :.=..':.:,, :. ..'.' -?:("" '-" '"- ' '-. -', '..... ,, .'.:...,'..:, "-',..'X,
{:-..-33.';:. ::. :v:- .'../:::. :i .:.. 4:.. ;:. ;. :.' . .:q :,.;'&.._. .'.:-..' _
*"", . . ,. .'..:::,-7'..::: '.;:'...',::?: ;'>. "-:' ''""" "::
577;:"'""'":'"'..,' :,,. '--::1:' '"'"" :':" '" '" ' '" -' ' ':::i:"'""' ..... "" ' ' "':;
I I
Slope Middleton
Hinchinbrook Magnetic
Island Anomaly Island
Figure 5 I
Aleutian
Trench
o - _ .- , --_ -= . . I , .
I
10 -
E
, _
20
--
Upper -
30 mantle _
i i i i i I
3OO 2OO 1 O0
Distance, km
Fig. 4. Lower figure shows a velocity model for the TACT line with raypaths of arrivals along the top
of the subducting oceanic upper mantle. Upper figure compares observed wide-angle data with calcu-
lated times (solid lines) for rays refracted in the upper mantle of the subducting oceanic lithosphere.
Large solid dots along velocity model indicate the projections of receiver locations used for the wide-
angle study. Boxed area of velocity model shows location of seismicity highlighted in Figure 5.
BROCHER et al. 245
p-
w
5O
NW SE
lOO
DISTANCE (Klvl)
10
3O
200
I I
6'3 Ikm/ /
o 20
DISTANCE (KM)
Fig. 5. Plot of regional Wadati-Benioff seismicity along a
transect parallel to the TACT line [modified from Page et
al., 1989]; inset below plots relocated hypocenters from a
differing, local seismicity study [Page et al., 1988] using
the velocity model inferred from the wide-angle data show-
ing the correspondence of the mid-crustal reflector with the
top of the Wadati-Benioff zone along a portion of the
TACT line shown in Figure 4.
erally aligned with the relative motion of the subducting
Pacific plate, consistent with results for Wadati-Benioff
seismicity north of Prince William Sound [Page et al.,
1989].
The dip of the prominent mid-crustal event between Mid-
dieton Island and Montague Island lies within the range of
dips determined for the low-angle slip plane which pro-
duced the great 1964 earthquake [Hastie and Savage, 1970;
Miyashita and Matsu'ura, 1978]. These estimates of the
dip of the slip plane, ranging between 4 and 7 , were
obtained from inversions of coseismi vertical and horizon-
tal displacements, and are considered to be the best con-
strained parameter inverted from these data [Miyashita and
Matsu'ura, 1978]. While these data do not provide tight
constraints on the depth of the slip plane, they are con-
sistent with a slip plane about 5 km above the prominent
mid-crustal reflection, and significantly more shallow than
the depth to the top of the subducting oceanic crust as
modeled in Figure 4.
Griscom and Sauer [1990] model a prominent magnetic
anomaly on the continental shelf south of Prince William
Sound (the Slope Magnetic Anomaly whose location is
shown on Figures 2 and 4) using a magnetized layer dip-
ping landward between 14 and 20 km depth (their profile
CC'). This dipping layer corresponds closely to the high-
velocity, reflective mid-crustal horizon defined seismically.
These magnetic rocks can be traced eastward in the mag-
netic data to outcrops of Eocene basalts in the "basement"
of the Yakutat terrane [Bruns, 1985]; therefore it would
appear that the basaltic basement of the Yakutat terrane
constitutes at least the upper part of the currently subduct-
ing crust in the northern Gulf of Alaska.
Based on the velocity model, magnetics, earthquake
hypocenters and inferred orientations of principal stress
axes, and slip models for the 1964 Alaska earthquake, we
hypothesize that the prominent mid-crustal reflector beneath
Prince William terrane represents the interplate decollement
between North America and the Pacific plate. By Pacific
plate we include the subducted portion of the Yakutat ter-
rane, which has been sutured to the Pacific plate for the last
5 Ma [Bruns, 1985]. The seismic data from Prince William
Sound thus provide an image of an important seismogenic
plate boundary. In our interpretation of previous TACT
results north of Prince William Sound [Fisher et al., 1989a;
Fuis et al., in press], the underthrust Yakutat terrane can be
traced landward from the Aleutian Trench for at least 200
km, and possibly more. This interpretation is compatible
with the over 300 km length of Yakutat terrane subducted
during the past 5 Ma at the 63 mm/y convergence rate
[Minster and Jordan, 1978]. Our model implies that slivers
of far-traveled terranes may be partially, if not completely,
subducted, and thus represents another mechanism for
recycling of the crust.
Acknowledgments. Supported by the U.S. Geological
Survey Deep Continental Studies Program. We thank E.
Flueh, W. Frank, G. Fuis, and J. McCarthy for reviewing
early drafts of the manuscript.
References
BABEL Working Group, Recording marine airgun shots at
offsets between 300 and 700 km, Geophys. Res. Lett. in
press.
Brocher, T. M., and M. J. Moses, Wide-angle seismic
recordings obtained during the TACT multichannel
reflection profiling in the northern Gulf of Alaska, U.S.
Geol. Surv. Open File Rep., 90-663, 40 p., 1990.
Bruns, T. R., Tectonics of the Yakutat block, an allochtho-
nous terrane in the northern Gulf of Alaska, U.S. Geol.
Surv. Open File Rep., 85-13, 112 pp., 1985.
Fisher, M. A., T. M. Brocher, W. J. Nokleberg, G. Plafker,
and G. L. Smith, Seismic reflection images of the crust of
the northern part of the Chugach terrane, Alaska: Results
of a survey for the Trans-Alaska Crustal Transect
(TACT), J. Geophys. Res., 94, 4424-4440, 1989a.
Fisher, M. A., T. M. Brocher, T. R. Bruns, and E. Geist,
Seismic reflections from a possible brittle/ductile transi-
tion within the accretionary wedge near the Eastern Aleu-
tian Trench, EOS Trans. AGU, 70, 1339, 1989b.
Fuis, G. S., E. L. Ambos, W. D. Mooney, and N. I.
Christensen, Crustal structure of accreted terranes in
southern Alaska -- Chugach Mountains and Copper River
246 IMAGING TI-IE PLATE BOUNDARY BENEATH S. ALASKA
Basin -- from seismic-refraction results, J. Geophys. Res.,
in press.
Griscom, A., and P. E. Sauer, Interpretation of magnetic
maps of the northern Gulf of Alaska, with emphasis on
the source of the Slope Anomaly, U.S. Geol. Surv. Open
File Rep., 90-348, 18 pp., 1990.
Hastie, L. M., and J. C. Savage, A dislocation model of the
1964 Alaska earthquake, Bull. Seismol. Soc. Am., 60,
1389-1392, 1970.
Jokat, W., and E. R. Flueh, On the use of airgun arrays for
seismic refraction investigations of the crust, First Break,
5. 440-447, 1987.
Jones, D. L., N.J. Silberling, P. J. Coney, and G. Plafker,
Lithotectonic terrane map of Alaska (west of the 41st
meridian), U.S. Geol. Surv. Misc. Field Studies Map.
MF- 1874-B, 1 sheet, scale 1:2,500,000 and text.
Levander, A. R., and N. E. Putzig, Continuous offset
seismic profile on the central California margin, J. Geo-
phys. Res., in press.
Mereu, R. F., D. Epili, and A. G. Green, Pg shingles: Prel-
iminary results from the onshore GLIMPCE refraction
experiment, Tectonophysics, 173, 617-626, 1990.
Minster, J. B., and T. H. Jordan, Present-day plate motions,
J. Geophys. Res., 83, 5331-5354, 1978.
Miyashita, K., and M. Matsu'ura, Inversion analysis of
static displacement data associated with the Alaska earth-
quake of 1964, J. Phys. Earth, 26, 333-349, 1978.
Page, R. A., G. Plafker, G. S. Fuis, W. J. Nokleberg, E. L.
Ambos, W. D. Mooney, and D. L. Campbell, Accretion
and subduction tectonics in the Chugach Mountains and
Copper River basin, Alaska: Initial results of the Trans-
Alaska Crustal Transect, Geoloev. 14. 501-505, 1986.
Page, R. A., K. A. Folgeman, C. D. Stephens, and J. C.
Lahr, State of stress in the subducted Pacific plate
beneath Prince William Sound, southern Alaska, Seismol.
Res. Lett., 59, 16, 1988.
Page, R. A., C. D. Stephens, and J. C. Lahr, Seismicity of
the Wrangell and Aleutian Wadati-Benioff zones and the
North American plate along the Trans-Alaska Crustal
Transect, Chugach Mountains and Copper River Basin,
southern Alaska, J. Geophys. Res., 94, 16,059-16,082,
1989.
Plafker, G., Regional geology and petroleum potential of
the northern Gulf of Alaska continental margin, in Geo-
logy and resource potential of the continental margin of
western North America and adjacent ocean basins -
Beaufort Sea to Baja California, Earth Science Series,
vol. 6, edited by D. W. Scholl, A. Grantz, and J. G.
Vedder, pp. 229-268, Circum-Pacific Council for Energy
and Mineral Resources, Houston, Texas, 1987.
Plafker, G., W. J. Nokleberg, and J. S. Lull, Bedrock geol-
ogy and tectonic evolution of the Wrangellia, Peninsular,
and Chugach terranes along the Trans-Alaska Crustal
Transect in the Chugach Mountains and southern Copper
River Basin, Alaska, J. Geophys. Res., 94, 4255-4295,
1989.
Stephens, C. D., R. A. Page, and J. C. Lahr, Reflected and
mode-converted seismic waves within the shallow Aleu-
tian subduction zone, southern Kenai Peninsula, Alaska,
J. Geophys. Res., 95, 6883-6897, 1990.
Trehu, A., Tracing the subducted oceanic crust beneath the
central California continental margin: Results from ocean
bottom seismometers deployed during the 1986
PG&E/EDGE experiment, J. Geophys. Res., in press.
LITHOPROBE REFLECTION TRANSECT OF SOUTHWESTERN CANADA: MESOZOIC THRUST AND
FOLD BELT TO MID-OCEAN RIDGE
Frederick A. Cook and John L. Varsek
Department of Geology and Geophysics
University of Calgary, Calgary, Alberta Canada T2N 1N4
and
Ronald M. Clowes
Department of Geophysics and Astronomy
University of British Columbia, Vancouver, British Columbia V6T 1W5
Abstract. Deep seismic reflection profiles continental shelf and slope, the data image
recorded between 1984 and 1989 provide a 1200 km terranes underthrust beneath the pre-Tertiary
long section that crosses the entire southern margin, structural deformation associated with
Canadian Cordillera, the active Cascadia the accretionary wedge, and the top of the
subduction zone, the oceanic Juan de Fuca plate, subducting oceanic crust. Across the Juan de
the Juan de Fuca ridge, and ends on the Pacific Fuca plate, the igneous crust produces prominent
plate. On the east, the cross section begins in continuous reflections from its upper surface,
the Mesozoic Rocky Mountain thrust and fold belt from the Moho, and from a newly discovered
where the reflectivity is dominated by upper seamount that may exhibit subcrustal magmatic
crustal sediments deformed during east-directed underplating. The oceanic Moho is imaged to
thin-skin thrusting. The profiles progress within 5 km of the east side of the Juan de Fuca
westward across the central Cordillera where ridge, is unclear beneath the ridge, and then
Mesozoic terrane accretion and basement-involved reappears west of the ridge where the section
crustal imbrication were followed by Eocene ends.
regional crustal extension. Throughout this
region, the seismic data show reflectivity at
all levels in the crust, a well defined Moho Introduction
that decreases in travel time from 12.0 s on the
east to about 10.5 s on the west over a distance The southern Canadian Cordillera is one of
of about 250 km, and crustal scale antiforms the best studied deformed belts in the world. It
associated with exposures of gneiss complexes. includes four major tectonic elements, the Rocky
On the western side of the mainland, the Mountain foreland thrust and fold belt, the
profiles cross Mesozoic accreted terranes and Intermontane belt composite terrane, the Insular
plutonic complexes of the Coast Mountains, belt composite terrane, and the Cenozoic
including the recent Garibaldi/Pemberton subduction complex (Figure 1). These elements
volcanic belt. In this region, faults associated record a history of continental evolution
with at least two stages of accretion are involving both constructive processes of
observed in the upper, middle, and lower crust, allochthonous terrane accretion, and destructive
and the Moho is commonly seen at about 11.0-11.5 processes of orogenic collapse and extension.
s. Near the coast, the data exhibit reflections The LITHOPROBE transect crosses the entire
to about 16.0 s, some of which are downdip from orogen and links to data that project westward
reflections observed beneath Vancouver Island into the Pacific ocean. Accordingly, it
and are likely associated with the Cascadia represents the first such cross section to
subduction complex. Offshore beneath the traverse from an ancient thrust and fold belt,
across an orogenic interior, across an active
subduction zone, and over a young oceanic plate
Continental Liohere: Dp Seisnfic Refltions to end near an active spreading ridge (Figure
Geodynanfis 22 1). This paper presents an overview of some of
o 1991 American Geophysical Union the preliminary results of the transect.
247
248 LITHOPROBE REFLECTION TRANSECT OF SOUTHWESTERN CANADA
o
128 120
........
128 120
//ridge
' Oceanic crust Csces-n Juan trua
Ceozol rete erres .
I a rle suu{io eie
InSular mslte lerrane . '
marg,n sheIt and toredeep
Autochthonous
Deep semmc prohie
used m cross sctio
Other deep seismic prohie AIIochthonous
Fig. 1. Map of the southern Canadian Cordillera with the
positions of the LITHOPROBE seismic reflection profiles (heavy
and light lines with end bars). Only the heavy and numbered
profiles that form a more or less continuous east-west transect
are described here. Other profiles, on the mainland, Vancouver
Island, and offshore, are discussed in more detail elsewhere
[Clowes et al., 1987; Hyndman et al., 1990; Cook et al., 1988].
In the east, the transect crosses the Rocky Juan de Fuca plate in the eastern Pacific ocean
Mountain foreland thrust and fold belt comprised (Figure 1). The Juan de Fuca plate consists of
of Proterozoic to Mesozoic miogeoclinal normal oceanic lithosphere and is bounded on the
sedimentary strata that were deformed and west by the Juan de Fuca ridge. The transect
translated eastward along east-verging thrust crosses this ridge and ends on the Pacific plate
faults during the Jurassic though Early (Figure 1).
Tertiary. The deformation of this region is
characterized as "thin-skin", as the faults LITHOPROBE Reflection Data
flatten above or at the surface of the
underlying craton [Bally et al., 1966]. To the Reflection data were acquired in the southern
west, the transect crosses into rocks that were Canadian Cordillera between 1984 and 1988.
accreted to North America during Mesozoic to Initial work in 1984 on Vancouver Island [Clowes
Early Tertiary orogenic activity that was et al., 1987] was followed in 1985 by offshore
responsible for the formation of the thrust and data acquisition carried out by the Geological
fold belt. This wide region includes the Survey of Canada [Hyndman et al., 1990] and by
Intermontane and Insular composite terranes, acquisition of data in the Rocky Mountain
each of which is an amalgamation of smaller foreland belt [Cook et al., 1988]. The transect
terranes. Two periods of compressional was completed with nearly 1000 km of data across
deformation, metamorphism, and plutonism, one in the Intermontane and Insular composite terranes
the Jurassic, and one in the Late in 1988.
Cretaceous/Paleocene, coincide with the The data presented here include only profiles
accretion of these composite terranes [Monger et that form a more or less continuous transect
al., 1982]. Much of the Intermontane composite across the Cordillera to the Pacific plate.
terrane was subjected to crustal-scale extension Additional profiles, such as cross lines and
during the Eocene [Parrish et al., 1988]. parallel lines, are described in more detail in
The transect crosses the western margin of the references cited above and others therein.
the Insular composite terrane, characterized by Although data acquisition and processing have
material accreted during Cenozoic subduction of been carried out with somewhat different
the Juan de Fuca plate, and proceeds onto the parameters from year to year, the sections
COOK ET AL. 249
TABLE 1. Legend for Cross Sections
Terranes and Assemblages
Label
North American Paleozoic
North American Proterozoic
Kootenay Terrane
Quesnel Terrane
Wrangellia
Georgia Basin
Pacific Rim Terrane
Crescent Terrane
Tofino Basin
Oceanic Sediments
PzNA
PNA
PKO, PZKO
1P zQN, MQN
P zWR-MWR
KTGB
MPR
eTCR
TTB
Os
Faults Label
Ashlu Creek
Ainsworth
Beavan (Monashee Mtns.)
Benard (Purcell Mtns.)
Beaufort Range
Cherryville (Monashee Mtns.)
Cowichan Lake (Vancouver Island)
Chapperon
Columbia River
Coldwater
Duffy Lake
Fraser River
Hall Lake
Kwoiek Creek
Marshall Creek
Miller Creek
Monashee Decollement
Okanagan Valley
Owl Lake
Pasayten
Quilchena Creek
Shuksan
Slocan Lake
St. Mary
Tofino (Vancouver Island)
Torrent (Rocky Mtns.)
Thomas Lake
Valkyr Shear Zone
West Coast
ACF
AF
BF
BF
BRF
CF
CF
CHF
CRF
CWF
DLF
FF
HLF
KF
MCF
MMF
MD
OVF
OLF
PF
QCF
SF
SLF
SMF
TF
TF
TLF
VSZ
WCF
Plutonic Rocks Label
Cretaceous to Tertiary granite
Jurassic to Cretaceous granite
Mid-, Early Jurassic granite
Kg, Tg, KTg
JKg
mJg, eJg
crust where they tend to flatten. Drilling and
industry reflection data show that they flatten
above the cratonic basement [Bally et al.,
1966], and the LITHOPROBE profiles show that the
basement can be followed westward beneath the
extensional Rocky Mountain trench and the
Purcell anticlinorium (Figure 2).
The Purcell anticlinorium consists primarily
of Proterozoic to Paleozoic strata that were
compressed and transported eastward along the
thrust faults that feed into the faults beneath
the Rocky Mountains [Price, 1981]. On its
western flank, the strata were arched and
rotated into a west-dipping panel that is
bounded on the west by rocks of the easternmost
accreted terrane, Quesnellia. This terrane,
part of the Intermontane composite terrane, is
underlain by shallow faults that can be
correlated with surface structures related to
the accretion boundary [Varsek and Cook, 1991]
and that are in turn underlain by North American
strata. Thus, seismic evidence indicates that
North American rocks project in the subsurface
to the west of the Purcell anticlinorium,
affirming interpretations that the easternmost
metamorphic complexes [Valhalla and Monashee
complexes; Parrish et al., 1988; Brown and Carr,
1990] are composed of metamorphosed North
American basement and cover.
Intermontane Composite Terrane
Profiles 5 through 18 (Figures 1 and 2) cross
the Intermontane composite terrane. On the
east, these lines traverse thin slivers of the

eastern edge of Quesnellia (Figures 1 and 2)
that are preserved in the hangingwalls of Eocene
normal faults [Parrish et al., 1988; Cook et
al., 1988]. The east-west portion of the
transect along profiles 7-9, however, traverses
rocks in the footwalls of some major normal
faults that are interpreted as North American
cover and basement which were arched during
compression, and then exposed during extensional
collapse of the orogen (Monashee complex and
Vernon antiform; Figure 2). In this region the
illustrated in Figure 2 are processed with crust exhibits remarkable reflectivity at all
similar post stack filters and migrations. For levels. Many features (e.g. MD in Figures 2
example, all of the profiles have had a and 3) can be correlated with surface structures
coherency filter applied, both to the unmigrated and followed to lower crustal travel times (8.0
and migrated (Figure 2a) data so that they s or more).
resemble line drawings but do not have inherent A regional extensional detachment (Okanagan
interpretational problems associated with Valley fault - OVF in Figures 2 and 3) can be
picking reflections. followed westward in the upper and middle crust
as a series of truncations of overlying
Rocky Mountain Foreland Belt reflectors. The fault is probably offset by
steep normal faults that bound the east and west
Crustal reflection data across the Rocky flanks of the Central Nicola horst (QCF and CWF
Mountain foreland belt (profiles 1-4 in Figure in Figures 2 and 3). The Nicola horst may also
2) confirm the interpretation of thin-skin be cored by Mesozoic compressional faults, but
tectonics for this region [Bally et al., 1966; it is not known if these faults are within
Cook et al., 1988]. Thrust faults and related accreted terrane rocks, or whether they involve
stratigraphy can be traced from the surface North American basement and cover. In any case,
along west-dipping trajectories into the shallow the suture between the Intermontane composite
OKANAGAN
LAKE
MONASHEE
MTNS a)
GUlCHON BATHOLITH MQN NICOLA HORST MQN IPZQN mJg E-PZKo IPz-MQN E-PZKo MQN
eJg CWF QCF CHF? OVF VERNON ANTIFORM CF BF CRF
I I I i I I I I I
-:-. ......... . ,.c-::.-_-'_:?:,,:,-.-<.._. .;<,,.., .:; j --.,.-.':,:.--,';..; - : :..:,..:.... ,....----._. ...-_,_.,,:...-.....- , .. ;-. .' :. -.:o
-.. ' ..'-.-..2::.
11 I I o I I' 9
ARROW
LAKE
o
o
I 8 I I "'7 I 15
0 50 100km
I , , ,,I ' I
OKANAGAN MONASHEE ARROW
b) LAKE MTNS LAKE
GUlCHON BATHOLITH MQN NICOLA HORST MQN IPZQN rnJg E-PzKc IPz-MQN I-PZKo MQN
eJg CWF QCF CHF? OVF VERNON ANTIFORM CF BF CRF
.ii...........:-....:....:...-.....-....................-....o..o, ..:..=:+:=?::.:`=7.:T:::..::.::T:.:.:T:..:`=:.:..::.:T::.:.=.:::.:.::.:`::=........=............ .....:......,_.,,_._- _ _ __. .,%,.._ _-. :-. _-. :::. 0
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ..... '"%:<:::'.'.::.' ::.: :'.:'.':.: ::.: ::.':'. :.'.'.:.". :::.' ::." ................. - - - --- ...... - b
.=;i,:...;i<:::.;::::::,;<=: o
I
... ..' -
I 11 10 i 9 I 8 I I 7 I
0 50 100km
I I
Fig. 3. (a) Enlargement of a portion of the migrated data from
the intermontane region to illustrate the quality of the
reflectivity. Migration "smiles" are obvious between lines. (b)
Interpretation of (a). Note the remarkable MohO along these
profiles and the crustal antiforms (Monashee Mountains, Vernon
antiform). interpretation is constrained by good correlations
with surface features.
terrane and North American rocks is viewed in that the Moho and lower crust acted as
this interpretation as a complex boundary that detachment zones. In terms of crustal
was arched during compression as faults cut to processes, this result suggests that rocks
deep structural levels during crustal scale formed at mid- to lower crustal depths have been
duplexing (Monashee complex), and that was then uplifted, unroofed and eroded without the
offset along Eocene extensional faults (Figure formation of any compensating roots. Because we
2). do not observe any such roots, ductile flow in
Throughout this region, the reflection Moho the lower crust appears to be an important
is a prominent feature that decreases in travel process.
time from about 12.0 s near the Valhalla complex The west side of the Intermontane composite
on the east to about 10.5 s under the Nicola terrane follows the complex Fraser River fault
horst on the west (Figures 2 and 3). The fact system (FF in Figure 2). This fault zone is
that the Moho is subhorizontal over such a large part of a Middle Tertiary right lateral strike
region implies that it formed either by magmatic slip system that has approximately 100 km of
intrusion that overprints the crustal layering offset, that juxtaposes Insular composite
above, or as a zone of detachment into which terrane on the west to the Intermontane
both extensional and earlier compressional composite terrane on the east, and that is
structures sole. The observation that many of subparallel to the Coast-Cascade orogen (Figure
the crustal reflections are listric into the 1). The subsurface expression of the Fraser
lower crust and Moho (e. g. west side of River fault is uncertain; reflections at about
Monashee mountains, Vernon antiform near 6.0 s are nearly continuous beneath it,
Okanagan lake, and the Nicola horst) indicates suggesting the fault trace is detached or
COOK ET AL. 253
flattens into the middle crust. Alternatively, Vancouver Island. If the deep E reflections
the fault zone may extend through the crust, as continue beneath Georgia Strait and are
offsets between similar orogen-parallel truncated against the crust-mantle boundary as
structures cannot be detected. illustrated in Figure 2, either the Moho is a
young feature overprinting structural continuity
Insular Composite Terrane to shallower features, or it is a relic
subduction scar that formed when accretion was
The Insular composite terranes were farther inboard. These interpretations imply
amalgamated outboard of the North American the E reflector may flatten at the Moho (Figure
continent during the Jurassic and Early 2), and that the area beneath profile 16 has
Cretaceous, and later accreted to North America been the site of lower crustal delamination
during the Late Cretaceous and Paleocene [Monger during formation of both the C and E structures.
et al., 1982; Gabrielse and Yorath, 1989; Alternatively, the sub-Moho event can be
Journeay, 1990; Monger, 1990] to produce the correlated to the E reflector on Vancouver
Coast-Cascade orogen. Profiles 18, 13 and 14 Island and offshore to profile 85-01, close to
exhibit distinctive dip domains at different the boundary between oceanic crust and overlying
crustal levels (Figure 2). The upper crust accretionary prism [Clowes, 1990] In this case
between 0.0 and 4.0 s is characterized by the E structure may represent strain
moderate to steeply east dipping reflections; partitioning from the contact between the
the middle crust between about 4.0 and 7.0 s is oceanic and continental plates into the
dominated by subhorizontal reflections, and the continental framework. Accordingly the interval
lower crust between about 7.0 and 11.5 s is between the E reflector and the subducting plate
characterized by both east and west dipping may be structurally isolated continental
reflections. The subhorizontal middle crustal lithosphere partly coupled to the downgoing
reflections appear to truncate both the upper oceanic plate.
crustal and lower crustal fabrics. Application The margin west of the Insular composite
of downplunge projections from exposed surface terrane includes two accreted terranes (Pacific
structures suggests that these patterns are most Rim and Crescent), the modern accretionary prism
easily interpreted as resulting from an early below the continental shelf/slope, and the
phase of primarily west-directed contraction Oligocene to Recent Tofino basin deposited on
that created the east dipping structures of the the other units. The basement to the basin is
upper crust, and a later phase of west-east marked by rough topography inherited from the
contraction that caused either west-directed terranes and accretionary wedge. Basin fill is
wedging of Intermontane rocks into the lower folded with amplitude decreasing upward
crust of the Insular terrane, or east-directed indicating that deformation has been continuous.
truncation of the older Insular terrane The sedimentary Pacific Rim terrane and volcanic
structures. crescent terrane have been clearly imaged on
Profile 16 borders the eastern shore of this and other profiles; they were thrust
Georgia Strait and exhibits a panel of east beneath the Insular composite terrane and each
dipping reflections from 3.5 to 15.0 s that can other in the Late Eocene [Hyndman et al., 1990].
be divided into four major structures, including The west end of profile 85-01 images the top of
two east dipping ramps located at about 4.0 and basaltic oceanic crust and overlying hemipelagic
7.0 seconds, an east facing wedge of high and turbiditic sediments. Landward dipping
reflectivity between 9.0 and 12.0 s, and an east thrusts at the toe of the accretionary wedge
dipping reflection zone between 13.0 and 15.0 s. flatten above the oceanic crust indicating that
The upper and middle crustal ramps are likely most of the sediment is being scraped off the
related to west directed thrust systems of Late downgoing plate [Davis and Hyndman, 1989].
Cretaceous and Early Tertiary age found on
eastern Vancouver Island [Monger, 1990]. The Pacific Ocean
reflective wedge in the lower crust is listric
into the Moho and is probably bounded by a roof Profiles 85-01, 85-09 and 85-03 extend across
and floor thrust. The upper boundary can be the Juan de Fuca plate to the Juan de Fuca
projected across Georgia Strait to the 'C' ridge, an active divergent plate boundary. On
reflector on Vancouver Island where it is all of these profiles, there are three
thought to be a decoupling zone, possibly reflecting zones. They include a thin veneer of
related to subduction erosion of accreted lower more or less horizontally layered sediments that
crust and mantle and Pacific Rim terrane [Clowes thicken eastward toward the Cascadia subduction
et. al., 1987]. If this interpretation is zone and are deformed at the accretionary wedge
correct, it appears the Moho acted as a (Figure 2), the rough surface of the oceanic
delamination zone by partitioning crust and basement (Figure 2a), and the subhorizontal
mantle rocks. oceanic Moho (Figure 2). On the western side of
Interpretations for the sub-Moho reflectors profile 85-09 a seamount that may be underlain
on profile 16 depend upon whether they are or by a subcrustal layered igneous complex [Calvert
are not continuous with the E reflection zone on et al., 1990] is visible.
The Juan de Fuca ridge is visible as a Insular belts). Initial processing was by
shallowing of the ocean basin on profile 85-03 Veritas Seismic, Geophoto Services Inc., and
and as a series of extensional fault blocks Western Geophysical; final processing was
(Figure 2). Oceanic sediments on the Juan de completed at the LITHOPROBE Seismic Processing
Fuca plate thin westward toward the ridge, and Facility at the University of Calgary with the
the oceanic Moho appears to rise beneath it. assistance of Dr. Kris Vasudevan. We gratefully
Moho reflections are occur west of the ridge in acknowledge the assistance of geohysicists and
the Pacific plate. geologists associated with the LITHOPROBE
Southern Cordillera Transect, particulary R. L.
Summary and Conclusions Brown, S. D. Carr, A. G. Green, J. M. Journeay,
E. R. Kanasewich, J. W. H. Monger, R. R.
The LITHOPROBE southern Canadian Cordillera Parrish, R. A. Price, and C. Spencer. LITHOPROBE
seismic reflection transect has been combined contribution No. 221.
with reflection data from the Pacific Ocean to
form a more or less continuous crustal section
from the Mesozoic/Early Tertiary Rocky Mountain References
thrust and fold belt on the east to the active
Juan de Fuca ridge and oceanic Pacific plate on
the west. The profiles show that east-verging Bally, A., P. Gordy, and G. Stewart, Structure,
thrust structures of the Rocky Mountains can be seismic data and orogenic evolution of the
followed westward beneath the Intermontane southern Canadian Rockies, Bull. Can. Petr.
composite terrane to the eastern side of the Geology, 1--4, 337-381, 1966.
Insular composite terrane in the Coast Brown, R., and S. Carr, Lithospheric thickening
Mountains. Throughout much of the Intermontane and orogenic collapse within the Canadian
region, the Moho is a remarkable reflecting Cordillera, in Proc. Pacific Rim Congress 90,
boundary that has a regional east dip of about Parkville, Victoria, Australia, I__I, 1-10,
1.5 s over 250 km and has little relief beneath 1990.
complex crustal structures that have up to 25 km Calvert, A., E. Hasselgren, and R. Clowes,
of structural relief. Crustal structures are Oceanic rift propagation-A cause of crustal
apparently listric into the lower crust or the underplating and seamount volcanism, Geology,
Moho and are not obviously overprinted by it. 18, 886-889, 1990.
The Insular composite terrane includes Clowes, R., LITHOPROBE-Multidisciplinary studies
complex structures formed during Late of continental evolution: An example from
Cretaceous/Paleocene, that were transposed and western Canada, in roc. Pacific Rim Congress
overprinted by subduction-related processes, 90, Parkville, Victoria, Australia, III,
including structural deformation and formation 83-97.
of igneous rocks, that continues today. The Clowes, R., M. Brandon, A. Green, C. Yorath, A.
Pacific ocean west of the Cascadia subduction Sutherland Brown, E. Kanasewich, and C.
zone includes typical oceanic crustal structure Spencer, LITHOPROBE southern Vancouver Island:
such as a recently discovered seamount, shallow Cenozoic subduction complex imaged by deep
Moho, thin ocean sediments overlying basaltic seismic reflections, Can. J. Earth Sci., 2--4,
basement, and both flanks of the active Juan de 31-51, 1987.
Fuca ridge. Cook, F., A. Green, P. Simony, R. Price, R.
These profiles thus provide the first crustal Parrish, B. Milkereit, P. Gordy, R. Brown, K.
reflection link between an ancient thrust and Coflin, and C. Patenaude, LITHOPROBE seismic
fold belt, two metamorphic core zones that reflection structure of the southeastern
resulted from collision of ancient accreted Canadian Cordillera: Initial results,
terranes, an active subduction complex, and a Tectonics, !, 157-180, 1988.
modern ocean basin with associated active ridge. Davis, E.E., and Hyndman, R.D., Accretion and
Along this transect it is possible to view the recent deformation of sediments along the
crustal relationships between the tectonic northern Cascadia subduction zone. Geol. Soc.
collage that now comprises western Canada and of America Bull., 101, 1465-1480.
the ongoing subduction processes that were Gabrielse, H., and Yorath, C., The Cordilleran
responsible for the formation and modification orogen in Canada, Geoscince anaa, 16,
of that collage. 67-83, 1989.
Hyndman, R.D., Yorath, C.J., Clowes, R.M., and
Acknowledgments. The LITHOPROBE program is Davis, E.E., The northern Cascadia subduction
funded by the Natural Sciences and Engineering zone at Vancouver island: seismic structure
Research Council of Canada and the Geological and tectonic history, Can. J, Qf Earth
Survey of Canada. Field data acquisition was by Sciences,27, 313-329, 1990.
Veritas Geophysical (1984-Vancouver Island), Journeay, J.M., A progress report on the
Enertec Geophysical (1985-eastern Cordillera), structural and tectonic framework of the
Geophoto Services Inc. (1985 and 1989 offshore) southern Coast belt, British Columbia, Geol.
and Sonics Exploration (1988 Intermontane and Surv. of Canada, Paper 0-1E, 183-195, 1990.
COOK ET AL. 255
Monger, J.W.H., Georgia basin: Regional setting southern Omineca belt, British Columbia and
and adjacent Coast mountains geology, British Washington, Tectonics, , 181-212, 1988.
Columbia, Geol. Surv. Canada Paper 90-1E, Price, R., The Cordilleran foreland thrust and
95-101, 1990. fold belt in the southern Canadian Rocky
Monger, J., R. Price, and D. Tempelman-Kluit, Mountains, Spec. Publ. Geol. $oc. London, ,
Tectonic accretion and the origin of two 4 2 7 - 4 4 8 , 1 9 8 1 .
metamorphic and plutonic welts in the Canadian Varsek, J., and F. Cook, Seismic reflection
Cordillera, Geoloq, 10, 70-75, 1982. geometry of a folded and detached accretionary
Parrish, R., S. Carr, and D. Parkinson, Eocene complex, Kootenay arc, British Columbia,
extensional tectonics and geochronology of the Geoloq, 1--9, 159-162.
SEISMIC STRUCTURE OF THE NORTHERN CASCADIA ACCRETIONARY PRISM:
EVIDENCE FROM NEW MULTICHANNEL SEISMIC REFLECTION DATA
G.D. Spence
Department of Physics and Astronomy
and Centre for Earth and Ocean Research
University of Victoria, Victoria, B.C., Canada V8W 3P6
R.D. Hyndman, E.E. Davis and C.J. Yorath
Pacific Geoscience Centre
Geological Survey of Canada, Sidney, B.C., Canada V8L 4B2
Abstract. Within the Cascadia accretionary prism west of Vancouver
Island, new marine seismic reflection profiles totalling 722 km provide
improved images which enable the three-dimensional variation of
structures along the margin to be determined. At the deformation front,
landward dipping thrusts spaced roughly 5 km apart are commonly
observed which may penetrate to near the top of the subdueting oceanic
crust. Along lines separated by 3 km, the amount of displacement on a
given fault is seen to vary significantly along the margin. Seaward
dipping faults occasionally develop, propagating upward from the frontal
thrust. In one region above gently-dipping oceanic crust, a shallow taper
sediment wedge has formed, implying high pore fluid pressures along the
detachment surface at the top of the oceanic crust. The detachment
exhibits a strong reflection probably indicative of high fluid pressure. The
accretionary prism is bounded by a landward dipping continental
backstop, formed by the marine volcanic Crescent Terrane. The base of
this terrane is imaged extending down to near the top of the subdueting
oceanic crust, so that little sediment is available for deeper subduction or
underplating. The seaward part of the Crescent Terrane appears to have
been uplifted, probably as a response to the accretion of prism sediments.
Introduction
The western Canadian convergent margin, where the Juan de Fuca
plate subducts beneath the North America plate, has been the subject of
numerous recent geophysical and geological studies. Hyndman et al.
[1990] present a summary of the recent onshore and offshore regional
geophysical work, which includes multichannel seismic reflection, shallow
seismic profiling, seismic refraction, gravity and magnetic surveys, plus
geothermal, seismicity and magnetotelluric measurements.
As part of the first phase of the multidisciplinary Canadian LITHO-
PROBE program in 1984, high-quality multichannel reflection data were
collected across Vancouver Island [Clowes et al., 1987; Green et al.,
1986, 1987]. In 1985, widely spaced offshore multichannel seismic lines
were recorded across the continental shelf and slope, as part of the
Geodynamics 22
Frontier Geoscience Program of the Geological Survey of Canada. Inte-
grated with the previous geophysical and geological data, the main results
of this offshore program [Hyndman et al., 1990; Davis and Hyndman,
1989; Davis et al., 1990; Hyndman and Davis, 1991] were
(1) the defmition of landward dipping thrusts at the deformation front
which cut through the sediment section to near the igneous ocean crust
basement, implying that few sediments are available for underplating or
deep subduetion.
(2) the delineation of two narrow terranes acefeted to the margin in the
Eocene (about 42 Ma) - the Mesozoic marine sedimentary Pacific Rim
Terrane and the Eocene marine volcanic Crescent Terrane.
(3) the imaging of the top of the downgoing oceanic crust to depths of
over 40 km beneath Vancouver Island.
(4) the recognition that the shallow bottom-simulating-reflector
observed beneath the lower continental slope marked the base of a layer
of methane hydrate that could be used as a measure of temperature
gradient and heat flow. Subsequently, it has been suggested that the
hydrate layer can be used as a semi-quantitative indicator of fluid flow
from within the accretionary prism.
In this paper we present new seismic data that provide additional
constraints on the structure and tectonic processes in the region. As part
of site surveys for the international Ocean Drilling Program, 722 km of
marine multichannel reflection lines were collected in 1989 across the
western Canadian convergent margin (Figure 1). The primary objective
of the program was to determine the three-dimensional variation of
structures along the Vancouver Island subduction zone. Of particular
interest were the geometry of the accretionary wedge and the acefeted
terranes, the variation in thrust and fold structures at the deformation
front, the nature of the basal thrust or detachment that could produce very
large earthquakes, and characteristics of the bottom simulating reflector.
This paper presents a preliminary description and interpretation of some
of the prominent features on the record sections.
Data Acquisition and Processing
Four of the new lines crossed the continental shelf (lines 1, 2, 6, 9; see
Figure 1). These were also recorded by the University of British
Columbia at two or more land refraction sites near the end of each line,
to provide important deep velocity information. Detailed grids of lines,
257
258 SEISMIC STRUCTURE OF CASCADIA ACCRETIONARY PRISM
49I
4800 ,
12800 ,
127o00 ' 126o00 ,
i i
12500 '
12400 '
'O JUAN DE FUCA
,, 1112
,:'romet ,.. Vancouv,r,s,and
, :.:.::.:..
\ .' ......
, .........:.:
\ , '.:.:.:. , / ...' .................... ..:.:.:.:
7 . ' '"--' ' .................... :'.. .............. ..........................
. - -... , :...:.:.:.:.:.:.:.:.:.:.: ::..:.:....:.:.:.:.:.: ...:.:.:.:.:.:.:.:.:.:........:.:.
, .' , / / . ,....... ..................... ::.:.:.:...,, ............. : ..............
. ..... /..:.: ......:.:.:.:.:.:.:..........-.........:.:.:............-.:.:
'7 ,' Cygnet
,' , .'-.' ''"'-.'--":......".....'. i:::::i:i: .... "') "' "-- .'.. '. :..-.':.::-:i:i::.!::i:!i
,, ' ':' "'"'"' ::':..:.__.'..:i':..::i::!:::i T'-'-"'""ii:i.i.:: if- ":'::::::: =========================== ' ":'::::i:i:i:i:!
.tO / . / [ /y, -,x '..'.-.-.-.-.. -.;....... :..:..:;::..
I I
Fig. 1. Location map for the 1989 Cascadia margin multichannel seismic
profiles (solid lines; segments shown in Figures 2-4 are indicated by the
thickened lines). Locations of older marine multichannel profiles (dotted
lines) and offshore exploratory wells are also shown. Hachured areas
indicate regions of positive magnetic anomalies on the continental shelf.
The shelf edge and lower slope region are defined by the 200 m and 2000
m bathymetry contours.
at approximately 3 km spacing, were recorded across the deformation
front in two regions covering sites proposed for scientific drilling by the
Ocean Drilling Program. Three new profiles (lines 3, 4, 5) were run
parallel to the 1985 Line 1, the extension of the main Lithoprobe land
seismic line, and two profiles (lines 7, 8) were obtained parallel to the
1985 Line 2. As well, two cross lines were collected in each region.
The airgun array source had a total volume of 125 L (7820 in). A
144-channel streamer recorded 36-fold data to a maximum offset of 3795
m. The shot and group intervals were 50 m and 25 m, respectively. Shots
were fired by an integrated navigation system controlled by Starfix.
Processing procedures included (1) true amplitude recovery, with the
application of exponential gain and spherical divergence corrections; (2)
f-k velocity filter to suppress low velocity backscatter; (3) zero-phase
source designature plus f-k demultiple; (4) derivation of stacking
velocities from semblance analyses for sub-bottom times less than about
5 s, combined with a general refraction velocity model for the area for
greater depths [Waldron et al., 1989]; (5) gapped deconvolution after
stacking; (6) f-k migration, using hand-smoothed NMO functions with
10% velocity reduction to prevent over-migration.
Preliminary Interpretation
Structures Near the Deformation Front
(i) Frontal folds and faults. Figure 2 shows the seaward portion of
Lines 3 and 4. The sediment sequence incoming from Cascadia Basin
comprises a rapidly-deposited Pleistocene turbiditc section (upper 1.5 s)
overlying a finer-grained pre-Pleistocene hemipelagic unit (lower 0.8 s).
At the deformation front, the sediment section is folded and faulted into
margin-parallel anticlinal ridges with the steeper limbs facing seaward
[Davis and Hyndman, 1989]. Thrust faults and related anticlines occur at
a spacing of about 5 km (LDF in Figure 2) and in some cases appear to
penetrate almost completely through the sediment section. Thrust faults
are defined both by relative reflector displacements across the fault, and
by reflections from the fault plane itself. Asymmetrical anticlines in the
fault hanging wall develop as material moves up the thrust ramp; the
steep seaward-facing slope is formed landward of where the thrust breaks
the surface or flattens near the seafloor. A gently landward-dipping limb
develops as the sediment package in the hanging wall is rotated. The
uplift of the hanging wall strata can be used to constrain the profile of the
thrusts. In some cases the layers are backtilted in a highly planar fashion
which requires the thrusts to be circular in profile [Davis and Hyndman,
1989].
The four profiles 89-03, 85-01, 89-04 and 89-05 in the southeastern
detailed area (Figure 1) have separations of only 3 km, so the three-
dimensional variation of deformation along the margin can be determined.
For a given line, the offsets of reflective turbiditc horizons are used to
calculate displacements along the thrusts. The velocity at the depth of a
given horizon is estimated from reflection velocity analyses [Davis et al.,
1990, Figure 9], and the local dip of the thrust ramp determined. Near
the base of the turbidites, dips range from 30 to 40 . The vertical uplift
of the thrust packet is measured, and displacements along the faults are
calculated by simple trigonometry. At this time only a single value has
been determined for each fault, although the data quality will probably
permit the variation in displacement with depth to be resolved.
From Table I, a systematic variation along strike is seen for the four
lines. Displacements along the most seaward fault LDF1 increase
substantially to the north from about 300 m on Line 89-03 to 100 m on
Line 89-05, while displacements along LDF2 decrease slightly to the
north from 1150 m to 950 m. The more landward fault LDF3 is not
clearly identified on Line 89-03, but on the other 3 lines there is a
northerly increase in fault displacement to over 300 m for Line 89-05.
SPENCE ET AL. 259
T
I
M4 T
E
(s)
LINE 89-04 V 5 km I
Deformation Front
SP 400 500 600 700 800 9x)
I I i I I BSR
LDF1 LDF2 LDF _
\
.
.
LINE 89-03
200 300
400 500 600 700
SP I i I I
BSR
2 LOl=2 LDF1 ....
Fig. 2. Migrated seismic reflection sections for segments of Lines 3 and
4 across the deformation front, aligned so that the positions of the main
frontal anticline are coincident. Vertical exaggeration is approximately
1' 1 at 3.0 km s '. LDF - landward dipping fault; SDF - seaward dipping
fault; BSR - bottom simulating reflector. The incoming oceanic sediment
section comprises an upper turbidire section (T) and a lower more
transparent hemipelagic section (H).
TABLE 1. Fault displacements (m) along landward dipping thrust faults
at the deformation front. Displacements are measured near the base of the
turbiditc layer.
Line
Fault 89-03 85-01 89-04 89-05
LDF1 280 400 620 1000
LDF2 1150 1000 1000 950
LDF3 - 50 250 330
On the frontal thrust of Line 89-03 (LDF1 in Figure 3), the shallowest
sediments above the thrust at SP 250 are undisturbed. A similar
relationship is seen on Line 85-01 [Davis and Hyndman, 1989]. One
explanation for this is that the fault is simply in the process of propagat-
ing upwards, and has not yet broken the surface. Deformation at an even
earlier stage than LDF 1 is possibly present over the region 1-2 km farther
seaward on Line 89-03, where some indications are seen of small fault
displacements and folding. The presence of a proto-thrust zone has also
been observed on the Nankai margin [Moore et al., 1990] and the Chile
margin [R. von Huene, pets. comm., 1991], and small-displacement
thrusts in the early stages of upward propagation are noted at the
deformation front off Mexico [Moore and Shipley, 1988]. An alternative
explanation for the undisturbed near-surface sediments on Line 89-03 is
that the motion on the fault has ceased, and thus the most recent activity
does not necessarily occur on the seaward-most fault [Davis and
Hyndman, 1989]. Although beyond the scope of this preliminary paper,
a thorough analysis is required of the variation in fault displacement with
depth and the deformation above the fault termination.
A seaward-dipping backthrust SDF has developed on Line 89-03
(Figure 2, SP 300-330). On the Oregon margin, similiar subsidiary
backthrusts propagate upward from the frontal thrust, and seaward-
dipping primary thrusts also occur [MacKay et al., 1989]. On the
Vancouver Island margin, other seaward dipping thrusts are observed on
the northernmost Line 9 (Figure 3), where a major fault approaches the
seafloor near SP 430, and on 1985 Line 85-04. Davis and Hyndman
[ 1989] noted deformation through the entire sediment section of the latter
line, on two seaward dipping thrust faults located 10-15 km landward of
the main frontal anticline.
On Line 89-03, the seaward-dipping fault is symmetric in profile to
the landward-dipping frontal thrust from which it appears to have
initiated. The two faults form a conjugate pair with a central block of
(s)
LINE 89-09
I 5 km
SP 100 200 300 400 500 600 700
I I I I I I I
800 900 1000 11 O0 650 700
I I I I I
BSR BSR
..;:.., .... ..... ,, .:..... ,..,;.. '... ", ....... "::'" ".:!,,,-..::.. .... :i;,;!i.' ; ......
................... . - --- .... -,_- ... ... -,... ,, ,.. :.: .,,::.;- .... .., .... . ,.,-:_. ---.___. &-?..-_. :
.,...",.,.,;:,,., , ;. ,,.', .. . ,., ..... ,,: ,,..., .... .,. ,. ...,:, ,., ,,
..... .... .... _... ..,.... ...<..- - ,..._, .. .
Fig. 3. Migrated seismic reflection sections for the landward segment of
Line 9. The top of the oceanic crust can be followed for 35 km landward
of the deformation front. Its dip is very small, so the sediment wedge
above the oceanic crust has a very shallow taper. The blowup shows
details of a broad anticline within the sedimentary wedge and a shallow
BSR cutting across lithology. Vertical exaggeration is approximately 1:1
at 3.0 km s '.
upward displaced sediments. On Line 89-04 (Figure 2) and Line 89-05,
the seaward-dipping fault is not seen, but the upper beds, which a
maintain a near-horizontal attitude, are vertically uplifted even more than
on Line 89-03 (Table 1). At the landward boundary of the central block
of sediments on the former two lines, deformation is accomplished by
folding instead of faulting. The different deformation mechanisms imply
that there is along-strike variability in the mechanical strength of the
sediments. If pore-fluid pressure is the key factor in controlling material
strength, then the localized variations in the style of deformation imply
that there may be localized variations of pore-fluid pressure.
(ii) Location of the ddcollement. For two-way times of more than 0.5-
1.0 s beneath the seafloor, the dips on most of the main thrust faults
decrease with depth. When the thrust penetrates to within the lower hemi-
pelagic unit, it is difficult to identify unambiguously the d6collement
where the thrust soles out. The difficulty arises because of the small
vertical offsets expected at low angles and because of the small number
of reflection horizons in the hemipelagic sediments. As well, there is
difficulty in locating the top of oceanic crust because reflections in the
region of the basement are typically low frequency with a period of 100-
150 ms, leading to an uncertainty in location of 200 m or more.
Davis and Hyndman [1989] present indirect arguments that the
d6collement on the Cascadia margin soles out near the top of the oceanic
crust, and with the new multichannel seismic data the arguments appear
to remain valid. First, as dicussed above, the backtilting of sediment
horizons in the hanging wall define the thrust profile. In the special ease
of a circular profile, the rearward limit of tilting defines the intersection
point between the thrust and the basal detachment. As on Line 85-01,
tilting on Line 89-03 (Figure 2) extends at least 8 km landward of the
main thrust LDF2, at which point an extrapolated circular fault profile is
close to the level of the oceanic basement. The second line of evidence
is the sediment budget. As calculated by Davis and Hyndman [1989], the
volume of material within the wedge is comparable to the volume of
material that has entered the subduction zone, implying that most of the
sediments have been off-scraped and accreted.
Other lines, for which deformation is on a horizontal scale of 5 km or
less, do not provide as reliable an estimate for the location of the
d6collement. Definition of the thrust plane often becomes ambiguous
within 0.5 s (--700 m) or more above the basement. In a restricted
sense, the conclusion that thrusts penetrate close to basement applies only
to the second thrust LDF2; for the frontal thrust LDF1 the detachment
level may be as high as the middle of the hemipelagic sediment section.
The d6collement may then step down landwards, as has been noted for
the Nankai margin [Moore et al., 1990], the Aleutian margin [McCarthy
and Scholl, 1985] and the Barbados margin [Westbrook and Smith,
1983]. This then provides the mechanism allowing underthrusting
sediment to be underplated to the accretionary prism.
In many subduction zones the d6collement clearly lies within the
sediments for significant distances landward of the deformation front.
Subducted sediments are traced for 5-7 km from the toe of the Middle
America Trench [Moore et al., 1988], at least 20 km in front of the
Aleutian subduction zone [McCarthy and Scholl, 1985], about 30 km
landward of the Nankai Trench [Moore et al., 1990], and up to 74 km
from the edge of the Barbados Accretionary Complex [Westbrook and
Smith, 1983].
SPENCE ET AL. 261
(iii) Anomalously thin wedge taper, Line 9. The deformation along
Line 9 (Figure 3) provides insight into some of the basic principles
involved in the formation of a basal detachment and accretionary wedge
critical taper [e.g. Davis et al., 1983]. Landward of the frontal anticline,
the water depth is nearly constant for almost 30 km; bathymetry charts
show a lobe-like region of uniform water depth. On the seismic section,
the top of the ocean crust can be followed clearly for 35 km landward of
the deformation front with a very small dip (about 3). Thus, the sediment
wedge here has an unusually small taper. According to critical wedge
theory, such a shallow taper must be associated with extemely low shear
strength of the basal d6co!lement. The main factor controlling sediment
strength is taken to be pore pressure, so that very high pore fluid pressure
is needed for the detachment surface at the oceanic crust to run far
seaward and for the overlying wedge to maintain such a low taper angle.
The presence of prominent reflections near the basement below the
deformation front (SP 300 to SP 500 near 5 s) corroborates this
interpretation. The reflections, which exhibit strong continuity and high
apparent amplitudes, could be produced at the top of a detachment zone
in the sediments just above the basement, where sediment velocities
decrease sharply due to the elevated pore pressures [Calvert and Clowes,
1990; Moore et al., 1990]. This will be studied by further true-amplitude
analysis, to determine if the reflections have reverse polarity as expected
for a velocity decrease and to see if the interface exhibits the appropriate
amplitude variation with incident angle.
On Line 9 the entire thrust package exhibits extensive deformation
involving both faulting and folding. Landward of the frontal anticline,
there is evidence for additional seaward dipping faults (e.g. SP 550) and
broad folds (e.g. as shown in the blowup of Figure 3), which are similar
to structures at the current deformation front on Line 9. Folding at some
locations in the upper sediments suggests that there may have been recent
deformation within the thrust sheet. For example, intense short-wave-
length synclinal folding occurs above 4 s at two locations (SP 550 and SP
630), while broad folds between SP 800 and SP 900 show deformation
at depths greater than 100-200 ms below the seafloor. The implication is
that the upper portion of the thrust sheet has sufficient strength to
transmit stresses up to 30 km from the deformation front, which could be
due to either pore pressure effects or to lithology. The upper section
comprises turbities and so is coatset than the underlying hemipelagic
sediments; the deeper part of the section should thus sustain higher pore
pressures and possess less cohesion.
Structure Beneath the Continental Shelf
The four new. seismic lines across the continental shelf provide
improved images showing the relationship between the accretionary
sedimentary wedge and the overlying terranes which now form part of the
continental framework. Of particular importance, a reflector that may
mark the Crescent thrust or lower boundary of the igneous Crescent
Terrane can be traced down to 6.5 s at SP 1950 on Line 6 (Figure 4).
The terrane acts as the landward dipping backstop to the aeeretionary
prism. This depth is very close to the top of the subducting oceanic crust.
Thus, if this interpretation is correct, very few sediments are available for
deeper subduction or continental underplating.
At the base of the accretionary prism on Line 6, very bright nearly
horizontal reflections are seen near 6 s between SP 1400 and SP 1700.
The reflections are located at the approximate depth expected for the top
of the oceanic crust beneath the outer continental shelf. However, because
of their high amplitude and smooth character, they are unlikely to arise
directly from the lithological contrast between the accreted sediments and
the rough surface of the dipping oceanic crust. The impedance contrast
producing the reflections may arise from a shear zone containing high-
pressure fluids. In a subduction zone environment, high strains exist and
fluid-rich sediments are present, perhaps trapped between the overlying
backstop and the oceanic crust.
The Crescent Terrane comprises Eocene oceanic basalts which have
been penetrated by exploratory wells in Torino Basin [Shouldice, 1971]
and which are exposed' on southern Vancouver Island and the Olympic
Peninsula. Its location beneath the shelf is marked by a magnetic high
[Hyndman et al., 1990] which can be traced parallel to the margin for
over 150 km (Figure 1). The position of the magnetic high corresponds
well with the location of the Crescent Terrane determined from the
seismic profile (Figure 4). The Crescent material appears to extend
somewhat farther seaward than the magnetic high, but the thickness of
these rocks may be small at their seaward edge. In contrast, the main
body of the Crescent Terrane appears to have a thickness of 3.5 s or
more (i.e. about 10 km). The section may have been thickened by
imbrication such as is observed on land within the Crescent voltanits
[Massey, 1986].
Further information on the mechanism of emplacement of the Crescent
Terrane is revealed by the new seismic data. As discussed by Hyndman
et al. [1990], one model of emplacement proposes that the subduction
zone stepped seaward after the complete oceanic lithosphere was faulted
and underthrust, while in a second model a sliver of oceanic crust was
detached from the top of the subdueting plate. The new seismic lines
show evidence for uplift of the seaward portion of the Crescent Terrane,
perhaps associated with thrusting of accretionary wedge sediments. As
typiried by Line 6 (Figure 4), most dips within the terrane are landward,
and a basement ridge at SP 1800 separates Tof'mo Basin into two
sections. On Line 2 (not shown) at the base of the landward section, the
onlap pattern of sediments at the top surface of the Crescent Terrane
clearly indicates a pattern of uplift in the west relative to the east. An
equivalent pattern is not seen at the base of the Tof'mo Basin section
seaward of the basement ridge. Thus, uplift or tilting since the initiation
of sediment deposition has probably occured by imbrication or duplexing
within the Crescent Terrane itself.
Conclusions
With the improved images from our new seismic data, the three-
dimensional variation of deformation along the margin can be determined.
Although it is often difficult to determine unambiguously the depth at
which thrusts sole out, the data suggest that some thrust faults penetrate
close to the top of the subducting ocean crust. Thrust faults and related
anticlines are spaced at roughly 5 km perpendicular to the margin.
Parallel to the margin, there are significant variations both in the amount
of displacement along a fault and in the style of deformation. Subsidiary
backthrusts propagating upward from the frontal thrust are sometimes
observed, but on nearby lines the equivalent faulting is absent and
deformation is accomplished by folding. Such variations imply that there
may be localized variations of pore-fluid pressure. In one region where
the sediment wedge taper is very small, high pore fluid pressures along
the detachment surface are inferred. Strong near-basement reflections
beneath the deformation front are also indicative of high fluid pressure.
The landward extent of accreted sediments is limited by the marine
volcanic Crescent Terrane, which dips toward the continent and can be
traced close to the top of the subdueting oceanic crust. High pore
pressures may exist at the base of the sediment wedge near the backstop,
where bright near-horizontal reflections are observed similar to those seen
beneath the deformation front of the shallow-taper wedge region. The
Crescent Terrane appears to have been tilted landward, perhaps associated
with the accretion of prism sediments. The pattern of tilting suggests that
uplift may have occured by imbrication within the Crescent Terrane.
Acknowledgements. The seismic program was supported by the
Geological Survey of Canada. Data collection was contracted to Digicon
(Canada) Inc., and we gratefully appreciate the efforts of the personnel
onboard the M/V GEOTIDE. Data processing was performed by
Geophoto Ltd of Calgary. Kristin Rohr at PGC provided useful advice
LINE 89-06
300- Magnetic Anomaly (nT)
200-
100 - i 5 km
SP 1400 1500 16,00 1700 1800 190 200_0 21tO0 2200 2300
2 .-..=. x;c'c-='----4c ' ..... - = -'-- .... ;.,...t.,-:. 't.' - - ,,--2 ;*-
7??. ".7.' .V.'.';; . ' U... i!::
I !9i':':/;.:/:.:/:.:.:.:W.474:S:8:8:S ..... , e}vP:::':'2i'';$.d?}?,.. Pacific Rim
:::::::;:7::7::}::.:}:}:::.::;:::?:;?::.:.:Y;V?.2R%V:.5 . v
7',"' ........ , ,,,--:--:,,,,:,,,,:.,,,:,,,,:.,,,:,,,,:,,:',,',',.','.,',' ',' ':,,-,.:-,i'.-.iL-,-,',,,-,.,.i,,, vvvvvvvvvvvvvvvvvvvvvvv"'
................................................... .vvv
Fig. 4. (a) Migrated seismic reflection section for the landward portion
of Line 6, above which is displayed the magnetic anomaly values along
the profile. Vertical exaggeration is approximately 2:1 at 6.0 km s 4. (b)
Interpretative cross section along Line 6, superimposed on a line drawing
of the major reflections from the seismic section.
during the processing of the data. The authors are particularly grateful to
R. von Huene for a substantive critical review. Support for GDS derived
from Grant URF0043000 of the Natural Sciences and Engineering
Research Council. Geological Survey of Canada contribution 57290.
References
Calvert, A.J. and Clowes, R.M., Seismic constraints on fluid flow within
the accretionary complex of western Canada (submitted).
Clowes, R.M., Brandon, M.T., Green, A.G., Yorath, C.J., Sutherland
Brown, Kanasewich, E.R. and Spencer, C., LITHOPROBE-
southern Vancouver Island: Cenozoic subduction complex imaged by
deep seismic reflections, Can. J. Earth $ci., 24, 31-51, 1987.
Davis, E.E. and Hyndman, R.D., Accretion and recent deformation of
sediments along the northern Cascadia subduction zone, Geol. $oc.
Am. Bull., 101, 1465-1480, 1989.
Davis, D., Suppe, J. and Darien, F.A., Mechanics of fold and thrust
belts and accretionary wedges, J. Geophys. Res., 88, 1153-1172,
1983.
Davis, E.E., Hyndman, R.D. and Viilinger, H., Rates of fluid expulsion
across the northern Cascadia Accretionary prism: constraints from
new heat flow and multichannel seismic reflection data, J. Geophys.
Res., 95, 8869-8889, 1990.
Green, A.G., Clowes, R.M., Yorath, C.J., Spencer, C., Kanasewich,
E.R., Brandon and A. Sutherland Brown, Seismic reflection imaging
of the subducting Juan de Fuca plate, Nature, 319, 210-213, 1986.
Green, A.G., Milkereit, B., Mayrand, L., Spencer, C., Kurtz, R. and
Clowes, R.M., LITHOPROBE seismic reflection profiling across
Vancouver Island: results from reprocessing, Geophys. J., 89, 85-90,
1987.
Hyndman, R.D., Yorath, C.J., Clowes, R.M. and Davis, E.E., The
northern Cascadia subduction zone at Vancouver Island: Structure
and tectonic history, Can. J. Earth Sci., 27, 313-329, 1990.
Hyndman, R.D. and Davis, E.E., A mechanism for the formation of
methane hydrate and sea floor bottom simulating reflectors by
vertical fluid expulsion, J. Geophys. Res. (in press).
MacKay, M., Cochrane, G., Moore, G.F. and Kulm, L.D., High
resolution seismic survey of the Oregon accretionary prism, Trans.
AGU, 70, 1345, 1989.
SPENCE ET AL. 263
Massey, N.W.D., The Metchosin Igneous Complex, southern Vancouver
Island; ophiolite stratigraphy developed in an emergent island setting,
Geology, 14, 602-605, 1986.
McCarthy, J. and Scholl, D.W., Mechanisms of subduction accretion
along the central Aleutian trench, Bull. Geol. $oc. Am., 96, 691-701,
1985.
Moore, G.F. and Shipley, T.H., Mechanics of sediment accretion in the
Middle America Trench off Mexico, J. Geophys. Res., 93, 8911-
8927, 1988.
Moore, G.F., Shipley, T.H., Stoffa, P.L., Karig, D.E., Taira, A.,
Kuramoto, S., Tokuyama, H., and Suyehiro, K., Structure of the
Nankai Trough accretionary zone from multichannel seismic
reflection data, J. Geophys. Res., 95, 8753-8765, 1990.
Shouldice, D.H., Geology of the western Canadian continental shelf,
Bull. Can. Petr. Geol., 19, 405-436, 1971.
Waldron, D.A., Clowes, R.M. and White, D.J., Seismic structure of a
subdueting oceanic plate off western Canada, in Studies of laterally
heterogeneous structures using seismic refraction and reflection data,
edited by A.G. Green, Geol. Surv. Can., Paper 89-13, 1989.
Westbrook, G.K. and Smith, M.J., Long d6collements and mud
volcanoes: evidence from the Barbados Ridge Complex for the role
of high pore-fluid pressure in the development of an accretionary
complex, Geology, II, 279-283, 1983.
FIRST DEEP SEISMIC REFLECTION TRANSECT FROM THE GULF OF LIONS TO SARDINIA
(ECORS-CROP PROFILES IN WESTERN MEDITERRANEAN)
B de Voogd 1, R Nicolich 2' J L Olivet 3' F Fanucci 4' . , . . , . , J. Burrus 5', A. Mauffret 11
G Pascal 12 A Argnani TM J M Auzende 3 M Bernabini 7 C Bois 5 L Carmignani 8 ! ! ! ! !
A Fabbri 6, I Finetti 2 A Galdeano 9 C Y Gorini 6,15 p Labaume 10 D Lajat 6 ! ! ! !
P Patriat 9 B Pinet 5 J Ravat 6 F Ricci Lucchi 13 and S Vernassa 6
! ! ! !
Abstract. A 500 km long ECORS-CROP volcanic origin for part of this relief
profile provides the first continuous cannot be excluded. In the deep water
transect across the Western Mediterranean domain, lateral variations in basement
Basin, between the Gulf of Lions and West topography and in sediment thicknesses are
Sardinian margins. In the Gulf of Lions, observed. The central part of the basin is
a 70 km wide zone where crustal thickness a structural low, characterized by
is reduced from about 25 km to less than significant structures within basement
5 km separates the shelf from a deep water coinciding with a central magnetic
domain where the crust is 4-5 km thick. anomaly. The structures of the eastern
The zone where most of the crustal part of the basin are taken as evidence
thinning occurs is characterized by for post-opening tectonic activity,
prominent lower crustal and possibly Moho already documented in western Sardinia.
reflect ions, observed nowhere else on the
margin. Half grabens bounded by listric
faults are observed across the Provencal
shelf. At the lower part of the slope,
tilted blocks are visible above a
continuous, subhorizontal reflector,
suggesting detachment faulting, though a
* ECORS-CROP chief scientists
1 Ecole Normale Suprieure, Paris,
France;
2 University of Trieste, Italy;
3 IFREMER, Brest, France;
4 University of Urbino, Italy;
5 Institut Francais du Ptrole, Rueil-
Malmaison, France;
6 Elf Aquitaine, France;
7 University of Rome, Italy;
8 University of Cagliary, Italy;
9 IPG Paris, France;
10 University of Montpellier, France;
11 University of Paris 6, France;
12 UBO, Brest, France;
13 University of Bologna, Italy;
14 CNR, Bologna, Italy;
15 University P. Sabatier, Toulouse,
France.
Continental Lithosphere: Dp Seismic Reflections
Geodynamics 22
Introduction
This paper presents preliminary results
of a survey carried out in the Western
Mediterranean basin (Figure 1) within a
French (ECORS)-Italian (CROP) joint
project, providing the first complete
crustal transect between the Gulf of Lions
and the western Sardinian rifted,
presumably conjugated, margins. The main
objectives of the project were to
determine the formation mechanism and
evolutionary history of the Provencal
Basin [ECORS, 1988]. First results of the
complete ECORS survey (lines NS, NW, and
ZZ; Figure 1) are given by Burrus et al.
(1991). This paper focuses on the slope
areas and the central deep basin.
Geologic Setting
The Provencal Basin (Figure 1) opened
as the result of the southeastward
drifting of Corsica and Sardinia [Le
Pichon et al., 1971a; Montigny et al.,
1981] in Aquitano-Burdigalian (Early
Miocene) time, following an Oligo-Miocene
rifting event well known in Western Europe
and in Sardinia. Previous data have
already enabled us to characterize the
evolution of the basin (review by Burrus
[1989]). Refraction experiments have
265
266 ECORS-CROP IN WESTERN MEDITERRANEAN
44
42
4O
MASSIF
CENTRAL
A A A
207 ^^;
I A A A A A A
A A A A A A A
A A A A A '
A A A A A -
A A A
A A A
A
$% 0o
4 6 8 10
Fig. 1. Location map of the three ECORS lines in the Gulf of Lions (NW, NS
and ZZ), the conjugated CROP line, the ESP's used here, and OGS line MS39.
Only ECORS NW and CROP are discussed in this paper. These 2 profiles form
a transect across the Provencal Basin. Contour lines are water depth,
every 400 m; major Oligo-Miocene basins, onshore and offshore, are
stippled; Sirocco (Si) and GLP2 are 2 wells located near ECORS NW; M is
the city of Montpellier; major faults are indicated' NPF is the North-
Pyrenean Fault zone, CF is the Cvennes fault, IF the Nimes fault, DF the
Durance fault. The trace and possible continuation of the NPF are
hatched. The salt symbols (open triangles) in the central basin indicate
the area of large salt diapirs (domains n3 and 4 discussed in text and
labelled on next figures).
established that crustal thickness has and Rossignol, 1977], and several drifting
been reduced from 20-25 km on the models have been proposed [Auzende et al.,
Provencal shelf [Sapin and Hirn, 1974] 1973; Bayer et al., 1973; Burrus, 1984].
down to 5 km at the base of the slope and In the Gulf of Lions, the Oligo-
in the abyssal plain [Le Douaran et al., Miocene rifting occurred in an area
1984]. In the Gulf of Lions, crustal previously affected by four main tectonic
thickness could have been greater than 25 episodes. The continental "basement" is
km at the end of the Eocene compression. formed by rocks which underwent Hercynian
In Sardinia, the Moho rapidly deepens from orogeny [Arthaud and Matte, 1977]. Late
20-25 km depth at the margin to 30 km
beneath the island [Egger et al., 1988].
Oceanic crust has been emplaced in the
center of the basin, although the
processes which led to this oceanization
are disputed' interpretation of the
magnetic patterns is ambiguous [Galdeano
Hercynian faulting is characterized by two
families of strike-slip faults, namely NE-
SW trending s inistral and E-W trending
dextral faults. Early Mesozoic
deformation consisted of block faulting,
subsidence, and basin formation [e.g.,
Curnelle and Dubois, 1986]. During Albo-
DE VOOGD ET AL. 267
Aptian time, a major reorganization of the deep basin and by a coeval Late Miocene
paleogeography around the Gulf of Lions
occurred, probably due to left-lateral
motion (about 400 km) along the North
Pyrenean Fault Zone, resulting from the
opening of the Gulf of Biscay and the
eastward drift of the Iberia plate [Le
Pichon et al., 1971b; Olivet et al.,
1984].
The Gulf of Lions crosscuts the
Pyreneo-Provencal fold belt, formed
erosional surface on the upper margin.
P lio-Quaternary shales seal this Tertiary
sequence. In the central basin, the salt
layer shows prominent diapiric structures
[Le Cann, 1987]. The prerift substratum is
represented by a Mesozoic carbonate
platform onshore and along the basin edge
[Arthaud and Seguret, 1981; Curnelle and
Dubois, 1986]. In the center of the Gulf
of Lions, the Tertiary sediments directly
between the end of the Cretaceous and the overlie Paleozoic granitic or metamorphic
Early Oligocene [Arthaud and Mattauer, rocks [Curnelle and Dubois, 1986].
1972]. This implies that collapsed Jurassic carbonates have been mapped in
Pyrenean structures exist beneath the the Nurra region [Cocozza et al., 1974].
Gulf, as evidenced by northward thrusting Little is known about the extension of the
encountered in wells and industrial Mesozoic cover on both margins.
seismic profiles [Arthaud and Seguret,
1981; Mauffret and Gennessaux, 1989]. The Data Acquisition and Processing
Oligocene rifting resulted in the
formation of a central graben in the Gulf The 570 km of ECORS seismic reflection
of Lions; its trend (typically N30) is data were acquired by CGG, and the 200 km
probably inherited from the Hercynian and of CROP data by OGS, both in October 1988,
early Mesozoic structures. The Pyrenean using standard industry equipment. A
and pre-Pyrenean structures observed on large airgun array (totalling 110 litre
the ECORS data and the question of their for ECORS, and 82 litre for CROP) was
Oligocene reactivation are discussed by recorded by a 3000 m long, 120 channels
Burrus et al. [1991]. streamer. The distance between shots was
In Sardinia, the Mesozoic and Pyrenean 50 m, resulting in 3000% coverage.
events are of minor importance. The The ECORS processing sequence, carried
Hercynian basement, as imaged on previous out by CGG, included' resampling at 8 ms,
surveys (OGS lines) generally exhibits a deconvolution, F-K filtering (slope 2000
landward-dipping fabric. Evidence is m/s), addition of every two traces,
found for the existence of a carbonate multiple attenuation, dynamic corrections
platform in the Mesozoic [Cocozza, et al., and stack. Water-bottom multiples were
1974]. The basement is dissected by two successfully removed in the deep water
major Tertiary grabens, a central north- areas, revealing the top of basement,
south graben (Oligo-Miocene) and the N130 initially hidden by the first multiple.
Campidano graben (Miocene to present).
Widespread Oligo-Miocene calc-alkaline
volcanism [Coulon, 1977] was
contemporaneous with the central graben
formation. After a break in volcanic
activity, spanning from middle to late
Miocene, more alkalic volcanism resumed in
This was achieved by raising all amplitude
values to a power less than 1, removing
multiples in the F-K domain, and restoring
amplitudes. A post-stack F-K migration was
performed on the northern half of the NS
and NW lines.
The CROP processing sequence, carried
Plio-Pleistocene, [Cherchi and Montadert, out by OGS, included' resampling at 8 ms,
1982]. In contrast with the Gulf of addition of every two traces, decon-
Lions, the system of grabens is here volution, dip-move-out, and F-K filtering
mainly onshore. The CROP profile reaches for multiple removal.
the northwestern fragment of the Sardinian General line-drawings of the ECORS-CROP
block, in an area of steep slope, similar NW-SE transect are presented in Figure 2.
to the Maures margin. Further south, the Complete line-drawings of the ECORS survey
morphology of the Sardinian margin is more can be found in Burrus et al. [1991].
like that of the Gulf of Lions [Thomas et
al., 1988] .
The stratigraphy of the Provencal Basin
is known from industrial seismic lines
tied to wells [Mauffret et al., 1973;
Finetti and Morelli, 1973; Cravatte et
al., 1974; Rhault et al., 1985]. The
Tertiary sedimentary infill comprises
Oligocene clastics confined in the
The Gulf of Lions Margin
Beneath the shelf and the slope,
reflections are observed down to the Moho.
The Plio-Quaternary sequence shows
numerous structures such as progradation
sets, growth faults and salt-tectonic
related features. On the upper margin, an
grabens, pre-Messinian sediments topped by irregular erosional surface is observed
the Messinian evaporitic sequence in the between SP 2200-3200 (Figure 2a) . In the
NW
5
r
Z
o
lO
Vg PCg ECORS NS I G
.o S,ROO o 170 ,, ?
MS39
SP n 500 1000
0 ' '
r
Z
o
m 10
2000 2500 30100
I I
3?0
20 km
I
40100 0
Fig. 2. Interpreted line-drawings of the NW-SE ECORS-CROP transect (see
Figure 1 for location). Circled numbers on top refer to the domains
defined in the text. Approximate positions of ESP's center point and
crossing with other seismic lines are indicated on top of each drawing. No
vertical exaggeration for velocity of 5 km/s.
2a. migrated portion of line ECORS NW; Vg and PCg are the Vistrenque and
Petite Camargue grabens.
2b. continuation of ECORS NW (unmigrated); here, rather than a full line-
drawing of the sedimentary units, major stratigraphic horizons are
indicated' P is intra-Pliocene, M1 top of evaporites, M2 top of salt, M3
base of salt, M4 base of evaporites, X and Y are intraMiocene horizons, B
is top of basement.
DE VOOGD ET AL. 269
deep basin, it is contemporaneous with the only clear evidence of listric fault-
evaporitic sequence observed south of SP controlled synrift deposition. Tilted
600 (Figure 2a) between horizons labeled synrift deposits, fanning upwards toward
M1 and M4 on Figure 2b. Evidence of the subhorizontal postrift sequence are
synrift deposition is found in most controlled by two major normal faults of
depressions. These deposits (up to 2 s opposite vergence. A narrow, elongated
thick in the troughs across the shelf) graben, visible just north of the
thin or disappear on top of the blocks and intersection with ECORS NS line (around
towards the deep basin. For that reason, SP1200 on Figure 2a), is possibly bounded
and because of the lack of a clear angular by steep normal faults. Here the flat
unconformity at the top of synrift graben infilling does not suggest block tilting.
filling, it is difficult to follow a A basement high (SP 1150, Figures 2a, 3)
break-up unconformity, at least in the is located south of this graben. Farther
shelf and slope areas. south, irregular basement topography
Few reflections are observed in the (SP1000 to 400, Figures 2a, 3) is
upper crust, except for several packages underlain by a non-deformed, gently
of south to southeast dipping reflectors. southward-dipping reflector (labelled R on
These dipping reflectors are better imaged Figure 3). Reflector R may be a flat
on ECORS line NS. They may be related to detachment surface, and the overlying
the Pyreneo-Provenal compressional irregular topography the envelop of
structures recognized onshore and in deformed tilted blocks.
wells, or may represent the edge of a
Mesozoic basin. Tying the ECORS survey to The Central basin and the Sardinia margin
industry profiles is necessary to discuss
the possible reactivation of these pre- The deep basin, defined as the area of
existing structures during rifting [Burrus very thin (5 km) crust between the Gulf of
et al., 1991] .
The lower crust is reflective beneath
the shelf, but in comparison with other
deep seismic data, these lower crustal
reflections are poorly organized, often
not very strong and superimposed on
numerous diffraction hyperbolae. Moho,
known from refraction data [Sapin and
Hirn, 1974] to be at 9 s (20-25 km),
coincides with the base of the lower
Lions and the Sardinian continental
slopes, can be divided into several
domains (labeled as circled numbers on
Figure 2) on the basis of stratigraphy,
basement topography, and salt
mobilization. Some of these domain
boundaries coincide with changes in the
pattern of magnetic anomalies.
The sedimentary cover of the deep basin
is composed of three main sequences
crustal reflections, but cannot be [Mauffret et al., 1973]. At the top, 2.5
identified as a continuous reflector below km of Plio-Quaternary shales within which
the shelf. The diffraction hyperbolae a Pliocene unconformity (labeled P,
drawn below the Moho (SP 1000-1400 on Figures 2, 4) is known from previous
Figure 2a) correspond to a migration surveys and can be traced on the deepest
velocity of 1500 m/s, implying side portion of the CROP and possibly of the
arrivals. A drastic reduction in crustal ECORS profiles. Underneath, the Messinian
thickness is observed on a 70 km wide zone evaporites can be divided into 3 units' an
(SP 400-1300, Figures 2a, 3) beneath the upper evaporitic sequence (between labels
lower continental slope. Here, a prominent M1 and M2), a salt layer (bounded by
and continuous series of reflections marks horizons M2 and M3), and a lower
the base of the crust as defined by ESP evaporitic sequence (between M3 and M4) .
(Expanding Spread Profile) data [Le These evaporites were deposited over a 3
Douaran et al., 1984; Pascal et al., km thick Early and Middle Miocene series.
1991]. Two prominent markers are indicated on the
Upper crustal extensional structures figures (labeled X and Y; Figures 2, 4)
appear diverse. The grabens of Vistrenque though their ages are uncertain. The top
and Petite Camargue (Figure 2a) show the of the prerift or oceanic basement is well
2c. CROP line (unmigrated); same stratigraphic horizons as in 2b- P is
intra-Pliocene, M1 top of evaporites, M2 top of salt, M3 base of salt, M4
base of evaporites, B top of basement.
2d. CROP line (unmigrated); same stratigraphic horizons as in 2b and 2c
though not all of these could be traced with confidence all the way to the
foot of the slope; again, M1 is top of evaporites, M4 base of evaporites,
and B top of basement.
ESP 203500 t 0
ESP202 SP n'1000 4' !
0 ' ,1' I I i ' ] I
I0
Fig. 3. Migrated portion of line ECORS NW across the zone where crustal
thickness is reduced, showing strong lower crustal reflections; The
position of Moho (labelled M) is from Pascal et al. [1991]. Note the
prominent reflections just above Moho. Between SP1000 and 400, the
basement topography is irregular, although the underlying reflector
labelled R shows no irregularity.
ESP 204 ESP 205
ESP 206
2.0
23
2.7
3.2
lO
lO KM
i I
Fig. 4. Portion of unmigrated line ECORS NW across the transition between
domain nl and 2. Crustal thickness is similar (about 5 km) on both side.
Seismic velocities are taken from Pascal et al. [1991]; stratigraphic
horizons are labelled as in Figure 2b' P is intra-Pliocene, M3 is base of
salt, X and Y are intra-Miocene horizons.
DE VOOGD ET AL. 271
defined (Figures 2b, c), but Moho Another major change occurs around ESP
reflections are difficult to distinguish 207, at the edge of the area of salt
from multiples. The different domains are diapirism (SP 550, Figue 2b). Domain n3
now described, going from northwest to is the central part of the basin, and
southeast. coincides with the large salt dome area
On the Gulf of Lions margin, the [Le Cann, 1987]. Here, the Tertiary
transition between the slope and the sequences thicken toward the center,
central basin may be associated with a suggesting a depressed area. In this
crustal discontinuity (SP 350 on Figures domain, the position of top of basement is
2a and 3). Between the slope and ESP 205 debatable, since the presence of numerous
(domain nl), the numerous reflectors salt diapirs has somewhat deteriorated the
present around 6.5-7 s suggest the seismic data.
presence of sedimentary infilling on a Going east, the large salt domes are
mildly structured basement. Southeast of replaced by smaller diapirs (domain 4;
ESP 205, top of basement is smoother and Figures 2c,5) . Top of basement is well
about 0.5 s shallower; the previous imaged and shallower than in domain 3.
sedimentary infilling is not seen, and the The eastern boundary of domain 4
seismic expression of the basement is corresponds to a structure (SP 1000,
different (domain n2). Figures 2c,5) which affects the basement
The transition between the zone of and the pre-Messinian sedimentary
structured, thin, presumably continental sequence.
basement (domain n 1) and the zone of East of this structure (domain n5;
smoother basement (domain n 2) is Figures 2d, 5), the pre-Messinian sequence
associated with an apparent change in the appears thicker than to the northwest, and
sense of basement flexure around SP 1900 has a different seismic character;
(Figures 2b, 4). In domain n 1 the organized reflections are often disrupted
subsiding crust tends to deepen basinward, by chaotic patches.
whereas in domain n 2 top of basement In domain n6 (Figure 2d), the
has a gentle dip toward the ProvenGal evaporites become significantly shallower
slope. Sedimentary sequences beneath and thinner; they probably disappear
the reflector labelled Y (Figure 4) thin around SP 3600. The top of basement is
out to the east, or possibly onlap on the irregular and becomes shallower as we
basement, suggesting lateral changes in reach the continental slope of Sardinia.
basement structure towards the east. Several southeast-dipping reflections are
, SP-8oo lOOO
3__' .................. ! ...... ]
M1
1300
I_
M1
M2
i...-- M3
M4
z
o
10
10 KM
10
Fig. 5. Portion of unmigrated CROP line showing the end of the salt dome
area (transition between domains n4 and 5). Major stratigraphic horizons
are labelled as in Figure 2' P is intra-Pliocene, M1 top of evaporites, M2
top of salt, M3 base of salt, M4 base of evaporites, and B is top of
basement. The irregular basement topography indicated on this
interpretation (dotted line) is most likely associated with the presence of
volcanics, labeled vx.
observed within basement. This domain n6 juxtaposition of salt diapirs and low
constitutes a basement high, the Sardinian velocity P lio-Quaternary sediments (4.4
'terrasse', considered as part of the vs. 2.3 km/s) have severely distorted
margin. The Western Sardinian margin is a seismic arrivals.
starved margin, with a slope much steeper
than in the Gulf of Lions. In the slope
area, half-grabens and tilted blocks are
observed (SP3500 on Figure 2d) .
Discuss ion-Conclusions
The interpretation of the ECORS-CROP
profile has been tied to OGS line MS39
(Figures 1, 6). The base of the Messinian
evaporites (labeled M4) is a fairly
prominent and continuous marker that can
be traced all across the basin. The
sagging of M4 and thickening of the
This preliminary interpretation of the Messinian sequence indicate that the
ECORS-CROP profile from the Gulf of Lions central basin may be a structural low
to the Sardinia margins permits us to (domain 3), though the top of the basement
highlight some global features of the
Provencal basin, as observed along this
transect, although it is too early to put
forward any definite basin formation
mechanism.
Basement topography
Top of basement can be identified on
most of the transect. The transect follows
ESP's 201 to 208 [Le Douaran et al.,
1984], which have been reinterpreted
[Pascal et al., 1991] to constrain the
velocity structure of the margin down to
the Moho. The existence of a paleo-
is difficult to identify under the salt
domes area. This central depression is
located between two domains (domains n2
and 4) where basement is slightly
elevated. This interpretation is
supported by possible reflections from the
top of the basement on either side of this
low (labeled B on left and right panel of
Figure 6).
Post-opening tectonic activity
The basement topography and lateral
variations in sediment thicknesses are
oceanic ridge as initially suggested by a taken as evidence for post-opening
basement high interpreted on ESP 208 by Le tectonic activity in the eastern part of
Douaran et al. [1984] is not supported by the Provencal Basin. On the Sardinian
the ECORS-CROP transect. Indeed, ESP 208 margin, this activity, studied with high
is located in the large salt dome area, resolution seismic profiles, is
and it seems that lateral velocity illustrated in the neo-tectonic map of
variations resulting from the Italy [Fanucci and Nicolich, 1984; AA.VV.,
SP-3100 2900 2700 2500 2200 21 O0
3.j i i_ 3 I I__ .. I !
. .... .. , ',.,' :, --. ...... ,
..... . ,
03 9- 9
Fig. 6. Selected portions of OGS line MS39, not migrated; see location of
MS39 in Figure 1. Same stratigraphic horizons as in Figures 2b, c,d- M4 is
the base of the Messinian evaporites; the two dashed lines drawn above M4
indicate the salt unit. Note the deeper position of M4 in the central part
by comparison with the east and west panels. Interpretation of ESP 215 is
taken from Pascal et al. [1991]. Top of basement is labelled B.
DE VOOGD ET AL. 273
1988]. The thermal asymmetry indicated by bathymetric contours on Figure 1) . This
the increase in heat flow from the Gulf of has to be clarified before discussing
Lions to the northwest Sardinian margin conjugate margins formation, but it is
may be related to postrift magmatic beyond the scope of this short paper.
activity [Burrus and Foucher, 1986].
Asymmetry in the tectonic heritage and in Acknowledgments. We thank the captain
basin formation may also affect the post- and crew of the Stormy (CGG) and of the
opening evolution. The axial part of the Explora (OGS), and B. Damotte (IFP) . We
basin, characterized by large magnetic are grateful to Elf-Aquitaine (France) who
anomalies, is significantly closer to the allowed access to numerous high quality
Sardinian continental slope. Oligo-Miocene seismic and well data in the Gulf of
volcanism is also more important on the
Sardinian side [Coulon, 1977].
Basin formation
'Typical' oceanic crust cannot be
readily recognized along the transect,
though the oceanic nature of the central
basin is not questioned. The relief
characteristics of a spreading ridge are
not observed, though features of uncertain
nature are imaged within basement at the
northwestern end of the CROP line (domain
n4 Figure 2c) The salt dome area t '
(domain n3 and beginning of domain n4)
coincides with the zone of large magnetic
anomalies of the central basin [Galdeano
and Rossignol, 1977; Burrus, 1984] .
Lions, and to OGS for permission to use
their seismic profiles. An anonymous
reviewer provided helpful comments. ECORS
was funded by Elf-Aquitaine, IFP, IFREMER,
and INSU-CNRS; this CROP survey was funded
by CNR.
References
AA. VV., Neotectonic map of Italy (sheet
3), C.N.R. Progetto Finalizzato
Geodinami ca-subprogetto neotettonica-
Roma, 1988.
Arthaud, F., and Mattauer, M.,
Presentation d'une hypothse sur la
gnse de la virgation pyrnenne du
Languedoc et sur la structure profonde
du Golfe du Lion, Comptes Rendus
The extension of the Gulf of Lions Aca. Sci., 274, 524-527, 1972.
margin has been accommodated by a variety Arthaud, F., and Matte, P., Dtermination
of structures, probably strongly dependent de la position initiale de la Corse et
on the prerift structural heritage. It is
difficult to assess whether the remarkable
crustal thinning observed below the slope
(Figure 2a) affects equally the upper
(transparent) and the lower (more
reflective) crust. The lower crustal
reflections observed in this zone (SP 300-
1000, Figure 2a) may be 1 or 2 km above
Moho [Pascal et al., 1991].
The Sardinian margin is also
characterized by an area of thinned
continental crust, covered by a thick pre-
Messinian series, probably including
volcanics. In contrast to the Gulf of
Lions, no bright lower crustal or Moho
reflection is observed beneath the slope.
The prominent series of northwest-dipping
reflectors imaged in the lower crust
beneath the slope in the Gulf of Lions
might suggest the presence of a
lithospheric shear zone [Wernicke, 1985]
along which Sardinia and Provence moved
before the drifting phase. Such a model
is however not clearly substantiated by
the data available. Models involving both
pure and simple shear, as well as crust-
mantle interactions may be better
candidates. It should also be noted that
these two margins may not be strictly
conjugate. The Gulf of Lions bears more
similarity to the wide Sardinian shelf
south of the CROP profile (indicated by
de la Sardaigne la fin de l'orognse
hercynienne grace aux marqueurs
gologiques ant-msozoques,
Bull.Soc.Geol.F., 7, 833-840, 1977.
Arthaud, F., and Sguret, M., Les
structures Pyrnennes du Languedoc et
du Golfe du Lion (Sud de la France),
Bull.Soc. Geol.F., 7, 51-63, 1981.
Auzende, J.M., Bonnin, J., and Olivet,
J.L., The origin of the Western
Mediterranean basin, J.Geo1.Soc.London,
129, 607-620, 1973.
Bayer, R., Le Moul, J.L., and Le Pichon,
X., Magnetic anomaly pattern in the
Western Mediterranean, Earth
Planet.Sci.Let., 12, 168-176, 1973.
Burrus, J., Contribution to a geodynamic
synthesis of the Provencal basin
(northwestern Mediterranean), Marine
Geol., 55, 247-269, 1984.
Burrus, J., Review of geodynamic models
for extensional basins and the example
of the Gulf of Lions (NW Mediterranean)'
the paradox of stretching,
Bull.Soc. Geol.F., 8, 377-393, 1989.
Burrus, J., and Foucher, J.P.,
Contribution to the thermal regime of
the Provencal basin based on Flumed heat
flow surveys and previous
investigations, Tectonophysics, 128,
303-354, 1986.
Burrus, J., et al., ECORS deep seismic
reflection study in the Gulf of Lions
(NW Mediterranean)' first results,
Tectonics, submitted.
Cherchi, A., and Montadert, L., Oligo-
Miocene rift of Sardinia and the early
history of the Western Mediterranean
basin, Nature, 298, 736-739, 1982.
Cocozza, T., Jacobacci, A., Nardi, R., and
Salvadori, I., Schema stratifico-
strutturale del Massicio Sardo-Corso e
minerogenesi della Sardegna,
Mem. Soc.Geol. Ital., 13, 85-186, 1974.
Coulon, C., Le volcanisme calco-alcalin
cnozoique de Sardaigne (Italie) .
Ptrographie, gochimie et gnse des
laves andsitiques et ignimbrites'
signification godynamique. Thse
,Univ. Aix-Marseille,358 pp.,1977.
Cravatte, J., Dufaure, P., Prim, M. and
Roubaix, S., Les sondages du Golfe du
Lion' stratigraphie et sdimentologie,
Notes et Mmoire delaCFP,2,209-274,1974.
Curnelle, R., and Dubois, P., Evolution
msozoque des grands bassins
sdimentaires francais; bassin de Paris,
d'Aquitaine et du Sud-Est,
Bull.Soc.Geol.F., 8, 526-546, 1986.
ECORS, Profil ECORS, Golfe du Lion'
Rapport d'implantation, Institut
Francais du Ptrole ref. 35941-1, 1988.
Egger, A., Demartin, M., Ansorge, J.,
Banda, E., and Maistrello, M., The gross
structure of the crust under Corsica and
Sardinia, .Tectonophysics, 150 (3), 363-
389, 1988.
Fanucci, F., and Nicolich, R., I1 mar
Ligure- nuove acquisizioni sulla natura,
genesi ed evoluzione di un "bacino
marginale",M.m.Soc.Geol. It. ,27,97-110,1984.
Finetti, I., and Morelli, C.,
Esplorazione geofisica dell'area
circostance il blocco sardo-corso,
Rend. $emin.Fa. Sci .Univ. Cagliari,
Suppl., 43, 213-238, 1973.
Galdeano, A., and Rossignol, J.C.,
Assemblage & altitude constante des
cartes d'anomalies magntiques couvrant
l'ensemble du bassin occidental de la
Mditerrane, Bull.Soc.Geol.F., 19, 461-
468, 1977 .
Le Cann, C., Le diapirisme dans le bassin
liguro-provencal (Mditerrane
occidentale). Relations avec la
tectonique et la sdimentation.
Implications godynamiques, Thse, Univ.
Bretagne Occ., 1987.
Le Douaran, S., Burrus, J., and Avedik,
F., Deep structure of the north-western
Mediterranean basin' results of a two-

