FACIES MODELS REVISITED

Edited by:
HENRY W. POSAMENTIER
Anadarko Petroleum Corporation, 1201 Lake Robbins Drive, The Woodlands, Texas 77380, U.S.A.
AND
ROGER G. WALKER
Roger Walker Consulting Inc., 83 Scimitar View NW, Calgary, Alberta T3L 2B4, Canada

Copyright 2006 by
SEPM (Society for Sedimentary Geology)

Laura J. Crossey and Donald S. McNeill, Editors of Special Publications

SEPM Special Publication 84

Tulsa, Oklahoma, U.S.A. September, 2006

 SEPM (Society for Sedimentary Geology) is an international not-for-profit Society based in Tulsa,
Oklahoma, U.S.A.. Through its network of international members, the Society is dedicated to the
dissemination of scientific information on sedimentology, stratigraphy, paleontology, environmental
sciences, marine geology, hydrogeology, and many additional related specialties.

The Society supports members in their professional objectives by publication of two major scientific journals, the
Journal of Sedimentary Research (JSR) and PALAIOS, in addition to producing technical conferences, short courses,
and Special Publications. Through SEPM's Continuing Education, Publications, Meetings, and other programs,
members can both gain and exchange information pertinent to their geologic specialties.

For more information about SEPM, please visit www.sepm.org.

ISBN 1-56576-121-9
2006 by
SEPM (Society for Sedimentary Geology)
6128 E. 38th Street, Suite 308
Tulsa, Oklahoma 74135-5814, U.S.A.
Printed in the United States of America
 FACIES MODELS REVISITED

Henry W. Posamentier and Roger G. Walker, Editors

CONTENTS

Facies models revisited: Introduction

ROGER G. WALKER .................................................................................................................................................................................... 1

Eolian facies models

NIGEL P. MOUNTNEY .............................................................................................................................................................................. 19

Fluvial facies models: Recent developments

JOHN S. BRIDGE ......................................................................................................................................................................................... 85

Estuarine and incised-valley facies models

RON BOYD, ROBERT W. DALRYMPLE, AND BRIAN A. ZAITLIN. ................................................................................................ 171

Deltas
JANOK P. BHATTACHARYA ................................................................................................................................................................ 237

A reexamination of facies models for clastic shorelines

H. EDWARD CLIFTON ............................................................................................................................................................................ 293

Facies models revisited: Clastic shelves

JOHN R. SUTER ........................................................................................................................................................................................ 339

Deep-water turbidites and submarine fans

HENRY W. POSAMENTIER AND ROGER G. WALKER ..................................................................................................................... 399

Index ............................................................................................................................................................................................................ 521

 FACIES MODELS REVISITED 1

FACIES MODELS REVISITED

ROGER G. WALKER
Roger Walker Consulting Inc., 83 Scimitar View NW, Calgary, Alberta T3L 2B4, Canada
e-mail: walkerrg@telus.net

ABSTRACT: The papers contained on this CD mostly originate from a session at the 2002 Annual Meeting of the Canadian Society of
Petroleum Geologists, repeated at the 2004 Dallas AAPG Meeting. The theme of both sessions was Facies Models Revisited, to see what
sort of progress had been made since the third (1992) edition of Facies Models, published by the Geological Society of Canada.
During the ten years between 1992 and 2002, there has been considerable progress in the understanding of modern and ancient
depositional environments. This additional complexity makes modeling much more difficult, and raises the problem of whether
modeling still serves a purpose. The original reasons for creating facies models still exista model is a point of comparison, it is a guide
for further observations, it serves as a basis for hydrodynamic interpretation, and most importantly, it acts as a predictor in new
situations.
Using submarine fans as an example, it is clear that progress during the last ten years (particularly in 3-D seismic) has highlighted
the inadequacy of all pre-existing modelsindeed, no comprehensive models have been proposed since the mid eighties. Yet with
continued and increasing exploration in submarine fan systems, predictive models are even more necessary. The traditional approach,
of distilling the features that modern and ancient systems have in common, is extremely difficult (and probably naive) in such diverse
and complex systems. Instead, it is necessary to identify all of the constituent building blocks of submarine fans (channels, point bars,
levees, splays, frontal lobes and so on), and try to identify the salient features of each. New models for particular situations can be
constructed by examining the relationships of the constituent building blocks. For example, sinuous channels, levees and splays may
be closely related in space, whereas frontal lobes are unlikely to be related to sinuous leveed channels (except for the channel that
ultimately feeds the lobe). A three-dimensional reconstruction can therefore be made by examining the building blocks that are closely
and commonly related, and also using information from the building blocks that are seldom or never found in juxtaposition.
These principles, discussed above for submarine fans, can be applied to all depositional environments, at all scales. The ideas are
elaborated in this introductory paper, and can be seen in the other contributions to this CD.

INTRODUCTION controlling physical processes of the elements represented in

Facies Models, a publication of the Geological Association
of Canada, first appeared in 1979 (Walker, 1979), with second and
third editions in 1984 and 1992 respectively (Walker, 1984; Walker
and James, 1992). In 2002, the Canadian Society of Petroleum
Geologists organized a session at their annual meeting entitled
Facies Models Revisited. The idea was to review progress in
facies modeling during the ten years since publication of the third
edition of Facies Models (Walker and James, 1992). The all-day
session was well received, and SEPM requested that a similar
session be organized at the annual AAPG meeting in Dallas (2004).
This CD includes most of the papers presented in Dallas. We
have taken advantage of the CD format by including abundant
full color illustrations of the examples discussed. The papers on
this CD cover clastic sediments only, and they are more compre-
hensive than the reviews in Facies Models. This partly reflects
the advances made during the last ten years, with increasing
recognition of the complexity and variability of depositional
environments.
This paper is organized in terms of increasing scale and
complexity. The concept of facies will be introduced first, fol-
lowed by facies associations and facies successions. Then the
stage is set for a discussion of facies models, with a final discus-
sion of future approaches to modeling.
FACIES AND FACIES MODELING
Facies modeling, as understood today, involves a synthesis of
information from ancient and recent depositional environments,
in an effort to understand the nature, scale, heterogeneity, and
each environment. The object is to identify the salient features of
recent sediments and ancient rocks, such that these features can
be identified, combined, and distilled into models that character-
ize that particular environment. Once a model is available, how-
ever simple and basic, it can be used to further our understanding
of natural systems. Perhaps the primary use is in the prediction
and interpretation of sandbodies in oil and gas reservoirs, but
increasing applications can be found in the movement of ground-
water through clastic sediments, and in environmental studies.
Facies models also play an important role in the understanding,
prediction and amelioration of coastal erosion, and Hurricane
Katrina in 2005 emphasizes the importance of incorporating
isolated extreme events into the formulation of facies models
this is a topic that has been underemphasized and is in need of
further development.
Facies
The concept of facies is a very old one, and was introduced
into geology by Nicholas Steno in 1669. It implied the entire
aspect of a part of the earths surface during a certain interval of
geological time (see Teichert, 1958). The modern usage was
introduced by Gressly (1838), implying the sum total of the
lithological and paleontological aspects of a stratigraphic unit.
Translations of Gresslys extended definition are given by
Teichert (1958) and Middleton (1973). The linkage of modern
and ancient environments probably dates back to Johannes
Walther in 1893. He suggested that the most satisfying genetic
explanations of ancient phenomena were by analogy with mod-
ern geological processes (quoted by Middleton, 1973, p. 981).

Facies Models Revisited

SEPM Special Publication No. 84, Copyright 2006
SEPM (Society for Sedimentary Geology), ISBN 1-56576-121-9, p. 117.
2 ROGER G. WALKER
FaciesA Working Definition
The most useful modern working definition of the term
facies was given by Middleton (1978), who noted that:
the more common (modern) usage is exemplified by de
Raaf et al. (1965) who subdivided a group of three forma-
tions into a cyclical repetition of a number of facies distin-
guished by lithological, structural and organic aspects
detectable in the field. The facies may be given informal
designations (Facies A etc.) or brief descriptive designa-
tions (laminated siltstone facies) and it is understood
that they are units that will ultimately be given an environ-
mental interpretation; but the facies definition is itself
quite objective and based on the total field aspect of the
rocks themselves . The key to the interpretation of facies
is to combine observations made on their spatial relations
and internal characteristics (lithology and sedimentary
structures) with comparative information from other well-
studied stratigraphic units, and particularly from studies
of modern sedimentary environments.
The term facies can be used in both a descriptive and an
interpretive sense. The definition above defines only the descrip-
tive facies. However, it may be useful, as a quick means of
communication, to tell a friend that you worked on a fluvial
facies. It is understood that you have made an interpretation of
the rocks you worked on, and that the term fluvial facies
encompasses a constellation of features including (in the fluvial
example) sharp-based fining-upward successions with lags at
their bases, thin siltstones with root traces, abundant trough and
planar tabular cross bedding, and the absence of marine indica-
tors. It is normally obvious from the context whether the term
facies is being used in a descriptive or an interpretive sense.
Facies Modeling
Although the term facies is old, the concept of facies modeling
is much younger. The term facies model was first used at a
discussion organized by Paul Potter at the Illinois State Geo-
logical Survey in 1958 (Potter, 1959). The purpose of the discus-
sion was to pool the knowledge and experience of the group
concerning three topics. These topics have a remarkably modern
ring to them, and could equally well have been chosen for a
research conference in 2005. They involved the existence and
number of sedimentary associations; the possibility of establish-
ing a model for each association that would emphasize the areal
distribution of lithological units within it; and the exploration of
the spatial and sequential relations between the associations.
Potter reports that a facies model was defined as the distribution
pattern or arrangement of lithological units within any given
association (a sedimentary association being a collection of
commonly associated sedimentary attributes). He continued
with another very up-to-date concept, that in the early stages of
geological exploration, the function of the model is to improve
prediction of the distribution of lithological types. One of Potters
conclusions remains true todaythe group discussion clearly
pointed out those areas of knowledge that permit generalization,
those areas in which problems are clearly recognized and stated
but for which definitive answers are not available, and the areas
in which the problems are not as yet clearly formulated.
In recent correspondence, Potter (personal communication,
2005) referred back to his 1959 discussion. A facies model was
defined as a commonly recurring sedimentary facies (that is
scale independent to a large degree)readers should note that
this is rather different from the modern usage, where a facies
model is on the scale of individual depositional environments
and may contain many different facies. However, Potter (per-
sonal communication, 2005) noted that his facies models were
created to improve prediction and understanding of how the
different lithologies that form a recurring facies are put together.
It is apparent that his facies are defined on a larger scale than
those discussed in this present paper. Finally, Potter noted that
improved prediction largely depends on relating the facies to
basin geometry and understanding the internal transport system
of the facies; together both help us understand the fabric of the
facies.
Despite the very forward-looking ideas expressed at Potters
conference, the term facies model did not catch on, perhaps
because there was insufficient information regarding deposi-
tional environments and architectures to make much progress.
The term was reintroduced in 1975 in my paper Generalized
facies models for resedimented conglomerates of turbidite asso-
ciation (Walker, 1975). At the time, I was unaware of Potters
earlier usage of the term. Shortly afterward I used the term again
in the first edition of Facies Models (Walker, 1979), not realizing
that the term would become so widely used in sedimentary
geology. In the second edition of Facies Models (Walker, 1984), it
is clear that facies are the smaller-scale building blocks of the
sedimentary record. Facies can be organized into facies se-
quences (today we would use the term facies successions), and
various sequences and successions in particular depositional
environments can be synthesized into models for that environ-
ment. Important points established in Walker (1984) include (1)
the term [facies] model has a generality that goes beyond a single
study of one formation, and (2) the facies model is a general
summary of a specific sedimentary environment, written in
terms that make the summary usable in at least four different
ways. These ways are discussed in this present paper.
The models discussed in the first two editions of Facies
Models (Walker, 1979, 1984) were static, inasmuch as they used
information from modern environments as seen today. The third
edition (Walker and James, 1992) examined models in the light of
responses to sea level change. In the last ten years, there has been
a large amount of new work on modern depositional environ-
ments. Also, in many cases (particularly turbidite systems) an
entirely new dimension has been added through 3-D seismic
studies. Consequently, the first questions to be asked at the Calgary
meeting in 2002 concerned the construction of facies models given
the size and diversity of the data base, and the appropriateness of
using such models in an increasingly complex world.
One of the themes of the Dallas symposium (2004) was to
examine these questions, and perhaps to shift the emphasis onto
the individual depositional elements within environments, and
the study of how these elements fit together laterally and verti-
cally. For example, submarine fans may be too complex for
simple models of entire fansnevertheless, most fans contain
similar elements (e.g., channel fills, splays, levees, frontal lobes
etc.) which can be stacked in various ways depending on relative
sea level fluctuations, local tectonics, variations in sediment
supply and other internal and external factors. Thus the various
descriptive entities have been reduced to the depositional-ele-
ment level, rather than the scale of the entire environment (sub-
marine fan, delta, shoreface etc.).
The best models embody large amounts of information from
as many examples as possible, modern and ancient. In generaliz-
ing this information, the resulting models can serve as reference
points for interpretations of new situations and examples, and as
a basis for making predictions from limited amounts of data in
new situations. The predictive aspects of facies models are impor-
 FACIES MODELS REVISITED 3

tant in subsurface exploration for oil, gas, and minerals, and,

increasingly, for studying and predicting the movement of ground
water through relatively unconsolidated surficial materials. In all
cases, the geometry and connectedness of the reservoirs or aqui-
fers is directly related to the original depositional environments.
Facies modeling can now be regarded as a mature science. The
basic facies, facies organizations and depositional controls of the
major environments are fairly well understood (perhaps with the
exception of submarine fans). However, depositional environ-
ments can always be subdivided (deltas, for example) and smaller-
scale models proposed for the various parts of larger-scale envi-
ronments (distributary mouth bars, interdistributary bays, pro-
grading beach ridge systems, etc.).

SCALES AND METHODS FOR DEFINING FACIES

Facies can be defined on many different scales. Whenever a
vertical stratigraphic section or core is measured, decisions have to
be made about what to include in each measurement unit. A simple
section might consist of 2 m of sandstone overlain by 2 m of shale.
The sandstone and shale units are different facies (they have a
different aspect), and the defining characteristic is their lithol-
ogy. In more detail, it might be decided to measure 1 m of cross-
bedded sandstone, 1 m of ripple cross-laminated sandstone, and 2
m of shale. In this case, the different sandstone facies are defined by
their sedimentary structures. It follows that the scale of subdivision
of a stratigraphic section into facies depends on:

1. the purpose of the study

2. the time available to make the measurements, and
3. the abundance of descriptive features in the rocks.
Purpose of the Study
If the objective of a study is a routine description and interpre-
tation on a large scale, the scale of facies subdivision may be fairly
broad. In contrast, if the goal is a detailed analysis and interpre-
tation of a thin stratigraphic unit, perhaps with a view to refining
cm scale
FIG. 1.Ripple cross lamination. Note symmetry of ripples (red
arrows), but unidirectional cross lamination (to the right,
yellow arrows). Symmetry suggests wave action, and cross
lamination suggests a superimposed unidirectional flow. These
are combined-flow ripples with relatively long periods of
mud deposition between sand emplacement. Discontinuity
of ripple layers suggests very limited sand supply. Pennsyl-
vanian, Devon, England.
FIG. 2.Ripple cross lamination. Red arrows show cross lamina-
tion dipping left, and blue arrows show lamination dipping
right. Note mudstone drapes (yellow arrows) between ripple
cross-laminated layers. Reversing flow directions with drapes
between layers suggests a tidal influence. Cretaceous, south-
ern Alberta. Core is 10 cm in diameter.
an existing model, the scale of facies subdivision must be much
more detailed. As an example, compare Figures 1, 2 and 3. They
all could be described as ripple cross lamination. However,
Figure 1 shows symmetrical ripple profiles with unidirectional
cross lamination, and can be interpreted as combined-flow ripples
(wave plus current). Figure 2 shows ripple cross lamination
alternating with thin mudstone drapes. The cross lamination is
directed both to the right and the left, and the assemblage of
features suggests alternating tidal currents with mud deposited
during slack tides. Figure 3 shows ripple cross lamination with-
out mudstone layers. Lamination is preserved on the ripple
crests, and sets of ripples climb on the backs of each other. This
lamination formed during active deposition from suspension,
perhaps during a waning flood stage.
These observations go beyond ripple cross lamination, and
involve sand/mud ratios, continuity of lamination within the
ripples and preservation (or not) of the entire ripple form. If the
purpose of the study is a detailed interpretation of sedimentary
environments, distinguishing Figures 1, 2 and 3 as different facies
is important. In a broader study, lumping all three Figures
together into one ripple cross-laminated facies might be suffi-
cient.
Time Available for the Study
With an entire day available for the description of one 18 m
core, it will be possible to subdivide the core into many different
4 ROGER G. WALKER

facies using varied and subtle criteria to distinguish the various

facies. In contrast, if a thick unit has to be studied over a wide area
in only two weeks of field work, the facies subdivisions will
necessarily be broader. This again is illustrated in Figures 1, 2, and
3, where more time would be necessary to distinguish and
describe the details of the ripple cross lamination, as opposed to
lumping all three figures together as ripple cross laminated.
The importance of time available and purpose of study are
also illustrated in Figure 4. With little time and a routine descrip-
tion on a large scale, the sandier-upward succession (red arrow)
could be described as one facieshummocky cross-stratified
sandstones interbedded with bioturbated mudstones. If more
time were available and more detail required, the succession
could be divided into three different facies (yellow arrows), a
lower muddier facies, a central facies in which the sandstone/
shale ratio is about 1, and an upper facies consisting almost
entirely of sandstone.

Descriptive Features

FIG. 3.Ripple cross lamination. Red arrows indicate preserva-
tion of lamination over the ripple crests, and yellow arrows
indicate forward movement and aggradation of the ripples.
This climbing-ripple cross lamination suggests rapid deposi-
tion from suspension during ripple formation, with no pauses
for mud deposition. Pleistocene, southern Ontario. Coin is 2.2
cm in diameter.
The degree of subdivision always depends on the abundance
of features in the rocks. A thick succession of interbedded sand-
stone and shales (thin-bedded turbidites) will be difficult to
subdivide because of the monotonous nature of the succession. In
contrast, a complexly interbedded succession of mudstones and
sandstones with a wide variety of sedimentary structures and
trace fossils, and various scales of interbedding of the lithologies

FIG. 4.Sandier-upward succession from the Cardium Formation (Cretaceous), Blackstone River, Alberta. Person circled for scale.
Red arrow shows overall sandier-upward succession, and yellow arrows suggest how the succession could be subdivided into
a lower muddier portion, a main central portion characterized by interbedded sandstones and mudstones, and an upper
dominantly sandy portion.
 FACIES MODELS REVISITED
(as might occur in an estuary fill), will allow detailed facies
subdivision on a small scale. Figures 1, 2 and 3 clearly present a
wealth of descriptive features on a small scale, whereas Figure 5
presents few descriptors within the thick package of thin-bedded
sandstones and shales.
Subdivision of a stratigraphic section into facies should not
be undertaken before gaining familiarity with the complete
section. It will then be apparent how much variability there is,
and how many facies should be defined to describe the unit
adequately. Most facies studies to date have relied on qualita-
tively assessed combinations of characteristic sedimentary struc-
tures and trace fossils (e.g., de Raaf et al., 1965; Williams and
Rust, 1969; Wilson, 1975; Cant and Walker, 1976; Scholle et al.,
1983; Walker, 1983). Statistical methods have also been used,
particularly where there is agreement among workers as to the
important quantifiable descriptive parametersfor example,
the proportions of different types of clasts in carbonate rocks
(Imbrie and Purdy, 1962; Klovan, 1964; Harbaugh and
Demirmen, 1964; Harbaugh and Merriam, 1968). Statistical
methods are less suited to clastic rocks, where the most impor-
tant descriptors (sedimentary structures and trace fossils) can-
not easily be quantified.
The facies concept can be extended beyond observable rock
types. Different seismic characteristics have given rise to various
seismic facies (Weimer, 1989), and different well log characteris-
tics have also given rise to log facies (Fig. 6). The gamma ray
FIG. 5.Thin-bedded classical turbidites, Devonian, South Australia. Stratigraphic top to left. The succession is characterized by its
homogeneity, with almost no change in facies from bottom to top.
5
profile in Fig. 6A shows a progressive upward shift to the left,
indicating cleaner rocks with fewer claysthe profile is com-
monly described as cleaning-upward, or more interpretively,
sandier-upward. The profiles in Figs. 6B and C are very similar,
and both show abrupt shifts to the left in the gamma ray profiles,
rather than the progressive shift seen in Fig. 6A. The profile is
commonly described as blocky. The log facies of Figs. 6A and
6B are the same, and differ from the log facies of Fig. 6A.
However, similar log facies may have very different interpreta-
tionsthe blocky profiles can indicate sharp-based channel-
filling sandstones, or sharp-based forced-regressive shorefaces.
Without other information, preferably core control, the two pos-
sibilities cannot be distinguished.
The pitfalls involved in the interpretation of such log facies are
highlighted by the fact that channel sandbodies and shoreface
sandbodies commonly trend at right angles to each other.
FACIES ASSOCIATIONS
In many studies, facies may have been defined in a detailed
way on a small scale, with minor subtle differences between the
facies (e.g., Walker, 1983). This subdivision may result in a facies
scheme where the descriptive differences outstrip our ability to
interpret the differences. It is therefore useful in such studies to
combine closely related facies into facies associations, or groups of
facies genetically related to one another and which have some
6 ROGER G. WALKER

05-03-50-8W5 07-20-51-11W5 05-28-09-09W4

B C

FIG. 6.Gamma-ray logs from three locations in Alberta. A shows a cleaner-upward succession, whereas B and C show
examples of a blocky gamma-ray signature. Scale in meters. Interpretations: 6A is a prograding shoreface from the Cretaceous
Second White Specks, 6B is a forced regressive shoreface from the Cretaceous Cardium Formation, and 6C is a channel fill from
the Cretaceous Viking Formation.

environmental significance (Collinson, 1969). These larger-scale
facies associations have also been termed architectural elements
(Allen, 1983), denoting the building blocks of the various deposi-
tional systems. It will be suggested later that the definition of
architectural elements is fundamental to the construction of
improved facies models in situations where the complexity of the
system (e.g., deltas) appears to discourage the development of
simple models.
It is now clear that some facies and architectural elements
occur universally, in recent and ancient sediments, and in many
different basins around the world. The first universal facies
scheme was proposed for turbidites (Mutti and Ricci Lucchi,
1972), and Miall (1977, 1985) has suggested a universal scheme for
fluvial deposits. For example, Mialls (1985) channel architectural
element (CH) consists of any combination of a series of defined
lithofacies which communally have a distinctive elongate chan-
nel geometry; it is part of the architecture of almost all modern
rivers and can be recognized in most ancient fluvial deposits.
Other examples of universal facies include:
1. sharp-based hummocky cross-stratified sandstones inter-
bedded with bioturbated mudstones (Fig. 7) (storm-domi-
nated shelf deposits),
2. giant (20 m +) sets of planar tabular cross bedding (Fig. 8)
(eolian environments),
3. thin-bedded turbidites with climbing ripples, convolute
lamination, and ripped-up mud clasts (Fig. 9) (the CCC
turbidites of Walker, 1985, interpreted as levee deposits),
and
4. cross bedding with bundles of mudstone drapes indicative
of deposition in a tidally dominated environment (Fig. 10)
(Visser, 1980).
Generally, the subtle differences that enable individual facies
to be defined may be the result of many small-scale local factors
affecting deposition. Architectural elements are the larger-scale
components of a depositional systemthey will tend to be more
general in nature, less influenced by local factors, and hence more
universal in their application.
FACIES SUCCESSIONS
This term implies that certain facies properties change pro-
gressively and systematically either vertically or horizontally.
 FACIES MODELS REVISITED
FIG. 7.Interbedded sandstones and mudstones, Kenilworth Formation, Book Cliffs, Utah. Red arrows show convex-upward
stratification typical of hummocky cross stratification (HCS), and black arrows show low-angle curved intersections of
stratification, also characteristic of HCS.
Examples include changing proportions of sand, giving rise to a
sandier-upward (or muddier-upward) succession (Fig. 4), chang-
ing grain size giving rise to a fining-upward (or coarsening-
upward) succession, or changing bed thickness giving rise to a
thickening-upward (or thinning-upward) succession. The im-
portance of recognizing such successions is that they place indi-
vidual facies into a context. Some individual facies, for example
medium-scale cross bedding (Fig. 11), may be descriptively and
hydrodynamically identical, but may actually represent different
depositional environmentsmedium-scale cross bedding can
form in many settings, including meandering or braided rivers,
tidal inlets, a shoreface dominated by alongshore currents, or in
an open marine tidal setting. The key to distinguishing the
environments lies in the context of the facies in vertical and
horizontal successionfor example, the shoreface cross bedding
might overlie offshore mudstones and be overlain in turn by
beach and nonmarine deposits. Thus the succession contributes
important information that the individual facies cannot.
The relationship between depositional environments in space
and the resulting stratigraphic successions developed through
time was first emphasized by Johannes Walther (1894, translated
in Middleton 1973) in his Law of the Correlation of Facies. Walther
stated that:
it is a basic statement of far reaching significance that only
those facies and facies areas can be superimposed primarily
which can be observed beside each other at the present time.
Application of this law suggests that in a vertical facies
succession a gradational transition (Fig. 12) from one facies to
another implies that the two facies represent environments that
were once adjacent laterally. If the contacts between facies or
facies associations are sharp and/or erosional (Fig. 12), there is no
way of knowing whether the two vertically adjacent facies repre-
sent environments that were once laterally adjacent. For example,
7
a fall of relative sea level may force a shoreface to prograde
rapidly onto an erosion surface, placing the shoreface sharply
and erosionally on shelf bioturbated mudstones. The erosion
surface represents the non-preservation of inner-shelf and shelf
shoreface transitional environments. Thus sharp breaks between
facies, marked for example by channel scours, thin bioturbated
horizons (Glossifungites surfaces) or thin pebble lags, may signify
fundamental changes in depositional environments, and per-
haps the beginning of new cycles of sedimentation (de Raaf et al.,
1965). Many of these sharp breaks (bounding discontinuities) are
now used to separate stratigraphic sequences and allostratigraphic
units (discussed below).
The relationships between facies within facies successions can
be shown qualitatively in facies relationship diagrams (Fig. 13; de
Raaf et al., 1965) or tabulated to show the numbers of observed
transitions. These numbers can be converted to probabilities, in a
technique known as Markov chain analysis. This technique is not
used as commonly as it once was, but interested readers are
referred to the second edition of Facies Models (Walker, 1984;
Harper, 1984).
FACIES MODELS
A facies model can be defined as a general summary of a
specific depositional environment, incorporating information
from recent sediments and ancient rocks. Two problems are
immediately apparent: scale and interpretation. The problem of
scale refers to the environment to be modeleddo we need a
model for barrier-island and lagoon systems (large scale), or
should we be separately modeling the lagoons and the shoreface
dunes of the barrier itself (small scale)? A simple answer
suggests that both scales need modeling. The smaller scale is
easier to define, describe, and distill, and a group of related
small-scale models can probably be combined into a larger-
scale model.
8 ROGER G. WALKER
FIG. 8.Giant (20 m) planar-tabular cross bedding from the Permian White Rim Sandstone, Canyonlands, Utah. Scale suggests an
eolian origin for the cross bedding.
The problem of interpretation concerns the integration of data
on ancient rocks into the facies model. For example, it is easy to
choose several examples of modern wave-dominated shorefaces
and compare their characteristicsthe data set is homogeneous.
However, incorporating data from ancient sediments into this
facies model involves making interpretationswe may be cor-
rect in many of the examples we choose, but some tide-dominated
shorefaces may end up in our data set, which would then be less
homogeneous.
In all modeling, a philosophical assumption must be made,
that there is system and order in Nature, and that geologists can
identify and agree upon a limited number of depositional envi-
ronments and systems. In a well argued alternative view,
Anderton (1985, p. 33) suggests that if, like me, you have a more
nihilistic view of life, the universe and everything, then you
have to admit an infinite number of environments, facies and
models.
For those of us who seek order in Nature, and who see value
in building facies models, the principles, methods, and motives
are shown in Figure 14. In this figure, a turbidite / submarine fan
example has been usedreaders should understand that in order
to present the basic ideas the details of submarine fan systems
have been very oversimplified.
The first assumption is that many modern submarine fan
systems have been studied, and many ancient turbidite systems
studied and interpreted. As a result of this work, we then
assume that we can make some general statements about how
20 m
such systems operate, rather than making statements only about
each individual example (see turbidite contribution by
Posamentier and Walker on this CD). These general statements
are obviously more powerful than countless statements involv-
ing only individual examples. The process of extracting this
general information is shown in Figure 14, using some modern
fans (Rhone, Amazon and Mississippi) and some ancient tur-
bidite examples (Wheeler Gorge, Frigg fan). Obviously, a better
generalization (model) would result if more examples were
used.
The entire wealth of information is first distilled, boiling away
the local details and concentrating the important features that all
examples share. The features that all examples share (in this case,
perhaps the monotonous alternations of parallel-bedded sand-
stones and shales) may be relatively easy to agree upon, but in
many cases separating local detail from general principles may be
more contentious. In this case, the conglomerates of Wheeler
Gorge are quite unlike the fine-grained sheetlike turbidites of the
outer parts of Mississippi Fan.
Answering these questions involves a thorough knowledge
of the literature along with extensive individual experience,
judgement, and discussion with other workers with different
experiences. Models are constantly being refined as more ex-
amples become available, as more distinct architectural elements
are recognized, and as depositional processes become better
understood. This is shown in Figure 14 by the feedback loop from
distillation to model, from model to comparison with more local
 FACIES MODELS REVISITED 9

A B C
FIG. 9.CCC turbidites from the Cretaceous Lange Formation, offshore mid-Norway. In A, note convoluted parallel and ripple cross
lamination. In B, the ripple cross lamination becomes progressively more convoluted upward. In C, the ripples are climbing
(arrow) and convoluted. Core is 10 cm in diameter.

examples, and incorporation of those new examples into the data
base. This in turn demands renewed distillation, and so on.
CHOICE OF ENVIRONMENTS TO MODEL
Facies models have traditionally been formulated for deposi-
tional systems that form obvious geographical entitiesfor ex-
ample, meandering rivers or deltas. In many environments, and
deltas are a good example, it is clear that many different deposi-
tional processes combine to give many different depositional
results. Thus deltas may be wave dominated, river dominated, or
tide dominated, all of which have very different geometries and
sandbody distributions (Bhattacharya, this CD). Yet they all
conform to one of the classic definitions of a delta, as a river-fed
depositional system that results in an irregular progradation of
the shoreline (Scott and Fisher, 1969).
The results of nearly fifty years of research (since publication
of Recent Sediments, Northwest Gulf of Mexico; Shepard et al.,
1960) have suggested that deltas may be too big and too complex
for the formulation of good facies models. Variability within
river-influenced deltas such as the Mississippi is enormous,
embracing many smaller environments that may deserve their
own models (distributary mouth bars, bays, distributary chan-
nels and levees, crevasses splays, beaches and barriers, lagoons,
and offshore shoals, among others). It should also be emphasized
that the basis of many facies models consists of the work done
during long periods of normal conditions. Hurricane Katrina
in 2005 may have altered the Mississippi Delta in ways that will
be preserved and may be recognized in the geological record
these effects will need to be built into future deltaic models.
Similar comments concerning facies complexity and the impor-
tance of rare catastrophic events can be made about submarine
fans and fluvial systems.
Depositional Elements Rather Than Environments
Instead of modeling obvious geographical entities, it may be
preferable to model discrete depositional elements. These ele-
ments may be found in several geographical settings. An obvious
example is the shoreface depositional element. It is controlled by
alongshore sediment supply under the influence of shoaling waves.
Variability of process, for example the relative importance of fair-
weather and storm processes, combined with the rate and caliber
of sediment supply, combine to give rise to a closely related set of
depositional products. If the immediate snapshot of the shoreface
at one time is combined with evolution of the depositional system
through time, emphasizing relative sea-level fluctuations, the depo-
sitional products may be gradationally based and sandier upward
(a normal prograding shoreface), or sharp based and sandy through-
out (the result of progradation during relative sea level falla
forced regressive sequence). Nevertheless, the shoreface is a rela-
tively easily defined and easily understood depositional element
(compared with the complexity of a delta). The shoreface element
occurs in several geographic environments, and contributes to
10 ROGER G. WALKER
FIG. 10.Set of cross bedding 20 cm thick from Cretaceous sandstones of Leighton Buzzard, England. Note mudstone drapes on the
foresets (yellow arrows), and at least one paired set of drapes (red arrow) indicative of deposition in a tidal setting. The mudstone
drapes form during slack tides.
several geographically based traditional facies modelsfor ex-
ample, wave-dominated deltas, prograding strandplains, barrier
islands and transgressive shoreline systems. Other examples are
discussed in the individual contributions on this CD.
Another excellent example of depositional elements within a
specific depositional environment has been presented by Miall
(1977, 1985). In this fluvial example, a set of architectural elements
has been defined, using capital letters to designate grain sizes (G
for gravel, S for sand, etc.). Lower-case letters were used to
indicate sedimentary features (f for flat bedded, for example),
resulting architectural element designations such as Gf. Mialls
catalog of elements for fluvial systems is very useful, particularly
for workers new to fluvial systems seeking guidance in what to
look for and what to describe. More experienced workers will be
sensitive to the possibility that there may be elements that do not
fall easily into Mialls catalog.
Elements Remain Constant
Geographic Environments Change Through Time
The shoreface depositional element discussed above remains
relatively constant through time. It progrades given sufficient
sediment supply, with details of the progradation being controlled
by wave and tidal processes, and changing relative sea level.
On a larger scale, geographically defined environments may
change dramatically through time, such that very few parts of
todays snapshot may be preserved in the geological record.
Barrier islands form an excellent example (Fig. 15; see contribu-
tion by Boyd et al. on this CD), where one of the most obvious
parts of the system involves the foreshore and barrier superstruc-
ture (Fig. 15). However, Rampino and Sanders (1980) have shown
by detailed coring studies (Fig. 16) that during transgression the
sand from the foreshore and barrier superstructure is moved (1)
seaward by storm waves to form new nearshore sand ridges (Fig.
16, red arrow), and (2) landward into the lagoon as washover
facies (Fig. 16, blue arrow). In a transgressive setting, the barrier
itself is not preservedthe resulting sedimentary record is shown
in the black rectangle of Fig. 16, and consists of thin lagoonal and
washover deposits, sharply overlain by thin nearshore sand
ridges. Any attempt to model a geographically defined barrier
island and lagoonal system is bound to encounter severe prob-
lems of facies preservation. It will be more fruitful to identify the
various depositional elements of the system, to study the pro-
cesses that control them today, and the ways in which they will
(or will not) be preserved in the geological record. Our model, for
purposes of comparison and prediction (see below), may end up
as a transgressive lagoon-washover model (Fig. 16), rather than
a barrier-island model exemplified by Figure 15.
THE USES OF FACIES MODELS
There has been little or no discussion in the literature of the
original four uses of facies models proposed by Walker (1979,
1984; Walker and James, 1992). The generality embodied in a
model, as opposed to a summary of one particular example,
enables the model to assume four main functions (Fig. 14):
 FACIES MODELS REVISITED
FIG. 11.Trough cross bedding, seen more or less parallel to flow direction (to the left), from the Cretaceous rocks in, Berry Gulch,
Colorado. 15 cm scale circled.
1. a norm, for purposes of comparison
2. a framework and guide for future observations
3. a predictor in new situations, and
4. a basis for interpretation
The Model As a Norm
Figure 17 shows a simple 2-D model of a fluvial fining-
upward succession. It was derived from data published by
Allen (1970) on more than one hundred examples of Devo-
nian fluvial successions in Britain, and redrawn to scale in
Figure 17. It is characterized by roughly equal proportions of
point-bar and vertical-accretion facies, both about 3 m thick,
and the entire fining-upward sequence can be considered as
a norm.
Let us then assume that during field work a new succession is
found with a 6-m thick point-bar succession and 3 m of vertical
accretion facies. By itself, this succession may be difficult to
interpret, but by comparison with the norm (Fig. 17) it is imme-
diately clear that the new point-bar succession is twice as thick as
the norm. This comparison opens new lines of thought and
interpretationwas the river unusually deep; was the rate of
subsidence unusually high; does the sandbody actually consist of
11
a stack of two 3-m-thick sandbodies? These are questions and
ideas that are possible only if the new example is compared with
a norm. A fourth question also arises: is the comparison valid, or
is an apple being compared with a norm for oranges? In the case
of Allens fluvial successions, the norm has probably been con-
structed from very homogeneous data. Comparisons with other
Devonian examples may be good, but comparisons with (say)
Cretaceous rivers from very high-accommodation settings may
be less useful.
The Model As a Framework for Observations
A good model summarizes all of the important descriptive
features of a particular system. For example, in Figure 17 the
fluvial fining-upward sequence contains a basal lag overlain
by various cross-bedded and parallel-laminated facies. These
in turn are overlain by ripple cross-laminated facies. The
vertical-accretion facies may contain root traces, desiccation
cracks, and caliche nodules. These are the basic descriptors of
this particular model, and they act as a guide for making
observations in new examplesis the succession the same; are
the proportions of facies the same; are any distinctive features
absent; or are there new features that are not included in the
current model? Mialls (1977, 1985) fluvial depositional ele-
ments (Gf, etc.) are also excellent examples of a framework for
future observations.
12 ROGER G. WALKER

Sharp and/or erosive facies boundaries

Gradational facies boundaries

within successions

FIG. 12.Cretaceous Mountain Park Formation, Alberta. Yellow arrows show three sandier-upward successions. Within these
successions, all of the facies boundaries are gradational. At the tops of the successions, contacts are sharp (yellow dotted lines).
At the horizon of the uppermost sandstone, the dotted lines show a sharp base and a sharp and erosional top (seen at right).

