Global Geomorphology

This page intentionally left blank

First published 1991 by Pearson Education Limited

Published 2013 by Routledge

2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
711 Third Avenue, New York, NY 10017, USA

Routledge is an imprint of the Taylor & Francis Group, an informa business

© Michael A. Summerfield 1991

All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or
by any electronic, mechanical, or other means, now known or hereafter invented, including
photocopYing and recording, or in any information storage or retrieval system, without permission
in writing from the publishers.

Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical
treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described herein. In
using such information or methods they should be mindful of their own safety and the safety of
others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of products
liability, negligence or otherwise, or from any use or operation of any methods, products,
instructions, or ideas contained in the material herein.

ISBN 13: 978-0-582-30156-6 (pbk)

British Library Cataloguing-in-Publication Data

Summerfield~ Michael A.
Global geomorphology
). Landforms
I. Title
551.41

Library of Congress Cataloging-in-Publication Data

A catalog entry for this title is available from the Library of Congress.

Set in 10111.5 Times

To my parents
This page intentionally left blank
Brief table of contents

Preface ' xv
AcknowLedgements xvii

Part I Introduction 1
1 Approaches to geomorphology 3

Part II Endogenic processes and landforms 31

2 Global morphology and tectonics 33
3 Landforms and tectonics of plate margins 57
4 Landforms and tectonics of plate interiors 85
5 Landforms associated with igneous activity 107

Part III Exogenic processes and landforms 127

6 Weathering and associated landforms 129
7 Slope processes and forms 163
8 Fluvial processes 191
9 Fluvial landforms 207
10 Aeolian processes and landforms 235
11 Glacial processes and landforms 261
12 Periglacial processes and landforms 293
13 Coastal processes and landforms 313
14 Climate, climatic change and landform 345
development

Part IV Endogenic-exogenic interactions 369

15 Rates of uplift and denudation 371
16 Tectonics and drainage development 405
17 Sea-level change 433
18 Long-term landscape develqpment 457

Part V Extraterrestrial landforms 481

19 Planetary geomorphology 483

Appendices 511
Index 525
This page intentionally left blank
 Detailed table of contents

Preface xv Part II Endogenic processes and landforms 31

Guide to the reader xvi Chapter 2 Global morphology and tectonics 33
Acknowledgemen,ts xvii 2.1 Global morphology 33
List ofreviewers xx 2.1.1 The geoid 33
Chapter heading plate captions xxi 2.1.2 Global hypsometry 34
2.1.3 Major morphological features 34
Part I Introduction 1 2.2 Earth structure 36
Chapter 1 Approaches to geomorphology 3 2.2.1 Seismic evidence 36
1.1 The science of landforms 3 2.2.2 Mantle, asthenosphere, crust and 36
1.2 The development of ideas 3 lithosphere
1.2.1 The age of Hutton and Lyell 4 2.2.3 Gravity anomalies 37
1.2.2 Developments in North America 5 2.2.4 Isostasy 38
1.2.3 The modem era 5 2.2.5 Crustal structure 39 .
1.2.4 Future directions 5 2.3 Development of ideas on global tectonics 40
1.3 Some key concepts 7 2.3.1 Continental drift 41
1.3.1 Endogenic and exogenic processes 7 2.3.2 Palaeomagnetic evidence 42
1.3.2 Geomorphic systems 8 2.3.3 Global seismicity 44
1.3.3 Magnitude and frequency 10 2.3.4 Sea-floor spreading 44
1.3.4 Equilibrium and evolution 11 2.4 Plate tectonics 46
1.3.5 Scale in geomorphology 12 2.4.1 Outline of the plate tectonics model 46
1.3.5.1 Temporal scale 12 . 2.4.2 Classification of plate boundaries 47
1.3.5.2 Spatial scale 12 2.4.3 Plate motion 48
1.3.5.3 Scale and causality 13 2.4.4 Mechanisms of plate movement 51
1.3.6 Explanation in geomorphology 16 Further reading 54
1.4 Methods of analysis 17 References 54
1.4.1 Direct observations 17
1.4.2 Space-time substitution 17 Chapter 3 Landforms and tectonics ofplate 57
1.4.3 Simulation 20 margins
1.5 Endogenic and exogenic factors 21 3.1 Convergent plate margins: general 57
1.5.1 Sources of energy 21 characteristics
1.5.1.1 Internal energy 21 3.1.1 Steady-state margins 57
1.5.1.2 Solar radiation 22 3.1.2 Collision margins 59
1.5.2 The hydrological cycle 22 3.2 Intra-oceanic island arcs 60
1.5.3 Climatic controls 23 3.2.1 General characteristics and formation 60
1.5.4 Human agency 24 3.2.2 . Morphological and structural elements 61
Further reading 27
References 28
x Detailed table of contents