ship seismic survey, Marine Geology, 55,
325-345, 1984.
Le Pichon, X., Pautot, G., Auzende, J.M.,
and Olivet, J.L., La Mditerrane
occidentale depuis l'Oligocne. Schema
d'volution, Earth Planet.Sci.Let., 13,
145-152, 1971a.
Le Pichon, X., Bonnin, J., Francheteau,
J., and Sibuet, J.C., Une hypothse
d'volution tectonique du golfe de
Gascogne, in' Histoire structurale du
golfe de Gascogne, Technip, Paris,
VI. 11.l-VI. 11.44, 1971b.
Mauffret., A., Fail., J.P., Montadert, L.,
Sancho, J., and Winnock, E.,
Northwestern Mediterranean sedimentary
basin from seismic reflection profile,
AAP Bull., 57, 2245-2262, 1973.
Mauffret, A., and Gennessaux, M.,
Compression, dcrochement et distension
sur le pourtour mditerranen nord-
occide.cal, Compt$ Rendus Acad. Sci,,
308, 961-967, 1989.
Montigny, R., Edel, J.B., and Thuizat, R.,
Oligo-Miocene rotation of Sardinia' K-Ar
ages and paleomagnetic data of Tertiary
volcanics, Earth Planet.Sci.Let., 54,
261-271, 1981.
Olivet, J.L., Beuzart, P., Auzende, J.M.,
and Bonnin, J., Cinmatique de
l'Atlantique Nord et Central, Cnexo, 54,
107 pp., 1984.
Pascal, G., et al., A revised
interpretation of ESP's in the Gulf of
Lions (NW Mediterranean) based on Tau-P
transformation, Geoph. J., submitted.
Rhault, J.P., Boillot, G., and Mauffret,
A., The Western Mediterranean Basin, in
Geological Evolution of the
Mediterranean Basin, edited by D.J.
Stanley, and F.C. Wezel, pp. 101-129,
Springer-Verlag, NewYork, 1985.
Sapin, M., and Hirn, A., Results of
explosion seismology in the southern
Rhone valley, Annales de Gophysique,
30, 181-202, 1974.
Thomas, B., Gennesseaux, M., and Lecca,
L., La structure de la marge occidentale
de la Sardaigne et la fragmentation de
l'ile au Cnozoique, Marine Geology, 83,
31-41, 1988.
Wernicke, B., Uniform-sense normal simple
shear of the continental lithosphere,
Can. J. Earth Sci., 22, 108-125, 1985.
STRUCTURE AND EVOLUTION OF ADJOINING SEGMENTS OF THE
WEST AFRICAN MARGIN DETERMINED FROM DEEP SEISMIC PROFILING
Jacques Wannesson
Institut Francais du Ptrole, BP 311, 92506 Rueil-Malmaison Cedex, France
Jean-Claude Icart and Jean Ravat
Elf-Aquitaine, Tour Elf, 92078 Paris La Dfense, France
Abstract. Previous geological and geophysical iii)-a late drift stage ranging in age from
work has established that the West African conti- Albian to Holocene, which corresponds to a gene-
nental margin between the Douala Basin (Cameroon) rally westward prograding sedimentary system.
and the mouth of the Congo River is divided into Shallow facies are dominant in the eastern part of
three segments separated by major transverse the basin, starting with thick Albian to Turonian
faults. Each segment has its own specific geody- carbonates, whereas basin facies with dominant
namic evolution, whose mechanisms were previously shales and turbidites are present in the western
hard to assess. A deep seismic survey in 1988 re- part. A major hiatus and subsequent erosion sepa-
vealed the crustal structure of the central and rates the Cretaceous-Eocene series from the over-
southern segments as well as of the faulted zone lying Neogene sediments.
separating them. Differences in the rate of crus- The transition from platform to basin facies
tal synrift thinning were revealed between these during most of the margin history, with two exam-
two segments, which can be explained by comparison ples during the Albian and the Senonian (figure
with the Brazilian conjugate margin during the 3), is shifted abruptly seaward by some 50 km
pre-opening stage. From this we consider that the south of Port-Gentil. This suggests the presence
fault zones separating the segments are transfer of a transverse discontinuity, splitting the
zones involving no intraplate discontinuity. margin into segments with different geological
evolutions. Between Cameroon and the mouth of the
Geological Setting Congo river, three main segments have been iden-
tified (figure 4), namely the Cameroon, the North
The South Atlantic Ocean has been formed by the Gabon and the South Gabon segments.
separation of Africa and South America since the
Early Cretaceous (Rabinowitz and La Brecque, 1979,
Emery and Uchupi, 1984). All through that time,
typical passive margins were built up on both
sides, which are documented by a great deal of
both academic and industrial data. The discovery
of major oil provinces on these margins has con-
tributed to a large extent to what is now known
about these margins.
In the Gulf of Guinea, Gabon (figure 1) is
one of these major oil provinces with initial
recoverable reserves estimated to be about two
billion barrels of oil. The geology of its Creta-
ceous and Tertiary marginal basin has been
divided into three main stages (Teisserenc and
Villemin, 1989, figure 2):
i)-a rift stage of Neocomian to Barremian age
(130 to 112 Ma) characterized by three fluviatile
and lacustrine sequences. Sediments are mainly
sands and shales including organic black shales.
Thicknesses are highly variable due to the block
faulting of the rift which was active throughout
this period.
Deep Seismic Survey
In 1988, a deep seismic survey was carried out
by the Institut Francais du Ptrole and Elf-Gabon
to investigate, in greater detail down to the
Moho, the structure of the North Gabon and South
Gabon segments and of the intervening disconti-
nuity known as the N'Komi fault zone. The SPOG
(Sismique Profonde Offshore Gabon) survey consis-
ted of five lines (figure 4), recorded to 15 sec.
two-way time (T.W.T.), with a total length of
about 700 km. Aboard the survey vessel Stormy
(CGG), a 6600 cu.in. tuned airgun array was used
together with a 3000 m/120 trace streamer.
Crustal features can be traced on all the
lines, especially the top of the lower crust and
the Moho (figure 5). The quality and the conti-
nuity of these features is not always good on the
seismic records, probably because of the scatter-
ing of seismic energy in the areas of extensive
diapirism. The lower crust does not appear signi-
ficantly layered. As in many other areas, the top
ii)-an early drift stage of Aptian age, which of the upper crust does not correspond to a sharp
mainly corresponds, above the breakup unconformity, reflector on the seismic record. The top of the
to an evaporitic series several hundred meters upper crust has been tentatively located in combi-
thick. nation with available ESP refraction data as the
top of the 6 km/sec. layer. Refraction data also
confirmed the underlying crustal boundaries.
On a regional scale, the southern segment is
Continental Lithosphere: Dp Seisnfic Refltions imaged by line SPOG 2 (figure 6). Line 2 shows,
Geodynanfics22 for over 150 km, a very progressive change from
1991 Amecan Geophysical Union moderately thinned crust (Moho at 27 km, crust
275
276 STRUCTURE OF THE WEST AFRICAN MARGIN
..
4/-
Gabon
Fig 1 Separation of Africa and South America; location of the Gabon Margin survey area.
17 km thick) beneath the shelf to a very thinned
crust (Moho at 16 km, crust 9 km thick) close to
the ocean-continent boundary, located 15 km be-
yond the SW end of the line by pre-existing seis-
mic lines. The crustal thinning affects the lower
crust more than the upper crust. The pre-salt
synrift series (pre-Aptian and older) is 4 to 5
km thick along the whole line and appears in
faulted blocks 15 to 20 km wide. The stratal
pattern in the synrift is subparallel. A typical
fan-shape pattern occurs only in places.
The northern segment, imaged by line 4 (figure
7), shows across less than 60 km a rapid change
from a slightly thinned crust (Moho at 30 km,
crust 22 km thick) beneath the shelf, to an ocea-
nic crust beneath the continental rise. Thinning
affects both the lower and the upper crust. The
stratal pattern in the synrift series clearly in-
dicates that the offset originated during the rift
period. An enlargement of this area on a time-mi-
grated section shows (figure 9), southeast of the
major fault zone, a local folding of the synrift
series (which is not a velocity pull-up artifact)
suggesting that some transpression occurred as a
result of shear movement between the two segments.
The offset and regional tilt of the pre-salt
surface, which is onlapped in places by early
Miocene strata, also suggest post-rift tectonic
movements.
Across the upper margin, line 5 (figure 10)
shows similar offsets of the Moho and of the top
of the lower crust beneath the transition zone
from the platform to the basin in the Cretaceous.
This offset is also interpreted as the deep
synrift series is of similar thickness and stratal expression of the N'Komi fault zone. The pre-salt
pattern as in the southern segment, but in the surface is not offset in this particular area as
area of maximum rate of thinning in the upper it is on line 1. The synrift series is cut into
crust, around SP 1500, it is cutnto smaller smaller blocks by closely-spaced normal faults.
faulted blocks. A typical oceanic'crust is present
in the most distal part of the line, but the exact Discussion
boundary between continental and oceanic struc-
tures is unclear. The SPOG deep seismic lines show for the first
The different crustal structures observed along time that abrupt changes in sediment facies and
each segment are also clearly imaged on the strike thicknesses distribution between the North Gabon
lines. and the South Gabon segments are closely superim-
Line 1, across the lower margin shows (figure posed on abrupt changes in deep crustal structures
8) around SP 1500 a large vertical offset of the which suggest different crustal evolutions on each
Moho and of the overlying units, which is visible side of the N'Komi fault zone. The different evo-
on the sections both in time and depth. This fea- lution of these two segments of the African margin
ture is interpreted as the expression of the N' could be interpreted as being the result of diffe-
Komi fault zone on the lower margin. The divergent rential extension between the North Gabon and the
WANNESSON ET AL. 277
Fig 2
5O
lOO
11o -
12o -
13o-
AGE
MAIN
EVENTS
LITHOLOGY
W E
BASIN PLATFORM
DEEP EROSION --
THE PLATFORM
AND THE SLOPE
BY CHANNELS
MANDOROVE
EROSION OF THE
PLAfFORM AND
THE SLOPE BY
CHANNELS
TU SECON RBIDITDE
SYSTEM
FIRST
TURBIDITE
SYSTEM
TRONG
OF THE SLOPE
GEOPRESSURED
RESERVOIRS
IKANDO FM
-- -- 111. ' .
__'
__ Ii1'. '.__''
_
CAP LOPEZ
GROW TH
FAULTING
DRAPING
ADDITIONAL
GRABEN AND
BLOCK
FAULT I N G
BASEMENT
EXTENSIONAL
BLOCK
FAU LTI N G
L L L L L
L. .L .... .I. .... L .... L .... L,
DENTALE
CRABE
MELANIA
ORGANIC SHALE
M BYA Mbr
MELANIA
[ UCINA Mbr
KISSENDA
BASAL
SANDSTONE
Million Years + + + + + + + + + BASEMENT
Stratigraphic column of the Gabon marginal basin (from Teisserenc and Villemin, 1989).
8 10E
++
++
++
LIBREVlLL ++7
..... :ii!ii?:ii:',' - +
..:.:-:.:.:.:.:':'::::li. '"'""' +
O :::::::::::::::::::::::::: ...'.-.:.:.:.:' -- + - :::::::::::::::::::::::: .i!!!iii:!:!:!:'
========================== i:i:!:i:i:!:!:i + +
=========================== !:i:!:!:i:!:!:ii!.
::::::::::::::::::::::::::::::::::::::::::::::::::::::::: + +
PORT GEN
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
:i:i:i:!:i:i B ) +
2__
'.:.:.:.:.:.::
':::::::::::
':::::::::::!ii!iiii!!
.:.:.:.:.:..::
":':':':':'.N:i:!:i:i:i:i:
::::::::::.'i:!:!:i:i
=====================
:' :' :' :' :'
..:.:.:.:.:.::::::::::::::::::
':::::::::::..:!
'.:.:.:.:.:..
+ 5
+
+ -
+
+ --t
+ +
+ +
+ +
+ +
+ -
+
+
+
+
--
+
+ +
+ +
+
4
+
LIBREVILLE
PORT GENTIL
:.:.:.:.:.:.:.:: ........... ...:.: '.:.:.:.
:':':':':':':':':'. :-;.i-;.:.:.:.:..
.......... A El
o .
:' Basinal facies
:::.:':... Carbonate Platform
i===! Lagoonal/continental facies
+
+
+
+
+
+
+
+
+
+
+ +
+
+ +
+
+ +
+
+
+
+ +
+ +
+
+
+
+ +
+ .
+
+ +
+ +
+
+
":'; Platform area
:!' Slope Facies with turbidites
Basement outcrops
Fig 3 Facies distribution of the Albian and Senonian formations in the Gabon marginal basin
(from Teisserenc and Villemin, 1989).
South Gabon segments. In this case, the N'Komi single rift to the south and a multiple-arm rift
fault zone has acted as an intraplate decoupling to the north. This includes the Reconcavo-Tucano
zone with horizontal displacement that must extend Basin on the American side and the Interior Basin
northeastward in the adjoining craton. However, of Gabon on the African side. If we add the exten-
there is no evidence of such a discontinuity in sion in these aborted rifts to that in the north-
the geology of Equatorial Africa (figure 4). On ern segment itself, the total extension is approx-
the other hand, the initial fit between Africa and imately the same as in the southern segment (P.
South America (figure 11) indicates that the Unternehr, ELF-IFP internal report, 1986). Hence,
N'Komi fault zone marks a boundary between a the N'Komi fault zone could be only a large trans-
CAMEROON SEGMENT
0 o
..
..................
-t-
-t- -t- -t-
-t- -t-
-t- -t- -t-
-I- -I- -I- -I-
+ -t- -I- -I- -I-
'1- -t- -t- -t- -t-
+ -I- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -t- -t- -t- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -t- -I- -I- -I-
-I- + -I- -I- -I- -I- -I- -I- -t- -I- -I- -I- -I- -I-
-I- -I- -I- + -I- -I- -I- -I- -I- -I- -I- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I- + -I-
-I- -I- -I- -I- -I- -I- -I- -I- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -I- + + -I-
-I- -I- -I- -I- + -I- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -I- -I-
-I- -I- -I- -I- -I- -I- -I- -I-
NORTH GABO
SE G IVIEN
"SPOG"
SURVEY AREA
SOUTH GABON SEGMENT
+ + -t-
+ + +
+ + + -t-
+ + + +
+ + + + -t- +
+ + + + +
+ -t- -t- +
+ + +
+ -t- +
+ + + -I.
+ +
+ + +
+ + + -t-
+ +
CONTINENT-OCEAN
BOUNDARY
5os
POINTE NOIRE
10OE
:'"'""'"-..' Cretaceous/Tertiary Basins
[ Congo Craton (+2000my)
'":'..' Precambr. Basins( 900-1500my )
'-"'. :.. :.:... Panafrican Tectonics ( 650my )
Kibarian/Eburnean
Tectonics ( 1100-1800my )
ij Tertiary Volcanics
Fig 4 Major geological features of Gabon and surrounding areas (from Teisserenc and Villemin,
1989); location of the SPOG deep seismic lines.
fer zone (Bally, 1981, figure 12). This means that movements within the continental crust, which
the overall extension is the same on each side of only took place during the rift stage.
the N'Komi fault zone, but not evenly distributed A tectonic heritage has been proposed to
along it. This results in local relative shear explain the onset and development of rifts. In
<
co
z
0
0
UJ
0
0
0
(30s) 11
-H
-H
o
o
o
o
o
I
o
03 N
O
O 4
O
O
O
rn oO
o
cc_
TWT (sec)
] z-
I
o
i I
I I
I ,
/
J ,,..
ii
/
\ /
I/l'l ,,' o
km
z
0
H
(os) iMi ,,,
wo
o
.-
Z o
TWT (sec)
\1
w
I!
/1
/i
I
I
/I
w
km
7
o
o
LU
z
o
o
o
o
o
o
z
o
o
-r 'Jr
+ + + + + r +
+ + + + + + r + +
+ + + + + + + + +
+ + + + + + + + + + + '-':.:'.i.":.'.
+ + + + ,- + +
+ + + + + + +
+ + + + + + + + +
+ + + + + + + + +
+ + + + + + + + + + + +
+ + + + + +
+ + + + + -
+ + + + +
+ ++/ +++++++ + +
+ + + + -'
+ + + + NGERA+
+ + + + + + + + f + + + +
+ + + + + + + +
+ + +
+ + + +
+ + +
4- + +
+ + GAMEROON
+ + + +
+ + + +
+ + + +
+ + + +
+ + +
_
+
- +
+
+ + +
+ + + +
+ + + + + +
a_ 4- + + +
+ + +
+ + +
+ + +
+ + + +
+ + + +
+ + +
+ + +
BRAZIL S,N
+ +
+
;- + +
BASIN
::..:..'. + Atlantic rifts +
+
! Gondwana craton L
----- Major cratonic lineaments
0 500kin
I , I , ,
+
+
+ +
+ + +
+ + + / EQ.GUINEA
+ + + /
+ + + -'
+ + + +
+ + +
+ + +
+ + +
ILLE
+++ \
+ GABON
+
+ +
- + + +
,+ + +
+ + +
+ + +
+ + + +
+ + + +
+
GONGO
+
MTE NOIRE
+
+ +
+
+ + +
-t- +
+
+
+ +
+
+
+
-t- +
+
+
+.
Fig 11 Initial fit between Africa and South Situation at the end of the rift period (Late
America (adapted from Szatmari et al, 1987) and Barremian/Early Aptian, 110 Ma).
development of the Equatorial Atlantic rifts.
fact, there is a good parallelism between older lopment of the Equatorial Atlantic rifts, where
zones of weakness in the American and African cra- two propagating systems have merged, one from the
tons and some rift features (Jatoba Basin, Araripe south, with a dominant divergent component and one
Basin, Benue Trough, figure 11). The N'Komi fault from the west, with a dominant shear component.
zone is also well integrated in this dominant NE-
SW paleotectonic pattern. Additional physical , Acknowledgments. The authors thank ELF-Gabon
geological and kinematic parameters must be consi- and IFP for permission to publish the seismic
dered however to better explain the complex deve- lines of the SPOG-1988 deep seismic survey.
SALVADOR NORTH LAMBARENE
RECONCAVO HORST GABON HORST INTERIOR
BASIN SEGMENT / BASIN
FUTURE
OCEAN-CONTINENT
BOUNDARY
Fig 12 Schematic block diagram of the Equatorial Atlantic rift across the N'Komi fault zone, inter-
preted as a transfer zone.
References nental margins, J. Geopys. Res. r 84 , 5973-6002,
1979.
Bally, A.W., Atlantic type margins, in Geology of Szatmari, P., Francolini, J.B.L., Zanotto, O., and
Passive Continental Margins, Amer. Assoc. of Wolff, S., Evoluao Tectonica da Margem Equato-
Petrol. Geologists Education Course Note Series
No 19, 1-48, 1981.
Emery, K.O., and Uchupi; E., The Synrift Super-
sequence and Crustal Boundary, in The Geology
of the Atlantic Ocean Springier Verlag,_ New.
York, 263-367, 1984.
Rabinowitz, P.D., and La Brecque, J., The Mesozoic
South Atlantic Ocean and Evolution of its conti-
rial Brasileira, Rev. Brasil. de Geocienci,
17 (2), 180-188, 987.
Teisserenc, P., and Villemin, J., The Sedimentary
Basin of Gabon. Geology and Oil Systems, in Di-
vergent/Passive Margins, Amer. Assoc. of Petrol
Geoloqists Memoir 48 Tulsa Oklahoma, 117-199,
1989.
REFLECTIVITY CAUSED BY TEXTURE-INDUCED ANISITROPY IN MYLONITES
S. Siegesmund, M. Fritzsche ' and G. Braun**
Institut fhr Geologic und Dynamik der Lithosphire der Universitit G/Sttingen,
3400 G/Sttingen, Goldschmidtstr.3 (FRG)
Abstract. The directional dependence of P- and S-wave and anisotropy (Jones and Nur, 1984; Fountain et al.,
velocities from mid-crustal mylonites in high-pressure 1984; Christensen and Szymanski, 1988). One of the lea-
situations has been investigated. From crystallographic ding hypotheses is that anisotropic fabrics (crystallograph-
preferred orientations of mineral components and the ic preferred orientation, here referred to as texture) could
respective elastic constants, the Voigt average was used produce the necessary impedance contrast to cause in-
to determine the complete 3-D elasticity tensor. With creased reflectivity. Seismic anisotropy due to the texture
increasing preferred orientatio. b due to progressive myloni- of constituent minerals of mylonites at high effective
tization, the anisotropy of Vr, increases from about 2% pressure, where the effect of open microcracks is elimi-
(protolith) to about 11% at the ultramylonitic stage. Shear nated, is on the order of up to 15% (Siegesmund and
wave splitting is at a minimum (< 1%) in the nearly iso- Kern, 1990). At upper crustal levels under conditions of
tropic parent rock; whereas, in the ultramylonite, it is brittle deformation the texture-induced anisotropy increa-
around 8%. ses, influenced by the occurrence of microcracks (Cram-
Synthetic seismograms were calculated, using the pin, 1987, Siegesmund et al., 1991).
reflectivity method, for models consisting of horizontal In this paper we investigate the seismic anisotropy of
layers with anisotropy increasing with depth, in a first mylonitic rocks which have been progressively deformed
attempt to model the seismic response of an anisotropic from metagranitoids to ultramylonites. Based on the rex-
medium with realistically derived elasticity tensors. ture of the rock-forming minerals, we computed the 3-D
INTRODUCTION
Seismic reflections from deep crustal levels in the cry-
stalline continental crust are commonly attributed to my-
lonite zones because in some cases these reflection surfa-
ces can be correlated to surface traces of major fault
zones (Smithson et al., 1979; Brewer et al., 1983). The
origin of these reflections are still in question, but many
hypotheses have been proposed such as zones of high
pore pressures (Jones and Nur, 1984), laminated structures
elastic and seismic properties of each mylonite using the
Voigt average.
Finally, we used these data for numerical calculation
of synthetic seismograms to investigate the effect of tex-
ture-induced anisotropic mylonite zones on reflectivity.
GEOLOGICAL SIGNIFICANCE AND SAMPLE
MODELLING
Mylonites are coherent rocks with, at least, microscop-
ic foliation, with or without porphyroclasts, characterized
by intense syntectonic crystal plastic grain-size reduction
of the protolith (Wise et al., 1984). Depending to the re-
* Geophysikalisches Institut der Universitit Karlsruhe, crystallization temperature of the respective rock-forming
7500 Karlsruhe, Hertzstr. 16 (FRG) minerals, shear zone rocks can be divided into LT- and
**Mineralogisch-Petrographisches Institut der Universitit HT-mylonites (Weber, 1990). LT-mylonites commonly con-
Kiel, 2300 Kiel, Olshausenstr. 40 (FRG)
1) present address: BP Exploration Dept. BPX-
London-XTC-Geophysics, Moorgate, London EC2 9BU
Geodynamics 22
tain cataclastically deformed minerals that were deformed
below their recrystallization temperature. In contrast, HT-
mylonites usually do not have clasts, because of suffi-
ciently high deformation temperatures. Mylonites show a
strong foliation and lineation and a distinct preferred
crystallographic orientation of those rock-forming mine-
rals, which have been strongly deformed by crystal-plastic
processes. In the ductile field, crystals mainly deform by
291
292 REFLECTIVITY CAUSED BY TEXTURE-INDUCED ANISITROPY IN MYLONITES
dislocation slip. Other processes can also take place, for
example dynamic recrystallization, dislocation climb, dif-
fusional flow, grain boundary sliding, superplastic defor-
mation, viscous flow, etc., (e.g., Schmid and Casey, 1986;
Urai et a1., 1986). Fabric patterns of mylonites are com-
monly monoclinic (asymmetric) with respect to the finite
strain direction (shape-preferred orientation, lattice-pre-
ferred orientation). This asymmetric reorientation of domi-
nant slip systems during deformation is regarded as an
indication of non-coaxial deformation based on the obser-
vation that slip planes and slip directions tend to rotate
parallel to the flow plane and the flow direction, respec-
tively. The effect of non-coaxial flow on the development
of mylonitic textures is still debated (e.g., Lister and
Snoke, 1984).
One of the main goals of structural geologists dealing
with mylonites is to establish the physical conditions of
the deformation and deformation history by textural and
microstructural studies. The interest of geophysicists
which has focussed special attention on mylonites is the
calibration of individual seismic reflections in terms of
geological structures, lithologies and fabrics. One of the
favored hypotheses suggests that changes in the magni-
tude of seismic anisotropy originates from the variations
in the degree of preferred orientation of minerals. Most of
the constituents in common mylonites are elastically an-
isotropic, and aggregates with a pronounced texture will
produce an elastically anisotropic rock.
In order to demonstrate the effect of different textures
on seismic anisotropy and crustal reflectivity, we selected
five rock samples which have been progressively deformed
from an metagranitoid to an ultramylonite. The composi-
tions of the protolith and ultramylonite are identical
(modelled as 35% biotite, 40% plagioclase and 25% quartz).
Schematic sketches (Fig. 1) illustrate the evolution of the
microfabric during mylonitization and the corresponding
progressive texture development of biotite (001), ranging
from random distribution to a strong pole maximum nor-
mal to foliation and lineation (parallel Z; Fig. 1). Such
texture transitions within mylonite zones are described,
for example, by Wenk and Pannetier (1990).
The data base used for our investigations is the set of
frequency diagrams of biotite (001) poles given by Hofler
(1989). The re-evaluation of the published pole figures to
compute the individual grain orientations were done by
digitizing with a scanner and vectorizing the contour
Progressive Mylonitization
Metagranitoid Mylonite
Ultramylonite
o
1
!Ii; 2
'" 3
'" 5
.. 7
.,
x
Increasing Maximum (00 I) Pole Density
Fig. 1. Schematically change of microstructures which illustrate the reduction in grain size
(taken from Higgins, 1971) and the development of biotite (001) concentration during
progressive mylonitization. XY is the foliation plane, lineation is parallel X; equal area
projection; contours in multiples of a random orientation.
$IEGESMUND ET AL. 293
lines. The image was further processed by deleting the CALCULATED SEISMIC PROPERTIES OF THE
contour lines and replacing their contour intervals with MODEL MYLONITES
the appropriate gray shading patterns. These gray patterns
were transformed to a random set of 1000 points of The directional dependence of wave propagation in a
representative density. The quartz and plagioclase aggre- polycrystalline aggregate depends on the single crystal ve-
gares are assumed to be randomly distributed, which is in locities, the volume fraction of the constituent minerals
general not realistic (see discussion in the following chap- and on their textures. These assumptions are valid for
ter). high-pressure conditions where the effect of microcracks
Z 7.0
6.0
05.0
Z.O -,
= /
"'""'l"'"'"'l"'"'"'l'"'"'"
//
//
.
:
.
--3.0
o /
/
/
>5.0

