Anda di halaman 1dari 367

Intra-stratal deformation in Quaternary deposits along the Black River, near Trenton, New Jersey, as portrayed by

Vanuxem (1842). The controversy surrounding these and similar structures in nearby Ordovician limestones is outline in
section 1.3.2.
The Geological Deformation of Sediments
The Geological
Deformation of
Edited by Alex Maltman
Institute of Earth 5tudies, University of Wales, Aberystwylh


First edition 1994
© 1994 Springer Science+Business Media Dordrecht
Originally published by Chapman & Hall in 1994
Softcover reprint of the hardcover lst edition 1994

ISBN 978-94-010-4314-4 ISBN 978-94-011-0731-0 (eBook)

DOI 10.1007/978-94-011-0731-0

Apart from any far dealing for the purposes of research or private study, or
criticism or review, as permitted under the UK Copyright Designs and
Patents Act, 1988, this publication may not be reproduced, stored,
or transmitted, in any form or by any means, without the prior permission
in writing of the publishers, or in the case of reprographic reproduction
only in accordance with the terms of the licences issued by the Copyright
Licensing Agency in the UK, or in accordance with the terms of
licences issued by the appropriate Reproduction Rights Organization
outside the UK. Enquiries concerning reproduction outside the terms stated
here should be sent to the publishers at the London address printed on
this page.
The publisher makes no representation, express or implied, with regard
to the accuracy of the information contained in this book and cannot accept
any legal responsibility or liability for any errors or omissions that may
be made.
A catalogue record for this book is available from the British Library
Library of Congress Catalog Card Number: 93-74887

i§ Printed on acid-free text paper, manufactured in accordance with

ANSIjNISO Z39.48-1992 (Permanence of Paper).

List of contributors ix 2.2.3 Pore pressure, effective

stress and porosity 44
Editor's preface xi 2.2.4 Primary consolidation and
Acknowledgements xv creep 48
2.2.5 Shear stress and deforma-
Chapter 1 Introduction and overview tion 51
Alex Maltman 2.2.6 Other stress paths and asso-
ciated strain states 60
1.1 General considerations 1
2.2.7 Over- and underconsolida-
1.1.1 Terminology: sediment, lith-
tion 64
ification and deforma-
2.3 Natural stress, strain and pore
tion 1
pressure 66
1.2 Mechanical aspects 6
2.3.1 General 66
1.2.1 General 6
2.3.2 Earth pressure and one-
1.2.2 Volume changes due to
dimensional compaction 67
burial 6
2.3.3 Natural shear stress and de-
1.2.3 Sediment strength 8
formation 68
1.2.4 Sediment deformation 9
2.3.4 Origin of overpressures in
1.2.5 Mechanical role of pore
fluids 12
sediments 70
2.4 Conclusions 71
1.2.6 Experimentation 16
1.3 Causes of deformation 16
1.3.1 Ice 16 Chapter 3 Glacial deformation
1.3.2 Disturbance in place 17 Tavi Murray
1.3.3 Gravitational mass move- 3.1 Introduction 73
ment 23 3.2 Subglacial conditions 74
1.3.4 Fluid-sediment movements 27 3.2.1 General 74
1.3.5 Tectonism 29 3.2.2 Bed type 74
1.3.6 Igneous activity 32 3.2.3 Thermal regime 75
1.4 Melanges as a case history 34 3.2.4 The ice-bed interface 76
Chapter 2 Mechanical principles of sediment
3.2.5 Bed thickness 77
3.2.6 Realistic basal conditions 77
3.3 Stresses arising from overlying
Mervyn Jones
ice 78
2.1 Introduction 37 3.3.1 General 78
2.2 Mechanics of particulate media in 3.3.2 Compressive stresses 78
theory and experiment 39 3.3.3 Shear stresses 79
2.2.1 Force, stress and strain 39 3.4 Sediment properties 79
2.2.2 States of stress and strain in 3.5 The sediment transport system:
a body 42 production, alteration and loss 81
vi Contents

3.6 Models of sediment properties and 4.3.6 Structures due to sediment

deformation 82 shrinkage 114
3.6.1 General 82 4.3.7 Structures due to sediment
3.6.2 Deformation of a homogen- wetting 117
eous sediment body 82 4.3.8 Deformation related to com-
3.6.3 Deformation of a structured paction 117
sediment body - anisotropy
4.3.9 Deformation related to early
and inhomogeneity 84
3.7 Basal processes as a control on chemical precipitation 118
deformation 85 4.4 Conclusion 124
3.7.1 Hydraulic processes 85
3.7.2 The consolidation--dilation Chapter 5 Mass movements
competition 86
Ole Martinsen
3.8 Effects of deformation 87
3.8.1 General 87 5.1 Introduction 127
3.8.2 Features arising from con- 5.1.1 General 127
solidation 87 5.1.2 Classification schemes 127
3.8.3 Features arising from shear 5.1.3 Basic theory 128
deformation 88
5.2 Falls 130
3.8.4 Development of sediment
5.2.1 Introduction 130
form 89
5.2.2 Rock falls 130
3.8.5 Sediments beneath surge-
5.2.3 Debris falls 131
type glaciers 91
5.3 Fluidal flows 133
3.9 Preservation of features 91
5.3.1 Introduction 133
3.10 Other types of glacial deformation 92
5.3.2 Turbidity currents 133
3.10.1 Pro glacial deformation 92
5.3.3 Flows related to volcanic
3.10.2 Deformation of frozen-sub-
eruptions 135
strates-basal-ice 92
5.3.4 Snow- and ice-generated
3.11 Conclusion 93
(avalanching) flows 138
5.3.5 Fluidized flows 140
Chapter 4 Sedimentary deformational struc-
5.4 Flows with plastic behaviour 140
5.4.1 Introduction 140
John Collinson
5.4.2 Debris flows 140
4.1 Introduction 95 5.4.3 Liquefied flows 142
4.2 Principles of physical disturbance 95 5.4.4 Grain flows 143
4.3 Physical deformation structures 99 5.5 Slumps 144
4.3.1 Partial loss of strength and 5.5.1 Introduction 144
density inversion 99 5.5.2 Process 145
4.3.2 Structures due to progress- 5.5.3 Products 147
ive loading of cohesive sed i- 5.6 Slides 152
ment 103 5.6.1 Introduction 152
4.3.3 Partial loss of strength and 5.6.2 Process 153
applied shear 105 5.6.3 Products 157
4.3.4 Structures related to up- 5.7 Creep 162
wards escape of pore water 5.7.1 Introduction 162
and sediment-water mixtures 108 5.7.2 Process 162
4.3.5 Synsedimentary faults 113 5.7.3 Products 164
Contents vii

Chapter 6 Tectonic deformation: stress paths 7.3.3 Transient fluid sources

and strain histories and sinks generated by
Dan Karig and Julie Morgan faulting 212
7.3.4 Response of poorly-lithified
6.1 Introduction 167 sediments 214
6.2 Stress paths during burial and up- 7.3.5 Response of well-lithified
lift of sediments in basins' 168 sediments 215
6.2.1 General 168 7.3.6 Chemical sources of fluid:
6.2.2 Laboratory studies of con- cementation, hydrocarbon
solidation 172 generation and mineral de-
6.2.3 Effects of geological pro- hydration 218
cesses not duplicated in lab- 7.4 Control of lithology and burial-
related consolidation on the per-
oratory experiments 175 meability of sedimentary units 220
6.2.4 Theoretical stress paths du- 7.4.1 General 220
ring sediment deposition 7.4.2 Permeability anisotropy
and unloading 177 resulting from consolida-
6.2.5 In situ measurements of tion 222
stress in sediments 181 7.4.3 Equivalent permeabilities
6.2.6 Geological implications and gross permeability an-
from theoretical and meas- isotropy 224
ured stresses 185 7.5 Permeability variations due to
6.3 Stress paths associated with defor- deformation in active tectonic
mation in accretionary prisms 187 systems: fractures, faults and
gouge 224
6.3.1 General 187
7.5.1 General 224
6.3.2 Lagrangian description of 7.5.2 Permeability changes in
sediment accretion 188 muddy fault zone materials 225
6.3.3 Large-scale variations 7.5.3 Permeability changes in
among prisms and mechan- fault zones in sands and
ical implications 189 sandstones 226
6.3.4 Diffuse strains in prism toes 190 7.6 Permeability changes at low effec-
6.3.5 Experimental studies 195 tive stresses 230
6.3.6 Implications for the defor- 7.6.1 General 230
mational histories of ac- 7.6.2 Extensional failure where
creted sediments from ob- the least principal effective
compressive stress is ten-
servation and experiment 200
sional 230
6.4 Conclusions 203 7.6.3 Open-fracture development
in the absence of regional
Chapter 7 Fluids in deforming sediments tensional stresses: load-par-
Kevin Brown allel extensional fractures 233
7.6.4 Stress amplification mech-
7.1 Introduction 205
anisms 234
7.2 Some basic hydrogeological con- 7.7 Effect of deformation on the tor-
cepts 207 tuosity of flow paths at different
7.3 Fluid sources and the nature of the scales 235
tectonic processes driving fluid flow 209 7.8 Discussion: transience and the inti-
7.3.1 General 209 mate coupling of hydrogeological
7.3.2 Consolidation and swelling 210 and tectonic processes 236
viii Contents

Chapter 8 Sediment deformation, dewatering and 9.2 Techniques of examination 261

diagenesis: illustrations from selected melange 9.3 Microfabrics 265
zones 9.4 Micro- to macroscopic structures 268
Tim Byrne 9.4.1 Shear zones 268
9.4.2 Slickensides 273
8.1 Introduction 239
9.4.3 Scaly clay and related fea-
8.2 Progressive deformation and de-
watering in the Nankai accretion-
tures 275
9.5 Macro- to mesoscopic structures 277
ary prism 240 9.5.1 Faults, folds and related
8.3 Progressive deformation of coher- 277
ent sediments in the Kodiak ac- 9.5.2 Liquefaction structures 285
cretionary prism 245
9.5.3 Dewatering structures 287
8.4 Progressive deformation of me- Recognition of sediment deforma-
lange terranes in the Kodiak ac- tion structures 292
cretionary prism 249 292
9.6.1 General
8.5 Deformation and fluid evolution in 9.6.2 Importance of recognizing
an accretionary sequence in west- pre-lithification deforma-
ern Washington 259 300
8.6 Conclusions 259 302
9.6.3 Outline of possible criteria
Chapter 9 Deformation structures preserved in 309
Appendix: List of symbols
Alex Maltman References 311
9.1 Introduction 261 Index 355

Kevin M. Brown Dan E. Karig

Scripps Institute of Oceanography, Geological Department of Geological Sciences, Snee Hall,
Research Division, University of California, San Cornell University, Ithaca, New York 14853-
Diego, 9500 Gilman Drive, La Jolla, California 4780, USA.
92093, USA.
Alex J. Maltman
Institute of Earth Studies, University of Wales,
Timothy B. Byrne Aberystwyth, Wales SY23 3DB, UK.
Department of Geology and Geophysics, Univer-
sity of Connecticut, 354 Mansfield Road, Storrs, Ole J. Martinsen
Connecticut 06269-2045, USA. Geologisk Institutt, Avd. A, Universitetet i Ber-
gen, Allegt. 41, 5007 Bergen, Norway.

John D. Collinson Julie K. Morgan

Collinson Jones Consulting, 56 Shropshire Street, Department of Geological Sciences, Snee Hall,
Market Drayton, Shropshire TF9 3DD, UK. Cornell University, Ithaca, New York 14853-
4780, USA.

Mervyn E. Jones Tavi Murray*

Department of Geological Sciences, University Department of Geophysics and Astronomy, Uni-
College, Gower Street, London WClE 6BT, versity of British Columbia, 2219 Main Mall,
UK. Vancouver V6T lW5, Canada.

*Present address: School of Geography, University of Leeds, Leeds LS2 9JT, UK.
Editor's preface

Sediments are now known to undergo deforma- shallow processes and for the pursuit of more
tion in a wide variety of geological circumstances. quantitative relationships. With these goals in
The deforming processes can happen on a vast mind, workers are increasingly drawing on the
scale and at all stages before the material be- principles and methods of the well-established
comes fully lithified. In fact, as exploration of the engineering discipline of soil mechanics.
earth continues, the widespread extent and im- All this is beginning to attract wider geological
portance of sediment deformation is still being interest. Yet to the newcomer, because progress
revealed, for example, below the oceans and has been rapid in recent years, the literature is
beneath ice sheets. At the same time, it is still already formidable. The information is scattered,
being realized just how varied are the resulting so even an expert on sediment deformation in a
structures, and how strikingly similar they can be certain setting may be unaware of analogous
to those produced by the deformation of deeply problems and successes in other environments.
buried rocks. At the same time, although the same basic prin-
However, there are few precedents to guide the ciples apply in the various geological regimes, a
geologist in interpreting structures that formed in subtly different terminology is evolving, which
unlithified sediments, or in understanding the can make the subject boundaries hard to cross.
mechanisms through which they arose. This is The divergent approach and nomenclature used
largely because structural geology has tradi- in soil mechanics add further to the confusion.
tionally been predisposed towards the deep- This book is a first attempt to approach these
seated deformation processes that operate on problems. It is written for structural geologists
rocks long after they have been lithified. The and others who wish to know more about sedi-
concern has been with the mechanics of rocks, ment deformation - i.e. the principles, processes,
particularly under elevated temperatures and terminology and the interpretational possibilities
pressures, almost to the exclusion of considering associated with preserved examples of the struc-
how unlithified sediments deform. Structures tures. Few of the works in each of the environ-
formed in sediments have tended in the past to ments of sediment deformation refer to parallel
have been viewed merely as localized, surficial enquiries in other contexts: I have attempted in
oddities that were not an integral part of the this book to bring together some of these dispa-
structural evolution of an area. This is now rate studies. The worker concerned with sediment
changing, at least in the geological settings where deformation in a particular geological setting can
the scale and significance of pre-lithification de- glimpse comparable studies in other regimes. The
formation have been realized. There is a new geologist used to considering rock deformation
appreciation of the importance of these early may see that structures preserved in sedimentary
structures and a concern with discovering how and metasedimentary rocks cannot simply be
they are produced. There is a realization that assumed to have originated through tectonic
they reflect a long and subtle gradation between stresses operating after lithification, at depth in
near-surface conditions and those at greater the earth's crust. At the very least, the book
depth. Hence the need is growing for a more should provide an entry into the voluminous
rigorous understanding of the mechanics of the relevant engineering and geological literature.
xii Editor's preface

The literature references cited are merely a what is being done now. To give a couple of
representative selection, but their number in illustrations of this in the present context, con-
some chapters provides an illustration of the sider first the existence of structures called sedi-
growing interest in this field of geology. I have mentary dykes. These structures were among
given priority to the most recent works, and tried the very first deformation features of any kind
to minimize duplication between chapters. Even to be recorded, and their origin as earthquake
so, references on some of the topics are so numer- liquefaction effects was essentially understood
ous that the selection in places is quite arbitrary. over a century ago. Yet articles are still published
Also rather subjective is the range of sediment today merely reporting the existence of these
behaviour and structures included here. Some of structures at a particular place, and perhaps
the processes that affect sediment at the same suggesting that they probably formed as a re-
time as or immediately after deposition could sponse to seismic activity. Most other structures
easily be viewed as deformation, but they are seem not to warrant such special treatment: it
largely excluded from the present treatment. To seems that some authors are simply oblivious of
give just two examples, processes such as debris the pedigree of studies into sedimentary intru-
flows and structures such as convolute lamina- sion. As a second, terminological, example, a
tions tend to fall into the realm of sediment- feature recognized back in 1914 was referred to
ology rather than structural geology, and so are then as ball-and-pillow structure, and the term
given less attention here than the deformations remains in use today. Yet along the way different
that arise substantially after sedimentary pro- authors have neglected the precedent and in-
cesses have ceased. Some triggering agents, such vented new names. They have seen fit to call what
as magmatic and volcaniclastic activity, have appear to be essentially similar structures such
not as yet been sufficiently analysed to warrant things as: flow rolls; slump rolls; birds-eye struc-
detailed review. A particular problem has been ture; slump casts; balled-up structure; roll-up
presented by the role of diagenesis in deforma- pebbles; flow casts; flow structure; kneaded struc-
tion. The mineralogical evolution of a sedi- ture; pseudonodules; hassock structure; slump
ment is no doubt critical to its deformational balls; and storm rollers. Therefore, in an effort to
behaviour, and the latter in turn is probably avoid this repeated 'reinvention of the wheel', I
highly influential on the further progress of make a point in Chapter 1 of outlining the
chemical change. Yet what information there is historical framework, and of mentioning some
in the book on this crucial interplay is meagre older review references.
and scattered among various chapters. This Each of the chapters following the intro-
comes about because knowledge on this complex duction has been planned as a state-of-the-
topic is so far glaringly sparse. Our understand- art review of a particular aspect of sediment
ing is simply still too poor to justify a more deformation. Chapter 2 focuses on theoretical
substantial treatment. principles and experimental studies. The follow-
I have written the first chapter as an overview ing chapters progress roughly from settings at
of the material presented in the book as a whole, the earth's surface and shallow levels of burial
and have tried to give a simple introduction to to the generally deeper environments where
the basic terms and concepts before they are lithification becomes complete. So by chapters 6
amplified and applied in later chapters. In these and 7, for example, much of the discussion deals
ways, the introduction is designed to provide with materials where the distinction between a
integration between the following chapters, sediment and a rock is blurred. Some of the
which naturally reflect the differing approaches ideas presented there are equally applicable to
and priorities of each of the authors. Also, the both materials. Chapter 8 discusses case histories
introductory chapter, despite the newness of of the close interaction between some of the
much of the information it reviews, is pervaded processes identified in earlier sections.
by a consciousness that many of the ideas have Some overlap and repetition between chapters
been around for a very long time. Geology is as are unavoidable as each one is meant to be fairly
guilty as any science of a certain arrogance about complete in itself. In places, different nuances of
Editor's preface xiii

emphasis or meaning serve to illustrate contrasts subject. I have tried to avoid any conflict of
in the approach of different workers. Within the information, even though the thinking of some
confines of each chapter, each contributor has of the contributors was evolving as the material
dealt with their subject matter largely as they was being written! After all, despite the great
have felt appropriate. Taken together, the chap- progress during the last 20 years or so in recog-
ters should provide a blend of the theoretical, nizing and understanding sediment deformation,
experimental and descriptive approaches to the the subject is still in its infancy.

Drs Jon Arch, Bobb Carson, Richard Cave and Geological Society of America, Boulder,
Antony Wyatt made helpful comments on early Colorado; Geological Society Publishing
drafts of various parts of chapters 1 and 9. Ruth House, Bath, Avon; Societe International des
Cripwell, Earth Sciences Editor at Chapman & Geologistes Ingineurs, Paris; Ocean Drilling
Hall, helped guide the book through its some- Program, College Station, Texas; Pergamon
times tortuous route to completion. Jeff Schmok Press Ltd, Headington Hill Hall, Oxford; Society
and Gordon Hamilton are thanked for comments for Sedimentary Geology, Tulsa, Oklahoma;
on Chapter 3. Springer-Verlag, Berlin; University of Chicago
The following publishers kindly gave per- Press, Chicago, Illinois.
mission for illustrations to be reproduced, The editor compiled much of the material
from the sources indicated in the figure cap- while on sabbatical leave at the University of
tions: Academie des Sciences, Montrouge, Minnesota. The help of the Geology Department
France; American Association of Petroleum staff and students there, and Peter Hudleston
Geologists, Tulsa, Oklahoma; Association of in particular, is gratefully acknowledged. The
Engineering Geologists, Palo Alto, California; Hartshorn family of Minnetonka made the
A.A. Balkema, P.O. Box 1675, Rotterdam; stay of the Maltman family both practicable
Blackwell Scientific Publishers, Osney Mead, and pleasurable. Finally, Jo, Alastair and Emily
Oxford; Elsevier Science Publishers BV, are thanked for their support throughout - and
Academic Publishing Division, Amsterdam; for time on the home computer.

Introduction and overview


1.1 GENERAL CONSIDERATIONS solid-state movement of grain material to sites of

lower stress. The last process, given the water-
1.1.1 Terminology: sediment, lithification rich environment of most sediments (Schutjens
and deformation 1991), is commonly fluid-assisted and called
pressure solution. This process becomes more
Sediments are accumulations of particles at the significant in the later stages of lithificatiem (e.g.
earth's surface, the result of physical deposition Bjfi)rlykke, Ramm and Saigal 1989; Ruilsback
or, in some cases, biological or chemical activity. 1993), where it can also help generate cementing
Deformation is a change in the bulk shape of the material (Tada and Siever 1989). Cementation
aggregate. The transformation of these loose, operates in parallel with the mechanical effects of
particulate aggregates into rock is termed lithifi- burial by helping to reduce the sediment porosity
cation. (One traditional use of the word rock to (Lasemi, Sandberg and Boardman 1990), a cen-
include all earth materials whether loose or not tral aspect of much lithification. In geotechnical
is unhelpful in the present context.) The change (soil mechanics) usage, all types of diagenesis are
from sediment into rock comes about through commonly referred to as 'cementation' and the
some combination of mechanical and chemical result is a 'structured' material. This terminology
processes. The mechanical aspects, essentially the is usually unhelpful in geology.
moving of the grains closer together so that there As indicated above, how such chemical
is increasing interaction, go hand-in-hand with changes interact with deformation processes has
the deformation processes that form the basis of been explored only in a sketchy manner (Bayly
this book and consequently are dealt with in 1993). At present the topic is insufficiently
detail in the following pages. understood to justify a separate treatment.
The chemistry of lithification, however, Consequently, it is emphasized at the outset that
although studied intensively in many ways, has this book does not deal explicitly with diagenetic
had little systematic analysis in the context of aspects of deformation. Some illustrations are
deformation, and forms only a minor part of the scattered throughout the book (section 4.4, for
present book. The various chemical processes example, discusses some effects associated with
that help turn a sediment into a rock are here growing concretions and section 6.3.5 documents
referred to collectively as diagenesis. In this some mechanical effects of grain bonding), but
usage, diagen<::sis is not synonymous with lithi- the progressive change of a sediment into a rock
fication but a fundamental component of it. and the deformation that arises before complete
Diagenetic reactions gradually increase the chem- lithification is treated here very largely in terms
ical bonding between grains, whether or not the of mechanical processes.
particles are being brought physically closer For many purposes of deformation studies,
(Lade and Overton 1991). The processes are therefore, the diagnostic property of a sediment
dominated by cementation - precipitation into a is that it changes its bulk shape principally in a
pore space; recrystallization - solid-state reorgan- particulate or granular way, by frictional grain-
ization of the grains; and diffusion mass transfer - boundary sliding. The particles are generally

The Geological Deformation of Sediments Edited by Alex Maltman Published in 1994 by Chapman & Hall ISBN 0 412 40590 3
2 Introduction and overview

the strongest part of the aggregate and simply of this spectrum of behaviour are amplified in
slide past each other, undergoing negligible section 2.2.5, and some specific illustrations from
deformation themselves. Any delicate clay do- nature are given in Chapter 8. The converse
mains probably undergo some modification, and situation, where rocks disintegrate through
there are circumstances in which even quartz weathering and return to the particulate state,
grains are affected (section 9.5.3), but most sedi- has received little attention from structural geol-
ment deformation is dominated by what has been ogists but is of great practical importance for
called independent particulate flow (Borradaile engineering geologists. Salt, although regarded
1981). In contrast, rocks deform by utilizing a for most purposes as a sediment, moves via
range of intragranular deformation mechanisms mechanisms that largely fall outside the defini-
(Figure 1.1). Hence, in this view, lithification tion of sediment deformation offered above
consists essentially of the processes that act to (Jackson and Talbot, 1993), and is not dealt with
curb progressively the role of grain-boundary here.
sliding. Such a mechanistic view of lithification has
shortcomings in addition to that of ignoring
the chemical processes. For example, during the
. Increasing course of a geological deformation event,
lithification physical conditions, such as strain rate and fluid
pressure, may fluctuate such that the relative
importance of intergrain and intragrain pro-
cesses varies. A sediment may find itself tempor-
arily under stress conditions that drive grain
breakage. The material could conceivably be
deforming at one time as if it were lithified, and
at some later stage, when conditions have altered,
it could go on to behave as a sediment. More-
over, a particular difficulty for geologists analys-
ing structures preserved in rocks is that the
grain behaviour at the time of deformation
may have been obscured by later events. There-
Increasing fore, the definition used here does little to aid
pore-fluid pressure
the difficult and frustrating business of distin-
Figure 1.1 The relation of frictional grain-boundary guishing between structures developed in rocks
sliding, characteristic of the deformation of unlithified
sediment, to the three main intragranular deformation
and those formed in sediments that were sub-
mechanisms that dominate post-lithification deformation, sequently lithified (section 9.6). However, a more
as conceptualized by Knipe (1986b), from which this practical definition remains elusive. The mechan-
diagram is simplified. (Used by permission of R. Knipe.) ically based approach used here is powerful in
that it allows geologists to capitalize on the
The highly gradational nature of these pro- principles well established in the science of soil
cesses means that materials can also be in transi- mechanics. In time it may become possible to
tional states, that is, partially Iithified or semi- incorporate chemical changes also.
Iithified. In these situations, grain-boundary A cameo of the mechanical effects of lithifica-
sliding remains dominant, but operates in con- tion, wholly qualitative but of practical use with
junction with a greater proportion of intra- modern sediments, is provided simply by the ease
grain deformation. Engineering geologists use with which materials can be indented and cut!
analytical methods and laboratory tests that are Unlithified sediment can be carved with a cheese-
different for sediments (e.g. Mohr-Coulomb knife. Firm sediments require greater finger
parameters and shear tests, section 1.2.3) and pressure to indent them than do soft materials,
rocks (e.g. unconfined compressive strength and and very firm sediments require thumb-nail
joint analysis), and often refer to the intermediate pressure. Bladed knives, progressively sturdier
state as weak rock. Mechanical repercussions with increasing lithification, are needed to cut
General considerations 3

partially lithified materials. Rocks have to be and so the two cannot be compared (Maltman
sawn. 1984). Also, the term soft-sediment lacks pre-
Deformation is bulk shape change, usually cision in that it mayor may not include partial
quantified, just as in rocks, as a strain (e.g. lithification and because softness is not a quanti-
Williams, Goodwin and RaIser 1994). However, fied parameter. The term hydroplastic is used
with sediments the accompanying volume by some workers, but also lacks rigour (section
changes are usually significant and have to be 4.2). Terminological aspects of these localized,
considered along with the deformation. Measure- early deformational processes are discussed by
ments of volume change are referred to as vol- van Loon (1992). One intriguing example of
umetric strain. These concepts are dealt with 'soft-sediment' being used purely to describe
systematically in Chapter 2. the state of material is its application to early
Deformation and bulk volume changes nor- structures formed in igneous rocks, by P.E.
mally come about as a response to an applied Brown, Chambers and Becker (1987).
force. Gravity is the overridingly important body In general, gravitational deformation is domi-
force in sediment deformation. Changes within a nant in the upper parts of a sedimentary pile and
sediment, in particular a fluid-pressure increase, tectonic forces are more significant deeper down,
can reduce the strength of the mass such that it is but there is much overlap and gradation. For
no longer stable in the earth's gravity field, example, buried sediments that have remained
and deformation may ensue. This is the basis of weak, perhaps because their porosity was unable
the various mass movements discussed in section to decrease, can be subjected to gravitational
1.3.3 and Chapter 5. Also, if sediments are deformation at depths of several kilometres (e.g.
deposited such that their density does not M.E. Jones and Addis 1985). Conversely, struc-
increase uniformly downwards in line with grav- tures forming through tectonic forces can form in
ity, instabilities can result which also induce near-surface sediments and can outcrop on the
deformation. A similar effect arises where the ocean-floor (Maltman et al. 1993a).
sediment density gradient becomes modified after A semantic awkwardness surrounds the term
deposition, for example by fluid changes, freez- tectonic, because it can also refer to the form of a
ing, or the addition of igneous material. These geological mass. Thus geologists talk about 'tec-
in-place processes are mentioned further in sec- tonic style' or the 'tectonics' of a district. The
tion 1.3.2 and amplified in Chapter 4. The results meaning is usually clear from the context. How-
can be striking in weak, water-rich sediments, ever, the situation can be confusing in the gla-
but the potential energies resulting from these ciological context, where structures that originat-
gravity driven instabilities are normally too ed through flow of the ice have been described as
small to be of significance for deeply buried tectonic (Sharp, Lawson and Anderson 1988).
sediments. The term glacio tectonic is used for both glacial
Surface forces causing deformation can arise structures and their production from ice, with the
locally, due, say, to a nearby igneous intrusion 'glacio' sometimes being dropped. Lea (1990), for
or from passing ice. They can also operate example, discusses 'glacial tectonism', and Rus-
regionally. In this case, the forces are usually of zczynska-Szenajch (1988) mentions processes
deep-seated origin - the result of lithospheric called 'tectonic detachment' and 'tectonic re-
physico-chemical changes - and they are called deposition', all of which result solely from ice
tectonic forces. The resulting effect is tectonic action. This can lead to situations, for example
deformation. This is not confined to rocks. As is in the Himalayas (Owen 1989), where some
made clear in section 1.3.5 and in later chapters, sediment structures are due to the tectonic
tectonic forces, albeit initiated deeply, can be processes associated with mountain uplift and
transmitted through loose, water-rich sediments others are the result of glacial processes, yet
and cause their deformation even at surficial both are referred to as tectonic! Similarly, intro-
levels. Some geologists have drawn a contrast duction of the term synsedimentary tectonics
between tectonic and so-called 'soft-sediment' (Thomas and Baars 1988) seems unhelpful in the
deformation. However, the former refers to a interests of clarifying the forces responsible for
force and the latter to a material condition, deformation.
4 Introduction and overview

The notion that tectonic deformation can take thought to occur in assocIatIOn with water. In
place in surficial, unlithified sediments may be a order to place traditional structural geology in its
revelation to some, in that it is out of line with full historical context, Historical vignette 1 con-
much structural geological thinking of the last siders briefly why the preoccupation with the
150 years. Yet once, much deformation was deformation of buried rocks came about.

Historical vignette 1: the influence of early thinking

When geological deformation structures were first the various associated contortions. They usually
observed, during the eighteen century, their origin invoked some sort of lateral pushing as an explana-
tended to be ascribed to one of two groups of tion, and it became increasingly common to
processes. They were thought either to be due in call on thermal changes as the source. In this way
some way to the effects of heat or they were the deformation became increasingly envisaged as
result of processes associated with water. It is occurring at depth, where the sediments were
interesting to speculate that it is because, in gen- sufficiently buried to enter the realm of thermal
eral, the former group won out - ideas in which heat effects. Hutton's colleague James Hall was a pion-
was invoked to engender the deforming forces and eer of studying such distorted beds and he was
therefore to operate at depth, and on rocks rather probably the first to attempt laboratory simulations
than subaqueous sediment - that a predisposition of the processes. In 1812 he reported his experi-
arose towards the deformation of buried, lithified ments to replicate folded structures in the labora-
rocks. tory, and the title is telling: 'On the vertical position
The importance ascribed by some early workers of convolutions of certain strata and their relation
to water-based processes came about not only with Granite'. Right from the start, his thinking
because of their significance in the lithological involved thermal effects at depth, and his experi-
and chronological classification systems of the ments simulated the load that was supposed to
time, and a literal belief in the biblical Flood, but overlie the deforming materials (Figure 1.2).
also because they allowed a ready visualization of The die was cast. As time progressed, processes
how deformation could happen. Disturbances of such as subaqueous gliding received occasional
strata from their original horizontal position were mention, but they gradually fell into abeyance.
generally envisaged as collapse phenomena rather When, towards the end of the nineteenth century,
than forced uplift, and erosion by water currents experiments demonstrated that rock subjected to
could easily be pictured as generating channels confining pressure and elevated temperature was
into which the sediments could fall and become capable of flowing, there seemed even less need to
disrupted. Who knows, if this approach had been question the belief that deformation structures
allowed to evolve, gravitational sliding and its necessarily occurred after burial. Deformation com-
significance, for example, may have been acknowl- monly became assumed to occur at depth.
edged much sooner by geologists, and may even This thinking also fitted with the stratigraphical
have become a conventional explanation for de- ideas of the time: sediments accumulated in huge
formation structures seen in sedimentary rocks? troughs, and only after a sufficient thickness had
However, as it happened, the importance of heat been reached was orogeny prompted - long after
gradually began to dominate geological thinking. most of the deposits had been buried and lithified.
James Hutton, for example, successfully demon- In Wales, for example, vast thicknesses of sedi-
strated the igneous origin of intrusive rocks such ments were envisaged as accumulating throughout
as basalt and granite, and he convinced many of Cambrian, Ordovician and Silurian times, and only
the role of the earth's heat in other geological then, at the end of the Early Palaeozoic, did the
processes. He argued that variable thermal ex- deforming stresses arise, to affect materials that
pansions and contractions could best explain the could have seen as much as 200 million years
deformation structures seen in rocks. elaspse since their deposition.
More and more writers of the time were report- The state of lithification of the deforming ma-
ing deviations of beds from the horizontal and terials was equally hazy, largely because the
General considerations 5

Figure 1.2 The pioneering experimental set-up of Sir James Hall involving, from the outset, the simulation of a
presumed burial load and deformation at substantial depth. (Reproduced from Hall 1812.)

relevant processes were very poorly understood. petrification and concretion, were variously men-
The word 'rock' was used very generally - loose tioned, but, again following Hutton's lead, heat
sands and clays, and even ice, were counted as was normally invoked as the chief lithifying agent. It
rocks; Lyell (1838) remarked on the 'insensible therefore followed that the deformation went hand-
passage from a soft and incoherent state to that of a in-hand with or followed lithification. It is
stone' all being encompassed by the single term instructive to quote directly from Hutton, in view
rock. It is still used this way by some geologists, but of the far-reaching influence of his ideas: 'The
for the present purposes the distinction between strata formed at the bottom of the ocean are
loose sediment and lithified rock is essential, as necessarily horizontal ... and continuous in their
the mechanics of deformation of the two materials horizontal direction or extent. There cannot be
are different. Interestingly enough, the term 'con- any sudden change, fracture or displacement. But
solidation' was used by Hutton and other early if these strata are cemented by the heat of
workers in a way approximating lithification, as fusion, and erected with an expansive power acting
some geologists would use it today (section below, we may expect to find every species of
1.2.2), but ideas on how it happened were vague. fracture, dislocation and contortion' (Hutton 1788,
Nebulous processes, with names like congelation, p.229).

Modern thinking, of course, corroborates thinking was profound. Structural geology in-
the fundamental role of thermal changes in the creasingly focused on the deformation of rocks
earth, but heat is not normally invoked today at depth, almost to the exclusion of other kinds of
to explain directly either lithification or de- deformation. Ironically, engineering geology
formation. However, the legacy of the earlier originated through the need to consider the
6 Introduction and overview

deformation of surficial materials under the in- tments both to the grains themselves and the
fluence of gravity! The fact that the two disci- pore spaces between them. A compacted sedi-
plines evolved in opposite directions helps ex- ment has simply lost volume, irrespective of how
plain why cross-fertilization has been delayed for it was achieved. Compaction could result from
so long. thermal changes adjacent to an igneous intru-
sion, for example, or through diagenetically in-
duced shrinkage of the sediment. Consolidation
1.2 MECHANICAL ASPECTS on the other hand, in line with soil mechanics
usage, is restricted to time-dependent mechanical
1.2.1 General reductions in volume. Normally it is pore volume
The physical principles of how sediments deform that is lost, in response to a burial load. The
are well established in the engineering discipline time-dependency comes about because the rate at
of soil mechanics, through efforts to understand which consolidation can proceed is governed by
the behaviour of foundations and slopes. the time it takes to dissipate the appropriate
(Geotechnical engineers and engineering geo- amount of pore water to maintain equilibrium
logists refer to particulate materials as soils (see section 1.2.5). Because of this, Schmoker and
rather than sediment.) Because sediments are Gautier (1989) were prompted to treat consolida-
particulate or granular materials - the solid par- tion in terms of time, but most analyses of
ticles are more or less dispersed within gaseous or consolidation involve the magnitudes of gravi-
liquid phases - the concepts of soil mechanics are tational load (see below and section 2.2.4).
of greater relevance to understanding sediment Squeezing out of the pore fluid dominates the
deformation than, say, fluid or continuum mech- early stages, called primary consolidation, but the
anics. later, fine-scale reductions - secondary consolida-
Many of the concepts and terms outlined be- tion - involve some adjustments to the grain
low and developed in Chapter 2 derive from soil framework. As noted earlier, numerical manipu-
mechanics principles, and they are applied and lation of any or all of these volume changes is
illustrated in later chapters. Recent introductions referred to as volumetric strain.
to soil mechanics and its basic terminology in- Therefore, in the terminology encouraged here,
clude Atkinson (1987), J.A. Barker (1981), F.G. consolidation is one kind of compaction. Both
Bell (1987), enkins (1984) and Whittow (1990). processes help convert sediment into rock. They
Further sources are given in Chapter 2. Exten- are therefore important contributors to lithifica-
sions of the ideas to the submarine situation, tion but they are not synonymous with that term,
which have come about with the growth of which, as emphasized in section 1.1.1, also in-
various aspects of offshore engineering, were cludes diagenetic (chemical) effects. Consolida-
reviewed by Bouma et al. (1982), Denness (1984), tion can come about in response to tectonic
Demars and Chaney (1990), Sangrey (1978, 1982) forces (e.g. Minshull and White 1989), but it is
and Poulos (1988). The behaviour of particulate normally the result of sediment burial. It is in the
aggregates from the point of view of applied latter context that the mechanics of consolidation
mechanics and material science, respectively, was are usually considered, but analogous processes
reviewed by Satake and Jenkins (1988) and operate in sediments being loaded by ice (Boul-
Shahinpoor (1983). ton and Dobbie 1993).
As sediment becomes progressively overlain by
further additions of material, in the earth's grav-
1.2.2 Volume changes due to burial ity field it becomes subject to an increasing load,
In geology, words such as compaction, consoli- of a magnitude that depends on the thickness and
dation and lithification are often used more or density of the overlying sediment. The load is
less interchangeably, but if we are to understand variously referred to in geology as burial, over-
how sediments deform it is necessary to adopt a burden or lithostatic pressure. In experimental
more rigorous usage. Lithification has already work it is simulated by the cell or confining
been defined (section 1.1.1). Compaction is useful pressure. In nature, it tends to be dominated by
as a general term for all permanent reductions in the vertical, downward component, and, being
the bulk volume of an aggregate, including adjus- directional, is better called a stress, although it can
Mechanical aspects 7

be of equal magnitude in all directions. The although occasionally wet water content is used:
latter situation, in analogy with the stress con-
figuration that exists in water at rest, is referred W
%= ,,:eight of pore water .
to as hydrostatic. It is useful to use pressure where weight of wet aggregate (1.4)
no direction is implied, and stress where a The latter has the advantage of allowing porosi-
directional component is meant. Deviatoric, or ties to be readily calculated from it if the relevant
differential stress, where shear stresses arise densities are known (Boyce 1976). Modern
because the stress magnitudes in the three, methods of determining water content are out-
orthogonal, principal directions are unequal, lined by Gilbert (1991).
becomes significant when considering deforma- Water content governs the consistency of sedi-
tion (section 2.2.2). ments. Most fine-grained materials, because of
The magnitude of horizontal, or lateral, burial the presence of clays and organic matter, are
stress depends on the ratio referred to as the K naturally plastic and easily achieve unrecoverable
value, where: deformation without fracturing, that is, they exist
within their plastic range. A drier sediment, how-
magnitude of horizontal stress
K (1.1) ever, may fall below its plastic limit, and a wetter
magnitude of vertical stress one can exceed its plastic range by having a water
If K = 1, the sediments are subject to hydro- content above its liquid limit, and flowing under
static stress and undergo isotropic consolidation. its own weight like a fluid. The difference in water
A more realistic situation is for the horizontal contents between the two limits is called the
stresses to be just sufficient to prevent lateral plasticity index. Such limiting values to plasticity,
strain in the sediment, a condition termed K o, sometimes called the Atterberg limits, can be of
the coefficient of earth pressure at rest (sections geological relevance, for example, in assessing the
2.2.6 and 6.2.3; Mesri and Hayat 1993). Ko con- stability of sediments on slopes (Einsele 1989)
solidation tends to induce anisotropy in the sedi- and the form of mass flows (section 1.3.3). Coarse
ment, with corresponding effects on strength and sediments, especially those without clay, tend to
stress-strain behaviour (Atkinson, Richardson lack plastic properties. Note that this usage of the
and Robinson 1987). Other implications of the term plastic differs slightly from that describing a
Ko condition have been explored by Daramola rheological behaviour (section 1.2.4), in that no
(1980), Garga and Khan (1991) and Kitamura yield stress or strain-rate requirement is implied.
and Shinchi (1988). Plasticity tends to increase with greater carbon
Increasing burial pressure tends to progressive- content (Bennett et al. 1985): Booth and Dahl
ly pack the grains closer together. Assuming that (1986) documented in a clayey silt from offshore
pore fluids are able to dissipate adequately, a southern California, a 20% and 14% increase in
critical matter that is developed later, this de- the liquid limit and the plastic limit, respectively,
creases the porosity (,,) or void ratio (e) of the with a mere 1% increase in organic carbon.
aggregate, where: Porosity reduction that maintains mechanical
equilibrium with the burial load is termed normal
volume of voids volume of voids consolidation. Numerous theoretical and empiri-
e= cal attempts have been made to quantify the
volume of aggregate' volume of grains
relationships (Baldwin and Butler 1985; Dzevan-
shir, Buryakovskiy and Chilingarian 1986; Bayer
,,=--; e=--. (1.2) and Wetzel 1989; section 6.2.1). Observations on
l+e 1-" porosity loss with burial on the modern ocean
In much of geology, " is multiplied by 100 to floor are reported by Faas (1982) and Shepard
express porosity as a percentage. Assuming water and Bryant (1983). It is clear that in nature
is occupying the pores, there takes place a con- porosity loss does not always keep pace with
comitant decrease in the water content (w), nor- progressive burial, which leads to underconsolida-
mally expressed in per cent by weight: tion. The most common explanation involves the
sediments being unable to expel the pore water at
weight of pore water a sufficient rate. Overconsolidated sediments, with
W= x 100 (1.3)
weight of dry grains porosity less than that expected from their burial
8 Introduction and overview

level, are also well known in nature (Mayne 1988; usefulness of this equation is reflected by its use
Rafalovich and Chaney 1991). A common expla- in several of the following chapters. (Some other
nation of this effect invokes diagenesis, as cemen- criteria for defining the failure of a sediment are
tation can independently reduce the porosity, mentioned in section 3.6.1). Strictly, the last two
though without an overall compaction (Ehren- terms of the equation are solely mathematical
berg 1989). Subaerial sediments commonly be- quantities and not physical attributes of the sedi-
come overconsolidated through geological uplift ment, but there are numerous observations that
and erosion of the overlying material. In de- point to a relation between the friction coefficient
tail, consolidation is a complex phenomenon, in- and the nature of grain surfaces (e.g. Koerner
teracting with compression of the sediment 1970; Frossard 1979). For this reason, angular
framework and proceeding at variable rates sediments, such as carbonates, give high friction
(Sridharan and Rao 1982; Znidaric and Schiff- angles and are relatively strong (Semple 1988). In
man 1982; Yong and Townsend 1986; Nagaraj, unsaturated sediments, increasing water content
Vats ala and Sriniva Murthy 1990; Griffiths and decreases the magnitude of cohesion and friction,
Joshi 1991). The influence on consolidation prog- and hence shear strength (Yoshida, Kuwano and
ress of different lithic grains in a sand framework Kuwano 1991). Note that discussion of the role
was investigated by Pittman and Larese (1991). of fluid pressure in these relationships is deferred
The various states of sediment consolidation are until section 1.2.5.
addressed in greater detail in sections 2.2.4, 2.2.7, The shear strength of sediments is sometimes
2.3 and 6.2. expressed in words such as hard, stiff, firm and
soft. Sediments that have undergone intense dis-
turbance are termed remoulded. The sensitivity of
1.2.3 Sediment strength a sediment is the ratio between shear strengths in
As sediment particles are being deposited they the undisturbed and remoulded states, a differ-
increasingly interfere with each other (Martinez ence that can be marked in clays. The ratios can
et al. 1987; Tan et al. 1990). The gradation exceed 4 in sensitive clays, and values in excess of
between the stage of hindered settling (Pane and 16 are a characteristic of quick clays. To give
Schiffman 1985; Toorman and Berlamont 1991) some quantitative view of strength: very weak,
and final deposition has been the subject of near-surface sediments typically have shear
sedimentological study (Mehta 1991) and much strengths of a few kilo pascals (kPa), increasing to
engineering research, in contexts ranging from hundreds of kPa with tens of metres of normal
sludge disposal to dredging (Yong and Townsend burial, and perhaps to a few MPa on lithification.
1984). As the particles are brought closer to- A strong sandstone might sustain a shear stress
gether, electrostatic bonds are initiated between of several hundred MPa. To complete the pic-
them, imparting a physical cohesion to the sedi- ture, a shear strength of around 100 kPa is
ment (Wetzel 1990). Also, movement between achieved at the base of a 10-m pile of dry sugar!
adjacent grains begins to involve intergrain fric- The strength varies by about 20 kPa depending
tion. For either or both reasons the sediment on whether the sugar is loosely or densely packed
acquires some resistance to an applied shear (Feda 1982).
stress. (The concept of shear stress, and its differ- Progressive packing of the sediment particles
ence from a principal stress, is explained in sec- during burial provides further opportunities for
tion 2.2.2). The maximum shear stress the sedi- intergrain bonding and increases the normal
ment can sustain before failure is called its shear stress acting across the grain boundaries. In gen-
strength. The time-honoured view of the relation eral, and still ignoring fluid pressure effects, it is
between these parameters utilizes the Mohr- the effect of the normal stress on the frictional
Coulomb equation: term that is responsible for increasing sediment
strength with burial. An observable repercussion
(1.5) of this is that clayey sediments, with their elec-
where" is the shear stress sustained at the point trostatically charged large surface areas, acquire
of maximum resistance, an the stress normal to greater cohesion than clastic materials and can
the plane of failure, c is termed cohesion and cjJ be stronger at shallow burial levels, whereas the
the angle of internal friction. The continuing latter are normally the stronger material at
Mechanical aspects 9

depth, as the frictional contribution begins to

dominate. For example, some surficial volcanic
sediments weather to give clay minerals that softening
appear to impart a particularly strong bonding
to the sediment, hence generating strong sedi-
ment at shallow levels of burial (Belloni and
Morris 1991). The proportion of clayey material
in the matrix of glacial sediments is a paramount Shear strain
influence on their deformational response to ice :;-1 gj
<I Q)
(section 3.4). However, in general, increasing
Dense sands or
c . <:: overconsolidated clays
burial reduces the ability of the grains to over-
come the frictional resistance to their sliding and 13 gj
ttl Q)
Loose sands or
___ ; normally consolidated clays
E 1t;
Q) Q)
the sediment grows in strength - the sediment OJ Q)

undergoes lithification. (a) g 0

1.2.4 Sediment deformation

The deformation of a sediment probably begins
with a small amount of elastic strain, an aspect
considered in some detail in Chapter 6. However,
for most sediments, the elastic range is distinc-
tively smaller than that of their lithified equival- PH?O Axial strain

Increasing overconsolidatron
ents (section 2.2.1) and its effects become ob- <I 3l
S:-r ttl

scured by the grain-boundary sliding that rapidly

Q) '-
ensues (Hardin and Blandford 1989). Many sedi- c..<::
ttl ~-.~~
llncreasing overconsolidation
.!: Q)
ments, and particularly dense sands and overcon-
solidated clays, show a high point in the stress-
E - - - - - - Drained, normal consolidation

strain curve, known as the peak strength, nor- (b).2 ~

mally equivalent to the shear strength. This is Figure 1.3 SynoptiC diagram of typical behaviour
followed by an overall strain softening as they patterns of deforming sediments. (a) Densely packed and
drop to some value of residual strength (Lupini, loose sands, and normally and overconsolidated clays
subject to shear. (b) Normally and overconsolidated clays
Skinner and Vaughan 1981), where additional subjected to drained and undrained triaxial deformation.
strains are achieved at virtually constant stress.
Loose sediments show no peak, and gradually and 50% give a transitional behaviour, the exact
attain a steady state; others show some degree of nature of which depends on both the precise
strain hardening. In the laterally confined axial- nature and amount of the clays. A 50% clay
shortening arrangement known as triaxial test- content gives a marked drop to residual, which
ing, the sediment behaviour depends crucially does not increase in conspicuousness with yet
upon whether the pore fluid is trapped in the higher clay proportions. The role of friction is
specimen - an undrained test - or allowed to illustrated by montmorillonite clays (¢ = 5°)
escape - a drained test (see Figure 2.26). A synop- showing the lowest residual strength, illites
sis of typical behaviours is given in Figure 1.3. (¢= 10°) having intermediate values, and kao-
In the residual state, any cohesion possessed by linites (¢ = 15°) showing the highest residual
the aggregate will have become negligible and the strength values of these three minerals (Skempton
behaviour governed by its frictional characteristi- 1985). R. Moore (1991) pointed out the signifi-
cs (Burland 1990). The drop to a residual strength cant influence of pore-water salinity on these
is well shown by clays, where it comes about values, and on the cohesion and plasticity of
through the reorientation of the platelets into clays. Moore's measurements, on subaerial clays,
zones of alignment. Skempton (1964) showed that showed an increase in residual strength of up to
sediments with 20% clay content behave essen- 21 % as the pore fluid was changed from fresh to
tially like sands or silts, proportions between 20 salt-water.
10 Introduction and overview

Attempts to document the quantitative stress- greater strains than are likely to be achieved in
strain behaviour of sediments during deforma- nature at the critical state, because of irregulari-
tion, largely by laboratory testing, have led to the ties and fabric changes inducing volume fluctu-
concept of stress paths (Shibuya and Hight 1987; ations, so it may be that they developed in the
Vaid, Chung and Keurbis 1990) and their por- residual condition. Here also the stresses remain
trayal on stress path diagrams. These are pres- approximately constant with increasing strain,
ented in two or three dimensions, and various but the fabric collapse that typically allows the
parameters are plotted on each of the axes high strains to be achieved in turn engenders
(Jamison 1992). The familiar Mohr-circle con- volume loss.
struction uses a two-dimensional example in r-(J Modifications to fabric are an integral part of
space, as pursued in section 2.2.5. More common sediment deformation but so far the contribution
in soil mechanics treatments are p-q diagrams, as that soil mechanics results can make to geology is
explained and applied in section 2.2.5. Further restricted, because the work is still largely quali-
applications of p-q stress path diagrams arise in tative. There are numerous descriptions of the
sections 6.3.5 and 8.2. interplay of consolidation and deformation with
One of the chief shortcomings of the Mohr- microscopic fabric (Clark and Gillott 1985;
Coulomb approach to sediment deformation is Rothenburg and Bathurst 1989; Wesley 1990)
that volumetric strain is not directly taken into and macroscopic features such as joints (Kinkal-
account, although significant volume change is die 1992), fissures (Costa-Filho 1984), and bed-
characteristic of many deforming sediments. For ding surfaces (Eigenbrod and Burak 1991), but
example, sands and normally consolidated clays attempts to quantify fabric changes and integrate
tend to pack more tightly in response to stress them with models of stress and volume changes
(e.g. Sladen and Oswell 1989), thus showing con- remain elusive. A further complication is the
tractive behaviour, whereas materials such as effect of time, for it is well known in soil mechan-
dense sands and overconsolidated clays show ics that fabrics can change even when the sedi-
marked dilative behaviour as they approach fail- ment is at rest - a condition known as ageing
ure (Hardin 1989; Shimizu 1982). Some geologi- (Schmertmann 1991; Tsuchida, Kobayashi and
cal repercussions of dilation were considered by Mizukami 1991).
Brodzikowski (1981). A further condition becomes relevant where
As a consequence of the importance of volume long periods are considered, namely the pheno-
change, p-q stress paths are commonly combined menon of creep. Here is a good example of a term
with the volumetric responses of a deforming being used in slightly different ways, according to
sediment into a three-dimensional diagram (sec- context. Sedimentologists and geomorphologists
tion 2.2.5). This approach has led to a series of tend to use it to indicate any very slow - perhaps
synoptic models, mostly based on that originally centimetres per year - gravitational downslope
developed at the University of Cambridge for movement (e.g. McKean et al. 1993), without any
clays, and referred to as Cam-clay theory particular implication of the stress conditions,
(Srinivasa Murthy, Vats ala and Nagaraj 1991). strain rate evolution or volume change. The
The concepts are now extended to sands (Been, treatment of mass movements in Chapter 5, for
Jefferies and Hachey 1991) and carbonate sedi- example, uses the term in this way (see, especially,
ments (Coop 1990). One aspect of such models is section 5.7). In addition, although the movement
their prediction that sediments reach a certain tends to differ from slope failure by being more
volume-stress combination that remains con- distributed, an analogous localized creep is men-
stant during additional deformation, a condition tioned in section 9.5.1 as an inferred explanation
known as the critical state. Hence these stress- for certain detachment structures. In some soil
volume analyses are commonly called critical mechanics usage, on the other hand, creep has a
state models. The principles are explained in stricter meaning of continuing long-term strain,
section 2.2.5 and, for example, by Wood (1990). not necessarily downslope, and typically showing
The applicability of critical state models to vari- a three-stage response to low-level stresses. Silva
ous submarine sediments is illustrated in sections and Booth (1985) suggested that sediments mov-
6.3.5 and 8.2. Many structures of geological inter- ing slowly downslope may actually be following
est, however, appear to have involved much this latter pattern, in that their laboratory testing
Mechanical aspects 11

of slope sediments from offshore northeastern linear response to shear stress it is termed a
USA reproduced the expected three-stage creep Bingham fluid (see Figure 3.4):
behaviour. Some implications of this work for
analysing slope stability effects are mentioned (1.9)
in section 1.3.3. Yet a further variation in the where )1p is the plastic viscosity, which incorpor-
meaning of creep is illustrated in sections 2.2.4 ates the excess shear stress sustained over that in
and 6.2.3, where the term is used synonymously Newtonian behaviour, and cy is the resistance to
with secondary consolidation, for the slow shear.
volume loss that may continue after the pore Most sediments do not show such a linear
fluid appropriate for the burial load has been response, partly because the viscosity itself varies
expelled. with increasing strain rate:
Much of the analysis of sediment deforma-
tion outlined above is based essentially on a
Mohr-Coulomb view of the behaviour. The value and so more complex relationships are required
of this stress-based approach is thoroughly estab- (see Barnes, Hutton and Walters 1991). Accord-
lished in soil mechanics and it will be used ing to whether the viscosity increases or de-
repeatedly in the following chapters. It has creases with strain rate, the behaviour is referred
proven an appropriate treatment for many to as shear thickening or thinning, respectively.
geological situations. An alternative approach, The change may be gradual (Figure 3.4) or form
however, is to view the deforming sediment a series of plateaux. Wet clayey sediments, for
as some kind of fluid and to consider its rheo- example, commonly show progressive shear thin-
logical properties. This method also has engineer- ning (Philip Harrison, personal communication).
ing proponents (Keedwell 1984, 1988; Vyalov Many mechanical analyses of geological sedi-
1986), but has so far been less widely used in ments use an elastic-plastic rheology. Here, an
geology. Examples of the rheological approach initial elastic (recoverable) strain gives way at a
arise elsewhere in this book; a brief introduction yield point to bulk plastic behaviour (unlimited,
follows. rate-independent strain beyond the yield stress).
Settling sediments are commonly regarded as The elastic strains, by definition, are not recorded
approximately Newtonian fluids (no shear in the sediments (section 6.1), and the latter mode
strength, and a shear-strain response propor- dominates in any prolonged deformation. The
tional to strain rate): range of elastic behaviour is usually considered
to be minimal in poorly lithified sediments, but
dy ,
(1.6) its importance grows with progressive increase in
dt =~' interparticle bonding. The strain response of a
where y is shear strain and t is time. The propor- wholly elastic solid to a given stress will, follow-
tionality term, )1, is termed the viscosity. In integ- ing Hooke's Law, depend on the proportionality
rated form (1.6) becomes: constant of the material, known as Young's
(1.7) modulus. The elastic strain at right-angles to the
direction of maximum strain, for example the
or lateral extension accompanying greatest shorten-
1 ing in the vertical, is given by the Poisson's ratio
Y=-' (1.8) of the sediment. These parameters are pursued
further in section 6.2.3; other parts of Chapter 6
where y is the change in shear strain with time. In and section 7.6.2 analyse the contribution of
reality, as much work on materials such as paint, elastic behaviour to sediment deformation.
drilling mud and wet cement has shown, the Much sediment deformation beyond the yield
behaviour of particle suspensions is much more point is somewhat rate dependent - most stress-
complex than this. As the settling approaches strain curves are curved, with the slope at any
completion the particles increasingly interfere particular point dependent on the rate of stress
with each other and the aggregate acquires increase - and so some rheological models use a
significant strength (section 1.2.3). If this viscous rather than a plastic behaviour. Here the
strengthened sediment is still able to show a strain rate is related to the stress level. Such
12 Introduction and overview

non-linear viscous fluids, provided the viscosities sediment (i.e. the pore water is continually expel-
are not extreme, are commonly represented by a led to keep pace with progressive burial), it is
power law of the form: called normal fluid pressure - and is that portion
of the burial load arising from the overlying
column of pore fluid (plus sea water in the case of
where x is called the consistency of the sediment submarine sediments). The remaining component
and n the power law index. is that sustained by the sediment particles and is
Realistic modelling of geological sediments called the effective stress, (J':
may well have to combine several models of
a' = a - PH20 (1.12)
rheological behaviour, and the power laws can be
complex. Kutter and Sathialingam (1992), for where (J is the total stress and PH20 the fluid
example, attempted to incorporate rate-depend- pressure. This symbolism for fluid pressure is
ency into their analysis of clay deformation, lead- used throughout this book, in line with much
ing to an elastic-viscoplastic model. Other, more geological usage. In soil mechanics literature the
or less complex formulations are illustrated in the symbol u is commonly used for fluid pressure. It
following chapters. Section 1.3.3, for example, is the effective rather than the total stress that
mentions viscous-based modelling of landslides, controls deformation (see later), and it is there-
section 3.6.1 discusses several flow laws that have fore best to modify the Mohr-Coulomb relation
been used for the base of ice sheets, and section (1.5) in terms of parameters with respect to effec-
5.1.3 describes some variations on the basic tive stress:
models in the context of mass movements. Thus,
'I: = c' + a~ tan <p'. (1.13)
for many geological deformations, the crucial
step in a rheological analysis is the choice of the A prime symbol added to a term indicates that
most appropriate model or combination of allowance is made for pore fluid pressure. Sedi-
models. The value of the analysis depends on its ment that shows contractive behaviour during
realism and on the simplicity or otherwise of the deformation (section 1.2.4) will undergo a simul-
resulting flow law. taneous, more or less transient, fluid pressure
increase (effective pressure decrease) while dila-
tive materials will experience an accompanying
1.2.5 Mechanical role of pore fluids decrease in pore pressure. Such effects on the
Much of the analysis of fluid movement in geo- pore fluid pressure and their embodiment in the
logical material has been in the realm of ground- concept of effective stress is a fundamental aspect
water studies (e.g. Nielsen and Johnson 1990), of the mechanical behaviour of sediments and is
petroleum geology (e.g. Robinson in press), and, amplified below. The importance in soil mech-
more recently, waste disposal (e.g. Devgun 1989; anics is emphasized by Wood (1991), and some
Valent and Lee 1976). Earlier literature on water further ramifications are discussed by Mayne and
resources from unlithified sediment was reviewed Stewart (1988), Sridharan and Rao (1979) and in
by Dudgeon and Yong (1969). However, the section 6.2.3.
crucial importance of pore fluids for soil engin- With burial, the pore fluid has to be expelled in
eering was established long ago, and it is clearly appropriate amounts if the pore volume is to
the same in geological sediments. The foregoing reduce in equilibrium with the additional load.
outline of sediment behaviour has ignored this Possible variations in porosity and void ratio
fundamental effect only for the sake of introduc- with vertical effective stress are given in Figures
ing topics systematically. General principles of 6.1-6.3. The ability of a sediment to dissipate
fluids in geology are discussed by Bear (1972) and pore fluid is a function of its permeability, as
Fyfe, Price and Thompson (1978). explained in section 7.2. This in turn depends on
In most geological sediments the fluid is the configuration of the pores and particles in the
water of some sort, and this supports some sediment (Beard and Weyl 1973; Ahuja et al.
portion of any applied load, giving rise to a pore 1989; Bloch 1991; Bryant, Code and Meller,
or fluid pressure. Both terms, together with pore 1993). Although most quantitative models of
fluid pressure, are widely used in geology and consolidation make the simplifying assumption
throughout this book. In normally consolidating of constant permeability, Schiffman (1982) has
Mechanical aspects 13

emphasized that this critical parameter is likely in shows that interbedded sands and muds should
nature to vary sensitively with different amounts develop a higher proportion of lateral drainage
of burial. The expulsion of pore water during than more homogeneous sequences (Magara
laboratory tests is known as drainage in soil 1976). Bredehoeft, Djevanshir and Belitx (1988)
mechanics and in geological sediments as dew- have documented lateral dewatering on the re-
atering (Burst 1976). The process is central to gional scale.
consolidation under the effect of gravity (AI- We come now to the profound implication that
Tabba and Wood 1991), but in geology, tectonic overpressuring has for sediment mechanics.
dewatering is being increasingly recognized in Overpressuring weakens sediments. Because the
addition (Carson and Berglund 1986; Shi and fluid is sustaining an extra part of the stresses
Wang 1985). acting across the aggregate framework, the effec-
If, after a given increment of loading, a sedi- tive stress decreases, the intergrain friction is
ment lacks sufficient permeability for the pore reduced (O'~ tan 4>' in equation 1.13), and hence so
fluid to escape sufficiently, the fluid, being vir- is the sediment strength. This is why, as stated
tually incompressible, has to sustain a dispropor- above, it is the effective rather than the total
tionate part of the load (Gretener 1981). This stress that governs deformation. Overpressured
pressure excess over the normal fluid pressure is horizons are commonly sites of shear failure (e.g.
termed excess fluid (or pore) pressure or, in geo- section 1.3.3 and Chapter 5). Overpressured
logy, overpressure, geopressure, hydropressure, ab- sediments may undertake intrusion (section 1.3.4)
normal or supernormal fluid pressure. Subnormal and diapirism (sections 4.3.2 and 2.3.3). The re-
or negative fluid pressure, sometimes referred to duction of effective stress and strength by over-
as underpressure (section 6.2.4), is also known in pressuring is a crucial factor in all settings of
nature (Belitz and Bredehoeft 1988). The amount sediment deformation. Two examples follow.
of excess pressure is often expressed as the fluid Lash (1985), in his analysis of Ordovician rocks
pressure ratio (.Ie), where: in the central Appalachian tectonic orogen of
pore fluid pressure
eastern Pennsylvania, used fluctuating effective
A= . ' (1.14) stresses to explain the differing structural evol-
total bunal pressure ution of neighbouring sand and mud horizons.
Excess pore pressure is sometimes defined with Initial deformation increased pore pressures in
respect to steady-state flow conditions (Gibson, sand beds contained within mud-rich horizons,
Schiffman and Whitman 1989); Audet and and the reduced strength led to flow and stratal
McConnell (1992) presented a mathematical disruption. The interlayered mud, because of its
analysis of the process. Normal fluid pressure is greater cohesion, became the stronger sediment
also commonly called the hydrostatic pressure, in and deformed only along discrete zones. These
both geology and engineering. The types of press- eventually interweaved to produce a scaly foli-
ure gradients likely in a sedimentary sequence are ation, which facilitated dewatering of the mud,
represented schematically in Figure 1.4. Real in thus further increasing its strength. Meanwhile
situ measurements are difficult to obtain and are sand beds that lacked mud envelopes had
still in their infancy (e.g. Hurley and Schultheiss drained adequately and maintained strength and
1990; Carson et al. in press). coherence throughout the early deformation. As
The pressure head in overpressured fluid is no deformation continued, the contrasts in drainage,
longer balanced out by the elevation head (sec- and hence effective stress, strength and deforma-
tion 7.2), as in the equilibrium condition, and so tion behaviour, progressively decreased, and the
the fluid attempts to move to lower values of the various lithologies converged to a more uniform
potential gradient (Chapman 1981). There are behaviour. Abbott, Embley and Hobart (1985)
numerous geological repercussions of this, es- found that those sediments in the South Pass
pecially if the fluid is able to flow with sufficient district of the Mississippi Delta that gave per-
velocity to entrain the host sediment. Note that meability measurements of only 4 x 10- 14 m 2
the fluid potential gradient, although upwards (expressed by those workers as hydraulic conduc-
overall, can be downwards locally. This is why tivity, see section 7.2), had shear strengths around
sedimentary dykes, for example, do not necessar- 3 kPa and were highly deformed. In contrast,
ily intrude upwards (section 1.3.4). Section 7.4.3 sediments with permeabilities of 1.8 x 10- 13 m 2
14 Introduction and overview