3.3 Continental-margin orogens and continental- 62 5.2.2 Volcanic activity through time 110
margin island arcs 5.2.3 Products of volcanic activity 111
3.3.1 The Andes 63 5.2.3.1 Lava forms 111
3.3.1.1 General characteristics 63 5.2.3.2 Tephra 112
3.3.1.2 Evolution of the central Andes 64 5.2.3.3 Pyroclastic flow deposits 113
3.3.2 The Sunda arc 65 5.2.3.4 Volcanic products associated with water 113
3.4 Intercontinental collision orogens 66 5.3 Volcanoes 113
3.4.1 General sequence of development 66 5.3.1 Types of eruption 113
3.4.2 Pre-collision history and configuration of 70 5.3.2 Volcano morphology 115
converging continental margins 5.3.3 Basalt domes and shield volcanoes 115
3.4.3 The Himalayas 70 5.3.4 Strato-volcanoes 117
3.4.3.1 Morphological and structural elements 70 5.3.5 Calderas 118
3.4.3.2 Models of development 71 5.4 Other extrusive igneous landforms 119
3.4.4 The Tibetan Plateau 73 5.5 Landforms associated with igneous 120
3.4.4.1 Morphology and structure 73 intrusions
3.4.4.2 Models of development 73 Further reading 124
3.5 Oblique-slip margins 75 References 125
3.5.1 The San Andreas Fault System 76
3.5.2 The Southern Alps 76 Part III Exogenic processes and landforms 127
3.6 Displaced terranes 77
3.7 Mesoscale and microscale landforms 79 Chapter 6 Weathering and associated landforms 129
associated with faulting 6.1 The weathering system 129
Further reading 80 6.1.1 The nature of weathering 129
References 81 6.1.2 Water in rocks and soils 129
6.2 Chemical weathering 130
Chapter 4 Landforms and tectonics ofplate 85 6.2.1 Chemical characteristics of rock-forming 130
interiors minerals
4.1 Landscapes of plate interiors 85 6.2.2 Chemical reactions: thermodynamics and 131
4.2 Mechanisms of epeirogeny 86 kinetics
4.2.1 Thermal models 86 6.2.3 Chemical weathering processes 132
4.2.1.1 Lithospheric vulnerability 89 6.2.3.1 Solution 132
4.2.1.2 Effects of hot-spot migration 90 6.2.3.2 Hydrolysis 133
4.2.2 Phase changes 90 6.2.3.3 Carbonation 133
4.2.3 Mechanical models 90 6.2.3.4 Oxidation and reduction 135
4.3 Continental rifts 92 6.2.3.5 Cation exchange 136
4.3.1 Rift structure and location 92 6.2.3.6 Organic processes 136
4.3.2 Rift formation 93 6.2.4 Products of chemical weathering 136
4.3.3 Rifting and uplift in Africa 94 6.2.4.1 The weathering mantle 136
4.4 Continental basins 95 6.2.4.2 Mineral stability and the formation of 137
4.4.1 Basin morphology 95 secondary minerals
4.4.2 Mechanisms of subsidence 95 6.2.5 Factors influencing chemical weathering 141
4.5 Passive continental margins 95 6.3 Physical weathering 144
4.5.1 Active rifting 96 6.3.1 Volumetric changes of the rock mass 144
4.5.2 Passive rifting 97 6.3.2 Volumetric changes within rock voids 146
4.5.3 Passive margins: a key research focus 101 and fissures
4.6 The break-up of supercontinents 102 6.3.2.1 Frost weathering 146
Further reading 102 6.3.~.2 Salt weathering 147
References 103 6.4 Lithology and weathering forms 148
6.4.1 Karst weathering forms 148
Chapter 5 Landforms associated with igneous 107 6.4.1.1 Minor forms 149
activity 6.4.1.2 Major forms 150
5.1 Extrusive and intrusive igneous activity 107 6.4.2 Other weathering forms 151
5.2 Volcanism 108 6.5 Duricrusts 153
5.2.1 Distribution of volcanic activity 109 6.5.1 Models of duricrust formation 153
 Detailed table of contents xi