The Model As a Predictor
This is without question the most important function of any
facies model. The basic idea is very simple: given one new piece
of information, it may be possible (1) to assign that information to
a particular model, and therefore (2) use the model to predict the
rest of the system.
As an example, some thin-bedded turbidites are shown in
Figure 9. The sandstones are characterized by climbing ripples
and convolute lamination, features that are more abundant in
thin-bedded levee turbidites than in thin-bedded basin-plain
turbidites (Walker, 1985). We may therefore make a preliminary
interpretation, assigning the beds in Figure 9 to the levee of a
deep-sea channel. To make further predictions, we select a model
for deep-sea channels (see Posamentier and Walker on this CD).
The one new piece of information (thin-bedded levee facies), plus
the appropriate model, suggests a system involving back-of-
levee facies, channel-margin facies, channel-fill facies (perhaps
coarser sandstones interbedded with mudstones), and possible
channel point-bar facies with morphological scars as the point
bar has shifted in position as the channel migrated (Kolla et al.,
2001; Posamentier and Walker on this CD).
There are two equally important steps in using models as
predictors. The first involves the correct interpretation of the new
piece(s) of information. If thin-bedded levee turbidites were
incorrectly interpreted as distal basin-plain facies, the wrong
model might be chosen for prediction. It follows that the second
important step involves selection of an appropriate model. As
another example, let us assume that our new piece of information
consists of the lag and cross-bedded sandstones at the base of the
succession in Figure 17. The appropriate model would be a
meandering-river modelan inappropriate model would be a
braided-river model or a tidal-inlet model.
In many cases, it will be possible to test the predictions made
from the one new piece of information, perhaps by examining all
of the nearby outcrops. But if thin-bedded turbidites similar to
those in Figure 9 are also characteristic of one core from a Miocene
submarine fan in offshore West Africa, testing the prediction of a
nearby channel sandbody may involve millions of dollars of
drilling costs. Clearly, the new data must be interpreted as
carefully and accurately as possible, and the most appropriate
model used for prediction.
The Model As a Basis for Hydrodynamic Interpretation
This use of facies models was originally prompted by the idea
that one individual turbidite may be difficult to interpret, whereas
many turbidites combined into a model (the Bouma sequence,
1962) would give a more consistent and general basis for inter-
preting depositional processes (Harms and Fahnestock, 1965;
 FACIES MODELS REVISITED 13

cross-stratified sandstones interbedded with bioturbated

mudstones (Fig. 7), b) thin-bedded turbidites with climbing
ripples, convolute lamination, and ripped-up mud clasts
(Fig. 9), c) sets of cross stratification thicker than 10 m (Fig.
8), that may be either planar-tabular or trough shaped (the
latter commonly with very long toesets), and d) cross-
bedded sandstones with paired mudstone drapes on the
foresets (Fig. 10). There are more and more examples of
universal facies being recognized, and these will form the
basic building blocks of the sedimentary record.

2. Modern processes, varying rates of sediment supply, and

fluctuations in relative sea level combine to form distinctive
depositional elements in recent sediments. Some of these
elements are also universal, and occur in many different
places around the world today. Examples include a)
shorefaces, b) eolian desert dune complexes, c) tidal inlet
tidal delta systems, and d) lagoonal and barrier washover
systems. Many other examples could easily be added to this
list, but perhaps with the exception of desert dunes the
depositional elements tend to be smaller rather than larger
in scale, and homogeneous in character.

Granted these two truths, the future of modeling may lie in

FIG. 13.Facies-relationship diagram from the Carboniferous
refining and agreeing upon the commonly occurring universal
Westward Ho! Formation, north Devon, England. Facies have
architectural elements in the geological record (a process akin to
been given descriptive names, and arrows show sharp and
distillation in Fig. 14). At the same time, there may be more
gradational facies contacts, and numbers indicate the occur-
refining and agreeing upon the combinations of processes that
rence of each transition. From de Raaf et al. (1965).

Walker, 1965). This interpretive usage of models is probably more

effective for small-scale models (the Bouma sequence for turbid-
ites), and less effective for large-scale systems (e.g., deltas). Also,
individual turbidity-current depositional processes are difficult Distill local examples

to observe in modern oceans, whereas work over the last 20 years

in nonmarine and shallow-marine depositional environments
has added enormously to our understanding of depositional Make a model
processes. Models have not been used extensively as a basis for
hydrodynamic interpretation, and this is probably their least
useful aspect.
Comparison
FUTURE APPROACHES TO MODELING

It was pointed out above that our knowledge of modern

environments has expanded tremendously in the last twenty-five
years (since the publication of the first edition of Facies Models).
Because of this, many modern environments may be perceived as
so complex that simple models and distillations may be impos-
sible, inappropriate, or both. In the absence of models, however,
there will be no norms, and no bases for making predictions. In
the hope that modeling of any sort is preferable to anarchy, I
emphasize two basic truths from which we might proceed. Prediction
1. Individual, small-scale facies can be identified in ancient
rocks, as shown in Figures 1, 2 and 3. Associations of facies
that commonly occur laterally and vertically adjacent to each
other can be combined into facies associations, and these
associations (or architectural elements) form the basis for a
descriptive subdivision of the stratigraphic record. The na- FIG. 14.Facies modeling, from Walker and James (1992). Note
ture of the bounding surfaces between associations/ele- the relationship between individual examples and their
ments is also importantthey may be gradational, or sharp distillation into a general model. Note how new examples
and/or erosional (Fig. 12). It is becoming apparent that some can be compared with the model (the norm), and then
facies, facies associations, and architectural elements are incorporated into the general data base (feedback). The
universal (i.e., they occur in many different places, in all parts model also serves as a guide for making observations and as
of the geological record). Examples include a) hummocky a predictor in new situations.
14 ROGER G. WALKER

FIG. 15.Block diagram of a barrier-island and lagoonal depositional environment, from Reinson (1992). Note that this is a snapshot
in time and that the various environments shown in the diagram may not be preserved if the barrier progrades or is transgressed.
For the transgressive setting, see Figure 16.

Washover

Superstructure
Foreshore
Nearshore
ridges

FIG. 16.Cross section of Cedar Beach, Long Island, New York. Barrierlagoon system rests disconformably on Pleistocene diamictite
(blue, conglomeratebreccia symbols). The barrier cuts down into the Pleistocene at a tidal inlet. Note the age dates of lagoonal
deposits, 7815 and 7130 years BP south of the barrier, 5055 years BP at Cedar beach, and 1015 and 300 years BP at the northern
edge of the lagoon. The sand of the barrier superstructure is not preserved during transgressionit is washed into the lagoon
(dark red) during storms (blue arrow) and is also moved offshore to form small ridges (red arrowsouthern end of cross section).
The record of barrier transgression is shown in the black rectangle, and consists of a thin sheet of lagoonal and washover facies
(blue, dark red, seaward of barrier) overlain disconformably by modern storm sands. The two black arrows at the right show two
erosion surfaces, one separating Pleistocene from Holocene sediments and one separating 7000+ year-old lagoonal and washover
facies from offshore sands forming today. From Rampino and Sanders (1970).
 FACIES MODELS REVISITED 15

7 nearshore ridges, lagoon, and washover. Each association will

have distinctive lithologies, sedimentary structures, and trace-
fossil assemblages. However, if the situation shown in Figure 15
CALICHE were to be preserved in the geological record, three of the associa-

MUDSTONES 2.86 m
NODULES tions would be closely related (the lagoon, washover, and near-

VERTICAL ACCRETION
6 shore ridges shown in the black rectangle in Fig. 16), and two
DESSICATION would be missing (barrier superstructure and foreshore). Signifi-
cantly, the lagoon and washover facies would lie on a major

FLOODPLAIN
CRACKS
transgressive surface of erosion, and the nearshore ridges would
be separated from the lagoon and washover facies by another
ROOT transgressive surface of erosion (Fig. 16). Thus the geological
5 TRACES interpretation would be based on (1) defining the facies associa-
tions present, (2) making a preliminary interpretation of the
associations, and (3) defining the relationships of the facies asso-
ciations and their bounding surfaces. Item 3 in this list is the crux
of the overall interpretation, because it involves building the
4 facies associations into a three-dimensional structure, indepen-
ALTERNATING dently of any preconceived facies model.
SSTS. AND MSTS.
1.00 m BUILDING MODELS
3 The ideas presented above suggest that we are moving away
RIPPLE from interpretation by reference to existing models. We are better
LATERAL ACCRETION
CROSS-LAM. able to recognize facies and architectural elements. We have more
0.95 m data expressing the complexity of modern environments. It there-
fore becomes less and less appropriate to use simple models to
2
POINT BAR

interpret complex geological situations.

TROUGH, P.T.
CROSS BED
1 1.71 m
LAG. 0.32 m
0 2. Carefully fit the elements into their 3-D framework. Which
METERS
FIG. 17.Fining-upward meandering-fluvial succession, com-
plied by Walker (1979) from data published by Allen (1970).
The average thicknesses shown for each facies are also com-
piled from Allens data.
form distinctive depositional systems in recent sediments. It
follows that the combination of architectural elements in ancient
sediments, and depositional systems in recent sediments, will
form the true basis for defining the building blocks of the sedi-
mentary record. As an example, I again refer to the shoreface,
where the architectural element is well defined in the geological
record and where there is a large body of data from recent
sediments. Questions remain, particularly concerning the con-
cept and definition of fair-weather wave base, and the preserv-
ability of many of the features seen in modern shorefaces. Never-
theless, the shoreface is an important example of a basic architec-
tural and depositional element, particularly in its role in defining
transgressions and regressions.
In many ancient examples from the geological record, there
may be no easy and direct comparison with existing facies mod-
els. It therefore becomes even more important to recognize indi-
vidual facies and facies associations (architectural elements), and
to determine which elements commonly occur together and
which never occur together. In Figure 15, at least five associations
could be recognized: barrier superstructure, barrier shoreface,
The solution is not to abandon models, which would result in
anarchy. The solution is to build your own interpretations, using
the following stages.
1. Recognize and define facies, facies associations, and archi-
tectural elements in the example you are studying. Some of
the elements may be universal, and some may be local to
your particular example.
ones occur together, and which are never found together?
Define the surfaces that separate the elements.
3. Attempt a preliminary interpretation of those elements that
allow it. Some will probably present features that have been
well studied and have agreed-upon interpretations (e.g.,
hummocky cross stratification), but other may be enigmatic
(e.g., thick structureless sandstones).
4. From whatever interpretation is possible, refer to the clos-
est existing model (the model as a norm). How does the
distribution of depositional elements in the model conform
to the distribution defined in your example? Lagoonal and
washover elements might suggest reference to a barrier-
island model, but the association of lagoonal, washover,
and nearshore-ridge elements defines one part of a dy-
namic barrier system rather than a barrierlagoon system
shown as a block diagram (e.g., Reinson, 1992, his fig. 3).
This approach might be even more important in submarine
fan systems, where there has been considerable work on
depositional elements but no simple fan models since the
1970s and early 1980s.
Models and Interpretations
The three-dimensional relationships of architectural ele-
ments that emerge from a new study of the hypothetical Beau-
fort Sea Formation add up to an interpretation of that formation,
16 ROGER G. WALKER
and NOT to a model for the Beaufort Sea Formation. There are
many papers in the literature that have titles along the lines of
The Beaufort Sea Formationa model for deposition in shal-
low marine environments. Presumably the idea is that if a
model is presented, a routine description of the Beaufort Sea
Formation may sound more interesting, and hence attract atten-
tion.
One formation may provide a superb case history, but one
formation does not make a model. The theme of facies modeling
is one of distilling many examples in the search for generality.
Small is beautiful because small is usually more homogeneous. I
therefore suggest that one important approach in the future will
be to define the pieces (modern and ancient) and to define the
relationships between the pieces. This is part of the distillation
process, but it begins with the pieces rather than with an initial
assumption of a geographically defined environment (delta,
whatever). Progress will truly be made when the geologically
defined pieces (ideally, the universally accepted architectural
elements) closely agree with the depositional elements defined
by modern processes.
ACKNOWLEDGMENTS
I thank Brian A. Zaitlin and Henry Posamentier for their
comments on the manuscript. I also thank the Natural Sciences
and Engineering Research Council of Canada for their support of
my research.
REFERENCES
ALLEN, J.R.L., 1970, Studies in fluviatile sedimentation: a comparison of
fining-upward cyclothems, with special reference to coarse member
composition and interpretation: Journal of Sedimentary Petrology, v.
40, p. 298323.
ALLEN, J.R.L., 1983, Studies in fluviatile sedimentation: bar complexes and
sandstone sheets (low sinuosity braided streams) in the Brownstones
(L. Devonian), Welsh Borders: Sedimentology, v. 33, p. 237293.
ANDERTON, R., 1985, Clastic facies models and facies analysis, in Brenchley,
P.J., and Williams, B.J.P., eds., Sedimentology: Recent Developments
and Applied Aspects: Oxford, U.K., Blackwell Scientific Publications,
p. 3147.
BOUMA, A.H., 1962, Sedimentology of Some Flysch Deposits: Amsterdam,
Elsevier, 169 p.
CANT, D.J., AND WALKER, R.G., 1976, Development of a braided fluvial
facies model for the Devonian Battery point Formation, Quebec:
Canadian Journal of Earth Sciences, v. 13, p. 102119.
COLLINSON, J.D., 1969, The sedimentology of the Grindslow Shales and the
Kinderscout Grit: a deltaic complex in the Namurian of northern
England: Journal of Sedimentary Petrology, v. 39, p. 194-221.
DE RAAF, J.F.M., READING, H.G., AND WALKER, R.G., 1965, Cyclic sedimenta-
tion in the lower Westphalian of North Devon, England: Sedimentol-
ogy, v. 4, p. 152.
GRESSLY, A., 1838, Observations gologiques sur le Jura Soleurois: Neue
Denksch. allg. schweiz., Ges. ges. Naturw., v. 2, p. 1112.
HARBAUGH, J.W., AND DEMIRMEN, F., 1964, Application of factor analysis to
petrologic variations of Americus Limestone (Lower Permian), Kan-
sas and Oklahoma: Kansas Geological Survey, Special Distribution
Publication 15, 50 p.
HARBAUGH, J.W., AND MERRIAM, D.F., 1968, Computer Applications in
Stratigraphic Analysis: New York, Wiley, 262 p.
HARMS, J.C., AND FAHNESTOCK, R.K., 1965, Stratification, bed forms and flow
phenomena (with example from the Rio Grande), in Middleton, G.V.,
ed., Primary Sedimentary Structures and Their Hydrodynamic In-
terpretation: Society of Economic Paleontologists and Mineralogists,
Special Publication 12, p. 84115.
HARPER, C.W., JR., 1984, Improved methods of facies sequence analysis, in
Walker, R.G., ed., Facies Models, Second Edition: Geological Associa-
tion of Canada, Reprint Series 1, p. 1113.
IMBRIE, J., AND PURDY, E.G., 1962, Classification of modern Bahamian
carbonate sediments, in Ham, W.E., ed., Classification of Carbonate
Rocks: American Association of Petroleum Geologists, Memoir 1, p.
253279.
KLOVAN, J.E., 1964, Facies analysis of the Redwater reef complex, Alberta,
Canada: Bulletin of Canadian Petroleum Geology, v. 12, p. 1100.
KOLLA, V., BOURGES, P., URRUTY, J.M., AND SAFA, P., 2001, Evolution of deep-
water Tertiary sinuous channels off shore Angola (west Africa) and
implications for reservoir architecture. American Association of Pe-
troleum Geologists, Bulletin, v. 85, p. 13731405.
MIALL, A.D., 1977, A review of the braided river depositional environ-
ment: Earth-Science Reviews, v. 13, p. 162.
MIALL, A.D., 1985, Architectural element analysis: a new method of facies
analysis applied to fluvial deposits: Earth-Science Reviews, v. 22, p.
261308.
MIDDLETON, G.V., 1973, Johannes Walthers law of the correlation of facies:
Geological Society of America, Bulletin, v. 84, p. 979988.
MIDDLETON, G.V., 1978, Facies, in Fairbridge, R.W., and Bourgeois, J., eds.,
Encyclopedia of Sedimentology: Stroudsburg, Pennsylvania, Dowden,
Hutchinson & Ross, p. 323325.
MUTTI, E., AND RICCI LUCCHI, F., 1972, Le torbiditi dellAppennino
settentrionale: introduzione allanalisi de facies: Societ Geologica
Italiana, Memorie, v. 11, p. 161199. English translation by T.H.
Nilsen, 1978, International Geology Review, v. 20, p. 125166.
POTTER, P.E., 1959, Facies models conference: Science, v. 129, p. 12721273.
RAMPINO, M.R., AND SANDERS, J.E., 1980, Holocene transgression in south-
central Long Island, New York: Journal of Sedimentary Petrology, v.
50, p. 10531079.
REINSON, G.E., 1992, Transgressive barrier island and estuarine systems, in
Walker, R.G., and James, N.P., eds., Facies Models: Geological Asso-
ciation of Canada, p. 179194.
SCHOLLE, P.A., BEBOUT, D.G., AND MOORE, C.H., 1983, Carbonate Deposi-
tional Environments: American Association of Petroleum Geologists,
Memoir 33, 708 p.
SCOTT, A.J., AND FISHER, W.L., 1969, Delta systems and deltaic deposition,
in Fisher, W.L., Brown, L.F., Scott, A.J., and McGowen, J.H., collo-
quium leaders, Delta Systems in the Exploration for Oil and Gas:
Texas Bureau of Economic Geology, Research Colloquium, various
pagination.
SHEPARD, F.P., PHLEGER, F.B., AND VAN ANDEL, T.H., 1960, Recent Sediments,
Northwest Gulf of Mexico: American Association of Petroleum Ge-
ologists, 394 p.
TEICHERT, C., 1958, Concepts of facies: American Association of Petroleum
Geologists, Bulletin, v. 42, p. 27182744.
VISSER, M.J., 1980, Neapspring cycles reflected in Holocene sub-tidal
large-scale bedform deposits: a preliminary note: Geology, v. 8, p.
543546,
WALKER, R.G., 1965, The origin and significance of the internal sedimen-
tary structures of turbidites. Yorkshire Geological Society, Proceed-
ings, v. 35, p. 132.
WALKER, R.G., 1975, Generalized facies models for resedimented con-
glomerates of turbidite association: Geological Society of America,
Bulletin, v. 86, p. 737748.
WALKER, R.G., ed., 1979, Facies Models: Geological Association of Canada,
211 p.
WALKER, R.G., 1983, Cardium Formation 3. Sedimentology and stratigra-
phy in the GarringtonCaroline area: Bulletin of Canadian Petroleum
Geology, v. 31, p. 213230.
WALKER, R.G., ed., 1984, Facies Models, Second Edition: Geological Asso-
ciation of Canada, 317 p.
WALKER, R.G., 1985, Mudstones and thin-bedded turbidites associated
with the Upper Cretaceous Wheeler Gorge conglomerates, Califor-
 FACIES MODELS REVISITED 17

nia: a possible channellevee complex: Journal of Sedimentary

Petrology, v. 55, p. 279290.
WALKER, R.G., AND JAMES, N.P., eds., 1992, Facies Models: Response to Sea
Level Change: Geological Association of Canada, 409 p.
WEIMER, P., 1989, Sequence stratigraphy of the Mississippi Fan (Plio-
Pleistocene), Gulf of Mexico: Geo-Marine Letters, v. 9, p. 185272.
WILLIAMS, P.F., AND RUST, B.R., 1969, The sedimentology of a braided river:
Journal of Sedimentary Petrology, v. 39, p. 646679.
WILSON, J.E., 1975, Carbonate Facies in Geologic History: New York,
Springer-Verlag, 471 p.
398 JOHN R. SUTER
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 399

DEEP-WATER TURBIDITES AND SUBMARINE FANS

HENRY W. POSAMENTIER
Anadarko Petroleum Corporation, 1201 Lake Robbins Drive, The Woodlands, Texas 77380, U.S.A.
e-mail: henry_posamentier@anadarko.com
AND
ROGER G. WALKER
Roger Walker Consulting Inc., 83 Scimitar View NW, Calgary, Alberta T3L 2B4, Canada
e-mail: walkerrg@telus.net

Abstract: Depositional environments of deep-water deposits commonly are complex and consequently do not neatly fit any single facies
model. Rather than developing specific models we discuss these deposits within the context of depositional elements and first principles
of process sedimentology. Depositional elements are described using 3D seismic as well as outcrop data. Detailed facies descriptions from
outcrops are then integrated with these depositional elements. Following the theme of this publication, we emphasize facies and
depositional environments rather than the mechanics of turbidity currents and related processes.
The spatial and temporal distribution of depositional elements is determined largely by characteristics of the shelf-edge staging area.
Such factors as grain-size distribution, sediment caliber, frequency of flow events, and magnitude of flows are all a function of conditions
at the shelf edge and upper slope. Sediments are supplied from the staging area to the slope and basin floor beyond. Turbidity currents
traverse the slope through canyons and slope channels. When these flows reach the basin floor they continue to remain confined by levees
for a certain distance. This distance is a function of grain-size distribution in the flow, flow magnitude, and flow velocity. Levee height
diminishes seaward, and eventually where levees can no longer effectively confine the basal sand-rich part of the flow the leveed channel
transitions into a frontal splay or lobe.
Relative sea-level change plays an important role in turbidite deposition, in that sea level is a major factor controlling conditions in
the outer shelf and upper slope. During relative sea-level lowstands, shorelines and consequently depocenters tend to be located at the
shelf edge. This sets up conditions favorable for delivery of sediments to the slope and basin floor. Conversely, relative sea-level
highstands commonly are associated with depocenters at the inner to middle shelf, resulting in a paucity of coarse sediments being
actively delivered to the shelf edge and ultimately to the slope and basin floor. Variations in grain size delivered to the shelf edge during
a cycle of sea-level change can vary predictably hence the temporal and spatial distribution of depositional elements in linked deep-
water environments can likewise be better understood within this context.

INTRODUCTION
The scope of turbidite and submarine-fan facies models is
vast, extending from individual beds a few centimeters thick to
entire submarine fans with volumes up to a million cubic kilome-
ters or more (for example, Indus fan area 1.1 x 106 km2, thickness
3+ km, hence volume of the order of 3 x 106 km3). The unifying
theme is the central role played by individual turbidity currents,
where each bed (a turbidite) is the result of a relatively short-lived
depositional event. The environment is consistently below storm
wave base, such that, once deposited, a turbidite is unlikely to be
reworked by other currents aside from the occasional strong
contour current. Figure 1 schematically illustrates an idealized
shelf to basin-floor physiography displaying most of the key
elements of the deep-water depositional environment.
We will briefly examine the history of turbidite and subma-
rine fan models and show that, perhaps more than in any other
depositional environment, technology (2-D and 3-D seismic data)
has influenced the definition of depositional elements and hence
the facies models (e.g., Posamentier and Kolla, 2003a). No single
model comes close to embracing the complexity of huge ancient
and modern submarine-fan systems, making the depositional
elements and their lateral and vertical relationships the basis for
interpretation and prediction. Our treatment of facies models will
take a first-principles approach that will focus on the linkage
between physical processes and associated depositional elements.
Many aspects of turbidity-current generation, movement,
and deposition were reviewed by Walker (1992) in the third
edition of Facies Models. This work will be referenced here but
not repeated in detail.
HISTORY OF FAN MODELS
The turbidity-current concept was introduced in 1950, in the
classic paper Turbidity currents as a cause of graded bedding
by Kuenen and Migliorini (1950). The paper was based mainly on
Kuenens experimental work both before and after the Second
World War. The idea that sand could be transported to great
depths in the ocean was very controversial at the time (Walker,
1973), and for many years there was considerable debate about
the existence of turbidites and, about their properties. It was
understood that modern fans existed, but their internal character-
istics were completely unknownindeed, Kuenens experiments
were much more concerned with the origin of submarine canyons
than the transport of sand onto the deep sea floor.
After a dozen years of observations, the first generalization
concerning turbidites was published by Bouma, (1962) (Fig. 2)
what is now known as the Bouma sequence for the internal
structures in individual turbidites. The sequence from Division A
(generally structureless) to division B (parallel lamination in
sand) and division C (ripple cross-lamination) was compared
with flume experiments and interpreted to represent waning
flow (Harms and Fahnestock, 1965; Walker, 1965). Division D
consists of thin laminae of silt and clay, and Division E is pelitic,
probably largely turbidity-current mud with a small proportion
of hemipelagic mud

Facies Models Revisited

SEPM Special Publication No. 84, Copyright 2006
SEPM (Society for Sedimentary Geology), ISBN 1-56576-121-9, p. 397520.
400 HENRY W. POSAMENTIER AND ROGER G. WALKER

Inner- to Mid-Shelf (Highstand) Delta

Shelf-Edge (Lowstand) Delta

Slump Scars

Canyon/Slope Channel

Staging Area Avulsion Node

Mass-Transport
Complex

Leveed
Channel
Crevasse
Splay

Frontal Splay
Sediment
Waves
Oxbow

FIG. 1.Schematic representation of shelf to deep-water physiography. The shelf staging area is connected to the deep-water
environment through slope channels and/or canyons. Depositional elements in the deep water include leveed channels, crevasse
splays, sediment waves, and frontal splays or lobes. (modified after Posamentier and Kolla, 2003a).

FIG. 2.The Bouma (1962) sequence for classical turbidites. Division D is placed in brackets because it is difficult to identify in
weathered or tectonized outcrops. Division E can be subdivided into two parts: turbidite mud E(t) and hemipelagic mud E(h).
In most beds, the turbidite mud predominates.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
Gradually, as more information became available on modern
fans, the first model was proposed by Normark (1970) in another
classic paperGrowth patterns of deep sea fans. The data base
was small, and the paper concentrated on the La Jolla and San
Lucas fans. The model showed leveed channels on the upper fan,
depositional (suprafan) lobes on the middle fan, and a smooth
surface on the lower fan. This model was based largely on
shallow-penetration seismic data.
Shortly afterward, Emiliano Mutti and colleagues proposed
fan models based exclusively on observations of ancient rocks.
Mutti and Ricci Lucchi (1972) proposed a model with an inner-fan
channel that branched into multiple channels on the mid-fan (but
without depositional lobes). In the same year, Mutti and Ghibaudo
(1972) showed a similar model but with lobes at the ends of the
channels. Normarks work was not cited, suggesting that the
modern and ancient fan models were derived independently
the proposed models suggested that modern fans and ancient
rocks behaved in very similar ways.
The channel-feeding-lobe models dominated turbidite stud-
ies for about 10 years (19701980). The literature on modern fans
and ancient rocks was formally brought together into the model
proposed by Walker (1978); the models proved to be popular but
also attracted considerable discussion (e.g., Nilsen, 1980). In
retrospect, the models clearly had severe limitationsthe distri-
bution of sand and mud on the fans was incorrect, and no
consideration was given to the influence of grain size or of local
and regional tectonics. Perhaps more importantly, the models did
not incorporate the influences of relative sea-level fluctuation.
With the advent of sequence stratigraphy, the fan models of
the 70s were updated first by Vail et al. (1977) and later by Mutti
(1985), Posamentier et al. (1988), and Posamentier et al. (1991),
who integrated the effects of relative sea-level fluctuation with
the channel-feeding-lobe models. This can be regarded as a
period of transition between the older, field-based models and
the rapidly evolving seismic-based modelsthe technology of
marine geology was overtaking the efforts of field geologists.
In 1982, the first side-scan sonar images of the Amazon Fan
were published by Damuth et al. (1982a), and Damuth et al.
(1982b). The presence of long, narrow, and sinuous channels
surprised most turbidite workers, as did the scale of the channel
levee complexes, which stood at least 200 m above the adjacent
fan surface.
However, 1985 can be considered the year in which the
emphasis shifted significantly from ancient rocks to large-scale
studies of modern fans. In that year, the first compilation of
modern fan studies was published (Bouma et al., 1985), with
discussion of the Amazon, Astoria, Bengal, Cap-Ferrat, Crati,
Delgada, Ebro, Indus, La Jolla, Laurentian, Magdalena, Missis-
sippi, Monterey, Navy, Rhne, and Wilmington fans. That vol-
ume also had a very useful fold-out that tabulated the quantita-
tive descriptors of the fans (channel dimensions and slopes, fan
sizes, and fan volumes).
Also that year, Droz and Bellaiche (1985) published a seismic
study of the Rhne Fan, showing the existence of meandering
channels, channellevee systems, and the lateral shifting and
stacking of these systems to make channellevee complexes.
They also showed large slump masses (acoustically transparent
units) up to 160 milliseconds thick that represented both failure
on the slope above the fan and failure of the back of the channel
levees.
The studies in the Bouma et al. (1985) compilation essentially
changed the direction and style of turbidite research, focusing on
modern fans rather than ancient-rock studies. Droz and Bellaiche
(1985), without using the term, essentially introduced the idea of
depositional elements. This approach was also applied by Weimer
401
(1989) in a classic study of the Mississippi Fan. Weimer (1991)
recognized a succession of seismic facies based on reflection
patterns (subparallel, wavy, hummocky, divergent, mounded,
and convergent) which could be interpreted in terms of mass-
transport complexes (hummocky, mounded), channel fill (dis-
continuous, high amplitude), levee (subparallel to convergent),
and basin floor (subparallel to parallel). Work up to 1991 was
collected in the volume Seismic Facies and Sedimentary Pro-
cesses of Submarine Fans and Turbidite Systems (Weimer and
Link, 1991). The seismic evidence presented was mostly high-
quality 2-D data. In that volume, Mutti and Normark (1991) first
systematized the depositional-elements approach. They defined
depositional elements as the basic mappable components of both
modern and ancient turbidite systems and stages that can be
recognized in marine, outcrop, and subsurface studies. These
features are the building blocks of fan models.
The past ten years have seen an explosion in 3-D seismic
studies, with a corresponding better understanding of deposi-
tional elements (e.g., Posamentier and Kolla, 2003a). The best
sources of this information are in the proceedings volume of a
Gulf Coast Section of SEPM research conference (Deep Water
Reservoirs of the World, Weimer et al., 2000) and a thematic
compilation of papers on deep-water systems in Marine and
Petroleum Geology (Mutti et al., 2003). Various classifications of
depositional elements were suggested, but no attempt was made
to formulate a general model for submarine fans.
ORIGIN OF TURBIDITY CURRENTS
Density currents flow downslope as gravity acts on the den-
sity difference between the flow and the ambient seawater (Fig.
3). The density difference can be due to any or all of the following:
the increased salinity of the flow, the cold temperature of the
flow, and the suspended sediment within the flow. A turbidity
current is a special case of a density flow, where the increased
density is due to sediment maintained in turbulent suspension
within the flow. The turbulence is maintained by the downslope
movement of the flow.
Turbidity currents can originate by two mechanisms: some
begin with large sediment slumps that accelerate and become
turbulent. Many of these slumps are triggered by earthquakes,
the most famous being the Grand Banks (Newfoundland)
earthquake, slump, and flow of 1929. The flow broke a series of
submarine telegraph cables, and reconstructions of the flow
mechanics (Piper et al., 1988) suggest flow velocities up to 20
m/s, flow thicknesses of several hundred meters, and a mini-
mum flow volume of 175 km3. The flow bypassed the entire
Laurentian Submarine Fan, and the deposit, in places over 1 m
thick, now covers a large part of the Sohm Abyssal Plain
(Walker, 1992).
Similarly, slumps off the delta of the Magdalena River in
Colombia have broken telegraph cables up to 100 km from the
delta. In the period 19321955, there have been 15 cable breaks,
averaging one every 1.5 years (Heezen, 1956). The flow of 1935
had an estimated volume of sand of 3 x 108 m3 (Heezen, 1956).
Turbidity currents off the fronts of major deltas may be large and
frequent.
In the case of the Congo (Zaire) river, where there is no delta,
a submarine canyon has its head within the estuary of the Congo
River. At times of peak river discharge in DecemberJanuary and
AprilMay, and during the years when the river is establishing a
new course among the estuarine sand bars (18921903 and 1925
1929), submarine cables have been broken seaward of the estuary
within the Congo Canyon (Heezen et al., 1964). These cables lay
close to the estuary, at the shelf edge, and in water depths as great
402 HENRY W. POSAMENTIER AND ROGER G. WALKER
HORIZONTAL
FIG. 3.Experimental turbidity current in a flume. Water depth is 28 cm. Note characteristic shape of the head and eddies behind the
head. Sediment is thrown out of the main flow by these eddies, and the body of the flow is about half the height of the head.
Experiment conducted by G.V. Middleton at Caltech.
as 2800 m, suggesting that sand swept into the canyon head from
the estuary continued down the canyon in flows powerful enough
to break cables in abyssal depths. Despite todays relative high
stand of sea level, there were 26 cable breaks between 1893 and
1937an average of one every 1.7 years.
Turbidity currents can also originate with delivery of river
flow charged with sediment directly onto the slope. During times
of river flood enough sediment can be entrained in the flow in
some instances to produce a mix that has greater density than sea
water, resulting in hyperpycnal flow down the slope (Mulder and
Syvitski, 1995; Mulder et al., 1998). With this mechanism, what
begins as inertial flow at the river mouth transforms into density
underflow and ultimately turbidity flow on the slope. Such flows
generally are of greater duration (i.e., days or weeks) than those
that originate from large sediment slumps (i.e., hours).
FAN BYPASSING AND DEPOSITION
ON MODERN ABYSSAL PLAINS
Turbidity currents traveling downslope may be moving at
several to many meters per second, at which velocities all of the
sand and finer sizes are in turbulent suspension. The flows
gradually decelerate to velocities of 12 m per second, when the
coarser sand fraction begins to be deposited from suspension.
During this period of deceleration, the flows may largely bypass
the slope and move long distances across the basin floor.
There are many studies of abyssal-plain deposition (Pilkey,
1988). The Grand Banks flow bypassed the Laurentian Fan at the
base of the slope and deposited a turbidite on the Sohm Abyssal
Plain, as discussed above. About 16,000 years ago, an even
larger flow bypassed the Hatteras Fan and deposited on the
Hatteras Abyssal Plain (western North Atlantic Ocean) (Elmore
et al., 1979). Deposition began about 120 km from the end of the
Hatteras Canyon system. This Black Shell turbidite (named
for the distinctive corroded shells contained in the deposit)
covers 44,000 km2 of Hatteras Abyssal Plain in a bed up to 4 m
thick, 500 km long, 200 km wide. The volume of the deposit is
between 100 and 200 km3. Characteristics of that deposit (grain
SLOPE
size, sorting, mud content, etc.) were discussed in detail by
Elmore et al. (1979).
One of the longest documented bypass systems is the Cascadia
Deep-Sea Channel (Nelson et al., 2000). The channel originates off
the coast of Washington, continues around the outer part of
Astoria Fan, cuts through the Blanco Fracture Zone, and ends on
the Tufts Abyssal plain. The turbidity-current pathway [traverses]
1000 km of Cascadia Basin and remained open throughout the
late Quaternary as shown by the presence of the 13 post-MA
(Mazama Ash, 7530 YBP) turbidites throughout the pathway in
all recent cores we have collected (Griggs and Kulm, 1968; see
also Nelson et al., 2000). Beds within the channel include thick (2
m) Pleistocene graded gravel-to-sand beds over 400 km from the
heads of the channel at the Washington coast. These long dis-
tances of bypass have significant implications regarding the
location of sand deposits in ancient basins, as discussed through-
out this review.
Kneller (1995) described the effects of waxing and waning
flows within individual events. Waxing flows, commonly at or
near the head of a turbulent flow, erode the substrate over which
they pass. Significant amounts of sediment can bypass the system
during this time. As the flow wanes, coarser sediments tend to
come out of suspension and be deposited in the area formerly
characterized as a zone of bypass. Consequently, even in slope
and proximal basin-floor areas, where sediment bypass and
erosion during waxing phase may be common, some sedimenta-
tion in the form of lag deposits almost always occurs.
TURBIDITE FACIESTHE BUILDING BLOCKS
There are several schemes for classifying the family of rocks
that occur in deep-water settings. The first was proposed by Mutti
and Ricchi Lucchi (1972) and was later simplified by Walker in
1978. Subsequent facies classifications have become more com-
plex, including the all-inclusive but unwieldy schemes of
Ghibaudo and Vanz (1987) and Pickering et al. (1986).
A detailed subdivision of features within individual beds was
proposed by Lowe (1982), based on interpretations of how sedi-
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 403

ment was deposited from sandy and gravelly high-density tur-
bidity currents. For sandy flows, division S1 is characterized by
traction structures, division S2 contains thin horizontal layers
showing inverse grading and basal shear laminations and divi-
sion S3 may be structureless or normally graded and it com-
monly contains water escape features. For gravelly flows, divi-
sion R1 consists of coarse gravel with traction structures, division
R2 consists of an inversely graded gravel layer, and division R3
consists of a normally graded gravel layer. Lowes (1982) scheme
is akin to a Bouma sequence (see below) for individual coarse
beds, rather than a facies classification of coarse-grained beds.
Because it is genetically based, the scheme may change as more is
learned about the flow and depositional mechanics of high-
density flow events.
In this review, we suggest that deep-water rocks contain a
variety of depositional elements (discussed below) and that these
elements contain distinctive assemblages of facies. We have
chosen to use the simple scheme of Walker (1978), which is
descriptive (except for the various deformed facies), and based on
grain size. The categories included in this scheme are (1) classical
turbidites, (2) structureless sandstones, (3) pebbly sandstones, (4)
conglomerates, and (5) various types of deformed rocks. We are
more concerned with the descriptive and environmental aspect
of the facies than with the mechanics of flow and deposition
(which are very difficult to study in flumes a few meters long
when in nature flows may change and evolve over distances of
hundreds of kilometers).
Classical Turbidites
This category includes all of those rocks originally considered
as turbidites in the 1950s and 1960sthe beds that give rise to
little or no controversy today. The facies includes thick monoto-
nous successions of alternating sandstones and mudstones (Fig.
4). The sandstones have sharp, flat bases, and the only erosional
features are normally on the centimeter scale. They include scour
marks (commonly flute casts) and tool marks (commonly groove
casts). Channeling on a scale greater than a meter is very uncom-
mon.
Internally, classical turbidites contain some or all of the
divisions first proposed by Bouma (1962) (Fig. 2). Division A
implies rapid deposition producing structureless sandstone in
the absence of any equilibrium bedforms, whereas divisions B
and C imply traction of grains on the bed to form parallel
lamination and ripple cross lamination, respectively (a waning-
flow successionHarms and Fahnestock, 1965; Walker, 1965).
Divisions D and E both imply deposition of fine-grained mate-
rial from suspension without traction on the bed. Note that
Bouma (1962) observed that division D (laminations of silt and
mud) was difficult to recognize in weathered or tectonized