.0
?.0 ..................................................... ......... ........
?.0 .0 5.0 .0 3.0 .0 5.0 .0 ?.0
VELOCITY IN K/SEC X
Z 7.0
6.0
o5.0
'.0
Z
6.0
7.0
b
7.0 6.0 5.0 +.0 3.0 .0 5.0 6.0 7.0
VELOCITY IN KH/SEO X
Fig. 2. (a) Phase velocity diagram for the protolith (layer I) in a vertical plane. Solid line
shows P-velocity, dotted line qSV-velocity and dash-dotted line qSH velocity. Note that P-
velocity for steep incidence angels differ most for those in the following layers. S-
velocities remain nearly unchanged; (b) Phase velocities for layer V (ultramylonite).
i i
IiDensity I Isot,ropic
Lithology Texture I, gJcm I Vp
Biotite (001) 2.7 2.8:6.456.556.65
Aniso, tropic i
Vp ', Vs ', vplvs
6.1 6.2 6.3 6.4 6.5 L9 4.0 1.5 1.6 1.7
Fig. 3. An idealized horizontally layered cross section (normal incidence) that includes a
mylonite zone and shows the dependence of seismic anisotropy on preferred mineral
orientation. The velocities and Vp/Vs ratios correspond to the propagation direction normal
to foliation and linearion. In addition, the isotropic reference model is also given.
is eliminated. We calculated the Voigt average to obtain verse direction. The transverse isotropic behaviour of a
the aggregate elastic constants for the textured polycrys- biotite single crystal, in addition to the orientation pattern
tals. The procedure is explained elsewhere (e.g., Crosson explains the nearly identical velocity parallel to the folia-
and Lin, 1971; Siegesmund, 1989). Comparison between tion plane.
texture-derived and laboratory seismic velocitiy measure- Quartz and plagioclase are assumed to be randomly
ments have shown that P-wave velocity calculations using distributed in our model rocks and are therefore referred
the Voigt average give a good agreement (Crosson and to as an elastically isotropic contribution. The experimen-
Lin, 1971). An extended study made by Siegesmund (1989) tal results of phyllosilicate-rich, quartz-feldspar rocks are
gives evidence that this observation is valid for the de- commonly transverse isotropic, thus giving evidence that
scription of P-, S-waves and shear wave splitting. the directional dependence is mainly controlled by the
Before considering the whole rock anisotropy it is im- mica group (Siegesmund, 1989).
portant to explain the contribution of the texture-related Mylonites commonly show three main types of quartz c
anisotropy of each mineral phase. The preferred orienta- -axis pattern (crossed girdle, single girdle and single ma-
rion of biotite (around 47% crystal anisotropy) and its high ximum pattern). Plagioclase textures of LT-mylonites are
volume percentage (35%) causes the highest velocity rare. Siegesmund (1989) measured plagioclase _[_(010) con-
anisotropy. The velocity is slowest in the direction normal centrations normal to foliation and lineation in a mylonite
to the lineation and foliation (Z), because the slowest P- from the Insubric Line.
wave direction of the single crystal is orientated parallel One of the investigated mylonites exhibited a single
to the c-axis and coincides with the strong maximum of maximum concentration of c-axes normal to the linearion
(001) poles in this direction, and is highest in the trans- within the foliation (Y); whereas, the plagioclase is ran-
SIEGESMUND ET AL. 295
domly orientated (Wenk and Pannetier, 1990). Such a velocity-depth function (Fig. 3). In addition to the fabric-
measured c-axes concentration has the highest P-velocity anisotropic relation an isotropic reference model is also
in roughly parallel this concentration direction, which cor- included. The given velocities correspond to the propaga-
responds to the single crystal velocities normal to z'- and tion direction normal to foliation and lineation.
z-planes and other lattice planes with normals closer to
the c-axis (Braun et al., 1991). The whole rock seismic COMPUTATION OF SYNTHETIC SEISMOGRAMS
anisotropy that corresponds to the biotite-related anisotro- WITH THE REFLECTIVITY METHOD
py will be reduced by the contribution of this quartz
single maximum pattern, because the high-velocity direc- To evaluate the reflectivity of horizontally stratified
tion of biotite (6.9 km/s) parallel Y is superimposed by media, where mylonitization increases with depth, which
the contribution of quartz (6.5 km/s). leads to different degrees of anisotropy, different numeri-
However, since only the different degrees of the bulk cal calculations were carried out using the reflectivity
rock seismic anisotropy (from the protolith to the ultra- method (Fuchs and Mueller, 1971) in an extended version
mylonite) is of importance for the first attempt to model to allow for anisotropic wave propagation (Nolte, 1988).
the reflectivity that is induced by the mylonitic fabric, The intention was to study the reflectivity of a model
the assumption for the calculation- exclusively biotite-re- made up by a progressively increased number of layers,
lated bulk seismic anisotropy - seems to be valid. , which differ in degree of anisotropy (I-V), underlying an
According to our model, we calculated the complete isotropic top layer (0) and to show that the synthetic re-
elastic and seismic properties of the five mylonitic layers. sponse of a 'realistic' mylonite zone can be modelled.
The phase velocities of P-, qSV- and qSH-velocities (in The model that was used for the computations, can be
the XZ plane) for layer I and V in order of increasing described as follows (see Fig. 4): the top layer is iso-
maximum (001) pole density and seismic anisotropy are tropic with 10 km thickness, a P-velocity of 6.1 km/s,
displayed in Fig. 2a, b. The effect of texture-induced an- 3.5 km/s S-velocity and 2.57 g/ccm density. The five an-
isotropy in an assumed model of horizontal layers (normal isotropic layers are 0.2 km in thickness each. All elasti-
incidence, no density change) from the metagranitoid city tensors were nearly of rotational symmetry and were
(layer I) to the ultramylonite (layer V, strongest point orientated such that the direction of highest P-velocity
maxima of biotite (001) poles) can be inferred from the was parallel to the profile along which the source and
(km)
10
isotropic
Fig. 4. Model used for the calculations of synthetic seismograms. The top layer (0) is iso-
tropic with Vp=6.1 km/s, Vs=3.5 km/s and density p=2.$? g/corn. Layers I-V are anisotro-
pic of 200 m thickness each, the degree of anisotropy increases downwards. For Vp. Vs
and of Layers I-V refer to Fig. 3. For the computations, layer V was treated as an aniso-
tropic halfspace.
receivers were aligned, i.e. the vertical plane containing either purely vertical for SV-waves and purely horizontal
both source and receivers was a plane of symmetry. for SH-waves.
The source used was of explosion type, source signal In the general case the plane of shear wave polariza-
was a Mueller-pulse with length of 0.2 s. The calculated tion may rotate during propagation through the medium,
frequency spectrum ranged from 0.2 to 25 Hz. The source polarization is then denoted as qSV (quasi-vertical) and
was located directly under the free surface, receivers at qSH (quasi-horizontal), respectively.
the surface. Synthetic seismograms for near and far The phase velocity diagrams in Fig. 2a-b show which
offsets were computed. angle of incidence velocity will show the greatest change
Results for the long offset range calculation for the across an interface. It is obvious that velocities for a
model viewed in Fig. 4 are shown in Fig. 5a, b. The given angel of incidence vary only slightly for consecutive
phases seen are the direct wave (1), the P-reflection (2) layers. Because the variation of density with depth is not
from the uppermost interface (isotropic[0]-anisotropic l-l]) significant (see Fig. 3), the reflectivity to be expected is
and the related P-S converted wave (3). Phases (4) and rather small. The biggest steps seem to occur for steep
(5) show the refracted waves from the same interface. incidence angles, we therefore calculated synthetic seis-
With the model parameters chosen, no phases associated mograms for short offsets taking the same model that
with a reflection from an anisotropic-anisotropic interface was used in the long offset calculations.
can be observed. No energy was recorded on the trans- The observed phases correspond to those identified in
verse component (not shown) as wave propagation took Fig. 5. The resulting seismogram (Fig. 6a-b) show the ad-
place in a symmetry plane. The polarization is therefore ditional phases (4) and (5). Phase (4) is caused by P-S
I-'
0N, 3eee. DISTRNCE R IN KH
a
b 08,N, 3eee. D ISTRNCE R IN KH
200
Fig. 5. (a) Radial component for the long offset range obtained for the model shown in
Fig. 4 . Source was of explosion type, reduction velocity 6.7 km/s. The first phase (1) is
a direct wave, phase (2) is the reflected wave from the top interface (isotropic[O]-
anisotropic[I]) and phase (3) the related P-S converted phase. Beyond 40 km offset,
refracted waves (4) and (5) can be observed; (b) Vertical component for the same model.
$IEGESMUND ET AL. 297
6
"' 3
I
Q
$ '1 '15 2 25
GnN, 3eee. DISTRNCE R IN KH
,
o
7
3
2
'1
o 5
b
RIN, 3000.
'1o 15 20 25
DISTRNCE R IN KH
Fig. 6. (a) Radial component for the short offset range. The model is the same as in Fig.
5. Phase (4) is caused by P-S conversion at the interface between layer II and III. Phase
(5) is the corresponding P-reflection, seen on the vertical component (b).
conversion of the incident P-wave at the interface be- stants and the volume content of each mineral phase, the
tween the second and third anisotropic layer (i.e. layer 3-D elastic properties were calculated. With increasing
[II-I and [III], see Fig. 3), the associated P-reflection (5) preferred orientation of biotite (001) poles, the Vp-aniso-
on the vertical component is also visible; note that it has tropy of the model mylonites increases from about 2%
the same sign as the isotropic[0-l-anisotropic[1-I reflection. (protolith, layer I) to about 11% in the ultramylonite (layer
These two phases could be ascribed to the above stated V). The shear-wave splitting anisotropy is at a minimum
interface, since they could not be observed in the results (< 1%) in layer I and at a maximum (ca. 8%) in the ultra-
obtained for a model that included only layers [0]-[II3 mylonite. According to the proposed geometry of horizon-
and, therefore, did not have the indicated interface. The tal layering, with the Z-axis as the symmetry axis, the
results shown in Figs. 5 and 6 do not yet resemble those shear wave splitting for all layers is small in this direc-
of laminated parts of the earth's crust, due to small re- tion due to the transverse isotropic behaviour. Because of
flection coefficients between layers. the comparatively low anisotropy and its small changes
across the interfaces, the results obtained for this source-
CONCLUSIONS receiver configuration with respect to the symmetry axis
of the elasticity tensors did not exhibit any spectacular
In this paper we presented a suite of mid-crustal mylo- features. The same holds true for calculations performed
nitic layers with different degrees of fabric anisotropy un- for a source-receiver profile which is rotated relative to
der high-pressure conditions. The mylonites provide objec- the symmetry axis (not shown); the difference due to
rive data that demonstrate the effect of fabric anisotropy nonsymmetric condition is only a slight change in the po-
on seismic anisotropy and reflectivity at deeper crustal larization of the events, which corresponds to the recor-
levels. From the textural data, the single crystal con- ded energy on the transverse component.
The most remarkable observation is the difference in H/Sfler, S., 1989. Texturanalyse an Metallen und Minera-
reflectivity for the short offset range that was obtained lien und Bestimmung von Meteorit-Mikrostrukturen mir
for the same model. The P-reflection at the mylonite[II]- Neutronenbeugung, tez7'chte der liDm/bzsc/Jzaa/e
mylonite[III] interface corresponds to the velocity de- 7/ch, A& ,274 (PHD Thesis Univ. Bonn), 109 pp.
crease parallel to the Z-direction which is correlated with Jones, T. and Nur, A., 1984. The nature of seismic reflec-
an increase in preferred orientation of the (001) mica tions from deep crustal fault zones, J. 6eo/Jhys. Res..,
poles. The reflectivity that is observed only for the steep aLo, 3153-3171.
incidence angle at these layer bounderies is caused alone Lister, G.S. and Snoke, A.W., 1984, SC-mylonites. J.
by the velocity contrast because the change in density 5Yzct. GeoZ, , 617-638.
with depths is rather small. Nolte, B., 1988. Erweiterung und Anwendung des Refiekti-
We feel that further seismic modelling, especially for vitiitsprogrammes f'dr anisotrope Medien, Z2/o./om? 7]w-
low-symmetry conditions (orthorhombic, monoclinic, tri- s/s, University of Karlsruhe, 120 pp.
clinic) with varying geometrical configurations and stron- Schmid, S.M. and Casey, M., 1986. Complete fabric ana-
ger reflection coefficients based on 3-D elastic data, lysis of some commonly observed quartz e-axis pat-
would provide more realistic information on the still ques- terns, Am. (eo/Jh2. Un. (eo/9t.: Alono: 75, 263-286.
tionable sources of the reflection surfaces at deeper Siegesround, S. 1989. Texturelie und strukturelle Eigen-
crustal levels. schaften mylonitischer Gesteine der Insubrischen Linie
Acknowledgments. The authors would like to thank K.
Fuchs (Karlsruhe), R. Meissner (Kiel) and K. Weber (GiSt-
tingen) for numerous discussions. Technical help was pro-
vided by B. Mackewic (GiSttingen).
REFERENCES
(Ivrea-Zone, Italien) und ihr EinfiuB auf die elastischen
Gesteinseigenschaften. Ein Beitrag zur Interpretation
seismischer in situ-Messungen, PHD thesis, Universitit
Kiel, 172 pp.
Siegesmund, S., Takeshita, T. & Kern, H., 1989. Anisotro-
py of Vp and V, in an amphibolite of the deeper crust
and its relationship to the mineralogical, microstructural
and textural charakteristics of the rock, Tectonophy-
sics., Z7, 25-38.
Braun, G., Siegesmund, S. and M. Dahms, 1991. The in- Siegesmund, S. and H. Kern, 1990. Velocity anisotropy and
fluence of quartz textures on the seismic anisotropy in shear wave splitting in mylonites from the Insubric
lower crustal granulites, J7. bDzct. GeoZ (in press). Line (Ivrea-Zone, Italy), ?rth.
Brewer, J., Matthews, D., Warner, N., Hall, J., Smythe, 29-47.
D. and Wittington, R., 1983. Birps deep seismic reflec- Siegesmund, S., Kern, H. and Vollbrecht. A., 1991. The
tion studies of the Britisch Caledonides, AOtzz'e, 7RX Effect of Oriented Microcracks on Seismic Velocities in
206-210. an Ultramylonite, Tectoaop/ysl'C& h'7, (in press).
Christensen, N.I. and Szymanski, D.L., 1988. Origin of re- Smithson, S.B., Brewer, J.A., Kaufinann, S., Oliver, J.E.
fiection from the Brevard Fault zone, J. (>ophys. Res., and Hurich, C.A., 1979. Structure of the Laramide Wind
R;', 1087-1102. River uplift, Wyoming, from COCORP deep reflection
Crampin, S., 1987. The geological and industrial implica- data, J. Geoph. Vs. Res.., 3'4, 5955-5972.
tions of extensive-dilatancy anisotropy, A/i?rz#'e, .7zva', Urai, J.L, Means, W.D. and Lister, G.S., 1986. Dynamic
491-496. recrystallization of minerals, Aaz (eo/Jhys. Un. Geo-
Crosson, R.S. and Lin, J.W., 1971. Voigt and Reuss pre- /9h. vs. g./ono: 35, 161-199.
diction of anisotropic elnsticity of olivine, J. Geo/Jh. rs. Weber, K., 1990. HT-mylonites: Fabrics, strain path and
A;'e:, ?, 570-578.
Fountain, D.M., Hurich, C.M. and S.B. Smithson, 1984.
Seismic reflectivity of mylonite zones in the crust,
eolR,.; I 195-198.
Fuchs, K. and Mkiller, G., 1971. Computations of synthetic
seismograms with the reflectivity method and compari-
metamorphic conditions. In: P. Bankwitz (Editor), De-
formation Processes and the Structure of the Lithosphe-
re, p 63.
Wenk, H.-R. and Pannetier, J., 1990. Texture development
in deformed granodiorites from the Santa Rosa mylonite
zone, Southern California, J. Szwt. OoZ ] 177-184.
sion with observations, (?eo/9t.?s. J. R. a3w: bbc., , Wise, O.U., Dunn, D.E., Engelder, J.T., Geisser, P.A.,
417-433. Hatcher, R.D., Kish, S.H., Odom, A.L. and Schamel, S.,
Higgins, M.W., 1971. Cataclastic rocks, U.X (?eoZ 5#. 1984. Fault-related rocks: suggestions for terminology,
?z'of' ?npe; '7, Washington, 97 pp. Geolo,; l 391-394.
CORRELATING CRUSTAL REFLECTIONS WITH GEOLOGIC OUTCROPS:
SEISMIC MODELING RESULTS FROM THE SOUTHWESTERN USA AND THE SWISS ALPS
R. K. Litak 1, R. H. Marchant 2, L. D. Brown 1, O. A. Pfiffner 3, and E. C. Hauser 4
Abstract. Outcropping geology can often provide direct and detailed
information concerning the relationship of seismic reflections to specific
geologic features, a crucial but difficult aspect of deep seismic
interpretation. In this study we present two very different examples in
which fortuitous geologic circumstances enable us to anticipate the
nature and geometry of reflectors to substantial depths. The approach
used is to construct geologic models based on surface mapping, for
which synthetic seismograms are calculated and then compared to nearby
deep seismic lines. In these cases, the geologic features observed at the
surface lead to reflection patterns quite similar to those actually observed
on the data.
In the Buck Mts. of Arizona, southwestern USA, rotation above a
Tertiary detachment fault has exposed an upper crustal block which now
represents a virtual cross-section of pre-Tertiary crust. Seismic modeling
based on this exposure reveals that the numerous (originally)
subhorizontal Proterozoic diabase intrusions observed in the Buck Mts.
generate strong synthetic reflections that are very similar in character,
continuity, and distribution to the anomalous Bagdad Reflection
Sequence (BRS) imaged by nearby COCORP and industry seismic lines.
We infer that the BRS, which permeates the upper crust for distances of
some 100 km and corresponds to depths of up to 15 km, is due to a
series of subhorizontal mafic intrusions similar to those in the Buck
Mts.
Another area where surface geology is especially helpful in
interpreting deep reflections is in the Penninic Alps of eastern
Switzerland, where the combination of relatively steep dips, high relief,
good exposures, and extensive fieldwork has permitted the construction
of contour maps of several Penninic nappe surfaces. Extrapolating these
contours to depth enables construction of a 3-D model of the upper crust
beneath the south end of Swiss seismic line NFP20-East. Results of
raytracing indicate that interfaces between nappe surfaces and
metasedimentary cover rocks can account for most reflections in the
upper six seconds, thereby constraining the 3-D structure of the upper
crust in this area.
1Department of Geological Sciences and Institute for the Study
of the Continents, Cornell University, Ithaca, New York 14853, USA
2Institute of Geology, BFSH-2, University of Lausanne, 1015
Dorigny, Switzerland
3Institute of Geology, University of Bern, Baltzerstr. 1, CH-
3012 Bern, Switzerland
4Institute for the Study of the Continents, Cornell University,
Ithaca, New York 14853, USA
Geodynamics 22
In both examples, reflections can be explained as caused by simple
lithologic contrasts: between mafic and silicic rocks (in Arizona), and
between silicic basement, metasediments, and ophiolites (in
Switzerland). These results emphasize the importance of lithologic
heterogeneity in producing deep crustal reflections, and provide possible
analogs for reflection patterns in areas where geologic control is less
certain.
Introduction
Since the advent of the COCORP (Consortium for Continental
Reflection Profiling) program in the mid-1970s, the increasing use of
deep crustal reflection profiling has made substantial contributions to
present understanding of the continental crust [e.g., Hauser and Oliver,
1987], and has now become a standard tool of crustal exploration.
Perspectives on continental structure and evolution are currently being
recast by results from deep seismic surveys around the world. Ever since
the earliest studies, however, the nature of the reflectors themselves has
remained controversial [e.g., Matthews et al., 1987]. A number of
hypotheses have been advanced to account for basement reflectivity; in
all likelihood a diversity of mechanisms are responsible. Although
substantial progress has been made toward identifying the nature of some
deep reflectors, important problems persist; some have even questioned
the ultimate comprehensibility of deep seismic data. These issues are
frequently difficult to address since direct evidence tying crustal
reflections to known geologic interfaces is often lacking.
Here we present results from two studies in which unusually
favorable circumstances appear to make it possible to correlate
reflections to specific geologic features with a relatively high degree of
confidence. Although they represent disparate situations both in terms
of geologic setting and investigative approach, in both instances seismic
modeling is the primary tool employed and the initial model is based on
specific geologic features exposed nearby. In interpreting the Bagdad
Reflection Sequence (Arizona), we model synthetic reflections from
mafic intrusions in the upturned upper-crustal block of the Buck
Mountains as analogous to reflections observed on nearby COCORP
data, thereby inferring that mafic intrusions are also the cause of the
COCORP reflections. In the Penninic Alps of eastern Switzerland,
downplunge projection of well-defined surface trends provides the basis
for an initial geometrical model of nappes beneath Swiss seismic line
NFP20-EAST. Comparison of synthetic and real reflection geometry
not only reveals that these nappe surfaces can explain nearly all upper-
crustal reflections in this area, but also allows iterative modification of
the original interpretation to better resolve subsurface structure at depth.
In these areas (at least), our results indicate that many crustal reflections
can be interpreted with confidence, and may, in conjunction with other
studies, provide insight into the interpretation of reflection patterns
observed elsewhere, and the fundamental question of the origins of
crustal reflections.
299
300 CORRELATING CRUSTAL REFLECTIONS WITH OUTCROPS
The Bagdad Reflection Sequence
In 1986, the Consortium for Continental Reflection Profiling
(COCORP) collected approximately 550 km of deep seismic reflection
data in Arizona from the Basin and Range across the transition zone onto
the Colorado Plateau (Fig. 1). Among the results, which are
summarized by Hauser et al. [ 1987], was the delineation of a remarkable
series of strong subhorizontal reflections from the upper crustal
basement, corresponding to depths as great as 15 km (Fig. 2). These
reflections, termed the Bagdad Reflection Sequence (BRS) by Hauser et
al. [1987], occur in narrow bands which are correlatable for tens of
kilometers, and are separated from each other by relatively non-reflective
zones. They extend for some 100 km across the transition zone, and are
notable for their unusually high amplitude.
Several factors argue that the BRS emanates from thin zones of
substantially different acoustic impedance from the surrounding rock, as
discussed by Hauser et al. [1987]. If so, two possibilities for the origin
of the BRS present themselves [Hauser et al., 1987]. First, the BRS
may be caused by Cenozoic detachment surfaces that are reflective due to
a mylonitic or shear fabric, or due to fluid-filled cracks associated with
brecciation. Second, these reflections may represent thin, high
impedance mafic intrusions into the gneissic and granitic country rock.
By taking advantage of an unusual basement exposure nearby, this study
examines the viability of the hypothesis that the BRS represents tabular
mafic intrusions. Our modeling results, in combination with other
analyses, strongly favor the mafic intrusion explanation over one
involving detachment surfaces.
Near the COCORP transect the Buck and Mohave Mountains have
both been described by Howard et al. [ 1982] as upper crustal blocks that
have been rotated by as much as 80 through a horizontal axis by
movement along Miocene detachment faults. Thus these rotated blocks
approximate cross sections of the upper crust as it was before Tertiary
extension, and provide perhaps some of the most direct evidence
available concerning what lies beneath the COCORP lines. As such,
they present an unusual opportunity to investigate the upper crustal
reflectors through the use of seismic modeling. Since the published map
of the Buck Mountains is more detailed and the block itself is less
disrupted than that of the Mohave Mountains, the former is used in this
study.
In the Buck Mountains, a variety of Precambrian granitic gneisses
comprise an antiform that has been intruded by thin zones of later
Proterozoic diabase (Fig. 3a). Tertiary contacts dip approximately 80 to
the southwest and are generally parallel to the diabase intrusions,
signifying that these intrusions were subhorizontal prior to Tertiary
extension, and the entire block has subsequently undergone rotation
through about 80 . The dipping, gradational contacts between the
gneisses, which are of similar lithology, would not be expected to
produce the strong, subhorizontal reflections observed. However, since
mafic rocks generally have substantially higher velocities and densities
than granitic gneisses, equivalent subhorizontal dikes may be expected to
generate strong reflections if they are present in the subsurface beneath
the COCORP transect. In fact, tabular diabase intrusions into granitic
gneiss have been found to cause similar reflections in the area of the
Siljan borehole [Juhlin, 1988, 1990]. Modeling the seismic response of
these dikes can test whether their geometric distribution is generally
consistent with the overall reflection pattern and seismic character
observed on the COCORP data.
Since the Buck Mountain range block has been rotated -80 , we have
applied an additional 10 rotation to transform the map view of present
structure (Fig. 3a) into a pre-Tertiary cross-section. This restored
section, buried 4.1 km in order to roughly match the depth range of the
principal seismic reflectors in the COCORP data, is the basis for the 12
by 6 km seismic model of the upper crust (Fig. 3b). Physical properties
are assigned based on well data, stacking velocities, and laboratory
measurements [Litak, 1990], and feature gneiss velocities ranging from
5.80 km/s to 6.05 km/s and a diabase velocity of 6.65 km/s, yielding a
gneiss/diabase reflection coefficient of approximately 0.1. It is
important to recognize that uncertainty in the absolute value of the
velocities will not affect the conclusions of this study. Rather, the
resultant synthetic reflection pattern is dependent merely upon the
premise that the gneisses possess similar values for acoustic impedance,
while the impedance of the dikes is substantially higher. The intrusions
were assigned a uniform thickness of 20 m based on their average
thickness observed in the field [K. Howard, pers. comm.]. At this
.;Cenozoic Volcanic Rocks Prterzic Crystalline
Rocks
..'.'.".Mesozoic Igneous Rocks Detachment Fault
'""-"Mz-Pz Sedimentary Cz Normal Fault
Rocks
Fig. 1. Generalized geologic map of west-central Arizona showing the
location of the Buck Mountains and the Phillips Kirkland State A-1
well. COCORP seismic profiles are depicted as heavy black lines;
dotted lines indicate the extent of the Bagdad Reflection Sequence on the
COCORP data.
LITAK ET AL. 301
Arizona I
0.0
1.0
2.0
3.0
4.0
5.0
6 km/s 2 km
Fig. 2. Stacked section of the upper five seconds of the western end of COCORP Arizona Line 1. The strong, horizontal-
to-gently-dipping reflections constitute the Bagdad Reflection Sequence. Amplitudes are corrected for spherical divergence
only.
thickness, in combination with the relevant velocity and frequency (25
Hz), tuning effects are insignificant.
After establishing reasonable velocities and layer thicknesses, pseudo-
wave-equation forward modeling was performed with the commercial
software package AIMS (m Geoquest International) in the manner
described by Reston [1988], and band-limited random noise added. In the
resulting synthetic seismic section, individual reflections appear to
coalesce into narrow bands which are broadly continuous for much of the
section. Lateral variations and an overall banded appearance strongly
resemble the BRS; the low impedance contrasts between the gneisses are
insufficient to produce observable reflections in this case [Litak and
Hauser, 1989; Litak, 1990].
In addition to comparing the overall geometry and reflection character,
it is enlightening to consider a detailed comparison of the synthetic
seismic section to the western end of Arizona Line 1, where the BRS is
best developed (Fig. 4). The similar scale of reflection features and
nonreflective zones between the BRS and the synthetic data is striking.
Moreover, in both cases sharp, continuous reflections with relatively
short, simple waveforms ('A' on Figure 4a,b) truncate or change abruptly
into more diffuse reflections ('B' on Figure 4a,b). What appear to be
cross-cutting reflections are even observed in both cases ('C' on Figure
4a,b). For the synthetic data, these characteristics can be ascribed to the
variation from isolated intrusions (continuous for hundreds of meters to a
couple of kilometers) to groups of proximate intrusions, some of which
have discordant dips, where interference has created a more complex
waveform. The fact that these effects are also evident on the COCORP
data suggests that the pattern of dikes used in the model can reproduce
the observed characteristics of the real data, indicating that a similar
mechanism could explain the BRS.
The proximity of the Buck Mr. exposures to the COCORP traverse
suggests subhorizontal mafic intrusions as perhaps the most obvious
candidates for the cause of the BRS. Modeling indicates that they in fact
(a) GEOLOGIC MAP OF BUCK MOUNTAINS
114J I f.gQ" _
riometers
,
114 07' 30"
Qa
...'....:..Fault, dashed where inferred,
dotted where concealed
42'
00"
54
50"
QUATERNARY
rs'lAIluv i urn
TERTIARY
IZVolc. & sed. rocks
(Miocene)
I;f'Silicic dike (Miocene)
PROTEROZOIC
IDiabase
[]Pegmatite
rLeucogranite gneiss
.J]Augen gneiss iPsO?S and
undivided gne'
F ine-grained biotite
granite gneiss
JAmphibolite
JGneiss, undivided
--t- An tifor m
After Howard et al (1982)
(b) SEISMIC MODEL
uJ
c
0 DISTANCE (km)
Fig. 3. a) Geologic map of the Buck Mountains. Precambrian granitic
gneisses comprise an antiform which has been intruded by later
Proterozoic diabase dikes. The entire block has been rotated -- 80 above
a Tertiary detachment fault. b) Seismic model produced by rotation of
(a) to pre-Tertiary vertical and burial to a depth of 4.1 km. Crosses
represent input data points to the model. Note that diabase intrusions,
which parallel the Tertiary contact, restore to near horizontal.
provide a viable explanation, since the synthetic seismic response of
dikes exposed in the Buck Mountains approximates the geometry,
character, and scale of the reflections comprising the BRS. Tertiary
detachment faults as exposed in the Buckskin Mountains and elsewhere
[Davis et al., 1980] appear to present the most geologically reasonable
alternative explanation. Space limitations prevent a full discussion of
this possibility here, but Litak [1990] argues that several lines of
evidence -- polarity [Goodwin et al., 1989], well data, similarity to the
Siljan reflections [cf. Juhlin, 1988], and the distinctive seismic character
of the BRS -- combine to make the detachment hypothesis considerably
less attractive.
Thus it appears that the Bagdad Reflection Sequence is most credibly
explained by the presence of tabular mafic intrusions. In addition to
possibly constraining models of Basin and Range extension, this result
could possess applicability to the interpretation of other laterally
extensive thin zones of high-amplitude reflections observed to occur in
the Precambrian in many areas. In other surveys where similar reflection
packages have been noted [e.g., Brewer et al., 1981; Pratt et al., 1989;
Hurich et al., 1989], emphasis has generally been accorded to a possible
origin involving buried Proterozoic sedimentary basins, volcanic flows,
or imbricate thrust sheets. In the Arizona transect, outcrops of
Precambrian metamorphic basement preclude such explanations.
Elsewhere, well data [Juhlin, 1988] and exposures of multiple episodes
of Precambrian mafic intrusions in shield areas [Halls and Fahrig, 1987]
suggest that these features may have a profound impact on reflection data
[MacDonald, 1988]. These results further emphasize the significance of
mafic intrusions as an important cause of reflections in the crystalline
basement of the upper continental crust.
Penninic Nappes
Another area where surface geology is crucial to the seismic
interpretation is the Penninic Alps of eastern Switzerland. Here exposed
geologic features are again used as the starting point for seismic
modeling; however, the connection is more direct as surface trends can be
projected downplunge beneath the seismic line. In the Penninic zone, a
stack of basement nappes up to 15 km thick are separated from each
other by thin zones of Mesozoic sedimentary cover [Pfiffner et al.,
1990a]. The Swiss seismic line NFP 20-EAST extends some 120 km
across eastern Switzerland, and crosses into the Penninic zone where the
Adula, Tambo, and Suretta nappes should be present in the subsurface
(Fig. 5). Among the results of this survey is the observation of
numerous strong reflections in the upper crust of this region; many of
which have been ascribed by Pfiffner et al. [1990b] to nappe surfaces.
The use of three-dimensional modeling to link reflections to observed
(a)
3.0
3.5
ARIZ I
SYNTHETIC
(b)
Fig. 4. Detailed comparison of a portion of Arizona Line 1 to a portion
of the synthetic section corresponding to the model in figure 3b.
Synthetic reflections are from 20 m thick intrusions, convolved with 12-
48 Hz Klauder wavelet, random noise added. Annotation indicates
features of similar seismic character, suggesting a similar origin.
Vertical scale is two way travel time in seconds.
.r- Variscan basement
Chiavenna ophiolite
Aar
Chur
BiJndner-
I schiefer
I
I
Study xxx St. ri'x
%%%%%%%%%
.. x x
Slmano.x,x
lO krn
Fig. 5. Generalized geologic map of the Penninic zone of eastern
Switzerland showing location of basement nappes and seismic line
NFP20-East; bold portion of line is used in this study.
geologic structures verifies this hypothesis and allows seismic mapping
of the deep structure of these features.
In the high relief of the Penninic Alps, detailed mapping of the
dipping nappe surfaces has yielded contour maps of the near-surface
structure for the Suretta, Tambo, and Adula nappes [Pfiffner et al.,
1990a]. Trends revealed by this mapping are the basis for a three-
dimensional geometrical model of the stacked nappes beneath line NFP
20-EAST. Contours were extrapolated downplunge in an effort to
determine if synthetic reflections from the nappes might have a gross
geometry similar to that of the observed reflections. Where preliminary
results were encouraging, the process was iterated to produce a best-fit
synthetic seismogram and a model consistent with known geology. The
final model has dimensions of 60 km (north-south) by 20 km (east-west)
and a maximum depth of 20 km.
For the Suretta nappe (the structurally highest), this process involved
extrapolation of up to 14 km in distance and 5 km in depth to complete
the entire model. However, extrapolation to the location of reflection
points (the downdip limit of seismic control), is somewhat less -- 9 km
in distance and 3 km in depth (Fig. 6). The deeper nappes involved
extrapolation over greater distances (up to 25 km laterally and 16 km in
depth for the base of the Adula); determination of their position is
therefore less well-constrained by surface data alone and more dependent
upon the modeling results.
LITAK ET AL. 303
Layers corresponding to other units exposed in the Penninic zone are
also incorporated into the final model; a total of 8 layers were digitized
for gridding by Sierra Geophysics modeling software. Velocities and
densities were primarily obtained from laboratory analysis of rock
samples in the area [Sellami et al., 1990]. Normal incidence ray-tracing
(zero-offset approximation) was performed, incorporating surface
locations equivalent to the southernmost 50 km of the seismic profile.
To ease computation, the synthetic trace spacing is 400 m versus 40 m
for the real data; i. e., each synthetic trace corresponds to ten real traces.
The resulting spike seismogram was convolved with a 10-45 Hz Klauder
wavelet to mimic the vibroseis sweep used in the survey, and displayed
with constant amplitude to facilitate comparison of reflection
geometries. Comparison to the real data suggested modifications to the
model structure, and the process was iterated to produce an optimum
match.
The resulting synthetic section exhibits reflection patterns remarkably
similar to the actual data, both in overall appearance and in detailed
geometry. In fact, nearly all prominent reflections in the first seven
seconds can be explained by reflections from the interfaces in the model
(Fig. 7). Some of the highest amplitudes recorded are north-dipping
reflections between one and four seconds at the south end of the line,
which can be ascribed to backfolding in the Suretta, Tambo, and
Chiavenna units. At deeper levels, a dramatic change in reflection
character of the Adula top can be attributed to the presence of carbonates
overlying Adula basement. The termination of the Simano nappe is
marked by a distinctive "bow-tie" pattern often indicative of a buried
Structure map, base Suretta nappe
km 740 750 760
170
160
150
140
x Reflection points
Co)tour interval 1 km
Fig. 6. Structure contours of the base of Suretta nappe. Solid contours
are based on surface mapping [Pfiffner et al., 1990a]; shaded contours are
extrapolated based on seismic results (this study). Synthetic reflection
points (X) depict the eastern limit of subsurface control. Southern
portion of seismic line midpoints is shown by heavy line. Contour
interval is one km. Coordinates are in Swiss National km-Grid system.
NFP20-East, south end
2.0
..; ... :e.,.8.,..-.-.::-.,,..'.>,,',-:c..'c=-., .-.?.-...-::..,:., ......... .-._.y.'---<.Z:'..';-.',-..::.z.?.:':.,,:,d:.x:'. ::::2 ..'&"-5'-?;''z"r-Y',-.-f ::".
'...':' :'..::-',?,<' :;: ;' :...::.x..:*2`.... .:L..:-;`......`..;.:....&.:.:.-..::..:: :..*...::.: : ....-*?...: :..: .:>...`....: :`-&`... .. -.....`:.v.: : z b..-.-;`` ..S-:.::;- .:: ,4.0
'...-.-':,'-wr..:..'--. .... .,r.-.:.: :.; :-:,.e: ..... .:::?"." ...x....:.....:?.:.z.::z....`.%.:::;.:.:..-`--..?-:::*..:`::`--:;-..: ' ':"G._ _..:i*>-..-.-'--;-':..?-'. '.:----'.. -"'-'--?'.--',.
,, ,.c.'. ,..-;";',,-'.....'>::-:d:: .:,, . z,o, :'::':. ..::- < .... '.: .;. <:.' ..... r?::-'-->-. ,-.-);.-.'-J---,,-.: .... '; :-..:..:.h:-:;,.,.--.:-.'-::.&:..; ./.' ..... .-.:,2:~..
6.0
10 km
NFP20-EAST. SOUTH PART. [dZTH SYNTHETIC OVERLAY
<- NORTH SOUTH - )
c Ho.
111 1:21 11 11 151 1,1 17'1 11 11 :201 ' 11 ''1 531
........... L ..................... ;;i ;; ',;; ;;: ',: ',;:: ', ', ',:: " :::: "::: ' 0.0
5 ..... ?;.. ......
r... '- (k:Z ': -, .i.[-':; '.'';-::' [-f.J ...... : ' '-..--'.. "z._ '.-. .. - - . ;: ' ' "
:'." ,.:'w,?..::w'2 ' '':(;..-9'2:. ---'.:_.. ...5':- i. ;;.''.,. "": ' .
.: . ..... ?-: ...- .,-. .-;,._;-m.. ,-...=:: ............ . '.,_.... .--_.. .'. ,:f.. -.:- -..
....... --... ...... . . .... ..,. .. ,. . ,. . .-- .-...2 ,...;_ . -... ...... . ..... . .... .- -.- ..... ---.. ;- ,- ;. .;z ,-
- .. ,,' ....... ':=. ,,-..,-..-:-:,-'.'- . -"., ':; .- ,..> ...' :.r- ' ... , .... -.-:;.... '' ' ..,x .... -' .... . ,..- ,- ..
.. . .:,; v.,.n-. ..- ......-..,x.- w.. ...... ... .-t.:-.- .......-...-...-.-.--...... ..... ; .... .- ... .-. -...- .
:-]-" :X:,' s::,':..:-.,.".Z.:r.:< :- c-:.-': O ": ' ::-/;: :,::,:"-.:-'.',..': .'.?::r..'....:r.'."-c,,, " -Z:cz=.-= n
' ..... - .... ............. ' ..... :'" ':-'- .... ' ......... ;ez .... ?--." ........... ." . ......... :-::.-. ,' .., . ......... k-;....,' ..... =-'::
': ;1 :.;- ::77 i . :.:..? ;,: :.- ...... - ;-c::.: W.%';:; :",:! '",- :.: _J: ';:. .:;; : :':Z ;<= _i *:' '-- :' -/;' ' '"' "'' '. ',-
. - .=.:,;..",,:.: -.... : ...... .. ..... .'.-.-,..%,. .'.. ;- , '** '';..' .x. .: .:. :-':..*-:.;:r-: '..'.--:..-'. ' ....... ....
:-, ;:::. ;i. , :, : - :': r:- ; ::: :- :.i :: :. ;. :?... '::' :-':: :.: .; .: ::?. : . ::'.' :--' '. :.?/.::: ' ? :':: '.'%.:' ',
:.::::.%:..t?.%.?...:?:;5..f?:.:::;:/.4.7?:`.: .'? .":z2?z:.::4z:;?' -:. '..z>' .:::, ?:':-i;';::?A5?:'.S . -2;- :5t ,,
-:I-.V. ;::.',.':-..' ,' ..... -:,-...-.:L;.: :::z':..,..- .... > .... ,' ........ ' ...... . ...... .-...,.. ;.:-:.-:-'.;- .',-,.'. Ihn, ."-.-.. '--,..": .... ..... ,.;?.::.< :;.........; o.
t.y:-' :?;::::?::' ::'7'7-.?.'' :5. ':; ::':.'t.-v:3; .. :: ::.' .=:.z.?: ::':'.' .: ;.?' . 5::'ZS.::x'F:t::::5?? ". ".?' ,. ;*': .',' ';"t '.'/'.'' ':." '?. ;z;? *',r *:Lt .-:.:::-dyt-c :..
:-'6:.':-;+'';; .;:?', ;.'::;..:.':.'.-;: o' .:. '. ::...i :...: "''-.'zo':+ ' s" .::,.::5 .,C;L'.:':xe
-:T:'5;{:'yLc5.Z';;:'.:;.: *' .*'.::_''.':';"':*, _:.t: C,. ;-.:,:.:?.t. ;L:::.::'.**.''.. '. ' ..... . .... :.- ...... *. . *.- '-,: .:;t:. .. : ;, .' '<:-,,: ;5,';:. :. .... ". =.._ =''--t --e .... :.. 8.0
10 km
Fig. 7. a) Southern portion of seismic line NFP20-East, final unmigrated stack. b) Same as (a), with constant amplitude
synthetic seismic section superimposed. Geometry of synthetic reflections matches that of most of the real reflections.
focus syncline; the synthetic data closely approximates the observed
reflection geometry. A very strong reflection between 3.2 and 4.0 s
from synthetic CMP 149 to 162 probably indicates the top of the Aar
Massif, although its extension to depth is somewhat problematic.
Amplitude variations are not considered here, but probably signify
variable lithology [Pfiffner et al., this volume]. More detailed
correlations between real and synthetic data can help to constrain the deep
structure of the Penninic zone [Litak et al., unpublished ms.].
Variations in reflectivity observed in this area may have important
implications for interpretation of deep crustal reflections. In particular,
LITAK ET AL. 305
it appears that the presence of carbonates between crystalline basement
nappes has a profound effect upon the reflection response. Conversely,
major basement faults may, at least in some cases, be essentially
nonreflective unless they juxtapose substantially differing lithologies,
and could remain seismically undetected.
Using surface geology to launch seismic modeling can not only
provide a quantitative test of interpretations, but also associate specific
geologic features with individual reflections. Under favorable
circumstances, this can be a very effective interpretational strategy.
Moreover, lessons from studies such as this may be of more general
applicability. For instance, in both areas, reflections are interpreted to
be primarily generated at lithologic boundaries: a volumetrically minor
amount of carbonate or mafic material within granitic gneiss is primarily
responsible for the observed reflection patterns, a result also obtained at
the Siljan borehole site [Juhlin, 1990]. Of course, this finding does not
preclude contributions to reflectivity from other sources (e.g. fluids,
shear zones), but does emphasize the importance of lithologic variations.
In these instances, the observed reflections are probably associated with
reflection coefficients of about 0.05-0.10.
In addition, distinctive reflection patterns associated with different
types of reflectors may be recognizable elsewhere. As discussed above,
the BRS may furnish an analog for "layered" upper crustal reflections
observed in many areas. Understanding of reflectivity in the Alps,
which is sometimes considered a prototype of continental collision,
could aid in interpretation of older mountain belts, such as the
Appalachians [Green et al., this volume] and Variscides [Meissner and
Bortfeld, 1990], and can perhaps be applied to even older, more cryptic
orogenies (e.g. Grenville). Although substantial variability in geology
and reflection patterns exists, illustrations of such examples provide
insight toward a comprehensive understanding of the causes and
significance of crustal reflections.
Acknowledgements. Discussions with K. Howard, L. Levato, S.
Sellami, J.-J. Wagner, and R. Olivier were extremely helpful. Litak was
partly supported by an Amoco Foundation Doctoral Fellowship; the
U.S. National Science Foundation grant EAR 89-16129 and the Swiss
National Science Foundation project NFP-20 also provided financial
support. Sierra Geophysics and Geoquest International provided
modeling software. Institute for the Study of the Continents contribution
no. 157.
References
Brewer, J. A., L. D. Brown, D. Steiner, J. E. Oliver, S. Kaufman, and
R. E. Denison, Proterozoic basin in the southern midcontinent of the
U.S. revealed by COCORP deep seismic reflection profiling,
Geology, 9, p. 569-575, 1981.
Davis, G. A., J. L. Anderson, E.G. Frost, and T. J. Shackelford,
Mylonitization and detachment faulting in the Whipple-Buckskin-
Rawhide mountains terrane, southoastern California and western
Arizona, in Cordilleran metamorphic core complexes, edited by M.
D. Crittenden, Jr., P. J. Coney, and G. H. Davis, Geological
Society of America Memoir 153, p. 79-130, 1983.
Goodwin, E. B., G. A. Thompson, and D. A. Okaya, Seismic
identification of basement reflectors: the Bagdad reflection sequence
in the Basin and Range province - Colorado Plateau transition zone,
Arizona, Tectonics, 8, p. 821-831, 1989.
Green, A. G., L. Levato, P. Valasek, L. DuBois, R. Olivier, S.
Sellami, J.-J. Wagner, B. Milkereit, and C. Spencer, Comparison of
reflection patterns in the Swiss Alps, northern Appalachians, and SE
Canadian cordillera, in Geodyn. Ser.(this volume), edited by R.
Meissner, AGU, Washington D.C., in press, 1991.
Halls, H. C., and W. F. Fahrig, eds., 1987, Mafic dyke swarms,
Geological Association of Canada, St. Johns, NF, Canada.
Hauser, E. C., and J. E. Oliver, A new era in understanding the
continental basement; the impact of seismic reflection profiling, in
Composition, structure, and dynamics of the lithosphere-
asthenosphere system, edited by K. Fuchs and C. Froidevaux,
Geodynamics Series,16, American Geophysical Union, Washington,
1987.
Hauser, E. C., J. Gephart, T. Latham, J. Oliver, S. Kaufman, L. Brown,
and I. Lucchitta, COCORP Arizona transect: strong crustal
reflections and offset Moho beneath the transition zone, Geology,
15, p. 1103-1106, 1987.
Howard, K. A., J. W. Goodge, and B. E. John, Detached crystalline
rocks of the Mohave, Buck, and Bill Williams mountains, western
Arizona, in Mesozoic-Cenozoic tectonic evolution of the Colorado
River region, California, Arizona, and Nevada, edited by E. Frost
and D. Martin, Cordilleran Publishers, San Diego, p. 377-390, 1982.
Hurich, C. A., H. Palm, D, Dyrelius, and Y. Kristoffersen, Deformation
of the Baltic continental crust during Caledonide intracontinental
subduction: views from seismic reflection data, Geology, 17, p.
423-425, 1989.
Juhlin, C., Interpretation of the seismic reflectors in the Gravberg-1
well, in Deep drilling in crystalline bedrock, Vol. 1, edited by A.
Bod6n and K. G. Eriksson, Springer-Verlag, Berlin, p. 113-121,
1988.
Juhlin, C., Interpretation of the reflections in the Siljan Ring area based
on results from the Gravberg-1 borehole, Tectonophysics, 173, p.
345-360, 1990.
Litak, R. K., Mafic intrusions and the origin of strong crustal
reflections: seismic modeling of the Bagdad Reflection Sequence,
Arizona, M. S. Thesis, Cornell University, Ithaca, NY, 1990.
Litak, R. K., and E. C Hauser, Mafic intrusions and strong crustal
reflections: A Seismic modeling study of the Bagdad Reflection
Sequence (abstr.), Eos Trans. AGU, 70(15), p. 400, 1989.
MacDonald, G. J., Major questions about deep continental structures, in
Deep drilling in crystalline bedrock, Vol. 1, edited by A. Bod6n and
Matthews, D., and the BIRPS group, Some unsolved BIRPS
problems, Geophysical Journal of the Royal Astronomical Society,
89, p. 209-216, 1987.
Meissner, R., and R. K. Bortfeld, eds., DEKORP-Atlas, Springer-
Verlag, Berlin, 1990.
Pfiffner, O. A., W. Frei, P. Valasek, M. Stituble, L. Levato, L. DuBois,
and S. M. Schmid, Crustal shortening in the Alpine orogen: results
from deep seismic reflection profiling in the eastern Swiss Alps, line
NFP 20-EAST, Tectonics, 9, p. 1327-1356, 1990b.
Pfiffner, O. A., E. M. Klaper, A.-M. Mayerat, and P. Heitzmann,
Structure of the basement-cover contact in the Swiss Alps, in Deep
Structure of the Alps, edited by F. Roure, P. Heitzmann, and R.
Polino, Mdm. Soc. gdol. Fr., Paris, 156; Mem. Soc. gdol. Suisse,
Zurich, 1; Vol. Spec. $oc. geol. It., Rome, 1, p. 247-262, 1990a.
Pfiffner, O. A., L. Levato, and P. Valasek, Crustal reflections from the
Alpine orogen: results from deep seismic profiling, in Geodyn.
Ser.(this volume), edited by R. Meissner, AGU, Washington D.C.,
in press, 1991.
Pratt, T., R. Culotta, E. Hauser, D. Nelson, L. Brown, S. Kaufman, and
J. Oliver, Major Proterozoic basement features of the eastern
midcontinent of North America revealed by recent COCORP
profiling, Geology, 17, p. 505-509, 1989.
Reston, T. J., Evidence for shear zones in the lower crust offshore
Britain, Tectonics, 7, p. 929-945, 1988.
Sellami, S., F. Barblan, A.-M. Mayerat, O. A. Pfiffner, K. Risnes, and
J.-J. Wagner, Compressional wave velocities of samples from the
NFP20-EAST seismic reflection profile, in Deep Structure of the
Alps, edited by F. Roure, P. Heitzmann, and R. Polino, Mdm.
Soc. gdol. Fr., Paris, 156; Mem. Soc. gdol. Suisse, Zurich, 1;
Vol. Spec. $oc. geol. It., Rome, 1, p. 77-84, 1990.
REFLECTIONS FROM A RELIC MOHO IN SCOTLAND?
David B. Snyder
BIRPS, Bullard Laboratories, Madingley Road, Cambridge CB3 0EZ, U.K.
Abstract. Prominent reflectors are observed at about 30 and 50 km
depths throughout much of northern Scotland. The deeper reflector results
from impedance contrasts here attributed to metasomatic intrusions: high-
pressure, relatively low-density minerals such as phlogopite or kyanite in
a local zone that has survived due to a decreased geothermal gradient
following the Caledonian orogen. This metasomatism may be localized
by a shear zone, a delamination scar or by a former base to the crust. The
position of the Moho is here assumed to be controlled largely by isostatic
forces driven by heat. Garnet formation in the gabbroic, deep lower crust
of Caledonian Scotland produced eclogitic metamorphic rocks with seismic
velocities of 8 kms-1. These rocks are part of the Scottish lithosphere
between 28 and 50 km depths that today forms the seismic uppermost
mantle, capped by the Moho at 28-30 km. The relic metasomatic zone at
50 km depth marks the former Caledonian Moho, the boundary between
the gabbroic/eclogitic mid-lithosphere and the peridotitic uppermost
mantle lithosphere in this alternative model.
Introduction
Various seismological and petrological observations of the past decade
have led some earth scientists to call for a new, broader definition of the
boundary separating crust and mantle beneath the continents [Griff'm &
O'Reilly, 1987; Oliver, 1988; Jarchow & Thompson, 1989; Mengel and
Kern, 1990]. Vertical changes in seismic velocity that are observed over
depth changes of 100-1000 m on refraction profries and then correlated
with density changes inferred from gravity anomalies do not necessarily
produce the more localized (< 200 m) impedance contrasts that cause
reflections at Moho depths on normal incidence and wide-angle surveys.
However, these vertical changes have been noted to coincide, within limits
of resolution, in most places [Mooney and Brocher, 1988].
The present discussion will be based on the xenolith model of the
lithosphere [Griffin and O'Reilly, 1987]. A felsic, granitic (mostly
feldspar and quartz) crust is separated from an ultramafic, peridotitic
(mostly olivine and pyroxene) manfie by a mafic, gabbroic (mostly
feldspar and pyroxene) layer. This middle layer has varying thickness
beneath the continents dependent on the lithospheric age, temperature,
density-depth distribution, and the amount of mixing between this layer
and the overlying granitic layer and underlying peridotitic layer. In
addition, the gabbro minerals can metamorphose to garnet eclogite to
produce new density-depth distributions [Furlong and Fountain, 1986;
Mengel and Kern, 1990].
New interpretation of recent deep seismic reflection data from
northwest Scotland has provided tentative age constraints on reflectors in
the uppermost mantle and at the Moho (Figure 1) [Snyder & Flack, 1990].
In the present paper the seismic attributes that can be determined for these
deep reflectors will be summarized and models for the origin of the
reflections discussed. Two similar impedance anomalies have large
vertical spacing, occurring at 30 and 50 km depths (Figure 2). The
shallower feature has been interpreted as the modern Moho [Flack et al.,
1990]. The deeper feature has not been successfully explained, but would
be consistent with observations and tectonic models if it represented a
'fossil' Moho inherited from Paleozoic times. Here a justification of how
a Paleozoic Moho formed at the greater depth and how it survived to the
present day will be attempted by simplistic petrological arguments
involving migrating metasomatic melts and metamorphism of the mafic
middle lithosphere in the specific convergent, orogenic environment
envisaged for the Scottish Caledonides during the last 400 Ma.
Mantle Reflectors in Northwest Scotland
Observations
The continental lithosphere of northwest Scotland contains the best
imaged and most thoroughly studied mantle reflectors so far discovered,
and the present work benefits from the availability of both near-vertical
and wide-angle seismic profiling [Flack et al., 1990; Faber and Bamford,
1979]. Grids of deep seismic profiles now cover two areas, a 200 by 120
lan rectangle north of Scotland and a triangular area 200 lan on a side west
of Scotland (Figure 1A), and reveal similar reflectors in the mantle
lithosphere beneath each grid. Palinspastic restoration of these reflectors
based on inferred surface displacements indicates that 400 or more Ma ago
these mantle reflectors traced a continuous feature (Figure lB) [Snyder and
Flack, 1990]. The mantle reflector appears to be composed of two parts, a
subhorizontal one at c. 50 lan depth (the 'W' reflector ) (Figures 1B & 2)
and another dipping at about 30 between depths of 30 and 50 lan (the
'Flannan' reflector) [Wamer and McGeary, 1987]. The dipping part
parallels the mapped Caledonian front of NW Scotland. The subhorizontal
part may be projected beneath most of Scotland by correlating wide-angle
reflections observed on LISPB with the near-vertical ones described above
(Figure lB) [Faber and Bamford, 1979; Barton, i.n press].
Mantle reflectors from offshore areas north of Scotland produce normal
incidence reflections (Figure 2) requiting impedance contrasts of c. 20%
within the mantle [Wamer and McGeary, 1987], that is, density and p-
wave velocity contrasts combine to produce reflection coefficients of about
0.1. LISPB super-critical-angle reflections beneath mainland northern
Scotland are high-amplitude phases that arrive closely after those from
diving rays refracted in the uppermost mantle [Faber and Bamford, 1979].
The observed high amplitudes can be generated from post-critical
reflections off a downward increasing impedance contrast within the
uppermost mantle [Barton, in review], although a zone of increased
velocity several kilometers thick with no strong velocity gradients above,
but a steep negative gradient below, can simultaneously explain the
observations of slow lateral propagation, the absence of head waves, and
the large amplitudes [Faber and Bamford, 1979]. Thus it has not been
possible to determine whether velocity increases step-wise with depth or
decreases within a local layer to produce these reflections.
Geodynamics 22
Viable Sources of Mantle Reflections
Numerous geologic models have been proposed to explain reflections
from the lithosphere. Anisotropy and compositional (SiO2 content)
307
308 RELIC SCOTTISH MOHO?
Fig.1. Location maps of northern Scotland showing (A) location of grids of deep seismic profiles, the LISPB refraction profile, and
locations where mantle reflections have been observed (star denotes location of sections in figure 2), and (B) map palinspasfically
restored at end-Caledonian times [Snyder & Flack, 1990]. Contours indicate depth to dipping mantle reflector in kilometers, dashed
where less certain; isolated numbers are depth to subhorizontal manfie reflector in kilometers. FCL, Fair Head-Clew Bay line; GGF,
Great Glen fault; HBF, Highland Border fault.
contrasts are preferred reflection sources among models synthesizing fault
zones within crystalline rocks of the upper crust [Christensen and
Szymanski, 1988]. Fluids or volatiles [Matthews and Cheadle, 1986],
shear zones [Reston, 1988], gabbroic intrusives [Warner, 1990a], and
layered ultramafic cumulates [Collins et al., 1986] have been favored as
sources of subhorizontal, layered reflections in the lower crust of western
Europe. Anisotropy and high-silica-content models use crustal minerals
inappropriate for the ultramafic uppermost manfie. Thus generally viable
models of mantle lithosphere reflectors, specifically appropriate for
Scotland, can be simplified into three non-exclusive groups: metasomatic,
igneous, and structural.
Volatile-rich mantle melts at very low melt fractions (<<1%) can
separate from their matrix to flux and metasomatize the mantle lithosphere
[Menzies et al., 1987]. Pre-existing structures such as shear zones or
lithologic boundaries may concentrate such melts to produce zones of
enhanced impedance contrasts [inferred from Downes, 1990]. Geochemical
studies of late-Caledonian lavas and post-orogenic granites in northern
Scoffand indicate a source region for the rocks that was rich in trace
elements such as Sr, Ba, and K, a source region related to Andean-type
sulxluction [e.g Thirlwell 1981; Watson 1984 and references therein].
Watson [1984] attributed the onset of magmatism about 435 Ma ago to
post-collisional tectonism dominated by strike-slip movement parallel to
the orogenic belt. Vertical uplift had been underway for some 50 Ma. An
elevated geothermal gradient in the asthenospheric wedge above the
sulxluction zone and upward migration of a small-percent mantle melt
partially melted the Caledonian root, a likely lower crustal source for the
magmas (Figure 3). In Watson's preferred model, melt from sulxlucted
oceanic crust and related aluminous sediments invades and metasomatizes
the uppermost mantle during active sulxluction. During cessation of
convergence and subsequent collapse of the orogenic belt [Dewey, 1988],
partial melting in this altered, elevated mantle [MCKenzie and Biclde,
1988] gives rise to volcanic and plutonic suites with arc characteristics,
but unusual trace-element content [Saunders et al., 1987].
Igneous models for mantle reflectors involve production of gabbroic
sills by uppermost mantle melting [Furlong and Fountain, 1986] or
kimberlitic or nephelinitic sills produced by melting near the wet solidus
[MCKenzie, 1989]. The layered structure observed in ophiolites may have
partly inspired these models of mafic sills in continental lower crust
SNYDER 309
2O KM
GRID 7
lO
20 s
DRUM
W x7 E
Fig.2. Deep seismic sections from north of Scotland showing high amplitude
Moho (M) and mantle (W) reflections. Triangles at top indicate where the
lines intersect.
[Collins et al., 1986; Jarchow & Thompson, 1989]. In Scotland, the
proximity of Paleozoic subduction zones described above [Watson, 1984]
and large-scale Tertiary igneous centres [Mussett et al., 1988] increases the
likelihood of some igneous process having produced the reflectors.
Magma intrusion must generate sufficient density and velocity contrasts
over a large region. The observed seismic waveforms constrain these
contrasts to occur in layers > 100 m thick that have boundary zones
<200.m thick [Warner, 1990b] over areas covering thousands of square
kilometers (Figure 1). Models attributing the mantle reflectors to
intrusions of asthenospheric melt where lithosphere delaminated
[Houseman et al., 1981; Warner, 1989] require that heat loss by diffusion
exceeds its supply by the upward migration of melt. For melt to freeze
continuously within a layer 500 m thick, conduction must control
temperature distribution with the thermal Peclet number, Pet << 1 (also
see Figure 4) [McKenzie, 1989].
Structural models generally involve shear or fracture zones [Reston,
1988] that do not produce observed near-vertical reflections solely by the
mechanical modification (e.g., alignment of olivine) of the mantle rocks
[Warner and McGeary, 1987], but may achieve sufficient impedance
contrasts if the fractures acted as conduits for hydrous metasomatic fluids
(high-pressure brines) [Sanders, 1991], mineralizing fluids, or metasomatic
magmas or if they form repositories of the residuals [Griffin & Brueckner,
1985; Warner and McGeary, 1987; Austrheim & M0rk, 1988; Downes,
1990]. Precise determination of the composition of such metasomatic
materials within the mantle remains unlikely, but materials of positive
impedance contrast (eclogite-facies rocks, FeTi oxides and heavy metals),
negative impedance contrast (brines of several percent volume), or high
seismic anisotropy (aligned phlogopite) remain viable sources.
Observations have not yet resolved whether the reflections originate
from a layer of anomalous impedance or a step in impedance. The
structural models of shear zones clearly require discrete zones, several
meters to hundreds of meters thick, of anomalous velocity and density.
Igneous sills would likewise form a local zone, unless the sills
pervasively invaded a region of the lithosphere several kilometers thick.
In that case the base or top of this invaded region might appear as a step in
velocity and density to propagating seismic waves. A metasomatic front
typically results in a single change in velocity and density unless the front
coincides with a pre-existing barrier to the migrating fluids and volatiles.
That barrier could act as a lid to concentrate the metasomatic melt
components into local zones with percentages perhaps higher than 1%.
Is the Scots' Moho Unique?
Formation of a Relic Moho
Regional eclogitic layers and relatively thin phlogopite-rich zones are
favored candidates to form rock layers with sufficient impedance contrast to
produce the bright reflections observed beneath northern Scotland. A layer
with much interstitial phlogopite, phengite, or aragonite [Austrheim,
1987] could have formed at 50-km depths during fiuxing of 'orogenic'
lower crust by volatile (K)-rich metasomatic magmas and remained
metastable since the time when a cooler geotherm representative of a
shield was established (Figure 5). A 'low-velocity layer' would result.
The geochemistry of late-Caledonian lavas (1-4 weight percent K20) found
in northern Scotland [Thirlwell, 1981] is also compatible with the
existence of such a layer [MCKenzie, 1989].
These metasomatic melts may have cooled in sills or dikes along pre-
existing or coeval shear zones active at c. 70 km depths, near the Moho of
the late-Caledonian orogeny [Warner, 1989]. To produce the present-day
observed geometries the reflective zones must have been uniformly
uplifted 20 km as a passive marker horizon over all of northern Scotland.
My alternative model uses metasomatic zones formed at Caledonian
Moho depths (50-60 km), at the boundary between the marie (gabbroic)
and ultramafic (pefidofific) lithospheric layers, to produce the present-day
reflectors at 50 km depth. During the end of and following the Caledonian
orogeny, much of the marie layer metamorphosed to eclogite and
isostatically rebalanced the lithosphere [Meissner and Wever, 1986]
following the disappearance of the Caledonian mountains. A present-day
velocity-depth profile for northem Scotland [Barton, in press] can be
compared with a reconstructed Caledonian one based on those from the
Pyrenees, Alps, or Rockies [Meissner, 1987; Laubscher, 1988] to
illustrate the necessary velocity changes (Figure 6).
Gamer-forming reactions, insensitive to temperature, occur during
metamorphism of the unusually deep lower crust of orogenic belts (Figure
5) [Kay et al., 1990; Mengel and Kem, 1990]. Whereas gabbroic lower
crust is typically gamer-poor at depths less than 20 km (with a density of
about 2.8 g cm-3), abnormally thick crust will have significantly higher
density than gabbro due to the formation of garnet as an equilibrium
metamorphic mineral at these c. 50 km depths. If garnet-forming
reactions proceed to completion so that garnet and aluminous pyroxene or
amphibole replace plagioclase with the obliterafion of the transient Al-rich
phases, the resulting 'eclogific' rocks are denser than mantle peridotitc
[Austrheim & MCrk, 1988]. However, these reactions seldom reach
completion.
Subsequent processes of partial melting and further depletion of any
remaining feldspathic (gabbroic) components from the volume of lower
crust between 20 and 50 km depths occur during the time that the
mountain belt disappears. Calculations of melt mobility in the
lithosphere suggest that melt of 1% by volume could leave the orogenic
lower crust in 1 Ma (Figure 4) [McKenzie, 1984], coinciding with
multiple emplacement of granites into the upper crust. Reduced lithostafic
pressure during orogenic collapse allows feldspar-rich melt in the orogenic
lower crust at 40-50 km depths [by analogy to MCKenzie and Bidde,
1988] to migrate to a new, post-orogenic, level of neutral buoyancy at
depths of 20-30 km. The melt can form sills and dikes in the modern
lower crust, and may initiate granitic plutonism into the upper crust
Caledonian orogeny
0 400 800
80
1200C
PhIOI
Collapse & uplift
p=3.0
I
I
I
w
0
Present day
400
E
800 1200C
p=2.8
(vp=6.4)
MOHO
(vp=8.0)
p=3.3
(vp=8.5)
INTRUSIVES: ';3 Volatile-rich Gabbroic ,(:) Granitic
Fig.3. Cartoon showing proposed evolution of the lithosphere beneath the Caledonian orogeny. Watson [ 1984] proposed that volatile-
rich melt from subducted oceanic crust and related aluminous sediments invaded and metasomatized the uppermost mantle during
active subduction. During orogenic collapse, partial melting in this altered mantle gave rise to volcanic and plutonic suites with arc
characteristics. Several proposed geotherms and some mineral (ol, olivine; ph, phlogopite) phase boundaries (see Figure 5) are
superimposed. Present-day velocities are those observed, and the associated densities balance isostatically. 'W', mantle reflector; MT,
Moine thrust.
[Watson, 1984] (Figure 3). If garnet-forming reactions near completion,
the former lower crust, now garnet-rich and depleted of most feldspar, will
be denser than peridofite and have a similar, homogeneous 'mantle'
seismic velocity.
Kay et al. [1990] suggest that this dual metamorphism and partial melt
process leads to delamination of the lower lithosphere, an aspect
permissible but not necessary to the Scotland model. Palinspasfic
reconstruction of the dipping mantle reflectors [Snyder and Flack, 1990]
suggests that delamination did not occur at the more likely 30 km level,
above the dense garnet-rich zone, during or after Caledonian collapse.
Delamination may have occurred at the 50 km level [Warner, 1989]
leaving local, negative velocity/density gradients due to eclogite/garnet
amphibolite overlying peridotite.
Observations from the Norwegian Caledonides and Mantle Xenoliths
This model combining metasomatism and metamorphism of lower
crust is supported by recent studies of xenoliths erupted in kimberlites
[Schulze, 1989] and magmas and of exposed sections of high-grade
metamorphic rocks in the Caledonides of Norway. The proposed model
predicts the relative abundance of garnet-rich, eclogific rocks in sub-Moho
rock assemblages and the local occurrence of phlogopite and related
residnal. of volatile-rich melts. Many eclogite xenoliths worldwide
[Arculus et al., 1989] have a basaltic composition whose diversity appears
to have been established when garnet was not present. These xenoliths
once had crustal density, but metamorphosed to mantle density during
continental convergence wherever thickened basaltic crust deformed in a
relatively cool environment with water present [e.g., Laubscher, 1988].
Evidence of partial melting or fluxing by volatile-rich metasomafic melts
occurs in the preferential disaggregation of peridotite attributed to
alteration of orthopyroxene to stmcturally weaker phlogopite prior to the
eruption of Group II kimberlites. K-rich fluids or magmas infiltrated and
metasomatized the garnevperidotite-dominated host rocks, resulting in
abundant secondary phlogopite in eclogites from several Group II
kimberlites localities [Schulze, 1989].
Eclogitic rocks occur within the Western Gneiss Region of the
Scandinavian Caledonides; most formed by in situ metamorphism within
the base of an overthickened crust 410-420 Ma ago [Griffin & Brueckner,
1985; Austrheim, 1987; Austrheim & M0rk, 1988]. Sr and Nd isotopic
data suggest that the rocks had complex crustal histories prior to the
eclogite-forming metamorphism and many samples cannot be derived from
the mantle [Griffin & Brueckner, 1985]. Eclogite-facies rocks with
abundant hydrous phases and relic granulite-facies rocks alternate on a
meter-scale, and 50-100 m thick zones of banded eclogitic rocks can be
followed for more than a kilometer in outcrop. Eclogites account for 30-
45% of the rock volume [Austrheim, 1987]. The metamorphism of
granulites to eclogites was likely facilitated by fluid infiltration along
shear zones that resulted in the more ductile eclogite flowing around
incompletely reacted coronitic granulites [Austrheim & M0rk, 1988;
Downes, 1990]. Thus eclogitic rocks have reached the surface in one
region of the former Caledonian orogeny, several hundred kilometers along
strike from Scotland where similar rocks are hypothesized to remain at 30-
50 km depth.
Discussion
Early models of the crust-mantle boundary used refraction results to
define a single, discrete increase in velocity, the Moho [Jarchow &
th keb,g Wo ..................
h=20km ...........
:: :.:+:.:.:
k e =( 1 mm) 2 f n C'1 ........
Ap = 700kg rrr 3 ..............
' 8
( 6
.o
4
o
2 it wet
mel ite
1 1 oo 1 oooo
Viscosity, (Pa s)
Fig.4. Plot showing the time required, th, before only 1/e of the original melt
remains within a layer between 30 and 50 km depths (thickness, h) for
variations in melt fraction, (l), and viscosity, It, [McKenzie 1985,1989]. Values
for n and C in the equation for matrix permeability, k(l) ' as discussed in these
references. Density contrast between melt and matrix, Ap, is assumed. Wo
is the volume flux of melt. Several temporally distinct melt migrations
between these depths, each representing several per cent melt, are permitted
within the --.10 Ma available (non-stippled area) during end of the Caledo-
nian orogeny.
Thompson, 1989, for review]. Gravity studies also adopted such a simple
boundary. Within the past decade, wide-angle reflections have demanded
more complexity of the Moho. Preferred models consist of a laminated
zone with alternating high and low velocities (6-7.5 km -1) superimposed
on a velocity gradient between depths of 15-30 km [Wenzel et al., 1987],
with individual layers each about 100 m thick [Paul & Nicollin, 1990].
Near-vertical reflection studies have meanwhile indicated impedance
SNYDER 311
variations on the scale of 50-150 m at similar depths to the impedance
contrasts inferred by earlier refraction and gravity models [Mooney &
Brocher, 1987]. Petrologists have suggested rock suites that provide
suitable velocities and densities to explain the seismic observations [e.g.,
Griffin & O'Reilly, 1987]. All these models place the depth of the Moho
within a few kilometers of one another. Beneath northern Scotland two
bright reflectors with Moho characteristics lie 20 km apart.
This paper presents interpretations built upon observations and
inferences made from deep seismic data from northern Scotland discussed
elsewhere [Snyder and Flack, 1990]. Presented in order of decreasing
certainty, these interpretations can be summarized thus:
(1) Two equally bright seismic reflectors occur beneath northern Scotland
at c. 28 and 50 km depths; the shallower one coincides with the modern
Moho.
(2) Successful palinspastic restorations indicate that the deeper reflector
occurs over most of northern Scotland and is at least 400 Ma old (late
Caledonian).
(3) Only in areas immediately northwest of Scotland has Penno-Triassic
extension occurred on a shear zone nearly coincident with part of the deeper
reflector but continuing down to 80 km depth.
(4) During the late Caledonian orogeny, garnet-poor gabbro in the
orogenic lower crust metamorphosed to garnet-pyroxene eclogites and a
small percent (<10%) melt migrated upwards to 20-30 km depths (as
observed today) leaving behind lower crust with upper mantle seismic
velocities and no layering because of the lack of a low-velocity
component.
(5) The 50 km reflector derives from a zone several hundred meters thick
with impedance contrasts formed by the juxtaposition of this eclogitic
lower crust against metasomatic melt residuals rich in phlogopite, kyanite,
or metastable non-eclogized crustal rocks. This reflector coincided with
the Caledonian Moho 400 Ma ago and represented a significant downward
increase in density that provided a level of neutral buoyancy for upwardly
migrating volatile-rich metasomatic magmas. This density filter
concentrated volatile-rich melts producing metasomatic minerals such as
phlogopite and kyanite [Lappin, 1978] in a localized zone that has
survived due to a decreasing geothermal gradient. Some aspects of
Scotland's lithosphere may indeed be unique in that a collapsed orogen
now occupies a position on a submerged continental margin and that
delamination of former lower crust did not occur due to insufficient
thickening, large-scale dismemberment of the orogen, or the unusual
lateral strength of the Caledonian lithosphere. The model described here is
40
80
120
TEMPERATURE
300 900 1500C
', ..................... ! ......... ./ I I
. ===============================
. oot :::::::::::::::::::::
. .,' ...................
'- r,, ........................................
:::::::::::::::::::::::::::::::::: w C
Fig.5. Phase diagram for phlogopite, kyanite (grospydite is a high pressure,
kyanite-bearing rock) and peridotite [Modreski and Boettcher, 1972; Lappin,
1978]. Garnet-in and plagioclase-out curves taken from Furlong and
Fountain [ 1986, Fig. 6]. Several representative geotherms, labelled Basin &
Range (B&R) and "orogen" from Griffin and O'Riley [1987] and for
continental shields with thermal lithospheric thickness (h) indicated
[MCKenzie, 1989] are superimposed as well as a likely location of a
metasomatism zone [MCKenzie, 1989].
10
.-.20
30
40
50
sent-day northern Scotland
.15
.23
"Caledonian"
.2 .7 .1
6 7 8
Velocity (krn/s)
Fig.6. Comparision of velocity-depth curves for present-day northern
Scotland (squares, light stippled pattern and associated heavy curve) [Smith
and Bott, 1975; Barton, in press] and the "Caledonian" mountain belt, to
show changes in velocity required to make former orogenic crust appear
seismically as uppermost mantle. The "Caledonian" curve was compiled
from analogous modern data from the Alps, Pyrenees, and Rockies [Meissner,
1987]. Black horizontal lines indicate the mixing required between two
fractional end member components to produce the present-day observed
velocity (solid circle); each low-velocity component is compatible with the
"Caledonian" velocity at that depth.. The shaded background pauern shows
an inferred velocity structure [Wenzel et al., 1987] that can: (1) produce the
observed average velocity, (2) incorporate the end-member velocities
modelled here, and (3) provide impedance contrasts required to produce
reflections observed in northern Scotland.
perhaps more complex than equally viable models of shear zones and
lithologic changes, but attempts to adequately encompass recent field
observations from throughout the Caledonides.
Acknowledgments. Comments from many colleagues helped clarify
ideas presented here. Specialized additional processing and display of the
seismic data utilized the Merlin SKS system running on the Bullard Labs
Convex computer. BIRPS is funded by its Industrial Associates Program
and the Natural Environmental Research Council. Cambridge Earth
Sciences contribution number 1948.
References
Austrheim, H., Eclogitization of lower crustal granulites by fluids
migration through shear zones, Earth Planet. Sci. Lett., 81,221-232,
1987.
Austrheim, H., and M. B. MOrk, The lower continental crust of the
Caledonian mountain chain: evidence from former deep crustal
sections in western Norway, in Kristoffersen, Y., Progress in studies
of the lithosphere in Norway, Norges Geologiske UndersOkelse Spec.
Publ. No. 3,102-113, 1988.
Arculus, R. J., J. Ferguson, W. Chappell, D. Smith, M. T. McCulloch,
I. Jackson, H. D. Hensel, S. R. Taylor, J. Knutson, and D. A. Gust,
Ecologites and granulites in the lower continental crust: Examples
from eastern Australia and southwestern U.S. A., in Eclogites and
related rocks, edited by D.C. Smith, Elsevier, NY, 1987.
Barton, P., LISPB revisited: a new look under the Caledonides of northern
Britain, submitted to Geophys. J. Int., July, 1990.
Christensen, N. I., and D. L. Szymanski, Origin of reflections from the
Brevard fault zone, J. Geophys. Res., 93, 1087-1103, 1988.
Collins, J. A., T. M. Brocher, and J. A. Karson, Two-dimensional
seismic reflection modeling of the inferred fossil oceanic crust/mantle
transition in the Bay of Islands Ophiolite, J. Geophys. Res., 91,
12520-12538, 1986.
Downes, H., Shear zones in the upper mantle-Relation between
geochemical enrichment and deformation in mantle peridotites,
Geology, 18, 374-377, 1990.
Dewey, J. F., Extensional collapse of orogens, Tectonics, 7, 1123-1139,
1988.
Faber, S., and D. Bamford, Lithospheric structural contrasts across the
Caledonides of northern Britain, Tectonophysics, 56, 17-30, 1979.
Flack, C. A., S. L. Klemperer, S. McGeary, D. B. Snyder and M. R.
Warner, The reflective upper mantle of the UK, Geology, 18, 528-
532, 1990.
Furlong, K. P., and D. M. Fountain, Continental crustal underplating:
Thermal considerations and seismic-petrologic consequences, J.
Geophys. Res., 91, 8285-8294, 1986.
Griffin, W. L., and H. K. Brueckner, REE, Rb-Sr and Sm-Nd studies of
Norwegian eclogites, Geochem. Geology, 52,249-271, 1985.
Griffin, W. L., and S. Y. O'Reilly, Is the continental Moho the crust-
mantle boundary?, Geology, 15, 241-244, 1987.
Houseman, G. A., D. P. MCKenzie, and P. Molnar, Convective
instability of a thickened boundary layer and its revelance for thermal
evolution of continental convergent belts, J. Geophys. Res., 86,
6115-6132, 1981
Jarchow, C. M., and G. A. Thompson, The nature of the Mohorovicic
discontinuity, Ann. Rev. Earth Planet. Sci., 17, 475-506, 1989.
Kay, R. W., S. Mahlburg Kay, and R. J. Arculus, Magma genesis and
crustal processing, in Geology and Geophysics of the Continental
Lower Crust, edited by D. M. Fountain, R. J. Arculus, and R. W.
Kay, Elsevier, in press.
Lappin, M. A., The evolution of a grospydite from the Roberts Victor
Mine, South Africa, Contrib. Mineral. Petrol., 66, 229, 1978.
Laubscher, H., Material balance in Alpine orogeny, Geol. Soc. Am. Bull.,
100, 1313-1328, 1988.
Matthews, D. H., and M. J. Cheadle, Deep reflections from the
Calendonides and Variscides west of Britain and comparison with the
Himalaya, in Reflection seisrnology: A global perspective, edited by
M. Barazangi, and L. Brown, Am. Geophys. Union Geodynam. Ser.,
13, 5-20, 1986.
MCKenzie, D., The extraction of magma from the crust and mantle, Earth
Planet. $ci. Lett., 74, 81-91, 1985.
MCKenzie, D., Some remarks on the movement of small melt fractions in
the mantle, Earth Planet. Sci. Lett., 95, 53-72, 1989.
MCKenzie, D., and M. J. Bickle, The volume and composition of melt
produced by extension of the lithosphere, J. Petrology, 29, 625-679,
1988.
Mengel, K., and H. Kern, Petrologic versus seismic Moho and crustal root
mysteries of the Hercynian orogen, in The European Geotraverse:
integrative studies, edited by R. Freeman, P. Giese, and St. MQller,
pp. 169-176, European Science Foundation, Strasbourg, France,
1990.
Menzies, M. A., N. Rogers, A. Trindle, and C. J. Hawkesworth,
Metasomatism and enrichment processes in lithospheric peridotites,
an effect of asthenosphere-lithosphere interaction, in Mantle
metasomatism, edited by M. A. Menzies, and C. J. Hawkesworth,
313-61, Academic Press, London, 1987.
Meissner, R., The continental crust: a geophysical approach, Orlando,
Fla., Academic Press, 426 p., 1987.
Meissner, R., and T. Wever, Nature and development of the crust
according to deep reflection data from the German Variscides in
Reflection seisrnology : A global perspective, edited by M. Barazangi,
and L. Brown, Am. Geophys. Union Geodynam. Ser., 13, 31-42,
1986.
Modreski, P. J., and A. J. Boettcher, The stability of phlogopite and
enstatite at high pressures: a model of micas in the interior of the
earth, Am. J. Science, 272,852-869, 1972.
Mooney, W. D., and T. M. Brocher, Coincident seismic
reflectionRefraction studies of the continental lithosphere: a global
review, Rev. Geophysics, 25, 723-742, 1987.
SNYDER 313
Musseli, A. E., P. Dagley, and R. R. Skelhorn, Time and duration of
activity in the British Tertiary Igneous Province, in Early Tertiary
volcanism and the opening of the NE Atlantic, Spec. Publ. Geol. Soc.
London, 39, 201-214, 1988.
Oliver, J. A., Opinion, Geology, 16, 291, 1988.
Paul, A., and F. Nicollin, Thin crustal layering in Northern France:
observations and modelling of the PMP spectral content, Geophys. J.
Int., 99, 229-246, 1990.
Reston, T., Evidence for shear zones in the lower crust offshore Britain,
Tectonics, 7, 929-936, 1988.
Sanders, I. S., Exhumed lower crust in NW Ireland and a model for crustal
conductivity, J. Geol. Soc. London, 148, 131-136, 1991.
Saunders, A.D., G. Rogers, G. P. Marfiner, D. J. Terrell, and S. P.
Verma, Geochemistry of Cenozoic volcanic rocks, Baja California,
Mexico: implications for the petrogenesis of post-subduction
magmas, J. Volcan. Geothermal Res., 32,223-245, 1987.
Schulze, D. J., Constraints on the abundance of eclogite in the upper
mantle, J. Geophys.l Res., 94, 4205-4212, 1989.
Smith, D. J., and M. H. P. Bott, Structure of the crust beneath the
Caledonian foreland and caledonian belt of the north Scottish shelf
region, Geophys. J. R. Astron. Soc., 40, 187-205, 1975.
Snyder, D. B., and C. A. Flack, A Caledonian age for reflectors within the
mantle lithosphere north and west of Scotland, Tectonics, 9, 903-
922, 1990.
Thirlwell, M. F., Implications for Caledonian plate tectonic models of
chemical data form volcanic rocks of the British Old Red Sandstone, J.
Geol. Soc., London, 138, 123-138, 1981.
Warner, M. R., Seismic reflections from a paleo base to the continental
lithosphere (abs.), EOS, 70, 1203, 1989.
Warner, M. R., Basalts, water, or shear zones in the lower continental
crust?, Tectonophysics, 173, 163-174, 1990a.
Warner, M. R., Absolute reflection co-efficients from deep seismic
reflections. Tectonophysics, 173, 15-23, 1990b.
Warner, M. R., & S. McGeary, Seismic reflection coefficients from
mantle fault zones, Geophys. J. R. Astron. Soc., 89, 223-230, 1987.
Watson, J., The ending of the Caledonian orogeny in Scotland, J. Geol.
Soc., London, 141. 193-214, 1984.
Wenzel, F., K-J. Sandemeier, and W. W'alde, Properties of the lower crust
from modeling refraction and reflection data, J. Geophys. Res., 92,
11575-11583, 1987.
THE REFLECTION MOHO ALONG THE COCORP NORTHWEST U.S. TRANSECT
Eleanor I. Prossen 1
Institute for the Study of the Continents
Cornell University, Ithaca NY 14853 USA
Abstract. Marked variations in the seismic reflection
character of the continental Moho occur on the COCORP deep
seismic reflection transect of the northwestern U.S. These
variations correlate with the four major tectonic provinces
crossed by the transect, the extended Cordilleran interior, the
Rocky Mountain thrust belt, the Precambrian craton, and the
western Williston basin. Variations in Moho structure and
reflectivity observed along this transect provide an important
constraint on processes and structures at the base of the crust.
The prominent reflectivity beneath both the Cordilleran interior
and Williston basin edge suggests that magmatic underplating
may have played an important role in the development of the
Moho in both regions.
Introduction
The COCORP (Consortium for Continental Reflection
Profiling) Northwest U.S. transect crosses the accreted
terranes of northern Washington, the region of documented
Eocene extension and exposed core complexes of the
Cordilleran interior, the Rocky Mountain thrust belt, the
Precambrian Wyoming craton, the Tertiary Bearpaw alkaline
intrusive complex, and the intracratonic Williston basin [Potter
et al., 1986; Yoos et al., in press; Latham et al., 1988] (Figure
1). This transect thus provides the opportunity to investigate a
range of reflection characteristics of the Moho in several
different, but contiguous, geologic provinces. To date,
characterizations of Moho variability have been generally
qualitative in nature and have often focused on one particular
'type' of observed reflection pattern.
1 now at: Scott Polar Research Institute, Cambridge
University, Lensfield Road, Cambridge CB2 1ER UK
Geodynamics 22
The Mohorovicic discontinuity was originally recognized as
an abrupt change in seismic P-wave velocities as obtained
from earthquake data. Modem seismic refraction and wide-
angle reflection observations variously suggest sharp,
gradational, or layered transitions at the boundary between
crust and mantle [e.g., Hale and Thompson, 1982; Braile and
Chiang, 1986; Mooney and Brocher, 1987]. In recent years,
near-vertical incidence seismic reflection surveys have
reaffkmed that the Moho represents a major boundary within
the earth, and yet suggest considerable lateral variability [e.g.,
Hale and Thompson, 1982; Oliver, 1982]. Although this
lateral variability has long been recognized [e.g., Steinhart and
Meyer, 1961], clear and detailed images of the structural
variation have remained elusive.
Coincident or adjacent reflection and refraction/wide-angle
reflection profiles in locations around the world show a
general correspondence in depth of refraction-defined Moho
and reflection-defined Moho [e.g., Mooney and Brocher,
1987]. However, differences in techniques and uncertainties
in resolution make it prudent to maintain the separate entities of
reflection Moho and refraction Moho, especially when only
limited data of any type are available for a region, as is.:the case
here. For this study, previously interpreted velocity-depth
models from refraction surveys in the northwestern U.S. and
southwestern Canada were converted to two-way travel time
and compared with the reflection data. It should be noted,
however, that these comparisons use data of varying ages and
distances from the reflection transect.
Two criteria are commonly used to identify deep reflections
as images of the Moho. First, the reflections typically mark a
prominent boundary between an overlying reflective zone
(crust) and an underlying non-reflective zone (upper mantle).
Second, the travel times to the reflection Moho are in general
agreement with available refraction results collected coincident
with or near the deep seismic reflection data. In this paper, we
show that Moho reflections generally agree in travel time with
features that were previously suggested by available refraction
data. In some regions, however, the reflection data simply
show the dying-out of a reflective zone (crust) at travel times
consistent with the depth to the base of the crust.
315
316 REFLECTION MOHO NORTHWEST U.S.
Data Acquisition and Processing
A total of 1050 km of seismic reflection data was collected
to form a transect across the northwestern U.S. (Figure 1).
These data were collected using the Vibroseis technique, and
were carried out over local roads using 4 or 5 vibrators in line
with a station interval of-100.6 m (330 ft) and off-end or
asymmetric split geophone spreads with a maximum far-offset
of approximately 10 km and a nearest group offset of 3 to 5
vibrator stations. Acquisition parameters were occasionally
varied along the transect.
Data processing for this study consisted of a standard series
of steps (see Zhu and Brown, 1987 for review of typical
processing of similar data). Some variations in the processing
sequences were dictated by variations in the data quality of the
individual seismic lines; in some cases, additional processing
steps were applied, typically before deconvolution. For
example, data with low-frequency noise (WA 7; ID 2; MT 1,2)
benefited from a low-cut bandpass filter (14-32 Hz). A
frequency-wavenumber filter [Yilmaz, 1987] was applied to
several of the lines (MT 1, 8) in order to remove low apparent-
velocity noise, improving the quality of the shallow data.
Editing of noisy traces improved the signal-to-noise ratio of
data collected in areas of high traffic noise. Lateral variations
in relative amplitude on the field data were removed by a
combination of noise-spike removal and inter-trace amplitude
balancing. The data were processed with a common datum of
1 km above sea level, and are displayed with an automatic gain
correction and coherency filter applied, and no vertical
exaggeration at a conversion velocity of 6 km/s. Although this
paper shows only unmigrated data, migrations of larger
portions of the transect have been presented in Potter et al.
[ 1986], Sanford et al. [1988], and Yoos et al. [in press].
Reflection Moho Variations
Cordilleran lnterior--A single Moho reflection
Moho reflections beneath the extended Cordilleran interior
are notable for their nearly constant travel time and apparent
continuity throughout the extended zone. Strong,
subhorizontal reflections occurring at 10.5-11.2 s on the
seismic profiles are prominent along portions of this segment
of the transect, with individual reflections laterally continuous
for up to 30 km. The Moho reflections are strongest on
Washington lines 1, 2, 5 and 8, and Idaho line 1, but are also
evident on Washington line 3. On the stacked sections, Moho
reflections characteristically are 2-3 cycle events having a
duration of 0.1-0.3 s, and locally dipping or horizontal events
occur immediately overlying them. Overlying dipping
reflections are most clearly seen on Washington lines 5 and 8,
and Idaho line 1 [Potter et al., 1986; Sanford et al., 1988]. In
general, the reflection Moho appears sub-horizontal
throughout the extended Cordilleran region, over a distance of
approximately 250 km. Apparent gaps in the reflection Moho,
in the form of vertical blank panels, can be related to surface
conditions [Prussen, 1989].
126
52 103 o
___----'52
'2. ' B'C.'-- \,',,'k,,, 3',, ALTA I SASK
'2 ' - , '
' . ' I oo
: 1 ....... MONT. 8
.... ', x 6
250 500
, I ,
< CENOZOIC VOLCANIC ROCKS MESOZOIC PALEOZOIC kilometers
_ACCRETED CRUST
CENOZOIC ACCRETED TER-
RANKS SEDIMENTARY ROCKS "CORE COMPLEXES" (PER-TERTIARY
BASEMENT Mz TERTIARY PLUTONS)
CASCADE VOLCANIC ROCKS PRE-MESOZOIC NORTH AMERICA WILLISTON BASIN
==2 MESOZOIC TERTIARY ISOPACHS
INTRUSIVE ROCKS (KOOTENAY ARC
SA SWEETGRASS ARCH
Fig. 1. Location map showing the COCORP Northwest U.S. transect and regional geologic features.
Seismic lines are shown with heavy lines and numerals.
PRUSSEN 317
Previous estimates of the depth to Moho within the
Cordilleran interior in northern Washington and southern
British Columbia range between 30 and 35 km [Hill, 1972;
Rohay, 1982]. Figure 2 compares Washington line 8 with the
velocity-depth model of Hill [ 1972] from an unreversed
refraction profile about 35 km east of the seismic reflection
line. The refraction results suggest two possible models for
the crust. The strong reflection Moho seen on the COCORP
lines closely corresponds with the refraction model which
yields a Moho two-way travel time of 11.0 s. Similar Moho
reflections appear on some of the data collected by the
Canadian Lithoprobe group about 100 km north of the
COCORP lines [Cook et al., 1988; Varsek et al., 1989], and
on a deep reflection line collected within the Omineca
crystalline belt about 350 km north of the COCORP data [Mair
and Lyons, 1976].
Rocky Mountain Thrust Belt--non-reflective Moho
COCORP seismic data from northwestern Montana in the
Rocky Mountain fold and thrust belt (ID 2, MT 1-2) show a
EXTENDED CORDILLERAN INTERIOR
s WASHINGTON LINE- t3 At
V P 300 2_00 1OO
G
M
Fig. 2. Unmigrated, coherency-filtered seismic section of
Washington line 8, a north-south oriented line within the
extended Cordilleran interior, shown with a comparison of the
refraction results of Hill [ 1972]. The 'M' indicates the
reflection Moho, and 'G' is the overlying reflection discussed
by Sanford et al. [1988]. The prominent reflection below 'M'
becomes a dipping lower crustal reflection upon migration.
PURCELL ANTICLINORIUM
W MONTANA LINE - 2: E
VP , 1OO 200 300 400
0 i i i i i i i i i
-;k-:-:,:(:'.-'-.'f'-c"y:-.*";' :-_ .: -,.- .C:.: ..;;....... t"-'--X: it--:_---?-'?-':'(
. .... - .... : :.-_-,,, r.,:...,.,-..:.-- ...... ;:---:'.-.,4,;:._?,'. -
-...--..-'-,-.,-:.....--_,: ....... :-- .... -: "' -'''-
'I ==''-::,-,k .... '="':-' .. ":'i-C''--J.g"--7.--:',2._ ' [-.?- 2-?-:'..:-:'-z:; .'-:-
' I `:......-``.`................`;.`-...`:.:::......`..:.`..;;::...:`!: 7.--:77.":::--':,;':-:.:.-2;:,:,-?-..; :7'='=-'"'--,. -,, .t.*
g::..--..,,::_-r_-.:&.::::-, , ---t .... -:.'L. "--.'- -..:; '. --::'-.'....---.<'-%-?.>:L :%.; '-
I :):'-:-'--'..' v., '-'--:;=--.-----:'--:.75 ;.: !'.---.-;'_:..::i;%;:('.5: ....
I0 krn 2 ---'--'-'- ! I .MT
Fig. 3. Unmigrated, coherency-filtered section of Montana
line 2. The crust beneath the western Rocky Mountain thrust
belt shows highly reflective crust down to travel times of 7-8
seconds, with almost no coherent lineups below that time.
highly reflective upper crust down to a travel time of 7.0-8.0
s. These reflections define structures within the thrust belt,
and the discontinuous, prominent reflection sequence that
marks the base of this reflective zone has been interpreted as
the basal d6collement of the thrust belt [Yoos et al., in press].
Below this sequence, the seismic section is notable for its
nearly complete lack of reflections (Figure 3). Regional
refraction data over the Rocky Mountain thrust belt have been
interpreted to indicate an eastward increase in depth to Moho
[Hales and Nation, 1973; Cumming et al., 1979]. The
velocity-depth model of Hales and Nation [1973] gives a depth
to Moho of 37.5 km, near the Montana/Idaho border, which
converts to a travel time of 12.3 s.
The absence of reflections below 7.0-8.0 s might be the
result of energy lost while penetrating the thick, highly
reflective fold-thrust belt [Yoos et al., in press] in the upper
crust [e.g., Waters, 1981]. It is not clear whether the lack of
lower crustal reflections beneath the thrust belt is a signature of
geologic differences we would expect between the thrust belt
and metamorphic hinterland, or whether it is simply an artifact
of complex raypaths.
Montana Plains (Craton)--Moho defined by a downward
cessation of crustal reflections
On the Montana plains data over the craton (MT 3-7), east
of the edge of the Rocky Mountain thrust belt, reflections
beneath the base of the sedimentary section (at 1.0-2.0 s)
extend downward to between 14 and16 s. Short,
discontinuous, subhorizontal reflections and occasional
diffractions occur throughout the crust. The Moho beneath the
eraton appears to be defined by a decrease in the density of
reflections, although locally, interpreted Moho reflections are
noted (Figure 4).
Regional refraction data from the north-central Montana
plains, on the eraton, have been interpreted in terms of a
variety of velocity-depth models [Meyer et al., 1961; Asada
and Aldrich, 1966]. The models from this older data set yield
two-way travel times to the Moho that vary between 12.5 and
18.2 s. Although comparison using these data sets is tenuous
because of the variation in the refraction models, the Moho
travel times calculated from the refraction models show general
agreement with the observed cessation of crustal reflections on
the COCORP data.
Williston Basin--Moho at base of reflection/diffraction band
Deep reflections beneath the western Williston basin are
unique along the transect. On Montana line 8, a 1.0-1.5 s
wide band of strong reflections and diffractions is observed
between 12.5 and 14.0 s. Individual events are laterally
continuous for up to 6 km. This band appears to be up to 1.0
s shallower at the eastern end of the line, occurring at 11.2-
PRECAMBRIAN CRATON
MONTANA LINE-4
W
VP 100 200 300 400
O- _, _.. __'_. _____
":--.-, ::.--' .---4.0 .:: 8.0 km/s %.,-<--.----;
'"F :. ':.." .:. --.. t : , --. ...z:.=.--.-.%--zz&;:"..:-:-- ::--- -----:""'-'-'----_. _"__ --- <_';---7.4:.-"7-'
-,:.:':,_L.. :l: :5:.'"i: ?-: _---:,:'_--':..: O:-. ':-/:":;--'::: :.- ::.-.::":,57.9 '?'!a. ';%1-;' -": -
-....-... .:: ..... -., . - -
5
lo-
M
15
Fig. 4. Unmigrated, coherency-filtered section of Montana
line 4 compared with the refraction results of Meyer et al.
[ 1961]. The refraction results show general agreement with
the cessation of reflections on Montana line 4. 'M' shows the
location of possible Moho reflections.
12.7 s from VPs 670-800. Some of the events within this
band are interpreted as diffractions on the basis of their
agreement with calculated diffraction curves, and also collapse
when migrated. Figure 5 compares Montana line 8 with recent
refraction data collected by COCRUST from the Williston
basin in southern Alberta. The velocity-depth model was
taken from an east-west trending line approximately 100 km
north of Montana line 8. The reflection / diffraction band on
Montana line 8 generally corresponds with the position of the
high-velocity zone (>7.0 km/s) modeled at the base of the
crust on the refraction data [Morel-t-l'I-Iuissier et al., 1987].
The anomalously high concentration of diffractions
indicates sharply defined (with respect to the signal
wavelength) discontinuities at the base of the crust in this
region. The refraction results, combined with the diffractions
on the reflection data, suggest a band of high-velocity discrete
bodies in the lowermost crust.
Discussion
Implications for Reflection Moho Development
The distinct differences in character of the reflection Moho
beneath the Cordilleran interior and the Williston basin are a
fundamental feature of the crust along this transect. The
observed variations can be compared to provide insights into
processes and structures at the crust-mantle transition.
The geometry of Moho reflections in the Cordillera suggests
a simple boundary for the base of the crust and may imply a
relatively sharp discontinuity 0.4-1.1 km thick (for a
conversion velocity of 7.0 krn/s). On the other hand,
elsewhere in the Cordilleran interior, a thicker and perhaps
more transitional zone is sugggested. Other workers have
suggested that the Moho may be a transition zone several
kilometers thick, inferred from both surface geologic relations
and geophysical information [Meissner, 1973; Hale and
Thompson, 1982; Braile and Chiang, 1986]. Although the
determination of a soling or truncation relation is difficult, the
termination of dipping reflections at the the reflection Moho, in
particular on Washington line 5 and Idaho line 1, suggests an
abrupt, non-transitional boundary that has developed via an
active process within the upper mantle. Similar, sharply-
defined Moho reflections observed in southern British
Columbia [Mair and Lyons, 1976; Varsek et al., 1989] may
suggest that the deep crust here has undergone similar
processes.
In contrast to the Cordilleran interior, beneath the western
Williston basin on Montana line 8, a 1.0-1.5 s band of
reflections and diffractions occurs with the reflection Moho
defined as the base of this zone. For a velocity of 7.0 krrds,
the thickness of this zone would be 3.5-5.3 km. The
refraction results, when viewed with the diffractions observed
on the reflection data, suggest a band of discrete high-velocity
bodies in the lowermost crust. These data thus suggest
addition of material, perhaps as discontinuous sills, to the base
PRUSSEN 319
WESTERN WILLISTON BASIN
MONTANA LINE -8
VP, 190 , 290 300 490 500 600
I I I _J I i
700 80O