6.5.2 Environmental controls 154 Further reading 203

Further reading 157 References 203
References 158
Chapter 9 Fluvial landforms 207
Chapter 7 Slope processes and forms 163 9.1 The fluvial system 207
7.1 Properties and behaviour of slope materials 163 9.2 The drainage basin 207
7.1.1 Factors determining the strength of slope 163 9.2.1 Channel network characteristics 208
materials 9.2.2 Areal and relief characteristics 209
7.1.2 Rock properties 165 9.3 River channels 210
7.1.3 Soil properties 166 9.3.1 Alluvial channels: plan form 210
7.2 Mass movement 167 9.3.1.1 Meandering channels 212
7.2.1 Slope stability 167 9.3.1.2 Braided channels 215
7.2.2 Mass movement processes 168 9.3.1.3 Channel pattern change 215
7.2.2.1 Creep 168 9.3.2 Alluvial channels: hydraulic geometry 215
7.2.2.2 Flow 170 9.3.2.1 At-a-station adjustments 215
7.2.2.3 Slide 172 9.3.2.2 Downstream changes 216
7.2.2.4 Heave 173 9.3.3 Alluvial channels: longitudinal form 216
7.2.2.5 Fall 173 9.3.4 The alluvial channel system 218
7.2.2.6 Subsidence 174 9.3.5 Bedrock channels 219
7.3 Gravity tectonics 174 9.3.5.1 Longitudinal profile 219
7.4 Water erosion and solute transport on slopes 175 9.3.5.2 Incised meanders 219
7.4.1 Hillslope hydrology 175 9.4 Fluvial depositional landforms 220
7.4.2 Rainsplash erosion 175 9.4.1 Floodplains 220
7.4.3 Slope wash 176 9.4.2 Alluvial fans 222
7.4.4 Soil erosion 177 9.4.3 River terraces 224
7.4.5 Solute transport 178 9.5 Fluvial activity through time 225
7.5 Rates of slope processes 178 9.5.1 Geomorphic thresholds, complex response 225
7.6 The slope system 179 and episodic erosion
7.6.1 Slope form 179 9.5.2 Magnitude and frequency of fluvial activity 227
7.6.1.1 Rock slopes 182 9.6 Fluvial systems in limestone terrains 228
7.6.1.2 Soil-mantled slopes 183 9.6.1 Surface drainage 228
7.6.2 Slope evolution 184 9.6.2 Subsurface drainage 229
Further reading 186 Further reading 231
References 188 References 232