Pressure (MPa) Pressure (MPa)

o 5 10 15 20 o 5 10
ot----.- Steeper gradient at shallow levels, because of rapid
porosity loss increasing density of aggregate
Narrow, poorly permeable, overpressured horizons
100 seawater with intervening horizons of sufficient lateral permeability
depth for overpressuring to be partially reduced

200 Permeability adequate for consolidation of these

\ ...._---'r-sediments, but insufficient to dissipate completely
the nearby overpressured horizons

400 Broad, poorly permeable overpressured

horizon, with broad transition zone

E 600
Fluid-pressure gradient
(J) 700 Effective-
0 gradient
800 Poorly permeable material:
overpressuring increases with Fluid-
depth potential'


Normal fluid-pressure gradient
(= 'Hydrostatic' pressure gradient)
(a) (b)
Figure 1.4 Patterns of pressure increase with depth. (a) Rates of increase of pressure and stress due to burial. The
gradient for normal fluid pressure is based on an overlying seawater depth of 250 m and a constant fluid density in the
entire column of 1025 kg m- 3 . The burial-stress gradient is based on a grain density of 2650 kg m- 3 and a porosity decrease
with depth following the average values for sand-silt-clay of Einsele (1989). The hypothetical fluid-pressure gradient
shown here illustrates the kinds of effects that can arise in a sequence of varied sediments. Compare with Figure 7.5.
(b) Effective-stress gradient and fluid-potential gradient derived from the curves in (a). Note that the fluid potential and its
gradient should properly be expressed in units of metres (the height that could be supported by a column of specified
fluid, normally water or mercury) rather than pascals. The fluid potential, or total head (H), is here a pressure head (<P),
arising from the excess fluid pressure ( PH2 0 See Figures 7.1 and 7.2, and section 7.2 for further explanation.
(From Maltman (1994) Reproduced with permission of Pergamon Press.)

showed strengths around 8 kPa and were vir- Although it is a long time since Karl Terzaghi
tually undeformed. In the former materials, the clarified how these kinds of factors prompted
temporarily trapped pore water was overpres- slope failure rather than any effect of lubrication,
sured and the reduced effective stress and sedi- it is surprising how often this latter notion is still
ment strength were allowing greater strains. invoked in geological explanations of moving
Overpressuring as a cause of hydraulic fracture sediment masses. Most silicate minerals are not
has been given great attention in rocks, but in lubricated by water, and most unlithified sedi-
sediments this aspect has so far been little ad- ments contain too much water for the thin-film
dressed (Murdoch 1992). phenomenon of lubrication to be of relevance. In
Mechanical aspects 15

fact, Savina (1983) remarked that invoking lubri- best simplified in current quantitative models
cation in mass movement betrays those geol- (Stevenson and Scott 1991), and from the fluid
ogists who lack an understanding of the mechan- comprising more than one physical phase (Parker
ical principles involved. It is fluid pressure that 1989). Permeability remains poorly investigated
holds the key, and the geological repercussions of in many geological situations. Vast numbers of
this relationship between permeability, overpres- on-land measurements have been collected in
sure, effective stress and sediment strength - and engineering hydrology, but the effects on permea-
hence intensity of deformation - are enormous, bility of, for example, inhomogeneities in the
and further examples arise frequently throughout sediment (Atkinson and Richardson 1987; Bethke
this book. 1989), anisotropy (Arch and Maltman 1990) and
Overpressuring can originate in numerous dilatancy (Harp, Wells and Sarmiento 1990) re-
ways in geological situations, such as through main to be understood. In situ submarine
mineral dehydration (Colten-Bradley 1987), bio- measurements are in their infancy (Fang, Lang-
pressuring resulting from organic decay (Nelson seth and Schultheiss 1993).
and Lindsley-Griffin 1987), and gas liberation Finally, fluids are undoubtedly central to dia-
from clathrate decomposition (Carpenter 1981). genetic reactions and their effect on sediment
Further mechanisms are indicated in section mechanics. As mentioned earlier, the interplay
2.3.4. Where the dissipation of pore fluid is great- between deformation and diagenesis has been
ly curbed, or increments of loading arise in addi- little explored. It may be that the chief effects of
tion to burial, say from the transmission of seis- deformation on diagenetic reactions are not di-
mic waves, the pore fluid may have to rect but through the production of features that
temporarily sustain the entire stress acting on the affect fluid distribution. For example, rates of
sediment. In this situation the internal friction is ionic diffusion through pore water have little
effectively reduced to zero, and the aggregate, direct sensitivity to pressure, whereas the permea-
now lacking resistance to shear, behaves as a bility reductions associated with fabric changes
liquid. The resulting phenomenon of liquefaction during consolidation will reduce diffusion mag-
causes a number of geological structures in sedi- nitudes by a half to a twentieth of the value in
ments (section 1.3.2) and has been much studied free solution (Manheim 1971).
by engineers (R.D. Davis and Berrill 1983; Seed In contrast, there is more documentation of the
1985; Richards, Elms and Budhu 1990). Similar progress of diagenesis directly affecting sediment
results can arise in sediments with the particular- deformation. Boudreau, Brueckner and Snyder
ly delicate fabric known as quick clays (Torrance (1984), for instance, noted diagenetic influences
1983). Marine geotechnology has been much con- on the pre-lithification deformation of Upper
cerned with overpressuring and liquefaction aris- Palaeozoic rocks in Nevada. The silica in clay-
ing through the height of storm waves temporar- rich siliceous sediments tends to remain as dia-
ily increasing the fluid head acting on the pore genetically immature opal-A and opal-CT longer
fluid (Kraft et at. 1985; Okusa 1985). In addition than in clay-poor sediments, which convert
to these direct consequences of overpressuring, at rapidly to quartz. Therefore, the former material
sites of high fluid-pressure gradients, the drag tended to deform in a more ductile manner than
force exerted on a grain framework by a flowing the latter, quartz-enriched and hence more brittle
fluid, termed the seepage force, may be sufficient sediment. At the same time, the diagenesis was
to exceed the sediment strength and to prompt enhancing the ductility contrast between the ad-
failure (Orange and Breen 1992). jacent materials, thus promoting boudinage and
Nevertheless, despite all these repercussions, a flexural-slip mechanism of folding. Moreover,
the behaviour of fluids in sediments is far from because loading during burial caused preferential
being fully understood. Although water is the dissolution of silica from the clay-rich layers, the
common fluid in nature, it can have variable remnant, now clay-enhanced, laminae were able
chemistry, the effects of which are little explored. to act as the slip surfaces during the folding. An
Moreover, biogenic gas will often be present in analogous example from Miocene rocks is given
near-surface materials (Wheeler 1989; Sills et at. in section 9.5.1. Maliva and Siever (1988) sugges-
1991). Further complications arise from the fluid ted that some mineral textures produced during
interacting with the sediment grains, an aspect at lithification are the result of stresses generated by
16 Introduction and overview

diagenetic reactions - a force of crystallization that the deformation was due to submarine slid-
induced by new mineral growth. Tribble (1990) ing. However, comparison of the shear stresses in
documented the progress of clay diagenesis in the the likely slope conditions with the appropriate
Barbados accretionary prism and, through its mechanical parameters derived from laboratory
effect on sediment physical properties, its possible testing also eliminated this possibility. The defor-
influence on the location of thrust faults and the mation had to involve tectonic stresses. More
basal decollement of the prism. Sample (1990) elaborate illustrations of applications of labora-
argued that carbonate cementation in the Upper tory testing to understanding sediment deforma-
Cretaceous sandstones of Kodiak, southern tion are given in sections 2.2.5, 6.2.2 and 6.3.5.
Alaska, played a major role in the dynamics of
the plate convergence. Sample envisaged a
'cementation front' moving through the evolving 1.3 CAUSES OF DEFORMATION
prism, which influenced sediment strength and
the styles of deformation, and which may even 1.3.1 Ice
have led to the large amount of underplating Historically, ice was one of the first direct agents
postulated for this plate junction. of sediment deformation to be documented, as a
result of noticing structures in deposits recog-
nized as glacial in origin (e.g. Sorby 1859). Exact-
1.2.6 Experimentation
ly how the ice produced deformation structures
Soil mechanics principles have been derived was debated for a time: Salisbury and Atwood
largely from laboratory testing. However, the (1897), for example, invoked the grounding of
vast majority of this kind of work is devoted to icebergs and ice rafts to explain examples of
routine evaluations for engineering construction intrastratal folds, whereas Johnston (1915, p. 43)
purposes, using the routine tests described in reported sand laminae that were 'minutely folded
handbooks such as Bowles (1978) and Vickers and overthrust by the overriding of the ice sheet'.
(1983). The specifications for executing standard The evolution of early views on glacial deforma-
tests are stated in publications such as ASTM tion is summarized by Aber, Croot and Fenton
(1991) and BSI (1991). The testing methods are, (1989). However, despite the impressive range of
however, under continuous review (Donaghe, ideas, the notion of ice bulldozing into sediments
Chaney and Silver 1988; Toolan 1988; Jewell and forcing various 'ice-thrust' structures be-
1989; Burland and Hight 1990; Atkinson, Lau came, for a long time, the dominant concept. The
and Powell 1991). Advances in instrumentation influential textbook of Flint (1971, p. 121), for
are expanding the role of field testing, which example, remarks that most deformation of gla-
offers less control but has the marked advantage cial sediment 'seems to have been made at or
of minimizing disturbance to the measured ma- near glacial termini'. In this view, the deforma-
terials (Demars and Nacci 1978; Graham, Crooks tion of sediment by ice is confined to some of the
and Bell 1983; Chaney and Demars 1985). Tech- peripheral areas of the glacier, and consequently
niques for permeability measurement are dis- the process received limited attention.
cussed by Aban and Znidaric (1989), Dunn and A major discovery of recent years is that some
Mitchell (1984) and Fernuik and Haug (1990). glaciers and parts of major ice sheets rest not on
The great value of laboratory testing in the bedrock, but on a layer of sediment (section 3.1).
present context is to provide quantitative limits This seems particularly true of the fast-flowing
for the interpretation and understanding of natu- zones of ice sheets known as ice streams, and it
ral processes. To give one example, Lucas and seems that shearing of the basal sediment layer is
Moore (1986) observed cataclastically deformed largely responsible for the high flow rates that
quartz grains in near-surface marine sediments have been observed (e.g. Alley 1991). This 'de-
off southern Mexico. Any possibility that the formable bed' hypothesis forms the basis of
grain breakage had arisen purely through burial Chaper 3. There, soil mechanics principles are
was eliminated on the basis that laboratory tests applied to an understanding of the subglacial
had shown the available burial stresses to be dynamics. The chapter also outlines the rheologi-
insufficient. More importantly, steepened bed- cal approach to quantifying the behaviour of
ding dips at the sample site raised the possibility these weak and highly strained subglacial sedi-
Causes of deformation 17

ments. The principles are remarkably parallel to till genesis (van der Meer 1987), and even to
those involved in understanding other environ- verify the glacial origin of deposits (Mahaney
ments of sediment deformation. For example, 1990). Soil mechanics methods have been applied
glaciologists are currently investigating sediment to understanding some of the structures, such as
consolidation (Boulton and Dobbie 1993), shear joints (Feeser 1988) and shear zones (Tsui,
strength and viscosity (Humphrey et al. 1993), Cruden and Thomson 1989). Sauer and Chris-
and hydrogeology (Murray and Dowdeswell tiansen (1988) measured the consolidation states
1992): all matters of topical interest at convergent of glaciolacustrine clays from central Sas-
plate margins and elsewhere (section 6.3). This katchewan in order to find likely burial loads and
new view of ice-sheet motion represents a major hence ice thicknesses at the time of glaciation,
change in glaciological thinking (Boulton 1986) and in the Netherlands the likely glacier dyna-
and the way in which structures in glacial depos- mics were deduced from geotechnical measure-
its are interpreted. Although caution has been ments by Schokking (1990).
advised against applying the deforming substrate It follows from the interaction of the kinds of
hypothesis too widely (Robin 1986; Clayton, glaciological processes indicated above that
Mickelsen and Attig 1989), it appears to be many structures are not the result of a single
gaining support, prompting a new interest in how agent. For example, Owen (1989) reported struc-
sediments deform in the glacial environment. tures in which both glacial and neotectonic defor-
In addition to the shearing effects expected at mations have been involved, and in the active
the base of a moving ice sheet, sediment deforma- Kleszcz6w Graben of central Poland, Brod-
tion comes about in response to ice loading zikowski, Krzyszkowski and van Loon (1987)
(Drozdowski 1982) and a host of glacial and have documented structures due to tectonic, gla-
periglacial processes, such as iceberg drop and cial and sedimentological interaction. Pederson
grounding (G.S.P. Thomas and Connell 1985), (1987) described structures resulting from both
repeated freeze-thaw cycles (Coutard and glacial ice and gravitational mass-flow, and the
Mucher 1985) and deglaciation (Vesajoki 1982). glacially generated gouge zones described by
The processes are complex and closely interre- Stauffer, Gendzwill and Sauer (1990) governed
lated. both further glacial deformation and post-glacial
Not surprisingly, therefore, the structures re- landslides. Brodzikowski and van Loon (1991)
sulting from glacial deformation are extremely reviewed effects of the interplay between sedi-
varied (e.g. Fischbein 1987). Their similarity with mentological and glacial processes.
structures formed in completely different environ- A number of glacial landforms and related
ments has long been noted. Salter (1866, p. 566), features are now thought to involve various kinds
for example, compared structures in the glacial of sediment deformation. Although geomor-
drift of southern England with 'those produced in phological phenomena are outside the scope of
our mountain ranges'. More recent examples this book, examples are mentioned in section
include comparisons of glacially induced struc- 3.8.4 because their genesis is related so closely to
tures with thin-skinned thrust belts (Croot 1987) the structures preserved within them. Recent col-
and with accretionary prisms (Aber, Croot and lections of works on landforms, their internal
Fenton 1989). Aber (1988) compared the origin of deformation structures and related features in-
some ice-related hills in Saskatchewan to the clude Croot (1988), van der Meer (1987) and
mudlumps of the Mississippi Delta, which are Goldthwait and Matsch (1989). Some further
themselves analogous to the diapirs associated examples of deformation structures from the gla-
with accretionary prisms. At the microscopic cial environment are mentioned in Chapter 9.
scale, authors such as Hart and Boulton (1991)
have compared highly sheared till fabrics to
mylonites, although in view of the crystal-plastic 1.3.2 Disturbance in place
processes that dominate in rocks, the grain-scale It is still being learned just how gravitationally
mechanisms must differ. Fabric studies have long unstable many near-surface sediments are - on
been an important aspect of glacial deposits. This the sea floor, in lakes and on land. Material lying
approach has helped interpret the origin of struc- on a slope has long been known to be vulnerable
tures (Menzies and Maltman 1992), distinguish to downslope motion, and the deformational
18 Introduction and overview

implications of these mass movements are dealt example, reported a 'singularly wrinkled struc-
with in the next section and in Chapter 5. How- ture' that occurred within otherwise uniform
ever, sediment instability without substantiallat- Devonian limestones at Gaspe, Quebec (Figure
eral transport is also now widely recognized, 1.5). The vexing nature of such 'intrastratal'
particularly in the near-shore setting, where it has disturbances is illustrated by the fact that the
immense repercussions for marine geotechno- precise origin of some examples would still be
logy. Such in-place disturbance gives rise to a debated today, although, hopefully, understand-
variety of deformation structures, and these are ing of the possible formative processes is now
described in section 4.3.1. The separation be- more rigorous. In earlier times, some of the
tween these and the structures dealt with in statements on inferred processes were merely
Chapter 5 is arbitrary: many natural structures speculative and intuitive, criteria were vague, and
involve a small amount of lateral motion, and are even the observations lacked the necessary preci-
part-way between in-place and true mass move- sion. Some of the discussions, however, were
ments. In the geological record, their precise more informed and perceptive. In the interests of
origin might well be unclear. learning from such past experience, Historical
The kinds of geological structures that would vignette 2 summarizes the exchanges of views on
now be ascribed to these processes have been a particular example of these structures: those
known for some time. William Logan (1863), for illustrated in the frontispiece to this book.

Figure 1.5 Sir William Logan's illustration of localized deformed beds at Gaspe, Quebec. (From Logan 1863.)

Historical vignette 2: Intrastratal disturbances at Trenton, New York

Lardner Vanuxem, at various times a farmer in Black River and for these latter structures he
Pennsylvania, a university professor in South provided more accurate illustrations - beautifully
Carolina, and a gold-miner in Mexico - and during engraved wood-cuts, two of which form the frontis-
his scientific training in Paris an associate of such piece herein. 'The layers show a series of con-
luminaries as Brogniart and Hauy - was for 4 years tortions of different kinds', Vanuxem declared, 'for
part of the official geological survey of New York. which no cause can reasonably be assigned but
His 1842 report which concluded this last enterprise different degrees of lateral pressure' (1842, p. 213).
mentioned some 'extraordinary contortions' in the He was unsure of the origin of the pressure, and
Ordovician limestones near Trenton Falls, in vaguely invoked either some differential crystalliza-
Oneida County, central New York State. He supplied tion between the layers or some accommodation
a rather stylized representation, reproduced here response to 'lateral pressure ... forcing whatever
as Figure 1.6a. He also reported a series of similar soft or yielding material was subjected to its
structures in Quaternary deposits along the nearby power' (p. 216). Vanuxem emphasized the contrast
Causes of deformation 19



Figure 1.6 (a) Lardner Vanuxem's stylized representation of the intrastratal contortions at Trenton, New York. (From
Vanuxem 1842.) (b) Reproduction of William Miller's depiction of the intrastratal disturbances at Trenton, New York.
(As shown in Miller, 1908, 1909 and 1922.)

between the disturbed and the normal horizons, but from some cause, very shortly after the period of
did not specifically mention the critical observation deposition' .
of the truncated upper and lower contacts of the About 50 years after Vanuxem's project, W.J.
disturbed horizon, although they are recorded in Miller was engaged in a survey of the same area,
the illustrations. and was also struck by the 'highly twisted or
The same structures caught the attention of contorted' structures at Trenton and the similar
Theodore White during his palaeontological work phenomena along the Black River (Miller 1908, p.
in the area for his doctoral thesis at Columbia 429). However, the standard of precision in illustra-
University. He thought Vanuxem's explanation to ting the structures was noticeably poorer (Figure
be inadequate, as well as a suggestion he had 1.6b), and Miller came to a different interpretation.
heard from W.O. Crosby, a professor at the Massa- He found no difference in the crystalline state of
chusetts Institute of Technology (MIT), that the the different limestones, as implied by Vanuxem,
cause may have been the yielding of this layer and proposed that· 'it is thought that the folded
under the tremendous weight of strata above' structure ... was in reality caused by a differential
(White 1895, p. 90). He cited observations such movement within the mass of the Trenton lime-
as the undeformed nature of the fossils and the stone', as a result of accommodation to nearby
insufficient degree of metamorphism as in- thrusting. Some layers were able to 'move more
compatible with the hypotheses offered, though easily and consequently faster' (p. 431), and the
he seemed unable to specify any alternative 'folded zones thus merely indicate horizons of
explanation. He merely commented that 'he was weakness along which the differential movement
inclined to think that the contortion took place, took place'. The idea, of course, is perfectly feasible
20 Introduction and overview

in general, but Miller made no attempt to assess the Grabau, fluent in German, was also conversant
truncations of the Trenton structures, or, especially with Heim's writings. Oddly enough, he had been
in view of the uniform 'crystallinity' he had empha- a student of Crosby at MIT, but he perceived the
sized, why the various limestone layers should significance of Heim's observations and espoused
have been so mechanically contrasting. Miller also the importance of subaqueous sliding. 'As the
noted that the 'locally folded clay layers between result of such gliding the strata must suffer much
non-folded layers along the Black River are to be deformation' (1913, p. 660), he wrote, and went
explained in a somewhat similar manner', although on, 'such deformations have all the characters of
here 'the movement of the upper over the lower orogenic disturbances due to lateral pressure,
masses may have been caused by ice action or by and indeed it has been suggested that some ex-
having been pulled down the hillsides by gravity' (p. tensive mountain folds and overthrusts may have
432). originated in this manner'. Grabau reviewed the
Later, Miller (1909) repeated his descriptions, Trenton structures, mentioning one explanation
and reproduced again the drawing in Figure 1.6b. In that was 'tectonic - lateral pressure having resulted
considering the origin of the structures, he recalled in the folding of certain strata while others took up
Logan's (1863) remarks on the similar features of the thrust without deformation', and another 'that
Gaspe (Figure 1.5): 'it would appear as if the layers they were due to squeezing out of certain layers
after their deposit had been contorted by lateral under the weight of the overlying rock masses'.
pressure, the underlying stratum remaining undis- Grabau concluded that in his view 'both are un-
turbed, and had been worn smooth before the supported by the detailed characteristics of the
deposition of the next bed. Where the inverted folds and their relationships to the enclosing layers'
arches of the flexures occur, some of the lower (p. 784) and argued in support of the subaqueous
layers are occasionally wanting, as if the cor- sliding hypothesis.
rugated bed had been worn on the underside as Miller (1915) could not accept this explanation.
well as the upper side' (Logan 1863, p. 392). Miller Moreover, he was aggrieved that Grabau 'neither
commented that 'it is difficult to see how such a states nor presents arguments against the present
lateral pressure could cause certain layers to be- writer's hypothesis' (p. 135). Apparently he felt it
come highly folded and broken while the layers was not represented by Grabau's statement on
immediately below them are undisturbed' (Miller tectonic causes. Miller's list of evidence in support
1909, p. 32). Logan had put his finger on the crucial of his hypothesis is rather subjective, and provides
matter of the truncated boundaries, but although two general illustrations of the way in which geo-
Miller had reproduced the remarks, the significance logists work. First, previous experience tends to
eluded him. 'The rubbed or worn character' was colour perception. Miller's geological experience
simply one repercussion of the entire sequence and perspective contrast with those of his oppo-
being differentially moved by faulting. nents. Having been born near the Lassen Peak
In contrast with these intuitive speculations, volcano, in California, and having worked mainly on
the next contribution was based on the results of igneous and metamorphic phenomena, he ap-
observed processes. Albert Heim in 1908 had proached the structures from a 'hard rock' view-
produced very detailed descriptions, in German, of point and may have been predisposed towards
two disastrous landslides that had entered Lake these kinds of processes. At this stage at least,
Zurich, and from which falling lake levels had given Miller seems to have had little feel for the sediment
the opportunity for the resulting structures to be sliding process. Ironically, he was later to work on
directly observed. Fritz Hahn had read the accounts some California landslides! Second, intuition alone
while working in Stuttgart, but had subsequently can mislead. To Miller, the deformed horizons
become temporary Curator at New York's Columbia appeared too thin to be due to slumping, given their
University, from where he happened to visit the areal extent, and he felt that 'the slope of the sea
Trenton structures. He immediately perceived their bottom ... was altogether too slight, a fact 'well-nigh
likely subaqueous pre-lithification origin, and wrote fatal to the hypothesis of submarine slumping' (p.
a detailed argument in support of this interpreta- 143). Regarding his feel for the rates of lithification
tion. Amadeus Grabau, a giant of the time in east- of limestone he felt that 'the corrugations could not
ern North American geology, took up Hahn's theme have been produced before deposition of the over-
in his monumental 1200-page textbook (1913). lying masses because the limestone layers at least
Causes of deformation 21

were comparatively hard and brittle when con- sand' (p. 382). Miller, however, remained uncon-
torted, as shown by the numerous sharp breaks'. vinced, and in 1922 reaffirmed that 'in the opinion of
He concluded that 'it would seem that Hahn has the writer differential movement, not always as an
fallen into a common error of making a single accompaniment to thrust faulting, is involved in
hypothesis or explanation altogether too inclusive many, if not most, cases of intraformational dis-
in its scope'. Hahn met a premature death, in 1914 turbance'. He also published for a third time the
at the age of 29, and never had the opportunity to drawing reproduced here as Figure 1.6b, and
reply. repeated his complaint that Grabau had discussed
The differential weighting idea continued to the Trenton structures 'without even mentioning the
receive some favour in interpreting structures like tectonic differential slipping hypothesis' (p. 590).
those at Trenton, and Kindle (1917) provided some Miller then reviewed intraformational structures
experimental justification. He pointed out that more generally, together with a range of possible
although 'various explanations have been offered, origins, including igneous processes. He went on to
the subject appears to be still open to experimental a long and illustrious career, becoming, for
investigation'. From the results of irregularly piling example, the first head of the new geology de-
different-density wet sediments in a tank, he argued partment at the University of California at Los
that 'disturbances of the bedding in strata of sand Angeles, but there is no evidence that he ever
may originate in the same general way through the became converted to the importance of subaqueous
juxtaposition of beds of quicksand and ordinary movements.

Near-surface sediments are at their most rapid sediment give rise to static loading. Addition of
stage of consolidation, and if the pore fluid is an extra external stress is called dynamic loading;
unable to escape adequately they are particularly individual loading events are described as mono-
vulnerable to two effects. First, layers of differing tonic. They can arise in nature for a host of
permeability will undergo different rates of po- reasons, including: an increment of rapid
rosity decrease, so that at anyone time denser sedimentation (G. Owen 1987); emplacement
sediments may be overlying less dense materials. above of a moving sand dune (B.G. Jones 1972),
Primary deposition can sometimes produce the slump sheet (Eva 1992) or igneous mass (Einsele
same arrangement. Because gravitational poten- 1982); even the footstep of a mammal (Lewis and
tial results from height above a datum times Titheridge 1978) or shock (supersonic) loading
mass, the overlying layer has the higher potential due to meteorite impact (Read 1988).
energy and will attempt to mobilize to produce Cyclic loading comes about most commonly
the more stable, greater density with depth, through the oscillatory transmission of seismic
configuration. Should the shear strength of any waves (Seed and Idriss 1982) and, in subaqueous
of the layers be approached, through further sediments buried (in typical water conditions) no
pore-pressure increments, which lower the effec- more than about 200 m, from the passing of
tive stress, or the addition of some extra load, the travelling pressure waves at the sea surface (Seed
potential mobility will be activated. The kinds of and Rahman 1978). Each stress rise in these
deformation structures described in section 4.3.1 repetitive processes may well be much less than
will be generated. The best analysis of these kinds in a monotonic event, but if each fluid pressure
of processes remains the work of Anketell, Cegla response is incompletely dissipated before the
and Dzulinski (1970) and Ramberg (1981). next pulse, the incremental accumulation of over-
Second, lowering of the effective stress is facili- pressure can be large. The idea is shown sche-
tated in these loose, near-surface materials, be- matically in Figure 1.7. A variant on this behav-
cause they are likely to deform initially by con- iour, called cyclic mobility, can induce strength
tracting, thus temporarily increasing the fluid loss even in relatively dense sediments. It arises
pressure and promoting further weakening. In where the periodic loads oscillate in nature be-
addition, pulses of greater pore pressure may tween compression and tension and create transi-
arise from mineral dehydration, aquathermal ef- ent pulses of overpressure (Castro 1975, 1987).
fects, etc. at depth, as discussed in sections 2.3.4 These loading phenomena are discussed further,
and 7.3.6. Such processes that are internal to the in the geological context, by Decker (1990).
22 Introduction and overview

Number of cycles ... loading is abrupt and rapid. This kind of behav-
iour is commonly included with liquefaction al-

: ~ IIIIHIII!!! IIIIWJ II! l!Iu lUll ~

Applied though, strictly, overpressuring is not involved,
shear and irrespective of the loading conditions thixot-
ropic sediment retrieves its strength when it re-
(as % burial
pressure) 4 turns to rest and the particles reflocculate.
8 I Liquefied sediments that find themselves in a
reversed density gradient will be highly unstable.
Pore Any perturbations at the interface between the
overlying, denser material and the underlying less
Burial dense sediment will, seeing as they are behaving
as fluids, act as Raleigh-Taylor instabilities, caus-
ing the irregularities to amplify until gravity
driven overturn can be achieved (Ronnlund
1989). There are similarities with movements in
salt (Jackson and Talbot 1994).
Hird and Hassona (1990) showed experimen-
(%) tally that, in principle, any sediment can liquefy,
but that rounded grains promote the effect and
Figure ·1.7 Schematic diagram of cyclic loading leading
that the presence of platy minerals, because of
to liquefaction. The horizontal axis represents the their cohesion and compressibility, retard it.
number of loading cycles with time. The fluid pressure Bornhold and Prior (1989) discussed the role of
accumulates incrementally until it equals the burial load, gas hydrates in promoting liquefaction and asso-
whereupon the sediment liquefies and the shear strain ciated subsidence. Vaid et ai. (1990) argued that
response is instantaneous. (Based on Decker (1990). Used
by permission of P.L. Decker.)
because particle grading affected the ability of a
sediment to show contractive deformation it
The rapidity of these loading processes can therefore influenced the tendency to liquefy. For
easily outstrip the ability of the sediment to example, in their experiments, high-porosity,
dissipate the nearly instantaneously pressured poorly graded sand resisted cyclic loading better
pore fluids, such that their pressure may rise to than a well-graded equivalent, whereas lower
equal the burial pressure, hence reducing the porosity material showed the opposite effect.
effective stress to zero. In the absence of signifi- In-place disturbance can come about for a
cant cohesion, the sediment now lacks shear wide variety of reasons, but most involve pore-
strength and temporarily exists as a fluid. This is fluid effects in one way or another. The funda-
the phenomenon of liquefaction. The sediment is mental mechanical role of pore water is well
available to generate the kinds of injection struc- illustrated by the commonness of deformation
tures discussed in sections 4.3.4 and 9.5.2. It will structures in wet sand dunes (Horowitz 1982) as
remain in this condition until the pore pressure is opposed to their scarcity in desert sands (McKee
reduced below the burial load. Cakmak and and Bigarella 1972). The other major role of pore
Heron (1989) illustrate the engineering approach fluid is, of course, in promoting diagenesis, and
to dealing with this crucial phenomenon. For this is discussed in sections 1.2.5 and 4.3.9.
some geological purposes (e.g. sections 4.2 and Another type of in-place deformation arises
5.1.2), it is necessary to distinguish between lique- through differential consolidation, such as in
faction, in which the particles are able to settle, producing the faults that are increasingly being
and fluidization, in which the upward flow of the reported from the modern sea floor (e.g. Buckley
fluid is sufficient to maintain the particles in and Grant 1985). Unusual variations on this
suspension. An effect analogous to liquefaction theme come about through the removal of ma-
comes about in the delicate but cohesive sedi- terial in some way, and the consequent founder-
ments that show thixotropy. Here, constant-rate ing of the beds above. Upright folds in the
loading destabilizes the sediment framework and Eocene sediments of the South Carolina coastal
disperses the particles in the pore fluid so that plain were claimed by Johnson and Heron (1965)
strength is lost. The effect is marked where the to be due to the 'uneven let down' of the beds
Causes of deformation 23

into the material below, from which carbonate, mirrors the exploration of the ocean floor. Benest
or possibly opaline silica, was being preferentially (1899, p. 394) noted the frequent breakages in the
dissolved by groundwater. Wardlaw (1972) de- submarine cables that were beginning to be laid
scribed deformation in an evaporite sequence on the floor of the ocean and inferred 'un-
that was prompted by the preferential dissolution suspected forces constantly in action and altering
of interlayered halite, anhydrite and carnalite the features of the sea bottom'. Continuing
beds; pure halite remained unaffected. Mohl and exploration of sea-floor bathymetry revealed all
Bakken (1968) illustrated what they call slump kinds of new features, and three decades later
structures that are due to collapse into underly- Shephard (1932, p. 226) was describing sub-
ing, burned coal-beds, thought to have been marine canyons and explaining them as the
ignited by lightning strikes! result of 'landslides caused largely by earth-
quakes accompanying faulting-action along the
continental slopes'. Modern oceanographic work
1.3.3 Gravitational mass movement has detected the gravitational movement of
The importance to society of slope movements sediments on an enormous scale (e.g. Bugge,
on hillsides is obvious (e.g. Brabb and Harrod Belderson and Kenyon 1988), and in all kinds
1989), and the effects have long been witnessed. of materials (e.g. J.G. Moore et al. 1989). The
Even geological reports of the phenomena go Agulhas slump off southeast Africa, for example,
back at least as far as Silliman's (1829) des- occupies an area of 80000 km 2 - considerably
criptions from the White Mountains of New more than, say, the area of England and Wales or
England. Today, similar gravity driven processes the state of Oklahoma (Dingle 1977). The
are being increasingly discovered on the ocean Neogene slides in the Gulf Coast basin described
floors, and they are even observed on other by Morton (1993) reach 600 m in thickness, but
planets (Shaller et al. 1989). The attendant ter- a striking aspect of many modern submarine
minology has had an intricate evolution, via slides is their thinness relative to the areal
'landslip' - still used for some English examples - extent. The slide northwest of the Canary Islands,
through to 'landslide'. At its simplest, the latter is for example, covers around 40000 km 2 of the
merely 'the movement of rock, earth, or debris Madeira abyssal plain, yet is mostly less than
down a slope' (Cruden 1991). In this usage there 20 m in thickness (Masson et al. 1992). Section
is no restriction on the mode of movement or on 5.5.2 draws attention to the apparent discrepancy
the setting - landslides can flow and they can be between the magnitude of these modern features
submarine. As Cruden (1991, p. 28) put it: 'like with those so far recognized in the geological
cowboy, landslide is another North American record.
word formed by two words which together mean Following the demise long ago of inferr-
something entirely different'. Terminology for the ing aqueous processes as agents of deformation
components of a landslide, from an engineering (section 1.1.1), the modern realization that some
viewpoint, is given by the IAEG (1990). Geo- major deformation features in rocks may after all
logists prefer the term mass movement, perhaps be the result of subaqueous instability was
because it is more obviously all-encompassing. prompted largely by reports on landslides into
The great variety of processes by which earth Swiss lakes (section 1.3.2). Shortly after, T.e.
masses move downslope in subaerial and sub- Brown (1913, p. 224) interpreted disoriented
aqueous environments is reviewed in Chapter 5. and folded Lower Palaeozoic conglomerates in
Most attention is paid there to the processes that Pennsylvania as having been 'slumped or slid
induce the deformation of existing sediments along the (sea) bottom due to gravity'. Two
rather than to those that give rise to fresh deposi- decades later, the importance of these long-
tion, and are hence essentially sedimentological discarded ideas was being restored. Archanguel-
in nature. sky (1930) emphasized the importance of mass
To many geologists, it is the submarine movements in the Black Sea, and illustrated
phenomena that are of greatest interest, because sequences of repeated stratigraphy and folds in
of their widespread occurrence, vast scale and rocks that he interpreted as the result of sub-
relatively recent discovery. In fact, progress in aqueous processes. Hadding (1931) remarked on
recognizing submarine mass movement closely the commonness of the results of subaqueous
24 Introduction and overview

sliding, and presented both a workable classifica-

tion scheme and a perceptive analysis of various
possible causes of instability. De Terra (1931)
described folds, shearing planes and faults in Factor of
rocks from Wyoming and the Himalayas that he safety
IF) 2
thought were due to 'sub-aquatic gliding'. He
also made the premonitious remark that 'the fact
that gliding can set free a tangential pressure
which causes structural forms similar to those
which are usually considered to result from
contraction of the crust seems to give us a useful .85 .90 .95 1.00
hint for any analysis of related features' (de Terra Pore-water pressure

1931, p. 213). Geology, however, has been slow to Burial pressure

return to recognizing the importance of large- Figure 1.8 Relations between factor of safety (F), slope
scale subaqueous deformation. The conceptual inclination (9) and overpressuring, as deduced by Prior et
al. (1981) for the Mississippi Delta. The fluid pressure
difficulties and the problem of distinguishing was calculated on the basis of
between the results of large-scale mass movement
and post-lithification tectonic deformation are c'-F(pzsin9cos9) 2
encapsulated in a famous controversy between PH,o + pz cos 9,
tan <\>'
two leading British geologists, which is recalled where c and <\>' are the likely effective Mohr-Coulomb
in section 9.6.1. parameters (see equation 1.13), and p and z are the density
The mechanics of downslope movement have and thickness, respectively, of the sediment. A factor of
received much analysis, particularly through the safety < 1 (stippled area) implies conditions of slope
importance of hazard prediction (Zaruba 1987), failure. A pore fluid pressure/burial pressure ratio of 1
implies liquefaction. (Reproduced with permission from the
and such engineering studies have provided a American Association of Petroleum Geologists.)
quantitative basis for understanding the geo-
logical processes. Historical reviews of the arc (Espinoza, Repetto and Muhunthan 1992).
engineering approach to the stability of on-land Analyses of submarine slopes are most commonly
slopes are given by Graham (1984) and Fleming approached by an 'infinite slope analysis', in
and Varnes (1991). Interestingly enough, Graham's which the failure plane is treated as planar and
discussion includes a further illustration of the parallel to the free surface, and a small volume is
're-inventing the wheel' point made in the present taken to adequately represent the total sliding
Preface, as certain fundamental contributions mass. Such a treatment is in some ways sim-
concerning curved failure surfaces on clay slopes plistic, but it provides a reasonable approxima-
were virtually forgotten for almost 80 years, tion, at least with our present knowledge of
before their resuscitation to become a keystone of sea-floor properties. An example of this approach
some modern slope analysis techniques. is incorporated in Figure 9.15.
Most analyses of slope stability are based on In the simplest view of the stresses acting on a
Mohr-Coulomb principles; the most widely used slope, if the sediment has an area A and a weight
are outlined by Kenney (1984). They assume that W (area x thickness x bulk density x gravitational
failure comes about when the shear stress acting acceleration), and slopes from the horizontal at
along the slope exceeds sediment strength, and an angle {J, then the shear stress operating along
that the resulting movement takes place on a the slope will be:
discrete plane of failure. A common specifica-
tion of slope stability uses a safety factor F - the W .
'=-:4 81TI
e, (1.15)
ratio of available shear strength to the likely
shear stresses (e.g. Graham 1984). Slopes are and the stress normal to the slope will be:
unstable where F falls below unity (see, for
example, Figure 1.8). A variety of analytical an =-:4 cos e. (1.16)
models are used for subaerial slopes, for example,
according to whether the geometry of the failure Therefore, assuming that a potential detach-
is analogous to a sliding block or some circular ment parallels the planar slope, the overlying
Causes of deformation 25

load acts both to strengthen the sediment by Bouma 1982), fault reactivation (Holler 1986),
increasing the stress normal to the failure plane salt movement (Almagor 1984), and because of
and to promote failure along the plane by in- erosion at the slope foot (Schwab et al. 1988).
creasing the shear stress. Although for slopes of Strength parameters will be evolving through
low and moderate inclination the strength (in- progressive consolidation and diagenesis (Charles
creasing through the cosine of the slope angle) 1982). Various effects owing to sea-level fluc-
will grow more rapidly with increasing burial tuations have also been invoked to explain sub-
thickness than the shear stress (increasing marine slope instabilities (Morton 1993).
through the sine of the slope angle), with steeper However, it is likely that most slope failure
slopes (in which sine values increase more rapidly comes about through overpressuring. Essentially,
than cosines) the opposite will be the case. this reduces the sediment strength, as outlined in
Hence, sediments on slopes above a certain in- section 1.2.5, and hence the resistance to failure.
clination will become increasingly unstable with Overpressuring in subaerial slopes commonly
progressive burial. comes about through artesian conditions, or after
Ross (1971) measured shear strengths of heavy rainfall (Reid and Baum 1992), which
sediments and slope angles around the Middle temporarily increases the elevation head of
America Trench. In line with the above relation- the pore fluid (Hutchinson 1989; van Genuchten
ships, the data suggested that whereas the shal- and de Rijke 1989; Reid, Nielsen and Dreiss
low slopes of the adjacent shelf bore sediments 1988). Vegetation may sometimes play a role
stable during burial - because the extrapolated (Terwilliger 1990). In a subaerial mudslide in
shear-strength increase with burial continued to Dorset, southern England, Allison and Brunsden
exceed the calculated shear-stress increase - on (1990) documented the rainfall-fluid-pressure
the steeper slopes of the trench flanks, shear relation and its influence on permeability. This
stresses exceeded the strength at a depth of a few in turn determined whether the style of slid-
tens of metres. That failure is in fact occurring at ing was dominated by what they called stick-
such depths is supported by the seismic profiles slip, graded-slip, surge or 'random' modes of
across the trench flanks, whereas there are no motion.
indications of failure on the sections across the In the subaqueous environment, biogenic gas
shelf. (Esrig and Kirby 1978), gas-hydrate melting
For many near-surface sediments, the contri- (Kayen and Lee 1990), and storm waves (Lu et al.
bution of cohesion to their strength is negligible. 1991) may be important, but many mass move-
In this case, incorporating (1.15) and (1.16) into ments are surmised to have been caused by
the basic Mohr-Coulomb relation (1.5) gives: earthquakes (Poulos 1988). Propagation of seis-
mic waves adds to the bulk shearing stress acting
: sin () = : cos () tan ¢. (1.17) on the sediment but a more crucial effect is the
instantaneous transmission of the stress to the
At the failure condition of shear stress equal- pore fluid. This creates spontaneous overpressur-
ling strength, the load terms cancel out, so that: ing and perhaps liquefaction, as explained in the
previous section. The oscillating seismic waves
may produce an incremental accumulation of
that is, the slope cannot exist at an inclination fluid pressure or cyclic loading (setion 1.3.2) and
(sometimes called the angle of repose) greater further promote slope instability (Jibson and
than the friction angle of the sediment. As dis- Keefer 1993). A further cause commonly invoked
cussed below, wet sediment may be unstable at to explain overpressuring is a rate of sedimen-
considerably lower inclinations. tation that outstrips the ability of the sediments
In these ways the potential for failure depends to dissipate the pore water adequately, as out-
on the interplay between slope angles and the lined in section 1.2.5. As an illustration, Aksu and
sediment strength. In a real geological situation, Hiscott (1989) pointed out that large-scale sliding
both the slope angles and the mechanical pro- is more significant at higher latitudes than near
perties are likely to be continuously evolving the Equator, because of the higher rates of sedi-
(Baraza et al. 1990). Slope inclinations can ment accumulation adjacent to continental ice
change in response to diapirism (Martin and sheets.
26 Introduction and overview

Prior et al. (1981) illustrated the importance of Although slope failure in general comes about
overpressures in the Mississippi Delta. Calcula- when the sediment exceeds its peak strength
tions on typical properties showed that an aver- (section 1.2.4), Skempton (1985) has argued that
age slope would reach failure if fluid pressure many unstable slopes will have already under-
ratios exceeded 0.85, and that the delta-front gone some failure and mass movement, and
slope of OS would fail at a ratio of 0.97 (Figure therefore the sediment will be exercising its resid-
1.8). Because these fluid-pressure values agree ual strength. This is very likely the case for the
directly with field piezometric measurements, the intermittent, far-travelled movements common in
authors argued that much of the delta must be in the oceans. However, verification in the latter
an inherent state of failure, due to the overpres- situation is hampered by the difficulty of acquir-
suring. They went on to argue that the strain ing reliable strength measurements.
softening associated with initial failure would An alternative approach to analysing slope
lead to in-place collapse, followed by upslope stability is to treat the sediment not in terms of
retrogression of the subsiding area. Their concep- Mohr-Coulomb failure but as a viscous material
tion is shown diagrammatically in Figure 1.9. (section 1.2.4). Such treatments view the mass
They further argued that conditions for subsi- movement either as fluid-like flow or as taking
dence would readily be met in the Mississippi place on a fluid-like substrate (Ohlmacher and
Delta, thus explaining the common observations Baskerville 1991). In analogy with other contexts,
of collapse depressions in the region. this approach does involve the problem of choos-
ing the most appropriate flow model to use.
Retrogression of scarp Sousa and Voight (1992), for example, modelled a
landslide using two different viscosities, to repre-
sent the mass deforming below and beyond its
yield point. Some treatments do not explicitly
take into account the crucial effects of fluid
pressure. Iverson (1985) took a non-linear viscous
flow model and included a Mohr-Coulomb term
Figure 1.9 Schematic representation of the upslope
retrogression of failure conditions and the block
in addition, together with some allowance for
subsidence that arises where the minimum subsidence effective rather than total stress. He suggested
ratio, that the required data on parameters such as
strain rate, fluid pressures and viscosity could be
llh = 1_ (1 _ 4Cu )~ estimated adequately from field observations.
hI UWh,
However, other workers have argued that even if
is exceeded. Cu is the average undrained strength of the such a flow model is reasonably realistic, knowl-
sediment and UW its bulk unit weight. In the Mississippi
Delta, common values of hare 0.3-3.0 m and of h, are
edge of these properties in nature is still too poor
7-10 m, so that hlh, is in the range 0.05-0.03 m. Using a to justify a viscosity based approach (e.g.
bulk density of 1400-1500 kg m- 3 and Cu values of Hutchinson 1986).
4.8-9.6 kPa for burial depths of 10 m, the critical subsidence The same shortcoming prompted Einsele
ratios are 0.07-0.15, suggesting that slope failure in the (1989) to use Atterberg limits (section 1.2.2),
Mississippi may enlarge by a retrogression mechanism.
(From Prior et al. (1981). Reproduced with permission from measured from retrieved cores, to assess stability
the American Association of Petroleum Geologists.) on submarine slopes. His data indicated that
even normally consolidated marine sediments
Another cause of slope failure is seepage at commonly have sensitivities between 4 and 20, or
sites of groundwater outflow, which causes ero- more, implying a high liquefaction potential even
sion (Robb 1990) and chemical dissolution (Paull at hundreds of metres below the sea floor.
& Neumann 1987), leading to undercutting and Hutchinson (1986) combined this approach with
eventual slope failure. The process is commonly Mohr-Coulomb principles, suggesting that the
referred to as sapping (Dunne 1990). It may masses termed slides are likely to form in low
be responsible, for example, for the headless sensitivity clays or friction-dominated clastic
submarine canyons that appear not to have ex- sediments. The failed material would still be
perienced current erosion (Orange and Breen around its peak strength for localized, rotational
1992). slides but at the residual state for more far-
Causes of deformation 27

travelled blocks. On the other hand, material tion. Examples of these are presented in Chapter 5.
forming slumps would exhibit lower strengths, Further instances of the kinds of structures pro-
and commonly comprise higher sensitivity, per- duced by mass movement as they are preserved in
haps normally consolidated material. (Terms for rocks are illustrated in Figures 9.17 and 9.37.
the morphology of mass movement are explained
in Chapter 5.) Flows showing plastic behaviour
must have operated at a water content within the
1.3.4 Fluid-sediment movements
plastic range of the sediment, whereas those with The escape of pore fluids in response to the
fluid behaviour are probably of high-sensitivity loading of sediments can itself produce deforma-
sediment which must have exceeded its liquid tion structures, especially if the sediment is in a
limit. Material with a relatively high plasticity liquefied state. Some principles of fluid behaviour
index will tend to give thicker flows with a lesser in sediments were sketched out in sections 1.2.5
run-out distance. and 1.3.2. The theoretical background to fluid
In addition to the relatively sudden slope failure movement and examples of how this interacts
implied in all the above models, it may be that with deformation are presented in Chapter 7.
much deformation associated with mass move- Many of the principles outlined there are of
ment is due to downslope creep (section 1.2.4). The relevance to any deforming wet sediment, but, as
process is little recognized so far in submarine the discussions in Chapter 7 illustrate, the drain-
deposits because the appearance of its results are age-deformation interplay is particularly impor-
undocumented in sediments or rocks, and they are tant in those materials that are sufficiently
probably too subtle to appear on most seismic bonded to allow some degree of brittle fracture.
records. The experiments of Silva and Booth Section 4.3.4 describes some of the structures that
(1985) suggested that creep is a significant process result from near-surface deformation.
off the northeastern USA coast. Annual rates of Historically, a theoretical analysis of how
movement are several centimetres a year or more, fluids move through sediments had to wait for
with slopes of over 10° undergoing metres of Henry Darcy's classical work on permeability in
movement, almost certainly leading to eventual 1856, after which the early efforts were
failure. Moreover, the work revealed a degrada- dominated by groundwater (King 1899) and,
tion of the sediment strength brought about by the later, petroleum geology (e.g. Fraser 1935). How-
creep, which probably further engenders the likeli- ever, observations concerning natural structures
hood of ultimate failure. This notion received greatly preceded all this - in fact reports of
support from field measurements carried out on sedimentary dykes form some of the earliest rec-
Ruanda hillsides by Moeyersons (1989). In addi- ords of structures in sediments. In 1821,
tion, he pointed out differences in orientation of W.T.H.F. Strangways published descriptions
friction-influenced surfaces of failure and the slip of structures along the River Puicova, near
lines that reflected shearing while the mass was St Petersburg, and drew attention to two dykes,
behaving plastically. Local adjustments to these labelled x and y on the map inset reproduced
differences were taken by Moeyersons to account here as Figure 1.10. These were 'infilled with
for certain small-scale structures seen in the diluvian gravel' (Strangways 1821, p. 386).
landslide such as en echelon microfissures and Charles Darwin noted sedimentary dykes in
'microdistortions'. If such features are preserved in Patagonia while on his voyage on the Beagle in
ancient sediments and rocks, they appear not to 1833. He described four examples in some detail
have been recognized yet. and remarked of one of them: 'the structure of
Some varieties of mass movement, such as falls this dyke shows obviously that it is of mechanical
and flows, appear to produce little in the way of and sedimentary origin', and, regarding its gen-
recognizable internal structure, although there esis: 'if we reflect on the suction which would
may be some alignment of clasts and features in result from a deep-seated fissure being formed,
any cohesive clay that is present (section 9.4.3). we may admit that if the fissure were in any part
Those masses of greater strength, however, and slightly open to the surface, mud and water might
especially slumps and slides, can develop a wide be drawn into it' (Darwin 1846, p. 150).
range of striking structures, some remarkably Early workers had difficulty coming to terms
analogous with other environments of deforma- with the igneous-like form of these structures and
28 Introduction and overview

I r ---------- .---------
: : I

';::========~I i···········::::::.::::::::::::::::::::::::::::::···.;
..... "....

Figure 1.10 Inset to a map by Strang ways (1821) of the Pulcova River area, near St Petersburg, depicting the two
sedimentary dykes described in the accompanying report.

the sedimentary content. Strickland (1840), for extrusion of mud diapirs at the surface of the
example, remarked of dykes in Jurassic rocks Mississippi Delta - the so-called mudlumps -
near Brora in northeastern Scotland (first re- and interpreted them as responses to differential
ported, though not interpreted, by Murchison burial loads. Reports of sedimentary dykes con-
1829) that 'however much this rock might re- tinued to proliferate, and they continue to this
semble an aqueous product ... it forms genuine day, in some cases without adding much to an
intrusive dykes' (p. 599). He went on, 'the understanding of their origin (Dionne and Shilts
sedimentary structure of this rock forbids us to 1974; see M.Y. Williams (1927) for a historical
refer it to igneous injection from below, and review). The structures are sometimes also refer-
notwithstanding the complete resemblance of red to as clastic dykes. Although it has been
these intrusive masses to ordinary plutonic dykes, advocated that the term 'dyke' be restricted to
we have no recourse left but to refer them to those discordant sheets known to have been
aqueous deposition, filling up fissures which had forcefully injected, because it is not always poss-
been previously formed' (p. 600). James Dana ible in the geological record to distinguish these
discovered a series of sandstone dykes in 1841 from passive fissure-fills it seems practical to
during his explorations around the mouth of the allow the term to be used in a broad sense. Some
Columbia River, in what was then British workers distinguish passive fissure-fills as neptun-
Oregon Territory. He, too, interpreted the dykes ian dykes. This is more useful than using the
as fissure-fills, 'probably formed after or during term, as has been done, merely as a further
the deposition of the sandstone while the region synonym for all sedimentary dykes.
was yet underwater', and added that 'their num- Modern discussions on sedimentary intrusions
ber and irregularity evince that these regions include the significance for mining operations
have been often shaken by subterranean forces' (Bauer and Hill 1987); the importance of distin-
(Dana 1849, p. 654). guishing between dykes and periglacial wedge-
Diller (1890), reporting Tertiary sandstone like structures for environmental reconstructions
dykes west of the Sacramento River in California, (Elson 1975); horizontal injection (Aspler and
deduced that the sand was injected into the Donaldson 1986) and the downward injection of
fissures, upwards as well as downwards, and sands (Chown and Gobeil 1990) and diamictites
probably as a result of seismic activity. This last (von Brunn and Talbot 1986). Sedimentary dia-
inference was facilitated by the contemporary pirism is known on a substantial scale, from both
direct observations of fluid and sediment expul- ancient rocks and the modern sea floor (e.g.
sion at the earth's surface during the Sonora Brown and Orange 1993; section 2.3.3). Sedi-
earthquake of 1887 (Goodfellow 1887). Although mentary dykes have been used by Snavely and
many of the contributions appearing around this Pearl (1979) to deduce palaeostress patterns, and
time were concerned with sandstone intrusions, Borradaile (1977) used folded examples as esti-
Gresley (1898) described what he called clay mators of post-dyke compaction. Clastic sills
veins, associated with the Upper Palaeozoic coals were discussed by Hiscott (1979), and the difficul-
of Pennsylvania, and Shaw (1914) described the ties of distinguishing between sills and normal
Causes of deformation 29

beds were considered by Archer (1984). All the sediments, where such localized drainage con-
postulated mechanisms by which the sediment duits are referred to as pipes (e.g. J.A.A. Jones
was able to flow in order to generate these 1990), entrainment and transport of sediment has
various intrusions involve some combination of been shown to progressively modify the permea-
the liquefaction and thixotropic processes out- bility (Harp, Wells and Sarmiento 1990). It may
lined in section 1.3.2. Where such mobilized be that the various enigmatic cylindrical struc-
sediments reach the surface they are said, conti- tures and unusually long 'trace fossils' that have
nuing the analogy with igneous features, to ex- for long perplexed geologists (e.g. Wnuk and
trude (e.g. Schwan et al. 1980), from where the Maberry 1990) may in some instances have a
sediment may undergo recycling (Talbot and von dewatering origin. Further examples of structures
Brunn 1987). The extruded sediment may form associated with the transfer of fluid and sediment
volcanoes, in both sands (Okada and Whittaker are given in sections 4.3.4 and 9.5.3.
1979) and muds (Dionne 1973). Neumann (1976)
reported sand volcanoes actually forming during
modern construction operations - apparently the 1.3.5 Tectonism
load that induced the overpressuring was a cater- Recognition of the tectonic deformation of sedi-
pillar tractor! ments has a relatively short history, perhaps
Escaping pore fluid may travel pervasively, because of the kinds of reasons surmised in
particularly in shallow-buried sediments (Sills section 1.1.1. However, recently there have been
and Been 1984) or via more localized routes. The major changes in appreciating the extent to
increasing recognition on the modern sea floor of which tectonic stresses affect sediments as well as
localized biological communities has been inter- deeply buried rocks. The change in outlook has
preted to indicate the widespread existence of come about for three chief reasons:
channelized fluid escape (e.g. Fujioka and Taira
1989). Many of these seem to be sited at the 1. a realization of the close interplay in many
outcrop of faults, which are likely to provide geological settings between sedimentation and
conduits if they are undergoing dilative shear (e.g. tectonics;
Buckley 1989), but not all the seismic traces 2. a new interest in neotectonics;
indicate the presence of faults. The sea-floor 3. the advent of ocean-floor exploration and
features known as pockmarks - circular depress- plate tectonics concepts.
ions commonly a few metres across - are in some Each of these points is commented on below.
cases thought to be sites of water seepage (e.g. In a sedimentary sequence undergoing burial
Whiticar and Werner 1981; Harrington 1985), the transition from gravitational to tectonic
although the overall genesis is probably more stress (section 1.1.1) as the dominant cause of
varied (Hovland and Judd 1988). The influence deformation is highly gradational. In many set-
of pre-existing fractures on sediment permeability tings the highest levels of a sedimentary pile are
has been explored in the soil mechanics context very largely under the effect of gravity. This is
by Bosscher, Bruxvoort and Kelley (1987); Walsh certainly the case with the high-level structures
(1981) gave a mathematical analysis of the effects considered in Chapters 4 and 5. However, the
of fracture width and roughness. Chapter 7 ex- role of tectonism in governing the patterns and
plores further the role of localized strain zones processes through which sediments are first laid
for sediment drainage. down is now widely accepted (Ingersoll 1988;
Preserved structures show that the velocity of Pickering and Taira 1994). Gravity and regional
the fluid expulsion is sometimes sufficient to tectonics work hand-in-hand to influence
entrain some of the host sediment, although the sedimentation.
physics of the process has received little attention Almost immediately after deposition a sedi-
in the present context. It has been much studied ment becomes subject to a smaller scale but
in fluid mechanics and sedimentology (Allen equally close interplay between gravitational and
1984), but the slow, non-turbulent conditions tectonic stress. Gravity is mainly responsible for
relevant to buried sediment are probably more driving the consolidation of the sediment, at least
analogous to the entrainment of crystals in a in extensional sedimentary basins, but lateral
moving magma (e.g. S. Blake 1987). In on-land tectonic stresses can play a significant part in
30 Introduction and overview

contractional settings. Horizontal tectonic stresses convergence in Lower Palaeozoic times. The
influence the extent to which the sediment fol- precocity and complexity of the structures that
lows a Ko or some other consolidation stress can develop in a compressive regime were em-
path during burial (section 2.2.6). The importance phasized by Labaume, Bousquet and Lanzafame
of these deep-seated stresses grows as sediments (1990) as a result of their investigations south of
undergo increasing lithification and extend the Mt Etna, Sicily.
range of their elastic behaviour. Such interactions Structures preserved from extensional tectonic
between burial consolidation, lateral tectonic regimes have also been documented. An example
stresses, and the elastic-viscous response of is the work of Montenat, Barrier and Ott
lithifying sediments are considered in detail in d'Estevou (1991) around the Strait of Messina,
section 6.2. Section 7.3 discusses the tectonic Italy. Part of their reconstruction of the regional
processes that are responsible for driving fluids extensional movements was based on distinguish-
out of the sediments as they are buried, deformed ing between neptunian dykes in the sense of
and progressively lithified. infilled fissures (section 1.3.4) and those dykes
Appreciation of the role of active tectonism in resulting from sediment liquefaction. Regarding
influencing sedimentation leads to the notion a strike-slip setting, the importance of sub-
that the tectonic stresses generating the regional horizontal fault motion in governing Triassic
structures can also have some deforming effect on sedimentary basins in Morocco was identified
the sediments that are being deposited (Higgs by Laville and Petit (1984) on the basis of pre-
1988). As an example, Rascoe (1975) related local- lithification fault features such as slickenline
ized deformation breccias and features that he orientations. Genna (1988) measured folds, sand-
termed gravity-slide faults and sand-blows, pre- stone dykes, cleavage and microshears to derive
served in Upper Palaeozoic rocks in Oklahoma, the stress regimes that were operative during
to the growth at the time of sedimentation of early, prelithification tectonism in the Pyrenees.
regional anticlines and domes in the underlying These too were indicative of strike-slip motions,
basin floor. He suggested that analysis of the perhaps accompanying regional flexure.
pre-lithification deformation structures should The subject of neotectonics includes tectonic
aid in evaluating the hydrocarbon potential of processes and structures that have arisen in re-
the area, as they helped detect the positive cent geological time up to the present. Slemmons
structures that were influencing the nature of (1991) and Pavlides (1989) discuss the defini-
the sedimentation. Rumsey (1971) provided a tion of the term, and Becker (1993) assesses the
general review, in the petroleum context, of length of period covered. The kinds of structures
how fractures in an unlithified sedimentary cover dealt with in this field are related to stresses of
can reflect regional features in the underlying lithospheric origin but are chiefly developed in
basement rocks. surficial, poorly lithified sediments (Stewart and
Understanding of pre-lithification tectonic Hancock 1994). Many of the relevant observa-
structures preserved in rocks is probably most tions so far have been on subaerial processes but
advanced in convergent-margin settings. As it is now proving possible to measure directly the
examples: Agar (1988) analysed faults, sheath patterns and magnitudes of tectonically induced
folds and broken formations in the Cenozoic strain accumulating in modern submarine sedi-
Shimanto Belt in order to decipher the kin- ments (Larsen, Agnew and Hager 1993).
ematics of convergence between the Phillipine Neotectonic faults and associated features,
Sea and Eurasian plates; G.F. Moore and Karig such as joints, form particularly intensively in
(1980) described shear structures that formed areas of earthquake activity, where they can
during plate convergence in the Sunda Trench; impinge on various aspects of society and pro-
Kimura, Koga and Fujioka (1989) reported veins, vide indications of seismic risk. Russ (1979), for
foliations and kink bands developed during example, analysed cross-cutting fractures and
bedding-parallel shear and consolidation of associated folds and sand dykes in the New
sediments between the Mariana and Yapp Madrid earthquake zone of the central USA in
trenches; and Knipe and Needham (1986) an effort to assess the seismic periodicity of the
discussed structures in the Southern Uplands region. Bullard and Lettis (1993) emphasized
of Scotland interpreted as the result of plate the importance of ductile features, such as folds,
Causes of deformation 31

in their analysis of earthquake risk in the Los At first, diagnosis of the structures in retrieved
Angeles basin. There, folds on the scale of several cores as being of tectonic origin rather than being
kilometres are forming on surficial sediments, in gravity- or drilling-induced was frustrating (von
association with blind contractional thrusting at Huene and Kulm 1973). Consequently, early
depth. The interaction of deep-seated stresses work was directed at assessing the effect of
with surficial sediments has even been argued tectonism on consolidation rather than deforma-
to be the fundamental influence on the geo- tion (e.g. Lee, Olson and von Huene 1973).
morphology of the earth's surface (Scheidegger However, the importance of tectonic stresses
and Ai 1986). Attempts to use joints and faults for both dewatering and deformation rapidly
as indicators of the stress system of sedi- became clear (Carson, von Huene and Arthur
mentary basins are reviewed in section 6.2.6. 1982). J.C. Moore and Karig (1976), reporting
Of course the most direct way of determining folds and foliations from the active Nankai
current stress configurations is to make in situ accretionary prism, presented one of the first
measurements. However, there are severe prob- systematic documentations of the tectonic origin
lems in applying the methods to particulate ma- of such structures. An outline of the overall
terials. Section 6.2.5 analyses the efforts that have structural style of accretionary prisms was
been made. provided by J.e. Moore and Lundberg (1986)
Regarding the repercussions of oceanographic and a summary is given here in section 6.3.3.
exploration and plate tectonics thinking, there Lundberg and Moore (1986) reviewed the de-
has been some direct sea-floor documentation of formation structures recovered to that time by
extensional tectonic features. S.J. Williams (1987), DSDP drilling at convergent margins. Examples
for example, has used seismic and shallow coring of tectonic structures in sediments from the ex-
methods to investigate extensional faulting in the ceptionally well-documented Nankai prism, off
east Atlantic, and Masson (1991) has discussed southwest Japan, are described in sections 6.3.4
extensional fracture patterns in sediments on the and 8.2, and illustrated in Figure 9.5.
outer walls of subduction-related trenches, as The tectonic deformation of sediments is a
seen on GLORIA (side-scan sonar) surveys. field that has proved especially fruitful for the
Masson concluded that the fracture orientation application of soil mechanics principles. Both
almost always reflects either the strike of the DSDP and Ocean Drilling Program (ODP)
the subducting slab or an inherited oceanic- have given some priority, in all tectonic settings,
spreading fabric. to investigating the geotechnical behaviour of the
Nevertheless, it is at the sites of plate con- sediments encountered. The physical properties
vergence where tectonically induced sediment of sediments are analysed routinely during the
deformation has emerged as a topic of import- shipboard operations, and large repositories of
ance. In his historical review, Carson (1983) data are to be found in DSDP and ODP cruise
remarked that a principal result of geological reports.
research at convergent margins has been the Nowhere have geotechnical relationships been
realization that deformation occurs remarkably pursued more vigorously than in the study of
early, such that older, highly deformed sequences accretionary prisms. For example, Karig (1986)
from this setting almost certainly have an early synthesized the data available at that time in an
history, which so far has been largely un- analysis of the mechanical state of sediments
appreciated. When seismic sections across what accreted into the Nankai prism. Examples of
have come to be known as accretionary newer data include the results of ODP Leg 110 to
prisms were first interpreted as a series of Barbados (Mascle et aI. 1988), Leg 131 to the
landward-dipping thrusts, it followed that these Nankai prism (Taira et aI. 1991), and Leg 146 to
had to be affecting high-level sediments. the Cascadia prism (Carson et aI., in press). In
Sampling of the materials by the Deep Sea some cases knowledge is sufficient to enable
Drilling Project (DSDP) verified that the stresses stress path diagrams (sections 1.2.4 and 2.2.5)
arising from plate convergence were being to be drawn, though with varying degrees of
transmitted through the wet, unlithified sedi- quantitative rigour. Such diagrams are used in
ments and producing a variety of deformation sections 6.3.5 and 8.2 to summarize two slightly
structures. different synoptic models of progressive sediment
32 Introduction and overview

deformation in the Nankai prism. The latter less dynamic features such as pillow lavas. Small-
example is set in the context of incipient melange er ratios, which would be expected if the magma
generation. Section 6.3 presents a more extended is entering buried sediment, have been inves-
treatment of stress paths for tectonically con- tigated less systematically, and much of the infor-
tracting sediments and introduces new experi- mation is merely descriptive.
mental data. In so doing it illustrates the new The interface between the two materials is
solicitude for the quantitative understanding of influential in both explosive and quiescent behav-
sediment deformation, in addition to summariz- iour because of the influence of geometric bound-
ing the current state of knowledge on the mech- ary instabilities that can be generated (Wohletz
anical behaviour of sediments in the compres- 1986). The presence at the interface of highly
sional tectonic environment. active perturbations can induce rapid instability
and violent mixing, leading to explosivity. In
more ordered intrusion, the perturbations are less
1.3.6 Igneous activity
unstable but serve to nucleate sites of attempted
The intrusion of magma into sediments has vari- mixing. The situation is analogous to the in situ
ous effects, and a major control is the degree of disturbances of differing density sediments men-
lithification - and hence mechanical resistance - tioned in section 1.3.2. Needham (1978) described
of the sediments. This influences the intensity of circular and rectilinear depressions, up to 250 m
any diapiric effect the magma may induce, but and 50 m across, respectively, in middle Pro-
it also controls the extent of magma-sediment terozoic sandstones of the Northern Territories,
interaction. For example, in Pliocene sediments Australia, which he interpreted as being exhumed
of Inyo County, California, Bacon and Duffield from below basalt flows. The instabilities that
(1978) noted that a basaltic intrusion passed arose from the reversed density gradient across
cleanly through several horizons of differing the base of the basalt flow had induced 'giant
lithology, and only interacted with the mud- load moulds', the crest areas of which show
silt layers. The intrusion formed a flat-lying 'giant convolute laminations'. In other situations,
wedge, with a progression towards it of in- the emplacement of igneous sheets can evidently
cipient structures in the more distant mud-silt, cause the sedimentary substrate to fail, for Dixon
through asymmetric isoclinal folds, to a chaotic (1990) recorded both magma-sediment breccia-
zone at the igneous contact. From a judgement tion, below an Ordovician andesite intrusion in
of the difference in structural dimensions and east-central Wales, and syn-emplacement, prob-
geometry the authors suggested that at the ably listric, gravity slides. Density instabilities
time of intrusion the silt and mud layers had a coupled with downslope slumping were invoked
viscosity ratio reaching as much as 20, and were by Lorenz (1982) to explain complex 'three-limb'-
therefore in a delicate state. Leat and Thompson style folds in the Dunnage melange of Newfound-
(1988), reporting the results of Miocene volcan- land.
ism in northwestern Colorado, noted that basalt Kokelaar (1982) has emphasized the import-
had no interaction at all with the Cretaceous ance of sediment fluidization at the margins of
rocks that were lithified at the time of intrusion, some igneous intrusions. He argued that al-
but that it mixed violently with the overlying, though fluidization would be curbed by burial
unlithified Miocene deposits. pressures greater than about 30 MPa, equivalent
Explosive reactions between sediment and to a depth of 3 km of sea water or 1.6 km of wet
magma have received greater attention from sediment, at levels higher than this, the fluidiz-
volcanologists than other kinds of interaction, ation process will induce complex deformation
because of their importance for generating vol- structures in the contact region while allowing
canic features at the land surface. Factors that the bulk of the host sediment to undergo minimal
control the degree of explosivity include the sedi- disturbance. A case history of fluidization and
ment to magma ratio (Kokelaar 1986) and the other interactions of silicic volcanism with shal-
geometry of the magma-sediment interface. The- low-buried sediments is provided by the features
ory and experiment suggest that a water-magma preserved in the Lower Ordovician rocks of
mass fraction of about 0.35 generates maximum Ramsey Island, southwest Wales (Kokelaar, Be-
explosive energy, with greater ratios leading to vins and Roach 1985).
Causes of deformation 33