. -.. .. _ . _ .,.._..:.
Fig. 5. Unmigrated, coherency-filtered section of Montana line 8, on the western edge of the Williston basin,
compared with the refraction results from line H of Morel-h-l'Huissier et al. [1987]. 'M' indicates the
reflection Moho at the base of this band.
of the crust. Furlong and Fountain [ 1986] suggested that
mantle-derived crustal material can add more than 10 km to the
thickness of the crust, via underplating of mafic magma,
although exact thicknesses and resulting seismic velocities are
dependent on the compositions and thermal histories of the
material. Mafic magma intruding the base of the crust as sills
might change the overall composition of the layer enough to
modify the seismic velocity observed on refraction data, and
cause the discontinuities observed on the reflection data. Other
regions have lower crustal zones which may be analogous to
the high-velocity layer below the Williston basin. U.S.
Glogical Survey seismic refraction data from the central
Columbia Plateau show an anomalous, high-velocity 'pillow'
(7.5 km/s) at the base of the crust beneath a thick sediment-
filled graben; this high-velocity layer is similar in shape and
velocity to those observed in several other continental rifts
such as the Mississippi Embayment, the East African Rift, and
the Salton Trough [Catchings and Mooney, 1988].
Studies of the subsidence history of the Williston basin
have postulated the existence of subcrustal loads to explain the
existence of the basin. Fowler and Nisbet [1985] postulate
that the essentially steady, non-exponential subsidence of the
basin can be explained by the presence of a mafic subcrustal
body undergoing transformation to eclogite, and calculate that
a sill-like layer of gabbro less than 5 km thick, with a diameter
of 400 km, would be needed. This model would be consistent
with the results observed from the refraction and reflection
data, in that a broad region intruded by material of gabbroic
composition could explain the observed velocity contrast, and
the high-amplitude and complexity of the lower crustal
reflections. Alternatively, the tectonic emplacement of a mafic
crustal slab during mid-Proterozoic plate convergence in the
region could also explain the anomalous layer [Fowler and
Nisbet, 1985; Green et al., 1985].
Unlike the data beneath both the Cordilleran interior and the
Williston basin, the reflection data from the Montana craton
show only sparse reflections at approximate Moho travel
times, as estimated from the refraction data. Similar reflection
character is observed on other data sets over stable
intracratonic areas having undergone little or no extension,
such as the Colorado Plateau and the Minnesota Precambrian
Shield [Meissner, 1986; Gibbs, 1986; Lundy, 1988].
Geochemical analyses of the sparse Eocene volcanics in the
Montana craton indicate mantle derivation and little crustal
contamination [O'Brien et al., in press]. This interpretation is
also supported by the absence of voluminous Cenozoic
volcanics as seen in the Cordilleran interior.
Numerous explanations of general Moho reflectivity have
been given in the literature: some have included
metasedimentary packages, cumulate igneous layering, lenses
of partial mantle melt, or a detachment or decoupling horizon
[e.g., Meissner, 1973; Hale and Thompson, 1982; Matte and
Him, 1988]. Seismic reflectivity modeling suggests that the
high amplitude of some lower crustal and Moho reflections
might not be caused by simple velocity contrasts, but can be
explained as the result of constructive interference [Braile and
Chiang, 1986; Hurich and Smithson, 1987]. A material
explanation for high-amplitude reflections could be provided
by voluminous early Cenozoic magmafism in the region [e.g.,
Armstrong, 1988] which might be manifested as discrete sills
or widespread intrusive zones at the base of the crust. In the
same sense, the prominent reflection Moho beneath the
extended Cordilleran interior can also be related to Cenozoic
magrnatism and extension in the region [Potter et al., 1986].
Wilshire [ 1990] presents a model for evolution of the crust-
mantle boundary in the southwestern Basin and Range which
suggests that extension affects both the upper mantle and
lower crust, and is temporally overlapped by multiple
magmatic episodes. This may be directly analogous to the
northwest Cordillera. Eocene extension in the northwest
Cordillera is recognized from structural and geochronological
evidence [for review, Parrish et al., 1988], and is
superimposed on deeply penetrating Mesozoic thrusts [Monger
et al., 1986; Brown et al., 1986; Price et al., 1985]. For the
Williston basin, no analogous surface geologic evidence
supporting major extension or rifting exists; however, the
geophysical data are consistent with somewhat thinner crust,
relative to the craton, beneath the basin.
Implications for a Younger Moho
Two opposing styles of Moho reflections have been
observed on deep seismic reflection data worldwide. Regions
where the data show crustal structures that crosscut and offset
the Moho, or simply show an offset, suggest the early
development of the reflection Moho as a 'rigid' non-dynamic
boundary [Matthews and Cheadle, 1985; Hauser et al., 1987].
Other data suggest that dipping crustal structures are
terminated at the Moho, where they are either truncated by the
Moho or approach the Moho asymptotically, possibly acting as
a d6collement surface [Nelson et al., 1985; Matte and Him,
1988]. In some regions where deep seismic data suggest that
the Moho is horizontally superimposed over earlier crustal
structure, surface geologic relations show that a period of
large-scale extension post-dates an older compressional
episode [Allmendinger et al., 1987]. The inference that can be
made for this last case is that, just as crustal extension was the
most recent major tectonic episode in these regions, the Moho
formed last as a dynamic boundary. On the other hand, basin
analysis of data from the Valencia trough [Watts et al., 1990]
suggests that the reflective lower crustal layer in the Iberian
crust was formed before mid-Tertiary extension, and was thus
unrelated to the last tectonic episode in the area. The
possibility that a more widespread extensional event, such as
that associated with the Tethys, caused the highly reflective
lower crust of Iberia cannot be ruled out [Wans et al., 1990].
Evidence from both the Cordilleran interior and western
Williston basin suggests the importance of magmafism in the
development of the reflection Moho. Whereas the reflection
Moho beneath both regions may be interpreted solely as the
result of magmafism, the sub-horizontal nature of the reflection
Moho and reflective lower crust in the Cordilleran interior may
have also developed in association with significant (-30%)
crustal extension in the region [Parrish et al., 1988]. In
contrast, the reflection Moho beneath the western Williston
basin may have developed through intrusion associated with
an incipient rifting event in the absence of significant crustal
extension.
Acknowledgments. C. J. Potter and other members of the
COCORP project provided useful comments throughout the
course of this study. I would also like to thank T. E. Jordan,
M. A. Fisher, S. L. Klemperer, and F. A. Cook for their
comments. This research was supported by National Science
Foundation grants EAR 84-18157 and EAR 86-09992. The
data were collected by crew 6834 of Petty-Ray Geophysical
and processed on the COCORP Megaseis TM system at Cornell
University. INSTOC Contribution 165.
References
Barazangi, L. D. Brown, and J. E. Oliver, Deep seismic
reflection characteristics of the continental crust, Geology,
15, 304-310, 1987.
Armstrong, R. L., Mesozoic and early Cenozoic magmatic
evolution of the Canadian Cordillera, in Processes in
Continental Lithospheric Deformation, edited'by S. P.
Clark, B.C. Burchfiel, and J. Suppe, Spec Pap. Geol.
Soc. Am. 218, 55-91, 1988.
Asada, T., and L. T. Aldrich, Seismic observations of
explosions in Montana, in The Earth beneath the
continents, Geophys Monogr. Ser., 10, edited by J.
Steinhart and T. Smith, pp. 244-250, AGU, Washington,
D.C., 1966.
Braile, L. W., and C. S. Chiang, The continental Mohorovicic
discontinuity: results from near-vertical and wide-angle
seismic reflection studies, in Reflection seismology: A
global perspective, Geodyn. Ser., 13, edited by M.
Barazangi and L. Brown, pp. 257-272, AGU, Washington.
D.C., 1986.
Brown, R. L., J. M. Journeay, L. S. Lane, D.C. Murphy,
and C. J. Rees, Obduction, backfolding and piggyback
PRUSSEN 321
thrusting in the metamorphic hinterland of the southeastern
Canadian Cordillera, J. Struct. Geol., 8, 255-268, 1986.
Catchings, R. D., and W. D. Mooney, Crustal structure of the
Columbia Plateau: Evidence for continental rifting, J.
Geophys. Res., 93, 459-474, 1988.
Cook, F. A., A. G. Green, P.S. Simony, R. A. Price, R. R.
Parrish, B. Milkereit, P. L. Gordy, R. L. Brown, K. C.
Coffin, and C. Patenaude, Lithoprobe seismic reflection
structure of the southeastern Canadian Cordillera: Initial
results, Tectonics, 7, 157-180, 1988.
Cumming, W. B., R. M. Clowes, and R. M. Ellis, Crustal
structure from a seismic refraction profile across southern
British Columbia, Can. J. Earth Sci., 16, 1024-1040,
1979.
Fowler, C. M. R., and E.G. Nisbet, The subsidence of the
Williston basin, Can. J. Earth Sci., 22, 408-415, 1985.
Furlong, K. P., and D. M. Fountain, Continental crustal
underplating: thermal consideratons and seismic-petrologic
consequences, J. Geophys. Res., 91, 8285-8294, 1986.
Gibbs, A. K., Seismic reflection profiles of Precambrian
crust: a qualitative assessment, in Reflection seisinology:
The continental crust, Geodyn. Ser., 14, edited by M.
Barazangi and L. Brown, pp. 95-106, AGU, Washington,
D.C., 1986.
Green, A. J., Z. Hajnal, and W. Weber, An evolutionary
model of the western Chumhill province and western
margin of the Superior province in Canada and the north-
central United States, Tectonophysics, 116, 281-322,
1985.
Hale, L. D., and G. A. Thompson, The seismic reflection
character of the continental Mohorovicic discontinuity, J.
Geophys. Res., 87, 4625-4635, 1982.
Hales, A. L., and J. B. Nation, A seismic refraction survey in
the northern Rocky Mountains: more evidence for an
intermediate crustal layer, Geophys. J. R. astr. Soc., 35,
381-399, 1973.
Hauser, E. C., J. Gephart, T. Latham, J. E. Oliver, S.
Kaufman, L. D. Brown, and I. Lucchitta, COCORP
Arizona transect: Strong crustal reflections and offset Moho
beneath the transition zone, Geology, 15, 1103-1106,
1987.
Hill, D. P., Crustal and upper mantle structure of the
Columbia Plateau from long-range seismic refraction
measurements, Geol. Soc. Am. Bull., 83, 1639-1648, 1972.
Hurich, C.A., and S. B. Smithson, Compositional variation
and the origin of deep crustal reflections, Earth Planet. Sci.
Lett., 85, 239-242, 1987.
Latham, T., J. Best, T. Chaimov, A. Barnes, J. Oliver, L.
Brown, and S. Kaufman, COCORP profiles from the
Montana plains: The Archean cratonic crust and a lower
crustal anomaly beneath the Williston Basin, Geology, 16,
1073-1076, 1988.
Lundy, J., Deep crust of the Colorado Plateau: The COCORP
reflection surveys, M. S. thesis, 67 pp., Cornell
University, Ithaca, New York, 1988.
Mair, J. A., and J. A. Lyons, Seismic reflection techniques
for crustal structure studies, Geophysics, 41, 1272-1290,
1976.
Matte, P., and A. Him, Seismic signature and tectonic cross
section of the Variscan crust in western France, Tectonics,
7, 141-155, 1988.
Matthews, D. H., and M. J. Cheadle, Deep reflections from
the Caledonides and Variscides west of Britain and
comparison with the Himalayas, in Reflection seisinology:
A global perspective, Geodyn. Ser., 13, edited by M.
Barazangi and L. Brown, pp. 5-19, AGU, Washington.
D.C., 1986.
Meissner, K. R., The "Moho" as a transition zone,
Geophysical Surveys, 1, 195-216, 1973.
Meissner, R., The Continental Crust: A Geophysical
Approach, 426 pp., Academic Press, Inc., New York,
1986.
Meyer, R. P., J. S. Steinhart, and W. E. Bonini, Montana,
1959, in Explosion studies of crustal structure, edited by
Steinhart, J. S., and R. P. Meyer, Carnegie Institute of
Washington Publications, 622, 305-343, 1961.
Monger, J. W. H. (principal compiler), G. J. Woodsworth,
R. A. Price, R. M. Clowes, R. P. Riddihough, R. Curde,
T. Hoy, V. A. G. Preto, P.S. Simony, P. D. Snavely, and
C. J. Yorath, Transect B2, Southern Canadian Cordillera,
North American continent-ocean transects, in Decade of
North American Geology Program, Geological Society of
America, Boulder, Colo., 1986.
Mooney, W. D., and T. M. Brocher, Coincident seismic
reflection/refraction studies of the continental lithosphere: A
global review, Reviews of Geophysics, 25, 723-742,
1987.
Morel-h-l'Huissier, P., A. G. Green, and C. J. Pike, Crustal
refraction surveys across the Trans-Hudson
orogen/Williston basin of south central Canada, J.
Geophys. Res., 92, 6403-6420, 1987.
Nelson, K. D., J. H. McBride, J. A. Arnow, J. H. Willemin,
J. Huang, L. Zheng, J. E. Oliver, S. Kaufman, and L. D.
Brown, New COCORP profiling in the southeastern United
States. Part I: Paleozoic suture and Mesozoic rift basin,
Geology, 13, 714-718, 1985.
O'Brien, H., A. J. Irving, and, I.S. McCallum, Eocene
potassic magmatism in the Highwood Mountains, Montana:
Petrology, geochemistry and tectonic implications, in
Eocene magmatism in the northwestern USA, edited by F.
Moye, Spec. Pap. Geol. Soc. Am., in press.
Oliver, J. E., Changes at the crust-mantle boundary, Nature,
299, 398-399, 1982.
Parrish, R. R., S. Carr, and D. L. Parkinson, Eocene
extensional tectonics and geochronology of the southern
Omineca belt, British Columbia and Washington,
Tectonics, 7, 181-212, 1988.
Potter, C. J., W. E. Sanford, T. R. Yoos, E. I. Prussen, R.
W. Keach, J. E. Oliver, S. Kaufman, and L. D. Brown,
COCORP deep seismic reflection traverse of the North
American Cordillera, Washington and Idaho: Implications
for orogenic evolution, Tectonics, 5, 1007-1025, 1986.
Price, R. A., J. W. H. Monger, and J. A. Rodick, Cordilleran
cross section; Calgary to Vancouver, in FieM guides to
geology and mineral deposits in the southern Canadian
Cordillera, edited by D. J. Templeman-Kluit, pp. 3-1 to 3-
85, Geological Society of America, Cordilleran Section,
Vancouver, B.C., 1985.
Prussen, E. I., Variations in the seismic reflection character of
the Moho along the COCORP Northwest U.S. transect-
Washington, Idaho, and Montana, M. S. thesis, 76 pp.,
Cornell University, Ithaca, New York, 1989.
Rohay, A., Crust and mantle structure of the North Cascades
range, Washington, Ph.D. thesis, 163 pp., University of
Washington, Seattle, Washington, 1982.
Sanford, W. E., C. J. Potter, and J. E. Oliver, Detailed three-
dimensional structure of the deep crust based on COCORP
data in the Cordilleran interior, north-central Washingon,
Geol. Soc. Am. Bull., 100, 60-71, 1988.
Steinhart, J. S., and R. P. Meyer, Explosion studies of
crustal structure, 409 pp., Carnegie Institute of
Washington, 1961.
Varsek, J., F. Cook, R. Clowes, C. Spencer, and E.
Kanasewich, Lithoprobe Reflection traverse of southern
British Columbia, Geol. $oc. Am. Abstr. Programs, 21,
153, 1989.
Waters, K. H., Reflection Seisinology, 453 pp., John Wiley
and Sons, New York, 1981.
Watts, A. B., M. Tome, P. Buhl, A. Mauffret, G. Pascal, and
B. Pinet, Evidence for reflectors in the lower continental
crust before rifting in the Valencia trough, Nature, 348,
631-635, 1990.
Wilshire, H. G., Lithology and evolution of the crust-mantle
boundary region in the southwestern Basin and Range, J.
Geophys. Res., 95, 649-665, 1990.
Yilmaz, 6., Seismic Data Processing, 526 pp., Society of
Exploration Geophysicists, Tulsa, 1987.
Yoos, T. R., C. J. Potter, J. L. Thigpen, and L. D. Brown,
The Cordilleran foreland thrust belt in northwestern
Montana and northern Idaho from COCORP and industry
seismic reflection data, Am. Assoc. Petrol. Geol. Bull., in
press.
Zhu, T. F., and L. D. Brown, Consortium for Reflection
Profiling Michigan surveys: Reprocessing and results, J.
Geophys. Res., 91, 11,477-11,495, 1986.
UPPER CRUSTAL POISSON'S RATIOS IN THE COLORADO PLATEAU
FROM MULTICOMPONENT WIDE-ANGLE SEISMIC RECORDING
Roy A. Johnson and Kenneth A. Hartman 1
Department of Geosciences, University of Arizona
Tucson, Arizona 85721
Abstract. Vertical- and horizontal-component recording of
wide-angle seismic data on the Colorado Plateau -60 km NNW
of Flagstaff, Arizona, indicates that upper crustal shear-wave
velocities are high with respect to compressional-wave
velocities. Calculations of Poisson's ratios using velocities
determined from plane-wave decomposition (slant stacking)
yield values that are significantly lower than 0.25. Poisson's
ratio in the upper 3-4 km of the crystalline basement (beneath
Phanerozoic sedimentary rocks) is 0.24, but decreases to 0.20
between about 6 and 15 km depth in the upper crust.
Significant decrease in Poisson's ratio with depth can be
explained by appropriate juxtapositions of wet granite, granitic
gneiss or felsi schist and closure of microcracks with
increasing depth. Elevated subsurface temperatures may play
a role in lowering Poisson's ratio. Evidence for anomalously
high fluid pressures, which would significantly increase
Poisson's ratio, is not found.
Introduction
The Colorado Plateau is an anomalously undisrupted block
of continental crust in the western U.S. that is surrounded on
all sides by evidence of intense compressional and extensional
deformation. Tertiary and Quaternary intrusive and extrusive
volcanism has encroached into the margins of the Plateau and
large Laramide and younger faults cut through the thick
Paleozoic and Mesozoic sedimentary rock cover, but this
enigmatic block maintains a distinct structural identity that
poses fundamental questions about its nature at depth.
Beyond limited exposures of Precambrian crystalline rocks
in the Grand Canyon and at the edge of the Plateau in the
Transition Zone of central Arizona, very little is known about
the nature of basement rocks even immediately below the
Phanerozoic cover. Where exposed, Precambrian basement
metamorphic rocks show a distinct NE-SW structural grain; to
at least some degree, this structural grain appears to control
later structures [Karlstrom et al., 1987]. This Precambrian
structural grain is presumed to be related to accretion of island
arc terranes during Proterozoic continental growth [Anderson,
1978, 1989; Karlstrom et al., 1987]. However, inferences about
the structure and composition of the upper crust must be
obtained principally from application of geophysical
techniques.
This paper focuses on results of analysis of upper crustal
P- and S-wave refraction arrivals from normal-incidence to
1Now at Amoco Production Company, Houston, Texas.
Geodynamics 22
wide-angle seismic reflection/refraction recording on the
Colorado Plateau approximately 60 km NNW of Flagstaff,
Arizona. Shear-wave data from this area are used in
combination with traditional compressional-wave data to
estimate Poisson's ratio and provide constraints on possible
upper crustal compositions and factors that affect
compositional estimates from seismic data, including temper-
ature and fluid content. Results from analysis of deeper
penetrating seismic waves are presented elsewhere (K. A. Hart-
man and R. A. Johnson, P- and S-wave velocity structure and
Poisson's ratio beneath the Colorado Plateau, northern Arizona,
submitted to Journal of Geophysical Research, 1991,
hereinafter referred to as K. A. Hartman and R. A. Johnson
(1991)).
Arizona PACE-Piggyback
Lithospheric Experiment ("APPLE")
In September, 1989, the University of Arizona deployed a
192-channel, 9.6-km-long recording spread as a "piggyback"
experiment to record 35 large chemical explosions detonated by
the USGS as part of the 1989 "Pacific to Arizona Crustal
Experiment" (PACE) project (Figure 1). Twenty-three PACE
shots were located along a main SW-NE transect extending
from Blythe, California to Tonalea, Arizona, two shots were
offset south of the main transect as "fan" shots, and ten shots
were located along a perpendicular cross line that extended
from the north rim of the Grand Canyon, southeast to Meteor
Crater.
The APPLE spread was deployed at the intersection of the
two shot lines, generally along the azimuth of the main (SW-
NE) transect (Figure 1). Station spacing was 50 m, with
vertical- and horizontal-component geophones interleaved at
alternating stations. Multielement, 20-m geophone arrays were
aligned along the azimuth of the main shots; within the
horizontal-component geophone arrays, elements were oriented
parallel to the azimuth of the main transect (N52W). Thus,
horizontal geophones preferentially recorded SV-wave
information from main-line shots, but principally recorded SH-
wave information from cross-line shots [Johnson and Hartman,
1989].
Discontinuous coverage was obtained over a distance of 460
km; concentration of shots in the Transition Zone provided
nearly continuous coverage over a 140-km range near the
boundary between the Colorado Plateau and Transition Zone.
Individual and composite shot records generally exhibit very
high signal-to-noise ratios (S/N) for both P- and S-wave
reflected and refracted arrivals. Although weak shear-wave
arrivals are recorded on the vertical-component geophones,
deployment of horizontal-component geophones resulted in
much higher S/N for shear arrivals than would have been
323
324 UPPER CRUSTAL POISSON'S RATIO
83
85 46
Colorado
Plateau
76
Flagstaff
82
23
20
Phoenix O
Basin and Range
Tucson
0 100 200
I I , I
Km
Fig. 1. Arizona PACE-Piggyback Lithospheric Experiment
(APPLE). 1989 USGS PACE shotpoint locations are numbered.
(Due to shot density some shots are not shown.) Main line
runs SW-NE; cross line runs SE-NW. The APPLE seismic
profile was located between shotpoints 34 and 71 and consisted
of 192 channels of P- and S-wave geophones.
possible with the deployment of vertical-component geophones
alone.
Velocity Estimation from Tau-p Analysis
An objective measure of Vp and V= can be obtained from
conventional time-distance (t-x) domain seismic data by
transformation to the delay time--ray parameter (tau-p)
domain. This process, also known as slant stacking, sums t-x
trace amplitudes along different linear slopes (ray parameters
or apparent slownesses) for all time intercepts (delay times)
[Johnson and Gilbert, 1972]. Summations along coherent
events yield large values in tau-p; apparent velocities are
obtained by inverting the ray parameter value: Vapparent -- l/p.
We used a tau-p transformation scheme adapted from Stoffa
et al. [1981 ] which also allows for semblance weighting. Noise
or events-with poor coherence in t-x (low semblance) are
reduced in amplitude in tau-p.
Linear refraction arrivals in the t-x domain ideally are
transformed to points in the tau-p domain (Figure 2), and
hyperbolic reflections are transformed to sums of ellipses
[Phinney et al., 1981]. Although reflections do provide
important constraints on velocity and structure and were used
in modeling, for determinations of upper crustal P- and S-
wave velocities, we used only refraction arrivals with high S/N
(K. A. Hartman and R. A. Johnson, 1991).
Apparent compressional-wave and shear-wave velocities
determined from tau-p analysis (e.g., Figure 2b) were
confirmed by visual inspection and conventional t-x velocity
analysis. The usefulness of tau-p analysis, however, is in its
ability to objectively identify events and in its utility for
estimating confidence bounds in velocity determinations [cf,
Dorman, 1983; Hawman et al., 1990; Hawman and Phinney, in
press]. Compressional-wave velocity estimates determined
from our analysis generally are accurate to less than +0.1 km/s,
and shear-wave velocities are accurate to less than _+0.04 km/s
(K. A. Hartman and R. A. Johnson, 1991).
Poisson's Ratio in the Upper Crust
Most previous wide-angle seismic studies of the Colorado
Plateau [Roller, 1965; Warren, 1969] recorded compressional
waves, which provide fundamental constraints on crustal
structure, but interpretation of possible compositions of crustal
rocks based on compressional wave velocities alone are highly
nonunique [Fountain and Christensen, 1989]. Determination
of accurate compressional- and shear-wave apparent velocities
is more appropriate for interpretations of crustal composition
and for inferences about other factors, such as temperature or
pore pressure, that could strongly affect velocities and thus
could bias compositional estimates. Vp/V= ratios are more
diagnostic of rock types [Fountain and Christensen, 1989] and
are obtainable from multicomponent seismic recording
experiments such as APPLE. We use Poisson's ratio, a,
expressed in terms of compressional- and shear-wave
velocities, as a basis for comparison:
a = (Vp'-2Vs')/2(Vp2-V ') (1)
A first-order approximation of Poisson's ratio in the upper
crust can be derived from simple analysis of reduced travel
time (Treaucea = T - X/Vreduction) plots for P- and S-wave first
arrivals. If a is assumed to be 0.25 in the upper crust, and
the upper crustal P-wave refraction phase, Pg, has a velocity
of 6 km/s, the appropriate shear-wave velocity for Sg will be
3.46 km/s. Compressing the time scale of the S-wave plot by
a factor of 1.73 (appropriate for a = 0.25) will render Pg and
Sg plots that should look "the same" if a truly is 0.25.
Analysis of shot 32, which is offset ~100 km from our
spread (Figure 1), shows that Pg arrives at about 0.85 s
(reduced time) with a velocity of ~6.0 km/s, but Sg phases
begin to arrive at about 0.35 s (Figure 3). If Poisson's ratio
was 0.25, Sg should have been observed at -1.47 s; "early"
arrival of Sg indicates that upper crustal a is appreciably less
than 0.25.
Detailed investigation using arrival velocities determined
from tau-p analysis provides more specific results. Shot
records with offsets less than 40 km, on which Pg and Sg
propagated through the uppermost part of the basement (below
Phanerozoic sedimentary units), yield Poisson's ratios of 0.24
_+ 0.01, whereas shot records from offsets greater than 100 km
(shot points 28, 31, 32 and 38) yield values of 0.20 _+ 0.02.
Consistency of measurements from multiple offsets argues that
these values are truly representative of upper crustal Poisson's
ratio and not anomalous values due to significantly different
travel paths for P- and S-waves. These results are also
consistent with estimates of "average" Poisson's ratio within an
interval from arrival times.
Estimates from ray tracing indicate that the longer offset
shots sample a depth range between about 6 and 15 km. Hence
Poisson's ratio decreases significantly below the upper part of
the crystalline basement. Uncertainties in shear-wave velocity
determinations from shots in the offset range from 40 to 100
km so far prevent unambiguous identification of a downward-
decreasing gradient if indeed one exists. Thus in the upper
crust on the southwest margin of the Colorado Plateau, the
decrease in Poisson's ratio may be abrupt or gradational, and
the depth of transition presently is not well resolved but occurs
below 5 km.
Upper Crustal Structure
Velocities determined for the upper crust of the
southwestern Colorado Plateau suggest that, to a first order, the
1. 000
2.000
3.000
1.000
2.000
3.000
4.000 4.000
Fig. 2. (a) Vertical-component response over 6 km aperture
for shot 31. Datum (t = 0) is 23.5 s. Note very high S/N and
high-frequency content. Slightly different moveouts for events
following first arrivals represent different apparent velocities
that can be resolved using tau-p analysis. (b)Slant stack of
data from Figure 2a. Top annotation (l/p) in km/s. Side
annotation in seconds from datum at 23.5 s (t = 0) for offset
JOHNSON AND HARTMAN 325
1.000
_.boo
q. ooo
I '
4. 1300 --- ' 4. 000
i. 000
2.000
3.000
of 140 km. Arrival times of high-amplitude zones correspond
to travel times in Figure 2a at 140 km. "Smear" of events is
result of finite wavelet length and oversampling in p. Velocity
is picked at maximum event amplitude. Variation of amplitude
around maximum provides estimate of velocity uncertainty.
Note events at 0.75 s, 0.9 s, 1.55 s, and 2.5 s corresponding to
events in Figure 2a.
upper crust can be subdivided into three layers (Figure 4).
The first layer, which extends from the surface to a depth of
about I km, comprises mainly Paleozoic and Mesozoic
sedimentary rocks that are well known from spectacular
exposures in the Grand Canyon and from limited borehole
information. Surface outcrops along the APPLE spread are
Kaibab Limestone which is underlain by thick sandstone,
mudstone, shale and carbonate units. Compressional wave
velocities, which range from 2.8 km/s to 4.0 km/s in this layer,
are typical of such sequences (Figure 4).
The uppermost basement layer extends from the base of the
sedimentary section to a depth of about 5 km. Within this
layer, Vp increases from 5.7 km/s to 5.9 km/s and V, increases
from 3.35 km/s to 3.45 km/s to maintain a Poisson's ratio of
-0.24. Absence of a significant change in P-wave velocity
near the base of this layer suggests that transition to the deeper
parts of the upper crust is gradational. Beneath about 6 km,
however, shear-wave velocity reaches 3.62 km/s, indicating
that a significant gradient in shear-wave velocity occurs within
a zone that could be as thin as I km (Figure 4).
The upper crust beneath the "basement" layer extends to a
depth of about 18 km. The base of this zone is marked by
onset of strong reflectivity which we assign to the middle crust
(K. A. Hartman and R. A. Johnson, 1991). Compressional-
wave velocity in the interval between about 6 and 15 km
increases from 5.9 km/s to 6.1 km/s, and shear-wave velocity
increases from 3.62 km/s to 3.70 km/s (Figure 4). The
resulting Poisson's ratio of 0.20 is surprisingly low, particularly
since it represents a decrease with respect to the uppermost
basement layer.
Discussion
Although composition exerts a dominant control on Poisson's
ratio in crustal rocks, numerous other factors may strongly
influence e, including pressure, temperature, fractures or
porosity, anisotropy, fluid content and pore pressure [Birch,
1961' Kern, 1978; Christensen, 1979; Christensen, 1989a,
1989b; Fountain and Christensen, 1989]. The simplest means
of explaining observed changes in velocity and Poisson's ratio
in the upper crust would be to identify typical upper crustal
rocks with appropriate laboratory-measured Vp and V, (and
thus ) values at appropriate pressures and assume that these
rocks represent compositions in the upper crust of the Colorado
Plateau. Observed P-wave velocities suggest that upper crustal
rocks could be quartzite, granite, amphibolite-facies granitic or
granodioritic gneiss, or felsic schist [Fountain and Christensen,
1989; Christensen, 1989a; Holbrook et al., in press; Brother et
al., in press]. However, shear wave velocities, and particularly
derived from both P- and S-wave velocities, tend to restrict
these possibilities to granite, granitic gneiss or felsic schist.
Quartzite, with its anomalously high shear-wave velocity with
respect to Vp, gives values too low to explain observations
and, although it does occur in isolated outcrops near Chino
Valley at the edge of the Plateau [Karlstrom et al., 1987;
Anderson, 1989], it is not likely the dominant constituent
lithology. If one considers only variations in lithologies, a
change in Poisson's ratio from 0.24 in the basement layer to
0.20 in the deeper part of the upper crust could be
accomplished by juxtaposition of granite or granitic gneiss over
felsic schist.
Lithostatic pressures to depths of at least 7 km are less than
200 MPa; between about 100 and 200 MPa, microcracks close
[Birch, 1961], causing an increase in Vp and V, but a decrease
in Poisson's ratio for water-saturated rocks [Christensen,
1989b]. Thus it is possible that the observed decrease in in
the upper crust of the Colorado Plateau occurs within a single
lithology and is due principally to closure of microcracks below
4 km. Appropriate lithologies that perhaps could best explain
this behavior are felsic schists, since some samples of these
rocks have Poisson's ratios of about 0.20 for pressures above
which microcracks close [Brother et al., in press]. Open
microcracks or fractures shallower than about 6 km could
increase Poisson's ratio to 0.24. Nevertheless, lack of a strong
compressional-wave velocity gradient, apparently coupled with
a significant shear-wave velocity gradient between the
326 UPPER CRUSTAL POISSON'S RATIO
103 101 100
O. 000
1. 000
O. 000
1. 000
2. 000 2. 000
Fig. 3. Reduced travel time plots of part of shot 32. (a) P-
waves, reduction velocity = 6.0 km/s. Note first arrivals for
P-waves at 0.85 s. First-arrivals (velocity 5.9 km/s) plot
nearly flat. (b) S-waves, reduction velocity = 3.464 km/s.
Time scale is compressed by factor of 1.732. If a in the upper
crust is 0.25, Figures 2a and 2b would look the same with P-
KM 103 102 101 100 KM
o. ooo o. ooo
1. 000
2. 000
3. 000
l. 000
2. 000
3. 000
wave and S-wave first arrivals near the centers of both plots.
However, note onset of S-wave first arrivals at -0.35 s. For
a of 0.25, S-wave first arrivals would occur at ~ 1.47 s. "Early"
arrival of shear waves indicates that upper crustal a is less than
0.25. S-waves exhibit dip toward smaller offset, indicating
velocities are higher than 3.464 km/s.
basement layer and deeper upper crust suggests that this
monolithologic situation is unlikely.
Elevated temperatures may also reduce Poisson's ratio.
Tertiary and Quaternary volcanism along the southern margin
of the Colorado Plateau is widespread [Reynolds, 1988], and
about 40 km south of the APPLE spread, the San Francisco
volcanic field has been active as recently as 900 years ago.
Travel paths for arrivals on our spread traverse a region that
has been thermally perturbed [Sass et al., 1982]. Compressional
and shear-wave velocities of rocks generally decrease as
temperature increases [Christensen, 1989a], but in normal
continental heat flow environments, increasing pressure with
depth tends to offset this effect [Fountain and Christensen,
1989]. Where heat flow is higher, such as in the southern
Colorado Plateau [Blackwell, 1978; Swanberg and Morgan,
1985; Sass et al., 1982], increasing temperature with depth may
cause an overall decrease in velocities [Fountain and
Christensen, 1989]. Very little information on the simultaneous
effects of temperature on compressional- and shear-wave
velocities is available, but limited data from Christensen
[1989a] indicate that for quartzite, granite, and gabbro, a shows
a marked decrease for increasing temperatures; samples of
eclogite and peridotitc show an increase of a with temperature.
In the presence of water at low pressure, increasing
temperature causes a decrease in Vp for granite while causing
little change in V, and thus results in a decrease in [Spencer
and Nur, 1976].
Decrease of Poisson's ratio with increasing temperature is
too slow at moderate temperatures to lower relatively
abruptly from 0.24 to 0.20 below 5-6 km for reasonable upper
crustal compositions. However, elevated upper crustal
temperatures beneath the Colorado Plateau may be a factor in
overall low observed values of . Since elevated temperatures
are a real possibility for our seismic travel paths, some
lowering of is likely and would tend to make estimates of
constituent lithologies somewhat too felsic.
If open fractures and microcracks occur in the upper crust
as seems likely, filling of this porosity with water or other
fluids would have a marked effect on Poisson's ratio. In fact,
for the upper crust, this effect may be extremely important.
Christensen [1989b, p. 794] shows plots of Poisson's ratio as a
function of confining pressure and differential pressure.
Curves for low fluid pressures show very rapid decrease of
for increasing pressure. Elevated pore pressures generally
Depth
(km)
Vp -- 2.8 - 4.o
Vp = 5.6 Vs = 3.30
Poisson's
Ratio
= 0.24
Vp = 5.9 Vs = 3.45
Vp = 5.9 vs = 3.60
Poisson's
Ratio
= O.2O
Vp = 6.1 Vs = 3.70
18
Fig. 4. Simplified velocity model of the upper crust in the
southern Colorado Plateau showing first-order vertical vari-
ations. Velocities are in km/s. Crystalline upper crustal rocks
exhibit marked decrease in Poisson's ratio below about 5 or 6
km. See text for discussion.
result in higher a. The implications of these relationships is
that, for a given lithology, a should decrease significantly in
the upper crust as pressure increases and microcracks close
with depth, "flattening out" as pressure (depth) increases
further [Christensen, 1989b]. If hydrostatic pressure is
maintained to a particular depth (at which permeability
becomes inhibited by partial closure of fractures and micro-
cracks), onset of high pore pressures could occur which could
sharply increase a. Such effects also would apply to cases
where different lithologies occur together.
Conclusions
Water is probably an important constituent in the upper
crust beneath the Colorado Plateau and may be an important
factor in the observed decrease of a with depth. The low
values for upper crustal Poisson's ratios suggest that water or
other fluids (such as COg) are not at anomalously elevated pore
pressures. Similarly, partial melts in the upper crust would
tend to give larger values of a than we observe [Christensen,
1989b] and therefore are unlikely.
Apparently the most likely causes for decrease of Poisson's
ratios in the upper crust beneath the Colorado Plateau as well
as for the generally low observed values (0.24-0.20) are
changes in lithology with depth coupled with closure of water-
filled fractures or microcracks. Temperatures that are probably
elevated with respect to a "normal" geothermal gradient also
contribute to lowering Poisson's ratios.
Reasonable lithologies that could account for our
observations on the southern Colorado Plateau under presumed
conditions of lithostatic pressure, elevated (but not extreme)
JOHNSON AND HARTMAN 327
geothermal gradients, and water-filled microcracks are felsic
schist, quartz-rich granite, or granitic gneiss overlain by schist,
felsic Precambrian metasedimentary rocks, or quartz-rich
granite or granitic gneiss. Although estimates of lithologies
from seismic velocities, even with determinations of a, are
nonunique, they provide perhaps the best means for estimating
crustal compositions in situ short of deep drilling.
Acknowledgments. Funding for field data acquisition was
provided by the Air Force Geophysics Laboratory through NSF
Grant EAR-8817241. Computational support was provided by
NSF Grants EAR-8804667 and EAR-9018552. Additional
support was provided by a Shell Oil Foundation Faculty Career
Initiation Award to Johnson. Basic data processing was done
on a Convex C-120 mini-supercomputer using SierraSEIS
software developed by Sierra Geophysics, Inc. L. Stephen
Sorenson developed codes used in tau-p analysis and some
other aspects of the data processing. We thank Nick Boyd, Ed
Criley, Chris Humphreys, Jill McCarthy, Scott Smithson, David
Steinke and Alan Tanner for technical and logistical help, and
Norman Meader and Kim Johnson for manuscript preparation.
References
Anderson, P., The island arc nature of Precambrian volcanic
belts in Arizona, Geol. Soc. Am. Abstr. Programs, 10, 156,
1978.
Anderson, P., Proterozoic plate tectonic evolution of Arizona,
Geologic Evolution of Arizona, edited by J.P. Jenney and
S. J. Reynolds, Arizona Geol. Soc. Dig., 17, 17-55, 1989.
Birch, F., The velocity of compressional waves in rocks to 10
kilobars, Part 2, J. Geophys. Res., 66, 2199-2224, 1961.
Blackwell, D. D., Heat flow and energy loss in the Western
United States, Cenozoic Tectonics and Regional Geophysics
of the Western Cordillera, Mem. Geol. Soc. Am. 152, 175-
208, 1978.
Brocher, T. M., W. J. Nokleberg, N. I. Christensen, M. A.
Fisher, and E. L. Geist, Seismic reflection/refraction
mapping of the faulting and regional dips in the eastern
Alaska Range, J. Geophys. Res., in press.
Christensen, N. I., Compressional wave velocities in rocks at
high temperatures and pressures, critical thermal gradients,
and crustal low-velocity zones, J. Geophys. Res., 84, 6849-
6857, 1979.
Christensen, N. I., Seismic velocities, in Practical Handbook of
Physical Properties of Rocks and Minerals, edited by R. S.
Carmichael, pp. 431-546, CRC Press, Inc., Boca Raton,
Florida, 1989a.
Christensen, N. I., Pore pressure, seismic velocities and crustal
structure, Geophysical Framework of the Continental United
States, Mem. Geol. Soc. Am. 172, 783-798, 1989b.
Dorman, L. M., Modelling and parameterization errors in body
wave seismology, Geophys. J. R. Astron. Soc., 72, 571-576,
1983.
Fountain, D. M., and N. I. Christensen, Composition of the
continental crust and upper mantle: A review, Geophysical
Framework of the Continental United States, Mem. Geol.
Soc. Am. 172, 711-742, 1989.
Hawman, R. B., R. H. Colburn, D. A. Walker, and S. B.
Smithson, Processing and inversion of refraction and wide-
angle reflection data from the 1986 Nevada PASSCAL
experiment, J. Geophys. Res., 95, 4657-4691, 1990.
Hawman, R. B., and R. A. Phinney, Analysis of sparse wide-
angle reflection data in the r(p) domain, Bull. Seismol. Soc.
Am., in press.
Holbrook, W. S., W. D. Mooney, and N. I. Christensen, The
seismic velocity structure of the deep Continental crust, in
Geology and Geophysics of the Lower Continental Crust,
edited by D. M. Fountain et al., in press.
328 UPPER CRUSTAL POIS$ON'S RATIO
Johnson, L. E., and F. Gilbert, A new datum for use in the
body wave travel time inverse Iroblem, Geophys. J. R.
Astron. Soc., 30, 373-380, 1972.
Johnson, R. A., and K. A. Hartman, The Arizona PACE-
Piggyback Lithospheric Experiment (APPLE): Vertical-
incidence to wide-angle reflection data from the Colorado
Plateau, Eos Trans. AGU, 70, 1205, 1989.
Karlstrom, K. E., $. A. Bowring, and C. M. Conway, Tectonic
significance of an Early Proterozoic two-province boundary
in central Arizona, Geol. Soc. Am. Bull., 99, 529-538, 1987.
Kern, H., The effect of high temperature and high confining
Iressure on compressional wave velocities in quartz-bearing
and quartz-free igneous and metamorphic rocks,
Tectonophysics, 44, 185-203, 1978.
Phinney, R. A., K. R. Chowdhury, and L. N. Frazer,
Transformation and analysis of record sections, J. Geophys.
Res., 86, 359-377, 1981.
Reynolds, $. J., Geologic Map of Arizona, Ariz. Geol. $urv.,
Tucson, Arizona, 1988.
Roller, J. C., Crustal structure in the eastern Colorado Plateau
Province from seismic refraction measurements, Bull.
Seismol. Soc. Am., 55, 107-119, 1965.
Sass, J. H., C. Stone, and D. J. Bills, Shallow subsurface
temperatures and some estimates of heat flow from the
Colorado Plateau of northeastern Arizona, U.S. Geol. Surv.
Open File Rep. 82-994, 112 Il., 1982.
Spencer, J. W., and A.M. Nur, The effects of Iressure,
temperature and Iore water on velocities in Westerly
granite, J. Geophys. Res., 81, 899-904, 1976.
$toffa, P. L., P. Buhl, J. B. Diebold, and F. Wenzel, Direct
mapping of seismic data to the domain of intercept time and
ray Iarameter: A Ilane wave decomposition, Geophysics,
46, 255-467, 1981.
Swanl;erg, C. A., and P. Morgan, Silica heat flow estimates and
heat flow in the Colorado Plateau and adjacent areas, J.
Geodynamics, 3, 65-85, 1985.
Warren, D. H., A seismic refraction survey of crustal structure
in central Arizona, Geol. Soc. Am. Bull., 80, 257-282, 1969.
VELOCITY-RESISTIVITY CORRELATIONS IN THE DEEP CRUST
Guy Marquis
Department of Physics & Astronomy, University of Victoria
Victoria, B.C. V8W 3P6 Canada
Roy D. Hyndman
Pacific Geoscience Centre, Geological Survey of Canada,
Sidney, B.C. VSL 4B2 Canada
Abstract. A compilation of approximately coincident
refraction velocity and magnetotelluric resistivity data shows a
general correlation of lower velocity with lower resistivity in the
deep crust. The data fit a range of models where both velocity
and resistivity are mainly a function of fluid porosity. Reduction
of velocity from an expected zero porosity mafic rock value of
about 7.2 km/s to commonly observed values of about 6.7 km/s
and explanation of the low resistivities require between 0.3 to
3 % porosity, depending on the pore geometry. Poisson's ratio
data may provide an important additional constraint on the pore
geometries.
Introduction
Over the past 20 years, a large number of crustal
geophysical profiles have provided information on the nature and
physical properties of the rocks in the lower crust. Important
observations include: (1) seismic refraction profiles have defined
lower crustal velocities often less than 7.0 km/s, less than
expected for the otherwise inferred dominant mafic composition
[Kay and Kay, 1986; Hyndman and Klemperer, 1989], (2)
magnetotelluric surveys have revealed low resistivity zones in
the deep crust [e.g. Haak and Hutton, 1986]. A possibly related
result is that the lower crust is also commonly seismically
reflective. While many hypotheses have been presented to
explain these properties of the lower crust, the only single
hypothesis that appears to reconcile both seismic and electrical
observations is the presence of up to several percent aqueous
fluids in the lower crust [Hyndman and Hyndman, 1968;
Shankland and Ander, 1983; Gough, 1986]. Hyndman and
Klemperer [1989] have pointed out that if the saline fluid
porosity explanation for conductive lower crusts is correct, it
should affect the seismic velocity as well, i.e. there should be a
relation between velocity and resistivity. Previous compilations
Geodynamics 22
have shown that the averages for Precambrian and Phanerozoic
areas are consistent with the porosity model; both lower crustal
resistivity and velocity averages from Phanerozoic areas are less
than those for Precambrian areas. In this article we present a
comparison of the velocity-resistivity correlation with a
compilation of approximately coincident magnetotelluric
electrical resistivity and refraction seismic velocity data.
Porosity models
The effects of porosity on the elastic and electrical properties
of rocks under lower crustal conditions have been discussed by
many authors (see Hyndman and Shearer [1989] for a review
and references). We have used the P-velocity vs porosity
relations for a rock containing randomly oriented ellipsoidal
pores of given aspect ratios. Models with different pore
geometries all give fairly similar results [Shearer, 1988], and in
any case the conclusions of this paper are not dependent upon
the choice of one velocity-porosity model. We have chosen a
zero-porosity velocity of 7.2 km/s, assuming a mafic
composition for the lower crust [Kay and Kay, 1986]. Our
choice is motivated by the observation of velocities up to 7.5
km/s in certain areas [e.g. Boland and Ellis, 1989]. Again, a
slightly lower zero-porosity velocity would not affect the general
conclusions of this paper, except for the calculated porosity and
aspect ratio values. We have also assumed that the low
resistivity of the deep crust is the result of saline fluid porosity,
although other conduction mechanisms such as graphite cannot
be excluded [Frost et al., 1989; Schwarz, 1990]. We have used
the empirical relation of Archie [Hermance, 1979] that fits many
laboratory measurements on crystalline rocks. A seawater
salinity pore fluid has been assumed. The effects of very well
connected, e.g. tubular, pore shapes has been approximated
using Archie's Law exponent 1.2. Exponents of 1.5 to 2.5
approximate the behaviour of reasonable equilibrium pore
geometries, although with a less abrupt transition from high to
low resistivity [Hyndman and Shearer, 1989].
329
330 DEEP CRUSTAL VELOCITY AND RESISTIVITY
Compilation of Geophysical Results
We present here a summary of a compilation of
approximately coincident seismic refraction velocities and
magnetotelluric resistivities. The detailed compilation and
references cannot be given within the space limitations of this
article and will be presented elsewhere. The refraction and
magnetotelluric survey pairs used were required to be in the
same tectonic unit, usually within horizontal distances of several
tens of km. The layers were restricted to the lower half of the
crust at depths greater than - 15 km. Some localities for which
there are particularly extensive coincident surveys are discussed
in further detail below, but first the compilation as a whole will
be analyzed in order examine the general correlations of lower
crustal layers.
There are two serious difficulties in obtaining resistivities
that correspond to velocity layers. The first is that
magnetotelluric surveys only effectively resolve layer
conductance or thickness/resistivity, rather than resistivity and
thickness [Edwards et al., 1981]. Thus, the resistivity required
if the lower crustal layer were 10 km thick (approx. the data
average) is also given. The second important problem is "static
shift" [e.g. Jiracek, 1990], caused by local near-surface galvanic
effects, resulting in a shift of the whole resistivity-depth profile.
For seismic refraction velocities, the most serious problem is
that lower crustal low-velocity layers have undoubtedly been
missed in older interpretations that were based only on first-
arrivals. Thus, only velocity data that have been obtained
through amplitude synthetic modelling have been used. This
minimizes the possibility that lower crustal layers that do not
produce first arrivals are missed in the interpretation.
Figure 1 presents a range of theoretical relations between
resistivity and velocity based on the resistivity-porosity and
velocity-porosity relations discussed above (see also Hyndman
and Klemperer [1989]), along with the data values from the
compilation. Although the data cover a wide range, there is a
systematic trend of decreasing resistivity .with decreasing
velocity, and most data fit between the two reasonable bounds of
aspect ratio and Archie's law exponent, 0.03:2.0 and 0.1:1.5
respectively. For a velocity of 6.6 km/s, these models require
porosities of 1.8 and 3.2% respectively. A model of very thin
well connected pores (0.01:1.2) is also given. It requires a
porosity of only 0.8% to reduce the velocity to 6.6 km/s. Such a
pore geometry gives a large increase in Poisson's ratio (i.e. 0.25
7.2
7.0
6.8
6.6
6.4
6.2
8.0
Velocity-Resis'tivity Relations
c
4 3 2 1 o
Log Resistivity (ohm-m)
Fig. 1. Velocity-resistivity relations for a porous mafic rock with saline water porosity. The
theoretical curves represent combinations of pore aspect ratio a (for velocity) and Archie's Law
exponent m (for resistivity). Closed symbols denote Precambrian and open symbols Phanerozoic
areas. Circles denote resistivity as reported by the original investigators, triangles resistivities for a 10
km-thick layer of the same conductance. Localities cited in Table 1 are as follows: C, Cascades; R,
Rio Grande Rift; T, Tangshan and V, Vancouver Island.
MARQUIS AND HYNDMAN 331
to 0.28 for 1% porosity) and may explain some data indicating
high Poisson's ratio in the lower crust. It is remarkable,
considering the limitations of the geophysical techniques and the
assumptions made, that the results are not more scattered. The
younger Phanerozoic areas again in general have lower
resistivities and velocities, indicating that if porosity is indeed
the main factor controlling these physical properties, the deep
crusts of Phanerozoic areas generally have higher porosity than
shields. This is to be expected if the crust dehydrates at a slow
steady rate, or dehydrates through recurring metamorphic
events.
Specific Areas
For most localities in the compilation, there are only a few
measurements and the seismic and electromagnetic surveys are
only approximately coincident. However, at a few localities
detailed multidisciplinary geophysical surveys have been carried
out and seismic refraction and magnetotelluric measurements are
along the same profiles. We summarize here briefly: Cascade
Mountains, northwest U.S.A.; Rio Grande Rift, southwest
U.S.A.; Tangshan, China; and Vancouver Island, southwest
Canada. The latter is a special case of a subduction zone. For
these locations, average or typical values of lower crustal
resistivity and velocity have been selected. Table 1 contains
several combinations of aspect ratio, Archie's law exponent and
porosity that can reconcile the seismic refraction and
electromagnetic results. Five Archie's law exponents (m= 1.2 to
2.5) have been chosen that cover a reasonable range for
crystalline rocks as measured in the laboratory, and the required
aspect ratios have been calculated. Both the resistivities given
by the authors and the 10 km resistivity have been used. In
addition porosity values are constrained to less than 5 %, since
we consider larger porosity at depth unlikely because of the
problem of porosity retention for long periods. The acceptable
combinations are in bold type in Table 1.
Cascades
A 400-km long seismic refraction profile has been shot
across the Cascade Range [Leaver et al., 1986], with
magnetotelluric stations along the same line [Stanley, 1984]. The
main feature is a 2-20 ohm m low-resistivity body at mid- to
TABLE 1. Parameters required to satisfy the seismic and electrical data
using p using
Locality Vp P Po m Aspect Porosity Aspect Porosity Poisson's
Ratio (%) Ratio (%) Ratio
Cascades 6.65 10 20
Rio Grande Rift 6.67 10 10
Tangsban 6.2 4 5
Vancouver Island 6.35 30 30
1.2 0.016 0.8 0.004 0.4 0.31
1.5 0.05 2.1 0.02 1.3 0.28
1.8 0.10 3.9 0.06 2.7 0.27
2.0 0.14 5.4 0.10 3.9 0.28
2.5 0.30 9.8 0.21 7.4 0.27
1.2 0.017 0.8 0.017 0.8 0.28
1.5 0.07 2.1 0.07 2.1 0.28
1.8 0.14 3.9 0.14 3.9 0.28
2.0 0.20 5.4 0.20 5.4 0.28
2.5 0.50 9.8 0.50 9.8 0.28
1.2 0.017 1.6 0.014 1.4 0.28
1.5 0.06 3.8 0.05 3.3 0.26
1.8 0.11 6.6 0.10 5.8 0.27
2.0 0.16 8.7 0.14 7.7 0.26
2.5 0.29 14.1 0.26 12.9 0.27
1.2 0.002 0.3 0.002 0.3 0.34
1.5 0.02 1.0 0.02 1.0 0.28
1.8 0.03 2.1 0.03 2.1 0.27
2.0 0.06 3.2 0.06 3.2 0.27
2.5 0.13 6.3 0.13 6.3 0.27
Vp, refraction velocity in km/s; p, resistivity as reported by the original investigators (fl m); Po,
resistivity for a 10 km-thick equivalent layer; m, Archie's Law exponent; Aspect ratio of randomly oriented
ellipsoidal pores. Poisson's ratio calculated assuming a zero-porosity value of 0.29.
332 DEEP CRUSTAL VELOCITY AND RESISTIVITY
lower crustal depths, corresponding to a 6.5 km/s velocity layer.
Under the assumptions described above, these observations fit a
porosity from 1.3 to 3.9%, with the effective ellipsoidal crack
aspect ratio of 0.02 to 0.1. The thin crack porosities require 0.4
to o.8%.
Rio Grande Riff
Sinno et al. [1986] have presented seismic refraction profiles
from the southern Rio Grande Rift, and Herinante and
Perdersen [ 1980] have obtained magnetotelluric results from El
Paso, Texas, about 30 km from the seismic refraction profile.
A lower crustal conductor of 10 ohm m at a depth of 25 km is
the most remarkable feature, approximately coinciding in depth
with a velocity of 6.67km/s. Porosities between 2.1 and 3.9%
and aspect ratios between 0.07 and 0.14 best reconcile the
experimental results. Thin cracks require 0.8 %.
Tangshan
Coincident seismic refraction and magnetotelluric results
have been obtained in the Tangsban seismic zone, in northeast
China [Liu Guodong, unpublished data, oral presentation,
University of Victoria, September 1990]. The most prominent
crustal resistivity feature is a 3-5 ohm m conductive layer in the
lower crust. Corresponding seismic refraction velocities are in
the range 6.0 to 6.4 km/s. These properties can be satisfied
with a 3.3 to 3.8% porosity, with thin cracks of aspect ratio
0.05 to 0.06. Thin cracks require 1.4 to 1.6%.
Vancouver Island
A detailed multidisciplinary geophysical profile has been
obtained across Vancouver Island, where the subdueting Juan de
Fuca plate provides a ready source of fluid for overlying crustal
porosity [Hyndman, 1988]. Crustal velocities have been
presented by Spence et al. [1985] and Drew and Clowes [1990].
Magnetotelluric data has been presented by Kurtz et al. [ 1990].
A high conductivity layer dipping landward with an estimated
resistivity of 30 ohm m corresponds to a layer with velocity of
6.35 km/s in the most recent interpretation. The porosity range
estimated here is 1.0 to 3.2 for aspect ratios from 1.5 to 2.0.
The thin crack model gives 0.3 %.
pore fluid salinity influencing the resistivity. A wide range of
pore geometries can satisfy the resistivity and velocity data. An
important additional constraint may be provided by shear wave
and Poisson's ratio data. Thin pores of small aspect ratios
(<0.01) will increase Poisson's ratio rapidly with increasing
porosity, while thicker pores will decrease it somewhat [Shearer,
1987]. For rocks of marie composition at deep-crustal
conditions, a zero-porosity Poisson's ratio value of 0.29 can be
used (see Goodwin and McCarthy [1990]). We have calculated
the expected value of Poisson's ratio for each porosity-aspect
ratio combination in Table 1.
A few estimates of Poisson's ratio are available from areas
where both P- and S-velocities are known with sufficient
accuracy. Several examples are noted here. Hall [ 1987] reports a
Poisson's ratio of 0.25 in Lewisian rocks, Scotland. This would
imply either quite thick pores, or a less marie lower crustal
composition. Controlled-source shear-wave experiments are still
at the development stage, but Ward and Warner [ 1990] have
obtained a deep-crustal Poisson's ratio of approximately 0.33 in
northern England. Cassidy and Ellis [personal communication,
1990], using P- to S- converted phases, have obtained a value
between 0.29 and 0.32 for the 'E' reflecting band under
Vancouver Island. With this additional constraint, very thin
pores and low porosities, often less than 1% (see Vancouver
Island entry in Table 1), can reduce the velocity and if they are
well interconnected, they can also explain the observed low
resistivities.
Conclusion
Magnetotelluric and seismic refraction results support a
common dependence on porosity for electrical resistivity and
seismic velocity. Modelling indicates that the porosity required
to reconcile both observed parameters is less than 3 percent,
depending on the pore geometry, which can in turn be better
constrained using Poisson's ratio data.
Acknowledgments. This work has benerited greatly from
discussions with Lawrie Law and Trevor Lewis. GM wishes to
acknowledge rinancial support from an NSERC Post-Graduate
Scholarship and a University of Victoria President's Research
Scholarship. Geological Survey of Canada Contribution 57390.
References
Discussion
The data presented above suggesting a correlation between
deep crustal velocity and resistivity provide some support for the
presence of fluids in the deep crust, particularly in younger
areas. While other explanations for the correlation can be
suggested such as graphite being associated with lower resistivity
rocks, it is difficult to explain the contrast between shields and
younger areas. The most important unknown in estimating the
amount of fluid required to satisfy both the velocity and
resistivity data is the pore geometry, although composition is
also a very important factor influencing the velocity, and so is
Boland, A.V., and Ellis, R.M., Velocity structure of the
Kapuskasing Uplift, northern Ontario, from seismic
refraction studies, J. Geophys. Res., 94, 7189-7204,
1989.
Drew, J.J., and Clowes, R.M., A re-interpretation of the
seismic structure across the active subduction zone of
western Canada, in Studies of laterally heterogeneous
structures using seismic refraction and reflection data,
edited by A.G. Green, Geol. Surv. Can. Pap. 89-13,
115-132, 1990.
Edwards, R.N., Bailey, R.C., and Garland, G.D., Conductivity
anomalies: lower crust or asthenosphere?, Phys. Earth
Planet. Int., 25, 263-272, 1981.
MARQUIS AND HYNDMAN 333
Frost, B.R., Fyfe, W.S., Tazaki, K., and Chan, T., Grain-
boundary graphite in rocks from the Laramie
Anorthosite complex: implications for lower crustal
conductivity, Nature, 340, 134-136, 1989.
Goodwin, E.B., and McCarthy, J., Composition of the
lower crust in west central Arizona from three-
component seismic data, J. Geophys. Res., 95, 20097-
20109, 1990.
Gough, D.I., Seismic reflectors, conductivity, water and
stress in the lower continental crust, Nature, 323, 143-
144, 1986.
Haak, V., and Hutton, R., Electrical resistivity in
continental lower crust, in The Nature of the lower
continental crust, edited by J.B. Dawson, D.A. Carswell, J.
Hall, and K.H. Wedepohl, Geol. Soc. Spec. Publ. 24, 35-
49, 1986.
Hall, J., Physical properties of Lewisian rocks: implications
for deep crustal structure, in Evolution of the Lewisian
and comparable Precambrian high-grade terrains, edited
by R.G. Park and J. Tarney, Geol. Soc. Spec. Publ. 27,
185-192, 1987.
Hermance, J.F., The electrical conductivity of materials
containing partial melt: a simple model from Archie's
Law, Geophys. Res. Lett., 6, 613-616, 1979.
Hermance, J., and Pedersen, J., Deep structure of the Rio
Grande Rift: a magnetotelluric interpretation, J. Geophys.
Res., 85, 3899-3912, 1980.
Hyndman, R.D., Dipping seismic reflections, electrically
conductive zones and trapped water in the crust over a
subducting plate, J. Geophys. Res., 93, 13391-13405,
1988.
Hyndman, R.D., and Hyndman, D.W., Water saturation
and high electrical conductivity in the lower continental
crust, Earth Planet. Sci. Lett., 4, 427-432, 1968.
Hyndman, R.D., and Klemperer, S.L., Lower crustal
porosity from electrical measurements and inferences
about composition from seismic velocities, Geophys.
Res. Lett., 16, 255-258, 1989.
Hyndman, R.D., and Shearer, P.M., Water in the lower
continental crust: modelling magnetotelluric and seismic
reflection results, Geophys. J. lnt., 98, 343-365, 1989.
Jiracek, G.R., Near-surface and topographic distortions in
electromagnetic induction, Surv. Geophys., 11, 163-203,
1990.
Kay, R.W., and Kay, S.M., Petrology and geochemistry of
the lower continental crust: an overview, in The Nature of
the lower continental crust, edited by J.B. Dawson, D.A.
Carswell, J. Hall, and K.H. Wedepohl, Geol. Soc. Spec.
Publ. 24, 147-159, 1986.
Klemperer, S.L., A relation between heat flow and the
seismic reflectivity of the lower crust, J. Geophys., 61,
1-11, 1987.
Kurtz, R.D., DeLaurier, J.M., and Gupta, J.C., The electrical
conductivity distribution beneath Vancouver Island: a region
of active plate subduction, J. Geophys. Res., 95, 10929-
10946, 1990.
Leavet, D.S., Mooney, W.D., and Kohler, W.M., A seismic
refraction study of the Oregon Cascades, J. Geophys. Res.,
89, 3121-3134, 1984.
Schwarz, G., Electrical conductivity of the Earth's crust and
upper mantle, Surv. Geophys., 11, 133-161, 1990.
Shankland, T.J., and Ander, M.A., Electrical conductivity,
temperatures, and fluids in the lower crust, J. Geophys.
Res., 88, 9475-9484, 1983.
Shearer, P.M., 1988. Cracked media, Poisson's ratio, and
the nature of the upper oceanic crust, Geophys. J., 92,
357-362, 1988.
Sinno, Y.A., Daggett, P.H., Keller, G.R., Morgan, P., and
Harder, S.H., Crustal structure of the southern Rio
Grande Rift from seismic refraction profiling, J. Geophys.
Res., 91, 6143-6156, 1986.
Spence, G.D., Clowes, R.M., and Ellis, R.M., Seismic
structure across the active subduction zone of western
Canada, J. Geophys. Res., 90, 6754-6772, 1985.
Stanley, W.D., Tectonic study of the Cascade Range and
Columbia Plateau in Washington State based upon
magnetotelluric soundings, J. Geophys. Res., 89, 4447-
4460, 1984.
Ward, G., and Warner, M., Implications for lower crustal
lithology from shear wave reflection data (abstract), in
Fourth international symposium on deep seismic reflection
profiling of the continental lithosphere, 36, 1990.
DETERMINATION OF THE FINE STRUCTURE OF SEISMIC CRUSTAL
BOUNDARIES FROM THEIR REFLECTED PHASES
Ramon Carbonell and Scott B. Smithson
Program for Crustal Studies Department of Geology and Geophysics
University of Wyoming, P.O. Box 3006, Laramie, Wyoming, 82071
Abstract The characteristics of the wave form of layers, intrinsic attenuation, scattering due to
a particular event in a reflection (normal small heterogeneities, three dimensionality and
incidence and/or wide-angle) record section are the source excitation function, interference with
due to complex interference of individual wave reflected phases from shallower levels, etc.
fields generated by different physical phenomena: Forward modeling of the reflected phase and/or
1) the intrinsic seismic response of the target its amplitude spectrum has been used for the
boundary, 2) scattering due to small determination of the fine structure of crustal
heterogeneities, 3) the source excitation boundaries [Meissner, 1965; Clowes and
function, 4) stratigraphic filtering, 5) Kanasewich; 1986, Louie and Clayton, 1987]. In
intrinsic attenuation, 6) three dimensionality, addition the characteristics of its amplitude
7) the interference with converted (Ps, Sp) spectrum and of its coda are indicative of
phases from shallower levels among others. layering (i.e., fine structure) in the lower
Spectral ratios and/or qualitative comparisons of crust [Paul and Campillo, 1989; Paul, 1990]
the amplitude spectrum of different phases are and/or the complexity of the target boundary
often used in studying intrinsic attenuation and [Mereu, 1990]. Numerical modeling of the seismic
crustal structure. However, this approach wave propagation phenomena demonstrates that the
disregards all the information contained in the seismic signature of a reflecting boundary is
phase spectra. Thus the deconvolution (spectral determined by its fine structure. For instance,
ratio) of the target event (SmS, PmP, etc) by an the seismic response of a thinly laminated zone,
immediately preceding arrival removes the at normal incidence, has frequency peaks and
contaminating effects due to the interfering troughs which are a function of the average
processes that can be regarded as filters and thickness of the layering [Berzon, 1965;
provides estimates of the reflection coefficients Khudzinskii, 1962; Fuchs, 1969; Starodubrovskaya,
and/or impedance contrasts. This can be done 1969]. The seismic response of a discontinuity
through O'Doherty-Anstey formula or by consisting of blocks with random velocities
considering mathematical developments that take present particular features in the amplitude
into consideration the reflection and spectrum of their reflected phases [Davydova et
transmission properties of the interfaces. These al., 1970; 1972]. Even small heterogeneities
estimates place important constraints on the within the target structure and the topography of
determination of the structure and geological the reflecting boundary will produce
nature of crustal discontinuities. characteristic seismic signatures [Ojo and Mereu,
1978; Gibson and Levander, 1988; Paul and
Introduction Campillo, 1988; Mereu, 1990].
Some of the disrupting wave fields that have
The study of the fine structure of crustal interfered with the seismic response of the
boundaries is commonly obtained from the dynamic target structure (e.g., Moho) can be removed by
characteristics of their reflected phases; deconvolving the reflected phase (e.g., PmP) by
however the recorded seismic wave field is not an preceding arrival. This procedure removes from
only the response of the actual structure (for the recorded (raw) arrival the effects of all the
instance the Moho transition zone) but a complex processes that affected the wave field which can
superposition of numerous phenomena such as: be regarded as filters. Spectral ratios have
stratigraphic filtering due to the overlying often been used to describe the Earth's crust
[Phinney, 1964], to estimate intrinsic
attenuation [Rebollar, 1984; Hobbs, 1990] and to
Continental Liosphere: Deep Seismic Refltions study the fine structure of crustal boundaries
Geodynamies 22 [Meissner, 1965 ]; however, the information
O l991AneficanGeophysicUnion contained in the phase spectrum has often been
335
336 DETERMINING FINE STRUCTURE OF CRUSTAL BOUNDARIES
disregarded. Berzon [1965] demonstrated the use structure. The overlying heterogeneities and/or
of peaks and troughs, appearing in the amplitude the topography of shallower reflecting boundaries
and in the phase spectra, to compute thicknesses strongly disrupt the wave fronts traveling
and velocities of the layering. The spectral through them, by deforming them and
ratio provides estimates of the reflection redistributing .their energy (focusing and
coefficients (impedance contrasts) for deep defocusing them) so that a single continuous
structures. This can be evaluated through the reflector can appear as a wide band of short
relationship between the spectral ratio and multicyclic, coherent events [Carbonell and
stratigraphic filtering [O'Doherty and Anstey, Smithson, 1990]. Reflections from out of the
1971; Menke, 1983] or by considering the plane will contribute to the recorded wave field
amplitude of the target and reference events as a [Blundell and Raynaud, 1986]. The events
function of the reflection and transmission attributed to such reflections are in most cases
properties of the target and reference indistinguishable from actual structure. The
structures. interference effects due to wave-type
superposition, scattering of the overlying
Contributions to the Recorded Wave field structure, and three dimensionality cannot be
represented by filters, and thus they cannot be
The convolutional approach models the seismic removed from the recorded signal.
response of a stratified medium by convolving the
source wavelet with the reflection coefficient Mathematical Developments
series, thus if the source function is known (for
instance when using vibroseis and in some cases Thickness estimates can be obtained from the
air guns) the reflectivity function can be amplitude and phase spectra. The time lag x
recovered from the recorded wave field by between the two phases reflected from the top and
performing the inverse operation, (i.e., bottom of a thin layer (of thickness z) is:
deconvolution). Unfortunately the source
excitation function is not always available, such
2 z cos 8
is the case of data acquired by explosion x = (1)
v
sources, particularly ripple fired quarry shots.
Since our targets are deep structures the
recorded wave field is far more complicated. The where 8 and v are the angle of incidence at
structures overlying the target will modify the the bottom interface and the velocity of the
wave field as it travels through them. Thus in layer respectively. The spectral components of
order to obtain physically significant both phases (each frequency) will interfere
information on the target structures from the constructively or destructively generating peaks
recorded signal we need to remove the (maxima) or troughs (minima), in the resulting
interference effects due to the overlying wave train. The maxima and/or minima appearing in
structures. The overburden geology will introduce the amplitude and phase spectra of this wave
a stratigraphic filtering due to intrabed train depend on the sign of the reflection
multiples generated within the layering coefficients at both interfaces and on the phase
[O'Doherty and Anstey, 1971; Schoenberger and shift introduced by the thickness of the thin
Levin, 1974, 1978; Menke 1983; Banik et al., layer (i.e., time lag between the two
1985a, b; Resnik, 1990]. This is a minimum phase reflections). A maximum in the amplitude spectrum
filter that depends on the impedance fluctuation of the resulting wave is the contribution of the
[Banik et al., 1985a]. two frequency components that are in phase;
A great variety of grain scale processes similarly, a minima is the contribution of two
contribute to intrinsic attenuation. The main frequency components that have opposite polarity
mechanism of attenuation is internal friction. (phase). Constructive and destructive
Intrinsic attenuation is assumed to be frequency interference require a phase shift of
independent, a valid approximation for high Q A0 = 2 n and A0 = (2n-1) (where n is an
(quality factor) crystalline rocks. The integer) respectively. Considering equation (1),
mechanisms of intrinsic attenuation are discussed the phase shift for constructive interference,
elsehere [Knopoff, 1964; Mason, 1964; Walsh, and the fact that the velocity of the layer is
1965; Savage, 1966; Jackson and Anderson, 1970]. larger or smaller than the velocities of the
The associative property of the convolution layers above and below, the location of the peaks
operation allows the combination of attenuation in the amplitude spectra is given by:
from intrabed multiples and intrinsic attenuation
(both can be regarded as filters) into a single (2 n - 1) v
2
operator. This operator represents the = (2)
contribution of the overlying structure, and can n 4 z cos 82
be removed from the recorded wave field by a
deconvolution procedure. Therefore by measuring the location of the maxima
Wave conversions (Ps, Sp) at shallower levels an minima on the amplitude and phase spectra we
interfere with the deeper Pp reflected phases can obtain estimates of the average thicknesses
masking the seismic response of the target of the layered structure by using (2) if the
average velocity is known. Knowledge on the As this is not the case (Figure 1) a more
average velocity can be acquired from one realistic approach consists in developing both
dimensional inversion methods. phases (al(t) and a2(t)) as function of the
The apparent attenuation observed at normal acoustic reflection and transmission
incidence of a wave train transmitted through a coefficients. In the frequency domain, these
layered structure which is due to intrabed phases (Ai() and A2()) can be written as:
multiples was obtained by O'Doherty and Anstey
[1971] and can be written in the frequency domain
as:
A(9,t) - R2(9)
= ex{ }
A(u,0) 2
A(,0) and A(9,t) are the amplitude at a
frequency 9 after having propagated a time 0
t
Ai( ) = R (9) S(9) D 1 exp { - 1
ppl
A2(9 ) = Rpp2(U ) T (9) (n) S(n) ppl Tpp2
9t t 2
D 2 e xp {- (O- 02 )
1 } (8)
(9)
and t respectively, R2(v) is the power spectrum where R (v), T (v), R 2(v) and T 2(v) are
coefficient anP{ansmon
the layered structure and x is its vertical an incident P-wave reflected as a P-wave) for the
travel time. The simple and more detailed first and second interface respectively, S()
derivations of this mathematical expression and is the source excitation function, D 1 and D
its physical significance can be found in are the spherical divergence factors and te
O'Doherty and Anstey [1971], Schoenberger and exponential terms in both equations account for
Levin [1974], [1978], Banik et al, [1985a, b], the attenuation. Then the reflection coefficient
Richards and Menke [1983], Resnik [1990] among of the target structure is:
others.
Menke [1983] extended equation (3) to a
randomly layered media, characterized by a mean R (u) -
velocity v, a mean impedance I with a rms pp2
fluctuation o I, and layers with average
thicknesses h obtaining
A2() R (9) D 1
.ppl
Al(9) Tppl() Tpp2(9) D 2
t
2
}
(10)
A(,t) = exp { - t } (4) Equation (10) can be applied to any angle of
A(9,0) v) incidence provided we know the appropriate
R 1() and T () of:. the reference interface.
1 I h__ I (5) Coecting foPlspherical divergence assuming
= [ [ v '
high Q (not an important factor in crystalline
terranes) and vertical incidence, expression
where 0(9) is the quality factor at a (1.14) reduces to:
frequency .
From equation (3) we can obtain the power
spectrum of the reflection coefficient as :
2
2 t in { A(,t) } (6)
R () = - t A(,0)
and from (4) and (5):
o I 2
[__)2 = _ 2v in { A(,t) } (7)
2 A(9,0)
ht9
A2(9 ) R (9)
pp1
pp2 A1 -Rppl pp2
expression (11) is a second order equation in
R 2 and we can estimate its value if we have
so information on the reflection and/or
transmission coefficients of the reference
interface. Brocher and Phinney, [1981] extended
this approach to - p transformed data.
Both theoretical developments involve the
normal incidence assumption. However, the weak
o dependence of the reflection coefficients with
As R2() and (_I)2 are physically offset at subcritical distances is not expected
meaningful quantiies this approach provides to be a significant source of error, at least not
information on the physical parameters of the as significant as the error introduced by three
target structure. Since equation (3) is an dimensionality and scattering due to the
attenuation factor ( i.e., apparent attenuation overburden geology. Since equation (6) is an
introduced by intrabed multiples), equation (6) attenuation factor ( i.e., apparent attenuation
requires assumptions that closely parallel the introduced by intrabed multiples), equation (7)
ones needed for the estimation of the intrinsic requires assumptions that closely parallel the
attenuation (i.e., Q, quality factor). Both ones needed for the estimation of the intrinsic
inverse equations ((6) and (7)) assume that the attenuation. Due to stability requirements a
second phase is a transmitted wave. smoothing operator (approximately a 5 point
10. 0 - - I , , I I ' ' ' I-[TT 1-!
15.0
6.0 ,o I.O 0.0 IO.O 20.0 30.0 40.0 50.0
AveraRe Th Ickfleoa
AveraBe veloci(y Peak Frequerlc. y F:fitl.Ate d re-i
(km/.) (hz) (m) (m)
7.6 11.9 159.7 150.0
10.0 r,,,|,,, 1.0 ........... I .... T-r 1
15.0 :.,
, 0.0 ,Jl , ,f,,,
.L 0.0 IO.O 20.0 30.0 40.0 5,0.0
, &.o o.o
15.o ijj
6.o i.o
.0 *,,, ,,, ,, [,,, r]-vrr
0.0 O.O :SO.O .10.0 40.0 SO.O
7. Ii .5 160.9 150.O
7.2 IO.O IUO.O 175.O
,,I,r.p 1 -r-rrT
$s.o o.o 1 .... l
. o.o lo.o 2o.o )o.o 40.0 $o.0
6.o
7.4
10.0
185 0
181.0
Veloc i t;y ( ke/o ) IP[equency ( hz )
Fig. 1. Amplitude spectra of the synthetic velocity, peak frequency, estimated and real
response of layered structures. P-wave velocity average thicknesses (right).
profiles (left), amplitude spectra (center),
operator) has to be applied to the amplitude and amplitude spectra of a seismic response of
spectra before performing the spectral ratio. the layered structures. This illustrates that
Estimates of the acoustic reflection and/or thicknesses obtained by using equation (2) agree
transmission coefficients for the reference with the real average thicknesses of the layering
interface are needed to use equation (11) while within 50 m. The first two models in Figure (2)
only knowledge on the average velocities and demonstrate the independence of the peak
travel times are needed to apply equation (6) and frequency of the amplitude spectra with respect
(7) these velocities can be obtained from one to the magnitude of the reflection coefficient.
dimensional inversion techniques. The estimated The synthetic data used to test equations (6)
reflection coefficient is interpreted as the and (11) was generated by using the same source
reflection coefficient needed for an individual function and the reflectivity algorithm, and
interface to produce the observed amplitude. models that consisted of two interfaces. Figure
(3a) shows the seismograms for one example where
Application the reflection coefficients were 0.105 and 0.245
for the top and bottom interfaces respectively.
We tested the previous formulae for the Before preforming the spectral ratio we added to
determination of average thicknesses and the seismograms some random noise with the same
reflection coefficients using synthetic data frequency band than the signal (5-25 hz), and
generated by an elastic reflectivity algorithm with a peak amplitude for the noise of 25% of the
[Kennett, 1983; Muller, 1985]. The source maximum amplitude of the signal. The comparison
function, in the frequency domain, is a tapered between the results obtained by (6) and (11) when
boxcar of unit amplitude between 5 to 25 hz. they are applied to synthetic data sets is show
Figure (2) shows a series of velocity profiles in Figure (3b) which reveals the that both
s(t] &l (t) &2 (t)
Dote Truce
, &l (t)
r &2 (t) Amplitude Spectrum
ir 1
Amplitude Spectrum
100