TOP

FIG. 4.Alternating beds of sandstone and mudstone, Devonian, Cape Liptrap, South Australia. Note the monotonous alternation
of sandstones and mudstones, and the very parallel nature of the bedding with no evidence of any topography on the sea floor.
404 HENRY W. POSAMENTIER AND ROGER G. WALKER
outcropsconsequently it may not be a useful or significant
part of the Bouma sequence.
Two sub-categories of classical turbidites have been sug-
gested by several workers: thin-bedded and thick-bedded. It
must be emphasized that there is a complete spectrum of bed
thicknesses and that their separation is arbitrary. The thick-
bedded turbidites (sandstones roughly in the range 10100 cm in
thickness; Fig. 5) tend to be composed of Boumas division A,
with fewer beds also containing divisions B and C (Walker, 1968).
Thin-bedded turbidites (Fig. 4) tend to lack Boumas division A,
and the sandstones contain only the BC or C divisions. The
nature of these divisions suggests two distinct types of thin-
bedded turbidites.
The simplest type of thin-bedded turbidite contains a single
set of ripple cross lamination, with or without division B parallel
lamination underneath (Fig. 6). In more complex thin-bedded
turbidites, the ripples in division C consist of climbing sets rather
than single sets, and the ripple cross lamination (and the parallel
lamination beneath) can be convoluted (Fig. 7). These complex
thin-bedded turbidites also commonly contain ripped-up mud-
stone clasts, and they have been termed CCC turbidites (Fig.
FIG. 5.Thick-bedded sandstones consisting mainly of Boumas
division A, separated by very thin siltstone partings. Height
of cliff about 180 m. Compare with Figure 4. Annot Sandstone
(Eocene), southern France.
8A; Walker, 1985); C for climbing, C for convolution and C for
clasts.
The presence of convolute lamination implies rapid deposi-
tion of sediment and trapping of pore fluid, such that the primary
structures are easily deformed. The climbing ripples imply depo-
sition of sediment from suspension while the ripples are moving
on the bed. Within the category of thin-bedded turbidites, the
simple beds imply traction on the bed and essentially no deposi-
tion from suspension, whereas the CCC turbidites suggest high
rates of deposition from suspension during formation of the
primary bedforms. It has been suggested that thin beds showing
high rates of deposition commonly form on levees. The thin beds
that show little evidence of rapid deposition from suspension
may indicate basin-plain settings, where the turbidity currents
have much less sediment left in suspension (Fig. 8B; Walker,
1985). The presence of ripped-up mudstone clasts supports this
interpretationthere is more likelihood of erosion associated
with confined flows in channels than in unconfined settings on a
distal basin plain.
Structureless Sandstones
There is an association in facies between classical turbidites
and structureless sandstones. Individual structureless sandstone
beds tend to be thicker (several tens of centimeters to a few
meters) than the sandstones in classical turbidites, and mudstone
partings between beds tend to be thin (centimeters) or absent (Fig.
5). The deposits of several flows may be amalgamated, the
amalgamation planes being denoted by (1) abrupt changes in
grain size, (2) layers of ripped-up mudstone clasts, or (3) simply
the disappearance of thin mudstone partings (Fig. 9). On a larger
scale, scouring on the scale of meters is commonly observed in
this facies (Fig. 9). It follows that the monotonous interbedding of
sandstones and mudstones, typical of classical turbidites, does
not occur in structureless sandstones. Stacks of amalgamated
beds without mudstone partings can be as much as 200 m thick,
as in the Annot Sandstone (Fig. 5).
Parallel lamination and ripple cross lamination are rare, and
the term structureless (now preferred to the older term mas-
sive) denotes this absence of primary sedimentary structures.
Graded bedding is present in some beds and not in others; its
presence may be largely a function of the range of grain sizes
available in the flow.
Although most beds lack primary structures, secondary
structures indicating dewatering during compaction of the bed
are common (Lowe, 1975). These include vertical or subvertical
fluid-escape pipes (Fig. 10), which can become contorted if the
bed is sheared by continuing turbidity-current flow during the
fluid escape (Fig. 11). If the escaping water encounters a crude,
incipient parallel lamination with variations in permeability,
the water may be forced to flow horizontally until able to break
through the less permeable layers and continue its vertical
escape. The curved upward edges of these laminates take the
shape of an irregular stack of dishes, hence the term dish
structure (Figs. 10, 12).
The association of this facies with classical turbidites suggests
that individual structureless sandstones are also the deposits of
turbidity currents. This interpretation is strengthened by the
presence of fluid-escape features, which indicate initial deposi-
tion of a fluid-rich sedimentwater mixture (rather than a more
rigid plug flow with grain-to-grain contacts and much less inter-
stitial water). Despite the thickness of individual beds and the
general absence of Bouma sequences, there is no compelling
observational or experimental evidence to reject turbidity cur-
rents in favor of speculative processes such as fluxoturbidity
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 405

FIG. 6.Thin-bedded turbidites beginning with Bouma divisions B and C. Sharp bases shown by yellow arrows, parallel lamination
by a blue arrow, and ripple cross lamination by red arrows. Note the absence of convolute lamination, climbing ripples, and
ripped-up mudstone clasts. Ordovician turbidites at Chutes Montmorency, Quebec.

CONVOLUTE
LAMINATION

CLIMBING RIPPLES

FIG. 7.Thin-bedded turbidites in the Chatsworth Sandstone (Cretaceous), Chatsworth (Simi Hills), California. Bases shown by red
arrows, and climbing ripples shown by yellow arrows. Convolute lamination is outlined in blue. Compare with Figure 6 (where
there is no climbing and no convolution).
406 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 8.Diagram showing two types of thin-bedded turbidites. One is characterized by single rows of ripple cross lamination without
climbing, and the other is characterized by climbing ripples, convolute lamination, and ripped-up mudstone clasts A). The CCC
turbidites are interpreted as levee deposits (see text), and the others as distal basin plain deposits B) (From Walker, 1985).

TOP

STRUCTURELESS SANDSTONES

CLASSICAL
TURBIDITES

SCOUR

AMALGAMATION

FIG. 9.Devonian turbidites in Germany. Note the classical turbidites (right) and the underlying thick-bedded structureless
sandstones. Yellow arrows show thin mudstone partings that disappear along strike (amalgamation), and the red arrow shows
a small scour.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
A)
A) DISH
DISH STRUCTURES
TOP
FIG. 10.Vertical fluid-escape pipes with overlying dish structures (outlined in yellow).
Ordovician Cap Enrage Formation, Gaspesie, Quebec.
currents (Dzulynski et al., 1959) or sandy debris flows
(Shanmugam et al., 1994; Shanmugam, 1996).
Pebbly Sandstones
As the coarse fraction within flows gradually increases, the
structureless sandstone facies grades through granule sandstones
into the pebbly sandstone facies. Graded bedding is common (Fig.
13) and readily observed because of the wide range of sizes present.
Internally, beds may show a crude horizontal stratification, and, in
rare cases, planar tabular and trough cross bedding may be present.
In the Cap Enrage Formation (Quebec), the trough sets are up to 50
cm thick, and trough widths seen in plan view are up to at least two
meters. Apart from structureless sandstone, the elements of the
Bouma sequence do not occur in pebbly sandstones, and hence the
Bouma sequence cannot be used as a descriptor in this facies
If blade-shaped or disc-shaped pebbles are present, they are
commonly well imbricated (Fig. 14). The features described
abovegraded bedding, cross bedding, and imbricationall
suggest turbulent flows in which grains are free to move relative
to one another, enabling the development of these features.
Making reasonable estimates of the turbulence of the flow and
particle settling velocities, it appears that a flow moving at 6 m/
s (the Grand Banks flow of 1929 near the toe of Laurentian Fan;
Uchupi and Austin, 1979) could suspend by fluid turbulence
alone clasts up to 2 or 3 cm in diameter. It therefore appears that
pebbles can be transported into deep water by turbidity currents
(flow velocities of 6 m/s or greater), and that such flows could
deposit graded, imbricated, and/or cross-bedded beds. In these
instances, despite the coarse nature of the beds, it is again not
necessary to appeal to alternative transport processes such as
fluxoturbidity currents and sandy debris flows.
Conglomerates
Conglomerates are not as common as the facies described
above, but they do make up an important part of the deep-water
407
STRUCTURES
VERTICAL
VERTICAL PIPES
PIPES
F IG . 11.Distorted
fluid-escape pipes
in a core from the
Cretaceous Lysing
Formation, off-
shore mid-Nor-
way. Well 6506/
12-4, 3240.6 m
depth. 5 cm
sedimentary record. Four distinct facies were recognized by
Walker (1975a), but the classification is based on a relatively small
sample and does not have the authority of the Bouma sequence
for classical turbidites. The features used to define the facies are
(1) the style of grading (normal or inverse), (2) the type of
stratification, and (3) the fabric. In combination, these features
define the four facies.
The first consists of beds which are normally graded and pass
upward into finer-grained stratified pebbly sandstone (Fig. 15).
The second consists of beds that show only normal grading (Fig.
16) without a stratified component. The third consists of beds that
begin with inverse grading and pass upward into normally
graded beds (Fig. 17). Finally, the fourth facies lacks any of these
features and is described as disorganized or structureless (Fig.
18).
The first three facies may also display clast imbrication (Figs.
14, 17). In the stratified parts of the graded-stratified facies,
clasts lie with their long axes transverse to flow, and the short
axis dips upstream. In the graded and inversely graded parts of
408 HENRY W. POSAMENTIER AND ROGER G. WALKER

3255.25 m 3255.44 m

FIG. 12.Dish structures out-

DISH
DISH STRUCTURE
STRUCTURE lined in yellow from the
Agat 35/3-4well, 3255.5
m, offshore Norway. Note
darker (less permeable)
layer at base of each dish,
and a few fluid-escape
pipes red arrows) where
fluid has broken through
vertically.

AGAT. 35/3-4

5 cm

the beds, however, clasts more commonly lie with their long
axes parallel to flow, with the long axes dipping upstream. This
fabric suggests that the clasts have not been rolling on the bed
(if clasts roll, long axes tend to be transverse to flow). A full
discussion of conglomerate fabrics has been given by Walker
(1975a).
Bed thicknesses in the conglomerate facies are very variable.
Individual graded beds can be over 10 m thick (Fig. 16), but
alternatively, beds may be only one or two pebble diameters in
thickness. The example from Point Lobos, California, in Figure
19, shows several thin conglomerate layers alternating with
sandstone layers. The pebbles probably did not constitute the
bulk of the flow, because they would not have given a sufficient
density contrast with the surrounding fluid to drive the flow. It
is more likely that the flows that transported the pebbles were
large sandy turbidity currents, and that the thin pebble beds
represent lags left behind by the main flows.
Despite the suggestions made above, interpretations of con-
glomerate facies remain somewhat speculative because of the
lack of large-scale experimental work.

EXOTIC FACIESOTHER TYPES OF DEPOSITS

IN DEEP-WATER ENVIRONMENTS

This category contains a variety of facies that do not fit into the
four facies described above. They are generally characterized by
poor sorting and lack of coherent bedding features. Some of the
main types are described below.

FIG. 13.Pebbly sandstone about 1 m thick showing overall
graded bedding. Annot Sandstone at Chambre du Roi, south-
ern France. Fixed eyebolts for rock climbers are circled for
scale.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 409

FLOW

F IG . 14.Graded con-
glomerate showing
clast imbrication (cen-
ter of bed above note-
book). Bed rests on de-
formed and slumped
mudstones. Tourma-
line State Surfing
Beach, north of San
Diego, California.

Pebbly Mudstones

Pebbly mudstones (Fig. 20) consist of granules and pebbles,

commonly along with distorted sandstone clasts, all embedded
in a deformed mudstone matrix. The term was introduced by
Crowell (1957), who suggested two mechanisms for their origin.
The first was emplacement by debris flows, wherein the strength
of the muddy matrix prevents the settling of the larger pebbles
and clasts during transport. Some debris flows may slide rapidly
on a basal layer of lubricating fluid (sea water) (Mohrig et al.,
1988), though it is likely that at least part of the moving mass
clearly is in contact with and erodes the substrate over which they
pass (Fig. 21; Posamentier and Kolla, 2003a) suggesting a more
complex rheology. As the flow velocity increases, there is a
tendency for sediment to be suspended at the head as well as the
upper parts of the flow. Rapidly moving debris flows therefore
may tend to transform at least in part into turbidity currents. The
transformation may be quite slow for muddy flows, but for sandy
debris flows moving at more than 1 to 2 m/s (the velocity at which
sand is carried in suspension) the transformation into a turbulent
turbidity current may be rapid, and take place over a short
distance.
A second mechanism for depositing pebbly mudstones
(Crowell, 1957) involves the passage of a sandy/pebbly turbidity
current over a bed of fluid-saturated, uncompacted mud. The
coarser material from the flow may be deposited on the muddy
surface and then quickly sinks into the uncompacted mud. The
pebblesandmud mixture may then flow for a short distance as
it dewaters, mixing the various grain sizes and then depositing an
ungraded, poorly sorted pebbly mudstone.

FIG. 15.Graded conglomerate in the Cap Enrage Formation,

Ordovician, Quebec. Note the large carbonate blocks in the
base of the bed, and the gradation into structureless pebbly
sandstone, stratified pebbly sandstone, and finally structure-
less sandstone. Bed is at least 8 m thick.
410 HENRY W. POSAMENTIER AND ROGER G. WALKER

BEDDING
BEDDING

TOP

FIG. 16.Graded conglomerate about 14 m thick, Eocene of Oregon. Close field examination showed a progressive decrease in
maximum and estimated mean grain size throughout the bed, and no internal planes or grain size changes that might suggest
an amalgamated bed.

INVERSE
INVERSE GRADED
GRADED
BEDDING
BEDDING

BASE
BASE

FIG. 17.Inversely to normally graded conglomerate, Cretaceous La Jolla Formation, California. Note also the well developed
imbrication with clasts dipping upstream (to the right).
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 411

TOP

STRUCTURELESS

FIG. 18.Conglomerate in the Ordovician Cap Enrage Formation, Quebec, showing no grading and no imbrication. The bed is
described as disorganized or structureless.

SANDSTONE
SANDSTONE LAYERS
LAYERS

TOP

FIG. 19.Thin conglomerate horizons separated by sandstone layers (shown by yellow arrows). Cretaceous, Point Lobos,
California.
412 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 20.Pebbly mudstones from Pigeon Point, California. This is the classic location where Crowell (1957) first discussed the
origin of pebbly mudstones. Note clasts scattered throughout the muddy matrix, along with rolled-up sandstone beds (above
notebook).

Slumps Slumps with Stratified Blocks

Slumps (Figs. 22, 23) comprise a large category of variously
deformed sediments. The term in its most general sense de-
scribes chaotic unbedded units meters to tens of meters in
thickness. The lithology may consist only of mudstone (Fig.
22), or it may involve pulled-apart or rolled-up sandstone beds
in a matrix of mudstone. In all cases, deposition was probably
fairly rapid, and mudstones form an important part of the
facies. Transport distances vary from a few meters (e.g., col-
lapse of channel walls) to hundreds of kilometers across basin
floors. Original depositional conditions may have involved
rapid deposition of sandstone with trapped pore water, fol-
lowed by mudstone deposition and sealing of the pore fluid.
With continued deposition, the lithostatic load would in-
crease, but, if the pore fluid could not escape, the fluid pressure
would also increase. Slumping would be initiated if beds fail
along a weak layer with high pore pressure. The sediment may
move a few meters during dewatering and deposit a unit that
consequently would be identified as a slump. Alternatively, if
the sediment moved a greater distance the resulting deposit
may be identified as a debris flow. Some slumps may move fast
enough so that much of the sediment is taken into suspension
with turbulence characterizing the flow, transforming the
mass into a turbidity current.
Slumps with stratified blocks (Fig. 24) are not uncommon. The
stratified blocks may be meters in diameter, and they consist of
interbedded layers of sandstone and mudstone (perhaps origi-
nally classical turbidites). In most cases the blocks were probably
not lithified, so that transport in a turbulent medium would
probably result in splitting of the blocks along the cohesionless
sandstone layers. This would destroy the stratified blocks and
probably would give rise to large mudstone clasts. One sugges-
tion is that the blocks were derived from an undermined, col-
lapsed channel wall, where the blocks subsequently were buried
by turbidity-current sediment before they could be transported
downchannel. Later in this paper, examples will be shown of
large blocks that have been rafted on top of mass-transport
complexes. Such flows are not turbulent, and the rafted blocks
may retain some stratification.
Slumps involving only one or two beds (Fig. 25) are fairly
common. The deposit is characterized by undeformed bedding
below and above the slumped horizon, and coherent but rolled-
up beds within the slump. The beds are commonly thin (a few
tens of centimeters maximum), and are associated with other thin
beds and with CCC turbidites (Walker, 1985) interpreted as
levee deposits. Thus the slumps may indicate rapid deposition of
beds on the levee (either the side facing the channel, though more
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 413

A N

Flow Direction
1.5 km

one km

S 50 m N
S
B
one km

50 msec

one km

SSW NNE

FIG. 21.A) Plan view of the erosional base of a mass-transport complex in the ultra-deep environment of the Makassar Strait,
Indonesia. Parallel to divergent erosional grooves are observed. B) Section-view image of mass-transport complex. The mass-
transport complex is characterized by chaotic seismic reflection character, with erosional scour in excess of 50 m locally, and up
to 1.5 km wide (modified from Posamentier et al., 2000). Seismic data courtesy of WesternGeco.

FIG. 22.Large slump resulting in almost complete disruption of bedding, Carboniferous Bude Sandstones at Efford, southwest
England.
414 HENRY W. POSAMENTIER AND ROGER G. WALKER

commonly on the back side of the levee), perhaps with trapped

pore fluid within the sandstones. The slopes associated with the
levee must have been sufficient to allow gentle sliding of just one
or two beds, without large-scale deformation of the underlying
sediment

CONTROLLING FACTORS
ON DEEP-WATER SYSTEMS

Here, we discuss some of the principles relevant to the depo-

sition of turbidites and the depositional elements within which
they occur. We will set the stage with regard to the context
within which deep-water depositional elements are deposited. A
sound understanding of process is key to the construction of
robust depositional models, enabling geoscientists to construct
models that will be applicable to their unique set of environmen-
tal circumstances. Such models can then be a useful predictor of

FIG. 23.Slump folds in Eocene slope mudstones of the Cozy Dell
Formation, Highway 33 north of Wheeler Gorge, California.
Scale shown by notebook.

FIG. 24 (below).Slump involving large stratified blocks in the

Upper Cretaceous Great Valley Sequence, Lake Berryessa,
California. The two heavy black lines show bedding (top to
the left)the bedding is parallel but the lines converge be-
cause the camera is pointed steeply up the cliff. The matrix is
a silty mudstone with a large variety of pebbles and cobbles.
The stratified blocks consist of layers of sandstone and mud-
stone. It is argued that these would easily be disintegrated
along the unconsolidated sandstone layers if there had been
significant transport. It follows that they may have collapsed
from a nearby channel wall and were buried before they could
break up. Slumped bed is about 7 m thick.

BEDDING
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
BEDDING
15
15 cm
cm
FIG. 25.Small-scale slump involving only two beds within otherwise flat-bedded succession, Eocene, Waitemata Group, New
Zealand.
spatial and temporal lithofacies distribution. A typical deep-
water depositional environment with associated depositional
elements is illustrated in Figure 1.
Sediment Staging Areas
The staging area can be defined as the shelf and/or upper-
slope location where turbulent flows originate (Fig. 1). This
staging area and the characteristics of the sediments delivered to
that area are all-important in dictating the nature of turbidity
currents and subsequently their deposits farther down-system.
In particular, the sand-to-mud ratio that characterizes these sedi-
ments plays an important role in determining whether long
leveed channels will develop down-system or whether short
leveed channels feeding frontal splays or lobes will characterize
downslope areas instead. The sediments that ultimately get in-
corporated into flows can be delivered to the staging area by
fluvial, eolian, or longshore-drift processes. Subsequent turbidity
currents originate as sediment failures, associated with seismic-
ity and slope instability. Alternatively, if rivers deliver sediment
directly to canyon heads, high-density flows within rivers can
continue directly into the deeper basin by density underflow (i.e.,
hyperpycnal flow). Such density underflows can transform into
true gravity flows farther down the slope.
The Significance of Sand-to-Mud Ratio within Flows
The initial sand-to-mud ratio within flows is dictated largely
by conditions in the staging area. The grain-size distribution in
these shelf-edge depocenters ultimately plays a critical role in the
415
style of turbidite deposition downslope, a relationship described
by Reading and Richards (1994). Posamentier and Kolla (2003a)
discuss how grain-size distribution exerts this control on the style
of turbidite deposition, and is schematically illustrated in Figure
26. This figure illustrates the relationship between total flow
height, the height of the high-density part of the flow, and levee
height, and the resulting transition between leveed channel and
frontal splay in the absence of a change in the gradient of the
slope.
As flows travel down-system, they progressively become
better organized, with finer sediments concentrating in the upper
part of the flow and coarser sediments concentrating in the lower
part of the flow. The result is that the upper part of the flow tends
to have a lower density and concentration than the lower part of
the flow. The tops of many turbulent flows are higher than the
associated levee crests (Fig. 27), the result of which is that the
lower part of the flow, where much of the sand-size sediment is
concentrated, is fully confined by the channel walls. In contrast,
the upper part of the flow, which rides well above the levee crests,
is largely unconfined by channel walls and hence is free to expand
laterally beyond the levee crests and onto the overbank. This
process of flow spillover results in the deposition of thin, fine-
grained turbidites (commonly CCC turbidites; Figs. 7, 8) on the
crests and backs of the levees, and it also results in progressive
impoverishment of mud within the total flow. In addition to mud
being lost from the flow by spillover, some sand may also be lost
from the flow because of sedimentation at the flow base and by
mixing with the remaining upper, lower-density part of the flow.
However, the amount of sand lost from the flow is volumetrically
significantly less than the amount of mud lost from the flow by
416 HENRY W. POSAMENTIER AND ROGER G. WALKER

Sand:Mud
Total Flow Height High
(i.e., Height of low-density +
high density columns)

Levee Height

Potential Overbank
Height Sediment Supply

Channel Complex
Single Leveed

(Frontal Splay)
Distributary
Channel
Low
Proximal Distance Down-System Distal
Effective Flow Height
Transition Point
(i.e., Height of high-density
column within turbidity flow)

FIG. 26.Schematic depiction of the interplay between sediment gravity flows, net sand, and levee height with distance down-
system. Note that the high-density part of the gravity flow is located progressively more closely to the levee crest with distance
seaward. A transition from leveed channel to frontal splay/lobe occurs when the high-density part of the flow (i.e., the sand-rich
part of the flow) reaches bankfull stage. Note also that the highest sand-to-mud ratio occurs there as well (modified from
Posamentier and Kolla, 2003a).

Levee crests

Potentially prospective
overbank deposit

Confined channel flow

Overspill

Flow stripping

FIG. 27.Schematic illustration of sediment gravity flow through a leveed channel (compare with Fig. 77). The cross-sectional view
illustrates that the flow top lies well above the levee crest. The part of the flow between the flow top and levee crest is unconfined
and systematically spills out of the channel. Enhanced spillover occurs at outer bends (by the process of flowstripping). These
locations constitute areas of preferred sand deposition in the levee environment.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
spillover. Thus the channel floor tends to aggrade somewhat
more slowly than the levee crests.
Also, somewhat counterintuitively, flows tend to become
sandier down-system (i.e., flows have a progressively higher
sand-to-mud ratio down-system) as a result of continual prefer-
ential shedding of muddier sediment due to spillover. Kolla and
Coumes (1987), Pirmez and Flood (1997), and Hiscott et al. (1997)
have observed that with increased distance down-system there is
a gradual increase of net sand deposited on levees, consistent
with a progressive impoverishment of mud within flows in the
down-system direction. The progressive loss of the upper or
mud-rich part of the flow results in a progressive decrease in
levee height down-system (Fig. 28). At some point down-system,
the high-concentration or sand-rich part of the flow reaches the
levee crests (bankfull stage). This is a critical location because
down-system from this point spillover is no longer associated
mainly with the muddy part of the flow; rather, sand-rich flows
are now directed across the overbank. Geomorphologically this is
expressed as a transition from a single leveed channel to a
distributary channel complex or frontal splay (Posamentier and
Kolla, 2003a). This location, referred to as the transition point, also
marks the location where the sand-to-mud ratio within the flow
is greatest. Downslope of this location the rate of sand being lost
from the flow exceeds the rate of mud being lost from the flow,
largely because the sand-rich part of the flow is now largely
unconfined. The increased cross-sectional area of the floor results
in decreased flow velocity and sand deposition.
The sand-to-mud ratio in the flow is critical to this analysis
insofar as changing this flow characteristic changes the location
FIG. 28.Arbitrary seismic section constructed along a levee
crest. Levee facies are characterized by low-amplitude dis-
continuous reflections and overlie a high-amplitude continu-
ous to discontinuous frontal-splay complex (see Figs. 64 and
116). The section, which is flattened on the levee top, shows a
progressive decrease in levee thickness from landward to
seaward. It also shows a consistently thicker levee along outer
bends than along inner bends. Most notably, the location
where the levee thickness approaches zero is also where the
confined flow within the leveed channel transitions into a
frontal splay/lobe. This location is referred to as the transition
point. Seismic data courtesy of WesternGeco.
West Decreasing levee height
Top Levee
Base Levee
Seaward
417
of the transition point (Figs. 29, 30). Thus, if a succession of sand-
rich flows is followed by a succession of mud-rich flows, the
transition point shifts seaward and the result is the superposi-
tion of a single leveed channel across an older frontal splay (Fig.
31). Under certain circumstances, where later flows are more
sand rich, the reverse can occur as well. As previously dis-
cussed, such changes in sand-to-mud ratio commonly originate
in the staging area, and are manifestations of changing propor-
tions and rates of delivery by rivers and shelf processes, of
different sediment sizes.
Slope and Basin Physiography
The morphology of the slope and basin floor influences the
deposition of turbidites in a variety of ways. Physiographic
factors include (1) sea-floor rugosity on a large scale such as fault
scarps and intraslope basins associated with salt tectonics or toe-
of-slope thrust faults, (2) small-scale sea-floor rugosity compris-
ing sea-floor irregularities associated with earlier depositional
events such as slides or debris flows, (3) the height of the available
relief from shelf edge to basin floor, (4) the gradient of the slope,
(5) the presence of significant breaks in slope such as those that
can occur where the slope meets the basin floor, and 6) the rate of
change of the slope.
Perhaps the most well documented example of the effect of
sea-floor rugosity on turbidite systems is the fill-and-spill model
of turbidite systems that characterizes salt-supported intraslope
basins (Prather et al., 1998). They described a scenario whereby a
string of intraslope basins would fill progressively from the
West East
n
wD
low
Flo
F Dirireecctitioon
X
X
Transition Point
East
100 msec
Inside Bend
five km
Outside Bend
418 HENRY W. POSAMENTIER AND ROGER G. WALKER

A Transition Point
Time 1
Leveed channel
Lev
ee C High Sand:Mud
rest
Height
(Early Lowstand)
Flow Top

Top of sand-rich
part of flow

Distance Down-System

Seaward Shift of Transition Point

Frontal splay

Time 2 Transition Point

Decreasing Sand:Mud
B
Leve Transition Point
e Cr
Height

est
Flow Top

Top of sand-rich
part of flow Distance Down-System

Time 3
Flow Top Low Sand:Mud
C Leve
(Late Lowstand)
e Cr Transition Point
est
Height

Top of sand-rich
part of flow Distance Down-System

FIG. 29.Shift of the transition point in response to differences in sand-to-mud ratio with sediment gravity flows. A) A high sand-
to-mud ratio is associated with a transition point that is significantly farther landward than is the case with a lower sand-to-mud
ratio B, C).

Transition Point Transition Point

Time of high sand:mud Time of low sand:mud

A'

Migration of transition point in response

to change from high to low sand:mud

Leveed channel

A A'

Frontal splay

FIG. 30.Superposition of leveed-channel system over a frontal splay/lobe, which would accompany a progressive muddying-up
of successive sediment gravity flows (compare with Figs. 31, 74, 116, and 163).
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
5 km
Late
Late Lowstand
Lowstand
Early
Early Lowstand
Lowstand
Late Lowstand
Early Lowstand
FIG. 31.Two examples of the superposition of late lowstand leveed-channel deposits over early lowstand frontal-splay deposits
(compare with Fig. 30). Seismic data courtesy of WesternGeco.
proximal basin to ever more distal basins through time. Within
each intraslope basin, turbidite deposition comprises a succes-
sion from sheet-bedded deposits at the base to leveed-channel
deposits near the top. Flows entering an intraslope basin would
initially encounter markedly concave-up topography. This con-
cave-up morphology results in a rapid deceleration of flow,
which in turn favors deposition of sheet-like frontal splay depos-
its (Posamentier and Kolla, 2003a). As the basin gradually fills,
the topography progressively becomes less concave up. The
response is an upward transition from frontal splays to leveed
channels. As a result, leveed-channel deposits tend to dominate
the upper part of intraslope basin fill. This basin-fill evolution
from frontal-splay dominated to leveed-channel dominated is an
autocyclic phenomenon occurring in response to the evolution of
local topography. Posamentier and Kolla (2003a) describe a ma-
trix of possible outcomes associated with varying slope concavity
and varying sand-to-mud ratio (Fig. 32A).
419
5 km
Late
Late Lowstand
Lowstand
Early
Early Lowstand
Lowstand
The effect of a slope to intraslope-basin transition is shown in
Figure 32B and C. Flow vectors on the slope are directed primarily
downslope parallel to flow. Upon encountering an intraslope
basin, laterally directed flow vectors are significantly enhanced.
This increases the likelihood of levee breaching and a resultant
distributive channel pattern and deposition of a frontal splay.
Sea-floor rugosity associated with fault scarps or other abrupt
changes in slope gradient can also have profound influence on
turbidite deposition. In general, the farther a flow travels without
any significant breaks in slope the more the sand grains tend to
become concentrated towards the base of the flow and mud tends
to concentrate towards the top. Thus grain-size sorting or segre-
gation within the flow progressively improves down-system
the flow becomes better organized. The presence of an abrupt
slope change in the path of the flow results in a perturbation (i.e.,
a hydraulic jump; Komar, 1971) within the flow and a consequent
tendency for abrupt increased flow disorganization. This pertur-
420 HENRY W. POSAMENTIER AND ROGER G. WALKER

C Slo
pe/
bas
B in
flo
or
tra
ns
itio
n

Time 1

Time 2

Time 3
FIG. 32.A) Matrix of possible responses of location of transition point to varying sand-to-mud ratio and varying slope curvature.
Increased slope curvature, a local or autocyclic parameter, results in a landward shift of the transition point. Increased sand-to-
mud ratio, an external or allocyclic parameter, has a similar effect (after Posamentier and Kolla, 2003a). B) Turbidity flow through
a leveed channel onto a basin floor. C) Upon encountering the basin floor, laterally directed flow vectors are significantly and
abruptly increased. This can result in deposition of a frontal splay on the basin floor. The degree to which a frontal splay forms
at the transition from slope to basin floor is a function of the abruptness of the slope change at this location.

bation results in poorer grain-size sorting within the body of the
flow down-system of the abrupt slope change. Consequently,
settings such as the base of a fault scarp or an abrupt slope-to-
basin transition can cause significant reorganization of flow. The
result of this reorganization is the tendency for the system to
change from relatively confined leveed channels to relatively
unconfined frontal splays (Posamentier and Kolla, 2003a) as
illustrated in Figure 33. The available relief from point of flow
origin to the basin floor can also play an important role in
determining the style of turbidite deposition. Assuming constant
slope gradient, the greater the relief from shelf edge to basin floor,
the longer the run of the turbidity flow, and therefore the greater
the tendency for concentration of sand towards the flow base.
This allows greater efficiency of levee construction and a greater
likelihood for levees to extend across the basin floor. All else
being equal, two flows of identical grain-size composition flow-
ing down two slopes with different length (each characterized by
the same gradient), one with relief of a few hundred meters and
the other with relief of a few thousand meters, can be associated
with a very different set of turbidite depositional elements on the
associated basin floor. As illustrated in Figure 34A the flow that
has reached the basin floor early in its run has not had the chance
to become organized from a grain-size distribution perspective,
hence the transition point is located significantly farther land-
ward. The basin floor in this instance is characterized by a
minimal leveed channel and a relatively widespread frontal
splay or lobe. Where the flow has had a long run before reaching
the basin floor (Fig. 34B) the flow is much better organized and the
transition point lies farther across the basin floor.
DEPOSITIONAL ELEMENTS
The integration of facies description and process sedimentol-
ogy leads to the identification of larger-scale depositional or
architectural elements and their linkage into depositional sys-
tems and ultimately depositional sequences. Depositional ele-
ments in deep-water systems are of the order of ten to a few tens
of meters in thickness, and may extend laterally for tens of meters
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 421

Confined flow (within leveed channel) Frontal splay

Lower-density
part of flow
Levee top

Higher-density part of flow

Location of hydraulic jump

Abrupt flow disorganization
and expansion

FIG. 33.Schematic depiction of flow expansion that occurs at the base of slope where the change of gradient is abrupt. Flow on the
slope is sufficiently well organized so that coarser-grained sediments are entrained in the flow base and are fully confined by
levees. Once the flow strikes the abrupt change in gradient that is located at the base of slope, the flow abruptly becomes
disorganized as it experiences a hydraulic jump. This abrupt increased disorganization causes a sudden increase in sand content
within the upper part of the flow and results in a situation where the higher-density part of the flow lies above the levee crests
and avulsions are likely to occur at this location. At this abrupt base-of-slope location, transition from leveed channel to frontal
splay is likely (compare with Fig. 32B, C).

