Relative Motions Between Oceanic

and
Continental Plates in the Pacific Basin

David C. Engebretson
Department of Geology
Western Washington University
Bellingham, Washington 98225

Allan Cox
Department of Geophysics
Stanford University
Stanford, California 94305

Richard G. Gordon
Department of Geological Sciences
Northwestern University
Evanston, Illinois 60201

Ser.

PER
@ 1985 by The Geological Society of America, Inc.

All rights reserved.

All materials subject to this copyright and included

in this volume may be photocopied for the noncommercial
purpose of scientific or educational advancement.

Published by The Geological Society of America, Inc.

3300 Penrose Place, P.O. Box 9140, Boulder, Colorado 80301

GSA Books Science Editor Campbell Craddock

Printed in U.S.A.

Library of Congress Cataloging-in-PublicationData

Engebretson, David C.
Relative motions between oceanic and continental
plates in the Pacific Basin.

(Special paper / Geological Society of America ;206)

Bibliography: p.
1. Plate tectonics-Pdc Ocean. I. Cox, Allan,
1926- . 11. Gordon, Richard G., 1953-
111. Title. IV. Series: Special paper (Geological
Society of America) ;206. 2

QE511.4.E54 1985 551.1'36'09164 8&&

ISBN 0-8137-2206-3
 Contents

Acknowledgments .................................................... v
Abstract ............................................................ 1
Introduction ......................................................... 2
The Model .......................................................... 2
Method and Assumptions ............................................. 2
Reconstructions Used to Build the Model ................................. 3
Errors and Uncertainties ............................................... 17
Motion between hotspots ............................................ 17
Errors in determining plate-hotspot motions ............................. 17
Errors in determining plate-plate motions ............................... 17
Uncertainty in the location of boundaries between
oceanic plates ................................................... 18
Summary of errors .................................................18
Reconstructions ......................................................18
Method, Data. and Assumptions ........................................18
Detailed Reconstructions ..............................................27
Present configuration ...............................................27
20 Ma (chron 6) ...................................................27
37 Ma (chron 13) ..................................................27
56 Ma (chron 25). 65 Ma (chron 29). and 80 Ma
(chron 33r) ..................................................... 27
110 Ma (Albian. middle of Cretaceous normal superchron) ................. 28
140 Ma (Berriasian. chron M14) ......................................28
Comparison with Global Circuit in the Neogene ............................29
Plate Tectonic Interpretation ...........................................29
Age and Thickness of Subducting Farallon Plate ............................29
Variations in Bathymetry Parallel to the Trench ............................32
MendocinoEscarpment .............................................34
Velocity of Convergence ..............................................34
Trajectories of Plates .................................................42
Geologic Interpretation ................................................42
Pacific-North American Convergence beginning at 5 Ma ..................... 42
Convergence and Magmatism, ..........................................44
Laramide Orogeny ................................................... 44
Appendix A-Information for Tracing the Causes of
Changes in Relative Plate Motions ................................46
Appendix B-Azimuth and Speed of the Linear Velocities
of Oceanic Platesfor North America and Eurasia ................... 50
References ...........................................................58
...
Ill
 Acknowledgments

Grateful acknowledgment is made to Cecil and Ida Green for fellowship support to
D.C.E.; to Shell Faculty Career Initiation Funds for support of R.G.G.; to the donors of the
Petroleum Research Fund, administered by the American Chemical Society, for partial
support of this research; to the National Science Foundation for support from grants EAR
79-19712 and EAR 8 1-07702; to Gulf, Mobil, Texaco, Arco, and Aminoil corporations for
support of this project through the Stanford Circum-Pacific Tectonics Affiliates Program; to
R. B. Hart for convincing us that the spreading history south of the Mendocino Fracture Zone
should be given serious consideration;to Michel Debiche for help in checking and revising the
tables; and to Benjamin Page and George Thompson for their encouragement and criticism
along the way.
2 Engebretson, Cox, and Gordon

INTRODUCTION thonous terranes great distances northward, as suggested by

paleomagnetic data from the terranes? (4) Are plate tectonic
A causal link with plate tectonics has been proposed for processes responsible for the widespread Pliocene orogeny re-
most of the major elements of Cordilleran tectonics, including corded at many localities along the western boundary of North
plutonism, episodes of strike-slip and normal faulting, basin for- America?
mation, terrane transport, and orogeny. The information needed
to test these hypotheses is a set of Euler poles describing the THE MODEL
former locations of oceanic plates relative to adjacent continents.
Because such a set of Euler poles can be used to generate maps Method and Assumptions
showing the former locations of oceanic plates adjacent to the
continental margin and relative plate velocities along the margin, This model is based on the assumption that the hotspots in
they provide the basic framework needed to relate plate tectonics the Pacific and Atlantic regions have remained fixed with respect
to Cordilleran geology. A full set of Euler poles describing the to one other, an approach first applied by Morgan (1972) to
history of displacement between Pacific basin plates and adjacent analysis of plate motions and by Coney (1972,1978) to analysis
continents has not yet been published. of the convergence history and tectonic evolution of western
Most attempts to link plate tectonics and geology along the North America during the Mesozoic and Cenozoic. Our relative
margin of North America are based on the important study of motion model (Tables 2,3, and 4) is based on previously deter-
Atwater and Molnar (1973) for the interval from 38 Ma to the mined reconstructions (Table 1). Once the displacementsof all of
present, which gives paleogeographic reconstructions (but not the plates relative to a common hotspot reference frame were
Euler poles) for the Pacific and North American plates. Similarly determined (Table 2), their motions relative to each other (Tables
Coney (1972) presented generalized reconstructions and conver- 3 and 4) were found by addition of finite rotations (Figure 1).
gence velocities for the late Mesozoic and Cenozoic but not the One advantage of this approach over that of constructing a global
Euler poles upon which these were based. Francheteau and oth- North America-Africa-India-Antarctic-Pacific circuit (Atwater
ers (1970) found the coordinates of Euler poles for Pacfic-Faral- and Molnar 1973) is that in the past one or more of these plates
Ion motion. Starting with these poles, Carlson (1982) estimated may have been divided by an undetected boundary into two
the angular rates of Pacific-Farallon motion and the rates of sub-plates (Gordon and Cox 1980; Suarez and Molnar 1980;
convergence of the Farallon and Pacific plates relative to North Morgan 1981; Duncan 1981) with an unknown amount of dis-
America during the Cenozoic, but no map reconstructions of the placement and unknown contributions to the global circuit. A
Farallon plate and no information about the Farallon plate were disadvantage of using a hotspot reference frame is that motions
given. In addition to these pioneering studies, many models have between hotspots in the Atlantic and Pacific basins and errors in
been presented for former locations of the ridges between oceanic plate-hotspot motions introduce errors into reconstructions, as
plates in the Pacific basin (e.g. Larson and Chase 1972; Hilde and discussed below. A comparison between reconstructions made
others 1976; Alvarez and others 1980). using the two approaches is given in a subsequent section.
In this paper we develop a quantitative plate tectonic model The important issue of whether hotspots move slowly
for the history of displacements between continental and oceanic enough with respect to one another to provide a useful mantle
plates in the Pacific basin during the past 180 Ma. Our results are reference frame has been the subject of active investigation during
expressed in the form of tables of Euler poles describing former the past decade. Morgan (1972) initially estimated that hotspot
locations of all of the known relevant plates, tables and diagrams motions of 5 km/m.y. were sufficient to explain the non-
of relative plate velocities, and map reconstructions showing parallelism of seamount chains on the Pacific plate. On the other
former locations of the North American, Farallon, Kula, and hand, Molnar and Atwater (1973), Burke and others (1973), and
Izanagi plates. In our analysis we emphasize how the bathymetry Molnar and Francheteau (1975) all concluded that the hotspots
and age of the ocean floor adjacent to the margin of North have had relative velocities of at least 15 km/m.y. However,
America have varied through late Mesozoic and Cenozoic time. Morgan (1981,1983) showed that, with only minor reinterpreta-
This plate tectonic model provides a basis for a discussion of the tion of the data from previous studies, plate reconstructions for
following questions about possible relationships between plate the last 200 m.y. are consistent with relative movement of only a
tectonics and major geologic events. (1) Do major episodes of few km/m.y. between the hotspots in the Atlantic and Indian
Mesozoic plutonism occur during times of unusually rapid plate Oceans. This is supported by Duncan's (1981) analysis of hot-
convergence? Are all episodes of rapid plate convergence asso- spots in the Atlantic and Indian Ocean basins and an analysis by
ciated with major plutonism? (2) Is the Laramide orogeny (80 to Henderson and Gordon (personal communication, 1981) of
40 Ma) due to a change in the absolute velocity of the North hotspots in the Pacific Ocean basin. Morgan's conclusion is
American plate, as suggested by Coney (1972), or is it more further supported by studies of instantaneous plate motions
closely related to changes in the rate of convergence and to a (Minster and others 1974; Chase 1978; Minster and Jordan
decrease in the age of the subducting plate? (3) Are displacements 1978) that detected no significant relative motion between the
derived from these plate models capable of transporting alloch- hotspots during the past 10 m.y., and by the observation that
 Geological Society of America
Special Paper 206
1985

Relative Motions Between Oceanic

and
Continental Plates in the Pacific Basin
ABSTRACT

A model is presented for the displacement history between western North

America, eastern Eurasia, and adjacent oceanic plates (Pacific, Farallon, Izanagi, Kula,
and Phoenix) for the past 180 million years. The model is based on the assumption that
the hotspots in the Atlantic region have remained fixed relative to the hotspots in the
Pacific basin (but not necessarily relative to the spin axis). The model uses a new
determination for relative motion between the oceanic plates of the Pacific basin.
The results show that in a broad sense the Kula and Izanagi plates moved in a
general south to north direction through the Pacific basin, implying rapid subduction
beneath Eurasia and right lateral oblique subduction with respect to North America. In
contrast, the Farallon plate swept in a general west to east trajectory across the basin
and was accompanied by rapid subduction beneath North America. The Kula and
Izanagi plates were capable of transporting allochthonous terranes rapidly northward
toward the paleopole. The Farallon plate was capable of transporting terranes bearing
Tethyan fauna eastward across the Pacific basin and juxtaposing those terranes against
the western edge of North America, with moderate displacementstoward the paleopole.
A set of maps showing reconstructed plate boundaries for the past 140 Ma provides
the basis for interpreting terrane displacement histories. The reconstructions are also
used to estimate the ages of the plates that were consumed at convergent plate boundar-
ies throughout the late Mesozoic and Cenozoic eras. The following events and relation-
ships are noted: (1) fast (greater than 100 km/m.y.) convergence of the Farallon plate
with respect to North America during Late Cretaceous and early Tertiary times (75 to 40
Ma); (2) rapid trenchward migration of western North America over the hotspots during
this same time interval; (3) rapid decrease in age of the subducting Farallon plate within
this interval of fast convergence; (4) synchroneity of these three processes with the
Laramide deformation; (5) decrease in Farallon-North America and North America-
hotspot velocities at about 40 Ma as the age of the subducting Farallon lithosphere
decreased rapidly to less than 30 m.y.
Our analysis shows that the age and bathymetry of the descending plates varied
markedly along strike of the trenches. Fracture zones on the Farallon plate, across
which large age offsets occur, were characterized by shallow (young) and deep (old)
ocean floor on opposite sides of the fracture zones. As the fracture zones migrated north
with the Farallon plate along the continental margin, buoyant young lithosphere capable
of producing uplift and erosion existed immediately adjacent to dense old lithosphere
capable of forming a bathymetric low. We speculate that some of the basins that formed
along the continental margin during late Cretaceous and Tertiary times may have origi-
nated in these source-sink pairs.
The relative velocities between continents and adjacent oceanic plates are shown on
a series of maps as arrows representing velocity vectors at selected points of tectonic
interest around the Pacific margin. Significant changes through time in these relative
plate velocities offer insights into the mechanisms that control the diversity of tectonic
styles found in the geologic record at the margins.
 Relative Motions Between Oceanic and Continental Plates

Figure 1. Global plate tectonic model showing the links between the major plates included in our
analysis. In addition, the motion of Eurasia (not shown here) was determined from that of North
America using the record of the opening of the North Atlantic (Srivastava 1978). South America and
Antarctica rotations (not shown here) are included in the tables. PA: Pacific plate. FA: Farallon plate.
KU: Kula plate. NA: North America plate. AF: Africa plate. HOT SPOTS: Those in the Pacific and
Atlantic basins, here shown schematically, are assumed to have remained at a constant distance from
each other.

when hotspots have shifted relative to the rotation axis, hotspots
on one side of the globe moved toward the rotation axis and those
on the other side moved away, as if the mantle as a whole were
moving relative to the rotation axis while the hotspots remained
fixed and at the same distance from each other relative to the
mantle (Morgan 1981; Gordon and Cape 1981; Gordon 1983;
Andrews in press). On balance, hotspots appear to us to form an
imperfect but still useful mantle reference frame.
Reconstructions Used to Build the Model
The reconstructions used to build the model come from a
variety of sources. The motion of the African plate relative to the
Atlantic hotspots (i.e. Africa-hotspot motion) was taken from
Morgan (1983). The ages of these reconstructions were adjusted
to be consistent with the magnetic reversal time scale of Harland
and others (1982), which is used throughout this paper. To obtain
the motion of North America relative to hotspots, the Africa--
North America relative motion determined by Klitgord and oth-
ers (1984 and personal communication) was combined with
Morgan's Africa-hotspot motion. The motion of Eurasia with
respect to the hotspots was then determined from North Ameri-
ca-hotspot motion and the Eurasia-North America reconstruc-
tions of Srivastava (1978). These reconstructions, as modified by
the interpretation of Cande and Kristoffersen (1977), showed that
the anomalies Srivastava identified as anomalies 31 and 32 are
actually anomalies 33 and 34. The timing of initial rifting of
Eurasia-Greenland from North America (-95 Ma) was esti-
mated by extrapolating the spreading rate between anomalies 33
and 34 out to the ocean-continent boundary as determined by
Srivastava.
The post-43 Ma Pacific-hotspot motion, that is, the motion
of the Pacific plate relative to the hotspots, is generally repre-
sented by a single Euler pole (Jarrard and Clague 1977). The
possibility that the Pacific plate began to move in a more north-
erly direction several million years ago is suggested (Cox and
Engebretson 1985) by a difference between the Euler pole for the
Hawaiian trend (Jarrard and Clague 1977) and the Euler pole for
present plate motions, as determined from observations spanning
the past several million years (Minster and Jordan 1978). A
4 Engebretson, Cox, and Gordon

TABLE 1. Sources of Euler poles used for displacement model. Age (Ma) Latitude Longityde Angle
Reference
From 1 To ON (Deg)
Age (Ma) Latitude Longitude Angle
?rom 1 To ON (Deg) PACIFIC-FARALLON N.o.M.@)
--
---
5 0 -19 2 206 5 -3.1 11
PACIFIC-HOTSPOTS
9 5 -77.9 68.0 -5.0 11
5 0 56 5 -75.1 4.7 5 17 9 -88 0 252.1 -6.5 11
28 0 66.6 -71.0 22.8 2 28 17 -80 6 245.1 -7.5 12
43 0 67 4 -70.5 33.2 2 37 28 -79.7 31.9 -13.1 12
74 0 52.2 -90.7 49.1 1 43 37 -79 7 31.9 -11.0 12
100 0 48.8 -81.6 72.1 1 48 43 -79 7 31.9 -8.0 12
115 0 53.9 -79.8 77.1 1 56 48 -77.4 358.4 -10.6 12
135 0 56 1 -77.2 86.5 1 61 56 -66.0 244.2 -3.8 12
145 0 61.1 -78.8 87.3 1 66 61 -66.0 244.2 -3.4 12
74 66 -66.0 244.2 -5.7 12
AFRICA-HOTSPOTS 74
85 -66.0 244.2 -7.8 12
5 0 50.4 323.1 -1.3 6 100 85 -64.9 236.0 -13.1 12
9 0 50.4 323.1 -2.4 6 119 100 -64.9 236.0 -16.1 12
17 0 50.4 323.1 -4.5 6 127 119 -9.2 195.7 -7.4 12
37 0 46.3 43.1 -9.8 6 135 127 -9.2 195.7 -7.4 12
48 0 46.2 43.1 -12.8 6 145 135 -25.3 227.3 -5.9 12
85 0 34.7 -56.9 -22.3 6 163 145 -25.3 227.3 -11.3 12
119 0 26.0 -44.7 -31.8 6 PACIFIC-FARALLON s.o.M.^
145 0 23.9 -45.8 -34.1 6
180 0 .5 -71.4 -22.7 6 85 74 -66.0 244.2 -7.8 12
100 85 -64.9 236.0 -18.9 12
AFRICA-NORTH AMERICA 119 100 -64.9 236.0 -23.9 12
37 0 70.5 341.3 -10.4 3 127 119 -9.2 195.7 -7.4 12
66 0 80.4 -8.6 -22.5 3
PACIFIC-KULA
85 0 77.1 -19.2 -29.4 3
119 0 66.3 -19.9 -54.3 3 43 0 0.0 0.0 0.0 12
127 0 66.1 341.0 -56.4 3 48 43 -18.0 291.4 -6.5 12
145 0 66.1 341.6 -59 8 3 56 48 -18.0 291 4 -10.4 12
161 0 67.2 -16.0 -64.7 3 61 56 -18.0 291.4 -1.7 12
180 0 67.0 -12.0 -75.6 3 66 61 -18.0 291.4 -3.0 12
74 66 -18.0 29 1:4 -7.7 12
NORTH AMERICA-EURASIA 85 74 -18.0 291.4 -10.7 12
37 0 68.0 129.9 7.8 9 PACIFIC-IZANAGI I
48 0 50.8 142.8 9.8 9
53 0 40.0 145.0 11.4 9 100 85 48.0 182.7 -22.6 12
85 0 70.8 150.9 20.3 9 119 100 48.0 182.7 -27.8 12
95 0 75.7 153.8 24.2 9 127 119 48.0 182.7 -12 3 12
180 0 75.7 153.8 24.2 9 135 127 48.0 182.7 -19.1 12
145 135 14.7 267.5 -8.2 12
AFRICA-SOUTH AMERICA
PACIFIC-PHOENIX
37 0 57.4 -37.5 -13.4 8
85 0 63.0 -36.0 -33.8 8
119 0 46.6 -31.1 -52.4 8
135 0 44.1 -30.3 -56.1 8
180 0 44.1 -30.3 -56.1 8