.
0 12 24 36 48 60
frequency (In)
0 12 2& 36 8 6o
frequency
(a) (b) ()
Fig. 2.a: Diagram showing the raypaths followed and the two phases considered, c: The amplitude
by two consecutive arrivals at) and a2(t ). . spectra of the two phases.
s(t) is the source function. : The seismogram
, -,,,, u- u ,_,4,, ,,-
0o1; "
! -
M 0. ' --
0.210
.0 2.$ $.0 ?.$ 10.0 12.$ 15.0 17.$
Xnc idence Arkie
Ibl
approaches were successful in estimating the
reflection coefficients. Equation (11) gives more
consistent results (with less scatter) than
equation (6) even with the presence of noise
(Figure 3b). The equivalent of equation (11)
applicable to x - p transformed data is more
complicated and its development and application
can be found in Brocher and Phinney, [1981]. The
estimated error for equations (6) and (11) is
strongly dependent on the signal-to-noise ratio.
The previous developments were used in
determining the structure and nature of the
Mohorovicic discontinuity in northwestern Nevada
from the wide- angle seismic reflection data set
of the Nevada PASSCAL 1986 lithospheric
experiment. The data set presents a high
amplitude, laterally continuous PmP phase (P-wave
reflected from the Moho) comparable to 'other
studies [Klemperer, 1987]. This Moho reflection
is a prime target for the application of
equations (6) or (11) to estimate the reflection
coefficients and to study the peak frequencies of
the amplitude spectra in order to estimate the
average thickness of the layering. The amplitude
Fig. 3. Spectral ratio results from a synthetic spectra of the PmP phase in individual traces and
test. a: Synthetic seismograms obtained by a the average amplitude spectrum of a group of
reflectivity method [Kennett, 1983; Muller, traces also at near vertical incidence present a
1985]. A: Reference phase, B: target phase. b: peak at 24 Hz (Figure 4a). This frequency and the
Reflection coefficients evaluated from the average velocity derived from one dimensional
synthetic seismograms as a function of angle of inversion schemes together with equation (2) give
incidence. (-) Reflection coefficients estimated an estimated average thickness for the layering
from O'Doherty and Anstey's formula equation of 85-250 m. The average velocity was obtained
(11). ( ) Reflection coefficient estimated by from the average velocity functions of Hawman et
using equation (15). al. [1990]. These same velocities functions were
o. 14.e #.
O.O _
,14 48 O O.O 14.0 #.
4.0
I
11.0
OttOet (kB)
(b)
frequency (hz)
I.I |J.O 14.0 #.J 11.J #.lo
x5.o "'1" '1 .... l'"'l"
lO.O
CJ S.O
e.o .-. ,,J ,,I,,,! ....
Reflection Coefficients
(c)
(a)
Fig. 4. a: Amplitude spectra for individual (true offsets are not represented). ( * )
traces (left) and average amplitude spectra for Reference phase, basement reflection. ( ) PmP
25 seismograms of the PmP phase at subcritical phase (P-wave reflected from the Mohorovicic
offsets for the seismic reflection data of the discontinuity). c: Percentage of the reflection
Nevada PASSCAL 1986 seismic experiment. b: coefficients obtained by equation (15) for the
Selected seismograms from the Nevada PASSCAL 1986 Nevada PASSCAL data set at subcritical offsets
Basin and Range Lithosphere seismic experiment 300 seismograms where used.
used to perform the spherical divergence reflection coefficient possible for a single
corrections in order to apply (11), for the interface at Moho depth required to produce the
estimation of the reflection coefficients. We observed reflected amplitude. This reflection
applied equation (11) to seismograms from shots coefficient can be partially a result of
recorded along the north-south and east-west constructive interference of the layered
lines; an example of the seismograms at different structure.
offset is shown in Figure (4b). We use offsets up As constructive interference due to layered
to 10.0 km (this corresponds to incidence angles geologic sequences can double [Spaargaren and
for the Moho of < 10), which assures the normal Warner, 1990] or even triple the reflected
incidence assumption. The basement reflection was amplitudes [Fuchs, 1969; Hurich and Smithson,
used as the reference phase, its reflection 1987] we studied the velocity fluctuations
coefficient was estimated to be of the order of required in a stratified medium to generate a
.2. The statistical distribution of the estimates amplitude reflection comparable to a reference
for the reflection coefficients obtained by event generated by a single interface with a
equation (11) for the Nevada PASSCAL data set for reflection coefficient of 0.2. This reflectivity
subcritical offsets up to 10.0 (for angles of modeling demonstrates that interlayering of high
incidence < 10 ) has a maxima around 0.24 0.02 (7.90.2 km/s) and low (below 6.5 km/s) velocity
(Figure 4c). This is interpreted as the minimum layers are required to generate a reflection with
plodel
lJ.o
6.0 o.o
!.! t.lt
o 3. . o
Velc)city TiN (s)
("/") ( a )
leode1
True lel&tive
Ampi I rude
e lojr True Relative
-pl ltude
s.e s.o ................ I .... I .... h,,J ...........
_ : _:
&.O I.O "- - , ,,;'l , ' ,,' . -"'-"-,
o 3. . o 3. 4.
Velocity TiM
( km/s )
(b)
layering of the reflecting structure. The
location of these peak and troughs is independent
of the reflection coefficients, and only depends
on the average thickness of the layered sequence.
Spectral ratios of the target event by a
reference arrival can give estimates of the
reflection coefficients thus constraining the
geologic nature of the target structure. All the
wave kinematic processes that disrupt the
reflected wave field and that can be regarded as
filters, such as intrinsic attenuation and
stratigraphic filtering due to the overburden
geology, are removed from the recorded wave field
by the spectral ratio technique. Estimates of the
4. ijj physical properties (i.e. reflection
coefficients, impedance contrasts) are obtained
from the spectral ratio approach through the
application of the O'Doherty and Anstey [1971]
equation and/or considering a development that
takes into account the acoustic reflection and
transmission properties of the structures. This
latter approach gives more reliable estimates.
Both developments are mathematically derived for
normal incidence. The application of these
approaches to the Nevada PASSCAL large aperture
seismic reflection recordings reveals a layered
structure for the Moho with average thicknesses
within the range of 85 to 250 m. and with an
equivalent reflection coefficient (for the whole
layered sequence) of 0.24 0.02. Reflectivity
modeling requires interlayering of low (below 6.5
km/s) and high velocities (7.90.2 km/s) to
generate reflected amplitude comparable to the
amplitude generated by an interface with
reflection coefficient of 0.2. These velocities
agree with mafic (mantel) material interlayered
with low velocity (probably partial melt)
granitic, intermediate or, more likely, basaltic
Fig. 5. Synthetic seismograms obtained by a composition rocks. This represents a first step
towards the estimation of the fine structure of
reflectivity algorithm for layered structures
deep crustal boundaries from seismic reflection
located below a reference interface with a
measurements.
reflection coefficient of 0.2. a:Seismic response
for structures with low velocities above 6.5. b:
Seismic response for layered structures with low Acknowledqments. This research was partially
velocities below 6.5 km/s. supported by NSF grant EAR-8519153.
References
amplitudes comparable to the amplitudes obtained
from the reference interface (Figure 5). These Banik, N.C., Lerche, I., Shuey, R.T.,
high and low velocity layers may be interpreted Stratigraphic filtering, Part I: Derivation
as representing a structure consisting of a crust of O'Doherty-Anstey formula, Geophysics, 50,
mantel mix comprised of an interlayering of 2768-2774, 1985a.
mantel (mafic, high velocity) material and Banik, N.C., Lerche, I., Shuey, R.T.,
partially melted rocks of granitic to
intermediate or, more likely, basaltic
composition [Smithson and Johnson, 1989]
Conclusions
The analysis of the kinematic characteristics
Stratigraphic filtering, Part II: Model
spectra, Geophysics, 50, 2775-2783, 1985b.
Berzon, I.S., The determination of a thinly
layered medium by the simultaneous use of
amplitude and phase spectrum characteristics
of the layer, Bulletin (Izv.) of the Academy
of Science USSR, 5, 363-367, 1965.
of the reflected wave form places important Brocher, T. Phinney, R.A., 1981, A ray parameter-
constraints on the fine structure and geologic intercep-time spectral ratio method for
nature of deep crustal boundaries. The features seismic reflectivity analysis: Journal of
of the amplitude and phase spectra (for instance Geophysical Research, 8--6, 7865-7873.
the location of the maxima and minima) give Blundell, D.J., and Raynaud, B., Modeling lower
information on the average thickness of the crust reflections observed on BIRPS profiles
in Barazangi, M., and Brown, L., eds., Menke, W., A formula for the apparent attenuation
Reflection Seismoloq: The continental crust, of acohstic waves in randomly layered media,
Volume 14: Washington, D.C., AGU, Geodynamic Geophys. J. R. astr. Soc. 7--5, p. 541-544,
Series, p. 287-295, 1986 1983.
Carbonell, R., and Smithson, S.B., Large scale Mereu, R.F., The complexity of the crust from
anisotropy within the Basin and Range refraction/wide-angle reflection data,
province, Geology, (in press) 1991. PAGEOPH 132, 269-288, 1990
Clowes, R.M., Kanasewich, E.R., Seismic Muller, G., The reflectivity method: a tutorial,
attenuation and the nature, of reflecting J. Geophys., 53, 153-174, 1985
horizons within the crust, J. Geophys. Res., O'Doherty R.F., and Anstey, N.A., Reflections on
7-5, p. 6693-6705, 1970. amplitudes, Geophys. Prosp. 19,430-458,1971.
Davydova, N.I., Kominskaya, I.P., and Michota, Ojo, S.B., and Mereu, R.F., The effect of random
G.G., The thickness and nature of seismic
discontinuities based on deep seismic
sounding data, Tectonophysics, 1--0, 561-571,
1970.
Davydova, N.I., Kosminskaya, I.P., Kapustian,
N.K., and Michota, G.G., Models of the
Earth's crust and M-boundary, J__. Geophys. 38,
369-393, 1972.
Fuchs, K., On the properties of deep crustal
reflectors, J. Geophys. 3-5, p. 133-149.,1969.
velocity functions on the travel times and
amplitude of seismic waves: Geophys. J. R.
astr. Soc. 84, p. 607-618, 1986.
Paul, A., Implications of the lamination of the
lower crust on seismic regional phases,
(abstr.) XXII General Assembly European
Seismological Commission, 1990.
Paul, A., and Campillo, M., Diffraction and
conversion of elastic waves at a corrugated
interface, Geophysics, 53, 1415-1424, 1988.
Gibson, B.S., and Levander, A.R., Lower crustal Paul, A., and Campillo, M., Thin crustal layering
reflectivity patterns in wide-angle seismic in northern France: observations and modeling
recordings: Geophys. Res. Let. 1-5, p. 617- of the PmP spectral content, Geophys. J. Int.
620, 1988. 99, 229-246, 1989.
Hawman, R.B., Colburn, R.H., Walker D.A., and Phinney, R.A., Structure of the Earth's crust
Smithson, S.B., Processing and interpretation from spectral behavior pf long-period body
of refraction and wide-angle reflection data waves, J. Geophys. Res. 69, 2997-3017, 1964
from the 1986, Nevada PASSCAL Experiment, J. Rebollar, C.J., Calculation of Q using the
Geophys. Res. 95 4657-4691, 1990. spectral ratio method in northern Baja
Hobbs, R.W., Effective Q determination using California, Mexico, Bulletin of the
frequency methods on BIRPS data, Seismological Society of America, 7--4, 91-96,
Tectonophysics, 173, 25-30, 1990. 1964.
Hurich, C.A., and Smithson, S.B., Compositional Resnik, J.R., 1990, Stratigraphic filtering,
variation and the origin of deep crustal PAGEOPH 132, 49-65., 1990.
reflections, Earth and Planetary Science Savage, J.C., Thermoelastic attenuation of
Letters, 85, 416-426, 1987. elastic waves by cracks, J. Geophys. Res. 71,
Jackson, D.D., and Anderson, D.L.,, Physical 3929-3938, 1966.
mechanisms of seismic-wave attenuation, Schoenberger, M., Levin, F.K., Apparent
Reviews of Geophysics and Space physics, , attenuation due to intrabed multiples,
1-63, 1970. Geophysics, 39, 278-291, 1974
Kennett, B.L.N., Seismic waves i__n stratified Schoenberger, M., Levin, F.K., Apparent
media, Cambridge Univ. press, 342, 1983 attenuation due to intrabed multiples, II,
Khudzinskii, L.L., On the determination of Geophysics, 43, 730-737, 1978.
parameters of layers of intermediate Smithson S.B., Johnson, R.A, Crustal Structure,
thickness from the spectra of reflected of western U.S. based on reflection
waves, Bulletin (Izv.) of the Academy of seismology, in Pakiser, L.C., Mooney, W.D,
Science USSR 5, 439-444, 1962. Geophysical framework of the continental
Knopoff, L., Q, Reviews of Geophysics and space United States, Geological Society of America
physics, 2, 625-660, 1964. Memoir, 172, Boulder, Co. 1989
Kominskaya, I.P., and Kapustian, N.K., The wave Spaargaren, B., and M. Warner, Constructive
field associated with a fine structured Moho interference: Geological mythology re-
in continents and oceans, J. Geophys. 45, examined, (abstr..) th. International
159-170, 1979. Symposium o__n Deep Seismic Reflection
Louie, J.N., and Clayton, R.W., The nature of Profiling of the Continental Lithosphere,
deep crustal structures in the Mojave desert, 1990.
California, Geophys. J. R. astr. Soc., v. 89, Starodubrovskaya, S.P., Makushkina, T.S., and
p. 125-132, 1987. Vilkova, Physical basis for the use of
Mason, W.P., Internal friction mechanisms that reflected wave forms in studying groups of
produces an attenuation in the Earth's crust thin beds, in Berzon, I.S., Seismic waves in
proportional to the frequency, J. Geophys. real media, Consultants Bureau, New York, 80-
Res. 74, 4963-4966, 1969. 85, 1969
Meissner, R., An interpretation of the wide angle Walsh, J.B., Seismic wave attenuation in rock due
measurements in the Bavarian Molasse basin, to friction, J. Geophys. Res. 71, p. 2591-
Geophys. Prosp. 24, 7-16, 1965 2599,'1965.
LOWER CRUSTAL LITItOLOGY FROM SI1EAR WAVE SEISMIC REFLECTION DATA
Gavin Ward
BIRPS, Bullard Laboratories, Cambridge CB3 0EZ, United Kingdom
Mike VVarner
Department of Geology, Imperial College, London SW7 2BP, United Kingdom
and the British Institutions Reflection Profiling Syndicate (Catherine Flack, Robert Hardy, Richard Hobbs, Simon Klemperer,
Drum Matthews, Carolyn Peddy, Richard Scott-Robinson, Dave Snyder and Nicky White)
Abslracl. WISPA (eardale Integrated S and P-wave Analysis)
was BIRPS' first major venture on land. The aim of this experiment
was to image the lower crust using both P and S-waves at near-normal
incidence, and to measure the variation of Poisson's ratio with depth.
Initial results were excellent, with raw S-wave shot gathers showing
clear lower crustal layering corresponding to the P-wave reflectivity.
Conventional CDP stacking of shear-wave data proved difficult be-
cause of severe S-wave statics problems.
The data are of sufficient quality to allow direct and accurate es-
timation of Poisson's ratio. gVe obtain an average value of 0.23
(Vr:V = 1.70) for the whole crust. However, this figure disguises
a clear and systematic change of Poisson's ratio with depth through
the crust; in the upper crystalline crust values are low, around 0.20
(Vr:V = 1.64), while in the layered lower crust values rise to an
average of 0.29 (Vr:V - 1.85).
These values are consistent with a geological model in which the up-
per crust is granitic, becoming increasingly marie with depth, leading
to a lower crust of marie amphibolite. Such a model is consistent with
shear-wave data at wide-angle, which may be explained by a layered
sequence in which the average S-wave velocity gradient is slightly neg-
ative with depth, but the P-wave velocity gradient is strongly positive.
The combined normal incidence and wide-angle shear-wave data sets
are not consistent with a granulitic lower crust, with a lower crust of
intermediate or acidic composition, nor do they require that S-wave
layering is produced by aligned amphiboles.
Introduction
In recent years there has been much discussion regarding the nature
and origin of lower crustal layering. Deep seismic profiles in several
parts of the world have shown bright layered reflections from the
lower crust beneath a relatively transparent upper crust (for example
Matthews [1986] and Meissner [1986]), mainly in areas of crustal ex-
tension. Several models have been suggested to explain this pattern
of reflectivity [arner, 1990], however as yet no single model can
definitively explain all the geological, geochemical and geophysical
observations.
Conventional P-xvave deep seismic data provide a clear image of
crustal structure, especially when combined with other methods such
Geodynamics 22
as seismic refraction and gravity. However, P-wave reflection data
do not give an accurate indication of the lithologies which comprise
the deep crust. The composition of a rock is better characterised by
its Poisson's ratio, derived from the ratio of P and S-wave velocities,
than by P-wave velocity alone [Kern, 1982].
Until recently [Liischen et al., 1989; Ward and Warner, 1989; Lfi-
schen et al., 1990; Johnson and Hartman, 1990] all controlled S-wave
measurements of the deep crust have been carried out by refraction.
Measurements at near-norlnal incidence are potentially lnore accurate
than those at wide-angle, since, for a one-dimensional earth, P and S
ray paths are identical. In contrast, at wide-angle, if Poisson's ratio
varies with depth, then P and S ray paths do not coincide. In princi-
ple, on near-normal incidence data, 1/: Call be obtained directly by
taking the ratio of travel times in P and S-waves between two reflec-
tors. This requires that distinct reflectors Call be uniquely identified
on both P and S reflection records. In contrast, wide-angle P and
S-wave data sets must be modelled separately, and then combined to
give values for Vr:V. Such a procedure is intrinsically less accurate.
Studies of wide-angle seismic data have led IIolbrook et al. [1988]
to suggest that S-waves would not be recorded fi'om the lower crust
at near-normal incidence. Therefore in a high risk experiment, the
BIRPS group decided to attempt to acquire coincident P and S-wave
reflections from an area of brightly layered lower crust. This paper
summarises the acquisition, interpretation and implications of that
experiment.
WISPA (VVeardale Integrated S and P-wave Analysis)
The investigation xvas to be BIRPS' first deep reflection project on
land. To better guarantee fulfillment of the aims of the project it was
decided to work in an area already known to exhibit bright P-wave
layering, preferably with a simple crustal seislnic character to ease
interpretation. During 1986 the British Geological Survey (BGS),
in the course of investigating the geothermal potential of the UK,
shot a number of P-wave vibrator lines to 12 s two-way travel time in
northeast England (J. Kilnbell, BGS, personal communication). On
several of these profiles, bright lower crustal layering was observed
between 8 and 10s. The reason these reflections are so clear is thought
in part to be due to the profiles crossing of the Weardale Granite.
From studies of other BIRPS profiles, most notably SWAT, it appears
that granites such as Weardale provide a high-Q window to the deeper
crust [Scheirer, 1988].
This area therefore became the site of WISPA (Figure 1), which was
acquired during 1988. BIRPS already had a profile just offshore from
343
344 StlEAR WAVE REFLECTION DATA
Newcastle
NORTH SEA
WISPA
10 km
Main Urban Areas
Approx. Extent of diesbrough
Weardale Granite
eig. 1. Location of Weardale Integrated S and P-wave Analysis. Dark
ines indicate BIRPS reflection profiles. Area shaded by crosses indi-
ates the extent of the Weardale granite constrained by gravity data.
this part of England, the NEC line [Klemperer and Matthews, 1987]
which, like the BGS lines, shows a brightly layered lower crust beneath
a transparent upper crust covered by a thin veneer of sediments.
The project comprised three short intersecting profiles located over
the strongest lower crustal layering observed by the BGS. The three
lines, oriented at approximately 120 o to each other, were designed to
allow study of, and compensation for, the effects of anisotropy. The
sources for the survey were based on the Camoufiet technique [Edel-
mann, 1985] with three dynamite shots, 2 m apart, at each shotpoint,
intended to provide polarised S-waves. Shot size was 20 kg at a depth
of 10m below bedrock surface. Shotpoints were located at a mean
spacing of 300m and fired into a 6 k,n three-component spread with
a nominal station spacing of 50m. The 10 ttz three-component geo-
phones were all carefully buried to reduce ambient noise levels and
improve coupling. Recording, performed by Compagnie G6ndral de
G6ophysique, was on 120 channels per component to 32 seconds. A
maximum CDP fold of cover of 20 was achieved by shooting through
a split spread.
Resulls
Initial results were excellent on all three lines, as illustrated by the
shot gathers shown in Figure 2. Strong S-wave reflections are observed
0 Distance (km) 6
0
? '. "-"-.'.' ..-"-:- - :.".4. --
... . .,_,- ,,. .':'' ...
. :: ..., ..' .. :.::...-':-.:.,:: ..: .
. . .. ._ .,.:.. .,' .. ?...." ... .........
. . .,......- .r .- 'a,.', ,,,'. . . .:,, ..-..
...-, ..... ,.,..:,. '.?.:?..., .. : .
l0
(a) Vertical component Shot Gather
Fig. 2. Shot gathers from WISPA corrected for spherical divergence
and statics with no further processing. (a) shows vertical component,
(b) the transverse horizontal component. The time scale of (b) has
S-'wave
refraction
Top of
Lower Crust
MOHO
Distance (km) 6
.__...:...i...:.._..:-_- .. 0
:_.....- -. -..
-:'...._.-..,,,._-'...: :;-[(;.....-.
::.-.;::-: .....
..::::.. "...
.....-

(b) Horizontal componen Sho Gaher
been compressed by V (Poisson's ratio of 0.25). Note the bright
S-wave reflections from the lower crust, and clear S-wave refraction.
between 13 and 17 s and it appears that these correspond to the bright
P-wave reflectivity between 8 and 10s. The seismic character of each
appears similar, with several brighter zones within the lower crust,
and the Moho marked by a rapid decrease in reflectivity. Analysis
of S-wave polarisations indicates that the shots did not behave as
anticipated (as for Liischen et al., [1990]) and most S-wave energy
was apparently produced by mode conversion at the bedrock surface
[Fertig, 1984].
After CDP stacking of the data, the P-wave profiles exhibit strong
layered reflections in the lower crust which dominate the sections (see
Figure 3(a)). However the large S-wave statics posed a severe prob-
lem to processing of the data. Considerable effort was put into field
statics and subsequent surface consistent derivation. The variation
in final statics across the spread was found to exceed twice the domi-
nant period of lower crustal layering in S-waves. This is not surprising
considering that the patchy glacial cover in the area showed a Vr:V
as high as 10. Figure 3(b) shows the S-wave stacked section which
corresponds to the P-wave stack in Figure 3(a). This S-wave stack
proves the ability to collect coincident P and S-wave reflections from
the layered lower crust.
(a) P-wave Stack
Fig. 3. True amplitude stacks of (a) P-wave and (b) S-wave data for
WISPA Line 1. Again S-wave time scale compressed by v/'. The char-
acter of the lower crust is similar in P and S-waves, arrows indicate
WARD AND WARNER 34:5
Poisson's ratios
Within WISPA, both shot gathers and stacks are of sufficient qual-
ity to allow estimation of Poisson's ratio variation within the crust.
The seismic character of the lower crust is similar in both P and S-
waves allowing correlation in time of key features. We can identify
accurately both the top and the base of the lower crustal layering,
allowing determination of average Poisson's ratio for both the up-
per and lower crust. In addition, we can locally correlate individual
bright events on both sections, which gives an indication of finer scale
variation.
Using the clearest of these markers, Figure 4 shows the variation in
Poisson's ratio with depth observed from these data. In common with
many previous estimates, the average for the crust is 0.23. This value
however conceals a clear and significant increase in Poisson's ratio
with depth through the crust; in the upper crystalline crust values
are low, around 0.20 (Vr:V -- 1.64), while in the layered lower crust
values rise to an average of 0.29 (Vr:V = 1.85). If our correlation
of events within the lower crust is correct, then Poisson's ratio may
exceed 0.3 near to the base of the crust.
0 Distance (km) 5
_
....
.:
_
... . .
correlations made to determine Poisson's Ratio (tentative correlation
marked by ?).
346 SHEAR WAVE REFLECTION DATA
P-wave
TWTT
(.)
10
Crustal
Average
0.23
Approx
Depth
(km)
3O
O. 17 Poisson's Ratio 0.33
Fig. 4. Poisson's Ratio as derived from stacked sections. The crustal
average is close to that of other studies. Note the significant jump
in average Poisson's Ratio between the upper crust and layered lower
crust.
These values take no account of the possible effects of anisotropy
on the travel times of the correlated reflections. In the lower crust,
S-wave anisotropy has been proposed due to preferred orientation of
minerals, most notably hornblende [Liischen et al., 1990]. The re-
sults of Siegesround et al. [1989] show that, if amphiboles are aligned
horizontally in the lower crust, then no significant splitting of shear
waves will occur at near-normal incidence. However, in the brittle
upper crust a degree of S-wave splitting is expected, caused by mi-
crocracks [Booth et al., 1985; Crampin et al., 1986; Crampin, 1987]
and/or preferred mineral orientation [Brocher and Christensen, 1990].
Such splitting would cause the marker at the top of the lower crustal
layering to yield a minimum value for Poisson's ratio of the upper
crust if the fast S-wave has been picked. Preliminary modelling in-
dicates that the value of 0.29 for the lower crust may be reduced a
little by the same effect, but the significant difference in Poisson's
ratio between the upper and lower crust will remain.
S-waves at Wide-angle
Several studies of the crust have been carried-out using both P
and S-waves at wide-angle, most notably in southwest Germany, to-
gether with conventional deep reflection profiling [Fuchs et al., 1987;
Liischen et al., 1987; Liischen et al., 1990]. The pattern of P-wave re-
flectivity in the Black Forest is similar, although not identical, to that
seen below Weardale, with a strongly layered lower crust between a
transparent upper crust and upper mantle. The crustal thickness is
comparable, and the results of Holbrook et al., [1988, Figure 10] indi-
cate that the variation in Poisson's ratio exhibits approximately the
same increase with depth as seen beneath Weardale. Lfischen et al.
[1990] have recorded near-normal incidence S-wave layered reflections
from the lower crust. Lfischen (personal communication) considers
that these Black Forest S-wave reflections are qualitatively different
to those seen on our Weardale data; we consider the data to be closely
analogous, and explain the superficial differences as due to the longer
offsets and different source parameters used in the Black Forest.
Sandmeier and Wenzel [1986] pointed out that at wide-angle, the
S-wave sections from the Black Forest do not resemble the P-wave sec-
tions (Figure 5). The bright P-wave reflections seen from the lower
crust between 40 and 60 km are missing, or at least considerably re-
duced in amplitude, on the S-wave sections. At shorter offsets the
signal to noise ratio in these wide angle data is probably not suffi-
cient to detect lower crustal S-wave reflections.
The velocity models produced by Holbrook et al. [1988, Figure 10]
would place the critical distance for P-wave reflections from the top
of the lower crust at less than 50 km, so that these reflections would
be largely post-critical. The velocity model produced by Wenzel et
al. [1987] place this critical distance at about 60km, making these
reflections pre-critical. The difference between the two interpretations
depends upon the velocity structure in the upper crustal low velocity
zone, and upon the velocities in the thin layers at the top of the lower
crust. Both these parameters are poorly resolved by the seismic data.
The variation of amplitude with offset seems, however, to suggest that
the reflections beyond about 40 km are post critical.
14
Horizontal Component
0 Distance (km) 100
Vertical Component
0
0 Distance (kin) 100
Fig. 5. Wide angle seismic data from the Black Forest [after Sand-
meier and renzel, 1990]. Lower panel shows vertical component;
upper panel shows radial component. Note the difference in reduc-
tion velocities and time scale. The strong P-wave reflections between
40 and 60 km are absent on the S-wave data.
The lack of wide-angle S-wave reflections has been explained in a
variety of ways. Holbrook et al. [1987] suggested shear reflections
were absent because of alternating extregne high and low Poisson's
ratio values in the lower crust, produced by the interleaving of quartz-
rich and mafic layers, giving the model of Figure 6(c). This model
was refined by Holbrook et al. [1988], then by Sandmeier and Wenzel
[1990]. The Weardale data and the results of Liischen et al. [1990]
show that bright lower crustal layering can be observed at near-normal
incidence using S-waves; impedance contrasts must therefore be large
in the lower crust for both P and S-waves.
A second mechanism proposed to explain the Black Forest data
[Liischen et al., 1990] uses velocity anisotropy to supplement the com-
positional layering in the lower crust of Sandmeier and Wenzel [1990].
They suggest that the velocity anisotropy of amphiboles, aligned hor-
izontally in discrete shear zones, would be sufficient to provide the
necessary impedance contrasts in the lower crust to give strong P-
wave reflections at wide-angle and near-vertical incidence, yet only
give S-waves at near-normal incidence. However the mechanism sug-
gested by these authors suggest that this effect only occurs at certain
azimuths; at others, wide-angle S-wave reflections would be visible.
The refraction profiles of the Black Forest (S, Z and U) appear to
show a similar lack of wide-angle S-wave reflections at a range of
azimuths [Holbrook et al., 1988].
We propose a simple, unified explanation of both the Black For-
est and WISPA data sets. Our model considers the effects of pres-
sure, temperature and composition on V r and V of typical crustal
rocks. Laboratory measurements have been carried out on whole
rocks [Christensen, 1979, 1982; Kern and Richter, 1981; Kern, 1982]
and individual minerals [Sumino and Anderson, 1984; Siegesmund et
al, 1989]. The main minerals thought to form the Earth's continen-
tal crystalline crust are quartz, feldspars, amphiboles and pyroxenes
in varying proportions. Because laboratory measurements on whole
rock samples provide only isolated points on a continuum, we have
calculated V r and V for isotropic aggregates of the main minerals.
We consider a simple model for the crust, in which there is a gradual
increase in mafic minerals with depth, from granite at the surface,
through mafic amphibolite, to mafic granulite above the Moho. V r
and V for such a model can be calculated after compensating for
the effects of lithostatic pressure and temperature. We have assumed
a crust of 30km thickness and a geothermal gradient of 25C/km.
Figure 7 shows the resulting velocity-depth profiles.
WARD AND WARNER 347
The values for Poisson's ratio obtained from WISPA clearly indicate
that the lower crust here can not contain a significant fraction of
pyroxene; that is, it can not be dominantly granulitic. The data also
show that the average composition of the lower crust must be mafic;
it can not be acidic or intermediate. If VCeardale and the Black Forest
can be considered analogous, then we can apply our Vp:V to the P-
wave model of Holbrook et al. [1988, Figure 10]. This will generate a
negative S-wave velocity gradient in the lower crust as in Figure 6(d).
Figures 8 and 9 show reflectivity modelling for a P-wave model that
is similar to that used by Holbrook et al. [1988, Figure 16]. Figure
8 shows the synthetic seismograms produced when Poisson's ratio is
constant; Figure 9 shows the results produced when Poisson's ratio
increases with depth such that the S-wave velocity gradient in the
lower crust is negative. Bright P-wave reflections, from the lower
crust, are seen between 40-60km offset in both models. The cor-
responding S-wave refelections are suppressed at these offsets in the
second model. At near normal incidence, the reflected S-wave ampli-
tudes are not strongly dependent upon the average velocity gradient,
and both models generate similar reflection amplitudes. These simpli-
fied models make no attempt to reproduce the spectral characteristics
of the data, nor to accurately model layer thickness and impedance
contrast in the lower crust.
Conclusions
We see bright S-wave reflections, at near-normal incidence, fi'om the
lower continental crust. These are similar in appearance to the bright
lower crustal P-wave reflections which have been widely reported. We
find an average Poisson's ratio of 0.23 for the whole crust; this is
similar to values obtained from other studies. Within the crystalline
crust there appears to be a systematic increase in Poisson's ratio with
depth, from values around 0.20 in the upper crust, to values of at least
0.29 for the lower crust. This implies significant vertical lithological
contrast, and is consistent with a simple nodel in which the upper
crust is dominantly granitic underlain by a marie, amphibolitic lower
crust [Kern, 1982]. Our data are not consistent with granulites in the
lower crust, or with acidic or intermediate compositions in the lower
We have shown that the increase of Poisson's ratio with depth is
sufficient to explain the difference between P and S-wave sections at
wide angle, without the necessity of introducing layering in Poisson's
ratio, or introducing layered anisotropy.
(a) P-wave model (b) Constant
Poisson's Ratio
4.0 km/s 9.0 2.0 km/s
S-wave Models
(c) Alternating (d)Smooth Varying
Poisson's Ratio Poisson's Ratio
0
6.0 2.0 km/s 6.0 2.0 km/s 6.0
25
Fig. 6. Velocity models presented to explain the reflection charac-
ter below the Black Forest. (a) the P-wave model of Sandmeier and
Wenzel (1986). (b) shows for reference the S-wave velocity model as-
suming a constant Poisson's Ratio for the crust of x/. (c) the model
of Holbrook et al (1988) with strongly alternating Poisson's Ratio.
(d) a model with slightly negative average S-wave velocity gradient
(smoothly increasing average Poisson's Ratio) in the lower crust.
SHEAR WAVE REFLECTION DATA
0 Composition %
'granite' Na
Plagioclas