Lower-density part of flow

A

Higher-density part of flow

Lower-density part of flow

Higher-density part of flow

FIG. 34.Schematic depiction of two similar sediment gravity flows at the point of initiation (i.e., similar sand-to-mud ratio, volume,
etc.), but facing slopes of the same grade but significantly different length. The sediment gravity flow that faces the short slope
A) has less distance available to it for sorting to occur than does the sediment gravity flow facing a long slope B). The result is that,
when the flow finally reaches the basin floor, the flow down the long slope is significantly better sorted, with coarser sediment
more concentrated near the flow base, than the flow down the short slope, where poorer sorting results in sands much higher up
in the flow.
422 HENRY W. POSAMENTIER AND ROGER G. WALKER

up to tens of kilometers. The elements are defined by (1) their
external geometries and (2) the internal facies within the ele-
ments. Our approach will follow that of Mutti and Normark
(1991), and our analysis of depositional elements will integrate
stratigraphic, geomorphologic, and facies observations, based on
seismic, outcrop, and borehole data. Moreover, the relationships
between depositional elements in time and space will be within
the framework of process sedimentology. We will start with
depositional elements that are observed in proximal settings
slope channels and canyonsand then move progressively far-
ther seaward down the slope and across the basin floor, where we
will examine leveed channels, overbank deposits (including sedi-
ment waves, crevasse splays, and planar levee deposits), frontal
splays (i.e., lobes or distributary channel complexes), and debris-
flow deposits (including debris-flow lobes, channels, and sheets).
In each instance we will suggest, based on process sedimentol-
ogy, what facies would be encountered in association with spe-
cific depositional elements.
Canyons and Slope Channels
Canyons and slope channels are the primary conduits for
sediments to travel from the shelf-edge staging area, across the
slope, and onto the basin floor. They can range in scale from a
few meters in depth and width (these would be referred to as
slope gullies) to ten or more kilometers wide and over a kilome-
ter deep (submarine canyons). They tend to be largely erosional
with significant incision into the substrate. In the context of this
discussion, the distinction we draw between canyons and slope
channels is that canyons consistently fully confine the flows that
pass through them, whereas slope channels only partially con-
fine the flows that pass through them (Fig. 35). The effect of
partial confinement is that some spillover from the tops of the
flows passing through slope channels occurs, resulting in the
construction of levees on the flanks of the channel. Levee con-
struction does not occur when the flows are fully confined, as
they are with canyons (though some canyons contain smaller
leveed channels within the confines of the canyon walls). An-
other distinction between canyons and slope channels is that,
whereas both can be deepened by the passage of turbidity
currents, canyons are more likely to widen by mass wasting on
the canyon walls than are slope channels. In the distal reaches of
canyons, as channel-wall relief diminishes, turbidity current
height eventually exceeds the height of the canyon walls, and
levees develop.
Canyons.
An example of a canyon is shown in Figure 36. There is no
evidence of levee construction on the flanks of the canyons,
suggesting that flows were fully confined within this feature. The
presence of sand within this canyon is largely at the base (Figs. 37
39), expressed as moderate- to high-sinuosity channel threads.
These channel deposits can be fully to partially preserved, the
latter illustrated by the segment of high-sinuosity channel depos-
its observed in the canyon terrace perched above the canyon floor
(Fig. 37). The canyon fill is inferred to be overwhelmingly mud
dominated, as evidenced by the seismic reflection character
observed within the confines of the canyon Figs. 37, 40). The
seismic reflection pattern of canyon fill commonly is character-
ized by moderate- to low-amplitude, discontinuous chaotic-con-
torted seismic reflections. This seismic reflection character com-
monly has been associated with mass-transport deposits such as
associated with slides and debris flows (Posamentier and Kolla,
2003a). The morphology of the top of the canyon fill is character-
ized by linear flow lines (Fig. 36A, C), further evidence for the
absence of turbulence in the flows responsible for at least the most
recent phase of canyon fill (Posamentier and Kolla, 2003a). How-
ever, there can exist isolated threads of channel sands, character-
ized by high-amplitude, continuous to discontinuous seismic
reflections at the canyon base or embedded within the canyon fill
near the canyon base.
The fill of many canyons is commonly fine grained, and it is
deposited after the canyon or channel has been abandoned. If
abandonment is due to cutoff of sediment supply during rela-
tive sea-level rise, the fill may consist largely of slump and slide
material from the canyon or slope channel walls with additional
contribution of hemipelagic mud and silt that gradually settles
over the area of the slope and in the canyon. The present-day
Mississippi Canyon appears to have filled in this way (Goodwin
and Prior, 1989). Consequently, it is likely that the preponder-
ance of canyon filling occurs only after the axial turbidity-

A B

Canyon Slope Channel or Gully

FIG. 35.Schematic illustration showing section views across a canyon A) and slope channel or gully B). Sediment gravity flows
within canyons are fully confined by canyon walls. Consequently no levee deposits are observed outside the canyon. In contrast,
sediment gravity flows within slope channels or gullies are not fully confined by channel walls so that levee deposits are observed
outside the channels or gullies.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 423

A B C

Flo
w
dir
ec
tio
n

Figure 42

10 km

FIG. 36.Seismic attributes of the modern sea floor in the vicinity of Mississippi Canyon, Gulf of Mexico; seaward is from upper left to
lower right. A) seismic dip azimuth mapthis map highlights the principal canyon as well as smaller tributary channels. The canyon
floor is characterized by long linear grooves. The walls of the canyon are characterized by delicate tributary networks of small gullies.
B) seismic dip magnitudethis attribute highlights the edges of the canyon floor as well as the edges of the tributary channels. C)
seismic reflection curvaturethis attribute highlights the drainage networks and drainage divides; it also clearly shows the arcuate
nature of the canyon walls in places (compare with Fig. 42). Seismic data proprietary to PGS Marine Geophysical NSA.

Mass-transport deposits

Flow direction

FIG. 37.Seismic cross section through Mississippi canyon, with low-sinuosity to moderate-sinuosity turbidite channels at the
canyon base. Note that the bulk of the canyon fill comprises mass-transport deposits. The axial channel shown here corresponds
to the channel segment shown in Figure 38A. The sinuous channel on the flank of the canyon fill is only partially preserved. Seismic
data proprietary to PGS Marine Geophysical NSA.
424 HENRY W. POSAMENTIER AND ROGER G. WALKER

A B
A B

1 km
C
C

1 km 1 km

Sinuosity = 1.2

Flo
w
Di
rec
tio Channel width = 200250m
n

Sinuosity = 1.5
10 km

FIG. 38.Seismic reflection amplitude images of the channel at the base of Mississippi canyon (compare with Figs. 37 and 39).
Sinuosity ranges from 1.2 proximally to 1.5 distally. Seismic data proprietary to PGS Marine Geophysical NSA.

FIG. 39.Coherence image of proximal Mississippi canyon channel. Note the discrete channel threads indicate meander loop
expansion (i.e., swing) and down-system migration (i.e., sweep). Seismic data proprietary to PGS Marine Geophysical NSA.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 425

current channels shut down. Once these conveyer belts cease
to be active, any sloughing off the canyon walls remains within
the confines of the canyon. Extensive slumping commonly
characterizes the canyon walls (Figs. 40, 41). Thus, the bulk of
canyon filling likely occurs as a result of mass wasting during
rapid sea-level rises and subsequent highstands, thus making
the fill predominantly part of the transgressive and highstand
systems tracts.
Some canyon systems characterized by supply of very coarse-
grained sediment can have extensive coarse-grained sediments
deposited within canyons (e.g., Pigeon Point Formation, Lowe,
1979; Carmelo Formation, Clifton, 1981, 1984). This is most com-
mon in active continental-margin settings where deep-water
turbidite systems are linked to short and steep fluvial drainage
systems.
The facies that would most likely characterize the bulk of the
canyon fill is that of a debrite, that is, a mud-dominated deposit
with minimal internal organization. Convolute bedding associ-
ated with isolated cohesive blocks can be present in some
instances. The isolated channel deposits (e.g., Fig. 37) observed
within the canyon would be characterized by true turbidite
facies, likely dominated by Bouma A and B units, with Bouma
C, D, and E units commonly lacking preservation potential
because of erosion by successive turbidity currents through the
channel. These channels commonly are not associated with
constructional levees, inasmuch as the canyon walls serve the
purpose of confining the flows in their entirety, thus precluding
the possibility of flow spillover and levee construction. In
isolated instances, small channels within the middle to upper
part of the canyon fill can have associated levees, all deposited
within the confines of the canyon walls. This situation can
develop when a turbidity flow travels across the relatively flat
floor of a partially filled canyon. Essentially, this constitutes an
underfit situation; the flows coming through the channel no
longer feel the canyon walls, hence they form their own
confining levees.
Some canyons can be completely mud filled and contain
channels that have no associated sand deposits. This commonly
occurs within tributary arms of larger canyon systems. Such
tributaries may have formed as a result of retrogressive slumping
on the canyon margin. The heads of these tributary systems do
not apparently link up with shelfal fluvial systems; hence there is
no significant sand input. These tributary systems commonly are
characterized by small slope gullies or rills, feeding larger, com-
monly straight axial channels. A tributary arm of Mississippi
Canyon, shown in Figure 42, is characterized by channels at the
base but notably no associated high-amplitude seismic reflec-
tions, suggesting a complete absence of sand in the channels.
Such features are thought to have formed by low-density, slow-
moving turbidity currents originating along the crests of steep
drainage divides.
Slope Channels.
Sand-prone slope channels such as that shown in Figure 43
have been described in some detail by Hackbarth et al. (1994),
Mayall and Stewart (2000), Kolla et al. (2001), and Posamentier
(2003a). In contrast with the canyon previously discussed, this
type of sediment conduit is associated with levee construction.
Such levees can be observed high up on the slope at least to
within 8 km of the shelf-edge staging area (Posamentier, 2003a).
The channel thread at the base of the slope channel illustrated by
Posamentier (2003a) is characterized by high-amplitude seismic
reflections and a moderate- to high-sinuosity channel pattern
that persists landward nearly to the toe of slope of a small shelf-
edge delta, which itself is laterally confined to the head of the
slope channel (Fig. 44). The presence of channel sinuosity nearly
to the slope channel head suggests the presence of turbulent
flow at least this high up on the slope, if not all the way to the
shelfslope break. Further, a time slice through the upper slope
and outer shelf reveals a protuberance of the shelf edge pre-
cisely where the slope channel is located (Fig. 45). The associa-

A B

Note scallop-shaped
slump scars on
canyon wall

Canyon Margin
Flow direction

FIG. 40.A) Transverse profile through Mississippi Canyon illustrating canyon cut as well as sand-prone channel fill at base of
canyon. Dotted line indicates location of time slice shown in B. B) Time slice through Mississippi Canyon illustrating arcuate walls
indicative of extensive slumping along canyon walls. Dotted line indicates location of time slice shown in A. Data proprietary to
PGS Marine Geophysical NSA.
426 HENRY W. POSAMENTIER AND ROGER G. WALKER

A B

on
cti
re
Di
Flo

w
w

Flo
Dir
ec
tio
n

5 km

FIG. 41.A) Curvature map of modern sea floor over Mississippi Canyon characterized by arcuate walls indicative of extensive
slumping. B) Time slice across Mississippi Canyon and 3D perspective view of interpreted arcuate wall of the canyon. Data
proprietary to PGS Marine Geophysical NSA

5 km

FIG. 42.Seismic reflection curvature of the sea floor along southwestern side of Mississippi Canyon, Gulf of Mexico (detail of Fig. 36).
Three larger channels up to one kilometer wide, as well as numerous smaller gullies and rills, are shown. Note that many of the
smaller gullies and rills originate along knife-edge drainage divides. Seismic data proprietary to PGS Marine Geophysical NSA.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 427

Flow direction

FIG. 43.3D perspective illumination of the Einstein Channel on the upper slope of the Desoto Canyon area, eastern Gulf of Mexico.
Note the presence of levees external to this channel as well as the sinuous nature of the channel pattern. The channel is
approximately 1.5 km wide, from levee crest to levee crest. Seismic data courtesy of VeritasDGC.

tion of a shelf-edge protuberance or delta with the slope channel
suggests a genetic link between shelf fluvial and distributary-
channel sediment delivery systems and the presence of turbid-
ite deposits on the slope. One can infer from this close connec-
tion between the fluvial distributary channels and the meander-
ing threads within the slope channel that the turbidity currents
which came through the slope channel may have originated as
density underflows (i.e., hyperpycnal flows) sourced by river-
borne sediments. Such processes likely would have been active
for days or weeks at a time, while the river was in flood. Direct
links between fluvial distributary channels and canyon systems
is also common. Most well-developed canyons are associated
with major river systems on the shelf (e.g., the Hudson canyon,
the Congo canyon, and the Mississippi canyon) (Posamentier
and Allen, 1999).
Slope channels commonly contain sand-prone facies at or
near the base of the conduit. Mayall and Stewart (2000) document
examples and propose a model for slope channel fill that contains
debris-flow deposits at the base overlain by amalgamated turbid-
ites. The fill culminates with late-stage, isolated underfit leveed
channels associated with diminishing flow discharge, located
within the master slope-channel walls. Other examples of com-
plex slope channel fill are documented in Kolla et al. (2001),
Sikkima and Wojcik (2000), and Abreu et al. (2003). The amalgam-
ated sandy facies are characterized predominantly by Bouma A
and B turbidites, whereas the underfit late-stage channels are
characterized by Bouma A to D turbidites. Climbing current
ripples can be observed most commonly within the levees asso-
ciated with the late-stage underfit channels.
Examples of Coarse-Grained Canyon Fills.
In unusual cases, the upper parts of large canyons may
contain significant quantities of sand. Two examples have been
reported from the Atlantic margin of Brazil, in studies based on
closely spaced log and core control. The sand-prone fill of a large
slope channel (canyon?) has been documented in the Carapeba
Pargo system on the Atlantic margin offshore Brazil (Fig. 46;
Bruhn and Walker, 1995), where a large slope channel can be
traced for at least 150 km. Turbidites have been studied in the
Carapeba and Pargo oilfields, which lie about 90 km down-
canyon from the updip erosional edge of Cretaceous rocks. The
slope-channel width is of the order of a few kilometers, the
precise width of the channel being hard to determine because
younger turbidites spread more and more widely over the upper
parts of the slope-channel margins. The thickness of the fill is
428 HENRY W. POSAMENTIER AND ROGER G. WALKER

Flow direction

Transverse
section
D

Axial
section

FIG. 44.A) 3D perspective view of the Einstein channel (approximately 2 km wide) with associated sand-prone channel fill at the
base and shelf-edge delta at the slope-channel head. Each delta-front shingle is shown in a different color. B) Axial section through
the Einstein slope channel illustrating the shingled nature of the slope-channel-head delta as well as the high reflection amplitude
of both the delta as well as the channel fill at the base of the slope channel. C) Transverse section through the Einstein channel
illustrating high-amplitude sand-prone deposits at base of channel. D) Seismic reflection amplitude of the deposits at the base
of the slope channel illustrating the presence of sinuous narrow channel threads, characterized by down-system migration of
meander loops. Seismic data courtesy of VeritasDGC.

about 260 m (Turonian) plus 24 m (Maastrichtian). The turbidites
within the slope channel form eight thinning- and fining-upward
successions. Individual beds at the bases of the successions can be
up to 12 m thick and consist of granule sandstones with scattered
pebbles (Fig. 47; the pebbly sandstone facies described earlier in
this paper). The successions are 27 to 140 m thick and can be
mapped as tabular or linguoid sandbodies 1 to 12 km wide in
which the younger turbidites become finer grained, thinner
bedded, and more discontinuous upsection and downcanyon.
The successions are stacked in an overall retrogradational back-
stepping pattern for at least 20 km, recording the fill of the slope
channel.
The slope-channel morphology can be traced at least 60 km
downslope from Carapeba (Fig. 46), thus defining the deposi-
tional site at Carapeba and Pargo as within the slope channel. The
depositional elements consist of the thinning- and fining-upward
successions, which appear to spread with a sheet-like geometry
from wall to wall (Bruhn and Walker, 1995, their Figs. 5 and 6). A
hypothetical continuous outcrop of this canyon fill, 200 m thick
and covering one km2, would be almost impossible to interpret as
a canyon fillthe only suggestion of channelization might be the
very thick beds and the coarse grain sizes.
The second example of canyon filling is also from the Atlantic
margin of Brazil, in Regencia Canyon (Bruhn and Walker, 1997).
The canyon can be mapped near the mouth of the Doce River (Fig.
48) and is up to 6 km wide. It can be mapped for at least 15 km, and
the fill is up to 1 km thick. Lagoa Parda field is about 2.8 km long
and 2.5 km wide, and it has 70 wells with an average spacing of 200
250 m. Seven wells are cored, with a total of 324 m of core. Because
of this unusually good control, individual beds can be traced from
well to well; where beds or groups of beds can no longer be
correlated, channel margins can be defined (detailed correlations
are shown by Bruhn and Walker, 1997, their fig. 13, and a schematic
correlation diagram is shown here in Fig. 49). The thirty-eight
channels so defined are 210 to more than 1050 m wide and over 1
km long. The channel fills range in thickness from 9 m to more than
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 429

e
ic
sl
e
m
Ti
Time slice

Shelf-Edge
Shelf-Edge Delta
Delta
B

FIG. 45.A) Perspective view of Einstein channel and associated shelf edge, Gulf of Mexico, from 3D seismic data. The surface shown
lies at the base of the channel levees and is Late Pleistocene in age. B) Time slice through the upper part of the section reveals a
protuberance of the shelf edge corresponding to a shelf-edge delta (compare with Fig. 57). Note that this protuberance
corresponds precisely to the location of the Einstein slope channel. Seismic data courtesy of VeritasDGC.

50 m. Relationships of channels and their stacking patterns are
shown in Bruhn and Walker (1997, their figs. 14, 15, and 16).
Channel fills comprise bouldery to pebbly conglomerates in
normally graded beds up to 6.4 m thick, along with graded
sandstone beds up to 3.8 m thick. Finer-grained facies include
bioturbated mudstones and thin-bedded sandstones, and mo-
notonous dark gray mudstones. Detailed correlations show that
these finer-grained facies are deposited as the levees of channels
filled with the coarser facies.
The channels can be grouped into three channel complexes
(CC), colored orange (channels 111), yellow (1335), and red
(3638) in Figure 49. Overall, the channel fills become narrower,
thinner, and finer grained from CC 1 (orange) to CC 3 (red).
Channel complexes 1 and 2 are deeply incised. Channel orienta-
tions in CC1 suggest flows from smaller tributaries along the
northwest margin of Regencia Canyon (Fig. 48), and there are no
levee facies associated with the CC1 channels. Orientations in
CC2 suggest flows generally from west to east along the axis of
the canyon. Levees are associated with channels 1935 (Fig. 48).
Slopes on the levees are up to 10 degrees, with slopes facing the
channels generally being steeper than slopes facing away from
the channels. Levees are also steeper on the left sides of the levees
(Coriolis effect in the southern hemisphere). In CC3, channel
orientations again suggest flow from the northwest margin of
Regencia Canyon, and asymmetrical levee growth appears to
have influenced channel switching (details in Bruhn and Walker,
1997). Channels are only 230280 m wide, with fills 1315 m thick.
Two main points emerge from this study. First, it is clear that
turbidity currents can deposit coarse sediment within canyon
heads, where the distance between flow initiation and sediment
deposition is only a few kilometers. Deposition may be strongly
influenced by the abrupt flattening of the gradient from canyon
margin to canyon floor, particularly in CC1 and CC3. Second, it
is clear that small (1020 m deep, 200300 m wide) leveed chan-
nels can develop within canyon headsthis makes the interpre-
tation of some limited outcrops very difficult. Specifically, the
channels at San Clemente, California (described below), have
coarse fills, muddy channel walls, and multiple incisions. They
have previously been described as mid-fan channels (Walker
(1975b), but this interpretation is revised below, partly in the light
430 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 46.Location map of Carapeba and Pargo fields, offshore Brazil. Note Brazilian coastline and city of Campos. Long serrated
line shows limit of occurrence of Upper Cretaceous rocks (after Bruhn and Walker, 1995).

of a resemblance between the San Clemente channels and the CC3
channels of Lagoa Parda. The general point to emphasize is that,
unless an ancient example can be fairly positively identified, it
cannot be used to construct general facies modelsusing the data
from San Clemente without being sure of its setting (canyon head
or basin floor) could lead to distorted syntheses of data during
modernancient comparisons and facies modeling.
Example of Slope-Channel Fill,
San Clemente, California.
A candidate for a slope-channel fill is observed in the
Capistrano Formation (Upper Miocene) at San Clemente, Califor-
nia (Fig. 50; Walker, 1975b; Campion et al., 2000; Camacho et al.,
2002). The channel complex cuts into mudstones containing some
beds of chert (similar to beds in the Monterey Chert). Bedding is
generally horizontal in these mudstones, but some beds are
broken by small soft-sediment faults suggesting movement on a
slopethus the beds outside the channel may be slope deposits.
Within the channel complex, beds are very well exposed.
They are also horizontal, and the channel complex can be traced
for over 500 m (Fig. 51). The Upper Miocene part of the cliff is 20
m thick, with indications that the nested channels may be as much
as 40 m thick (Walker, 1975b). In addition to horizontal beds,
there are eight prominent inclined surfaces, seven of which are
draped with mudstone (Fig. 51). The surfaces have dips ranging
from 5 to 18 degrees, and the mudstone drapes vary from 30 cm
to 2 m in thickness.
The interpretation of the San Clemente section is that it
comprises a series of nested channels, separated by dipping mud-
draped surfaces. There were no outcrop indications that the
channels might be sinuous, although excavations of channel
walls suggested that the strikes of the walls varied from 230 to 300
degrees (Walker, 1975b).
The fill of the channels consists of interbedded sandstones
and mudstones. The grain size is up to coarse sand, with many
beds containing scattered granules and pebbles. Sandstone bed
thickness is variable; in some of the channels it is typically a few
tens of centimeters (Fig. 52), whereas in other channels beds can
be almost 1 m or thicker (Fig. 53). The thicker beds tend to be
amalgamated, without mudstone layers (Fig. 54). Amalgamation
surfaces are characterized by grain-size changes and bedding
irregularities that suggest loading of the upper sandstone into the
lower one (Fig. 54).
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
10 cm
FIG. 47.Core photo of pebbly sandstone, Carapeba Field. Core
sleeves are 1 m long, top to left. The core shows one pebbly
sandstone bed, with prominent fluid-escape features in the
uppermost two core sleeves.
The finer layers between the sandstones consist of gray silt-
stones and mudstones with abundant bioturbation. On the in-
clined mudstone drapes, gray silty mudstones are irregularly
interbedded with very fine-grained brown claystones (Fig. 55).
The different fine-grained facies were not examined micro-
paleontologically, but commonly in Tertiary turbidites in Califor-
nia the gray silty mudstones contain transported shallow-water
foraminifera whereas the brown claystones contain deeper-wa-
ter benthonic foraminifera. It therefore is assumed that the in-
clined mudstone drapes consist of silt and mud introduced by
turbidity currents (the gray layers), as well as hemipelagic brown
claystones deposited between turbidity currents.
These relationships are particularly well displayed in channel
6 (Fig. 56). The lower three beds pinch out rapidly toward the
431
channel margin. In the case of the lower two beds, thin sandy
layers a few millimeters thick continue up the channel margin
and eventually disappear. It appears that the turbidity currents
transported sand close to the bed, and that silt and mud (with very
thin sand layers) was draped higher up over the channel margin.
Note also that potential reservoir rocks in the channel pinch out
laterally against the channel margin within a few meters.
The suggestion of soft-sediment disturbance in the beds out-
side the channel complex, and the proven occurrence of small
leveed channels within larger incisions (Lagoa Parda, Brazil;
Bruhn and Walker, 1997), suggests that the Capistrano Formation
at San Clemente can be interpreted as a slope-channel fill.
3D Seismic Examples.
In contrast with outcrop data, good-quality 3D seismic data
can afford a comprehensive view of depositional elements and
their relationship to each other, though, of course, ground-truth
calibration may be lacking. In a well-documented example from
the eastern Gulf of Mexico, Posamentier (2003a) illustrates the
evolution of a linked shelf-edge delta and slope channel. The
shelf and slope environment became an active depocenter likely
in response to sea-level lowering and associated forced regres-
sion resulting in a seaward shift of the depocenter across the
shelf and to the shelf edge. Once the depocenter reached the
shelf edge a lowstand shelf-edge delta formed. This lowstand
delta, shown in axial view in Figure 57, is characterized first by
successive downstepping of the delta plain, likely associated
with falling sea level (i.e., the early lowstand systems tract of
Posamentier and Allen, 1999, or the falling stage systems tract of
Plint and Nummedal, 1998), and later aggradation of the delta
plain, likely associated with rising relative sea level (i.e., the late
lowstand systems tract of Posamentier and Allen, 1999, or
simply the lowstand systems tract of Plint and Nummedal,
1998). Numerous small slope channels or gullies are observed
on the surface that marks the base of this lowstand delta com-
plex (Posamentier, 2003a) (Fig. 58). These slope gullies are not
uniformly distributed along the entire breadth of the slope
shown, but rather tend to cluster in one area. The area where the
gullies are clustered directly coincides with the location of the
shelf-edge protuberance, suggesting a genetic link between
deltaic distributary channels and downdip slope gullies
(Posamentier, 2003a). These slope gullies are observed only at
the base of the lowstand delta complex; within and to the top of
the delta a single larger slope channel can be observed (Figs. 43,
44). This larger slope channel seems to have captured most of
the flow from the shelf systems at the expense of the numerous
smaller slope gullies that originally were present and seems to
represent the culmination of an evolution from many slope
gullies to a single slope channel within a single depositional
sequence.
Basin-Floor Leveed Channels
Basin-floor leveed channels commonly are genetically linked
with canyons or slope channels. Examples of basin-floor leveed
channels are illustrated in Figures 59 and 60 and are described by
Posamentier et al. (2000), Peakall et al. (2000), and Posamentier
(2003b). Aggradation of channel-fill deposits can occur both on
the basin floor (Fig. 61) and on the slope. The Joshua channel,
described by Posamentier (2003b), built a channel ridge that
stands c. 65 m above the adjacent basin floor. Likewise, the
channel shown in Figure 61 stands c. 90 m above its adjacent basin
floor. Sinuosity can be variable, from segments that are nearly
straight to segments that can display sinuosity of up to 3.0.
432 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 48.Location map of Regencia Canyon, State of Espirito Santo, Brazil. The canyon head is just onshore, close to the present day
Doce River mouth, 550 km northeast of Rio de Janeiro.

FIG. 49.Diagram of channels within the Lagoa Parda field, Regencia canyon head. Orangechannel fills of unstratified bouldery
to pebbly conglomerate and very coarse-grained sandstone. Yellow and redunstratified coarse-grained sandstone and parallel
stratified medium-grained sandstone. Greeninterbedded bioturbated mudstones and thin-bedded sandstones, interpreted as
levee facies. Note decrease in channel size and grain size upward. Lower channels do not have levees, but channel horizons 26
and higher have associated levees.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 433

San Clemente
N
0 0.5 km

Capistrano Formation
San Clemente
Outcrop Belt State Beach

L o Afte
0 160 km

ca r W
(
tio a
n o lke
y
dar

f C r, 1
oun

h a 97
k B
Par

n n 5)
els
1
San Francisco

8
Pa
Pa
cif

cif
ic

ic
Oc

Oc
ea

Los Angeles
n

ea
n
Study Area

FIG. 50.Location of the outcrop at San Clemente, California, south of the parking lot (from Campion et al., 2000).

40

Bioturb. Mdstns. Bioturb. Mdstns. Pleistocene

Parking
20 Lot Cgls. WIDE GULLY Cgls. Terrace

Mainly massive ssts Mainly massive ssts U. Miocene

Capistrano Fm
0
0 Meters 50 100 150 200
8 7
40

Path from Campground

Bioturb. Mdstns. at Top of Valley

20 Cgls.
Cgls.

CU
Path
0
210 250 300 350 Restrooms 400
6 5 4 4 3 Path Tunnel
40 2
Lifeguard

Bioturb. Mdstns.

20 Cgls.
Lower cliff edge

Bioturb. Mdstns. Bioturb. Mdstns.

0 Bioturb. Mdstns.
450 Poorly exposed 500 550 600 650
Tunnel 2 1 Tunnel

Mudstone Drape

FIG. 51.Measured section of the cliff at San Clemente. Distances are in meters south of the parking lot. Regional bedding is
horizontal, and heavy dipping black lines indicate mudstone drapes on channel walls (numbered 2 through 7). Channel 6 is shown
with red arrow and is discussed in the text. From Walker (1975b).
434 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 52 (above).Four thin but coarse-grained sandstone layers

separated by gray (turbidite) mudstones. Part of the fill of
channel 7, San Clemente.

Meander-loop migration is a common attribute of deep-water

channels, with both meander swing and sweep being common
features (Fig. 62). This can be observed both in section as well as
in map view (Figs. 6264). In many instances, the seismic expres-
sion of meander loops in plan view can be characterized by a
series of scrolls (Fig. 62). Somewhat less common are meander-
loop cutoffs with associated oxbows (Fig. 65). Internal scour at
successive channel bases within an overall aggradational amal-
gamated channel package is very common. The base of basin- 15 CM
floor channel complexes is erosional with at least some incision
into the substrate (Fig. 64).
In association with meander-loop channel migration, lateral-
accretion deposits can be observed on seismic sections (Figs. 66
69). Abreu et al. (2003) present a well-illustrated example of lateral-
accretion sets in a deep-water meandering channel offshore west
Africa. Such features are similar in stratigraphic architecture to
fluvial point bars (Fig. 70), though the formation process can be
quite different. Nonetheless stratigraphic compartmentalization
from a petroleum flow-unit perspective is quite similar.
Lateral-accretion surfaces have been observed in outcrop as
well, as in the Ross Sandstone and overlying Gull Island Forma-
tion of western Ireland (Lien et al., 2003). Figure 71 shows a cliff
outcrop with prominent lateral-accretion surfaces characterized
by alternating layers of sandstone and mudstone. These lateral-
accretion surfaces have a vertical relief of about 5 m. In the Gull
Island Formation, lateral-accretion surfaces can be seen in a
continuously sandy succession (Fig. 72).
Lateral accretion commonly can be seen in 3-D seismic images
(Figs. 6669). In Figure 62 the point bars show a series of scrolls

FIG. 53.Graded coarse sandstone at San Clemente, part of the

fill of channel 8.
DEEP-WATER TURBIDITES AND SUBMARINE FANS 435

FIG. 54.Two coarse sandstone beds amalgamated along a
loaded contact (black line). Quarter for scale. Note gray
(turbidite) mudstones above and below.

that may be comparable to the mudstone-draped lateral-accre-

tion surfaces of the Ross Formation or to the channel-wall drapes
from San Clemente (Fig. 51).
In both the San Clemente and the Ross examples, the lateral-
accretion surfaces contain prominent mudstone layers. These
could form vertical and/or lateral baffles to fluid flow in a
reservoir situation. Meander-loop (point bar) settings in the
subsurface (Fig. 62) may therefore present engineering problems
with respect to fluid flow within the meander loops.
The process responsible for meander-loop migration may
involve a single flow event with lateral accretion developing in a
way similar to point-bar development in fluvial systems. That is,
a prolonged flow event would produce an undercut bank along
outer bends of a meandering channel, and at the same time
accretes sediment on inner bends, forming lateral-accretion sets
such as shown in Figures 71 and 72. An alternative explanation
for how deep-water meander loops migrate down system is
shown in Figure 73. This process involves a succession of discrete
flows, so that, with each flow event, cutting or erosion accompa-
nies the passage of the head and body of the turbidity flow, and
fill accompanies the tail. Each flow event would preferentially
erode outer bends more deeply, thus resulting in lateral shift of
channel axis (i.e., meander swing). Moreover, because the thal-
weg would exercise maximum erosive force on the outer bend
just down system of the channel bend itself, the meander loop

FIG. 55 (below).Part of the drape on the wall of channel 2. The

gray layers are silty and easily disaggregated, and represent
mud introduced by the turbidity currents. The brown layers
are clay and contain a benthonic fauna. They represent hemi-
pelagic deposition between turbidity currents. Arrows indi-
cate truncation surface of the margin of channel 2 (see Fig. 51).
436 HENRY W. POSAMENTIER AND ROGER G. WALKER

REGIONAL BEDDING

RESERVOIR SECTION

FIG. 56.Channel 6 at San Clemente, California. Note the lateral pinchout of the lower three beds against the channel margin (red
arrows). In detail, the beds can be traced part way up the wall, where they are represented by very thin (< 1 cm) sandy layers. The
yellow dotted lines mark the mudstone drape on the channel wallnote that the lower two sandstones pinch out into the drape.
In the subsurface, the thick turbidites in the channel fill might make up a reservoir sectionbut note how rapidly this section
would disappear along strike.

one km

100 msec

one km
FIG. 57.Shelf-edge delta in close proxim-
100 msec ity to Einstein Channel. Note the suc-
cessive downsteps of the delta top, in-
Late lowstand dicating that the delta prograded un-
der the influence of falling sea level.
This is an excellent example of deposits
om plex associated with forced regression. The
sta nd c slightly progradational to aggrada-
e low tional section that caps the delta repre-
Bas sents deposition during the latter
phases of a sea-level lowstand, when
sea level is slowly rising. Seismic data
courtesy of VeritasDGC.
Early lowstand
(forced regression)
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 437

A C

Shelf Shelf

Slo Slo
pe pe

one km
B D (approximately)

FIG. 58.A, C) Clustering of slope gullies on two late Pleistocene surfaces in the area of the Einstein Channel, Gulf of Mexico, derived
from 3D seismic data. In each instance the gully clustering is at the base of a lowstand slope succession, and in each instance the
top of the lowstand slope succession is characterized by a single, larger slope channel, not coincidentally where gully clustering
at the base was greatest. B, D) illustrate seismic reflection profiles parallel to and transverse to the slope gullies. The relief of these
gullies commonly is less than 20 m. Seismic data proprietary to PGS Marine Geophysical NSA.

Sea floor dips gently Sea floor dips steeply Sea floor dips gently
(< 1 degree) to the East (~ 6 degrees) to the East (< 1 degree) to the East

Flow Direction

2 km
N
FIG. 59.Seismic reflection dip-azimuth map of deep-water leveed channel system in the Makassar Strait, Indonesia. The sea floor
dips to the right at a very low angle (less than 1 degree) with the exception of a seaward-dipping segment as indicated. This steep
dip is associated with the surface expression of a syndepositional base-of-slope toe thrust. Note the abrupt increase in channel
sinuosity and decrease of channel width seaward of this toe thrust. Note also the extensive sediment waves on both sides of the
leveed channel (compare with Fig. 77) (after Posamentier et al., 2000). Seismic data courtesy of WesternGeco.
438 HENRY W. POSAMENTIER AND ROGER G. WALKER

Flow
Dire
ction

5 km

FIG. 60.Seismic reflection dip-magnitude map of Joshua channel. The channel itself is part of a channel belt that lies within a larger
leveed channel (shown by red arrows). (Posamentier, 2003b). Seismic data courtesy of WesternGeco.

Levees

Flow direction

Crevasse splay 4 km
(approximately)
Leveed channel

FIG. 61.3D view of the top of a Pliocene leveed channel with crosscutting seismic section.
Note the crevasse splay deposited upon the lower overbank. Seismic data courtesy of
WesternGeco.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 439

Flo
w
dir
ect
ion
Sweep

3
2
1
Swing

FIG. 62.Seismic horizon slices illustrating meander-loop

swing and sweep within the Joshua Channel, Gulf of
Mexico. Channel threads indicate order of formation one km
(channel 1 is first, channel 3 is last). Seismic data courtesy
of WesternGeco.

n
r e ctio
w di
Flo

Sweep
1

3 one
one km
km
4

5 Amplitude extraction - horizon slice (-36 msec)

FIG. 63.Seismic horizon slice through the deep-water leveed-channel system shown in Figure 59, in the Makassar Strait, Indonesia
section. Down-system migration of meander loop is well expressed. Channel threads indicate order of formation (channel 1 first,
channel 5 last) illustrating channel sweep. Seismic data courtesy of WesternGeco.
 Passive channel fill 440
Channel fill
A
A Flow direction
C

Levee
Levee
A'

Frontal splay deposits

A' Channel
Channel base
base A

one km

D
HENRY W. POSAMENTIER AND ROGER G. WALKER

No vertical exaggeration

one km
FIG. 64.Seismic reflection profiles across the channellevee system shown in Figures 59 and 63 (illustrated in plan view in A) and section view in B D). This section
shows the channel fill to be characterized by high-amplitude reflections, inferred to be sand-prone (C). The transverse profile also is shown without vertical
exaggeration to reinforce the observation that interpretation of the architecture of these types of deposits can be greatly facilitated by squashing the profile. Seismic
data courtesy of WesternGeco.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 441

A B
Incipient neck cutoffs

Flo
w
di r
ec
tio
n
D
five km

C D

one km one km

FIG. 65.Meander-loop development within the Joshua channel, eastern Gulf of Mexico, as observed on seismic reflection horizon
slices (compare with Fig. 62). A) Time slice that illustrates two incipient cutoffs of meander-loop necks. B) Slightly higher
(approximately 12 m) time slice that illustrates cutoff of meander-loop necks. C, D) Time-slice details of the neck cutoffs and
resulting oxbows shown in Parts A and B. Seismic data courtesy of WesternGeco.

would therefore display down-system migration (i.e., meander
sweep). Figure 73 illustrates how flow velocity vectors tend to be
greater along outer bends especially just down-system from each
meander bend, thus tending to cause a down-system shift of
successive channel axes (i.e., meander sweep). The resulting
architecture of cut and fill can readily appear as lateral-accretion
architecture on seismic data, whereas the outcrop expression
would be one of cut and fill accompanying a progressive shift of
channel axes. Within each cut and fill, it is possible that smaller-
scale true lateral accretion such as shown in Figures 71 and 72 can
be present.
Avulsion events in deep-water channels have been docu-
mented using 3D seismic data (Posamentier and Kolla, 2003b).
Figure 74B, C illustrates a late lowstand leveed channel that
underwent an avulsion event. Avulsion events tend to be asso-
ciated most commonly with levee breaches or crevasses on
outer channel bends (Figs. 61, 75), though in isolated instances,
because of flow perturbations, crevasses can form on inner
bends as well.
Channel-fill lithofacies in these basin-plain leveed channels
would be similar to that which is encountered within slope chan-
nels. Those basin-floor channels characterized by more significant
aggradation would allow greater preservation of waning-phase
turbidites, so that Bouma A to C units would be most common.
Levee and Overbank Deposits
Channel levees commonly are deposits with concave-up,
gull-winged shapes. In general, the facies common to this depo-
sitional element include CCC turbidites, lenticular bedding and
small-scale erosion, slumping involving one or two beds, and
large-scale chaotic failures. Seismically, levees tend to be charac-
terized by low-amplitude continuous to discontinuous reflec-
tions (Figs. 64, 67, 69, 74).
Because the tops of turbidity currents associated with channels
commonly ride higher than the channel walls, there is continual
spillover onto surrounding areas (Fig. 27). When this occurs, flows
abruptly become less confined, and in response to this flow expan-
442 HENRY W. POSAMENTIER AND ROGER G. WALKER

A
Flo
w
Dir
ec
tio
n

one
one km
km

100 ms

one km

Lateral accretion

Slice level

FIG. 66.Seismic reflection horizon slice illustrating A) plan view and B) section view across a high-sinuosity deep-water leveed
channel, Gulf of Mexico. Down-system migration of meander loops (A) and lateral-accretion deposits (B) are clearly shown.
Seismic data courtesy of WesternGeco.

sion (i.e., increase in cross-sectional area), flow velocity abruptly
decreases. Lowered flow velocity results in decreased sediment-
carrying capacity; that is, the flow is less capable to transport
sediment, and rapid sedimentation out of suspension takes place.
As noted before, thin-bedded turbidites in levees commonly con-
tain climbing current ripples (Bouma C facies), convolute lamina-
tion, and ripped-up mud clasts (CCC turbidites). These sedimen-
tary structures are more common in proximal overbank/levee
settings than in any other turbidite environment of deposition.
Enhanced spillover occurs at outer bends of channels as the
upper parts of flows tend to continue in a straight-line trajectory
whereas the lower parts of flows tend to follow the curving
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 443

A x y
FIG. 67.Seismic reflection horizon slice illustrat-
ing A) plan view and BC) section view across
the Joshua Channel shown in Figure 60. The
section views (B and C) clearly illustrate mean-
der-loop migration and lateral accretion. The
lateral accretion in this system is characterized
by a strong aggradational component. Seismic
X' data courtesy of WesternGeco.
ction
one
one km
km Flow Dire Y'

B C
x X' y Y'

Slice level

100 msec one km

FIG. 68.A) Seismic reflection horizon slice

and B, C) section views across a leveed
A channel system, Gulf of Mexico. Both the
horizontal as well as the vertical sections
clearly show down-system meander-loop
X'
X' migration as well as lateral accretion (see
yy yellow boxes and line-drawing interpreta-
tions). The lateral accretion in this system is
characterized by minimal aggradation. Seis-
mic data courtesy of WesternGeco.

x Lateral accretion x

n
ctio
w dire
Flo
x
x y y

Y'
Y'
one
one km
km

B x
x X'
X'

Slice level

100 msec
C yy Y'
Y'
Slice level

one km
444 HENRY W. POSAMENTIER AND ROGER G. WALKER

A 55 km x y FIG. 69.Seismic reflection horizon

km
slice through a meandering slope
channel, A) offshore Nigeria, il-
lustrating down-system migration
of meander loops. The two seis-
mic reflection cross sections B, C)
illustrate the lateral-accretion
tio
n packages that are associated with
irec meander-loop evolution. Seismic
d
w data courtesy of VeritasDGC.
Flo
x' y'

x' x
B
Lateral accretion

Slice level

y' y

C
Lateral accretion

100 msec 5 km

A B
Flo
n wd
tio ire
irec ctio
n
d
ow
Fl

C
nn
iioo
cctt
irree
i
dd
ooww
FFll F IG . 70.Fluvial examples from
Wyoming showing plan-view
expression of lateral accretion,
meander-loop migration, and
meander-loop cutoffs. Channel
widths are approximately 500 m.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 445

LATERAL ACCRETION

BEDDING

FIG. 71.The Ross Sandstone in Rehy Cliff, Shannon Estuary, western Ireland. Regional bedding here is horizontal, and the dips in
the sandstone and mudstone layers in the upper part of the cliff are interpreted to represent lateral accretion on a point bar-
like feature in a submarine channel. Vertical relief of the lateral accretion surfaces is about 5 m.