Chapter 8 Fluvial processes 191 Chapter 10 Aeolian processes and landforms 235
8.1 Drainage basin hydrology 191 10.1 Aeolian activity 235
8.1.1 Channel discharge 192 10.1.1 Global distribution of aeolian landforms 235
8.1.1.1 Spatial variatioris 192 10.1.2 Wind characteristics 236
8.1.1.2 Temporal variations 192 10.1.3 Effect of surface characteristics 237
8.1.2 Runoff generation 194 10.1.4 Sediment entrainment and transport 237
8.1.3 Channel initiation 195 10.2 Aeolian erosion 239
8.2 Open channel flow 195 10.2.1 Deflation and abrasion 239
8.2.1 Resistance to flow 196 10.2.2 Erosional landforms 239
8.2.2 Laminar and turbulent flow 196 10.2.2.1 Small-scale forms 240
8.2.3 Flow regimes 197 10.2.2.2 Intermediate-scale forms 240
8.2.4 Velocity of flow 197 10.2.2.3 Large-scale forms 242
8.3 Fluvial erosion and sediment entrainment 198 10.3 Depositional landforms 242
.8.3.1 Erosion of bedrock channels 198 10.3.1 Basic depositional forms 242
8.3.2 Sediment entrainment 199 10.3.1.1 Ripples 243
8.4 Fluvial transport and deposition 200 10.3.1.2 Dunes 244
8.4.1 Modes of fluvial transport 200 10.3.2 Ciassification of dune morphology 246
8.4.2 Sediment transport 201 10.3.3 Classification of free dunes 246
8.4.3 Fluvial deposition 202 10.3.4 Global occurrence of free dune types 248
xii Detailed table of contents

10.3.5 Development of free dunes 248 12.2.1.3 Mass displacement 298

10.3.5.1 Transverse-type dunes 251 12.2.1.4 Frost cracking 298
10.3.5.2 Linear dunes 252 12.2.2 Chemical weathering 299
10.3.5.3 Other types of free dune 254 12.2.3 Mass movement 299
10.3.6 Development of impeded dunes 254 12.2.4 Nivation 299
10.3.7 Fine-grained deposits 256 12.2.5 Fluvial activity 300
Further reading 257 12.2.6 Aeolian activity 301
References 258 12.3 Periglacial landforms 301
12.3.1 Patterned ground 301
Chapter 11 Glacial processes and landforms 261 12.3.2 Ground-ice phenomena 302
11.1 Glacier characteristics and dynamics 261 12.3.2.1 Ice wedges 302
11.1.1 Glacier distribution and classification 261 12.3.2.2 Pingos 303
11.1.2 Characteristics of glacier ice 263 12.3.2.3 Palsas 304
11.1.3 The glacier mass balance 263 12.3.2.4 Thermokarst 305
11.1.4 Glacier motion 264 12.3.3 Depositional forms related to mass 307
11.1.4.1 Mechanisms of ice movement 264 movement
11.1.4.2 Glacier flow 265 12.3.4 Asymmetric valleys 308
11.1.4.3 Short-term glacier fluctuations 267 12.3.5 Cryoplanation terraces and 308
11.1.5 Glacier morphology . 268 cryopediments
11.1.5.1 Large-scale forms 268 Further reading 309
11.1.5.2 Small-scale surface features 269 References 310
11.2 Glacial erosion 269
11.2.1 Mechanisms of erosion 269
11.2.2 Debris entrainment and transport 271 Chapter 13 Coastal processes and landforms 313
11.2.3 Erosional landforms 272 13.1 The coastal environment 313
11.2.3.1 Forms associated with unconfined 273 13.2 Waves, tides and currents 315
ice flow 13.2.1 Waves 315
11.2.3.2 Forms associated with channelled 273 13.2.1.1 Wind-generated waves 316
ice flow 13.2.1.2 Storm surges 320
11.2.3.3 Forms associated with periglacial 274 13.2.1.3 Seismic sea waves 320
action 13.2.2 Tides 320
11.3 Glacial deposition 275 13.2.3 Currents 321
11.3.1 Mechanisms of deposition 276 13.3 Coastal processes 323
11.3.2 Depositional forms 277 13.3.1 Destructional processes 323
11.3.2.1 Forms parallel to the direction of 277 13.3.1.1 Shoreline weathering 323
ice flow 13.3.1.2 Coastal erosion 323
11.3.2.2 Forms transverse to the direction of 280 13.3.2 Constructional processes 324
ice flow 13.3.2.1 Sediment movement and deposition 324
11.3.2.3 Forms lacking consistent alignment 280 13.3.2.2 Organic activity 325
11.4 Fluvioglacial erosion and deposition 281 13.4 Coastal landforms 325
11.4.1 Glacial meltwater 281 13.4.1 Destructional forms 325
11.4.2 Fluvioglacial denudation 282 13.4.1.1 Cliffs 325
11.4.3 Fluvioglacial deposition 284 13.4.1.2 Shore platforms 326
Further reading 287 13.4.2 Constructional forms 327
References 288 13.4.2.1 Beaches 327
13.4.2.2 Barrier islands 330
Chapter 12 Periglacial processes and landforms 293 13.4.2.3 Spits, baymouth bars and cuspate 331
12.1 The periglacial environment 293 forelands
12.1.1 Characteristics of permafrost 294 13.4.2.4 Tidal flats 331
12.1.2 Distribution of permafrost 294 13.4.2.5 Deltas 331
12.2 Periglacial processes 296 13.4.2.6 Coastal dunes 336
12.2.1 Frost action 296 13.4.2.7 Reefs 340
12.2.1.1 Frost weathering 297 Further reading 340
12.2.1.2 Frost heaving and thrusting 297 References 341
 Detailed table of contents xiii