'amphibolite ................
'granulite' Ca Pyroxene
100 1.7[ Vp/Vs 1.90
Fig. 7. The effects of pressure, temperature and composition on Vr:.
The compositional model ranges from a 'granite' near-surface to a
mafic 'amphibolite' at 25km. Pyroxenes are added at deep crustal
levels to illustrate the effect for a 'granulite'. For such compositions,
6.0 Vp (km/s) 7.0
\vp
3.5 Vs (km/s) 4.0
25
MOHO
V r and V were calculated for an isotropic mineral aggregate and cor-
rected for lithostatic pressure and temperature (at 25C/km). For
a typical P-wave velocity gradient in the crust, the average S-wave
gradient would be negative.
14
0 Distance (km) 100
Vertical Component
0
0 Distance (km) 100
Fig. 8. Reflectivity synthetic seismograms for a model with constant
Poisson's ratio (insert).
14
0 Distance (km) 100
Vertical Component
0 Distance (kin) 100
Fig. 9. As Figure 8 for a model in which the S-wave velocity gradient
is negative in the lower crust (insert). Note the similar appearance of
the P-waves to those shown in Figure 8, and the lack of strong S-wave
reflections from the lower crust.
WARD AND WARNER 349
Acknowledgements We would like to thank all those fi'iends and as-
sociates of BIRPS who assisted during the acquisition and processing
of WISPA, and the BIRPS Advisory Conmfittee for agreeing to back a
high risk experiment. BIRPS is funded by the Natural Environment
Research Council. Thanks are extended to the Jeff Kimbell of the
British Geological Survey for making their seismic data in Weardale
available to us. Drilling and shooting was performed by Ritchies
Ltd. Acquisition and initial processing of WISPA was by Compagnie
Gn4ral de Gophysique. G.W. was sponsored by Shell International
Petroleum Co. Ltd. for the duration of the project.
References
Booth, D.C., Crampin, S., Evans, R., and Roberts, G., Shear wave
polarisations near the North Anatolian Fault, I: evidence for aniso-
tropy-induced shear-wave splitting, Geophys. J. R. astr. Soc., 83,
61-73, 1985.
Brocher, T. M., and Christensen, N. I., Seismic anisotropy due to
preferred mineral orientation observed in shallow crustal rocks in
southern Alaska, Geology, 18, 737-740, 1990.
Christensen, N. I., Compressional wave velocities in rocks at high
temperatures and pressures, Critical thermal gradients and crustal
low-velocity zones, J. Geophys. Res., 84, 6849-6857, 1979.
Christensen, N. I., Seismic Velocities, in Handbook of Physical Prop-
erties of Rocks, Volume II, edited by R. S. Carmichael, CRC Press,
Boca Raton, Florida, pp. 1-228, 1982.
Crampin, S., The geological and industrial implications of extensive-
dilatancy anisotropy, Nature, 328, 491-496, 1987.
Crampin, S., Booth, D.C., Krasnova, M. A., Chesnokov, E. M.,
Maximov, A. B., and Tarasoy, M. T., Shear wave polarisations in
the Peter the First Range indicating crack-induced anisotropy in a
thrust fault regime, Geophys. J. R. astr. Soc., 8, 401-412, 1986.
Edelmann, H. A. K., Shear-wave energy sources, in Seismic Shear
Waves, Part B. Applications, edited by G. Dohr, Geophysical Press,
London, pp. 134-177, 1985.
Fertig, J., Shear-waves by an explosive point-source: the earth surface
as a generator of converted P-S waves. Geophys. Prospecting, 32,
1-17, 1984.
Fuchs, K., Bonjer, K.-P., Gajewski, D., Liischen, E., Prodehl, C.,
Sandmeier, K.-J., Wenzel, F., and Wilhelm, tI., Crustal evolution
of the Rheingraben area: I. Exploring the lower crust in the Rhe-
ingraben Rift by unified geophysical experiments, Tectonophysics,
11, 261-275, 1987.
Holbrook, W. S., Gajewski, D., and Prodehl, C., 1987. Shear-wave
velocity and Poisson's ratio structure of the upper lithosphere in
southwest Germany, Geophys. Res. Lett., 1, 231-234, 1987.
Holbrook, W. S., Gajewski, D., Krammer, A., and Prodehl, C., An
interpretation of wide-angle compression and shear wave data in
southwest Germany: Poisson's Ratio and petrological implications,
J. Geophys. Res., 93, 12081-12106, 1988.
Johnson, R. A., and Hartman, K. A., Constraints on P and S-wave
velocity structure in the Transition Zone and Colorado Plateau
from multi-component wide-angle seismic data, paper presented at
the Fourth International Symposium on Deep Seismic Reflection
Profiling of the Continental Lithosphere, Bayreuth, 4-7 September,
1990.
Kern, H., and Richter, A., Temperature derivatives of compressional
and shear wave velocities in crustal and mantle rocks at 6 Kbar
confining pressure, J. Geophys., 9, 47-56, 1981.
Kern, H., Elastic-wave velocity in crustal and mantle rocks at high
pressure and temperature: the role of the high-low quartz transition
and of dehydration reactions, Phys. Earth Planet. Inter., 29, 12-23,
1982.
Klemperer, S. L., and Matthews, D. It., Iapetus suture located be-
neath the North Sea by BIRPS deep seismic reflection profiling,
Geology, 15, 195-198, 1987.
Ltischen, E., Wenzel, F., Sandmeier, K.-J., Menges, D., Ruehl, Th.,
Stiller, M., Janoth, W., Keller, F., Soellner, W., Thomas, R.,
Krohe, A., Stenger, R., Fuchs, K., Wilhehn, H., and Eisbacher,
G., Near-vertical and wide-angle seismic surveys in the Black For-
est, SW Germany, J. Geophys., 62, 1-30, 1987.
Lfischen, E., Sandmeier, K.-J., Wenzel, F., Nolte, B., and Fuchs,
K., Shear-wave information on the nature of the deeper crust in
southwest Germany, Annales Geophysicae, Special Issue, 302, 1989.
Lschen, E., Nolte, B., and Fuchs, K., Shear-wave evidence for an
anisotropic lower crust beneath the Black Forest, southwest Ger-
many, Tectonophysics, 173, 483-493, 1990.
Matthews, D. H., Seismic reflections from the lower crust around
Britain, in Nature of the Lower Continental Crust, edited by J. B.
Dawson, D. A. Carswell, J. Hall and K. H. Wedepohl, Geol. Soc.
London, Spec. Publ., , pp. 11-22, 1986.
Meissner, R., Twenty years of deep seismic reflection profiling in Ger-
many - a contribution to our knowledge of the lower Variscan crust,
in Nature of the Lower Coninental Crust, edited by J. B. Dawson,
D. A. Carswell, J. Hall and K. H. Wedepohl, Geol. Soc. London,
Spec. Publ., , pp. 1-10, 1986.
Sandmeier, K.-J., and Wenzel, F., Synthetic seismograms for a com-
plex crustal model, Geophys. Res. Left. 13, 22-25, 1986.
Sandmeier, K.-J., and Wenzel, F., Lower crustal petrology from wide-
angle P and S-wave measurements in the Black Forest, Tectono-
physics, 173, 495-505, 1990.
Scheirer, D., Seismic attenuation in the continental crust SW of Eng-
land, EOS, Trans. Am. Geophys. Union, 69, 1318, 1988.
Siegesmund, S., Takeshita, T., and Kern, H., Anisotropy of Vp and
in an amphibolite of the deeper crust and its relationship to the
mineralogical, microstructural and textural characteristics of the
rock, Tectonophysics, 157, 25-38, 1989.
Sumino, Y., and Anderson, O. L., Elastic Constants of Minerals, in
Handbook of Physical Properties of Rocks, Volume III, edited by R.
S. Carmichael, CRC Press, Boca Raton, Florida, pp. 39-138, 1984.
Ward, G. S., and Warner, M. R., S-wave images of the layered lower
continental crust, Annales Geophysicae, Special Issue, 52, 1989.
Warner, M. R., Basalts, water, or shear zones in the lower continental
crust? Tectonophysics, 173, 163-174, 1990.
Wenzel, F., Sandmeier, K.-J., and ]Vilde, W., Properties of the lower
crust h'om modelling reh'action and reflection data, J. Geophys.
Res., 92, 11575-11583, 1987.
LAYERS THICKNESSES IN THE LOWER CRUST:
MODELLING AND SPECTRAL ANALYSIS OF BIRPS DATA
Anne Paul * and Richard Hobbs **
Abstract. Most of the in situ velocity models
proposed for the reflective lower crust of W-Europe
include thin layers with thicknesses concentrated in a
very narrow range: 50-200m. Within this hypothesis,
constructive interference should induce a strong
focussing of reflected energy in narrow frequency bands
at normal incidence. Based on this argument, a simple
method is proposed to evaluate thicknesses of lower
crust lamellae from detailed spectral analysis of the
reflections in the time-frequency domain. Its resolution
is tested on synthetic shotpoint gathers computed from
complex crustal models. The applicability to real data
is tested on more than 200 shotpoints from the WAM
profile and WISPA experiment of the BIRPS dataset.
In WAM, the spectra of a group of reflections show
focussing of reflected energy around 17 Hz, suggesting
that most of the associated reflecting boundaries have
a nearly constant spacing of 90 to 110 m. However the
frequency spectra show clear variations with both
traveltime (i.e. depth) in the lower crust and distance
along the profile. This suggests that the stucture of the
lower crust is more disorganized than the velocity
models proposed in previous studies. In WISPA, some
focussing of energy could also be detected for a few
reflections. But since the frequency bandpass is very
different from WAM, the frequency of resonance
appears to be shifted towards higher frequencies (28-
32Hz). This difference has two possible origins: 1) the
dominant thickness is smaller in WISPA (i.e. 50-70m)
and interference is induced at 30Hz, 2) it is greater
(150-170m) and focussing is observed at the first
harmonic of the frequency of resonance (9-10Hz). This
* Laboratoire de G4ophysique Interne et
Te c t onop hysi que,
BP 53X, 38041 Grenoble Cedex, France.
** BIRPS, Bullard Laboratories, Madingley Road,
Cambridge CB3 0EZ, GB.
Geodynamics 22
351
ambiguity shows that the absence of low fi'equencies in
deep seismic profiling is a critical limitation-to such
analyses.
Introduction
The origin of the reflectivity of the lower
continental crust imaged by many deep seismic profiles
in Western Europe remains very controversial (see Mac
Carthy and Thompson (1989) for a review of the main
hypotheses). A significant improvement of the
geological models proposed would be possible with a
better knowledge of the crustal velocity structure. The
question of the estimation of in situ physical properties
of buried rocks from seismic data has been a subject of
interest for a long time in oilfield exploration
(Velzeboer, 1981). But in crustal studies, such
measurements have mainly been conducted on wide-
angle data. Sandmeier and Wenzel (1986) have derived
a complex velocity model from amplitude modelling of
a wide-angle shot recorded in the Black Forest. Paul
and Nicollin (1989) have used spectral analysis of the
Moho reflection to give some trends on layers
thicknesses and velocities in the lower crust in
northern France. Both studies lead to the same type of
velocity structure characterized by a thin layering at
hectometric scale and strong velocity contrasts in the
lower crust. Such detailed quantitative analyses have
not been conducted on normal incidence data although
they are much more abundant and often of better
quality than wide-angle recordings. Previous work on
the estimation of lower crustal physical parameters
include the measurement of Q around Britain by Hobbs
(1990), the evaluation of reflection coefficients for lower
crustal reflectors by Warner (1990) and AVO studies
by Peddy (1991). But no confirmation of the results
from wide-angle data analyses, at least for layers
thicknesses and velocities, has come from normal
incidence data. This paper attempts to reduce this gap.
Methodology
This study is centred around the energy distribution
of the reflections from the Dower crust in the time
352 LAYERS THICKNESSES IN THE LOWER CRUST
vnocrn,
d
VILOCITY
t
m 30 t
b
Fig. 1. Perspective views of the spectral energy of the fiat between 10 and 35 Hz. Model b was inspired by
normal-incidence reflections computed for 2 different Sandmeier and Wenzel's velocity structure (layers
crustal structures. The synthetic source spectrum is thicknesses of 120 4- 30 m).
(two-way traveltime TWTT) and frequency domains. computed for different lower crustal velocity structures.
This time-frequency analysis is conducted by Fourier Our aim is to check the influence of the models
transform in a moving window 400 ms wide. In the parameters on the reflected energy. Synthetic
synthetic examples shown in Fig. 1, the reflectors on seismograms are computed for a fiat-layered crust with
top and base of a homogeneous lower crust reflect all the Discrete Wavenumber Method developed by
the frequencies of the incident wave with the same Bouchon (1981). Responses are computed in the
strength, whereas the spectral response of a layered frequency domain, and time-domain seismograms are
lower crust is more complicated. The explanation for synthesized after convolution by a fiat source signal.
this difference is straightforward. In model b, reflected Fourier spectra are computed in moving windows
energy is focussed (or attenuated) by constructive (or between 5 and 10 s TWTT. In each time-window, the
destructive) interferences between the reflections on total amount of reflected energy is evaluated, giving
the top and base of the thin layers At normal
incidence, constructive interferences induce a strong 5o
focussing of reflected energy at the frequency of
resonance of a layer: f0 --v/4h and its harmonics:
fn----(2n+l)f0, where h and v are the thickness and the m 4o
interval velocity respectively. This simple relationship
could lead to an estimation of the ratio v/4h by
o 3o
picking the frequency of maximum reflected energy in
the time-frequency maps The frequency of resonance is
r
plotted as a function of the layer thickness for a range o 2o
of possible lower crustal interval velocities in Fig. 2. It
reveals two important limitations of the method of
10
energy maxima picks. First, the influence of the
velocity is much too weak to enable any estimation of
velocity by measurement of frequency. The second
point is that the bandpass of seismic data is limited to
8 - 40 Hz, allowing resonances to be detected only for
layers with thicknesses in the range 40 to 250 m.
Tests on synthetic data
0 100 200 300 400 500
LAYER THICKNESS (m)
Fig. 2. Frequency of resonance of a thin layer at
normal incidence as a function of its thickness. The 2
curves correspond to the extremal bounds of the range
This section deals with the study of the distribution of possible layers velocities for lower crustal rocks (6.0
of reflected energy in synthetic shotpoint gathers and 7.5 km/s).
PAUL AND HOBBS 353
the curve of energy variation with TWTT plotted
along the top of the maps. In order to compensate for
the effects of the variations of total energy with time,
the spectra computed in each window are normalized
by the corresponding total energy. This gives time-
frequency maps of energy distribution, expressed in
percentage of the total energy. The average
distribution of spectral energy is plotted on the right-
hand side of the maps.
Fig. 3 shows the maps obtained for 6 models with
different distributions of layers thicknesses and
velocities. In model 1, thicknesses are random in the
range 90- 150 m in the whole lower crust and
velocities alternate from low to high values. As
V.LOCIT (kin/,)
0 6 7 8
1o 1
o
VELOCITY (km/s)
6 7 8
o
- 20
3o
VELOCITY (krn/s)
o, 7 a
1o 3
30
6 8 10
max=9.6% TIME (s)
_,40
'a0
20
[10
0
VELOCITY
o. _ _ _? ....
lo 4
3o
VELOCITY
o 7
.o
30
VELOCITY (krn/s)
1o 6
1
30
8 10
TIME (s)
0_ I , , , I , , ,
max= 7.8% m!u ")
Fig. 3. Maps of energy distribution obtained for different lower crustal velocity structures. Models
i to 4 have organized velocity structures (slatilar thicknesses and systematic alternations of
velocities from low to high values). Models 2, 3 and 4 have layers with thicknesses departing from
the average value (120 4- 30 m) between 20 and 23 km depth. Models 5 and 6 have disorganized
velocity structures. In model 5, thicknesses are spread out between 50 and 500 m. The systematic
alternation between low and high values is suppressed from model 6. The black colour
corresponds to the maximal concentration of energy indicated below each map.
lOO
1 2 3 4 5 6
o
Fig. 4. Relative levels of total energy reflected between
5 and 10 s TWTT for models I to 6.
expected, reflected energy is concentrated in a narrow
frequency band around the frequency of resonance of
the layers (12 Hz) and its first harmonic (36 Hz).
Focussing at 12 Hz is less clear for the later reflections
after 7.5s. A possible explanation is that most of the 12
Hz energy incident on top of the lower crust is
reflected by the first layers and is not transmitted to
deeper levels. In model 2, layers thicknesses are
multiplied by 2 in the middle part of the lower crust.
Between 6.5 and 7.2 s, energy is concentrated at a
lower frequency (7 Hz) and its harmonic (21 Hz). A
symmetric result is obtained with models 3 and 4
where layers with average thickness of 50 m (resonance
at 29 Hz) and 25 m (resonance out of the bandpass) are
inserted. The conclusion is that, even in complex
models, layers with similar thicknesses only reflect
energy within a narrow frequency band, but this can be
observed only when numerous layers have very similar
thicknesses. The introduction of some disorganization
results in very different energy distributions (Fig. 3).
Model 5 has the same velocity distribution as model 1,
but layers thicknesses are spread out in a large range
betwen 50 and 500 m. Reflected energy is spread out in
the whole frequency bandwidth and the maximal
concentration (5.1%) is weaker than in organized
models. The effect of the suppression of systematic
alternations from low to high velocities from model i is
examined with model 6. Since thicknesses belong to a
narrow distribution (90- 150 m), energy is less
scattered than for model 5.
The spectral energy of the reflections from the lower
crust is characterized by its distribution within the
frequency spectrum and its total amount. From the
results obtained with synthetic seismograms, we can
0 20 40 60
FREQUENCY(Hz)
Fig. 5. Examples of average trace spectra computed for WISPA (c) profiles, in the time-window of the reflective
3 shotpoint gathers of the NEC (a), WAM (b) and lower crust.
PAUL AND HOBBS 355
0 I .... I .... ! , , _
7 8 9 10
max=12.4% TIIE (s)
I I I I I I I I I I I I I I I I I I I I I I I I I I I [.
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
M)00 '000 BOO0 000 10000
TIME ()
contribution of constructive interferences is less
important as far as the total energy is concerned. But,
if energy focussing is detected in the spectrum of the
lower crustal reflections, this can be considered as a
good indication of the presence of a great number of
layers with similar thicknesses, and this should allow
mapping of layers thicknesses in the lower crust.
Application to real data
Preliminary precautions. Similar to other spectral
methods, the application to real data requires a good
signal-to-noise ratio and a wide frequency bandpass.
Fig. 5 shows 3 examples of spectra computed on the
whole time-window of the lower crust for 3 shotpoint
gathers of the NEC (Freeman et al.. 1988), WAM
(Peddy et al., 1989) and WISPA (Ward and Warner,
1989) profiles. The NEC data set cannot be used for
this spectral analysis because of the disastrous spectral
effect of the sea-bottom reverberations in shallow water
(60 m)(Fig. 5a). Much better conditions are reached on
the WAM (water depth of 110 m) and WISPA
(recorded on land) profiles. Further, it is necessary to
check that the spectra of the lower crustal reflections
are representative of the spectral signature of the lower
crustal reflectors, and not of the signature of some
parasitic filters. For WAM data, the signature of the
air-gun array and the response of the hydrophones
were well-known and we could check that they were
fiat in the range 8 - 45 Hz. A second filter is due to
water bottom reverberations which produce the
Fig. 6. Map of the average distribution of energy discontinuous aspect of the spectra, but it has no effect
between 6.2 and 10 s TWTT for the 3000 traces from on the relative amplitudes of the peaks we are
zone C of the WAM profile. The 60 traces of one of the interested in. A third filter is the response of the
shotpoint gathers used are plotted at the same time sediments and upper crust. To get rid of the rapid
scale for comparison. lateral variations in the response of the shallower levels
of the crust, we computed average spectra on a very
large number of adjacent traces covering a length of
conclude that the distribution is mainly controlled by 2.5 km. The last parasitic effect is the strong
the distribution of layers thicknesses and that the absorption of high frequencies in the sediments
velocity , or at least the strength of the velocity affecting WAM data (Fig. 5b). We checked that the
contrasts, has a weaker influence. Fig. 4 shows that the low Q factor cannot explain by itself the large
total amount of reflected energy is mainly controlled by difference in amplitude between the 3 low frequency
the strength of the velocity contrasts between the peaks and the others.
layers. In models 1 to 4 where strong velocity contrasts
are associated with strong constructive interference Application to WAM. 4 zones of 50 shotpoints each
between layers with similar thicknesses, the total have been selected in the eastern part of the profile,
energy is strong. But it remains strong for model 5 where the lower crust is very reflective. A single
though constructive interference are spoiled by a wide average spectrum was computed from all the 3000
distribution of thicknesses. It appears that energy traces of each zone. Although the reflectivity of the
decreases as soon as systematic alternations from low lower crust appears very clearly in the 4 zones on the
to high velocities are suppressed (model 6). The first stacked section, the individual traces have very
consequence is that the strong reflectivity of the lower different signal-to-noise ratios. Fig. 6 shows the average
crust observed in many places is mainly due to the distri'bution of energy in the time-frequency domain
presence of strong velocity contrasts, and that the obtained for the zone which has the best signal-to-noise
[.i.] ............... :'..=.::.!:Ji!i.'i,:::i:.:,11.:!:!!:-.,i'-i.,:..,,:: ................ .:----, .........
o
7 8 9 10
max=10.6 TIE
Illllllll IIIllllll Illllllll IIIIIIIII
6;000 '000 BOO0 9000 10000
TIME (m)
Fig. 7 Map of the distribution of energy computed
from the 3000 traces from zone D of the WAM profile.
ratio. The energy map displays a great vertical
heterogeneity in the spectral response of the lower
crust. This result conflicts with the velocity structures
proposed in previous studies (e.g. Sandmeier and
Wenzel (1986))where the thickness of the lamellae was
constant. A group of reflections located between 7.5
and 8.0 s exhibits a very narrow frequencial response
limited to the medium ray (17 Hz). This focussing is
probably an indication that these reflections are the
result of constructive interferences in a group of layers
with very similar thicknesses around 100 m. The other
reflectors, including the Moho, have a more
disorganized structure as indicated by their wider
frequency spectra. The results obtained in another zone
are presented in Fig. 7. Though the noise level is
higher, clear reflections with a lower frequency content
are visible between 7.5 and 9.5 s. The comparison with
Fig. 6 indicates that the response of the lower crust
Application to WISPA. The last example of energy
map was computed from the 180 traces of a single-
shotpoint of the WISPA dataset (Fig. 8). Since the
experiment was conducted on land, the spectra are
continuous i.e. void of the parasitic filter experienced in
the marine data. Another striking difference with WAM
is that the high frequencies are not strongly affected by
anelastic absorption because of the lack of a low Q
sediment cover. Finally, an acquisition low frequency
cutoff was applied below 15 Hz making it impossible to
detect constructive interferences in layers 100 m thick.
But, once again, a group of reflections with a very
narrow frequency content (28-30 Hz) can be detected
between 9.0 and 9.5 s. It could indicate that energetic
reflections are produced by constructive interferences
in a group of layers with very similar thicknesses
around 60 m. But we may be observing the first
harmonic of the frequency of resonance, in which case
the dominant thickness of the layers would be 180 m.
8 9 10
amax-9.7% TIME (s)
YOO0
i I I I I I I I I I I I I I I I I I I I t I I I I I I
I I I I I I I I t I I I I I I I I I I I { I I I I I I I I I I I I I I I
8000 9000 10000 11000
TIME (m)
also varies with distance along a profile. No result was Fig. 8. Map of the distribution of energy computed
obtained in the two other zones since both spectra from the 180 traces of shotpoint 11 of the WISPA
were dominated by the 10 Hz noise. experiment.
PAUL AND HOBBS 357
Conclusions References
Through a few examples from the WAM and
WISPA datasets, we have shown that the detailed
mapping of the spectral energy of the reflections in the
time-frequency domain is a useful tool to increase our
knowledge of the deep crustal reflectors. In a few cases,
energy appeared to be strongly focussed in narrow
frequency bands. We think that this concentration
Bouchon, M., A simple method to calculate Green's
functions for elastic layered media, Bull. seism. Soc.
Am., 71,959-971, 1981.
Freeman, B., S.L. Klemperer and R.W Hobbs, The
deep structure of northern England and the Iapetus
Suture zone from BIRPS deep seismic reflection
indicates that constructive interferences in groups of profiles, J. geol. Soc. Lond., 145, 727-740, 1988.
layers with similar thicknesses are present in the lower Hobbs, R.W., Effective Q determination using
crust. However, both the vertical and lateral variability frequency methods on BIRPS data, Tectonophysics,
are important, indicating a strong heterogeneity of the
structure of the lower crust.
The narrowness of the bandpass of seismic data is a
critical limitation to such studies. Only layers with
173, 25-30, 1990.
Mac Carthy, J. and G.A. Thompson, Seismic imaging
of extended crust with emphasis on the western
United States, Geol. Soc. Am. Bull., 100, 1361-1374,
thicknesses in the range 40 to 250 m induce 1988.
constructive interferences between 8 and 40 Hz. A Paul, A., and F. Nicollin, Thin crustal layering in
significant improvement of the power of resolution of northern France: observations and modelling of the
spectral methods would require a broadening of the PM P spectral content, Geophys. J. Int., 99, 229-246,
bandpass. Since high frequencies are generally absorbed 1989.
before reaching the lower crust, any effort towards high Peddy, C.P., B. Pinet, D. Masson, R. Scrutton, J.C.
frequencies is useless. But a lowering of the low
frequency cutoff to 5 Hz could be important to detect
layers thicker than 150 to 200 m in the lower crust.
Acknowledgements. This study was supported by
the National Environment Research Council, through a
visiting fellowship granted to AP, and by Institut
Sibuet, M.R. Warner, J.P. Lefort, and I.J. Shroeder
(BIRPS and ECORS), Crustal structure of the
Goban Spur continental margin, Northeast Atlantic,
from deep reflection profiling, J. geol. Soc., Lond.,
146, 427-437, 1989.
Peddy, C.P, Amplitude-versus-offset studies of the
reflective lower crust, this volume, 1991.
National des Sciences de l'Univers (INSU-CNRS) and
Institut Franqais pour la Recherche et l'Exploitation de Sandmeier, K.-J., and F. Wenzel, Synthetic
la Mer (IFREMER) through the financial support of seismograms for a complex crustal model, Geophys.
'Action d'accompagnement ECORS'. Part of the Res. Lett., 13, 22-25, 1986.
numerical simulations were performed at the Centre de Velzeboer, C.J., The theoretical seismic reflection
Calcul Vectoriel pour la Recherche. The BIRPS project response of sedimentary sequences, Geophysics, 46,
is supported by the Natural Environment Research 843-853, 1981.
Council and industry. The NEC and WAM profiles Ward G., and M. Warner, S-wave images of the layered
were shot by Geophysical Company (GECO). The lower continental crust (abstract), Annls Geophys.,
WISPA data were shot by Compagnie G4n4rale de EGS X/V General Assembly special issue, 1989.
G4ophysique (CGG). Cambridge Earth Science Warner, M., Absolute reflection coefficients from deep
contribution number 1819. seismic reflections, Tectonophysics, 173, 15-23, 1990.
CONSTRUCTIVE INTERFERENCE- GEOPHYSICAL MYTHOLOGY RE-EXAMINED
Bastiaan Spaargaren and Mike Warner
Department of Geology, Imperial College, London SW7 2BP, United Kingdom
Abstract. It is almost universally assumed by deep reflection seis-
mologists, that a multi-layered sequence of rocks, with only modest
acoustic impedance contrasts, can produce large apparent reflection
coefficients by constructive interference. It is also generally believed
that a quarter wavelength layer produces a factor of tvo increase in
amplitude, and that many fine layers can produce large effects. For
the effective seismic wavelets used in deep marine profiling none of
these assumptions is correct, and constructive interference is much
less important than is generally assumed.
Computer modelling using simple one-dimensional layered models,
including the effects of transmission losses and multiple reflections,
has been used to test the effects of constructive interference with a
variety of wavelets. A wide range of models has been tested, including
both deterministic and random models, in which variable layer spac-
ing, variable reflection coefficients and variable acoustic impedance
contrasts were generated.
The effectiveness of constructive interference depends strongly upon
the wavelet used; it is greatest for a long multi-cyclic wavelet, and is
least effective for a minimum-phase short-duration wavelet of the type
used in marine profiling. For a quarter wavelength layer, a typical
marine wavelet produces only a factor of 1.5 increase in apparent re-
flection coefficient, compared with a factor of 2 appropriate for a sine
wave or for the zero phase Ricker wavelet. Using marine wavelets on
the suite of multi-layered models, the largest increase in amplitude we
were able to obtain was just over a factor of 2. Constructive interfer-
ence from a plane layered model, using a realistic wavelet, can never
produce the large increases in amplitude often stated or assumed in
the seismic literature.
Introduction
Seismic reflections froin the deep continental crust often have sur-
prisingly large amplitudes [Warner, 1990]. Although acoustic imped-
ance contrasts between commonly occurring crystalline rock types are
sufficiently large to explain such bright reflections (see for example
Hurich and Smithson [1987]), it is often assumed that constructive
interference must play an important role in explaining these large
amplitudes [Fuchs, 1986; Klemperer et al., 1987; Roy-Chowdhury et
al., 1988]. Indeed it has become almost axiomatic among the deep
reflection profiling community, that bright reflections are predomi-
nantly the result of constructive interference produced froin a layered
sequence which contains only modest intrinsic impedance contrasts.
In this paper we examine the effects of constructive interference on
reflections by considering an exhaustive range of layered models and
incident wavelets. We are, here, interested only in constructive inter-
ference produced by a one-dimensional layered sequence, and do not
concern ourselves with constructive interference produced by focusing
Geodynamics 22
in two and three dimensions. Neither do we concern ourselves with
effects at wide-angle where the physics of constructive interference
is quite different from that at near-normal incidence. Our modelling
was performed in the time domain; it asstimes a plane wave normally
incident upon a one-dimensional, layered, elastic medium. We have
included transmission effects and the effccts of all multiples.
The range of models, and particularly the range of incident wavelets
we have considered, is more comprehensive than that used in earlier
studies [Fuchs, 1969; Blundell and Raynaud, 1986; Hurich and Smith-
son, 1987]. We find that constructive interference is considerably less
effective in generating large amplitudes than is generally assumed.
This is particularly so for marine surveys which use tuned airgun
arrays. We caution against the simplistic application of earlier pub-
lished results to deep seismic data, without taking account of the
wavelets which were used to generate those results.
Wavelets and models
During the initial stages of modelling, it rapidly became clear that
the type of incident wavelet used greatly influenced the effectiveness
of constructive interference. The greater the number of cycles present
in the wavelet, that is the more ringy the wavelet, the larger is the
potential for significant constructive interference.
Figure i shows examples of four types of wavelet that we have
used. The first wavelet has an infinitely long, multi-cyclic, periodic,
sinusoidal waveform. This is the type of wavelet used in the widely
quoted paper by Fuchs [1969]. It is clearly non-physical.
The second wavelet shown in Figure I has a symmetric, short-
duration, zero-phase waveform. This is analogous to the wavelets
used by Hurich and Smithson [1987]. It is the appropriate forin of
wavelet to use for an ideal vibrator survey after correlation. ttowever,
a zero-phase wavelet of this sort will not be representative of the true
effective wavelet, even in a vibrator survey, after it has been modified
by various approximately minimum-phase processes which occur in
the earth; these include anelastic attenuation, random scattering, and
short-period, near-surface multiple generation.
The third wavelet shown is of short duration, it is anti-symmetric,
and is approximately minimum phase. It is an idealized approxima-
tion to the fourth, which shows the real wavelet used on one of the
recent BIRPS marine surveys [Warner, 1990]. This fourth wavelet
shows a far field recording, in deep water, of a large tuned airgun
array. It includes the free-surface ghost, and has been modified to
include the effects of a theoretical earth attenuation filter which as-
sumes a constant Q of 1000 and reflection froin a depth of about 25
km.
The models tested in this study included regular, randomly gener-
ated, and hybrid models as shown in Figure 2. The number of layers,
the thickness of individual layers, and the intrinsic reflection coeffi-
cient of individual interfaces, were all varied. Layer thicknesses were
varied using a variety of statistical models which included bi-modal
and non-stationary distributions. Each type of model in turn was
examined using a wide range of parameters. In addition, for the sta-
359
360 CONSTRUCTIVE INTERFERENCE
FAAAAAAAAAAA
/VVVVVVVVVVV
I
k
Fig. 1. The four main wavelets used in the modelling.
tistical models, multiple runs were made using each set of parameters,
so that the maximum reflection coefficients obtained, and quoted here,
represent approximately the maximum of the appropriate ensemble,
and not just the maximhum of a particular sample run.
Intrinsic reflection coefficients were varied directly, using a similar
approach and a similar range of statistical models to those used to
vary the thickness, and also indirectly, by using the statistical models
to determine the acoustic impedance of individual layers. Towards
the end of this study we attempted to steer the models towards those
we expected to produce the largest effects, and to examine patho-
logical cases, in an attempt to maximize the amount of constructive
interference produced.
Impedance
i
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Fig. 2. The three main types of models used. The first is a periodic
sequence of layers; it contains multiple layers of uniform thickness
with the layer thickness equal to the layer spacing. The second is a
sequence where the thickness of the layers varies randomly. Different
probability distributions have been examined. The third type is a
sequence where each layer has a random acoustic impedance. Again
different probability distributions were tested.
Results
Single Layer
It is generally assumed that constructive interference produced by
a single thin layer, embedded in an otherwise uniform medium, will
produce a two-fold increase in reflection amplitude for a layer that is
a quarter of a wavelength thick. This assumption is incorrect.
Figures 3a and 3b show the effect of constructive interference for
a single thin layer for the wavelets shown in Figure 1. For all the
wavelets the maximum increase in amplitude occurs at a quarter
wavelength. For the sinusoidal and symmetric wavelets the maxi-
mum increase in amplitude is indeed by a factor of two. However,
for the anti-symmetric wavelet, and most significantly for the airgun
array, the increase in amplitude is only by a factor of one-and-a-
half. For an intermediate, mixed-phase, short-duration wavelet the
increase in reflection coefficient lies between 1.5 and 2. At a certain
thickness, the reflections from the upper and lower interfaces become
resolvable. Therefore the reflected signal is only ever analysed for its
peak-to-trough amplitude within one dominant period of the incident
wavelet, and no constructive interference occurs for thick layers.
It is clear that, for this most simple of layered models and for a real-
istic wavelet, in this case that of a tuned airgun array, the conventional
view significantly overestimates the contribution of constructive in-
terference. We will see that this phenomenon is repeated for more
complex models.
Regular Layering
Figure 4 shows how the apparent reflection coefficient, produced by
constructive interference within a regular sequence of quarter wave-
length layers, varies with the number o layers within the sequence.
For wavelets with only a few oscillations, i.e. the symmetric, anti-
0.2 2.0
.... 1.0 o
"1 ............ 7 '
/
/ I
1.o
Layer Thickness
Fig. 3a. Variation of apparent reflection coefficient with thickness
of a single layer using the sinusoidal and the symmetric wavelets.
The horizontal axis is thickness of the layer scaled by the dominant
wavelength of the incident wavelet. The vertical axis is the maxi-
mum apparent reflection coefficient of the layer, (i.e. the maximum
peak-to-trough amplitude reflected, divided by the amplitude of the
incident wavelet). The inherent reflection coefficient of the interface
at the top of the layer is 0.1, so greater than 0.1 implies constructive
interference, less than 0.1, destructive. Note the factor of 2 enhance-
ment.
0.15 _ _ 1.5
0.10
o
.9
r- 0.05
m
SINGLE
LAYER
MODEL
0 i i i
0 0.25 0.5 0.75
Layer Thickness
0
1.0
Fig. 3b. Variation of apparent reflection coefficient with thickness of a
single layer using the anti-symmetric and the BIRPS wavelets. Again
the inherent reflection coefficient of the interface at the top of the
layer is 0.1. Note that only 1.5 times enhancement is the maximum
reached.
symmetric and BIRPS wavelets, the results quickly reach a maximum.
ttowever, for long-duration periodic wavelets like the sinusoid shown,
the apparent reflection coefficients produced continue to increase with
increasing number of layers. By using an infinite-duration periodic
wavelet, the effects of constructive interference can be made to be ar-
bitrarily large; indeed, apparent reflection coefficients of greater than
one are easily obtainable.
Fuchs [1969], who has been widely quoted as showing the im-
portance of constructive interference in the deep crust, uses a non-
physical, infinite-duration, periodic sinusoid to generate large ap-
parent reflection coefficients from multi-layered sequences. Figure
4 shows clearly that such results cannot be applied in a direct way
1.2 _ 12
1.0
0.8
0.6
0.4
0.2
!,,-] i 111.9:e] i 1 -
. I:'i
.... m .............. m ........ m
0 I I I I 0
0 5 10 15 20 25
Number of Layers
10
Fig. 4. Variation of apparent reflection coefficient with number of
layers present in a periodically layered sequence. The reflection co-
efficient of the upper interface is agah 0.1. Note that the sinusold
reaches apparent reflection coefficients of greater than 1.
SPAARGAREN AND WARNER 361
to deep crustal seismic data which have been acquired using wavelets
with quite different properties. This figure shows the importance of
choosing a wavelet that is qualitatively appropriate, if results are to
be obtained that quantitatively match those from real experiments.
Figure 5 shows how the effects of constructive interference vary
with layer thickness in a regularly layered sequence of 25 layers. The
symmetric wavelet generates a maximum increase in reflection coeffi-
cient of just over 2.5 times. However, again, for the realistic wavelets
the maximum effect is less; the BIRPS wavelet produces a lnaximum
increase in reflection coefficient of only a factor of two.
Hurich and Smithson [1987] performed similar modelling, and man-
aged to obtain larger increases. For a wavelet which they call the
"10-80 Hz bandpass wavelet", they obtained an increase by a factor
of greater than 3. They suggest that the increase of the bandwidth
of the wavelet is responsible for the increase in constructive interfer-
ence. We have tried to confirm their apparent correlation between
bandwidth and effective constructive interference, but have been un-
able to reproduce any such effect.
0.25 / _ 2.5
i W]!=II;! [I,!
0.2 i % 2.0
i i -_. % i
o 1.5
0.15
-
0 0.5
0 i i i 0
0 0.25 0.5 0.75 1.0
Layer Thickness
Fig. 5. Variation of apparent reflection coecient with layer thickness
for a periodically layered sequence of 25 layers. efiection coecients
of the upper interfaces are 0.1. Note that only the symmetric wavelet
reaches over 0.2.
We suggest that the effect reported by Hurich and Smithson [1987]
is not related to varying bandwidth, but may instead have been pro-
duced by applying a steep-sided, zero-phase, band-pass filter to their
original wavelet. Such a non-physical filter will produce ringing in the
wavelet; this non-physical wavelet can then produce additional con-
structive interference. We reiterate that a short-duration minimum-
phase wavelet can produce only a maximum two-fold increase in re-
flection coefficient from a regularly layered sequence.
Random Layering
We have examined a wide range of random models in which both
layer thicknesses and layer contrasts were varied. In general the re-
sults were closely analogous to those produced by regular layering.
The symmetric wavelet produced the largest effects, and the BIRPS
wavelet produced results which were close to, and a little higher, than
those produced by the anti-symmetric wavelet. For the whole range
of stationary random models that we have tested, the largest increase
in reflection coefficient obtained for the BIRPS wavelet was under a
factor of 2.2.
362 CONSTRUCTIVE INTERFERENCE
Figure 6 shows results from a typical run which produced a max-
imum in constructive interference close to our all-time maximum of
2.2. We interpret the observation that random models are somewhat
more effective at producing constructive interference than are regular
models (2.2 vs 2), by noting that the BIRPS wavelet has a short tail
following the main wavelet, and that this tail has a period that is dif-
ferent to that of the main wavelet. In order to make this tail interfere
effectively with the rest of the wavelet we require a model which con-
tains a similar irregularity. We also suspect that some of our random
models are more effective at ordering the multiple sequence so that
it interferes constructively with the primary reflections. Both these
effects are small for short duration wavelets.
Non-stationary Models
The models considered above are all stationary models; that is,
their statistical properties do not change with depth. We have also
examined a wide range of non-stationary models; such models are
likely to be appropriate for the earth. We have been particularly
interested to examine the properties of models which have bright re-
flectors near the surface and weaker reflectors at depth. In such a
model, the multiples generated at shallow depths may modify the ef-
fective wavelet, so that it becomes more like the extended, multi-cyclic
wavelets which we have seen are capable of generating large effects
by constructive interference. Typically, the bright shallow reflectors
in such a model will correspond to the sea surface and sea bottom,
and to shallow sedimentary layers.
For models in which the intrinsic reflection coefficients increase or
are approximately constant with depth, we obtained results analogous
to those discussed under random layering. That is, we see constructive
interference generating maximum increases in amplitude of a little
over a factor of two.
In contrast, for models in which intrinsic reflection coefficients de-
crease markedly with depth, we do indeed find that constructive inter-
++ _ + + . +
0.2 q- -r q_ -I- 0.2
I A BIRPS
o i i i o
0 500 1000
Trial Number
Fig. 6. Results from 1000 calculation runs for a random thickness
model having a Gaussian probability distribution of layer thicknesses.
The sequence contained 25 layers with a mean of just over wave-
length and a standard deviation of of a wavelength. The incident
wavelet was the BIRPS wavelet. This model gave the highest amount
of constructive interference for the stationary models, just under a
factor of 2.2.
ference becomes more effective. That is, the weaker, deeper reflectors
have amplitudes that are enhanced by the multiples generated higher
in the section. Not surprisingly, the greater the contrast between shal-
low and deep reflection coefficients, the greater is this effect. When
the shallow reflectors are intrinsically ten times brighter than the deep
reflectors, we were able to generate a maximum constructive interfer-
ence of about 2.9; when the shallow reflectors are one hundred times
brighter, we are able to generate a nmximum of about 3.5.
We also see a competing effect. When the upper reflectors are too
bright, greater than about 0.2-0.3, we begin to see a reduction in
amplitude at depth, rather than an increase, because transmission
losses come to dominate the effect of constructive interference. For
deep crustal work, where apparent reflection coefficients are often
around 0.1, it is clear that the effects of constructive interference can
not be greatly enhanced over that predicted by the simple stationary
models discussed above.
Conclusions
We have shown that the effectiveness of constructive interference
generated by a one-dimensionally layered sequence depends critically
upon the seismic wavelet used. A long-duration, multi-cyclic wavelet
can produce significant enhancement of reflected amplitudes by con-
structive interference; the short-duration, minimum-phase wavelets
typically used in modern marine seismic surveys do not.
For the airgun arrays used by BIRPS the maximum increase in am-
plitude that can be produced by a single quarter wavelength layer is
1.5, not the factor of 2 commonly assumed. The maximum increase in
amplitude that can be produced from a regularly layered sequence is
close to 2. The maximum increase in amplitude that can be produced
from any model is about 2.2, provided only that the reflections being
studied are bright (reflection coefficients greater than about 0.03).
It is clear that the deep crust produces bright reflections on deep
seismic profiles because it contains bright reflectors.
Acknowledgements. The BIRPS source signatures used in mod-
elling were provided by GECO. We would like to thank Richard Hobbs
for useful discussions. B.S. would like to thank Shell International
Petroleum Co. Ltd. for a research scholarship.
References
Blundell, D. J., and Raynaud, B., Modelling lower crustal reflections
observed on BIRPS profiles, A GU Geodynamics Series, 13, pp.
287-295, 1986.
Fuchs, K., On the properties of deep crustal reflectors, Z. Geophys.,
35, 133-149, 1969.
Fuchs, K., Reflections from the subcrustal lithosphere, AGU Geody-
namics Series, ld, pp. 67-76, 1986.
Hurich, C. A., and Smithson, S. B., Compositional variation and the
origin of deep crustal reflections, Earth Planet. Sci. Left., 85, 416-
426, 1987.
Klemperer, S.L., and the BIRPS group, Reflectivity of the crystalline
crust: hypotheses and tests, Geophys. J. R. Astron. Soc., 89, 217-
222, 1987.
Roy-Chowdhury, K., Phinney, R. A., and Pan, J., Consequences of
multiple scales of layering for the crustal seismic response, paper
presented at the International workshop and symposium on seismic
probing of continents and their margins, Canberra, 1-8 July, 1988.
Warner, M. R., Absolute reflection coefficients from deep seismic re-
flections, Tectonophysics, 173, 15-23, 1990.
SEISMIC REFLECTIVITY PATTERNS: COMPARATIVE INVESTIGATIONS OF EUROPE AND NORTH AMERICA
Petra Sadowiak ' Rolf Meissnerand Larry Brown:
Abstract. Reflectivity maps of Europe and the are observed, and systematic correlations to tec-
United States of America have been compiled to tonic units found, on both continents. The reflec-
summarize major variations in the reflection tivity patterns are correlatable over large distan-
character of the crust. Significant differences in ces and provide a new and robust basis for recog-
reflectivity - largely independent of seismic nizing tectonically distinct crustal units.
technique - are observed and systematic cor-
relations with tectonic setting are found. Reflec- Seismic Reflectivity Patterns
tivity, concentrated in the lower crust as bands or
lamellae, is predominant in post-orogenic exten- A systematic comparative study of the reflecti-
sional areas, i.e. regions of orogenic collapse. vity of the continental crust now seems justified
Stable crustal units of Precambrian age seem to be by the great amount of deep seismic reflection data
characterized by more diffuse reflectivity, mostly accumulated in recent years. The locations of the
decreasing with depth, and by frequent diffraction- investigated reflection lines in Europe and the
like events. Compressional areas are marked by ramp United States are shown in Figures 1 and 2. In
and flat structures (thin-skinned tectonics) or general, stacked sections are preferred to migrated
special reflection-diffraction patterns (thick- sections: sometimes problems arise by migrating
skinned tectonics), further by seismic "crocodiles" deep seismic reflection data [Warner, 1987], and
and seismic duplexes. Major shear zones are charac- diffractions as a specific seismic pattern cannot
terized by deep-reaching, dipping reflection zones. be observed in migrated sections. Migrated sec-
tions, however, are consulted when considering dip-
Introduction ping events. If it is questionable whether energy
penetrates to the deeper parts of the crust at all,
In recent years many deep seismic reflection data are omitted. The studied profiles suggest re-
profiles have been collected by several groups, flectivity patterns that can be classified as in
e.g. COCORP and the USGS in the United States of Figure 3. Examples are given by Sadowiak et al.
America, LITHOPROBE in Canada, DEKORP in Germany, [1991].
BIRPS in Great Britain and ECORS in France. They Seismic LAMELLAE and multiple BANDS OF REFLEC-
show that the continental crust exhibits signifi- TIONS (Figures 3a, b) are observed often in Western
cant differences in reflection character, that are and Central Europe [BIRPS and ECORS, 1986; Lschen
independent of data acquisition and processing et al., 1987] and in the Basin and Range Province
parameters [Meissner et al., 1990; Trappe et al., of the western United States [Hauge et al., 1987,
1988]. Allmendinger et al. [1987] proposed that Klemperer et al., 1986]. The reflectivity is more
specific reflectivity patterns were related to or less concentrated as subhorizontal reflections
tectonic regimes. This has been extended by in the lower crust, often accompanied by a mainly
Sadowiak et al. [1991] for Europe and by Brown transparent upper crust with only some dipping
[1991] for the United States of America. reflections. Occasionally only one single band of
This paper presents reflectivity maps of Europe reflections is present at Moho-level (Mohoband),
and the United States, showing major variations in e.g. in southwestern Georgia and northern Florida.
reflection character. Similar reflectivity patterns DIFFRACTIONS are mostly observed in the middle
lInstitute for Geophysics, University Kiel,
Olshausenstr. 40-60, 2300 Kiel, Germany
:Institute for the Study of the Continents,
Cornell University, Ithaca, N.Y. 14853
Geodynamics 22
and lower crust, either dominating in clusters of
hyperbola-shaped events (Figure 3c), as in many
parts of the midcontinental United States [Brown et
al., 1983], or appearing together with inclined up-
per crustal reflections (Figure 3d). The latter re-
flection-diffraction pattern was first recognized
on DEKORP line 2-S in the area of the suture be-
tween the Variscan Moldanubian and Saxothuringian
units in southern Germany [Sadowiak et al., 1989].
Additional examples from other regions are sum-
marized in Sadowiak and Wever [1990].
363
364 REFLECTIVITY PATrERNS IN EUROPE AND NORTH AMERICA
o'
$O'Ai .
10E
Figure 1. Location map of investigated deep seis-
mic reflection profiles in Europe.
short and inclined, are concentrated in the upper
crust, and the Moho is not clearly defined (Figure
3f).
The SEISMIC DUPLEX (Fig. 3g) and the RAMP AND
FLAT STRUCTURE (Fig. 3h) have been observed in
several instances. The first pattern is formed by
reflections of rhomboidal shape in the upper and
middle crust. It appears on KTB line 5 in the Ger-
man Oberpfalz area [DEKORP Res. Group, 1988], on
BIRPS DRUM line north of Scotland [Snyder, 1990]
and on COCORP New Hampshire line 6 [Ando et al.,
1984]. The RAMP AND FLAT STRUCTURE is a composition
of low- and high-angle reflections in the upper and
middle crust. It is observed on profiles approach-
ing old, stable crustal units, e.g on lines close
x (km)
'
!c)
0
Res. Group, 1988], Alps [Bayer et al., 1987; Frei 5
o/< ;--''-
et al., 1989], Pyrenees [Choukroune et al., 1989],
and Carpathians (Tomek, personel communication,
1989) of Europe and the Cordillera in Washington
[Potter et al., 1986] and Wyoming [Smithson et al.,
1979; Allmendinger et al., 1982] of the United
States, seismic CROCODILES [Meissner, 1989] can be
identified. They consist of clearly diverging and
rather strong, planar reflections in the middle and
lower crust (Fig. 3e).
REFLECTIVITY DECREASING WITH DEPTH appears on
reflection lines in the midcontinental United
States [Gibbs, 1986] and in Sweden [Dahl-Jensen et
al., 1987] and Norway. The reflections, mostly
lb)
x (km L
5 -- '""
o
(d}
x (km) xlkm)
._.0
' If)
(e)
xlkm) I x(km)
5 ::"'--:---
x (km) x (km)
it ;::; o. 1 10,--:- -' ....
_
x (km)_
F_F4 =;
Figure 3. Classification of seismic reflectivity
patterns: (a) lamellae, (b) bands of reflections,
(c) diffractions, (d) diffractions accompanied with
dipping upper crustal reflections, (e) seismic cro-
codiles, (f) decreasing reflectivity with depth,
Figure 2. Location map of investigated deep seis- (g) ramp and flat structure, (h) seismic duplex,
mic reflection profiles in the United States of (i) deep-reaching, steeply dipping reflection
America. zones.
$ADOWIAK ET AL. 365
to the London-Brabant Massif in Belgium [DEKORP
Res. Group, 1991]. Profiles in the United States
approaching the NorthAmerican craton from the west
in the Rocky Mountain area [e.g., Potter et al.,
1987] and from the east in the Appalachians [e.g.,
Cook et al., 1979] also exhibit ramp and flat
patterns.
Some lines in the United States exhibit deeply
penetrating, steeply DIPPING REFLECTION ZONES
(Figure 3i). This has been observed for example on
the COCORP Ohio lines [Culotta et al., 1990].
The distributions of the predominant reflecti-
vity patterns are summarized in reflectivity maps
of Europe and the United States (Figures 4 and 5).
Lamellae or bands of reflections in the lower crust
are widespread in Western and Central Europe and in
the western United States, while diffuse, sometimes
decreasing with depth, reflectivity and diffrac-
tions dominate the crust of the midcontinental
United States.
120
0 , 100 90 80 70 /
Figure 5. Reflectivity map of the United States of
America, (1) lamellae, (2) bands, (3) diffractions,
Correlation of Reflectivity Patterns to Tectonic (4) decreasing reflectivity, (5) seismic croco-
Settings diles, (6) seismic duplex, (7) ramp and flat struc-
ture, (8) deep-reaching, steeply dipping reflection
Figures 6 and 7 show simplified tectonic maps of zones.
Europe and the United States. Whereas in both
figures (a) shows mainly Precambrian crustal units,
(b) emphasizes extensional areas of post-orogenic
collapse and rifts.
6o /.
6o v.
10'E
20'E
'60 N
60'N
SO
50'N
50 N '
20 E
0
Figure 6a. Simplified tectonic map of Europe
Figure 4. Reflectivity map of Europe, (1) lamel- showing Precambrian crustal areas: (1) Precambrian
lae, (2) bands, (3) diffractions, (4) decreasing shield crust, (2) buried Precambrian massifs; (3)
reflectivity, (5) seismic crocodiles, (6) seismic pre-orogenic massifs; (4) Fennoscandian Border
duplex, (7) ramp and flat structure. Zone; (5) thrusts.
366 REFLECTIVITY PATTERNS IN EUROPE AND NORTH AMERICA
6 0 '
r
' -60 N
[Wever and Sadowiak, 1991], seismic crocodiles and
occasionally seismic duplexes [Price, 1981, 1986].
- On a more local scale, the old, stable London-
Brabant Massif, squeezed between the Caledonian and
Variscan orogenies, shows compressional reflecti-
vity patterns in a generally low reflective crust
[Bouckaert et al., 1989].
- Major crustal shear zones in the eastern part
of the United States are marked by deep-reaching,
steeply dipping reflection zones [Pratt et al.,
1989; Culotta et al., 1990].
In spite of the still limited coverage of deep
seismic reflection profiles, this initial attempt
of mapping the lateral distribution of predominant
:
,0
E ...... ', -'
O'
o 20 E
Figure 6b. Simplified tectonic map of Europe
showing extensional areas of post-orogenic col-
lapse: (1) post-Caledonian collapsed areas, (2)
post-Variscan collapsed areas; (3) rifts; (4) Figure 7a. Simplified tectonic map of the United
Alpine Deformation Front. States showing Precrian crustal areas [after
Bally et al., 1989]: (1) thrusts, (2) Paleozoic
platform cover, (3) Precrian shield and sedi-
ments, (4) Colorado Plateau, (5) Laride basement
By comparing the reflectivity and the tectonic
uplifts, (6) Paleozoic basement uplifts. Stable
maps the following points become evident:
craton: (2) and (3), metastable craton: (4), (5)
- Lellae and bands of reflections in the lower
and (6).
crust are widespread in young extensional areas of
post-orogenic collapse. There is a suggestion that
the younger the extension, the denser the reflec-
120 10 10 9'0 o 80
only Moho band reflectivity, while Cenozoic exten-
sional areas show pronounced lellae or multiple
bands of enhanced reflectivity over much of the
0 o
Great Britain and Norway and the Keweenawan Rift in
the midcontinental United States are dominated by
diffractions in the middle and lower crust F ',,. .,{ //
- The investigated Precrian areas are charac-
terized by more diffuse reflectivity, sometimes }, ,. decreasing with depth and by diffraction-like
event s. Spe c i f ic Moho r e f 1 ect ion s are unc oon or 2 5
weak in those regions, at least in the United
States . 20 0 00 B0
- Young and ancient compressional areas are fre-
ently marked by reflector geometries suggesting Figure 7b. Simplified tectonic map of the United
overthrusting or indentations: rp and flat struc- States showing Precrian crustal areas [after
tures or diffractions coined with dipping reflec- Bally et al., 1989] showing extensional areas: (1)
tions in the upper crust, according to thin- or Precrian, (2)Paleozoic, (3)Mesozoic, (4)Ceno-
thick-skinned tectonic enviroents respectively zoic; (5) intra-cratonic basins; (6) thrusts.
$ADOWIAK ET AL. 367
reflectivity patterns indicates how reflectivity It is also clear from this exercise that much
can serve as a guide to subsurface geology and its more reflection coverage is needed, both to test
utility in matching lower crustal structure to sur- correlations such as these and to fill in the con-
face tectonic expression. sidered gaps that exist even in Europe and North
America. As for the rest of the world, much of it
remains terra incognita as far as even preliminary
Discussion and Conclusions surveying is concerned.
This exercise suggests that seismic reflectivity Acknowledqments. This work was financially
patterns are correlatable over large distances and supported by the German BMFT under grant RG 8703 6
provide a new method for identifying geologically (DEKORP), by the DFG under grant ME 335/77-5, and
and tectonically distinct crustal units and ter- by the German DAAD. We thank European collegues for
ranes, young and old suture zones, and other ter- their cooperation in showing original seismogramms,
rane boundaries. Seismic reflectivity patterns can owing special thanks to the BIRPS group in Cam-
often be seen as independent of the field techni- bridge (UK), W. Frei (Switzerland) and C. Tomek
que. These clearly correspond to variations in sub- (CSFR). We thank A. Green for critical comments and
surface geology, and not to any artifact of acqui- reviewing. This is contribution No. 420 from Insti-
sition or data processing, although in some cases tut fr Geophysik, Kiel.
problems with energy penetration can not be totally
excluded.
Systematic correlations of seismic reflectivity
patterns and tectonic units are suggested for Eu-
rope and the United States. From these results it
seems that young extensional crusts of collapsed
areas are characterized by strong lower crustal
reflectivity and rather tranparent upper crusts,
References
Allmendinger, R., Brewer, J., Brown, L.D., Oliver,
J., and S. Kaufman, COCORP profiling across the
Rocky Mountain front in southern Wyoming, Part
II: Precambrian basement structure and its influ-
ence on Laramide deformation, Geol. Soc. Am.
contrasting with the poorly reflective and dif- Bull., 9--3, 1253-1263, 1982.
fractive lower crust of many Precambrian crustal Allmendinger, R., Nelson, D., Potter, C., Bara-
sections. zangi, M., Brown, L.D., and J. Oliver, Deep seis-
However, the recent BABEL data [BABEL Working
Group, 1990, 1991] for the Proterozoic Baltic
Shield found a highly diffractive crust adjacent to
regions with strong lower crustal reflectivity, si-
milar to the bands of reflections in Phanerozoic
Western Europe. It has been suggested that these
lower crustal reflectors of the Baltic Shield are
old Proterozoic features [Gal and Gorbatschev,
1987] that were never disrupted or destroyed by
subsequent tectonic events. Tectonic stability to-
mic reflection characteristics of the continental
crust, Geoloq, 15, 289- 386, 1987.
Ando, C.J., Czuchra, B., Klemperer, S., Brown, L.,
Cheadle, M., Cook, F., Oliver, J., Kaufman, S.,
Walsh, T., Thompson, J., Lyons, J., and J.L. Ro-
senfeld, Crustal profile of a mountain belt:
COCORP deep seismic reflection profiling in New
England Appalachians and implications for ar-
chitecture of convergent mountain chains, Am.
Assn. Petr. Geol. Bull., 68, 819-837, 1984.
gether with the absence of high stress levels and BABEL Working Group, Evidence for Early Proterozoic
variable stress orientations in the interior of the plate tectonics from seismic reflection profiles
shield [Mller et al., 1991] might be factors that in the Baltic Shield, Nature, 348, 34-38, 1990.
allowed such reflections to survive in the Baltic BABEL Working Group, Reflectivity of a Proterozoic
Shield while not in other cratons. shield: examples from BABEL seismic profiles
The difference in reflectivity of the North across Fennoscandia, in this volume, American
American craton and the Baltic Shield might be due Geophysical Union, Washington D.C., 1991.
to widespread Proterozoic anorogenic magmatism in Bally, A.W., Scotese, C.R., and M.I. Ross, North
the former America [Emslie, 1978; Van Schmus et America: Plate-tectonic setting and tectonic ele-
ments, in The Geoloqy of North America - An Over-
al., 1987]. This mechanism might disrupt pre-
existing reflections by repeated dike injection view, edited by A.W. Bally and A.R. Palmer, Vol.
[Nelson, 1991]. LITHOPROBE deep seismic reflection A, Geol. Soc.America, Boulder, Colorado, l-15, 1989.
data in Canada also show some reflections in the Bayer, R., Cazes, M., Dal Piaz, V., Damotte, B.,
lower crust [Green et al., 1990], and Moho reflec-
tions are observed in cratonic areas of Australia
[Finlayson, 1982] and South Africa [Durrheim et
al., 1987] (B. Rodney, personel communication,
1990). These sections, however, suggest more and
stronger reflections in the upper than in the lower
crust [e.g. Wu and Mereu, 1990] in contrast to the
transparent upper crusts and highly reflective
Elter, G., Gosso, G., Hirn, A., Lanza, R., Lom-
bardo, B., Mugnier, J.-L., Nicolas. A., Nicolich,
R., Polino, R., Roure, F., Sacchi, R., Scarascia,
S., Tabacco, I., Tapponier, P., Tardy, M., Tay-
lor, M., Thouvenot, F., Torreilles, G., and A.
Villien, Premiers resultats de la traversee des
Alpes occidentales par sismique reflexion verti-
cale (Programme ECORS-CROP), C. R. Acad. Sci.
lower crusts in young extensional areas of orogenic Paris, 305, Serie II, 1461-1470, 1987.
collapse. Because different cratons have undergone BIRPS and ECORS, Deep seismic reflection profiling
different evolutionary histories, perhaps a broader between England, France and Ireland, J. qeol.
range of reflectivity patterns should be expected. Soc. Lond., 143, 45-52, 1986.
368 REFLECTIVITY PATrERNS IN EUROPE AND NORTH AMERICA
Bouckaert, J., Fock, W., and N. Vandenberghe, First Gibbs, A.K., Seismic reflection profiles of Pre-
results of the Belgian geotraverse 1986 (BELCORP)
Ann. Geol. Belq., 111, 279-290, 1989.
Brown, L.D., A new map of crustal "terranes" in the
United States from COCORP deep seismic reflection
profiling, Geophs. J. Int., 105, in press, 1991.
Brown, L.D., Serpa, L., $etzer, T., Oliver, J.,
Kaufman, S., Lillie, R., Steiner, D., and D.W.
Steeples, Intracrustal complexity in the United
States mid-continent: preliminary results from
COCORP surveys in northeastern Kansas, Geoloqy,
1__1, 25-30, 1983.
cambrian crust: a qualitative assessment, in Re-
flection Seismoloqy: The Continental Crust, edit-
ed by M. Barazangi and L. Brown, Geodyn. Ser.,
vol. 14, American Geophysical Union, Washington
D.C., 95-106, 1986.
Green, A., Milkereit, B., Percival, J., Davidson,
A., Parrish, R., Cook, F., Geis, W., Cannon, W.,
Hutchinson, D., West, G., and R. Clowes, Origin
of deep crustal reflections: seismic profiling
across high-grade metamorphic terranes in Canada,
Tectonophsics, 173, 627-638, 1990.
Choukroune, P., and ECOR$ Team, The ECOR$ Pyrenean Hauge, T., Allmendinger, R., Caruso, C., Hauser,
deep seismic profile: reflection data and the E., Klemperer, S., Opdyke, S., Potter, C., San-
overall structure of an orogenic belt, Tectonics, ford, W., Brown, L., Kaufman, S., and J. Oliver,
, 23-39, 1989. Crustal structure of western Nevada from COCORP
Cook, F.A., Proterozoic thin-skinned thrust and deep seismic reflection data, Geol. Soc. Am.
fold belt beneath the Interior Platform in north- Bull., 9--8, 320-329, 1987.
west Canada, Geol. Soc. Am. Bull., 100, 877-890, Klemperer, S., Hauge, T., Hauser, E., Oliver, J.,
1988. and C. Potter, The Moho in the northern Basin and
Cook, F.A., Albaugh, D.S., Brown, L.D., Kaufman, Range province, Nevada, along the 40 transect,
S., Oliver, J., and R.D. Hatcher, Thin-skinned Geol. Soc. Am. Bull., 9--7, 603-618, 1986.
tectonics in the crystalline southern Appalachi- Lschen, E. et al., Seismic reconnaissance studies
ans: COCORP seismic reflection profiling of the of the continental deep drilling KTB-location
Blue Ridge and Piedmont, G. eoloqy, !, 563-567,1979. Black Forest, SW Germany, J. Geophs., 6--1, 1-10,
Cook, F.A., Simony, P.S., Coflin, K.C., Green, 1987.
A.G., Milkerheit, M., Price, R.A., Parrish, R., Meissner, R., Rupture, creep, lamellae and croco
Patenaude, C., Gordy, P.L., and R.L. Brown, diles: happenings in the continental crust, Terra
Lithoprobe southern Canadian Cordilleran tran- Nova, 1, 17-28, 1989.
sect: Rocky Mountain thrust belt to Valhall Meissner, R., Wever, Th., P. Sadowiak and DEKORP
gneiss complex, Geophys. J. Roy. Astr. Soc., 8--9, Res. Group, Reflectivity patterns in the Variscan
91-98, 1987. mountain belts and adjacent areas: an attempt for
Culotta, R. C., Pratt, T., and J. Oliver, A tale a pattern recognition and correlation to tectonic
of two sutures: COCORP's deep seismic surveys of units, Tectonophysics, 173, 361-378, 1990.
the Grenville province in the eastern U.S. mid- Mller, B., Fuchs, K., Mastin, L., Gregersen, S.,
continent, Geoloqy, 1-8, 646-649, 1990. Pavoni, N., Stephansson, O., Ljunggren, C., and
Dahl-Jensen, T., Dyrelius, D., Juhlin, C., Palm, M.L. Zoback, Regional patterns of stress in Eu-
H., and L.B. Pedersen, Deep reflection seismics rope, J. Geophys. Res., in press, 1991.
in the Precambrian of Sweden, Geophys. J. Roy. Nelson, K.D., A unified view of craton evolution
Astr. Soc., 89, 371-378, 1987.
DEKORP Research Group, Results of the DEKORP 4/KTB
Oberpfalz deep seismic reflection investigations,
J. Geophys., 62, 69-101, 1988.
DEKORP Research Group, Results of the DEKORP 1
(BELCORP-DEKORP) deep seismic reflection studies
in the western part of the Rhenish Massif, Geo-
phys. J. Int., in press, 1991.
Durrheim, R.J., Seismic reflection and refraction
studies of the deep structure of the Agulhas
motivated by recent deep seismic reflection and
refraction results, Geophys. J. Int., 105, in
press, 1991.
Potter, C.J., Sanford, W.E., Yoos, T.R., Prussen,
E.I., Keach, II, R.W., Oliver, J.E., Kaufman, $.,
and L.D. Brown, COCORP deep seismic reflection
traverse of the interior of the North American
Cordillera, Washington and Idaho: implications
for orogenic evolution, Tectonics, , 1007-1025,
1986.
Bank, Geophys. J. R. astr. Soc., 8-9,395-398, 1987. Potter, C.J., Allmendinger, R.W., Hauser, E.C., and
Emslie, R.F., Anorthosite massifs, Rapakivi gra-
nites, and Late Proterozoic rifting of North
America, Precambrian Res., !, 61-98, 1978.
Finlayson, D., Seismic crustal structure of the
Proterozoic North Australian Craton beneath Ten-
nant Creek and Mount Isa, J. Geophs. Res., 87,
10569-10578, 1982.
Frei, W., Heitzmann, P., Lehner, P., Mueller, St.,
J.E. Oliver, COCORP deep seismic reflection tra-
verses of the U.S. Cordillera, Geophys. J. Roy.
Astr. Soc., 89, 99-104, 1987.
Pratt, T., Culotta, R., Hauser, E., Nelson, D.,
Brown, L., Kaufman, S., and J. Oliver, Major Pro-
terozoic basement features of the eastern midcon-
tinent of NorthAmerica revealed by recent COCORP
profiling, Geoloqy, 17, 505-509, 1989.
Olivier, R., Pfiffner, A., Steck, A., and P. Va- Price, R.A., The Cordilleran foreland thrust and
lasek, Geotraverses across the Swiss Alps, Na- foldbelt in the southern Canadian Rocky Moun-
ture, 340, 544-548, 1989. tains, in Thrust and Nappe Tectonics, edited by
Ga&l, G., and R. Gorbatschev, An outline of the N.J. Price and K. McClay, Geol. Soc. Lond. Spec.
Precambrian evolution of the Baltic Shield, Pre- Publ., No. 9, Blackwell, Oxford, 427-448, 1981.
cambrian Res., 3--5, 15-52, 1987. Price, R.A., The southeastern Cordillera: thrust
$ADOWIAK ET AL. 369
faulting, tectonic wedging, and delamination of flectivity pattern and its relation to geologic
the lithosphere, .Struct. Geol.,,239-254,1986. province, Geophys. Prosp., 36, 265-282, 1988.
Sadowiak, P., and Th. Wever, Reflection-diffraction Van Schmus, W.R., Bickford, M.E., and I. Zietz,
seismic patterns at crustal suture zones, Tec-
tonics, , 1495-1513, 1990.
Sadowiak, P., Voss, J., and R. Meissner, 3D
modelling of diffractions observed on deep re-
flection line DEKORP 2-S, Geophys. Prosp., 37,
623-637, 1989.
Sadowiak, P., Wever, Th., and R. Meissner, Deep
seismic reflectivity patterns in specific tec-
Early and middle Proterozpoic Provinces in the
Central United States, in Proterozoic Lithosphe-
ric Evolution, edited by A. Kr6ner, Geodyn. Ser.,
vol. 17, American Geophysical Union, Washington
D.C., 43-68, 1987.
Warner, M., Migration - why doesn't it work for
deep continental data?, Geophys. J. Roy. Astr.
Soc., 89, 21-26, 1987.
tonic units of Western and Central Europe, Geo- Wu, J., and R.F. Mereu, Seismic reflectivity pat-
phys. J. Int., 105, in press, 1991.
Smithson, S.B., Brewer, J., Kaufman, S., Oliver, J.
and C. Hurich, Nature of the Wind River thrust,
Wyoming, from COCORP deep reflection data and
from gravity data, Geoloqy, , 648-652, 1979.
Snyder, D.B., The Moine Thrust in the BIRPS data
set, J. Geol. Soc. Lond., 147, 81-8fi, 1990.
Trappe, H., Wever, T., and R. Meissner, Crustal re-
terns of the Kapuskasing structural zone, Ab-
stract volume, Fourth Int. Symp. on Deep Seismic
Reflection Profilinq of the Continental Lithos-
phere, 207, 1990.
Wever, Th., and P. Sadowiak, Crustal suture zones:
seismic signature and structural interpretation,
in this volume, American Geophysical Union,
Washington D.C., 1991.
CRUSTAL SUTURE ZONES: SEISMIC SIGNATURE AND STRUCTURAL INTERPRETATION
Thomas Wever and Petra Sadowiak
Institute for Geophysics, University, Olshausenstr. 40-60, 2300 Kiel, Germany
Abstract Collision of continents always re- et al., 1979], or the North Variscan Deformation
sults in a juxtaposition of different terranes. Front in Germany [Meissner et al., 1981], the R/D
At the surface the prominent sutures can be easi- pattern has not been observed Here, a ramp-and-
z uwsu uz ' tls W OiO9 and tecto- fill rhiLecture is dominnh
nics on either side. At depth only indirect map-
ping is possible. We inspected suture-crossing
near-vertical reflection seismic profiles from
three continents. The strong similarity of seis-
mic patterns leads us to the conclusion that the
typical pattern may be a diagnostic tool for
identifying continent-continent collisions in
areas where outcrops are hidden by e.g., post-
collision sediments. The pattern is not observed
in association with thin-skinned tectonics. We
discuss various models for the interpretation of
A Case Study Along DEKORP 2-S
DEKORP 2-S was the very first DEKORP profile
[DEKORP Res. Group, 1985]. Explosive energy sour-
ces also served for piggy-back recordings in the
area of DinkelsbHhl in southern Germany [Milke-
reit et al., 1986]. These additional recordings
allowed a 3-D interpretation in a zone of promi-
nent hyperbolic events in the middle and lower
the observed seismic signature across thick-skin- crust [Sadowiak et al., 1989]. The seismogram in
ned sutures. Figure 4 depicts that part of profile DEKORP 2-S
which portrays the diffractions. The dipping re-
Observations flections in the upper crust trace upwards to the
(sediment covered) subcrop of the suture, and
Deep seismic near-vertical reflection profiles downwards to the diffractions. The approximately
across moderate- to steep-dip (more than 25 ) su- planar inclined reflections originate from the
ture zones between different crustal terranes or
provinces reveal a distinct seismic reflectivity
pattern of dipping reflections in the upper crust
and related diffractions in the lower crust. This
"reflection/diffraction" (R/D) pattern is sketch-
ed in Figure 1. Figures 2 and 3 show the locali-
ties where it can be observed. For a detailed
description of the examples (DEKORP and KTB pro-
files in Germany, BIRPS lines around Britain, and
USGS and COCORP lines in the USA) we refer to Sa-
dowiak and Wever [1990]. In addition to the Euro-
pean and North American data, this pattern can be
seen on Australian profiles [Goleby et al., 1988,
1990].
Across shallow-angle thrusts marking orogenic
fronts, e.g., in the southern Appalachians [Cook
Now at: Forschungsanstalt der Bw. fHr Wasser-
schall- und Geophysik, Klausdorfer Weg 2-24,
2300 Kiel 14, Germany.
0 20 x (km)
0 i I I
Geodynamics 22
e 1991 American Geophysical Union
Fig. 1. Cartoon showing the characteristic (R/D)
pattern of diffractions in the lower crust and
accompanying dipping reflection in the upper
crust.
371
372 SEISMIC SIGNATURE AT CRUSTAL SUTURE ZONES
10W 0
I I i
50N
I I
0 10OE
10OE 20E
60N
-50N
Fig. 2. Location map of R/D patterns in Europe.
"1": DEKORP 2-S Dinkelsbhl [DEKORP Res. Group,
1985], "2 ": KTB 8401 Black Forest [Lschen et
al., 1987], "3": DEKORP 2-S Spessart Mnts. [Behr
and Heinrichs, 1987], "4": BIRPS' line 'WINCH'
[Hall et al., 1984].
layer along which the Moldanubian province was
thrust to the north over the $axothuringian pro-
vince.
Using 3-D traveltime modeling Sadowiak et al.
[1989] showed that the diffractions are caused by
horizontally extended line diffractors oriented
parallel to the suture.
Until recently line DEKORP 2-S was the only
example where such detailed investigations were
possible. However, as more reflection data across
mapped or inferred suture zones become available
the distinct pattern shown in Figure 1 becomes
more common. The observation of this pattern
seems to be restricted to thick-skinned sutures
(i.e., suture dips exceeding 25).
Discussion
However, upper crustal reflections may also
lose their strength - unless they are reactivated
as they become older. Young sutures (and a few
older ones) still show such dipping reflections
throughout the whole crust and diffractions in
its middle part. An important example is the less
than 10 Ma old northern Carpathian collision (C.
Tomek, 1989, personal communication) where first
signs of overprinting (horizontal reflections in-
tersecting older inclined ones, interpreted as
the newly equilibrated Moho) are observed. A sig-
nificant aspect of the R/D pattern is the allian-
ce with thick-skinned sutures. This fact, togeth-
er with the detailed analysis of the DEKORP 2-S
suture gives some hints on the origin of the dif-
fractions. Their elongated source, paralleling
the suture might be caused by:
(1) Elongated dikes. Their volume makes up a
large portion of the continental crust [see e.g.,
Taylor and McLennan, 1985]. If constrained to the
lower crust they are possible candidates for the
line diffractors, similar to the salt walls per-
colating sedimentary basins. However, dikes are
not genetically related or restricted to suture
zones and are thus unlikely to be the main cause
of the diffractions.
(2) The termination of upper crustal brittle
faults (partly bi- or multifurcating) in a duc-
tile middle/lower crust. This is related to the
suture zone itself and is a possible explanation
(see Figure 5).
(3) Magmatic intrusions spreading along thrust
planes between the collided units. They represent
another suture zone-related interpretation of the
diffractions (see Figure 6). In addition, magmas
of multiple intrusions might stack in the middle
crust as argued by Kampfmann [1988]. He proposed
that the 'diffractions' originate at the tops of
/., O
Thrusting of one continent or terrane over an
other results in prominent shear zones. In gene-
ral, their traces are preserved only in the upper
crust. The disappearance of inclined shear zones
in the deep crust probably results from post-col-
lisional collapse [Meissner et al., 1991] which
also causes the orogenic roots to disappear. Du- -
ring such processes a complete re-orientation and
re-organization of structures occurs in the weak 120
lower crust. Consequently, subhorizontal lower
30 ...
20 o 0 o 100 o 90 t50 70
I, I ! I, I I
110 o 100 o 90 80 o
...z. 0 o
-30 o
crustal reflectivity is younger than reflections Fig. 3. Location map of some examples in North
in the brittle upper crust in such areas. A chan- America. "5": USGS Maine [Hutchinson et al.,
ge of lower crustal reflectivity is not to be ex- 1988], "6": COCORP 'Georgia 13/14' [Nelson et
pected across former suture zones affected by al., 1985], "7": COCORP 'Washington 1/2' [Potter
post-collisional collapse. et al., 1986; Sanford et al., 1988].
10
A
WEVER AND $ADOWIAK 373
N $
80 km 5 0 km
10
N S
80 km 5 0 km
0 i .L- 0
5 5
10 10
B
Fig. 4. Part of the unmigrated seismogram of profile DEKORP 2-S in the DinkelsbQhl area
(km 40-92), (a) without interpretation, (b) including interpretation.
374 SEISMIC SIGNATURE AT CRUSTAL SUTURE ZONES
! o' o o o o
Fig. 5. Mechanism to create line diffractors that
are elongated parallel to a suture zone: termina-
tion of upper crustal brittle faults, possibly
bi- or multifurcating in the upper and middle
crust as observed e.g., along COCORP's Wind River
lines [Smithson et al., 1979]. M: Moho, RLC= re-
flective lower crust, S: sediments.
multiple intrusions and represent reflections.
Here, the ratio of seismic wavelength to diameter
of intrusion is a critical parameter= one might
either record diffractions or reflections from
the tops of intrusive bodies.
(4) An alternative, and in our view the most
probable, explanation is interwedging. During
collision, material of different composition (and
hence theology) is juxtaposed along the suture
(Figure 7). These differences may lead to an in-
terfingering at different depths and can cause
the clustered appearance of the observed diffrac-
tions.
Fig. 7. Interwedging of two terranes (T1, T2) du-
ring collision resulting in an interfingering of
material of different composition and rheology
along the suture as mechanism to create line dif-
fractors parallel to a suture zone. Abbreviations
as in Figures 5 and 6.
A definitive explanation for the occurrence of
prominent high-energy diffractions is not yet
possible. However, the causes can be limited. For
a better understanding of the tectonic processes,
specially designed high-resolution 3-D experi-
ments across suture zones with known diffraction
clusters are needed to allow traveltime and am-
plitude modeling.
We postulate a genetic relation of the R/D
pattern and thick-skinned suture zones, similar
Fig. 6. Intrusion of magma along the fault plane to the observed ramp-and-flat pattern associated
between the collided terranes forming a magmatic with thin-skinned thrust sheets which obviously
sheet of limited extend in the middle crust, an- can only develop in areas with a prominent rigid
other suture zone related interpretation of line block on which the thin veneers are thrust.
diffractors. Multiple intrusions are possible and
might be responsible for the clustered appearance Acknowledqements. Continuous support and fun-
of diffractions. FP: fault plane, I: intrusion, ding of the project DEKORP by the German Bundes-
other abbreviations see Fig. 5. ministerium fur Forschung und Technologie (BMFT)
WEVER AND SADOWIAK 375
under grant RG 8703 6 is gratefully acknowledged. Lschen, E., et al., Seismic reconnaissance stu-
Part of this work was financed by the Deutsche dies of the deep drilling location KTB-Black
Forschungsgemeinschaft under grant Me 335/77-5. Forest, SW Germany, J. Geophs., 6--2, 1-30,
Comments of Alan Green and Larry Brown helped to 1987.
improve the presentation of our ideas. This is Meissner, R., H. Bartelsen, H. Murawski, Thin-
contribution No. 417 from Institut fur Geophysik, skinned tectonics in the northern Rhenish Mas-
Universitt Kiel. sif, Germany, Nature, 290, 399-401, 1981.
References Meissner, R., Wever, Th., $adowiak, P., Continen
tal collisions and seismic signature, Geophys.
Behr, H. and T. Heinrichs, Geological interpre- J. Int, 105, 15-23, 1991.
tation of DEKORP 2-$: A deep seismic reflec- Milkereit, B., R. Bittner, R. Meissner, Off-line
tion profile across the Saxothuringian and pos- acquisition of crustal reflection and refrac-
sible implications for the Late Variscan struction data, Geophys. Res. Lett., 1--3, 1161-1164,
tural evolution of Central Europe, TectonoDhv- 1986.
sics, 142, 173-202, 1987. Nelson, K.D., J. Arnow, J. McBride, J. Willemin,
Cook, F., D. Albaugh, L.D. Brown, $. Kaufman, J. J. Huang, L. Zheng, J. Oliver, L. Brown, and $.
Oliver, and R. Hatcher, Thin-skinned tectonics Kaufman, $., New COCORP profiling on the south-
in the crystalline Southern Appalachians: CO- eastern United States. Part I: Late Paleozoic
CORP seismic reflection profiling in the Blue suture and Mesozoic rift basin, Geoloqy, 1-3,
Ridge and Piedmont, Geoloqy, !, 563-567, 1979. 714-718, 1985.
DEKORP Res. Group, First results and preliminary Potter, C., W. Sanford, T. Yoos, E. Prussen, R.
interpretation of deep reflection seismic re- Keach, J. Oliver, $. Kaufman, and L.D. Brown,
cordings along profile DEKORP 2-$, J. Geophys., Deep seismic profiles across the interior of
5--7, 137-163, 1985. the North American Cordillera, Washington and
Goleby, B., C. Wright, C. Collins, and B. Ken- Idaho: Implications for orogenic evolution,
nett, Seismic reflection and refraction profil- Tectonics, , 1007-1025, 1986.
ing across the Arunta block and Ngalia and Ama- $adowiak, P. and Th. Wever, Reflection-diffrac
deus basins, Aust. J. Earth $ci., 3_5,275-294,1988. tion seismic pattern at crustal suture zones,
Goleby, B., B. Kennett, C. Wright, R. Shaw, and Tectonics, , 1495-1513, 1990.
K. Lambeck, Seismic reflection profiling in the $adowiak, P., J. Voss, and R. Meissner, 3D mo-
Proterozoic Arunta block, central Australia: delling of diffractions observed on deep re-
processing for testing models of tectonic evo- flection line DEKORP 2-$, Geophys. Prosp., 3-7,
lution, Tectonophysics, 173, 257-268, 1990. 623-637, 1989.
Hall, J., J. Brewer, D.H. Matthews, and M. War- Sanford, W., C. Potter, and J. Oliver, Detailed
ner, Crustal structure across the Caledonides three-dimensional structure of the deep crust
from the 'WINCH' seismic reflection profile: based on COCORP data in the Cordilleran inte-
Influences on the evolution of the Midland Val- rior, north-central Washington, Geol. $oc. Am.
ley of Scotland, Transac. Royal $oc. Edinburqh, Bull., 100, 60-71, 1988.
7--5, 97-109, 1984. Smithson, $., J. Brewer, $. Kaufman, J. Oliver,
Hutchinson, D., K. Klitgord, M. Lee, and A. Tre- and C. Hurich, Structure of the Laramide Wind
hu, U.S. Geological survey deep seismic reflec- River Uplift, Wyoming, from COCORP deep re-
tion profile across the Gulf of Maine, Geol. flection data and from gravity data, J. Geo-
$oc. Am. Bull., 100, 172-184, 1988. phys. Res., B84, 5955-5972, 1979.
Kampfmann, W., A study of diffraction-like events Taylor, $., McLennan, $., The Continental Crust:
on DEKORP 2-$ by Kirchhoff theory, J. Geophys., Its Composition and Evolution, 312 pp., Black-
6-2, 163-174, 1988. well, Oxford, 1985.
DEEP SEISMIC PROFILG AND CONTINENTAL EVOLUTION
K. D. Nelson
Institute for the Study of the Continents, Snee Hall, Cornell University, Ithaca, NY 14853, U.S.A.
Abstract. Key results from continental deep seismic
reflection and refraction profiling can be incorporated into a
generalized uniformitarian theory of continental evolution
[Nelson, 1991]. This theory provides a framework within
which to view variations in the reflection character of the
continental crust and Moho, variations in crustal thickness
and mean crustal velocity, and the relationship of each of
these parameters to tectonic province. In brief....
Island Arcs and Oceanic Plateaus
Juvenile continental crust is produced in island arcs, such
as the Marianas, and to a lesser extent in oceanic plateaus,
such as the Iceland/Faereos ridge [e.g. Hamilton 1981,
Reymer and Schubert, 1984; Schubert and Sandwell, 1989].
In both settings the crustal formation process is grossly
similar to that which produces "normal" oceanic crust at mid-
ocean ridges, and to a first approximation, both types may be
considered as simply anomalously thick oceanic crust (figure
1A). Like oceanic crust, the bulk composition of newly
formed crust (above Moho) in both island arcs and oceanic
plateaus is basaltic [e.g. Conrad and Kay, 1984; Kay and
Kay, 1990; DeBari and Sleep, 1991]. This statement is
critical to the issue of continental evolution, because the
continents as a whole are considered to have intermediate
bulk composition [e.g. Smithson et al., 1981; Taylor and
McLennan, 1985]. As with oceanic crust, the Moho in both
island arcs and oceanic plateaus lies within the crustal colunto
produced by crystallization of basaltic magma (downward
transition from cumulate gabbroic rock to cumulate ultramafic
rock), whereas the base of the igneous crust lies roughly
25% deeper in the lithosphere and corresponds to a
downward transition from cumulate ultramafic rock to
harzburgite tectonite. This boundary is geophysically cryptic
because the rocks on both sides have essentially the same
velocity and density.
Delamination at Collision Zones
Through the process of collision, island arcs and oceanic
plateaus are amalgamated into continents [e.g. Taylor and
White, 1965; Dewey and Bird, 1970; Hamilton, 1979],
which themselves periodically rift apart and collide together
Geodynamics 22
[e.g. Wilson, 1966]. There is growing evidence that during
"hard" collisions a substantial fraction of the lower crust on
the underthrusting plate becomes detached from the upper
crust and, along with its underlying mantle lithosphere, is
subducted back into the mantle (figure lB). This inference
contrasts with the conventional view that continental crust is
too buoyant to be subducted [e.g. McKenzie, 1969]. It also
provides a mechanism to account for the apparent discrepancy
between the composition of primitive continental crust
produced at island arcs and oceanic plateaus, and that of the
continents as a whole. In essence preferential subduction of
the relatively mafic lower crust at collision zones is a
mechanical refining process that "pushes" the continents
toward more silicic composition. Primary evidence for lower
crustal subduction is afforded by the recent ECORS and NFP
deep reflection profiles in the Alps. These show that the
exterior basement massifs and Pennine zone in the Alps are
detached in the mid-crust from the seismically laminated
lower crust of northern Europe, which is descending as an
apparently coherent unit toward the southeast [Pfiffner et al.,
1988; Nickolas et al., 1990]. As the present cross-sectional
area of the crust beneath the Alps (above Moho), deduced
from wide-angle reflection and refraction data, is
considerably less than would be predicted based on
palinspastic restoration of the surface geology, this lower
crust is, evidently, being subducted back into the mantle
[Butler, 1986; Laubscher, 1988]. Presumably upon reaching
the appropriate depth, it converts to mafic eclogite in much
the same manner that oceanic crust on the downgoing plate at
a marine subduction zone does. While only loosely
constrained, available experimental data suggest that this
transformation should occur at roughly 50 km depth (Wood,
1988), which is roughly the maximum Moho depth observed
beneath the Alps.
Incorporation of mafic eclogite in the lithospheric root
(downgoing slab) beneath a collision zone contributes to the
mots negative buoyancy relative to the surrounding mantle.
When plate convergence across the collision zone ceases this
lithospheric root may be expected to break away and founder
into the deeper mantle (figure 2). This is essentially the
"delamination" process proposed by Bird [1978] and
promoted in various forms by subsequent workers. The
version outlined here differs slightly from earlier
presentations in that the incorporation of maflc eclogite into
the root, rather than simply its temperature contrast with
respect to the surrounding mantle, is considered to contribute
significantly to its negative buoyancy. Because of this the
entire mantle root is likely to delaminate, rather than only a
377
378
SEISMIC PROFILING AND CONTINENTAL EVOLUTION
A.
B.
Island Arc
Collision
C. Young Stable Crust
O.
E.
30 km
Granite/Rhyolite Province
Shield
I ..... / ..... t ,_ / 'c;'"m'uiai';";'l'tr:a'm'';'i';; ..................................................... [ ........................... i ......................................... :l ............ I:::::::"::: 50 km
Fig. 1. Schematic illustrating stages of continental evolution
(see text for discussion). Stipple - mafic lower crust,
irregular black lines in C, D, and E - basaltic dikes and sills,
lower, advectively thickened, thermal boundary conduction
layer [as argued by Houseman et al., 1981; England and
Houseman, 1989]. Prima facie evidence for lithospheric
delamination beneath collision zones is afforded by
teleseismic data that show that such roots presently exist
beneath actively shortening orogens (e.g. Alps), but not
beneath older "collapsed" ones (e.g. Variscan belt of
Northern Europe)[Babuska et al., 1988].
CMB - crust/mantle boundary. (A) modified after Kay and
Kay (1985). (B) modified after Buffer (1986).
When delamination occurs, the overlying orogenic belt
may be expected to undergo rapid terminal uplift accompanied
by extension in the upper crust, and melting in the lower
(development of "collisional granites") [Dewey, 1988].
Uplift results from the replacement of the relatively dense
lithosphere root beneath the orogen by less dense
asthenosphere. Crustal melting results from the combined
effects of advecting hot asthenosphere directly against the
Fig. 2. Schematic illustrating whole-lithosphere delamination
at cessation of collision (see text for discussion). Top panel
illustrates situation before delamination, bottom panel
illustrates situation during delamination. Light stipple - mafic
lower crust, dark stipple - lower crust and down-dip oceanic
crust converted to eclogite, diagonal rule - mantle lithosphere,
heavy black lines - basaltic dikes and sills. Untilled arrows
denote asthenosphere advecting upward to replace foundering
lithosphere slab (black arrow).
base of the crust [Bird, 1978], and the intrusion of basaltic
melt into the lower crust produced by decompression melting
of the rising asthenosphere [McKenzie and Bickle, 1988].
The grossly intermediate composition crust acts as a density
filter for these basaltic liquids, and hence they tend to pond
within the lower crust and along the crust/mantle boundary
[Herzburg et al., 1983].
Assuming that delamination accompanies the cessation of
plate convergence, the orogen will then collapse under its
own weight. When these processes are complete the result is
"young stable crust" (figure 1C). This type of crust is
typically about 30 km thick when its upper surface has been
reduced to sea level, has intermediate bulk composition (has
evidenced by mean crustal velocity), and is pervasively
overprinted by late extensional structures. On seismic
reflection sections it also commonly, though not universally,
exhibits prominent lamination in the lower portion, and a
sharp subhorizontal Moho that crosscuts preexisting terrane
boundaries preserved in the overlying crust. Examples of
NELSON 379
this type of crust which have been examined by deep
reflection profiling include essentially all of that underlying
northern Europe and its adjacent continental shelf, stabilized
during the Caledonian and Vailscan orogenies [e.g. Brewer
and Smythe, 1986; Bois et al., 1987; Meissner et al., 1987],
and the Appalachian Piedmont, Coastal Plain, and adjacent
continental shelf of eastern North America [e.g. Ando et al.,
1983; Nelson et al., 1987; Hutchinson et al., 1988; Keen et
al., 1989]. The Basin and Range province [Allmendinger et
al., 1987] and the interior of the northwestern U.S and
Canadian Cordillera [Potter et al., 1986; Cook et al., 1988]
also fit into this category, though as yet the mantle thermal
anomaly (thin lithosphere) supporting the regional elevation
of these areas, has not yet decayed.
Following this scenario it may be inferred that the lower
crustal lamination observed in young stable regions is due
largely to disseminated mafic sills that were injected into the
lower crust during delamination. As discussed by Warner
[ 1990a], mafic sills in intermediate composition country rock
can easily produce the relatively high reflectivity that has
locally been determined for this lamination. Subhorizontal
stretching fabric in the lower crust may also play a role, but,
based on a variety of arguments is unlikely to be the dominant
contributor [Warner, 1989b; Nelson, 1991]. Moho in these
regions will be marked, locally, by either a ductile high strain
zone or mafic sill separating intermediate composition crust
above from fertile mantle (Lherzolite) below. Flow in the
lower crust, facilitated by partial melting during delamination,
results in this boundary assuming a smooth horizontal attitude
(gravitational equipotental surface).
Cratonization by Basaltic Underplating
Comparison of seismic refraction results from continental
regions throughout the world indicates that there is
considerable variation in crustal thickness (above Moho) and
mean crustal velocity, both within and across crustal age
provinces. However, when comparison is restricted to
regions of near-zero elevation it is apparent that the crust
beneath Precambrian shields and platforms is on average
roughly 40% thicker and 6% faster than crust stabilized in
Late Paleozoic to Recent time [Meissner, 1986; Wever, and
Sadowiak, 1989]. There are basically two ways to view this
observation: Either somewhat denser and therefore
correspondingly thicker crust (on average) was produced
during the Precambrian, or some process has acted long-term
to thicken the crust by the addition of relatively dense
material. While the former, non-uniformitarian, hypothesis
cannot be dismissed outright, there is abundant geologic
evidence to support the latter view.
In particular, virtually all the earth's exposed Archean
shields are riddied with overlapping sets of post-orogenic
basaltic dikes [Halls and Fahrig, 1987]. Individual swarms
commonly cover thousands of km 2, and are separated in
time by tens to hundreds of millions of years. The majority
of the dikes are iron-enriched quartz tholeites, implying that
the liquids had considerable residence time in crustal or
shallow subcrustal magma chambers [Tarney and Weaver,
1987]. By implication, the basaltic dike swarms are the
surface manifestation of episodic underplating (or "inplating")
of the continents by basaltic magmas. Similarly, large
regions of crust consolidated during Proterozoic time are
typified by anorogenic granites and associated rhyolites - the
1.5-1.3 Ga anorogenic Granite/Rhyolite province of southern
380 SEISMIC PROFILING AND CONTINENTAL EVOLUTION
North America, and the 1.8-1.55 Ga post-orogenic granites
within the Svecokarelian domain of the Baltic shield are well
known examples [Van Schmus et al., 1987; Gorbatchev and
Gaal, 1987]. These are generally regarded as having been
produced by crustal anatexis [e.g. Nelson and Depaolo,
1985], which in turn requires the advection of substantial heat
into the crust, most likely via the intrusion of basaltic melt
from the underlying mantle.
In essence then, it is suggested that any piece of
continental crust, given long enough residence time on the
surface of the earth, will be episodically underplated by
basaltic magmas. The first major underplating event
postdating initial stabilization will cause widespread melting
of the grossly intermediate composition lower crust. The
resulting silicic melts migrate upward, producing an
anorogenic granite/rhyolite province (figure 1D). The volume
of silicic melt produced by this process will be approximately
equal to the volume of basaltic melt introduced to the crust
[Kay et al., 1989]. Along with redistributing the light
melting fraction within the preexisting crust, the introduction
of mafic magma at the base of the crust results in the
establishment of a new deeper Moho, and yet deeper
crust/mantle transition. Intrusion of the upper crust by both
mafic dikes and silicic plutons disrupts preexisting lithologic
layering and produces point diffractors interspersed with
seismically transparent regions.
Subsequent underplating events continue this process,
episodically establishing new deeper Mohos and yet deeper
crust/mantle transitions, and further injecting the preexisting
crust with mafic dikes. Because the light melting fraction is
removed from the lower crust during early underplating
events, subsequent ones do not produce significant anatexis.
However, they do produce isostatic uplift due to crustal
thickening, which in turn results in progressive erosion of the
granite/rhyolite province and underlying supracrustal
sequences. By adding basaltic material to the base of the
crust and removing silicic material, via erosion, from the top,
the net effect is to push the crust "back" toward mafic
(denser) bulk composition and, because of the isostatic
requirement, thicken it. The end result of successive episodes
is relatively thick shield-type crust, in which crystalline
basement (largely tonalite gneiss) fiddled with mafic dikes is
exposed over wide areas, and the lower one third to one half
of the crust (above Moho) is composed of a composite
gabbroic underplate exhibiting P-wave velocity greater than 7
km/s (figure 1E). As with oceanic crust, Moho in this setting
likely corresponds to a gradafional downward transition from
cumulate gabbroic rock (mafic granulite) to cumulate
ultramafic rock. Similarly, the petrologic crust/mantle
boundary will lie somewhat deeper in the lithosphere and
correspond to a geophysically cryptic boundary between
ultramafic cumulates and underlying ultramafic tectonite. The
reflective character of the shield crust will be determined
largely by the varying distribution of post-orogenic plutonic
bodies within it; most notably mafic dikes - overlapping point
diffractions, mafic sills - high amplitude planar reflections,
and intermediate to silicic plutons - transparent zones. Only
the very largest orogenic structures are likely to be
discernable through this overprint. Deep reflection profiles
from cratonic regions that, at least arguably, exhibit these
characteristics include the new BABEL profries in the Gulf of
Bothnia [Klemperer et al., 1990], and COCORP and
LITHOPROBE profiles within the interior of North America
[Nelson, 1988; Green et al.,1988].
Implications
The view of continental evolution outlined here has several
general implications:
1) Interpreters of deep seismic reflection data from the
continents should beware of the prevailing "structural
geologist's bias" when viewing these data - this is the
common, though generally unspoken, assumption that most
features observed on deep seismic sections are a
manifestation of crustal strain. While this is undoubtedly true
in many cases, the global reflection data set may contain
equally interesting, and perhaps ultimately more important,
information regarding the character, volume, and distribution
of magmatic additions to the crust. How to extract this
information is a key problem.
2) The relatively high conductivity characteristic of the
lower crust in cratonic regions, is likely produced by a small
proportion of trapped volatiles (brine) carried into the crust
via basaltic underplating. Basaltic magmas typically carry on
the order of 1% volatiles [Charmichael et al., 1974], and
therefore, in principle a 10 km thick basaltic underplate
carries the equivalent of a 100 m thick layer of brine into the
crust. Only a small proportion of this brine, trapped in an
interconnected "wet" zone at the top of the underplate, would
be necessary to produce the observed conductivity [Jones,
1987]. Rocks in exposed granulite terranes in general, and
supposed upthrust deep crustal sections in particular, do not
exhibit this conductivity [e.g. Boland and Ellis, 1989],
because they do not expose the lower crustal underplate [e.g.
Fountain et al., 1990].
3) Substantial underplating of the continents, coupled with
the requirement to maintain near-zero continental freeboard
since the beginning of the Proterozoic, implies substantially
greater, overall, post-Archean continental growth than is
generally thought to be the case [Nelson, 1991]. At present,
the actual shape of the continental growth curve is poorly
constrained largely because of the lack of an agreed upon
estimate of the proportion of continental crust added via
underplating. This is an area where seismic studies may
contribute substantially.
4) Implicit in the view of continental evolution presented
here is the idea that the sub-cratonic lithosphere has a
complex spatial and temporal evolution. This is true
regardless of whether the lithosphere is considered to grow
solely by conductive cooling [e.g. Sclater et al., 1981], or
alternatively, partially or entirely by chemical differentiation
("tectosphere" of Jordan, 1975, 1989). In either case,
delamination and underplating imply local wholesale deletions
of pans of the lithosphere, and mass additions to other pans,
at intervals long after the overlying crust initially formed. It
seems entirely likely that such events have left discernable
imprints within the subcontinental lithosphere in the form of
major high strain zones, melt segregations, metasomatic
fronts, etc. [e.g. Warner, this volume]. Mapping these
features and relating them spatially and temporarily to the
evolution of the overlying crust is an obvious frontier for
deep seismic profiling.
Acknowledgments. The author's research in reflection
seismology and continental structure has been supported
largely by U.S. National Science Foundation grants to the
COCORP Project, most recently EAR-8916129. Institute for
the Study of the Continents contribution 164.
NELSON 381
References
Barazangi, L. D. Brown, J. E. and Oliver, Deep seismic
reflection characteristics of the continental crust, Geology,
15, 304-310, 1987.
Ando, C., F. Cook, J. Oliver, L. Brown, and S. Kaufman,
Crustal geometry of the Appalachian orogen from seismic
reflection studies, in Contributions to the Tectonics and
Geophysics of Mountain Chains, edited by R. Hatcher, H.
Williams,, and I. Zietz, Geological Society of America,
Memoir 158, 83-102, 1983.
Babuska, V., J. Plomerova, and J. Silen'y, Structural
model of the subcrustal lithosphere in central Europe, in
Composition, Structure, and Dynamics of the Lithosphere-
Asthenosphere System, edited by K. Fuchs, and C.
Froidevaux, American Geophysical Union Geodynamics
Series, 1_6, 239-251, 1988.
Bird, P., Initiation of intracontinental subduction in the
Himalaya, Journal of Geophysical Research, 8_3, 4975-
4987, 1978.
Bois, C., B. Damotte, A. Mascle, M. Cazes, A. Him, and
B. Biju-Duval, Operations and main results of the ECORS
project in France, Geophysical Journal of the Royal
Astronomical Society, 89, 279-286, 1987.
Boland, A. V. and R. M. Ellis, Velocity structure of the
Kapuskasing uplift, northern Ontario, from seismic
refraction studies, Journal of Geophysical Research, 94,
7189-7204, 1989.
Brewer, J. and D. Smythe, Deep structure of the foreland to
the Caledonian orogen, NW Scotland: Results of the
BIRPS WINCH, Tectonics, 5, 171-194, 1986.
Butler, R. W. H., Thrust tectonics, deep structure and
crustal subduction in the Alps and Himalayas, Journal of
the Geological Society of London, 143, 857-873, 1986.
Charmichael, I. A. E., F. J. Turner, and J. Verhoogen,
Igneous Petrology, McGraw Hill, New York, 739 p.,
1974.
Conrad, W. K. and R. Kay, Ultramafic and mafic inclusions
from Adak Island: Crystallization history and implications
for the nature of primary magmas and crustal evolution in
the Aleutian arc, Journal of Petrology, 2_5, 88-125, 1984.
Cook, F. A., A. G. Green, P.S. Simony, R. A. Price, R.
R. Parish, B. Milkeriet, P. L. Gordy, R. L. Brown, D.
C. Coffin, and C. Patenaude, Lithoprobe seismic reflection
structure of the southeastern Canadian Cordillera: Initial
results, Tectonics, 7, 157-180, 1988.
DeBari, S. M. and N.H. Sleep, High-Mg low-A1 bulk
composition of the Talkeetna island arc, Alaska:
Implications for primary magmas and the nature of arc
crust, Geological Society of America Bulletin, 103, 37-47,
1991.
Dewey, J. F., Extensional collapse of orogens, Tectonics, 7,
1123-1140, 1988.
Dewey, J. F. and J. M. Bird, Mountain belts and the new
global tectonics, Journal of Geophysical Research, 7_5,
2625-2647, 1970.
England, P. and G. Houseman, Extension during continental
convergence, with application to the Tibetan plateau,
Journal of Geophysical Research, 94, 17561-17579,
1989.
Fountain, D. M., M. H. Salisbury, and J. Percival, Seismic
structure of the continental crust based on rock velocity
measurements form the Kapuskasing uplift, Journal of
Geophysical Research, 95, 1167-1186, 1990.
Green, A. G., B. Milkereit, A. Davidson, C. Spencer, D.R.
Hutchinson, W. F. Cannon, M. W. Lee, W. F. A gena, J.
C. Behrendt, and W. J. Hinze, Crustal structure of the
Grenville front and adjacent terranes, Geology, 16, 788-
792, 1988.
Gorbatschev, R. and G. Baal, The Precambrian history of
the Baltic shield, in Proterozoic Lithospheric Evolution,
edited by A. Kroner, American Geophysical Union
Geodynamics Series, 17,149-160, 1987.
Halls, H. C. and W. F. Fahrig (eds.), Mafic Dyke Swarms,
Geological Society of Canada Special Paper, 34, 503 p.,
1987.
Hamilton, W. B., Crustal evolution by arc magmatism,
Philosophical Transactions of the Royal Society of
London, Series A, 301,279-291, 1981.
Hamilton, W.B., Tectonics of the Indonesian Region, U.S.
Geological Survey Professional Paper 1078, 345 p. 1979.
Herzburg, C. T., W. S. Fyfe, and M. J. Carr, Density
constraints on the formation of the continental Moho and
crust, Contributions to Mineralogy and Petrology, 84, 1-5,
1983.
Houseman, G. A., D. P. McKenzie, and P. Molnar,
Convective instability of a thickened boundary layer and its
relevance for the thermal evolution of continental
convergent belts, Journal of Geophysical Research, 86,
6135-6155, 1981.
Hutchinson, D. R., K. D. Klitgord, M. W. Lee, and A.M.
Trehu, U.S. Geological Survey deep seismic reflection
profile across the Gulf of Maine, Geological Society of
America Bulletin, 100, 172-184, 1988.
Jones, A. G., MT and reflection: An essential combination,
Geophysical Journal of the Royal Astronomical Society,
89, 7-18, 1987.
Jordan, T. H., Some speculations on continental evolution,
in Crust/Mantle Recycling at Convergence Zones, edited
by S. R. Hart and L. Gulen, NATO ASI Series C, 258,
259-276, 1989.
Jordan, T. H., The continental tectosphere, Reviews of
Geophysics, 13, 1-12, 1975.
Kay, R. W. and S. M. Kay, Basaltic composition xenoliths
and the formation, modification and preservation of lower
crust, in Exposed Cross-Sections of the Continental Crust,
edited by M.H. Salisbury and D. M. Fountain, NATO
ASI Series C, 317,401-420, 1990.
Kay, S. M. and R. W.Kay, Role of crystal cumulates and the
oceanic crust in the formation of the Aleutian arc,
Geology, 13, 461-464, 1985.
Kay, S. M., V. A. Ramos, and P. Sruoga, Late Paleozoic to
Jurassic silicic magmatism at the Gondwana margin:
Analogy to the Middle Proterozoic in North America?,
Geology, 17, 324-328, 1989.
Keen, C., C. Peddy, B. de Voogd, and D. Matthews,
Conjugate margins of Canada and Europe: Results from
deep reflection profiling, Geology, 17, 173-176, 1989.
Klemperer, S., Seismic reflection signature of a Proterozoic
shield: Project BABEL [abstract], Annales Geophysicae,
Special Issue, 39-40, 1990.
Laubscher, H., Material balance in Alpine orogeny,
Geological Society of AmericaBulletin,100,1313-1328,1988.
McKenzie, D. P., Speculations on the consequences and
causes of plate motions, Geophysical Journal of the Royal
382 SEISMIC PROFILING AND CONTINENTAL EVOLUTION
McKenzie, D. P. and M. Bickle, The volume and
composition of melt generated by extension of the
lithosphere, Journal of Petrolov, 29, 625-680, 1988.
Meissner, R., The Continental' Cuut, A Geophysical
Approach, International Geophysics Series, 34, Academic
Press, New York, 426 p., 1986.
Meissner, R., T. Wever, and R. Bittner, Results of
DEKORP 2-S and other reflection profiles through the
Variscides, Geophysical Journal of the Royal
Nelson, B. K. and D. J. DePaolo, Rapid production of
continental crust 1.7-1.9 b.y. ago: Nd isotopic evidence
from the basement of the North American midcontinent,
Geological Society of America Bulletin, 96, 746-754,
1985.
Nelson, K. D., A unified view of craton evolution motivated
by recent deep seismic reflection and refraction results,
Geophysical Journal International, 105, 25-35, 1991.
Nelson, K. D., (ed.), The COCORP Atlas, Volume 1,
Institute for the Study of the Continents, Cornell
University, Ithaca, New York, 1988.
Nelson, K. D., J. H. McBride, J. A. Arnow, D. M. Wille,
L. D., Brown, J. E. Oliver, and S. Kaufman, Results of
recent COCORP profiling in the southeastern United
States, Geophysical Journal of the Royal Astronomical
Society, 89, 141-146, 1987.
Nickolas, A., A. Him, R. Nicolich, R. Polino, Lithospheric
wedging in the western Alps inferred from the ECORS-
CROP traverse, Geology, 18, 587-591, 1990.
Pfiffner, O. A., W. Frei, P. Finckh, and P. Valasek, Deep
seismic reflection profiling in the Swiss Alps: Explosion
seismology results for line NFP 20-EAST, Geology, 16,
987-990, 1988.
Potter, C., W. Sanford, T. Yoos, E. Prussin, W. Keach, J.
Oliver, S. Kaufman, and L. Brown, COCORP deep
seismic reflection traverse of the interior of the North
American Cordillera, Washington and Idaho: Implications
for orogenic evolution, Tectonics, 5, 1007-1026, 1986.
Reymer, A. and G. Schubert, Phanerozoic addition rates to
the continental crust and crustal growth, Tectonics, 3, 63-
79, 1984.
Schubert, G., and D. Sandwell, Crustal volumes of the
continents and of oceanic and continental submarine
plateaus, Earth and Planetary_ Science Letters, 9_2, 234-
246, 1989.
Sclater, J. G., B. Parsons, and C. Jaupart, Oceans and
continents: Similarities and differences in the mechanisms
of heat loss, Journal of Geophysical Research, 86, 11535-
11552, 1981.
Smithson, S. B., R. A. Johnson, and Y. K. Wong, Mean
crustal velocity: A critical parameter for interpreting crustal
structure and crustal growth, Earth and Planetary_ .Science
Letters, 52t, 323-332, 1981.
Tarney, J. and Weaver, B. L., Geochemistry and
petrogenesis of Early Proterozoic dyke swarms, in Mafic
Dyke Swarms, edited by H. C. Halls and W. C. Fahrig,
Geological Society of Canada Special Paper, 34, 81-94,
1987.
Taylor, S. R. and S. M. McLennan, The Continental Crust,
Blackwell, London, 312 p., 1985.
Taylor, S. R. and A. J. R. White, Geochemistry of
andesites and the growth of continents, Nature, 208, 271-
273, 1965.
Van Schmus, W. R., M.E. Bickford, and I. Zietz, Early
and Middle Proterozoic provinces in the Central United
States, in Proterozoic Lithospheric Evolution, edited by A.
Kroner, American Geophysical Union Geodynamics
Series, 17, 43-68, 1987.
Warner, M. R., Absolute reflection coefficients from deep
seismic reflections, Tectonophysics, 173, 15-23, 1990a.
Warner, M. R., Basalts, water, or shear zones in the lower
continental crust?, Tectonophysics, 173, 163-174, 1990b.
Wever, Th. and P. Sadowiak, A relationship between crustal
thickness and mean crustal velocity, Tectonophysics, 170,
159-163, 1989.
Wilson, J. T., Did the Atlantic close and then reopen?,
Nature, 211,676-681, 1966.
Wood, B. J., Computation of multiphase equilibrium,
Reviews in Mineralogy, 17, 71-95, 1988.
MULTIPLE SUPPRESSION IN DEEP WATER
Robert 3.3. Hardy ! and Richard W. Hobbs
BIRPS, Bullard Laboratories, Madingley Road, Cambridge CB3 0EZ, U.K.
Abstract. The quality of marine deep seismic strategy to allow cascading of several multiple
data is degraded by shot-generated noise and suppression techniques. The effectiveness of the
multiple reflections. In water greater than 300m strategy has been discussed by Larter et
deep, such noise is often of greater amplitude
than that of'the lower crustal targets. Here we
illustrate the effectiveness of a formalized
strategy for multiple suppression using re-
processed portions of the BIRPS WAM and WIRE
surveys. Several dip-filtering techniques are
successfully used for the suppression of water-
layer multiples. Adaptive predictive
deconvolution and wave-equation modelling are
used to suppress peg-leg multiples, with varying
degrees of success.
Introduction
In a recent tutorial paper, Hardy and Hobbs
[1990] described a strategy for the testing of
al. [1990] for a sedimentary target on the
Antarctic ice shelf. In this paper we demonstrate
the strategy using deep water BIRPS seismic data,
acquired to image the lower crust. The lines
referred to in the subsequent text are shown
dashed on the map of Figure 2.
Example 1: Wire Line 5
Figure 3 shows a shot gather, CMP gather and
brute stack of shot points 101-301 from WIRE Line
5 (location shown in figure 2). A semblance
velocity analysis (not shown) was used to help
identify multiples. The analysis was averaged
over 10 CMP gathers to increase discrimination of
the main events (both multiples and primaries)
from noise. To provide a better estimate of the
multiple suppression techniques on seismic data. multiple energy in the section, Figure 3d shows a
The basic premise was that the usual trial and stack with NM0 correction applied at a multiple
error testing approach was wasteful of time and stacking velocity function. Water-layer multiples
computer resources, and that a formalized target dominate the section to 5s but are partially
orientated approach was required to overcome suppressed by stacking at primary velocity, as
these shortcomings. The strategy is summarized, shown by Figure 3c. A high amplitude gently
in diagrammatic form, by Figure 1. 0val boxes dipping mid-crustal event is seen at 5.5s on
indicate areas of data input and output. Diagonal Figure 3c with a stacking velocity of
boxes indicate the general groups of multiple 5000-7000m/s, followed by a train of peg-leg
suppression techniques, arranged in order of multiple reflections which can be seen to about
increasing cost and complexity, (top right to 10s. These peg-leg multiples mask expected energy
bottom left) such that the cheaper techniques are returned from the lower crust and have such a
tried first. Rectangular boxes on Figure 1 high stacking velocity that they cannot be
indicate specific multiple suppression suppressed by moveout based methods.
techniques. The design of the strategy shows that Suppression of water-layer multiples. Trials
once a technique has been shown to be effective, showed that while f-k filtering suppressed water-
then the results can be re-entered into the layer multiples on the CMP gathers, the result
after stacking was equivalent to that of applying
an inner trace mute. Figure 4a shows a great
1 Amerada Hess Limited, Berkeley Square House, improvement on Figure 3c due to the prestack
Berkeley Square, London W1X deconvolution and inner-trace mute function
applied; a weak sub-horizontal primary emerges at
4.4s after the processing. An autocorrelation
ContinenmlLithosphere: Dp Seisnfic Refltions function derived from 4-8s is appended at the
GeodynamJcs22 base of the sections on Figure 4, and is used to
1991AmeficanGeophysic Union show any remnant periodicity. A few strong
383
384 MULTIPLE SUPPRESSION IN DEEP WATER
CMP STACK
VELOCITY
ANALYSIS
Does stacking
look promising ?
wciglted
stacks
post st(:k
Is there
sufficient
differential
NMO ?
f- k or other
moveout filters
Is there
sufficient
periodicity ?
--predictive
or adaptive
dcconvolution
YES
Is there
a dipping
or complex
seabed ?
CMP domain
wave equation
.(VELOCITt ANALYSIS/
YES
]61 ,
wave equation _
Fig. 1. Strategy for multiple suppression proposed by Hardy and Hobbs [1990].
multiples before 2s have not been suppressed on Adaptive deconvolution. We tested a
Figure 4a because the stacking fold is reduced by deconvolution which adapts the operator to
the first-break mute pattern. varying multiple periods within the data by
Suppression of mid-crustal multiples. In designing a separate filter for each sample on
order to improve signal-to-noise ratio, the the trace (similar to the Griffiths [1977]
section was stacked with'just the wavelet method). The adaptive deconvolution incurs a
compression operator and inner trace mute applied substantial increase in processing time. It was
(Figure 4a) and several post stack predictive expected that a prestack application would be
deconvolution operators were used in an attempt more effective, as this is where the time-varying
to suppress the peg-leg multiples. Trials of nature of multiples is most pronounced; also
operator gap and length showed an operator gap stacking should suppress any artifacts generated
48ms and length 550ms performed best (Figure 4b). by the deconvolution process. In fact, while the
The autocorrelation functions show that multiples prestack adaptive deconvolution performed much
have been suppressed, but there is still better than a conventional statistical
substantial remnant energy confusing any possible deconvolution, remnant energy was still present
primaries from the lower crust. Post-stack and the prestack application was too expensive to
deconvolution has failed due to incorrect be justified. When applied post-stack, the
multiple amplitudes and wavelet distortion by the peg-leg multiples were severely attenuated, as
stacking process; some form of adaptive shown in Figure 4c. Unfortunately the suspected
deconvolution is required. primary at 4.4s was also partially attenuated and
HARDY AND HOBBS 385
NSDP-84
WINCH
BIRPS LINES 1981 - 1989
Fig. 2. Location map showing B1RPS deep seismic lines (WAM, and WIRE 5) are used in this
data as of January 1989. Portions of the dashed paper. * = shotpoint numbers.
SHOT
0 I I
2
P
(a)
CMP STACK STACK
I) @ 1500 m/s 0
, . 8
(b) (c) (d)
CMP
Fig. 3. Preliminary analysis panel for WIRE Line 5. P is location of primary at 5.5s.
(a) Shot gather; (b) CMP gather; (c) Brute stack; (d) Stack at 1500m/s to emphasize
multiples.
0 z .0
Fig. 4. Wire Line 5 stacks to compare deconvolution operators. An autocorrelation
function derived from a 4-8s time window is appended at the base. (a) Same panel as in
Figure 3c after wavelet compression deconvolution and inner trace mute; (b) As (a) but
also with post-stack deconvolution operator, lag 48ms length 550ms; (c) As (a) but with
adaptive deconvolution applied post-stack.
there is little evidence of any primary energy serious problem in the interpretation of the
after the adaptive deconvolution. It is oceanic continental crust transition section of
difficult, in this instance, to decide whether the WAM profile. The main area chosen to test
the predictive or adaptive deconvolution produces multiple suppression is in about 1500m or 2s
a more pleasing result. water depth and is shown in Figure 6a. Possible
Summary. For the WIRE Line 5 example, primary reflections are seen from the lower crust
prestack wavelet compression, weighted stack and but are lost to the west amidst interference
post-stack long-period multiple deconvolution
substantially improved the image of the upper
crust (sequence 4 followed by sequence i on
Figure 1). The image of the lower crust was
improved, but the presence of remnant peg-leg
multiple energy and high noise levels prevented
clear imaging of any lower crustal reflectivity.
Example 3: WAM
As reported by Hardy and Warner [1990]
long-period multiples were considered to be a
from multiples.
Karhunen-Love filtering. Using an f-k filter
for multiple suppression [Rhu, 1982] is often not
completely satisfactory because it degrades
primaries, both during application and because an
inner trace mute is required [Hardy et al.,
1990]. The Karhunen-Love (KL) transform can be
used to isolate coherent events in a suite of
traces. An NMO correction for water-layer and
peg-leg multiples was applied to the CM? gather
shown in Figure 5a. The NMO corrected data were
KL transformed and several of the KL principal
components (the coherent energy) were muted off
HARDY AND HOBBS 387
lO
(a) (b) (c) (d) (e) (f)
Fig. 5. Comparison of multiple suppression techniques on WAM CMP gather. (a) CMP with
t 2 gain correction applied. WL = water layer multiple, PL = peg-leg multiple, IPL =
Interlayer peg-leg multiple. (b) The CMP transformed to the KL domain. Solid line
indicates muting function chosen to produce the multiple suppressed gather (d); (c)
Multiples suppressed by the KL transform with mute technique; (d) CMP 1120 after
multiple suppression by KL transform. (e) Multiple model produced by wavefield
extrapolation. (f) result of adaptively subtracting (e) from (a).
before reconstruction, as described by Jones and result of component filtering is a near perfect
Levy [1987]. The multiple NM0 function was then suppression of known multiples; the intervening
removed before further processing. An effective regions e.g. between 5 and 6s have been
multiple suppression was achieved by muting 10 completely untouched by the filter, preserving
principal components, but this also destroyed low any possible primaries. As a result, the final
frequency primary information. result in Figure 5d shows several amplitude
The KL domain shown in Figure 5b, shows that discontinuities where the multiples have been
the NM0 corrected multiple energy is not removed. A solution is to run the KL transform on
compressed into the first few principal tapered overlapping windows which, as suggested
components because of an imperfect NM0 correction by Jones and Levy [1987], is recommended in order
and wavelet changes with offset. to focus the covariance matrix on those coherent
In an adaption of the method, we obtained a components within the window.
better multiple suppression whilst minimizing A particularly useful method of comparing the
damage to primaries. This was achieved by effectiveness of multiple suppression techniques
selectively muting out components identified as is to compare the amount of energy a particular
multiples in the KL domain. The mute pattern technique has removed. For the KL method, this is
chosen is shown by a solid line on Figure 5b. The shown in Figure 5c; Figure 5c summed to Figure 5d
resulting gather is shown in Figure 5d. The would be identical to Figure 5a.
West F__a_st
5 Km
.-.--:.' .:--- .... -J.x::'.: .~.-?:;...2. _-.-.-.---- '
: .. --_.--_ _ _.
_. -. . _ -.. ._..::..- ._ ....
_.-'.:.? -- : :-
11 11
4150 4250 4 0 (b) (c)
( a ) Sh otpoi n t
Fi. 6. Comparison of multiple suppression on k multiple suppression; (c) after wave-equation
stacked sections. (a) Stacked panel; (b) after f- multiple suppression
Wave-equation suppression. Methods of wave-
equation based multiple suppression have been
demonstrated by Wiins [1988]. We used an f-k
domain operator to extrapolate the CMP Eathers
through an extra pass of the water-layer, to
produce a model of the water-layer and pe-le
multiples shown in Fiure 5e for the WAM CMP
Eather. The effects of the water bottom
shows that the results are excellent for the
water-layer and peg-leg multiples. Some residual
multiples between 4-5s and 6-7s are double and
triple bounce multiples from the near-surface
reflector which are not predicted by the
wavefield extrapolation used. The third multiple
period between 8.5-9.5s has been efficiently
suppressed. In general, the gather is
reflectivity were not accounted for in this characterised by hih frequency noise and some
extrapolation, so there is considerable amplitude residual enery on the far offset traces. Both of
discrepancy between the data and the model these can be suppressed by the subsequent
(Fiure 5a and Fiure 5e) especially at the third application of an f-k filter, however, for this
multiple period. The low amplitude artifacts on data example stackin effectively suppressed this
the far offset traces at 4s are due to spatial remnant enery.
aliasin E and can be reduced by extrapolatin Stacked sections. The precedin discussion
interleaved CMP Eathers and interpolatin the has described in detail the effects of multiple
missin near offset traces. suppression on CMP Eathers. Unless the aim is to
In the second phase of the technique, the study events in the CMP domain, e.. for
model Eathers are adaptively subtracted from the amplitude-versus-offset analysis then the
input Eathers, the result of which is shown in artifacts enerated in that domain may not be
Fiure 5f. In this two phase least-squares important on stacked sections. Fiure 6 compares
process, timin errors in the model, due to the stack with no multiple suppression applied
incorrect water bottom picks, are first corrected (Fiure 6a) with the best f-k stacked section
and then wavelet shapin is applied. Fiure 5f (Fiure 6b) and the wave-equation processed
HARDY AND HOBBS 389
section (Figure 6c). All sections have been (SKS) running on the Bullard Laboratories Convex
processed with a post-stack deconvolution for C120. The wave-equation processing was produced
wavelet compression, a bandpass filter and 1500ms using the Geovecteur software by CGG, and
window automatic gain control. For this deep redisplayed with SKS. The KL transforms were
water example, the data has not been subjected to performed using the LUCID software from Logica,
multiple suppression processing before the first running on a SUN SPARC. We are grateful to Ian
multiple, to preserve the seabed. The f-k section Jones and the staff of CGG in London for the use
is the result of extensive testing of inner and
outer trace muting functions. The outer trace
mute was required to suppress wrap-around
artifacts generated by the f-k technique. To the
east of the section, the primaries are clearly
seen. It is not possible, however, to
unambiguously extrapolate them laterally across
the section because there is no certainty that
the energy to the west is not remnant multiple
energy.
Summary. For this WAM data example, f-k and
wave-equation based techniques have been used to
successfully suppress severe long-period
multiple contamination. However, severe
interpretational difficulties arise in
of their software. We thank many colleges from
industry and academia who have discussed multiple
suppression techniques with us. Robert Hardy
thanks Shell UK, Cambridge Department of Earth
Sciences, Darwin College and Amerada Hess Limited
for financial support during the above study.
Cambridge Earth Sciences contribution number
1950.
References
Griffiths, L.J., Smolka, F.R and Trembly, L.D.
1977. Adaptive deconvolution: a new
technique for processing time-varying
seismic data. Geophysics 42, 742-759.
interpreting the deeper part of the section where Hardy, R.J.J., Warner, M.R. and Hobbs, R.W.
the signal-to-noise ratio is low, because remnant
multiple energy is confused with potential
primary energy.
Conclusions
This paper has presented a strategy for applying
conventional long-period multiple suppression
techniques. Where the location of primary events
was uncertain, due to poor signal-to-noise ratio,
the remnant multiple energy left behind after
multiple suppression may be of comparable
amplitude to primaries and makes a final
interpretation ambiguous, because conventional
techniques fail to discriminate between primary,
multiple and remnant multiple energy. In such
data, the problem of remnant multiple energy
makes multiple suppression unlikely to achieve
positive results. Rather, some scheme of multiple
identification is required, such as those
proposed by Hardy et al. [1989, 1990]. The
problem of remnant multiple energy is
particularly pertinent to deep seismic data,
because it is the 6-12s of data which are most
affected for deep water data examples; however,
the problem is present on all data to which
multiple suppression techniques have been
applied.
Acknowledgements. The seismic sections were
produced using the GECO Seismic Kernel System
1989. Labeling long-period multiple
reflections. Geophysics 54, 122-126.
Hardy, R.J.J., Hobbs, R.W., Warner, M.R. and
Jones, I.F. 1990. The reliability of
multiple suppression. First Break 8, 297-
304.
Hardy, R.3.3. and Hobbs, R.W. 1990. A strategy
for long period multiple suppression. First
Break, 9, 139-144.
Hardy, R.3.3. and Warner, M.R. 1990. Problems
of imaging the crust in deep water.
Tectonophysics 173, 141-143.
Jones, I.F. and Levy, S. 1987. Signal-to-noise
ratio enhancement in multichannel seismic
data via the Karhunen-Loeve transform.
Geophysical Prospecting 35, 12-32.
Larter, R.D., Cunningham, A. P. and Hardy,
R.3.J. 1990. The sea-floor multiple problem
in multichannel seismic reflection data
acquired on the Antarctic continental
shelf: its causes and treatment. 1990.
Interna t. workshop on Antarctic offshore
seismic stratigraphy (ANTOSTRAT): Overview
and extended abstracts: USGS open file
report, pp90-309.
Ryu, 3.V. 1982. Decomposition (DECOM) approach
applied to wave field analysis with seismic
reflection records. Geophysics 47, 869-883.
Wiggins, 3.W. 1988. Attenuation of complex
water-bottom multiples by wave-equation
based prediction and subtraction. Geophysics
53, 1527-1539.
A COMPLETE WAVEFORM INVERSION AND ITS APPLICATION TO
ECORS DATA
Satish C. Singh
Institut de Physique du Globe de Paris, 4 Place Jussieu, F-75252 Paris Cedex 05, France
Michel Dietrich
Laboratoire de Gdophysique Interne et Tectonophysique, Universit6 Joseph Fourier
B.P. 53X, 38041 Grenoble Cedex, France
Fran(;ois Chapel
Elf Aquitaine, 64018 Pau Cedex, France
Abstract. A normal procedure in seismic waveform
inversion is to minimize the misfit between observed and
calculated wavefields. As this misfit function is highly
non-linear, the only way to be sure of arriving at the
global solution is to use a Monte-Carlo search. This is
feasible if the number of unknown parameters is not too
large. If it is large, as is the case with seismic inverse
problems, it is extremely expensive to use the Monte-
Carlo search to optimize for a single misfit function over
the whole model space. However, if one uses a number
of misfit functions depending upon the sensitivity of the
model parameters on the data, one can implement the
Monte-Carlo method efficiently. Fortunately, this is pos-
sible as the different wavelengths of the model parameters
influence the different parts of the data. Here, we propose
a three-step strategy for seismic waveform inversion. In
the first step, we explore the complete model space for
a certain number of parameters using the zeroth-order
relationship between model parameters and data. The
model so obtained is in the region of the global minimum,
but may be far from the global solution. We then de-
fine upper and lower bounds on this solution which are
obtained from the uncertainty in the solution. Then, for
a certain number of parameters, which should represent
the gross features of the model, we explore the bounded
model space using a Monte-Carlo search. The data are
related to the model parameters with a first-order ap-
proximation. Finally, we use a local non-linear waveform
inversion to arrive at the final solution. The data are
now related to model parameters with a second-order ap-
Presently at BIRPS, Bullard Labs, University of Cam-
bridge, Cambridge CB3 0EZ, U.K.
Geodynamics 22
proximation. To demonstrate the theory, we treat a one-
dimensional problem in the (r, p) domain. The numerical
example of synthetic data supports the theory. Finally,
the last step of the inversion scheme is applied to marine
reflection data collected by ECORS group from the Bay
of Biscay, France. Inversion result predicts a laminated
lower crust with velocity contrast varying from 50 m/s to
500 m/s of 100 m to 500 m thick layers.
Introduction
It has been recognized that the inversion of seismic
data can be treated in two steps: (1) estimation of long
wavelength of the velocity and (2) estimation of short
wavelength of impedance (or velocity) contrast [Jannane
et al., 1989]. In the sense of classical seismic data pro-
cessing, these two steps correspond to the estimation of
velocity using velocity analysis, and migration of seismic
data using these velocities. The literature is rich with pa-
pers on migration and velocity analysis [e.g., Berkhout,
1980; Hubral and Kreys, 1982]. There has been a great
deal of work on inverting the short wavelength parame-
ters [e.g., Cao et al., 1990]. Unfortunately, we have not
achieved much more than the results of migration as the
solution is very sensitive to the background velocity (long
wavelengths). For the long wavelengths, there are two
main approaches: coherency analysis and focusing analy-
sis. Theoretically, both approaches are similar, but they
differ in practice. The idea is to maximize the coherent
signals. The focusing analysis is more sensitive to the
noise in the data, but could be applied to the complete
data space. The coherency method is more robust, but
one needs vertical travel times for a certain number of
layers. Landa et al. [1989] have used the coherency ap-
proach in the time-distance domain. In this approach,
velocity of a particular layer is estimated by knowing the
velocities of the previous layers. Therefore, this is a kind
of layer-stripping method where the velocity of the first
layer is computed first and then used to get the velocity
391
392 WAVEFORM INVERSION
of the second layer and so on. Any error in the veloc-
ity estimation in the previous layers will propagate as we
proceed downwards. Hence, the errors are propagative in
nature and the solution for the last layer will be badly af-
fected. To avoid the propagation of errors with depth, the
velocity could be searched simultaneously for all the lay-
ers. A global solution to the long wavelengths of velocity
could be obtained using a Monte-Carlo search. The short
wavelengths of velocity could then be obtained using a
non-linear waveform inversion.
Here we propose a strategy, which we call complete
waveform inversion, where both the long, as well as the
short, wavelengths of the velocity are estimated in the
same framework. Our idea is to divide the misfit function
into three parts. Each part of the misfit function depends
on a different approximation used in relating data with
model parameters and a search is made in the complete
model space. In the next section, we will give the theory
for one-dimensional media where the background veloc-
ity is estimated from the Monte-Carlo inversion and the
short wavelength variation in velocity is obtained using
waveform inversion of the gradient type. The mathe-
matical details about the complete theory will be pub-
lished elsewhere. A synthetic example is presented in
the last section. Finally, we apply the theory to Etude
de la Crofite Continentale et Oc(anique par Rdfiexion et
Rfraction Sismique (ECORS) data from the Bay of Bis-
cay (Aquitaine Basin), France.
Complete Waveform Inversion
When the lateral variation in model parameters is not
too strong, one can assume the medium is locally 1-D. For
a horizontally stratified medium, the reflection events in
the (t, x) domain form coherent events along a curve. For
a single layer, this curve is a hyperbola. For multilayered
cases, it is not a hyperbola. However, for small offsets,
it is commonly assumed that the hyperbolic assumption
is an acceptable approximation. On the contrary, in the
(r,p) domain, the coherent reflection events align along
an ellipse, and the difference of the delay time between
two consecutive reflectors is also an ellipse. This valuable
property can be used efficiently to implement a velocity
inversion which will be accurate if the subspace is truly
a horizontally stratified medium. The travel time in the
(r, p) domain can be written as
f0 z (1
r( z, p) - d( v 2 (() _ p2, ( 1 )
where v(z) is velocity at depth z and p the horizontal
slowness. Since the data are in the time domain, we can
replace the depth parameter with the vertical travel time
TO
r(ro,p) - dt x/'l - v2(t)p , (2)
where r0 = r(p = 0) is the vertical travel time. Equation
(2) is simple and fast to compute and could be used to
perform velocity inversion. The complete wavefield can
be computed using the reflectivity method [Kenneff and
Kerry, 1979]. As it is simple to distinguish the distinct
coherent signal along an ellipse, equation (2) can be used
to invert for the distinct velocities as a function of time
and then waveform inversion can be used for the high fre-
quency part of velocities from complete wavefields. Thus
we can divide the model parameters into two parts: long
wavelength and short wavelength. As the long wavelength
part is related to the travel time with the zeroth and
the first-order approximations, we parameterize the long
wavelengths with a sparse grid. The short wavelengths
relating to the waveform are parameterized by a dense
grid. The choice of sparse grid depends upon the number
of strong arrivals in the seismograms. Thus the inversion
can be performed in three steps. The long wavelength
part is searched in the complete model space in the first
two steps. In the third step, we use a non-linear iterative
algorithm to invert for the short wavelengths.
Step 1
The first step is to pick up a certain number of re-
flections and their vertical travel times. Each reflection
corresponds to a layer in the time domain. We assume
that each is a single layer and compute energy along an
ellipse as a function of velocity. The energy function is
defined as
E(v(ro)) /.+a. f
= dt { dp
dr--Ar
u(r0V/1 - v2(ro)p 2 - ro + t,p)} 2, (3)
where u(r,p) is the wavefield, r is estimated from equa-
tion (2), and 2At is the length of the time window. Usu-
ally the length of the source wavelet is taken as the win-
dow length. We explore all possible velocities which could
maximize the energy function. The velocity correspond-
ing to the maximum energy is the RMS velocity. For
each To, we obtain a root mean square (RMS) velocity.
The maximum energy function is computed for each layer.
This function is used to normalize the energy function in
each layer. In this way, it is ensured that all the layers
inverted in the next step have equal weights, otherwise
the layers with the maximum energy will dominate the
solution. From the RMS velocity thus obtained, one can
calculate the interval velocity. Even though these RMS
velocities are not accurate they are close enough to the
true solution. Knowing the RMS velocity, we define lower
and upper bounds on the model parameters. The bounds
can be estimated using uncertainties in the interval ve-
locity computation and signal to noise ratio. The total
uncertainty is defined as
8V -- 8VRM$ n t- 8Vnoise.
(4)
A more precise meaning to equation (4) will be given else-
where. Here, it is simply assumed that the true model lies
in the bounded model space.
Step 2
Step 2 consists of exploring the bounded model space
for all the unknown parameters simultaneously. We start
with a random velocity in the space of possible solutions
and calculate the energy for each layer. Then, each of
these energies is normahzed by the corresponding max-
imum energy computed in the last step. The final cost
function is the sum of the normalized energies. The nor-
malization ensures that all the parameters have equal
weights on the cost function. However, one can change
the weight of a particular parameter depending upon the
a priori information. Our aim is to search for a model
which maximizes the cost function. If the initial model is
close to the true model, a local optimization technique is
used. If it is far from the true one, one should use a global
search. We perform a Monte-Carlo search to find the best
solution. This is achieved by an exhaustive search in the
bounded model space. We begin with a certain number
of random models instead of a single model and in each
iteration one parameter is randomly modified. We accept
the model if the energy function is greater than the previ-
ous ones. After a certain number of iterations, it is found
that we have practically filtered out all the possible lo-
cal minima except for a few. Using a number of models,
different minima are localized in the model space. Then
we perform a local optimization (simplex method) start-
ing from the solution of a random search to arrive at the
global minimum. One can use the Monte-Carlo search
to find the global minimum, but we do not see the point
in searching the complete bounded model space once we
have targeted all the possible minima. If the model is
parameterized properly, the final solution is the global
solution for the long wavelengths. If this velocity is used
as a reference velocity in the waveform inversion, we can
then find the complete solution of the model space.
SINGH ET AL. 393
Step 3
A nonlinear iterative waveform inversion is performed
in the (r,p) domain. It is based on a generalized least-
squares formalism. It utilizes the fast convergence prop-
erties of the conjugate gradient algorithm and achieves
computational efficiency through the use of analytical
solutions for the calculation of the reference and per-
turbational wavefields. The method is implemented in
the frequency-wave slowness domain, and can be read-
ily adapted to various source-receiver configurations. A
detailed description of the method is given by Kormendi
and Dietrich [1990]. The inversion scheme yields satisfac-
tory results as long as the "correct" velocity trends are
introduced in the starting model. If the parameters in the
last step are chosen properly, the smoothed version of the
result should be the "correct' velocity trend for waveform
inversion.
Synthetic Examples
To show the robustness of the method, we consider a
complex marine model. The P-wave velocity profile is de-
rived from an actual sonic log which has been discretized
with 303 layers having a thickness of 4 m. The density and
shear wave velocity distributions are deduced from com-
pressional velocity through empirical relationships [Gard-
ner et al., 1974]. Thirty seismograms for slowness ranging
from 0.0 s/km to 0.3 s/km are calculated using a reflec-
tivity method [Kennett and Kerry, 1979]. Figure I shows
the synthetic pressure field obtained in the (r, p) domain.
A Ricker wavelet with dominant frequency of 30 Hz is
used as a source wavelet.
I :3 1./1 1.(i f.8 2
Fig. 1. Synthetic seismograms for the velocity model shown in Figure 4 (thin line) in the (r, p)
domain.
The model for P-wave velocity is shown in Figure 4
Fthin line). To retrieve this model from the synthetic data
igure 1), the first step is to identify the dominant re-
ections in Figure 1. We pick nine reflections and their
vertical travel times. We assume that nine nodes are suf-
ficient to represent the long wavelength of the velocity.
Special care has to be taken to avoid multiples which
should be identified by preliminary analysis. We define
a window of 10 sampling points at the vertical travel time
and compute the RMS velocity using the first step. We
also calculate the RMS energy for each reflector. From
these velocities, we define upper and lower bounds on the
model parameters. In this case, we simply assume that
the error in the inverted velocity increases linearly with
depth. We choose 10 random models (Figure 2) in the
bounded model space, and perform 500 iterations of the
Monte-Carlo algorithm. We find that the solution is con-
verging but is still far from the true solution. Starting
from these solutions we perform a local optimization with
a simplex algorithm, to converge to the final solution (Fig-
ure 3). The sharp jumps shown in Figures 2 and 3 are the
node points. We see that all the starting models have
not converged to the same solution, but all of them are
in the same region except for the fifth layer, which is the
thinnest. As the layer is thin, the probability of interfer-
ence with the other primary signals or multiples and/or
shear waves is maximum and leads to strong uncertainty
in the solution. To avoid this, one should exclude this
layer from the unknown parameters. We chose to assume
that the solution with the highest energy function is best.
(Figure 4, dashed line).
Now the final step is to perform a waveform inversion
starting from the solution of the last step. To make the
algorithm robust and to filter out the high frequency con-
tent of the velocity, we smooth the resulting velocity with
a eight-point running average filter. As the model derived
in the previous step does not contain any information
about the velocity after 1.2 km, we simply assume that the
velocity after 1.2 km is the model velocity. Since there are
no reflections from the water layer, we fix the velocity in
the water layer at 1.5 km/s and invert for 303 parameters
(P-wave velocity). The program is designed to invert for
P-wave and S-wave velocity and density but we assume
that the S-wave velocity and density are known, and we
Final Velocities
Time (s)
Fig. 3. Result of inversion using 500 iterations of a
Monte-carlo search and then simplex method starting
with velocities in Figure 2.
only invert for P-wave velocity. The starting model for
the waveform inversion is shown in Figure 4 (dashed line)
along with the corresponding synthetic seismogram for
p = 0.0 s/km is shown in Figure 5a. We perform five
separate runs with 10 iterations in each run. First, we
take the first 10 traces and perform the inversion. A syn-
thetic seismogram from the inverted model is shown in
Figure 5b. The cross-correlation coefficient between the
data and synthetic is 0.85. It is clear that we have gained
some high frequency in the data. All the parameters are
kept the same in the second run except for a precondition-
ing which is applied to enhance the effect of the deeper
part of the model parameters on the misfit function. The
synthetic seismogram for p = 0.0 s/kin after the inversion
is shown in Figure 5c. In the third run, we take the first
20 traces and the resulting synthetic for p = 0.0 s/km is
shown in Figure 5d. For comparison, the data for p = 0.0
s/kin are shown in Figure 5e. One can see that the true
seismogram matches very well with the synthetic. In run
4, we used 15 traces beginning from trace number 6 to
20, and in run 5, we used 10 traces from trace number
11 to 20. The resulting velocity is shown in Figure 6
(solid line). The cross-correlation between the data and
Initial Velocities
Time (s)
Fig. 2. 10 randomly generated starting velocities for in-
version of the data in Figure 1.
Inverted and Real Model
' ' i
Depth (kin)
Fig. 4. The best solution (dashed) of Figure 3 and true
velocity profile (thin line).
.. ct
Fig. 5. Synthetic seismograms at various stages of inver-
sion and the original data for p = 0.0. (a) Starting
model ho,vn in Figure 4 (dashed), (b) after 10 iter-
ations, (c) after 20 iterations, (d) fte 3O iterations
and (e) original data.
the synthetic is 0.9973 for p = 0.11 s/km. The position
and amphtude of the velocity profile have been retrieved
accurately. Some of the high frequency part could not be
retrieved because of the bandwidth of the source wavelet.
We also see that the velocity between 0.54 km and 0.92
km does not match very well with the model velocity al-
though all the peaks have been positioned properly. If we
analyze the model (real) carefully, we find that there is a
kind of smooth gradient in the velocity profile (between
0.54 km to 0.92 km) which could neither be retrieved by
the coherency inversion (at the second step) nor by the
waveform inversion (at the third step). In the second
step, we assumed that the velocity is constant in the sec-
ond layer and in the third step we added high frequencies
on top of it and ignored the velocity gradient. One could
think of inverting for a gradient in the second step which
uses the travel time information, particularly when the
Run 5
Depth (km)
Fig. 6. Result of inversion (thick line) after the 5th run
(50 iterations), true model (thin hne) and starting
model (dashed).
SINGH ET AL. 395
layer is thick. However, Farra and Madariaga [1988] show
that the vertical velocity variations within a layer can not
be resolved from the travel times of a reflection profile.
This leaves us only with the choice of using the waveform
inversion of post-critical data.
Inversion Of ECORS Data From The Bay Of Biscay
The results presented in this section are an applica-
tion of the third step of the inversion method to a com-
mon shot gather recorded in the Bay of Biscay for the
French ECORS Program in 1984 (Figure 7). Analysis of
the ECORS profiles acquired along the Aquitaine shelf in
the Bay of Biscay showed that the lower crust is highly
reflective on each side of the Parentis basin. The lower
crust displays a series of strong subhorizontal reflectors.
This is in sharp contrast to the upper crust which is rel-
atively non-reflecting [Pinet et al., 1987]. These observa-
tions confirmed those made by the BIRPS, DEKORP, and
ECORS groups in recent years in Variscan and Caledo-
nian (western) Europe [see, e.g., Matthews and Cheadle,
1986; Hobbs et al., 1987; Meissner et al., 1987; Bois et
al., 1987], and by the COCORP group in the Basin and
Range province of the USA [Klemperer et al., 1986].
Early studies by Fuchs [1969] and Meissner [1973] have
shown that deep seismic reflections can be satisfactorily
interpreted as the seismic response of stratified zones with
alternated high and low velocity layers. The discontin-
uous patterns of deep crustal reflections have been ex-
plained either by near-surface generated noise, by un-
dulating interfaces or by lenticular inclusions producing
focusing and defocusing effects or tunneling phenomena
[Fuchs and Schultz, 1976; Gibson and Levander, 1987;
Andrier, 1989]. However, the precise nature and origin
of lower crustal reflectors have not yet been firmly es-
tablished. Seismic measurements alone cannot provide a
definite answer to these questions; nevertheless, they rep-
resent the main technique to investigate the deep crustal
structure.
The chief objective of the seismic method is to charac-
terize more completely crustal reflectors in terms of veloc-
ity contrast and shape. This has recently been attempted
by using a trial-and-error forward modeling approach to
match synthetic seismograms with deep seismic sound-
ing data [Sandmeier and Wenzel, 1986; Paul and Nicollin,
1989]. Calculations performed by Andtier [1989] in the
frequency range 5 Hz to 50 Hz showed that the Bay of Bis-
cay (ECORS) near-vertical reflection data are compatible
with a laminated structure whose elementary layers have
a characteristic thickness of 150 m with velocity contrasts
of up to 500 m/s. On the other hand, classical (r, p) ve-
locity analysis of the Expanding Spread Profiles recorded
for the ECORS Bay of Biscay Program failed to give any
detailed information on the velocity structure in the lower
crust, mainly because of the relatively low frequency con-
tent of the wide-angle seismic data. The kinematic inver-
sions in the (r, p) domain merely allowed definition of a
20 km thick, almost homogeneous zone, where the P-wave
velocity varies between 6.5 km/s and 7.0 km/s [Marillier
et al., 1988].
The method employed here to infer the P-wave veloc-
ity structure in the lower crust is a full waveform inver-
sion of the data which involves a systematic search for the
best fitting model in terms of flat layers. Because of the
resolution problem mentioned above with large aperture
seismic data, the inversion algorithm has been applied to
a common shot gather of the mainly near-vertical seismic
reflection profile shown in Figure 7. The GECO 65 shot
displayed in Figure 8 has been chosen for its good signal-
to-noise ratio and high frequency content which allow a
clear definition of the lower crustal reflections between 8
s and 11 s two-way travel time.
Several assumptions were made to apply the algorithm
outlined to the GECO 65 common shot gather. It was
basically assumed that the lower crustal reflections can
be modeled as the response of a horizontally stratified
medium to plane waves propagating only in the vertical
direction (p = 0.0 s/km slowness component). This ap-
proximation is justified by the fact that the spherical di-
vergence of the wavefronts at depths of 20 km to 30 km
is negligible considering the high velocities in the lower
crust (7 km/s) and the small aperture of the receiver ar-
ray (3 km). In addition, the inversion was limited to lower
crustal reflections between 6 s and 12 s by assuming a per-
fectly transparent upper crust. We finally assumed that
AtMORicAN
,LS
A ROCHELLE
lBS1
LANDES
HIGH
deep reflection profile
CAN, expanding spread profile
, common midpoint
drilling well
Fig. 7. Location of the seismic profiles recorded dur-
ing the ECORS-Gascogne 84 cruise. The arrow to the
north of the main near-vertical seismic reflection profile
indicates the position of shot GECO 65 which was used
in the inversion.
500
Offset (m)
1000 1500 2000 2500 3000
Fig. 8. GECO 65 common shot gather. A dynamic equal-
ization has been applied to the data for their display.
Inversion was performed with the true amplitude data.
the density is constant and equal to 2.8 g/cm 3 in the lower
crust in order to estimate the P-wave velocity rather than
the P-wave impedance.
A "horizontal" partial stacking of the 120 original
seismograms by groups of 11 traces was used to improve
the coherency of the reflections. The resulting traces, D1
to D10, are presented in Figure 9a along with the trace
labeled D which corresponds to the summation ofal1120
seismograms. The traces presented in Figure 9a clearly
display the energetic reflections of the lower crust, but
also some strong lateral variations in amplitude and phase
of the signals.
The choice of the source time function is one of the
most critical issues in waveform inversions. In the ab-
sence of any good estimate of the source wavelet, we used
a zero-phase Ricker wavelet with a prevailing frequency
of 20 Hz. This procedure entails errors in the phase of
the signal, and also neglects the reverberations in the su-
perficial water and sedimentary layers (which are not ac-
counted for in our crustal model). For these reasons, the
primary purpose of the present inversion was to obtain
a correct estimate of the overall velocity log rather than
SINGH ET AL. 397
(a)
{ 8
o
8
o
Dll
D1o
D9
D8
D7
D6
Ds
D4
D3
D2
D
8 10 12
Time (s)
(b)
o 0 10 20 30 40 50 60 70
Frequency (Hz)
Fig. 9. (a) Partial stacking of the 120 traces of shot
GECO 65. Trace Dll corresponds to the summation
of all seismograms. (b) Amplitude spectrum of trace
Dll (continuous line), and amplitude spectrum of the
Ricker wavelet used in the inversion (dashed line).
to infer the exact position of the reflectors. In practice,
the discrepancy between the actual source wavelet and
the Ricker signal was reduced by weighting the ampli-
tude spectrum of the data by the spectrum of the Ricker
wavelet, as shown on Figure 9b. This filtering was aimed
at concentrating the energy of the data in the frequency
range 10-35 Hz. Another benefit of this processing was to
attenuate the high peak of energy around 6.5 Hz on the
frequency spectrum, which is associated with multiple re-
flections in a 60-m thick water column.
Although the simplifying assumptions made in the de-
scription of the medium do not allow us to account for lat-
eral inhomogeneities, separate inversions were performed
with traces D1 to Dll to detect possible lateral changes
over short horizontal distances. The lower crust has been
represented by a stack of 280 elementary layers having a
thickness of 75 m. The starting velocity model- labelled
Model 0 in Figure 10 - features a constant velocity zone
between 17 km and 23 km depth, and a linear gradient
zone between 23 km and 38 km depth corresponding to
an increase in velocity from 6.5 km/s to 7.5 km/s.
The crustal models derived from the inversion of traces
D1 to Dll after 30 iterations are presented in Figure 10.
It is seen that all models obtained show a laminated struc-
ture with alternating high- and low-velocity layers super-
imposed on the starting model. The average thickness
of the lamellae is 280 m, which represents approximately
4/5 of the mean wavelength at 20 Hz, and the velocity
contrasts at the interfaces are typically of the order of
200 m/s. However, the lateral variability of the veloc-
ity profiles is difficult to analyze with all approximations
made. In general, the estimated models show similar pat-
terns with energetic reflections between 23 km and 28 km
depth, but in detail, they are badly correlated. In reality,
the observed differences merely reflect the lateral varia-
tions already noticed on seismograms D1 to D11, and can-
not be seriously exploited. We can also notice some long
wavelength artifacts, especially on models 3, 9, and 10,
which emphasize the fact that inversion of single traces
is usually a poorly constrained problem. On the other
hand, it is seen in Figure 11 that the real data have been
correctly modeled in the inversion process, and that the
fit of the data is fairly accurate even for weak amplitude
events. Indeed, the correlation coefficients calculated at
the last iteration between the data D1 to Dll and the
synthetic seismograms S1 to Sll are all between 82% and
88%.
Keeping in mind that the only reliable information
which can be retrieved from a waveform inversion is
bounded by band-width of the seismic signals, the range
of confidence is limited to wavelengths between 200 m and
700 m. With that restriction, one may conclude that the
results of the inversion of shot GECO 65 are in fairly good
agreement with previously published studies in terms of
the thickness of the lamellae and velocity contrasts at the
interfaces.
Conclusions
A strategy for complete waveform inversion is pro-
posed. In this method, the long wavelength of velocity
is estimated from the travel time of reflected events. A
global search is performed in a bounded velocity space
using a Monte-Carlo algorithm. The bounded velocity
model is obtained from the standard velocity analysis.
Once the long wavelength of the velocity is known, a non-
linear waveform inversion is performed starting from this
solution. The nonlinear inversion adds the short wave-
length to the velocity over the long wavelength solution.
Therefore, combining two approaches leads to a complete
solution. However, if there is a smooth gradient in veloc-
ity, it could neither be retrieved by velocity inversion nor
by the waveform fit. This leaves a gap in the complete-
ness of the solution. In the future, we shall fill this gap
using wide angie reflection data.
The synthetic example presented shows a good fit be-
tween the initial model and the inverted model. The last
step of our inversion algorithm is applied to ECORS data
from the Bay of Biscay. The waveform fit between the
data and the synthetic after inversion has approached
WAVEFORM INVERSION
i I ' I I I
i , , i i -- c)o
I I , I I i
i I I I I
O
I t ' I I I
(tu) t[docI
f I i ,i J
i I I I I I
I i I ..I I I
,
i i J i i i -
.
! I I i i i
, I , J , CO
SINGH ET AL. 399
0.885
Sll
D10
0.871
SlO
D 9
0.851
S9
D8
0.824
S8
D 7
0.868
S7
D6
0.874
S6
Ds
0.863
S5
D4
0.882
D3
0.868
D 2
0.852
D1
0.862
S o
Fig. 11. Real and synthetic seismograms after 30 iterations. Seismogram So is the theoretical
response of the a priori model 0. The oscillations in seismogram So are due to the represen-
tation of the velocity gradient in the starting model by a stack of thin homogeneous layers.
Synthetic traces S to S represent the best fit of data D to D. The numbers displayed
beside the seismograms are the correlation coefficients of the pairs of traces (Si, Di).
88%. Inversion of the data predicts a set of laminated
layers in the lower crust, which is in agreement with pre-
viously published results. However, it should be noted
that the thickness and the velocity of the layers change
randomly and are not constant as proposed by previous
workers. The thickness of the layers varies from 100 m to
500 m and the velocity contrast from 50 m/s to 500 m/s.
The inverted result gives a precise thickness and velocity
contrast of each layer which could be used to estimate the
petrological composition of the lower crust.
Acknowledgments. The work was supported by Elf
Aquitaine. The subroutines on optimization were pro-
vided by Zvi Koren. Isabelle Picard was always available
in solving problems related to computers. Discussions
with Evgeni Landa have been very useful. Department
of Earth Sciences, Cambridge, publication number 1939.
References
Andrier, J.-F., Etude de la crote infdrieure litde partit
des enregistrements sismiques de la campagne ECOR$-
Gascogne, Thkse de Doctorat, Universitd de Bretagne
Occidentale, Brest, 1989.
Berkhout, A. J., Seismic migration, imaging of acoustic
energy by wavefield extrapolation, Elsevier Publ., 1980.
Bols, C., M. Cazes, A. Him, A. Mascle, P. Matte, L. Mon-
taderr, and B. Pinet, Crustal laminations on deep seis-
mic profiles in France and neighbouring areas, Geo-
phys. J. R. astr. $oc., 89, 297-304, 1987.
Cao, D., W. B. Beydoun, S.C. Singh, and A. Tarantola,
A simultaneous inversion for background velocity and
impedance maps, Geophysics, 55, 458-469.
Farra, V., and R. Madariaga, Non-linear reflection tomog-
raphy, Geophys. J., 95, 135-147, 1988.
Fhchs, K., On the properties of deep crustal reflectors, J.
Geophys., 35, 133-149, 1969.
Fucks, K., and K. Schultz, Tunneling of low-frequency
waves through the subcrustal lithosphere, J. Geophys.,
J, 175-190, 1976.
Gardner, G. H. F., L. W. Gardner, and A. R. Gregory,
Formation velocity and density: the diagnostic basis for
stratigraphic traps, Geophysics, 39, 770-780, 1974.
Gibson, B. S., and A. R. Levander, Imaging deep reflec-
tors in the presence of signal generated noise: Geo-
phys. J. R. astr. $oc., 89, 19-30, 1987.
Hobbs, R. W., C. P. Peddy, and BIRPS Group, Is
lower crustal layering related to extension ? Geo-
phys. J. R. astr. $oc., 89, 239-242, 1987.
Hubral, P., and T. Kreys, Interval velocities from seismic
reflection time measurements, SEG Publ., 1982.
Jannane, M., and ETG, Wavelengths of earth structures
that can be resolved from seismic reflection data, Geo-
physics, 5, 906-910, 1989.
Kenneff, B. L. N., and N.J. Kerry, Seismic waves in a
stratified half-space, Geophys. J. R. Astron. $oc., 57,
557-583, 1979.
Klemperer, S. L., T. A. Hauge, E. C. Hauser, J. E. Oliver,
and C. J. Potter, The Moho in the northern Basin and
Range Province, Nevada, along the COCORP 40N
seismic-reflection transect, Geol. $oc. Am. Bull., 97,
603-618, 1986.
Kormendi, F., and M. Dietrich, Non-linear waveform in-
version of plane-wave seismograms in stratified elastic
media, Geophysics, (press), 1991.
Landa, E., W. B. Beydoun, and A. Tarantola, Reference
velocity model estimation from prestack wave forms:
coherency optimization by simulated annealing, Geo-
physics, 5, 984-990, 1989.
Maxillier, F., A. Tomassino, P. Parrtat, and B. Pinet,
Deep structure of the Aquitaine shelf: constraints from
expanding spread profiles on the ECORS Bay of Biscay
transect, Mar. Pert. Geol., 5, 65-74, 1988.
Matthews, D.H., and M. J. Cheadle, Deep reflections
from the Caledonides and Variscides west of Britain
and comparison with the Himalayas, in Reflection $eis-
mology: A Global Perspective, Geodynamic Series, 13,
edited by Barazangi, M., and L. Brown, pp. 5-19, Amer-
ican Geophysical Union, Washington, D.C., 1986.
Meissner, R., The 'Moho' as a transition zone. Geo-
phys. $urv., 1, 195-216, 1973.
Meissner, R., Th. Wever, and R. Bittner, Results of
DEKORP 2-S and other reflection profiles through the
Variscides, Geophys. J. R. astr. $oc., 89, 319-324, 1987.
Paul, A., and F. Nicollin, Thin crustal layering in North-
ern France: observations and modeling of the PMP
spectral content, Geophys. J. R. astr. $oc., 99, 229-246,
1989.
Pinet, B., L. Montadert, and the ECORS Scientific
Party, Deep seismic reflection and refraction profil-
ing along the Aquitaine shelf (Bay of Biscay), Geo-
phys. J. R. astr. $oc., 89, 305-312, 1987.
Sandmeier, K. J., and F. Wenzel, Synthetic seismograms
for a complex crustal model: Geophys. Res. Left., 13,
22-25, 1986.
GEOMETRICAL MIGRATION OF LINE-DRAWINGS:
A SIMPLIFIED METHOD APPLIED TO ECORS DATA
Guy S4n4chal and Franqois Thouvenot
Laboratoire de G4ophysique Interne et Tectonophysique, Observatoire de Grenoble,
I.R.I.G.M., B.P. 53X, 38041 Grenoble, France
Abstract. The geometrical migration of line-
drawings can be an intermediate solution to the
problem of migrating vertical-reflection data. Our
method, based on the principle of the common
tangent, quickly produces a depth-migrated line-
drawing that is readily usable for interpretation.
Introduction
Migration is the process by which effects of dip
and diffraction are removed. For deep seismic
reflection data, conventional methods often produce
poor results and, moreover, pre- or post-stack digital
migrations are always time-consuming and require a
great computer capacity (Sattlegger and Stiller,
1974; Warner, 1987). Consequently, vertical-
reflection data are often represented unmigrated,
and many interpretations are based on line-drawings
of unmigrated time sections where the geometry of
the dipping reflectors is largely distorted.
An alternative solution is to migrate the line-
drawing, for instance using a ray-tracing method
(Rayhand, 1987). Unger (1988) described an
application of ray theory to migrate line-drawings,
considering a velocity that increases linearly with
depth. We tried to go here one step further into the
direction of simplification, using a very primitive,
but ecient, geometrical technique. This method is
based on the principle of the common tangent of two
spherical wavefronts (Hagedoorn, 1954). Velocity
models with horizontal and vertical variations can be
taken into account, provided the variations are
smooth enough.
Description of the process
Consider the geometry of a spherical wavefront
as reflected from a dipping interface with source and
Geodynamics 22
receiver in the same place (zero offset). On the
unmigrated section, the reflection will be positioned
along the vertical of the source-receiver pair, while
the true location of this reflection is where the
reflector is a tangent to the wavefront.
Suppose now an unmigrated reflective segment
delimited by two points (Fig. 1). For each point, we
can draw the corresponding spherical wavefront,
using the intersection between the vertical of the
point and the surface. The true position of this
reflective segment will correspond to the common
tangent of these two wavefronts, where reflection
will be at normal incidence.
The input data are digitized unmigrated line-
drawings, with each reflector being digitized into
small individual reflective segments. The digitization
step takes into account the lateral resolution, which
decreases with depth as a consequence of the-
wavefront geometrical expansion with time (Sheriff
and Geldart, 1982). Therefore, the length of any
reflective segment should exceed a few ten metres in
surface and a few hundred metres at depth.
Such a geometrical construction requires the
same vertical and horizontal scale, and we first have
to compute the unmigrated depth section. To
transform the time section into a depth section, we
define a velocity model, which is a grid of estimated
velocities with horizontal and vertical variations
(Fig. 2). For the subsurface, we use velocities
obtained during the pre-stack process (velocity
analysis); for the deep crust, we use velocities
provided by deep-seismic-sounding data (long-range
refraction and wide-angle reflection). The velocity
anywhere in the model can be approximated using a
double linear interpolation between neighbouring
grid velocities, and the unmigrated depth section
can be computed.
The migration process uses simple
trigonometrical relations for computing the true
position of each reflective segment (Fig. 3):
Consider the dipping segment A1A 2 (dip ),
positioned beneath the surface (dip c). The
coordinates (Xl,Zl) and (x2,z2)of the points A 1 and
401
402 GEOMETRICAL MIGRATION OF LINE-DRAWINGS
2
SURFACE
0 I I I\
I } x
WAVEFRONT 2 I [ \ \ / /
xNa I { " ', ///
UIGRATED SEGMENT
REFLECTOR
(UCaATEV)
z
Fig.1 Geometrical migration of a reflective
segment. Reflector A moves to B.
A 2 are computed using the digitized coordinates of
the time section and the velocity model. Next, we
compute the coordinates (Xl,Zrl) and (x2,Zr2) of the
points O 1 and O 2 situated on the surface, on the
vertical of the two points limiting the reflective
segment. The distance O10 2 is:
D ----%/ d2-Fp 2,
with d=x2--x 1 and p=Zr2--Zrl.
Let H----R2--R1, with R 1 being the radius of
circle C 1 and R 2 the radius of circle C 2. We can
express dips c and :
Applying trigonometrical relations to triangle
0102T allows us to express angle "y in terms of H
and D
] ,
and the dip of the migrated segment is O=ffq-o. The
coordinates of the two points B 1 and B 2 that define
the migrated segment are:
B 1 xlm----Xl--R lsinO , zlm:zrlq'R lCOSO,
X2m:x2--R2sinO , Z2m----Zr2q'R2cosO.
The above equations are functions of the surface and
reflector dips. Reflector with excessive dips cannot
be migrated, which is the case for less than 1% of
the data. When c=0 (no altitude variation), /
cannot exceed 45 .
The comparison between an unmigrated time
section (Fig. 4a) and a migrated depth section (Fig.
4b) shows the transformation due to the process. Of
course, dips are increased and reflectors moved
laterally up-dip. This is especially clear for reflectors
around 8 s TWT on the left side of the section.
Moreover, entangled reflectors on the right side of
the section (5-6 s TWT) are migrated into a more
continuous reflective band.
Application to ECORS data
We processed the ECORS-CROP Alps data
that provide a 320-km long section of the crust from
the eastern Massif Central, France, to the Po plain,
Italy (Bayer et al., 1987; Nicolas et al., 1990). The
input data is the line-drawing of the unmigrated
time section (Figs 5a and 6a). This line-drawing was
produced by merging two stacks (near traces and far
traces), and totals some 20,000 reflectors. We
computed the unmigrated depth section, referring all
depths to sea level, and giving special attention to
altitude corrections. The velocity model was
determined using velocities given by the contractor
(C.G.G.) for the subsurface and those determined by
deep seismic sounding for the lower crust. The final
result is the migrated depth section (Figs 5b and
6b), represented with the same vertical and
horizontal scale. The more stippled aspect of these
two pictures is due to the splitting of each reflector
into several reflective elements during the process.
Data gaps on the line-drawing are due to locally-
poor signal-to-noise ratio on the unmigrated section.
Such gaps can also appear in case of a sudden
change in the reflector dip --at least if we stick to a
simple reflected ray geometry where diffraction is
not taken into account. Some reflective segments
can however be migrated in these blank areas during
the process.
Migration is clearly indispensable beneath the
Vanoise Massif (Fig. 6) where we can observe a two-
band reflective zone. This structure corresponds to
the Frontal Pennine Thrust that separates the
external Alpine realm from the inner Alpine realm.
These two bands merge into a single highly-reflective
horizontal zone under the Gran Paradiso Massif.
The close-up of Figure 4 shows the wedge geometry
of the Belledonne massif that seems to have been
both overthrusted and underthrusted by material
from the inner Alpine realm.
Advantages and limits
The main drawback of this method is the
subjectivity of the hand-picked input data. An
SINICHAL AND THOUVENOT 403
ECORS-CROP ALPS
DISTANCE (KM)
120 130 140 150 160 17'0 180
.
(.3 3 21
L
gO 4- 5.88