One of the few attempts to analyse the quanti-

tative controls on magma entering buried sedi- 1.0
ment was prompted by the successful drilling by
the DSDP, in the Guaymas Basin, central Gulf of 0.8
California, of a series of active sills within wet, S
unlithified sediments (Einsele 1982). One of the 0 0.6
sills encountered was still hot! The two major g
effects of the sills on the adjacent sediment were 0.4
thermally driven diagenetic changes and load-
induced consolidation. The two are interrelated.
Not only does the increased temperature of the
pore water promote diagenetic reactions, es-
020 40 60 80
pecially the dissolution of opaline silica and the Present porosity (%) (1]2)
partial dissolution of carbonate, but above a Figure 1.11 Diagram based on equation (1.19) to relate
critical threshold of burial pressure and tempera- the amount of water loss (expressed as a column
ture the pore water transforms into steam. The height, hpj from a sediment undergoing igneous intrusion
expansion associated with this reaction increases (and where the known initial and present porosities are
the fluid-pressure gradient away from the igneous fl1 and fl2' respectively) to the thickness of an invading sill.
For example, if the porosity were reduced from 80 to
contact, thus accelerating water expulsion. The 70%, then from the fl2+10 curve and the present 70%
steam film provides some insulation, so that the porosity value, (=0.5, and hp =0.5h 2. If h2 (the present
rate of heat loss is reduced sufficiently for mineral height of the relevant sediment section) has been reduced
and organic reactions to reach completion, from 15 to 10 m by the introduction of a 5-m sill, the
column of pore water expelled will equal 0.5 h2' giving
producing baking within the contact zone. Also, a similar value of 5 m. Thus, in conditions such as this, and
the effective stresses may be sufficiently reduced providing the intrusion is sufficiently slow, pore water
to prompt hydraulic fracture in the contact zone, can be expelled to make room for the igneous material.
leading to some localized drainage and enhanced (From Einsele (1982). Used with permission of
diagenesis. G. Einsele.)
Einsele (1982) attempted some approximate 30-40 m in thickness and covering a 1 km 2
quantitative assessment of the consolidation in- area will expel 3 x 107 m 3 of pore water. The
duced by these sills, using the relation: drainage probably takes place along induced
fractures rather than completely pervasively,
hp=h1-hz=hz ( 100-'11 =f'hz, (1.19) leading to hot springs at the sea floor. Einsele
calculated, using the physical properties meas-
where, expressing the fluid pressures as heights of ured at the Guaymas location and other likely
water column supported: hp = height of expelled values, that at burial depths greater than about
pore water; hl = original height of pore water; 200 m the pressures would only allow the
h2 = present, reduced height of pore water; magma to intrude as dykes, which would use
1] 1 = original porosity; and 1] 2 = present, reduced the tensional fissures induced by ocean spread-
porosity. Some representative values are given in ing. Above this level the pressures would allow
Figure 1.11. The results suggest that the water horizontal intrusion and, following the pro-
expelled from the sediment is very roughly the cesses outlined above, a sequence of sills to be
same as the sill volume, so that the mechanical built up, separated by sediment hardened by
emplacement of the sill (as opposed to the ther- the heat of the underlying sill. Einsele's con-
mal effects) does not have to significantly deform ception of this successive emplacement is en-
the host material, nor produce topographic ef- capsulated in Figure 1.12. In summary, although
fects at the sea floor. Some intrusions, of course, some intrusions are manifestly diapiric, in
are diapiric; the strain patterns around a forceful, poorly consolidated sediments it is the thermal
laterally invading sheet are probably very similar effects in the vicinity of the magma that common-
to the effects around salt modelled by Yu, Lerche ly induce the greatest physical effects on the
and Lowrie (1991). As an illustration of the huge host material; consolidation can allow magma
amounts of fluid expulsion that igneous emplace- to be emplaced with little mechanical dis-
ment can entail, according to Einsele (1982), a sill turbance.
34 Introduction and overview

ment, in recent years diapirism has also been

invoked to explain the origin of melanges. Geo-
logical settings such as convergent plate margins
commonly involve all three groups of processes,
acting on materials at various stages of lithifica-
tion. The interaction of a range of mechanical
processes is intimate.
Melange studies have a long history. Geo-
logists have long been intrigued by the occur-
rence of unusually coarse breccias and con-
-t glomerates, commonly with a mixture of clast
types, more or less isolated, and perhaps of exotic
11 provenance. Many examples showed some degree
~ of deformation, and this was usually taken to be
symptomatic of the generating processes. In line
with the thinking on deformation alluded to in
section 1.1.1, some form of deep-seated crush-
ing was normally invoked by way of origin.
Murchison (1829), for example, interpreted the
Jurassic 'boulder beds' at Helmsdale, northeast
Scotland, as resulting from upheavals associated
with the nearby Helmsdale granite intrusion: 'in
Figure 1.12 Einsele's conception of igneous emplacement breaking through those submarine deposits
and water loss from shallow-buried sediment, based on which might not perhaps have been originally in
DSDP drilling of the Guaymas Basin, Gulf of California. contact, it has fractured and dislocated their beds
1 = lower, older sill; 2=upper, younger sill; 3=possible
steam zone, at contact with wet sediments; 4=zone of
as to have prepared them for reconsolidation in
porosity reduction due to upper sill; 5=shrinkage cracks the state of a brecciated rock' (p. 354). Early
acting as vents for escaping steam; 6=possible transfer of explanations of exotic blocks in Alpine deposits
dissolved chemical species from magma to sediment; also involved deeply buried shearing. Subsequent
7 = permeable layer with lateral water flow; 8 = sandy base thinking invoked ice as the transporting agent,
of overlying turbidites; 9 = buried older hydrothermal
deposits; 10 = zone of earlier porosity reduction in
but the belief gradually returned that subaqueous
association with emplacement of older, lower sill; instabilities most readily accounted for many of
11 =zones of water loss. (Adapted from Einsele (1982). the examples. The historical development is
Used with permission of G. Einsele.) summarized by Bailey, Collet and Field (1928)
in the light of their studies on what they inter-
preted to be submarine landslips, preserved in
1.4 MELANGES AS A CASE HISTORY Palaeozoic rocks near Quebec City, Quebec.
As appreciation of the extent of submarine
None of the agents outlined in the preceding instability in the modern oceans grew, more
sections operates in isolation. Mass movements, formations of these mixed rocks were interpreted
for instance, can be prompted by and interact in this light, and in 1956 the term olistostrome
with glacial movements and tectonic stresses; was coined. However, a purely gravitational
fluid motions are involved in all the deforma- origin was not always clear in practice, and in
tions. These interactions between different some quarters the term began to incorporate the
agents and processes are perhaps nowhere possibility of tectonic processes also having been
better illustrated than in the materials known as involved. Back in 1919, Edward Greenly had
melanges. Indeed, it is the very closeness of the coined the term melange for a unit of mixed
interplay, and the difficulty with ancient rocks in Anglesey, North Wales, which he be-
examples of isolating the contributing factors, lieved were the result of intense, tectonically
that have engendered the long-standing debate induced shearing. The term has enjoyed a renais-
on the origin of melanges. For example, as well as sance in recent decades, but with a variety of
tectonic shearing and downslope mass move- implications, together with debate over the extent
Melanges as a case history 35

to which Greenly meant it to be confined to fluctuations may be even more spasmodic than
situations known to be of tectonic origin. Some this three-stage sequence.
authors feel its value is as a purely descriptive K.M. Brown and Orange (1993) further argued
term, so that there can be sedimentary melanges. from experimental and natural evidence that the
Indeed the 'type' melange in Anglesey is now transition from independent particulate flow
interpreted as a sedimentary rather than a tec- (section 1.1.1) of the first stage to the period of
tonic product. Part of the interpretational diffi- cataclasis took place while the sands retained a
culty is that disruption and mixing of sediments porosity of between 31 and 41 % and that there-
by gravitational processes and diapirism are pro- fore the effective stresses could not have exceeded
moted in active tectonic environments, where the 5 MPa. Stratigraphical evidence requires that
products are likely to be overprinted by tecton- the processes occurred at burial depths greater
ism, which can act on the deposit at any stage of than 2 km. Therefore, in order to produce such
its progress to lithification. The problem is par- low effective pressures at that depth, the fluid
ticularly well shown at convergent plate margins, pressures must have approached litho static
in the materials known as shale-matrix melanges. conditions. Such values might be expected in
The matter is discussed in Chapter 8. conditions of diapirism, and serve once again to
The application of soil mechanics principles to emphasize the crucial importance of pore-fluid
understanding melange formation has helped ex- pressures in governing sediment deformation.
plain some of the features observed in these It is not relevant here to explore the termino-
materials and has also emphasized the para- logical minefield of melanges, diapirs and olisto-
mount role of pore-fluid pressure. Brandon stromes. The importance here revolves around
(1989), for example, working in the Pacific Rim two points.
melange complex of Vancouver Island, was able
to distinguish between a unit of layer-parallel 1. The terminological difficulties arise because
fragmentation lacking in primary sedimentary of the very intimacy, in the kinds of environments
structures and units comprising disordered, dis- in which melanges are likely to form, of gravi-
membered asymmetric folds with well-preserved tational and tectonic processes. That is: melanges
sedimentary structures. He interpreted the latter illustrate the range, interaction and complexity
horizons as the product of retrogressive shear of the kinds of processes by which natural sedi-
failure (section 1.3.3) and relatively coherent ments deform.
downslope mass movement, and the former as 2. The interpretational problem is exacerbated
resulting from in situ liquefaction and intense when dealing with ancient examples, because
lateral flowage under conditions of cyclic loading criteria for distinguishing between gravitational
(section 1.3.2). and tectonic melanges and identifying the state
As a further example, K.M. Brown and Orange of lithification are at best subtle - melanges
(1993) recognized three stages in the evolution of illustrate the kinds of difficulties faced by field
an on-land diapiric melange complex associated geologists attempting to interpret deformed sedi-
with the Cascadia accretionary prism. Each mentary rocks.
period resulted from a major change in condi- These two matters are taken up in Chapter 8.
tions of pore-fluid pressure, causing the material There, examples of melanges are presented to
to deviate from critical state failure. The first act as case histories of the difficulties of dealing
stage took place under low effective stress, which with sediment deformation in nature. The second
allowed primary disruption of the sediments and concern, of recognizing pre-lithification deforma-
turbulent intrusion. The subsequent development tion long after the materials have been turned
of broad shear zones involving cataclasis into rocks is pursued further in section 9.6. How
evidenced a period of higher effective stress and does the geologist go about diagnosing structures
strain hardening. A later rise in fluid pressure as the result of sediment deformation? After all,
allowed dilation and strain softening, especially application to the geological record of the prin-
at the diapir margins, and the final emplace- ciples and ideas outlined here and amplified in
ment of the melange mass. The intricate geo- the following chapters assumes that the geologist
metry shown by mineralized veins in the melange knows he is dealing with structures formed in
(section 8.5) suggests that the fluid pressure unlithified sediments.

Mechanical principles of
sediment deformation

2.1 INTRODUCTION grain boundaries and the redistribution of min-

eral material as grain overgrowths (Rutter 1976,
The majority of sediments originate as weak 1983; Gratier 1987). In such cases, lithification
materials that deform readily, but which become not only enhances rock strength, it is also a
stronger during burial and residence at depth. consequence of rock deformation. Sediments
This process of lithification eventually converts may, however, compact (lose bulk volume) dur-
the sediments into rocks (section 1.1.1), which ing burial without the development of strength
exhibit pronounced strength. Lithification is a due to intergranular bonding.
progressive change: the transition from unlithi- The term consolidation is used here in its
fied through partially lithified sediment to rock is specific geotechnical sense, as defined in section
highly gradational. 1.2.2. For a particular stress state the sediment
Sediments are of great concern to geotechnical consolidates until an equilibrium pore volume is
and foundation engineers, who normally refer to developed (Figure 2.1a). The term normal con-
the materials as soils (Lambe and Whitman solidation is commonly used to describe the re-
1979). The mechanical behaviour of such ma- duction in equilibrium pore volume that accom-
terials is described in the literature of soil mech- panies progressively increasing stress (Figure
anics (Schofield and Wroth 1968; Atkinson and 2.1b). This terminology is used throughout this
Bransby 1978), a subdiscipline of civil engineer- book, although some authors also refer to this
ing. The principles of soil mechanics form the process as 'compression'. The reader is referred to
basis of this chapter. However, although the term Atkinson and Bransby (1978) for a comprehen-
soil is very widely used in the engineering litera- sive discussion of the use of these terms in
ture, sediment is preferred here and throughout geotechnics. Consolidation results in a progress-
the present book because this term is so firmly ive densification of the sediment in the presence
established in geology. In this context, sediments of increasing stress, and typically accompanies
can be regarded as aggregates of mineral particles sediment burial (Burland 1990). Consolidation is
in frictional contact (section 1.1.1), although some only inhibited during sediment burial by low
examples, such as clays, may have weak interpar- permeabilities (which prevent or substantially re-
ticle bonds. tard the dissipation of excess pore fluid pressures)
In this chapter, only the mechanical behaviour and integranular cementation (which resists fric-
of sediments will be considered, although this tional sliding of grains). Although consolidation
closely interrelates with diagenetic changes. For has been described extensively in soil mechanics
example, lithified strength often develops because (Atkinson and Bransby 1978; Lambe and Whit-
compression of grain boundaries enhances sol- man 1979), it has received little attention in the
ution and diffusion of the mineral material, literature of sedimentology and structural geo-
allowing grain interpenetration, serration of logy (but see Rieke and Chilangarian 1974; M.E.

The Geological Deformation of Sediments Edited by Alex Maltman Published in 1994 by Chapman & Hall ISBN 0 412 40590 3
38 Mechanical principles of sediment deformation

compaction. If the same material were to be

deposited on a steeply inclined slope, then the
action of gravity would promote downslope
g movement rather than compaction. In
o geomechanical terms the material would tend to
fail and the downslope sliding would be a shear
deformation. Shear deformation in particulate
materials is not normally accompanied by com-
(a) Time paction, so that early slumping will cause a
different fabric to develop. Thus, two sediments
that were identical at the moment of deposition
can, by virtue of the physical environment in

which they were deposited, evolve in quite differ-
ent ways and develop very different fabrics (Fig-
ure 2.2). A good example of this can be found in
the Danian Chalks in the Central Graben of the
North Sea. Here, deposition on slopes has led to
(b) Effective stress reworking of some parts of the sequence through
Figure 2.1 Illustrations of consolidation and normal
slumping, debris flows and turbidity currents
consolidation. (a) The response of the pore volume of a (Kennedy 1986). Fabrics developed at this early
saturated permeable clastic sediment to the application stage have influenced the subsequent compaction
of two increments of load. This is the change in pore history, even though the rock is now buried to a
volume that accompanies the time-dependent dissipation
of an excess pore fluid pressure, and is termed
depth of approximately 3 km. Today, the re-
consolidation. (b) The relationship between increasing worked chalks still preserve larger primary pore
effective stress and equilibrium (fully consolidated) sizes which, where not infilled with sparry calcite,
pore-volume states in a similar sediment. This is the provide the best hydrocarbon reservoir quality,
compression curve that is the locus of the end-point such as in the large Ekofisk oil field.
pore-volume states for each consolidation increment.
See also Figure 2.20. Vertical compaction

Jones and Addis 1986; M.E. Jones et ai. 1991).

Consolidation constitutes a major facet of mech-
anicallithification and in many sediments it is the
major mechanism responsible for pore-volume
changes, the development of stress and pore-
pressure regimes, and the development of early
fabrics (most notably in mud rocks). It brings the
mineral grains closer together, increasing the fric-
tional strength (section 1.2.3) and prompting
chemical bonding. Figure 2.2 The influence of location on the response of
Sediments, from the moment of deposition, are a sediment to gravitational loading. These early fabrics
will be preserved during the geological history of the
subject to body forces, especially the action of resulting sedimentary rock and will influence subsequent
gravity, but with increasing burial tectonic activ- deformation.
ity can also become important. The response of
any sediment to these forces depends upon its It follows from this introductory discussion
physical properties, the magnitude of the forces that irrespective of depositional environment and
and its physical location. For example, a clastic sediment type there are two ubiquitous responses
sediment deposited slowly on a flat sea floor will to the action of applied forces. From the moment
compact under gravitational loading through ex- of deposition, all sediments will develop coexist-
pulsion of pore fluid and grain rotation (normal ing and related states of stress and strain. These
consolidation). If the sediment contains clays or stress-strain states are diverse and complex, and
other platy minerals then a fabric will develop evolve during burial, diagenesis and tectonism.
which reflects this passive state of gravitational They generally exert a strong influence on the
Mechanics of particulate media 39

nature of any fabrics that develop in the sedi- simultaneously within the material (M.E. Jones et
ment, and are themselves strongly influenced by ai. 1991). As the forces change in magnitude
the evolving nature of those fabrics. To under- and/or orientation (as may occur during sedi-
stand the deformation characteristics of sedi- ment burial), the stress-strain state progressively
ments and weak sedimentary rocks it is necessary changes, maintaining equilibrium (generally a
to understand this complex interrelationship be- condition where no displacements are occurring)
tween the applied forces, stress state, strain and between the fabric of the sediment and the forces.
fabric. The action of a force thus causes the sediment to
undergo changes in shape, which constitute the
state of deformation or strain (Figure 2.3a) How-
2.2 MECHANICS OF PARTICULATE ever, the sediment is able to resist the develop-
ment of strain to an extent that depends upon
MEDIA IN THEORY AND EXPERIMENT both its strength and its geometrical factors (Fig-
ure 2.3b). This resistance constitutes the state of
2.2.1 Force. stress and strain stress (Janbu 1985). Thus a weak, unlithified
When a sediment or sedimentary rock is subject sediment may achieve a particular strain state in
to forces, states of stress and strain develop response to a weak force and the resulting

(a) F

1i Q)

11 al



(b) Confining sleeve ~


Figure 2.3 Illustrations of force, stress and strain. (a) The development of strain in an unconfined, right-cylindrical
specimen due to the application of a force (F). 10 is the original length and 11 the deformed length. (b) The effects of the
application of a force (F) to a confined specimen. The confining sleeve prevents lateral strain and a horizontal compressive
stress develops to maintain equilibrium. The vertical stress in the rock sample is also related to the strain state and is
illustrated schematically in this figure as the progressively compressing spring. (c) Idealized force (and hence stress)
and strain relationships for a strong and a weak sediment. This shows the variation in compressibility and can be likened
to changing the stiffness of the spring in (b). (d) Idealized development of vertical stress and vertical shortening in a
sediment (stippled) deposited on a horizontal surface and subject only to gravitational loading. This is equivalent to the
strain states that develop in the specimen illustrated in (b) as the axial force is increased.
40 Mechanical principles of sediment deformation

stresses will be small, whereas a lithified sedimen- ments using an oedometer (a one-dimensional
tary rock will require a much larger force to compression apparatus, Figure 2.4a). Experi-
achieve the same state of strain, and the asso- ments on the initial compaction of sediments
ciated stresses will then be large (Figure 2.3c). have been conducted (Been and Sills 1981; Edge
In formal mechanics, stress is defined as the and Sills 1989; Leddra, PetIey and Jones 1992)
force acting per unit area: and reveal relationships between stress and strain
similar to that shown in Figure 2.4b. This curve
(J= - (2.1) shows that as the sediment achieves equilibrium
A' strain states for the increasing load, the in-
where (J is the stress, F is the force and A is the crement of stress that develops per increment
area over which the force is acting. Strain (e) is of strain increases. This means that as the sedi-
most simply expressed as elongation and is de- ment deforms (through compression of its pore
fined (to keep contractional strains positive) by space) its compressibility progressively decreases.
equation (2.2): As the sediment equilibrates to the increasing
load it is becoming stronger and more able to
10- 11 resist subsequent deformation. This is why the
e=-- (2.2)
II ' stress increase is greater with increasing strain.
If the force acting upon the sediment in the
where 10 is the original length and 11 is the oedometer is removed, then the stress decays
deformed length of the object subject to strain with the force but some strain state will generally
(Figure 2.3a). In geomechanics it is the convention be preserved, giving an unloading curve of the
that stresses that act in a compressive manner and type illustrated in Figure 2.4b. The deformation
strains that are compressive are regarded as posi- resulting from the application of the force is
tive. The states of strain and stress that develop largely irrecoverable. This is a feature of the
when a sediment is subject to the action of a force deformation of sediments and other particulate
are related by a physical property termed the materials (Atkinson and Bransby 1978; Burland
compressibility. This is an inverse measure of 1990).
strength and is defined as the strain per unit stress: If a similar compression test were to be con-
C = !lei!l(J, (2.3) ducted on a lithified sedimentary rock, and pro-
viding that the applied force does not promote
where C is the compressibility; fle is the in- failure of the bonding responsible for lithification,
cremental increase in strain and fl(J is the in- then a relationship between stress and strain of
cremental increase in stress. the type shown in Figure 2.4c will be observed.
Thus when gravity acts on a sequence of unlithi- This is a linear relationship (generally only quasi-
fied sediments, a vertical compactional strain and linear for real materials) and shows that the
a vertical stress develop, both of which increase compressibility is nominally constant. When the
with depth. This burial stress is a product of the force is removed the strain is seen to recover as
density of the sediment and the depth of burial the stress decays. This style of materal behaviour
(Figure 2.3d) and is given by equation (2.4): is normal for strong materials and is described as
elasticity. Elastic materials obey Hooke's Law
(Jv= p'g' z, (2.4)
(Jaeger and Cook 1979). All rocks exhibit elastic-
where (Jv is the vertical stress in the sediment ity at small strains, but in highly compressible
column, p is the density of the sediment, g is materials such as unlithified sediments the range
gravity and z is the depth. If for simplicity we of this kind of behaviour is extremely limited
assume that the deformation resulting from the (Burland 1989). This leads to the important ob-
action of gravity is one-dimensional and in the servation that unlithified sediments and weak
vertical direction, then the vertical strain will also sedimentary rocks normally exhibit strongly non-
increase with depth as the vertical stress in- linear stress-strain relationships (Uriel and Ser-
creases, and its magnitude will depend upon the rano 1973; Vaughan 1985; M.E. Jones and
compressibility of the sediment. Leddra 1989; M.E. Jones et al. 1991) over the
It is possible to reproduce this relationship range of stresses encountered during deposition
between stress and strain in laboratory experi- and initial burial.
Mechanics of particulate media 41

If) Loading ~~


~ ~

ii5 Unloading

(b) (c) Strain

Figure 2.4 Consolidation, stress and strain. (a) Schematic illustration of an oedometer of the type used by Leddra,
Petley and Jones (1992) and others to study the gravitational compaction of sediments. Load is applied to the top piston
and pore fluids are expelled through a porous plate at the base of the sample. Apparatus of this type is routinely used
in geotechnics to study the consolidation of sediments. (b) A typical stress-strain curve for a weak saturated clay deforming
under oedometric loading. (c) Stress-strain curve for a sedimentary rock deforming within its elastic range.

The difference between the linear elastic defor- ment, the frictional contacts between the grains
mation of lithified sedimentary rocks and the are far weaker than the atomic bond strengths
non-linear deformation behaviour of particulate within the individual mineral crystals. The action
sediments is due to the location and nature of the of a force causes grains to slide past one another
strain in the material. Elastic deformations occur (Figure 2.5b), mobilizing only small stresses so
at an atomistic scale. The bonds between individ- that elastic distortion of atomic bonds is mini-
ual atoms or molecules in parts, or all, of the rock mal. This strain is accomplished with little work
are stretched or shortened like springs (Figure energy being stored in the system, so there is little
2.5a). The strain is located in these distorted or no potential for the strain to recover when the
bonds and the rock must store a considerable force (and associated stress) is removed.
amount of work energy to maintain the strain. This mechanical distinction between unlithified
This is facilitated by the stress that is mobilized sediments and sedimentary rocks is fundamental
with the strain. When the force promoting the to the behaviour of the materials and also ac-
strain is removed, the stress decays and the strain counts for the almost independent development
energy is released as the distorted bonds revert to of the otherwise related geotechnical disciplines
their original positions. In an unlithified sedi- of soil mechanics and rock mechanics. However,
42 Mechanical principles of sediment deformation

(a) sediments deform, to review what is meant by the

terms principal stress and shear stress and how
they are related. The significance of principal
stresses, which act normal to planes that do not
allow the mobilization of shear stresses, is illus-
trated in the following idealized example.
Consider the situation illustrated in Figure
2.6a. A rigid crystalline rock is resting on its flat
base on a smooth horizontal surface of a similar
material (this could, for example, be the idealized
~~ (b)
situation of a glacial erratic deposited on a
smooth glaciated surface). The rock is at rest, and
Figure 2.5 Recoverable (elastic) and irrecoverable neither it, nor the surface below, have strained
(inelastic) strain. (a) Schematic representation of the
location of strain in the interatomic bonds of elastic rock
appreciably as a result of its presence. None the
materials. The bonds are represented by springs, which less, the erratic is exerting a load on the rock
lengthen or shorten according to their orientation in which is mobilized as a stress that is proportional
response to the applied load. (b) Schematic representation to its mass acting over the area of contact. Both
of the location of strain in a particulate sediment subject to the boulder and the rock below will be subject to
increased load. The grains slide to new positions and,
when the load is removed, remain there because there is
small elastic strains. This stress is acting normal
no potential to return them to their original sites. Elastic to the contact surface and has no component that
behaviour in deforming sediments is discussed more fully can be resolved to act parallel with it because in
in Chapter 6. this situation no forces are acting to move the
boulder across the surface. It therefore remains in
the behaviour of a completely unlithified sedi-
position pressing down on the rock below. There
ment and of the equivalent fully lithified sedimen-
is no shear stress acting on the surface and the
tary rock represent the ends of a broad spectrum.
gravitational stress acting normal to it is, there-
Most naturally occurring sediments and sedi-
fore, a principal stress. Now consider the situ-
mentary rocks fall within this spectrum and ex-
ation in Figure 2.6b. Here the surface is inclined
hibit material behaviour which involves a combi-
to the direction of the gravitational force. The
nation of elastic and particulate deformations
vertical stress acting through the boulder is
depending upon the stress magnitude (Vriel and
still a principal stress but intersects the surface
Serrano 1973; Vaughan 1985; M.E. Jones and
Leddra 1989; M.E. Jones et al. 1991). Elastic
deformation is discussed further in section 6.2.

2.2.2 States of stress and strain in a body

Sediments are generally deposited in situations
where they constitute part of a laterally extensive
continuum. It is therefore incorrect to treat stress
and strain as one-dimensional quantities because
in real systems the reaction of a sediment to the
application of a force is three-dimensional and
will involve the generation of stresses and/or
strains in all directions. (b)
The formal description of the stresses acting
upon a cube that is at rest can be found in any Figure 2.6 Principle stresses and shear stress. (a) The
number of texts (e.g. Ramsay 1967; Jaeger 1969; development of a normal stress that is also a principal
Jaeger and Cook 1976; Bayly 1992). The treat- stress between a gravitationally loaded block and the
ment leads to the definition of the stress tensor in horizontal surface upon which it is resting. (b) The
development of normal and shear stresses between a
terms of principle stresses and shear stresses. gravitationally loaded block and the inclined surface on
There is no need to repeat the analysis here, but which it is resting. In this case, the greatest principal
it is necessary, before considering in detail how stress is still vertical.
Mechanics of particulate media 43

obliquely. Now two components can be resolved, tant. By making the assumption that only the
one acting in the plane of the surface (the shear greatest and least principal stresses are import-
stress) and the other normal to it (the normal ant, it is possible to represent the state of stress
stress). The shear stress develops because a com- by a Mohr's circle construction. An example of
ponent of the gravitational force is trying to such a plot is given in Figure 2.7, which shows
cause the boulder to slide downhill, whilst the clearly how the magnitude of the shear stress
normal stress is due to the component of the depends on the difference between the principal
gravitational force acting across the surface, try- stresses and the orientation of the surface on
ing to compress it. The magnitudes of the shear which the stress is resolved.
and normal stresses vary with the inclination of
the surface with respect to the principal stress. In
all stress systems there are three principal stresses
which are mutually perpendicular. These may
have the same magnitude, or can be very differ-
ent, and one or more may be zero. In the example
in Figure 2.6a, as one principal stress (the
greatest principal stress) is acting normal to the
contact surface the other two lie on that surface
and will be small.
The magnitude of normal and shear stresses
depends not only on orientation but also on the
difference between the principal stresses. If all
three principal stresses are equal, the shear stress Figure 2.7 The standard Mohr's circle construction
is zero and the normal stress is equal to the defining the states of magnitudes and normal stress
principal stress. The greater the difference be- a
acting on surfaces inclined at an angle to the greatest
principal stress within a body subject to principal
tween the maximum and minimum principal stresses 0",>0"3. (See Atkinson and Bransby (1978) or
stresses, the greater the shear stress. The magni- almost any structural geology or geomechanics textbook
tudes of the shear and normal stresses acting on a for a discussion and description of this construction.)
surface as illustrated in Figure 2.6 are given in
equations (2.5) and (2.6): The states of strain associated with the mobil-
0"1 -0"3 ization of normal and shear stresses are different.
1"0= - - s i n 20, (2.5) Normal stresses accompany changes in dimen-
sion in the direction parallel to the stress. Gene-
rally they are associated with a reduction in
(2.6) volume when the deformation is compressive
(Figure 2.8a). Shear stresses are associated with
where To is the shear stress resolved on a surface changes in shape and need not be accompanied
whose normal is inclined at angle () to the by volume change (Figure 2.8b). Thus a spectrum
greatest principal stress, (10 is the stress acting of strain states exists with entirely volumetric
normal to the surface and (11 and (13 are, respect- strain and entirely shear (distortional) deforma-
ively, the greatest and least principal stresses. tions as its end members. During deformation a
Inspection of equations (2.5) and (2.6) reveals that sediment will experience some combination of
only the greatest and least principle stresses are volumetric and distortional deformation, which
considered. This two-dimensional analysis is a will depend upon the character of the stress
useful approximation that will be used through- system. This diversity of natural stress systems
out this chapter. It must, however, be realized means that in nature any strain state may exist.
that where natural stress systems have been Thus sediments deposited on unstable slopes or
measured, it is generally the case that all three affected by syndepositional faulting will develop
principal stresses have different values (Teufel et fabrics associated with large shear deformations
al. 1992). In these real systems, deformations but may experience little or no compaction (vol-
occur in three dimensions and the magnitude of umetric strain) at the time (Figure 2.2), whereas
the intermediate principal stress is then impor- those deposited and buried in flat-lying quiescent
44 Mechanical principles of sediment deformation

0"1 = <T3 principal stress. It should, however, be noted that
V1 » V2 except where the earth's surface is horizontal,
gravity will induce shear stresses parallel to the
surface. These shear stresses are responsible for

downslope movements in large sediment bodies,
such as deltas, and for many of the fabrics asso-
ciated with landslides, slumps, debris flows
and sediments deposited by turbidity currents
(Chapter 5).
(a) CT1
2.2.3 Pore pressure, effective stress
(J1 » Cf3 and porosity
V1 = V2
Porous sediments and sedimentary rocks are
CT3 two-phase systems, consisting of a skeleton of
mineral grains and pore spaces filled with water,
air, brine, hydrocarbons or some combination of
these fluids. Generally, except at very shallow
depth, the pore space is filled with water, brine or
a liquid hydrocarbon with a compressibility that
is comparable with or less than that of the
(b) mineral skeleton. When these materials are sub-
ject to load, the resulting total stress (0") is par-
Figure 2.8 (a) The action of isotropic stress, leading to
a reduction in the volume of a sediment but without the
titioned between the effective stress that acts
distortional components of strain associated with shear through the mineral grains (0"') and the pore fluid
stress. (b) The deformation developed in an identical P(H 2 0)' Thus, in liquid-saturated sediments there
sample to that in (a) due to the development of the is generally an effective stress that acts through
same principal stress 01' In this case, however, the
lateral stress has been kept small and the sample has
the mineral grains and a pore fluid pressure
developed large resolved shear stresses. The result is acting in the pores (Figure 2.9). In extreme cases,
that the sample has shortened, but has distorted (through when the pore pressure is equal to the total stress,
lateral expansion), so that although the axial and shear the effective stress can be zero. For a sediment
strains are large, the volumetric strain is negligible. which is permeable and porous, and subject to
gravitational loading, the pore pressure increases
basins will compact under gravitational loading with depth in a similar manner to the vertical
but experience little bulk shear deformation stress, except that the rate of increase is some-
(Figure 2.2). what less because the density of water, brine or
At this point it is sensible to summarize the oil is less than the density of the mineral compo-
main points of this discussion.
1. Principal stresses act normal to planes on Ground surface
which the shear stress is zero.
2. There are always three principal stresses,
which are mutually perpendicular.
3. The magnitude of a shear stress depends on
the difference between the greatest and least
principal stress and on orientation with respect
to the principal stresses.
4. Principal (and other normal) stresses cause
compressional deformations (or elongations)
whereas shear stresses promote sliding or distor-
tion (shear strains).
5. At or near to the earth's surface, the stress Figure 2.9 Representation of the state of effective stress
induced by the gravitational force is always a in a saturated porous sediment in the subsurface.
Mechanics of particulate media 45

nents of the rock (Figure 2.10; see also Figure pressure caused by the increased load dissipates,
1.4). All compressional deformations of this sedi- allowing the effective stress to increase (Figure
ment are accompanied by changes in the effective 2.11). This process is dependent upon the per-
stress (a /), and are therefore associated with meability of the sediment and is termed primary
either an increase in total stress (a) (which may be consolidation. It is described more fully below.
due to progressive burial or tectonic loading) or a I
change in pore fluid pressure P(H 2 0) (Terzaghi I~
1936, 1943). The effective stress is most simply liil
c l!ll
expressed as the difference between the total .~
stress and the pore fluid pressure according to Ci5 I~
____ 1°
the Terzaghi equation: 0---'----- --I
(2.7) Time
Application of load
This equates with equation (1.12); the concepts
Figure 2.11 The relationship between pore-pressure
were introduced in section 1.2.5. dissipation and the development of strain with time in a
low-permeability sediment. For thick layers of shale,
Vertical stress the time axis of this type of plot may extend to several
million years.

Pore fluid pressure therefore determines the

magnitude of effective stresses acting within a
sediment. It has two components:
where P H20h is the hydrostatic component of the
pore pressure caused by the overlying column of
fluids extending to the earth's surface; and P H20e
is the component of pore fluid pressure that is in
excess of the hydrostatic gradient. At equilibrium:
pressure vertical stress
(Fluid density - 1100 kg m-3 ) (Grain density - 2700 kg m-3 ) (J' = (J- P H20h , (2.9)
Figure 2.10 Schematic illustration of the development but during sediment burial, equilibrium may not
of pore pressure, total vertical stress and vertical
effective stress with depth in a brine-saturated, porous
be attained and the effective stress changes con-
sediment subject to gravitational loading beneath a tinuously with time:
laterally extensive horizontal surface. (See also Figure
1.4.) d(J' d(J dPH2Dh dPH20e
-=----+--. (2.10)
dt dt dt dt
An increase in total stress will therefore cause a
change in effective stress only if the stress change Sediments undergoing burial may sustain large
is experienced by the mineral skeleton of the pore fluid pressures, or may become overpres-
sediment. This will occur readily in sediments sured through other mechanisms (see sections
that are of high permeability and able to drain. 2.2.3 and 2.3.4), with the result that even in the
Under these conditions, an increase in the load presence of large total stresses (due to substantial
acting on the sediment will lead to a reduction in burial or the action of large tectonic forces), the
pore volume as the total and effective stresses effective stress may still be very small.
increase and pore fluid is expelled. If, however, When sediments are subject to increased effec-
drainage of pore fluid is prevented, the pore tive stress, pore volume is lost and although the
volume of the sediment can decrease only if the actual detail of this process varies with the type
pore fluid compresses in the pore spaces. For of sediment and with the stress path (see below),
aqueous fluids under small loads, the compressi- the basic relationship is that shown in Figure
bility is insignificant, and an increase in pore fluid 2.12. This is a plot of porosity (the ratio between
pressure will occur. In this case strain will accu- pore volume and total rock volume) against
mulate only if the excess component of the pore the mean effective stress (defined as the mean
46 Mechanical principles of sediment deformation

For sedimentary rocks that have become lithi-

fied and are elastic at low stresses, the change in
pore volume with increasing effective stress is
~omplicated by the bonding stiffness. The behav-
~our of these materials is best explored by regard-
mg the effective stress as a resistance to deforma-
tion that is mobilized in the mineral skeleton as a
co~sequence of its rigidity (Janbu 1985). This
Figure 2.12 The relationship between porosity and
mean effective stress (p') for a porous clastic sediment resIstance may be viewed as containing three
deforming by compression of the pore space. See also components:
Figure 6.1.
of the three principal effective stresses p' =
where r c is the resistance due to strong inter-
(0'1 +0'2+0'3)/3). Similar compaction curves have
been determined for many sediments during granular bonding, ri is the resistance due to
geotechnical studies. These are commonly repre- intergranular friction, and r m is the resistance due
sented as straight-line plots of void ratio (the to weak intergranular chemical effects, such as
van der Waals bonds, water absorption and
ratio between pore volume and the volume of
osmosis. In particulate sediments the effective
mineral grains) against the natural logarithm of
stress is dominated by the frictional resistance
the mean stress (Figure 2.13). These plots show
(s.ands) or by the chemical resistance (clays) (At-
that the major mechanism of compaction is loss
of pore volume, and that for a given sediment kmson and Bransby 1978; Lambe and Whitman
under particular effective stress conditions there 1979). In porous lithified rocks, strong inter-
granular bonding is present and resistance to
is a maximum pore volume that can exist. As
deformation is greater while these bonds are
~he effective stress is increased the pore volume
intact (M.E. Jones et al. 1990).
IS decreased. Thus, increase in effective stress is
The elasticity produced by strong intergranu-
accompanied by an increase in sediment
lar bonding at low effective stresses is destroyed
strength due to progressive densification. Be-
once the strength of the mineral skeleton (either
cause sediments of the type considered here are
the cement bonds or the component grains) is
virtually inelastic, unloading does not recover
exceeded. When this occurs the sedimentary rock
the pore volume lost during loading. Figures
yields (Figure 2.14), and its compressibility in-
2.12 and 2.1~ both illustrate a typical unloading
creases. At yield, strong bonding resistance is
path, sometImes referred to as an elastic swell
progressively destroyed by fracturing of grains or
line (e.g. Figure 7.4), which reveals the limited
mtergranular bonds, and the source of the effec-
elas~icity of ~articulate materials. The only ex-
tive stress must progressively change to become
ceptIOn to thIS general behaviour is where the
located in intergranular friction and/or the cohe-
sediment contains a high concentration of swell-
sion due to weak chemical bonding. During yield
i~g clays. However, as bonding between the par-
tIcles grows and the sediment increases in degree
of lithification, the range of elastic behaviour
increases (Chapter 6).

Figure 2.14 Stress-strain curve for an oedometric
experiment conducted on a bonded but porous
Log mean effective stress, In p'
sedimentary rock. This curve shows the linearly elastic
Figure 2.13 The data in Figure 2.12 replotted in the response (initial steep section) at low stress and the
~tan~ard void ~atio (e)-natural log p' diagram widely used ~ost-yi~ld compaction behaviour with the strength
In sOil mechanics. See also Figures 6.2 and 7.4. increasing as the sediment densifies.
Mechanics of particulate media 47

the rock is effectively changing from a strained

material of the type illustrated schematically in
Figure 2.5a to a strained material of the type
illustrated in Figure 2.5b. This is a change of
state. The weakened sediment is unable to sup-
port the effective stress and must either harden Strain Strain Strain
(through compaction) or reduce the effective
stress. Generally, the strain energy released by Figure 2.16 Stress-strain, pore-pressure-strain and
the change in state is transferred to the pore compressibility-strain curves for a cemented porous
fluid as an excess pore pressure, which serves to sediment that yields. These curves show how the
support the porous fabric of the sediment by compressibility increases to a maximum at yield. This
causes the increase in pore pressure because the
reducing the effective stress it must sustain. This effective stress cannot be instantaneously maintained.
transient pore pressure then dissipates and the Dissipation of the pore pressure is accompanied by
sediment consolidates until it is sufficiently den- compaction of the pore volume and densification,
sified to again support the effective stress leading to a recovery of low compressibility. The
effective stress increases as the sediment densifies.
present at yield. The process is shown schemati-
cally in Figure 2.15 and in the stress-strain,
compressibility-strain and pore-pressure-strain
curves in Figure 2.16. After yield, porosity
changes with the effective stress in a manner
similar to the equivalent unlithified sediment
(Figure 2.17).


Figure 2.17 The relationship between porosity and

mean effective stress for a porous sedimentary rock that
yields. The substantial loss of porosity following yield
is due to the densification necessary to sustain the
effective stress state (Figures 2.15 and 2.16).

Figure 2.15 Mechanistic interpretation of the change in

state that accompanies yield in a cemented porous The term Cm/C. is very small in unlithified sedi-
sediment. Pre-yield strain energy increases in the ments, where the effective stress is controlled by
intergranular bonds, which behave like springs (this
ri and/or r m' and equation (2.7) is valid. How-
energy is stored in the interatomic bonds (Figure 2.5a)). At
yield the bonds fail, but the energy in the system must ever, in fully lithified sediments this term may
remain unchanged, so a transient increase in pore approach unity. Destruction of cement bonding
pressure occurs as the instantaneous response. This pore in a sediment at yield causes [1-(Cm/C s )]PH 2 o
pressure then dissipates (if the rock is sufficiently to approach P H20 causing an excess pore
permeable) and the grains compact (Figure 2.5b).
fluid pressure with a maximum value P H20 -
{[I-(Cm/C.)] P H20 }pre-yield' The magnitudes of
To see the effect of equation (2.11), it is necess- the greatest effective stresses sustained by a
ary to rewrite equation (2.7) in a more precise lithified sediment are therefore dependent
manner in order to consider the compressibility upon the nature of the deformation resistance
of both the mineral particles (C m) and the min- mobilized by the mineral skeleton, the ability of
eral skeleton (C.) of the sediment (Bishop 1976): the sediment to sustain an excess pore fluid
pressure (which depends on permeability of the
(1'=(1-(1- Cm)p
(2.12) sediment and adjacent rocks), and the volumet-
ric strain.
48 Mechanical principles of sediment deformation

A thorough familiarity with the effective stress Inflection point on each curve is
principle is essential to an understanding of the the yield point
deformation of fluid-saturated sediments. In un-
lithified sediments, unique relationships exist be-
---- \ Common post-yield compaction
tween the maximum sustainable pore volume and _ _ \ / trend irrespective of initial
the maximum sustainable effective stress. Put , \ porosity
simply, compacted pore-volume states can be ~
sustained at low effective stresses, but larger than ~~
equilibrium pore volumes are never sust~ined ~t ~===-~-~~
any maximum effective stress. If the sedIment IS -",~~.
on its normal consolidation line (Figure 2.1b), -
....:::::::-.......... -
then any increase in effective stress is al~a~s p'
accompanied by a decrease in pore volume. SImI- Figure 2.19 The effect of pore-infilling cement (without
larly, the pore volume can decrease only if the other fabric change) on the post-yield behaviour of a
effective stress is increased.
If the sediment has become lithified, the situ-
mally represents an equilibrium condition (Fig-
ation is more complex. Strong intergranular
ure 2.1 b). This generally has been regarded. as
bonds and pore-filling cements allow a range of
independent of the rate of change of effectIve
pore volumes to be sustained by ~he. sedime~t stress and of the duration over which the par-
prior to yield. Often quite porous, hthlfied, sedI- ticular effective stress state is maintained (Atkin-
mentary rocks will sustain large pore volu~es son and Bransby 1978). Change in effective stress
until the effective stress increases to a sufficIent
is not, however, rate insensitive. Increase in effec-
degree to promote yield. At yield, such rocks
tive stress occurs either because of increase in the
undergo a significant reduction in strength. Either
total stress acting upon the sediment, or because
the effective stresses or the pore volume - or, most of a decrease in the pore fluid pressure (equation
typically, both - must evolve rapidly to maintain
2.7). Generally, in geological situations, it. is pro-
mechanical equilibrium. After yield these rocks gressive growth of the total stress due to mcreas-
behave in a manner similar to the equivalent ing burial depth or tectonic processes that causes
particulate material. If the porosity of such sedi- effective stresses to increase. The total stress acts
ments has been reduced by pore-filling cements, on both the mineral skeleton and the pore fluids,
then the yield stress is progressively increased but the instantaneous response of a saturated
(Figure 2.18), although if the grain .fabric is not sediment to an increase in total stress is an
appreciably altered by the cementatIOn the post-
increase in the pore fluid pressure. This incr~ased
yield behaviour remains similar (Figure 2.19). pressure causes fluids to flow fron: the sedIment
and its pore volume to progreSSIvely decrease,
2.2.4 Primary consolidation and creep whilst the effective stress increases until the equi-
librium state of consolidation is re-established.
The progressive decrease in pore volume that The development of a new equilibrium effective-
accompanies an increase in effective stress nor- stress condition is thus dependent upon the per-
meability of the sediment, the permeability of
the drainage path to a low-potential drainage
boundary (such as the earth's surface) and the
length of the drainage path. Changes in effective
stress and the associated volumetric strains are
therefore sensitive to, and dependent upon, the
rate of pore-pressure dissipation. This process is
primary consolidation and is illu~t~ated sche-
matically in Figure 2.20. The ongmal forma~
Porosity description of this phenomenon was by Terzaghl
Figure 2.18 The relationship between yield stress and (1936), and it has subsequently been widely in-
porosity for a sediment showing progressive infilling vestigated in soil mechanics (Atkinson and
of its pore space with a mineral cement. Bransby 1978; Lambe and Whitman 1979).
Mechanics of particulate media 49

Initial water

Water under

In spring

In water
No flow of water ~--------~=====----- ~

Spring compresses
Water pressure decreases

Figure 2.20 Schematic representation of consolidation. In this diagram the sediment is represented by a sealed container
filled with water (the pore fluid) and a spring (the mineral skeleton). A force applied to the movable lid of the container
causes a reaction in the pore fluid, and the force is resisted by the generation of pore pressure. When a valve in the lid
of the container is opened (representing the permeability of the rOCk), fluid flows from the container and the pressure
decreases while the spring compresses. The stress in the spring increases through time (the effective stress) while the
pore pressure decreases. Eventually the applied force is balanced entirely by the stress in the spring and the system
is in equilibrium. The 'sediment' is now fully consolidated for the applied conditions.