B C

2:1 Vertical exaggeration

FIG. 72.A) Lateral-accretion surfaces in a sandstone from the Gull Island Formation (immediately above the Ross Formation) of
western Ireland. Thickness of sand body about 20 m. B) Two-to-one vertical exaggeration of outcrop photo with line-drawing
interpretation C).
446 HENRY W. POSAMENTIER AND ROGER G. WALKER

A
B

Flow direction
Time 1

Time 2
Slice Level
Time 3

Time 1 Time 2 Time 3

C
D
Flow-velocity contours Turbidite deposits
Time 1 Time 1 Cut Initial phase
Fill Late phase

Time 2 Time 2

Time 3 Time 3

FIG. 73.Moderately sinuous deep-water channel, eastern Gulf of Mexico. Cross section view shows apparent lateral accretion A).
Seismic time slice shows meander loop expansion (i.e., swing) as well as down-system meander-loop migration (i.e., sweep)
B). Schematic cross section illustrates flow velocity contours with highest velocity shown in yellow. Note the progressive
lateral shift of channel axis through time in response to highest flow velocity occurring on the right side of each channel C).
Schematic depiction of successive turbidite channel fills. Note that with the progressive lateral shift of the channel axis through
time, only the left side of each channel (with the exception of the final channel position) is preserved D). Seismic data courtesy
of WesternGeco.

channel path. This process of enhanced spillover by which the
upper part of the flow tends to shear off and decouple from the
lower part of the flow has been referred to as flow stripping (Piper
and Normark, 1983). One result of flow stripping is that levee
crests on outer channel bends are consistently higher than those
on inner channel bends (Fig. 76). Flow-stripping deposits in
certain instances can take the form of transverse sediment waves
(Fig. 7780). These sediment waves consistently appear to be
thicker on their proximal flanks, suggesting up-system migration
of wave crests (Fig. 78). The sedimentology of some sediment
waves has been documented by Migeon et al. (2000) in the Var
fan-channel turbidite system (Fig. 79). These waves also are
characterized by up-system wave-crest migration and have formed
transverse to flow direction. Cores taken from the updip, crestal,
and downdip limb of one of the Var sediment waves reveal
significantly greater amounts of sand present in the updip limb of
the sediment wave and the least amount in the downdip limb
(Fig. 79). Grain sizes up to medium sand were observed in the
updip limb. Other sediment waves observed in mid- to upper-
slope settings, which are not associated with turbidity-flow chan-
nels but rather with oceanic currents such as loop currents, likely
are not sand prone (Fig. 81).
Outcrop ExampleSimi Hills, California.
An excellent example of levee deposits in outcrop is the
Cretaceous Chatsworth Formation, Simi Hills, California, which
has been described by Link et al. (1984). Several hundred meters
of the Chatsworth Formation can be viewed along the north side
of the freeway (Highway 118) that traverses the Santa Susana
Pass (Fig. 82). Thick-bedded, channelized turbidites can be exam-
ined in Chatsworth State Park, with channellevee facies exposed
just to the south.
The levee facies is characterized by interbedded sand-
stones and mudstones in beds up to a few tens of centimeters
thick. Sandstone grain size is up to medium sand. Beds in the
basal parts of thinning-upward successions tend to be struc-
tureless and amalgamated (Fig. 83), but the thinner beds at the
tops of successions are commonly lenticular (Fig. 84) and are
characterized by CCC turbidites (Fig. 8). Climbing ripples, and
climbing ripples that become convoluted, are well displayed
in Figure 7. A scour with ripped-up mudstone clasts is shown
in Figure 85. The combination of these features suggests an
environment in which minor erosion is common, yet indi-
vidual beds are fairly coarse though relatively thin. Also, the
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 447

Avulsion node
A B

Flo
wd
irec
tion

G
C D
Leveed channel

Slices
AF

Frontal splay
Condensed section

100 msec one km

E F

FIG. 74.Succession of horizon slices through a deep-water turbidite system, eastern Gulf of Mexico. AF) Horizon slices illustrate
the evolution of this system from an initial frontal splay (F), eventually evolving into an isolated leveed channel (A). The cross-
section view (G) illustrates this evolution as well. Seismic data courtesy of WesternGeco.

climbing ripples and convolute lamination suggest rapid depo-
sition from suspension. The combination of features described
suggests rapid deposition of relatively coarse sand, yet in a
setting in which the beds are consistently relatively thinthat
is, high on a channel margin, or on the back side of a levee. In
this setting, turbidity currents may in places be erosive (Fig.
85), and the bedding lenticularity may be a function of depo-
sition in minor erosional scours and hollows (Fig. 84), or due
to the pinching of beds against a depositional topography (the
channel margin or levee).
Evidence for local slope instability within the levee deposits
is in the form of slump deposits involving a few beds, with
undeformed beds above and below. Such slumps are commonly
associated with CCC turbidites, and in this context probably
represent the local sliding of one or two beds, either on the back
of the levee away from the channel or down the channel wall
toward the channel. In many instances these slumps are sub-
seismic in scale. However, a seismic-scale example of slumps
into a channel is illustrated in Figures 76 and 80, and slumps on
the distal side of a levee in Figure 86.
The important architectural elements in these settings consist
of CCC turbidites, lenticular bedding styles, and slumps involv-
ing just one or two beds. Singly or in combination, these elements
combine to characterize the upper parts of channel margins, or
the backs of levees. These two settings can best be distinguished
by their context, as shown in the next example.
Outcrop ExampleWheeler Gorge, California.
An outcrop example of distal thin-bedded turbidites overlain
by a channellevee complex is observed in the Cretaceous Wheeler
Gorge channel systems north of Ojai, California (Rust, 1966;
Walker, 1975b). The beds are vertical and are preserved as a sliver
along the Santa Ynez Fault of the Transverse Ranges. They occur
as an isolated outcrop, and cannot be related to other Cretaceous
deposits in southern California. This example emphasizes the
importance of the vertical relationships of depositional elements
in establishing an overall interpretation of the depositional set-
ting. Once the overall setting has been established, individual
depositional elements can be reexamined in the light of the
overall setting.
The section can be subdivided into three parts (Walker, 1975b):
(1) a lower section of zebra-striped mudstones with 1-cm-scale
graded siltstones (about 250 m thick; Fig. 87); (2) a central section
that consists of three conglomerate-to-sandstone packages (105
m thick; Fig. 88); and (3) an upper section of interbedded sand-
 B 448

A A'
one km
Note positive relief

C
Channel fill eroded
50 m B B'

A Avulsion channel
Avulsion node

Note positive relief

A
A Note negative relief

B
B
Note negative relief
HENRY W. POSAMENTIER AND ROGER G. WALKER

B'
B'

A'
A'
Knickpoints
Flow direction

FIG. 75.A) Shaded oblique perspective view of channel shown in Figure 76. Two knickpoints are shown in association with an avulsion channel. The channel down-
system of the kickpoints remains unfilled and is characterized by a concave-up profile C). Upstream of the knickpoints the channel is filled with sand as indicated by
the convex-up profile B). Seismic data courtesy of WesternGeco.
 Relief: ~ 67 m 625 m
Avulsion node 2
A A' A
B Fl
ow
di
re
ct
io
n

100 msec

Avulsion node 1

Avulsion node 2
A' D
A

Slump scars
DEEP-WATER TURBIDITES AND SUBMARINE FANS

five km

FIG. 76.Images of Pleistocene Joshua channel, eastern Gulf of Mexico (Posamentier, 2003b). The perspective view B) and detail C) reveal the elevated nature of the channel
fill, suggesting a differential compaction effect associated with predominant sand fill of the channel. Note that levees are highest on outer bends. Note also the small
slump scars that characterize the inner levee slope. The channel is part of a channel belt that has accreted to a level 65 m above the adjacent basin plain D). Internally,
the channel is characterized by lateral accretion coupled with vertical aggradation A). Seismic data courtesy of WesternGeco.
449
 A 450

Flow direction
one km

B Sediment waves

C
HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 77.Leveed channel on the basin floor of the Makassar Strait, Indonesia C), characterized by overbank sediment waves illustrated on the associated dip magnitude
map A). Sediment waves are best developed on outer bends of the channel. Compare with the schematic illustration B). Seismic data courtesy of WesternGeco.
 A Direction of wave migration
A'

100 m Accretion

Current direction

one km

A A'
DEEP-WATER TURBIDITES AND SUBMARINE FANS

FIG. 78.Sediment waves associated with slope channels offshore Nigeria. Note that the levee crests migrate progressively up-system, suggesting landward migration
of these sediment waves. Seismic data courtesy of VeritasDGC.
451
 Overspill direction 452

A B
E7 45 E7 55
NW KN127 KN126 KN124 SE
Depth
(meters)

2400
N43 30

Wave migration direction

2500
Overspill

KN127
KN126
KN124

2400
N43 20

2500
1 km

Profile NIC34

KN127 KN126 KN124

HENRY W. POSAMENTIER AND ROGER G. WALKER

1m
F IG . 79.Overbank sediment
waves from the Var system,
Mediterranean Sea, shown in
plan view A), section view B),
and core C) (from Migeon et al.,
2000). Note the up-system wave
migration (B) as well as the pres-
ence of sand that characterizes
primarily the upstream limb of
these waves.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 453

Sediment waves

Flow direction

Slump scars

one km

FIG. 80.Sediment waves in the overbank of the Joshua channel, eastern Gulf of Mexico, shown on a curvature map extracted from
the upper bounding surface of the channellevee complex. Note also the small-scale slump scars on the inner levee. Seismic data
courtesy of WesternGeco.
454 HENRY W. POSAMENTIER AND ROGER G. WALKER

A B

Direction of wave migration

Shelf/slope break
Shelf/slope break
D
C

one km

one km

FIG. 81.Sediment waves on the middle to upper slope of the eastern Gulf of Mexico. These waves appear to be migrating obliquely
upslope. They are not associated with any nearby channel or canyon. Wavelength ranges from 400 to 600 m and wave amplitude
ranges from 5 to 10 m. Seismic data courtesy of VeritasDGC.

FIG. 82.Outcrop of Cretaceous Chatsworth Sandstone on north side of Highway 118, Santa Susana Pass, California. Note subtle
changes in bed thickness and sand/shale ratio from highway to skyline.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 455

FIG. 83.Sharp-based (yellow arrow), thick-bedded sandstones, with one prominent amalgamation surface (red arrow). The
irregular horizontal marks were made by an excavator! Total thickness of structureless sandstone about 3 m. This photograph
was taken about 20 years ago at the back of a trailer parkit is unlikely that the outcrop still exists; if it does it is probably
overgrown. Chatsworth Sandstone near Chatsworth, California.

FIG. 84.Interbedded sandstones and mudstones in the Chatsworth Sandstone, stratigraphically a few meters above the structureless
sandstones of Figure 83. Note the lenticularity of many of the beds (arrows). Thickness of section about 5 m.
456 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 85 (above).Scour outlined in yellow, with coarser-grained

structureless sandstone fill and abundant ripped-up mud-
A stones clasts. Chatsworth Sandstone near Chatsworth, Cali-
fornia.

FIG. 86.Slump scar on levee of deep-water channellevee

complex in Makassar Strait, Indonesia. Seismic data courtesy
of WesternGeco.

C
B
Flow direction

Slump scar

one km
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 457

Cgl. Seq. 1
Mudstones
Zebra Mudstones FIG. 87.Pace-and-compass map of
the lower part of the Wheeler
Gorge section on Highway 33,
north of Ojai, California. Note that
bedding is essentially vertical, and
that by rotating the map it also
Mudstones, 90 m serves as a measured section. The
section is slightly disrupted by
faulting, but the thicknesses shown
probably are fairly close.

Zebra
Mudstones, 120 m

Slump

50 m
North
Santa Ynez Fault

FIG. 88.Pace-and-compass map of the central channel complex at Wheeler Gorge, California. The channel complex can be divided
into three separate conglomerate sequences, interpreted as one channel fill.
458 HENRY W. POSAMENTIER AND ROGER G. WALKER
stones and mudstones with abundant soft-sediment deformation
(more than 120 m thick; Fig. 89).
The zebra-striped mudstones consist almost entirely of color-
graded siltstones 1 to 2 cm thick alternating with mudstones (Fig.
90). Right at the base, these beds are slumped in a section some 5
meters thick. The entire package is faulted, but it is estimated that
there are about 5500 of the graded siltstone beds. None of the
characteristics of levee deposits is present (i.e., no climbing
current ripples, no mud rip-up clasts, etc.); the depositional
environment is postulated to be very distaleither at a fan fringe
or lateral but very distal to a channel.
The central section comprises three stacked conglomerate-to-
sandstone packages. The base of the first sharply overlies the
zebra-striped mudstones and is characterized by large flute casts
(Fig. 91) and associated groove casts, some of which retain
pebbles at the ends of the grooves. There are also abundant
ripped-up clasts of the zebra mudstones within the basal channel
fill (Fig. 92). This lowest package is about 25 m thick, and it
consists of conglomerates and interbedded structureless sand-
stones and is interpreted as channel fill. The conglomerates are
characterized by abundant mudstone clasts, probably derived
from adjacent channel walls. The second package is approxi-
mately 50 m thick. There is a basal conglomerate, but the bulk of
the succession consists of graded conglomeratesandstone beds
(Fig. 93) and graded pebbly sandstones (Fig. 94). The uppermost
part of this package overlies these channel-fill deposits and
comprises a 15-m-thick succession of thin-bedded turbidites and
mudstones. Convolute lamination (Fig. 95), bedding lenticularity
(Fig. 96), and minor erosion surfaces (Fig. 96) with small ripped-
up mudstone clasts suggest that this is a channel-margin or back-
of-levee facies.
FIG. 89.Pace-and-compass map of the section above the channel complex at Wheeler Gorge, California. Two separate levee
complexes have been identified (both within the slumped turbidites).
As mentioned above, context is one of the most important
characteristics that allow architectural elements to be assigned to
depositional environments. The interbedded sandstones and
mudstones directly overlie the graded coarser facies of the chan-
nel fill. If the channel had migrated laterally, shifting the main
coarse-sand depocenter elsewhere, the channel margin could
migrate over the channel fill. A seismic example of this relation-
ship is shown in Figure 64B.
The third conglomerate-to-sandstone package is about 30 m
thick, and consists of conglomerates (25 m) that grade into
structureless sandstones (5 m) (Fig. 97). These are in turn overlain
by thin-bedded classical turbidites (Fig. 97). The base of this
package is spectacular, and erodes about 5 m into the underlying
beds. Large stratified sandstonemudstone clasts have been ripped
from the channel wall and deposited in the channel, with a sand
and conglomerate matrix (Fig. 98). Conversely, coarse sand from
the channel has been injected into the beds of the channel wall as
sills (Fig. 99)the sill in this photograph terminates abruptly, but
smaller sills splay from the top and bottom of the main sill. In
places, the sills feed dikes, clearly showing that these coarse
deposits are not original beds in the channel wall. Well developed
sills and dikes have also been documented in the subsurface (the
Tertiary Alba Field in the North Sea; Hurst et al., 2005) as well as
in outcrop (Surlyk and Noe-Nygaard, 2003; Hurst et al., 2005).
The uppermost part of the section exposed in Wheeler Gorge
consists of about 120 m of interbedded sandstones and mud-
stones. The lower 80 m consists of classical turbidites, with beds
averaging about 10 cm in thickness. Overall, there is a thinning-
upward succession. Individual beds have small ripped-up mud-
stone clasts, and convolute lamination is common. There is no
soft-sediment deformation. The upper 40 m contains similar
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 459

TOP

FIG. 90 (above).Very thin-bedded zebra-striped mudstones

below the main channel complex (Fig. 87) at Wheeler Gorge,
California. The striped appearance is the result of the stacking
of a large number of thin graded siltstone (pale) to mudstone
(dark) beds. Flat lamination and ripple cross lamination are
rare.

classical turbidites, but soft-sediment deformation is abundant.

A few beds a meter or so in thickness are completely disrupted,
but more commonly the deformation is restricted to a few beds
that show distinct soft-sediment folds (Figs. 100, 101). Convolute
lamination is also present, and in places there are distinct trends
in bed thicknesstwo thinning-upward successions, one thick-
ening-upward succession, and one thickening-to-thinning up-
ward succession.
The sedimentary folds involving one or two beds, and the
presence of convolute lamination, suggest that these thin-bed-
ded turbidites are channel-margin or levee deposits. Unlike the
channel-margin deposits described above, these thin-bedded
turbidites are not closely associated with thick, coarse-grained
channelized deposits. They are interpreted to represent back-of-
levee deposits associated with a channel system different from
that described abovea channel that is not exposed at Wheeler
Gorge (Fig. 102).

FIG. 91.Large flute casts on the base of the lowest conglomerate

sequence (Fig. 87), Wheeler Gorge, California. The bed is
vertical, flow direction is shown by the yellow arrow, and the
highway tunnel gives the scale.
460 HENRY W. POSAMENTIER AND ROGER G. WALKER

MUDSTONE
CLASTS

BASE OF CONGLOMERATE

TOP

BASEMENT CLASTS

FIG. 92.Very sharp base of conglomerate sequence 1 on the underlying zebra-striped mudstones. Note pale-colored basement clasts
and abundant ripped-up mudstone clasts.

TOP

FIG. 93.Graded conglomerate-to-sandstone bed, as part of the fill of conglomerate sequence 2 (Fig. 88), Wheeler Gorge, California.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 461

TOP
TOP
FIG. 94.Graded pebbly sandstone, part of conglomerate se-
quence 2 (Fig. 88). Irregular base shown by yellow dotted line.
Wheeler Gorge, California.

Outcrop Example of Levee FailureNew Zealand.

Probably the best example of levee failure and collapse is in
the Waitemata Group (Eocene, New Zealand; Ballance, 1964).
The large-scale soft-sediment deformations have been described
by Gregory (1966). The best outcrop is on Whangaparaoa Head,
where there is about 4 km of continuous coastal exposure in high
cliffs. The facies consists dominantly of thin-bedded classical
turbidites, with soft-sediment folding occurring on many dif-
ferent scales. Some of the deformation consists of stacked dis-
harmonic folds that can be seen in the cliff (Fig. 103) and on the
wave-cut platform (Fig. 104). In other places, entire sections of
turbidites appear to have been rotated, as in Figure 105, where
the regional dip of the Waitemata Group is only 20 degrees. The
beds in this rotated section contain CCC turbidites (Fig. 106) and
slumps involving one or two beds (Figs. 25, 107); both of these
features strongly suggest a levee origin for the thin-bedded
turbidites, and hence a levee-failure origin for the disharmonic
folding and rotation. Locally, sections of thick-bedded turbid-
ites suggest associated channel fills (Figs. 108, 109).

FIG. 95 (below).Convolute lamination in the uppermost part of

the fill of conglomerate sequence 2 (Fig. 88), Wheeler Gorge,
California.

TOP
462 HENRY W. POSAMENTIER AND ROGER G. WALKER

TOP

FIG. 96.Bedding lenticularity (red arrow) in the interbedded sandstones and mudstones, uppermost part of conglomerate sequence
2 (Fig. 88). Yellow dotted line shows scouring and more bed lenticularity. Wheeler Gorge, California.

CONGLOMERATE, SEQUENCE 3

STRUCTURELESS SANDSTONES

THIN-BEDDED TURBIDITES

FIG. 97.Overview of conglomerate sequence 3 (Fig. 88) after a forest fire. Note overall change from conglomerate, via structureless
sandstone, into classical turbidites. Section is about 30 m thick. Wheeler Gorge, California.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 463

LEVEE, SEQUENCE 2
TOP

FIG. 98 (above).Base of conglomerate sequence 3 (Fig. 88),

shown by yellow dotted line. Conglomerate cuts into levee
deposits of sequence 2 (yellow arrow). Channel fill consists of TOP
basement clasts, as well as stratified clasts plucked from the
channel wall (pink arrows). Green arrow shows a sill of coarse
sandstone injected into the fine-grained levee deposits.
Wheeler Gorge, California.

Crevasse Splays
Another form of overbank deposit is the crevasse splay. In
contrast with sediment waves, which are associated with flow
over the levee crest and onto the overbank, the crevasse splay is
associated with flow through the levee and into the overbank
environment. Because the crevasse splay involves flow through
the levee, this flow taps deeper into the main flow than simple
spillover flow, which taps only into the upper part of the flow. As
a result, flow through a breach in the levee is sourced by the more
sand-prone part of the main flow.
A typical crevasse splay is characterized by a short channel
leading away from the main channel and feeding a smaller
distributary channel system (Figs. 61, 110112). Note the plan-
view similarities between deep-water turbiditic and shallow-
water, continental-shelf deltaic crevasse splays (Fig. 113). A cre-
vasse splay can be considered a failed avulsion channel. The
distinction is that, in the case of an avulsion channel, flow is
permanently diverted through the crevasse and associated chan-

FIG. 99.Sill of coarse sand and granules (see Fig. 98) injected into
interbedded sandstones and mudstones of the wall of channel
3. Note very abrupt termination of the sill. Wheeler Gorge,
California.
464 HENRY W. POSAMENTIER AND ROGER G. WALKER

REGIONALTOP
CONVOLUTE LAMINATION

FIG. 100.Distinct soft-sediment folds involving two beds, with undisturbed bedding above and below. Note convolute
lamination in the uppermost bed of this small thickening-upward succession. Levee facies (Fig. 89) above the channel complex
in Wheeler Gorge.

REGIONAL TOP

FIG. 101.Re-folded fold involving one bed in an otherwise well-stratified succession. Regional top shown by yellow arrow, and
way-up of the folded bed is shown by yellow arrows. Levee facies (Fig. 89), Wheeler Gorge, California.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 465

Wheeler Gorge

ps
rbank slum
Levee-ove
retion
ateral acc
L

Channel-Levee
System 2
retion
ateral acc
L
Vertical stacking

retion Channel-Levee
al acc
Later System 1

Basin floor zebra-striped mudstones

Fact Interpretation

FIG. 102.Interpretation of the Wheeler Gorge section. The vertical line (fact) shows the location of the Wheeler Gorge section
the rest of the diagram is a reconstruction based on Wheeler Gorge data plus models derived from channel shingling patterns in
fans such as the Amazon and the Rhone.

FIG. 103.Large scale, disharmonically stacked soft-sediment folds in the Eocene Waitemata Group, Whangaparaoa Head, New
Zealand. Cliff is about 30 m high.
466 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 104.View from cliff down onto the wave-cut platform, showing disharmonic soft-sediment folding. Width of section at
base of photo about 60 m.

REGIONAL DIP = 20o

70 m

FIG. 105.The Eocene Waitemata Group has a regional dip of about 20 degreesthis photograph shows a section some 70 m thick
rotated by slumping to an angle of about 70 degrees. It lies just north of the disharmonic folds of Figure 101. The nature of the thin-
bedded turbidites in this outcrop is shown in Figures 106 and 107.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 467

CLIMBING
RIPPLES

TOP

CONVOLUTE
LAMINATION

FIG. 106.Thin-bedded turbidites with convolute lamination and climbing ripples, exposed on the wave-cut platform as part of
the 70 m thick rotated section shown in Figure 105. Whangaparaoa Head, New Zealand.

SLUMPED BEDS

FIG. 107.Thin-bedded turbidites, with one deformed horizon involving about 2 thin beds in a soft-sediment fold. Same location
as Figure 105.
468 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 108.Thick-bedded turbidites in the Eocene Waitemata
Group at Whangaparaoa Head, New Zealand. 15 cm scale is
circled.

nel, whereas with the crevasse splay the flow diversion is tempo-
rary and relatively short-lived. In some instances, as revealed by
3D seismic data, the early stage of a crevasse splay is character-
ized by a levee breach that feeds a field of transverse sediment
waves (Fig. 114). As the system becomes progressively better
organized, the sediment waves are overlain by a gradually ex-
panding distributary channel network (Posamentier and Kolla,
2003b; Van Wagoner et al., 2003). Channels within this distribu-
tary network are associated with low-relief, probably sand-prone
levees.
The most distinctive sedimentary structure associated with
crevasse-splay deposits is climbing current ripples. As with
overbank sediment waves, flow expansion, which occurs when
the confined flow within the main channel cuts through the levee
and becomes unconfined within the overbank environment, re-
sults in rapid sedimentation from suspension. In addition, be-
cause of erosion through the levee, locally derived mud rip-up
clasts can be common. The stratigraphic architecture of crevasse-
splay deposits is characterized by amalgamated turbidites near
the apex of the splay, becoming less amalgamated with distance
away from the splay apex.
In general, crevasse splays are far more common in basin-
floor environments than in slope environments. On the slope,
gravity-flow vectors tend to be directed parallel to flow, down the

FIG. 109 (below).Thick-bedded turbidites, with a suggestion of

a thinning-upward succession, in the Eocene Waitemata
Group, Whangaparaoa Head, New Zealand.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 469

n
tio
ec
dir
ow
Fl
Five km

FIG. 110.Seismic reflection horizon slice illustrating a crevasse splay associated with a channellevee complex on the basin floor of
the Gulf of Mexico. Note that several crevasses seem to have formed along this outer channel bend, feeding multiple small
channels characterized by bifurcation. Seismic data courtesy of WesternGeco.

one km
FIG. 111.Seismic reflection horizon slice illustrating a crevasse splay associated with a channellevee complex on the basin floor of
the Gulf of Mexico. Seismic data courtesy of WesternGeco.
 A 470

one
one km
km
C
Leveed
Leveed Channel
Channel
Line
Line D
D

Crevasse Splay

Line
Line E
E
one km

Leveed
Leveed Channel
Channel
HENRY W. POSAMENTIER AND ROGER G. WALKER

Crevasse
Crevasse splay
splay

Crevasse
Crevasse splay
splay
E

FIG. 112.Crevasse splay associated with a basin-floor channel/levee complex, Gulf of Mexico. Seismic reflection horizon slices A, B) illustrate deeper and shallower
slices through this system respectively. The crevasse splay is illustrated in perspective view in Part C) (see Fig. 61). Cross-section views through the principal
channel D) and the crevasse splay E) illustrate the high-amplitude character of these depositional elements. The high-amplitude character suggests the presence
of sand in these deposits. Seismic data courtesy of WesternGeco.
 A
B

one km

C
DEEP-WATER TURBIDITES AND SUBMARINE FANS

one km
one km

FIG. 113.Comparison between A) a shallow-water deltaic crevasse splay (from the Mississippi River delta) and B, C) two deep-water crevasse splays (Figs. 110 and
114) suggesting a strikingly similar morphology and likely similar depositional processes. Seismic data courtesy of WesternGeco.
471
 A B C 472

Flow direction

Sediment waves

D E
HENRY W. POSAMENTIER AND ROGER G. WALKER

one km

FIG. 114.Seismic reflection horizon slices from base A to top E) showing the morphological evolution of a leveed-channel crevasse splay. Note that the early
morphology appears to be characterized by transverse sediment waves as indicated by arrow (A). As the system evolves it appears to become better organized,
developing a distributary pattern (B through E). Seismic data courtesy of WesternGeco.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 473

relatively steep slope (i.e., slope angle commonly from 2 to 4
degrees), rather than normal to flow parallel to contour, thereby
diminishing the likelihood that sufficient force is exerted on
channel walls (i.e., levees) to form crevasses and hence crevasse
splays. Consequently the crevasse-splay sand habitat is an un-
common element in slope settings.
Frontal Splays (i.e., Lobes)
Leveed channels commonly are associated with terminal
lobes, or frontal splays (Figs. 115120). The terms frontal splay,
distributary channel complex, and lobe or lobeform can be used
somewhat interchangeably. Each term describes a depositional
element that lies at the end of a leveed channel and tends to be fan
or lobe shaped in plan view. Each term implies a different
perspective on this feature: (1) frontal splay has process signifi-
cance, wherein a flow spreads out or expands, (2) distributary
channel complex has map-pattern significance, wherein succes-
sive channel bifurcation results in a distributive channel network,
and (3) lobe or lobeform has morphologic significance wherein
the deposit has a fan-shaped or lobate planform.
As discussed above, frontal splays form at the transition
point, the location where levee height is no longer capable of
fully confining the sand-prone or high-concentration part of the
turbidity flow. At that location the sand-prone part of the flow
readily flows over the levee and establishes new courses on the
associated overbank. Also at this location, the down-system
trend towards increasing sand-to-mud ratio in the flow changes
to a decreasing sand-to-mud ratio, as shown schematically in
Figure 26. Up-system of the transition point mud is preferen-
tially lost from the flow by overbank spillover, whereas down-
system of the transition point sand is lost more rapidly from the
flow by sedimentation in the overbank and within distributary
channels.
Frontal splays have much in common with crevasse splays
both sedimentologically and architecturally. Both are distin-
guished by relatively common current ripples and sheet-like
bedding geometry with shallow channels and sand-prone, low-
relief levees. Both tend to be more amalgamated near the splay
apex and less so distally. Preservation of interbedded mudstones
becomes progressively more common with increasing distance
away from the transition point, which corresponds to the frontal

Frontal splay

Transition point

Flow direction

five km

FIG. 115.Frontal splay at the end of a leveed channel system on the basin floor of the Makassar Strait, Indonesia. The transition point
between leveed channel and frontal splay is located where levee height has diminished to below seismic resolution (compare with
Fig. 28). Seismic data courtesy of WesternGeco.
474 HENRY W. POSAMENTIER AND ROGER G. WALKER

Y Y'

A
100 msec
Slice level
Frontal splay
one km

one km Y'

FIG. 116.Frontal splay at base of channellevee complex, Makassar Strait, Indonesia. This splay complex is characterized in section
view by continuous to discontinuous high-amplitude seismic reflections and in plan view by extensive bifurcation. Note that, as
with the channel-levee complex illustrated in Figure 117, the frontal splay immediately underlies a moderate- to high-sinuosity
solitary leveed channel. Seismic data courtesy of WesternGeco.

splay apex. However, several characteristics distinguish frontal
splays from crevasse splays: (1) Frontal splays tend to be larger,
insofar as they involve the entire flow discharge in contrast with
only part of the flow discharge (i.e., that part of the flow that is
temporarily diverted through the crevasse) associated with cre-
vasse splay. (2) Crevasse splays commonly tend to be associated
with levee-derived rip-up clasts. These clasts tend to be mud
prone. (3) Crevasse splays lie in close proximity to channel levees,
where the flows have passed through a breach. (4) Slope instabil-
ity with resulting bed convolution and slumping is more com-
mon in crevasse-splay settings, insofar as they are deposited on
potentially steeper slopes of the overbank environment.
Most commonly, frontal splays are deposited on basin floors
or on the floors of intraslope basins, with leveed channels and
channel complexes being the more common form of turbidite
element encountered in slope environments. In some instances,
however, frontal splays can be observed on slopes as well, espe-
cially in intraslope basins (Prather et al., 1998), but also in some
instances on open slopes as well (Fig. 121).
The splays shown in this paper are mostly channelized, with
channel widths from about 100 m to smaller than can be imaged.
The channels appear to be multiple (possibly braided), al-
though it is not clear if more than one channel is occupied and
active during any single turbidity current. Other images show
that in more distal settings, the top of the frontal splay may be
essentially smoothany topography is too small to be imaged.
Outcrop ExampleCounty Clare, Ireland.
The outcrop example of a frontal splay discussed here is from
the Carboniferous Ross Sandstone of western Ireland (Figs. 122,
123; Martinsen et al., 2000; Lien et al., 2003). The Ross is about 460
m thick, and details of the regional geology, stratigraphy, and
paleotectonic setting are given by Lien et al. (2003). It can be
informally divided into lower and upper parts, 170 m and 290 m
thick, respectively.
Frontal splays are best seen in the lower Rossthe base is
gradational from the underlying Clare Shale. Initial turbidites are
very thin and widely spaced stratigraphically, but overall, through-
out the 170 m thickness, the beds tend to thicken upward (Lien et
al., 2003). In detail, however, there are no systematic thickening-
upward or thinning-upward trends (on the scale of ten meters, or
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 475

Slice level

Transition point

FIG. 117.Cross section A) and horizon slice B) views of a deep-water leveed channel feeding a frontal splay, Gulf of Mexico. Note
the clear location of the transition point where the leveed channel transitions into a frontal splay. The frontal splay is characterized
by continuous to discontinuous high-amplitude seismic reflections in section view, and by extensive bifurcation and braiding (?),
forming a complex distributive network in plan view. Seismic data courtesy of WesternGeco.

a few tens of meters), and no readily recognizable channels. Beds
appear to be continuous for at least 200 m and of constant
thickness as far as they can be traced in the cliffs (Fig. 124). These
characteristics suggest that deposition was from unconfined
turbidity currents that could spread on the sea floor with no
apparent topographic obstructionscharacteristic of the distal
parts of frontal splays. The absence of bed-thickness trends
suggests that each turbidity current was unrelated to the previ-
ous ones, with the deposits of larger and smaller flows interbed-
ded. The overall upward bed thickening suggests seaward step-
ping of the frontal splay over a long period of time.
The upper Ross is characterized by thickening-upward pack-
ages, channel fills, slumpslide horizons, and turbidites with no
bed-thickness trends. Lien et al. (2003) showed that the thicken-
ing-upward packages (Figs. 125, 126) represented aggrading
spillover lobes developed during the lateral migration of chan-
nels. The top surfaces of many of the packages show giant
erosional features resembling flutes (Fig. 127). The packages are
superficially very similar to the thickening-upward sequences
first described by Mutti and colleagues (Mutti and Ricci Lucchi,
1972) and interpreted to result from the progradation of deposi-
tional lobes. However, in the case of the Ross thickening-upward
packages, the intimate relationship with channel fills suggests
that they are spillover (sediment waves?) rather than frontal
lobes.
The channels in the upper Ross vary from about 10 to over 25
m and consist dominantly of amalgamated thick-bedded turbid-
ites. No consistent patterns of bed-thickness change were ob-
served. Two areas define the relationships of channels to pack-
ages bestKilbaha Bay and Cloonconeen.
476 HENRY W. POSAMENTIER AND ROGER G. WALKER

Frontal
Frontal splay
splay

Flow direction

Transition point

Leveed channel
five km

FIG. 118.Horizon slice view of a deep-water frontal splay, Gulf of Mexico. The transition point is located where the levees associated
with the up-system leveed channel have diminished in height to below seismic resolution. The frontal splay is characterized by
extensive bifurcation, forming a complex distributive network. Seismic data courtesy of VeritasDGC.