Chapter 14 Climate, climatic change and 345Chapter 16 Tectonics and drainage development 405
landform development 16.1 Active and passive tectonic controls on 405
14.1 Climate and landform development. 345 drainage
14.2 The significance of climatic change 348 16.2 Passive tectonic controls 405
14.3 The record of climatic change 349 16.2.1 Drainage patterns 405
14.3.1 The Cretaceous to Neogene record 349 16.2.2 Fracture and joint control 409
14.3.2 The Quaternary record 350 16.2.3 River capture 410
14.4 Effects arising predominantly from 353 16.2.4 Transverse drainage 411
temperature changes 16.3 Tectonics and drainage adjustment 412
14.4.1 Landscapes of deglaciation 353 16.3.1 Effects of warping 412
14.4.1.1 Relict landscapes of glacial erosion 353 16.3.2 Effects of faulting 413
14.4.1.2 Relict landscapes of glacial deposition 355 16.3.3 Drainage evolution in rift valleys 414
14.4.1.3 Glacio-isostasy 355 16.3.4 History of the Colorado River 414
14.4.2 Relict periglacial landscapes 355 16.4 Active tectonics and channel adjustment 416
14.5 Effects arising predominantly from 359 16.4.1 Field examples 416
precipitation changes 16.4.2 Experimental studies 417
14.5.1 Change in fluvial sys~ems 359 16.5 Global tectonics ~d continental drainage 418
14.5.1.1 Underfit streams 361 16.6 Drainage development in orogens 420
14.5.1.2 Long-term river metamorphosis 361 16.6.1 Drainage of the Zagros Mountains 420
14.5.2 Aeolian systems 364 16.6.2 Drainage of the Himalayas 422
Further reading 365 16.7 Drainage development associated with 423
References 366 passive margins
16.7.1 Drainage history of the African continent 424
Part IV Endogenic-exogenic interactions 369 16.7.2 A model of passive margin drainage 426
development
Chapter 15 Rates of uplift and denudation 371 16.7.3 Variation in passive margin drainage 426
15.1 Tempo of geomorphic change 371 evolution
15.2 Rates of uplift 371 Further reading 427
15.2.1 Methods of measurement and estimation 372 References 430
15.2.2 Spatial and temporal variations 375
15.2.2.1 Orogenic uplift 375 Chapter 17 Sea-level change 433
15.2.2.2 Epeirogenic uplift 378 17.1 Global and regional sea-level change 433
15.3 Present fluvial denudation rates 379 17.2 Evidence for sea-level change 434
15.3.1 Methods of measurement and estimation 379 17.2.1 Quaternary sea levels 434
15.3.1.1 Solid load 380 17.2.2 Pre-Quaternary sea levels 434
15.3.1.2 Solute load 382 17.2.2.1 Continental evidence 434
15.3.1.3 Estimation of volumetric changes 382 17.2.2.2 Seismic stratigraphy 435
15.-3.2 Rates of mechanical and chemical 383 17.3 The record of Quaternary sea-level 437
denudation change
15.3.3 Relative importance of mechanical and 386 17.4 The record of pre-Quaternary 438
chemical denudation sea-level change
15.4 Long-term fluvial denudation rates 388 17.5 Mechanisms of global 439
15.4.1 Methods of estimation 388 sea-level change
15.4.1.1 Estimates from sediment volumes 388 17.5.1 Changes in ocean water volume 439
15.4.1.2 Erosion of dated surfaces 389 17.5.2 Changes in ocean basin volume 440
15.4.1.3 Fission track and radiometric techniques 389 17.5.2.1 Variations in the volume of mid-oceanic 440
15.4.2 Variations in rates 390 ridges
15.5 Factors controlling fluvial denudation rates 392 17.5.2.2 Sediment accumulation on the 441
15.6 Rates of aeolian denudation 395 ocean floor
15.7 Rates of glacial denudation 397 17.5.2.3 Effects of orogeny 442
15.8 .Comparison of rates of uplift and 398 17.5.2.4 Desiccation of small ocean basins 442
denudation 17.5.3 Geoidal effects 442
Further reading 400 17.5.4 Explaining the long-term record of 443
References 401 global sea-level change
xiv Detailed table of contents