Terzaghi identified that the magnitude of the For a single layer draining in one direction,
excess component of the pore pressure (PH20J normal to its surface, the pore-pressure distribu-
caused by an instantaneous increase in total tion with time follows the form illustrated in
stress from time, to, to any time, t (t greater than Figure 2.21, and compaction proceeds at a rate
to) and depth (z) in a sediment layer draining to determined by the associated progressive changes
its upper surface, is described by the equation: in effective stress. The sediment is described as
fully consolidated (for the in situ stress) when all
d 2 P H20. dP H20. excess pore pressures are dissipated. The ap-
cv~=~. (2.13)
proach to this condition is seen as a progressive
Cv is termed the coefficient of consolidation and is reduction in layer thickness as the volumetric
given by strain increases (Figure 2.22).
Consolidation accompanies the deposition
K and burial of all clastic sediments owing to the
Cv = C UWw (2.14)

where K is the hydraulic conductivity of the Pore pressure

sediment, C is its compressibility and UWw is the
unit weight of the pore fluid. Solutions to equa-
tion (2.13) of the form P H20 • = PH20.(z, t) provide
the variation of excess pore pressure with time in
the consolidating layer, from which the variation
in effective stress can be determined. Equation
(2.13) requires that the total stress remains con-
stant with time. During sediment burial, and/or
tectonism, this is unlikely to be the case. Under
these circumstances the consolidation equation
takes the form given below:
Figure 2.21 Pore pressures in a single layer of
(2.15) consolidating clay at different depths (z) for times (t=O,
t= t1, t= t2, and t= infinity) after the onset of consolidation.
50 Mechanical principles of sediment deformation

c Moreo.ver, Burland. (1990) reports that naturally

E consohdated matenals are able to sustain larger
~ pore volumes than materials consolidated in the
'0 laboratory. This may be a rate effect, or it may be a
gj consequence of natural hardening of the mineral
~ skeleton due to chemical interactions between
~ ~-----------------------
~ains. and ce.mentation of the material. As empha-
slZed m sectIOn 1.1.1, the effect of diagenesis on
Figure 2.22 Change in thickness of the consolidating these mechanical changes has been little explored.
layer (Figure 2.21) with time. It h~s been the ~ommon experience of experi-
mentahsts that sedIments and sedimentary rocks
progressive ~ncrease i.n vertical (gravitational)
stress as bunal depth mcreases (Figure 2.10). In can accumulate additional volume strain even
sands and other highly permeable sediments the after the sediment is fully consolidated to the
time required for full consolidation is extre:nely applied stresses (Atkinson and Bransby 1978; de
Waal 1986; Johnson, Rhett and Seimers 1989'
short and burial is not accompanied by the de-
velopment of excess pore fluid pressures, unless Rh~tt 19~~; Anderse~, ~oged and Pedersen 1992):
ThiS additional stram IS variously referred to as
the sand becomes sealed by overlying sediments of
secondary consolidation or creep. Bishop and
low permeability. In clays and other sediments of
Lovebury (1969) undertook experiments on re-
low. permeability, the rate of pore-pressure dissi-
moulded London clay which revealed that creep
pation may be slower than the rate of increase in
was still continuing, although at an extremely
vertical stress due to burial. These sediments (and
also more permeable sediments sealed by low- slow rate, 3 years after primary consolidation was
completed. Compaction experiments on chalk
permeability layers) will become progressively
(Johnson, Rhett and Seimers 1989; Rhett 1990;
overpressured as they are buried, and the resulting
Andersen, Foged and Pedersen 1992) have re-
exce~s pore p~essures may be sustained over ap-
vealed that creep continues for periods of the
preciable penods, both during and following
ord~~ of 1 week to 2 months after the sample is
burial.. T~s mechanis~ of overpressure develop-
eqUlhbrated to the applied stresses. The mechan-
ment is Widely recogmzed by geologists (Yassir
isms of secondary consolidation or creep are
1989a) but is far from being the only cause of
uncertain and probably diverse. Possible can-
excess pore fluid pressures in sediments (section
didates include: grain surface diffusion; time-de-
pendent crack generation associated with a redis-
C~nsolidation of a sediment under gravitational
tr~bution of stored strain energy; and diffusion in
loa?mg has been described as self-weight consoli-
microfractur~s, with stress corrosion weakening
datlOn and has been investigated experimentally
the fracture bps. Note that the term creep in this
by Been and Sills (1981). They have concentrated
usage differs from that in other geological con-
~n the very low stress (large pore volume) condi-
texts (e.g. section 5.7; see section 1.2.4).
tions that accompany initial deposition and have
~us, although consolidation is a widely rec-
observed the changes in effective stress as slurries
ogmzed and well-understood process in geotech-
deposit sediment particles under gravitational
nical engineering, its application to natural sedi-
loading. The consolidation of clays and silts at
ment burial needs careful consideration.
somewhat larger stresses has been widely studied
Although natural sediments clearly experience
in soil mechanics (Gibson 1958; Gibson, England
consolidation as they are buried, the natural
and Husey 1967; McClelland 1967; Skempton
process is complicated by:
1970; Burland 1990). However, consolidation un-
?er the larger stresses more typically encountered 1. Chemical interactions between grains, which
m ~eology, SUC? as may exist in sedimentary may strengthen the sediment and reduce the
basms, has received less attention (Yassir 1989; equilibrium volumetric strain;
Charlez and Heugas 1991; Leddra, Petley and 2. long-term creep effects, which may contribute
Jones ~992). In .investigations of natural geological appreciable additional volumetric strains over
consohdatIon, it should be noted that experimen- long periods;
tal.observations may not be completely represen- 3. other sources of pore pressure (see section
tative of consolidation in nature (section 6.2.3). 2.3.4).
Mechanics of particulate media 51

The results of these complications are manifest shear stress and strain. The present section turns
when real sediments and sedimentary rocks are to a consideration of these fundamental aspects
examined. In many cases, consolidation fabrics of deformation. Almost all natural deformations
are overprinted by the effects of pressure sol- involve some component of shear.
ution, whereas in other examples extremely high The magnitude of the shear stress sustained by
porosities are preserved at great burial depths, a sediment or sedimentary rock is proportional
either as a result of strong grain bonding, or to the difference (q) between the greatest (O"d and
through the presence of high pore fluid pressures. least principal effective stresses (0'3) (section
Although laboratory data from geotechnical stu- 2.2.2). This difference is a stress invariant referred
dies indicate that even very thick (2-3 km) sedi- to variously as the differential or deviatoric stress
mentary sequences should be fully consolidated (equation 2.16). It is proportional to the diameter
within 5 million years, in nature, high-porosity, of the Mohr stress circle and is therefore, at the
highly overpressured formations are encountered specified stress conditions, a direct measure of the
in sediments that have been buried to these shear stress that will develop in favourable orien-
depths for 50 million years or more (van den tations within the material (Atkinson and
Bark and Thomas 1980; Chapman 1983). Thus Bransby 1978). Only when the material is subject
the evolution of effective stress conditions in to purely isotropic stress (a state that rarely exists
basins and other large accumulations of sedi- in nature) will the deviatoric stress be zero.
ments is far more complicated than indicated by
simple gravitational consolidation models for q= O"~ -0"3 (2.16)
burial and compaction (M.E. Jones and Addis As the deviatoric stress is the difference be-
1984, 1986). tween two effective stresses, and as it is directly
The topic of consolidation history is consider- proportional to the shear stress, it follows that
ed further in sections 2.3 and 6.2. However, it is the magnitude of the shear stress is independent
worth reiterating the statement made in the pre- of the magnitude of the pore fluid pressure.
vious section that sediments which become over- However, shear-stress magnitudes are limited by
pressured will be subject to small effective stresses the shear strength of the sediment or sedimentary
irrespective of the total stress state. The behav- rock. Materials can never mobilize or sustain
iour of such sediments will depend upon the shear stresses that exceed their shear strengths
stress history but, in the case where small effec- because shear failure will serve to maintain the
tive stresses have been maintained throughout shear stress at or below the failure condition.
the history of the sediment, the strength (in the This has been demonstrated dramatically in a
absence of diagenetic cements) will remain very number of laboratory studies on weak, saturated
low. Consolidation state, in the absence of inter- sediments and leads to the important concept of
granular bonding, has a crucial influence on the critical state.
strength. Laboratory studies on weak sediments, such as
remoulded London Clay (Schofield and Wroth
1968), remoulded Gault Clay (Parry 1960) and
2.2.5 Shear stress and deformation mud volcano clays (Yassir 1989a), reveal that at
any given mean effective stress (p') (equation
Principles (2.17) for the case where 0'2 = 0' 3), shear failure of
Isotropic consolidation leads to reduction in bulk the material occurs at constant deviatoric stress
volume of a sediment but not to deformation in (q), irrespective of the strain. At this stress condi-
the sense of a change of shape (see section 1.2.2). tion, the sediment is theoretically able to deform
Although volumetric strains are very common in at constant pore volume to infinite strain, or, in
nature, they are by no means ubiquitous, and the case of real systems, until geometrical
they do not necessarily accompany changes in changes force a change in the stress system or
shape. Many of the common expressions of sedi- stress magnitude. This condition is termed the
ment deformation (such as faults, landslips, dia- critical state. The stress-strain curves of mud-
pirs and slumps) occur with little or no volumet- volcano clay samples deforming in this manner
ric strain. These changes in the shape of the are shown in Figure 2.23. These curves illustrate
sediment body are due to the mobilization of the accumulation of strain at constant stress. The
52 Mechanical principles of sediment deformation

50.0 mon practice in soil mechanics, is to use a dia-

Consolidation stresses
gram in which the deviatoric stress is plotted
m 40.0
against the mean effective stress (Figure 2.25a).
Plots of this type are often referred to as stress
path diagrams, because they trace the stress his-
tory of the sediment in terms of both the effective
~ __________________ 23MPa and shear stresses it sustains. A stress path in p'-q
space is directly equivalent to a family of Mohr's

(5 0.0 ~::::;:=::,:=~5M~pa~_..----'-_--.-_-r----'
stress circles except that each circle is represented
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 by a single point with the coordinates p' and q.
Axial strain (%) The quantity p' defines the position of the circle,
Figure 2.23 Stress-strain curves for a mud-volcano and q its diameter (Farmer 1983). (See Atkinson
clay subject to undrained shear deformation following
consolidation to the isotropic consolidation stresses 40
indicated against each curve. (After Yassir 1989a.) m
experiments were conducted such that the state 6<J)
of volumetric strain was held constant during ~
1;) 20
shear deformation. This conforms to ideal critical u
state behaviour (Atkinson and Bransby 1978). .~

The sediment is behaving as a perfectly plastic .:; 10
material in the sense that it is undergoing a 0

permanent change in shape without fracturing

and without hardening or softening. While at this (a)
critical state condition, any attempt to increase
the deviatoric stress supported by the material
simply results in an increase in strain rate. 0.56
, U'l +2u3 .Q
p= 0.28
3 1J
·0 Isotropic consolidation line
::> 0.14 CSL
If identical samples of a saturated sediment
that is capable of exhibiting critical state behav- 0~---1~0--~2~0--~3~0---4~0~~5~0--~6~0-­
iour are consolidated to different mean effective (b) Mean effective stress (MPa)
stresses (p'), the deviatoric stress (q) required to
induce shear failure at the critical state is ob-
served to increase with the consolidation stress.
In other words, consolidation to greater effective
stresses serves to increase the shear strength of
the sediment. The results of such experiments can Drained triaxial loading
be represented by their stress-strain curves (Fig- /
ure 2.23) or as a family of Mohr's circles descri- /
bing the stress states at failure (Figure 2.24), but / /Undrained
the most convenient representation, and the com- / triaxial Isotropic consolidation
// loading

(c) Mean effective stress

Figure 2.25 (a) The stress paths in deviatoric-

stress-mean-effective-stress space for the experiments

represented in Figure 2.23, and the critical-state line
(CSL) defined by those experiments. (b) Pore-volume
states at failure as a function of the mean effective stress.
10 20 30 40 50 60 70 80 90 100 (c) Representation of the stress paths followed during (i)
isotropic consolidation followed by drained triaxial loading
Figure 2.24. Mohr's circles at failure for the experiments and (ii) isotropic consolidation followed by undrained
represented in Figure 2.23, showing the failure envelope. triaxial loading.
Mechanics of particulate media 53

and Bransby (1978) for a comprehensive dis- causative forces dissipate or are relieved by the
cussion of this representation of effective stress deformation, or until changes in geometry or
history.) Stress path diagrams form the basis of deformation mechanism promote a change in
the approach adopted in Chapter 6 for analysing either the stress state or the sediment's strength.
the tectonic deformation of sediments. Inspection of the critical state line in Figure
The shear strength of a sediment can be deter- 2.25a illustrates how, at low consolidation stress
mined by a series of laboratory experiments, each (low effective stress), sediments are prone to
conducted at a different consolidation stress (the large-scale shear deformation. Obvious examples
effective confining pressure), in which the devi- of this type of behaviour in nature include the
atoric stress is increased until failure occurs. The failure surfaces developed at shallow depths in
results of these experiments can be represented landslips and syndepositional faults, the mass
by the family of Mohr's circles that represent the movement of unconsolidated and lightly con-
stress states at shear failure, and a line can be solidated materials in flows, slides and slumps.
constructed that is tangential to and enveloping Under larger effective stresses, greater deviatoric
these circles and which defines the failure envel- stresses are necessary to promote shear failure at
ope (Figure 2.24). The sediment cannot sustain the critical state, but deformation under these
stress states that lie outside this envelope. Alter- conditions may take place in some fault gouges
natively, the same experimental results can be and account for the mobilization of shales to
represented by a line in the stress-path diagram. form diapiric structures.
This is the locus of the ends of all stress paths on Similarly, anomalously high pore pressures
which the sediment attained the critical state will allow low effective stresses to be maintained,
(Figure 2.25a). This (normally) straight line which so that deformation at the critical state under low
connects the ends of the individual stress paths at consolidation stress may occur even in environ-
the critical state is referred to as the critical state ments where the total stresses are large. Thus,
line. The critical state line is a representation of although the magnitude of the deviatoric stress is
the shear-failure criterion for the sediment. independent of the pore fluid pressure, the mag-
Strictly, the critical state concept applies only nitude of q at which a sediment is able to deform
to homogeneous sediments that behave in a per- plastically to large strains is strongly dependent
fectly plastic manner. This is rarely true of nat- upon the mean effective stress. Low shear
ural materials. Most natural sediments exhibit an strengths will characterize sediments that have
initial shear strength that lies above the critical sustained large excess pore pressures throughout
state (at least at low consolidation stresses). This burial, irrespective of the burial depth. Converse-
is generally referred to as the peak strength. Once ly, sediments that develop excess pore pressures
the peak strength has been exceeded, the devi- owing to some secondary mechanism of pore-
atoric stress sustained by the sediment is seen to pressure generation may fail at the critical state if
decay until the critical state is attained. For many the value of the deviatoric stress remains un-
real sediments the critical state corresponds to a changed while the mean effective stress is reduced
residual strength condition (Farmer 1983; by the increase in pore pressure.
Leddra, Petley and Jones 1992). The stress path diagram (Figure 2.25a) allows
Sediments can only normally sustain stress the relationship between shear stress and mean
states that lie below the critical state line. Any effective stress to be illustrated along with the
attempt to increase the effective stresses acting conditions for shear failure of the sediment (the
upon the sediment to shear-stress states outside critical state line). It is also the case that increase
of the critical state simply results in increased in mean effective stress leads to a reduction in
strain. The critical state line for any sediment is pore volume (see sections 2.2.3 and 2.2.4) and
thus part of a state boundary surface, dividing hence the critical state line can also be represen-
stress-space into attainable and unattainable re- ted on a plot of pore volume against mean
gions. It illustrates how the maximum deviatoric effective stress (Figure 2.25b). This is equivalent
stress that can be sustained by a sediment is to Figure 2.12, except that in this case the re-
limited by its physical characteristics and consoli- corded pore-volume states are those sustained
dation state. Sediments subject to stress states at during shear failure of the sediment. This
the critical state will deform until either the curve therefore represents the locus of end-point
54 Mechanical principles of sediment deformation

pore-volume states for samples consolidated to Applied vertical

different mean effective stresses and subject to
deviatoric stresses corresponding to the critical
state. Figure 2.25b represents an important rela-
tionship, as it illustrates the mechanism (pore-
volume reduction) by which consolidation to
greater effective stress increases the shear
strength of a sediment, thereby reducing the
likelihood of shear failure.
In Figure 2.25c, two paths are shown linking
the end point of consolidation with the critical Constant
state line. One of these shows an increase in mean confining ~~~~~~~~
effective stress with increasing deviatoric stress
and is a straight line; the other is a curve showing (b)
decreasing mean effective stress. The former cor-
responds to the case where loading to failure is
accompanied by full dissipation of any excess
pore pressures, so that the deformation can be
described as drained. Typically this path is fol- I Vertical effective
lowed in a laboratory triaxial experiment (Figure ~=====:::::::::;
2.26a) where the sample is consolidated under an If)
Pore fluid pressure
Applied vertical ~


Axial strain (%)

Volume Figure 2.26 Contd.

Increase in isotropic effective stress (O"t = 0'3) at low pore

displaced pore pressure. The consolidation stress (cell or confin-
fluid ing pressure, 0'3) is then held constant while the
pressu;~re~~~~~~~~M3¥l. axial stress (O"d is increased. The rate of deforma-
Valve open tion is adjusted until pore-pressure dissipation
(which depends on the coefficient of consolidation
(Cv ) (equations 2.13 and 2.14)) equals or exceeds
the rate of pore-pressure generation. Any pore
pressures associated with the consolidation that
_ - - - - - - - - Vertical effective accompanies axial deformation are thus fully
stress dissipated. The second curve corresponds to the
situation where drainage of pore fluids is com-
pletely prevented during the application of the
deviatoric stress - undrained deformation. Here,
axial deformation of the sample is not accom-
panied by continued consolidation (Figure 2.26b).
Pore fluid In this experiment, pore pressure increases with
If-------------pressure the deviatoric stress. The result is that the mean
Axial strain (%) effective stress is reduced, the effective stress being
Figure 2.26 (a) Schematic representation of a drained
dependent upon the pore pressure, while the de-
triaxial experiment. (b) Schematic representation of viatoric stress increases as the axial stress increases
an undrained triaxial experiment. until shear failure occurs at the critical state.
Mechanics of particulate media 55

Figure 25 (a and b) represents projections of a surface and deformation on this surface may
three-dimensional surface on to a pair of the involve components of both volumetric and shear
three axes: mean effective stress, deviatoric stress strain (Loe, Leddra and Jones 1992). Only when
and pore volume (void ratio). The surface is the sediment has been unloaded (either by in-
shown in three-dimensions in Figure 2.27. It is crease in pore pressure or through reduction of
referred to as a yield surface (Atkinson and the total stress) can paths within the internal
Bransby 1978; M.E. Jones, Leddra and Potts volume of this figure, or on to the Hvorslev
1990; M.E. Jones et al. 1992), and is seen to surface, be followed. These states are known as
consist of two complex curving surfaces that overconsolidated (section 2.2.7) and contrast with
intersect at the critical state line. These surfaces the normally consolidated states that define the
are known in soil mechanics as the Roscoe and Roscoe surface. The figure defines the only
Hvorslev surfaces (Atkinson and Bransby 1978). stress-pore-volume states in which the sediment
They are geometrically and, to a degree, physical- can exist and its surface is therefore a state
ly equivalent to a Mohr-Coulomb failure surface boundary surface for the sediment.
for porous rocks (M.E. Jones, Leddra and Addis The critical state concept and the yield surface
1987; M.E. Jones and Leddra 1989; Abdulraheen, figure provide a framework within which all
Roegiers and Zaman 1992; Loe, Leddra and aspects of the deformation of porous sediments
Jones 1992). can be described and interrelated effectively. The
Roscoe surface links the normal-consolidation
Roscoe surface Deviatoric line for the sediment (which plots on the zero
Critical state line stress deviatoric stress plane), where strain is entirely
volumetric, with the critical state line, where
deformations are entirely through shear. Unload-
ing is not normally accompanied by significant
recovery of the volumetric strain (but see section
6.2.4), so unloading paths leave the Roscoe sur-
face to proceed to lower mean effective and
Drained triaxial
deviatoric stresses without expansion of the pore
volume. These 'overconsolidated' states also in-
fluence reloading. During reloading of an over-
consolidated sediment, little change occurs in the
pore volume until the previous maximum stress
Void ratio state is attained. At this condition the stress path
will have arrived at a point on the Roscoe sur-
Figure 2.27 The state boundary surface for a particulate face, or will have attained the critical state. These
sediment (after Atkinson and Bransby 1978) showing the
Roscoe surface (where normal consolidation occurs),
relatively simple concepts of consolidation and
the critical-state line and the Hvorslev surface. The shear, and the relationships between them, are
paths followed in a drained and an undrained triaxial quantitatively valid for clays and fine carbonates,
experiment starting from the same stress-pore-volume and for sands where grain size and angularity do
condition are illustrated. Paths that track into the interior
of this figure, or those that reach the Hvorslev surface,
not cause dilation during shear deformation. For
are associated with overconsolidation. coarse, angular grained, clastic sediments, the
critical state concept must be modified to allow
The yield-surface diagram (Figure 2.27) pro- for the hardening that accompanies dilation dur-
vides the means to represent the potentially com- ing shear. When such materials maintain high
plex deformation path that a sediment may fol- porosities, increases in pore volume are minimal,
low during its geological history, in terms of but as the structure becomes more densely
progressive effective stress, deviatoric stress and packed, the effects of dilation become more pro-
pore-volume states. Deformation at the critical nounced. None the less, the yield surface figure
state has already been described and represents still preserves the geometry shown in Figure 2.27,
shear deformation that proceeds at constant although the deformation paths for dilatant ma-
deviatoric stress. All normally consolidated sedi- terials will tend to be more complex. This ap-
ments deform along paths that lie on the Roscoe proach to describing sediment deformation thus
56 Mechanical principles of sediment deformation

provides a unified model for material behaviour cemented to particulate material as the rock
from the moment of sediment deposition, yields. The following interpretation is offered as
through repeated cycles of burial and uplift, until an explanation of this variable behaviour, and is
diagenetic cementation renders the sediment a believed to be a more realistic model for the
lithified rock. deformation behaviour of most natural sediments
(Leddra, Petley and Jones 1992) than the
Experimental deformation of a real sediment idealized critical state model described above.
Shear deformation of many real sediments is Nevertheless, it is the critical state model that
complicated by the elastic stiffness caused by provides the framework for this interpretation.
intergranular bonding at low stresses. This aspect The stress-strain curves for the chalk experi-
of material behaviour has received attention only ments that produced the stress paths in Figure
in recent years, but studies have shown a consist- 2.28 are shown in Figure 2.29. The curves have
ent style of deformation behaviour irrespective of been divided into three sets according to the
the sediment type or strength. Such investiga- shear-failure behaviour exhibited by the sample.
tions include those of: Vaughan (1985), Vaughan, For samples consolidated under low mean effec-
Maccarini and Mokhtar (1988), Addis (1989), tive stresses, the stress paths pass beyond the
M.E. Jones and Leddra (1989), Leddra (1990), critical state line, and then return to it before
Leddra, Goldsmith and Jones (1990), Leddra and reaching the condition where deformation can
Jones (1990), M.E. Jones, Leddra and Potts proceed to large strains at nominally constant
(1990), Charlez and Heugas (1991), Steiger and stress. After deformation, these samples always
Leung (1991), Leddra, Petley and Jones (1992) show evidence of brittleness, with fractures con-
and Loe, Leddra and Jones (1992). A family of taining disrupted coccolith material separating
stress paths recorded from deformation experi- areas of undeformed rock (Figure 2.30a). The
ments on a weak but bonded sedimentary rock paths followed by samples consolidated to inter-
(chalk) is illustrated in Figure 2.28. The curves mediate mean effective stresses reach the critical
show a marked change in behaviour with increas- state line, but then show strain softening, follow-
ing consolidation stress, which is more complex ing the critical state line towards the origin before
than the ideal critical state behaviour described the critical state condition of deformation at
above. The result is typical of bonded mudrocks, constant stress is attained. These samples are
carbonates and sands, and can be explained in rather more pervasively deformed than those
terms of the effects of stiffness due to intergranu- illustrated in Figure 2.30a, but still contain dis-
lar cementation, and the change of state from crete failure surfaces along which strains became
localized (Figure 2.30b). The third group of
samples shows strain hardening at the critical
state. Deformation is pervasive with few, or no,
obvious failure surfaces and severe disruption of
the coccolith fabric occurs throughout the sample
(f) (Figure 2.30c), which becomes barrel shaped.
The three different responses during loading to
<..l shear failure exhibited by these chalk samples are
820.0 a consequence of the stress conditions pertaining
.~ when the bonding responsible for the elasticity
010.0 exhibited by this rock underwent yield. Under
low consolidation stresses, the volumetric strains
O.O-t-'------,"-----+---.L__,L----I---~-- are small, a consequence of the elastic stiffness
0.0 10.0 20.0 30.0 40.0 50.0 60.0 imparted by the intergranular bonding. Distor-
Mean effective stress (MPa) tional strain associated with the increasing shear
Figure 2.28 Stress paths recorded from a family of stress as the samples approached failure was also
undrained triaxial experiments on chalk samples with a largely resisted by the bonding stiffness of the
similar pre-test porosity, which were consolidated to
different mean effective stresses prior to undrained
fabric. The sample thus reaches the critical state
loading. The origin of each curve indicates the in an intact, bonded condition. The critical state
consol idation stress. defines a failure characteristic of disaggregated
Mechanics of particulate media 57

Because intact chalk is stronger than dis-

8 aggregated chalk, the Mohr-Coulomb envelope
lies above the critical state line, in the low con-
solidation stress regime. Loading to the Mohr-
Coulomb envelope necessitates the rock storing
appreciable strain energy in elastic distortion of
the intergranular bonds (refer to Figure 2.5a). At
brittle failure, this energy is dissipated on the
fracture surface, promoting disaggregation of ad-
jacent coccoliths and sliding. A chalk gouge
rapidly develops on the fracture surfaces. This
O~-----r-----.------'-----' consists of disaggregated chalk, which is unable
o 5 10 15 20 to support stresses beyond the critical state con-
Axial strain (%) dition. As this gouge forms it dominates the
20 stress-strain characteristics of the entire sample,
and the stress state evolves towards the critical
state. Shear strain becomes strongly localized
£' 15 about the shear failure surfaces, and the intact

~ sections of the sample are effectively unloaded.
They remain in the bonded state and respond in
11; 10 a passive manner during subsequent deforma-
(.) stress tion. The large strain behaviour that is seen in the
"g ~--
.s: stress-strain curves (Figure 2.29a) is entirely due
0 5 to frictional sliding in the disaggregated material
within the failed zones.
The behaviour exhibited by this chalk is a
0 typical response of a rock deforming by sliding
0 5 10 15 20 on a weak fracture surface. The initial failure at
(b) Axial strain (%)
the Mohr-Coulomb envelope can be equated to
the peak strength, whereas the steady-state con-

[L 60
t Increasing
dition that occurs at large strains (the critical
state) is equivalent to the residual strength (see
section 1.2.4). Thus shear failure under low con-
If) solidation stress corresponds to a change in state,

u:; 40 with part of the rock material evolving from an

"g intact, bonded material to a particulate frictional
1;j material. This change of state is seen in the stress
20 path diagram and the stress-strain curve as the
transition from peak- to residual-strength condi-
0 The peak strength corresponds to the initial
0 5 10 15 20
disruption of the cemented fabric of the rock,
(c) Axial strain (%)
which has developed as a consequence of diagen-
Figure 2.29 Stress-strain curves for the samples of esis and compaction during its geological history.
Buster Hill chalk that were deformed in undrained On the time-scale of the laboratory experiment,
triaxial compression at (a) low consolidation stress, (b) or a natural deformation such as a rock fall or
intermediate consolidation stress and (c) high
consolidation stress. (After Leddra 1990.)
slump, the change is irreversible. Once the rock
has attained a residual strength (critical state)
chalk and is not a property of the intact material. condition, this will dominate its shear strength
The stress paths are thus able to pass beyond the until other (diagenetic?) processes impose further
critical state line, and shear failure occurs on the changes in state. For example, diagenetic cemen-
Mohr-Coulomb envelope for the intact chalk. tation or mineralization of the fractures would
58 Mechanical principles of sediment deformation

Figure 2.30 (a) Macroscopic appearance of Buster Hill chalk specimen that was deformed in undrained triaxial
deformation experiments at low consolidation pressures (confining pressures). (b) Macroscopic appearance of Busler
Hill chalk specimen that was deformed in undrained triaxial deformation experiments at intermediate consolidation
pressures (confining pressures). (c) Macroscopic appearance of Buster Hili chalk specimen that was deformed in
undrained triaxial deformation experiments at high consolidation pressures (confining pressures). (After leddra 1990.)

re-establish the intact (peak) strength of the sion of the mineral framework so that shear
material. loading proceeds in the presence of appreciable,
Laboratory experiments conducted on sand- stored, strain energy. As the deviatoric stress is
stones and cemented mudrocks (Leddra, Petley increased, the bonded coccolith framework
and Jones 1992; Petley et al. 1992) show that continues to resist distortional strain but the
these materials behave in a manner similar to the additional strain energy associated with this
chalk reported above. In each case, peak strength resistance promotes yield of the framework be-
pertains to the failure of the intact, bonded rock fore the critical state is attained. These chalk
fabric, and is followed by reduction in deviatoric samples reach the critical state line in a partially
stress as strain becomes localized into conjugate dis aggregated condition and are therefore un-
shears deforming at the critical state. Strain local- able to support deviatoric stresses that exceed
ization of this type is important in most natural the critical state. At the critical state, strain is
deformations. At the scale of a landslip, the localized initially in the failed regions of the
strength of the entire system becomes dominated rock and is accompanied by pronounced strain
by the shear-failure surface at its base. Activation softening. Unlike the samples deformed after
from the intact rock occurs at peak strength but consolidation at low stresses, this softening is
with increasing shear strain, displacements within not caused by a change in the failure condition
the slipped mass become controlled by the resid- but because the proportion of dis aggregated
ual strength of the failure surface. At a larger chalk increases with the shear strain. As more of
scale, in deltas and extensional basins, strain the material passes from intact to disaggregated
becomes localized on syndepositional faults. state, the deviatoric stress needed to sustain
Once established, the residual shear strength of shear failure at the critical state is reduced, and
these fault surfaces will control the magnitUde of the sample strain softens along the critical state
the deviatoric stresses throughout the basin. line. These samples show a progressive transi-
The chalk samples consolidated to intermedi- tion from a peak to residual condition, but in
ate pressures (Figures 2.28, 2.29b and 2.30b) were this case the transfer occurs entirely on the criti-
still in an intact, bonded, state at the onset of cal state line and is far less dramatic than ob-
shear loading. However, the consolidation stress served in samples consolidated under low
wiHhave caused a considerable elastic compres- stresses.
Mechanics of particulate media 59

In many respects the style of shear failure tion of the individual mineral grains. At the
exhibited by these samples, which underwent critical state, shear deformation is accom-
progressive disaggregation at the critical state, is modated by grain sliding and rotation, and is
the most dramatic expression of shear deforma- completely pervasive. The samples are seen to
tion. In the North Sea, chalks developing this harden as they deform, because the deformation
type of failure have been reported to flow into improves grain packing in the dis aggregated ma-
oil wells (Leddra and Jones 1990), almost to the terial, reducing the ability of the grains to slide.
total exclusion of hydrocarbons; high-porosity This hardening promotes further grain fracturing
chalks involved in rock falls mobilize a similar in chalks (Addis, personal communication). The
style of deformation and become debris flows, change to a stiffer arrangement of grains appears
which behave as if the chalk had undergone to be accompanied by slight dilation of the test
liquefaction. In laboratory experiments, weakly sample. The hardening is observed during load-
cemented, highly porous mudrocks behave in ing under undrained conditions and necessitates
an equivalent manner, and these are also known an increase in the mean effective stress. This
to invade oil wells in a manner akin to the requires that the increasing load on the sample is
North Sea chalk. Geologically, this type of preferentially supported by the mineral grains
shear failure is probably a less frequent occur- and not by the pore fluid. A proportion of the
rence than the strain-localized failures de- excess pore fluid pressure generated by the load
scribed above. However, the development of must therefore be dissipated. Dilation of the
flows and slides, the pervasive restructuring of sample due to distortional strain provides this
North Sea chalk during resedimentation (Ken- mechanism. The increase in effective consolida-
nedy 1986), and the progressive mobilization of tion stress that the hardening represents is an
mudrocks into diapiric structures, almost cer- attempt, on the part of the material, to maintain
tainly involve this particular type of strain-sof- its consolidation state, while deforming at the
tening behaviour. critical state.
In sandstones, the equivalent failure state is The response of these chalk samples to in-
less easy to observe because in laboratory experi- creased deviatoric stress, following appreciable
ments pervasive disaggregation renders the test compaction due to the high consolidation stress,
material an incohesive powder. If such failures illustrates the importance of consolidation in
occur in sandstones they may be associated with promoting sediment strength. Sediments that
phenomena such as sand invasion of oil wells and have come to equilibrium with large effective
the mobilization of weak sand failures in land- stresses have low porosities and are therefore
slips. The former is known to be a frequent stiff materials. Increased consolidation reduces
occurrence; the latter has been reported by the probability that shear failures will be dra-
Brunsden (personal communication) for the west- matic events. Conversely, the results presented
ward extension of the Black Yen landslip, Lyme above can be interpreted as indicating that
Regis, Dorset. sediments and sedimentary rocks that support
The third group of chalk samples (Figures 2.28, large pore volumes and large pore pressures are
2.29c and 2.30c) were consolidated to large those most vulnerable to large-scale shear defor-
stresses. The consolidation stresses exceeded their mation at low effective stress. Cement bonding
yield strengths (section 2.2.3), and as a conse- can facilitate the preservation of high porosi-
quence the samples experienced a permanent ties during consolidation, rendering weakly
volumetric strain prior to shear loading. These cemented porous sediments highly sensitive to
samples were therefore in a compacted, particu- deformation.
late state (although not necessarily completely In summary, shear deformation of sediments
dis aggregated) prior to the application of the is normally accompanied by large strains, and
deviatoric stress. Because of the high consolida- in many cases steady-state plasticity (the critical
tion stresses, the increase in deviatoric stress state) dominates behaviour. Existing fabrics
necessary to promote failure at the critical state is cause deviation from the ideal behaviour (com-
also large, so that loading to the critical state will pare, for example, the results of Yassir (1989b)
be accompanied by further deformations and, at with those of Leddra, Petley and Jones (1992) in
least in the case of chalk, some brittle deforma- Figure 2.31), whereas intergranular cementation
60 Mechanical principles of sediment deformation

2.2.6 Other stress paths and associated
36.0 strain states
~30.0 The strain states associated with consolidation
~ under isotropic stress and shear deformations at
~ 24.0 the critical state represent two extremes. The
.§ 18.0 former are entirely volumetric and occur in the
absence of shear stress whereas the latter are
~ 12.0 purely distortional. The stress paths associated
with the development of these different strain
6.0 states reflect the differences. Isotropic consolida-
0.0 +-~:='--.--..------r--'----::-J~-r--=-"I--=-,
tion is accompanied by increasing isotropic effec-
0.0 6.0 12.0 18.0 24.0 30.0 36.0 42.0 48.0 54.0 tive stress whereas the critical state represents
Mean effective stress (M Pal deformation in the presence of the maximum
shear stresses the material can support for its
consolidation state. Between these extremes lies a
continuous spectrum of possible stress-deforma-
tion paths, any of which can be followed by
sediments deforming in nature. It is inappropri-
ate here to examine the characteristics of all the
deformations that accompany loading along
these different paths. The reader is referred to
Atkinson and Bransby (1978) for a rigorous treat-
ment with regard to sediments, and to Loe,
Leddra and Jones (1992) for a detailed investiga-
tion of the relationship between stress path and
strain states for a weak, porous rock (chalk). At
Mean effective stress (MPa) the time of writing, equivalent data for other
(b) weak sedimentary rocks are limited.
Figure 2.31 The stress paths associated with undrained In the remainder of this section discussion will
shear deformation of (a) a remoulded clay (after Yassir be restricted to the one geologically important
1989a) and (b) a naturally, compacted, weakly stress path not considered above. This is the
cemented clay. The origin of each curve indicate the uniaxial strain or Ko condition. It is the path
consolidation stress (after Leddra, Petley and Jones
characterized by one-dimensional consolidation
with zero lateral deformation of the material. The
has a more pronounced effect, and may cause path lies on the Roscoe surface between the
catastrophic failures under certain conditions isotropic consolidation and the critical state lines
(Leddra and Jones 1990; Figure 2.28). Thus lith- and represents the condition where consolidation
ification progressively complicates the shear proceeds in conjunction with sufficient distor-
behaviour of sediments and sedimentary rocks tional strain to continuously recover any lateral
and will exert a major influence on the specific compaction (Figure 2.32). Thus a state of zero
expressions of shear strain, as for example along lateral strain is maintained.
a fault passing through rocks with different fab- The Ko stress path is of particular importance
rics and bonding strengths. Generally, shear de- because it is widely regarded as the path followed
formations are facilitated by large pore-fluid by most sediments undergoing burial in the pres-
pressures and large pore volumes, although ence of uniform gravitational loading. Under
grain size, shape and strength also exert a such circumstances, the sediment is constrained
profound influence. Shear deformations tend to to compact in a one-dimensional manner because
produce pervasive fabrics in materials that are all of the adjacent sediment is compacting in an
both weak and uncemented but exhibit strain equivalent manner at an equivalent rate. Lateral
localization and brittleness in intact, bonded strains are thus unable to develop and the stress
materials. state is forced to follow the Ko condition. Near to
Mechanics of particulate media 61

be a constant for most uncemented, normally

consolidated sediments.

The magnitude of the Ko ratio depends upon
the type of sediment. Generally, the finer grained
consolidation and more clay rich the material the larger the Ko
Ko path value. In sands, the ratio commonly has values
Critical state line around 0.3-0.4, whereas in clays it may be 0.7 or
greater. Silts lie between these extremes. This
Mean effective stress variation with grain size, and to some extent with
grain shape, leads to an obvious association
between Ko and the friction angle for the sedi-
Critical state line
ment. Typically, for uncemented, normally con-
solidated sediments the Ko ratio is found to vary
according to:
Ko= I-sin 4/, (2.19)
where cjJ' is the effective friction angle.
Normally, Ko is greater in overconsolidated
sediments (see below), and in clays Ko ratios may
Isotropic exceed 1.0, depending upon the degree of over-
Mean effective stress consolidation (Figure 2.34). This means that lat-
eral stresses may exceed vertical stresses in sedi-
Figure 2.32 (a) The uniaxial strain or Ko stress path in
relation to the isotropic consolidation and critical-state ments that have been subject to unloading during
lines. The inset sketches illustrate how bulk strain would uplift, for example as a consequence of erosion.
accumulate in a laboratory specimen deforming on the This occurs because although the vertical stress
Ko stress path. (b) The relationship between the Ko path decays as the overburden is removed, the lateral
and the critical-state line in the coordinates of the
stress-path diagram.
stress is to a degree 'locked' into the grain frame-
work of the sediment, so that although it does
the earth's surface, in all sedimentary basins decrease as the vertical stress decreases, it does so
where there is not appreciable surface relief, the less efficiently. This effect is most pronounced in
sediments are likely to be under stress condi- clays and shales, where unloading and associated
tions close to Ko. With increasing depth the overconsolidation may cause the Ko ratio to
presence of non-gravitational forces associated exceed 4.0, giving rise to large incremental
with basin extension or contraction will pro- changes in lateral stress for small changes in
gressively move the sediment away from this vertical stress. Such effects are even more pro-
stress path. nounced when heavily overconsolidated clays are
The quantity Ko is often referred to as the subject to reloading.
coefficient of earth pressure at rest (Brooker and The fact that the Ko ratio is dependent upon
Ireland 1965). It has important practical applica- the lithology of the sediment, its grain fabric, and
tions in soil mechanics in the design of excava- previous stress history, has important conse-
tions and retaining walls (Figure 2.33). Ko is quences for in situ stress states in uncemented
defined as the ratio between the horizontal (lat- sediments. Under gravitational loading, an alter-
eral) and vertical effective stresses (where such nating sequence of sands, silts and clays will
stresses are principal stresses, as is normally the develop a vertical stress that is a function of their
case close to the earth's surface) which maintain density and the burial depth. Vertical stress
the condition of no lateral deformation (equation therefore increases in a fairly regular manner
2.18). The ratio is normally determined in the with increasing depth, usually shown as a straight
laboratory in a one-dimensional compression ex- line (Figure 2.35). By way of contrast, the hori-
periment on the material of interest. It is found to zontal stress varies from bed to bed as the Ko
62 Mechanical principles of sediment deformation

Stress Stress a~ = pgz

PH 2 0 = pwgZ To resist lateral expansion
a~ = a v - P H20 of quarry wall,
ah = Ko a~ where
"""""""""""".,., ah 2-
E g K o =-, - 1 - sin
<I> ~
.r:: .r::
c.. c..
~ v
o 1 -v

crh = Kopgz + crT

, \ a v = pgz a v = pgz (Compressional
ah = Kopgz· crh= regime)
(Extensional regime)* z
(i) (ii)

(]~ aH


If Ko = 1 then ai = as
generally Ko - 0.2 (coarse sand)
--> 0.9 (clay)

Figure 2.33 Schematic representation of the quantity Ko (i) with respect to a deep excavation (ii) and the burial of an
element of sediment in a subsiding and filling sedimentary basin (iii). (cp=friction angle; v=Poisson's ratio)

ratio changes, so that the horizontal stress is The reasons for this are twofold. Firstly, the
strongly dependent upon the lithology. In addi- technology required to conduct reliable, routine,
tion, if effective stresses are considered, the pres- Ko experiments on stronger sedimentary rocks
ence of overpressured horizons will give rise to has been available only in recent years. Secondly,
further irregularities (Figure 1.4). Thus geological the number of occasions where the compaction
boundaries may also be important stress discon- behaviour of porous lithified rocks has been the
tinuities (Figure 2.35). These differences in lateral subject of engineering investigation has been few.
stress are of little consequence while the sedimen- The exception to this has been in the field of
tary succession is subject to one-dimensional hydrocarbon reservoir engineering. Reservoir en-
gravitational loading. With the development of a gineers have for a long time concerned them-
uniform horizontal component of stress due to selves with the isotropic compression of porous
the onset of tectonism, however, these same vari- sedimentary rocks, and in recent years this work
ations in lateral stress will play an important role has been widened to consider compression of
in the sequential development of shear failure, these materials under conditions of uniaxial
and in the initiation of deformation fabrics, strain.
folding and faulting. A further difficulty with investigating porous
The deformation of sedimentary rocks under rocks is that the Ko ratio changes with the state
Ko conditions has been studied less than the of strain. Under low stresses, when the majority
deformation of completely unlithified sediments. of the intergranular bonds in the material is
Mechanics of particulate media 63

500 Sandstone

r0- Shale
6 400
''@"" Sandstone

t5 300 Shale Overconsolidated

> .s::
<D Sandstone a.<D
~ 200 Cl
Iii Shale
~ 100 Horizontal
L-_--.-_---,-_ _,--_--.-_---.s_tr_ess (kPa) Sandstone
0 Vertical
100 200 300 400 500
Horizontal stress
effective stress
Figure 2.35 In situ stress states in a sequence of
sediments of varying lithology subject to gravitational
1.5 loading under Ko conditions and assuming normal pore

fluid pressures.

1.0 intact intergranular bonds decreases, and the

sedimentary rock becomes appreciably less stiff.
Lateral deformation can now be prevented only if
0.5 the lateral stress increases in proportion to the
vertical effective stress until the yielding process is
complete. After this, the population of free grains is
o L--~---.--_.--.---r-­
o 100 200 300 400 500 sufficient to allow the rock to consolidate. Further
Horizontal effective stress (kPa) deformation occurs on the Roscoe surface and is
accompanied by large compactional strains. Dur-
ing this process, the Ko ratio is typically about 0.3
or less in the elastic regime, 1.0 at yield, attaining a
value similar to the equivalent uncemented sedi-
15 ment in the post-yield regime (Figure 2.36a-c).
The behaviour represented by this path is
explained in Figure 2.37 (a and b). Here it is seen
1.0 that the initial behaviour corresponds to an elas-
tic loading path that reaches the Mohr-Coulomb
failure surface. The imposed condition of zero
lateral strain prevents shear failure and appreci-
able destruction of the intergranular cement
bonds occurs. This weakens the sediment but not
3 5 7 9 11 to the extent that pore-volume compaction can
Overconsolidation ratio begin (Figure 2.37a). This condition of progress-
Figure 2.34 The relationship between Ko and ive failure with small volumetric strains is main-
overconsolidation ratio (ratio between present effective tained by the imposed condition of zero lateral
stress and past maximum effective stress). (Based on
Atkinson and Bransby 1978.) deformation, and the path moves through the
space contained by the yield surface. Destruction
intact, the rock is stiff and able to resist distor- of intergranular bonding continues until suffi-
tional deformation. Under these conditions Ko is cient grains are freed to allow a more cata-
low and the vertical stress can increase appreci- strophic loss of pore volume. This increase in
ably for small changes in lateral stress. However, volumetric strain corresponds to the onset
as the elastic strain increases, the proportion of of deformation on the Roscoe surface. The rock
64 Mechanical principles of sediment deformation

60.0 50.0
6 50.0
38% porosity ~40.0
UJ 6 38% porosity
~ 40.0 0-
ui 30.0
~ 30.0 ~
~]ij 20.0 .g 20.0

o ~ 10.0
.~ 10.0
O.Of==,--..,-----,,-----,--,----, O.O-+-L---r---,----,----,,---,---.,----,
0.0 20.0 40.0 60.0 80.0 100.0 120.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
(a) Vertical effective stress (MPa) Mean effective stress, p' (MPa)

Yield = Disaggregation of
Mean disaggregated
intact structure
path, all chalks
(Ko - 0.6)

"11 =28%
tl 30
"11 =38%



"11 =40%

o 15 30 45 60 75 90
(c) Mean effective stress, p' (MPa)
Figure 2.36 (a) A plot of greatest against least principal effective stress for a Ko experiment conducted on a high-porosity
chalk sample. (b) The same data set recorded as a stress path in the coordinates p' and q. Note that the plots in both
(a) and (b) show the change in slope of the stress path that accompanies the change in Ko as the sample passes through
yield and its deformation resistance changes. (c) Stress-path plots for a family of chalk samples with different initial
porosities (r)) due to variations in the amount of pore infilling cement. Note that the pre-yield and post-yield stress paths
are common to all samples, but that the stress path during yield is dependent upon the porosity. Samples with the
lowest porosities yield at the highest stresses and require the greatest range of yield stress to transfer to the post-yield

has now become sufficiently dis aggregated to and Jones (1992). Natural Ko stress paths are
allow it to once again behave as if it were the discussed in further detail in section 6.2.
original particulate sediment. For a more de-
tailed discussion of this behaviour the reader is
referred to M.E. Jones and Leddra (1989), Gold- 2.2.7 Over- and underconsolidation
smith (1989), M.E. Jones, Leddra and Potts The term overconsolidation, as mentioned briefly
(1990), M.E. Jones et al. (1991) and Loe, Leddra in the previous section, refers to the state that a
Mechanics oj particulate media 65

1.00 depth than that prevailing. This is a direct conse-

/48% porosity quence of the preservation of pore fluid pressure
and pore volume during burial. Such sediments
are overpressured, undercompacted, and support
large pore volumes. They are therefore amongst
'§ 0.60 the weakest of geological materials (see Yassir
/38% porosity 1989b), and easily attain shear failure at their
critical state because they have experienced only
0.40 limited consolidation (Figure 2.25a).
In contrast to underconsolidated sediments,
28% porosity overconsolidated materials tend to be brittle.
0.20-+---,.--.---.,-----,--,.-----,----, They have pore volumes and mineral frameworks
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 that have been compacted to a greater effective
(a) Mean effective stress (M Pal stress than they now sustain and this imparts a
ko stress Critical state line Deviatoric stiffness. The brittle behaviour of overcon-
Mean paths stress solidated clays deforming in shear has been re-
effective stress ported by Skempton (1966) and Skempton and
Isotropic Petley (1967). These results have been used subse-
consolidation line quently in an interpretation of the natural defor-
mation fabric of the Joes River Formation scaly
clay from Barbados (Enriquez-Reyes and Jones
1991). Further laboratory studies on the relation-
ship between previous-consolidation fabric and
Roscoe surface shear behaviour of clays have been reported by
Leddra, Petley and Jones (1992).
Overconsolidation results from a reduction in
the effective stress experienced by a sediment, and
Hvorslev surface this, in addition to imparting a brittleness, causes
(b) Void ratio
the increase in the Ko ratio discussed in the
Figure 2.37 (a) The change in void ratio that previous section. It is important to note that
accompanies Ko deformation of a porous sediment (chalk). whereas reduction in effective stress may be due to
(b) The data in Figures 2.36c and 2.37a replotted in the
coordinates of p', q and void ratio (after Leddra 1990).
uplift and erosion, and therefore removal of the
This diagram shows how the elastic path intercepts the total stess, it can also be due to the introduction of
Mohr-Coulomb surface (equivalent to the Hvorslev surface pore fluid at elevated pressures. This means that
of fully particulate sediments) of the yield figure and then where overpressures in sediments result from the
moves into the volume defined by the surface, passing secondary generation of excess pore fluid pressure
beneath the critical-state line before emerging on the
Roscoe surface. It is important to note that it is not until the
(section 2.3.4), the sediment will become overcon-
path reaches the Roscoe surface that appreciable solidated. Overpressured sediments can therefore
volumetric strain occurs. be either underconsolidated, in which case they
will behave in an extremely weak and ductile man-
sediment attains when it is unloaded from some ner, or overconsolidated, in which case they will be
previously greater effective stress. All sediments brittle. Secondary overpressures are probably far
that have been buried and then returned to the more common in sedimentary basins than has been
surface are therefore overconsolidated. Under- generally recognized, and the widespread occurr-
consolidation is a term used in geology that has ence of deformation textures associated with over-
no equivalence in engineering. It is, however, a consolidation in mudrocks may be an indication of
most useful geological concept and refers to the how frequently shear deformation of overconsoli-
specific situation where a sediment has preserved dated, overpressured sediments occurs at depth.
high pore pressures during burial and has there- In many respects the mechanical characteris-
fore not attained the equilibrium effective stress tics of overconsolidated sediments are indis-
for its burial depth. Underconsolidated sediments tinguishable from those of weakly cemented
thus behave as if they are buried to a shallower sedimentary rocks. However, overconsolidation is
66 Mechanical principles of sediment deformation

sedimentary rocks. However, overconsolidation is Mohr-Coulomb or

simply a mechanical process - a consequence of Hvorsleve surface (in section)
stress history - whereas cementation (or other
causes of intergranular bonding) can result from
anyone of a spectrum of diagenetic processes. L
Overconsolidation is always a consequence of
reduction in effective stress, although normally or
underconsolidated sediments can become
cemented and behave, at least until yield, as if
overconsolidated. Preservation of pore space due Mean effective
to bonding in underconsolidated materials can \ stress (P')
produce spectacular deformation, with dramatic Roscoe surface (in section)
failure if the yield strength of the material is Figure 2.38 The various effects of cementation and
exceeded. Such materials are especially sensitive to overconsolidation on the Ko stress path for a compacting
energetic small strain events, such as the passage of sediment (see text for full explanation). This diagram
shows the two paths projected on to a constant void ratio
seismic waves. Failure of the porous, weakly section of Figure 2.37b.
cemented structure produces high transient pore
pressures and reduces the effective stress to a low consolidation pressure (Addis 1987). The pre-
value. The disaggregated rock develops large shear consolidation pressure is the stress at which an
displacements in an extremely short period, behav- overconsolidated material upon reloading re-
ing almost as if it had undergone liquefaction. turns to a normally consolidated condition. It is
Examples of such deformations are: the slope equivalent to the previous maximum effective
failures associated with highly sensitive, quick stress that the material had experienced. The
clays (Lambe and Whitman 1979); some slope change from overconsolidated to normally con-
failures triggered by earthquakes (Lambe and solidated material causes a change of slope in the
Whitman 1979); the Aberfan coal tip flow in Wales stress-strain curve at the pre-consolidation
(reviewed in Bromhead 1992); and the mobiliz- pressure. This is geometrically equivalent to the
ation of very high-porosity chalk to flow, both in yield stress of a bonded material.
the toe regions of rock falls and during transient The interplay between consolidation fabric,
loading during fluid production in some North Sea bonding, stress history and the timing and extent
oil wells (Leddra and Jones 1990). In the last type of overpressure development in sedimentary ma-
of occurrence, the chalk can flow into, and fill, a terials is an area in which much important and
large part of the well volume before the shear fruitful research waits to be done. The deforma-
deformation event stabilizes (Leddra and Jones tion fabrics we observe in sedimentary rocks
1990). These are all expressions of extremely depend fundamentally on the interaction of fabric
dynamic shear failure provoked by the extreme and effective stress history. At present, fabric
sensitivity of the material to small strain deforma- interpretation is hampered by a lack of quantitat-
tion. It is probably not correct to regard this ive laboratory measurements. Such experiments,
behaviour as liquefaction (section 1.3.2), although and appropriate numerical modelling, are now
phenomenologically it is extremely similar. possible, and the integration of field and labora-
No reliable criteria are available for distin- tory data should provide for important advances
guishing overconsolidated from bonded ma- in our understanding of natural sediment and
terials, indeed it may be the case that no undis- sedimentary rock deformation.
turbed natural materials are ever devoid of
intergranular bonding (see Burland (1990) and
the discussion in Leddra, Petley and Jones 2.3 NATURAL STRESS, STRAIN AND
(1992)). In the absence of microscopic identifica- PORE PRESSURE
tion of bonding it is possible that differences in
the Ko behaviour may provide evidence (Figure 2.3.1 General
2.38). Bonded materials tend to have a low Ko
before yield whereas overconsolidated materials The styles of deformation being introduced in
exhibit a high Ko before they reach their pre- this chapter are those associated with volume
Natural stress, strain and pore-pressure 67

change (compaction) and shape change (shear). changes in the effective stresses acting upon the
Compaction typically occurs as a porous sedi- reservoir material and may mobilize compac-
ment is buried, though it may be resisted during tional deformations. The most spectacular, recent
natural loading by cementation and pore-press- example of human-induced compaction has oc-
ure generation (Skempton 1970; Been and Sills curred in the Ekofisk Reservoir in the North Sea,
1981; Vaughan 1985; Addis 1987; Yassir 1989a). where the associated sea-floor subsidence has
Shear deformations can be a consequence of proved to be a major engineering problem (Aam
tectonism, where they give rise to faulting, folding 1988; Potts, Jones and Berget 1988). The pattern
and fabric development in the material. However, of displacements that have been mobilized as a
gravitational energy is responsible for most of the consequence of changes in pore fluid pressure in
shear deformation experienced by near-surface, the Ekofisk Reservoir is illustrated in Figure 2.39.
unlithified sediments. This deformation is entirely a consequence of the
Gravitationally induced shear deformations in- passive loading of the reservoir by its overbur-
clude the following. den, due to dissipation of the excess pore fluid
pressure in the chalk reservoir. It is one-dimen-
1. Mass movements in materials unable to sus- sional, and occurring at a strain rate of about
tain the modest shear stresses generated by sur- 10 - 12 to 10 - 13 S - 1. Transfer of the compactional
face relief (Bruns den and Prior 1984; Bromhead displacements from the reservoir to the surface
1986; Hutchinson 1988; Chapter 5). mobilizes shear and stretching deformation of the
2. Where compactional displacements in the sub- overlying materials (M.E. Jones, Leddra and
surface become localized on to pre-existing Potts 1990).
planes of weakness as they are transmitted
through the sediment mass. This has occurred
during compaction of oil fields resulting from 2.3.2 Earth pressure and one-dimensional
fluid extraction (Hamilton and Mechan 1971; compaction
Castle, Yerkes and Young 1973; M.E. Jones,
Leddra and Addis 1987) and, on occasions, has Compaction under gravitational loading gen-
given rise to seismic events. erally occurs under conditions of one-dimen-
sional consolidation - the Ko condition de-
3. When changes in surface relief due to sediment
deposition, uplift or erosion alter the equilibrium scribed above. Under this condition the vertical
between gravitational stress and strain state with- stress increases regularly with depth, as does the
pore fluid pressure, although if permeability is
in mountains, accretionary prisms and deltas
(M.E. Jones 1991). low, excess pore pressures may also develop.
4. During one-dimensional compaction of sedi- During burial, the strain condition will be domi-
nantly one-dimensional, with flattening in the
ments, because the Ko ratio is rarely 1.0 and the
equilibrium state of stress must therefore involve vertical direction and lateral expansion restric-
a deviatoric component (Brooker and Ireland ted or prevented by the earth pressure
1965; M.E. Jones and Addis 1986). (= Ko' O"D. The major fabrics that result depend
strongly on the rock type. In sands, silts and
At greater depth, gravity becomes relegated to carbonates with fairly equant grains, compac-
providing the passive component of the stress tion is simply accompanied by pore-volume re-
system, inducing compaction but with major duction, whereas in clays and other sediments
shear deformations being generally associated composed of platy minerals a pronounced
with the mobilization of tectonic stress (M.E. planar, bedding-parallel fabric may develop
Jones and Addis 1986; Chapter 6). (Petley et al. 1992; section 9.3). The mechanism
An understanding of the stress-strain charac- by which this strain is accommodated is expul-
teristics of sediments compacting under passive sion of the compressed pore fluid. This is ac-
gravitational loading is aided by examination of companied in the pre-yield regime by elastic
what happens when humans extract fluid from distortion of any intergranular bonds, at yield
the subsurface and cause depletion of the natural by breaking of the bonds (or bonded grains if
pore fluid pressure. Fluid extraction from hydro- they are weaker), and in the post-yield regime
carbon and water reservoirs, which causes a by grain-boundary sliding, including grain rota-
reduction in pore fluid pressure, will give rise to tion.
68 Mechanical principles of sediment deformation

~ l + J J J J II ~ I ~ ~ ~ ~ f ~ I I I,
\~ ~ \
\ , \ \ ~ + + ~ ~ l J J J I I I I I II
" , " \\"",,, \ \ \ \ \ \ \ ~ ~ + ~ + t + + ! I ~ t t t J J J J J J J f f II I " I I , "f
" " \ " \ \ ~ ~ + + ~ ! t t J J J I , f I If ' '

~ ,~ ~ ~ ~ " " , \P ~ I
~'+ + ~ ttl + + I J I J I J I f ~ , ~ ~ ~ , ~'; , ~ . :
I'I 'I' I
~ ~ \, t ~ J 1ft f Displacemen~
j j
. "\ \ '" , I I I I I I I I I I I \ I 1 f I ":

" , ~ + l
I " " . '

"'''1"1''/11'/1111 J'I II JJIljJ Iljllllll

,,' \ , I , , , , I' I I
, •

I ' I 'I
I t
I I 1 "
I 'vector
. , 'I

•;l;i!ii.iJUJ,j ld II IIII II J 1J JJ j .1Jh!Jlll.lJUlllu!WjJ;!illlW;);! ;; i (

.----_._------ Reservoir .. Seal e ,-'_....:1....:k--'m-'-----J
Figure 2.39 Finite-element prediction of the pattern of displacements occurring over the compacting Ekofisk oil field in
the North Sea,

Compaction of this type must affect all clastic The possible importance of diffuse strains in
sediments subject to burial and should account accretionary prism toes is assessed in section
for a large proportion of the pore volume lost 6.3.4. Movement on faults controls the state of
during diagenesis. Laboratory experiments have shear stress and the geometry of the sediment
demonstrated that sands, fine carbonates and body, causing surface subsidence due to lateral
clays will continue to compact due to pore- extension and reduction in slope angles. Shear
space compression even when the vertical stress deformation serves to maintain equilibrium be-
exceeds 70 MPa. Such deformations are poten- tween the imposed forces, the strength of the
tially rapid, providing excess pore pressures are sediments and their geometry. Active faults will
able to dissipate. Once excess pore pressures are deform at the critical state, protecting the inter-
substantially dissipated, or if dissipation is pre- vening host materials from shear failure. Under
vented by low permeabilities, slower deforma- these circumstances, strain within the sediment
tion mechanisms, such as diffusive mass transfer body will become strongly polarized, with gravi-
(Rutter 1983; Gratier 1987), will tend to drive tational loading causing pervasive consolidation
continuing compaction. Given sufficient time, leading to a reduction in pore volume and/or
these mechanisms may overprint the fabrics of the generation of excess pore pressures, while
primary consolidation and establish or re-estab- shear deformation remains localized on the fault
lish intergranular bonding. Further consolida- planes. This is akin to the separation of defor-
tion will then be prevented until burial stresses mation paths observed in the chalk deformation
have increased sufficiently to promote yield of experiments described in section 2.2.5.
the bonded material. These concepts are pur- Although tectonic compression generally in-
sued further in section 6.2. volves deformation of stronger geological ma-
terials (fully lithified sediments and crystalline
rocks), in some situations it is also an important
2.3.3 Natural shear stress and deformation cause of sediment deformation. The strain in
Viewed simplistically, shear deformation can be weak porous sediments in the deeper parts of
regarded as a consequence of either reduction or accretionary complexes at convergent plate mar-
increase in the lateral earth pressure with re- gins will be laterally compressive (section 6.3).
spect to the equilibrium Ko condition. Reduc- Similarly, during continent-continent and arc-
tion in earth pressure will cause extension in continent collision, accretionary material, late
basins and mass movements on slopes whereas pre-orogenic and synorogenic sediments will be-
increase in earth pressure will result in lateral come deformed. Current orogenic events, such
compression of a basin. In active extensional as the arc-continent collision occurring in
basins, deltas, accretionary prisms and during Taiwan (Brunsden and Lin 1991; Lin 1991),
slope failure, the major expressions of shear de- are rapid, and high pore pressures, generated
formation are the fault systems which develop. because of shear deformation (Yassir 1989b),
Natural stress, strain and pore-pressure 69

lateral consolidation and increased surface elev- water and gas pressures from their interiors
ation, will not dissipate readily. Sediments asso- (Yassir 1989a). Shale diapirism is one of the
ciated with plate convergence and orogenesis major natural expressions of pervasive shear de-
will tend to be overpressured and either preserve formation in sediments. Diapiric shales generally
large pore volumes and behave in an undercon- have a texture consisting of an anastomosing
solidated manner, or, if they were previously array of polished, curved surfaces with strong
consolidated, will become overconsolidated ow- mineral grain alignments, between which the
ing to secondary pore-pressure generation. Ma- fabric is far less ordered. The origins of this
terials in either of these states are prone to shear scaly fabric (section 9.4.3) have been the subject
deformation under low deviatoric stresses, al- of considerable attention, although most authors
though the strain response will depend on both agree that shear deformation is the major factor
lithology and previous history. Under these cir- (J.e. Moore et al. 1986). Similar fabrics can be
cumstances shear deformation at the critical produced during experimental shear deforma-
state will occur. This may be localized along tion of both overconsolidated (Skempton 1966;
thrust faults or, if more pervasive, may facilitate Skempton and Petley 1967) and weakly
the development of shear zones, diapirs and cemented shales (Leddra, Petley and Jones 1992;
some expressions of folding. Petley et at. 1992). The fabric also occurs in
It is important to note that although defor- non-diapiric shales that have been subject to
mation of sediments during collision orogenesis shear deformation, such as in the Carboniferous
appears to take place in the presence of low of the southern North Sea (Petley et at. 1992)
effective stresses, it is not the case that the total and non-diapiric melange deposits in Sarawak
stresses need be small. This is important because (Clenell 1992). The presence of a scaly fabric in
although high pore fluid pressures may protect an argillaceous rock may simply be an indicator
the material from mechanical compaction due that the sediment has experienced shear strain.
to consolidation, and maintain a low shear If this is the case, it is not surprising that the
strength, it is not protected from pressure- fabric is normally pronounced in diapiric shales.
dependent phase changes. Most pressure- Shale and mud diapirism is of particular in-
dependent phase changes during diagenesis and terest because within the diapir pervasive shear
low-grade metamorphism are driven simply by deformation must have occurred under steady-
increase in the total litho static pressure and are state or strain-softening conditions at the critical
independent of effective stress. A sediment sub- state. From the experimental data and discussion
ject to large total stresses and pore pressures given in section 2.2.5, it would appear that such
may therefore acquire the mineralogical signa- pervasive deformation requires very specific con-
ture of high-pressure diagenesis or incipient ditions, with failure of the intact, bonded fabric
metamorphism while retaining the ability to de- of the shale occurring close to the critical state,
form readily in the presence of low shear but beyond the intersection of the Mohr-
stresses. Coulomb and critical state failure lines. (If failure
Shale and mud diapirism is a common feature had occurred at lower consolidation stresses,
of both deltas and accretionary prisms (O'Brian shear displacements would have become local-
1968; Chapman 1974; Dailly 1976; Mascle, Bon- ized on to fault surfaces, whereas in the presence
hold and Renard 1976; Barber, Tjokrosapoetra of larger consolidation stresses, shear failure
and Charlton 1986; sections 4.3.2 and 7.1). It is would have been accompanied by hardening.)
known, for example, in the Mississippi Delta The materials most prone to diapiric activity
(Figure 4.10), in accretionary complexes such as will be overpressured shales in which bonding
the Oyo Complex, Nias, Indonesia and the Joes has preserved large pore volumes. Such materials
River Formation, Barbados; and in the active are stiff while the bonding is intact, and will
orogenesis of the western foothills region of the regain stiffness after shear deformation if the
Central Mountains of Taiwan. Shales of this shear stress falls below the critical state line.
type may have moved upwards for several kilo- This fits with field observations from Trinidad,
metres in the crust, and when these diapirs where oil wells drilled through diapiric shales
pierce the earth's surface they often produce have encountered materials that are stiff and
mud volcanoes, which appear to dissipate excess deform into uncased wells in a blocky manner,
70 Mechanical principles of sediment deformation

showing a high density of discrete shears (Yassir recovered from depth in North Sea oil wells
and others, personal communication 1986). This suggest that the overpressures would dissipate in
fabric is clearly seen in core samples of the 1-5 million years, even through 3 km of overbur-
diapiric shale and is similar to the deformation den (unpublished data). These shales also have
of the chalk sample illustrated in Figure 2.30b. compacted fabrics and a pronounced bedding-
Shale diapirism, therefore, is not an indicator parallel fabric. This suggests that the overpres-
that the sediment has behaved in a fluid manner. sures either owe their genesis to an event which
This is an important lesson - ductility (plasticity) occurred later in the history of the sediment,
is not equivalent to fluidity. Sediments behave in following significant compaction, or that the
a ductile manner when deforming in shear at the mechanism responsible for the overpressures has
critical state, but this behaviour pertains to the been active over a considerable period of time.
prevailing stress conditions. Thus shear deforma- Watts (1983) argued that overpressures in the
tion to large strains can occur in any shale if Cretaceous and Tertiary sequence in the North
appropriate stress conditions are attained. Sea central graben were sustained by the high
capillary pressures required for oil to migrate
into the chalk and for natural gas to migrate into
2.3.4 Origin of overpressures in sediments the overlying shales. In Taiwan, Miocene deltaic
Pore fluid pressures, particularly fluid pressures sediments, folded along the western margin of the
in excess of the equilibrium hydrostatic pressure, currently active tectonic belt contain hydrocar-
exert a fundamental influence on the mechanical bons with extreme overpressures. These sedi-
behaviour of sediments, and sedimentary rocks. ments have not been subject to rapid, deep or
The timing of overpressuring within the deforma- sustained burial but have been tectonized and
tion history of the sediment is also important in now lie at the margin of the 3 km-high mountain
so far that consolidation state, and the extent to chain. The large excess pore pressures are in part
which a sediment is under- or overconsolidated, due to the nature of the pore fluid (a hydrocar-
strongly influences subsequent deformation. Its bon gas), but the recent tectonic deformation,
origin in sediments and sedimentary rocks must thermal history and geomorphological setting
therefore be considered in any discussion of natu- will also be contributing to their development
ral stress and deformation states in sediments. (Chan 1964).
Overpressures are an almost ubiquitous fea- Yassir (1989a) has reviewed a number of poss-
ture of sedimentary basins, deltas, and accretion- ible causes for overpressure generation which are
ary complexes, and are often sustained for very not related to the trapping of fluids during burial.
considerable periods. Their origin is often at- These are as follows: tectonic deformation of the
tributed to the combined effects of low permea- sediment; artesian conditions; dehydration of hy-
bility and rapid burial. Overpressures of this type drous minerals during burial and invasion by
are well known from deltaic sequences (section fluids released during dehydration of more deeply
2.3.3), but are generally restricted to mudrocks, buried rocks (Bruce 1973, 1984; Magara 1975a;
or other lithologies that have become sealed by Pittman and Reynolds, 1989); geothermal heating
low-permeability mudrocks and then rapidly (Barker 1972; Magara 1975b; Barker and Hors-
buried (Magara 1971; Chapman 1972, 1981; Car- field 1982; Daines 1982); capillary pressures due
stens and Dypvik 1981). However, this is not the to the migration and gravitational segregation of
only cause of overpressure development; other multi phase fluids, generally gas-oil-brine systems
mechanisms are also important. This is particu- (Fertl 1973; Hedberg 1974; Archer and Wall
larly the case in sediment sequences where over- 1986); and osmotic effects (Hanshaw and Zen
pressures have been maintained for appreciable 1965; Young and Low 1965).
periods (Bredehoeft and Hanshaw 1968; Han- The origins of overpressures in sediments and
shaw and Bredehoeft 1968). sedimentary rocks are thus diverse and may
Few sediments are actually impermeable, so occur at different stages during the history of the
the persistence of overpressures generated by sediment. The presence of overpressures in a
sediment burial over appreciable periods of time naturally occurring sediment are thus not necess-
seems unlikely. Measured coefficients of consoli- arily evidence for rapid burial, low permeability
dation from highly overpressured Eocene shales and underconsolidation. Many, maybe the ma-
Conclusions 71

jority, of overpressured sediments owe the pres- is favoured by increasing deviatoric stress, and at
ence of the excess pressure to a mechanism of shear failure at the critical state will cause large
secondary pore-pressure generation. In such sedi- strain deformations at constant pore volume and
ments, the state of consolidation will depend constant deviatoric stress. Shear deformation is
upon the magnitude and timing of secondary also favoured by large excess pore fluid pressures,
pore pressure generation with respect to the which prevent consolidation. Both shear and
development of intergranular bonds. consolidation are strongly influenced by the
lithology and fabric of the sediment, and by its
previous stress and pore-pressure history.
2.4 CONCLUSIONS Sedimentary rocks are the result of lithification
of sediments, and exhibit significant elasticity at
This chapter has been concerned primarily with low stresses and small strains. With increasing
the mechanics of how sediments deform, with stress this elasticity is progressively destroyed
stress systems, the role of pore-fluid pressures and the rock undergoes a change of state, its
and the influence that deformation may have on behaviour reverting to that of the equivalent
the properties of the material. Although the prin- particulate sediment. The behaviour of the rock
ciples have been presented with reference to the- at and beyond yield depends on the stress path
ory and laboratory tests, the final part of the and on previous consolidation history. At low
chapter has given some glimpses of the relevance stresses, or in the presence of high pore pressures,
of the ideas to the explanation of natural sedi- yield at the critical state is accompanied by
ment deformation phenomena. These latter as- pronounced bifurcation of the deformation path.
pects are explored in the following chapters. Segments of the rock undergo disaggregation and
Much remains to be learnt about the mechanical become ductile shear zones, or faults deforming
behaviour of large sediment bodies before the in an ideal critical state manner, while the shear
generalizations presented above can be quanti- stresses decay in the remainder of the rock which
fied. The careful and informed application of then behaves in a passive manner. Similar bifur-
modern geotechnical theories and methods ap- cation of the deformation path appears in land-
plied to such problems will provide a basis slips and large sediment bodies such as deltas or
through which this quantification can be achieved. extensional basins. In the presence of larger con-
The principles, concepts and ideas presented in solidation stresses, shear deformation is more
this chapter lead to the following broad con- pervasive and leads to a form of ideal plasticity,
clusions. although dilation may also cause work harden-
Sediments are amongst the most deformable of ing.
geological materials, although this depends cru- Deformation of sediments and sedimentary
cially on their previous deformation history, and rocks is thus entirely dependent upon the nature
the pore fluid pressure. They tend to deform by a of the effective stress systems affecting the sedi-
combination of volume change, due to expulsion ment, and therefore on the pore-pressure history.
of pore fluids in response to an increasing iso- Large pore pressures preserve the pore volume
tropic component of the applied stress system, but facilitate shear deformation in sediments. If
and shear due to imbalance in the. applied the pore pressures are primary, the sediment will
stresses. Compaction is generally pervasive, nor- be underconsolidated and respond in a plastic
mally driven by the gravitational components of manner, whereas if they are secondary, a stiffer,
stress (the passive stress system) and often con- overconsolidated, type of shear response will oc-
strained to be one-dimensional owing to the cur. For all these reasons, the role of fluids and
presence of adjacent compacting rocks and the pore pressures in deforming sediments is men-
generation of earth pressure. Shear deformation tioned frequently in the chapters that follow.