At Kilbaha Bay (Fig. 123) almost horizontally dipping beds
are exposed for 1.5 km along the cliff, with a composite strati-
graphic thickness of about 40 m. There are 20 thickening-upward
packages (Fig. 128). In several locations the thick-bedded por-
tions of the packages can be traced laterally and seen to split into
non-amalgamated, thin-bedded turbidites with mudstone part-
ings between beds. The Kilbaha Bay channel (Fig. 128) has a
visible depth of incision of about 3 m, with a fill of at least 6 m. The
base is characterized by a layer of mudstone clasts, with at least
two stratified sandstone blocks up to 3.5 m long and 35 cm thick;
these are interpreted to represent channel-wall collapse, as de-
scribed above from Wheeler Gorge. Most importantly, on the
eastern side of the channel fill, thick-bedded amalgamated chan-
nel-fill sandstones split into thinner beds and grade laterally into
adjacent packages (Fig. 128). This relationship suggests that the
packages are in some way related to channel filling (and not lobe
progradation)the relationships are seen even better at
Cloonconeen Point.
The channel at Cloonconeen Point (Figs. 129, 130) has a
minimum fill thickness of 15 m. There appear to be at least three
separate incisions, and the fill consists of separate thick-bedded
turbidites close to the margin but an almost completely amalgam-
ated succession closer to the channel center. The uppermost part
of the fill consists of about 3 m of amalgamated sandstones, but
these beds can be walked out laterally for about 100 m (Fig. 130),
where they split progressively and pass into thin-bedded turbid-
ites separated by mudstone partings. These thin-bedded turbid-
ites are organized into a thickening-upward package identical to
the packages outside the channel. This relationship strongly
suggests that the thickening-upward packages represent aggra-
dation of spillover lobes rather than frontal splays developed
downstream from channel mouths. Both the Kilbaha and the
Cloonconeen channels suggest that the thickening-upward pack-
ages are closely related to channel filling. Lien et al. (2003) also
illustrate lateral-accretion surfaces, suggesting that the channels
can migrate laterally (Figs. 7173).
These relationships between architectural elements suggest
the following interpretation. Where the channel is far from a
particular depositional site, the only overbank deposits to reach
that site consist of mudstones (Fig. 131). During lateral channel
movement toward the site, overbank deposition may consist of
thin-bedded turbidites. The closer the channel approaches, the
thicker and more amalgamated the overbank succession become.
Closest to the channel, turbidity currents may be scouring (form-
ing the giant flutes) and bypassing without depositing.
The detailed relationships between all of the architectural
elements are shown in Figure 131. In phase 1 the channel is active
and is migrating laterally, forming lateral-accretions deposits on
one side and eroding a cut bank on the other side. In phase 2
during channel filling, note the lateral shift of the cut bank,
resulting in thicker and thicker overbank deposits.
Integrated Interpretation of the Ross Sandstone
Lien et al. (2003) suggested that the upper Ross consists
mainly of sinuous channels stacked into sinuous channel belts.
 North South
Levee crests
A

100
Slice level msec

Levee Levee
one km

Channel

North South
B Frontal splay margin
Frontal splay margin

Slice level

100
msec
A
A B
B
DEEP-WATER TURBIDITES AND SUBMARINE FANS

one km

one km

FIG. 119.Seismic reflection cross-section views through the frontal splay B) and the feeder leveed channel A) associated with the system shown in Figure 118. Note the
single leveed channel clearly shown in section view (A), contrasted with the high-amplitude continuous to discontinuous seismic reflection character of the frontal
splay. The section view through the frontal splay also shows that this system is characterized by numerous small channels that seem to coalesce into a sheet-like
morphology. Seismic data courtesy of VeritasDGC.
477
478 HENRY W. POSAMENTIER AND ROGER G. WALKER

Transition point

FIG. 120.Leveed channel feeding frontal
splay in the Bering Sea offshore Alaska
(from Kenyon and Millington, 1995).

Flow direction

5 km

Frontal splay

FIG. 121 (below).Leveed channel feeding frontal splay on the

mid-slope, eastern Gulf of Mexico. This 3D perspective
view illustrates a reflection amplitude map draped onto
the upper bounding surface of the channellevee complex.
Note that in the final stages of this channel-frontal splay
Leveed channel complex a solitary leveed channel flows across the top of
the frontal splay. Seismic data courtesy of VeritasDGC.

Flow direction

Transition point

Frontal
Frontal splay
splay
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 479

FIG. 122.Location map for the Ross Sandstone of western Ireland. Rectangle shows location of detailed map in Figure 123 (after
Lien et al., 2003).

FIG. 123.Detailed location map for the Ross Sandstone (yellow) and the overlying Gull Island Formation (brown), western
Ireland (see Fig. 122) (after Lien et al., 2003).
480 HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 124.Classical turbidites in the Lower Ross Sandstone. Note the bedding continuity and absence of any thickening- or
thinning-upward successions. Cliff north of Ballybunion, about 10 m high.

FIG. 125.Thickening-upward packages shown by yellow arrows in the upper Ross Sandstone, south side of Ross Bay. Note
succession from mudstones into thin-bedded turbidites, with thicker-bedded amalgamated turbidites in the upper parts of the
packages.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 481

Scoured surface

FIG. 126.Thickening-upward packages at Kilbaha Bay. Note the mudstones at the base, overlain by thin-bedded turbidites, overlain
in turn by amalgamated thick-bedded turbidites. The uppermost surface is scoured, but the scours are not visible in this picture.

11 m
m

FLOW DIRECTION

FIG. 127.Giant flute at Ross Bay. The shape of the flute is shown by the yellow arrows, and the flow direction by the red arrow.
The flute scours into amalgamated sandstones with a top surface covered in sinuous-crested ripples. The flute itself is partly
filled with mudstones and thin-bedded turbidites, with one surface showing straighter-crested ripples. There are no ripples
on the steeply dipping walls of the flute. Note the scalethe flute is nearly 2 m wide.
482 HENRY W. POSAMENTIER AND ROGER G. WALKER

LATERAL PASSAGE FROM THICK-BEDDED TO THIN-BEDDED TURBIDITES

10 m

CHANNEL GIANT SCOURS

FIG. 128.Cliff face at Kilbaha Bay. Beds can be walked out along the cliff and in the wave-cut platform. Mudstones and thin-bedded
turbidites are yellow, and amalgamated sandstones orange. Note lateral passage from amalgamated turbidites into thin-bedded
turbidites in several locations (blue arrows). Note also the giant scours (black arrows). The thick-bedded amalgamated turbidites
that fill the channel also pass laterally into thinner beds.

THICKENING-UPWARD PACKAGE LATERAL TO CHANNEL

THICKENING-UPWARD PACKAGES THICK-BEDDED SANDSTONES

CUT BY CHANNEL CHANNEL FILL

FIG. 129.Diagram of channel at Cloonconeen Point. Exposure is almost 100 percent, but details of the channel wall are partly
obscured by rock rubble and minor tectonics. Channel is filled with amalgamated structureless sandstones and cuts into adjacent
thickening-upward packages. Note that the uppermost amalgamated sandstones of the channel fill can be walked out laterally
into an overbank thickening-upward package.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 483

SANDIER-UPWARD PACKAGE

CHANNEL FILL

FIG. 130.Channel margin at Cloonconeen point. Photographer is standing in the middle of the channel fill (two yellow arrows). The
channel margin is outlined in yellow (see Fig. 129), and the lateral passage from thick amalgamated channel sandstones to a
thickening-upward package (red arrow) is shown with a dotted yellow arrow. Person for scale at end of dotted yellow arrow.

PHASE 1
Lateral-accretion distance Lateral-accretion distance

Lateral-accretion distance

Thin-bedded turbidites

Vertical
aggradation
mostly mudstone

Overbank spill during lateral Lateral migration with minimum Sandstones and mudstones on gently
migration of channels channel-floor aggradation dipping lateral accretion surface

Amalgamated beds
PHASE 2 Thick- to thin-bedded turbidites
Thin-bedded turbidites
Giant flutes within amalgamated sandstones Thick-bedded fill on Surface of erosional
Mudstones
and on erosional bypass surface cut-bank side of channel bypass with giant flutes Onlap of vertical
accretion deposits
Mudstone drape on
Ross Bay bypass surface

Clooconeen Rehy Cliff

Complete
sandier-upward package Rinevella Lateral-accretion deposits
Sandstones pass laterally into mudstones
as they onlap the lateral-accretion surface

FIG. 131.Composite diagram for the upper Ross Sandstone, showing the development of a channel. In phase 1, note the lateral
shifting, with lateral accretion on one side, and a cut bank on the opposite side. Thickening-upward sequences form as the
channel migrates, with mudstones deposited when the channel is far away (blue dotted line), thin-bedded turbidites as the
channel approaches (red dotted line), and thick-bedded amalgamated turbidites when the channel is closest (green dotted
line). As the channel migrates, the opposite bank may receive lateral accretion deposits. In phase 2, the main channel is filled
with structureless amalgamated turbidites, probably resembling those of Figure 68.
484 HENRY W. POSAMENTIER AND ROGER G. WALKER

The thickening-upward packages represented overbank spill,
and the non-packaged turbidites represented deposition on a
smooth basin floor far from any channels. The exact position of
the channels and channel belts was not discussed. The model is
shown in Figure 132, where three individual sinuous channels
are shown diagrammatically within a sinuous channel belt. At
the end of the sinuous channel belt, turbidity currents spread
laterally to form the unchannelized lower Ross turbidites.
In the light of the seismic images presented here, particularly
those that show channelization in the proximal parts of frontal
splays, it is possible that the entire Ross Sandstone represents a
frontal splay. The lower Ross would represent the smooth outer
portion of the splay, with the overall thickening-upward succes-
sion resulting from gradual progradation of the splay. The upper
Ross would then represent the proximal channelized part of the
splay, with the non-packaged parts of the upper Ross being
deposited on smoother parts of the splay.
Importance of the Ross Formation Case History.
The importance of this case history is that it illustrates how a
stratigraphic unit can first be subdivided into architectural ele-
mentsnon-packaged turbidites, thickening-upward packages,
channel fills, and slumpslide depositional units. The channels
are the easiest individual element to interpret, and the interpre-
tation of the thickening-upward packages follows from their
detailed relationships to the channel fills. Reliance on earlier
models that would suggest prograding depositional lobes would
give an alternative (and probably incorrect) interpretation. The
suggestion that the upper Ross might represent the proximal part

Area of non-packaged
and poorly packaged
tabular turbidites

Area of
non-packaged and
poorly packaged
tabular turbidites

Sinuous
Sinuous belt
belt
reestablished
reestablished after
after
sea-level
sea-level fluctuation
fluctuation Sinuous belt width ~ 5 km
(marine
(marine band)
band)
Sinuous belt of
channels and packages

FIG. 132.Overall interpretation of the Ross Sandstone. Individual sinuous channels in the upper Ross (shown as red, green and
blue) are stacked into sinuous channel belts (yellow). At curves in the individual channel, overbank spill can create thickening-
upward packages (red, green and blue lobes). Thus the sinuous channel belt consists of two main depositional elements
channel fills and overbank spills. At the downstream end, the channels lose their topography and feed a smooth basin plain
(the lower Ross). The passage from lower to upper Ross implies progradation of the sinuous channel belts over the smooth
basin plain.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
of a frontal splay is based on stratigraphic context and compari-
son with 3-D seismic images. Thus an interpretation has been
constructed that relies on defining architectural elements, defin-
ing their particular stratigraphic relationships, and comparing
this construction with seismic images of the deep sea floor. In this
way, we build our own model or interpretation without reliance
on a preexisting modelthe theme developed by Walker (this
volume) in the introduction to this volume.
Debris-Flow DepositsMass-Transport Complexes
Various forms of debris-flow and mass-transport deposits
comprise common depositional elements in many deep-sea envi-
ronments (Fig. 133). These deposits can assume a variety of
shapes and sizes. A characteristic that seems common to all is
their highly erosional nature and their contorted, chaotic, and
structureless internal architecture. Erosional relief at the bases of
such deposits can exceed 250 m in some instances (Figs. 134, 135).
Where clasts were embedded in the flow base, erosion of the
substrate is marked by long, linear striations or grooves that tend
to diverge in the downsystem direction (Figs. 21, 136140). Clast
sizes can range from cobbles to clasts larger than houses (Fig. 141).
In some instances, small outcrop-scale scour can also be observed
beneath mass-transport deposits (Fig. 142).
Large mass-transport deposits have been observed to origi-
nate at shelf edge (Fig. 143) as well as mid-slope locations (Fig.
144), where fault scarps as high as 54 m mark the point of slope
failure. Smaller mass-transport deposits originating on the flanks
100 msec
one km
MTD
FIG. 133.Seismic section from the eastern Gulf of Mexico illus-
trating the extensive nature of mass-transport deposits. At
this location, the mass-transport deposits constitute approxi-
mately 45% of the total section shown. Seismic data courtesy
of WesternGeco.
485
of salt domes and mud volcanoes also have been observed, and
on the flanks of turbidity-flow channel levees (Figs. 80, 86).
Because in most instances mass transport originates in low-
energy environments, the associated deposits of such processes
tend to be mud-rich.
In plan view these deposits can be lobate or channelized (Figs.
145148). In some instances mass-transport deposits can oppor-
tunistically use the channel of an earlier-formed turbidity-flow
channel (Fig. 149). The seismic character in both planar horizontal
or vertical section commonly is chaotic to contorted (Figs. 135,
150153), though in some instances they can be characterized by
large-scale convolute bedding (Fig. 154).
Where debris flows encounter obstructions or where flow
velocities diminish abruptly, as is common near their termini,
internal compressional structural features (i.e., low-angle thrust
faults) can be common (Figs. 151,155, 156). On the basin floor, a
commonly observed aspect of mass-transport deposits is an
apparent large-scale channelization that is not so much related to
erosion of a channel and subsequent fill as it is a consequence of
plowing of the sea floor by a plug of debris analogous to the
effect of a shovel pushing through a snow layer (Figs. 157, 158). In
instances such as these it is possible to calculate the travel distance
of the mass-transported debris (Fig. 159). Lateral compression
caused by multiple phases of mass-transport events also has been
observed (Figs. 152, 160).
The hallmark of the lithofacies of mass-transport deposits is
its lack of organization. Commonly they comprise mud-sup-
ported conglomerates, though they can also occur as pure mud-
stone containing muddy rip-up clasts. In isolated instances they
can also be relatively sand-rich (Jennette et al., 2000).
The upper bounding surfaces of mass-transport complexes
can vary from smooth to highly rugose. In some instances this
surface rugosity can exert an influence over subsequent turbidity
-flow deposits (Fig. 161), whereas in other instances, because the
short wavelength of the bathymetric lows, only ponding of
mudstones seems to occur (Fig. 162).
SEQUENCE STRATIGRAPHY
A typical deep-water depositional sequence on the basin floor
has been proposed by Posamentier and Kolla (2003a) as consist-
ing of basal debris-flow material, overlain by sand-rich frontal-
splay deposits, in turn overlain by isolated leveed-channel de-
posits and finally by debris-flow deposits and a condensed
section. This sequence is in part a distillation of the work of
Weimer (1991), Piper et al. (1997), Pirmez et al. (1997), Manley and
Flood (1998), Maslin et al. (1998), Beauboeuf and Friedmann
(2000), Brami et al. (2000), and Winker and Booth (2000) (Fig. 163).
It is relatively unlikely that each of these stratigraphic units
would be observed at any given location; debris-flow deposits are
most common on basin floors and within canyons, frontal splays
are most common on basin floors, and leveed channels are com-
mon in both slope and basin-floor environments. In contrast,
condensed sections are widespread and ubiquitous. Figure 74G
shows a section of a deep-water depositional sequence, character-
ized by frontal-splay deposits overlying a condensed section (thin
stratigraphic unit corresponding to the transgressive and high-
stand systems tract of the preceding sequence). In this example,
frontal-splay deposits gradually give way to an isolated leveed
channel. We suggest that this transition from frontal splay to
leveed channel during the waning phase of a deep-water deposi-
tional sequence occurs because of a progressive decrease in sand-
to-mud ratio within the flows, which originate at the shelf-edge
staging area. This diminished overall sand content results in a
seaward shift of the transition point as discussed above.
 A 486
Line C B

Line D

Flow direction one km

five km

South North
C

Slice level

100 msec
five km
West East
D
HENRY W. POSAMENTIER AND ROGER G. WALKER

Flow direction
Slice level

Terminal wall

FIG. 134.A) Mass-transport complex in the basin floor environment, eastern Gulf of Mexico, which lies within a broad, flat-floored channel C) that ends abruptly
at a terminal wall D). These sediments likely did not travel far but rather comprise a mass of material that was pushed from behind and slid to some degree
along a decollement surface at the base (note the relatively flat base in section view (C and D). The complex rheology of this deposit is illustrated by the chaotic
nature of the seismic reflections within the mass-transport complex as well as the tongues of sediments that extend beyond the terminal wall B). The relief of
the channel is approximately 240 m. Seismic data courtesy of WesternGeco.
 12.5 km

240 m

Flo
wd
irec
tion

30 km
DEEP-WATER TURBIDITES AND SUBMARINE FANS

Terminal wall

FIG. 135.Perspective-view image of the base of the mass-transport complex shown in Figure 134. Seismic data courtesy of WesternGeco.
487
488 HENRY W. POSAMENTIER AND ROGER G. WALKER

Flow
Flow direction
direction

FIG. 136.Grooves beneath mass-transport complex deposited
in a continental-slope environment (image courtesy of D.
Mosher).

FIG. 137 (below).Grooves beneath mass-transport deposit at

Five km the base Paleocene, North Sea, in a basin-floor environment
(after Wilson et al., 2005).

Flow
Flow direction
direction

10 km
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 489

one km

Grooves

FIG. 138.Dip-azimuth map of surface at the base of a channelized mass-transport deposit. Note the grooves that characterize this
surface and the tendency for these grooves to diverge down-system. This suggests that internal deformation characteristic of
flow rather than slide processes have occurred. Seismic data courtesy of WesternGeco.

Basal grooves

Outrunner blocks

one km
FIG. 139.Grooves at the base of a mass-transport complex, basin floor, eastern Gulf of Mexico. Note the outrunner blocks (yellow
arrows). Compare with Figure 154, a seismic slice through the middle of this mass-transport deposit. Seismic data courtesy of
WesternGeco.
490 HENRY W. POSAMENTIER AND ROGER G. WALKER

Line 1

100 msec

Line 2

1
2
Line 3 3
1 km

FIG. 140.Linear grooves at base of levees of slope channel shown in Figure 43. The events responsible for eroding these grooves
occurred just prior to levee construction, suggesting that these mass-transport events represent the earliest phases of a sea-level
lowstand. Seismic data courtesy of VeritasDGC.

Vertical section

Horizontal section

Debris Clasts Near Top of

Mass Transport Complex

52
0
m

FIG. 141.Seismic reflection time slice as well as section view through a mass-transport complex with a single megaclast over 500
m wide highlighted. This clast is observed to be rafted near the top of the mass transport complex.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 491

Debris-flow
deposit

Base of
debris flow

FIG. 142.Small-scale grooves and striations on top of sandstone
deposit at base of mass-transport complex, Borneo, Malaysia.

FIG. 143 (below).Shelf-edge-detachment slump scars offshore

Indonesia. These slump scars likely represent the point of
detachment or staging area of sediments that traveled down
the slope, possibly transforming from slump to slide to flow
with increased distance from the shelf edge.
5 km Shelf edge
Shelf edge

Slump scar

1 km
 492

Shelf edge Slump scar

A
B
Slope channel

Basin floor
C

Slump scar
HENRY W. POSAMENTIER AND ROGER G. WALKER

Scar Characteristics
Relief: 45 m
Width: ~ 16.4 km
Length: ~ 52 km
Area: 916 km2
Volume: 41.2 km3

FIG. 144.Perspective A), dip-azimuth B), and section C) views across the slope of the eastern Gulf of Mexico illustrating massive slump/slide scars and evacuation
of massive amounts of material onto the basin floor to the south of this area. Seismic data courtesy of VeritasDGC.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
1 km
FIG. 145.Debris flow off flank of submarine mud volcano.
The principal driver in deep-water sequence evolution is
relative sea-level changeprimarily the effect it has on character-
istics of the staging area, including sand-to-mud ratio, sediment
caliber, and depth of storm wave base. The stratigraphic, geomor-
phic, and sedimentologic expression of deep-water deposits is
directly linked to the conditions at the shelf edge. It is there that
sediments are delivered by rivers and other processes from
hinterland areas. The locus of sedimentation, or depocenter,
shifts seaward and landward as a direct result of sea-level fall or
rise, respectively. During sea-level fall, forced regression occurs
on the shelf (Posamentier et al., 1992) and depocenters shift
rapidly toward the shelf edge. Even before the shoreline reaches
the shelf edge, instability of the upper slope has been exacerbated
both by lowered wave base and by unloading of a wedge of water
493
(i.e., corresponding to the magnitude of sea-level fall). Another
factor that may play a role in destabilizing the slope at this time
is the shifting of oceanic currents that could accompany relative
falls of sea level. Such instability can commonly result in mass
failure in the mid- to upper-slope environment (Fig. 144) and
subsequent deposition of a mass-transport complex on the slope
and basin floor.
Ultimately, when river mouths are close enough to deliver
sediments directly to the outer shelf or upper slope, deep-water
turbidity currents become common across the slope and associ-
ated basin floor. During periods of relative sea-level fall, erosion
by incising rivers results in increased sediment delivery to the
shoreline, but more importantly increased sand-to-mud ratio of
sediments delivered there. Whether the sediments are delivered
directly from rivers by density underflow (i.e., hyperpycnal flow)
or are initially deposited and then later remobilized as slumps
transforming to slides and ultimately flows is not clear, and likely
both occur. At these times of relative sea-level fall, relatively
sand-rich flows tend to build relatively short leveed channels
transitioning into relatively large frontal splays.
When relative sea-level rise resumes, initially stationary shore-
lines with aggrading coastal deltas and plains (Fig. 57) and then
later transgressing shorelines result in a progressive decrease of
sands available for transport to the deep-water environment. At
these times coarser sands tend to be preferentially deposited
behind the shoreline within incised valleys, back-barrier lagoons,
and delta plains, producing gravity flows with potentially lower
sand-to-mud ratio late in the relative sea-level lowstand and
subsequent transgression. In response to this relative increase in
mud content within flow events, levee construction becomes
more efficient and leveed channels can extend significantly far-
ther basinward. The transition point is observed to shift signifi-
cantly farther seaward late in a sea-level cycle. This results in a
juxtaposition of frontal splays overlain by solitary leveed chan-
nels corresponding to early and late lowstand systems tract
times, respectively (Figs. 31, 74, 163, 164).
During subsequent periods of shoreline transgression, mini-
mal amounts of river-supplied sediments reach the shelf edge
and turbidite deposition largely ceases. Mass-transport deposi-
tion may continue at this time because of reequilibration of
residual oversteepened, upper-slope deposits associated with
lowstand shelf-edge delta deposition (Booth, 1979) or
oversteepening of canyon walls associated with lowstand canyon
or slope-channel cutting (Figs. 37, 165). A typical depositional
sequence within a canyon is illustrated in Figs. 37 and 165. In most
instances canyons are filled with relatively minor amounts of
channelized turbidites (Figs. 3739) and an overwhelming amount
of mass-transport deposits that constitute the bulk of the canyon
fill that forms during times when active turbidity flows have
ceased (Fig. 165).
CONCLUSIONS
The complexity of deep-water deposits is apparent from the
examples and discussions above. Large-scale, complex systems
are particularly difficult to distill into simple models that can act
as norms, predictors, and guides for future observations (Walker,
this volume).
In the 1970s, little of this complexity was understood, and
models such as those of Normark (1970, 1978), Mutti and Ricci
Lucchi (1972), Mutti and Ghibaudo (1972), and Walker (1978)
summarized observations from modern fans and ancient rocks,
and combined such observations into models based on modern
and ancient data (Walker, 1978). The models made little attempt
to incorporate data from turbidity-current processes (laboratory
 494

MTC
MTC lobes
channel

Line 1

Levees
Line 1
Deflated MTC lobe
HENRY W. POSAMENTIER AND ROGER G. WALKER

MTC
5 km

FIG. 146.Seascape of the lower slope and basin floor, DeSoto Canyon area, eastern Gulf of Mexico. Note the varying types of mass-transport deposits, including a
leveed mass-transport deposit, a lobate mass-transport deposit characterized by positive relief, and a flattened or deflated mass-transport deposit showing no
positive relief and characterized only by a debris field. Seismic data courtesy of VeritasDGC.
 A Knickpoint Channel depth = ~ 26 m
Channel width = ~ 1.2 km

Debr
is-flo
w ch
anne
l

Turbidity-current channel

B C
DEEP-WATER TURBIDITES AND SUBMARINE FANS

five
five km
km

FIG. 147.A) Dip-magnitude, B) curvature, and C) time-structure maps of a debris-flow channel. This channel is characterized by low sinuosity and the presence of
a well-defined knickpoint at its head. Seismic data courtesy of WesternGeco.
495
496 HENRY W. POSAMENTIER AND ROGER G. WALKER

A B Line 1
five km

MTC Lobes

n
tio
rec
Di
l w
Fo
Structural Ridge

South North FIG. 148.Illuminated sea floor based on 3D seismic data

A) and interpretation of mass-transport deposits B) on
C five km
the basin floor of the Makassar Strait, Indonesia. Sec-
tion view C) through several shallowly buried mass-
transport lobes of various age. Note that these elon-
gate lobes represent flow complexes imbedded within
a low-amplitude seismic reflection package sugges-
tive of hemipelagic to pelagic sedimentation. Because
of the draping effect of the hemipelagic and pelagic
sediments, these buried lobes all have sea-floor ex-
pression. Seismic data courtesy of WesternGeco.

FIG. 149.Seismic reflection horizon slice showing an

opportunistic mass-transport deposit within an
earlier-formed turbidity-current leveed channel.
Seismic data courtesy of WesternGeco.

one km
B

Flo
wd
irec
tion

one km
 South North
one km
A

Lobe 4
Slice level

100
msec
Lobe 1 Lobe 3

B
five km
DEEP-WATER TURBIDITES AND SUBMARINE FANS

Lobe 3
Flow direction

Lobe 2

Lobe 4

FIG. 150.Seismic reflection cross-section A) and plan-view B) images illustrating a succession of amalgamating mass-transport deposits. Note the typical chaotic to
transparent reflection character within these deposits, suggesting a severely disrupted and discontinuous stratigraphic architecture. Seismic data courtesy of
WesternGeco.
497
 Compressional ridges 498
A

five km

one km
100 msec
B
Flow direction

Flow
Flow Direction
Direction
Terminal wall

C D
HENRY W. POSAMENTIER AND ROGER G. WALKER

15

100 msec

one km
FIG. 151.Seismic reflection plan view A) and section views BD) through a mass-transport complex, Gulf of Mexico. The transverse lineaments are the plan-view
expression of thrust faults caused by the mass flow abutting against the terminal wall. Thrust faults are measured at approximately a 15 degree dip. Seismic data
courtesy of WesternGeco.
 A

5 km C
(approximate)

Event 3
DEEP-WATER TURBIDITES AND SUBMARINE FANS

Event 1

Mass-transport complex

Toe-thrust faults
Event 2
Lateral
compression
FIG. 152.A, B) Oblique and C) section views through an amalgamated mass-transport complex. At least three event units can be recognized. Internal deformation in
the form of compression-associated low-angle thrust faults can be observed within events 2 and 3 (B). Both toe thrusts as well as thrusts associated with lateral
compression caused by later flows are seen. Seismic data courtesy of WesternGeco.
499
 500

2 km

Nigeria

2 km

Indonesia
HENRY W. POSAMENTIER AND ROGER G. WALKER

2 km

Gulf of Mexico
FIG. 153.Cross-section view through three mass-transport deposits. Each is characterized by chaotic and contorted seismic facies pattern. Top line, seismic data
courtesy of VeritasDGC. Middle and lower lines, seismic data courtesy of WesternGeco.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
FIG. 154.Seismic horizon slice through mass-transport deposit characterized by convolute bedding, basin floor of the eastern Gulf of
Mexico. Compare with Figure 139, a seismic slice at the base of this mass-transport deposit. Seismic data courtesy of WesternGeco.
or modern ocean observations), and thus remained observational
rather than process based. In retrospect, we note that the Walker
(1978) model showed a very simplistic and commonly incorrect
distribution of sand and mud on the fan, and also ignored the
effects of different grain-size distributions, tectonic settings, and
relative sea-level fluctuations.
Relative sea-level fluctuations were built into the models of
Mutti (1985), Vail et al. (1977), and Posamentier and Vail (1985).
This model essentially combined three aspects of older Mutti
models into one evolving model, with sandy detached lobes
forming during lowstands of relative sea level. During relative
sea-level rise, the fan evolved into one where the channels were
attached to smaller depositional lobes. At highstand, with a much
reduced sediment supply, either condensed sections or finer-
grained channels and levee formed, with little or no lobe forma-
tion.
The 1980s marked the time when geophysical observations
became important, particularly side-scan sonar and shallow-
penetration seismic profiles across fan surfaces. This information
was very different in nature and scale from the rock observations
and poorer-quality side-scan data from the 1970s. Because the
data were so differentfor example, data on the meandering
channels from the Amazon Fan (Damuth et al., 1982a; Damuth et
al., 1982b) and the data on channel, levee, and large-scale slump
from the Rhne Fan (Droz and Bellaiche, 1985), it was difficult to
combine older models with this new datain fact, there are no
published attempts to do so.
At around the same time, 3-D seismic data from Tertiary and
Modern fans was becoming available to oil companies, but few of
these studies made their way into the public literature until the
1990s. The seismic data revealed superb three-dimensional rela-
tionships between all of the depositional elements, but again it
was difficult to relate seismic-scale data to outcrop data. With the
advent of seismic geomorphology (Posamentier and Kolla, 2003a)
the integration of such observations was made possible. Because
501
one km
the seismic data revealed so well the geometrical relationships
between elements, it became possible to infer and better under-
stand the processes that gave rise to the various depositional
geometriesprocesses such as channel meandering and avul-
sion, and lobe development by crevassing, overbank spill, or flow
expansion at the ends of channels.
Thus any attempts at modeling of turbidite systems must
incorporate information of four different types and scales: pro-
cesses (on the flume experimental scale as well as the scale of
modern fans), 3-D seismic studies, observations of ancient rocks,
and numerical models. There is still a rather incomplete linkage
between these different sources of data, and we have tried to
show how some of these linkages might be made. Our conclu-
sions are based in fact as much as possible but are nevertheless
somewhat process and conceptual in flavor. Future fan models
will evolve as more work is done relating experimental studies to
ancient rocks and relating the geometry of large outcrops to 3-D
seismic data derived both from modern fans and Cretaceous
Tertiary turbidite systems. As the seismic studies evolve, so the
large-scale processes of fan development will be better under-
stood, and the future conceptual basis for modeling will be better
established.
In this paper, we have integrated facies with depositional
elements and then discussed modern and subsurface 3-D seis-
mic examples to show how the depositional facies fit into the
various depositional systems. Because in most instances ground
truth is lacking in studies based on 3D seismic, we have inte-
grated information from outcrop examples to help provide
insights to assist in interpreting the 3-D seismic images (for
example, the San Clemente and Ross outcrop examples help to
calibrate the seismic images of meander loops; Figs. 51, 71, 72).
Rather than create a model that would serve as a template, we
have focused on first principles based largely on process sedi-
mentology. In this way, we leave it to geoscientists to build their
own models using depositional elements as building blocks in
 A B
B C
C 502
A A B A
A A'
A'

Terminal wall

Slice
C'
C' level
B'
B'

1 km

A' one km C
A'

Leveed channel
Turbidite
frontal splay

Mass-transport complex

D B B'
B B'

100 m
HENRY W. POSAMENTIER AND ROGER G. WALKER

E C
C C'
C'

one km

FIG. 155.Example of the terminus of a mass-transport deposit, eastern Gulf of Mexico. This deposit shows evidence of having been transported across a decollement
surface that likely was located within a condensed section at the base of a frontal-splay complex. In response to compression against a terminal wall, internal
deformation in the form of thrust faulting occurred. The plan-view A) as well as the section views BE) the clearly show the mass-transport unit entraining earlier-
deposited, sand-prone leveed-channel and frontal-splay deposits. Seismic data courtesy of WesternGeco.
 A B Transport
direction

Small thrust faults

C
D
DEEP-WATER TURBIDITES AND SUBMARINE FANS

Transport
direction

Transport
direction

FIG. 156.Seismic time slice in the coherence A) and amplitude B) domains illustrating the plan view and oblique view C) expression of thrust faults within mass-transport
deposits. D) Seismic reflection profile across mass transport deposits characterized by low-angle thrust faults. Seismic data courtesy of WesternGeco.
503
504 HENRY W. POSAMENTIER AND ROGER G. WALKER
Trench
Thrust faults
FIG. 157.Small-scale analog for mass-transport deposits shown
in Figures 134, 135, and 155. The snow shovel slides on a hard
base, pushing snow before it. The semi-rigid snow pack
deforms internally predominantly by low-angle thrust fault-
ing as the mass slides on the underlying decollement surface
and is compressed against the snow pack before it. The trench,
which is formed in the wake of the shovel, is characterized by
steep walls produced by shearing of the flowing mass.
space and time. On the basis of a sound understanding of local
physiography, sediment flux, and sediment caliber, integrated
with process sedimentology, predictive facies models can be
constructed. Lateral as well as vertical facies successions can be
predicted.
Sediment accumulates in staging areas, moves down can-
yons or slope channels, and ultimately travels onto the basin
floor through leveed channels and frontal splays. Principal
deep-water sand habitats include slope-channel complexes,
basin-floor channel fills, crevasse splays and sediment waves in
overbank settings, and frontal splays. Canyons also can be sites
of confined leveed channel deposits, though the bulk of canyon
fill commonly is mud prone. To a lesser extent, sand deposition
also occurs within levees in response to repeated spillover
across levee crests. Where slopes are characterized by high
rugosity, such as areas underlain by mobile salt or complex toe-
thrust ridges, deposition of a broad range of depositional ele-
ments can occur within relatively small areas. Intraslope basins
can be characterized by the full range of sand habitats, insofar as
they comprise, in microcosm, slope and basin-floor physiogra-
phy. Figures 166168 illustrate and summarize the various
significant sand habitats on the basin floor. Figures 169 and 170
illustrate the significant sand habitats on the slope.
The role of grain-size distribution within a flow as well as in
a succession of flows is critical to understanding the evolution of
depositional elements, especially those on the basin floor. The
more mud in the system, the farther seaward the leveed-channel
depositional element extends. As flows travel seaward they are
preferentially impoverished of their mud content through con-
tinual spillover of the upper parts of the flows into the overbank
environment. At the point where the levees are low enough to
allow sand from the higher-density, lower part of the flow to spill
over the levees, the system morphology transitions from leveed
channel to frontal splay. The frontal splays are commonly chan-
nelized in their proximal parts, but the channels are shallow and
have small levees. Channel paths probably switch rapidly. In the
distal parts of the frontal splays, flows become even less channel-
ized and more sheet-like.
Mass-transport deposits, largely debrites, are common com-
ponents of deep-water environments (Fig. 133). These deposits,
which are largely mud prone, commonly originate in mid-slope
to upper-slope environments. They commonly directly overlie
erosional surfaces characterized by long grooves and striations.
Near their termini as well as along their margins, they are
characterized by internal compressive deformation in the form of
imbricate thrust faults. Mass-transport complexes likely are char-
acterized by complex rheology that reflects frequent transitions
from turbulent to laminar to plug flow within a single flow
event.
In summary, we urge caution in adopting overly simple facies
models in deep-water environments. The deep water is a poten-
tially complex depositional setting. The degree of complexity of
the facies model desired should be dependent on the quality of
data available and should be built upon depositional elements
observed and inferred. However, as the geoscientist well knows,
the devil is in the details!
ACKNOWLEDGMENTS
The authors acknowledge Anadarko Petroleum Corporation
for permission to publish. Permission to publish seismic data
from VeritasDGC, Western Geophysical, PGS Geophysical is
gratefully acknowledged. We also wish to thank the various
colleagues with whom we have had lengthy conversations about
the world of deep-water sedimentation through the years. These
include V. Kolla, P. Weimer, W. Normark, H. DeV. Wickens, B.
Kneller, and W. Morris. This paper benefited from and was
significantly improved thanks to the comprehensive reviews of
Todd Greene and Octavian Catuneaunu.
REFERENCES
ABREU, V., SULLIVAN, M., PIRMEZ, C., AND MOHRIG, D., 2003, Lateral accretion
packages (LAPs): an important reservoir element in deep water
sinuous channels: Marine and Petroleum Geology, v. 20, p. 631648.
BALLANCE, P.F., 1964, The sedimentology of the Waitemata Group in the
Takapuna section, Auckland: New Zealand Journal of Geology and
Geophysics, v. 7, p. 466499.
BEAUBOEUF, R.T., AND FRIEDMANN, S.J., 2000, High-resolution seismic/
sequence stratigraphic framework for the evolution of Pleistocene
intra slope basins, Western Gulf of Mexico: depositional models and
reservoir analogs, in Weimer, P., Slatt, R.M., Coleman, J., Rosen, N.C.,
Nelson, H., Bouma, A.H., Styzen, M.J., and Lawrence, D.T., eds.,
A B

C D

Slice level
DEEP-WATER TURBIDITES AND SUBMARINE FANS

FIG. 158.Seismic time slice and section views A, B, D) of mass-

transport deposits characterized by thrust faults induced by
compressional stress regime. The interpreted thrust faults C)
formed as a result of country rock being pushed and com-
pressed by stress directed from the right by a mass flow E).
505
 Distance traveled = 6.6 km 506
[Assumption: 30 faults @ 220 m/fault]

Terminal wall

Restoration of fault results in 220 m of lengthening Distance Traveled = 0 km

100 msec
HENRY W. POSAMENTIER AND ROGER G. WALKER

Terminal wall
one km

FIG. 159.Seismic profile through mass-transport deposit shown in Figure 149. Approximate palinspastic reconstruction results in 220 m of lengthening for each thrust
fault. With approximately 30 faults along an axial profile within the study area, restoration results in extension of approximately 6.6 km at the up-system limit of
the deposit, whereas it can be assumed that the sediments near the terminal wall would have moved only minimally. Seismic data courtesy of WesternGeco.
 A Turbidite channel B

Lateral
compression

MTD Transport
margin direction

C
DEEP-WATER TURBIDITES AND SUBMARINE FANS

Lateral
compression

FIG. 160.A) Illuminated horizon at the top of a mass-transport deposit illustrating a lateral compressional bulge. B, C) Seismic reflection profile illustrating lateral
compression associated with a later phase of mass transport.
507
508 HENRY W. POSAMENTIER AND ROGER G. WALKER

A Transition point

Transition point

Onlap of frontal splay

Transport
direction

FIG. 161.Influence on turbidites by rugosity atop mass-transport deposits. A) Seismic time slice that shows the transition point
of a frontal splay, with extensive frontal-splay deposits seaward of that location. B) Seismic section that illustrates the onlap
of frontal-splay deposits against a bathymetric high associated with the irregular top of a mass-transport complex. Seismic data
courtesy of WesternGeco.