1 ANTARCTICA-AFRICA
 neianve Motions Between Oceanic and Continental Plates 5

TABLE 1. (Continued) probability that their pole was biased by the inclusion in their
Age (Ma) Latitude Longitude Angle analysis of hotspot tracks as old as 10 Ma. The motion of the
From 1 To O
E (Deg) Pacific plate relative to the hotspots for 43 Ma to 5 Ma was
determined using the Pacific-hotspot Euler pole of Jarrard and
Clague (1977). The 28 to 5 Ma angle was determined by requir-
123 -3.9 -27.8 45.5
145 -2.4 -32.7 55.4
ing that the Hawaiian hotspot lay beneath Midway Island at 28
Ma, its radiometrically determined age (Dalrymple and others
AUSTRALIA-ANTARCTICA 1980). The 43 to 28 Ma angle was determined by locating the
10 0 8.7 35.6 -6.7 10 Hawaiian hotspot beneath Yuryaku Guyot at the bend in the
20 0 9.0 12.1 -12.0 10 Hawaiian-Emperor Seamount Chain at 43 Ma, its radiometri-
37 0 11.7 31.8 -20.5 10 cally determined age (Dalrymple and others 1980). The pole of
43 0 11.5 31.0 -23.6 10 motion from 74 to 43 Ma is essentially the Emperor pole of
85 0 15 37.0 -28.0 4 Clague and Jarrard (1973), adjusted slightly to provide a match
PACIFIC-ANTARCTICA at the Emperor-Hawaii elbow with our track for 43 to 0 Ma.
Our rotations for 74 to 43 Ma and from 43 to 0 Ma move Tenji
10 0 72.0 -70.0 9.8 10 'Seamount from its present location in the northern Emperor
20 0 71.3 -73.2 15.4 10 Chain to the present location of the Hawaii hotspot.
37 0 74 8 -56.9 28.0 10
For motion from 74 to 100 Ma, we used a stage pole (36ON,
43 0 75.1 -51.3 32.6 10
56 0 71.6 -57.5 40.1 10 76OW) that Epp (1978) found by fitting small circles to the three
67 0 71.7 -41.0 53.8 10 northwest-striking seamount chains north of the Hawaiian Chain.
Small circles about this pole are also a good fit both to the
INDIA-AFRICA Emperor Chain north of 50 and to the northern Line Islands,
40 0 13.3 54.1 -20.8 7 which we assume to have been originally formed in Late Cre-
50 0 12.9 45 3 -30.1 7 taceous time (Schlanger and others 1984). The rotation angle for
63 0 18 8 26.2 -38 4 7 this stage was determined by finding the angle that placed the
85 0 18.7 25.8 -56 0 7 intersection of the Line Islands and the Mid-Pacific Mountains
123 0 24.5 33.5 -59 0 7 over the present-day location of Henderson and Gordon's (1981)
145 0 29.6 36.1 -56.8 7 Easter plate hotspot (22OS, 114OW).
(1) Based on spacing of isochrons north of the Mendocino Fracture For pre-100 Ma motion of the Pacific plate relative to hot-
Zone. spots, Morgan (1972) proposed a very simple model, a single
(2) Based on spacing of isochrons south of the Mendodno Fracture stage pole located at the present north pole. The model we used
Zone. Entries other than those listed are the same as for Farallon N.O.M.
(Henderson and Gordon 1981) resembles Morgan's but is some-
what more detailed. As in Morgan's (1972) model, the Mid-
References:
(1) Henderson and Gordon (1981); (2) Jarrard and Clague (1977); (3) Pacific Mountains are interpreted as the older continuation of the
Klitgord et al. (1982); (4) Koenig (1980); (5) Minster and Jordan Line Islands Chain, which is assumed to be a hotspot track.
(1978); point on error ellipse for Euler pole closest to equator; (6) However, the Mid-Pacific Mountains were not fit by a single
Morgan (1980); (7) Norton and Sclater (1979); (8) Sibuet and Mascle Euler stage pole (as in Morgan 1972) but by three sub-stages
(1978); (9) Srivastava (1978); (10) Stock and Molnar (1982); (11) corresponding to the west-southwest trend of the easternmost
Wilson et al. (1982); (12) this paper.
(and presumably youngest) Mid-Pacific Mountains (100-115
Ma), the northwest trend of the central portion of the Mid-Pa-
cific Mountains (115-135 Ma), and the southwest trend of the
possible consequence of this change in direction of motion can be westernmost (and presumably oldest) portion of the Mid-Pacific
seen in the trend of the chain through the islands Oahu, Maui, Mountains (135-145 Ma). These rotations are consistent with the
and Hawaii, which is rotated 5 to 10 degrees clockwise relative to interpretation that much of the Hess and Shatsky rises are the
the trend of the older islands. We do not know whether this trace of a single hotspot, active at the same time as the hotspot
change in trend is due to irregularitiesin the pattern of volcanism that formed the Mid-Pacific Mountains. The youngest pre-100
forming on a uniformly moving plate or to a recent change in the Ma segment of this hotspot track, corresponding to the eastern-
direction of Pacific plate motion. most Mid-Pacific Mountains, passes through the west-southwest
In the present analysis we have adopted the latter view and trending platform of the southern Hess Rise that lies just north of
divided the post-43 Ma motion into three stages. The motion for and sub-parallel to the Mendocino Fracture Zone. The interme-
the past 5 m.y. was based on Minster and Jordan's (1978) model diate segment of the hotspot track, corresponding to the central ,
AMI-2 for present Pacific-hotspot motion. For the present anal- Mid-Pacific Mountains, begins with Northwestern Ridge on Hess
ysis we selected a pole at the southern limit of the confidence Rise near the southwestern edge of the rise, crosses the Emperor
limit listed for the Minster and Jordan (1978) pole because of the Chain, and then joins Shatsky Rise where it meets the Emperor
 Engebretson, Cox, and Gordon

TABLE 2. Stage poles and total reconstructions relative to the hotspots.

Stage Total ~econstructions(')
Age Lat Long Angle Lat Long Angle
(Ma) ON (D eg) ON (Deg)
- -

AFRICA [AF-AHS]^
5 -50 143 1.3 -50 143 1.3
9 -50 143 1.1 -50 143 2.4
17 -50 143 2.1 -50 143 4.5
37 -43 132 5.3 -46 137 9.8
48 -46 137 2.9 -46 137 12.7
85 -2 1 108 10.6 -35 123 22.3
119 -2 149 11.4 -26 135 31.8
145 0 116 2.6 -24 134 34.1
180 32 359 19.8 0 109 22.7
NORTH AMERICA [NA-AF-AHS]
5 35 122 .5 35 122 .5
9 35 121 .4 35 122 .9
17 34 119 .8 35 120 1.8
37 32 104 30 33 110 4.7
48 51 115 3.3 40 112 7.9
66 46 91 7.9 44 102 15.6
85 41 59 6.0 46 90 21.0
119 55 21 19.3 59 63 37.4
127 40 41 2.2 59 61 39.4
145 43 52 3.5 58 59 42.8
161 41 6 13.7 59 42 54.4
180 40 7 21.4 58 25 73.8
EURASIA [EU-NA-AF-AHS]
5 -85 70 .7 -85 70 .7
9 -83 66 .5 -85 68 1.2
17 -84 67 1.1 -84 67 2.3
37 -62 77 2.6 -72 76 4.9
48 42 10 3.0 -40 38 4.1
53 36 9 2.8 -10 27 5.3
66 -8 104 4.9 -13 64 8.0
85 -30 88 5.6 -2 1 73 13.2
95 15 51 2.5 -15 70 15.2
119 55 21 13.6 21 58 21.8
127 40 41 2.2 23 57 23.8
145 43 52 3.5 26 57 27.2
161 41 6 13.7 36 45 38 2
180 40 7 21.4 42 33 57.7
SOUTH AMERICA [SA-AF-AHS]
5 76 320 .5 76 320 .5
9 77 322 .4 76 32 1 .9
17 76 325 .8 76 322 1.8
37 70 47 2.6 76 21 4.3
48 77 60 2.2 77 34 6.5
85 53 58 12.6 61 52 18.6
119 22 347 11.2 54 13 26.2
135 3 360 3.5 49 9 28.4
145 0 116 1.O 49 12 28.2
180 32 359 19.8 43 3 47.3
 Relative Motions Between Oceanic and Continental Plates

TABLE 2. (Continued)
Stage Total ~econstructions(~)
Age Lat Long Angle Lat Long Angle
(Ma) ON OR (Deg) ON Peg)

PACIFIC [PA-PHs]
5 57 285 4.7 57 285 4.7
28 69 289 18.2 67 289 22.8
43 69 289 10.4 67 290 33.2
74 22 269 20.2 52 269 49.1
100 48 299 24.0 49 278 72.1
115 68 162 8.0 54 280 77.1
135 75 273 10.0 56 283 86.5
145 24 162 7.0 61 281 87.3
180 - - - 61 281 87.3
FARALLON N.o.M.^ [FA-PA-PHs]
5 33 244 5.1 33 244 5.1
9 -64 2 2.1 12 255 4.1
17 -12 299 2. 3 27 1 6.0
28 10 295 4.1 5 281 9.9
37 -58 14 7.6 -30 301 11.8
43 -60 25 7.3 4 9 317 16.6
48 -70 340 6.5 -56 319 22.7
56 -55 333 8.8 -56 323 31 4
61 -27 288 4.8 -52 318 35.3
66 -23 290 4.5 4 8 315 39.1
74 -22 293 7.4 4 4 312 45.5
85 7 316 10.9 -35 318 53.0
100 1 325 15.9 -28 324 66.3
115 -14 14 6.5 -30 328 70.8
119 -12 355 2.4 -30 330 72.9
127 -8 280 6.8 -27 327 77.2
135 -8 284 6.8 -24 324 82.1
145 2 206 1.9 -23 323 81.2
163 -28 328 11.3 -24 323 92.5
180 -28 328 10.7 -25 323 103.1
FARALLON s.o.M.(')
 Engebretson, Cox, and Gordon
TABLE 2. (Continued)
Stage Total ~econstructions(~)
Age Lat Long Angle Lat Long Angle
(Ma) ON OE (Deg) ON (Deg)

IZANAGI I PZ-PA-PHs]
85 - - - 27 298 80.7
100 27 246 30.1 37 288 103.6
115 * 23 229 24.8 46 278 120.0
119 23 243 6.8 47 275 125.5
127 22 246 14.1 49 269 137.3
135 17 249 20.9 50 260 154.3
145 75 276 6.9 50 262 160.5
IZANAGI I1 [IZ-PA-PHs]
130 - - - 56 282 84.1
135 -22 302 8.9 50 278 86.1
142 -36 285 9.8 44 272 87.3
PHOENIX [PH-PA-PHs]
85 - - - 51 274 58.8
100 41 107 8.9 59 280 59.8
115 9 128 18.5 78 281 56.4
119 10 131 3.8 82 280 56.5
127 9 134 7.7 89 212 57.3
135 9 138 7.7 82 115 58.9
145 5 151 11.6 71 119 62.0
163 -2 1 140 15.1 57 114 62.8
180 -2 1 140 14.3 44 112 66.4
PACIFIC (GLOBAL PLATE CIRCUIT) [PA-AN-AF-AHS]
5 67 290 4.0 67 290 4.0
9 67 289 3.2 67 289 7.2
10 67 288 .8 67 289 8.0
17 59 280 2.8 65 286 10 8
20 65 28 1 1.2 65 285 12.0
37 76 334 10.0 71 302 21.8
40 72 338 1.8 71 305 23.5
43 84 218 2 ,O 73 304 25.4
48 54 252 2.5 72 296 27.6
50 74 255 1.1 73 295 28.7
56 63 30 1 4.0 71 295 32.6
63 70 9 9.6 72 312 41.6
67 61 3 5.5 71 319 46.7
ANTARCTICA [AN-AF-AHS]
5 -87 309 0.9 -87 309 ' 0.9
9 -90 297 0.7 -87 303 1.7
17 -87 280 1.5 -87 291 3.2
37 -8 1 86 3.3 -87 76 6.5
40 -90 165 0.5 -87 78 7.0
48 -23 174 2 .O -77 166 8.0
50 -2 139 0.5 -74 161 8.1
63 59 77 1.2 -70 148 7.2
85 0 5 4.5 -65 39 7.4
119 -4 1 324 17.4 -50 339 23.1
123 -27 328 2.8 -48 338 25.7
145 3 290 8.6 -36 326 30.2
 Relative Motions Between Oceanic and Continental Plates

TABLE 2. (Continued)
(1) Stage poles in a fixed hotspot reference frame. The rotations given are those that will restore the
plates to their positions at times shown in the left-hand column from their positions at the next youngest
times shown in the left-hand column. For example, to restore Africa from its present position to its 5 Ma
position rotate 1.3O counterclockwiseabout the pole with lat. 50Slong. 143OE. Similarly, to restore
Africa from its 5 Ma position to its 9 Ma position rotate l.1counterclockwiseabout the pole with lat.
50Slong. 143OE.
(2) Total reconstruction poles describing the rotation necessary to restore the plates from their present
positions to their positions at the times given in the left-hand column relative to a fixed hotspot frame of
reference. For example, to restore Africa to its position at 85 Ma, rotate 22.3O counterclockwise about
the pole with lat. 3S0S,long. 123OE.
(3) Elements of circuits used to calculate poles are as follows. AF, African plate. AHS, hotspots
producing tracks on Africa and North America. AN, Antarctica plate. EU, Eurasia plate. FA, Farallon
plate. IN, India plate. IZ, Izanagi plates I and 11. NA, North American plate. PA, Pacific plate. PH,
Phoenix plate. PHs, hotspots producing tracks on the Pacific plate.
(4) See note (I), Table 1.
(5) See note (2), Table 1.

Chain at -37O N. The oldest segment of this hotspot track, cor-
responding to the westernmost Mid-Pacific Mountains, trends
southwestward from 37ON and 162OEto the southwest comer of
Shatsky Rise, which we take to be the oldest comer of the rise.
In summary, from 145 to 135 Ma the Pacific plate moved to
the southwest parallel to the western Mid-Pacific Mountains.
From 135-1 15 Ma the Pacific plate moved northwest, parallel to
the central portion of the Mid-Pacific Mountains. From 115 to
100 Ma the Pacific plate moved west-southwest, parallel to the
eastern portion of the Mid-Pacific Mountains. From 100-74 Ma
the Pacific plate moved rapidly to the northwest, parallel to the
northern Line Islands and the three northwest-trending seamount
chains just north of the Hawaiian Chain. From 74-43 Ma the
plate moved rapidly northwestward, parallel to the Emperor
Chain. From 43 Ma to the present, the Pacific plate moved
rapidly in a generally west-northwest direction, parallel to the
Hawaiian Chain, this motion being divided into three small sub-
stages as described above.
The above model allows reconstructions of the Pacific plate
relative to hotspots back only to 145 Ma, whereas reconstructions
are available back to 180 Ma for most of the major plates of
interest. In order to develop a specific model for pre-145 Ma plate
interactions that can be tested against geologic observations, we
extended our Pacific-hotspot model back to 180 Ma by conjec-
turing that the Pacific plate was fixed with respect to the hotspots
between 180 Ma and 145 Ma. The rationale for this assumption
was based on two lines of evidence suggesting that the Pacific
plate had little or no absolute motion prior to mid-Early Creta-
ceous time. First, Larson and Lowrie (1976) presented paleo-
magnetic evidence, based on the skewness of magnetic lineations,
that the Pacific plate moved very slowly, perhaps with clockwise
rotation about a local Euler pole, during latest Jurassic and earli-
est Cretaceous times. Second, it appears that the spreading centers
surrounding the Pacific plate migrated over several hotspots in
earliest Cretaceous time, yet no hotspot tracks older than earliest
Cretaceous now appear on the Pacific plate. If the Pacific plate
had not been approximately fixed with respect to the hotspots
prior to Early Cretaceous time and if all of the numerous hotspots
located near the spreading ridges of the Pacific that were active in
the Early Cretaceous were also active during the Late Jurassic,
then we would expect to observe Late Jurassic hotspot tracks on
the Pacific plate now, which is not the case. Our model for the
interval 180 to 145 Ma, although highly speculative, has the
virtue of making specific predictions which, when compared to
the geology of the circum-Pacific margin, may permit substantia-
tion or refutation of the proposed model.
The relative motion of the oceanic plates (Pacific, Farallon,
Izanagi, Kula, and Phoenix) was determined by Engebretson and
others (1984a) using the magnetic lineations and fracture zones
shown in Figure 2. These plate-plate motions were combined
with the Pacific plate-hotspot motions to obtain the motions of
the oceanic plates with respect to the hotspots. The relative mo-
tions of the Pacific and Farallon (Juan de Fuca) plates younger
than 17 Ma are those of Wilson and others (1984).
The spacing between isochrons commonly changes across
fracture zones in the north Pacific because of ridge jumping or
asymmetric spreading (Menard 1984). Where this has happened,
the spreading rates inferred from the isochrons are not consistent.
In effect, the method of analysis used by Engebretson and others
(1984a) averaged the varying spreading rates and provided error
estimates based on data dispersion. The determination of Faral-
Ion-Pacific spreading during the Cretaceous superchron was a
special case. Data available mainly from plate segments north
and south of the Mendocino Fracture Zone imply that spreading
rates are quite different, Rather than find the average, we deter-
mined two different rates to bracket the range of uncertainty. The
resulting Euler poles and quantities derived from them are listed
in the table under "Farallon N.O.M." for the rates determined
from isochrons north of the Mendocino and "Farallon S.O.M."
for rates determined south of the Mendocino Fracture Zone.
In most of our analyses we use the preferred model of Enge-
bretson and others (1984a), which is based on the spacing of
isochrons on the north side of the Mendocino Fracture Zone
under the assumption that spreading from 118 to 83 Ma occurred
 Engebretson, Cox, and Gordon

TABLE 3. Stage poles and total reconstructions relative to North America.