Glacial deformation

3.1 INTRODUCTION ice masses and, further, that many palaeo-ice

masses overlay unlithified sedimentary material
The world is an icy place. Some 10% of the rather than bedrock. Because access to the sub-
earth's land surface is currently covered by gla- glacial environment is difficult, much of this evi-
cier ice; at glacial maxima this coverage has been dence comes from the now exposed beds of past
much greater. Moving glacier ice is a powerful glaciations. Retreat of ice following the 'Little Ice
geomorphological agent and many parts of the Age' maximum is continuously exposing sedi-
world have been distinctively altered by the ac- ments that were covered by ice until very recent-
tion of glacier ice. It is at the base of the glacier ly. Together with the sedimentary cover of the
or ice sheet, where ice and substrate are in mid-latitude regions that lay beneath Quaternary
intimate contact, that geomorphological activity ice, these materials provide evidence of the basal
and change is greatest: it is here that most glacial conditions within the sediments beneath these ice
sediment deformation occurs. Modern research masses. If ice were to advance again in Europe
has shown that the theoretical principles outlined and North America to occupy positions of previ-
in the previous chapters are relevant to both the ous maxima, it would do so over an extensive
movement of the ice sheets and to the under- sediment cover. Sediments are estimated to have
standing of the structures that are produced. underlain some 70-80% of the mid-latitude ice
Thus glacial movement provides a fine example sheets outside of their central regions (Boulton
of how the mechanical principles of sediment and Hindmarsh 1987).
deformation are applicable in a particular geo- It is not known what proportion of modern
logical environment. This is the subject of the glaciers overlies unlithified sediments; the basal
present chapter. environment lies hidden beneath tens to
Early glaciological observation was concen- hundreds of metres of glacier ice. However,
trated in the alpine regions, where basal condi- evidence for the widespread presence of sedi-
tions of ice overriding hard, lithified material ments beneath modern ice masses has accumu-
were inferred from the accessible margins of lated in recent years. The first indications came
mainly small, mountain glaciers (Clarke 1987a). from direct observation of active sediment de-
Deglaciated rock with striated and polished sur- formation beneath the margin of a modern gla-
faces provided further evidence of the impressive cier in Iceland (Boulton and Jones 1979). These
erosive power of glacier ice in intimate contact observations were followed by the collection of
with bedrock. Detailed theories for glacier high-resolution seismic results from Ice Stream
motion by internal ice deformation and sliding B, West Antarctica, which revealed what has
over a rigid bedrock surface were developed. been interpreted as a metres-thick layer of
Thus, until recently, the glaciologists' view of this saturated sediment beneath this active ice
basal region was of a clean interface between ice stream. Subsequently, unprecedented interest
and rough, lithified bedrock. However, evidence has arisen in sedimentary glacier beds and in the
has mounted that this picture is unlikely to be mechanics of the deformation processes within
accurate for a significant proportion of current them.

The Geological Deformation of Sediments Edited by Alex Maltman Published in 1994 by Chapman & Hall ISBN 0 412 40590 3
74 Glacial deformation

Despite the logistic difficulties of gaining access as supra- or proglacially in debris flows, or en-
to glacier beds, sediment samples have been re- glacially in layers of debris or debris-rich basal
trieved from under the ice, and direct in situ ice. Furthermore, the wide range of glacial set-
measurements of deformation have been per- tings, both geographical and thermal, results in
formed. Initially, these measurements were made the possibility of both glacial and glacio-aqueous
close to the glacier margins, where tunnels were deformation occurring either before or after de-
used to access basal sediments, allowing deriva- position. The dynamic nature of the pro glacial
tion of the first flow law for subglacial sedi- environment - where sediment is commonly re-
ments (Boulton and Jones 1979; Boulton and worked after deposition and structure is created
Hindmarsh 1987). Further upglacier, where basal and destroyed by marginal processes such as
conditions are probably more characteristic of dessication, dewatering and overriding - results
the glacier bed as a whole, borehole access has in complex and potentially hard-to-interpret
allowed the collection of sediment samples (En- sediments in which processes may be only subtly
gelhardt et al. 1990; Blake, Clarke and Gerin recorded (Clarke 1987b). No-one who has spent
1992) and the emplacement of sensors to measure time in the forefield of a glacier will dispute either
directly the deformation of basal sediments and the activity or complexity of the environment.
the sliding of ice over this unlithified material This review is concerned largely with the sub-
(Fahnestock and Humphrey 1988; Blake and glacial deformation of terrigenous sediments.
Clarke 1989; Kohler and Proksch 1991; Blake, Such glacio tectonic deformation is defined as
Clarke and Gerin 1992; Humphrey et al. 1993). the direct result of glacier motion or loading
Where ice overlies predominantly soft and (INQUA Group on Glacial Tectonics; see also
water-saturated sediments, deformation may oc- section 1.1.1). Marginal, depositional and prog-
cur in response to gravitationally driven forces. lacial processes will be discussed only briefly, and
Both compressive and shear stresses operate on largely in the context of the preservation of
the sediment. Compression results from the verti- features developed subglacially. Subglacial sedi-
cal component of stress due to the ice overbur- ment deformation and the resulting sedimentary
den, and can cause dewatering and consolidation. structures will be developed from the perspective
The process is in many ways analogous to the of sediment properties, the physics of till mechan-
burial pressure that arises through sediment ac- ics (Clarke 1987b) and the processes operating at
cumulation (section 1.2.2), and has recently been the base of the ice.
analysed by Boulton and Dobbie (1993). Shear
stresses are driven by changes in ice-surface and
sediment-surface elevation, and may result in 3.2 SUBGLACIAL CONDITIONS
shear deformation within the basal material. This
deformation can contribute directly to ice-surface 3.2.1 General
motion. Unlithified sediments undergoing active
deformation have been identified beneath a Basal conditions beneath a real ice mass over-
number of modern ice masses, including: Ice lying an unlithified bed are complex (Figure 3.1).
Stream B, Antarctica (Alley et al. 1986; Blanken- The sediment is inhomogeneous, rheologically
ship et al. 1986; Engelhardt et al. 1990); Columbia complex, and parameters such as fluid pressure
Glacier, Alaska (Fahnestock and Humphrey may vary both spatially and temporally, poten-
1988; Meier 1989); Trapridge Glacier, Yukon tially within a brief time-scale (e.g. Clarke,
Territory (Clarke, Collins and Thompson 1984; Meldrum and Collins 1986). This section dis-
Blake and Clarke 1989); Urumqi Glacier No.1, cusses some idealized scenarios in order to indi-
China (Echelmeyer and Wang Zhongxiang 1987); cate the range of conditions that may occur
Breidamerkurj6kull, southeast Iceland (Boulton beneath ice masses, and it defines the termino-
and Jones 1979; Boulton and Hindmarsh 1987); logy used in the rest of this review.
and StorglacHiren, Sweden (Kohler and Proksch
In addition to subglacial deformation, sedi- 3.2.2 Bed type
ments within the glacial environment may de- The bed underlying an ice mass may consist of
form in a number of other distinct settings, such sedimentary material or bedrock, but the nature
Subglacial conditions 15

Figure 3.1 The nature of the

ice-sediment interface near the
margin of a temperate glacier.
Note the wide size distribution of
the sediment, the presence 01 very
coarse material and the high
water content of the sediment. (a)
Clean ice-sediment interface. (b)
Glacier sole covered with loose,
wet sedimentary material.

of the material underlying the ice may not be stresses but is not necessarily actively deforming
the same at all locations and also may vary at a given time.
temporally. The definitions used here are as Active deformable bed: an unlithified sedimen-
follows. tary bed that is underoing current deformation,
driven by forces resultant from the overlying glacier
Hard bed: the glacier bed consists solely of ice and/or the glacial hydrological system, and
bedrock material. resulting in a contribution to ice-surface motion.
Deformab!e bed/amlitiuHied boo: a glacier bed
consisting of unlithified material, which has the
potential for active deformation under the 3.2.3 Thermal regime
appropriate subglacial conditions and driving In general, the role of temperature in sediment
76 Glacial deformation

deformation is to affect modestly the stress sediment may then cause deposition and erosion
configuration (section 6.2.3), but in glacial de- of sediment by melt-out or freezing-on of sedi-
formation the influence is major. The thermal ment (Menzies 1981).
conditions govern the extent to which ice rather
than water is present in the sediment pores. A
sedimentary bed underlying a glacier may be 3.2.4 The ice-bed interface
ice-infiltrated or ice-free, a factor that will be It is largely the nature of the interface between
determined largely by the basal temperature the glacier ice and the underlying sedimentary
and basal melt rate. Temperatures recorded be- bed that will determine to what extent the ice
neath glacier ice range from -13 to -18°C slides over the bed or deforms the sediments
(Hansen and Langway 1966; Paterson 1976) beneath it. Theoretically, three situations can be
beneath cold-based ice masses, to the pressure envisaged.
melting point beneath warm, wet-based ice
masses. In this context three scenarios may be 1. Complete decoupling of the ice and sedi-
envisaged. ments: the ice and sediment are buffered from
one another and motion will be by slip alone
1. Frozen bed: water contained in the bed (Figure 3.2a). Decoupling could result from the
is below the pressure melting point and pre- presence of a water film or cavities at the
sent largely as ice. Such a frozen bed was interface (Shoemaker 1986; Lingle and Brown
observed beneath Urumqi Glacier No.1, China 1987).
(Echel-meyer and Wang Zhongxiang 1987), 2. Complete coupling of the ice and deform-
where a frozen debris layer has been observed able bed: no slip occurs between the sedimen-
to have a viscosity less than glacier ice and to tary material and overlying ice (Figure 3.2b);
undergo enhanced deformation. There is, in this situation the ice and sediments must
however, no clear distinction between sedi- move as a continuum. Such coupling could result
ment-rich ice and ice-rich sediment; that is, from the infiltration of ice into the bed (Shoe-
between basal ice and a frozen sedimentary maker 1986). If no sliding of the ice occurs,
bed. then ice will infiltrate into the underlying sedi-
2. Unfrozen bed: water contained in the ments by the processes of Darcian flow and ice
sediments is at, or above, the pressure melting regelation (Boulton and Hindmarsh 1987) un-
point and is present in the liquid phase. Such less basal melt rates equal or exceed the infiltra-
a bed would behave in a characteristically dif- tion rate.
ferent way from a frozen bed, and its viscosity 3. Partial decoupling of the bed and sedi-
would be much lower. Such an unfrozen bed is ments: the bed is incompletely coupled to the
believed to exist under Ice Stream B (Blanken- overlying ice and both sliding and deformation
ship et al. 1986; Engelhardt et al. 1990). The occur (Figure 3.2c). Partial decoupling also
sediments comprising such a bed are normally covers active sliding, such as the ploughing of
assumed to be water saturated. sediments by clasts entrained in the ice (N.E.
3. Partially frozen bed: water at, and within, Brown, Hallet and Booth 1987), and active
the bed is present in both the solid and liquid grinding and glacier motion due to deforma-
phases. Such a bed may result from ice infil- tion within a thin coarse layer of sediment at
tration into sediments, such as would be expected the base of mountain glaciers (Robin 1989).
to occur if the pore water pressure within the Measurements beneath glaciers suggest that
bed is low (Shoemaker 1986; Boulton and partial decoupling does occur. At Trapridge
Hindmarsh 1987). It is likely that on a larger Glacier both sliding and bed deformation have
scale most beds will vary spatially in thermal been measured in situ, with the former account-
regime and, therefore, will be partially frozen. ing for some 40% of measured surface motion
Furthermore, where debris-rich basal ice or (Blake, Clarke and Gerin 1992). At Breidamer-
frozen sediments overlie unfrozen sediments, kurj6kull, ice infiltration has been observed,
spatial and temporal variation of the freezing although basal slip accounted for ~ 10%
front within the sediment may occur. Migration of the glacier motion (Boulton and Jones 1979;
of the interface between frozen and unfrozen Boulton and Hindmarsh 1987).
Subglacial conditions 77

Velocity ~


Ice-sediment interface
Lower sediment interface

(a) Rock/immobile sediment





Figure 3.2 Schematic diagrams of the influence of the ice-bed interface on surface velocity: (a) complete
decoupling of the ice and sediments (sliding only); (b) complete coupling of the ice and sediments (active deformation
only); and (c) partial decoupling of the ice and sediments (sliding and active deformation). Diagram assumes no slip
at the lower interface and a linear velocity-depth profile within the active sedimentary layer. (After Alley et al.

The above discussion has assumed an idealized structure has been observed in Icelandic glaciers
transition from glacier ice to basal material in the (Boulton, Dent and Morris 1974; Sharp 1984).
vertical direction. For a real glacier, the ice-bed If the sediment layer is thin with respect to the
interface is unlikely to be this clearly defined active layer (Figure 3.3b), then deformation oc-
everywhere, and bands of sedimentary material curs down to bedrock and erosion of the lower
will probably exist incorporated into the ice bedrock surface is possible. This bed type is
above both frozen and unfrozen beds. Such sedi- considered to exist beneath Ice Stream B, at the
mentary bands were observed at Urumqi Glacier site of the Up Stream B field camp (Alley et al.
No. I and Shoestring Glacier, Washington (Brug- 1987a).
man 1983; Echelmeyer and Wang Zhongxiang
3.2.6 Realistic basal conditions
It is possible for the base of the glacier to vary
3.2.5 Bed thickness spatially or temporally between these bed types
A sediment layer is thick with respect to the (Menzies 1989a). A spatial variation between a
actively deforming layer if sediment defor- hard and soft bed would occur, for example, if
mation occurs above a layer of immobile there are bedrock protuberances that equal, or
sediments (Figure 3.3a). The resulting two-layer exceed, the sediment layer thickness, and the bed
78 Glacial deformation



(b) Rock
Figure 3.3 Schematic diagrams representing: (a) a thick, deformable bed where sediment thickness exceeds the active
layer thickness; (b) a thin active bed where the sediment layer thickness equals the active layer thickness.

was therefore hard in some places and deform- incompressible - or in the intrinsic properties of
able in others. Equally, a surge-type glacier, such the sediment, but it dramatically influences the
as Trapridge Glacier, which overlies a sedi- response to stress.
mentary bed (Clarke, Collins and Thompson
1984), probably experiences a dramatic tem-
poral increase in the activity of its deform- 3.3.2 Compressive stresses
able bed when it passes from quiescent to Direct loading by glacier ice gives rise to com-
active surge behaviour (Raymond 1987). On a pressive stresses, and potentially results in con-
shorter time-scale, evidence for such temporal solidation of sediments resulting from the loss
switching arises from in situ strain measure- of water and spatial rearrangement of solid
ments made within the basal sediments beneath particles (Mathews and MacKay 1960), in direct
Trapridge Glacier. Although these represent analogy with the submarine accumulation of
downglacier flow over longer time periods, a sedimentary pile (Boulton and Dobbie 1993).
they also show short periods of both zero Such consolidation is normally accompanied
and negative strain (Blake, Clarke and Gerin by alteration of sediment properties, including
1992). increased shear strength, and decreased com-
pressibility and permeability (sections 1.2.2 and
2.2.3). Loading a sediment causes the pore
3.3 STRESSES ARISING FROM pressure within it to rise. Pore water then starts
OVERLYING ICE to be expelled from the sediment at a rate deter-
mined by the sediment permeability and volume
compressibility, until the excess pore pressure
3.3.1 General is dissipated. As water is expelled the effective
Glacial deformation of the sediments underlying stress increases correspondingly.
an ice mass results from stresses that arise from Consolidation of a sediment can lead to de-
the overlying ice. As in other contexts of sedi- watering structures (section 9.5.3), and potentially
ment mechanics, it is the effective stress, rather to overconsolidation if changes in pore pressure
than the total stress, that controls deformation, occur repeatedly. Such cyclic pore-pressure
consolidation and swelling of the sediment (e.g. changes are common in the subglacial environ-
Skempton 1970; sections 1.2.5 and 2.2.3). A ment, and may drive alternating consolida-
change in pore-water pressure alone causes vir- tion and swelling of the sediment in a non-
tually no change in volume - as water is virtually reversible precess that leads to a reduction in
Sediment properties 79

sediment porosity known as ratchetting. This 3.4 SEDIMENT PROPERTmS

process results from a reorganization of par-
ticles within the sediment body (Schofield and The grain-size distribution of sediments is a
Wroth 1968; Clarke 1987b) and can be accom- fundamental control on sediment deformation
panied by the development of anisotropy. Large properties, and on the physics operating with-
water-pressure variations are common within in sediments at the glacier bed. Sediments pro-
the basal hydraulic system, on a variety of duced by the direct erosion of lithified material
time-scales, ranging from fluctuations on a by glacier ice typically contain particles spann-
scale of minutes to hours through diurnal to ing a large size distribution, and are frequently
annual fluctuations (e.g. Clarke, Meldrum and characterized by high specific surfaces due to
Collins 1986). However, it is not clear how the presence of fine materal. In situ, these
well the hydraulic system communicates with sediments may have high water contents and
pore water within the basal sediments. porosities (Ronnert and Mickelson 1992),
particularly where active deformation is occur-
ring (Figure 3.1). Deformation results in high
3.3.3 Shear stresses strains within basal sediments, often as the
The deviatoric component of the effective stress result of high pore-water and low effective
tensor (section 2.2.2) can result in the shear pressures. Furthermore, sediment properties are
deformation of sediments and the formation of altered by ice-basal processes, which are con-
structures resulting from high strain. Because trolled by conditions at the bed. These basal
deformation results from the spatial rearrange- conditions are, in turn, affected by changes in
ment of sediment particles, properties such as sediment properties. The subglacial sediment
shear strength and permeability will be affected. system is characterized by a subtle interplay
The response of sediments to applied shear stress and feedback between sediment properties and
will be dependent on basal conditions. Shear basal processes.
forces must be transmitted from ice to sedi- The bimodal size of rock fragments produced
ment; hence coupling across the ice-sediment by the erosion processes of mechanical crush-
interface must occur, at least locally. Coupling ing and abrasion (Dreimanis and Vagners 1971)
across the ice-sediment interface is unlikely to is reflected in the subglacial environment by
be uniform, and so the shear stress transmitted the classic grain-size distribution of subglacial
to a portion of the glacier bed may vary both sediment - rich in both fine and coarse material.
spatially and temporally. The heterogeneity of The fine component reflects the terminal mode,
the resulting deformation is probably more or modes, of the parent lithology (Dreimanis
marked than in other geological environments. and Vagners 1971; Haldorsen 1981), typically
Deformation may be pervasive (e.g. Boulton 2-9 phi (0.25 mm to 2, and the proportion
and Hindmarsh 1987), and occur through a of coarse to fine material reflects the travel
significant depth of basal material, or, alter- distance of material. Increased travel distance
natively, discrete deformation may occur (Kamb results in a depletion of clasts with resultant
1991), resulting in localized planes or regions of progressive fining of sediments downglacier
high strain with little strain occurring in sur- (Dreimanis and Vagners 1971).
rounding sediment. Clasts entrained at the This bimodal distribution means that sub-
ice-sediment interface may plough through the glacial sediment may be considered to consist
underlying sediments if the sediment yield of clastic material suspended in a matrix of
strength is exceeded only locally (N.E. Brown, saturated fines (Clarke 1987b). The coarse
Hallet and Booth 1987). Ploughing represents material is typically elongate and relatively
an intermediate state between sliding (no coup- rounded (Boulton 1978; Lawson 1979), reflect-
ling of ice to sediment) and pervasive defor- ing the frequency of collision and abrasion events
mation (good coupling across the ice-sediment that result in comminution where deforma-
interface) (Alley 1989a). tion occurs. These processes of comminution
Further effects of deformation on sediment and result in the production of matrix material. The
the structures that may be developed will be fine material that makes up the matrix may
discussed in section 3.8. be relatively infrequently rich in clay-sized
80 Glacial deformation

material (> 9 phi, < 2 ,um), and especially clay actively deforming layer and support significant
minerals, unless an ice mass has advanced over shear stress.
marine-sorted sediments (Milligan 1976). Marine As the proportion of matrix material increases,
sedimentation tends to result in sediment sort- clasts become effectively isolated from one an-
ing, with progressive fining of sediment material other within the matrix and interactions between
in ice-distal locations. This, accompanied by larger particles become rare. Evidence from soil
chemical breakdown of sediments in the marine mechanics suggests that approximately 30% of
environment, means that ice masses overriding the solid particles by mass must be matrix
material deposited below eustatic sea-level may material to effectively isolate clasts from one
have significant proportions of clay-sized par- another (Smart 1985). The matrix material with-
ticles and clay minerals in the matrix material. in subglacial sediment may be either silt- or
Ice that is not advancing over previously de- clay-rich. The deformation behaviour of a sedi-
posited sediments, but is producing a sedimen- ment in which clasts are isolated will depend
tary bed by primary erosion processes will re- largely on the percentage of clay-sized material;
quire very long erosion distances to produce as the amount varies, changes in the character
significant clay-sized material, especially as the of the forces dominating deformation occur. At
terminal modes for many minerals are silt-sized the free surface of solid particles of any size, a
rather than clay-sized (Dreimanis and Vagners potential field exists associated with unsatisfied
1971). Clay minerals are likely to be uncommon, intrasolid bonds. This potential field gives rise
because chemical weathering rates resulting in to surface forces that result in cohesion/adhesion
the formation of clay minerals will be sup- bonding. Summed over the surface of a solid
pressed at the low temperatures and relatively particle, the magnitude of these forces will be
low chemical activities associated with the sub- insignificant compared with other forces, such
glacial environment; in contrast, physical abra- as weight, for silts, sands or larger particles. As
sion processes will be relatively rapid. Particle the surface area to volume ratio increases with
sizes in the sediments at the margins of many decreasing particle size, the importance of
modern glaciers: e.g. Matanuska Glacier, Alaska such bonding increases, typically becoming
(Lawson 1979); Trapridge Glacier (Clarke 1987b); significant compared with other forces at a
sediments from Icelandic and some Spitsbergen particle size of ~ 1-2,um (i.e. for clay-sized
glaciers (Boulton 1976); sediments from beneath particles). Consequently two types of matrix
Ice Stream B (Engelhardt et al. 1990); and sedi- can be envisaged: silt-rich or clay-rich. Within
ments underlying past ice masses: e.g. Norwegian a silt-rich sediment, deformation will be domin-
sediments (Haldorsen 1983); sediments from ated by physical frictional forces between par-
central Finland (Haldorsen 1981; Virkkala 1969); ticles, and the properties of individual particles
those from the Canadian Shield (Scott 1976); controlling physical deformation will be similar
and central and northern Sweden (Haldorsen to those controlling deformation of a clast-
1981); indicate that many land-based glaciers rich sediment. As the particle size decreases
and ice sheets overlie sediments that are silt- and the sediment becomes clay-rich, deforma-
rather than clay-rich. tion will be dominated by surface effects and
As indicated above, the proportion of clastic to factors, such as mineralogy and pore-water
matrix material will be fundamental in determin- impurities, that perturb conditions at particle
ing the deformation characteristics of the sedi- surfaces.
ment. Where a continuous, or quasi-continuous The control of particle-size distribution on
skeleton of clasts exists, deformation will be deformation results from: (i) the changing role of
dominated by physical interactions and frictional surface forces with particle size; and (ii) the effect
forces between this clastic material. The physical of particle size on permeability (e.g. Freeze and
properties of the bulk sediment will be controlled Cherry 1979). The permeability/water-through-
by the large size and inert nature of the clasts. flow dependence will be discussed in section 3.7.
Deformation of such a sediment is controlled Note that particle mineralogy also affects the
by particle size, shape, surface texture and size magnitude of surface forces and, hence, values of
distribution. Furthermore, clast size may be the Mohr-Coulomb friction and cohesion coeffi-
such that one or more particles may bridge the cients ¢ and c, the residual strength (e.g. Mitchell
Sediment transport system 81

1976} and the porosity control on deformation 3.5 THE SEDIMENT TRANSPORT
(Terzaghi 1955). Because different mineralogies SYSTEM: PRODUCTION,
produce different sized characteristic abrasion
and crushing products, physical comminution ALTERATION AND LOSS
results in changes in mineralogy with particle size
within a natural sediment. Sorting by size will, as In the glacial environment sediment production,
a result, create mineralogical contrasts and fur- deformation and deposition cannot be separated
ther affect sediment properties. conceptually. A brief outline follows of the sedi-
Subglacial sediment is inhomogeneous on ment system, from initial production to final
many scales. Firstly, its wide range of particle deposition.
sizes gives rise to an inherent inhomogeneity - Much as the basal hydraulic system transports
compare the size of a clay particle with a large water beneath a glacier or ice sheet, the basal
glacial erratic. However, further heterogeneity is system of an ice mass overlying an active deform-
superimposed on the sediment by internal struc- able bed transports sediment. Measurements sug-
ture and the spatial grouping of particles by gest that 88% of ice-surface motion may result
depositional and mobilization processes (Boulton from sediment deformation (Boulton and Hind-
1976; Menzies 1989a). Not only is coarser ma- marsh 1987), so the sediment transport rate may
terial preferentially deposited at local surface be significant. At Ice Stream B, where the basal
perturbations, and particularly where other large sediments are of the order of 6 m thick and the
particles have lodged, but the resulting clusters surface velocity is '" 500 m a -1 (values at the Up
of coarse material are further resistant to re- Stream B field site; Alley et al. 1989b), sediment
mobilization or erosion by water. Finer material, transport will occur at the rate of '" 600 m 3 a - 1
in contrast, both couples readily to water, per unit width of grounding line (Alley et al.
resulting in the flushing of fines (Clarke 1987b), 1989a).
and mobilizes more readily under many condi- For an ice mass in equilibrium, the rate of
tions away from the ice margin. Furthermore, transport of sediment must equal the sediment
many sediment properties, such as shear and production rate. Where a glacier is producing
residual strength, viscosity and permeability, are its own sedimentary bed by primary erosion of
controlled, or affected by changes in particle-size lithified material, the sediment transport system
distribution. Because deformation occurs prefer- may be considered analogous to the glacier it-
entially at local sediment weaknesses (Alley self. Upglacier primary production of sediment
1989b), such inhomogeneities become self-per- occurs by direct erosion of bedrock beneath a
petuating. thin bed, the sediment layer thickens, and sedi-
Deformation results from the spatial rearran- ment flow is extensive. Downglacier deposition
gement of particles within the sediment. As this occurs, the immobile layer thickens and sediment
rearrangement occurs, continuing deformation flow is compressive (Boulton 1987). Between
will progressively affect sediment properties. these regions a conceptual sediment equilibrium
Shear deformation in many sediments is likely to line may be defined where no depositon or ero-
cause dilation and particle alignment accom- sion occurs. This line may not be fixed in either
panied by the development of anisotropy, al- time or space.
though those that have undergone extremely Sediment production for an ice mass over-
high strains, such as may be common in the lying a soft bed is less efficient than for a hard
subglacial environment, may be characterized by bed. The primary production of sediment from
homogenization and mixing (Hart and Boulton underlying lithified material can occur only at
1991). a bedrock interface (either ice-bedrock or
Therefore anisotropy and inhomogeneity oc- sediment-bedrock). A thick deformable bed has
cur due to the deformation of sediment and no access to a bedrock interface, and so no
its deposition, with structure developing at all ability to produce sediment. An ice mass over-
scales. In a thick bed such changes in properties lying such a bed, and producing this bed by
result in a two-layered structure, with mobile, erosion of underlying bedrock (rather than ad-
deforming sediments lying above an immobile vancing over previously deposited material),
lower layer. must have a source region with either a hard or
82 Glacial deformation

thin bed where intimate contact with bedrock 3.6 MODELS OF SEDIMENT
occurs. Further downglacier the sediment layer
will thicken until a region is reached where de-
position occurs and active sediment deforms
above previously deposited sediments - the re- 3.6.1 General
gion of thick bed.
As this sediment is transported it will under- The deformation of sediment beneath a glacier
go comminution by intrasediment processes. results in changes of structure, particle arrange-
Sediment generation at the bed of an ice mass ment and water content. If we are to be able to
can occur by the processes of plucking, abra- decode the resultant structure, and to reconstruct
sion, bedrock crushing and water-aided erosion the physical processes and conditions that have
(Rothlisberger and Iken 1981; Drewry 1986; produced it, we must develop the science of till
Iverson 1991). Typically, the primary processes mechanics (Clarke 1987b). This section summar-
of sediment production (plucking and rock crush- izes some current models of sediment deforma-
ing) will produce larger fragments and a coarse tion and physical properties.
particle size distribution, whereas subsequent
abrasion and erosion produce fine material. 3.6.2 Deformation of a homogeneous
Thus, the particle distribution will tend to sediment body
become finer down glacier (Dreimais and
Vagners 1971). Accompanying this variation in The simplest models of sediment beneath an ice
particle-size distribution will be changes in bulk mass consider it to be a homogeneous material
properties. with no internal structure. Typically the sediment
Sediment output from the basal system occurs is considered to obey a failure criterion, although
in a region of deposition. Provided the bed re- some models set this to be zero. At applied
mains unfrozen, sediment will be initially deposit- stresses below failure no deformation is consi-
ed from the base of the deforming layer. Such dered to occur. Once the criterion is exceeded,
deposition occurs when the frictional forces re- deformation begins, usually considered to be gov-
sulting from the underlying bed are greater than erned by a fluid-type viscous flow response.
local shear stresses. Lodgement will occur prefer- Subglacial sediment overlies much of the mid-
entially at surface roughnesses of underlying bed- latitudes of Europe and North America, on
rock or sediment, where these frictional forces are which substantial development and building has
largest, for example, where a large particle has occurred. As a result, the engineering properties
previously been deposited. Consequent groupings of these soils and sediments have been tested
of larger particles result and may form boulder extensively by geotechnical engineers. The result-
clusters or boulder pavements within sediments ing geotechnical theory is often used to predict
(Boulton 1976). Sorting as a result of deposition the failure stress of subglacial sediment; frequent-
tends to reinforce internal inhomogeneities with- ly the Mohr-Coulomb criterion is used (e.g. Boul-
in the sediment, and large-scale features, such as ton and Jones 1979; Boulton and Hindmarsh
drumlins, can be initiated by this process (Boul- 1987; Alley 1989b; section 1.2.3). The physical
ton 1987). properties of glacial sediment reflect the mode of
Over the last 10 ka, basal sediments deposited deposition, internal and sedimentary characteris-
at the grounding line of Ice Stream B have tics, such as particle-size distribution, and min-
created a delta tens of metres thick and tens of eralogy. As a result, the values for parameters
kilometres long (Alley et al. 1989a). Smaller ice such as friction angle and cohesion quoted, or
masses and outlet glaciers that overlie, or par- used in models, vary widely (e.g. Boulton and
tially overlie, active deforming beds will also Dent 1974; Beget 1986; N.E. Brown, Hallet and
produce a zone of deposition at the margin. Booth 1987; Clarke 1987b).
Meltback at the snout of such a glacier re- Other failure criteria are used in soil mechan-
vealing thick sediments does not necessarily ics. The Hvorslev failure criterion defines poro-
reflect basal conditions further up the glacier sity-dependent cohesion and friction functions
flowline, but may characterize marginal condi- (e.g. Das 1985) such that shear strength varies
tions only. with water content. The friction angle typically
Models of sediment properties and deformation 83

varies only slightly with porosity (Gibson 1953), Shear Fluid with yield
but the cohesion may be sensitively dependent on stress stress and shear
this parameter, and is often written: thinning
Bingham fluid
c'(ry)=co exp(-B -ry-), (3.1)
l-ry Fluid with yield stress
and shear thickening
where Co is the cohesion at zero porosity, '1 is the Shear thinning
porosity and B is a constant. This failure cri-
terion is used by Clarke (1987b), with the depen-
dence of the friction angle on porosity taken to Newtonian
be negligibly small. The Tresca and von Mises
criteria use different combinations of the princi- Shear thickening
pal stresses to define failure conditions for soils.
There is no evidence that these would prove o~--====~----------
superior to the well known Mohr-Coulomb cri- Shear rate (1)
terion, and to my knowledge have not been used
Figure 3.4 Schematic stress-shear rate curves for
for subglacial sediments. idealized fluid flow, illustrating the terminology used in the
In fact, the traditional approach of soil mech- rheological approach to modelling sediment deformation.
anics may not turn out to be entirely appropriate
for sediments under the conditions of high poros- than a Newtonian or plastic behaviour. Other
ity, and often high strain, typical of ice-basal fluid models are typically non-linear with
environments. Geotechnical engineers are pri- strain rate, with the sediment displaying either
marily interested in non-deforming sediments, or shear thickening or shear thinning (Figure 3.4), or
in very small strains - deformation would typi- some more complex behaviour.
cally result in some form of failure of a human- Evidence for non-linear flow behaviour in sub-
built structure, a clearly undesirable outcome! A glacial sediments arises from a number of sour-
clear-cut concept of failure stress has an obvious ces. In situ measurements of strain rates beneath
place in soil engineering, although creep, in the the margin of a glacier in Iceland resulted in the
sense of a time-dependent strain that can occur extrapolation of an effective-pressure-dependent
below the failure condition, is acknowledged. It flow law (Boulton and Hindmarsh 1987) of the
may be that on a glaciological time-scale, sedi- form:
ment deformation will occur at all applied
stresses (Barnes and Walters 1985; Boulton and (3.3)
Hindmarsh 1987).
In this view, once deformation has begun, the
sediment is typically considered to deform as a where a, band Lb are constants, and T represents
viscous fluid with a characteristic flow behaviour. the shear stress. However, marginal conditions
Often a linear Newtonian or Bingham fluid is are characterized by high shear and low normal
used (e.g. Boulton and Hindmarsh 1987; Alley et stresses, and by what may be higher than average
ai. 1987b; Menzies 1989b) such that sediment water contents; basal conditions here
may be unlike those further from the terminus. At
1 Trapridge Glacier, away from the ice margin, in
11=- (a' -T y), (3.2) situ measurements of sediment deformation
(made using tilt-cells inserted through boreholes
where 11 is the resultant strain rate and J1 is the in the glacier ice into basal sediments) show
Newtonian sediment viscosity, assumed constant temporal variations of strain rate that appear
in this model (Figure 3.4). A Bingham fluid shows poorly correlated with fluctuations in basal water
no variation of viscosity with either basal condi- pressure (Blake, Clarke and Gerin 1992). These
tions or sediment strain rate. The measurements temporal variations should result from variation
obtained from beneath the Lower Columbia Gla- in either applied shear stress, or effective pressure
cier, Alaska, by Humphrey et ai. (1993) indicated driven by pore-pressure variation, if the Boulton-
that the till rheology had to be more complex Hindmarsh equation (3.3) is universally valid.
84 Glacial deformation

Unfortunately neither of these parameters has non-linear dependence of either failure strength
been measured except close to the ice margin or residual strength on strain rate (Skempton
(Boulton et al. 1979; Boulton and Hindmarsh 1964, 1985; Kamb 1991). Purging coarse material,
1987). Because pore-water pressures within the which is required for most small-scale laboratory
sediment have yet to be measured beneath a experiments, may turn out to be a major problem
realistic thickness of glacier ice, basal water pres- when extrapolating results of such experiments to
sure measured just above the ice-sediment inter- the subglacial environment.
face is often used as a surrogate variable. The
shear stress is usually assumed to be constant 3.6.3 Deformation of a structured sediment
and to be driven by ice and bed surface slopes body - anisotropy and inhomogeneity
alone. The validity of these assumptions is un-
known. Sediment is not well represented by a homo-
Subaerial debris flows also display non-linear geneous fluid; structure exists at many scales: (i)
deformation behaviour (e.g. Iverson 1985; Phil- subglacial sediments are inhomogeneous. Such
lips and Davies 1991). Such debris flows often inhomogeneity is reflected in spatial changes in
consist of material with a wide particle-size dis- sediment internal structure, in particle-size dis-
tribution, although size distributions appear to tribution and in particle arrangement. (ii) Sub-
be unimodal and denuded in both fines and very glacial sediments are anisotropic. The anisotropy
coarse material compared with tills (Johnson and results from previous deformation and deposition
Rodine 1984; Pierson and Scott 1985; Major and processes.
Pierson 1990). Porosities vary widely (Pierson Sediment properties are expected to be both
and Scott 1985) and the sediment mayor may inhomogeneous and anisotropic, with deforma-
not be saturated. For these reasons, direct ex- tion occurring preferentially at weaknesses within
trapolation of flow laws derived for such material the sediment. Localized perturbations in sedi-
to sediment deforming in the subglacial environ- ment properties can cause regions where the
ment may be inappropriate, especially since con- failure criterion is exceeded, although this is not
ditions of effective and pore pressure differ con- the case for the entire bed. In these regions
siderably. deformation will occur preferentially, and result
Further evidence for non-linear variation of in relative motion between parts of the bed.
strain rate with applied stress comes from labora- The deformation of subglacial sediment, result-
tory study. Sediment shear strength is dependent ing in property changes and anisotropy, is dis-
on the work done during deformation, and is cussed by both Clarke (1987b) and Menzies
made up of components related to friction (1989b), who predict a hysteresis behaviour for
and cohesion effects, particle rearrangement and the sediment. Three states are assumed: an immo-
dilation (Rowe 1962), and hence is deformation- bile pre-deformation state; an active mobile state;
mechanism dependent. Therefore, it is important and an immobile post-deformation state (Men-
that such laboratory studies should realistically zies 1989b). Sediment in the post-deformation
reflect basal conditions, and that failure within a state is not assumed to have the same properties
laboratory apparatus should occur by similar as that in the pre-deformation state. This predic-
mechanisms to those causing deformation be- tion is consistent with fabrics seen in naturally
neath a glacier. Laboratory experiments on deformed sediments, which suggest that an initial
concentrated suspensions, including suspensions assumed random structure becomes non-random
of matrix material from subglacial sediments, after deformation, such that the particles are
deformed at high porosity suggest non-linear aligned parallel to ice flow (e.g. MacClintock and
deformation behaviour dependent on strain rate, Dreimanis 1964). After deformation most flow
with both shear thickening and shear thinning parameters will be anisotropic in behaviour with
occurring (Cheng and Richmond 1978; Beazley respect to directions both parallel and perpen-
1980; Murray 1990; Major 1993), although strain dicular to the direction of active deformation.
rates are orders of magnitude higher than those While there is clearly feedback between the
occurring subglacially. At lower porosities, deformation, basal conditions and sediment
measurements using subglacial sediment in a geo- properties, this is as yet unrecognized by models
technical apparatus may also result in a of sediment deformation. Interpretation of sedi-
Basal processes 85

ment structure ansmg from deformation must pressure and effective pressure, and thus on de-
require consideration of both the anisotropic formation.
and inhomogeneous nature of subglacial sedi- An ice mass overlying a significant thickness of
ments. After all, deformation of homogeneous sediment may have a hydraulic system with a
sediment provides only minimal structural fundamentally different nature than an ice mass
information! overlying lithified bedrock (Figure 3.5). Because
bedrock is often highly impermeable, water flow-
ing beneath a glacier with a hard bed does so
3.7 BASAL PROCESSES AS A largely at the interface between ice and bedrock,
either in channels (Rothlisberger 1972; Nye 1976),
CONTROL ON DEFORMATION linked cavities (Walder 1986; Kamb 1987), or as a
thin sheet beneath the ice (Weertman 1972). In
3.7.1 Hydraulic processes contrast, the sediments comprising a unlithified
bed may be both relatively permeable and active-
Sediment deformation is controlled by both ly deforming. Although both sheet (Alley 1989b)
porosity and effective pressure. These para- and channelized flow (Shoemaker 1986; Alley
meters are in turn controlled by the amount of 1989b) have been hypothesized to occur, water
water present at the bed of an ice mass, its distri- transport can also occur within the bed by advec-
bution, throughput rate and throughput mechan- tion with deforming sediment, by Darcian flow
isms. The configuration of the hydraulic system is through the sediment, or preferentially through
a fundamental control on the distribution of inhomogeneities within the sediment (Walder
water, and on sediment porosity, pore-water and Fowler 1989).

INPUT: Meltwater, rain water

- - - - - - - - - - - - - - - - Supraglacial streams

THROUGHPUT: Englacial channels, ICE

INPUT: grain boundary flow _ OUTPUT:
Englacial. Englacial outlet streams
basal melting THROUGHPUT: R-channel

~""""f- ,T,f-!~9,~<2tl~~L !"l,-9~~~~~L",

INPUT: Meltwater, rain water

• THROUGHPUT: Englacial channels, Supraglacial streams
INPUT: grain boundary flow
basal melting THROUGHPUT: R-channel Englacial outlet streams

.~~------------ THROUGHPUT: Sheet flow _ OUTPUT

I~~ii~alii!i~.~ 6~~~~~HPUT M,octioc

_ Sheet flow - OUTPUT

(b) ~~ ~ ~~~~cFr'{Q~~~9,~Q~~S, ~/~TERlAL." ,~
Figure 3.5 Schematic diagrams showing throughflow processes in hard and soft bed hydrological systems. (From Murray
and Dowdeswell 1992.)
86 Glacial deformation

The basal hydraulic system evacuates water to Since deformation results from the spatial re-
the margin unless it is lost, for example, through arrangement of particles within the sediment,
an underlying aquifer (Alley 1989b). How the deformation alters sediment properties. Shear de-
basal water is distributed between the flow mech- formation may lead to dilatancy and thus to
anisms discussed above will strongly affect basal increased sediment permeability. Consolidation,
conditions, and the location of water at the bed. in contrast, decreases porosity and hence perme-
Water transported by channelized or sheet flow ability. These processes act in competition, con-
at the ice-sediment interface may not affect the trolling sediment permeability. Resulting local
underlying sediments greatly; an ice mass that changes will affect both the rate and route of
evacuates its basal water largely by these mech- water movement through the sediment and may
anisms is analogous to a hard bed ice mass, and also affect sediment porosity and pore-water
may have little communication with the underly- pressure. Variation in porosity has been shown to
ing sediments. However, because sediments are control sensitively both shear and residual
permeable, water can flow into, and through, the strength (Henkel 1960; Maltman 1987), and to
bed. Because basal sediments are not homo- control viscosity (Beazley 1980; Murray 1990),
geneous such water flow will be localized. Where with the resistance to deformation decreasing
sediments are anisotropic, permeabilities will also with increasing porosity. Thus, the distribution of
be anisotropic. The inhomogeneous nature of the water at the bed is a critical control for the
basal sediments suggests that water flow and the deformation of basal sediments.
resulting water distribution within the sediments
will also be inhomogeneous. Channelized flow
3.7.2 The consolidation-dilation
within basal sediments represents one such hete-
rogeneous water-content distribution. Water flow competition
may also occur through connected regions of The processes of consolidation and dilation act in
high local permeability, which could result, for competition in a sediment beneath a glacier or ice
example, from changes in particle-size distribu- mass. A lodged, inactive sediment experiences
tion or sediment dilation. overburden that tends to consolidate it. After
Open channels within basal sediments will retreat of an ice mass, the sediment will be
tend to close by creep and viscous flow (Alley overconsolidated and may preserve dewatering
1989b) and the flow of water through such chan- structures and a horizontal fabric associated with
nels may be unstable. In contrast, preferential consolidation. Shear deformation will, in con-
flow through sediment is likely to be a stable trast, tend to dilate sediment, increasing its water
process. Water flowing through non-uniform content. The material changes that accompany
sediments will take the path of least resistance dilation resulting from shear deformation, includ-
through the bed and will tend to follow paths of ing the increase in porosity, tend to weaken
high hydraulic conductivity through connected sediments. At Skalafellsj6kull, southeast Iceland,
regions of enhanced porosity and reduced tor- the upper, dilated layer is reported to be much
tuosity (Arch and Maltman 1990; Murray and weaker than the underlying layer as measured by
Dowdeswell 1992). Water flow by processes that shear-vane tests (Sharp 1984). These sedimento-
are essentially Darcian will be enhanced along logical changes mean shear deformation will tend
sediment unconformities, shear planes and other to be self-perpetuating. Consolidation also tends
lines of weakness. These water paths may tend to to be self-reinforcing; loss of water strengthens
reinforce themselves in positive feedback, as the sediment and makes shear deformation less like-
flow itself can remove fines thus increasing the ly. The feedback loop is complicated, and made
local sediment permeability. more subtle, by changes in the pore-water pres-
Heterogeneous water flow through channels, sure and permeability that accompany these
or along preferred paths through the sediment, processes. As in other geological settings, as
will create inhomogeneities of porosity and, po- sediment consolidates, permeability is reduced,
tentially, fluid pressure gradients between regions evacuation of water through sediments is re-
of the bed. Thus the hydraulic system itself has tarded and pore-water pressures may rise. Defor-
the potential to create conditions for heterogen- mation might be facilitated by this increase in
eous deformation within the sediment. pore-water pressure while consolidation is
Effects of deformation 87

occurring. In contrast, shear deformation accom- 1988; Feeser 1988; Owen and Derbyshire 1988);
panied by dilation may increase permeability, at the macroscale through structures such as
facilitating the dissipation of pore pressure. These deformation zones and internal sediment sorting
feedback systems are as yet incompletely under- (e.g. N.E. Brown, Hallet and Booth 1987); and
stood in the glacial environment. at the very large scale through mega-structures
such as flutes, drumlins, moraines and tunnel
valleys (Boulton 1987). Deformation structures in
3.8 EFFECTS OF DEFORMATION homogeneous sediments, or arising from homo-
geneous deformation, may be particularly diffi-
3.8.1 General cult to decipher, whereas those in inhomo-
geneous sediments will be emphasized by con-
The deformation of sediment beneath a glacier
trasts in properties and characteristics that will
results in the spatial rearrangement of particles
better record the results of deformation. Where
within the sediment, and potentially in the
sediments are inhomogeneous, deformation will
development of three-dimensional form. Such vary spatially. Resulting structures will reflect
form develops at the upper and lower interfaces localized changes in deformation, and will be
of the sediment as deformation or deposition
rendered visible firstly by variation in properties
cause spatial changes in the thickness of the
and secondly by changes superimposed by the
sediment layer and, hence, the formation of
deformation process.
ice-parallel or ice-transverse features. Although
In the following sections the record that will
this book is concerned with structures formed
result within basal sediments from the processes
within sediments, in the glacial situation these
of consolidation and shear deformation at all
structures are so closely associated with the
scales is described.
surface form of the sediment mass that a brief
outline of sediment form is given here. Together,
the spatial arrangement of solid matter and the 3.8.2 Features arising from consolidation
three-dimensional form compose the structure of
the sediment (Figure 3.6). This structure develops Sediment structures resulting from consolidation
in response to a complex set of driving processes include those of dewatering, overconsolidation
and range of basal conditions. The structure and jointing (Harrison 1958; Boulton 1976;
developed reflects the deformation history of the Dowdeswell and Sharp 1986; Feeser 1988;
sediment. Menzies and Maltman 1992). Dewatering may
Sediment parameters + Basal conditions + Processes => Sediment structure

Mineralogy Pressure Consolidation

Particle size/shape Temperature Loading/unloading
Pore water chemistry Hydraulic system Freezing/thawing
Sediment generation
Pressure, temperature and time effects

Figure 3.6 Processes determining the structure of a subglacial sediment.

Most structures developed subglacially result lead to the alignment and intrusion of fines, fluid
from sediment inhomogeneity and its reinforce- injection features, layer rupturing and localized
ment by inhomogeneous deformation and strain- fluvial sorting (Lowe 1975; Derbyshire, Edge and
induced anisotropy. Structure occurs at all scales; Love 1985; Nocita 1988; Owen and Derbyshire
at the microscale through structures in matrix 1988). Sediment mixing or the development of
material (e.g. Derbyshire 1978; Love and consolidation laminations may occur as pore
Derbyshire 1985; Arch, Maltman and Knipe water is expelled from the sediment. Particle
88 Glacial deformation

alignment can result within the sediment and

individual grains may become crushed as the
solid skeleton readjusts to occupy a smaller
volume (Owen and Derbyshire 1988), with grain
size and grain-size distribution playing an
important role in the features developed.

3.8.3 Features arising from shear

The shear deformation of sediment requires that
particles move relative to one another, either
sliding over, or rotating past one another (Schaef-
fer 1990). Such shear deformation may result in
dilation accompanied by an increase in porosity
(Reynolds 1885), particularly while it is being
initiated. Evidence for dilation of actively de-
forming sediments beneath glaciers arises from
the structure of Icelandic tills, and the high
porosity inferred for sediment beneath Ice Stream
B (Engelhardt et al. 1990). The effects of this
dilation have been observed at the margin of
Icelandic glaciers with thick sedimentary beds,
where it results in a two-layered structure
(Kozarski and Szupryczynski 1973; Sharp 1984;
Boulton and Hindmarsh 1987). At Breidamerkur- (b)
jokull, the upper layer that has undergone shear Figur. 3.7 Some examples of the microscopic effects of
deformation is dilated, with a porosity reported deformation during the deve!opment of shear zones in
to be ",,0.4. Below this lies a layer of largely direct shear tests. The sediments were collected from the
undeformed sediment with porosity '" 0.2-0.3 margin of a Norwegian glacier and were then deformed
in the laboratory. (a) Dilation of sediments. Long side of the
(Boulton and Dent 1974) that appears to have photograph is 11.3 mm in length. Void space shows black.
deformed along discrete planes. Laboratory de- (b) Particle alignment in scanning electron micrograph.
formation experiments of sediments collected at Scale bar at top is 20 IJm in length. (From Murray and
the ice margin have also resulted in dilation Dowdeswell 1992.)
(Figure 3.7a; Boulton, Dent and Morris 1974;
Murray and Dowdeswell 1992), although it is shearing (Boulton and Hindmarsh 1987). Particle
suggested that such dilation might be suppressed alignment results from the rotation of elongate
by the occurrence of particle crushing at high particles to occupy positions of minimum torque,
normal stress (Boulton, Dent and Morris 1974). If either parallel or perpendicular to deformation.
such particle crushing occurs, then this feature, Particle alignment occurs within both the clastic
normally associated with high consolidation, material (e.g. Dowdeswell and Sharp 1986) and
could also be associated with shear deformation. the matrix (Sitler and Chapman 1955; Ostry and
Deane 1963; Derbyshire 1978; Owen and
Particle alignment Derbyshire 1988), although alignment within the
Particle distribution is non-random in space and matrix material will be heavily controlled by the
previous deformation is recorded as a preferred presence of clasts. The direction of flow of matrix
particle orientation (Figure 3.7b). Deformation material will be deflected close to larger particles
is characterized by particle alignment (e.g. as the matrix flows relative to these clasts. Fur-
Love and Derbyshire 1985) and structures thermore, where sediments are coarse, matrix
elongated by high strain in the direction of motion may be dominated by clast dynamics,
Effects of deformation 89

resulting in irregular matrix flow and consequent been suggested to be deformational in ongm
sediment mixing (Engelhardt, Harrison and (Boulton 1987; Boulton and Hindmarsh 1987;
Kamb 1978; Hart and Boulton 1991). Menzies 1989b).
Deformation, and hence particle alignment, Flutings (Figure 3.8b) are commonly observed
can occur either along discrete slip planes, result- where glaciers override deformable sediments.
ing in the formation of shear zones (Sitler and Such features are thought to be formed where
Chapman 1955; Owen and Derbyshire 1988; sec- glacier ice flows past boulders at the ice-sediment
tion 9.4.1) in which intense particle alignment interface, forming a cavity at the downstream
may occur, or as bulk deformation involving side. This cavity fills with soft sediment and forms
moderate thicknesses of sedimentary material. a flute (Boulton 1976). If such flutings become
Shear zones may be slickensided (Boulton and attenuated and form streamlined features, drum-
Hindmarsh 1987; section 9.4.2); considerable linoid features can result. A similar process of
strain can be accommodated by shear along such sediment intrusion is thought to produce cre-
slip planes. vasse-fill ridges, which form at the margin of
The alignment of particles, and in particular surge-type glaciers (Figure 3.8c). Weak, water-
the alignment of matrix particles, will result in saturated sediments are intruded into bottom
anisotropy within the basal sediments and in crevasses within the ice forming features trans-
their deformation properties. Deformation, water verse to flow (Sharp 1985).
throughflow and other processes will be facili- Drumlins are streamlined features formed at
tated in a preferred direction resulting from the the base of ice masses. Sedimentary evidence for
deformation-derived anisotropy (Murray and their mode of formation is equivocal; one set of
Dowdeswell 1992). theories suggests that they arise from spatial
variation in subglacial deformation resulting
Shear deformation of inhomogeneous material from changes in sediment rheology or basal con-
The shear deformation of inhomogeneous ditions (Smalley and Unwin 1968; Evenson 1971;
material will commonly be localized. Deforma- Boulton 1979, 1987; Smalley and Piotrowski
tion of low-viscosity material around more 1987). Alternative hypotheses suggest that the
competent inclusions can result in streamlined formation of these features results from: infilling
features, analogous to drumlins, lending credence of subglacial cavities; deposition from sediment-
to the formation of at least some of these features rich basal ice; or preferential lodgement.
by inhomogeneous deformation (Figure 3.8; If drumlins arise from spatial inhomogeneities
Hart, Hindmarsh and Boulton 1990). At very within the deforming sediment there is no short-
high strains such inhomogeneities may become age of opportunities for their formation. Regions
attenuated and preserved only in the horizontal of inhomogeneity within the sediment can result
direction, resulting in a secondary layering being from regions of changing sediment-size distribu-
formed within sediments (Hart, Hindmarsh and tion, porosity, or basal stress. The most obvious
Boulton 1990). At extreme strains, even this cause of sediment inhomogeneity is the occur-
layering may disappear as lamination thick- rence of a large boulder within a much finer
nesses are attenuated to the average grain size matrix. In this case the drumlin core will be
within the layers, resulting in mixing (Boulton completely undeformable, although the boulder
1987). must be moving more slowly than the surround-
ing material in order to produce a streamlined
form. Rock-cored drumlins have been reported
3.8.4 Development of sediment form by Hill (1973) and Minell (1973). Regions of high
Large-scale features of sediment form arise on permeability resulting from a local coarsening of
scales from decimetres to the scale of land- the sediment will potentially have lower pore
forms. They arise from large-scale changes in pressures, and hence higher effective pressures. As
sediment properties, basal boundary conditions, this tends to increase the shear strength of the
or glacier forcings. Although there is contro- sediment, this region of sediment will have less
versy as to the genesis of many large-scale tendency to deform than the surrounding sedi-
forms, features such as glacier flutings, tunnel ment and may form a stiff inhomogeneity within
valleys, Rogen moraines and drumlins have the sediment. Drumlins formed in this manner
90 Glacial deformation


Figure 3.8 Some examples of the macroscopic effects of
deformation. (a) Inhomogeneous deformation of material
around a competent inclusion of material incorporated from
the lower, undeforming, region. Some streamlining of the
inclusion has occurred. Intense shearing of sediment
material has resulted in the formation of a secondary,
deformation-induced layering. Inhomogeneous deformation
on a larger scale could have lead to the formation of
subglacial mega-scale features such as drumlins. Feature is
exposed at West Runton, East Ang!ia, UK. Pen for scale.
(From Hart, Hindmarsh and Boulton (1990). Reproduced
(b) with permission from J.K. Hart.) (h) Fluting at the margin of
will have a coarse-grained core. Drumlins cored Trapridge Glacier, Yukon Territory, a quiescent surge-type
glacier. Flutes initiate at single boulders at the ice-sediment
by coarse grained material have been reported by interface. (c) Crevasse-filled ridge at the margin of
Evenson (1971), Whittecar and Mickelson (1979) Trapridge Glacier.
and Stanford and Mickelson (1985). Where a
predominantly silt-rich bed exists with local fin- in relative motion between parts of the bed.
ing, drainage may occur around rather than Stationary, or slowly moving, portions of the bed
through clay-rich areas due to the lower perme- form an obstacle to faster flowing sediment, re-
ability of the finer material, and hence fine ma- suIting in erosion of material upstream of the
terial can also act as a drumlin core (J.T. Wilson obstacle, and deposition downstream; a stream-
1938). lined form thus occurs.
Contrasts in sediment rheology (Figure 3.8a), In order for sediment inhomogeneity to be
or in the stress that is driving deformation, results acting as a core for drumlin deposition it must be
Preservation of features 91

moving more slowly than the surrounding sedi- production of a high-porosity, low-viscosity
mentary bed. Although this is possible in the bulk 'slurry' at the glacier bed (A.S. Jones 1979;
sediment, due to the relative displacement of Clarke, Collins and Thompson 1984) could facili-
portions of the bed, the bed forms produced will tate the high velocities of surging. Collapse of the
be more likely to survive erosion and transport if drainage system must result from either channel
the inhomogeneity occurs at or near the base of closure or a large-scale reduction in sediment
deformation (e.g. Menzies 1989a). Such in- permeability. Clarke, Collins and Thompson
homogeneities near the interface between deform- (1984) suggest that such a change in permeability
ing and undeforming material are more likely to could result from consolidation or be caused by
initiate streamlined bed forms than those within substrate deformation.
the bulk sediment. Irregularities in this interface Certain sediment forms have been identified as
may also act as zones for drumlin initiation characteristic of surge-type glaciers, such as cre-
(Boulton, 1987; Smalley and Piotrowski 1987). vasse-filled ridges (Sharp 1985, 1988), and both
fluting and remoulding of sediments also occur,
suggesting the presence of highly mobile sedi-
3.8.5 Sediments beneath surge-type glaciers ments. These features may be destroyed or rewor-
A surge-type glacier is an ice mass that experien- ked by subsequent surges unless the surge is a net
ces a cyclical flow regime. The cycle is character- depositional process, or unless climatic amelior-
ized by a quiescent phase of ice stagnation or low ation results in successively decreasing ice ad-
surface velocity and an active phase of high vances. It is extremely unlikely that surface fea-
surface velocity. The cycle is not related to clima- tures would survive subsequent surges, although
tic forcing or external parameters but is inherent features lower in the sediment layer might do so.
within the glacier system (Meier and Post 1969). Net deposition of sediment is likely at the margin
Where a surge-type glacier overlies deformable unless all the sediment transported by the basal
sediments, the cyclic flow regime should drive system is removed, for example, by flushing out
cycles of deformation within the sediments. These with basal water. Although the termination of the
sediments may then provide a record of surge surge of Variegated Glacier, Alaska, was marked
activity. by a release from the bed of a pulse of extremely
The active phase of surging is considered to turbid water (Kamb et al. 1985), it is unlikely that
result from either fast sliding or fast subglacial flushing would remove all of the sediment trans-
deformation rather than from fast creep within ported by a soft-bed glacier.
the glacier ice (Clarke 1987c). The mechanisms by
which a surge is initiated are, however, not well
understood, although surges are typically con- 3.9 PRESERVATION OF FEATURES
sidered to result from the destruction of a well-
developed drainage system, whether ice overlies a The glacial environment is poor for the preserva-
hard or deformable bed (Clarke, Collins and tion of features within sediment. Preservation
Thompson 1984; Kamb et al. 1985; Kamb 1987). requires a non-erosional environment, which will
The breakdown of the drainage system results in occur only near the ice margin (Boulton 1987).
water accumulation at the bed, as water input to Features formed in the extensional region of net
the basal system will exceed water output from it. erosion have little chance of survival. Shear de-
Rapid velocities then occur as a result of either formation leads to mixing of sediments (Kemmis
enhanced sliding on a water film, as obstacles are 1981; Boulton 1987) and the very high strains
submerged (Weertman 1969), or the activation of imposed on subglacial sediments attenuate fea-
a highly mobilized, water-saturated layer of de- tures rapidly. Hart et al. (1990) suggested that
formable sediments (A.S. Jones 1979; Clarke, Col- strains of greater than 700% may occur. Boulton
lins and Thompson 1984). (1987) estimated that at Breidamerkurjokull a
Where a surge-type glacier overlies a deform- feature within the basal sediments that started
able, unfrozen substrate, the stable, quiescent perpendicular would be rotated so that it was
water drainage is likely to be by channelized or dipping 1-6 upglacier within 1 year.