5

- .
7
.
I-- 8
i
>- 9
.< -5.88 ',
z 10 5.88
! 11- :
0 ' 6.18 :
32:12 6.18

- 13
.

15
.

17

18

19
20
120 130 140
OBSERV^TOIRE OE GRENOBLE
CI60GNE V. 3.02, 16-1"1^R-1990
200 210 220
I I

.55-.----.67----.60 ...4.0.-____,-4.$0 4.25_
.oe 4. ?e : 4. ?e ! 4.45 s. so i s. oo
5.o : 5. e '5.75
5.88 '5.88
5.88
' 5.88
5.88
,6.18 ' 6.0
6.20
6.25

___ 4.$5
5.80
.5.90
i I I I
150 160 170 180 190 200
VELOCITY GRID (KM/S)
6.25
2 0
Fig.2 Example of velocity grid. (Surface is drawn using the elevation static corrections,
datum plane at altitude 2400 m). Values are velocities in km/s, averaged from the
surface.
o
i 1
i 2
i 3
i 4
i 5
I 7
I 8
-- 9
-- 10
I 12
--13
--14
115
i 16
I 17
218
I 19
20
220
automatic line-drawing can also be used if the
signal-to-noise ratio is acceptable. We follow Stiller
and Thomas (1990) in their statement that both
methods complement each other and form a basis for
a solid interpretation.
A second drawback is the use of spherical
wavefronts, which may appear an oversimplification
of the problem. To test this point, we checked that
the velocity variation between the unmigrted and
migrated positions of the reflective element is
usually negligible: a velocity variation higher than .3
km/s is observed only for .5% of the data.
Conventional digital migrations are always very
sensitive to the velocity model (Yilmaz and
Chambers, 1984; Berkhout, 1984), and in particular
those based on ray-tracing methods result in
migrated sections with unstable dips. The program
proposed by Unger (1988) uses a linear velocity
increase with depth and doesn't take into account
altitude variations, two conditions that are not
acceptable in the Alps. In our method, velocity is of
course an important parameter to transform the
time section into a depth section, but the migration
itself is not so much sensitive to velocity variations.
The computing time amounts to a few seconds
for processing several thousands of reflectors,
whereas conventional digital methods require several
hours. Very little computer memory is required, to
such an extent that even a microcomputer can
handle the work. The process can be interactive,
which gives the possibility to test different velocity
models and different line-drawings. For instance,
several data files can be used to enrich the final
section suitably.
Conclusion
The fastness of the process and the advantages
described hereover make this method a good
Fig.3 Trigonometric construction showing the
relations between the dip of the unmigrated
reflector A 1A2 (), the dip of the surface (c)
and the dip of the migrated reflector B1B 2 ().
0
o
9
10
ll
12
13
90
SERVATOIRE gREnObLE
CiGNE V.3.04, 28-AR-190
DISTANCE (KM) DISTANCE (KH)
90 200 210 220 230 190 200 210 220 230
I I I -5 I I
o ' ', , '" . ..... " :5':.---
-': :'- o ., ......:. ":%,-.-.'::,'. . ..-. +:..:, :- :. -'-
--- ,.- '- , . -Z-= ' ': .-' ':;. ,.' '-' ',':.. ........ -:.-::--2;:-'-'-:-
-k ...... -- - . .' :,:?.,..-,%: -' .' - .: . -
- n- . :7, -' '9 : a . '-4:7:%:-." 'qf.'-:- '- . '.-" '..
- - r-k.- . L 4 o
-- _ ." t '. '- .' ' ,,. -.: ?: ,. '/..-.
- . . ' : .- ;2., ,.--,..q[. - , .. ,-.--'
, - ..... _ :y. --..
....... - ..... ' ' - ' =. '..'-' -_=.e,=. m:C-,;
--"' . ;._-. 7 '
_ 20
. -'- r .... . -- 9
_.-- . -- ' 10
_ _ .
.... 11 25
..
12 . . . : _ .--.
13 30
o o a;o o so o o o o
BEFORE MIGRATION AFTER MIGRATION
Fig.4 Unmigrated time section (a) and migrated depth section (b) of part of the ECORS
Alps profile (ALP1). The two bands correspond to the Frontal Pennine Thrust.
alternative solution to the problem of migrating
deep seismic data. These results are of course not
claimed to be better than those obtained by other
migrations (e.g. Damotte et al., 1990). Simply, we
believe that a comparison between several migration
methods is highly commendable, especially at great
depth where it allows us to confirm or reject some
structural interpretations.
References
Bayer, R., Cazes, M., Dal Piaz, G.V., et al., Premiers
rdsultats de la traversde des Alpes occidentales
par sismique rdfiexion verticale (programme
EGORS-CROP), C. R. A cad. Sci. Paris, 305,
1461-1470, 1987.
SlNlCHAL AND THOUVENOT 405
o
o
o o
0 0
0 0
U")
0 0
(33S) 3NI 1 iVri-Orll
o o
c
o o
z
oo o
- - Z o
o
o o
o o
o o
o
o in o in o in o it) o v1 o
o it) -- -- c e4 c c ,r - u)
o
i i i o
.r
(,.. . - ..
-: ;".., ....;.
::.'" .-" ;'4'.'..
-
... :;;i! ,'-"
;.6;: ' :...' .:;.';.'
:.!';?-'"' :4,,', ;,
.,,..' . ...''; ....

!':?;!

. .
, .;..,

.. I" - ;.- .'.'..? .-
: : :-if.? :' '..<
, ' '? .' '2.
. -' ' -.-:4 ' '
.
..:....-: :..:. '(::..., '
".::;:".:Ji ":'
t'. ".'".: i,' '
-'.. .;.:.:?... . .....
.-.
:{ ' ..: ,-..: (. .: . . o .
...........
" .; ..:..'. '.' z..,.:
!; :. ... ..
. .- ,.
'":; . .', e%'.- !i, ?..-".,
k,':.'/ . --/7';-" c:.'.;.;'
'.' ' "-"
" ' :.'" fii{'
k ' '-'--'
.i. ...-i... .... -..?.:.:..,
.. ' .;' .4..::' ,, 71. .. o
;:: .: ......f..c...,:.,.,.:.y..., , .
i:. :: '%' '...,.,..,,_ :,:,,!;?... ..,.,:,....:.:,,,,:.{ ..... ,i.r,..x.-.
.?..,..: .i.:;..-:; ;..:' i '
:"., .X.',"..',.c:;tit :,.-:..
:: t...:; -:-';-:t-.*i.:.C:.'
.. . .?:.,.,,.,i -,... ;....; 'i:
.-, -. ..t..'7,:-
-'..:';,c;- .' ..: t'i," "'. e':''-',C...
,x' ,.
:i';..;!' .f:?d.,::./?.: .
:,. , .: .FL.: .':::'.:. '-
.....,: :.&: ..;. -
./. :,.,t.'. -:.;..
o
LL!
o o
,,, ._<
I1 o u') o u') o o o ly)
HidSO
SINICHAL AND THOUVENOT 407
Berkhout, A.J., Seismic migration: imaging of Sattlegger, J.W., and Stiller, P.K., Section
acoustic energy by wave field extrapolation, migration, before stack, after stack, or
Elsevier, Amsterdam, 1984. inbetween, Geophys. Prosp., 22, 297-314, 1974.
Damotte, B., Nicolich, R., Cazes, M., et al., Mise en Sheriff, R.E., and Geldart, L.P., Exploration
0euvre, traitement et pr6sentation du profil seismology (vol. 1), Cambridge University Press,
plaine du Pb-Massif central, In: Roure, F., 1982.
Heitzmann, P., and Polino, R., Eds., Deep Stiller, M., and Thomas, R., In DEKORP Atlas,
structure of the Alps.- Mdm. Soc. gdol. Ft., Meissner, R., and Bortfeld, R. K., eds., 7-10,
Paris, 156, 65-76, 1990. 1990.
Hagedoorn, J.G., A process of seismic reflection Unger, J.D., A PC Program for Migration of Seismic
interpretation, Geophys. Prosp., 2, 85-127, 1954. Reflection Profiles, Geobyte, 42-52, 1988.
Nicolas, A., Hirn, A., Nicolich, R., et al., Warner, M., Migration - Why doesn't it work for
Lithospheric wedging in the western Alps deep continental data?, Geophys. J. R. astr.
inferred from the ECORS-CROP traverse, Soc., 89, 21-26, 1987.
Geology, 18, 587-590, 1990. Yilmaz, O., and Chambers, R., Migration velocity
Raynaud, B., A 2-D, ray-based, depth migration analysis by wave-field extrapolation,
method for deep seismic reflections, Geophys. Geophysics., 49, 1664-1674, 1984.
J., 93, 163-171, 1987.
ENERGY- AND POWER-SECTIONS IN SEISMIC INTERPRETATION
R. Bittner and W. Rabbel
Institut fr Geophysik, Universitt Kiel, Leibnitzstr. 15, D-2300 Kiel.
Abstract. In deep reflection seismics, scat- anomalies within the earth's crust are often
tering and diffraction of seismic waves are often observed in deep seismic profiling. The resulting
observed. These phenomena cause, for example, scattered appearance of reflected/diffracted
offset dependent reflectivity pattern and phase arrivals generally causes problems in seismic
distortions which cause additional processing data processing and interpretation. Many examples
problems. In this context it is known that the can be found, e.g., in the COCORP- and DEKORP-
phase- and group-characteristics of reflected/ Atlas (COCORP, 1988; Meissner and Bortfeld,
diffracted arrivals may deviate from another. 1990). Wever (1988) counted average reflector
Compared to the phase behavior, the group- or lengths of 1 to 4 km for the profiles which are
energy-pattern often represents the more stable now shown in these documents. Most of the single
average reflection response of an interface. phases of the stacked sections could hardly be
Using data from the DEKORP 2S profile (South- correlated for longer distances than a few
Germany) we illustrate above problems and aspects kilometers. Where unique horizons cannot be fol-
of enhancing group characteristics as an ad- lowed, local average-properties of the wavefield
ditional tool for seismic processing and inter- (accumulations of reflection elements, trends of
pretation. Three algorithms are shown which focus their dip etc.) usually take the role of in-
on the energy and on temporal changes of the dividual phases as a basis of the tectonical or
energy density of a seismogram. The corresponding geological interpretation.
types of sections are: Using the complex trace approach, Shtivelman
et al. (1986) have shown that the phase- and
- power section (CMP-stack of temporal derira- group-characteristics of arrivals which have been
tires of seismic trace envelopes), reflected from irregular interfaces may deviate
- perigram sections (CMP-stack of high-pass fil- from one another while for smooth interfaces both
tered trace envelopes) and characteristics coincide. The group or energy
- energy-sections (CMP-stack of traces with part of the signal which is represented by the
squared amplitudes averaged within a moving trace envelope contains information on the aver-
time window). age reflector structure. Its phase behavior, on
the other hand, is strongly influenced by the
In comparison to conventional stacked sec- local reflector structure. As mentioned above,
tions, the resulting section types have lower many interpretations of deep seismic reflection
resolution. But they are, on the other hand, not data are intuitively based on average properties
so sensitive to details of velocity structure and of the observed wave fields. In the present paper
small static time shifts. we try to quantify one of the possible average
Reliable information on the energy pattern of properties, namely energy, by some simple
a seismic section can, therefore, be obtained algorithms and test their meaningfulness with a
even in early stages of processing. Focussing on field example.
zones of changing energy level, power-, perigram-
or energy-sections enhance the dominant features
of the seismic image of the subsurface and thus
contribute to simplifying the interpretation.
Introduction
Energy of a Seismic Trace
Engelhard (1986) has shown that the energy of
ground particle movement is proportional to the
square of the seismic trace envelope if the amp-
Discontinuous reflector elements, finely litude of the trace is proportional to the par-
layered structures and irregular velocity ticle velocity (which is usually the case for
land geophones). With the help of a Hilbert
transform the envelope is easily computed as the
ContinenmlLiosphere: Dp Seisnfic Refltions argument of the complex equivalent of the real
Geodynarrfics22 seismic trace (Taner et al., 1979). Formulae are
1991AmeficanGphysical Union given in Table 1.
409
410 ENERGY- AND POWER-SECTIONS
Seismic Trace
Hi|bert-Transform
Complex Trace
: s(t)
: h(t) : H (s(t))
: z(t) : s + i h
Envelope I ! : e(t) = ( s 2 + h 2 )
Power Trace : P(t) = de(t)/dt
Energy Trace
t._
= I )2
: E(t) , (s(t)
N
j=t 2
Perigram
: P(t) = e(t) - b(t)
N
t+
! E e(t) ( moving average ) : b(t) =
N
j=t- E
N N
--+ 1 g t g NSAMP
2 2
NSAMP = samples / trace , N = windowlength
Compared to a normal seismic trace the en- the temporal derivative of the envelope. It in-
velope is a smooth time series with only positive dicates variations of the energy level, suppres-
values (see Fig. 1, trace# 1 and 2), a feature ses the d.c.-component of the envelope trace and
which is not appropriate for presentation in therefore reverberations of the original seismic
seismogram sections. Therefore, we prefer using trace and transforms the envelope to a
I 2 3 4 5
1) Normal trace
2) Envelope
3) Power trace
4) Energy trace
5) Perigram
wavelet-like signal (Fig. 1, trace#3). Since
the envelope corresponds to the seismic ener-
gy we call its temporal derivative the
"power trace".
For the same reason, i.e. better graphical
presentation of envelopes $htivelman et al.
(1986) defined the so-called perigram which is
the envelope minus a moving average (Fig. 1,
trace#5). The resulting high pass effect depends
on the length of the moving time window within
which the average is computed. The appear-
ance of perigrams is quite similar to that of
the power-trace.
A last and simpler approach is the "energy
trace". It consists of the squared amplitudes of
the seismic input trace, which can optionally be
averaged over a moving window (Bittner, 1989) in
order to approximate the envelope function
(Fig. 1, trace#4). Its only advantage is short
computation time.
Formulae for the computation of perigrams,
power- and energy- traces are listed in Table 1.
It is evident that these three types of traces
are independent of the sign of the arrivals of
the original seismic trace.
A Field Example
Fig. 1. Seismic trace and its energy related The DEKORP 2S reflection line crosses three
attribute traces. Formulae are given in Table 1. main Variscan fold belts (Fig. 2). For our inves-
BITrNER AND RABBEL 411
OmmmO km
Fig. 2. Location map of the DEKORP (D-1 to D-4)
and KTB reflection profiles. The location of
the selected segment of the D-2 profile in the
south is the rectangle.
1 = Basement of the Rhenohercynian, 2 = Basement
of the Saxothuringian, 3 = Basement of the Mol-
danubian, F = Frankfurt/Main, HH = Hamburg,
K = K61n, M = Mnchen, NVDF = North Variscan
Deformation Front
At the top of the reflective lower crust (Fig.
3 at about 5 s TWT, see arrows), for instance,
strong reflections with opposite dip directions
are present on all offset panels. Their relative
amplitudes, however, are changed drastically from
offset to offset. The right dipping reflector, in
particular, is strong until 4 km offset, vani-
shes more or less at 8-10 km and appears again at
13-15 km offset. At present we do not have a
satisfactory explanation for this observation.
Another noticeable feature is the varying appea-
rance of the reflections at Moho level at about
9.5 s TWT which become more and more diffuse or
disrupted with increasing offset.
The greatest similarity to the final stack
(Fig. 4b) is shown (as expected) by the near
offset panels. We regard these seismogram
examples as typical for three dimensional
scattering media. Problems of CMP processing
arising from this situation are obvious:
- Since the reflection amplitudes vary abruptly
with offset it is not possible to get "true
amplitude stacks".
- Reflected, dipping reflected and diffracted
phases often require multi-valued different
stacking velocities.
- Static inaccuracies, phase shifts caused by
varying geophone coupling conditions etc. con-
tribute additionally to more or less random
variations of stacked amplitudes.
It is, therefore, not surprising that the 16-fold
stacked section (Fig. 4a, "brute stack"), which
corresponds to the offset panels of Fig. 3, has
lost information in particular at the top of the
lower crust and at Moho level, where a gap now
occurs. Phase-shifts leading to amplitude de-
struction after stack can partly be corrected by
refined NMO-velocities, residual statics and
coherency-enhancing processes. In the final stack
(Fig. 4b, processed at the DEKORP Processing Cen-
tigations we choose a 12 km long segment of its ter, Clausthal-Zellerfeld, which included: decon-
southern part near Dinkelsbhl (Fig. 2, see rect- volution, frequency filtering, extensive veloci-
angle) which lies within the oldest Variscan belt ty analysis and application of residual sta-
unit, the Moldanubian. A rather transparent upper tics) arrivals of about 4 km correlation length
crust, a highly reflective lower crust and strong can now be observed. Also the reflections at Moho
diffracted arrivals are the essential features of level are now continuous over the whole distance
this section (Figs. 3, 4). Since the DEKORP 2S range. This stack contains a maximum of available
profile was recorded with explosives throughout coherent energy and will be used for reference
(DEKORP Research Group, 1985), the initial sig- later. For the interpreter, however, in this
hal/noise ratio was high enough to allow us to particular case it is difficult to discriminate
select single coverage offset panels of high dominant phases from those of minor importance in
signal-to-noise ratio. Fig. 3 shows four sections order to derive a structural picture.
covering the offset ranges of 1-3 km, 2-4 km,
8-10 km and 13 - 15 km, respectively. The traces
are NMO-corrected with a simple smooth velocity Energy-, Power- and Perigram-Sections
depth function. (EPP-Sections)
The common appearance is more or less un-
changed throughout the four sections of Fig. 3, As described above, perigrams, power- or
revealing a high level of signal energy up to the energy-traces serve as tools for the visual-
maximum offset. The details of the reflected/dif- ization of the energy content of a seismic trace
fracted phases, however, depend strongly on the with time. Since the energy trace has only posi-
source-receiver offset (Fig. 3). tire values, a CMP-stack of the corresponding
T
w
T
10-
T
w
T
Is] -
10-
Fig. 3. Single coverage NMO-corrected offset panels for the selected segment of the
DEKORP 2S profile. The range of the source-geophone offsets is indicated. Note the
variation of the reflectivity pattern with offset in particular at 5 s and 9.5 s TWT,
respectively.
traces will be resistent against local variations For Fig. 5 and 6a the same simplified velocity
of stacking velocities, inaccuracies of static depth function was used for NMO-correction as for
corrections and other phase distortions. Energy- the brute stack (Fig. 4a). No residual static
stacked sections or the derived power- or peri- corrections were applied. Compared to the brute
gram-sections (in the following simply called stack (Fig. 4a), the corresponding EPP- stacks
"EPP-sections") display those parts of standard (Fig. 5, 6a) show a continuous Moho-band. Also
seismic sections which are stable with respect to midcrustal reflections are more pronounced. Com-
minor variations of the processing parameters. pared to the final stack (Fig. 4b), the EPP-
They enhance dominant features of a stack. stacks give a simplified image of the lower
The EPP-sections of Fig. 5 and 6 show the same crust, emphasizing those horizons which are domi-
segment of the DEKORP 2S profile as Fig. 4. nant in terms of energy. Even reflections in the
Brl'TNER AND RABBEL 413
DEKORP 2S
Brute Stack
DEKORP 25
T
W
T
Is]
lO
A B
21on
Fig. 4. Conventional stacked CMP sections of the selected segment of the DEKORP 2S
profile. (a) brute stack, (b) final stack of the DEKORP processing centre Clausthal-
Zellerfeld, F R G.
DEKORP 2S DEKORP 2S
T
W
T
Is]
lO
Energy Stack
Power Stack
"'.' ':'": :'.-"'-". :"2"'. '2 :".',; ....... .... ':':"' '' ' "": ..... 0 2 km
B
Fig. 5. (a) CMP stack of energy traces and (b) CMP stack of power traces corres-
ponding to the brute stack (Fig.4a). The stacking parameters are unchanged.
DEKORP 2S DEKORP 2S
Perigram pre-stack ..... 'i"?":':':,. 7:: Perigram - post-stack '.. '
"': ;::72;:,,., ' .......
?'. " ,.,':'...... '" .. .. :,.,;.: :':'".,','"-:.::- ..... : ..... ,:: .
' ' :.' :'7., '"' "' ..::,: .',:_j.: ..:" ...:,,;.,:-z,, 'L..:, : .... ."'- c: " .;'C.;: ;,.,. ,4' ..,
T
W
T
Is]
lO
Fig. 6. (a) CMP stack of perigrams corresponding to the brute stack (Fig.4a, the
stacking parameters are unchanged). (b) Perigrams of the final stack (Fig.4b). The
algorithm was in this case applied post-stack.
upper part of the crust at about 3s TWT and at above mentioned stability of EPP-sections to
4-5 s TWT are clearly presented. the stacking process. Residual statics, refined
The difference between brute and final stack stacking velocities etc. do not change the re-
(Fig. 4a and 4b, respectively), for example sults significantly. Compared to the conventional
at Moho-level, is striking. It results from an stacks, however, the EPP-sections suffer from a
insufficient temporal adjustment of the phases to loss in resolution. It means, on the other hand,
be stacked in case of the brute stack. The cor- that reliable information on the energy or group
responding perigram-sections (Fig. 6a and 6b, characteristics of a seismic section is already
respectively) obviously show greater similarity. available at an early stage of processing.
Although there are some differences between the
"brute" and "final" perigram sections (e.g. the
reflector at 9 s on the left side of the section) Conclusions
most of the dominant phases are observed in com-
parable quality on both sections. These are in Three types of attributes have been presented
particular: which describe the energy content of a seismic
trace and which can be handled like seismic tra-
- A continuous "Moho reflection" at 9.5 s TWT, ces: "Perigrams", "power-" and "energy-traces".
- right dipping reflections between 8 and 9 s In comparison to conventional stacked sec-
TWT, tions, the corresponding new types of section
- reflection/diffraction elements around 6 s TWT have lower resolution. They are, on the other
- reflections on top of the lower crust around hand, insensitive to details of the velocity
5 s. structure, small static time shifts or other
phase distortions. Reliable information on the
This good correspondence, which is similar for energy pattern of a seismic section can, there-
all EPP-type sections, is an indication of the fore, be obtained even in early stages of proces-
BITrNER AND RABBEL 415
sing. Emphasizing zones of changing energy DEKORP Research Group, First results and prelimi-
level, power-, perigram- or energy-sections en-
hance the dominant features of the seismic image
of the subsurface. They may thus be an additional
aid for simplified interpretation of seismic
reflection profiles.
Acknowledqements. We acknowledge the con-
nary interpretation of deep-reflection seismic
recordings along profile DEKORP 2-South, J.
Geophys., 5--7, 137-163, 1985
Engelhard, L., Determination of the attenuation
of seismic waves from actual field data, in
DGMK report 254, Absorption of seismic waves,
83-119, 1986
tinuous support of the project DEKORP by the Meissner, R . and Bortfeld, R . (eds.), DEKORP-
Bundesministerium ffir Forschung und Technologie Atlas, Springer- Verlag, Heidelberg, 1990
(BMFT) under grant RG 8703. Publication No. 422 Shtivelman, V., Landa, E. and Gelchinsky, B.,
from Institut ffir Geophysik, Universitt Kiel. Phase and group time sections and possibilities
for their use in seismic interpretation of
References complex media, Geophysical Prospectinq, 3--4,
508- 536, 1986
Bittner, R., Siqnal/Noise-Verbesserunq enq ver- Taner, M.T., Koehler, F. and Sheriff, R. E.,
messener seismischer Daten unter Verwendunq ei- Complex trace analysis, Geophysics, 4--4,
nes Enerqie-Mehrspurfilters, PhD thesis, Uni- 1041-1063, 1979
versity of Kiel, F R G, 1989 Wever, Th., Untersuchunqen zum Reflexionsverm6qen
COCORP, COCORP-Atlas, Cornell University, Ithaca, der kontinentalen Kruste,PhD thesis, University
USA, 1988 of Kiel, F R G, 1988
IMPROVED INTERPRETATION OF KAPUSKASING CRUSTAL REFLECTION DATA
BY SELECTION OF THE MOST EFFECTIVE REPROCESSING SEQUENCES.
Z. Hajnal, D. Scott, B.I. Pandit and B. Reilkoff
Department of Geological Sciences, University of Saskatchewan,
Saskatoon, Saskatchewan, Canada, S7N 0W0
G.F. West
Department of Physics, University of Toronto,
Toronto, Ontario, Canada, M5S 1A7
Abstract. Reprocessed and migrated seismic 1989, Ernst and Halls, 1984; Percival and Card,
sections of lines 1 and 6 of the Kapuskasing 1983].
reflection experiment displayed a highly Since 1984, LITHOPROBE has supported a number
reflective upper crust with a significant number of multidisciplinary programs aimed at defining
of previously unrecognized laterally continuous the structural geometry and tectonic evolution of
reflectors. Five of these reflectors imaged major this window to the deep crust of the Superior
low angle thrust faults with geometry revealing a Province. An important constituent of these
ramp-and-flat style of deformation. Three of investigations is the 1987-1988 reflection survey
these events are representing the imbricate fault comprised of lines 1, 2, 3, 4 and 6 (Figure 1).
system of the Kapuskasing uplift. The other two A regional analysis of this information was
underlying subparallel reflections and an reported by Percival et al. [1989]. A more
equivalent number of major near-vertical faults detailed synthesis of lines 2, 3 and 4 was
were detected at mid-crustal depth. The combined described by Geis et al. [1990]. The regional
structural attitudes of these features established interpretation was based on unmigrated line
a local subsidence framework which affected the drawings taken from sections which were subjected
entire crust under the central region of the only to rudimentary signal processing sequences.
Kapuskasing structural zone. Herein, we describe reprocessing of lines 1 and
6, giving a direct comparison to subsurface
Introduction structural features recognized by the earlier
studies. A number of specific evidences
The Kapuskasing Structural Zone (KSZ) cuts illustrate that processing sequences designed on
obliquely across the Archean Superior Province spectral properties of coherent seismic events on
(Figure 1) over a distance of at least 400 km in field records, coupled with a close monitoring of
northern Ontario. It is outlined by major gravity the output from each of the signal enhancement
and aeromagnetic anomalies [Wilson and Brisbin, operations, leads to a clear improvement to both
1965] and by three distinct low to high signal levels and continuity of the coherent
metamorphic grade lithotectonic zones (2-3 kbar events on the final seismic sections.
greenstone-granite, 5-6 kbar tonalite-gneiss, 7-9 Reprocessing illuminates the spatial complexity of
kbar granulite-gneiss megalayers) [Percival, the previously identified structural elements and
1990]. Earlier interpretations have postulated permits a recognition of the characteristic
that the KSZ is an indication of a pronounced attitude properties of the KSZ related major
upwarp of the Conrad discontinuity [Wilson and geotectonic domains. The ubiquitous appearance of
Brisbin, 1965], a graben associated with mid- diffraction patterns reveals the rough and
Proterozoic rifting [Innes et al., 1967], a suture irregular nature of the reflecting interfaces of
[Gibb, 1978], a failed arm of a plume generated at this complex structural environment.
a triple rift junction [Burke and Dewry, 1973],
and a broad sinistral transcurrent fault zone Initial Analysis
[Watson, 1980].
Recent evidences support crustal thrusting and The commercially stacked seismic sections of
brittle uplift of granulites above a mid-crustal lines 1 and 6 (Figures 2, 3) displayed a number of
decollement (=15 km depth), leading to crustal identifiable reflective zones with regionally
shortening of about 70 km and thickening of the variable image quality [Percival et al., 1989].
present lower crust over a 200 km-wide zone The variations in reflection character were
[Percival, 1990; Fountain et al., 1990; Geis et attributed to alterations in the geological
al., 1990; Percival et al., 1989; Boland and Ellis terrane. Segments of both profiles (H on Figures
2 and 3) that cross into the Abitibi belt were
associated with west and northwest dipping
prominent reflections extending from the surface
ContenmlLiohere: DpSeisnfcRefltions down to nearly 13 seconds two-way travel time.
Geodynamis22 With dips of these events larger than 20 , they
o 1991 Amedcan Oeophysil Union project to the surface beyond the south end of the
417
418 INTERPRETATION OF KAPUSKASING REFLECTION DATA
>
+ + /+ +
5
--!- --F
+ -/
;::,,,,
;:>
i..,d
c)
LINE 1
HAJNAL ET AL. 419
CMP
1
2
6
COMMERCIAL SECTION
Fig. 2. Final seismic section of line 1 produced illustrates the scatter of CM points. There is no
by the original processing parameters. F to I direct correlation between decrease of signal level
represent zones of reflectivity recognized by and increase in distribution of reflection points.
Percival et al. [1989]. Binning diagram on the top
survey lines. Truncations of reflections (I"-I", to reflect the local surface structural geometry.
Figure 2 and 3) were interpreted as the extensions The abundance of reflective zones in the upper 8-
at depth of the Ivanhoe Lake Fault Zone (ILFZ). 10 second portion of both profiles were judged
A zone of sub-horizontal reflections at the comparable to patterns observed in other
northern end of line 1 (Figure 2) was linked to Precambrian areas [Gibbs, 1986].
the overlying gently undulating high-grade
paragneiss. The gently curving set of subparallel Reprocessing
events (G, Figure 2) were correlated to the
overlying structural dome of tonalitic gneiss. Since local ambient noise conditions are
The shallow, gently west-dipping reflections (J, assumed to be independent from trace to trace
Figure 3) of the north-west end of line 6 appeared [Mayrand and Milkereit, 1988], an algorithm
lOOO
i
7
rn
8
m
lO -
11 -
lO
11
_
-12
-13
-14
COMMERCIAL SECTION
Fig. 3. Final seismic section of line 6 generated with the
original processing parameters. H to J are zones of reflectivity
as defined Percival et al. [1989].
HAJNAL ET AL. 421
adopting this principle was developed to reject background noise, and hyperbolic primary
traces contaminated with noise bursts or reflections. (See Table 1 for field recording
exhibiting less than minimal signal levels. About parameters.) Spectral properties of these waves
5% of the traces were eliminated by this were monitored on a significant number of field
technique. records. A thorough analysis of the
All field records exposed five coherent characteristic features of these coherent events
patterns (Figure 4); first breaks, ground roll, permitted the design of an effective multi-step
vibrator phase loss generating harmonic (AGC, predictive deconvolution, FK filtering)
distortions [Martin & White, 1989], near-offset signal enhancement procedure. A strict adherence
0 0 0 0
i i i i
3
5
6
7
FIELD RECORD
Fig. 4. A typical field record (Line 1, Record harmonic distortion, 4 = near-offset background
740, V pat. CMP 2437). 1 = first break energy, noise, 5 = primary reflection.
2 = ground roll, 3 = vibrator phase loss generated
TABLE 1 Data acquisition parameters
Source Recording
- 4 Mertz 18, 20,072 kg vibrators
- Sweep: 14 s, linear, 12-52 Hz
Pelco Advance 1, Model 5
controller, 8 sweeps per VP.
- Drag: 127.5 m (vibrators equally
spaced, moving 12.5 m per
sweep)
- VP spacing: 100 m
- Geophones: 0Y0 14Hz 0.7 damping
- Group spacing: 12 geophones spaced evenly
over 50 m
- Station spacing: 50 m
- Spread Design: Asymmetric split spread,
60 station NW
13 st at ion gap
180 station SE
- Recording system: Two recording trucks, each
with a Texas Instrument DFS V
andI/Ofullprecisioncorrelator
stacker
- Filters: 64 Hz high cut
60 Hz notch (when needed)
- Recording Times: 18 s listen
32 s raw record
18 s final record
UNNORHRL'IZED RHPLITUDE 5PECTRUH
SCRLING-tO,'-4
RAW DATA
:3.0
1.5
0 I
15 30 45 60
,
75
FREQ. (HZ}
UNNORHRLIZED RHPLITUOE $PECTRUH
SCRLING-IO,,-3
DECON FILTERED
10.0
5.0
0 .----' - I I ---'-- .... -
0 15 30 45 60 75
FREQ. (HZ}
Fig. 5. Spectral properties of the primary reflections of the sample
record (Fig. 4) before and after prestack processing.
HAJNAL ET AL. 423
0 0 0 0 -.
i
6
PROCESSED RECORD
Fig. 6. Display of sample shot record in Figure 4 after AGC, Predictive
deconvolution and FK filtering.
to the original order of this many-sequence seconds and events which were not recognizable at
operation is mandatory to achieve optimal signal 2.8 seconds and 5 to 8.0 seconds, can now be
enhancement levels. The ultimate outcome of this traced through the record (Figure 6). The single
process was a broad band, spectral balanced set of deconvolution based spectral whitening processes
records (Figures 5 and 6). [Yilmaz, 1987] are also effective in broadening
Frequency content of all traces was raised over the frequency spectra [Milkereit et al., 1990] but
the range of 18 - 48 Hz (Figure 5). Primary they do not eliminate coherent noise patterns.
reflections, which were barely visible at 3.8 The above multi-step operation not only enhances
LINE'6
NW SE
--500
--450 o