Stage Total ~econstructions(~)
Age Lat Long Angle Lat Long Angle
(Ma) ON OE (Deg) ON Peg)

HOTSPOTS [AHS-AF-NA~~)
5 -35 302 .5 -35 302 .5
9 -35 30 1 ,4 -35 302 .9
17 -34 299 .8 -35 300 1.8
37 -32 284 3.0 -33 290 4.7
48 -50 297 3.3 -40 292 7.9
66 -48 272 7.9 -44 282 15.6
85 -49 236 6.0 -46 270 21.0
119 -69 194 19.3 -59 243 37.4
127 -50 205 2.2 -59 241 39.4
145 -49 218 3.5 -58 239 42.8
161 -62 162 13.7 -59 222 54.4
180 -60 161 21.4 -58 205 73.8
AFRICA [AF-NA]
37 -71 161 10.4 -7 1 161 10.4
66 -85 233 12.4 -80 171 22.5
85 -67 140 7.1 -77 161 29.4
119 -55 150 25.9 -66 160 54.3
127 -57 171 2.2 -66 161 56.4
145 -64 169 3.4 -66 162 59.8
161 -66 205 5.1 -67 f64 64.7
180 -60 180 11.0 -67 168 75.6
EURASIA [EU-NA]
37 -68 310 7.8 -68 310 7.8
48 -6 330 3.4 -5 1 323 9.8
53 6 326 2.6 -40 325 11.4
85 -78 104 12.1 -7 1 331 20.3
95 -78 104 4.3 -76 334 24.2
180 - - - -76 334 24.2
SOUTH AMERICA [SA-AF-NA]
37 21 302 4.3 21 302 4.3
66 -9 302 4.9 5 301 8.9
85 60 230 1.2 11 298 9.3
119 -57 27 1 13.3 -29 283 18.4
127 -35 272 1.8 -30 282 20.2
135 -17 289 1.7 -29 283 21.8
145 -64 169 1.9 -33 279 22.4
161 -66 205 5.1 -39 272 25.7
180 -60 180 11.0 -49 252 33.0
PACIFIC [PA-PHs-AHS-AF-NA]
 Relative Motions Between Oceanic and Continental Plates
\
TABLE 3. (Continued)

Stage poles^ Total ~econstructions^

A&e Lat Long Angle Lat Long Angle
(Ma) ON '3 Peg) ON '33 Peg)

5 28 248 5.2 28 248 5.2

9 ' -62 347 2.4 4 262 4.6
17 -17 299 3.3 -6 276 7.3
28 -2 29 1 5.4 -5 283 12.7
37 -56 358 8.3 -31 298 16.5
43 -59 7 8.6 -46 309 22.7
48 -64 333 7.9 -5 1 312 30 1
56 -53 315 11.6 -52 3 13 41.8
61 -33 281 6.9 -49 308 47 9
66 -30 282 6.6 -46 306 53.8
74 -29 277 9.2 -42 302 62.0
85 1 290 11.7 -36 304 70 8
100 -16 294 15.3 -31 305 85.2
115 -63 327 8.7 -35 303 92.4
119 -42 305 2.7 -35 303 95.1
127 -12 24 1 8.0 -3 1 300 99.3
135 -7 246 7.7 -27 298 103.8
145 -34 190 3.5 -27 295 104 0
161 -57 247 15.2 -27 288 115.0
163 -58 226 2.3 -27 287 116 4
180 -54 22 1 19.7 -25 277 128.4
FARALLON s.o.M.@)
85 1 290 11.7 -36 304 70.8
100 -22 304 20.0 -33 307 90 3
115 -49 332 14.2 -37 306 103 4
119 -38 316 4.25 -37 307 107.6
127 -12 24 1 8.0 -33 304 111.3
135 -7 246 7.7 -30 302 115.4
145 -34 190 3.5 -29 300 115.4
161 -57 247 15 2 -28 293 126.2
163 -58 226 2.3 -28 292 127.6
180 -54 22 1 19,7 -25 282 138.9
KULA [KU-PA-PHs-AHS-AF-NA]
43 - - - 56 284 32.0
48 -7 303 9.4 43 286 36.9

~1 2:;
56 -7
-6
-3

27
5
298
278
285
288
295
15.4
5.6
6.6
11.5
18.0
28
25
22
19
19
285
283
282
281
285
48.1
52.9
58.8
69.8
87 3

I
 Engebretson, Cox, and Gordon

TABLE 3. (Continued)
Stage poles(^ Total ~econstructions(~)
Age Lat Long Angle Lat Long Angle
(Ma) ON (Deg) ON OR (Deg)
-- -- - - - -

IZANAGI I [IZ-PA-PHS-AHs-AF-NAI
85 - - - 19 285 87.3
100 4 219 31.2 30 269 103.4
115 -5 197 . 27.4 38 252 112.2
119 -1 204 7.3 40 248 116.0
127 9 206 15.2 43 240 126.4
135 9 209 21.6 45 230 142.4
145 54 188 6.1 47 230 147.8
PHOENIX [PH-PA-PHs-AHS-AF-NA]
85 - - - 32 261 59.7
100 -20 123 8.1 35 255 53.7
115 -25 Ill 20.6 43 236 37.8
119 -32 112 4.6 44 229 34.7
127 -19 112 7.9 47 211 31.2
135 -19 Ill 7.6 48 190 29.3
145 -23 121 11.9 38 162 30.7
161 -58 118 25.2 -1 137 35.1
163 -61 123 3.7 -5 134 37.0
180 -64 121 31.1 -30 121 58.8
PACIFIC (GLOBAL PLATE CIRCUIT) [PA-AN-AF-NA]
10 60 290 7.9 60 290 7.9
20 44 282 4.1 54 286 11.9
37 65 304 9.3 59 293 21.1
40 46 310 1.5 58 294 22.5
43 62 265 1.4 58 293 23.8
50 20 261 3.4 55 285 26.4
56 24 276 3.9 52 282 29.7
63 68 309 7.7 55 287 37.1
66 60 316 3.4 55 290 40.4
67 69 314 1.2 55 290 41.5
(1) Conventions as in Table 2.
(2) See note (3), Table 2.
(3) See note (I), Table 1.
(4) See note (2), Table 1.
 Relative Motions Between Oceanic and Continental Plates

TABLE 4. Stage poles and total reconstructions relative to Eurasia.

AFRICA [AF-NA-EU]
37 -22 151 6.7 -22 151 6.7
48 -49 169 5.6 -35 157 11.8
53 -4 1 164 3.5 -37 158 15.3
66 -2 1 276 1.6 -41 164 15.2
85 17 184 1.9 -36 168 16.2
95 -26 184 . 4.3 -34 172 20.3
119 -54 182 18.3 -44 173 38.0
127 -54 202 2.2 -45 174 40.0
145 -61 202 3.4 -47 175 43.2
161 -60 233 5.1 -50 177 47.4

NORTH AMERICA MA-EUl

INDIA DN-AF-NA-EUI
 Engebretson, Cox, and Gordon

TABLE 4. (Continued)
Stage ~ o l e s ^ Total ~econstructions(~)
Age Lat Long Angle Lat Long Angle
(Ma) ON OE (Deg) ON Peg)
f

Stage poles^ Total ~econstructions^

Age Lat Long Angle Lat Long Angle
(Ma) ON OE Peg) ON OE Peg)

PACIFIC [PA-PHs-AHS-AF-NA-EU]
43 - - - 69 282 36.9
48 7 255 3.4 65 273 38.5
53 -4 238 4.1 62 263 39.8
66 21 274 13.2 51 262 50.4
74 28 267 7.6 48 261 57.4
85 51 284 13.1 48 265 70.1
95 46 277 9.4 47 266 79.4
100 23 266 3.6 46 265 82.7
115 -10 186 7.6 48 258 83.0
119 -4 213 1.6 48 256 83.7
127 41 212 3.5 49 255 86.6
135 52 211 3.5 50 254 89.7
145 -10 162 7.2 54 247 88.9
161 -59 195 13.7 50 234 82.6
180 -58 194 21.4 41 215 76.5
FARALLON N.o.M.^ [FA-PA-PI-IS-AHS-AF-NA-EU]
5 38 244 5.6 38 244 5.6
9 -57 358 1.6 27 255 4.7
17 13 298 2.6 23 271 6.8
28 24 293 5.0 23 280 11.6
37 -53 11 6.6 -6 298 11.4
43 -71 32 7.9 -38 307 14.3
48 -75 322 7.2 -50 307 20.5
53 -69 29 1 6.4 -55 303 26.6
56 -37 316 3.8 -53 306 30.2
61 -12 289 6.4 -46 304 35.1
66 -8 290 6.2 -40 304 40.1
74 -5 29 1 8.9 -33 303 47.5
85 22 309 13.2 -23 310 56.0
95 8 316 10.7 -19 313 65.3
100 -11 315 5 .O - 19 314 70.2
115 -63 336 8.6 -23 312 76.3
119 -40 322 2.7 -24 312 78.9
127 -6 264 8.0 -20 309 84.4
135 -1 269 7.7 - 17 307 90.1
145 -30 216 3.5 -16 305 90.6
161 -51 268 15.2 -18 297 101.8
163 -52 250 2.3 -18 296 103.3
180 -48 245 19.7 -17 287 116.4
FARALLON s.o.M.^
 Relative Motions Between Oceanic and Continental Plates

TABLE 4. (Continued)
Stage ~ o l e s ( l ) Total ~econstructions(~)
Age Lat Long Angle Lat Long Angle
'Ma) ON OE (Deg) ON OE (Deg)
--

127 -6 264 8.0 -25 3 13 95.2

135 -1 269 7.7 2 1 312 100.4
145 -30 216 3.5 -20 309 100.8
161 -51 268 15.2 -2 1 302 112.1
163 -52 250 2.3 -2 1 301 113.5
180 -48 245 19.7 - 19 292 126.3
KULA [KU-PA-PHs-AHS-AF-NA-EU]
43 - - - 69 282 36.9
48 -5 305 8.0 57 284 39.6
53 -8 297 7.6 48 283 43.3
56 8 307 6.0 43 284 47.8
61 17 289 6.1 40 283 53.4
66 17 297 7.1 37 284 59.8
74 22 304 12.8 33 286 71.7
85 38 316 20.9 31 293 91.0
IZANAGI I [IZ-PA-PHs-AHS-AF-NA-EU]

IZANAGI II [IZ-PA-PHs-AHS-AF-NA-EU]

PHOENIX [PH-PA-PHs-AHS-AF-NA-EU]
85 - - - 48 265 70.1
95 28 139 4.8 52 264 70.2
100 -2 1 153 2.8 53 261 68.9
115 -28 138 20.5 62 236 57.1
119 -35 140 4.6 63 227 54.7
127 -22 137 7.9 64 208 52.6
135 -22 136 7.6 63 189 51.5
145 -26 147 11.9 54 168 51.9
161 -60 151 25.2 28 149 46.7
163 -62 157 3.7 24 147 46.8
180 -66 157 31.1 -9 136 55.9
PACIFIC (GLOBAL PLATE CIRCUIT) [PA-AN-AF-NA-EU]
10 70 287 9.4 70 287 9.4
20 65 278 5.3 68 283 14.7
37 77 317 12.0 72 295 26.5
40 75 257 1.2 73 294 27.7
43 60 198 1.6 74 290 29.0
48 29 229 2.4 73 276 30.5
50 18 205 1.1 73 270 30.9
53 33 227 1.5 72 264 32.0
56 49 291 2.8 71 267 34.5
63 71 338 10.3 71 282 44.2
 Engebretson, Cox, and Gordon

TABLE 4. (Continued)
Stage Total ~econstructions(l)
Age Lat Long Angle Lat Long Angle
(Ma) ON Peg) ON (Deg)

(1) Same conventions as in Table 2.

(2) See note (3), Table 2.
(3) See note (I), Table 2.
(4) See note (2), Table 2.

Figure 2. Magnetic lineations discussed in text and shown in subsequent figures. After Figure 1 in
Engebretson and others (1984a), which lists the sources of data.