Darcian flow at sediment inhomogeneities. The Even if a feature survives to the glacier margin
destruction of these stable flow paths and the the pro glacial environment is very active, features
92 Glacial deformation

may commonly be reworked, and hence partially

or completely destroyed. Marginal processes
such as debris flows and mudslides are common
(Lawson 1979). Such processes will be especially
active in response to the high sediment-water
contents that will arise when ice retreat causes
blocks of ice to melt in situ, a process that was
probably common at the margins of the great ice
sheets at the ends of each of the ice ages. Mter
deposition and subsequent ice retreat, events that
are not necessarily contemporaneous, modifica-
tion of the sediment continues (Boulton and Dent
1974). Dessication, dewatering and flushing of
fines from sediment will occur both syn- and
post-deposition. With all of these opportunities Figure 3.9 Proglacial deformation of inhomogeneous
for alteration it is surprising that features are ever material resulting in folding. Feature is exposed at
Trimingham, East Anglia, UK. (From Hart (1990).
preserved! Reproduced with permission from J.K. Hart.)
Minimum requirements for preservation of
features or structure within the sediment are:
(i) a non-erosional environment, at least locally; deformation has been reviewed by Hart and
(ii) inhomogeneity in sediment; and (iii) rela- Boulton (1991).
tively stable marginal conditions. Luckily the
first occurs close to the ice margin, and the 3.10.2 Deformation of frozen-substrates--
second is superimposed by the deformation pro- basal-ice
cess itself, as changes in sediment properties
caused by deformation are recorded within the A glacier or ice mass may overlie frozen sedi-
sediment. ments. The lower portion of the glacier ice may
be basal ice, which may be rich in debris and
formed in close interaction with the bed. Such
3.10 OTHER TYPES OF GLACIAL ice is often laminated. A frozen substrate is there-
fore not a distinctive situation, and a glacier
DEFORMATION may show a gradation from basal ice to frozen
or unfrozen sediments without a sharp interface.
3.10.1 Proglacial deformation Melt-out of both basal ice and frozen sediment
may result in features within the unlithified
The deformation of unfrozen sediments by the material. Traditionally it was thought that
action of glacial ice is not limited to shear de- glaciers frozen to their beds would show no
formation of water-saturated sediments beneath basal slip and that their motion would be limited
an ice mass. Deformation of sediments also oc- to internal deformation of the ice. However,
curs subaerially in the proglacial environment. where significant debris exists within ice, it
Sediment continuity implies that the shear appears that the ice is weakened and may de-
forces applied to sediment, which are very high form more readily than the bulk of glacier ice
marginally, do not cease suddenly at the ice (Mathews and MacKay 1960; Brugman 1983;
margin (Hart 1990) but rather continue into Echelmeyer and Wang Zhongxiang 1987).
the glacier forefield. Compressive deformation Deformation rates hundreds of times greater
of the sediment results. Typically push moraines than in clear glacier ice are quoted (Echelmeyer
are formed at the ice margin transverse to ice and Wang Zhongxiang 1987), although it is
flow. Deformation in the pro glacial environ- not obvious why high concentrations of debris
ment is characterized by folding and thrusting, weaken ice, especially as low concentrations
and in East Anglia has resulted in the folding of foreign material appear to strengthen it
of laminated sediments beyond the ice margin (Nickling and Bennett 1984). The deformation
(Hart and Boulton 1991; Figure 3.9). Proglacial occurs as discrete slip along sediment bands
Conclusion 93

with an extremely high debris content (Brugman 3.11 CONCLUSION

1983), or as homogeneous deformation (Holds-
worth and Bull 1970), or both, and typically results Deformation processes, and their effect on the
in both particle alignment and banded debris. structure, fabric and physical properties of the
Deformation of debris-rich basal ice or frozen sediments, reveal subtle feedback systems be-
substrate results in strong particle alignment tween the sediment and subglacial conditions.
that persists after deposition (Dowdeswell and Clearer identification of these feedback systems,
Sharp 1986). During melt-out, formation of dis- and quantification of the effects of subglacial
continuous lenses is likely to occur (Paul and deformation on basal conditions will lead to a
Eyles 1990), which may produce weak strati- better understanding of both the conditions and
fication. Post-deposition sediments, thus, are the sediment structures, and of the response and
poorly sorted, underconsolidated and have a interplay between the sediment and the deforma-
strong particle alignment. tion process.

Sedimentary deformational structures


4.1 INTRODUCTION those structures that develop as a result of inor-

ganic, physical disturbance and with some prod-
This chapter is concerned with those structures ucts of early chemical processes in the sediments.
that develop as a result of deformation early in Specifically excluded are structures caused by
the burial history of a sediment. Historical as- organic activity within the sediment or at the
pects, including numerous literature references, depositional surface, such as burrows, roots and
are discussed by van Loon (1992). Some struc- animal surface tracks and trails. Also excluded
tures form very soon after or even during deposi- from detailed discussion are those structures that
tion and are sometimes referred to as penecon- develop as a direct result of sediment transport
temporaneous structures. In those cases, the and deposition. Depositional lamination,
deforming stresses commonly relate to the same whether due to fall-out of material from suspen-
processes that deposited the sediment. Other sion or the sorting processes associated with the
structures come about when some compaction migration of bedforms, is not discussed in detail
and/or diagenesis has been established. At what (for discussion of such features see Allen 1984,
point sedimentological processes give way to 1985; Collinson and Thompson 1989). However,
those producing features of concern to structural some appreciation of the geometry of such struc-
geology, and hence which structures to include in tures is essential to the full interpretation of
the present chapter, is a somewhat arbitrary deformed structures, as it is commonly those
judgement. Mass movements are dealt with sep- same depositional structures that are deformed.
arately, in the following chapter. All of these Early deformation, in many cases, has the
processes take place in loose, highly porous ma- effect of modifying the original structures and
terials and consequently, just as in other settings lamination that developed as the sediment was
of sediment deformation, the role of the inter- laid down. Many original structures may be still
granular fluid is paramount. recognizable, and their growth and geometry
Almost all sediments, with the possible excep- may have exerted a close control on the deforma-
tion of organically bound carbonates and deep- tion. In other cases, the deformation may modify
water oozes, are susceptible to disturbance after or destroy depositional structures to the point
deposition. This may happen at or close to the where they are no longer recognizable. In certain
depositional surface or following relatively shal- cases deformation may create new structures in
low burial. Deeper burial may also lead to defor- sediment that was depositionally featureless.
mation, as is the case with salt diapirism and
growth faulting, although the structures pro-
duced tend to be on a larger scale and be grada- 4.2 PRINCIPLES OF PHYSICAL
tional with structures related to tectonic defor- DISTURBANCE
mation. Early sedimentary deformation can
occur in a wide range of sedimentary environ- Before looking at the kinds of structures pro-
ments, ranging from aeolian dune sands to deep- duced during early sedimentary deformation it is
water turbidites. This chapter deals only with appropriate to consider briefly some aspects of

The Geological Deformation of Sediments Edited by Alex Maltman Published in 1994 by Chapman & Hall ISBN 0 412 405903
96 Sedimentary deformational structures

sediment deposition, and to recall some of the grain fall and grain flow are likely to have con-
mechanical principles outlined in Chapter 1. trasting packing, the former being somewhat
Most sediments laid down by physical processes more closely packed than the latter, which have
are delivered to the sediment surface either by high initial porosity.
fall-out from suspension or as the result of bed- Initial packing and porosity and the contrast
load transport. Sediments from suspension are in these properties between adjacent layers at a
most commonly of finer grain sizes - silt and clay variety of scales are amongst the prime causes of
- though coarser sediment may be delivered from potential instability in sediments and their sus-
suspension by particularly powerful currents. ceptibility to deformation. Initial conditions of
Clay and silt-grade sediments commonly have a deposition are not, however, the only controls on
high proportion of platy or acicular mineral sediment instability. After all, most of the sedi-
particles, which tend, on deposition, to have very ments seen in the stratigraphical record appear to
loose packing and consequently high initial po- have maintained their initial depositional con-
rosity. In addition, the presence of significant figurations without subsequent deformation
proportions of clay minerals means that the sedi- other than normal burial compaction. For sedi-
ment will begin to develop cohesive strength ments to become deformed shortly after deposi-
soon after deposition. Coarser suspension deposits tion, prevailing stresses must exceed sediment
are likely to be rather poorly sorted and also to strength for a period of time long enough for
have been deposited rather rapidly. Both these stresses to achieve perceptible deformation. In
features tend to give rise to loose packing and most cases, the stresses are the result of gravity,
high initial porosity. either acting vertically on layers of contrasting
Bedload sediments are always silt grade or density or as a shear component parallel with an
coarser, and tend to be better sorted and less rich inclined sediment surface. In less common cases,
in platy mineral grains. In the case of aeolian contractional stresses associated with volume
transport, grains are confined within a narrow loss and compressive stresses associated with
range of sizes spanning fine to very coarse sand. crystal growth may apply.
Particles transported as bedload may be deposi- The strength property involved is usually the
ted in one of two ways. The first is by becoming shear strength of the sediment, although in some
lodged in a stable packing position on the bed cases tensional strength may be relevant. As
over which they are rolling or bouncing. In that indicated in sections 1.2.3 and equation (1.5),
case the particle is most likely to lodge amongst shear strength is a function of particle cohesion
neighbouring grains of similar size so that the and intergranular friction, and the pore fluid
accumulating layer is fairly well sorted and the pressure (sections 1.2.5 and 2.2.3; equation 1.13).
packing stable. Fluctuations in current strength Therefore, shear strength can be reduced by a
may lead to the accumulation of parallel layers of loss of cohesion or a reorganization of grain
contrasting grain size. Progressive transport and packing to reduce tan cp, or through an increase
deposition under these conditions, as on a sandy in pore fluid pressure. Cohesive properties of a
flat beach subjected to moderate wave swash and sediment are largely a function of grain size
backwash, may lead to well-sorted sediment in a and the proportion of fine-grained, commonly
closely packed framework with low potential for clay, particles. These are inherent properties
instability. Higher depositional rates may lead to of the sediment and are not readily changed.
less good packing and higher initial porosity. The frictional component is most readily re-
The second way in which bedload sediment duced by an increase in pore fluid pressure, so
may accumulate is in the lee of a bedform such that more of the support of overlying sediment
as a ripple, dune or sandwave. There, deposi- falls upon the pore fluid and less upon grain
tion may result from grains falling directly to contacts.
the bed having been thrown over the crest of the Excess pore fluid pressures commonly develop
bedform in saltation or suspension. Deposition in near-surface sediment as a result of an initially
may also result from avalanche or grain flow high porosity being maintained during burial,
down the lee side due to oversteepening of the through the sediment having too low a permea-
upper slope by the addition of grains delivered bility to allow fluid to escape and thereby main-
mainly through rolling. Layers produced by tain the normal fluid, or hydrostatic, pressure
Principles of physical disturbance 97

(section 1.2.5). This is most likely in materials fluid, the elevated pore pressure may lead to a
dominated by fine-grained sediment, silts and complete or partial loss of strength and the
clays, particularly where they are rapidly deposi- sediment-fluid mixture may behave rheologically
ted. Such successions are underconsolidated and as a fluid or a plastic. Further aspects of the rapid
the excess pore fluid pressures (overpressured loading of sediments are mentioned in section
conditions) are likely to be sustained for long 1.3.2.
periods of time. Deformation processes may be Total loss of strength, the result of the pore
slow and long-lasting. The origins of overpressur- fluid pressure reaching lithostatic values, is
ing and the various ramifications for sediment known as liquefaction (section 1.3.2). The process
deformation are discussed in sections 1.2.5 and of fluid loss whereby grain contacts are re-
2.3. established and frictional strength restored will
In coarse-grained, loosely packed sediments, be by intergranular flow to the sediment surface.
loss of strength is likely to be short-lived (G. If this is sufficiently vigorous the moving fluid
Owen 1987). It may result from loose initial may carry particles with it and support them
packing or be triggered by some form of shock, in the fluid, the process known as fluidization.
causing instantaneous loading of the pore fluid, Fluidization, as a result of sediment dewatering
and a tendency for the grain packing to readjust is, of necessity, short-lived, as it depends upon
(Figure 4.1). Closer grain packing means excess the finite volumes of the excess pore fluid ex-
fluid. The increase in pore pressure consequent pelled. In natural examples it tends to be local-
upon the shock may result in loss of strength ized in narrow pipes and channels, within which
until the fluid has escaped through the permeable particles are completely reworked from their
sediment, usually to the surface, or the triggering original depositional packing. The grains may be
effect has subsided. During the relatively short significantly translated within the sediment, with
interval of time between shock and loss of excess elutriation of finer particles.

Shear strength
Shear strength (i.e. resistance To tighter packing
imparted by to deformation)
of sediment
grain-grain greatly reduced
contacts \ I
Un- I
disturbed Resedimentation
restoration of grain
I contacts; dilation
(9 if undergoing
Z shear
(/) Figure 4.1 Model for the loss of
shear strength in sands, its
restoration through the
DEFORMATION r-----t"""''------~r---t_--... TIME
establishment of tighter packing
MECHANISM Solid Viscous fluid I Plastic Solid and the types of deformation that
might prevail during these
None Flow None transformations. (After G. Owen
STYLE fracture
98 Sedimentary deformational structures

Such sediment-water mixtures are commonly to the preservation of most deformational

taken to behave rheologically as a Newtonian structures. Such structures represent, in effect,
fluid (sections 1.2.4 and 3.6.2), in which case snapshots taken at the point in the deformational
deformation will begin as soon as applied shear process where strength was re-established.
stresses exceed zero. The character of the de- In one group of structures, the sediment does
formation will therefore be controlled by the not lose strength through the breakdown of
viscosity of the mixture. For liquefied and intergranular friction, but rather it is disrupted
fluidized units, however, the dissipation of excess very locally through the effects of tensional
pore pressure and the draining of pore fluid will stresses. In some cases these act parallel with
both give rise to eventual and commonly rapid the depositional surface, such that the tensile
re-establishment of frictional contact between strength of the near-surface material is exceeded
grains and the return of strength to the sediment- and the sediment breaks up into a system of
fluid mixture. cracks. In order that the broken surface may
Where the sediment is less well sorted and be preserved it is also necessary that the sedi-
especially where there is a considerable content of ment has significant cohesive strength, which
clay particles, the sediment-water mixture will commonly means a significant content of clay-
have a cohesive component to its strength. It may grade material. In other cases, tensile strength
then behave as a rheological plastic (hydroplastic is exceeded locally within some larger deforma-
of Figure 4.3), which can sustain a certain value tional package, with the result that extensional
of shear stress before deformation begins. Once faults develop (Figure 4.2).
deformation does begin, it will continue until the The stresses generated by crystal growth,
applied shear falls below some critical value, especially of evaporite minerals at or close to
probably the shear strength of the sediment-fluid the sediment surface, are sufficiently large to
mixture (Figure 4.1). cause deformation of both the precipitating
For both fluid and plastic states, the rapid mineral layer and the host sediment in which
re-establishment of shear strength is crucial precipitation is taking place.

Exceed strength Reduced yield strength Liquidize
Brittle Plastic High pore
fluid pressure
1 Elevated
1 Creep Liquefied Fluidized
Gravitational body force
Slides Slumps Debris flows
on slope

Unequal confining load Growth Loaded Shale ridges

I Loaded ripples and sole marks
Clastic dykes
faults ripples lamination Sand volcanoes
>- Continuous lamination
~'"ffi ~v
C\l C
cW'" Within a single ~~ ,<,,0 Dish Water-escape
.Q"O 53 layer ::J
",<vv structures pipes and pillars
l§~:c ... (J)
.-.0 C\l
> Multiple layers c c
Load casts
C\l ~ WW
... 01
(J) not pierced Ex
::J Multiple layers
Clay diapirs
I ISalt domes Bali and piliow/pseudonodules
o E Overturned
~ (J)
Current drag (/)~
(/)Ul Dish structures
Vertical ~v Pillar structures
Desic~ation ,<,,0
Physical syrfferll'sis
crac s
Q; ~oncretion
.c Chemical Crystal
0 growth
Biological turbation

Figure 4.2 A classification scheme for sediment deformation based on the interaction of loss of strength and the
nature of the applied force. (Modified from G. Owen 1987.)
Physical deformation structures 99

4.3 PHYSICAL DEFORMATION situation and reduction of shear strength of the

layers can lead to foundering of the denser layer
into the less dense layer, with the resultant for-
mation of a set of structures that can all broadly
Structures resulting from post-depositional de- be attributed to loading.
formation may be classified in a variety of Load casts are most commonly seen as defor-
ways based on different geometric and genetic mations of the interface between a sandy layer
criteria (Figures 4.2 and 4.3). This account adopts and an underlying layer of finer grain size,
essentially genetic criteria, which take into ac- commonly mud or silt. Where the interface is
count both the degree to which sediment strength seen, as on the base of a lithified sandstone
has been lost and the nature of the forces that bed, the sand is deformed downwards in a series
produce the deformation during the period of of, commonly, rather equidimensional rounded
reduced strength. lobes (Figure 4.4). Between the lobes are narrower


leld strength -+-iIf-------Negli

Relative pore ~----<Uo :->U
fluid velocity : ! 0
, i
Flow I :

structure nar~----f.--+-- Turbulent-_~

%H 2 0 ------+-------+------~
Rate H 2 0 -------+-------+--------~
, '
Primary Preserved
i: .. I.. NO~ preserved
Figure 4.3 Processes of sediment
deformation and the associated
structures , ,
physical characteristics, behaviour
Negligible--!f- Minor--.j_~ignificant ----l~
Elutriation of
and products. (After Lowe 1975.)
finer grains 'Hydroplastic' here refers to
Intrusions "ene",'''" con~ordant "I'" :Generally ___-.t
the state of a sediment-water
mixture with a significant yield
strength, that strength being the
!f-DiSh result of cohesive or frictional
forces. It is synonymous with
'rheological plastic' of the text.

4.3.1 Partial loss of strength and density grooves or channels, which represent zones of
inversion upward-moving mud (Figure 4.5). The sandy
lobes are commonly referred to as load casts
Structures falling within this group are character- whereas the narrower muddy zones are flame
ized by essentially vertical movements and are structures, so called from their appearance in
the results of gravitational body forces. They vertical section (see below). Although most com-
were introduced in section 1.3.2. The structures monly equidimensional in plan view, load casts
commonly result from the occurrence, within may also be elongate, often with a preferred
a sequence of sedimentary layers, of density orientation. This may reflect a gradient to the
inversion, so that a more dense layer overlies sediment surface such that there is a component
a less dense layer. This is a potentially unstable of downslope movement associated with the
It"""-,,wa structures

ing convolute lamination
Haluszczak see below). In some cases it
may be to discern the nature of the
original as in the case of loaded
ripples, where the cross-lamination can be seen.
Where the sediments show flame
structures show lamination
gin, with greater contortion the centre.
In most cases, load casts occur on the base of
a continuous bed of coarser sediment.
In some cases, however, the
foundering leads to the
figure 4.4 Load casts on the base of a sandstone bed coarse layer so that it occurs
in an interbedded sandstone-mudstone sequence.
~oad bans or floating in a matrix
The load casts do not appear to show any preferred
elongation. Bude Formation, Upper Carboniferous, north of finer sediment (Figure 4.6; Eschmann
Cornwall. 1992).

figure 4.5 Thin-bedded, line-grained sandstones interbedded with dark mudstones. The upper sandstones have very
small load casts on their bases, whereas the somewhat thicker beds in the lower half of the photograph show more
extensive loading, almost developed to the point 01 detachment. Complex flames of mud extend up between the load balls,
which have highly convoluted internal lamination. The mineralization of ihe small faulls suggests that they are later
tectonic features and unconnected with soft-sediment deformation. Bude Formation, Upper Carboniferous, north Cornwall.

loading. It may result from the process Diameters of load

accentuating some earlier elongate feature of the isolated,
layer interface, such as erosional sole marks or ness of layer
ripple forms in the upper layer. cases where loading has accentuated ""l"'~_pv",,,,
In vertical section, load casts commonly show structures, that not
internal contorted lamination (Figure 4.5). Close In some cases of
to the edge of the structures, lamination tends to difference in size
Physical deformation structures 101

Figure 4.6 Isolated sandy load balls or pseudonodules

floating in a unit of slurried siltstone that has
undergone total liquefaction. Pseudonodules are around
5 cm across. Sude Formation, Upper Carboniferous,
Figure 4.7 Convolute bedding in well-laminated sands.
north Cornwall.
The sharp anticline and the more gentle synclines are
and sand appears to have sunk into similar sand. typical of relatively simple convolute bedding and
lamination. The core of the anticline is characterized by
This occurs most commonly in thick-bedded, more intense deformation and the whole unit is truncated
amalgamated turbidites. It appears that quite above by an erosive contact. Recent river terrace
subtle differences in texture combined with the deposits, Tana Valley, Finnmark, Norway.
rapidity of the deposition lead to the instability.
Sand-on-sand loading is also produced in inter- also preferred routes for upward escape of excess
tidal sand bodies as the result of the entrapment pore water (Figure 4.8). In originally cross-bed-
of air during the rising tide. Overpressuring lead- ded sands, it is sometimes the case that deforma-
ing to loss of strength, as explained in section 4.2, tion is confined within a single set, with deformed
is commonly thought of in terms of pore water, laminae truncated by the overlying bounding
but in this situation compression of trapped air is surface. In other cases, the bounding surfaces
thought to lead to strength loss and consequent themselves may be involved in the deformation,
density inversion (De Boer 1979). Such load with multiple sets clearly having deformed to-
structures, within a more or less uniform lithol- gether (Lang and Fielding 1991).
ogy, are gradational in character, with convolute Some of the largest examples of convolute
lamination and bedding. bedding occur in sands deposited as aeolian dunes
Convolute bedding and convolute lamination are and which were liquefied beneath the water table
structural styles that occur most commonly with- following burial (e.g. Doe and Dott 1980). These
in a single bed, usually of silt- or sand-grade examples may involve sedimentary units up to
material. The terms are to some extent inter- several tens of metres thick (Figure 4.9). Deforma-
changeable, with a loosely defined size limit sep- tion commonly crosses set-bounding surfaces and
arating them. Deformation at the scale of deci- only rarely is truncated above by a bounding
metres or above would be 'bedding' and at the surface. More commonly, deformation dies out
scale of centimetres 'lamination'. The structures vertically both upwards and downwards. In some
are seen most commonly in vertical section, al- cases, deformed cross-bedding passes gradation-
though they also appear in plan view. ally into more or less structureless sand within
They commonly involve original depositional which it is possible to detect only very weakly
lamination being folded and contorted. In some defined, disturbed lamination. It appears that
cases the folds may be upright and cuspate, with these structure less units represent the most ex-
rounded synclines and sharper anticlines (Figure treme state of deformation, with the convolute
4.7). In other cases a more chaotic style prevails. cross-bedding being an intermediate state between
Where sharp anticlines are seen, these may reflect structureless sand and undisturbed cross-bedding.
both an internal foundering within the deformed Aeolian sands tend to be very wen sorted and
layer, akin to flame structures (see above), and to have high initial porosity, properties which
'102 Sedimentary deformational structures

may make them susceptible to loss of strength

when waterlogged (e.g. Doe and Dott 1980;
Horowitz 1982; Collinson, Bevins and Clemmen-
sen 1989). Total or partial liquefaction may be
triggered by a shift in the water table, seismic
shocking or rapid sediment loading. Exception-
any, convolute bedding in aeolian sands may
result from a rapid rise of the water table through
the dune sands as a result of rapid inundation of
a desert area by the sea (e.g. Glennie and Buller
In some beds that display convolute lamina-
tion or bedding, the intensity of deformation
increases upwards, often with undisturbed
lamination towards the base. Although folds are
often upright, overturning in a preferred direction
also occurs. In plan view, the lamination is com-
monly seen to be in the form of basins and
intervening ridges.
Convolute bedding and lamination records the
internal foundering of liquefied sediment layers
upon themselves, commonly in conjunction with
active upward escape of pore water. Normally
this appears to have been driven purely by gravi-
tational forces normal to the deforming layer.
Where anticlinal folds are overturned in a prefer-
red direction, these forces may have operated in
Figure 4.8 Convolute bedding within coarse-grained,
cross-bedded channel sandstones. The sharp overturned
conjunction with a down-slope component of
anticline suggests water escape, perhaps coexisting mass movement or with an applied shear from
with shear at the sediment surface. Roaches Grit, Upper water flowing across the top of the layer (see also
Carboniferous, Staffordshire. overturned cross-bedding).

West East
• • • • • • • • •~~~·-····-----------l
Indistinct bedding I

200 I
100';--0--0-- -0
~ g ~ g
'Flgure 4.9 Somewhat simplified diagram of large-scale convolute bedding in aeolian dune sandstones. Note that
contorted units pass upwards into units with indistinct bedding. An active water table was probably responsible for the loss
of strength, the contorted bedding reflecting short-lived and partial loss, the indistinct layers total liquefaction with loss
of lamination. (After Doe and Dott 1980.)
Physical deformation structures 103

Convolute lamination is commonly rather ir- The use of liquefaction structures for the charac-
regularly distributed in a sediment unit, often terization of palaeoseismicity is discussed further
localized within an otherwise intact bed. It ap- in section 9.5.2.
pears that in many cases the liquefaction and loss
of strength were spontaneous consequences of the 4.3.2 Structures due to progressive loading
early post-depositional state of the sediment layer of cohesive sediment
(e.g. Allen 1977), although external triggering
may be invoked where the disturbance is very Where fine-grained cohesive sediments are pro-
widespread or occurs in an unusual setting. In gressively but fairly rapidly buried by younger
certain turbidite sandstones, convolute lamina- sediments they may develop overpressure, and
tion seems to be especially associated with the lateral and vertical flowage may ensue. Gravi-
Bouma 'C' division of ripple cross-lamination tational instability may lead to the development
and appears to have developed virtually syn- of diapiric structures. The most common types of
chronously with deposition. structure produced by this type of behaviour are
The deformation of aeolian dune sands men- the mudlumps or mud diapirs which rise through
tioned above is an example where introduction of prodelta and mouth-bar sands of rapidly pro-
water is the destabilizing agent. In other cases, grading deltas. These are particularly well seen
heavy wave action on shallowly submerged aque- off the mouths of the major distributary channels
ous bedforms may cyclically load the sediment to of the present-day Mississippi Delta, where muds
instability (Dalrymple 1979; section 1.3.2). Where from a burial depth of many tens of metres rise
convolute bedding is confined within single cross- up through later sediments and, in many cases,
bedded sets, it is likely that the loss of strength emerge above water level as low temporary is-
was a product of rapid deposition and the grain- lands (Figure 4.10; Morgan, Coleman and Gag-
size characteristics of the sediment. Where defor- liano 1968). These are commonly eroded over a
mation involves several sets and their bounding number of years by wave action and the muds
surfaces it is more likely to have been triggered dispersed for redeposition. Internally, the diapiric
by an external agent, probably an earthquake muds commonly show steeply dipping discrete
(e.g. Reimnitz and Marshall 1965; Davenport and shear surfaces and a characteristic small-scale
Ringrose 1987). Where seismic activity is thought brecciation. There are many analogues with the
to have been the triggering mechanism, the areal large-scale diapirs and diatremes found, for
distribution of convolute lamination and bedding example, at convergent plate margins (section
within sands of fluvial channel origin has been 7.1). In addition, large-scale mud diapirs occur on
used as an indicator of which faults were syn- the lower parts of prograding continental slopes
depositionally active (Allen 1986a; Leeder 1987). off the fronts of major muddy deltas such as the

Vertical exaggeration =
2~ x I

1800m 1500 1200 900 600 300 0

Figure 4.10 Diapiric mUdlumps in the modern Mississippi delta, illustrated by a section along the axis of one of the
distributary channels. Note the high-angle reverse faults associated with the crests of the diapirs and the marked thickening
of the mouth-bar sands into the areas between the diapirs. Contrast the scale of these structures with the large diapirs
seen on the lower parts of progradational continental slopes (Fig. 4.26) (after Morgan, Coleman and Gagliano 1968).
"'104 Sedimentary deformational structures

Mississippi and Niger. These are associated with ularly well displayed, with both penetrative
the lateral flowage of deep, overpressured clays and non-penetrative types recognized (Pulham
from beneath the deltaic depocentre towards the 1989).
free surface (section 4.3.5). As well as deforming both the diapiric muds
The close association of mudlump diapirs with and the overlying sediments, this type of behav-
the site of most active and rapid sedimentation in iour also affects local depositional patterns, es-
the mouth bar, and the coincidence of mudlump pecially the thickness of overlying units. As dia-
emergence with periods of flood deposition, sug- pirs move upwards, mud is withdrawn laterally
gests that their growth is rapid and closely re- from an area around the mudlump or shale ridge,
lated to sediment load applied to the overpres- and this leads to the development of local depo-
sured layer. centres between and around the diapirs. Mouth-
The emergent mudlump islands are thought to bar sediments associated with diapirs show pat-
represent peaks or spines on the crests of shale terns of spectacular thickness change as a result
folds or ridges, which are initiated by the basin- (Figure 4.10).
ward lateral flowage of clays in front of the At a larger scale, diapiric deformation occurs
advancing mouth bar. As folds are progressively in association with the burial of thick units of
overtaken by this progradation, the movements evaporitic halite, which deforms in a viscous
take on a more. vertical component and, in a fashion to form large salt domes, pillars and
sense, the diapiric movements which then domi- sheets. Associated features, such as crestal faults
nate are akin to those associated with sediment and rim synclines, are broadly similar to those of
loading (see above). Ancient examples of mud- mud diapirs although on a larger scale. Deforma-
lump diapirism within deltaic sediments are well tion ofthis type takes place after deep burial and
known from the rock record (Figure 4.11), those occurs so long after deposition that it can hardly
of the Namurian of Country Clare being partic- be considered 'early'.

rilouth-b~r sands in a deltaic sequence. The mud is

derived from a unit 10 m thick. in the base coarsening unit and the whole deformation is confined
within that progradational interval. Other diapirs within the same succession are non-penetrative. Figure (arrowed) for
scale. Central Clare Group. Upper Carboniferous. County Clare. western Ireland.
Physical deformation structures 105

At a smaller scale and in other settings, defor- material moves. Slumps are distinguished
mation of this type is probably rather rare. It has, from slides in that the moving material is itself
however, been suggested for folding and convol- subjected to internal plastic deformation as it
ution in Devonian siltstones underlying a moves. The relative movements of different
laterally restricted set of large-scale cross-bed- masses of material within the moving layer
ding, probably of aeolian dune origin (B.G. Jones give rise to a spectrum of deformational struc-
1972). The folding in the siltstone is most intense tures. These range from extensional structures
under the axial part of the overlying dune set and with normal fault geometries in the upslope
asymmetric folds verge away from the axis (Fig- head region of the slump, to compressional
ure 4.12). It is suggested that the progressive folds and thrusts in the toe region. The
advance of the aeolian dune loaded the cohesive patterns may be very complex and difficult to
silts and caused them to flow laterally before they decipher in terms of processes. The topic is
were fully buried. dealt with in Chapter 5, and numerous illus-

Fold amplitude decreasing, wave length constant
~---------------------- ..
Folds contorted due to
loading by sandstone
. West

........... :."
. .

... ... '.:".';""
. -- ' . '. ' . ,- .
' . . .. .

.. " - ' ,'.

Preserved erosion surface


? Metres 2

Figure 4.12 Schematic reconstruction of the development of contortion in a siltstone unit as the result of the advance of a
large aeolian sand dune across its surface. Langra Formation, Upper Devonian, central Australia.

4.3.3 Partial loss of strength and applied

trations are provided of the resulting struc-
shear tures.
Structures described here are the product of a Overturned cross-bedding is a common style of
temporary loss of strength during a time when deformation in sandstones of shallow marine and
the weakened layer was subjected to a shear fluvial origin where powerful currents were in-
force, due either to the sediment resting on a volved in the migration of dune-scale bedforms.
slope (i.e. a downslope component of gravity Overturning of foreset laminae is confined within
body force) or to the effect of an overriding single cross-bedded sets, and is always in the
flow on the upper surface of the layer. The first direction of the foreset dip (i.e. down current)
situation gives rise to the phenomenon of mass (Figure 4.13). In some examples, the overturning
movement, which was introduced in section is restricted to the uppermost part of the set,
1.3.3 and is comprehensively reviewed in Chap- whilst in others virtually the full thickness of
ter 5. Perhaps the most commonly seen conse- the set is involved (Figure 4.14). In all cases
quences of the downslope movement of near- the degree of steepening and overturning in-
surface sediments are slump folds, the result of creases upwards through the set. Although most
the movement of weakened sediment on a cases involve simple overturned folds, some
slope, in response to the downslope component examples show more complex folding and/or the
of gravity. Within slumped masses, the greatest loss of definition of lamination (Allen and Banks
shear strain is commonly concentrated on the 1972; Hendry and Stauffer 1977; Doe and Dott
basal slip surface above which the slumping 1980).
106 Sedimen.tary deformational structures

Figure 4.13 Overturned

cross-bedding in coarse
sandstones of probable shallow
marine origin. Kap Holbrek
Formation, Lower Cambrian,
Danmarks Fjord, N.E. Greenland.

~o )
: bl
- - - -SOem -----

·)_ _ _--'-25 em
o SOcm

SO em

:11 ---

Figure 4.14 Variations in the geometry of overturned cross-bed foresets from a variety of settings. Note the variable
position of the lold axis. (After Allen and Banks 1972.)
Physical deformation structures 107

The structures result from the loss of strength crease upwards (Figure 4.1Sc; Allen and Banks
of the cross-bedded set very soon after it was 1972; Allen 1984). In Figure 4.1Sb, a tabular set
deposited, so that the overriding current, which of thickness h, which was instantaneously
produced the bedform in the first place, was still liquefied with a grain concentration, gc, is
active. The boundary shear stress exerted by this sheared by an overriding current such that the
current was able to shear the weakened sediment velocity of the surface layer is U. The instan-
layer and cause essentially laminar (viscous) taneous velocity (u) of sedimentary particles at a
shear within the layer (Figure 4.15b). Re-estab- height Yo above the bed is determined by the
lishment of shear strength within the deforming slope of the velocity profile, itself a function of the
layer appears to have taken place from the bot- applied shear and the viscosity (p) of the sediment
tom upwards as excess pore fluid escaped (Figure dispersion. The total displacement (X) of a par-
4.1 Sa). As a front of reconsolidation migrated ticle (Figure 4.1Sb) at a particular height above
upwards, higher levels in the set would be subjec- the base of the bed is an integral of its changing
ted to progressively longer durations of shearing, horizontal velocity (u) as the base of the liquefied
and hence the degree of overturning would in- layer moves upwards at velocity v. The resultant

200 400 600 800 1000


0.1 e
I<D 0.2
<U 0.3
~ (c) x

Base of bed
(a) 0.6

Convex-up Rectilinear



\ "

(b) Base of initial bed (d)

Figure 4.15 Mechanisms in the formation of overturned cross-bedding. a. The upward migration through time of the
front of redeposition following initial loss of strength. b. Definition diagram of simple shear operating on the liquefied
bed as the front of redeposition moves upwards. c. Deformed foreset trace of an initially planar foreset following
simple shear and progressive base-upwards resettlement. d. Influence of original foreset shape on the form of the
overturned foresets. (After Allen and Banks 1972; Allen 1977, 1985.) The symbols in (b) and (c) are discussed in the
text; further details are given in Allen and Banks (1972).
108 Sedimentary deformational structures

values of X through the bed determine the shape plan view, are roughly equidimensional dishes of
of the overturned foresets. A suite of different a few centimetres diameter (Figure 4.16). Pillar
curves may be generated depending on values structures have more or less vertically elongated
attached to particle concentrations in the zones which cross the dishes and often stem from
liquefied and consolidated layers, the particle their upturned edges.
settling velocities (Us) within the liquefied layer The clay particles, whose enrichment defines
and the shear rate. The curve generated on Fig- the dishes, are thought to have been carried
ure 4.15c derives from an initially planar foreset. upwards with the escaping pore waters and to
More complicated shapes arise from the defor- have been filtered out at slight inhomogeneities -
mation of initially more complex foresets (Figure possibly incipient lamination - within the sand.
4.15d). Once initiated, the reduced permeability of a
clay-enriched layer would tend to filter out more
4.3.4 Structures related to upwards escape clay and also force the escaping waters sideways.
of pore water and sediment-water mixtures Vertical escape would be increasingly around the
margins of the developing dishes, dragging them
As wen as the upwards deflection of depositional upwards, and concentrated in vertical conduits -
lamination sometimes seen in association with the pillar structures. These may exhibit a slightly
convolute lamination and bedding, the escape of cleaner texture owing to the elutriation of some
excess pore fluid to the sediment surface may also of the clay.
generate structures in its own right and some- In late Quaternary glacial outwash deposits,
times in sediment which otherwise would be Cheel and Rust (1986) described dish structures
structureless. as the last of a series of liquefaction products.
Very rapid deposition of sandy sediment from Their inferred development is illustrated in
suspension, as when a heavily loaded, power- Figure 4.17. Initially, local fluidization of low
ful current decelerates rapidly, will commonly permeability sediment at the base of the unit
lead to the trapping of large volumes of pore generated convolute lamination, with disrupted
water. This will tend to escape upwards as the anticlines evolving into ascending vertical dia-
sediment settles and the escape may take several pirs or coalescing flames. In the overlying, less
forms. weakened sediment, ball-and-pillow structures
Dish-and-pillar structures are subtle and rather formed, owing to the penetration of the rising
uncommon features which reflect the local clay diapirs and the foundering of more dense ma-
enrichment of small zones within sand (Lowe and terial above. The lower burial pressures at the
Lopiccolo 1974; Lowe 1975). The dish structures top of the layer allowed the fluid pressure to
are small, concave-upwards features which, in decrease, so that the rising material passed from

Figure 4.16 Dish structures in

thick, massive turbidite sandstone
bed. Note how the rather flat dishes
in the lower part of the unit pass
upwards into shorter, more curved
dishes higher in the bed. The
upturned ends of the dishes
connect with pillar structures which
acted as the main conduits of water
escape. Eocene, San Sebastian,
Physical deformation structures 109

t Water source t
(a) (b) (c) (d) Preserved

..... .
: : : : : :: Massive and t Upward flowing sediment-water
: : : : : :: stratified sand I mixture

I'.: ;: ~· · · · .~· .·I Silty fine sand

: Sinking blocks of

I sediment

Total thickness of sequence is up to approximately

10 m. No horizontal or vertical exaggeration.

Figure 4.17 A model for the development of convolute bedding, ball-and-pillow structures and dish structures within
glacial outwash deposits. The excess pore fluid pressure may have resulted from shocking or from the melting of buried
ice within the sediment. The genetic model is based on an idealized composite model (e) for the distribution of
the structures in the beds. (After Cheel and Rust 1986.)

a fluidized to a liquefied state. The final stages clasts. This apparently gave rise to pebble clus-
of upward movement of fluid in the liquefied ters and pocket structures, which are bowl-
sand generated dish structures at the top of the shaped concentrations of granules and pebbles a
unit. few centimetres wide (S.Y. Johnson 1986). In
Lowe (1975) envisaged dish-and-pillar struc- addition, some elongate pebbles may be reorien-
tures as developing optimally under conditions tated subparallel with the direction of water
of moderate water escape (Figure 4.18). Under escape.
conditions of normal upward seepage, original Sheet dewatering structures are similar to pillar
structures are undisturbed, whereas at high structures but are more linear in plan view and
rates of escape the liquefaction of the bed may may occur independently of dish structures in
be so great that major internal foundering otherwise structureless sands (Laird 1970). They
of sediment masses (internal loading) occurs. are often characterized by cleaner sand than that
Dish structures form in the middle parts of a of the host sediment and may be closely spaced
bed while the more rapid water escape to- within the bed. They may also coalesce vertically
wards the top leads to liquefaction and to the upwards, sometimes defining horizontal zones
development of convolute lamination (Figure (Figure 4.19).
4.18). In the case of the above structures, the escap-
In coarse-grained, poorly sorted sediments that ing pore waters locally modify the texture of the
appear to have been deposited very rapidly, sediment by preferentially enriching or depleting
water escape during a phase of liquefaction ap- it in particular components. The water escape
pears to have led to downward settling of coarser does not involve the wholesale translation
110 Sedimentary deformational structures


-; ... / >

(a) (b) (c)

Figure 4.18 Schematic variation

in water escape structures in
response to rate of water escape
and original depositional facies: I,
medium to coarse-grained
cross-stratified sandstones; IIA,
turbidites with non-cohesive
Bouma C and D divisions; liB,
turbidites with cohesive Bouma C
and D divisions. Left-hand column (d) (e) (f)
indicates undisturbed depositional
structures, preserved as a result of
low water-escape rates. Middle lIB
column, at moderate escape rates,
gives rise to dish structures and to
small water-escape pipes in
non-cohesive sediment with
convolute lamination disturbing the
upper parts of more cohesive units.
Highest rates of water escape
(right-hand column) give rise to
more general liquefaction and loss
of lamination and to more strongly (9) (h) _____ (i)
penetrative, vertical water-escape
pillars. (After Lowe 1975.)
Water escape rate

of sediment, but such bulk movement does, how- material in its overall composition in order to
ever, occur. The resultant structures are described impart some cohesion to the material. Dykes
below. vary in width up to several tens of centimetres
Sediment injection structures are increasingly and they may extend over many metres verti-
recognized in the rock record, mainly in the cally (Figure 4.20). Both dykes and sills may
form of dykes, although sills and associated ex- wedge out rapidly. Dykes may become folded,
trusive features are also sometimes found. His- sometimes into tight ptygmatic folds when the
torically, they were amongst the first sediment host has undergone significant compaction
deformation structures to be recognized and on burial (e.g. Truswell 1972). It is possible
some early investigations are recalled in section that sandstone dykes may be confused with
1.3.4. Sedimentary dykes occur in a range of sandy fills of desiccation and synaeresis cracks if
depositional settings, from deep-water turbidite seen only in vertical section (see below). As re-
associations (e.g. Truswell 1972) to subaerial marked in section 1.3.4, such passive fissure-fills
mass-flow deposits (e.g. Collinson, Bevins and are sometimes called neptunean dykes. In plan
Clemmensen 1989). They are also intruded into view, dykes are commonly random in orienta-
a range of host lithologies, whose main pro- tion but some sets show a polygonal patterns
perty must have been to have had strength at whereas others may show a preferred orienta-
the time of injection. In practice, this means tion.
that the host must have at least some clay Both dykes and sills are commonly made up
Physical deformation structures 111

of rather structureless sandstone, although

some show a marginal foliation parane! with
the wans, reflecting shearing during intrusion.
Where intrusions broke through to the contem-
poraneous sediment surface, sand volcanoes
and/or extruded sand sheets may occur (see be-
Sand dykes and sins are the result of the
forcible intrusion of liquefied sand into a cohe-
sive host. The intrusion is probably most com-
monly from below, although it may in some
cases have been sideways. In all cases the source
of the liquefied sand must have been buried by
less permeable cohesive sediment, in some cases
for a considerable period of time prior to intru-
sion. Some sandstone dykes, which cut across
and therefore post-date septarian carbonate
nodules (see below) in the mudstone, must have
been intruded quite a long time after deposition
(e.g. Martin and Hudson 1989). In fact, there is
no reason, in principle, why sedimentary dykes
cannot arise in deeply buried sediments, pro-
vided the overpressuring is sufficient to over-
Figure 4.19 Sheet dewatering structures in a thick, come any strength increase due, for example, to
massive turbidite sandstone bed. Note the upwards cementation or compaction that occurred dur-
coalescence of the sheets, with merging taking place at
discrete levels. Skipton Moor Grits, Upper Carboniferous,
ing the burial history. Section 9.5.2 mentions
Yorkshire. some examples that are thought to have formed

Figure 4.20 Vertical sandstone

dyke cutting through poorly sorted
conglomerate of mass-flow
origin. Weak foliation parallel to
the walls of the dyke suggest
shearing of liquefied sand during
emplacement. Moraenes€
Formation, Precambrian, north
112 Sedimentary deformational structures

at substantial burial depths in association with They occur on the tops of both sandstone and
regional tectonism. more muddy beds that have been overlain by
The process of liquefaction may have resulted fine-grained sediment deposited from suspension
from sudden shock, such as earthquakes (Gill and Kuenen 1958).
(Reimnitz and Marshall 1965), or the sudden Sand volcanoes are broadly conical mounds,
emplacement of the cohesive layer itself, as when ranging in diameter from a few centimetres to a

Figure 4.21 Schematic block diagram of the geometry of submarine sandy gulley fills in the Harelv Formation
(Jurassic) of east Greenland. The gulleys were cut into slope mudstones and filled my mass-flow processes. The
wing-like features on the flanks of the sandbodies are sand sheets intruded into the muds, probably as a
result of overpressuring of the succession during burial. (After Surlyk 1987.)

subaerial mass flows are deposited on water- few metres (Figure 4.22). They commonly show
logged sands (e.g. Collinson, Bevins and Clem- many of the features of igneous volcanoes but
mensen 1989). In other cases, the progressive on a much smaller scale, with central craters
burial of sands by fine-grained sediments may and flanking flow lobes. In vertical section they
lead to increasing overpressure in the sands show an axial pipe analogous to an igneous
until they eventually break out into the sur- vent, and inclined layers dipping radially away
rounding sediments. This appears to have been from the axis. In some cases it may be possible
the case in certain ancient channelized deep- to trace the pipe downwards into the underlying
water sands (e.g. Surlyk 1987; Newton and layer and identify the source bed from which the
Flanagan 1993; Thomas, Collinson and Jones sand had been expelled. In order for a sand
1992; Figure 4.21). volcano to form, a source bed must have be-
Sand volcanoes and extruded sand sheets are come sufficiently overpressured for liquefied
relatively rare features in the geological record sediment to break out upwards to the deposi-
but are sometimes found on bedding surfaces. tional surface. Some type of shock may have
Sand volcanoes are preserved mostly in relative- been involved in the liquefaction process. As the
ly deep-water sediments involving turbidites and sediment-water mixture was extruded at the
slump deposits, whereas sheets are more wide- surface, the excess water would rapidly escape
spread, occurring even on subaerial surfaces. and the flow lobes would become frozen in a
deformation structures '1

depends on observation of a transition from

an intrusive dyke to a concordant bed (Figure
4.23). In some cases the distinction may not be
totally dear, as some extruded sheets may be
confused with sills. In plan view, extruded sheets
commonly show signs of internal shearing in
the form of weakly developed and folded foli-
ation (Figure 4.24) and their upper surfaces
may be modified by regular hydrodynamic
bedforms, such as current and wave ripples.
Extruded sand, in the form of a mobile slurry
probably lost mobility very rapidly as the excess
pore fluid drained to the surface, causing
the extruded flow to 'freeze' (Hesse and Read-
Figure 4.22 Sand volcanoes on the upper bedding ing 1978; Collinson, Bevins and Clemmensen
surface of a sandstone bed. The sand of the volcanoes 1989). Modern examples are documented from
may have been sourced from the bed itself or from areas subjected to intense seismic shocking
underlying layers. Note the axial vents and the
preservation of del icate flow lobes on the flanks of the
(e.g. Swanson 1964; Reimnitz and Marshall
volcanoes, attesting to the quiet nature of the water at 1965).
the site of the extrusion. Ross Formation, Upper
Carboniferous, County Clare, western Ireland.
43.5 Synsedimentary faults
Faulting of very early post-depositional origin is
quite common, occurring both in otherwise un-
deformed sequenes as wen as in association with
other deformation features. The results are not
always readily distinguishable from later faults,
although a useful rule of thumb is that early
faults are seldom mineralized and commonly
show smearing of sediment along the fault sur-
face. This matter is discussed further in section
Both extensional and contractional faults oc-
cur. Where the features occur in otherwise unde-
formed strata, they are most likely to be of
Figure 4.23 Folding in the transformation of a sandstone
extensional origin, that is, normal faults. Throws
dyke (bottom right) into an extruded sandstone sheet range up to a few metres, although more com-
(top leit). The sand is intruded into poorly sorted monly they are at the scale of miHimetres or
conglomerate of mass-flow origin which, itself, must centimetres. The extensional stress field may re-
have been in a very mobile state at the time of
flect contemporaneous tectonic activity or it
intrusion-extrusion. Moraenesi Formation, Precambrian,
north Greenland. may result from mass-movement of sediment on
a slope. The head zones of slump and slide
way somewhat analogous to the sieve lobes that sheets are commonly characterized by such ex-
are so commonly found in the stockpiles of tensional movements.
sand-washing plants in quarries (e.g. Carter Within larger, more complex deformation,
1975). Very quiet conditions must have ensued such as convoluted layers or mud diapirs, both
in the basin for such delicate features to have extensional and contractional faults may occur,
survived until they were blanketed by fine sedi- reflecting the local stress field within the larger
ment from suspension. feature, probably at a late stage in its develop-
Extruded sand sheets are sometimes indistin- ment when shear strength was being re-estab-
guishable from beds deposited as part of the lished.
normal sequence. Their true identity commonly At the larger scale, syndepositional growth

Figure 4.24 Irregularly folded foliation in an extruded sandstone sheet, suggesting internal flowage within the
sheet prior to its consolidation. Moraeneslt Formation, Precambrian, north Greenland.

faulting occurs, driven by gravity acting over a pressure. Volume loss of sediment may also lead
prograding delta slope and possibly driven by to deformation through the development of ex-
mass flowage at depth. Such faults, at the smaller tensional stresses. Layers of muddy sediment at
end of their size range, may be confined to single the depositional surface may sometimes suffer a
progradational cycles, as in the Namurian of volume loss that cannot be funy accommodated
County Clare, western Ireland (Rider 1978; Pul- by a loss of thickness. The result is that a ten-
ham 1989), or to a small number of cycles, as in sional stress field is established within the
the Triassic of Svalbard (Edwards 1976; shrinking layer. Once the tensile strength of the
Figure 4.25). The largest occur at the scale of the layer is exceeded it will begin to crack. For a
continental slope and involve packages of pro- horizontal layer, the tensional stress is likely to
gradational cycles, as in the Niger Delta (Figure be homogeneous and the resultant crack pattern
4.26) and the Gulf Coast of the southern USA will be a system of broadly similar polygons,
(e.g. Weber and Daukoru 1975; Winker and commonly hexagons. The diameters of the poly-
Edwards 1983). AU such faults have listric geo- gons seem to scale roughly with the thickness of
metries and complex associations of synthetic the cracked layer and some surfaces may show
and antithetic faults. Sequences thicken markedly coexisting large and smaB crack patterns. When
into the hangingwaH areas and there is common- the surface is sloping or where there are hetero-
ly significant ron-over into the fault. geneities within the layer, the crack pattern may
be influenced by these, with rectangular and
elongated crack patterns developing (Donovan
4.3.6 Structures due to sediment shrinkage and Archer 1975).
Structures dealt with above are an the product of Subsequent deposition of contrasting, com-
loss of shear strength as a result of increased pore monly coarser grained, sediment above the
Limit of exposure

Scale (approx.) (no v.e.) FACIES D 5th sequence
Faults Depositional contacts Stratification 150
Massive sandstone. EB'3 4th sequence
Growth Post- Sharp Grad- Erosional Horizontal Cross 100
depositional ~ational _ _ ._ Stratified sandstone ~ o 3rc1 sequence
'\ ,'\. -- ¢:'
Ob~ed _
~ --:'l.,--- ._~._j._""::--:':-- --- 50 Shale [3J lI[] 2nd sequence
Inferred \ \ \ \
\ ' \' ~ 1st sequence
00 100 200 300 (m)

Figure 4.25 Succession of small-scale growth faults developed in deltaic sediments. There is slight overlap between the profiles from top to bottom. The
faults stayed active during several phases of progradation, which was from left to right. Note the listric geometries of the faults and the thickening and roll-over
in the hangingwall areas. The wedges of massive sand are infills of depressions that developed owing to continued movement on the faults after the last
progradation. The apparent antithetic fault (10) may reflect a cuspate fault trace in plan so that faults 9 and 10 are the same, with a component of displacement
out of the face. Triassic, Edgeoya, Svalbard. (After Edwards 1976.)
deformational structures

Closely spaced Down-te-basin faults, Shale diapirs. antithetic and Shale diapirs,
NIGER DELTA down-to-basin growth faults, more rollover erestal faults, back-to-back faults major cQunterregiona! faults
minor roilover (Oligocene trend) (Miocene trend) (Mio-Plio-Plelstocene trend)

J 25 000 it

figure 4.26 Large-scale cross-section through the Niger Delta complex showing extensive and deep-rooted growth
faulting driven by basinward flowage of deeply buried, overpressured muds. Individual faults were active over long periods,
involving many progradational cycles of the delta. The faults displace progressively younger sediments in a basinward
direction. Near the free surface of the continental slope, major shale diapirs reflect the basinward flowage of the
buried muds. (Aller Winker and Edwards 1983.)

Figure 4.27 Desiccation cracks in present-day muds. Figure 4.28 Casts of lenticular cracks on the base of a
Note the relationship between the diameter of the crack sandstone bed. Such cracks have been interpreted as
polygons and the depth of the cracking. The thin mud being due to synaeresis, although some similar forms
layer associated the smaller scale cracks is rolling up into may result from evaporite pseudomorphs. Independence
potential mud clasts susceptible to future erosion. Fjord Group, Proterozoic, northeastern Greenland.
Sunglasses, bottom right for scale.
mud-cracked layers commonly ron up into con-
cracked surface will give sandy passive infins vex upwards biscuits, which may then be subjec-
to the cracks. These commonly taper down- ted to erosion as mud clasts by the wind or by
wards but may be folded as a result of dif- subsequent aqueous currents. These may become
ferential compaction of the host and crack-fiU concentrated as mud-flake conglomerate layers
lithologies. above the cracked surface.
Shrinkage cracks occur in two main forms. So-called synaeresis cracks are thought to de-
Desiccatioll cracks result from sustained drying velop in subaqueous settings where the contrac-
of a surface mud layer on an exposed river bed, tion of the surface layer appears related to
lake floor or tidal flat (Figure 4.27). The cracks changes in the volume of days, possibly as a
tend to be paraHel sided and to have well-devel- result of changing water chemistry. Synaeresis
oped polygonal patterns. It is not uncommon for cracks tend to be lenticular in plan and less
two or more scales of crack to coexist, smaller inclined to develop into dear polygonal patterns
cracks reflecting the drying of a thin surface layer, (Donovan and Foster 1972; Figure 4.28). The
and larger systems reflecting desiccation pen- details of the processes are not well known and
etrating to a greater depth (Figure 4.27). Thin the criteria for distinguishing cracks of this origin
Physical deformation structures 117

from those due to desiccation are somewhat

ambiguous. Recently, Astin and Rogers (1991,
1993) have challenged the subaqueous origin for
lenticular cracks of this type. They suggest that
many, if not most, examples are the result of
subaerial desiccation, with the lenticular forms .. _-_ ............ .
being pseudomorphs of evaporitic gypsum crys-
tals. Trewin (1992) and Barclay, Glover and
Mendum (1993) have questioned this reinter-
pretation and suggested that cracks in organic- .......
rich lacustrine mudstones are hardly likely to be ..............•..................•.

other than subaqueous.

A more localized type of crack pattern, poss-
ibly of similar origin to synaeresis cracks, is
that associated with certain carbonate con-
cretions developed in mudstones. These con- (c) (d)
cretions, called septarian nodules, are character-
ized by a pattern of irregular but somewhat
radial cracks in their core. They are further
described in section 4.3.9, together with explana-
tions of their origin which do not involve sedi-
ment shrinkage.

4.3.7 Structures due to sediment wetting

These structures are features of some aeolian Figure 4.29 Styles of deformation in the foresets of
dune sands, especially those in coastal dune aeolian dunes, including rotation, folding and brecciation.
fields. When winds are actively supplying sand to The establishment of tensile and cohesive strength in
the lee side of a dune, the grains are non-cohesive the sands, probably as a result of wetting, allowed the
sands to behave in this way rather than as the more usual
and accumulate by grain fall, grain flow or as the non-cohesive flow. (After McKee 1979.)
result of ripple migration. During inactive per-
iods, the sediments of the lee slope may become In finer grained and organic-rich sediments,
moistened by rainfall, dew or salt spray, and, as a compaction is much more significant and is com-
result, the sand of the surface layer may develop monly characterized by a general flattening of
some cohesion and tensile strength. Resting on a depositional features normal to bedding or
steep slope, this material may begin to move lamination. There may be significant thickness
down slope and, in the process, buckle and/or change, up to an order of magnitude in the case
brecciate (McKee 1979; McKee and Bigarella of very organic materials such as peat. Not only
1972). The result is the development of lenses of is the thickness of bedding and lamination reduc-
deformed sand within the steeply dipping foresets ed but also fissility may be induced by the re-
that normally typify the lee side of a dune (Figure orientation of platy particles. Plant fragments
4.29). and thin-shelled animal fossils may be reduced to
bedding surface impressions. Deformation is per-
haps more apparent in vertical section when
4.3.8 Deformation related to compaction there are inhomogeneities in the sediment, be-
In terms of structures produced, compaction tween which differential compaction has occur-
plays a variable role during burial, depending red. Such inhomogeneities, which may occur
upon lithology. In coarser sediments, compaction across a range of scales, may be depositional in
is of relatively minor importance and might be origin, as where sandy ripple lenses float in a
detected only through slight reductions in the matrix of silt or clay, or they may be the result of
inclination of dipping depositional surfaces such early organic disturbance, as when sand-filled
as cross-bed foresets. burrows occur in a finer matrix. In such cases, the
118 Sedimentary deformational structures

ssw NNE



Figure 4.30 Seismic section across a major deep-water channel-fill sandbody. The top of the channel bodies are more
conspicuous because later compactional deformation has caused convex upwards folding of both the top surface of the
sand and the bedding in the overlying mudstones. Alba Formation, Alba Field, North Sea.

coarser sediment compacts less than the matrix, 4.3.9 Deformation related to early chemical
with the result that laminae in the muds are precipitation
deformed around the coarser body. The burrow
or ripple lens itself may also be subjected to One of the most obvious ways in which sedi-
some change of shape as a result of the compac- ments are modified after deposition is by the
tion. Originally circular burrows may be flat- suite of chemical processes known as diagenesis
tened to elliptical cross-sections, and ripple (section 1.1.1). These processes often proceed
lenses that were depositionally plano-convex without leading to any very obvious deforma-
may become convex both upwards and down- tion, although, as has been made clear from the
wards. Section 4.3.9 outlines how deformation outset of this book, there is probably a close
arises from differential compaction around dia- interlinking with mechanical processes, a topic
genetic nodules. which is complex and poorly understood.
At a larger scale, sand-filled channels, which Some mineral precipitation may be associated
originally had flat tops and convex-downwards with overpressuring, such as that inferred for the
bases, may develop distinctly convex-upwards curious bedding-parallel calcite veins spectacu-
tops as a result of differential compaction. Such larly developed in Jurassic rocks of southern
changes of shape are particularly apparent in England, and known as 'beef' structure (Stoneley
seismic sections (e.g. Newton and Flanagan 1983).
1993; Thomas, Collinson and Jones 1992; Figure Where fine-grained sediments have become
4.30). Inhomogeneities related to chemical pre- locally cemented as concretions or nodules early
cipitation and the differential compaction asso- in their burial history (e.g. Mozley and Bums
ciated with them are dealt with in the following 1993), these hardened patches may resist compac-
section. tion and preserve bedding or lamination close to
Physical deformation structures 119

its depositional state. Compaction of surround- an indication of the relative burial depths at
ing sediments commonly leads to deformation of which the cementation processes began. For
lamination around the concretion (Figure 4.31). example, Craig (1985) compared the separation
Some measure of compaction may be deduced of laminations as they bow around septarian
from comparison of lamination thickness within nodules in Lower Palaeozoic rocks south of
and outside the concretions. In turn this may give Aberystwyth, Wales. Typically, the greatest dif-
ference in separation was in the range 52-58%.
Assuming that this represents porosity loss due
to compaction, the implication of such large
values is that the concretionary growth must
have begun within no more than a few metres of
the sediment surface. Conversely, those nodules
that show little deflection of the laminations are
inferred to have initiated after greater burial and
volume loss. The idea is summarized in Figure
As well as leading to physical deformation
through controlling differential compaction, con-
cretions may be regarded as a type of sediment
deformation in their own right. The mineralogy
Figure 4.31 Chert concretions with some compactional
of the concretion-forming cement may vary
draping of bedding in the surrounding limestone. Portfjeld considerably. Carbonate minerals (e.g. calcite,
Formation, Lower Cambrian, north Greenland. siderite), sulphides (pyrite), sulphates (e.g. barite,

No Continuous Continuous No Continuous

deposition deposition deposition deposition deposition
_____ ~bs~L _ _ _ _ _ ~U ________ .J.:U ___ ~bs~!L __ ~U ____ Water.