Sea-level change and landscape

17.6 445 Part V Extraterrestrial landforms 481
development
17.6.1. Drainage basin response 445 Chapter 19 Planetary geomorphology 483
17.6.2 Coastal effects 449 19~'1 Approaches to planetary geomorphology 483
17.6.3 Oceanic islands 450 19.2 The Moon 484
Further reading 453 19.2.1 Impact cratering 485
References 454 19.2.2 Volcanism and tectonics 485
19.2.3 Surface materials and processes 486
Chapter 18 Long-term landscape development 457 19.2.4 History of landscape development 487
18.1 Models of landscape evolution 457 19.3 Mars 488
18.1.1 The Davisian cycle of erosion: 457 19.3.1 Impact craters 489
peneplanation 19.3.2 Volcanic and tectonic features 489
18.1.2 The Penck model: uplift and 460 19.3.3 Weathering 491
denudation related 19.3.4 Slopes and mass movement 492
18.1.3 The King model: pediplanation 461 19.3.5 Aeolian processes and landforms 492
18.1.4 The Btidel model: etchplanation 462 19.3.6 Periglacial processes and landforms 494
18.1.5 Classic models of landscape evolution: 463 19.3.7 Polar terrains 494
summary and assessment 19.3.8 Channels 495
18.2 Landscape stability and change 464 19.3.9 History of landscape development 497
18.2.1 The Hack dynamic equilibrium model 464 19.4 Mercury 497
18.2.2 The dynamic metastable equilibrium 464 19.5 Venus 498
model 19.6 The moons of the outer planets 500
18.2.3 Landscape sensitivity 465 19.6.1 The Galilean moons of Jupiter 500
18.2.4 Time-dependent and time-independent 466 19.6.2 The moons of Saturn 501
landforms 19.6.3 The moons of Uranus 503
18.3 Denudation chronology 468 19.6.4 The moons of Neptune 504
18.3.1 Principles of denudation chronology 468 19.7 Comparative planetary geomorphology 505
18.3.2 Continental denudation and the marine 469 19.8 Global geomorphology 506
stratigraphic record Further reading 507
18.3.3 Regional case studies 469 References 508
18.3.3.1 South-east England 469
18.3.3.2 The Appalachian region 471 Appendix A Units of measurement 511
18.3.3.3 Southern Africa 472 Appendix B Dating techniques 513
18.3.3.4 The East Australian Highlands 473 Relative dating 513
18.4 Tectonics, climate and landscape 475 Geochronometric dating 513
development Radiometric dating 513
18.4.1. Continental drift and changing 475 Fission track dating 514
morphoclimatic regimes Palaeomagnetic dating 514
18.4.2 Orogenesis and changing morphoclimatic 475 Varve dating 515
regimes Dendrochronology 515
18.4.3 Plate tectonics, climate and weathering 475 Amino acid racemization dating 515
18.4.4 Tectonics, climate and landscape 477 Appendix/C Geological time scale 517
development: a summary Appendix D Particle size 519
Further reading 477 Appendix E Behaviour of materials 521
References 478 Appendix F Geological structures 523
Faults 523
Folding 524