Deep-Water Reservoirs of the World: Gulf Coast Section SEPM, 20th
Annual Research Conference, p. 4060.
BOOTH, J.S., 1979, Recent history of mass-wasting on the upper continental
slopes, northern Gulf of Mexico, as interpreted from the consolidation
states of the sediment, in Doyle, L.J., and Pilkey, O.H., eds., Geology
of Continental Slopes: SEPM, Special Publication 27, p. 153165.
BOUMA, A.H., 1962, Sedimentology of Some Flysch Deposits; A Graphic
Approach to Facies Interpretation: Amsterdam, Elsevier, 168 p.
BOUMA, A.H., NORMARK, W.R., AND BARNES, N.E., eds., 1985, Submarine Fans
and Related Turbidite Systems: New York, Springer-Verlag, 351 p.
BRAMI, T.R., PIRMEZ, C., ARCHIE, C., HEERALAL, S., AND HOLMAN, K.L., 2000,
Late Pleistocene deep-water stratigraphy and depositional processes,
offshore Trinidad and Tobago, in Weimer, P., Slatt, R.M, Coleman, J.,
Rosen, N.C., Nelson, H., Bouma, A.H., Styzen, M.J., and Lawrence,
D.T., eds., Deep-Water Reservoirs of the World: Gulf Coast Section
SEPM Foundation, 20th Annual Research Conference, p. 104115.
BRUHN, C.H.L., AND WALKER, R.G., 1995, High-resolution stratigraphy and
sedimentary evolution of coarse-grained canyon-filling turbidites
from the Upper Cretaceous transgressive megasequence, Campos
Basin, Brazil: Journal of Sedimentary Research, v. B65, p. 426442.
 A B

C Scarp relief up to 22 m
D
Relief up to 32 m
DEEP-WATER TURBIDITES AND SUBMARINE FANS

one km

FIG. 162.Rugose upper surface of a mass-transport deposit. Small-scale accommodation atop the mass-transport deposits shows no apparent influence on
subsequent turbidite deposition.
509
 510

B
HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 163.A) Seismic reflection profile across deep-water deposits on the basin floor of the Makassar Strait, Indonesia, illustrating a stratigraphic succession with mass-
transport deposits at the base, overlain by frontal-splay deposits, leveed-channel deposits, a further mass-transport deposit and ultimately a condensed-section
deposit. B) Idealized cross section and well logs through a deep-water depositional sequence (after Posamentier and Kolla, 2003a). Seismic data courtesy of
WesternGeco.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS 511

Relative sea level

High

Lowstand interval

Low Time

Condensed section
Interval that shoreline is
located near the shelf edge
(turbidity currents dominate)
Interval of upper-slope
instability (mass-transport dominates)
Leveed channels dominate
(relatively low sand:mud)
Frontal splays dominate
(relatively high sand:mud)

Interval of upper-slope instability Condensed

(mass-transport dominates) section

FIG. 164.Schematic depiction of sediment transport events associated with relative sea-level change. As sea level begins its fall,
lowered wave base results in slope disequilibrium conditions, which favor mass-transport events. Once sea level has lowered
to the point where river mouths are in close proximity to the shelf edge, direct and indirect delivery of turbidites to the slope
and basin floor is facilitated. During the early stages of this process, associated with the interval of relative sea-level fall, shelf
valleys are incised and canyons can form. This results in sediment bypass of the shelf, which favors delivery of a relatively sand-
prone sediment load to the deep water. During the late stages of sea-level lowstand, when sea level is slowly rising, sediments
(preferentially coarse-grained sediment) tend to be trapped within earlier-formed incised valleys, resulting in muddier
turbidites at that time. This progression favors an evolution from sand-rich frontal splays to mud-rich isolated leveed channels.
Finally rapid sea-level rise again is associated with slope disequilibrium and deposition of mass-transport deposits. When sea
level finally stabilizes, background deposition of hemipelagic and pelagic sediments dominates in the deep-water environ-
ment, forming a condensed section.

BRUHN, C.H.L., AND WALKER, R.G., 1997, Internal architecture and sedi-
mentary evolution of coarse-grained turbidite channellevee com-
plexes, Early Eocene Regencia Canyon, Espirito Santo Basin, Brazil:
Sedimentology, v. 44, p. 1746.
CAMACHO, H., BUSBY, C.J., AND KNELLER, B., 2002, A new depositional model
for the classical turbidite locality at San Clemente State Beach, Califor-
nia: American Association of Petroleum Geologists, Bulletin, v. 86, p.
15431560.
CAMPION, K.M., SPRAGUE, A.R., MOHRIG, D., LOVELL, R.W., DRZEWIECKI, P.A.,
SULLIVAN, M.D., ARDILL, J.A., JENSEN, G.N. AND SICKAFOOSE, D.K., 2000,
Outcrop expression of confined channel complexes, in Weimer, P.,
Slatt, R.M, Coleman, J., Rosen, N.C., Nelson, H., Bouma, A.H., Styzen,
M.J., and Lawrence, D.T., eds., Deep-Water Reservoirs of the World:
Gulf Coast Section SEPM Foundation, 20th Annual Research Confer-
ence, p. 127150.
CLIFTON, H.E., 1981, Submarine canyon deposits, Point Lobos, California,
in Frizzel, E., ed., Upper Cretaceous and Paleocene Turbidites, Cen-
tral California Coast: Pacific Section SEPM, Guidebook to Field Trip
no. 6, p. 7992.
CLIFTON, H.E., 1984, Sedimentation units in stratified resedimented con-
glomerate, Paleocene submarine canyon fill, Point Lobos, California,
in Koster, E.H., and Steel, R.J., eds., Sedimentology of Gravels and
Conglomerates: Canadian Society of Petroleum Geologists, Memoir
10, p. 429-441.
 512
10 km

450 m

1170 m
Mass-transport deposits

18

Seismic section
HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 165.Seismic section across Mississippi canyon, along with perspective shaded relief. Seismic data proprietary to PGS Marine Geophysical NSA.
 Mud plug

Blocky to fining upward

50 m
one km Amalgamated
channel fill
DEEP-WATER TURBIDITES AND SUBMARINE FANS

FIG. 166.Basin-floor leveed-channel sand habitatschannel. Seismic data courtesy of WesternGeco.

513
 514

50 m

one km

Crevasse splay
Sediment waves
Thin-bedded overbank
HENRY W. POSAMENTIER AND ROGER G. WALKER

FIG. 167.Basin-floor leveed-channel sand habitatslevee. Seismic data courtesy of WesternGeco.

 50 m
one km

Frontal splay
Frontal splay
DEEP-WATER TURBIDITES AND SUBMARINE FANS

FIG. 168.Basin-floor frontal-splay sand habitats. Seismic data courtesy of WesternGeco.

515
 516

Onlap at channel wall with shale drape

HENRY W. POSAMENTIER AND ROGER G. WALKER

Slope-channel lateral-accretion package

Slope-channel amalgamated fill

FIG. 169.Slope-channel sand habitats. Seismic data courtesy of VeritasDGC.
 Canyon filled with predominantly mass transport deposits

one km

one km

Partially preserved channel

(erosional remnant) Coarse-grained canyon-axis fill
DEEP-WATER TURBIDITES AND SUBMARINE FANS

one km

Axial leveed channel Canyon-fill onlap onto canyon wall

FIG. 170.Slope-canyon sand habitats. Seismic data proprietary to PGS Marine Geophysical NSA.
517
518 HENRY W. POSAMENTIER AND ROGER G. WALKER
CROWELL, J.C., 1957, Origin of pebbly mudstones: Geological Society of
America, Bulletin, v. 68, p. 9931009.
DAMUTH, J.E., KOWSMANN, R.O., FLOOD, R.D., BELDERSON, R.H., AND GORINI,
M.A., 1983a, Age relationships of distributary channels on Amazon
deep-sea fan: implications for fan growth pattern: Geology, v. 11, p.
470473.
DAMUTH, J.E., KOLLA, V., FLOOD, R.D., KOWSMANN, R.O., GORINI, M.A., AND
BELDERSON, R.H., 1983b, Distributary channel meandering and bifur-
cation patterns on the Amazon deep-sea fan as revealed by long-
range side-scan sonar (GLORIA): Geology, v. 11, p. 9498.
DROZ, L., AND BELLAICHE, G., 1985, Rhne deep-sea fan: morphostructure
and growth pattern: American Association of Petroleum Geologists,
Bulletin, v. 69, p. 460479.
DZULYNSKI, S., KSIAZKIEWICZ, M., AND KUENEN, P.H., 1959, Turbidites in
flysch of the Polish Carpathian Mountains: Geological Society of
America, Bulletin, v. 70, p. 10891118.
ELMORE, R.D., PILKEY, O.H., CLEARY, W.J., AND CURRAN, H.A., 1979, Black
Shell turbidite, Hatteras Abyssal Plain, western Atlantic Ocean:
Geological Society of America, Bulletin, v. 90, p. 11651176.
GHIBAUDO, G., AND VANZ, V., 1987, Proposta di classificazione delle facies
torbiditiche e di un metodo practico per la loro descrizione sul
terreno: Giornale di Geologia, v. 49, p. 31/43.
GOODWIN, R.H., AND PRIOR, D.B., 1989, Geometry and depositional se-
quences of the Mississippi Canyon, Gulf of Mexico: Journal of
Sedimentary Petrology, v. 59, p. 318329.
GREGORY, M.R., 1966, Sedimentary features and penecontemporaneous
slumping in the Waitemata Group, Whangaparaoa Peninsula, north
Auckland, New Zealand: New Zealand Journal of Geology and
Geophysics, v. 12, p. 248282.
GRIGGS, G.B., AND KULM, L.D., 1970, Sedimentation in Cascadia deep-sea
channel: Geological Society of America, Bulletin, v. 81, p. 13611384.
HACKBARTH, C.J., AND SHAW, R.D., 1994, Morphology and stratigraphy of
a mid-Pleistocene turbidite leveed channel from seismic, core and
log data, northeastern Gulf of Mexico, in Weimer, P., Bouma, A.H.,
and Perkins, B.F., eds., Submarine Fans and Turbidite Systems: Gulf
Coast Section SEPM Foundation, 15th Annual Research Conference,
p. 127133.
H ARMS, J.C., AND F AHNESTOCK, K., 1965, Stratification, flow phenom-
ena and bed forms (with an example from the Rio Grande), in
Middleton, G.V., ed., Primary Sedimentary Structures and Their
Hydrodynamic Interpretation: SEPM , Special publication 12, p.
84115.
HEEZEN, B.C., 1956, Corrientes de turbidez del Rio Magdalena: Sociedad
Geogrfica de Colombia, Boletn, v. 51/52, p. 135143.
HEEZEN, B.C., MENZIES, R.J., SCHNEIDER, E.D., EWING W.M., AND GRANELLI,
N.C.L., 1964, Congo submarine canyon: American Association of
Petroleum Geologists, Bulletin, v. 48, p. 11261149.
HISCOTT, R.N., HALL, F.R., AND PIRMEZ, C., 1997, Turbidity-current overspill
from the Amazon Channel: texture of the silt/sand load, paleoflow
from anisotropy of magnetic susceptibility and implications for
flow processes, in Flood, R.D., Piper, D.J.W., Klaus, A., and Peterson,
L.C., eds., Proceedings of the Ocean Drilling Program: Scientific
Results, v. 155, p. 5378.
HURST, A., CARTWRIGHT, J.A., DURANTI, D., HUUSE, M., AND NELSON, M.,
2005, Sand injectites: an emerging global play in deep-water clastic
environments, in Dor, A.G., and Vining, B.A., eds., Petroleum
Geology: Northwest Europe and Global Perspectives: Geological
Society of London, Proceedings of the 6th Petroleum Geology Con-
ference, p. 133144.
JENNETTE, D.C., GARFIELD, T.R., MOHRIG, D.C., AND CAYLEY, G.T., 2000, The
interaction of shelf accommodation, sediment supply and sealevel
in controlling the facies, architecture and sequence stacking pat-
terns of the Tay and Forties/Sele basin-floor fans, Central North Sea,
in Weimer, P., Slatt, R.M, Coleman, J., Rosen, N.C., Nelson, H.,
Bouma, A.H., Styzen, M.J., and Lawrence, D.T., eds., Deep-Water
Reservoirs of the World: Gulf Coast Section SEPM Foundation, 20th
Annual Research Conference, p. 402421.
KENYON, N.H., AND MILLINGTON, J., 1995, Contrasting deep-sea deposi-
tional systems in the Bering Sea, in Pickering, K.T., Hiscott, R.N.,
Kenyon, N.H., Ricci Lucchi, F., and Smith, R.D.A., eds., Atlas of Deep
Water environments: London, Chapman & Hall, p. 196202.
KNELLER, B.C., 1995, Beyond the turbidite paradigm: physical models
for deposition of turbidites and their implications for reservoir
prediction, in Hartley, A.J., and Prosser, D.J., eds., Characterisation
of Deep Marine Clastic Systems: Geological Society of London,
Special Publication 94, p. 3149.
KOLLA, V., AND COUMES, F., 1987, Morphology, internal structure, seis-
mic stratigraphy and sedimentation of Indus Fan: American Asso-
ciation of Petroleum Geologists, Bulletin, v. 71, p. 650677.
KOLLA, V., BOURGES, P., URRUTY, J.M., AND SAFA, P., 2001, Evolution of
deepwater Tertiary sinuous channels offshore, Angola (West Af-
rica) and implications to reservoir architecture: American Associa-
tion of Petroleum Geologists, Bulletin v. 85, p. 13731405.
KOMAR, P.D., 1971, Hydraulic jumps in turbidity currents: Geological
Society America, Bulletin, v. 82, p. 14771488.
KUENEN, P.H., AND MIGLIORINI, C.I., 1950, Turbidity currents as a cause
of graded bedding: Journal of Geology, v. 58, p. 91127.
LIEN, T., WALKER, R.G., AND MARTINSEN, O.J., 2003, Turbidites in the
Upper Carboniferous Ross Formation, western Ireland: recon-
struction of a channel and spillover system: Sedimentology, v. 50,
p. 113-148.
LINK, M.H., SQUIRES, R.L., AND COLBURN, I.P., 1984, Slope and deep-sea
fan facies and paleogeography of Upper Cretaceous Chatsworth
Formation, Simi Hills, California: American Association of Petro-
leum Geologists, Bulletin, v. 68, p. 850873.
LOWE, D.R., 1975, Water escape structures in coarse-grained sedi-
ments: Sedimentology, v. 22, p. 157204.
LOWE, D.R., 1979, Stratigraphy and sedimentology of the Pigeon point
Formation, San Mateo County, California, in Nilsen, T.H., and
Brabb, E.E., eds., Geology of the Santa Cruz Mountains, California:
Geological Society of America, Cordilleran Section, Field Trip
Guidebook, p. 1729.
LOWE, D.R., 1982, Sediment gravity flows: II. Depositional models with
special reference to the deposits of high-density turbidity currents:
Journal of Sedimentary Petrology, v. 52, p. 279297.
MANLEY, L., AND FLOOD, R.D., 1998, Cyclic sediment deposition within
Amazon deep-sea fan: American Association of Petroleum Geolo-
gists, Bulletin, v. 72, p. 912925.
MARTINSEN, O.J., LIEN, T., AND WALKER, R.G., 2000, Upper Carboniferous
deep-water sediments, western Ireland: analogs to passive margin
plays, in Weimer, P., Slatt, R.M., Coleman, J., Rosen, N.C., Nelson,
H., Bouma, A.H., Styzen, M.J., and Lawrence, D.T., eds., Deep
Water Reservoirs of the World: SEPM, Gulf Coast Section, 20th
Annual Bob F. Perkins Research Conference, Houston, p. 533555
(available only on CD).
MASLIN, M., MIKKELSEN, N., VILELA, C., AND HAQ, B.U., 1998, Sea-level
and gas-hydratecontrolled catastrophic sediment failures of the
Amazon fan: Geology, v. 26, p. 11071110.
MAYALL, M., AND STEWART, I., 2000, The architecture of turbidite slope
channels, in Weimer, P., Slatt, R.M, Coleman, J., Rosen, N.C.,
Nelson, H., Bouma, A.H., Styzen, M.J., and Lawrence, D.T., eds.,
Deep-Water Reservoirs of the World: Gulf Coast Section SEPM
Foundation, 20th Annual Research Conference, p. 578586.
MIGEON, S., SAVOYE, B., AND FAUGERES, J.-C., 2000, Quaternary develop-
ment of migrating sediment waves in the Var deep-sea fan: distri-
bution, growth pattern and implication for levee evolution: Sedi-
mentary Geology, v. 133, p. 265293.
MOHRIG, D., WHIPPLE, K.X., HONDZO, M., ELLIS, C., AND PARKER, G., 1998,
Hydroplaning of subaqueous debris flows: Geological Society of
America, Bulletin, v. 110, p. 387394.
 DEEP-WATER TURBIDITES AND SUBMARINE FANS
MULDER, T., AND SYVITSKI, J.P.M., 1995, Turbidity currents generated at
river mouths during exceptional discharges to the world oceans:
Journal of Geology, v. 103, p. 285299.
MULDER, T., SYVITSKI, J.P.M., AND SKENE, K.I., 1998, Modeling of erosion
and deposition by turbidity currents generated at river mouths:
Journal of Sedimentary Research, v. 68, p. 124137.
MUTTI, E., 1985, Turbidite systems and their relations to depositional
sequences, in ZUFFA, G.G., ed., Provenance of Arenites: NATO-ASI
Series, Dordrecht, The Netherlands, Reidel, p. 6593.
MUTTI, E., AND GHIBAUDO, G., 1972, Un esempio di torbiditi di conoide
sottomarina esterna: le Arenario di San Salvatore (formazione di
Bobbio, Miocene) nellAppennino di Piacenza: Memorie
dellAccademia delle Scienze di Torino, Classe di Scienze Fisiche,
Matematiche e Naturale, Serie 4a, no. 16, p. 141.
MUTTI, E., AND NORMARK, W.R., 1991, An integrated approach to the study
of turbidite systems, in Weimer, P., and Link, M.H., eds., Seismic
Facies and Sedimentary Processes of Submarine Fans and Turbidite
Systems: New York, Springer-Verlag, p. 75106.
MUTTI, E., AND RICCI LUCCHI, F., 1972, Le torbiditi dellAppennino
settentrionale: introduzioni allanalisi de facies: Societ Geologica
Italiana Memorie, v. 11, p. 161199.
MUTTI, E., STEFFENS, G.S., PIRMEZ, C., ORLANDO, M., AND ROBERTS, D., EDS.,
2003, Thematic set: Turbidites: models and problems: Marine and
Petroleum Geology, v. 20, p. 523933.
NELSON, C.H., GOLDFINGER, C., JOHNSON, J.E., AND DUNHILL, G., 2000, Varia-
tion of modern turbidite systems along the subduction zone margin
of Cascadia Basin and implications for turbidite reservoir beds, in
Weimer, P., Slatt, R.M., Coleman, J., Rosen, N.C., Nelson, H., Bouma,
A.H., Styzen, M.J. and Lawrence, D.T., eds., Deep Water Reservoirs of
the World: SEPM, Gulf Coast Section, 20th Annual Bob F. Perkins
Research Conference, Houston, p. 714738 (available only on CD).
NILSEN, T.H., 1980, Modern and ancient submarine fans: discussion of
papers by R.G. Walker and W.R. Normark: American Association of
Petroleum Geologists, Bulletin, v. 64, p. 10941101.
NORMARK, W.R., 1970, Growth patterns of deep-sea fans: American Asso-
ciation of Petroleum Geologists, Bulletin, v. 54, p. 21702195.
PEAKALL, J., MCCAFFREY, W.D., KNELLER, B.C., STELTING, C.E., MCHARGUE,
T.R., AND SCHWELLER, W.J., 2000, A process model for the evolution of
submarine fan channels: Implications for sedimentary architecture, in
Bouma, A.H., and Stone, C.G., eds., Fine-Grained Turbidite Systems:
American Association of Petroleum Geologists, Memoir 72 and SEPM,
Special Publication 68, p. 7388.
PICKERING, K.T., STOW, D., WATSON, M., AND HISCOTT, R.N., 1986, Deep-
water facies, processes and models: a review and classification scheme
for modern and ancient sediments: Earth-Science Reviews, v. 23, p.
75174.
PILKEY, O.H., 1988, Basin plains: giant sedimentation events, in Clifton,
H.E., ed., Sedimentologic Consequences of Convulsive Geological
Events: Geological Society of America, Special Paper 229, p. 9399.
PIPER, D.J.W., AND NORMARK, W.R., 1983, Turbidite depositional patterns
and flow characteristics, Navy submarine fan, California borderland:
Sedimentology, v. 30, p. 681694.
PIPER, D.J., SHOR, A.N., AND HUGHES-CLARK, J.E., 1988, The 1929 Grand
Banks earthquake, slump and turbidity current, in Clifton, H.E., ed.,
Sedimentological Consequences of Convulsive Geological Events:
Geological Society of America, Special Paper 229, p. 7792.
PIPER, D.J.W., PIRMEZ, C., MANLEY, P.L., LONG, D., FLOOD, R.D., NORMARK,
W.R., AND SHOWERS, W., 1997, Mass-transport deposits of the Amazon
Fan, in Flood, R.D., Piper, D.J.W., Klaus, A., and Peterson, L.C., eds.,
Proceedings of the Ocean Drilling Program, Scientific Results, v. 155,
p. 109146.
PIRMEZ, C., AND FLOOD, R.D., 1997, Morphology and structure of Amazon
Channel, in Flood, R.D., Piper, D.J.W., Klaus, A., and Peterson, L.C.,
eds., Proceedings of the Ocean Drilling Program, Scientific Results, v.
155, p. 2345.
519
PIRMEZ, C., HISCOTT, R.N., AND KRONEN, J.K., 1997, Sandy turbidite succes-
sions at the base of channellevee systems of the Amazon Fan re-
vealed by FMS logs and cores: unraveling the facies architecture of
large submarine fans, in Flood, R.D., Piper, D.J.W., Klaus, A., and
Peterson, L.C., eds., Proceedings of the Ocean Drilling Program,
Scientific Results, v. 155, p. 722.
PLINT, A.G., AND NUMMEDAL, D., 1998, The falling stage systems tract:
recognition and importance in sequence stratigraphic analysis, in
Gawthorpe, R.L.G., and Hunt, E., eds., Sedimentary Responses to
Forced Regression: Geological Society of London, Special Publica-
tion.
POSAMENTIER, H.W., 2003a, A linked shelf edge delta and slope channel
turbidite system: 3D seismic case study from the eastern Gulf of
Mexico, in Roberts, H.H., Rosen, N.C., Fillon, R.H., and Anderson,
J.B., eds., Shelf Margin Deltas and Linked Down Slope Petroleum
Systems: Global Significance and Future Exploration Potential: Gulf
Coast Section of SEPM Foundation, Proceedings of the 23rd Annual
Bob F. Perkins Research Conference, p. 115134.
POSAMENTIER, H.W., 2003b, Depositional elements associated with a basin
floor channellevee system: case study from the Gulf of Mexico:
Marine and Petroleum Geology, v. 20, p. 677690.
POSAMENTIER, H.W., AND ALLEN, G.P., 1999, Siliciclastic Sequence Stratigra-
phyConcepts and Applications: SEPM, Concepts in Sedimentology
and Paleontology, no. 7, 210 p.
POSAMENTIER, H.W., ALLEN, G.P., JAMES, D.P., AND TESSON, M., 1992, Forced
regressions in a sequence stratigraphic framework: concepts, ex-
amples, and exploration significance: American Association of Petro-
leum Geologists, Bulletin, v. 76, p. 16871709.
POSAMENTIER, H.W., ERSKINE, R.D., AND MITCHUM, R.M., JR., 1991, Models for
submarine fan deposition within a sequence stratigraphic frame-
work, in Weimer, P., and Link, M.H., eds., Seismic Facies and Sedi-
mentary Processes of Submarine Fans and Turbidite Systems: New
York, Springer-Verlag, p. 127136.
POSAMENTIER, H.W., JERVEY, M.T., AND VAIL, P.R., 1988, Eustatic controls on
clastic deposition Iconceptual framework, in Wilgus, C.K., Hastings,
B.S., Kendall, C.G.St.C., Posamentier, H.W., Ross, C.A., and Van
Wagoner, J.C., eds., Sea Level Changes: An Integrated Approach:
SEPM, Special Publication 42, p. 109124.
POSAMENTIER, H.W., AND KOLLA, V., 2003a, Seismic geomorphology and
stratigraphy of depositional elements in deep-water settings: Journal
of Sedimentary Research, v. 73, p. 367388.
POSAMENTIER, H.W., AND KOLLA, V., 2003b, Anatomy of a deep-water
channel avulsionExample from the basin floor of the Desoto Can-
yon area, Gulf of Mexico (abstract): American Association of Petro-
leum Geologists, Annual Meeting, Abstracts Volume, p. A140.
POSAMENTIER, H.W., MEIZARWIN, WISMAN, P.S., AND PLAWMAN, T., 2000, Deep
water depositional systemsUltra-deep Makassar Strait, Indonesia,
in Weimer, P., Slatt, R.M., Coleman, J., Rosen, N.C., Nelson, H.,
Bouma, A.H., Styzen, M.J., and Lawrence, D.T., eds., Deep-Water
Reservoirs of the World: Gulf Coast Section SEPM Foundation, 20th
Annual Research Conference, p. 806816.
POSAMENTIER, H.W., AND VAIL, P.R., 1985, Eustatic controls on deposi-
tional stratal patterns: SEPM, Research Conference no. 6, Sea Level
ChangesAn Integrated Approach, October 2023, 1985 (Abstract
and Poster).
PRATHER, B.E., BOOTH, J.R., STEFFENS, G.S., AND CRAIG, P.A., 1998, Classifica-
tion, lithologic calibration and stratigraphic succession of seismic
facies of intraslope basins, deep-water Gulf of Mexico: American
Association of Petroleum Geologists, Bulletin, v. 82, p. 701728.
READING, H.G., AND RICHARDS, M., 1994, Turbidite system in deep-water
basin margins classified by grain size and feeder system: American
Association of Petroleum Geologists, Bulletin, v. 78, p. 792822.
RUST, B.R., 1966, Late Cretaceous paleogeography near Wheeler Gorge,
Ventura County, California: American Association of Petroleum
Geologists, Bulletin, v. 50, p. 13891398.
520 HENRY W. POSAMENTIER AND ROGER G. WALKER

SHANMUGAM, G., 1996, High-density turbidity currents: are they sandy Processes of Submarine Fans and Turbidite Systems: New York,
debris flows?: Journal of Sedimentary Research, v. 66, p. 210. Springer-Verlag, p. 323347.
SHANMUGAM, G., LEHTONEN, L.R., STRAUME, T., SYVERTSEN, S.E., HODGKINSON, WEIMER, P., AND LINK, M.H., eds., 1991, Seismic Facies and Sedimentary
R.J., AND SKIBELI, M., 1994, Slump and debris-flow dominated upper Processes of Submarine Fans and Turbidite Systems: New York,
slope facies in the Cretaceous of the Norwegian and northern North Springer-Verlag, 447 p.
Seas (6167N): implications for sand distribution: American Asso- WEIMER, P., SLATT, R.M., COLEMAN, J., ROSEN, N.C., NELSON, H., BOUMA,
ciation of Petroleum Geologists, Bulletin, v. 78, p. 910937. A.H., STYZEN, M.J., AND LAURENCE, D.T., eds., 2000, Deep-Water
SIKKIMA, W., AND WOJCIK, K.M., 2000, 3D visualization of turbidite sys- Reservoirs of the World: SEPM, Gulf Coast Section, 20th Annual Bob
tems, Lower Congo Basin, offshore Angola, in Weimer, P., Slatt, F. Perkins Research Conference, Houston (available only on CD).
R.M., Coleman, J., Rosen, N.C., Nelson, H., Bouma, A.H., Styzen, WILSON J., WALL, G., KLOOSTERMAN, H.J., CONEY, D., CAYLEY, G., WALKER, J.,
M.J., and Lawrence, D.T., eds., Deep-Water Reservoirs of the World: AND LINSKAILL, C., 2005, The discovery of Goldeneye: in Dor, A.G.,
Gulf Coast Section of SEPM Foundation, 20th Annual Research and Vining, B.A., eds., Petroleum Geology: North-West Europe and
Conference, p. 928939. Global Perspectives: 6th Petroleum Geology Conference, Proceed-
SURLYK, F., AND NOE-NYGAARD, N., 2003, A giant sand injection complex: ings, p. 199216.
the Upper Jurassic Hareelv Formation of East Greenland: Geologia WINKER, C.D., AND BOOTH, J.R., 2000, Sedimentary dynamics of the salt-
Croatica, v. 56, p. 6981. dominated continental slope, Gulf of Mexico: integration of obser-
UCHUPI, E., AND AUSTIN, J., 1979, The stratigraphy and structure of the vations from the sea floor, near-surface and deep subsurface, in
Laurentian Cone region: Canadian Journal of Earth Sciences, v. 16, Weimer, P., Slatt, R.M., Coleman, J., Rosen, N.C., Nelson, H., Bouma,
p. 17261752. A.H., Styzen, M.J., and Lawrence, D.T., eds., Deep-Water Reservoirs
VAIL, P.R., MITCHUM, R.M., JR., AND THOMPSON, S., III, 1977, Seismic of the World: SEPM, Gulf Coast Section, 20th Annual Bob F. Perkins
stratigraphy and global changes of sea level, part 3: relative changes Research Conference, Houston, p. 10591086.
of sea level from coastal onlap, in Payton, C.E., ed., Seismic Stratig-
raphyApplications to Hydrocarbon Exploration: American Asso-
ciation of Petroleum Geologists, Memoir 26, p. 6381.
VAN WAGONER, J.C., BEAUBOUEF, R.T., HOYAL, J.C.J.D., DUNN, P.A., ADAIR,
N.L., ABREU, V., LI, D., WELLNER, R.W., AWWILLER, D.N., AND SUN, T.,
2003, Energy dissipation and the fundamental shape of siliciclastic
sedimentary bodies (abstract): American Association of Petroleum
Geologists, Annual Meeting, Abstracts Volume, p. A175.
VAN WEERING, T.C.E., NIELSEN, T., KENYON, N.H., KATJA, A., AND KUIJPERS,
A.H., 1998, Large submarine slides on the NE Faeroe continental
margin, in Stoker, M.S., Evans, D., and Cramp, A., eds., Geological
Processes on Continental Margins: Sedimentation, Mass-Wasting,
and Stability: Geological Society of London, Special Publication 129:
p. 527.
WALKER, R.G., 1965, The origin and significance of the internal sedimen-
tary structures of turbidites: Yorkshire Geological Society, Proceed-
ings, v. 35, p. 132.
WALKER, R.G., 1967, Turbidite sedimentary structures and their relation-
ship to proximal and distal depositional environments: Journal of
Sedimentary Petrology., v. 37, p. 2543.
WALKER, R.G., 1973, Mopping up the turbidite mess, in Ginsburg, R.N.,
ed., Evolving Concepts in Sedimentology: Baltimore, The Johns
Hopkins Press, p. 137.
WALKER, R.G., 1975a, Generalized facies models for resedimented con-
glomerates of turbidite association: Geological Society of America,
Bulletin, v. 86, p. 737748.
WALKER, R.G., 1975b, Nested submarine channels at San Clemente,
California: Geological Society of America, Bulletin, v. 86, p. 915
924.
WALKER, R.G., 1978, Deep-water sandstone facies and ancient submarine
fans: models for exploration and stratigraphic traps: American
Association of Petroleum Geologists, Bulletin, v. 62, p. 932966.
WALKER, R.G., 1985, Mudstones and thin-bedded turbidites associated
with the Upper Cretaceous Wheeler Gorge conglomerates, Califor-
nia: a possible channellevee complex: Journal of Sedimentary
Petrology, v. 55, p. 279290.
WALKER, R.G., 1992, Turbidites and submarine fans, in Walker, R.G., and
James, N.P., eds., Facies Models; Response to Sea Level Change:
Geological Association of Canada, p. 239263.
WEIMER, P., 1989, Sequence stratigraphy of the Mississippi Fan (Plio-
Pleistocene), Gulf of Mexico: Geo-Marine Letters, v. 9, p. 185272.
WEIMER, P., 1991, Seismic facies, characteristics and variations in channel
evolution, Mississippi Fan (Plio-Pleistocene), Gulf of Mexico, in
Weimer, P., and Link, M.H., eds., Seismic Facies and Sedimentary
 INDEX 521
Index
bioturbation 57, 60, 119, 133134, 179, 195, 201, 203, 207, 212,
A 249, 251, 253, 258259, 262, 265, 311312, 314, 318, 320, 336,
abrasion 23, 137 350, 363, 369, 371, 385, 389, 431
Abu Dhabi 40 Blackhawk Formation 303304, 321, 333
accommodation 22, 58, 60, 63, 171, 199, 211, 344, 350351 Bouma sequence 1213, 369, 399, 403404, 407
accumulation rate 52, 66, 71 boundary layer 2223
accumulation surface 32, 50, 5253, 55, 57, 6668, 74 bounding surface 13, 15, 1920, 22, 32, 34, 3840, 48, 57, 60, 63,
adhesion plane beds 32 67, 73, 159, 179, 453, 478, 485
adhesion ripples 32, 39 braid delta 256
adhesion strata 30, 32, 38 braiding 91, 9698, 102, 112, 138, 475
adhesion warts 32, 39 Brazil 62, 252, 256, 361, 427428, 430432
aerodynamic configuration 53, 55, 60 breaker bars 295, 299300, 311
Al Liwa sand sea 40 brinkline 31
algal growth 33 brittle failure 33
allocyclic 19, 175, 179, 189, 237, 281, 420 burrow 33, 4547, 65, 107108, 122123, 126, 129134, 193, 203,
allogenic 50, 60, 67, 73 219, 248249, 251, 259, 261, 265267, 280, 314, 329, 363364,
alluvial architecture 129, 141145, 147148, 150, 153156, 159 385
alluvial fan 24, 50, 71, 80, 85, 119, 135, 137, 139, 140141, 145, bypass 22, 50, 52, 55, 57, 5960, 66, 69, 71, 199200, 216217,
148, 156157 246, 267, 280, 344, 359, 365, 401402, 476, 483, 511
alluvial valley 87, 119, 138, 141, 145, 217, 249
anastomosing 96, 134135, 138, 139, 160
angle of climb 5253, 5556, 62, 6669, 73, 7577 C
angle of internal friction 33 Caliente Range 321, 331332, 337
angle of repose 27, 3132, 89, 91, 104, 122, 272 California 50, 249, 302, 304, 307, 319321, 323325, 328329, 331
anhydrite 65 332, 335, 337, 377, 405, 408412, 414, 429431, 433, 436, 446
animal activity 46 447, 454464
animal trackway (footprint) 33, 43, 46 canyon 8, 113, 279, 304, 361, 363, 373374, 399402, 415, 422
antidune 8586, 90, 95, 99, 104, 119 432, 454, 485, 493494, 504, 511512, 517
arid 19, 22, 24, 7071, 73, 107, 119, 124, 129130, 133, 139, 140, capillary fringe 45, 53, 55, 65, 68, 126
156157 carbonates 157, 200
Arizona 60 Carmel Formation 60, 62
armored lag 26, 58 Carmelo Formation 425
Askja sand sheet 29, 36 CCC turbidite 6, 9, 404, 406, 415, 441, 458459, 461, 474
Askja, Iceland 39, 44, 50, 55 Cedar Mesa Sandstone 3638, 4344, 4647, 58, 65, 6870, 77
Atchafalaya 241, 243, 246, 250, 272, 347, 352, 355, 381 cement 22, 55, 6265, 71, 73, 136, 220, 279, 314, 344, 350, 364, 409
Australia 5, 32, 44, 47, 65, 176177, 185, 221, 226, 249, 255, 340, cementation 63, 65
342, 360, 366, 369371, 390391, 403 channel 12, 57, 9, 12, 24, 63, 68, 71, 8587, 91, 93124, 126,127,
autocyclic 19, 34, 57, 175, 179, 243, 348, 359, 419420 129, 133148, 150151, 156160, 171, 176177, 179180, 184,
autogenic 73 186, 188196, 199200, 202204, 206208, 210211, 216217,
avalanche 27, 3032, 40, 104, 107 220, 225, 227, 237, 240241, 243244, 246247, 249250, 252
avalanche strata 3031 253, 257, 260, 262263, 265272, 274275, 279, 281, 285, 300,
avulsion 119, 121122, 134143, 147, 150151, 156160, 237, 243, 312313, 316317, 331, 353, 361, 367369, 372, 399404, 412,
249, 252253, 266267, 274, 285, 421, 441, 448, 463, 501 414451, 453454, 456459, 461, 463465, 468470, 472478,
482485, 486, 489490, 493, 495496, 501502, 504, 510511,
513514, 516
B channel belt 85, 87, 91, 97, 102, 108112, 115122, 126, 134139,
Bahrain 50 141144, 146, 151, 156, 158, 270, 438, 449, 476, 484
bar and trough system 316 channel fill 2, 6, 91, 96, 100101, 104, 107110, 112, 114, 117118,
barchan (barchanoid) 28, 32, 34, 4043, 53, 63 122, 133, 141, 159, 257, 265, 401, 425, 427432, 436, 440, 446,
barrier island 10, 200, 240, 244, 260, 271, 297298, 300, 303, 381, 389 449, 457458, 461, 463, 475476, 482484, 504
basalt 25, 6264 channel migration 86, 9697, 99, 103, 119, 137, 141, 176, 190, 253,
baseline of erosion 58 266, 434
bedform 1920, 22, 2730, 32, 3440, 43, 4553, 5557, 6264, 66, channel pattern 91, 93, 9699, 102103, 116, 118119, 134135,
68, 70, 7273, 76, 85, 8991, 102, 104105, 113, 174, 221, 243, 137138, 140, 143144, 147148, 156159, 186, 203, 253, 419,
270271, 295, 306307, 311, 313, 361, 363, 366, 371372, 385, 425, 427
403404 chenier 243, 268, 271272, 355
bedform behavior 34 climate 19, 22, 43, 4647, 58, 60, 63, 69, 71, 73, 80, 85, 107, 119,
bedform climb 22, 43, 50, 52, 5657, 62 121122, 124, 126, 130, 133135, 139, 141, 143, 147149, 151,
bedform migration 19, 22, 34, 39, 4849, 52, 56, 66, 68, 73 155160, 171, 178179, 199, 216, 225, 241, 254, 307
bedload 23, 8687, 89, 94, 9799, 104, 119, 137, 182183, 186, 191, climate change 22, 58, 60, 69, 71, 73, 122, 135, 141, 143, 148149,
243, 267, 295, 353, 361 151, 156158, 241
bedload sheet 8687, 89, 9899, 104 climatic cyclicity 71