VI Concretion body
I:::=:: I Laminae
constantthickness· ...•.... Deformed and
t Timespan of

of laminations but thinned laminations t concretionary

la~, "pa:~::o:~::::_ m" of "dlmoo'


compaction and depth and between Differences most marked in

sediment porosity and depth, assuming core of the concretion
continuous deposition at the sediment
water interface
Compaction at these depths too
little and too slow to allow
much deflection or thickness change

Decreasing sediment porosity/sediment compaction rate

Figure 4.32 Schematic depiction of the changes in lamination thickness and separation adjacent to and within concretions
formed at different burial depths. (From Craig (1985). Used with permission of Jon Craig.)
120 Sedimentary deformational structures

gypsum and anhydrite) and silica (flints and the nodules. The extent to which developing
chert) are all common concretion-forming min- concretions can cause deformation in the host
erals. They may be precipitated within the host sediment by exerting their own growth force
sediment simply to locally fill the original pore has long been debated (Maliva and Siever
space, or the concretion may grow either re- 1988). Deformation within concretions is ex-
placively, preserving many of the host's original emplified by the features known as septarian
features in a changed mineralogy, or displacively, nodules. These are characterized by roughly
pushing the host sediment aside and thereby radial or polygonal cracks within the concre-
deforming it. tion (Figure 4.34). The cracks are typically
Some concretions follow and to some degree filled with calcite, although dolomite, ankerite,
pseudomorph earlier features of the host sedi- quartz, barite, celestite and pyrite are found,
ment. They may follow burrows or root traces in various degrees of mutual replacement. The
or may nucleate around a fossil fragment. In isotopic composition of the minerals may
some cases the concretions may infill a pre- provide a record of the palaeohydrology
existing cavity, caused by dissolution of some (Desrochers and AI-Aasm 1993; Wilkinson
original soluble material. In the case of con- 1993).
cretions formed as a result of pedogenic pro- Astin and Scotchman (1988) invoked over-
cesses, the concretions may be distributed to pressuring during shallow burial to explain the
define a distinct profile, reflecting a period of septarian fractures and the textures of their min-
sediment starvation during which the near-sur- eral infiH, but other accounts have emphasized
face sediments were subjected to the ground- the role of nucleii of decaying organic matter
water regime of the soil. In semi-arid settings, (Allison 1988). The alkaline, reducing microen-
the concretions are most likely to be of calcite vironments that are created around the nucleii
(caliche or calcrete soils) (e.g. Allen 1986b; Fig- favour the precipitation of calcium from the sea
ure 4.33). In humid, water-logged soils, siderite water. Recent explanations of the process have
is the more likely concretionary mineral, as in involved intermediate soapy materials (Duck
coal measure seat-earths (e.g. Besly and Fielding 1990), the calcium stearates being derived from
1989). the saponification of fatty acids (Berner 1968;
The growth of certain concretions, particular- Hesselbo and Palmer 1992). This not only ex-
ly those of calcite within mudstones, may estab- plains the localization of the concretions but
lish secondary stress fields, which, in turn, lead also the origin of the septarian structure, as
to further local deformation within and around the subsequent conversion of the calcium soap
to calcite involves a volume reduction and
leads to shrinkage and cracking of the cemented
Another striking structure within some con-
cretions is known as cone-in-cone. The feature
typically consists of opposing hemispheres of
nested cones, with their apices directed inwards
(Figure 4.35). The origin has long been debated,
with either a gravitational consolidation stress
or an outward force of concretionary growth
commonly being invoked to explain the arrange-
ment of the cones. More recent views have em-
phasized the mutual interference of bundles of
Figure 4.33 Calcite concretions in siltstone in fluvial growing calcite fibres, growing with a preferred
overbank setting characteristic of a caliche palaeosol. crystallographic orientation normal to the sedi-
The progressive downwards reduction in the intensity and mentary laminations (Marshall 1982). Many
interconnection of the nodules from a discrete surface is
typical of a mature profile resulting from an extended
workers have surmised that the apical angle of
period of sediment starvation. lower Old Red Sandstone, the cones reflects the plasticity of the host sedi-
lIanstephan, Dyfed, Wales. ment at the time of growth. The relatively minor
Physical nPHIl"l'TIn. structures 1

Figllre 4.34 Septarian nodules (concretions). Bedding-normal appearance of septarian vein systems in eroded con-
cretions. Silurian, Traeth-yr-ynys lochtyn, west Wales. (Photographs provided by Jon Craig.)

effect of cone-in-cone nodules in west Wales on pattern, perhaps originating from precursor des-
the laminations in the adjacent rocks, together iccation cracks (Eugster and Hardie 1978; Figure
with their incorporation of diagenetic chlorite, 4.37). Displacive growth of gypsum and anhyd-
was interpreted by Craig (1985) to indicate initi- rite nodules within earlier sediment, as in pres-
ation of the structure at greater burial depths ent-day sabkhas, pushes aside the host sediment
than the other concretions in the district (Figure and in extreme cases this process may lead to
4.36). the development of crocken-wire texture (Figure
Significant stresses are also set up within 4.38). Layered anhydrite may develop quite in-
near-surface sediments as a result of the growth tense folding as a result of the horizontal com-
of displacive evaporite crystals, concretions and pressive stresses associated with crystal growth.
layers. As precipitation proceeds, a surface layer This so-caned e[lterolitilic structure may cause
of evaporite mineral may be pushed up into extensive small-scale folding confined within a
pressure ridges with a broadly polygonal plan single layer (Figure 4.39). Such deformation may
_... -
.. =- ~~\rti.-{'i ,; -- - - - -


.'-. -~/' -',

.-" f\', 1",
." f"\
''\f' ,. .,. "
.-- LX -
-;-Y7IA~·""" If~~'f~ ~- _., -..-
IT.... '---L..J\


Figure 4,35 Cone-in-cone nodules (concretions). as portrayed by Gresley (1894). Drawings 3, 9 and 17 are perspective views showing the nested arrangement of the cones;
the remainder are either plan or section views.
cm '80% F
Few ti:
m z
10 m

I I' ynte CO~E -IN -c~~t_in_cone -r -STRUCTJRE' 'I ~~Y)0 p
100 II development in I I I
m II compacted rim I I 1

35% ~II I
fu I I
(/) I I
1000 125 %
m ~'0,."'-."'-.'0,."'-.'0,."'-."'-."'-."'-.~"'1~ "'~~
- Coarse
"'-".'-".'-".'-".'-".'-".'-". "'-. . "'-."'-."'-.-".'-".'-."
c R f 1I quartz R I R ~
:.;z R I o
() Vein quartz .
al grows into Matnx replaced by LL
o o
z- R dolomite fill Silica ~ ::J
oal (j)
-E I from main body 1
3000 110 %
m oQJ

~""""""1''''''''''''''''''''r~~ ~""""""""""'r""""""""~ ~""""""""""~""""""""~""""""""""""'l0.."""""""~,,,,'-':j
UU"(J z
-c ~
f-o ? N
0:;:::; R i ( Some matrix silica replaced by idiomorphic
-al F
en- '. R i dolomite rhombs
<5<- ~
o 1 1 I (contain undigested quartz inclusions) o
.0 ..J
ro I I I
QJ + + +
~'Lozenge-shaped dolomites TECTONIC DEFORMATION
deposited along vein
quartz boundaries

Figure 4.36 Chart summarizing the diagenetic evolution of concretions in southwest Ceredigion, Wales, (From Craig 1985.)
Depth of burial axis is not to scale and all numbers are approximate, (Used with permission from Jon Craig.)
124 structures

Figure 4.37 Polygonal pressure ridges deveioped on

the surface of a playa lake as the result of the
precipitation of evaporitic halite. Death Valley, California.
(After Hunt st al. 1966.)

Figure 4.39 Enterolithic folds in a gypsum layer,

probably a pseudomorph after evaporitic anhydrite.
Lower Purbeck (Jurassic), Worbarrow Trout, Dorset.
(Reproduced from Shearman (1980). with permission
from Editions Technip.)

also be detected in pseudomorphs after precursor

evaporites, the transformation occurring at quite
an early stage. Chert and quartz pseudomorphs
after anhydrite nodules (e.g. Chowns and Elkins
1974) and after complex sodium silicate evap- Figure 4.38 Chicken-wire texture developed by the
growth of displacive anhydrite nodules, which have
orites in the soda lakes in East Africa Eu-
pushed aside the host sediment into thin sheets
gster 1969) are examples of this. between nodules. Lower lias (Jurassic), Aquitaine
Basin, southwest France. (Reproduced from Bouroullec
(1980), with permisSion from Editions Technip.)
throughout the stratigraphical record. Some
The structures that are developed or modified types of particularly those which in-
early in the post-depositional history of sedi- vOlved depositional rates or which are
ments are widely but irregularly developed located in areas of seismic activity, are
Conclusion 125

perhaps more susceptible to deformation but indicators. Some potential strain indicators, such
those relationships are far from clear-cut. That as burrows or bedforms, may have undergone
being the case, it is essential for the scope, significant modification through early processes
scale and style of deformation structures to be of alteration prior to becoming involved in any
constantly borne in mind when trying to un- later disturbance. Equally, some deformation
ravel the structural history of an area or of a structures, particularly more regular and simple
particular succession. This is particularly import- ones, such as desiccation cracks, could be used
ant in trying to identify and work with strain with care as strain indicators in their own right.

Mass movements

5.1 INTRODUCTION overview of the spectrum of mass movement

processes and products, but emphasizes slumps
5.1.1 General and slides. Both modern and ancient examples
are used.
Mass movements are significant geological pro- The theoretical background to how such weak
cesses which have an important impact on hu- sediments deform has been outlined in Chapter 2.
man life. They are environmental hazards, both The present review concentrates on the qualitative
on land and in the sea, and justify intensive macroscopic and morphological aspects of
attention. Much research has been related to mass movements, as seen in nature. For the
human or industrial loss and welfare, and the reasons mentioned above, those movements that
importance of further efforts cannot be under- are mainly depositional in nature, such as falls and
estimated. The effects on modern communities of flows, are treated only briefly, and more attention
avalanches, major rock falls and mudslides are is given to movements that modify sediments
well known from recent catastrophies (Bolt et al. already deposited, as these are unarguably defor-
1975; Press and Siever 1978; Voight 1978a; Brun- mation processes. Even these cannot be treated
sden and Prior 1984), and mass movements can here in detail. Rather, this general summary of the
seriously affect the stability of offshore installa- important aspects of each process or group of
tions (Prior & Coleman 1982). The importance in processes is designed to provide the reader with a
the context of this book is that mass movements sense of the wide variability that exists.
present a very widespread situation in which
sediments are subject to deformation.
For the geologist, gravitational mass move- 5.1.2 Classification schemes
ments span a wide range of processes, both Mass movements have been classified on the
subaerially and subaqueously (Brunsden 1971; basis of a wide variety of factors, including pro-
Schumm and Mosley 1973; Voight 1978a,b; cess and rheology, product, climate, type of ma-
Saxov and Nieuwenhuis 1982; Brunsden and terial moved, local geology and triggering mech-
Prior 1984; Allen 1985; Morton 1993). At what anisms (Ladd 1935; Ward 1945; Dott 1963;
point in this spectrum the processes are regarded Crozier 1973; Middleton and Hampton 1976;
as deformational is a very subjective judgement. Nardin et al. 1979; Hansen 1984; Pierson and
All the products provide important information Costa 1987). Many of these schemes are complex
on the depositional setting, and in the ancient (Varnes 1978), and frequently difficult to use,
record are powerful tools, both for sedimen- particularly in the field. Some schemes are con-
tological and stratigraphical analysis, yet all cerned only with subaqueous gravity flows (Lowe
could be argued to involve deformation of one 1979), and do not include slope failures such as
kind or another. The processes involved in slides and slumps. Others aim to classify all
slumps and slides are deformational in any view, subaqueous processes, whether gravity-driven or
and the resulting products of relevance to struc- not (Pickering et al. 1986). Geologists are often
tural geology. This chapter therefore presents an concerned that classification schemes should be

The Geological Deformation of Sediments Edited by Alex Maltman Published in 1994 by Chapman & Hall ISBN 0 412 40590 3
128 Mass movements

simple, that they concentrate on descriptive and

morphological factors, and that they be easily CREEP
applicable. A useful scheme for mass movements Slow intergranular
frictional sliding
should also point the user towards the genesis of with quasi-static
grain contacts
the particular unit observed.
The scheme suggested by Kruit et al. (1975), SLIDE
Translational (as shown)
and further developed by Rupke (1978) and Stow or rotational;
coherent mass
(1986), has probably come the furthest in deve- with minor
loping a classification scheme within which pro- internal deformation

cess and product are considered in a simple and SLUMP

Coherent mass
easily applicable way. The scheme, based on mass with considerable
internal deformation
movement rheology, was simplified by Nemec (discrete, non-pervasive
(1990), who divided the movements into six ca-
tegories (Figure 5.1a). This classification is simple FLOW
(with plastic
and logical, both for subaqueous and subaerial behaviour)
RemouJded mass.
processes, because it shows a range from slow Non-turbulent but possibly
with transient, large-scale
movement of coherent masses (creep), with little turbulent churning

or no relative movement of individual grains

('quasi-static' grain contacts), through increasing- FLOW
ly turbulent movements, to rapid mass movement (with fluidal
of grains which move almost to entirely indepen- Fully turbulent

dently of other grains (e.g. falls of debris).

Although the various mass movement pro- FALL
Solitary grains or loose
cesses are presented under individual headings, it grain assemblages

is important to note that they represent parts of a (a)

process continuum (Figure 5.1 b), and in many

instances one may evolve into another with time, CREEP
Debris flows
or the depositional effects of one type may trigger Graln flows
other processes. Thus, no one mass movement :::J

process should be considered entirely indepen-

.~ I / /' Liquefied flows
o() SLUMPS (/ _
dent of the others. rn FLOWS (with plastiC /snow-sedlment flows
Another useful and more detailed way of clas- rn

sifying mass flows was suggested by Pierson and


behaviour) L----
FlUidized flows
D- FLOWS (with flUidal c---
Costa (1987) based on the mean velocity of the behaviour) ~'Volcanic' flows
flow and sediment concentration in per cent FALLS -Debris falls ~ ..
"'-ROCk falls Turbidity currents
(Figure 5.2). This scheme also emphasizes flow or
movement rheology, and from the products im- Figure 5.1 Classification scheme of mass movements:
(a) details six categories, based on movement rheologies;
portant quantitative estimates of flow velocity (b) emphasizes that the categories comprise a process
and sediment concentration can be made. continuum. (From Nemec 1990.)

here as fluidal flows (section 5.3) are defined as

5.1. 3 Basic theory showing approximately Newtonian behaviour.
To develop an understanding of mass movements Distinguished from fluidal flows are those mass
it is necessary to recall the rheological principles movements that require the applied shear stress
mentioned in section 1.2.4. Newtonian fluids, like to have exceeded a yield point before initiating
water, have no shear strength and deform im- the irreversible deformation. At their simplest,
mediately once a shear stress is applied. The where the shear rate increase with increasing
amount of flow depends on the viscosity of the shear stress is linear, these materials are showing
material, which may decrease (shear thinning) or Bingham behaviour (Figure 3.4), but they are
increase (shear thickening) with increasing rate treated here as being generally plastic materials
(section 3.6.2). The mass movements described and termed plastic flows (section 5.4).
Introduction 129

Sediment concentration (vol. %)

Sediment concentration (vol. %) a A B C 100
Fast - inertial r----,--V'"e:;::lo:::cci:;::tie::::s-::n::::ev=er:--r------.-----,
ABC 100 forces 10 2 measured or Sturzstrom
Fast - Inertial ;----T--,-V"'el",occ::iti"'es:-:n:::ec::ve"'r....:;:-------r---;Fc:-IU-c:id;::ize-::d:---l dominant estimated
forces 10 2 measured or granular flow
dominant;:im::::a::::te~d_+_ _ _ 1" Hyperconcentrated
10 ' avalanche
inertial streamflow
10 ' Hyperconcentrated slurry flow
streamflow Inertial
- - - - - granular flow 10 0
toO Normal Grain flow
streamflow 10- 1

.s Debris flow
i':' 10-3
u 10-3
c Earth flow
No mechanism
~ 10-4 Viscous 10-4 to suspend sediment
:;; slurry flow
granular flow
10- 5

No mechanism
to suspend Slow-
Slow _ 10- 7 sediment viscous/ 10- 7
viscous frictional
frictional forces
forces 10-8 dominant
dominant 10-8
Solifluction Mass creep
FLUID TYPE Newtonian 1 Non-Newtonian
FLUID TYPE Newtonian I Non-Newtonian
Water j Water + fines i Water + air
+ fines
l Water + fines -' Water + air
+ fines
1 J Granular flow
CATEGORY Streamflow I Slurry flow I Granular flow FLOW
CATEGORY Streamflow Slurry flow
I Plastic
J Plastic

Figure 5.2 Classification scheme based on flow velocity and sediment concentration. (a) Rheological classification of
sediment-water flows. The vertical boundaries A, Band C are rheological thresholds and functions of grain-size
distributions and sediment concentration. Boundary A marks the initiation of importance of yield strength, boundary B
marks the sudden rapid increase in yield strength and onset of liquefaction behaviour. and boundary C marks the end of
liquefaction behaviour. (b) Application of existing flow nomenclature to (a). (From Pierson and Costa 1987.)

Generation of mass movement requires a resistance of a rock or sediment, and vice versa.
slope, which may vary in inclination from less Most importantly, overpressuring (section 1.3.3)
than 0.10, such as on modern delta fronts (Prior is a very common process of reducing shear-
and Coleman 1978b), to vertical and overhang- ing resistance and causing slope failures. Slope
ing surfaces, where rock falls may occur. Mass oversteepening, e.g. by high sedimentation
movements arise where the shear stress acting rates, undercutting or retrogressive sliding, is
on the sediment or rock exceeds the material's another common process of reducing shearing
shear strength or shearing resistance. Following resistance.
the formulation of Terzaghi (1962), the shearing In the following six sections, each major
resistance is a function of the cohesion, stress- group of mass movement is dealt with, in order
dependent internal friction and pore-fluid press- of decreasing amount of solitary grain move-
ure of the material, as given by the Mohr- ment and turbulence. Thus, falls are summarized
Coulomb relation (equation 1.13). This equation first, then gravity flows, starting with those
is fundamental to understanding the initiation of with fluidal behaviour, and moving on to flows
mass movements. with plastic behaviour. As mentioned above,
From the Mohr-Coulomb relation, a decrease slumps, slides and sediment creep are reviewed
in cohesion, friction or normal stress at the in more detail, emphasizing their deformational
potential slip plane will decrease the shearing aspects.
130 Mass movements

5.2 FALLS mate (subaerially); and (iii) bedrock fracturing

and structural deformation (see extensive review
5.2.1 Introduction by Whalley 1984). Carson and Kirby (1972) rec-
ognized several types of rock fall, their forma-
The term falls refers to downslope movement of tion largely depending on the scale of the falling
solitary grains or relatively loose assemblages of fragments (Figure 5.3). Rock falls are probably
grains where each particle moves more or less more common in subaerial settings than in sub-
independently of others. In subaerial settings the aqueous, largely because of the relative lack of
process is constrained to rock falls from rocky significant slopes in the latter. However, in steep
headwalls, a very common process is mountain- fjords (Prior and Bornhold 1990) and on frontal
ous terrain. In subaqueous settings, falls occur margins of carbonate platforms (James and Gin-
in two modes: rock falls essentially similar to sburg 1979; Mullins and Cook 1986), rock fall
those in subaerial settings; and debris falls, may be a significant process.
which are produced by similar processes to rock
falls, but are formed from previously deposited Bedrock type Massive, homogeneous rocks,
sediment. such as some igneous rocks, are probably less
prone to producing rock falls than layered and
heterogeneous rocks, such as metamorphic and
5.2.2 Rock falls most sedimentary rock packages, unless con-
siderably influenced by other factors. Rocks that
Process are layered are more susceptible to breaking
Rock fall occurs when slabs or pieces of bedrock apart, particularly along bedding planes, and
come loose from their original position and fall some lithologies are more easily weathered than
down slopes or mountain sides (Figure 5.3). The others. However, there is no simple correlation
volume of the rock fall may vary from single between rock strength and probability of rock
small pieces to cascades of solid rock, which fall formation (Abele 1972).

may cause extensive damage. In addition to Fine-grained lithologies will tend to produce
steep slopes, three factors are important for the rock falls with smaller fragments than coarser
formation of rock falls: (i) bedrock type; (ii) cli- and more massive lithologies (Whalley 1984).

.:, T

••••• T
Failure surface


(a) (b) (c)

(d) (e) (f)

Figure 5.3 Types of rock fall, dependent on size and number of falling fragments. (Based on Carson and Kirby 1972.)
Type (a) is characterized by a large fragment bounded by a well-defined, vertical fracture. Type (b) is characterized by
numerous smaller fragments, also bounded below by a relatively well-defined surface. Type (c) is characterized by a
sliding rock mass above a well-defined, inclined but non-vertical basal shear surface. It generally compares with (a) and
(b), but contrasts with (a) in the nature of the basal contact and with (b) in the number of fragments involved. The
characteristics of type (d) are freely falling, larger fragments and no extensively developed basal contact. Type (e) is
similar to (d). but only very small fragments are involved. Type (f) is termed 'toppling failure' (de Freitas and Watters
1973). where the fragments are large and only move a short distance. with the maximum movement occurring at the top
of the fragment. It has some resemblance to (a) and (d). (From Whalley 1984.)
Falls 131

Furthermore, the rock falls formed from massive be expected. The debris is generally poorly
rocks tend to have much larger volumes than sorted and has an openwork, clast-supported
rock falls formed from fine-grained rocks, al- texture.
though there is probably no one-to-one relation- A first-look, qualitative assessment of the risk,
ship. This is largely because fine-grained rocks frequency and common quantity of subaerial
generally weather more easily and form rock rock falls in a particular area can often be made
falls continuously, whereas massive rocks tend by viewing the expression of the debris cone. A
to break off in larger volumes and generate vegetated, major tongue of debris at or beyond
more voluminous rock falls. the base-of-slope point suggests that rock falls
are infrequent but of major scale when they
occur. Conversely, a non-vegetated talus built
Climate Climate is a critical factor for rock-fall
up on to the rock wall suggests small but fre-
generation in subaerial settings. Temperate and
quent rock-fall events (Figure 5.5).
humid climates are probably the ideal condi-
Rock-fall products range in scale from cones
tions for rock-fall formation. Rain-water perco-
less than 10 cm wide, which can be observed on
lating continuously in cracks and advancing
any rocky slope, to rock falls of more than
weathering, lubricates potential slip planes in
106 m 3 in volume. There seems to be an inverse
the rocks and significantly diminishes the fric-
relationship between the scale of the rock fall
tional resistance to slippage. Periods of high
and frequency, so that the daily rock falls gen-
rainfall are therefore particularly critical periods
erally are small, whereas infrequent ones are of
for rock-fall formation. larger scale (Rapp 1960; Gardner 1970). There
Temperatures fluctuating around freezing also
are numerous examples of historically recorded
promote formation of rock falls. Repeated
rock falls that have caused extensive damage
freezing and thawing of percolating water causes
(J0rstad 1968), either because of the impact of
rapid volume changes in cracks and joints (Mar-
the rock fall itself, or as a result of associated
tini 1967; S.E. White 1976). During freezing per-
events such as catastrophic waves when the
iods the water expands in volume, causing an
rocks fell into the sea.
expansion of the cracks, whereas the subsequent
Rock falls are also a dominant process along
thaw causes volume decrease. These rapidly al-
carbonate platform margins and constitute an
ternating stress conditions lead to rapid propa-
important component of facies models from
gation of the cracks and frequent slippage and
these areas. Erosional slopes around the
rock-fall formation.
Bahamas and off the Belize coast in the Cari-
bbean are dominated by rock falls (James and
Fracturing and structural deformation Pre- Ginsburg 1979; Schlager and Ginsburg 1981).
existing joints and fractures will enhance rock-
fall generation (e.g. Gene; 1993), particularly if
5.2.3 Debris falls
their strike orientation parallels that of the rock
wall. This is a common feature in granitic ter- Process
ranes (Selby 1977). Furthermore, if the dip of the
Debris falls (a term introduced by Holmes 1965)
rock wall exceeds that of the fractures, particu-
are straightforward downslope movements of
larly large and frequent rock falls may occur,
dispersed debris, with single, freely moving
because the bedrock may be inherently unstable
grains responding to the downslope pull of
and prone to break-off.
gravity (Nemec 1990). The momentum within
these falls is transferred by 'streaming' (in
Products contrast to flowing) (Campbell 1989), which is
Rock falls most commonly produce a debris entirely controlled by the gravity pull on the
cone or talus cone at the foot of the slope particles. Grain collisions are only of sub-
(Figure 5.4). Such cones can have a wide variety ordinate importance. Debris falls are transi-
of shapes, largely depending on the extent of the tional to cohesionless debris flows and may
area from which the rock falls are produced. change into such flows if slowed, for example
Thus, a spectrum from narrow, wedge-shaped on a gentler slope (Nemec, 1990; see also
debris masses to wide, rectangular masses is to below).
132 Mass movements

Figure !ii.4 Rock fall and snow

avalanche cones from near
Moraine Lake, close to Lake
Louise, Alberta, Canada. Several
cones overlap, each originating
from re-entrants in the rocky
headwall. Avalanche tracks,
produced from snow-sediment
avalanches, are developed on
several cones. The mountainside is
around 300 m high. The cone on the
right has more big boulders on i!,
suggesting a greater influence
from rock-fall processes.
Because the movement of the individual particles
{!\, Type 1 is controlled mainly by gravity, the larger par-
\ Frequent small ticles tend to move the farthest, causing a distinc-
\ rock fall events
tive downslope coarsening of debris fall deposits
(Figure 5.6).
Debris falls tend to produce elongate tongues of
(a) coarse sediment which are coarsest at their
downslope end (Figure 5.6). Furthermore, de-
posits of successive debris falls may coarsen up,
Type 2 because the increased bed roughness and fric-
Infrequent large tional slow-down, where former debris falls were
rock fall events deposited, enhance deposition of coarse debris in
Abundant vegetation succeeding debris fans (Nemec 1990). This is
\ provided that no deposition of fines and
consequent smoothing of the depositional surface
Texturally, debris fan deposits tend to be open-
work gravels which may show normal grading
Figure !ii.S End-members of talus cones: (a) shows a
non-vegetated talus cone along the rock wall, reflecting Depositional characteristics of
frequent but relatively small rock-fall events; (b) shows a debris falls (schematic)
vegetated talus cone at the foot of the rock wall slope,
which reflects infrequent but large rock-fall events.

Debris fans may form if slope oversteepening

occurs due to rapid sedimentation, retrogressive
sliding or slumping, or wave undercutting. This Large

can destabilize an unconsolidated, cohesionless

sediment that is either at rest or slowly moving.
Failure will then occur, causing rapid avalanches Figure 5.6 Depositional characteristics of debris falls.
controlled by debris fan processes (Nemec 1990). (From Nemec 1990.)
Fluidal flows 133

(Figure 5.6) due to decreasing flow competence Coarse-grained Medium-grained Fine-grained

with time as the largest clasts are deposited first (grain flow, (classical)
and possibly overridden by the finer tail of the fluidized flow)
event (Nemec 1990). Debris fall clasts may also
occur as isolated clasts or groups of outsized
pebbles, for instance within or between turbi- Ie.. .~11
dites deposited basinwards of the main debris fall E
deposits (Figure 5.6). The clasts may be 'out- D
T7 -

runner' clasts rolling into otherwise sandy de- T6 E2

posits. C T5 -
B T3 E
T2 1
r---<./-,.--- To-
5.3.1 Introduction
Fluidal flows comprise mass-movement processes '. _..lL_ _---L_ _----LL--_-'-_ _-----"_ _-'--_ __
that are fully or dominantly turbulent during Figure 5.7 Models of turbidity current deposits based
flow. They show a Newtonian behaviour. Some on grain size. (Modified from Stow 1986; based on Bouma
flows, such as high-density turbidity currents, (1962), Piper (1978), Stow (1979) and Lowe (1982).)
may attain a non-Newtonian behaviour due to
Stow 1979; Stow and Shanmugam 1980), fluid
very high particle concentrations, and this is
turbulence and associated effects such as bound-
commented on below. Three main categories of
ary layer shearing are the dominant transporting
flow will be described: turbidity currents, some
and grain-supporting mechanisms.
flows related to volcanic eruptions, and snow and
For the coarse-grained model, other processes
ice flows (avalanching and entraining clastic sedi-
are also important. Lowe (1982) divided the grain
sizes transported by turbidity currents into three
categories, of which the second and third are
5.3.2 Turbidity currents important for the coarse-grained model.
Process 1. Clay to medium-grained sand that can largely
be suspended by flow turbulence alone;
Turbidity currents are generated from sediment
2. Coarse sand to small pebble-sized gravel,
suspensions with an excess gravitational poten-
which is suspended by flow turbulence, with
tial to flow down slopes. The main grain support
settling hindered due to high particle concentra-
mechanism, particularly in low-density turbidity
currents, is the fluid turbulence. During steady tions and by buoyancy created by the interstitial
mixture of fine-grained sediment and water;
state of the flow, a state of autosuspension (Bag-
nold 1962; Pantin 1979) may exist. The sus- 3. Pebble- and cobble-sized clasts constituting
more than 15% of the grain mixture will be
pended sediment is put into motion due to the
supported by the combined effects of fluid turbu-
gravitational pull downslope, the motion causes
turbulence because of shearing along the bed, lence, hindered settling, matrix buoyant lift and
and this further causes suspension. In this man- dispersive pressure resulting from grain colli-
ner, the flow of the turbidity current is largely sions.
maintained by its own motion. Generally, particle concentrations around 20-
Fundamental work on turbidity current flow 30% have been considered to divide low-density
mechanisms has suggested that these flows from high-density flows (Bagnold 1954; Middle-
should be treated as several grain-size popula- ton 1967; Wallis 1969). Lowe (1982), however,
tions (Bagnold 1954; Middleton 1967; Walker considered this division to be somewhat arbit-
1977, 1978; Lowe 1982). There are four models rary, because for turbidity currents made up of
for classification of turbidites, depending on the fine-grained sediment (category I) flows appear
grain size involved (Figure 5.7). For the three to be stable over the whole range of possible
finer grained models (Bouma 1962; Piper 1978; grain-size concentrations. Coarse-grained flows
134 Mass movements

with particle concentrations below 20%, how-

ever, may be unstable and collapse rapidly. I . ,I" .
: .. ,1'.... J \ !.. .:, '\
'1 ( '. " ....
:-~ ,- \ I"
" ,. I • I

.-, "
Products "- , . . . . " ' , .. - 40' ••

Turbidity currents mostly produce sharp-based -: . _.... "-

beds, which generally fine upward, and within -./
"- - ,; ..... - ........
~ ,
which there are successions of sedimentary struc-
tures suggesting waning flow conditions with
-- ~ (J)
J. '-:../ --
- I

-----. -'
time. Deviations may occur (see also below). The Suspension ::s
most well-known product of deposition from
turbidity currents is the Bouma sequence. The B
(parallel laminated) and C (cross laminated) divi-
....--" _ .... til
sions of the Bouma sequence (Figure 5.7; Bouma
1962) are formed by traction, which has also been
discussed extensively, for example by Walker . -.-
. -:-:-
(1965), Middleton (1967, 1970), Allen (1970) and
Middleton and Hampton (1973, 1976). The over-
lying D and E divisions of the Bouma sequence
are formed from further slow-down of the flow,
which initiates deposition from suspension of the
finest grains. (Figures 5.7). Some traction may
occur during deposition of the D division, which
produces the fine lamination caused by textural
sorting related to near-bed shear (Walker 1965;
Stow and Shanmugam 1980; Hesse and Chough
In very mud-rich turbidity currents, waning
flow will cause a fining-upward succession essen-
tially similar to the C, D and E divisions of the Figure 5.8 Idealized vertical section through a coarse-
Bouma sequence (Figure 5.7; Piper 1978; Stow grained turbidite, showing the various divisions and their
depositional mechanisms. (From Lowe 1982.) Lowe's
and Shanmugam 1980). However, this succession symbols for the divisions are given to the left of the section.
is dominated by repetitive sharp-based laminae,
the lowest of which may have very low-angle, 1972; Mutti and Ricci Lucchi 1972; Walker 1978).
climbing, fading ripples (Figure 5.7, divisions The dunes will usually not be very organized,
T 0-2)' Upwards, the laminae fade into wispy, owing to the relatively short-lived nature of the
convolute ones, and at the top into graded mud flow and its unsteadiness (Lowe 1982).
(Figure 5.7, divisions T 3-8; Stow and Shan- Lowe (1982) suggested that cross-beds in tur-
mugam 1980). The repetitive laminae are thought bidites were formed from high-density flows. This
to form within the same muddy turbidity current is controversial, because tractional movement of
as a result of multiple phases of boundary layer sediment would most easily occur from low-
shear causing textural sorting (Stow and Bowen density flows (W. Nemec, personal communica-
1980). tion 1992). For example, Hiscott and
Coarse-grained turbidity currents go through Middleton (1979) documented cross-bedding on
three stages of deposition, according to Lowe the top of massive to stratified turbidites in the
(1982): (i) a tractional stage; (ii) a traction-carpet Ordovician Tourelle Formation on Gaspe Penin-
stage; and (iii) a suspension sedimentation stage sula, eastern Canada. These authors suggested
(Figure 5.8). that the cross-bedding was formed from rework-
ing by the low-density tail of an initially high-
Tractional stage A depositing turbidity current density turbidity current.
can form bed forms such as dunes and plane
beds, owing to flow interaction with the surface Traction-carpet stage As the turbidity current
on to which it is depositing (Govier and Aziz continues to deposit sediment, the increasing
Fluidal flows 135

concentration of particularly coarse-grained par- within normally graded beds (Winn and Dott
ticles near the base of the flow leads to an 1978; Hein 1982; Clifton 1984). Postma, Nemec
increasing amount of grain collisions (Bagnold and Kleinspehn (1988) explained this feature,
1956; Shook and Daniel 1965). This causes deve- based on an experimental study, by large clasts
lopment of a layer near the bed maintained by being transported along a rheological boundary
dispersive pressure and fed from the settling sedi- between two distinctly different parts of the flow
ment above. Turbulence will be suppressed in (Figure 5.9). The lower and basal part of the flow
this layer, eventually leading to inverse grading is a pseudo laminar, inertia layer, along which the
(Figure 5.8a and b). In many traction-carpet clasts 'glide' and are partly submerged. The clasts
deposits, several superimposed, inversely graded are driven downslope by faster moving, turbulent
layers are observed, suggesting that as one carpet shear stresses set up by the top layer of the flow,
froze and was deposited, a new one formed which is fully turbulent (Figure 5.9; Postma,
rapidly above (e.g. Hiscott and Middleton 1979). Nemec and Kleinspehn 1988).
Traction carpets probably do not develop in sand
finer than coarse grained, owing to the insignifi-
cant dispersive pressure between grains of these 5.3.3 Flows related to volcanic eruptions
size grades (Lowe 1982).
Suspension sedimentation stage The final stage During volcanic eruptions, subaqueous and, par-
of the deposition is direct sedimentation from ticularly, subaerial catastrophic mass flow pro-
suspension, which occurs when sediment fall-out cesses can occur, with a tremendous impact on
rate is so high that there is no time for formation surrounding areas (Fisher and Schminke 1984).
of traction structures or a traction carpet. The The dramatic effects of the Mount St Helens
suspension deposits correspond to the Bouma A eruption in Washington, USA, on 18 May 1980
division, and can be graded (Figure 5.8a) or are easily recalled. Also, the pre-historic destruc-
massive. The grading is either a distribution- tion of the city of Pompeii due to an eruption of
grading, if late-stage turbulence has slowed down Vesuvius was a dramatic event, caused at least in
the deposition of the fine material, or a coarse-tail part by mass flow processes related to a volcanic
grading, if the sediment settled mainly as a non- eruption. Three processes are summarized here
turbulent cloud (Middleton 1967). Arnott and which appear to be the most common mass flow
Hand (1989) challenged the view that the massive processes related to volcanic eruptions: (i) debris
Bouma A division results from suspension avalanches; (ii) pyroclastic flows (or nuees ar-
sedimentation. Based on experimental work they dentes); and (iii) lahars. A detailed account of
suggested that the massive character resulted volcaniclastic mass movements is given by Fisher
from deposition during upper flow regime trac- and Schminke (1984), and case studies are de-
tion and suppression of parallel lamination due scribed by Ballance and Hayward (1983) and
to a very high rate of grain fall-out from suspen- J.D.L. White and Busby-Spera (1987).
SIOn. Many of the flows probably attain a turbulent
Lowe (1982) termed these three sedimentation to quasi-turbulent state, due to the admixture of
stages Sl-S3, for coarse sand and gravel, the S3 hot air, water, volcanic material and bed rock,
being equivalent to the Bouma A division, and although some, particularly debris avalanches,
R 1 -R 3, for pebbly and cobbly flows. A tractional are prone to be transitional into rock falls, slides
stage is not likely to occur in the coarsest flows, and debris flows. Siliciclastic debris flows are also
because grain collisions and formation of a dis- common in relation to volcanic eruptions (Koro-
persive pressure will dominate (R.Y. Fisher 1971; sec, Rigby and Stoffel 1980), and are considered
Walker 1975). Therefore, cross-bedding will below under 'Debris flows'.
probably not form at the bases of such flows
(Figure 5.8a), and only R2 and R3 divisions are Debris avalanches
usually found. Debris avalanches occur when large portions of a
In several studies it is observed that outsized volcanic caldera are displaced downslope under
clasts up to an order of magnitude larger than the gravitational forces, as a result of caldera
normal grain size can occur in turbidites, also fragmentation (S.G. Evans and Brooks 1991;
136 Mass movements

Figure 5.9 Drawing of a

high-density turbulent flow
showing development of a basal
laminar flow zone able to carry
larger clasts on top and within the
middle of the entire flow. This
model explains the occurrence of
'outsized' clasts in relatively
fine-grained turbidites. (Based on
laboratory experiment by G.
Postma; from Postma, Nemec and
Kleinspehn 1988.)

Smith 1991; Palmer and Neall 1991}. Numerous

such events are recorded historically (Siebert
1984). Although debris avalanches generally may
be large, single-event mass movements, this pro-
cess can also be cyclic, with recurrence intervals
around 150-200 years (Beget and Kienle 1992).
Debris avalanches can exceed 1 km 3 in volume
and leave large, amphitheatre re-entrants in the
volcanic cones (Figures 5.10 and 5.11). There is
often a close correspondence in volume between
the debris avalanche deposits and the missing
parts of the volcanic calderas, suggesting that
gravitational mass movement is the main process
of removing material from the calderas (Siebert
1984). The morphologies of the debris avalanche

~ Bozymalanny
Hope (British Columbia)
Figure 5.11 Dyke orientation in relation to generation
and transport direction of debris avalanches. Most dykes
are oriented parallel to the areal maximum horizontal
compression (MHC) direction, and perpendicular to

~ St Helens
Madison Canyon (Montana)
transport direction. Downslope from the fragmented cone,
the debris avalanche forms numerous small hills (see
text for discussion). (From Siebert 1984.)

craters are asymmetrical, concave-up forms that
resemble slump and slide scars (Figures 5.10 and
~ Frank (Alberta)
5.11), and which differ from the more concentric
depressions of collapse calderas.
Debris avalanches seem to have a systematic

orientation in relation to principal stress direc-
tions in a volcanic terrane. Dykes tend to parallel
the maximum compressional stress orientation,
Galunggung Goldau (Switzerland)
o 1 km 0 0.5 km and the movement of the debris avalanches is
generally at right angles to this (Figure 5.11;
Figure 5.10 Comparisons of profiles of volcanic Moriya 1980; Siebert 1984). Therefore, the debris
mountains affected by debris avalanching (left) and avalanches are probably controlled by structural
non-volcanic mountain sides affected by rock slides (right).
Note the concave-upward lower boundary of the debris
weaknesses along dykes and related fractures,
avalanches and their larger volumes (note difference in and are triggered during particularly violent
scale). (From Siebert 1984.) eruptive events.
flows 1

The avalanches can attain speeds up to and result is a nuee ardente, or pyroclastic flow surge,
pro bably exceeding 100 km hr - 1. They have great which flows down the volcano side with tremen-
mobility, and can be transported large distances. dous speed and can cause incredible damage. The
The mean ratio between vertical drop and travel flows are fully turbulent, keeping the volcanic
distance of numerous examples of debris ava- material funy suspended, and the flows may en-
lanches was calculated by Siebert (1984) to be train material that they flow over on their path.
0.11, illustrating this point. It is likely that their Thus, they may be the dosest subaerial analogue
mode of transport, at least temporarily and in- of subaqueous turbidity currents, although the
itially, can approach that of debris faUs (see turbulent medium is not only water but an ad-
above). During continued transport they may be mixture of water, gas and hot air.
able to achieve a quasi-turbulent state depending The deposits of pyroclastic flows are unsorted
on the content of hot air and water, which can masses of volcanic, angular fragments, which
enhance turbulence, although water is not an have a distinctive appearance of being welded
essential component for explaining the long run- together. The reason for this is that a substantial
out distances of some debris avalanches (McEwen component of the flows is glass, and when cooled
1989). Direct observation of the Mount St Helens it gives the welded appearance. Pyroclastic flow
debris avalanche suggested that this was initiated deposits are usually called weided tuffs or ignim~
as a slide and evolved into a flow (Voight et aT. brites, and can attain thicknesses up to 100 m and
1983). The frequent preservation of undeformed, cover large areas (Press and Siever 1978).
primary and laminated fragments of the volcanic Some parts of the deposits may not be welded,
cone in some debris avalanches has been inter- and are characterized by chaotic masses of vol-
preted to suggest laminar plug flow (Voight et al. canic debris (Figure 5.12).
1981). Thus there is still a great deal of uncertainty
connected to flow processes in debris avalanches.
The deposits of debris avalanches are generally
very hummocky or hilly masses of unsorted and
brecciated volcanic debris. However, it is common
that the size of the material and the size of the
hummocks decrease away from the source of the
avalanche (Siebert 1984). The hummocks often
seem to occur in a predictable fashion within a
debris avalanche deposit The largest hummocks
or hills occur near the axis of the deposit, while
the hills may be slightly elongated, and their long
axes radiate away from the volcanic cone in the
transport direction of the avalanche (e.g. Aramaki
1963; Glicken 1982). The hills may occur in
clusters, separated by flat areas. The inter-hill
areas may be dominated by lahar deposits and
reworked material (Ui 1983). The hills are thought
to form by flow segregation around large frag-
ments in the flow (Crandell 1971).

Pyroclastic flows (nuees ardentes)

Pyroclastic material is volcanic material ejected
into the air during violent eruptions. Water and
gas are natural components of magma, and when
released from the volcano together with signifi-
Figure 5.12 Unwelded par! of pyroclastic flow deposit of
cant amounts of pyroclastic material, clouds of andesilic composition from Milos, Greece, in the Aegean
excess density in relation to the surrounding air Sea. Note the chaotic texture of boulders of volcanic debris
can be set up which will respond to the gravi- set in a finer grained tuff matrix. (Photograph courtesy of
tational force along the sides of the volcano. The Harald Fumes.)
138 Mass movements

Pyroclastic flows can cause tremendous dam- around freezing during the winter months are
age. In 1902, on the island of Martinique, a nuee more likely to produce avalanches than arid and
ardente flowed down from the volcano Mont stable cold regions. Regions with frequent tem-
Pelee into the town of St Pierre at a speed of perature changes are generally also windy and
100 km h -1. The flow was estimated to have had therefore local accumulations of snow may be
an internal temperature of 700-1000°C, and it built up in failure-prone areas near or on slopes.
killed 30000 people (Bolt et al. 1975). Therefore, coastal mountainous regions in tem-
perate to arctic regions may be some of the
Lahars highest risk areas for avalanches.
Lahars form when pyroclastic flows or debris Shifting conditions between wet and dry snow
avalanches flow into a standing body of water decrease the stability of the snow column. The
such as river or lake. The entrainment of fluid boundaries between different snow types are gen-
into the pyroclastic flow or avalanche results in a erally very sharp, and such boundaries may
transformation from a fully turbulent flow to a eventually form slip planes, for example, if wet,
cohesive, laminar flow, which in most ways com- new-fallen snow accumulates above earlier frozen
pares with 'normal' debris flows (see below). snow (Mellor, 1978).
Lahars have great competence, and may carry
large boulders several kilometres.
Steep mountain slopes The inclination of moun-
Lahars may also form when lakes close to
tain slopes is also a critical factor for producing
volcanic craters breach, and intermixed volcanic
avalanches. Generally, a 30° slope is needed to
material and water flow down adjacent slopes. In
produce avalanches (Perla 1978, figure 7), unless
Java in 1919, when the volcano Kelut erupted, a
extreme conditions are present, for example where
lahar formed when the crater lake was breached,
the underlying sediment fails as a result of heavy
flowing down the volcano sides and killing 5500
rainfall. Therefore, relatively steep slopes are
people (Press and Siever 1978). In glacial regions,
needed to trigger avalanches, although they seem
lahars form when pyroclastic flows or lava melt
to be able to flow on much less inclined slopes once
surrounding ice during eruptions.
generated. Slush flows, which are entirely
The deposition and cohesive freezing of lahars
saturated with water (up to 35-40%; Mellor 1978),
is probably a result both of temperature reduc-
may initiate on much less steep slopes than dry
tion and other factors that reduce the applied
snow avalanches. These flows behave as mud
shear stress as slope angle reduces. Thus, al-
flows, and slopes as low as 10-12° are sufficient to
though lahars possess many of the same charac-
cause flowage (Rapp 1960; Mellor 1978). Slush
teristics as 'normal' debris flows, their flow poten-
flows also may develop from more turbulent drier
tial is to a large degree controlled by the ability of
snow flows if they flow into stream channels and
the flow to retain its temperature.
become saturated with water.
Two mechanisms primarily control the incor-
5.3.4 Snow- and ice-generated (avalanch- poration of clastic material into the avalanches:
ing) flows avalanche turbulence and basal shear. Ava-
lanches of relatively dry snow are probably more
Description and formation turbulent but have less basal shear than slush
Mass movements of snow and ice are important flows and avalanches of wet snow and ice. There-
not only because they move snow and ice, but fore, the latter may be more likely to entrain
also because these movements induce mass larger pieces of rock and vegetation during move-
movements of clastic sediments, and it is this ment. Mixed snow-sediment avalanches may be
respect that will be concentrated on here. (For a fully turbulent because of lubrication by snow
thorough review of mass movements of pure between entrained sediment.
snow and ice see Perla (1978) and Mellor (1978).) Mixed mass movements of snow and sediment
Two factors are essential for the initiation of are mostly relatively unsorted and correspond in
avalanches: (i) climate; and (ii) steep slopes. several ways to subaerial debris flows (see below).
They can behave as surging flows, with one event
Climate Regions with high amounts of precipi- being dominated by several surges (Blikra and
tation and frequent temperature fluctuations Nesje 1991).
Fluidal flows 139

At their upslope end, snow-sediment avalanches
tend to leave a concave-downslope depression
very similar to slide and slump scars (Perla 1978;
Blikra and Nesje 1991). This depression tends to
narrow downslope into a rather narrow ava-
lanche track, which is mainly a bypass zone of
sediment, snow and ice (Figure 5.13a). At the
downslope end, the avalanche track widens on
to a debris cone or scree cone (Figure 5.13a and
b), or an alluvial fan if significant alluvial pro-
cesses also take place. In the case that the scree
cone occurs at the land-water interface, a scree-
apron delta may develop, with a pronounced fan
shape if developed from a point. If there are
several closely spaced avalanche tracks, a 'ramp'
of debris may occur of some lateral extent.
The deposits of individual avalanche events are
mostly unsorted masses of debris, with large tree
trunks and boulders to fine sand and mud being
entrained into the avalanche (Rapp 1960; Luck-
man 1971). The deposits of mixed flows after
snow melt may resemble those of subaerial debris
flows, but the presence of large tree trunks and a
largely openwork, clast-supported framework Interpretation

(Figure 5.13b) of a very heterogeneous clast-type Snow-

mixture may suggest deposition from a mixed 5
snow-sediment flow. In talus screes, deposits of
snow avalanches can be rather easily differenti- 4

ated from those of rock falls. Whereas rock-fall Debris flow

deposits are rather well-sorted and characteristi-
cally coarsen downslope, the avalanche deposits 3
Debris flow

are unsorted masses of debris, often with a signifi-

cant portion of finer material at the downslope
end (see Figure 5. 13b). Debris tails may also 2 Snow-
occur (Rapp 1960), which extend un slope and
downslope from large boulders embedded in the
scree (Luckman 1971). Debris flow
The preservation potential of such flows seems I Snow-
relatively small because of the generally high- o ~~~,~,~,,~~,~, I I 1---r-,...,-,h--.-,,-1

relief regions that they tend to occur in. Most of fmcvcgr2 4 8 1632 0 102030405060708090100

the deposits will probably be redeposited after (b) mea~~~~:~~~:I(cml Max. clast size (em)

reworking by valley rivers, renewed avalanching Figure. 5.13 (a) Plan-view sketch of the B0ndergjerde fan
or other mass flow types, or be resedimented in in Skorgedalen, Norway. The fan is built by deposition of
lakes or in the sea either by gravity flows or wave debris flows and mixed sediment-snow flows. Note the
narrow avalanche track extending from the top of the
reworking. diagram on to the western part of the fan, and the relatively
Mountain regions of Norway provide many narrow avalanche boulder tongues. (From Blikra and
excellent examples of avalanches of mixed snow Nesje 1991). (b) Vertical section from the gravel pit on the
and sediment. The Sunnm¢>re region, western lower part of the B0ndergjerde fan. Note the interbedding
between debris flow deposits and mixed sediment-snow
Norway, is a particularly high-risk area for such avalanche deposits. The former are generally coarser
mass movements because of high mountain grained, less sorted, and possess an openwork framework
slopes and abundant precipitation and fluctuating of clasts. (From Blikra and Nesje 1991.)
140 Mass movements

temperatures around freezing during the winter mechanism (see below). Mud flows are analogous
months. An excellent example is provided by to debris flows processwise, the main difference
avalanche deposits on the Bv'ndergjerde fan in being that they do not carry large volumes of
the Skorgedalen Valley (Figure 5.13a; Blikra and debris. In the definition of A.M. Johnson and
Nesje 1991). There, elongate avalanche boulder Rodine (1984), debris flows are a process where
tongues up to 80 m wide and 250 m long were 'granular solids, in general only admixed by mi-
deposited by mixed snow-sediment avalanches. nor amounts of clay, entrained water and air,
The individual flow deposits are 1-1.8 m thick move readily on low slopes'.
and contain openwork boulder and cobble beds There is a critical thickness Tc for initiating or
interbedded with true subaerial debris flows, stopping debris flows, assuming Coulomb behav-
which are matrix-supported and generally finer iour of the debris (A.M. Johnson and Rodine
grained than the avalanche deposits (Figure 1984):
5. 13b).
(e/UW sin 0)
T c =----'---'----------'--- (5.1 )
(1- tan r/>/tan 0)
5.3.5 Fluidized flows
Fluidized flows are sediment flows where the where c is the cohesion, ¢ is the angle of internal
moving grains are entirely supported by pore friction, UW is the unit weight of the debris and
fluid moving upward (Lowe 1976a, 1979). De- e is slope angle of the surface (and base) of the
posits of fluidized flows have seldom been ob- debris flow. Thus, for instance, more cohesive
served, mainly because this flow mode is unstable debris flows can attain greater thicknesses,
and probably represents only one stage of a whereas greater slope angles will favour thinner
broader evolution during flow development or flows (Figure 5.14; Van Steijn and Coutard 1989).
flow deposition. Lowe (1982) states that fluidized
flows probably decelerate to become liquefied 5
flows or accelerate and become fully turbulent
turbidity currents. Therefore, fluidized flows are 4~~...

not treated further here.

3. 2
1-" ill'" SO
5.4.1 Introduction ill'" 0°

Flows with a plastic behaviour include debris c/j3 (cm)

flows, liquefied flows and grain flows. There is a
close linkage between fluidal and plastic flows, 3 4 5
and particular flows may show characteristics of
UWb = 1.35
both flow behaviours (Nemec 1990). Specifically,
fluidal flows may develop into plastic flows upon
deposition (Lowe 1979, 1982), or plastic flows
E 2
.. '.; ',.,,',

may develop into fluidal flows when accelerating ~ ,-,..

.. R
(Nemec 1990). Below, end members only are rl ;\\'>ll~~IIl'
summarized, but the transitional aspect is im- 3

5.4.2 Debris flow

Process Figure 5.14 The relationship between Tc (critical
thickness), Rm (radius of the largest clast), c (cohesion), uw
Debris flows are cohesive to non-cohesive masses (unit weight of the debris), UWb (unit weight of boulder or
of relatively unsorted debris that can flow on large clast) and cp (angle of internal friction) of debris
very low slopes, depending on the grain support flows. See also equation 5.1.
Flows with plastic behaviour 141

The ratio between the critical thickness and sur- developed, relatively steep snout at their down-
face slope angle is important for the internal slope end (Figure 5.16; A.M. Johnson and Rodine
shear characteristics of the flow. When the sur- 1984). Most often, the flows have superimposed
face slope angle is at the critical value, all the 'waves', which tend to move at higher velocities
internal debris will behave rigidly (Figure 5.l5a). than the main body of the flow (Figure 5.16).
When the surface slope angle is greater than the Subaqueous debris flows often may also be sheet-
critical value, the thickness of the flow will be less like, since the entrainment and rapid mixing of
than the critical thickness. In this case, there will the sediment with water may cause lateral flow
be a laminar shear zone at the base of the flow, expansion and sheet development.
with a rigid plug above (Figure 5.15b). In debris
flow channels the same relationships will apply as
shown in Figure 5.15a and b, not only in the
depth profile (Figure 5.15c and d) but also in plan Lateral deposits t
'- Medial deposits
view (Figure 5.15e). Waves
Subaerial debris flows, in particular, tend to Cross sections (c)
move down channels and therefore have a very ----~

elongate plan-view expression. Their downslope (b)

margin is often lobate, and they have a well- -------~

/Lobe (a)
8 8
Velocity profile
Figure 5.16 Schematic view of a debris flow, showing
waves and deposits formed by successive waves of debris.
(From Johnson and Rodine 1984.)

(a) (b) Most debris flows tend to show a matrix-

supported texture, with the largest clasts being
8 positioned towards the top of the flows. This is
caused by dispersive pressure between the clasts
in a buoyant matrix with some cohesive strength,
causing the largest clasts to move towards the
top of the flow. There is usually a density differ-
ence between the debris and the dense matrix,
which also gives the largest clasts buoyancy
(Rodine and Johnson 1976). Hampton (1979)
showed that the debris flow buoyancy is caused
by two factors: (i) high density of the matrix, and
(ii) loading of the pore fluid by clasts or matrix
causing overpressure that buoys up the clasts.
Velocity profi Ie Hampton further showed that even in grain-
(e) -'Plug'- matrix mixtures with a grain percentage up to 90,
Figure 5.15 Ideal flow of debris in channels: (a) and (b)
the largest clasts could be supported. Rodine and
show flow in infinitely wide channels, whereas (c), (d) and Johnson (1976) further showed that in poorly
(e) show flow in semi-circular channels. (a) The surface sorted debris flows, mobility was sustained when
slope angle is at critical value so that the thickness of the matrix was as low as 5%.
flow is equal to critical thickness. All the debris is rigid. (b)
The surface slope angle is greater so that the flow
During flow, debris flows tend to sweep clean
thickness is less than the critical thickness. There is a their pathway and thus effectively entrain large
'rigid' plug of debris above, and a zone of shear or laminar clasts (A.M. Johnson and Rodine 1984). The
flow below. The velocity profile is shown. (c) Transverse coarsest debris is generally carried in the snout of
section of semi-circular channel filled with debris. The the flow, so that an upslope fining is commonly
critical radius Rc is less than the radius of the channel. (d)
Longitudinal section of (c). (e) Plan view of (c) showing observed (Figure 5.16). Therefore, the finer and
velocity distribution expected on debris flow surface. (From more fluid upslope debris sometimes remobilizes
Johnson and Rodine 1984.) underlying coarser debris. Finer debris flows tend
142 Mass movements

to travel further than coarse, but because fine and show inverse grading. Subaqueous flows appear
fluid flows readily incorporate coarser material, a as lower-concentration flows. They are usually
simple coarse to fine gradation in a particular finer grained due to the lower competence caused
depositional setting (e.g. alluvial fans) should not by water incorporation (Figure 5.18b). Grading is
always be expected (A.M. Johnson and Rodine more common and tends to be inverse to inverse-
1984). to-normal. Clast imbrication is also more com-
Flowslides are a particular kind of violent and mon in subaqueous than subaerial flows (Glop-
very rapid debris flow (Rouse 1984). These are pen and Steel 1981).
mixtures of sediment, water and air which may Debris flows are common on many modern,
originate from very large landslides or rock- coarse-grained delta slopes, on continental slopes
fall events and in which the excess pore pressures (e.g. Masson, Huggett and Brunsden 1993) and
are very high so that the sediment strength is were also important on many ancient slopes
greatly reduced. The flows can travel at enor- (Nemec and Steel 1988, and references therein).
mous speeds, up to 500 km h - 1. It is believed that Fjord deltas in British Columbia, Canada, are
the flows travel upon cushions of fluid or air generally associated with relatively high slope
(Shreve 1968), which greatly reduce the basal gradients, and debris flows are readily formed,
friction. These flows can also travel uphill, and both on the fjord head deltas (Figure 5.19), and
are known to have become airborne (Plafker on the very steep underwater deltas along the
and Ericksen 1978) when overtopping hills. In fjord margins (Bornhold and Prior 1990; Prior
1962, two catastrophic avalanches in the high and Bornhold 1990).
Peruvian Andes caused initiation of very rapid
debris flows or flowslides. These flowed for 16 km 5.4.3 Liquefied flows
and attained speeds up to 280 km h - 1. In two
towns, Yungay and Ranrahirca in the Rio Santa Process
valley in Peru, 22000 people were killed by the Liquefied flows are sediment flows where the
flows (Plafker and Ericksen 1978). The Huas- moving grains are partially supported by up-
can'm flow, destroying Yungay on 31 May 1970, ward-displaced pore fluid generated by grain
travelled at a speed of 480 km h - 1. It bounced settling (Lowe 1976a, 1979). They are convenient-
off the gorge walls, causing superelevation at ly discussed in the present section although they
each impact point, overtopping a 150 m-high are strictly not a rheological plastic. Two ways of
hill and becoming airborne before descending generating liquefied flows are by sediment slump-
on Yungay (Plafker and Ericksen 1978; Rouse ing or by spontaneous liquefaction (Terzaghi
1984). 1947) on slopes that exceed 3-4 (Lowe 1976a).

Liquefied flows may move as, and subsequently

Products deposit sediment from, laminar suspensions.
The textural difference between subaerial and However, if they accelerate, the liquefied flows
subaqueous debris flow deposits is often pro- can become turbulent and readily evolve into
nounced and is important for interpretation of high-density turbidity currents (Inman 1963;
depositional environment (Gloppen and Steel Lowe 1976a).
1981). Subaerial debris flows tend to be more Turbidity currents can, in their very latest
clast-rich and less muddy (Figure 5.17a) than stages of flow, become laminar and attain char-
subaqueous flows (Figure 5.17b) due to the acteristics of liquefied flows. This is because
general lack of incorporation of water and sub- flow deceleration causes particle settling and
aqueous muds during flow (Gloppen and Steel hyperconcentration, preventing turbulence and
1981; see also section 5.3.4). Subaerial flows are displacing pore fluid upward, which in turn
generally also coarser, and may display a clast- causes temporary and incomplete grain support.
supported framework. The bed thickness to Liquefied flows may be unable to travel very
maximum particle size ratio is relatively low, and long distances because the grain-support mech-
generally less than 3 (Figure 5.18a). The fabric anism is insufficient to maintain prolonged move-
may be unordered, or elongate clasts may lie ment. Experimental work suggests that the travel
subparallel or parallel to bed boundaries. Grad- distance of liquefied flows depends on grain size,
ing is rare, although the lower part of the bed can velocity of the head of the flow and bed thickness
Flows with plastic behaviour 143


Figure 5.17 (a) Subaerial debris

flow deposit from the Devonian
Hornelen basin, western Norway.
Note the high clast-to-matrix ratio.
Lens cap for scale. (Photograph
courtesy of Ron Steel.) (b)
Subaqueous debris flow deposit
from the Eocene St Lloreny del
Munt fan-delta complex, Catalan
Coastal Range, Spain. Note the
steeply imbricated clasts
(a(p)a(i)-fabric), suggesting band
shearing in the latest stages of flow
(Nemec 1990, figure 33), and the
lower clast-la-matrix ratio than in

(Lowe 1976a). The finest and thickest flows are suggesting non-turbulent conditions during de-
able to travel the farthest, up to an estimated position. In several cases, water escape structures,
maximum travel distance of 10 km (Figure 5.20; such as dish and pillars, are present (Lowe and
Lowe 1976a). LoPiccolo 1974; Lowe 1975; section 4.3.4).
Deposits of liquefied flows tend to be massive 5.4.4 Grain flows
beds, generally without any evidence for traction-
produced structures (Figure 5.21). However, in Process
some cases, the top zone may be cross-laminated, Grain flows are flows of granular solids on a
suggesting late-stage reworking from the tail of slope where the grain support mechanism is,
the depositing flow or, alternatively, shear be- by dispersive pressure, caused by grain col-
tween the flow and the ambient water (Lowe lisions, and in which the interstitial fluid is the
1982; Martinsen 1987). Grading may be present, same as the ambient fluid (Lowe 1976b). Modified
as normal, coarse-tail grading (Lowe 1982). In grain flows are flows where excess density of
contrast to massive turbidites, liquefied flow the interstitial fluid, ambient fluid current-drag
deposits show no sole marks, such as flutes, or escaping pore fluid contributes to grain
144 Mass movements

Nibbevatnet fan Density modified grain flows are cohesionless

(cm) /
flows of granular solids in which the excess mass of
21.S large grains is supported by the excess density
MPS = 1.S 20
of the matrix (Lowe 1976b, 1982). This usage of the
. ..
(= 0.92" % [
40 • 0 20 MPS term grain flow is confusing, as these flows can be
better termed cohesionless debris flows, as the only
-r :;
rfl-r, difference between them and true cohesive debris

/ %[cUWo flows is the nature of the matrix. Therefore, these
o 40 SO
O!o-------'----:2::'::0--'------:-40 MPS (cm) types of flows are included under debris flows.
BTh SO Distal (1) O.6
(=0.76 Products
BTh = 2.5 20
True grain flows are common both in subaerial
MPS j / % MPS
0/../ and subaqueous settings. They are limited to
40 0 20
slip-faces of sandy dunes and bar forms of largely
.i~· 20
any scale, or to inclined, destabilized surfaces, for
j %LdJj
o 40
BTh example in slump scars that approach the angle of
o MPS (cm)
repose. On the lee sides of aeolian dunes, grain
flows occur as relatively narrow, elongate tongues
Nibbevatnet fan of sand, and they probably have the same ge-
BTh (cm) ometry on subaqueous dunes. Grain flows form
SO when the slip-face inclination increases above the
BTh =6.2 (.) angle of repose, either due to increased deposition
60 + or from erosion. Their downslope extent depends
BTh + +
MPS = 5.1 (+)
+ +-+ +
+ on slope inclination relative to angle of repose.
Where the inclination becomes less than the angle
of repose, grain flows will freeze and deposit, and
+ VIce versa.
0+' + + +.+
In cross-section, grain flows are characteristi-
20 •
• + •• + cally lenticular, a few centimetres thick and on
-. +
+ •
slip-faces of bar forms are most often coarser than
o :. the surrounding sediment (Figure 5.22). This is
o 2 4 6 S 10 12
because they are supplied from higher up on the lee
MPS (cm)
BTh = 2S.1 MPS = 5.3 face by mass movement and are not subject to
(b) + interm • distal size-sorting processes imposed on the normal
Figure 5.18 Bed thickness (BTh) to maximum particle saltational or suspended load, which also is de-
size (MPS) ratio of (a) subaerial and (b) subaqueous debris posited on the lee face.
flows from the Devonian Nibbevatnet alluvial fan in the
Hornelen Basin, western Norway. Note the much lower
BTh/MPS ratio for subaerial flows. Note also the difference
in scale for the subaqueous flow diagram. (From Gloppen 5.5 SLUMPS
and Steel 19S1.)
5.5.1 Introduction
collisions in sustaining the flow (Lowe 1976b). Slumps are downslope movements of sediments
True grain flows cannot exceed 5 cm in thick- above a basal shear surface where there is signifi-
ness, largely because dispersive pressure alone be- cant internal distortion of the bedding (Stow 1986).
comes insufficient in maintaining movement and Nevertheless, the bedding should be recognizable.
support of the grain dispersion. Thicker beds must There is a continuous transition between slides (see
have had one or several additional grain-support next section), slumps and plastic flows, and some
mechanisms. Lowe (1976b) suggested that the term units may show characteristics of all three modes
grain flow should not include flows where other of transport (Bakken 1987; see also below). There-
grain-support mechanisms operate. Other authors fore, careful analysis is required to fully under-
have also discussed usage of the term grain flow stand the movement behaviour of the deformed
(Stauffer 1967; Middleton and Hampton 1973). unit, and to categorize it satisfactorily.
Slumps 145

Arcuate scarps
Kitimat Delta Local depositional lobes
Dock facilities Shallow

Side wall fan delta

I Smooth fjord floor
'//,"''' . \" Hummocky

Out-runner blocks
Figure 5.19 Conceptual drawing of a submarine debris flow in Kitimat Arm, British Columbia, Canada. Note the
heterogeneous nature of the flow and the lobate pressure ridges in the toe region. (From Prior, Bornhold and Johns 1984).