smoothly without ridge jumps. The model proposes that the
spreading rate during the Cretaceous normal superchron is similar
to that before 118 Ma and after 83 Ma. Picks of chron 34 (85
Ma) on several ship tracks north of the Murray Fracture Zone are
consistent with a spreading rate from chron 34 to chron 32b
(85-74 Ma) that is the same as that from chron 32b to chron 25
(74-56 Ma). This model requires an eastward jump of the Paci-
fic-Farallon ridge south of the Mendocino Fracture Zone during
the normal superchron because the Cretaceous quiet zone on the
south side of the Mendocino Fracture Zone is much wider than
on the north. This is part of a pattern in which isochrons on
opposite sides of the Mendocino Fracture Zone showed very little
offset at the time of chron M29 and increasingly large offsets up
to and through the Cretaceous quiet zone and chron 32, subse-
quent to which the offsets show no further systematic increase. If,
on the other hand, spreading on the south side of the Mendocino
occurred smoothly, without ridge jumping, then the ridge north
of the Mendocino Fracture Zone must have jumped during this
time and the true Farallon-Pacific spreading rate recorded on the
south side was substantially greater than that proposed by our
preferred model. The effect of this may be seen by comparing
tables labeled "Farallon N.O.M." and "Farallon S.O.M."
 Relative Motions Between Oceanic and Continental Plates
Errors and Uncertainties
Confidence limits for the Euler poles describing displace
ments between North America and adjacent plates are difficult to
estimate because of the heterogeneity of the data and the uncer-
tainty inherent in the assumption that the hotspots have remained
fixed. Previous workers addressed important aspects of the prob-
lem of calculating confidence limits for Euler poles and propagat-
ing them through plate circuits (Atwater and Molnar 1973;
Molnar and Atwater 1973; Hellinger 1979; Stock and Molnar
1982, 1983; Pilger 1981; Engebretson and others 1984a), but
no one has presented a rigorous analysis of the total error in
reconstructions based on the global circuit or the hotspot refer-
ence frame. The following discussion is thus of necessity semi-
quantitative in nature.
Errors and uncertainties of the following types will be dis-
cussed: (1) errors arising from the assumption that hotspots have
not moved relative to each other when, in fact, they have moved;
(2) errors in determining plate-hotspot motions; (3) errors in
determining plate-plate motions; and (4) uncertainties in the lo-
cation of boundaries between oceanic plates.
Motion between hotspots. Several lines of evidence point
to an upper limit of 3 km/m.y. for the relative velocities between
the hotspots in one ocean basin like the Atlantic or Pacific where
multiple hotspot tracks on one plate provide a measure of the
internal consistency of plate-hotspot motion (Morgan 1983;
Henderson and Gordon personal communication, 1981). Since
several hotspot tracks are used to determine the motion of large
plates like the Pacific, errors due to motion of these hotspots
relative to each other will be cancelled partially by averaging.
Larger errors are probably introduced by systematic motions be-
tween entire groups of hotspots in different ocean basins. If we
assume that the mean position of the Atlantic group of hotspots
has moved with respect to the mean position of the Pacific group
of hotspots in a constant direction for 100 million years at 3
km/m.y., then the cumulative error in the displacements calcu-
lated by our model will be 300 km.
Errors in determiningplate-hotspot motions. These er-
rors may well be the largest in our analysis. The main uncertainty
arises from the possibility that a hotspot track has been correlated
with the wrong hotspot. Moreover if the inaccuracy of the recon-
struction age is taken into account, this type of error probably has
a strong directionality. The motion along hotspot tracks is well
defined because the tracks are narrow, whereas progression along
the track is generally poorly defined because ages along the track
are generally poorly known. The motion of the Africa-North
America plate pair relative to the Atlantic hotspot system appears
to be reasonably well determined back to 180 Ma (Morgan 1981,
1983). W. J. Morgan (personal communication 1981) estimates
that the uncertainty in early reconstructions of North America
with respect to the hotspots is 500 km. Recognizing that smaller
errors are to be expected in reconstructions for relatively recent
times, and assuming an approximately linear increase in errors
with increasing age, a plausible estimate of errors in North Amer-
17
ica-hotspot reconstructions is 300 km at 50 m.y., 500 km at 100
m.y., and 700 km at 145 m.y. Errors for reconstructions of the
Pacific plate with respect to the hotspots are probably similar to
these values back to 74 Ma, substantially larger from 74 to 145
Ma, and of unknown magnitude prior to 145 Ma.
Errors in determiningplate-plate motions. Because the
Farallon, Kula, Izanagi, and Phoenix plates have been subducted
along with any hotspot tracks they may once have carried, de-
termining their motions in the hotspot reference frame requires
using plate-plate motions between these plates and the Pacific
plate. Errors in these motions contribute to the total error in the
motion between these oceanic plates and North America and
Eurasia.
In determining the relative motion between these oceanic
plates, the largest uncertainties are associated with attempting to
trace the motion of the Izanagi, Kula, Farallon, and Phoenix
plates relative to the Pacific plate through the Cretaceous normal
superchron, a time from chron MO (1 19 Ma) to chron 34 (85
Ma) when recognizable magnetic anomalies were not recorded
on the ocean floor. For the Pacific and Farallon plates, the anom-
alies and fracture zones that formed just before or just after the
Cretaceous normal superchron are connected by the through-
going Mendocino Fracture Zone and the relative motion poles for
time intervals on either side of the Cretaceous normal superchron
suggest that no great discontinuity in plate motion occurred dur-
ing the superchron (Engebretson and others 1984a). In the ab-
sence of similar data for Kula-Pacific, Izanagi-Pacific, and
Phoenix-Pacific motion, we simply extrapolated the known an-
gular velocities of stages immediately before and after the Cre-
taceous normal superchron into the midpoint of the superchron.
Having adopted the model of Woods and Davies (1982) for an
85 Ma beginning of the Kula plate, we extrapolated the Pacific-
Izanagi relative motion forward in time into the Cretaceous nor-
mal superchron and ended it at 85 Ma. The presence of a
magnetic bight in the isochron pattern back to chron M29 (164
Ma) indicates that spreading occurred between the Pacific-Faral-
Ion and Pacific-Izanagi plates at least back to chron M29. Pacif-
ic-Farallon spreading before chron M26 was estimated by
extrapolation from the next youngest stage, as was that for Pacif-
ic-Izanagi spreading before chron M16. Any changes in relative
motions that may have taken place during the extrapolated inter-
vals are unrepresented by our model.
Engebretson and others (1984a) estimated errors for the
angular velocities corresponding to their stage poles describing
the motion between the Pacific and adjacent oceanic plates, but
not for their finite rotations, each of which is the sum of a set of
stage poles. Because the error in one stage covaries with the errors
in adjacent stages, it is impossible to combine these estimates of
errors in a simple straightforward manner to yield estimates of
errors in the displacements. Taking a different approach, Enge-
bretson and others (1984a) estimated the magnitude of the errors
in their displacement model by constructing a synthetic Mendo-
cino Fracture Zone using their finite rotation poles to rotate the
present intersection of the Gorda spreading center and the Men-
18 Engebretson, Cox, and Gordon
docino Fracture Zone back to earlier locations. Comparison of
this synthetic fracture zone with the observed location of the
Mendocino Fracture Zone and the age progression of magnetic
isochrons along it revealed a maximum difference between the
two of less than 200 km during the last 165 m.y. Because the
Pacific-Farallon spreading center lay near the west coast of
North America, the errors are not multiplied by geometrical ef-
fects described by Stock and Molnar (1983). Therefore, 400 km
at an age of 100 Ma is probably a pessimistic estimate of the
errors in plate-plate displacements in our model.
Uncertainty in the location of boundaries between
oceanic plates. The history of convergence along a continental
margin depends critically upon knowing which converging oce-
anic plate lay adjacent to the margin, yet this is commonly the
least well-constrained element of plate models. This ambiguity in
plate geometry, although difficult to quantify, may be more signif-
icant than the accumulated errors in the motions between plates
discussed earlier. Consider, for example, the Farallon-Kula di-
vergent boundary, across which rates of divergence can be rea-
sonably well determined by combining finite rotations deduced
from the Pacific-Farallon and Pacific-Kula spreading histories.
Because all of the oceanic lithosphere produced by spreading
from this boundary has been destroyed by subduction, the loca-
tion of this boundary is almost totally unconstrained by direct
observation. Therefore at a given point along the west coast of
North America, the main uncertainty in the rate and direction of
convergence stems less from statistical errors in determining plate
velocities than from not knowing whether the Kula or Farallon
plate was adjacent to the point.
Summary of errors. To summarize, we estimate that the
different errors present in a 100 Ma reconstruction are the
following:
(1) A 500 km error in the location of the North American
plate with respect to the group of hotspots in the Atlantic region
due to imperfect knowledge of bathymetry and age progressions
along hotspot tracks,
(2) A 500 km error in the location of the Pacific plate with
respect to the group of hotspots in the Pacific region due to
imperfect knowledge of bathymetry and age progressions along
hotspot tracks;
(3) A 100 km (= 1 km/m.y. x 100 m.y.) error in the loca-
tion of the North American plate with respect to the hotspots due
to possible wander between hotspots in the Atlantic region;
(4) A 100 km (= 1 km/m.y. x 100 m.y.) error in the loca-
tion of the Pacific plate with respect to the hotspots due to possi-
ble wander between hotspots in the Pacific region;
(5) A 300 km (= 3 km/m.y. x 100 may.) error in relative
location due to wander between hotspots in the Atlantic and
Pacific regions; and
(6) A 400 km error in the displacements between the Pa-
cific and adjacent oceanic plates.
These uncertainties are all presumably independent and can
be combined to obtain the total uncertainty from the following
relationship:
where OT is the total uncertainty and the or; are the individual
uncertainties listed above. This yields a combined uncertainty in
location of 900 km, our best estimate of the error in our recon-
structions for 100 Ma. We think that this error can be signifi-
cantly decreased as more information regarding ancient hotspot
tracks is obtained.
Not included in the above analysis are the uncertainties in
the location of the boundaries between the Farallon, Kula, and
other oceanic plate pairs for which isochrons and fracture zones
are not preserved. For example, at 70 Ma the velocity of conver-
gence of an oceanic plate with North America at the latitude of
San Francisco may have either of two quite different values,
depending on whether the oceanic plate was Farallon or Kula.
The resulting uncertainty is difficult to express quantitatively. The
information needed to constrain the location of these boundaries
lies not in the ocean basins but along continental margins where
the geologic record may reflect changes in the motions of oceanic
plates and the passage of triple junctions. A first step in looking
for correlations between plate motions and geology along the
continental margin is to construct alternative models showing
possible former locations of the plates and their convergence
rates. One such model is described in the next section.
RECONSTRUCTIONS
Method, Data, and Assumptions
Plate reconstructions for the Pacific basin are shown in Figure
3 in a reference system in which the hotspots are held fixed. These
reconstructions were generated by rotating magnetic anomalies
and fracture zones from their present locations on the Pacific
plate back to their earlier locations using the Pacific-hotspot
poles listed in Tables 1 and 2. These and the other poles used in
our reconstructions can be traced to original sources in Tables
1-4, which list the finite plate rotations that occurred between
stage boundaries-times of marked change in the motion of one
or more of the plates. Known plate boundaries based on iso-
chrons and fracture zones now preserved on the Pacific plate are
depicted with solid double lines (ridges) or solid single lines
(transforms). Speculative ridges and transforms in regions where
neither of a plate pair is preserved are depicted with dashed lines
drawn perpendicular and parallel to tangents of small circles
about the appropriate relative-motion pole for the time of the
reconstruction. The locations of these boundaries are highly
uncertain.
As we venture further and further back into the geologic
record, there are fewer and fewer constraints to guide the drawing
of ancient boundaries between once enormous plates of which
only small remnants of boundaries have survived. We have been
able to remove some of the arbitrariness in drawing these bound-
aries by making plots of velocity fields describing plate motions in
Relative Motions Between Oceanic and Continental Plates
Engebreison, Cox, and Gordon
Relative Motions Between Oceanic and Continental Plates
Engebretson, Cox, and Gordon
Relative Motions Between Oceanic and Continental Plates
Relative Motions Between Oceanic and Conti~ntalPlates
140 180 220 260
Figure 3f
Relative Motions Between Oceanic and Continental Plates
Engebretson, Cox, and Gordon
 Relative Motions Between Oceanic and Continental Plates
the hotspot reference frame. These have proved very useful in
determining whether the Mesozoic boundaries between paleo-
plates like Izanagi and Farallon were ridges, transforms, or sub-
duction zones.
In order to make successive reconstructions mutually con-
sistent we have employed forward and backward modeling con-
sistent with first-order rules of plate tectonics. For example, if at a
given time we specify the location of a ridge between two diverg-
ing plates, for earlier and later times we use forward and back-
ward modeling to find the location of the ridge, under the
assumption that spreading was symmetric and continuous. Maps
of our reconstructions are presented in Figure 3 where the
present-day outlines of continents serve as a familiar reference for
showing the displacements of continents relative to hotspots. Of
course in early Tertiary and Mesozoic time, the actual margins of
the continents were substantially different from the present ones.
Detailed Reconstructions
Present configuration. The isochrons and fracture zones
shown on the Pacific plate in Figure 3a comprise the basic data
set used to make the reconstructions. These isochrons, which give
previous locations of the ridge system around the Pacific plate,
are reconstructed in Figure 3 using known motions of the Pacific
plate relative to the Hawaiian (H in Figure 3) and other Pacific
hotspots. North America was rotated back to earlier positions
over North American and Atlantic hotspots including Y, the
Yellowstone hotspot, under the assumption that these hotspots
existed at these earlier times and remained fixed relative to Pacific
hotspots. Arrows showing the direction and velocity of plate
motions in this and succeeding figures are scaled to show the
distance the plate would move in 10 m.y. based on the stage pole
angles listed in Table 2. The 0-5 Ma stage was used for Figure 3a.
20 Ma (chron 6). The velocity fields for this reconstruction
are based on the Euler poles for the stage from 17 to 28 Ma
(Table 2). Note that the part of the Pacific plate being subducted
beneath the Aleutian arc at this time was former Kula plate
(Fig. 3b). The location of the extinct Pacific-Kula spreading
center, assumed to have died at 43 Ma, is shown by the (x-x-x-x-)
pattern. This extinct ridge arrived at the Aleutian trench at - 10
Ma, and, since it would have been only 30 million years old at
that time, might have been sufficiently buoyant to have produced
compressive tectonics in the Aleutian arc.
37 Ma (chron 13). In our reconstruction, the Pacific plate
first came into contact with North America at about 37 Ma
(Fig. 3c). The Pacific-Farallon ridge, as shown in this reconstruc-
tion, lay slightly east of the present margin of the continent-an
impossible plate geometry. However the amount of the overlap
was smaller than the expected errors in reconstructions. More-
over the overlap would have been smaller if, in the model, we
had closed the Gulf of California and restored the continental
margin to a more easterly 37 Ma location to account for the
effects of Basin and Range extension. In a subsequent section, this
result is compared to the analysis by Atwater and Molnar (1973)
based on the global plate circuit.
27
56 Ma (chron 25), 65 Ma (chron 29), and 80 Ma (chron
33r). The location of the Kula-Farallon boundary during the
Late Cretaceous and early Tertiary, which is not constrained by
marine data, is crucial to the present analysis. We have adopted
Woods and Davies's (1982) interpretation that the Kula plate
broke from the ancestral Farallon plate at about 85 Ma and that
subsequent Kula-Pacific spreading produced the east-west mag-
netic lineations south of the Aleutian trench. From the broad
range of possible locations for the Kula-Farallon boundary at the
time of the break, we have considered two end members, a
northern option and a southern option, in which the boundary
between the Kula and Farallon plates is placed either as far north
or as far south as is plausible.
For the northern option the location of the Kula-Farallon
ridge was specified at 56 Ma on the basis of the following geo-
logic reasoning. In northern Washington and British Columbia
the presence of northwest-trending faults with large amounts of
Cretaceous dextral motion suggests a tectonic driving force with a
strong northward component. This would be consistent with a
Kula plate that extended at least as far south as northern Wash-
ington in latest Cretaceous time (Plumley 1980). Accordingly, for
the northern option the model constrains the Kula-Farallon
spreading center at 65 Ma to intersect the continental margin in
southern Washington and connected this control point to the
Pacific-Farallon-Kula triple junction with a single great circle
ridge. Using the known Euler poles for Kula-Pacific and Faral-
Ion-Pacific motion to derive a pole for Kula-Farallon motion
and assuming symmetrical spreading, we modeled the generation
of oceanic lithosphere by Kula-Farallon spreading and found that
the eastern termination of the Kula-Farallon ridge ranged be-
tween northern California and northern Washington from 85 to
56 Ma.
In exploring the southern option, we placed the initial break
between the Kula and Farallon plates along a set of ridges and
transforms trending subparallel to the continental margin. Figure
3f shows a hypothetical boundary at 80 Ma, shortly after the
presumed time of the break between the two plates. Because the
orientation of the Kula-Farallon ridge inferred from velocity
analysis is approximately parallel to the trend of old Farallon-
Pacific fracture zones on the Farallon plate, it is possible that
Kula-Farallon spreading began along these fracture zones, as
shown in Figure 3f. Beginning with the assumption that the break
was as shown for 80 Ma, the reconstruction at 65 Ma was found
using calculated Kula-Farallon Euler poles assuming symmetri-
cal spreading.
Between 65 and 56 Ma the Euler poles describing Kula-Far-
allon motion shifted markedly (Engebretson and others 1984a).
A corresponding change in the orientation of the Kula-Farallon
ridge is shown between the 65 and 56 Ma reconstructions. In the
absence of data, no attempt was made to model details of this
plate reorganization.
The shaded areas in the 80 and 65 Ma reconstructionscould
be either Kula or Farallon plate. The northern option places
Farallon plate adjacent to the western United States from 85 to
28 Engebretson, (.70% and Gordon
56 Ma, producing rapid plate convergence that may have been
associated with Laramide tectonism (Engebretson and others,
1984b). The southern option places the Kula plate along the
continental margin, offering a mechanism for rapid northward
transport of terranes. Further geologic consequences of selecting
the southern option for the Kula-Farallon ridge are discussed by
Page (1984) and Page and Engebretson (1984). By 56 Ma the
part of the Kula plate adjacent to California in the southern
option had been subducted and replaced by the subducting Faral-
Ion plate, shifting the locus of Kula-Farallon divergence to the
north and eliminating most of the difference between the north-
em and southern options (Figure 3d).
In our early Tertiary reconstructions (65-56 Ma) the Yel-
lowstone hotspot lies west of the continent in the vicinity of the
Kula-Farallon ridge. Although differing from each other in de-
tail, several plate tectonic models advanced to account for the
ages and distributions of early Tertiary seamounts in the Wash-
ington and Oregon Coast Range (Simpson and Cox 1977;
Duncan 1982; Wells and others 1984) have in common the
presence of the Yellowstone hotspot and the Kula-Farallon
spreading center at about the location shown in our reconstruc-
tion for 56 Ma.
Similarly, the location of the Kula-Pacific boundary is not
constrained by ocean floor data. We have followed Hilde and
others (1976) in making the boundary a northwest-trending
transform. In our model, the trend of the transform (although not
its location) is determined by the Euler pole for Kula-Pacific
motion. An alternative model would be to extend the well-
documented short segment of Kula-Pacific ridge adjacent to the
Kula-Pacific-Farallon triple junction westward as an east-west,
dominantly ridge-type Kula-Pacific boundary. In this model,
however, the motion of the Pacific plate relative to hotspots
would be toward the Pacific-Kula ridge, an unlikely geometry
with no analogue in the present motions of plates. With the
Pacific boundary drawn as shown in Figures 3e and 3f, the Pacific
plate was moving toward a subduction zone along its northwest
boundary where presumably a slab attached to the plate was pro-
viding slab pull to drive the plate. This geometry is characteristic
of all of the present plates that have rapid velocities relative to
hotspots (Gordon and others 1978).
110 Ma (Albion, middle of Cretaceous normal super-
chron). Although magnetic anomalies older than 85 Ma in the
northwestern Pacific have been attributed to spreading between
the Pacific plate and an ancestral Kula plate, Woods and Davies
(1982) pointed out that there is no evidence that the post-85 Ma
and pre-85 Ma plates were ever continuous in space or time.
They proposed that the older plate to the west be called the
Izanagi plate. The analysis of Enbegretson and others (1984a)
supports this interpretation. In our reconstructions we will refer to
pre-85 Ma plates west of the ancestral Farallon plate as the
Izanagi plate or plates.
The following plate geometry is proposed for the northern
Pacific basin prior to 85 Ma. To the east lay an ancestral Farallon
plate including the northern segment that broke away at 85 Ma to
form the Kula plate. To the west lay the Izanagi plate or plates.
The following information is available concerning the nature of
the boundaries between the three plates. The time of this recon-
struction, 110 Ma, lies within the Cretaceous normal polarity
superchron (118 to 83 Ma), so no isochrons are available to mark
plate boundaries. However, well defined isochrons werepro-
duced by Pacific-Farallon spreading both before and after the
Cretaceous normal superchron, so there is little doubt that the
Pacific-Farallon boundary was a ridge at 110 Ma. Its location
was found by interpolation (Figure 3g). The Pacific-Izanagi
boundary produced a magnetic anomaly at the beginning of the
Cretaceous normal superchron and therefore was a ridge at 119
Ma. Its location at 110 Ma as found by extrapolation from the
119 Ma position is shown in Figure 3g.
To investigate the nature of the Izanagi-Farallon boundary,
velocity fields in a fixed hotspot reference frame were plotted for
both plates (Figure 3g). The two velocity fields converge along
almost any boundary drawn between the Izanagi and Farallon
plates except for a short segment of boundary south of the Izana-
gi-Farallon rotation pole (Figure 3g). Therefore the boundary is
dominantly convergent. To determine which of the two plates is
subducting along this boundary, we noted that the velocity of the
Izanagi plate in the hotspot reference frame was rapid and that it
was directed toward the boundary whereas that of the Farallon
plate in the vicinity of the boundary was low. Rapidly moving
present-day oceanic plates are all attached to downgoing slabs
and are moving toward the trenches where the slabs are being
subducted (Gordon and others 1978), whereas present-day upper
plates generally have small velocities relative to hotspots. There-
fore it seems likely that the Izanagi plate was the subducting plate
and the Farallon the upper plate at the boundary between them,
as shown in Figure 3g, although as noted earlier the location of
this boundary is highly uncertain. The location of the Izanagi-
Farallon boundary is not well determined and could lie well to
the east of the position shown in Figure 3g.
Our reconstructions for 110 and 80 Ma suggest that a major
change took place in the plate geometry of the northwest Pacific
basin during this interval. One scenario consistent with the sparse
amount of available information would be for the Pacific-Izanagi
ridge and the Izanagi-Farallon subduction zone to die at some
time after 119 Ma (MO) and for the new Kula-Pacific transform
boundary to originate in the general vicinity of the old Izanagi-
Farallon subduction zone. The timing of these hypothetical
events, which would have occurred at some time or times during
the Cretaceous normal superchron (118-83 Ma), is uncertain.
The onset of rapid Pacific-hotspot motion toward the northwest
at 100 Ma suggests that at this time the geometry of the Pacific
plate was changed by addition of a subduction zone along its
northwest boundary. One possibility is that at this time the old
Izanagi plate became part of the Pacific plate as the Izanagi-Pa-
cific ridge died.
140 Ma (Berriasian, chron M14). The Pacific-Izanagi
and Pacific-Farallon boundaries are known to have been spread-
ing centers at this time because of the presence on the Pacific
 Relative Motions Between Oceanic and Continental Plates