4OO
Surface Elevation
-- 500
rrl
o o
::3 (1)
-- -500
Near- Surface Model
-- lO
s.
--lO
--:20
--30
Refraction Statics
Fig. 7. First break energy derived near-surface
static corrections along line 6. Near-surface
complexity and magnitude of corrections are similar
for line 1. For the near-surface model, velocities
are in m/s. On the statics profile, L = long
wavelength static, S = short wavelength static.
the upper end of the signal spectra but it also
removes a variety of the coherent noise
configurations (Figures 4 and 6).
The marginal overburden cover and small
elevation variations in most regions of the
Canadian Shield are not necessarily indicative of
uniform near-surface acoustic environments.
Detailed analysis of refracted first break energy
of field records along both profiles revealed
abrupt bedrock alterations, demanding specific
station-related static correction applications
(Figure 7). Static corrections, including short
and long wavelength, were derived by an automated
iterative procedure [Hampson and Russell, 1984].
Velocities of the near surface layers were derived
also by computation of iterative best fit to first
break energy. The constructive influence of these
corrections on the stacked data quality is
distinctly visible (Figure 8). Continuity and
clarity of several reflection trends were
significantly improved in comparison to the
conventional elevation plus residual static
corrections.
All past and present signal enhancement
processing sequences are tabulated in Table 2. Two
additional data specific operations, delay
corrections and migrations were implemented in the
present scheme. A constant 16 ms time adjustment
was required to correct for an instrumentally
induced delay on the far offset traces (121-240)
arising from a master-slave configuration of two
120-channel recording systems. Both final stacked
sections were migrated with finite difference
depth migration procedures. Migration velocities
were derived from refraction information [Boland
and Ellis, 1989]. Lateral variations in
velocities, as indicated by the refraction
results, were incorporated in the computational
HAJNAL ET AL. 425
CMP CMP
2954 2855 29,54 28,55
1 ........... ,,, .... ............... ,.., .... . ....... .., ,.-...,1-- 1.5 .., ................ ......................... , ......... - ....
:','!;'':'?::::;:,:,,:;;:;:,;;;i:!:;:::;:. ::::::::::::::::::::::::::::::::::::::::::::::::: ;':;ii :;:::,;;:; :i;:.:;i, .:::i?:;.:ii:fiii::'::!;::i:;:::i.i:i:.;:.:...11; :i,
',::": ,':;,:'.:,:.:,;,;::::::'.:; ',,..?".::,.,:';':.'t;-.,,.?.:,,:;:;,.,;;:,::':..;': .i::;:;:.::;;.::::::`.!!:..!;;:::c.":?:>..".:!;;:;::.. :::.. :.,
:,:i;,"':'; " ,.: ': ':,:.!'.,-.::';zi::,,ii',,,i::;i:::.b..ili ':
:e'...e:
'" ............. ;i " ' :'"';,i7,:';" ;"',.,.'r:,.::;'i;
':'.-.'. ,.,:..".;e.,'?_-. ,.-_7.';?".':.. ,... ".,,.'"..,,:? ".'c..-: ',i :': .: ?,.e;.'.;:,%-r:,,', . t, :?::.i,.:.'i:'i'.' :5!f::. ;. ::::...'7.. !::
,- , .7,'".,' ';,, %'.'..;, .', ;. I' ....... !.".:".'.';;'.',.',,'
. ' ' e...?.'.7'. (sec.) '..i::'::::.':.::'.'?--':::",,.-o..',...'::
, . ..,. ,d.
....... .....
.,.:,,., .,;; ,.$..::!.?.r:.i:; i
4.5 ";'lZ6.'_ 'knr..' '?,::::,,.%_'
Elevation+Residual Statics Refraction+Residual Statics
Fig. 8. A comparison of influences, on stacked all reflections improved by the application of
data quality, of the earlier and present near- refraction and residual static corrections.
surface static corrections. Lateral correlation of
processes. Final parameter selections were based presentation (Figure 10a), the increase in signal
on the level of tie correlation of arrivals at the amplitude levels and continuity becomes very
intersection of the two survey lines (Figure 9). evident. The marginally traceable subparallel
As expected, the correlations of corresponding reflections of Figure 10a are, in reality,
events are excellent only in the neighbourhood of portions of a number of complex diffraction
the intersection since the directions of the two patterns. It is clearly illustrated after
profiles are quite different and the point of reprocessing that the majority of these events
crossing is located in a region of significant were generated at segments of highly irregular
structural disturbance. discontinuities. Signal levels are hardly
When the reprocessed upper segment of zone H perceivable in the lower crustal section of zone
(Figure 10b) is compared to its original H of line 1 (Figure 11a). After reprocessing
Original Sequence
TABLE 2 Processinq parameters
Reprocessing Sequence
- Demultiplex
- Amplitude recovery: 2.0 sec. AGC
- Geometry: Crooked line
- Elevation statics:
Datum=400 m ASL
- Mute: First breaks
- CMP gather
- Velocity analysis / NMO
- Trim statics
- Stack: 64 fold
- Trace equalization: Time variant
- Filter: Bandpass 8/12 - 52/62
- Demultiplex
- Trace edit
- Amplitude recovery: 2.0 sec. AGC
- Geometry: Crooked line
- Delay correction: -16 ms static
Traces 121-240
- Refraction statics: Two-layer model
Datum=440 m ASL
- Instrument dephasing
- Predictive decon: Operator=600 ms
Prewhitening=0.01%
Lag=2nd crossing
- Filter: Bandpass 18/20 - 45/47 Hz.
- Filter: FK reject 1500 - 5500 m/s
- Mute: First breaks
- CMP gather
- Velocity analysis / NMO
- Residual statics (max. 16 ms)
- Automatic trace edit: Threshold=0.25
- Stack: 60 fold
- Trace equalization
- Filter: Bandpass 18/20 - 45/47
- Migration: Finite difference 24 step size
- Filter: Coherency - 9 trace group
ASL = Above Sea Level
LINE INTERSECTION
Line 6 Line 1
NW SEN
Fig. 9. Intersection of line 1 and 6. Correlation in the direction of the lines and complexity of
of events are directly visible in the centre of the subsurface structures along both profiles,
diagram where the lines cross. Due to differences dissimilarity is increasing with increasing offset.
(Figure 11b), similar to the overlying regions diffraction bow-tie arrival forms, from the near-
(Figure 10b), hyperbolic diffraction events appear surface to beyond 4 seconds, illustrates that most
across this segment of the section from 7 to of the energy has reflected from a large number of
beyond 10 seconds. The interlacing of these buried focuses scattered through all levels of the
patterns reveals the extent of abrupt acoustic upper crust. Understanding of the true nature of
irregularities to deep, lower crustal depths. these reflections most likely requires not just
The intricacy of the subsurface structural comprehensive migration of this data set but
setting of the KSZ is best demonstrated by the establishment of the out-of-plane characteristics
reprocessed segment of the original zone G (Figure of these features as well. According to present
12 a and b). The complex arrangement of evidences [Percival et al., 1989], the development
CMP
HAJNAL ET AL. 427
o
CMP
Fig. 10. Enlarged 1 to 4.5 s segment of zone H of line 1. After implementation of the
original processing parameters (a) only poorly traceable and barely visible NW dipping
linear arrival patterns can be recognized, mainly on the right side of the figure.
Reprocessing (b) was able to enhance amplitude levels throughout the section and led to
the exposure of the true hyperbolic characteristics of the interfering diffraction arrival
patterns.
lO
lO
Fig. 11. 7 to 10.5 s portion of zone H of line 1. Original processing (a) uncovered only
faintly distinguishable reflectivity on the right half of the diagram. Reprocessing (b)
again makes the complex diffraction arrivals visible across the entire section. A large
complex buried focus appeared in the lower central segment of the section beyond 9.5 s
(depth in range of 30 km). A side view of this example more clearly illustrates the
complex interfering properties of the diffraction patterns.
CMP
HAJNAL ET AL. 429
CMP
Fig. 12. 1 to 4.5 s portion of zone G, line 1, between 1400 and 2200 CMP points.
Insufficient processing (a) produced misleading and simplistic looking arrival sets.
Between CMP 1500 and 1700, events appear subparallel, no coherence from 1700 to 2050 CMP
and inadequately developed diffractions are visible at the right margin. Enhanced
processing (b) brings to light several sets of bow-tie forming diffractions. The entire
section is dominated by interfering events.
Time (sec.)
I I I I I I I
HAJNAL ET AL. 431
of these structures occurred prior to 2.45 Ga ago. events, the upper 5 seconds of the section can be
It is very intriguing, therefore, to perceive that subdivided along the subparallel lines of C, D and
these features are preserved at great depth, that E. Near the south end of the profile, most of the
seismic imaging can recognize them and, very above reflectors either abruptly terminate or
importantly, proper survey and signal enhancement change attitude at a set of near-vertical faults
procedures can resolve them. located in the vicinity of the intersection with
line 6. These steeply dipping faults, H 1 and 2'
Discussion cut the central region of the crust causing
visible changes in the dips of most of the
Structures emerging after the migration of line reflective zones. The entire regional pattern of
1 are dominated by variable but consistently reflectivity is continuous under the area of
north-northwest dipping subparallel reflectors surface exposure of the ILFZ. There is no
(Figure 13). At the south end of the line, where immediate comparable surface geologic
the survey crossed rocks of the Abitibi belt, a manifestation for the sudden absence of
recurrent set of steeply dipping subparallel reflectivity in the top 5 seconds between CMP 2100
reflectors prevail from near-surface to beyond 10 and 2450. Toward the north, beyond this zone of
seconds. Continuity of these reflections is transparency, distinct bands of reflections
excellent, specially from 1.1 seconds to 5 persevere from the near-surface (A) to over 5
seconds. Following sudden alterations in regional seconds (E), more or less through the remainder of
attitude and amplitude characteristics of these the sections. Across significant distances,
:- GRANITE
r---1 TONALITIC GNEISS
:,,'- MAFIC GNEISS
PARAGNEISS
TONALITIC GNEISS
' ::?.: PARAGNEISS
MAFIC GNEISS
-- DIORITE
t----5 km-- Line 6: Migrated
Fig. 14. Interpreted and migrated section of line 6. H, I"-I"
and J refer to the original interpretation of Percival et al.
[1989]. Unmarked subhorizontal events image features generated
by the subsidence process.
lateral continuity of many reflectors can be comparable ramp-and-flat geometry reveals a
established by visible correlation, particularly definite inter-relationship with the overlying
in zone D-E. The absence of coherent events at imbricate thrust system. In addition, all the
the extreme north end of the section are mainly identified events tend to dip toward the centre
the result of interference created by the region of the structure where the refraction
migration generated noise. survey [Boland and Ellis, 1989] imaged about a 5
The overall structural intricacy outlined on km increase in crustal thickness. The relative
line 1 (Figure 13) is reproduced with remarkable motion along faults H 1 to H 3 (Figures 13 and 14)
fidelity by the comparably processed version of also exhibits a comparable down sliding of the
line 6 (Figure 14). Major reflectors such as D hanging walls demonstrating a form of subsidence
and E tie to the similarly marked events of line as a consequence of the upward thrusting of middle
1 at the intersection of the two profiles within to lower crustal rocks in the KSZ.
a few millisecond accuracy. A shallow high angle
fault is also recognized on both profiles at the Conclusion
crossing. The southeast end of profile 6 is also
dominated by a strong band of relatively steeply The reprocessed sections of lines 1 and 6
dipping reflectors from the near-surface down to revealed many previously unrecognized features and
9 seconds where they become masked by migration structural complexities of the Kapuskasing
noise. The existence of these very deep events Structural Zone. These profiles demonstrate that
was established from the unmigrated segment of the systematic, data-oriented signal enhancement
data. In the central portion of the section, the procedures can lead to high quality seismic
attitude of all horizons changes to near sections in the difficult seismic environment of
horizontal. It is clearly evident that rocks the Canadian Shield.
which form the upper segment of the crust in the Both profiles imaged a number of low angle
Abitibi belt extend through the entire line. southeast verging thrust fault systems which have
Comparably to properties observed at the south end a ramp-and-flat geometric configuration. The
of line 1 (Figure 13), deep crustal faults (, H ,
H3) also fragment the interior portion te geometry of these imbricate faults are very
comparable to ones which are observed in younger
crust beneath line 6 (Figure 14). If these deep sedimentary fold and thrust belts.
faults of the two profiles are interconnected, as The mid and lower crustal reflectors (D,E) and
their spatial positions suggest, it would appear the mid-crustal high angle faults (H to H ) reveal
that a curvilinear, subparallel crustal fault that the entire crust in the KSZ was af
cted by 1
system exists east of the ILFZ in the contact the thrusting process. The uplift and
boundary zone of the Abitibi belt. transportation of denser, high velocity material
The complexity of the structures revealed by on the top of the crust produced an anomalous load
the two migrated seismic sections distinctly which was subsequently compensated by a form of
indicate that the Kapuskasing uplift is a subsidence under and to the west of the Chapleau
laterally changing feature. Comprehensive three- Bock at the southern end of the Kapuskasing
dimensional understanding of the entire structure structure. The remnant of this root is exposed by
of the region will require an examination of the thickening of the crust in the area from 40-43
similarly processed reflection data from all to 50-53 km [Boland and Ellis, 1989].
survey lines of the experiment. However, the The complex low angle imbricate thrust system
predominant north-northwest trending dip of the that has emerged from this analysis differs
reflecting horizons on the present sections, as significantly from the earlier models of Percival
well as on profiles 2, 3 and 4 [Geis, et al., et al. [1989], Cook [1985] and Percival and Card
1990], and the characteristic pattern of [1983] but it displays considerable similarity to
reflection bands on these data sets, reveals gross structures presented by Geis et al. [1990].
similarities throughout the southern half of the The earlier interpretation of the domal
Kapuskasing structure. Reflections A, B and C are structures (Zone G, Figure 3) within the Wawa
also and consistently associated with reflectors gneiss terrane rocks [Percival et al. 1989] is not
which divide lithostructural units with differing substantiated by the reprocessed migrated data
geometry. sets.
A, on both sections (Figures 13 and 14), appear The intricate deformation comprised of tight
concordant with the geometry of the overlying folding and faulting is very likely responsible
surface rocks (paragneiss, line 1; tonalitic for the seismically transparent zone at the
gneiss, line 6). Therefore, it likely represents southern end of line 1. More elaborate migration
a fault bounded stratigraphic contact. It procedures than were implemented here may resolve
projects to the surface at the northwestern margin some of these complex features.
of the ILFZ along line 6. The same association
with this major structure is not recognizable on Acknowledqements. This project has been
line 1. The east verging subparallel and north- supported from funds provided by the Canadian
northwest dipping events A, B and C are Lithoprobe Program.
interpreted as low angle imbricate thrust fault
planes of the KSZ. 'C' appears to be the References
easternmost detachment of the fault zone.
Association of this fault system with the surface Boland, A.V., and Ellis, R.M., Velocity structure
exposure of ILFZ is very evident on Figure 14. It of the Kapuskasing uplift, Northern Ontario,
is not as clear at the south end of line 1 (Figure from seismic refraction studies, J. Geophys.
13). If C is extended under the ILFZ, as it is Res., 9--4, B7189-B7204, 1989.
marked, then it is a blind thrust projecting into Burke, K., and Dewry, J.F., Plume-generated
the Abitibi belt. This scenario would indicate triple junctions: key indicators in applying
that in this southern segment, the fault which plate tectonics to old rocks, J. Geol., 8--1,
marked the base of the uplift has not yet been 406-433, 1973.
recognized. Cook, F.A., Geometry of the Kapuskasing structure
Reflections D and E are mapping structures from a Lithoprobe pilot reflection survey,
below the entire Kapuskasing structural zone. The Geoloqy, 1--3, 368-371, 1985.
HAJNAL ET AL. 433
Ernst, R.E., and Halls, H.C., Paleomagnetism of
the Hearst dike swarm and implications for the
tectonic history of the Kapuskasing structural
zone, northern Ontario, Can. J. Earth Sci.,
21, 1499-1506, 1984.
Fountain, D.M., Salisbury, M.H., and Percival,
J., Seismic structure of the continental crust
based on rock velocity measurements from the
Kapuskasing uplift, J. Geophys. Res., 95,
Bl167-Bl186, 1990.
Geis, W.T., Cook, F.A., Green, A.G., Milkereit,
B., Percival, J., and West, G., Thin thrust
sheet formation of the Kapuskasing structural
zone revealed by Lithoprobe seismic reflection
data, Geoloqy, 18, 513-516, 1990.
Gibb, R.A., A gravity survey of James Bay and its
bearing on the Kapuskasing Gneiss Belt,
Ontario, Tectonophysics, 45, 7-13, 1978.
Gibbs, A.K., Seismic reflection profiles of
Precambrian crust: a qualitative assessment
in Reflection Seismoloqy: The Continental
Crust, edited by M. Barazangi and L. Brown,
Geodynamics Series 14, Amer. Geophys. Union,
pp. 95-106, 1986.
Hampson, D., and Russell, B., First-break
interpretation using generalized linear
inversion, C.S.E.G. Journal 20, 45-54, 1984.
Innes, M., Goodacre, A., Weber, J., and
McConnell, R., Structural implications of the
gravity field in Hudson Bay and Vicinity,
Can. J. Earth Sci., , 977-993, 1967.
Martin, J.E., and White, R.E., Two methods for
continuous monitoring of harmonic distortion
in Vibroseis signals, Geophys. Prosp., 3--7,
851-872, 1989.
Mayrand, L.J., and Milkereit, B., Automatic
editing and true-amplitude stacking of seismic
data, Can. J. Earth Sci., 25, 1811-1823, 1988.
Milkereit, B., Spencer, C., and Mayrand, L.J.,
Migration and amplitude analysis of deep
seismic reflection data: processing results
of CCSS data sets II and III, in Studies of
Laterally Heteroqenous Structures Usinq
Seismic Refraction and Reflection Data, edited
by A.G. Green, Geo. Survey of Canada, Paper
89-13, pp. 151-164, 1990.
Percival, J.A., A field guide to the Kapuskasing
uplift, a cross section through the Archean
Superior Province, in Exposed Cross Sections
of the Continental Crust, edited by M.H.
Salisbury, NATO Advanced Study Institute, in
press, 1990.
Percival, J.A., and Card, K.D., Archean crust as
revealed in the Kapuskasing uplift, Superior
Province, Canada, Geoloqy, 11, 323-326, 1983.
Percival, J.A., Green, A.G., Milkereit, B., Cook,
F.A., Geis, W., and West, G.F., Seismic
reflection profiles across deep continental
crust exposed in the Kapuskasing uplift
structure, Nature, 342, 416-420, 1989.
Watson, J., The origin and history of the
Kapuskasing structural zone, Ontario, Canada,
Can. J. Earth Sci., 17, 866-876, 1980.
Wilson, H.D.B., and Brisbin, W.C., Mid-North
American ridge structure, Geol. Soc. Amer.,
Spec. pa D . 87, 186-187, 1965.
Yilmaz, O., Seismic data processinq.
Investiqations in Geophysics v. 2, edited by
S.M. Doherty, pp. 147-151, Society of
Exploration Geophysicists, Tulsa, 1987.
COMPARISON OF COINCIDENT HIGH-RESOLUTION
WIDE-APERTURE AND CDP PROFILING ALONG THE
SOUTHWEST COAST OF NORWAY
S. J. Deemer and C. A. Hurich
Institute of Solid Earth Physics, University of Bergen,
Allegaten 41, N-5007 Bergen, Norway
Abstract. Comparison of coincident wide-aperture and
CDP data recorded in southwestern Norway shows that
distinct crustal features defined by each data set may have
little correspondence. In some areas where the upper crust
is seismically transparent in near vertical incidence images
wide-aperture data indicate there are distinct and extensive
velocity heterogeneities. A pervasive upper-mid crustal
velocity boundary interpreted from WA reflections generally
does not correlate with CDP reflection horizons. Conversely,
areas of dense reflectivity in CDP profiles may not be
imaged at wide angle. Moho interpreted from WA
reflections can correspond with an abrupt lower crustal
reflectivity boundary within several hundred ms or less over
broad (100 km) areas but clearly does not always coincide
so closely, particularly in areas with structural complexities
in the Moho.
using an airgun source and land receivers have been
collected recently by different scientific groups in diverse
geologic environments including: the North American Great
Lakes [GLIMPCE seismic refraction working group, 1989],
profiles parallel to the southwest coast of Norway
[Norwegian Lithosphere Project data presented in the
following], east coast of Britain [West and Long, 1989],
TACT and EDGE [Brocher and Moses, 1990, Moses et. al.,
1989] profiling in Alaska, the Bothnian Bay [BABEL
Working Group, 1990], the east coast of Greenland [Gohl
et al., 1990], western Canada [Hole et al., 1990], and the
coast of Oregon and California [Trehu et al., 1990, Moses
et al., 1990]. These data sets represent part of a rapid large
increase in coincident profiling experiments as foreseen by
Mooney and Brocher [1987] in a recent compilation of such
work.
Introduction
Experiment
There has been much discussion in recent years about the
value of coincident wide-aperture (WA) and CDP profiling
as the two data sets provide complimentary information
about the area sampled by the wavefield [Mooney and
Brocher, 1987, Braile and Chiang, 1986]. High resolution
wide-aperture data sets employing large airgun arrays and
land recorders coincident in geometry with marine CDP
profiles can create much better possibilities for constraining
concepts of the nature of the crust than either method by
itself. The ease and inexpense of obtaining very close shot
spacing makes the airgun an extremely attractive seismic
source. Land receivers in arrays with multi-component
seismometers add to the information obtainable and improve
possibilities for signal enhancement processing of WA data.
Strong P-S conversions in areas of thin sediments and hard
water bottom provide a rich source of converted shear
waves. Although rare only a few years ago [Mooney and
Brocher, 1987], many high-resolution coincident data sets
Geodynamics 22
High resolution wide-aperture recording coincident with
CDP profiling was carried out in the Spring of 1987 by the
Institute of Solid Earth Physics, University of Bergen,
within the framework of the International Lithosphere
Project [Hurich and Kristoffersen, 1988] utilizing the high-
quality acquisition system of the survey vessel Mobil Search
(Figure 1). The source was a tuned wide airgun array of 48
airguns with a volume of 7300 cu in. The streamer was 4.5
km tong and group spacing was 50 m. The wide-aperture
profiles were recorded simultaneously on the University of
Bergen's Western Norway Earthquake Network. All
recordings were on vertical 1 Hz seismometers well
insulated from environmental noise with an analog filter that
standardly limits signal frequencies to less than 12 Hz.
Data
The wide-angle data collected are of high resolution with
a 50 m trace spacing. It was possible to create traces that
were longer (50-70 s) than the approximate shot interval (20
s) from the continuously recorded signal because the shots
were fired on a constant distance interval (50 m) so that the
435
436 COINCIDENT WIDE-APERTURE AND CDP PROFILING
60N

,6000
Bergen
5000
ooo
3000
' L"'--.. Stavanger
Seismometer . v 6OE
60N
Fig. 1. CDP Line ILP-11 (SP: shotpoint) recorded along
the southwest coast of Norway with land stations of the
Western Norway Network used to record reversed. coincident
wide-angle profiles parallel to shore. See Hurich and
Kristoffersen [1988] and Klemperer and Hurich [1990] for
regional geology.
time interval was slightly different for each shot. An
irregular shot interval means that arrivals from later shots
will not be coherent. If shots are spaced at a constant time
interval it would be of little or no use to make traces
longer than the shot interval because coherent arrivals from
later shots would obscure the weaker late arrivals of the
previous shot. An advantage is that small shot intervals
optimal for CDP profiling can be used and the profiles can
be recorded simultaneously. The long traces allowed
inspection of S-phases on the vertical seismometer
recordings at offsets up to 250 km. The high resolution of
these wide-angle data sets provides much more detailed
information than standard land refraction surveys,
particularly when the S-phases are identifiable. The best
quality wide-angle profile recorded on the Karm0y station
(Fig. 1) is presented in Figure 2a with a line drawing of
major phases from the reduced version (Fig. 2b). The WA
data have been processed with a 4-I stack of adjacent traces
after NMO correction with a single average velocity (6.4
km/s) designed to flatten the Moho reflection. The NMO
correction was removed for display both without (Fig. 2a)
and with (Fig. 2b) a reduction correction. The WA data set
clearly images Pg and Sg (including several secondary
arrivals PcP,ScS), PMP and SMS, and Pn and weak Sn.
Upper mantle P-wave refractions are strong at least to 300
km offset and S-wave refractions are visible over a limited
offset range of 220-250 km.
A line drawing of the unmigrated CDP profile with
features from the WA model superimposed is displayed in
Figure 3. The CDP data were recorded to 16 s two-way-
time with a 50 m trace spacing. Standard processing for
enhancement of shallow and deep events has been applied
to the data [see Hurich, 1991]. The upper 3-4 s is generally
DEEMER AND HURICH 437
featureless, but mid to lower crustal reflectivity varies
distinctly along the line and an abrupt termination of
reflectivity marks a distinct 'reflection' Moho. Details of the
reflectivity will be discussed as the comparison with the
WA data is made.
Comparison of Data sets
A simplified P-wave structural model developed from the
wide-angle data using the asymptotic ray theory technique
20-
25-
KARMOY STATION
OFFSET(KM)
60 7O 0 9O 100 0
PcPb ' ' ' ",,
: :.""!,'. ?':i:': !, ::-i'!. :'":
.: --...,,:-%....... : .
.., ...... ...... - :,. -. :. :: ..-?'.'.'...:;i':.,.:, '..
':;. :... '.": - .:: ' ' ..... *- :'Z.'--:'::- ~- 'p'"
:it ' ::'i.,: '":.-:; -. -':. ' .... ".::..'-.. ::', .. ::-_'i': ..... :.:'-..,. .,
:,;.:-: ,.',,! .-'-,.-.:: -._'- : , --:-'. '. ....... "- '-Y',".,??,'-r.?::"-: .,,,
!:%":: ',:<L:.:. .-: '. :: ' ' ':::';:'.-"- - . .'?-{.:.:.'".2--'.':::-
:.-s:i t'-:. ?:.t.-:-'.::,. ....... ' :-...- . ' ..::. :' -".:":...":-' - -t t_,..:.:.St'_L.? :-:',';'': ..-,."-:: ' I
: .... . ........... , ,'.. .... .... ,-. .... .'x. ........ ; -:.-.,. :... ::.-<... :.::.%-,,x.::,_.,.._<.,-z _.
..... -. '-'.:: !.27& .?::'.,}'.",,..'-.';[' .'_---':...-.':':....':-",'. - '..:'.- :,.- % ':'- :' . .::.: .'!: "'. , ' 'i:: :: ':t: :,:..7.'-:"
'5":!'.?; ':::%iT-....:.',,::':Y:':'.'."i."'::'i.::,',':i:::':..":t?i::::" '.-L ' :..'?..'}::i" .:.,,.-.'- '%:-?'.L:: 5-'.: ':':.'-:T
': ": ?": :t:,i'.:t[:!.,?::'-'::,:::5-:'::;v.'.%-'"-t.." ' ::',::_:.-.-..,: ':.:i"'.'.':'.. : '- -': .'.-" .'- '..::.:"
:":"& ::.sf'"-'.:...,ivur' "'::'.' :,:'.' ';:"'S."5'<..-'" :., :..':."::, -.:. ..... .:':':'.. :." ':-:'.'- .... ::'!
!, Y,-;: ':'; .'..:):"':-:_'";:::. ..: ........ ::'":5 ;':;,:?;7..:i;'":":,..::?::: :::" :.':.:::' '.':" .'.::. ':.:'.'. "... 7.1., ':-' !
:;.,, :.,.'.' :.%v....:':.-.-::':'_..::.i..:.t;';: ;:.'::-.t: :?:.t!' ';:.:'.:', .-..;: ...--:::. :-"'.".i-.--- '?,:..i.' "'-"?.
;.': ?'..': ..':.."::::,t".; :::,'-:.%::':':-.:.'..i:':'.- -:': ..-"--"'::.,, ...... ,' t. "' ...',,.'' "?.'.i.. '"'- .":" :..' .:z,,: ?.....:,: . ' . -: :
,<'.':: .."::: .".;:'..:L'.':g.!:'..L'. .::..-...'.--' .'?..'".. ,-": :<,' '.;.":- ..... -:r Y:r,-:-.,:.?,::...-.'~:"' ;:.'-. :.:.}'.,.., :.'::.
'.:st".', ?', ?..i :':.-; is.,.t!..;.:.?:: :. :: .,:E: _ ,- .i:?.,":?:?..:. ::- ::::.?.'.,:.:.-) .... ,'?. .... 5'":':..." ::. i::, : '.." 'r' :-',:
.r:-.:. < .,? ;:: ';.: .... :,:"z,.'.':::.:,.; .:-:..'..'-.-'-..-.-'.-.-',- ::,:..-{..,.2..-..:. ,..:-:.. :.;.,x ...., , ....... :.z.:. ...:: : ....,.:::..
.:::.t: .:.-t: t:- ..... """?:..:_'<.:-.,..':':'::...-.....:-..:..:.:.:..!q., .,. ..... i.' .... ....."!...-.:i" '..' .!:: .... ;:';:!" i
.-. . ;:5 'r:' ':,,.. : ,;','"'i.'...',,.t:- : .% ..... ". ':. '.". z,.'--ic- ' :,-: :,. ' -4 ." ', .:'.;'-2:. '.7 '1; :r. 'v' :; -
' '-' :' '?' "' S '" :':" "% "'";'""' .-,k'':', '- - -'. , ~-. -. .... 4-., .. .,.,.. ;...;-- ..:': ,." ,'-:.., '
-:: .:.. 2...,.: S M :,..,;;:'::,:..::.::.:,:::......-::'..,::.:.,.. :;t;::.:.,'. .'i-'. . '2...
,':' :'\' ";:.'. :i! -' . ..%::. ::.:'.: ...,.,::'....!,.?.',"q;; 't%.,, ." ...',. ';.),:'" ?,;'"'.. '.' 't'-::'- -:':":'.. L,.: -,..',.'.
"' ?:?:. ;.::t:i.;.'..:::':::.;'.7-.St?".')..'::::'! :. i'!':-:'i;::; :-:.':':: ::.'.--:5': ,:::;'.: ..':.' -'.:: ': ::. :)'.'::::? 't
..:-: .:.::.:...: ..:-.. ..: -.... z:.: .,.:_-.:::,.t...... :'-'.'.''..::'t',z'--'--:...' ',". .;:.' : .... _,. "'" -. : ": ..... . ....
;;' ::::::!.:! ;::.:!'. ::: :i'?,..t: :' :::t::':::'.'...:.:,:,:r'..-,-,:.-,::x:.:::..: -... :,:. -, ..... : .... .:.. ....
:':':i',!;:'.i: ':::.";'!.'"::'5:'>::i. :'-';.':t.: L)i.--: :i".',:...<;:',t;.'::; '.%-::/:,.,.:-..::-: :.' .' ', -'..-.:. .'..:-. ; '. '. :":
,.'." .;. :...:.'-;.o:-..?z:',t'.:-,,':.t,';, ..: t.'...', .;,'. '." ." .jt"..,-,,.,c'.'...-'.", - , .':.--. ? q'.- '' :::. . '.. ','.. : ;.
:L.i.: '.-.:. ',:ii ';:.?'!:.'. ': 7 ;: .::'.t, '.:-'.:.i-:..,,.."--i:-it '.'.z'_,:(g"'t.'-.' :..i,,-,t'..:,;:..; t.'$-. :' :. t' .... :'... ''?' '":' '"' "i":i !
5 ..... ".'. ,.'!.:-..':!: ' , '.- ' .-:.-.: .-'-: ,':':7: ',<.!..;-i:'.:X-':'::'.' :':',:::.-': "'?'-5,-'.K,?,,, ," . ,h,. : .. :'-. '..-
..... "..,:.. ':. . ..- ..:::, , . ".. - r ' :,-.- ,:.. :--.'.:: - ::,,,- '- ::;,.- z'-..' .,- >,<:-..N,',2.:e.,,.,',-.'. ...- .,,'-' "-;<-.'- '. ' :'-t
'-: :: ', !.. :.', :z ,:-..' '." :-.: i-:.::.' .... . :.:tt "'.'?, './, ':t:' '::'- ": '3 "' ' ":.)'_q:'.?..'.:"L".'.:. ';:.x . ." ':'.' . ':":.::..
!:? .'"i; 5:!i:':,?.' :;.'!?:.;'!,71.; ..;'. '5:1':!':. ,:..:..: - - i- :.::. :. :;:":"' .< ::"-t :--.-.-.' .: :?: .,.'.,.%
.:: :z.t . :'t,::;. ':i.:'!':i:": .... i?.-..'",. ': "?':: '-: ;:! .:.:!i: .':...i: :: t':. ":'-. :..-':,,:': :',',, .,.':
?:.. :.:':-..-b'/.'...:.' .:5.-:..,..:..:.!: .;.:.;:: - . .:..!:.......-.:.. '.[:. '; . ., , : :..-,, ,,
:.. i.,.; :,,'2?[ ;.'. .a;.- '': ::r ": ':;' :- "..' ,' .' ' ':-.;,- ';.',.: ";..r' 't. : : *. '.',t ' ",: ' .... -'
Fig. 2. Wide-angle profile recorded on the KarmOy land
station. a) True amplitude plot over limited offset range
showing major events. Time axis unreduced. b) Line
drawing of major phases. Time axis reduced by 6.0 km/s.
o
-7
-2.
8.
,,
:
18'
23.
28.
Distance (km l
5,0 100 150 O0 250 300
........ i .... i .... , .... , i i i i -7
Pg . ---- _._.
c . SS
SS
18
23
.28
Sn
Fig. 2. (continued)
[Pajchel, 1989] is overlain on the CDP profile line drawing,
to illustrate the combined information (Fig. 3).
Interpretations of deep crustal CDP data rely on changes in
the reflectivity patterns to characterize changes in crustal
material. Reflectivity is commonly used to define tectonic
boundaries, different crustal domains typified by changes in
bulk crustal reflectivity patterns and the crust-mantle
transition [eg. DEKORP Research Group 1990].
Determination of deep crustal velocities from standard
marine CDP data is impossible because of the limited
recording aperture, making wide-aperture recordings
imperative for a full understanding CDP data. Velocity and
Poisson's Ratio information constrains compositional
interpretations and helps to distinguish whether or not
reflectivity changes represent major compositional
boundaries.
Upper Crust
The wide-aperture data profiles have shingled,
discontinuous first arrivals suggestive of velocity reversals
in the upper crust (0-10 km) (Fig. 2a and b offset 40-80
km). Several events are recorded at pre-critical offsets but
most phases from the upper crust are discernable at near-
and post-critical offsets. The lack of reflected signal in the
short offset range can be a result of small reflection
coefficients, but a lack of continuous reflectors in the
vicinity of the receiver is an alternative possibility. Some of
the shingled events have been modelled with planar surfaces
tens of kilometers long (Sh, Fig. 3) suggesting that small
reflection coefficients are more likely than a lack of
continuous reflectors. The low average velocity in the upper
crust here, as interpreted from the 6.0 km/s group velocity
of the shingled arrivals, requires that there be thin relatively
high velocity layers embedded in a lower velocity crust.
This contrasts with the interpretation of similar data by
Mereu et al. [1990] that showed that shingled arrivals in
high resolution data from the GLIMPCE project could be
qualitatively explained by moderate increases in velocity
producing reflections visible only in the near-critical range
where amplitudes are highest.
The CDP data in the area of clearest and densest
shingles (SP 4800-3800, Fig. 3) shows almost no reflectivity
except for an unrelated dipping package interpreted as a
shear zone (Hardangerfjord shear zone:HFSZ) [Hurich and
Kristoffersen, 1988, Klemperer and Hurich, 1990]. The shear
zone coincides with the northern termination of the shingle-
producing features. The recording environment along the
coast with variable thickness of low-velocity sediments (0-
300 m) overlying high velocity bedrock was found to be
extremely problematic. What is seemingly a transparent
upper crust as imaged by the CDP data may actually be
reflective, but because of the near-source environment,
reflected energy may not be visible over the source-
generated noise [Hurich, 1991 this volume]. The wide-
aperture data contains water bottom reverberations, but is
not affected by most of the noise problems in the CDP data
because of the long offsets and the receiver location on land
instead of in the water. The vertical-incidence reflection
coefficients may also be small as the shingled events in the
WA profile had no observable reflection energy at small
offsets.
Comparison of the data sets also illustrates the limitations
geometry imposes on wide-angle imaging of features such
20
40
6o
lOO
12.o
Solund
S P 7500 7000 6500
Karmoy
HFSZ
6000 5500 5000 4500 4000 t' / 3500
------ _ -- .,-_____ _ _ ?-
_ _ -.: ...... . ,, _...:-...__-...._ r '"h --
....r. ..... ,.% .: '- MI - --: ............... ---- ;.:7. ...... ''",,,
Scale: 0 10km
Une:ILP-11
Fig. 3. ILP-I l line drawing between Solund and Karm0y
stations at approximately 1'1 (assuming average 6.0 km/s)
<N
with superimposed features from wide-angle modelling.
Labels described in text.
as the upper crustal dipping reflection packages in the CDP
data (HFSZ, Fig. 3). Shear zones are particularly difficult to
image with wide-angle data [Mooney and Brocher, 1987]
without more closely spaced receivers providing reversed
data for the shallow crust.
Shallow anomalous bodies with seismic properties of
mafic intrusions (Vp=6.8) (MI Fig. 3) were modelled from
the WA data because of a spatial coincidence with strong
and steep gravity and magnetic anomalies and the presence
of nearby gabbroic intrusions on land [Sigmond et. al.,
1984]. They produce strong distortions of arrivals in the
wide-angle data such as the locally very high apparent
velocity obtained on some first breaks (app. 40 & 80 km
Fig. 2b) but they are not imaged in the CDP profiles. The
bodies are well constrained laterally, but thickness and
velocity are arbitrary trade-offs as the model is constrained
only by a time delay. The depth position is also
approximate, but evidence from both the seismic and
potential field data limits them to quite shallow. In contrast,
strong diffractions in the CDP data (D Fig. 3) above and
below the HFSZ indicate other possible anomalous bodies
or sharp structural discontinuities not evident in the wide-
angle data.
Upper-Lower Crust Transition
The WA wavefield contains a reflection/diving wave
from a rapid increase in velocity from about 6.1 to about
6.4 km/s representing a fundamental mid crustal boundary
(Fig. 3 ULCT, Fig. 2b PcP). This change is not imaged in
the vertical incidence data in the south where the midcrust
is essentially nonreflective. The well-defined lower crustal
reflective package (lc 1) is considerably deeper than the mid-
crustal velocity jump. This is in contrast to the general
trend Wever [1989] found in a worldwide survey of nearly
coincident wide-aperture and vertical incidence profiles
where a mid crustal velocity jump interpreted from WA
data usually occurred deeper than the onset of significant
reflectivity in nearby CDP data. At the north end there may
be a rough correlation between the velocity jump and the
onset of significant discontinuous reflectivity, but the noise
studies mentioned previously indicate that the onset of
middle crustal reflectivity may be an artifact of the
acquisition. A striking correspondence does seem to occur
over a limited area at about $P 5800 where the velocity
boundary defined by wide-angle reflections abruptly
changes depth at the same place where the onset of
significant CDP data reflectivity changes (Fig. 3). Since the
WA reflection at both ends of the profile was not imaged
at near-zero offset, although visible pre-critically, one could
infer that the transition is gradual rather than abrupt along
most of its length and therefore not directly imageable with
the CDP technique.
Lower Crust
In the WA data no strong laterally continuous reflections
were recorded between the mid crustal arrival associated
with a rapid increase in bulk crustal velocity and the Moho
reflection. The apparent velocity of Pg increases smoothly
until no longer distinguishable from the post-critical PMP but
remains low (< 7.0 km/s) out to 250 km offset where PMP
arrivals are no longer visible.
Crustal reflectivity varies distinctly in the CDP profile.
South of SP 4000, a dense zone of reflections occurs
between 8 and 11 s. North of SP 4000 there is little crustal
reflectivity besides the area around the HFSZ and near
Moho to about SP 5300. North of SP 5300 zones of
discontinuous reflections (lc2) alternate with transparent
areas and north of SP 7000 reflections are distributed more
uniformly throughout except for a transparent zone (lc3)
over a thin package of reflections at Moho (lc4) depth.
The discontinuous reflectivity imaged throughout the
profile length by the CDP method is manifest in the WA
data by low amplitude events crisscrossing the profile with
variable apparent velocity and limited lateral extent (2-15
km). Short reflectors (usually < 1.5 km) may in general be
hard to see at wide-angle except in cases of extraordinarily
dense occurrence and highly contrasting properties such as
in the Black Forest [Gajewski and Prodehl, 1987, Wenzel
et al., 1987]. Finite difference modelling [Gibson and
Levander, 1988a] of the Black Forest data shows that
velocity contrasts of very limited lateral extent (5 2.0 km)
can produce diffraction events over a large offset range in
wide-angle recording which should be correlatable with high
resolution data. The Karm0y WA profile had far fewer
arrivals than the Gibson and Levander [1988a] model which
may partly be a function of a less reflective crust but also
may be related to noise levels. Complications which arise
in the analysis of real data such as a ringy wavelet, water-
column multiples and mode conversions may prevent
recognition of wide-angle reflections from small bodies if
the velocity contrasts are moderate or low.
Moho and Upper Mantle
The Moho is strictly defined in terms of wide-angle
reflection and refraction arrivals but CDP data is now often
interpreted as having a vertical incidence reflection Moho
defined as the base of the latest package of reflections or as
the latest extensive thin band of reflectors [Klemperer et al.
1986]. Both types of boundaries are seen in the CDP data
from Norway. Package lcl (Fig. 3) is an example of a
reflective zone which has an abrupt cutoff and the latest lc2
to the north ($P 6000) is an example of an isolated thin set
of arrivals which occur deepest in the section. Coincident
high resolution WA data has not usually been available in
previous studies [Mooney and Brocher, 1987] to make a
careful comparison of the Moho to the vertical incidence
reflectivity. In Figure 3, the Moho defined by the wide-
aperture data (M) generally coincides with the base of the
lowermost reflective package along the line. The weak
events below the dominant cutoff in reflectivity are often
dipping and migrate to within or very near the base of the
reflective package and those that are flat are few and
isolated compared to shallower reflectivity patterns. An
exception occurs in the south around $P 4500-3800 where
the unmigrated CDP data has a large-scale 'bow-tie' feature
and the correspondence with Moho is not clear in the
unmigrated data. South of SP 4000 Moho appears to be a
bit shallower than the base of the dense lower crust
reflectivity. In such areas this relationship may be expected
because of multiple scattering [eg. Gibson and Levander,
1988b] producing vertical incidence energy later than the
last primary reflection.
In the north, the WA Moho reflection (profile not
shown) is visible beginning at offsets of about 60 km with
a critical distance at 90-100 km which when combined with
a model depth of about 30 km is indicative of a thin crust-
mantle transition. Braile and Chiang [1986] have illustrated
clearly how the highest post-critical amplitude occurs at
greater distances as the crust-mantle transition thickness
increases. Shear waves are reflected from the Moho and a
PMP multiple, but no PMS, which is evidence that the
boundary is thin but not sharp. The correspondence with the
CDP reflection Moho is very close in the north except
perhaps the apparent Moho step at SP 6700 where the WA
reflection is modelled better by a gentle slope.
At the southern end the image is more complicated. A
very weak WA reflection arrival (PlcP Fig. 2b), similar to
the Moho reflection in the north but visible beginning at 35
km offset, comes from a boundary unimaged in the CDP
data and earlier than the deepest significant vertical
incidence reflectivity. At about SP 4250 a strong reflection
is observed (PP* Fig. 2a) from the vicinity of the crust-
mantle boundary and near the change in Moho depth seen
in the CDP data. The distinctive event is much stronger
than PlcP and is too strongly curved to be a reflection from
a subhorizontal interface. Recent reports of other high
resolution data also describe events with similar character of
strong amplitude and large curvature associated with an
abrupt change in depth of the CDP reflection data Moho
[West and Long, 1989, Heikkenen et al., 1990]. In the
Karm0y data set the arrival times are most closely fit with
a reflection from a dipping interface below a shallower
boundary. Results of raytrace and finite difference modelling
up to this point show that the anomalously high amplitude
in the later branch cannot result from a simple boundary
which is kinked or offset. There must be an anomolous
velocity contrast. The large amplitude is explainable by
various scenarios in this environment such as tuning in an
underplated body or shearing and serpenfinization or
eclogitization. The CDP data in the vicinity of the Moho
step has an approximately 1 s thick dipping package of
discontinuous events with no anomalously high amplitudes
which occur after the Moho arrival. Whatever the source of
the anomalous P-wave reflection, it also reflects the shear
energy (SS' Fig. 2b) and a PP multiple (PP"), but there
is no sign of a P-S conversion. The mantle is basically
transparent in the CDP data except for a few short
subhorizontal events (ml and m2) that are similar in lateral
extent to the crustal reflections and so are probably not
extensive enough to be imaged by wide-aperture data. Pn-
amplitudes are very weak relative to later arrivals, but at
offsets greater than 210 km in the Karm0y data set the
refracted mantle P-wave amplitudes increase and are larger
than any later arrivals. This can be explained by a change
in the velocity gradient beneath Moho so that the weak
refracted wave resembles a head wave in a constant velocity
layer at the crust-mantle boundary but deeper acts as a
diving wave where there is a weak gradient. A weak S-
refraction from the mantle (Sn) is visible over a short offset
range (220-250 km) illustrating further the importance of
high resolution WA data because Sn has been seldom
imaged in land refraction surveys [Gajewski et al. 1990].
Poisson's Ratio
The recording of shear waves in the high resolution
wide-angle data provide more information about bulk
composition through the calculation of Poisson's ratio ()
[eg. Holbrook et al. 1988]. The ability to interpret shear
waves is increased in two ways: the close trace spacing
makes the typically weaker and more irregular shear waves
easier to correlate and the waterborne source can often
produce greater shear wave energy in the form of
waterbottom conversion than explosives on land. The
interpretation of ( provides large-scale compositional
estimates which are impossible to infer from CDP data [eg.
Holbrook et al. 1988]. Figure 4 shows the digitized shear
wavefield reduced by 3.46 km/s with the time axis divided
by 1.732 overlying the primary wavefield reduced by 6.0
km/s. The plot gives a qualitative look at the variation of
Poisson's ratio in the crust. If the crust and upper mantle
everywhere had (=.25 there would be an identical
correspondence between shear and primary wave arrival
times. S-waves in the shallow crust come in several hundred
ms earlier than P-waves, indicating a ( considerably less
than .25. PMP and SS arrivals at post-critical offsets
correspond more closely indicating an increase in ( with
depth. Sn arrivals come in before Pn suggesting that the
upper mantle cannot be more than (=.25. In the data from
Norway, the low values of ( (.24-.26) coupled with
moderate P-velocities (6.0-6.8 km/s) are representative of
quartz-rich felsic (granite) to mafic (tonalite) gneisses. The
ability to better identify crustal reflections in the P- and S-
wavefields increases the possibility of imaging blocks of
distinctly different (, such as has been done in Germany,
[Holbrook et al. 1988] but on an even finer scale.
Discussion and Conclusion
Benefits of recording wide-aperture data with stationary
receivers on land and a marine airgun source include
repeatability and good coupling of the source, although
rough water bottom topography and variable sediment
thickness, such as along the coast of Norway, can have
variable affects on signal quality. Where there is a strong
water- or sediment-basement acoustic impedance contrast
abundant shear waves can be generated by P-S conversion.
The airgun source is inexpensive and convenient to use
compared to explosives and is environmentally more
acceptable. Profiles can often be made straighter and have
more flexibility in orientation although lines at an angle to
the shore may need to be recorded on an OBS. Recording
o
-7.0
-45,
0.5.
rj"J 3.0.
,5.5
o,-.d
8O
Distance (km)
50 100 150 200 i . . . i .... i , , , , i
250 300
...... Shear waves
-- Primary waves
Fig. 4. Wide-angle P-wavefield reduced by 6.0 km/s with S-wavefield superimposed after
reducing by 3.464 km/s and dividing time by 1.732 for relative comparison of arrivals
assuming Vp/Vs=l.732 (Poisson's ratio= .25). See Fig. 2a and b for identification of
arrivals.
on an existing earthquake network has the advantage of
seismometers that are usually well-shielded from
environmental noise, but of course there is no guarantee that
the locations are optimal for geologic targets, and they may
not have the benefit of arrays and multi-components. Using
the earthquake network in western Norway provided very
inexpensive and easy to record data. Near coastal areas,
network seismometers of this sort can be used for affordable
wide-angle reconnaissance work.
A small shot spacing makes the correlation of weak
events possible. Errors in modelling can be reduced by the
more certain phase identification and onset picks and in
some cases recognition of relatively weak events that cannot
be identified in wide-aperture data with large receiver
spacings. In order to properly constrain the interpretation of
more details in the WA seismic image of the upper crust
smaller receiver spacing than was available here is
necessary. High resolution data, particularly if recorded on
in-line arrays, provides signal enhancement opportunities
through various stacking schemes. Small trace spacings
provide data suitable for more sophisticated analysis
techniques such as WA migration. More detailed models are
needed in answering some of the questions concerning
crustal structure arising from seismic observations such as
the composition of crustal regions with variable reflectivity
patterns or transparent zones bounded by reflections or the
nature of the crust-mantle boundary. Higher resolution helps
to more accurately pick S-wave arrivals which are typically
less coherent than P-waves. Whenever more secondary shear
arrivals can be picked, Poisson's Ratio can then be
calculated in more detail for inferences about crustal
composition.
Observations made in the KarmOy wide-angle recordings
that perhaps would not have been recognized in more
traditional land recordings with widely spaced receivers
include Pg and Sg shingles and a weak reflection from near
the base of the crust which merges with a strong reflection
from a dipping boundary near Moho. The detailed image
has allowed modelling of shallow intrusive bodies that may
have been ignored earlier, and it was possible to pick weak
irregular shear wave arrivals. Greater accuracy in WA
modelling allows a detailed comparison between lower
crustal CDP reflectivity and the onset of Moho.
Comparison with coincident CDP data shows that in
some areas where the upper crust is seismically transparent
in near vertical incidence images there may actually be
distinct and strong velocity heterogeneities. A pervasive
upper-mid crustal velocity boundary interpreted from WA
reflections generally does not correlate with CDP reflection
horizons. Moho interpreted from WA reflections can
correspond with an abrupt lower crustal reflectivity
boundary within several hundred ms or less over broad (100
km) areas but clearly does not always coincide so closely,
particularly in areas with structural complexities in the
Moho.
Acknowledgements. The data collection was possible
through a generous gift of recording time with the Mobil oil
company's survey vessel Search. Wide-angle recording was
organized at the University of Bergen by Lisbeth Engell-
SOrensen and funded by the Norwegian Lithosphere Project.
The work done represents Norwegian Lithosphere Project
contribution no. 136.
References
BABEL Working Group, Evidence for early Proterozoic
plate tectonics from seismic reflection profiles in the
Baltic shield, Nature, 348, 34-38, 1990.
B raile L. W., C. S. Chiang, The Continental Mohorovicic
Discontinuity: Results from Near-Vertical and Wide-
angle Seismic Reflection Studies, Seismic Reflection
Studies, Reflection Seismology: A Global Perspective,
ed. M. Barazangi and L. Brown, AGU Geodynamics
Series 13, 257-272, 1986.
Brocher T. M., M. J. Moses, Wide-angle seismic recordings
obtained during the TACT multichannel reflection
prof'fiing in the northern gulf of Alaska, USGS Open-
File-Report 90-663, 1990.
DEKORP Research Group (Represented by R. Meissner,
Th. Wever, P. Sadowiak), Reflectivity patterns in the
Variscan mountain belts and adjacent areas: an attempt
for a pattern recognition and correlation to tectonic
units, Tectonophysics 173, 361-378, 1990.
Gajewski D., R. Stangl, K. Fuchs, K. J. Sandmeier, A new
constraint on the composition of the topmost continental
mantle-anomalously differnet depth increases of P and S
velocity, Geophys. Jour. Int., 103, 497-507, 1990.
Gajewski D. and C. Proriehl, Seismic refraction investigation
of the Black Forest, Tectonophysics, 142, 27-48, 1987.
Gibson B. S., A. R. Levander, Lower crustal reflectivity
patterns in wide-angle seismic recordings, GRL, 15, 617-
620, 1988.
Gibson D., A. R. Levander, Modeling and processing of
scattered waves in seismic reflection surveys,
Geophysics, 53, 466- 478,1988.
GLIMPCE Seismic Refraction Working Group, GLIMPCE
Seismic Experiments, EOS, 70, 841 & 852, 1989.
Gohl K., N. G. K. Boyd, S. B. Smithson, R. B. Hawman,
Seismic Wide-angle and Normal-incidence Studies of
Precambrian Crust in North America and Greenland,
Terra Abstracts 2, 152, 1990.
Heikkenen P., P. Lindblom, U. Luosto, K. Komminado, J.
Yliniemi and the BABEL Working Group, Land
Recording of BABEL Profiles in Finland, Terra
Abstracts 2, 197, 1990.
Holbrook W. S., D. Gajewski, A. Krammer, C. Prodehl,
Interpretation of Wide-Angle Compressional and Shear
Wave Data in Southwest Germany: Poisson's Ratio and
Petrological Implications, JGR, 93, 12081-12106, 1988.
Hole J. A., R. M. Clowes, R. M. Ellis, Three-Dimensional
Travel Time Modelling in the Queen Charlotte Basin,
Offshore West Coast Canada, EOS Trans.,71,1589,1990.
Hurich, C. A., this volume, 1991.
Hurich, C. A. and Y. Kristoffersen, Deep structure of the
Caledonide orogen in southern Norway: new evidence
from marine seismic reflection profiling, Norges
Geologiske UndersOkelse Spec. Pub., 3, 96-101, 1988.
Klemperer S. L., T. A. Hauge, E. C. Hauser, J. E. Oliver,
C.J. Potter, The Moho in the northern Basin and Range
Province, Nevada, along the COCORP 40 N seismic
reflection transect, Geol. Soc. Amer. Bull., 97, 603-618,
1986.
Klemperer S. L., C. A. Hurich, Lithospheric structure of the
North Sea from deep seismic reflection profiling,
Tectonic evolution of the North Sea Rifts, Ed. D. J.
Blundell, A.D. Gibbs, Oxford Science Publications, 37-
63, 1990.
Meissner R. and N.J. Kusznir, Crustal viscosity and the
reflectivity of the lower crust, Annales Geophysicae 5B,
365-374, 1987.
Mereu R. F., D. Epili, A. G. Green, Pg shingles:
preliminary results from the onshore GLIMPCE refraction
experiment, Tectonophysics, 173, 617-626, 1990.
Mooney W. and T. M. Brocher, Coincident Seismic
Reflection/Refraction Studies of the Continental
Lithosphere: A Global Review, Reviews of Geophysics,
25, 723-742, 1987.
Moses M. J., T. M. Brocher, M. A. Fisher, M. Talwani, J.
I. Ewing, Wide-Angle Recordings of the 1989 EDGE
MCS Survey Along the Eastern Aleutian Arc-Trench
Transect, Alaska, EOS Trans., 70, 1339, 1989.
Moses M. J., T. M. Brocher, S. D. Lewis, Wide-Angle
Seismic Reflection/Refraction Studies of Crustal Structure
in the vicinity of the 1989 Loma Prieta Earthquake, EOS
Trans., 71, 1631, 1990.
Pajchel J., Report on Release 3 of the Ray-tracing Package,
Norsk Hydro Internal Report, 1989.
$igmond, E. M. O., M. Gustavson, D. Roberts,
Berggrunnskart over Norge, 1:1 million, Norges
Geologiske UndersOkelse, 1984.
Trehu A., T. Holt, Y. Nakamura, T. Brocher, Preliminary
results from the 1989 Oregon onshore-offshore seismic
imaging experiment, EOS Trans., 71, 1588, 1990.
West, T. E., R. E. Long, Crustal wide-angle reflections from
airgun shots, Geophys. Jour., 96, 587, 1989.
Wenzel F., K. Sandmeier, W. Walde, Properties of the
Lower Crust from Modelling Refraction and Reflection
Data, Journal of Geophysical Research, 92, 11575-11583,
1987.
Wever Th., The Conrad discontinuity and the top of the
reflective lower crust- do they coincide?, Tectonophysics,
157, 39-58, 1989.
SOURCE-GENERATED NOISE IN MARINE SEISMIC PROFILES: THE LIMITS OF
REFLECTION DETECTABILITY IN THE UPPER CRUST
C. A. Hurich
Institute for Solid Earth Physics, University of Bergen
Allegaten 41, N-5007 Bergen, Norway
Abstract. Experiments demonstrate that source-generated
noise in marine seismic profiles may mask reflections in the
upper crystalline crust and limit the completeness and
reliability of some deep seismic data. The level of source-
generated noise depends on the size of the acoustic contrasts
in the near-surface geology. Areas in which source-
generated noise is important are not limited to the well-
known hard-bottom situations and may be common in
marginal platform areas such as those of northwest Europe.
The standard processing sequence commonly applied to
marine deep seismic data has limited effectiveness for
attenuating source-generated noise. In the experiments,
reflections from acoustic impedance contrasts commonly
encountered in crystalline rocks (0.025 - 0.10) are unimaged
or poorly imaged in the first 6 to 8 km of the crust but are
better imaged with increasing depth. These experiments
suggest that the commonly observed transparent upper crust
and onset of reflections in the middle crust may be the
result of limitations of the seismic reflection technique
rather than crustal geology.
Introduction
Exclusive of reflections from sedimentary basins, deep
seismic reflection profiles in northwest Europe most often
show the general pauem of a strongly reflective lower crust
and a seismically transparent upper crust. Although
certainly not ubiquitous, this commonly observed pattern has
led to considerable investigation and speculation concerning
the differences in composition, rheology, and insitu
properties between the upper and lower crust. A number of
workers have offered credible explanations for the strong
reflections often observed in the lower crust [for a review
see Blundell, 1990] but the apparent lack of reflections in
the upper crust has received only minor attention.
Crustal models derived by comparing the reflection
response of different crustal levels assume that the observed
reflection pattern is the result of the physical and geometric
Geodynamics 22
properties of the rocks and not the result of the seismic
reflection method. This assumption is important for
reflection profiles that traverse a variety of different
geologic domains with variable reflection character in the
lower crust and a nearly uniformly transparent upper crust.
The implication of such a regionally transparent upper crust
is that either the different geologic domains all have the
same seismic reflection signature or in some cases the
seismic method is not providing reliable data for the upper
crust. This situation is particularly poignant for areas in
which exhumed lower crustal rocks are presently at the
surface and are apparently seismically transparent
[Klemperer et al., 1987]. This implies either that the strong
lower crustal reflections result from processes that have not
affected the exhumed lower crustal terrane, the process of
exhumation destroys the reflectivity [Wever et al. 1987] or
the mechanism of lower crustal reflections is effective only
in situ. It is apparent that an understanding of the cause
for the lack of reflections in the upper crust is critical to
the formulation of crustal models based on reflection
patterns.
In the course of our program of marine seismic profiling
we have observed a number of cases in which, based either
on geology or on reflections observed in nearby land-based
profiles, we expected to detect reflections in the upper crust
but failed to do so. Review of the field data suggests that
high levels of source-generated noise in the first 3-5 s two-
way-time (TWT) may be masking reflections if they are
present in the upper 8-12 km of the crust. In this paper I
present several experiments designed to evaluate the effects
of source-generated noise on marine seismic acquisition and
discuss the results and implications of these experiments.
Source-Generated Noise
Noise in marine seismic surveys generally falls within
two categories: 1) noise associated with the acquisition
system; 2) source-generated noise. For a modem seismic
acquisition vessel under ordinary operating conditions, noise
associated with the acquisition system (instrument noise,
boat noise, streamer noise) is very low with respect to the
energy level of the airgun array and will not be considered
443
444 LIMITS OF REFLECTION DETECTABILITY
further in this discussion. Source-generated noise, which
includes essentially the entire wavefield generated by the
source except for the primary P-wave reflections (multiples
of both reflections and refractions, scattered waves and
mode-convened waves), potentially has energy levels
approaching that of the airgun array and will be the focus
of this discussion.
For a tuned marine source array the level of source-
generated noise depends mainly on the amount of energy
trapped in the near-source environment for a significant time
after the shot and generally reflects the characteristics of the
near-surface geology. Two environments that are commonly
associated with high levels of source-generated noise are
areas with large reflection coefficients at the water bottom
(i.e. hard-bottom areas) and areas with variable water-
bottom topography and significant shallow velocity
heterogeneities. The former situation traps energy in the
water column in the form of multiples with a variety of
raypaths while the latter situation results in backscattering
of energy within the water column or just below the water-
bottom [eg. Larner et al., 1983]. These two situations can
of course occur in concert. As will be shown in this paper
areas that produce significant levels of source-generated
noise are not limited to these two well known situations and
may be quite common in marginal platform areas. In all
cases the source-generated noise is at least partly coherent
although variations in coherence can result from interference
between various noise trains. The amplitude of the noise is
high with respect to most reflections because of high
reflection coefficients and relatively short travel paths
encountered in the near-surface environment.
Experiments
To evaluate the possible effects of source-generated noise
on the apparent reflectivity of the upper crust I have
examined three different marine localities. All three of
these are localities in which, based either on geology
exposed nearby onshore or on reflections observed in other
seismic data, we expect to observe reflections in the upper
crystalline crust. In one of these areas crystalline basement
is exposed directly at the water bottom, in the other two
areas basement is covered by 800 to 1300 m of sedimentary
rocks. All of the seismic data used in the experiments are
part of a larger data set collected along the southwest coast
of Norway.
My approach to evaluating the effects of source-
generated noise is to add synthetic reflections to actual shot
records, evaluate the signal amplitudes with respect to the
ambient amplitude in the shot records and perform a
common-midpoint (CMP) stack with a standard processing
sequence to determine the detectability of the synthetic
reflections in the presence of the actual recorded wavefield.
The synthetic seismograms were constructed using well-
determined refraction information to constrain near-surface
velocities and by using a reasonable range of reflection
coefficients for reflectors in the crtstalline basement. The
synthetics were calculated by dynamic ray-tracing and
include the effects of geometric divergence. The synthetic
seismograms were calibrated to the actual seismic data by
comparison of the amplitude of the water bottom reflection
on the near (200-600 m) traces. The proper calibration of
the synthetic reflections is critical to the results of the
experiment but in all cases the velocity contrast at the water
bottom is well constrained by the refraction information and
care was taken to insure the consistency of the calibrations.
For reasons of proximity, it is convenient to discuss two
of the test areas together and the third area separately.
Figure 1 shows the relative location of the first two
localities used in the experiment. ILP-7 is an area with
crystalline rocks at or very near the water bottom while the
area of ILP-7A has about 800 m of moderately consolidated
sedimentary rocks overlying the basement. The two
different areas have different near-surface geology and
consequently differing near-surface velocity distributions. At
ILP-7 the upper crust is seismically transparent to about 3.5
s TWT while at ILP-7A the upper crust is transparent to
about 2.5 s. The apparent thickness of the transparent
upper crust correlates directly to the presence and absence
of sedimentary rocks overlying the crystalline basement
suggesting a causal relationship.
Figure 2 illustrates representative shot gathers from ILP-
7 (SP 280) and ILP-7A (SP 560), the velocities and
reflection coefficients determined from the refraction
analysis are included for reference. In both shot records the
majority of the high amplitude arrivals in the first 4.0 s
TWT can be identified as various forms of source-generated
noise. Both shot records show a wave-train with a group
velocity approximating the velocity of the direct wave in the
water. The apparent phase velocities are similar to those of
the water bottom reflection. This wave-train is the result of
ILP-7A ILP-7
SP-560 SP-280
Fig. 1. Test locations ILP-7 and' ILP-7A. The lateral
extent of the data used in the stacking tests are marked by
the lines with arrows. The CMP data had a nominal fold
of 45.
HURICH 445
SHOT RECORDS
NO GAIN
N SP-560
SP-28 0 ; , ,
SP793
Fig. 2. Representative shot records from the three test areas. No processing has been applied
to the records. The data were acquired with a 7300 in 3 airgun array, a 4.5 km streamer and a
50 m group and shot interval. Velocity functions and reflection coefficients determined from the
refractions are shown at the bottom of the shot records. Small numbers are velocity in km/s and
the large numbers the reflection coefficients.
pegleg multiples in the water column. In SP 280 the first
arrival, which represents a refraction or headwave in the
crystalline basement, reverberates for at least 3.0 s TWT.
Although the first breaks do not reverberate in SP 560, both
long- and short-period multiples from within the sedimentary
column are dominant down to 3.0 s TWT. The only
primary reflection identifiable in SP 280 is the water bottom
reflection while in SP 560 the water bottom reflection and
the reflection from the sediment/basement contact are
identifiable. It is clear that source-generated noise
dominates the wavefield in the first 4.0 s TWT in both of
the areas.
A comparison between ambient amplitudes and synthetic
reflection amplitudes (reflection coefficient (RC)=0.05) for
the two shot records are shown in Figure 3. I use the term
ambient amplitude to refer to the RMS amplitude over a 32
ms running window of the actual data. Comparisons are
shown for the near-, mid- and far-traces of the two shot
records. For both shot records, the reflection amplitudes are
significantly lower (20-36 dB) than the ambient amplitudes
in specific limited windows. These windows can be
identified as having anomalously high ambient amplitudes
and correlate directly to the zones of strong source-
generated noise on the shot records (Fig. 2). To obtain a
first-order estimate of the signal/noise (S/N) improvement
resulting from further processing, particularly CMP stacking,
I also show on Figure 3 the theoretical reflection amplitude
resulting from assuming a S/N improvement proportional to
the square-root of the number of samples (root-n) during
CMP stacking. The calculation of the theoretical S/N
improvement takes into account decreased CMP fold in the
first few seconds of the data due to muting. Clearly, the
theoretical S/N improvement based on root-n considers only
uncorrelated noise and does not estimate the effect of the
other processing steps on the correlated noise. However,
the root-n prediction does provide a useful first-
approximation on the travel time at which the source-
generated noise ceases to dominate the data and also
provides an estimation of the amount of S/N improvement
required from the other processing steps to overcome the
correlated noise. Comparison of the times for which the
theoretical signal amplitude becomes equal to or greater than
the ambient amplitude shows that for the near- and mid-
traces the crossover occurs about 1.0 s earlier for SP 560
(ILP-7A) than for $P 280 (ILP-7). This is consistent with
the original stacked data that shows reflections 1.0 s earlier
on ILP-7A than ILP-7 (Fig. 1).
To determine the limits of reflection detectability in the
fully processed data I added synthetic reflections to 100
shot records (5 km of surface coverage) for both areas and
applied the full processing sequence (Fig. 4). Synthetic
reflections from simple, horizontal boundaries with reflection
coefficients representative of those expected in crystalline
rocks [Hurich and Smithson, 1987; Warner, 1990] and
SHOT 280
NEAR MID FAR
o
Rc os .. Rc-o5 c-
2
3
FAR CMP STACK CMP STACK STACK
SHOT 560
NEAR MID FAR
0
2 NEAR ST,CK 0 DI) ACK RC- TACK
Fig. 3. Amplitude decay plots for the near, mid and far
traces of shot records from areas ILP-7 (SHOT 280) and
ILP-7A (SHOT 560). The fine line shows the amplitude of
synthetic reflections with a reflection coefficient of 0.05.
The heavy line shows the relative reflection amplitude
assuming a S/N improvement proportional to the square-
root of the number of traces ,involved in the CMP stack.
The arrows at the left indicate the crossover time of the
ambient amplitude and the estimated CMP stack response.
Note that between 1.8 s and 2.6 s the far offset traces may
contribute signal if the mute is chosen carefully.
reflection times appropriate for the upper crust allow
evaluation of the limits of reflection detectability in the
presence of source-generated noise. The results of this
exercise are shown in Figure 5. As predicted by the
analysis of the shot records, reflections are visible for
smaller travel times and smaller reflection coefficients in
ILP-7A than in ILP-7. The shallowest of the three
reflections is only visible in the case of ILP-7A and
RC=0.1. The deepest reflection is most visible in all cases.
It is evident that for the cases of ILP-7 and ILP-7A source-
generated noise severely limits the detection of reflections
originating from contrasts that may reasonably be expected
in the crystalline upper crust.
The third area that was used for testing is shown in
Figure 6 and is designated ILP-12. The area of ILP-12 has
about 1300 m of sedimentary rocks overlying crystalline
basement; the sediments are of a different age, and the
water bottom topography is smoother than the area of 7A.
ILP-12 provides an interesting test location because a
dipping band of strong reflections (lower right quadrant) is
truncated abruptly at 4.0 s TWT without any apparent
geological reason. It is, of course, of interest to the
interpretation of the seismic data whether or not the
structure that produces the band of reflections actually
terminates and is not present in the upper crust.
A shot record (SP 793) from ILP-12 and the associated
velocity function is illustrated in Figure 2. As in the other
examples, the wavefield in the first 3.0 to 4.0 s TWT is
dominated by multiples originating in both the water column
and the sediment column. Multiples that probably originate
within the sediment column are particularly strong on the
mid-traces and are the dominant arrival to at least 5.5 s
TWT. Primary reflections from the water bottom and the
sediment/crystalline contact are also visible. It is
remarkable that the near-surface velocities in the ILP-12
area are so similar to the ILP-7A area given that the areas
are several hundred kilometers apart and the age of the
near-surface sediments is different. Apparently the two
areas have quite similar possibilities for the production of
source-generated noise even though they differ geologically.
Comparison of the synthetic reflection amplitude and the
ambient amplitude for SP 793 (Fig. 7) shows a similar
although not identical pattern to $P 560 of ILP-7A. The
cross-over times predicted by the root-n estimation are
PROCESSING SEQUENCE
GAIN RECOVERY
GAP DECONVOLUTION
GATE 0 - 6.0 s
OPERATOR 248 ms
GAP 32 ms
SHOT DOMAIN ARRAY SIMULATION
5 TRACE WEIGHTED MIX
TIME VARIANT
CMP GATHER
NORMAL MOVEOUT CORRECTON
MEAN STACK
GAP DECONVOLUTION
GATE 0 - 6.0 s
OPERATOR 248 ms
GAP 32 ms
BAND PASS FILTER
0.0 - 4.0 s 7/10 - 40/50 HZ
4.0 - 6.0 S 7/10- 35/40 HZ
AGC
1200 ms
Fig. 4. Processing sequence for CMP data. The processing
sequence for the data examples and the experimental
examples were identical.
HURICH 447
(S-280)
RC-.025 RC-.05 RO-. 10
Fig. 5. Stacking tests for 1LP-7 and 1LP-7A. The tests were carried out for three different
reflection coefficients (0.025, 0.05, 0.10). The dots mark the reflection times of the three
synthetic reflections.
somewhat less for SP 793 than for SP 560, particularly for
the far offset but the S/N improvement required to
overcome the source-generated noise is still in the range of
20-30 dB. The differences in the crossover times reflect
differences in the patterns of the noise rather than the
amplitude of the noise. Note in particular that the ambient
amplitudes of the mid-trace decline more slowly in SP 793
than in SP 560 reflecting the dominance of multiples in the
mid-traces.
The stacking test for ILP-12 (Fig.8) shows that, for the
reflection coefficients tested, reflections at 4.0 s TWT
should be detectable; for a RC=0.05 a reflection at 3.0 s
TWT may be detectable and for RC=0.1 reflections at 3.0
s TWT should be clearly detectable. The reflection at 2.0
s TWT was only detectable for RC=0.1 and then only
marginally so. Subhorizontal multiple trains are
predominant even after pre-stack and post-stack predictive
aleconvolution.
ILP- 12
SP-793
3
Fig. 6. Test location ILP-12. The location of the shot record (SP 793) and the test profile are
indicated. See ext for discussion of reflections A and B.
Discussion the modelled sediment/basement contact with a reflection
coefficient of 0.43.
The results of the stacking tests in all three areas The tests presented herein suffer from an over-
examined indicate that source-generated noise cm limit simplification of the geological situation in most crystalline
reflection detectability in the upper crust if the near-surface rocks. The synthetic reflections were generated using
geology has moderate to large acoustic impedance contrasts. horizontal, simple boundaries, a situation rarely encountered
The obvious question arises: if source-generated noise is in field observations. The simplification affects the
important for limiting reflection detectability, why do we experiment in two different ways. First, as suggested
see reflections in sedimentary basins? The stacking tests do above, because most reflectors occur as layers rather than
not address this question as the stacking velocities were simple boundaries, constructive interference may enhance
chosen specifically to best stack the synthetic reflections reflection amplitudes, and the apparent reflection coefficients
relieving the test stacks from any bias due to improper may be larger than for simple boundaries. Thus,
stacking velocities. Figures 1 and 6 show that the constructive interference can make reflections detectable at
sedimentary basins show clear images in the same areas smaller travel times than predicted by the experiments. On
where the stacking tests demonstrate that source-generated the other hand, reflectors in crystalline rocks almost never
noise limits reflection detectability in the upper crust. I occur as flat interfaces of large lateral extent. Rather they
suggest that the order in the sedimentary rocks, laterally occur as small, individual elements of variable dip with
extensive, subhorizontal layers, results in constructive considerable lateral variation. This character generally
interference and high apparent reflection coefficients. This decreases reflection amplitude and continuity and
view is supported by noting that the reflections in the consequently detectability. Because of the simplicity of the
sedimentary section on a true amplitude version of ILP-12 experiment with respect to the actual situation the results
stacked with a velocity function appropriate for both the must be regarded as semiquantitative but the general
actual data and the synthetic reflections, are comparable to conclusions remain valid.
HURICH 449
w 4
NEAR
SHOT 793
FAR
RC-.O RC-.05 ' I ......

STACK CMP STACK ~ . CMP STACK
Fig. 7. Amplitude decay curves for area ILP-12 (SHOT 793). The notation is the same as in Figure 3.
Source-generated noise is not a unique explanation for interpretation of the data. Consider the situation examined
the occurrence of a transparent upper crust. Undoubtedly, in the area of ILP-12. Strong reflections that extend
lateral variations in upper crustal reflectivity due to through the lower and middle crust are apparently truncated
geological changes occur as do variations in lower crustal at 4.0 s TWT and not visible in the upper crust (A in Fig.
reflectivity leg. Blundell, 1990]. However, in the case of 6). Another set of reflections (B in Fig. 6) is visible
the many marine seismic profiles that cross a variety of between 3.5 s and 2.5 s. If the termination of the
crustal provinces and suggest a regionally transparent upper reflections is real, it is important information for the
crust, source-generated noise must be considered in the interpretation of the data. The stacking tests indicate that
RC-.025
ILP- 12
(SP-793)
RC-.05
RC-.10
Fig. 8. Stacking tests for ILP-I. The notation is the same as in Figure 5.
unless the reflections originate in fairly large contrasts they
will be obscured by the source-generated noise to at least
3.0 s TWT suggesting that reflection package B must result
form larger contrasts than reflection package A. In this
particular case there are two other seismic profiles in the
area, one 10 km closer to shore and another 25 km further
offshore. The offshore profile shows reflections continuing
up to the sediment/basement interface while the nearshore
profile shows truncation of the reflections similar to ILP-
12 [Hurich and Kristoffersen, 1988; Klemperer and Hurich,
1990]. Unless the geology in the crystalline basement
changes considerably in the 35 km between the profries,
source-generated noise must play an important role in
obscuring reflections from the upper part of the crystalline
basement in the area of ILP-12. The decrease in the level
of source-generated noise in the area offshore from ILP-12
reflects lateral variations in.the basin margin sedimentary
sequence.
The processing sequence that was applied in the
experimental stacking was the same as that applied to the
data of Figures 1 and 6 and very similar to the general
processing sequence for deep seismic data discussed by
Klemperer [1989]. Such a sequence actually applies a
minimum of effective processing to the first 3-4 s of the
data. The variety of raypaths in the water and sediment
column results in interference of the various multiple trains
and variations in periodicity that render deconvolution less
effective. Mutes are generally chosen quite deep decreasing
the CMP fold for the shallow data. Time-variant array
simulations generally are designed to have a minimum effect
in this time range. In addition, for a 50 m group interval,
wavetrains with apparent velocities near the velocity of
water are spatially aliased for frequencies higher than 15
Hz, and array simulation and F-K filtering are limited by
the spatial sampling. Because the effectiveness of the
processing is limited, accurate velocity analysis is most
difficult in the upper crustal region where it is most
important. Thus it seems that the standard acquisition and
processing scheme is not particularly effective for
attenuating source-generated noise, and we need to explore
alternatives to make our deep seismic profiles more
complete and reliable data sets.
Conclusion
The experiments demonstrate that source-generated noise
can limit detectability of reflections from the upper
crystalline crust and may account for at least some of the
observations of a regionally transparent upper crust. Near-
surface geology that may produce significant source-
generated noise is not confined to the well-known hard-
bottom areas but includes geological sections that are
common in marginal platforms such as those of
northwestern Europe. In areas of apparently regionally
transparent upper crust, reprocessing of the data may reveal
a significant amount of upper crustal reflectivity. Clearly,
lateral variations in reflectivity must be expected in the
upper crust; large intrusive terrains cannot be expected to
be very reflective, however many metamorphic terrains may
very likely have a reflection signature that is not presently
being imaged in many marine deep seismic profiles.
Models of crustal evolution that rely on comparison of the
patterns of reflectivity in the upper and lower crust must
take into account the limitations of the seismic reflection
method particularly with respect to upper crustal reflectivity.
Acknowledgments. I thank the Captain and crew of
M/V Mobil SEARCH and Mobil Norway whose generous
donation of ship and processing time provided the data used
in this study. Ernst Flueh helped to improve the focus and
usefulness of this paper. This work is supported by the
Norwegian Council for Science and Humanities. Norwegian
ILP Contribution No. 137.
References
Blundell, D. J., Seismic images of the lithosphere, J. Geol.
Soc. London, 147,895-913, 1990.
Hurich, C. A., and Smithson, S. B., Compositional variation
and the origin of deep crustal reflections, Earth Planet.
Sci. Lett., 85, 416-426, 1987.
Hurich, C. A., and Kristoffersen, Y., Deep structure of the
Caledonide orogen in southern Norway: new evidence
from marine seismic reflection profiling, Nor. geol.
unders. Special Publ. 3, 96-101, 1988.
Klemperer, S. L. and the BIRPS group, Reflectivity of the
crystalline crust: hypotheses and tests, Geophys. J. R.
astr. Soc., 89, 217-222, 1987.
Klemperer, S. L., Processing BIRPS deep seismic reflection
data: a tutorial review, in Digital Seismology and Fine
Modeling of the Lithosphere, edited by R. Cassinis, G.
Nolet, and G. F. Panza, 229-257, Plenum, 1989.
Klemperer, S. L., and Hurich, C. A., Lithospheric structure
of the North Sea from deep seismic reflection profiling,
in Tectonic Evolution of the North Sea Rifts, edited by
D. J. Blundell and A. Gibbs, Oxford University Press,
1990.
Lamer, K., Chambers, R., Yang, M., Lynn, W., and Wai,
W., Coherent noise in marine seismic data, Geophysics,
48, 854-886, 1983.
Warner, M. R., Absolute reflection coefficients from deep
seismic reflections, Tectonophysics, 173, 15-23, 1990.
Wever, Th., Trappe, H., and Meissner, R., Possible relations
between crustal reflectivity, crustal age, heat flow, and
viscosity of the continents, Ann. Geophysicae, 5B, 255-
266, 1987.

Continental Lithosphere - Deep Seismic Reflections PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Continental Lithosphere - Deep Seismic Reflections PDF

Uploaded by

Copyright:

Available Formats

Continental Lithosphere:

You might also like