5.5.2 Process were discussed in the context of mass movements

Slumping is a common process, particularly on by Middleton and Southard (1978) and Hampton
subaqueous slopes, and especially where there is (1979).
a significant input of fine-grained sediments. The Once initiated, the shear surface will propagate
slumps form above a basal shear surface (Figure upslope in a radial fashion from its nucleation
5.23), the depth to which is decided mainly by the point (G.D. Williams and Chapman 1983; Farrell
pressure gradient in the sediment. Where the 1984), leading to the formation of a scoop-
pore pressure approaches or balances the normal shaped, concave-downslope depression or slump
stress induced by the weight of the overburden, scar, often with an irregular outline (Martinsen
the shear strength is sufficiently reduced to allow 1989). The shear surface is probably initiated as a
slippage along a basal shear surface, given a slope-parallel feature, but at some point steepens
sufficiently high shear stress. The magnitude of to intersect the sediment surface. The transition
the shear stress acting on a slope can be viewed may occur at lithofacies boundaries (Figure 5.24),
as: or at sites of pore-pressure jumps where material
strength contrasts are present (Crans, Mandl and
i = pgsh tan e, (5.2) Harembourne 1980).
The moving slump deforms intensely inter-
where p is the sediment density, g the acceler- nally, and produces a wide variety of deforma-
ation due to gravity, s the solidity (the comple- tional structures. Folds, boudins, microfaults, in-
ment of porosity), h is sediment height (thickness) ternal shear surfaces and faults are all common
and (] is the slope inclination (see also equation structures in slumps (Figure 5.23). The occur-
1.15). Failure at the basal surface occurs when rence of these structures suggest that the slumps
the shear stress exceeds its strength, which is go through a main phase of plastic/ductile defor-
given by equation (1.5). The above equations mation whereby folds and boudins are formed.
146 Mass movements

(a) 103 irregularities. Idealized models of slumps and

slides (Lewis 1971; Allen 1985) show the de-
formed units to have a well-defined upper exten-
102 sional zone and a downslope contractional zone.
However, it is quite likely that a significant
amount of lateral compaction will occur when
slumps, particularly those that are fine-grained,
5-:>"">:>" D = 0.00625 em come to a halt, thus preventing development of
downslope or toe contractional zones (Crans,
1.0 10 103 104 10 5
Mandl and Harembourne 1980; Garfunkel 1984).
Farrell (1984) suggested that if slumps halt first
(b) 10 3
at their downslope margin, a contractional strain
wave will propagate upslope through the slump,
en overprinting any earlier formed structures by
E contractional structures. In contrast, if the
~ slumps halt first at their upslope margin (for
instance due to initial pore-water escape there),
",," <0"" >:>
.c: D = 0.0125em an extensional strain wave will propagate down-
"c:J Ud;sp" = 0.033 slope through the slumps, causing extensional
" Ux
structures to overprint earlier formed structures.
1.0 10 102 103 104 10 5 Slumps form on low slopes, as little as 0.10 or
(c) 103 less (e.g. Prior and Coleman 1978b). Very shallow
slopes are most common when slumps are for-
med from sediment finer than sand. This is be-
cause the internal deformation in slumps is gen-
erally ductile or plastic, a condition which is
promoted by the interstitial water being more
D = 0.10cm
easily retained in the pore spaces of less per-
Ud;sp_ = 0.092
meable, finer sediments.
Ux Slumps can range in thickness from 0.5 m
(Martinsen 1987) to several hundreds of metres,
10 102 103 10 4 for example, some major slumps on continental
Distance (m)
slopes (Dingle 1977; Jansen et al. 1987). There is a
Figure 5.20 Diagrams showing the relationship between
flow distance and velocity (U) of the head of liquefied flows
clear contrast in the scale of slumps observed on
of quartz spheres for silt to very fine sand (a), very fine modern continental margins to those observed in
to fine sand (b) and coarse to very coarse sand (c) in ancient successions (Woodcock 1979a). The con-
flow depths of 10, 50, 100, 500 and 1000 cm. 0 is the tinental-margin examples are several orders of
thickness of flow head. The shaded parts represent magnitude larger in cross-sectional area than
approximate ranges of natural liquefied flows. (From Lowe
1976a, which contains further details.) ancient examples. This scale difference may have
several causes. One explanation is that the size of
The ductile phase is followed by a very late brittle outcrops of ancient slumps may be far too small
phase where the faults form. It is common to see to detect the extremely large slumps. Another
strain overprinting, where early formed folds are explanation is that most ancient slump examples
truncated by late faults (Figure 5.25; Martinsen come from deltaic successions, whereas continen-
1989; Martinsen and Bakken 1990). tal-margin successions, which may preserve the
The slump folds are mainly sheath folds, for- exceptionally large slumps, have poor preserva-
med by simple shear (Figure 5.26; Martinsen, tion potential. The deformed units observed in
1989), although buckle folds can also occur most modern delta areas are generally less than
(Woodcock 1976). Faults can be both exten- 40-50 m in thickness (for example, in the Missis-
sional and contractional (Figure 5.25), and sippi Delta; Prior and Coleman 1978a,b), which
may either occur randomly within the slumps or corresponds to the scale of deformed units in
be related to local obstacles or shear surface some ancient successions (Martinsen 1989).

Figure 5.21 Probable liquefied

flow deposits from the Gull Island
Formation (Upper Carboniferous),
County Clare, western Ireland.
Note the internally massive
character and rippled top,
particularly of the lowest bed, and
flat base 01 this bed and the bed
above, suggesting laminar flow
conditions, at least in the final
stages of flow. Measuring tape is
10cm long.

Figure 5.22 Ancient grain-flow

len\icles in the aeolian Entrada
Sandstone (Middle Jurassic), near
Gallup, New Mexico, USA. Nole the
difference in geometry between
the grain lail deposits, which thin to
the left, the onlapping wind-ripple
lamination, and the lenticular,
coarser grained grain flow
deposits. Lens cap lor scale.

Slumps (and slides) may be triggered by a cause of the induced shock (Martinsen 1987,
variety of processes. Seismic triggering (Seed 1989).
1968; Leeder 1987), cyclic wave-loading (Henkel
1970; Suhayda et al. 1976), high sedimentation
rates and methane generation causing over- 5.5.3 Products
pressuring (Whelan et al. 1976; Prior and Slumping produces lenticular units of deformed
Coleman 1978b) and slope oversteepening sediment of varying scale, bounded below by a
(Martinsen 1989) are an possible initiators of distinct shear surface. Slumps tend to be charac-
gravitational sliding. It has also been suggested terized by a somewhat chaotic appearance, where
that deposition of turbidites upon a poorly primary bedding is distorted into isolated rafts or
consolidated sediment may trigger failure, be folded beds. Folds tend to occur throughout the
148 Mass movements

Tributaries units, although they are most common in the

basal and middle parts because of the increased
shear there. The slumps generany have very
sharp lateral margins, which in effect are strike-
slip or oblique-slip faults, depending on whether
the slump margins maintain parallelism to the
movement direction or are skewed. The upslope
slump scar may have tributaries or show evi-
dence of retrogressive failure, as a result of foot-
wall unloading during initial slump movement
(Martinsen 1987, 1989). The slump scars are
generally filled by fine-grained sediment, or ex-
Figure 5.23 Idealized slump model showing the most
panded by current erosion and fined by material
common location of structures associated with slumps. such as coarser grained, turbidity-current depos-
Based on field observations in the Upper Carbonif.erous its. In the latter case it may be impossible to
Gull Island Formation (western Ireland). (From Martinsen differentiate the slump scar from an erosional
1989.) channel.
Slumps form hummocky sediment masses in
Sea level plan view, particularly at their downslope ends if
toe zones are developed. The topography may be
organized into lobate, convex-downslope ridges
('pressure ridges'; Figure 5.19; Prior, Bornhold
and Johns 1984) which probably are the surface
manifestation of toe-zone thrust faults (Martin-
sen and Bakken 1990).
Numerous authors have described slumps in
the literature (e.g. Helwig 1970; Lewis 1971; Far-
rell 1984; Hill 1984; Postma 1984; Hein 1985;
Farrell and Eaton 1987, 1988; Martinsen and
Bakken 1990; Collinson et al. 1991). Some repre-
sentative examples are described below, both
Figure 5.24 Basal shear surfaces may start to climb
upsection at facies boundaries, here depicted in an from modern sediments and the ancient record.
idealized delta-front setting. (Based on Crans, Mandl and
Harembourne 1980, figure 9a.) No scale implied. The Fisherstreet slump (Namurian), County
Clare, Ireland
The Kinderscoutian (Namurian, Upper Carbon-
iferous) Gull Island Formation exposed in
County Clare, western Ireland (Figure 5.27),
represents slope deposits laid down on an in-
herently unstable basin slope (Martinsen 1987;
Collinson et al. 1991). More than 75% of the
550 m-thick formation was deformed by syn-
depositional processes, and most of the deforma-
tion was by slumping. One extremely well-ex-
posed slump occurs at Fisherstreet in north Clare
(Figure 5.27).
The slump is approximately 20 m thick and it
Figure 5.25 Late thrust fault truncating earlier slump fold retains this thickness for about 4 km of outcrop.
in the Fisherstreet slump, Gull Island Formation (Upper It consisted mainly of mud but included some
Carboniferous), County Clare, western Ireland. This
relationship of faults truncating folds is commonly sand beds toward the top. Plan-view exposures
observed in slumps (see text for discussion). Notebook, show a very organized style of folding (Figure
20 cm long (arrowed), for scale. 5.28a and b). The folds are non-cylindrical, in-
Slumps 149

Figure 5.26 Eleven stacked,

recumbent sheath fold hinges in
the Gull Island Formation (Upper
Carboniferous) in a slump at
Killard, central County Clare,
western Ireland. In some slumps,
primary bedding may be shortened
to less than 15% of the original
due to superimposed sheath folds.
Note the changing orientation of
the fold axes. Compass (circled) for
scale (10 cm long).


(al km



Figure 5.211 (a) Plan-view map of major slump fold axes

(dashed lines) in the Fisherstreet slump to show regularity.
The change in orientation toward the east may have been
( " 0 10km caused by increased lateral shear in this region.
(b) 1 =stereonet plots of fold axes ([8]) and axial plane
poles (x) of the Fisherstreet slump (n=23);
2=palaeocurrent rose diagram for overlying turbidites.
Figure 5.27 Map of locations in County Clare, western Note the similarity in inferred palaeoslope attitude (toward
Ireland, described in the text. the ENE), suggesting that there was relatively little
rotation of fold axes. This is consistent with relatively little
dined to recumbent, and tight to isoclinal with movement and/or internal deformation of the slump.
dearly thickened hinge zones (Figure 5.29). The
folds resemble early development stages of large- The axial planes dip to the WSW (Figure 5.28b,
scale sheath folds. The fold axes, allowing for the diagram 1).
non-cylindricity, are consistently oriented NNW- Normal and reverse faults are also common,
SSE, normal to regional palaeocurrent measure- and these always truncate the folds where super-
ments from turbidites (Figure 5.28b, diagram 2). imposed (Figure 5.29). The faults occur mainly as
150 Mass movements

Figure 5.29 Recumbent, isoclinal slump fold formed by simple shear and cut by later extensional fault (arrowed).
Fisherstreet slump, Gull Island Formation (Upper Carboniferous), County Clare, western Ireland.

single faults, i.e. they do not seem to be grouped failure, which occurred before 30000 years BP
into 'families'; they sole out into a common was the largest, displacing 3880 km 3 of primarily
decollement. Extensional faults and contractional young, unconsolidated sediments. The second
faults occur in groups, but show no systematic and third failures occurred in the Holocene be-
areal distribution. Nevertheless, both the exten- tween 8000 and 5000 years BP. The last two
sional and contractional faults strike systemati- events displaced a total of 1700 km 3 of sediments,
cally NNW ~SSE (parallel to fold axes from and cut into sediments as old as the Palaeogene.
slumps); the extensional faults always dip to the The second failure occurred in 3500 m of water
ENE, and the reverse faults always dip to the depth and affected sediments over a distance of
WSW. 800km.
Sandy turbidites drape the top of the slump, fiU These slope failures were described as slides by
in lows, and thin over the preserved highs. In a Jansen et al. (1987), but are probably more cor-
few places, sand volcanoes occur on top of the rectly termed slumps because of the high degree
turbidites. These were probably formed from of internal deformation in cored sediments. Most
water escape from the slump triggered by the of the sediments are transitional from slump to
deposition of turbidites (Martinsen 1989). debris flow deposits, and there are also associated
The Fisherstreet slump is probably a relatively turbidites (Jansen et al. 1987). The turbidites may
organized slump, because little variation of the have been carried as turbulent suspensions on
orientation data occurs. This may suggest that top of the displacing sediment masses, or, alter-
the slump moved only a short distance, or that it natively, could have been triggered by it. Numer-
retained its internal coherence to a large degree. ous fine-grained turbidites and creep deposits
The latter may be the result of the main deforma- occur in the head region of the slump scar. These
tion and shear taking place along the basal shear were probably triggered by both the instability
plane, perhaps as a result of a high internal and increased slope angles created by the down-
strength. Many slumps are more chaotic and slope movement of the failed sediments (Jansen et
show a large variability in the orientation data al. 1987).
over a small area (Martinsen 1987). Nevertheless, The slump scar is elongate, striking SE~ NW
the example from Fisherstreet serves to show the (Figure 5.30). The first slump encompassed the
fold style and strain overprinting that is common largest area, and the two succeeding failures are
in many slumps (Farren 1984). superimposed on this (Figure 5.30), with the last
event being the smallest in terms of area. This
The Storegg(l slumps, Norwegian continental
nested pattern suggests clearly that the triggering
margin of the two last events was related to the morpho-
A large slope failure area on the continental slope logy created by the large first event.
west of central Norway was described by Jansen This example serves to show the enormous
et aZ. (1987), based on seismic and shallow-core volumes involved in recent and subrecent mass
data. Three periods of massive slope failure were movements on continental slopes. Other such
recognized (Figure 5.30), involving a total volume examples include the slope failures off Nova
of 5580 km 3 of Cenozoic sediments. The first Scotia (Heezen and Drake 1964), Brazil (TC.
Slumps 151




Shetland Islands

.~ ................. Last slide 58°
Figure 5.30 Map of the Storegga
slumps, offshore central Norway.
Note the nested pattern of the latest
- - - - - - Second slide and smallest slump superimposed
. - . _ - First slide on the second slump, which is
superimposed on the largest and
0° earliest slump. (From Jansen et a/.

Moore et al. 1970), southern Africa (Dingle 1977, of terminology has been applied to these features
1980), northwestern Africa (Jacobi 1976), the Ca- but, in essence, they range from growth faults
nary Islands (Embly and Hayes 1974) and New bounding slides with little internal deformation
Zealand (Barnes and Lewis 1991). A review of to mudflows with penetrative internal. deforma-
large submarine slides and slumps is given by tion of the sediment (Figure 5.31; Roberts,
D.G. Moore (1978). Suhayda and Coleman 1980). Slumps (sensu
stricto) are common, and comprise the 'bottle-
Mississippi delta-front slumps neck' failures, and probably also some of the less
There is a wide variety of deformational features deformed 'mudflows' of, for example, Roberts,
described from the delta-front area of the Suhayda and Coleman (1980) (Figure 5.31).
Mississippi River (Coleman and Garrison 1977; All the slumps have a characteristic mor-
Prior and Coleman 1978a,b; Roberts, Suhayda phology, with an upslope head region that is
and Coleman 1980). Some of these features lobate and concave downslope. In some cases,
do not involve lateral translation (for instance the head region narrows downslope into a
the collapse depressions and the mud diapirs, 'bottleneck' from which the slump area widens
section 4.3.2), but most other features show into a toe zone, where the slump may overflow
downslope movement of sediment. A wide variety its margins on to the adjacent undisturbed
152 Mass movements

Interdistributary bay Undisturbed seafloor Undisturbed seafloor

Bottleneck slide Elongate retrogressive Coalescing mudflow noses Shelf edge slump
Collapse depression slide Mudflow gully Overlapping toe lobes Deep normal fault

- ~-~l3moulded sedim-en(-= -!Tludflow floor~-=-=-~-=-=-=

- - ---=-=-=-=- - - - - ---- =_Pleistocene silts-and clays-- --_-_-
Bar slumps
Peripheral slumps
Incipient mud diapir Continental slope mud diapir
Figure 5.31 Overview sketch of instabilities on the Mississippi delta-front. Most instabilities range from slumps to flows of
muds. (From Prior and Coleman 1982.)

sediment (numerous examples appear in the Cratsley and Whelan 1976; Whelan et al. 1976).
conceptual model in Figure 5.31). In other cases, The slope of the Mississippi delta-front is generally
the head region of the slump is dominated very low (less than 2°, e.g. Prior and Coleman
by several 'tributaries', which may be highly 1978b (see Figure 1.8)), so a significant reduction of
skewed in orientation to the main slump direc- the sediment shear strength is needed for instabil-
tion (Figure 5.32). In ancient successions this ity to occur.
pattern is important when palaeoslope directions This example shows that slumps can occur on
are measured. Tributaries are probably related very low slopes, provided that the conditions are
to retrogressive failure, in which the upslope favourable. It also shows that a range of slump scar
scarp formed by the initial slump movement morphologies is to be expected when analysing
becomes a 'free' end, and causes unloading of the ancient deposits.
adjacent footwall, leading to renewed failure (see
Figure 1.9). A local slope then exists between the
adjacent sediment and the slump scar, irrespec- 5.6 SLIDES
tive of the original orientation of the main delta
5.6.1 Introduction
The slumps leave a chaotic deposit character- Slides are downslope displacements of sedi-
ized by floating blocks in a deformed sediment ments above a distinct shear surface where there
mass (Roberts, Suhayda and Coleman 1980). is little or no internal deformation of the
Whether the deposits should be termed slumps or transported material (Figure 5.33; sensu Stow
flows is largely semantic, but to some degree 1986). Slides are fully transitional from slumps,
depends on the movement distance of the de- and in cases the differentiation may be semantic.
formed sediment. The further the movement, the However, an arbitrary boundary between slides
more likely it is to turn into a flow. and slumps may be at the point where the original
Overpressure in the sediment caused by high bedding is folded and/or disrupted. Thus, in
sedimentation rates (hindering normal pore- slides, the original bedding can be slightly rotated
water escape), together with methane (CH 4 ) along fault planes, such as into hangingwall
from degradation of the high amount of organic anticlines, but not deformed as a direct response
material in the sediment, are probably the pri- to simple shear or buckling. A special case may
mary mechanisms for causing instability (Roberts, occur where layer-parallel slip has caused intra-
Slides 153

10-25 C. 24000-12000 yrs p)

Contour inlerval2501t (76m) I

Figure 5.32 Plan view of major slump scar on the Mississippi delta-front. Note the highly skewed orientation of the
scar tributaries in relation to the main scar. This pattern is probably caused by a local excessive slope gradient being
set up as a result of destabilization and footwall unloading following removal of the sediment in the major scar. (From
Coleman, Prior and Lindsay 1983.)



Internal exten-
sional zone
Internal shear fold

Leading-edge f;;ids
Figure 5.33 Idealized model of a slide showing the most common location of associated structures. (From Martinsen 1989;
based on field observations in the Gull Island Formation (Upper Carboniferous), County Clare, western Ireland.)

stratal sheath folds to form. These failed sediment lational slides; Allen 1985) (Figure 5.34). This
masses are still termed slides because the main means that sediments displaced by growth faults
part of the bedding is retained in its original and shelf-edge fault systems are termed slides,
configuration. because of the focus on their bounding discon-
The following discussion encompasses sedi- tinuities.
ments which at their upslope ends and bases Slides span a great range of slope instabilities,
are bounded by clearly defined fault planes. including such diverse features as bank collapse
Therefore, sediments moved for only short dis- features in river channels, subaerial mudslides,
tances compared with their thickness are in- delta-front growth faults, shelf-edge faults and
cluded (rotational slides; Allen 1985), together submarine glide-blocks and olistoliths. The fol-
with sediments moved for relatively long dis- lowing general outline of slide theory is discussed
tances compared with their thickness (trans- only briefly as they have much in common with
154 Mass movements

shear surface. The downslope toe region is

usually dominated by contractional deformation
and has a convex downslope and characteristi-
- ~ " Rotational cally lobate form.
~-",-- The bead region is usually dominated by ex-
tension along a group or 'family' of listric faults
(Crans, Mandl and Harembourne 1980) which
tend to sole out at a common level or basal
decollement. In some large slides the fault fami-
lies occur at several orders of magnitude, so that
one small family may be entirely enclosed in the
hangingwaH of a larger order fault family (Figure
figure 5.34 Schematic illustration of the difference 5.35). Quite commonly, antithetic extensional
between rotational and translational slides. Rotational faults occur, which may be either listric or planar.
slides are only moved for a very short distance relative to Antithetic faults show downthrow in the opposite
their thickness. thus the sliding motion is mainly rotational. sense to the master faults, which can significantly
Translational slides are moved for relatively long
distances, or affect longer areas of the slope compared
confuse measurements of palaeo slope (Martinsen
with their thickness, so that a translational mode of 1987). The existence of fault families in slides is an
movement is dominant (see also Allen 1985). essential difference from slumps, where most
faults are single faults (see above).
slumps. The main difference between these two The central part of slides generally shows little
mass movements is simply the degree of internal evidence of the sliding itself, and the basal shear
deformation. zone can be extremely difficult to detect, particu-
larly in fine-grained sediments. In places, there
5.6.2 Process may be evidence for internal slip between beds in
the form of sheath folds or microfaults. These
In their most simple form, slides are spoon- structures are important to detect and provide
shaped features with a three-part morphology of important information on the translated nature of
upslope head region, middle 'rigid' zone and the sediment when other features are not present.
downslope toe zone (Figure 5.35; Brunsden 1984; The margins of the slides are dominated by
Gawthorpe and Clemmey 1985; Martinsen 1989). strike-slip deformation, but the width variability
The upslope head region is concave downslope of the slide scar can cause both transpressional
and dominated by extensional deformation, and transtensional movement if the slide scar
whereas the middle region is mainly translational narrows or widens. Any evidence for strike-slip
and may behave in a rigid fashion and not show motion is important to detect, because it shows
any particular strain signature above the basal that the slide margin observed is a lateral margin,
and not a head or toe region.
Contractional deformation dominates the toe
region, and this is most commonly expressed in
the form of thrust faults. The thrusts form classic
duplex and imbricate zone geometries (Figure
5.36a and b; Lewis 1971; Dingle 1977; Martinsen
and Bakken 1990), which often may be very
difficult to differentiate from the familiar post-
lithification structures. In plan view, the mor-
phological expression of the contractional toe-
zone is sometimes that of downslope lobate
ridges, or 'pressure ridges' (Figure 5.19; Roberts,
figure 5.35 A family of two growth faults (labelled 1 and
2) at the upslope margin of a major slide complex at
Suhayda and Coleman 1980; Prior, Bornhold
Edgeii\ya, Svalbard, Norwegian Arctic. Note the and Johns 1984). Martinsen and Bakken (1990)
listric shape of the faults and the growth of several beds suggested that three different types of contrac-
across (2). The section shown is 150 m thick. tional zones occur in slides (Figure 5.37). The
Slides 155
w E


o 3m
N~ vertical exaggera:ion

Figure 5.36 (a) Contractional slide zone in the Upper

Carboniferous Gull Island Formation at Failure Point,
Gull Island area, County Clare, western Ireland (see
Figure 5.27 for location). (From Martinsen and Bakken
1990; based on Marlinsen 1987.) (b) Picture of righi-hand
part of (a). The ramping thrust (marked 'roof thrust' in
(b) (a» and the sale thrust are indicated. Compare with (a) for
location of other structures. Backpack for scale.

No vertical exaggeration. o 3m
thick package zone is the most complex and is
Scale only
suggestive Backthrust Second-order zone most similar to mountain-belt thrust zones, al-
though the scale is several orders of magnitude
smaller. Thin package zones seem to develop at
the frontal margin of heterogeneous slides, par-
o 3m
ticularly where sand and mud is closely interbed-
Extensional fault
L '_ _- "

ded (Figure 5.37; Martinsen and Bakken 1990).

Sheared Basal contractional zones are not related to toe
mUdstone Sheared mudstone
zones but rather to obstacles, such as ramps, at
I the base of slides. Such zones can be mistakenly
2 interpreted as toe zones and their context can be
Main slump o 3m understood only through careful analysis.
In several instances, slides do not seem to have
clearly defined toe zones, and some slides are
'open-ended'. This is probably caused by a high
degree of lateral compaction so that the strain
3 piles is taken up by porosity reduction rather than
Figure 5.37 Models for contractional zones in slides. by thrust-zone formation (Crans, Mandl and
(From Martinsen and Bakken, 1990; based on Martinsen
Harembourne 1980; Mandl and Crans 1981;
156 Mass movements

Garfunkel 1984; Martinsen 1987; Barnes and stress component will be oriented in the direction
Lewis 1991). This is probably particularly com- of progradation due to the progressive loading in
mon for slides in muds and clays. that direction. Therefore, the thrust faults will
Toe zones of slides may be confused with show propagation in an offshore direction, parallel
contractional zones formed by gravity spreading. to the generally expected sliding direction of slides
Gravity spreading forms as a result of loading a and slumps down the delta slope. Pedersen (1987)
substratum by an overlying medium (Figure 5.38; listed criteria to distinguish gravity sliding from
for example progradation of sandy delta-front gravity spreading (Figure 5.38), but in areas of
sediments over muddy prodelta deposits). During insufficient exposure, the two may be inseparable.
spreading of the load, a lateral stress component Gravity spreading occurs on a variety of scales,
forms which causes the loaded medium to be from minor deltas to plate-convergent zones (see
pushed sideways and upwards along thrust faults Pedersen (1987) for a discussion). Galloway
(Bucher 1956; Galloway 1986; Pedersen 1987). In (1986) also discussed gravity sliding and gravity
progradational settings, such as deltas, the lateral spreading in Cenozoic sediments of the Gulf of
Mexico. The Niger Delta shows excellent
Sliding Spreading
examples of gravity spreading caused by loading
of the deltaic sediments on offshore fine-grained
deposits (Weber and Daukoru 1975).
Because both slides and slumps can occur on
1~ very low slopes, piling up of deformed material
can easily generate local slopes of orientation
highly skewed to the regional slope. Therefore,
both extensional and contractional structures
2[~7l with very variable orientations should be ex-
pected to occur on, for example, delta slopes.
These can occur particularly where main slide
scars are fed by tributaries (Coleman et al. 1983).
1 3 2 1
Significant departures can occur from the
3FA~ idealized model discussed above. Commonly, the
slide will be a heterogeneous mass that moves
above a heterogeneous substratum. Therefore,
local stress regimes can be set up within the slide
mass, and cause both local extensional as well as
contractional zones to form. Consequently, ex-
Figure 5.38 Conceptual diagram to show the difference tensional zones are not constrained only to the
between gravity sliding and spreading. The criteria head region of a slide, neither are contractional
for distinguishing gravity sliding are listed in the left
column and include: 1, listric normal faults at the trailing
zones confined to the toe region of a slide (Mar-
end of deformed sheet; 2, flat-lying thrust sheet disturbed tinsen 1989; Martinsen and Bakken 1990).
by extensional deformation; 3, diverticulation, whereby the A particular kind of slide occurs when exotic
upper stratigraphical units are displaced further than the rock units slide downslope. These slide blocks are
lower; 4, exposure of a 'peel-off' region in the rear end of termed olistoliths and are particularly common in
the thrust fault region. Four criteria for distinguishing
gravity spreading are listed in the right column and carbonate terranes (Conaghan, Mountjoy and
comprise: 1, an imbricate fan formed by listric, splay, Edgecomb 1976; Bosellini 1984), and can be of
thrust-faults in front of the overlying spreading mass; 2, gigantic size, up to several cubic kilometes in
a duplex of imbricate thrust sheets formed in the deformed volume. Melange terranes are other settings
and overthrust sediments - because of loading, boudins
may be formed beneath the spreading unit; 3, in the frontal
where slide blocks are common (Swarbrick and
part of the gravity-spreading deforming system, Naylor 1980; Naylor 1982), but they can also be
syndeformational basins are formed with progressively found in purely clastic marine settings (e.g. In-
younger sediments away from the front of the spreading eson 1985). Naylor (1982) listed characteristics of
mass; 4, due to increasing overpressure in the
decollement, water-escape structures and mud-diapirs
exotic blocks displaced by gravity sliding, focus-
occur in the frontal region of the gravity spreading system. ing on field relations and orientations of the
(From Pedersen 1987.) blocks, external morphology and host-sediment
Slides 157

, Draping by host Talus breccias at The Gela slide itself occurs immediately above

sediment sides and top only
the tip of the Gela nappe. It covers 1500 km 2 , and
- ... ' has a strike-parallel, very elongate shape. In
. . . . . cross-section (Figure 5.41) the slide appears as an
undisturbed mass, where the original bedding is
, .. =;:i2' ..
well-preserved, as shown by the continuous and
/. parallel reflectors in seismic sections.
Trincardi and Argnani (1991) recognized four
Soft-sediment emplacement· Glide-induced shear brecciation
related structures at base only and folding concentrated at the main parts of the slide and related features,
base. Upper and lateral mar-
gins undeformed. Breccia matrix namely slide head, slide toe, lateral ramps and a
of intraformational sediment chaotic unit which probably filled the slide scar
Figure 5.39 Schematic illustration of principal features of (Figure 5.41). The slide head is dominated by one
submarine slide blocks. (From Ineson 1985; based on extensional fault that slopes 15° at the most
Naylor 1982.) towards the basin, but it is a listric fault and soles
deformation. Ineson (1985) pointed out that in out underneath the slide. The slide scar, the
complex tectonic terranes it may be difficult to surface expression of the detachment, is continu-
differentiate between gravity slides and thrusting. ous along strike for at least 120 km. The relative-
In particular, the internal deformation of the ly long extent along strike is probably related to
gravity-slid block may be very similar to tectoni- the fact that the slide occurs directly above the
cally generated structures (Figure 5.39). However, Gela nappe, which seems to extend for a similar
gravity sliding may be indicated by increasing distance (Figures 5.40 and 5.41). Only one exten-
internal disruption towards the base of the block, sional fault is found in the head region. This also
and by the style of deformation, which is mainly may be related to its position immediately above
by simple shear under low overburden (Figure the nappe.
5.39; Ineson 1985). Immediately downslope from the head, the
slide appears undeformed. This central propaga-
5.6.3 Products tional or translational zone is typical of many
slides, and represents the 'rigid' central zone
Gela submarine slide, Sicily foredeep where little or no deformation occurs above the
The Gela Basin, southwest of Sicily in the Medi- basal detachment. In several instances, it is im-
terranean (Figure 5.40), is filled with up to possible to detect evidence for sliding within this
2500 m of marine sediments that shallow upward, zone because deformation is confined to the basal
and are of Pliocene to Quaternary age. The basin shear zone, which itself may only be a few mil-
is situated at the front of the Gela nappe, and is a limetres thick.
foredeep basin to the Maghrebian fold-thrust Further downslope, the slide toe region shows
belt (Trincardi and Argnani 1991). well-developed contractional deformation with
thrusts, folds and imbrication zones. The thrusts
caused thickening of the slide body in this region
(Figure 5.41), producing a positive feature on the
sea bottom (Trincardi and Argnani 1991).
Lateral ramps formed along the sides of the
slide. Local detachments have formed where
abrupt changes take place in their distance from
the basal slip surface. The slid material is de-
formed and folded above the lateral ramps to
accommodate the geometry of the basal slip
~ Possible debris flow deposits
~ Front of Gela Nappe
Above the head-zone roll-over, a chaotic unit
Schematic section fills in the depression created by the downslope-
13° 00' 13° 30' translated material. Trincardi and Argnani
Figure 5.40 Location of the Gela slide, Sicily foredeep. (1991) interpreted this unit as a debris flow de-
(From Trincardi and Argnani 1991.) posit, possibly formed by the locally increased
158 Mass movements

Contraction Propagation Extension

A I Frontal ramps and related Undeformed layers IRoll-over Slide scar 1A'

SW 2 km

r-----d Late-Pleistocene
------ prograding wedge
I. ...... ... ......1 flow
Possible debris
~. ...... Slide-involved
L:.:::....J unit
rrmmm Foredeep onlapping-filling
WillilllJ sequence
Foredeep basement ~GelaNappe

Figure 5.41 Cross-section (A-A' in Figure 5.40) of the Gela Slide. See text for discussion. (From Trincardi and Argnani

slope-angle due to the formation of the slide scar. combinations of these. Slides are especially com-
This relationship of slumps or debris flows filling mon in the upper part of the slope succession,
slide scars is also observed in ancient slide de- which is particularly muddy and fine-grained.
posits (see below). One of the slides, informally named the Point
The Gela slide is an example of a simple slide of Relief mudslide, is well exposed in a cliff-
where the tripartite division of extension, transla- section along the Atlantic Coast (Figure 5.27).
tion and contraction is well displayed. There is The slide disturbs more than 35 m of mudstone,
no evidence for strain overprinting in the internal and occurs approximately 100 m below delta-
extensional and contractional zones. Strain over- front sediments of the overlying deltaic cycle. The
printing may occur if heterogeneities are encoun- depositional setting was therefore probably a
tered (for example, uneven or ramping decolle- distal prodelta environment.
ments), if the slide stopped abruptly so that a The largest deformation features are two west-
compressional strain wave propagated upslope erly dipping, normal master-faults (EF and WF;
through the slide, or if the sliding motion was Figure 5.42a). The sediments east of EF are
hindered in the head zone, causing an extensional undeformed, thus the two faults probably repre-
strain wave to propagate downslope through the sent the head region of a slide. Displacement can
slide (Farrell 1984). only be determined along WF, and is around
12 m. Both faults are concave-up and listric, but
Point of Relief slide, County Clare. Ireland no decollement is seen as the faults extend below
The 2100-m-thick Namurian (Carboniferous) sea-level. Between the faults, several minor scale
succession of County Clare, western Ireland, re- faults occur, both with synthetic and antithetic
cords the fill of a deep, symmetrical ENE-WSW displacement (Figure 5.42a). Some of the bedding
elongate trough which formed above the Iapetus is clearly rotated, but the displacement on these
Suture as a response to N-S extension (Rider faults does not exceed a few metres. There is
1974; Collinson et al. 1991). The fill is a shallow- normal drag of the bedding in the hangingwall of
ing upward succession, from deep basinal shales, EF, but this is not so clearly developed along
through turbidites and a slope succession, into WF. The pre-lithification nature of the faults is
deltaic cyclothems. The slope succession, the Gull clearly displayed by the truncation of WF by the
Island Formation, is 550 m thick, of which 75% overlying beds (Figure 5.42a).
is deformed as a result of sediment deformation Other deformation features occur in the cliff,
processes (Martinsen 1989). The deformation which are related indirectly to the sliding. A very
style is either slumping, sliding, water escape or contorted zone occurs near the base of the cliff,
Slides 159


Beds thicken toward fault

Figure 5.42 (a) Drawing of the Point of Relief (for location, see Figure 5.27) slide in the Gull Island Formation (Upper
Carboniferous). Gull Island region, County Clare, western Ireland. (From Martinsen and Bakken 1990; based on Martinsen
1987.) (b) Photograph of areas labelled WF and EF on (a). Note fill in hangingwall. EF is seen in the lower
right-hand corner. The cliff is 50 m high.

and this is cut by WF and therefore pre-dates the temporal evolution of increased remoulding with
slide. Above this, a wedge-shaped bedset occurs time (sliding followed by slumping and/or debris
which records infill of the depression in front of flows) suggests that sliding is an important pro-
WF (Figure 5.42a and b). The beds above this cess in forming local excess slopes. These in turn
wedge are slumped and folded into an isoclinal, prompt further failure and other more intense
recumbent fold which covers the entire thickness types of deformation.
of the slump, but is truncated on the top. It is
probable that the slump moved in the slide scar Kidnappers slide. New Zealand margin
and was banked up against the slide-scar wedge. The Kidnappers slide occurs off the eastern mar-
The same may be true for the overlying slump, gin of New Zealand's North Island. This is a
which also thins towards WF. convergent plate margin with the sea floor slop-
The Point of Relief slide shows the importance ing around 1-50 into the Pacific Ocean, where
of interaction and superimposition of sliding, the Pacific plate is being subducted underneath
slumping and in situ deformation (the contorted the Indo-Australian plate (Figure 5.43). In the
zone at the base of the cliff). The similarity to the early Holocene, the lowstand sediments of the
Gela slide (Trincardi and Argnani 1991; see last glacial age failed on the upper continental
above), where debris flows filled the depression slope seaward of Cape Kidnappers. The slide
in front of the master fault, is significant. The was originally investigated using seismic data by
160 Mass movements

Figure 5.43 Location map of the Kidnappers slide area, offshore North Island, New Zealand. (From Barnes and
Lewis 1991.)

Lewis (1971), but recently has been re-examined deep-seated normal faults, and no basal slide
by Barnes and Lewis (1991). surface can be detected (Figure 5.44b). The dif-
The Kidnappers slide is a composite unit of ferences between these two areas and the deep,
several slides and slumps dominated by exten- rotational slumps illustrate important along-
sional, rotational failures (Figure 5.44a). The slope variation in slide style (Figure 5.44b). The
minor deformed units range in thickness from 20 latest phase of deformation encompasses transla-
to 140 m, and slope failure is thought to have tional sliding in the central slide area, which has
occurred in several phases. The main failure deformed some of the earlier-formed normal
probably took place in early Holocene time, but faults.
failure occurred more or less continuously from Only extensional zones can be documented
mid-glacial time up to the present. The entire within the entire slide area, and no evidence is
slide area is around 720 km 2 in area, and the found for contraction. Therefore, only a series of
total displaced sediment volume is around normal fault scarps is observed, which produces a
33 km 3 . very uneven sea-bottom topography. The ab-
Three phases of deformation can be discerned. sence of contractional faults may be explained by
The earliest phase comprised formation of now- the underlying large-scale tectonics. Imbricate
buried slumps, bounded upslope by normal thrust faults occur within the sediment package,
faults. The middle phase was the most volumi- producing growth anticlines (Figure 5.44a and b).
nous, and the slide area can be divided into two The upslope anticline appears to grow the fastest,
areas, northern and central (Figure 5.44b). In causing an increase of slope inclination toward
addition, deep rotational slumping occurred. The the basin. This has probably induced a general
central slide block was the largest, covering ap- state of extension within the entire sediment
proximately 500 km 2 , and can be described as a package, so that all the stress is resolved by the
large sheet failure affecting 20-70 m of sediment, normal faults. In addition, lateral compaction
where a lower basal slide surface can be recog- may have contributed to the prevention of com-
nized. The northern slab is affected mainly by pressional structures.
Slides 161

, Shelf edge




Lachlan Ridge

Recent basin
subsidence faults

Figure 5.44 (a) Cross-section of part of the Kidnappers slide, showing the relation of extensional zones to
thrust-controlled growing anticlines of the offshore imbricate-thrust wedge. (b) Various cross-sections along-slope of
the Kidnappers slide. Note the along-slope variation in structural style. (From Barnes and Lewis 1991.)
162 Mass movements

The Kidnappers slide is important because it

shows that 'open-ended' slides may be quite com-

mon. In addition, the variation in structural style
along the slope is an important contribution,
indicating that the slide is in fact a complex and -~
composite unit of several deformed units and

" Weight .",.,

several deformational styles. Weight Contracting r /
(a) force I /
Range of Contraction
5.7 CREEP particle ...
.. ", ...
5.7.1 Introduction
Creep, used here in the conventional geomor-
phological sense (see section 1.2.4), is the slowest-
moving mass movement and involves the slow
translation of sediment or soil down a slope
(Radbruch-Ha1l1978). Creep takes place at a rate
that cannot be observed directly with the naked (b)
eye (McKean et al. 1993). The moving material Figure 5.45 Diagram to explain the origin of
may move only a few centimetres a year or even slope-induced physical creep. See text for full explanation.
less, but rates up to 20 cm day-l have been (From Allen 1985.)
measured (Radbruch-Hall 1978). Nevertheless, it
is an important process both subaerially and the same happens, only the contracting force is
subaqueously because it can lead to substantial oriented in the opposite direction to the expan-
land waste, and, in addition, creep often triggers sion force. The resultant downslope motion thus
other and more violent mass movements, such as defines a zig-zag pattern as a consequence of
landslides (Heim 1932; Muller 1964). repeated expansion and contraction (Figure
5.45b; Allen 1985). Organic activity by animals
5.7.2 Process and plants also contributes to creep.
Thus, this kind of creep is controlled mainly by
Subaerial creep climate (Allen 1985; Feda 1992), and the controll-
Subaerial creep can be divided into two classes ing mechanisms differ between climatic regions. In
(Carson and Kirby 1972): rheological creep and tropical and subtropical areas, wetting and drying,
physical/organic creep. The former relates directly heating and cooling, and organic activity control
to soil clay minerals and is caused by continuous creep, whereas in mountain areas and arctic
breaking and re-establishment of clay mineral regions, creep is mainly controlled by freezing and
bonds under the influence of gravity. Each indi- thawing. Solifluction is a special and relatively
vidual bond break is insignificant, but the cumu- sporadic (but also relatively rapid) form of creep
lative effect on a soil mass is important in causing which combines the effects of freezing and thawing
downslope movement. with rheological creep, because of internal slippage
Physical creep is more important than during thaw due to wetting and soil saturation.
rheological creep and occurs on a seasonal or Rates of creep depend to some degree on
diurnal basis in soils. Three mechanisms can be climate and vary between a maximum of
distinguished (Allen 1985): (i) heating and cooling 0.3ma- 1 (0.01-0.1ma- 1 is more usual) in
of particles; (ii) wetting and drying; and (iii) solifluction-dominated arctic and mountainous
freezing and thawing of interstitial water in soils. areas, 0.002-0.01 m a -1 in temperate continental
These mechanisms all produce contraction and areas, a maximum of 0.002 m a - 1 in temperate
expansion of the soil. During expansion or heave, maritime climates and up to 0.005 m a -1 in tropi-
the particle moves along the resultant vector cal rainforests (Allen 1985; see also Brunsden
between the directions of expansion (directed 1979).
upwards normal to the surface) and the particle Soil creep decreases below the surface, and can
weight force (Figure 5.45a). During contraction, reach depths up to 10 m with solifluction. More
Creep 163

commonly, the maximum depth is less than 1 m. oriented parallel to valley walls (Figure 5.47; e.g.
In temperate and warmer regions, creep rarely Ferguson 1967). The joints at the top of the
exceeds depths of a few tens of centimetres (Allen valley slope are commonly more developed than
1985). Vertical profiles of creep rate were pub- those at the base of the slope. A gouge or
lished by Fleming and Johnson (1975) and are mylonite-like zone can be developed at the base
shown in Figure 5.46. Convex-up, rectilinear and of the valley slope, suggesting that the entire
inflected profiles occur, but the last two may be valley slope is moving outwards, towards the
most common (Allen 1985). valley centre (e.g. Ferguson 1967; US Corps of
Engineers 1973).
Rock creep can range from less than 2 cm a - 1
:,,: ,'~

... :•.... ::: ..
: . \ : ,.......•...
.. .: ..•....
(Huffman, Scott and Lorens 1969) to more
. . •. :.:'.:. •.:':.:.:.;;.> .•• :.: .• / '< ..\; ....... .

than 20 cm day-1 (Muller 1968). It can affect

U(y) U(y) the bedrock up to depths of 300 m (Nemcok
\" 1972).
(a) (b) (c)

Figure 5.46 Variation in velocity profiles in creeping

Subaqueous creep
soils. Uy represents the creep rate in the y direction, On subaqueous slopes, contraction and expan-
parallel to the land surface. (From Allen 1985; based on sion of sediment cannot explain creep generation
Fleming and Johnson 1975.)
because continuous or semi-continuous mechan-
isms producing this behaviour do not exist.
Periglacial environments are particularly Rather, it is thought that slow, intergranular
prone to mass movement of weathered debris frictional sliding of non-cohesive sediment better
because of seasonal freeze and thaw, water- explains creep (Nemec 1990). The strain rate
saturated sediment from meltwater and frozen must be low, preventing the development of well-
ground at depth, which prevents deep perco- defined slip planes.
lation of water (Brunsden 1979). A particular Creep on subaqueous slopes has not been
kind of creep occurs in periglacial environments studied in detail. It may be an important process
and is called gelifluction (sensu Baulig 1956). This on steep, coarse-grained shorefaces or deltas
occurs where weathered, thawed debris saturated where the slope inclination may be up to 35° (W.
with meltwater flows over frozen subsurface beds. Nemec, personal communication). In these set-
Gelifluction can be more rapid than creep or tings, creep can be important in stabilizing the
solifluction due to freeze-thaw cycles, but in slope, as the slow sediment movement reduces
places the two processes can operate at the same the gradient, which may reduce the risk of more
time (Brunsden 1979). massive slope failure. It is unlikely that creep is
Rock glaciers are also a periglacial phenome- an important process on low-angle delta slopes of
non and occur as lobes of angular boulders fine-grained sediment. The sediment is unlikely to
below cliffs. Their shape is similar to glaciers behave in an intergranular frictional manner, and
and they occur preferentially on northern slow mass movement is probably rather by slid-
slopes near the snow line in mountain areas mg.
(Flint 1971). Rock glaciers form and move as One way of initiating subaqueous creep is
a result of interstitial ice and snow 'lubricating' perhaps by 'gravitational winnowing' or expul-
grain contacts under the influence of gravity sion (Nemec et al. 1984; Postma 1984), in which
(Wahrhaftig and Cox 1959). This results in the interstitial fines in heterolithic mass-flow de-
downslope creep, which can reach up to posits liquefy and are expelled from the steep
150 cm a - 1. Ridges, lobes and crevasses form that snout of the flows. This renders the coarse, re-
are analogous to those formed in glaciers (Flint maining fraction more susceptible to creeping
1971). because the overall frictional resistance will be
Rock creep is a common process on slopes reduced when the pore spaces are emptied of
where loose or solid rock is influenced by gravi- matrix. Finally, flow of pore water downslope
tational pull (a detailed account was given by will exert a stress or drag on grains which can
Radbruch-HallI978). This generally causes faults lead to creep. This may be important on steep
and fractures to form, which most commonly are slopes.
164 Mass movements

EJ. 347.5.
EJ. 338.3 (1110.0) (1140.0)

Figure 5.47 Cross-sectional view of a valley in the Alleghany Plateau region, eastern USA, showing vertical jointing of
valley sides and bulging and fracturing of the valley bottom as a result of rock creep. (From Radbruch-Hall 1978; based
on Ferguson 1967.)

5.7.3 Products
Subaerial soil creep produces characteristic de-
posits (Cotton and Te Punga 1955; Flint 1971;
Feda 1992) and morphologies, which can have
important environmental effects. Creep can lo-
cally increase the slope inclination, causing mud-
slides and debris flows to be generated. More
commonly, creep results in such well-known ef-
fects as the bending of trees, moving of fences and
formation of creep tongues (Figure 5.48). The
tongues lack sorting, and are mainly fine-grained.
Sometimes, clasts up to boulder size are included.
The larger clasts may be imbricated, and are
characteristically angular and of local prov-
enance (Flint 1971).
The creep tongues are generally less than 1 m
thick, but on flatter areas they may attain greater
thicknesses, particularly where several tongues
are superimposed. The tongues are smooth, es-
pecially where the soil is relatively permeable,
allowing downward percolation of rain-water
and prevention of tongue destruction from rain-
water runoff (Flint 1971). Gelifluction produces
similar deposits to solifluction, but is constrained
to treeless areas.
Rock creep produces blocks and disrupted
bedding (Figures 5.49), and is a prime initiator of
rock fall. The rocks deform mainly by buckling in
a convex-upward fashion. This causes the rocks
below buckles to be steepened, commonly Figure 5.48 Solifluction tongues on a grassy, cultivated
prompting the rocks upslope to penetrate under- slope, near Ulven in Os, western Norway. Note the high
neath the downslope part of the buckle (Figure number of tongues, which generally are 10-30 cm across.
Creep 165

movement, perhaps by enhanced creep (Heim

At Monte Toc, in the Italian Alps, a mass of
limestones slid into the Vaiont Reservoir in 1963.
The slide was initiated by creep, and caused a
flood wave 100 m high to form, which overtop-
ped the local dam and flowed into the Adige
Valley. Almost 2000 people were killed in the
Figure 5.49 Conceptual sketches to show steps in the
buckling of a sandstone bed above mudstone due to rock town of Longarone (Muller 1964). These two
creep. (From Radbruch-Hall 1978.) examples show the importance of creep in
generating catastrophic mass movement. There-
5.49). Apparent thrust faults may therefore form fore, monitoring creep can be an important prac-
(Radbruch-Hall 1978). tice in risk analysis of mass movements.
Rock creep may initiate major catastrophic Although subaqueous creep is known to occur
mass movement events. In 1806, at Goldau, (Hill, Moran and Blasco 1982), it is not clear
Switzerland, a block of Tertiary conglomerate to what extent it produces characteristic deposits
slid down a 20° slope (Heim 1932) and 457 in the way that, for example, turbidity currents
people in the village were killed. Ter-Stepanian or debris flows do. The effects of subaqueous
(1969), citing Zay (1807), pointed out that creep creep may be difficult to recognize but more
had occurred for 20 years at the site before research needs to be carried out to assess the
the slide. In addition, animals were restless for characteristics and overall importance of this
several hours before the slide, suggesting early process.

Tectonic deformation:
stress paths and strain histories

6.1 INTRODUCTION hence mineralization and fault behaviour. There

are also very practical reasons for understanding
The traditional endeavours of structural geo- the evolution of stress in the shallow crust, and
logists are now being extended, as indicated in especially in sediments. These include assess-
section 1.3.5, to shallow levels of the earth's crust ments of the stability of boreholes, the behaviour
and to situations where incompletely lithified and manipulation of subsurface reservoirs, and
sediments have been subject to tectonic stresses. the suitability of sediments for the foundations of
Such stresses, although by their nature of deep- structures.
seated origin, can extend to the highest levels of Geomechanical models are now becoming
sediment piles, and, as explained in section 2.2.7, more realistic and reliable, but they require ad-
the particulate deformation that characterizes equate knowledge of stresses at specific mechan-
sediments can persist to substantial depths of ical states, such as at brittle failure. However,
burial. Much of the sediment that is affected by because stresses are not preserved in rocks and
tectonic stress is in the broad and ill-defined area sediments, they cannot be determined directly
of partial lithification. Here, as this chapter illus- but must be deduced from the sequences of
trates, the mechanical effects of diagenesis be- structures and strain fabrics that happen to have
come important, as do the elastic responses of the been preserved. Such interpretive analyses are
sediment. very difficult for several reasons. First, there is
One objective in structural geology is the inter- still only a poor understanding of applicable
pretation of deformation histories of geological rheologies that couple strain to stress along geo-
bodies and of the stress paths responsible for logical deformation paths. Second, deformation
those histories. This task, incorporating the paths involve both elastic and inelastic strains,
evolution of volumetric strains, is particularly and only the inelastic components are preserved
relevant to particulate sediments and their in the rock fabric. Third, there may prove to be
equivalents in the geological record. The prin- non-unique couplings of stress paths to deforma-
ciples were outlined in section 2.2.5. This chapter tion histories. In addition, changes in the stress
focuses on the deformation that sediments are tensor, rotations of the rock element in the stress
likely to encounter in settings where tectonic field, and changing physical properties along the
stresses operate, and it discusses the subject deformation path greatly complicate the deduc-
largely in terms of deformation paths. Such tion of a stress history.
stress and strain histories are important not The understanding of the nature of stress and
only for a proper understanding of sediment of stress paths that produce deformation of sedi-
behaviour but because, for example, the roles of ments has advanced remarkably over the past
pore fluids and the resulting mechanical be- decade, both from laboratory experimentation
haviour influence patterns of fluid flow and and from in situ measurements. Both approaches

The Geological Deformation of Sediments Edited by Alex Maltman Published in 1994 by Chapman & Hall ISBN 0 412 40590 3
168 Tectonic deformation

are valuable, but both have serious limitations and integrated far better than has so far been the
and present only partial solutions. case.
Mechanical experiments on sediments in the Data from both these techniques can be ap-
laboratory are relatively easy to perform, with plied to two structural regimes that fall near the
modern servo-controlled systems interfaced to ends of a spectrum of common deformation
computers. Such equipment is fairly reliable and paths most likely to affect sediments. These paths
the experiments are inexpensive in comparison are the uniaxial strain that accompanies deposi-
with in situ measurements. Laboratory experi- tion in a stable basin and the plane strain in
ments also permit uniform sediment to be subjec- accretionary prisms that form at some conver-
ted to varying conditions to better isolate func- gent plate margins. Moreover, these deformation
tional dependencies. In these experiments, close paths are those most familiar to the present
control of stress and strain can be maintained authors. They will also serve as a framework
over the range of stress applicable to purely upon which to base concepts that apply to the
mechanical deformation of sediments. Examples general mechanical behaviour of sediments. This
have arisen earlier in this book, especially in chapter reviews and discusses data from labora-
section 2.2.5. tory based experiments and in situ measurements
There are, however, serious problems with the from these two regimes of tectonic sediment de-
experimental approach to mechanical behaviour. formation. The two contrasting paths also pro-
Experimental strain rates are many orders of vide a framework within which to illustrate some
magnitude faster than natural strain rates, which of the more general aspects of sediment deforma-
precludes adequate measurement of creep or vis- tion.
cous (time-dependent) effects. Diagenetic pro-
cesses are not easily mimicked in mechanical
experiments, and these will be shown to have 6.2 STRESS PATHS DURING BURIAL
very pronounced effects on behaviour. A most
serious limitation is that only simple stress paths
can be applied during experimental deformation.
Only a very few machines permit plane strain or 6.2.1 General
true triaxial tests. Perhaps the simplest deformation path in geo-
In situ measurements are, in ,effect, the logy is the loading of sediments in basins and
monitoring of natural experiments, which would in other tectonically quiescent environments by
seem to be an approach preferable to laboratory subsequent deposition. It was introduced in
experimentation, but most of these measurements section 2.3.2. This path might be considered
are difficult, time-consuming and presently are non-tectonic but, because literally all sediments
very expensive. Techniques such as hydrofractur- are deformed in some way during burial before
ing, although rapidly improving, are still subject being tectonically deformed, the burial path
to problems of reliability and accuracy. Technical constitutes an essential part of the total de-
problems with in situ measurements can be se- formation. Moreover, even the most stable
vere, particularly in environments with higher basins are subjected to tectonic stresses of litho-
differential stresses. Where data from in situ tech- spheric origin (Zoback and Zoback 1989) and
niques are taken to be the results of natural the deformation from these stresses cannot be
experiments they must be analysed with caution, adequately understood except against a basis of
because boundary and initial conditions are often simple burial loading. Burial approximates a case
difficult to document. For example, rates of lat- of uniaxial (vertical) strain in that there is very
eral strain are often unknown. The effect of little lateral strain in these settings, although this
lithological variations among the data can be a assumption will be addressed later in some detail.
problem because, in most cases, variatioos in a Deviatoric stresses are usually low in basins, but
spatial field must serve as substitutes for tem- yet can lead to failure: jointing and normal fault-
poral changes along a stress path. At this time ing being good examples of extensional failure in
and probably far into the future, both approaches this setting.
will have to be pursued concurrently. Moreover, Despite this apparent simplicity, the state of
the two approaches will have to be co-ordinated stress in basins can vary widely as a function of
Stress paths in basins 169

such variables as stress history, lithology, thermal more familiar with an approximate empirical
conditions and pore-fluid pressure (Evans and rel~tionship between porosity (11) and depth (z) in
Engelder 1989). The stress history of a sediment basIns:
is crucial because the loading path (consolida- 1'/ = 1'/0 exp ( - Bz) (Athy 1930), (6.2)
tion) is dominantly an inelastic process, whereas
unloading is more nearly elastic, and the two where 110 is the porosity at the surface and B is a
have quite different stress-strain relationships. lithology dependent constant.
Unloading can result from erosion, cooling, in- Consolidation in natural sediment sections
creased pore-fluid pressure, or extensional tec- would be better related to O'~ than to z (Figure
tonic strain. Variability of each of these and other 6.1~, because O'~ is not related simply to depth,
factors leads to a very broad range of stress states OWIng to non-linear vertical variations in bulk
in basins and thus to uncertainty concerning the density and in pore fluid pressures. Moreover,
interpretation of stress measurements even in there are significant departures from idealized
that environment. Nevertheless, simple uniaxial porosity versus O'~ relationships in most basins,
consolidation in a basin with hydrostatic pore reflecting variations in lithology, cementation
pressures and a normal thermal gradient can be and stress ratios, as will be discussed later.
viewed a~ a fundamental or reference stress path, If the vertical effective stress is changed as a
from WhICh much could be learned about the sediment follows a strain path that is elastic, for
intrinsic mechanical behaviour of sediments. example by erosion or by pore-pressure increase
As outlined in section 2.1, the uniaxial consoli- the vertical strain (By) can be calculated fo;
dation of sediments entails the plastic yielding and an isotropic sediment using the uniaxial strain
volu~e reduction of the grain framework by the equation:
effectIve stress due to the overburden. This vol-
Ul~e reduction requires the expulsion of pore
flUId, the rate of which is controlled by the
where v is Poisson's ratio (section 1.2.4), E is the
permeability of the sediments in the system. Con-
Young's modulus, and LiO'~ or LiO'~ represents the
solidation is a compactive process generating
change in horizontal effective or vertical effective
def?rmation that is largely irreversible. Although
stress. Thus By depends on LiO'~ as well as on
sedIments undergoing consolidation are defined
several elastic parameters, which will be shown to
by inelastic stress-strain relationships, they are
vary widely under the range of geological condi-
capable of elastic deformation when effective
stresses are reduced and the stress path lies
Horizontal stresses during uniaxial consolida-
within the yield envelope. Only this component
tion are dependent upon lithology and O'~, but a
?f st~ain is recoverable upon unloading, although
large amount of uncertainty surrounds this rela-
It WIll be shown that such elastic strain is not
tionship. For the relatively low stresses and short
always linear with respect to stress. Elastic and
time periods characterizing the consolidation of
inelastic responses are related, in that elastic
soils under a construction load, geotechnical en-
para~eters, such as the Young's modulus (E)
(sectIOn 1.2.4), depend on the state of consolida- gineers have c?ncluded that the ratio O'~/O'~,
termed Ko (sectIon 2.2.6), is constant for a given
tion through reduction of porosity. A summary
sediment. Complications regarding the magni-
of the symbolic notation used in this chapter is
tude of Ko in nature were discussed in section
provided in Table 6.1.
2.2.6. Whether Ko remains constant over the
Vertical stresses responsible for both elastic
stress range applicable to mechanically con-
and inelastic strain are governed by the effective
solidated basinal sediments or whether it is con-
overburden stress because the system has a free
stant over geological time spans are also points
upper surface. The compactive vertical strain
induced during uniaxial consolidation of soils has of considerable debate.
During elastic deformation, changes of hori-
been related empirically to this stress by:
zontal stress are dependent on O'~ as well as on
e=LlogO"~ (e.g. Wood 1990), (6.1) Poisson's ratio (v). For isotropic materials:
~here e is void ratio and L is a constant depend-
Ing on factors such as lithology. Geologists are
170 Tectonic deformation

Table 6.1 Symbolic notation used in formulae

Maximum and minimum principal stresses
q Differential stress: 0"1 - 0" 3
Vertical stress
Horizontal stress (H subscript denotes maximum and h subscript denotes minimum horizontal
stress where horizontal stress is not uniform)
p' Mean effective stress
Effective consolidation stress: maximum vertical effective stress to which the sediment has been
Note: primed superscript denotes effective stress; un-
primed denotes total stress

Elastic or related parameters

c Strain, with subscripts as for stress
Young's moduli (;) in vertical and horizontal directions
Constrained Young's modulus
Shear modulus
Poisson's ratios, where the first subscript denotes the direction of the induced strain and the second
is the direction of the applied stress (and strain) e.g. Vhv = - ch/cv
[3 Bulk compressibility; [3=-, where V is volume and P is pressure
Compressibility of mineral grains
Constant, relating porosity and depth
Thermal compressibility;