plate of magnetic anomalies produced along these ridges. Veloc- PLATE TECTONIC INTERPRETATION
ity fields for the Izanagi and Farallon plates in a fixed hotspot
reference frame show that the motion of the two plates was Age and Thickness of the Subducting Farallon Plate
oblique, suggesting that the fcanagi-Farallon boundary was dom-
inantly transform in character (Figure 3h). In going from our The importance of the age and thickness of subducting
reconstruction for 140 Ma to our reconstruction for 110 Ma, we plates in determining the style of subduction has been recognized
were guided by a suggestion of Uyeda and Yamano (1984) that a by many workers. Comparing the present-day tectonic styles of
subduction zone may originate along the locus of an earlier the overriding plates at convergent plate boundaries with the age
transform. of subducting lithosphere at such boundaries, Molnar and At-
water (1978) found that interarc spreading tends to occur behind
Comparison with Global arCuit in theNeogene trenches subducting old lithosphere (>I00 m.y.) whereas
collision-like compressive tectonism tends to occur behind
A key event in the Tertiary plate tectonic history of western trenches subducting young lithosphere (<50 m.y.). This led them
North America was the first arrival of the East Pacific Rise at the to suggest that the age of the subducted lithosphere may be a
continental margin. The location of the rise at or slightly before principal factor causing different tectonic styles and geologic de-
the time of arrival is marked by anomaly 8 (28 Ma), the last to velopment in the overriding plates at subduction zones. They
form prior to the first contact of the Pacific plate with North further suggested that the compressional Laramide tectonic re-
America (Atwater 1970). Our model is compared with that ob- gime of western North America during the Late Cretaceous and
tained using the North America-Africa-India-Antarctica-Pacific early Tertiary orogeny was caused in part by subduction over a
circuit (Atwater and Molnar 1973)in Figure 4 and in Tables 2-4. long interval of time of young, presumably buoyant lithosphere of
The location of anomaly 8 is plotted in Figure 4 for our model the Farallon plate.
(solid line) and the global circuit (dashed line) relative to a fixed In order to determine the age of the Farallon plate as it was
North America. Both reconstructions show the ridge located east being subducted, we constructed a set of synthetic Farallon and
of the present coastline of North America at 28 Ma. (The present Kula plates using the anomalies on the Pacific plate and our poles
North American coastline is shown only for geographic refer- for Farallon-Pacific and Kula-Pacific motion, assuming symmet-
ence.) The next step (that we have not taken) will be to modify rical spreading. The maps of the synthetic plates reflect the pat-
the coastline to allow for Basin and Range extension, the opening tern of isochrons and fracture zones present on the Pacific plate.
of the Gulf of California, widespread strike-slip faulting, and For clarity in the following discussion, the complement on the
other tectonic events that have changed the shape of the continen- Farallon plate of the Murray Fracture Zone, for example, will be
tal margin. Because an error-free reconstruction would place described as the Murray-F Fracture Zone. From the resulting set
anomaly 8 just west of the continental margin restored to its of maps for different geologic epochs, we estimated the age of the
configuration at 28 Ma, an overlap of this type shows that the Farallon and Kula plates at the time when they descended into
error in the reconstruction is at least as large as the overlap. Even the trench. Age estimates for four locations are given in Figure 5.
without moving the continental margin eastward to account for At 65 Ma the Farallon plate being subducted was -100
the westward motion that occurred during Basin and Range ex- m.y. old. Before this, the age of the Farallon plate being sub-
tension, the amount of the observed overlap is, in fact, within ducted (if indeed the Farallon plate existed prior to 165 Ma) is
earlier error estimates in our reconstructions. unknown because no anomalies formed by Pacific-Farallon
The difference between the two reconstructions is less than spreading are preserved. Before 163 Ma, if Pacific-Farallon
their probable errors and therefore the two reconstructions are spreading occurred at the same rate as from 163 to 135 Ma, then
not significantly different. Within experimental errors both mod- the age of the subducting Farallon plate was as shown by open
els fit the Neogene history of northward migration of the Mendo- circles in Figure 5. Under this assumption the age of the subduct-
cino (Pacific-Farallon-North America) triple junction inferred ing lithosphere at 85 Ma just prior to the onset of the Laramide
-
from the history of volcanism, crustal thinning, and basin forma- orogeny was 150 m.y. If the rate of Pacific-Farallon spreading
tion in California (Dickinson and Snyder 1979; Pilger and He- prior to 163 Ma was lower than afterwards, then at 85 Ma the
nyey 1979; Fox 1981; Zandt and Furlong 1982), although the fit age of subducting Farallon lithosphere was less than 150 m.y. but
of the global circuit appears to be somewhat better. On the other probably still greater than the age of 4 0 m.y. that Molnar and
hand, in Paleogene time the global-circuit model is grossly incon- Atwater (1978) associate with compressive tectonics behind sub-
sistent with the hotspot model because the global circuit fails to duction zones.
predict the elbow- in the Emperor-Hawaiian seamount chain The rifting of the Farallon plate to form the Kula plate
(Morgan 1981; Duncan 1981). The large uncertainties in Neo- beginning at 85 Ma (Woods and Davies 1982) caused a major
gene reconstructions based on both the global-circuit and hotspot discontinuity in the age of Farallon lithosphere, with old Farallon
models point to the importance of acquiring geologic and paleo- lithosphere in direct contact with new Farallon lithosphere
magnetic data from the continental margin capable of further formed at the Kula-Farallon spreading center. When this discon-
checking and constraining the plate tectonic models. tinuity contacted the west coast of North America, the age, thick-
30 Engebretson, Cox, and Gordon

Figure 4 (this and following page). Reconstructions of the Pacific plate relative to a fixed North
America plate. Heavy solid line: the chron 8 (28 Ma) Pacific-Farallon isochron obtained using our fixed
hotspot model. Heavy dashed line: 28 Ma isochron obtained using the plate circuit through Antarctica.
Light dashed band: approximate location of the active ridge at the time of the reconstruction. The
rotations used in the global circuit (Table 4) are: North America-Africa, Klitgord and others (1984 and
personal communication 1984); Africa-Antarctica, Norton and Sclater (1979); Antarctica-Pacific,
Stock and Molnar (1982).
Relative Motions Between Oceanic and Continental Plates
 Engebretson, Cox, and Gordon

Northern
Vancouver
I
(m.y.1
10 laland -I Kula_^i_Paclflc ~ . 4
I
Farallon 1 a
Fnad
r a/ o
l lro n (a 1

Pacific

(m.y.1
100 0
'
1
- F a r a l l o n
I (b)

150-- G U Y ~
AQ*
lo*-
(m.y.1 P a c i f i c (d)
50--
a , .
fr u . 1 r . . , I v
1 0 id0 6'0

Figure 5. Age at the time of subduction of plate adjacent to the western edge of North America as a
function of time. Open circles: present ages of Farallon plate which, if it had not been subducted, would
have been > 163 Ma, as estimated by assuming that rate of Farallon-Pacific spreading before 163 Ma
was the same as from 163 to 135 Ma. (a) Northern Vancouver Island, Lat. 51NLong. 232OE, (b) Cape
Mendocino, Lat. 40NLong. 236OE, (c) Los Angeles, Lat. 34ON, Long. 242OE, and (d) Guaymas, Lat.
28ON, Long. 24g0E.

ness, and density of the lithosphere being subducted underwent a irregularly in age and thickness as Kula-Farallon transform faults
sudden transition from old, thick, and dense to young, thin, and arrived a t the trench.
buoyant as lithosphere created at the Pacific-Farallon spreading /

center was replaced by crust created at the Kula-Farallon spread-
ing center. If the northern option for the locus of the initial
Kula-Farallon break is correct, this major age discontinuity
propagated southward from Vancouver Island, reaching Cape
Mendocino at 65 Ma, as shown in Figure 5. If a more southerly
option is correct, young crust could have arrived at the subduc-
tion zone starting at 85 Ma at any point from Vancouver Island to
Mexico. The region of young crust would slowly widen, jumping
Variationsin Bathymetry Parallel to the Trench
The age of the portion of the Farallon plate being subducted
at any given time varied greatly along strike, producing large
variations in the depth of the Farallon plate adjacent to the
trench. An example based on our reconstruction of the Farallon
plate for 52 Ma (Figure 6 ) illustrates this point. The Farallon
plate fracture zones, which mirror those now on the Pacific plate,
 Relative Motions Between Oceanic and Continental Plates 33

T KIII A - F A R A I I nu R I ~ G F a z f t

Figure 6. (a) Reconstruction of the plates at 52 Ma showing the age of the Farallon plate along line
A-A'. Arrows are linear velocities for the points located at the tails of the arrows. (b) Predicted depth of
the Farallon plate along line A-A' from (a) using the measured depth versus age compilation of Sclater
and others (1971).
34 Engebretson, Cox, and Gordon

offset isochrons by large distances, thereby juxtaposing oceanic t I I r I 50

crust of different ages. Ocean floor ages on the reconstructed
Farallon plate at 52 Ma were interpreted using curves of depth NORTH AMERICA
versus age for the North Pacific (~claterand others 1971) to
produce a profile of the bathymetry along the ocean floor adja-
-
cent to the continent (Figure 6b).
These bathymetric and plate reconstructions suggest a com-
plex sequence of events that may be related to the development of
- - 0 - 40
sedimentary basins. Consider a fracture zone like the Murray-F 4
on the Farallon plate with young lithosphere to the north and old
to the south, migrating northward along the coast. At a fixed
- 8
-

point along the margin the adjacent subducting plate was first
young and buoyant, producing uplift and erosion until the frac-
ture zone moved past the point. The subducting plate abruptly
-
became old and dense, providing subsidence conditions suitable
for the development of a sedimentary basin. This geometry ex- -
isted on the Farallon plate between the Surveyor-F and Mendo-
cino-F fracture zones and between the Murray-F and
Molokai-F fracture zones as they migrated northward with the , I I I I I I 1 620'~
220 230 240 250
Farallon plate along the continental margin. Local highs and lows a
along the trend of the fracture zones undoubtedly added to the
complexity of the process.
The direction of motion of the Farallon plate relative to
North America was nearly parallel to the trend of the fracture
zones, much as if the topographic escarpments of the subducting
oceanic plate had formed guiding indentations in the upper con-
tinental plate. As a consequence, the fracture zones appear to
have migrated very slowly along the continental margin during
the early Tertiary, providing ample time for the process of basin UJ

formation.
Mendocino Escarpment. Figure 7a shows the recon-
structed locations of the Mendocino Fracture Zone at 4 m.y.
intervals. The topographic relief across this escarpment is large
because of the great difference in the ages of juxtaposed crust
north and south of the fracture zone. To assess the changing relief
of the Mendocino Escarpment as it migrated northwestward
along the continental margin, we used the techniques described
earlier to make maps of the Farallon plate reconstructed to its
positions at earlier times. At 28 Ma, the Farallon plate north of
the Mendocino was approximately 20 m.y. old while directly to
the south there was an active ridge. Depth versus age data for the
North Pacific (Sclater and others 1971) indicates that the height
of the escarpment was approximately 1.3 km with the north side
down. Successive locations of this plow-like escarpment as it
migrated northward are shown in Figure 7b. With time the es-
carpment became smaller and at about 12 Ma it switched polarity
(Figure 7b), so that today it faces to the south rather than to the
north. The age of the Pacific plate to the south of the Mendocino
Fracture Zone is approximately 28 m.y. whereas that of the Juan
de Fuca plate to the north is approximately 5 m.y., producing a
south-facing escarpment about 1 km high adjacent to the conti-
nental margin.
The space and time association of sedimentary basins with
former positions of the Mendocino triple junction was noted by
28 24 20 16 12 8 4 0
TIME (Ma) b
Figure 7. (a) Palmlocations of the Mendocino Fracture Zone in fixed
North America coordinates at times shown to the left of each paleoloca-
tion. (b) Predicted height of the Mendocino Escarpment near the anom-
aly 8 lineation as a function age. The escarpment faces north for positive
heights and south for negative heights. Numbers in parentheses are the
approximate palmlatitudes of the escarpment.
Dickinson and Snyder (1979) and interpreted in terms of the
geometric instability of the triple junction. Another relationship
that may have caused the formation of basins in the wake of the
triple junction is the presence of a younger, buoyant, bathymetri-
cally high source region on one side of the fracture zone directly
juxtaposed against an older less-buoyant, deep section on the
other side. The northward migration of this source-sink pair
astride the fracture zone may well have contributed to the
formation of basins along the continental margin.
Velocity of Convergence
Also of geologic interest are the velocities of convergence of
oceanic plates relative to continental plates that were found from
the Euler poles given in Tables 2 to 4. Each stage pole describing
 Relative Motions Between Oceanic and Continental Plates
the displacement of one oceanic plate relative to a continent was
used to estimate a quasi-instantaneous angular velocity vector.
The orientation of this vector was taken as that of the stage pole,
and its length was obtained by dividing the finite angle of the
stage pole by the time spanned by the stage. An estimate of the
instantaneous velocity of the oceanic plate relative to a fixed
point on the continent was then found as the cross product of the
quasi-instantaneous angular velocity vector and the position vec-
tor of the fixed point.
The genealogy of the convergence vectors for all of the plate
pairs in the north Pacific basin is traced in detail in Tables 2 to 4
and in Appendix A. Because convergence vectors are found from
chains of linked plate and hotspot motions, the vectors change at
each time corresponding to a time boundary between stage poles
that describe the motion of any of the relevant plates. For exam-
ple, Farallon-North America convergence vectors were found by
adding stage poles for North America-Africa-Atlantic hotspot
motion and Farallon-Pacific-Pacific hotspot motion. The Faral-
Ion-North America convergence vectors change direction and
length, at times corresponding to the stage boundaries for all of
these plate motions. Stages for short time intervals commonly
have poles that vary erratically from the poles of adjacent stages
because of the relatively large magnitude of the errors of short
stages. Where practical we reduced the number of small stages by
eliminating or shifting stage boundaries in the input data sets so
that as many of the input stage as possible had common time
boundaries. This could be done without introducing large errors
where the poles for adjacent stages in the input data set were
closely spaced but could not be done where input stage poles
were widely spaced. The pattern that emerged is one of small,
generally insignificant changes in the vectors over times of the
order of 5 to 10 m.y. superimposed on larger changes over times
of -20 to -50 m.y. It is the latter that we believe reflects true
changes in the angles and rates of convergence between con-
tinents and adjacent oceanic plates.
The geologist interested in possible correlations between
geologic events on a continent and individual peaks and troughs
appearing on plots of convergence velocity versus time will want
to begin by assessing the origin and the reliability of individual
features on the velocity plots. Tables 2 to 4 and Appendix A
provide the basis for making this assessment. In Appendix A the
relative velocities for each plate pair in the circuits used to find
convergence velocities are displayed as plots of velocity versus
time at a point appropriate for each plate pair. These plots show
the genealogy of the peaks and troughs appearing in the final
plots of convergence velocity versus time. These genealogy dia-
grams show that some of the peaks and troughs probably describe
real changes in convergence velocity whereas others are probably
artifacts resulting from some of the errors discussed earlier. Ex-
amples of the use of these plots to trace the origin of apparent
changes in convergence velocity are given in Appendix A.
Velocity vectors that show the direction and speed with
which oceanic plates have converged with North America and
Eurasia are shown graphically in Figure 8. Each arrow along the
35
continental margin shows the convergence velocity of the adja-
cent oceanic plate at the time indicated, which is between the two
times that bound the stage used to estimate the velocity vector.
where there is some ambiguity as to which plate was converging
with the continent at a given locality, two arrows show the con-
vergence velocities of both oceanic plates. Numerical values of
the azimuth and speed of convergence for selected localities
around the north Pacific basin are listed in Tables 5 and 6.
Along the western margin of North America, convergence
at mid-latitudes in the Late Jurassic and Early Cretaceous was
-
dominantly east-west at the high rate of 100 km/m.y. (Figure 8,
maps for 175,150, and 125 Ma). In the Late Cretaceous, conver-
gence was oblique in a northeast-southwest direction (Figure 8,
maps for 100 and 75 Ma). In the Eocene, convergence was
oblique and at a more rapid rate than at any other time for which
data are available (Figure 8, map for 50 Ma). From mid-
Oligocene time to the present, the convergence rate of the shrink-
ing Farallon plate was small and the dominant plate interaction
was that of strike-slip motion along the growing transform
boundary with the Pacific plate.
The motion of the Farallon plate relative to North America
is displayed in another way in Figure 9, which shows the normal
and tangential components of convergence at a fixed point near
present-day San Francisco. In resolving the convergence vector
into normal and tangential components, it was assumed that the
trend of the continental margin was N40WSimilar diagrams
showing the normal and parallel components of motion for all of
the oceanic and adjacent continental plates of the North Pacific
are in Engebretson (1982).
These normal and tangential velocities are fairly well con-
strained during times when the geomagnetic field was reversing
(163-1 18 Ma and 83-0 Ma) but not during the Cretaceous nor-
mal superchron (1 18-83 Ma) nor during the normal superchron
prior to 163 Ma. As noted earlier, the curves for ages > 163 Ma
are based on extrapolation of the post-163 Ma Farallon-Pacific
spreading rate. The curves shown for the interval (1 18-83 Ma)
correspond to the preferred model of Engebretson and others
(1984a) for Farallon-Pacific spreading during the Cretaceous
normal superchron (entries labeled "Farallon N.O.M." in the
tables), which is based on the ocean floor record on the north side
of the Mendocino Fracture Zone. If, on the other hand, the ridge
jump occurred north of the Mendocino Fracture Zone during this
time, then the true Farallon-Pacific spreading rate is substantially
greater, (entries labeled "Farallon S.O.M." in the tables). The
result of using the spreading rates determined south of the Men-
docino is an increase in the North America-Farallon convergence
rate of about 40 km/m.y. during the time of the normal super-
chron, approximately 118-83 Ma.
The dashed line in Figure 9a shows the convergence velocity
of the Kula plate if the southern option is taken for the locus of
the initial Kula-Farallon break. This places the Kula plate oppo-
site San Francisco beginning at 85 Ma, the time of the break. A
somewhat surprising result of the present analysis is that the
subduction velocity of the Kula plate, that is, the convergence
 Engebretson, Cox, and Gordon