Facies Models Revisited

SEPM Special Publication No. 84, Copyright 2006
SEPM (Society for Sedimentary Geology), ISBN 1-56576-121-9, p. 521527.
522 INDEX

climb 3, 36, 40, 52, 53, 5556, 62, 6669, 7576 Devonian 5, 11, 44, 63, 112, 114115, 123, 129, 131, 157, 238, 253
clinoform 237, 239, 250, 262, 271276, 280, 362, 363, 367, 385 254, 274, 403, 406
coastal 1, 22, 27, 43, 4547, 50, 5758, 60, 62, 73, 119, 139, 147, 150, Dhahran 45, 50
157159, 171, 175177, 179182, 191193, 197, 215, 217218, diachronous 58, 199
221, 226227, 237, 240, 249, 254, 258, 293, 297298, 300, 303, diagenetic cement 55
305306, 318320, 323324, 327,332, 344, 347, 349, 350, 357, diastem 58, 191, 194, 390
367, 461, 493 directional variability 27
coastal classification 180, 182, 218, 221, 226 distributary channel 9, 199, 202203, 237, 241, 243244, 246, 249
cohesive 22, 32, 44, 97, 425 250, 252253, 257, 260, 262263, 265272, 274275, 279, 417,
cold-climate desert 47 422, 427, 431, 463, 468, 473
Colorado 11, 40, 60, 62, 201202, 210, 373, 391 distributary mouth bar 3, 9, 237, 241, 246, 250, 253, 263, 344
Colorado Plateau 60, 62 diurnal 27, 34, 156, 361, 364
compaction 58, 142, 143, 158, 404, 449 draa 2731, 35, 43, 51, 6263
compound bar 86, 91, 94, 98, 100102, 104105, 108 drift potential 28, 34
contorted bedding 46 drift potential (resultant, RDP) 28, 34
core 3, 5, 9, 12, 39, 50, 70, 73, 116, 201, 203, 218, 255, 272, 275277, drift direction (resultant, RDD) 28, 34
304305, 319, 363, 367, 379, 381382, 386, 407, 427, 431, 452 dry eolian system 19, 5253, 55-56, 6062, 68, 74
crawling trace 33, 129 dry interdune 31, 43, 45, 66, 68, 78
creep 23, 2627, 37 Duero Basin 46
Cretaceous 34, 6, 912, 36, 6263, 153, 157, 174, 177, 180, 184, dune 13, 1920, 2223, 25, 2736, 3848, 5057, 60, 6263, 6576,
186, 191192, 196197, 199200, 202, 204, 206208, 237, 243, 78, 8591, 9495, 98100, 104, 107, 109, 112, 115116, 119, 121
248, 253, 261263, 266268, 270271, 273274, 276281, 283, 122, 139, 187, 191, 227, 268, 295, 312, 316317, 319, 329, 361,
285, 302303, 333, 341, 374376, 405, 407, 410411, 414, 427, 365366, 370, 372
430, 446447, 454, 501 dune element 40, 43, 4546, 54, 6263, 6667, 69
crevasse splay 91, 115, 119124, 126127, 134135, 137, 139, 270, dune flank 27
400, 422, 438, 463, 468474, 504 Dunvegan Formation 180, 191, 248, 261, 266, 268270, 274, 276
crinkly laminae 47 282
critical climbing 52 dynamic facies model 19, 22, 67, 69, 73, 77
cross bedding (cross-bedding) 2, 68, 1011, 19, 32, 34, 3839, 42,
190, 249, 258, 261, 265, 298, 307, 311313, 317, 319, 322323,
331, 363364, 366, 371, 373, 375, 380, 389, 407 E
cross strata 32, 3335, 3841, 4850, 52, 55, 63, 86, 8990, 94, 98 Ebro delta 252
100, 104, 107, 109, 112, 121122, 127, 268 ejecta 23
cross stratification 7, 13, 15, 32, 78, 88, 104, 195, 249, 258, 260, 281,
England 3, 10, 13, 4344, 47, 62, 6566, 191, 215, 364, 413
381 Entrada Sandstone 48, 62, 65
crust 33, 46, 129, 134, 148, 149, 153, 157, 353 entrainment 19, 2223, 25, 50, 53, 119, 351
curl 33, 45, 46 eolian 6, 8, 13, 1920, 22, 23, 2540, 4348, 50, 5258, 6063, 65
current ripple(s) 33, 107, 126, 258, 260, 375, 427, 442, 458, 468, 473 76, 78, 80, 139, 153, 157, 270, 272, 329, 415
eolian plane beds 43
eolian system, wet 19, 53, 5557, 6162, 6869, 73, 75
D eolian system, wetdry 74
damp interdune 32, 38, 45, 47, 53, 65, 68 ephemeral 46, 50, 62, 119, 122, 126, 129, 133, 139, 153, 156, 176,
Danube 240, 244, 249, 255, 259260, 267, 270 180, 237
deflation 19, 22, 26, 28, 43, 4748, 52, 55, 5761, 6566, 6970 erg 22, 24, 4546, 50, 53, 5758, 60, 62 63, 65, 6667, 6971, 75
delta 23, 6, 910, 13, 16, 50, 85, 107, 115, 119, 121122, 134135, 77, 80, 88, 143145, 176, 180, 239, 257, 272, 297
138139, 143, 153, 157158, 173174, 176, 178180, 182, 184 erg construction 22, 50
185, 189191, 194, 196, 199204, 206208, 210, 215220, 225, extra-erg 46
237268, 270282, 284285, 293, 320, 324, 339, 342, 344345, estuarine lithosome 189, 190
347, 349350, 352353, 355, 360, 363, 367, 371373, 375, 381 estuary 5, 171, 173, 176177, 180, 182191, 193, 195, 198201, 206
383, 390, 401, 425, 427429, 431, 436, 463, 471, 493 208, 212, 215221, 225226, 237, 256, 265, 268, 282, 285, 344,
delta front 184, 246247, 249251, 253, 256, 257, 259, 261262, 363, 372, 401402, 445
266268, 270272, 279 Etendeka igneous province 62
delta, lacustrine 115, 119, 121122, 139, 249, 276 Etjo Formation 36, 6263
delta plain 157158, 184, 237, 244, 249250, 253257, 265, 270, eustasy 147, 158, 254, 281, 297, 339
272, 355, 381, 431, 493 evaporite 33, 47, 65, 124, 129, 156157
density 2223, 37, 89, 129, 153, 179, 203, 241, 243, 246, 258, 347, evaporite precipitation 33, 47
369, 370, 401403, 408, 415416, 421, 425, 427, 493, 504
density current 369, 401
depositional elements 2, 911, 13, 1516, 24, 183, 399401, 403, F
414415, 420, 422, 428, 431, 447, 470, 484485, 501, 504 facies architecture 19, 69, 172, 237238, 240, 255, 258, 264, 268,
depositional model 64, 6668, 85, 86, 99100, 111, 119, 121124, 270, 281, 388
135, 171, 293, 355, 414 facies association(s) 1, 57, 13, 15, 54, 71, 298299
desert 13, 19, 2223, 33, 3940, 4547, 5051, 58, 63, 66, 129 facies definition 2
desert rose 33 facies model(s) 13, 610, 1213, 1516, 19, 22, 39, 6667, 6970,
desiccation cracks 11, 33, 4546, 57, 107, 119, 122, 126, 133, 156 72, 73, 75, 77, 85, 171172, 174180, 182183, 189, 191, 212,
 INDEX 523

215218, 220, 223, 225, 237, 282, 293298, 300, 302, 304305, Great Sand Dunes 40
311, 322, 324, 339, 341, 345, 347, 363, 369, 379380, 390, 399, ground-penetrating radar 22, 85, 179, 270
430, 504 growth fault 243, 256259, 263, 274, 279
facies modeling 13, 13, 16, 6667, 177, 430 Guerrero Negro 46, 58
facies succession(s) 12, 67, 71, 177, 179, 183, 215, 237, 240, 255, Gulf of Mexico 9, 147, 200, 237, 239240, 243, 253, 272, 275, 305,
257258, 260261, 264, 267, 272, 277281, 304, 504 339, 341, 348349, 352, 356, 369, 372, 378, 383, 390, 423, 426
fair weather (fair-weather) 9, 15, 250251, 294, 296, 300, 311314, 427, 429, 431, 437, 439, 441443, 446447, 449, 453454, 469
317320, 323324, 326, 328, 336, 356, 360, 369, 377 470, 475476, 478, 485486, 489, 492, 494, 498, 501502
fenestral porosity 33 gutter casts 317, 323, 367, 374
Ferron Sandstone 262263, 268, 271, 283 gypsum 33, 46, 50, 65, 129
flake 33, 46
flash flood 46
flood basalt 62, 64 H
flood basin 115, 117, 119121122, 126, 134, 136137 halite 33
flood deposit 107, 137, 255256, 259, 377 hard-pan 46
flood plain 8587, 96, 110, 113, 115, 117123, 126, 129130, 134 Helsby Sandstone 4344, 62, 6668
139, 141145, 147, 150, 153, 156157, 159160, 179, 186, 192, high-energy coast 293, 304, 311314, 319320, 324
194, 196, 201, 210, 237, 255, 267 Holocene 14, 43, 113, 116, 134, 143, 157, 217, 247, 293, 302, 339,
flow concentration 52 343, 349, 351, 367, 379
flow fluctuation 34 homopycnal 241, 249
flow reattachment 27 humid 22, 43, 58, 107, 121, 124, 139, 156
flow separation 27, 93, 95 hummocky cross-stratification 301, 317, 370371, 373374
fluid mud 243, 246, 253, 311, 345, 355 hyperpycnal 241, 246247, 249, 253, 258, 282, 344, 347348, 350,
fluvial 2, 6, 911, 15, 19, 27, 43, 46, 50, 6263, 6566, 6871, 8588, 353, 356, 358, 370, 374, 402, 415, 427, 493
111, 113, 129, 131135, 141144, 146150, 153, 156160, 171 hypopycnal 241, 243, 249,250, 258, 344345, 348, 352353, 355,
172, 174176, 178180, 182183, 186, 189192, 194, 195202, 358
204205, 207, 209212, 215218, 225226, 237, 240, 242, 249,
253, 255256, 258, 260262, 265, 270, 272, 277, 279282, 331,
339, 344, 347348, 350, 353354, 359, 415, 425, 427, 434435, I
444 Iceland 29, 36, 39, 44, 47, 50, 55
fluvial inundation 46 ichnofacies 129130, 193, 195, 219221, 251, 259, 265, 280281,
fog 65 300, 302, 305, 369, 373, 375, 377, 381382, 385386
forced regression 257, 260, 279, 293, 299301, 318, 320322, 324, ichnology 171, 195, 218, 281, 379, 385, 386, 390
349, 354, 431, 436, 493 impact zone 27
forcing parameter 34, 71 incised valley 145, 159160, 171175, 177, 179180, 184, 191193,
foreset 10, 13, 30, 39, 41, 43, 47, 4950, 238239, 249250, 272274, 195201, 206, 208, 215217, 219220, 223, 253254, 279280,
320, 329331 341, 372, 374, 389, 493, 511
foreset azimuth 39, 43, 49 inclined strata 8687, 99105, 107108, 110111, 113114
fossil 45, 15, 47, 123, 129134, 157, 182, 190, 203, 207, 209, 212, Indian Ocean 360
215, 218219, 249, 300, 304, 359, 379 infiltration capacity 46
friction 2223, 33, 112, 183, 186, 241, 243, 246, 249, 253, 258, 267, interdistributary bay 3, 243, 249, 253, 268, 270
270, 350, 356, 360 interdune 1920, 22, 27, 3133, 38, 40, 43, 4548, 5157, 60, 6263,
frontal splay 250, 270271, 399400, 415417, 419422, 447, 473 6571, 7376, 78
478, 484485, 493, 504, 508, 511 interdune flat 38, 43, 45, 5253, 55, 57, 62, 6869
Frontier Formation 261, 273274, 277, 279, 284 interdune hollow 38, 51, 56, 6668
interdune migration surface 38, 48, 56
interglacial 43, 156158, 201, 346
G intra-erg 46, 63
Gallup Sandstone 265 inverse grading 27, 30, 37, 247, 403, 407
gas 1, 3, 19, 7071, 79, 85, 123, 129, 133134, 157, 159, 202, 210, isochronous 58
221, 237238, 244, 266, 278, 341, 362, 383, 407 isolated bedforms 45
geomorphic element 22, 40, 47 isolated sandstone bodies 299
Gilbert delta 250
glacial 19, 22, 43, 50, 97, 135, 139, 156158, 201, 246, 293, 302, 346,
366 J
Gondwana 62 Jackson Group 320, 328, 330
graded bedding 258, 399, 404, 407408 Jafurah area 46
grain packing 32 Jurassic 48, 60, 62, 65, 133, 153, 354
grainfall 3033, 36, 38, 4041, 78
grainfall strata 3032, 40
grainflow 27, 3132, 37, 4041, 4344, 78 K
grainflow, amalgamated 31 Kelso Dune Field 50
grainflow strata 3132, 44, 78 Kuiseb River 27
Gran Desertio 40 Kuwait 50
524 INDEX

Mekong 243, 251, 253, 261, 270

L Mexico 9, 19, 40, 4546, 50, 62, 147, 157, 200, 237, 239240, 243,
lacustrine 27, 50, 107, 115, 119, 121122, 124, 127, 129130, 139, 253, 265, 272, 274v275, 298, 305, 309, 339, 341, 348349, 352
146, 156, 180, 200, 238, 240, 249, 268, 276 353, 356, 369, 372, 378, 383, 390, 423, 426, 427, 429, 431, 437,
lacustrine delta (see delta, lacustrine) 439, 441, 442, 443, 446, 447, 449, 453454, 469, 470, 475476,
lag 2, 7, 915, 26, 28, 30, 43, 47, 50, 58, 63, 91, 148149, 172173, 478, 485486, 489, 492, 494, 498, 501502
180, 182, 193, 201204, 217, 221, 237, 239240, 249, 256, 260, migration 1920, 22, 2728, 30, 32, 3435, 3740, 43, 4849, 52,
264, 267, 270, 275, 280, 284, 297, 314, 320, 322, 349, 353, 365, 5556, 58, 6263, 66, 68, 70, 7375, 85, 86, 8991, 9697, 99,
368369, 402, 408, 430, 432, 493 101105, 109112, 115, 119, 121122, 124, 137, 141, 143, 148,
Lagniappe delta 275 159, 176, 179180, 184, 190191, 210, 243, 253, 266, 268, 272,
lagoon 7, 910, 1315, 173, 180, 182, 215, 217, 237, 239240, 242, 279, 297, 300, 306307, 313, 362363, 372, 379380, 383, 385,
249, 256, 260, 264, 267, 270, 493 387, 390, 424, 428, 434, 435, 439, 441444, 446, 451452, 475,
lagoonal 10, 1315, 237, 240, 264, 267 483
lake 24, 3233, 4547, 50, 5758, 60, 62, 71, 80, 102, 113, 119, 121 migration direction 28, 34, 3940, 48, 49
124, 127, 129130, 134135, 137139, 143, 153, 156157, 180, migration speed 34
184185, 191, 197, 201202, 206, 238239, 246, 249, 253, 414 Milankovitch 58, 62, 157, 158
Lake Lucero 50 mineralogy 22, 157, 379, 388
laminar flow 25 Mississippi 89, 9697, 116, 121, 157159, 174, 176, 180, 191, 197,
lateral accretion 434435, 441, 443446, 449, 483 201, 237238, 240, 242243, 246, 249250, 253254, 258, 260,
lee 20, 2728, 3135, 3739, 48, 5253, 91, 104, 216, 305, 311, 315, 264, 270, 272, 285, 344, 347, 349, 352, 355, 371, 373, 381382,
363 390, 401, 422427, 471, 512
lee-side depression 27, 31, 38 Mississippian 180, 191, 197, 201
Leman Sandstone 62, 78 mixed eolian system 65, 69
Lena River 252 Mobile Bay 215, 352
levee 12, 6, 9, 12, 91, 115, 119124, 126, 134, 137, 139, 179, 189, moisture 22, 43, 45, 5253, 65, 126, 130
199, 243, 270, 399401, 404, 406, 412, 414422, 425, 427, 429, mojave 22, 43, 45, 5253, 65, 126, 130
431432, 437443, 446447, 449451, 453, 456, 458459, 461, morphology 19, 27, 3839, 42, 45, 50, 55, 62, 6667, 70, 73, 171,
463464, 468470, 472478, 485, 490, 493, 494, 496, 501502, 174176, 178, 182184, 215, 227, 237238, 240242, 253, 255
504, 510511, 513514 256, 316, 323, 339, 390, 417, 419, 422, 428, 471472, 477, 501,
leveed channel 1, 400, 401, 415417, 419422, 427, 429, 431, 437 504
438, 441443, 447, 450, 473478, 485, 493, 496, 504, 511 mud curl 4546
liquefaction 3233, 43 mud drape 46, 50, 107, 195, 198, 203, 205, 262, 267, 361, 373, 375
lobe 1, 2, 91, 9697, 104, 119, 139141, 158, 237, 240242, 244, 246, mud flake 33, 46
250, 255256, 260, 262, 264, 267268, 270271, 274275, 277 muddy substrate 46
281, 340, 343, 381, 390, 399401, 415418, 420, 422, 473, 475
476, 484, 496, 501
loess 23 N
low accommodation 175, 191, 199, 202, 209 Namib Desert 45, 47, 5051, 63
low-energy coast 317319, 324, 333 Namib Sand Sea 27, 32, 35, 4546, 50, 52
Lower Cutler Beds 36, 45, 63, 68 Namibia 27, 29, 32, 3637, 4546, 55, 58, 6263
lower shoreface 300301, 318, 320323, 329, 340 Navajo Sandstone 35, 62, 68
lower-stage plane bed 89, 95, 104 nearshore circulation cells 295, 299, 311, 316, 321
lunate megaripples 307, 312, 316317, 319, 323, 327 nearshore profile 313, 320
Nebraska Sand Hills 43, 46
negative sediment budget (see sediment budget negative)
M nested reactivation surface 34
Macaronichnus 300 net sediment budget (see sediment budget net)
Mahakam 250 neutral sediment budget (see sediment budget, neutral)
marine flooding 57 New Mexico 19, 45, 50, 62, 265, 272, 353, 372
marine regression 60, 149, 159, 297 Niger 47, 261
marine, shallow 16, 19, 50, 159, 176, 246, 280, 293, 297, 302, 308, Nile River 174
320, 324, 349, 386 Nile delta 238, 249, 270
marine transgression 58, 62, 149, 156, 159160, 199, 297, 300, 323 North Sea 19, 58, 62, 70, 7881, 358, 375
mass-transport complex 401, 412413, 485491, 493, 498499, 504, numerical modeling 220221
508
Mauritania 43
meandering 7, 9, 12, 15, 91, 9698, 101102, 108, 112, 116117, O
135, 138, 140, 142, 156, 175, 186, 188, 190191, 196, 199202, oblique migration 35, 63
209210, 212, 268, 401, 427, 434435, 444, 501 Oman 25
meandering channel 97, 101, 188, 401, 434435, 501 Ophiomorpha 265, 300, 382
mechanical loading 33 optically stimulated luminescence 22, 58, 85
mega-bedforms 30 orbital forcing 58, 158
megaripple 27, 29, 47, 306307, 311 312, 315317, 319, 323, 327, Oregon coast 45, 312, 320
370 Ormskirk 62, 68
 INDEX 525

overbank 71, 94, 107, 115, 119, 121, 124, 136138, 140, 143, 156, ravinement 189191, 194, 199201, 206, 209, 218, 224, 251, 279
158, 160, 194, 199200, 210211, 415, 417, 422, 438, 441442, 280, 297, 300, 349, 351, 372, 376, 381382, 386, 388
450, 452453, 463, 468, 473474, 476, 482484, 501, 504 reactivation surface 3435, 38, 48, 65, 195, 261262, 375
recognition criteria 197, 217, 361
reg 26
P regional climate 46, 58, 157
Padre Island 41, 4546, 305, 308, 320, 326 regression 15, 60, 113, 149, 159, 180181, 192193, 197, 215, 257,
Page Sandstone 60 260, 277, 279, 293, 296, 297, 299301, 318, 320322, 324, 337,
paleoenvironment 19, 57 349, 354, 431, 436, 493
paleosol 33, 46, 68, 113, 126, 128130, 156160, 193194, 196, 201 remote sensing 85
202, 209211, 267 reptation 23, 2627
parabolic 33, 40 reservoir prediction 70, 73
Paradox Basin 63 resultant drift direction (see drift direction, resultant)
Paraba do Sul 252, 259, 271 resultant drift potential (see drift potential, resultant)
Paran Basin 62 reversing dune 40
parasequence 196197, 202, 268, 279, 294, 297, 300, 321, 383, 386, rhizoliths 33, 47, 5758, 60, 69, 133134
388 Rhone (Rhne) (River, delta, fan) 8, 259260, 274, 401, 465, 501
particle size 22, 25 rip channel 312, 316317
pebble lag 7, 43, 193, 201202 rip current(s) 178, 293, 295, 311, 313314, 317, 320321, 357, 366
Pennsylvanian 3, 36, 63, 68, 180, 253254, 272, 274 ripple 34, 6, 9, 1113, 2733, 3641, 43, 4547, 52, 56, 78, 8587,
perched water table 46 8990, 9495, 98100, 104, 107, 112, 119, 122123, 126127,
permafrost 63 191, 203, 207, 210, 248249, 258, 260262, 298, 301, 306307,
permeability 46, 85, 107108, 110, 144, 201205, 207, 210, 211 311320, 323, 327, 331, 363365, 369370, 373375, 399, 403
212, 404 406, 427, 442, 446447, 458459, 467468, 473, 481
Permian 8, 19, 3638, 4347, 58, 60, 6263, 65, 6871, 7781, 180 ripple indices 27
pinstripe lamination 31, 36 ripple strata 3031, 3638, 41, 43, 45, 78
planar strata 86, 89, 99, 104, 122 ripple trough 27, 30, 37
plane beds 32, 43, 47, 86, 89, 95, 99, 104, 119, 122 rippleform laminae 30
plant colonization 46, 58 river delta 252, 273, 367, 471
plant root(s) 33, 63, 121, 126, 129, 133134, 156 river mouth 176177, 189, 199, 215216, 218, 226, 240, 249, 253,
playa 45, 47, 71, 139 255, 257258, 267, 282, 304, 344345, 347, 350, 356, 402, 432,
Pleistocene 4, 14, 43, 113, 157158, 198, 223, 238, 320, 323324, 493, 511
328, 329, 335, 349, 375, 379, 402, 429, 437, 449 river-influenced 9, 240
Pleistocene terrace deposits 328329 root structure 33, 45, 47, 63, 330
plinth 27, 32, 36 Ross Sandstone 434, 445, 474, 476, 479480, 483484
Po delta (Po River) 243, 247, 252 Rotliegend 19, 58, 7071, 73, 78, 8081
point bar 1, 12, 91, 94, 96, 98102, 104, 109, 118, 186, 212, 434
435, 445
polygon 33, 57, 60 S
polygonal fracture 57, 60 sabkha 24, 47, 50, 55, 60, 62, 7172, 80
pore-water pressure 32 Sahara 47, 52
potential sediment load 50 Salima sand sheet 47
pour-in texture 30 salt crust 33
primary airflow 27, 55 salt flat 24, 47
prodelta 184, 190, 199, 218, 237, 239, 248249, 251, 253, 255, 257 salt-growth 47
258, 260, 262264, 267268, 270272, 274, 280, 347 salt precipitation 47
progradational shoreline 297, 299 saltation 23, 27, 30, 3738
prograding strand plain 298 saltation cloud 23
Proterozoic 60, 65, 186 saltation path length 27, 37
pseudomorph 33, 46, 65 San Clemente 429431, 433436, 501
pumice 25 San Miguel Formation 280, 285
sand sea 20, 22, 27, 32, 35, 40, 4546, 50, 52, 40
sand transport direction 2728, 33
Q sand-accumulating bedform 28
Quaternary 22, 43, 58, 73, 134, 156158, 180, 199, 210, 239, 272, sand-transporting bedform 28, 34
339, 341, 343, 348, 354, 359, 363364, 371374, 375, 381, 385, sandflow 31, 34
388, 389, 390, 402 sandsheet 29
Queensland 65, 371 satellite 22, 24, 85, 221, 340342, 352, 369, 391
Saudi Arabia 4546, 50
scarp recession grainflow 31
R scour pit 39, 48
raindrop imprint 33, 122 sea level 2, 7, 910, 13, 22, 50, 58, 73, 122, 124, 141, 148, 150, 152
rainfall 3033, 36, 38, 4041, 46, 78, 130, 149 153, 156160, 171, 173175, 179182, 184185, 189192, 197
526
202, 215218, 220, 224225, 227, 238239, 246, 254256, 279,
281282, 285, 293294, 296297, 299303, 306, 324, 337, 339,
346349, 354, 359, 390, 399, 401402, 425, 431, 436, 490, 493,
501, 511
seabed imagery 221
seasonal 27, 34, 46, 68, 119, 121, 129130, 133, 156, 241, 249, 360,
363
secondary airflow 27, 35
sediment availability 19, 50, 6870, 73, 7677, 121, 148
sediment budget, negative 47, 57
sediment budget, net 40, 52, 55, 57, 59, 61, 67
sediment budget, neutral 57
sediment compaction 58
sediment flux 22, 57, 67, 149, 179180, 182183, 218, 220, 504
sediment supply 2, 910, 13, 19, 22, 27, 30, 50, 55, 6465, 69, 85,
91, 122124, 135, 137, 138, 141, 144145, 148149, 151, 156
159, 171, 179182, 191192, 197, 199200, 216, 225, 250251,
255, 258, 279, 281, 297, 303, 341, 344, 347, 349350, 355356,
374, 380, 390, 422, 501
sediment transport 23, 50, 52, 8586, 9195, 98100, 107, 112, 119,
135139, 141, 145, 147149, 158, 176, 179, 186, 217, 221, 242,
258, 295, 320, 343, 350352, 360, 362, 366, 369, 370, 511
sediment transport rate 50, 52, 9193, 95, 99, 135,136, 139, 141,
145, 148149, 179, 242
sediment waves 400, 422, 437, 446, 450451, 453454, 463, 468,
472, 475, 504
sedimentary basin 137, 141, 147149, 180, 202, 211, 219, 220
seif 28, 144, 146
seismic 12, 5, 70, 85, 115116, 118, 141, 144145, 159, 171, 175,
179, 192, 197, 203, 206, 216, 218, 220221, 223, 238, 239, 256
257, 267, 270, 272, 274275, 305, 362364, 367368, 372, 375
376, 383385, 399, 401, 413, 415, 417, 422426, 428429, 431,
434, 437444, 446447, 458, 468470, 473477, 484486, 489,
490, 496498, 500501, 503508, 510, 512
semiarid 22, 70, 119
separation cell 27
sequence 11, 13, 22, 27, 32, 40, 48, 50, 52, 5758, 60, 6669, 71, 73,
77, 9192, 100101, 103104, 107, 115116, 122, 124, 126127,
129130, 134, 137, 141145, 156, 157, 158160, 171, 175, 179,
189, 191200, 202, 209, 211, 215216, 218219, 221, 237239,
241, 253, 255256, 260, 266, 268, 274, 279, 281283, 294295,
297, 299300, 318, 320322, 325, 328339, 341, 348, 351, 369,
376, 383, 385389, 399401, 403404, 407, 414, 420, 431, 457,
459463, 475, 483, 485, 493, 510
sequence boundary 57, 189, 191194, 197200, 202, 209, 215216,
218, 348, 388389
sequence stratigraphy 159, 171, 175, 239, 274, 281282, 299, 341,
385386, 401, 485
serir 26
setdown 295, 299, 321
setup 295, 299, 321, 350, 357, 366367
shadow zone 27, 37
shallow marine (see marine, shallow)
shear stress 23, 25, 8889, 9195, 112, 119, 137, 139, 243, 350, 360,
368
Shikaoda Formation 65
shoreface 180, 184, 189190, 196, 199200, 202, 207208, 210, 215,
217218, 221, 224, 237, 249251, 258260, 265, 267, 271272,
278279, 281282, 293305, 307314, 317324, 327, 329333, top-truncated delta 251, 279, 281
339341, 349350, 355, 357, 366367, 370374, 380381, 385,
390
shoreface profile 299300, 304, 308309, 323, 327, 339
shoreface-attached ridges 305, 310, 370372
Silurian 63, 157, 179
Skeleton Coast 27, 29, 32, 46, 55
INDEX
slab slide 33, 37, 44
slipface 27, 28, 3132, 40, 6263, 127
slope channel 399400, 422, 425, 427429, 431, 437, 441, 444, 451,
490, 504
slump degradation grainflow 31
slumping 32, 412, 425426, 441, 466, 474
soft-sediment deformation 32, 4344, 258, 260, 274, 458459, 461
soil 119, 126, 128130, 133, 138, 156, 160, 193, 353
sole marks 367, 374
South Atlantic 62
southern North Sea 19, 58, 62, 70, 7881, 358, 375
Spain 4546, 252, 319320, 324, 327329, 335
spatial complexity 73
sphericity 25
staging area 399400, 415, 417, 422, 425, 485, 491, 493, 504
static facies model 67, 174
stepped transgression 297, 300, 321
Stokes surface 19, 57, 59
storms 14, 86, 90, 139, 241, 247, 249, 253, 258, 293, 295, 313314,
317319, 339, 350, 356357, 359360, 366367, 369370
stoss 27, 29, 3637, 52, 91, 305
strataset 8587, 9091, 102, 108, 110, 115, 121123, 137
stratigraphic modeling 22
stratigraphic organization 175, 198, 200, 281
subsidence 11, 46, 5758, 61, 65, 67, 119, 135136, 138139, 141
142, 147151, 153, 156, 159, 237, 257, 274, 279, 282, 297, 302,
339, 342, 381
superimposed bedform 20, 30, 35, 39, 40, 48, 102, 104105, 366,
385
superimposition 27, 35, 38, 40, 43, 48, 62
supersurface 22, 48, 55, 5760, 62, 66, 67, 6971
supratidal flat 65
surface creep 23, 26
surface stabilization 22, 55, 58
surface trace 45
suspended load 23, 97, 119, 255, 311, 345, 353
swaly cross-stratification 301, 311
systems tracts 193, 199, 216218, 238, 281282, 339, 348349, 359,
386, 388, 390, 425
T
tectonism 58, 8586, 119, 122, 124, 135138, 141, 143144, 151,
156, 158160, 293
teepee 33, 47
Tnr 47
Tertiary 63, 132, 134, 148, 153, 157, 359, 383, 431, 458, 501
Texas 41, 4546, 180, 200201, 238, 249, 259, 281282, 285, 308, 310,
312, 320, 324, 326, 328, 330, 336, 349, 353, 369,371, 373, 378
tide 3, 810, 171172, 176179, 181184, 186, 189191, 196, 197,
215, 217218, 225, 237241, 243, 249256, 261262, 266267,
270, 273, 277, 279280, 282, 285, 293, 307, 344, 349350, 359
364, 366, 372, 375, 380
tide-dominated 8, 171, 177179, 183184, 186, 189191, 196, 215,
217218, 237, 240241, 252254, 256, 267, 270, 277, 282, 344,
362363, 372
tide-influenced 239, 241, 251, 254, 256, 261262, 266, 270, 273,
279280
topset 238239, 249, 251, 260, 272, 279, 281
trackway 33, 43, 4647, 134
transgression (transgressive) 10, 1415, 58, 60, 62, 149, 156, 159
160, 171172, 176, 179182, 184, 189192, 194202, 207, 210,
215218, 224, 237238, 240, 251, 255256, 260, 265, 277278,
280282, 284285, 293294, 297, 299300, 303305, 308, 321,
 INDEX 527

323, 332, 339, 343344, 348349, 351, 354, 359, 362364, 367,
371, 374, 379, 380381, 383, 386388, 390, 425, 485, 493 W
transition point 417418, 420, 473, 475476, 485, 493, 508 Waitemata Group 415, 461, 465466, 468
translatent rippleform stratification 30 Walthers Law 237, 293
transport 2, 19, 2223, 2528, 3031, 3335, 38, 40, 45, 47, 5053, water flow 86, 91, 9394, 99, 107, 119, 137, 312
55, 57, 63, 67, 8586, 89, 9195, 97100, 107, 112, 119, 122, water table 19, 22, 33, 4547, 50, 52, 53, 5563, 6570, 7475, 77,
129, 133139, 141, 145, 147149, 15859, 171, 176, 179180, 85, 126, 130, 133, 138, 160, 193
183, 186, 196, 200, 217, 221, 225, 242243, 253, 258, 295, 298 water-table elevation 46
299, 303, 306, 311, 314315, 320, 343345, 348, 350353, 356, wave base 15, 294296, 318, 328, 369, 399, 493, 511
359363, 365366, 369371, 373374, 377, 390, 399, 401, 407 wave energy 183184, 250, 254, 305306, 311, 317321, 323324,
409, 412414, 422423, 431, 442, 485491, 493494, 496511 328, 343, 356, 366
transport capacity 31, 50, 55, 171, 179, 243 wave ripple 46, 107, 260, 312, 331, 365, 370
transport distance 25, 412 wave-dominated 810, 171, 173, 177, 179, 183186, 189191, 193
transport rate 27, 30, 50, 52, 57, 89, 9195, 99, 135136, 139, 141, 194, 196197, 199, 200, 202, 207, 215, 225, 241, 249, 251252,
145, 148149, 179, 242 255257, 260, 265, 267, 270272, 281, 285, 293, 295, 300, 311,
transverse 22, 28, 3235, 3943, 45, 5051, 53, 6263, 67, 70, 72, 318, 344, 349, 363364, 366, 374, 376, 381
95, 104, 139140, 407408, 425, 428, 437, 440, 446447, 468, wave-influenced 237, 240, 242, 244, 250, 253, 255256, 258262,
472, 498 267, 270, 279281
Triassic 4344, 47, 60, 62, 6567, 153 wavy laminae 32, 46
Tsondab Sandstone 63 wet grainfall 33
turbidite 2, 46, 89, 1213, 203, 247, 250, 279, 370, 399407, 412, wet interdune 38, 46, 51, 6263, 66, 69, 71, 7475
414415, 417, 419420, 423, 425, 427428, 431, 434436, 441 wetting front 32
442, 446447, 458459, 461462, 466468, 474476, 480484, Wheeler Gorge 8, 414, 447, 457465, 476
493, 501, 508509, 511 White Sands 19, 45, 50
turbidity current 356, 369, 373, 399404, 407409, 412, 415, 422, Wilmslow Sandstone 62, 65
425, 427, 429, 431, 435, 441, 447, 474,476, 484, 493 wind gustiness 25
wind power 50, 52
wind regime 19, 22, 34, 40
U wind reversal 34
unidirectionality index 28 wind strength 27
unit bar 86 87, 91, 93102, 104105, 107110 wind velocity 2223, 27, 37
United Arab Emirates 25 wind-ripple stratification 30, 36
universal facies 6, 13 Wingate Sandstone 62
upper shoreface 265, 300, 311314, 317, 321, 323, 327, 329332 winnowing 26, 243, 271
Utah 78, 3638, 4348, 58, 60, 6263, 65, 6870, 77, 186, 199, 238, wireline log 70, 73, 115116, 197, 203, 209
243, 248, 262263, 268, 271, 278279, 283, 302304, 321, 333, 373

Y
V Yellow Sands 62
valley 46, 87, 93, 119, 121, 134139, 141, 143, 145, 147148, 157
160, 171175, 177, 179180, 182185, 189204, 206212, 215
221, 223226, 237, 249, 253254, 265, 267270, 278280, 282, Z
300, 341, 353, 372374, 383, 389390, 414, 493, 511 Zechstein Sea 62
valley fill 145, 147, 158160, 171172, 174, 180, 183, 191, 193202, zibar 47
207, 211, 216, 219220, 224, 253, 265, 267268, 279280
vegetation 22, 3233, 43, 47, 50, 55, 63, 86, 91, 9798, 103104,
108, 110, 119, 121, 126, 134, 141, 147149, 156158, 270
velocity gradient 23
ventifact 26
Volga basin 269
528 INDEX