Figure 8 (this and facing page). Convergence velocity vectors of oceanic plates relative to adjacent
continental plates. Each arrow at a point along a continental margin shows the convergence velocity of
the adjacent oceanic plate at the time indicated, which is between the two times defining the stage used
to estimate the velocity vector. Two arrows at certain points show convergence velocities of two oceanic
plates when either may have been adjacent to the point. P: Pacific plate. F: Farallon plate. I: Izanagi
plate. K:Kula plate. The Izanagi vectors (I) at 175 Ma are based on highly speculative extrapolations of
the oldest stage poles for Izanagi-Pacific motion from 145 to 135Ma. (a) Convergence velocities along
North American margin. (b) Convergence velocities along Eurasian margin.
 Engebretson, Cox, and Gordon

TABLE 5. Linear velocities of oceanic plates relative to North America

plate at selected sites; azimuth in degrees, speed in km/m.y.

Site Site Site

Central Aleutian Arc Eastern Aleutian Arc KO< A Island
Lat. 52 * N 1 Long. 187'E Lat. 56 * N 1 Long. 202 ' E Lat. 58 * N Long. 208O E
Pacific Plate Pacific Plate Pa 1c Plate
Age(Ma) 1
Vector Age(Ma) 1 Vector Age(Ma) Vector
from 1 to 1
azimuth speed 1 'rom 1 to 1 azimuth 1 speed from to -ZZq&z
0 1 5 1 318 1 87, 0 1 5 I 329 1 77 0 5
5 9
9 17
17 28
28 37
37 43
*43 48
*48 1 66 1 29 1
100 *48 66 35 I 99
*66 I 74 1 12 1 90 *66 174 1 24 1 88 Kula Plate I< a. Plate
Kula Plate Kula Plate Vector Age(Ma) Vector
Age(Ma) 1 Vector Age(Ma Vector from to
from 1to 1azimuth speed1 43 48
48 56
48 56 347 203 56 61
66 61 66
66 74
160 74 85 347 1 167
74 1 85 1 319 1 179 Farallon Far allon late N.o.M.^
Izpnagi Plate Izanagi Plate Vector Age(Ma) Vector
Age Ma Vector from to
*74 85
85 100 12 Ill
100 115 36 32
115 115 110 21 59
119 119 127 57 108
127 135 50 102

*
*I35 1 145 I 281 2 I 135 145 105 39
Farallon Plate N.o.M.~') Farallon Plate N.o.M.~') 145 161 67 92

-
Speed
Vector 161
163 1180 178 1
120
161
163
163
180
80 115

Farallon Plate s.o.M.~~ Farallor

114 Age(Ma) Vector
Age Ma Vector
96
18 from 1 to azimuth speed
41
Ill 109 100 115
107 104 115 110
39
85
Ill
163 1 180 1 61 117 163 180 1 70 I 122 1
Farallon Plate S.O.1 fF= Farallon Plate s.o.M.^
Vector
azimuth speed
109
25
58
 Relative Motions Between Oceanic and Continental Plates

TABLE 5. (Continued)

Site Site ite Site

Southern Chugach Mtns. Queen Charlotte Islands Juan d Puca Strait Klamatli Mountains
Lat. 60 'N 1 Long. 220-E Lat. 54 'N 1 Long. 228 * E Lat. 48 * N Long. 236' E ~ t42O. N 1 Long. 236' E
Pacific Plate Pacific Plate Pal c Plate Kula Plate
Age(Ma) 1 Vector Vector Vector Age(Ma) Vector
rom 1 to 1 azimuth 1 soeed

*37
'32836
izimuth speed

349I 37
rom
*43
*48
to
48
56
I<) Plate *56 61
Age(Ma) Vector *61 66 32 131
*66 74 24 138
rom to 358 1 140
*74 85
*43 48
Farallon ate N.O.M.~'J
*48 1 66 1 46 1 98 *48 1 66 1 53 1 94 *48 56
Age(Mal Vector
Kula Plate Kula Plate *56 61

qi!i ;:;1
rom -
- azimuth - speed
1 Age(Ma) 1 Vector
*61 66
Age(?) Vector
'rom 1 to, 1 azimuth 1 speed
-66
*74
74
85
PI:1 3 1 140
Farallon ate N.O.M.~~)
0
5
9
50
51
31
33
47
45
T Vect - 17 27 50
141 28 42 65
145 azimuth -speed
37 47 91
74 155 66 22 149 60 43 *43 55 135
74 85 358 157 74 150 46 43 *48 49 144
Farallon F iwallon Plate N.o.M.~" 32 46 152
- - Plate N.o.M.^ - *56 53
Age('M Vect a Vect - 29
36
51
59
*61
*66
51
54
145
126
rom
- azimuth -
rorn azimuth -speed 41 81 *74 25 106
"74 17 *74 23 113 52 125 *85 34 Ill
85 22 85 29 113 47 138 100 55 52
100 37 100 46 42 53 152 115 45 73
115 28 115 37 67 51 146 119 84 91
119 69 119 77 103 55 127 127 77 82
127 63 127 70 95 28 108 136 126 39
135 114 135 120 39 35 112 145 84 104
145 74 145 79 98 52 48 161 95 127
161 87 161 91 119 43 71
-
163 99 128
16] Farallon
180 90
late S.0.1
123
ate S.O.M. %==
-
85
78
126
97
89
39
-J

Farallon -F==
ate S.0.b
Vector
Vector Vector
azimuth azimuth 1 meed
84
96
102
123
azimuth speed

99 126
ate S.0.h 7 35 112
I
- Age1 Vector
-
'rom
85
100
-
115 33 110
 Engebretson, Cox, and Gordon

TABLE 5. (Continued)
Site Site Site
Los Angeles Basin Baja California Southwestern Mexico
Lat. 34 'N 1 Long. 242 'E Lat.26-N 1 Long.248-E Lat. 18 * N 1Long. 256.E
Pacific Plate c Plate Farallon Plate N.o.M!"
Age(Ma) 1 Vector Vector 4 4 a) 1 Vect -
1
- to azimuth speed
'rom
48 161
66 127
296 60 61 119
*17 I28 I 306 1 51 298 1 56 66 99
Kula Plate , Plate 74 72
Vector 85 87
Age Ma Vector 100 64
zq-&Z 115 73
119 62
127 49
135 38
348 1 122 145 102
*74 354 1 131 ate N.o.M.~') 161 127
Farallon ate N.o.M.^ [a)_ Vector- 163 1 180 I 111 126
&-
M -
Vect -to izimuth - w==
Farallon Plate S.O.M! -
Age Ma Vector
--
rom azimuth 37
43
52
56
*17 28 47
28 48 76 48 62
37 53 Ill 56 56 100 116 105
*43 59 151 61 60 115 119 46 116
*48 53 154 66 58
*56 57 148 74 61
*61 55 141 85 25
*66 58 121 100 38
*74 26 98 115 62
85 37 106 119 52
100 59 57 127 102
115 49 75 135 94
119 92 81 146 135
127 84 71 161 91
136 131 39 163 101
145 87 105 -180 106
161
163 180
98
103
129
129
-
allon late S.0.h
Vector
2
Id)
Farallon late S.0.t 7
- to
Vector 100
ZKZp-z 115
-119 118

*Presence of this plate at this place and time is questionable.

(1) Farallon plate velocity determined from isochrons north of the
Mendocino Fracture Zone.
(2) Farallon plate velocity determined from isochrons south of the
Mendocino Fracture Zone.
 Relative Motions Between Oceanic and Continental Plates

TABLE 6. Linear velocities of oceanic plates relative to fixed Eurasia

plate at selected sites; azimuth in degrees, speed in km/m.y.
Site Site Site Site
Eastern Koryakskiy Mtns. Shelikhova Sea Sakhalin Island Honshu Island
Lat. 62'N 1 Long. 175* E Lat. 60'N I Long. 155'E Lat. 53 'N 1 Long. 142' E Lat. 40 * N 1 Long. 140.E
Kula Plate Pacific Plate Pacific Plate Pacific Plate
Age(Ma) Vector Age(Ma) Vector Age(Ma) 1 Vector Age(Ma) 1 Vector
'rom to azimuth speed rom to azimuth speed -
rom rom 1 to 1 azimuth speed
43 48 325 164 0 5 299 94 0 0 1 5 1 288 114
48 53 334 158 5 9 290 72 5 94
53 56 319 . 220 9 17 289 72 9 94
56 61 331 135 94

-
17 28 292 73 17
61 66 32 157 28 37 292 65 28 84
66 74 316 177 37 43 319 42 37 55
74 85 301 203 $43 48 348 75 $43 73
Izaiiagi Plate $48 53 8 90 $48 89
Vector *53 66 325 112 $53 108
$66 1 74 1 328 I 102 $66 104
Kula Plate $74 I85 I 294 1125 $74 1 85 1 292 1 131
Kula Plate Kula Plate
'100 Vector Vector
"15
-
speed
153
"119 127 146
'127 135 213
'135 4 1 18 136 287
FE ate N.O.M.~~] 156 303 123
296 137
Vector -
azimuth
310
*
116 Izanagi Plate
177
208 159 .

Izai ci Plate Izanagi Plate

317 94 Vector
14 10 Vector Age(Ma) 1 Vector
328 37 azimuth 1 speed from 1 to 1 azimuth 1 speed
4 Ill 338 1 234
357 107 200
55 38 200 * 100
28 75 197
41 100 284 $119
41 108 *I35 I 145 1 344 1 29 $127
r
Farallon late S.O.1 -
Vector
Farallon Plate N.O.M>^
Vector -*
a
Far;
la)
327 1 (lJ39
ate N.O.M.
Vector
$135 1 145 1 314 1
Farallon Plate N.O.M..
d-a Vector -
I
--
from t o azimuth 1
speed from to azimuth speed
$85 95 276 1 104 $85 95 275 89
293 $95 100 $95 100 279 54
306 * 100 115 * 100 115 107 27
$115 119 $115 119 181 2
$119 127 $119 127 342 97
*Presence of this plate at this place * 127 135 * 127 135 333 93
and time is questionable.
(1) Farallon plate velocity deter- $135 145 * 135 145 31 38
$145 161 $145 161 35 64
mined from isochrons north of the
Mendocino Fracture Zone. $163 I 180 I 28 1 99 $161 163 $161 163 40 95
(2) Farallon plate velocity deter- Farallon Plate s.o.M.~) -180 - 26 1 180 35
rallon Plate s.o.M.~"
102
mined from isochrons south of the Age(Ma) 1 Vector Farallon late S.O.M. ,zJ 94
Mendocino Fracture Zone. Vector Vector
azimuth speed
267 112
264 84
182 29
42 Engebretson, Cox, and Gordon
175 150 125 100 75 50 25
Time (Ma)
" O
Time (Ma)
Figure 9. (a) Solid line: component normal to a trench axis trending
N40 of the convergence velocity of the Farallon plate relative to
North America, shown for a point near San Francisco. Dashed line:
normal velocity of Kula plate during time when it was opposite San
Francisco, assuming the southern option for the locus of the initial Kula-
Farallon ridge. Thick line: age of the Farallon plate that was entering the
trench at the latitude of San Francisco, ages greater than 163 m.y. being
an extrapolation to earlier times of post-163 Ma Pacific-Farallon spread-
ing. Dots: motion of North America relative to hotspots. (b) Component
tangential to the trench axis of convergence velocities of Farallon and
Kula plates relative to North America for a point near San Francisco.
Some of the more rapid peaks and troughs may be artifacts resulting
from combining stage poles for slightly different time intervals and there-
fore may not be significant (see text).
velocity normal to the trench, is nearly as great as that of the
Farallon plate. Of course this result is sensitive to the trend of the
trench, which is not well determined. An interpretation of these
results is given in the next section.
Trajectoriesof Plates
From their stratigraphic records it is clear that many of the
allochthonous terranes found near the rim of the Pacific basin
show that at least parts of their geologic histories were spent in
open-ocean environments. Moreover, paleomagnetic and fossil
evidence from these terranes leave little doubt that they have
traveled great distances since their origin. The displacement
model presented here can be used to determine whether known
oceanic plates could haveacted as the transport system that
brought these far-traveled terranes to their present locations.
Because terranes begin and end their journeys at different places
and different times and because they generally travel on different
sets of plates, a custom-made travel history is needed for each
terrane.
A few general observations are useful. The Kula and Izanagi
plates moved in a generally northerly direction through the Pa-
cific basin. They comprise a plate transport system capable of
moving terranes rapidly toward the paleopoles. Along the margin
0
of North America, intervals characterized by extensive dextral
faulting in the Late Cretaceous and early Tertiary may be asso-
a
ciated with the presence of the Kula plate along the margin. In
contrast, the Farallon plate swept across the Pacific basin in an
easterly direction and was usually (but not always) characterized
by rapid subduction beneath North America. The Farallon plate
comprised a plate transport system capable of moving terranes
great distances from west to east across the Pacific basin, as
required for terranes bearing Tethyan fauna.
Plate motion can be characterized by a trajectory showing
the path of a point that is riding with the plate. One such trajec-
tory (Figure lOa) shows the motion of a point near present-day
San Francisco that has moved with the North America plate from
180 Ma to the present as seen from a fixed hotspot reference
frame. Seen at 10 m.y. increments the coastline of western North
b America conveys an impression of the direction of motion of
North America and its speed over the hotspots, with wider spac-
ing of the coastline curves corresponding to higher speeds. The
high velocity of the continent over the hotspot at about 60 Ma is
apparent in both Figures 9.and 10. The significance of this inter-
val of rapid motion is discussed in the last section.
The trajectory of a terrane that formed near the Pacific-Far-
allon ridge at 160 Ma and subsequently rode with the Farallon
plate is shown in Figure lob. A general analysis of terrane trajec-
tories based on the present plate model is given by Debiche and
others (in preparation).
GEOLOGIC INTERPRETATION
In the present section we will review some of the more
obvious correlations between the present model for Pacific plate
tectonics and major geologic events in western North America,
including the Laramide orogeny. Some of these correlations are
discussed in more detail by Engebretson and others (1984b), Page
and Engebretson (1984), Wells and others (1984), and Wallace
and Engebretson (1984).
Pacific-North American Convergence Beginning at 5 Ma
Our model incorporates a change in the motion of the Pa-
cific plate relative to hotspots at 5 Ma from a west-northwesterly
direction to a more northwesterly direction (Cox and Engebret-
son 1985). As a result of this more northerly motion, our model
produces a component of Pacific-North America motion per-
pendicular to the transform boundary between the two plates.
The resulting component of normal motion between the Pacific
Relative Motions Between Oceanic and Continental Plates 43

Figure 10. (a) Motion of North America with respect to the hot spots for
the past 180 Ma at 10 may.intervals. The dark line is the trajectory of a
point (S)near present-day San Francisco. Small dots show positions of S
at 5 m.y. intervals and large dots show positions at 30 m.y. intervals.
(b) Sequence of dots shows the trajectory of a point on the Parallon plate
as the, Farallon plate moved across the Pacific basin. This point is located
on anomaly M29 (164 Ma) where"e anomaly is offset by a small
fracture zone. Coordinates remain fixed relative to North America.
44 Engebretson, Cox, and Gordon
and North America plates during the Pliocene is consistent with
the following geologic observations in California, which seem to
require compression transverse to the Pacific-North America
plate boundary at about this time (B. M. Page, personal commun-
ication 1982); (1) withdrawal of the sea from the San Joaquin
Valley and most of California around 3 Ma; (2) uplift of the
Diablo, Santa Cruz, and Santa Lucia Ranges, all parallel to the
plate boundary, within the past 2 to 3 m.y.; (3) development of
many Pliocene or Quaternary folds in the Sur-Obispo belt paral-
lel to the plate boundary; (4) rise of the nonbuoyant Mount
Diablo piercement in the past 2 to 3 m.y.; (5) rise of the Kettle-
man Hills anticline with its axis parallel to the Pacific-North
America plate boundary beginning around 1 Ma; (6) formation
of thrust faults, some of which are parallel to the plate boundary,
along both sides of the Santa Clara Valley within the past 1 m.y.;
(7) thrusting of strata of the lower continental slope over Quater-
nary sediments in the relict trench west of San Luis Obispo and
Point Sur.
Convergence and Magmatism
Cretaceous plutons of the Sierra Nevada were emplaced
during the interval 120 to 80 Ma (Chen and Moore 1982). The
important question of how this interval correlates with the rate of
Farallon-North America convergence depends critically upon
the model used for Farallon-Pacific spreading during the Cre-
taceous normal superchron (1 18-83 Ma). Figure 9a shows the
convergence rate corresponding to the preferred model of Enge-
bretson and others (1984a). As discussed earlier, this is based on
the assumption that no ridge jumping occurred north of the Men-
docino Fracture Zone during the normal superchron and that the
spreading rate changed smoothly. During the interval of intense
plutonism from 120 to 80 Ma, the rate of convergence, while
moderately high (50 to 100 km/m.y.), was lower than during the
subsequent pulse of very rapid Farallon-North America conver-
gence between 75 and 40 Ma (Figure 9a), which reached a peak
convergence rate of 150 km/m.y. and was not accompanied by
extensive plutonism.
The alternative model for Farallon-Pacific spreading dis-
cussed earlier is based on the assumption that no ridge jumping
occurred south of the Mendocino Fracture Zone during the Cre-
taceous normal superchron. The latter requires a westward jump
in the Pacific-Farallon spreading center to the north of the Men-
docino Fracture Zone. The net result is to increase the Farallon-
North America convergence rate by about 40 km/m.y. during
the normal superchron (1 18-83 Ma). As a result, the average
convergence rate during the episode of Sierra Nevada plutonism
is - 100 km/m.y., with more rapid convergence from 100-85 Ma
than from 120-100 Ma. In this model the cessation of plutonism
occurs at about the same time as a drop in convergence rate to
below 100 km/m.y., followed as before by a pulse of rapid
convergence unaccompanied by extensive plutonism. A generali-
zation consistent with both models is that plutonism is associated
with a moderate (>SO km/m.y.) to high (> 100 km/m.y.) rate of
convergence, but that a high rate of convergence does not neces-
sarily produce plutonism.
Both of the above models assume the northern option for the
locus of the initial break between the Kula and Farallon plates. In
the southern option for the initial boundary, at the time of the
break (-85 Ma) the Farallon plate became the Kula plate over
the latitude range of the Sierra Nevada. The approximate correla-
tion in time with the cessation of plutonism suggests the possibil-
ity that the latter was due to a marked decrease in convergence
rate at the time of change from easterly Farallon-North America
convergence to more northerly Kula-North America conver-
gence. Surprisingly, however, the convergence rate of the Kula
plate (dashed line in Figure 9a) was only slightly less than that of
the Farallon plate. Thus, whichever plate was opposite the Sierra
Nevada, the interval of time after the cessation of Sierra plu-
tonism was characterized by a high rate of convergence (>100
km/m. y.).
The weakest link in the plate circuit that provides the basis
for the above analysis is motion of the Pacific plate relative to
hotspots. The rapid increase in convergence rate at 74 Ma is
primarily the result of a change in motion of the Pacific plate
relative to hotspots (Appendix A). The increase is not strongly
dependent upon the model used for Farallon-Pacific spreading
nor upon the location of the initial Kula-Farallon break. Because
Pacific-hotspot motion is well-documented after 74 Ma but less
welldocumented before that time, it seems likely that con-
vergence was rapid during the time interval after the episode of
intense plutonism (120 to 80 Ma) but it is less certain what the
rate was during the episode of intense plutonism.
Laramide Orogeny
Many causal links have been proposed between plate tecton-
ics and the Laramide orogeny, including a change in the absolute
motion of North America, a decrease in the age of the subducting
plate, a change in the rate of convergence, a shallowing of the
angle of subduction, and the arrival of allochthonous terranes. In
examining these possible mechanisms, we find that the question
of the timing of plate tectonic processes offshore and continental
tectonic processes onshore is crucial.
Between 75 and 40 Ma magmatism ceased in the arc along
the western edge of the Cordillera but continued sporadically
inland at distances up to 1500 km from the arc (Lipman and
others 1972; Cross and Pilger 1978; Keith 1978; Dickinson and
Snyder 1978; Lipman 1980, 1982). This magmatic lull was not
synchronous throughout the western arc. It began first near the
latitude of Montana and Idaho (-45O N) and between 75 Ma
and 40 Ma, migrated south to the latitude of southern Arizona
and New Mexico (-30 N). Magmatism resumed in the western
arc in the north by -45 Ma and in the south by -25 Ma (Lipman
1980,1982).
The magmatic lull and increase in the arc-trench distance
during the Laramide orogeny have been attributed to anoma-
lously shallow-dipping subduction of the Farallon plate beneath
 Relative Motions Between Oceanic and Continental Plates
North America (Coney 1972; Burchfiel and Davis 1975; Coney
and Reynolds 1977; Cross and Pilger 1978; Dickinson and
Snyder 1978). This interpretation is supported by the paleodepths
of the Benioff zone inferred from the geochemistry of Cretaceous
and Tertiary andesitic rocks (Lipman and others 1972; Keith
1978, 1982).
If the northern option for the locus of initial rifting between
the Kula and Farallon plates is correct, the subduction of young
lithosphere began near Vancouver Island at about 85 Ma and
then migrated to the south (Figure 3). This correlates fairly well
with the timing and the southward migration of the locus of the
magmatic lull described above. However this model also predicts
the northward migration of an analogous magmatic lull in Can-
ada, and neither the pattern of magmatism observed in the United
States nor a lull in that pattern is observed in Canada (Lipman
and others 1972).
An alternative mechanism to account for shallow-dipping
subduction associated with magmatism well inland from the arc
is the subduction beneath North America of an aseismic ridge
riding on the Farallon plate (Livaccari and others 1981; Hender-
son and others 1984). According to Henderson and others
(1984), the aseismic ridge first collided with North America near
the present-day location of Cape Mendocino. The ridge migrated
southward rapidly, accounting for the southward migration of
tectonic and magmatic events associated with the Laramide orog-
eny. This model is consistent with both the northern or southern
option for the locus of initial Kula-Farallon rifting.
Our analysis (as summarized in Figure 9) supports Coney's
(1978) suggestion that the rate of convergence was high during
the Laramide orogeny. An important new result is the steep rise
in convergence rate beginning at -74 Ma, very close to the onset
of the Laramide orogeny, followed by a rise during the last pan of
the Laramide orogeny to the remarkably high rate of 150
km/m.y. The rapid increase in convergence rate at 74 Ma was
primarily the result of a change in Pacific-hotspot motion at this
time, as noted earlier. This interpretation depends strongly on our
model for Pacific-hotspot motion but not on the particular model
used for Farallon-Pacific spreading or the locus of the initial
Kula-Farallon break.
Two other relationships shown in Figure 9a may be related
45
to the Laramide orogeny. The first is the motion of the North
American plate relative to hotspots. The component of the mo-
tion normal to and toward the trench along the western margin of
North America (dotted curve in Figure 9a) was more rapid than
at any time before or after the orogeny, in accord with the hy-
pothesis that the Laramide orogeny was related to rapid motion
of the North American plate toward the trench (Coney 1978;
Cross and Pilger 1978; Engebretson and others 1984b). The sec-
ond is the age of the Farallon plate arriving at the trench, which
decreased rapidly during the Laramide orogeny. In a general way,
this result supports the hypothesis that the Laramide orogeny was
related to subduction of young lithosphere (Molnar and Atwater
1978; Cross and Pilger 1982; Engebretson and others 1948b).
However, our model, using the northern option, indicates
that at the beginning of the Laramide orogeny, the age of the
lithosphere arriving at the trench was 100 m.y. or more, an age
that Molnar and Atwater (1978) have associated with extensional
tectonics, suggesting that the onset of the Laramide did not corre-
late with the first subduction of young lithosphere. It was not
until half way through the orogeny that lithosphere as young as
50 m.y., the age associated by Molnar and Atwater (1978) with
compressional tectonics, arrived at the subduction zone. The
onset of the Laramide event appears to be more closely associated
with the rapid younging of the lithosphere arriving at the subduc-
tion zone than with the first arrival of young ( 4 0 m.y.) litho-
sphere when the northern option model is used.
The hypothesis that the Laramide orogeny was produced by
the arrival at the subduction zone of thickened buoyant litho-
sphere carrying aseismic ridges or oceanic plateaus (Livaccari and
others 1981; Henderson and others 1984) is consistent with ter-
rane trajectories based on the present model.
If our analysis had shown that only one of the plate tectonic
processes proposed as a cause for the Laramide orogeny had been
operating during the time 80 to 40 Ma, the single cause of the
Laramide orogeny would be obvious. Instead, our analysis shows
that almost all of the proposed processes may have been occur-
ring simultaneously. Our conclusion is that more than one and
perhaps all of the above plate tectonic processes, acting simul-
taneously and in concert, produced the Laramide orogeny (En-
gebretson and others 1984b).
46 Engebretson, Cox, and Gordon

APPENDIX A: INFORMATION FOR TRACING Genealogy of the Linear Velocities

THE CAUSES O F CHANGES
IN RELATIVE PLATE MOTIONS The origin of calculated linear velocities can be traced to
observed plate-plate or plate-hotspot motions in Figures A1 and
Relative linear velocities between the selected localities in A2. Linear velocities are shown at selected sites for plate pairs of
North America and Eurasia and adjacent oceanic plates are listed interest. Each graph displays the direction and speed of relative
in Tables 5 and 6 and are shown in Appendix B. Some of the motion as a function of time for the labeled plate pair. Motions
apparent changes in velocity in these figures are real and some are based on direct observation are marked with stars. The other
probably artifacts resulting from errors of several types. To aid motions can be traced to these. The discussions that follow, which
the geologist who needs to assess the reliability of individual are specific to sites where the linear velocities were calculated,
features in linear velocity diagrams before attempting to correlate trace the origin of several types of change in plate interactions.
these with geologic events, the genealogy of the linear velocity The velocities cited refer to those shown in Figure Al. The fol-
plots is displayed in figures A-1 and A-2. Using these figures, the lowing abbreviations are used: NA-North America plate, SA-
geologist can trace changes in linear convergence velocities back South American plate, AF-African plate, FA-Farallon plate,
to their sources at different links in the plate-hotspot circuit. KU-Kula plate, HS-hotspot reference frame.
The basic element used in determining the history of plate
displacements and velocities is an Euler stage pole; a stage being a HS-NA Motion
time interval during which it is assumed that plate-plate or plate-
Feature: Between 161 and 145 Ma, a stepwise decrease in
hotspot motion was constant. Short stages in the poles used to velocity of NA-HS motion from over 100 km/m.y.
describe the convergence of a certain plate pair, for example, the
to less than 25 km/m.y.
Farallon-North America plates, are due to either the existence of
Genealogy: (1) Decreases in AF-NA spreading rates at 161 and
a short stage used to describe one plate pair in the circuit, or to the
145 Ma.
fact that two of the boundaries at different links in the circuit
(2) Moderate decrease in speed and abrupt direc-
happened to have times that are closely spaced. The latter cir-
tional change of AF-HS motion at 145 Ma.
cumstance would not arise if the same set of isochrons is used to
Feature: At 119 Ma an increase in velocity from less than 25
describe all plate motions in all basins. In practice, this is not the
km/m.y. to greater than 50 km/m.y.
case, either because the motion of different plate pairs changed at
Genealogy: Slow spreading rate in AF-NA motion between 145
different times or because, in different ocean basins, different
and 119 Ma followed by moderate increase at 119
isochrons could be identified more easily and these are used to
Ma.
determine Euler poles. Similarly the times used to define stage
Feature: Between 48 and 37 Ma, a stepwise decrease and
boundaries for plate-hotspot motion depended in part on when
counterclockwiseshift of NA-HS velocity, the speed
the direction of plate motion actually changed and in part on the
decreased from approximately 50 km/m.y. to less
quality of bathymetric data and radiometric dates along the hot-
than 10 km/m.y. with a shift from a westerly to a
spot track. When all of the plate pairs are combined, as in the
southwesterly direction.
North America-Africa-hotspot-Pacific-Farallon circuit, the
Genealogy: (1) Moderate decrease in AF-NA spreading rate at
number of stage boundaries in the resultant North America-
37 Ma in conjunction with a small directional
Farallon motion is the number of noncoincident stage boundaries
change.
for all of the links of the circuit.
(2) Moderate directional change in AF-HS motion
Because the Euler poles for short stages are found by sub-
at 48 Ma.
tracting two poles, each with its errors, the errors for short stages
are relatively large. Errors are probably increased by uncertainties FA-NA Motion
in dating. As a result, where stage boundaries are closely spaced,
the calculated stages tend to have erratic Euler poles and erratic Feature: Between 161 and 145 Ma a stepwise decrease in
linear velocities. Geologists using these tables should be cautious FA-NA relative motion from approximately 125
in interpreting stages as short as 10 m.y. and even more skeptical km/m.y. to less than 50 km/m.y.
about shorter stages. Genealogy: (1) Stepwise decrease in HS-NA motion at this time
Because the displacement histories presented in this paper (see earlier discussion).
were constructed using available stage poles from many sources, (2) Decrease and directional change in FA-HS mo-
it would be unwieldy to evaluate critically the significance of tion at 145 Ma resulting from the initiation of
each stage pole used as input for the model. The reader is referred PA-HS motion at that time (see text).
to the original sources for additional information (see Table 1). Feature: Between 145 and 100 Ma a period of moderately
Relative Motions Between Oceanic and Continental Plates
Engebretson, Cox, and Gordon

1-1 HLUON d0 lSU3 '030

I- HLUON 40 lSU3 '030

 Relative Motions Between Oceanic and Continental Plates
slow (approximately 60 km/m.y.) FA-NA conver-
gence. Variations within this interval lie at the limit
of resolution of the method.
Genealogy: (1) Slow HS-NA motion beginning at 145 Ma (see
earlier discussion).
(2) Initiation of southerly PA-HS motion at 145
Ma.
(3) Directional changes in PA-HS motion at 135
and 115 Ma.
(4) Moderate directional change in FA-PA spread-
ing at 135 Ma.
(5) Small directional change and moderate increase
in FA-PA motion at 119 Ma.
Feature: At 100 Ma a marked increase in FA-NA conver-
gence from approximately 60 km/m.y. to greater
than 100 km/m.y.
Genealogy: Abrupt increase (nearly 100 km/m.y.) in FA-HS
motion resulting entirely from a change in direction
and velocity of PA-HS motion at 100 Ma.
Feature: At 74 and 66 Ma stepwise increases in FA-NA con-
vergence rates from approximately 100 km/m.y. to
approximately 150 km/m.y. accompanied by a
clockwise shift in direction.
Genealogy: (1) At 74 Ma a small directional change in FA-HS
motion as a result of a clockwise directional change
and increase in velocity of PA-HS motion.
(2) At 66 Ma an increase in HS-NA velocity result-
ing from an increase in AF-NA spreading rates.
Feature: Between 43 and 28 Ma stepwise decreases in
FA-NA convergence rates.
Genealogy: (1) Decrease in FA-HS velocity at 43 Ma that re-
sulted from the combination of an increase in
FA-PA spreading rates and a counterclockwise shift
in PA-HS motion marked by the bend in the Ha-
waiian-Emperor seamount chain.
(2) Decrease at 37 Ma in the FA-PA spreading rate.
(3) Decrease at 28 Ma in the FA-PA spreading rate.
(4) Decrease in HS-NA velocity at 37 Ma (see ear-
lier discussion).
(IZ, KU,PA)-NA Motion
In our reconstructions the Izanagi plate was never in contact
49
with North America, and the Kula plate broke away from the
Farallon plate at 85 Ma. Therefore, after 85 Ma KU-NA and
PA-NA motion are of primary interest.
Feature: Between 74 and 56 Ma stepwise decrease in
KU-NA convergence accompanied by clockwise
shift in direction.
Genealogy: (1) Stepwise (probably continuous) decrease in
KU-PA spreading rates at 74,66, and 61 Ma.
(2) Change in PA-HS motion at 74 Ma as discussed
earlier.
Feature: Between 56 and 43 Ma an abrupt (approximately
100 km/m.y.) increase in KU-NA convergence.
Genealogy: Marked increase in KU-PA spreading rate between
56 and 43 Ma, ending with the death of KU-PA
ridge.
NA-EU Motion
Feature: At 95 Ma abrupt changes in NA-EU (Figure A2).
Genealogy: Initiation of spreading ridge in the Arctic, accom-
panied by a change in the direction of spreading.
HS-EU Motion
Feature: Between 53 and 48 Ma a short-lived increase in
HS-EU velocity.
Genealogy: This feature, which is probably spurious, is the result
of combining a reported change in NA-EU motion
at 53 and 48 Ma with a reported change in HS-NA
motion at 48 Ma.
SA-NA Motion
Feature: Between 85 and 65 Ma a markedly different direc-
tion of SA-NA motion.
Genealogy: (1) The anomalous direction of motion during that
period is probably not significant because the veloc-
ity is close to zero. The origin of the low velocity
during this period is a simultaneous change in
NA-AF and SA-AF motion at 85 Ma.
(2) The 66 Ma increase in SA-NA velocity resulted
from a change in AF-NA divergence at this time.
50 Engebretson, Cox, and Gordon

APPENDIX B: AZIMUTH AND SPEED OF THE B2. These figures were developed to display when major changes
LINEAR VELOCITIES OF OCEANIC PLATES in plate interactions occurred. Because of uncertainty in the loca-
FOR NORTH AMERICA AND EURASIA tions of plate boundaries at certain times, velocity vectors for
more than one plate pair appear on some of the plots. Earlier
Tables 5 and 6 give the azimuth and speed of linear veloci- convergence velocities are known with much less certainty than
ties of the oceanic plates relative to North America and Eurasia are later convergence velocities.
respectively. This information is also plotted in Figures Bl and

Weotorn Aleutian Are Central Aleutian Are

at. 8 4 . 0 Long. 1 7 2 . 0 Lat. 83.0 Long. 1 0 7 . 0

Figure Bl (this and following pages). Speed and azimuth of linear velocities relative to North America at
selected sites. Where it is uncertain which of two plates was adjacent to a site at a given time, velocity
arrows for both plates are shown.
 Relative Motions Between Oceanic and Continental Plates
B a a t e r n A l e u t i a n Are Kodlak l a l a n d
Lat. 86.0 Long. 1 0 2 . 0 Lat. 5 8 . 0 Long. 2 0 0 . 0
 Engebretson, Cox, and Gordon

Southarn Chugaah Mta. Quean C h a t i o t t e l a l a n d 8

L i t . 60.0 Long. 2 2 0 . 0 f t. 6 4 . 0 Long. 2 2 6 . 0
 Relative Motions Between Oceanic and Continental Plates
Juan de F u a a S t r a i t K l a m a t h Mto.
Lat. 46.0 Lonq. 336.0 Lat. 4 2 . 0 Long. 2 3 6 . 0
I
 Engebretson, Cox, and Gordon
Low Angelem B8wIn
Let. 94.0 Long. 242.0
 Relative Motions Between Oceanic and Continental Plates

outhwee Mexleo
f t. 16.0 Long. 2 0 0 . 0
 Engebretson, Cox, and Gordon
t. K o r y a k k l Mt
Lat. 0 2 . 0 Longd 1 7 6 . 0 8hellkhova 8 e a
r3 Let. 0 0 . 0 Long. 1 6 6 . 0
1 I

Figure B2. Speed and azimuth of linear velocities relative to Eurasia at selected sites. Where it is
uncertain which of two plates was adjacent to a site at a given time, velocity arrows for both plates are
shown.
 Relative Motions Between Oceanic and Continental Plates 57

8kRoll l o l n Honohu l o l a n d
Lot. 6 3 . 0 Long. 1 4 2 . 0 Lt 40.0 Long.140.0
I
 Engebretson, Cox, and Gordon
REFERENCES CITED
Alvarez, W., Kent, D. V., Premoli Silva, I., Schweikert, R. A., and Larson, R. L.,
1980, Franciscan complex limestones deposited at 17O South palwlatitude:
Geological Society of America Bulletin, v. 91, p. 476-484.
Andrews, J. A., True polar wander: Journal of Gwphysical Research, in press.
Atwater, T., 1970, Implicationsof plate tectonics for the Cenozoic tectonic evolu-
tion of western North America, Geological Society of America Bulletin,
v. 81, p. 3513-3536.
Atwater, T., and Molnar, P., 1973, Relative motion of the Pacific and North
American plates deduced from seafloor spreading in the Atlantic, Indian,
and South Pacific Oceans, in Kovach, R. L., and Nur, A., eds., Proceedings
of the Conference on Tectonic Problems of the San Andreas Fault System,
13: Stanford, Stanford Univ. Publications, p. 136-148.
Burchfiel, B. C., and Davis, G. A., 1975, Nature and controls of Cordilleran
orogenesis, western United States: Extensions of an earlier synthesis:Ameri-
can Journal of Science, v. 275-A, p. 363-397.
Burke, K., Kidd, W.S.F., and Wilson, J. T., 1973, Relative and latitudinal motion
of Atlantic hotspots: Nature, v. 245, p. 133-137.
Cande, S. C. and Kristofferson, Y., 1977, Late Cretaceous magnetic anomalies in
the North Atlantic: Earth and Planetary Science Letters, v. 35, p. 215-224.
Carlson, R. L., 1982, Cenozoic convergence along the California coast: A qualita-
tive test of the hot-spot approximation: Geology, v. 10, p. 191-196.
Chase, C. G., 1978, Plate kinematics: the Americas, East Africa, and the rest of
the world: Earth and Planetary Science Letters, v. 37. v. 353-368.
Chen, J. H., and Moore, J. G., 1982, Uranium-lead isotopic ages from the Sierra
Nevada batholith, California: Journal of Geophysical Research, v. 87,
p. 4761-4784.
Clague, D. A,, and Jarrard, R. D., 1973, Tertiary Pacific plate motion deduced
from the Hawaiian-Emperor chain: Geological Society of America Bulletin,
v. 84, p. 1135-1154.
Coney, P. J., 1972, Cordilleran tectonics and North America plate motion: Amer-
ican Journal of Science, v. 272, p. 603-628.
-1978, Mesozoi~CenozoicCordilleran plate tectonics, Cenozoic tectonics
and regional geophysics of western Cordillera: Geological Society of Amer-
ica Memoir, v. 152, p. 33-50.
Coney, P. J., and Reynolds, S. J., 1977, Cordilleran Benioff zones: Nature, v. 270,
p. 403-406.
Cox, A., and Engebretson, D. C., 1985, Change in motion of Pacific plate at 5
m.y. B.P.: Nature, no. 313, p. 472-474.
Cross, T. A., and Pilger, R. H., Jr., 1978, Constraints on absolute motion and
plate interaction inferred from Cenozoic igneous activity in the western
United States: American Journal of Science, v. 278, p. 865-902.
Dalrymple, G. B., Lanphere, M. A., and Clague, D. A., 1980, Conventional and
MAr/39~rK-Ar ages of volcanic rocks from Ojin (Site 430), Nintoku (Site
432), and Suiko (Site 433) seamounts and the chronology of volcanic prop-
ogation along the Hawaiian-Emperor chain: Initial Reports of the Deep Sea
Drilling Project, no. 55, p. 659-676.
Debiche, M. G., Cox, A., and Engebretson, D. C., n.d., Terrane trajectory analy-
sis, in preparation.
Dickinson, W. R., and Snyder, W. S., 1978, Plate tectonics of the Laramide
orogeny: Gwlogical Society of America Memoir, no. 151, p. 355-366.
-1979, Geometry of triple junctions related to San Andreas transform: Jour-
nal of Geophysical Research, v. 84, p. 561-572.
Duncan, R. A., 1981, Hotspots in the southern oceans-an absolute frame of
reference for motion of the Gondwana continents: Tectonophysics, v. 74,
p. 29-42.
-1982, A captured island chain in the coast range of Oregon and Washington:
Journal of Geophysical Research, v. 87, p. 10,827-10,837.
Engebretson, D. C., 1982, Relative motion between oceanic and continental
plates in the Pacific basin [Ph.D. dissertation]: Stanford University, 21 1 pp.
Engebretson, D. C., Cox, A., and Gordon, R. G., 1984a, Relative motions be-
tween oceanic plates of the Pacific basin: Journal of Gwphysical Research,
V. 89, p. 10,291-10,310.
Engebretson, D. C., Cox, A., and Thompson, 0. A., 1984b, Correlation of plate
motions with continental tectonics: Laramide to Basin-Range: Tectonics,
V. 3, p. 115-1 19.
Epp, D., 1978, Age and tectonic relationships among volcanic chains on the
Pacific plate [Ph.D. thesis]: University of Hawaii, Honolulu.
Fox, K. F., Jr., 1981, Melanges and their bearing on late Mesozoic and Tertiary
subduction and interplate translation at the western edge of the North Amer-
ica plate: U.S. Gwlogical Survey Professional Paper, 1198.
Francheteau, J., Sclater, J. G., and Menard, H. W., 1970, Pattern of relative
motion from fracture zone and spreading rate data in the north-eastern
Pacific: Nature, v. 226, p. 746-748.
Gordon, R. G., 1983, Late Cretaceousapparent polar wander of the Pacific plate:
evidence for a rapid shift of the Pacific hotspots with respect to the spin axis:
Geophysical Research Letters, v. 10, p. 709-712.
Gordon, R. G., and Cape, C. D., 1981, Cenozoic latitudinal shift of the Hawaiian
hotspot and its implications for true polar wander: Earth and Planetary
Science Letters, v. 55, p. 37-47.
Gordon, R. G., and Cox, A., 1980, Calculating palwmagnetic poles for oceanic
plates: Geophysical Journal of the Royal Astronomical Society, v. 63,
p. 619-640.
Gordon, R. G., Cox, A., and Harter, C. E., 1978, Absolute motion of an individ-
ual plate estimated from its ridge and trench boundaries: Nature, v. 274,
p. 752-755.
Harland, W. B., Cox, A., Llewellyn, P. G., Pickton, C.A.G., Smith, A. G., and
Walters, R., 1982, A Geologic Time Scale: Cambridge, Cambridge Univer-
sity Press, 131 p.
Hellinger, S. J., 1979, The statistics of finite rotations in plate tectonics [Ph.D.
thesis]: Massachusetts Institute of Technology, 172 pp.
Henderson, L. J., and Gordon, R. G., 1981, Oceanic plateaus and the motion of
the Pacific plate with respect to hotspots: EOS Transactionsof the American
Geophysical Union, v. 62, p. 1028.
Henderson, L. J., Gordon, R. G., and Engebretson, D. C., 1984, Mesozoic aseis-
mic ridges of the Farallon plate and southward migration of shallow subduc-
tion during the Laramide orogeny: Tectonics, v. 3, p. 121-132.
Hilde, T. W., Uyeda, S., and Kroenke, L., 1976, Evolution of the western Pacific
and its margin: Tectonophysics, v. 38, p. 145-165.
Jarrard, R., and Clague, D., 1977, Implications of Pacific island and seamount
ages for the origin of volcanic chains: Reviews of Geophysics and Space
Physics, v. 15, p. 57-76.
Keith, S. B., 1978, Paleosubduction geometries inferred from Cretaceous and
Tertiary magmatic patterns in southwesternNorth America, Geology, no. 6,
p. 516-521.
-Paleocene subduction rates determined from K20/Si02 ratios in magmatic
rocks and their application to Cretaceous and Tertiary tectonic patterns in
southwestern North America: Geological Society of American Bulletin, no.
93, p. 524-532.
Klitgord, K. D., Popenoe, P., and Schouten, H., 1984, A Jurassic transform plate
boundary: Journal of Geophysical Research, no. 89, p. 7,753-7,772.
Larson, R. L., and Chase, C. G., 1972, Late Mesozoic evolution of the western
Pacific Ocean: Geological Society of America Bulletin, no. 83,
p. 3627-3644.
Larson, R. L.,and Lowrie, W., 1975, Paleomagnetic evidence for motion of the
Pacific plate from leg 32 basalts and magnetic anomalies: Initial Reports of
the Deep Sea Drilling Project, no. 32, p. 571-577.
Lipman, P. W., 1980, Cenozoic volcanism in the western United States: implica-
tions for continental tectonics: Continental Tectonics, Geophysics Study
Committee, National Research Council, Washington, p. 161-174.
-1982, Cenozoic volcano-tectonic transitions in the Rocky Mountains: corre-
lations with plate models (abstract): EOS Transactions of the American
Geophysical Union, no. 63, p. 91 1.
Lipman, P. W., Prostka, H. J., and Christiansen, R. L., 1972, Cenozoic volcanism
and plate-tectonic evolution of the western United States: I. Early and mid-
 Relative Motions Between Oceanic and Continental Plates
die Cenozoic: Philosophical Transactions of the Royal Society of London A,
no. 271, p. 217-278.
Livaccari, R. F., Burke, K., and Sengor, A.M.C., 1981, Was the Laramide orog-
eny related to subduction of an oceanic plateau? Nature, no. 289,
p. 276-278.
Menard, H. W., 1984, Evolution of ridges by asymmetrical spreading: Geology,
v. 12, p. 177-180.
Minster, J. B., and Jordan, T. H., 1978, Present-day plate motions: Journal of
Geophysical Research, no. 83, p. 5331-5354.
Minster, J. B., Jordan, T. H., Molnar, P., and Haines, E., 1974, Numerical
modeling of instantaneous plate tectonics: GeophysicalJournal of the Royal
Astronomical Society, no. 36, p. 541-576.
Molnar, P., and Atwater, T., 1978, Interarc spreading and Cordilleran tectonics as
alternates related to the age of subducted oceanic lithosphere: Earth and
Planetary Science Letters, no. 41, p. 330-340.
-1973, Relative motion of hot spots in the mantle: Nature, no. 246,
p. 288-291.
Molnar, P., and Francheteau, J., 1975, The relative motion of 'hot spots' in the
Atlantic and Indian Oceans during the Cenozoic: GeophysicalJournal of the
Astronomical Society, no. 43, p. 763-774.
Morgan, W. J., 1972, Plate motions and deep mantle convection: Geological
Society of America Memoir, no. 132, p. 7-22.
-1981, Hotspot tracks and the opening of the Atlantic and Indian Oceans, in
Emiliani, C., ed., The Sea, v. 7, New York, Wiley Interscience, p. 443-487.
-1983, Hotspot tracks and the early rifting of the Atlantic: Tectonophysics,
no. 94, p. 123-139.
Norton, I. 0..and Sclater, J. G., 1979, A model for the evolution of the Indian
Ocean and the breakup of Gondwanaland: Journal of Geophysical Re-
search, no. 84, p. 6803-6830.
Page, B. M., 1984, Which plate, Farallon or Kula, interacted with the California
margin 75-45 MYBP?, edited by D. G. Howell, D. L. Jones and A. Cox, eds.,
in Proceedings of the Circum-PacificTerrane Conference: Stanford, Stanford
University Publications, Geological Sciences, 18, p. 161-163.
Page, B. M., and Engebretson, D. C., 1984, Correlation between the geologic
record and computed plate motions for central California: Tectonics, no. 2,
p. 133-155.
Pilger, R. H., Jr., 1981, Plate reconstructions, aseismic ridges, and low-angle
subduction beneath the Andes: Geological Society of America Bulletin, no.
92, p. 448-456.
Pilger, R. H., Jr., and Henyey, T. L., 1979, Pacific-North American plate interac-
tion and Neogene volcanism in coastal California: Tectonophysics, no. 57,
p. 169-209.
59
Plumley, P. W., 1980, Paleomagnetism of Tertiary intrusive rocks in the Oregon
Coast Range: timing and mechanism of tectonic rotation [M.S. thesis]: West-
ern Washington University, Bellingham.
Schlanger, S. O., Garcia, M. O., Keating, B. H., Naughton, J. J., Sager, W. W.,
Haggerty, J. A., Phillpotts, J. A., and Duncan, R. A., 1984, Geology and
geochronology of the Line Islands: Journal of Geophysical Research, no. 89,
p. 11,261-1 1,272.
Sclater, J. G., Anderson, R. N., and Bell, M. L., 1971, The elevation of ridges and
the evolution of the central eastern Pacific: Journal of Geophysical Research,
no. 76, p. 7888-7915.
Simpson, R. W., and Cox, A., 1977, Paleomagnetic evidence for tectonic rotation
of the Oregon Coast Range: Geology, v. 5, p. 585-589.
Srivastava, S. P., 1978, Evolution of the Labrador Sea and its bearing on the early
evolution of the North Atlantic: Geophysical Journal of the Royal Astro-
nomical Society, no. 52, p. 313-357.
Stock, J., and Molnar, P., 1982, Uncertainties in the relative positions of the
Australia, Antarctica, Lord Howe, and Pacific plates since the Late Cre-
taceous: Journal of Geophysical Research, no. 87, p. 4697-4714.
Stock, J., and Molnar, P., 1983, Some geometrical aspects of uncertainties in
combined plate reconstructions: Geology, v. 11, p. 697-701.
Suarez, G., and Molnar, P., 1980, Paleomagnetic data and pelagic sediment fades
and the motion of the Pacific plate relative to the spin axis since the Late
Cretaceous: Journal of Geophysical Research, no. 85, p. 5257-5280.
Uyeda, S., and Yamano, M., 1984, Softness of the lithosphereand change of plate
motions due to collisions, in D. G. Howell, D. L. Jones, and A. Cox, eds.,
Proceedings of the Circum-Pacific Terrane Conference, Stanford, Stanford
University Publications, Geological Sciences, 18, p. 195-196.
Wallace, W. K., and Engebretson, D. C., 1984, Relationships between plate
motions and Late Cretaceous to Paleogene magmatism in southwestern
Alaska: Tectonics, no. 3, p. 295-315.
Wells, R. E., Engebretson, D. C., Snavely, P. D., Jr., and Coe, R. S., 1984,
Cenozoic plate motions and the volcano-tectonic evolution of western
Oregon and Washington: Tectonics, no. 3, p. 275-294.
Wilson, D. S., Hey, R. N., and Nishimura, C., 1984, Propogation as a mechanism
of reorientation of the Juan de Fuca ridge: Journal of Geophysical Research,
no. 89, p. 9,215-9,255.
Woods, M. T., and Davies, G. F., 1982, Late Cretaceous genesis of the Kula plate:
Earth and Planetary Science Letters, no. 58, p. 161-166.
Zandt, G., and Furlong, K. P., 1982, Evolution and thickness of the lithosphere
beneath coastal California: Geology, no. 10, p. 376-381.
MANUSCRIPT
ACCEPTED
BY THE SOCIETY
JUNE
11,1985
 Typeset by WESType Publishing Services, Inc., Boulder, Colorado
Printed in U.S.A. by Malloy Lithographing, Inc., Ann Arbor, Michigan