Index 525
 Preface

If it is to present its subject matter coherently a textbook to incorporate the results of newly applied dating techni-
must have a point of view. The perspective of this book is ques and of research on the offshore sedimentary record.
that an adequate appreciation of landform genesis must en- The examination of surface processes and the landforms
compass a knowledge of the large-scale framework of land- they create still accounts for the bulk of the text, but the
scapes as well as an understanding of the smaller-scale title of the book is intended to convey the global perspec-
processes which create individual landforms. An emphasis tive that I wish to present.
on small-scale, surface processes and their associated land- Geomorphology has grown in scope and depth over recent
forms has been pervasive in geomorphology since the 1960s, years to the extent that even in a fairly lengthy text I have
to the point where the larger-scale aspects of landform been forced to be selective in the topics discussed. Perhaps
genesis, and in particular the role of internal mechanisms in the major omission is applied geomorphology, but to have
influencing the development of major morphological done this topic justice would have extended the length of
features, have come to be regarded as almost incidental to the book by a further 20 or 30 per cent. With its emphasis
the main thrust of research in the subject. To borrow an. on the way the large-scale components of landscapes' develop
analogy from Richard Lewontin, commenting on the rela- over the long term this text is concerned essentially with
tionship between population genetics and evolutionary bio- the development of naturally-created landforms, an under-
logy, 'this has led to a kind of auto mechanics concerned standing of which is, of course, vital if we are correctly to
with carburettor settings and tyre pressures, but not with assess the impact of human activity. A second significant
how the car was manufactured or where it is is going'. omission is submarine geomorphology; this again is largely
There are growing signs that this situation is changing for reasons of space since the history of the ocean basins
and this book attempts to redress the balance by giving due and the operation of submarine processes raise some quite
weight to problems of long-term, large-scale landscape distinct issues which I felt could not be adequately tackled
development. In particular I have attempted to integrate without a fairly lengthy treatment.
ideas on global tectonics fully into landscape analysis and
 Guide to the reader

My aim is that this book should be used as a resource have decided, therefore, to rely on verbal description and
which provides both basic background information and discussion, but I have included a mathematical represen-
guidance towards the more advanced study of particular tation of selected concepts and processes in boxes through-
topics. In order to make the book more readable I have ,out the text. Furthermore, the text An Introduction to Quan-
omitted references from the text, but each chapter con- titative Geomorphology: An Exercise Manual by Larry
cludes with a detailed guide to further reading and a list of Mayer (Prentice-Hall, Englewood Cliffs and London, 1990)
references. These literature guides will be useful to those provides a mathematical treatment of a number of the
pursuing topics for essay or project preparation. The refer- topics covered in this book along with a range of relevant
ences have been selected on the basis of their importance, field and laboratory exercises.
readability and accessibility, although the nature of the Part I of the book (Chapter 1) provides an introduction to
coverage in the literature of some subjects means that I some of the major concepts applied in the analysis of land-
have not always been successful in meeting these criteria. scapes; other important concepts are introduced at appro-
I have assumed that readers have a basic knowledge of priate points in subsequent chapters. Part II, comprising
the physical environment including the main types of climate Chapters 2-5, examines the effects of internal processes on
and vegetation, and the major rock types and rock-forming the form of the landscape, while Part III- (Chapters 6-14)
minerals. An introductory course in physical geography or looks at the wide range of surface processes and their
physical geology would provide the appropriate background. associated landforms. The chapters in Part IV (15-18)
Nevertheless, I have attempted to define all but the most consider the ways in which internal and external geomorphic
elementary technical terms and these are printed in bold processes interact and the book concludes in Part V with a
type in the text. Where possible, terms have been defined survey of planetary geomorphology. Extensive cross-
where they first occur in the text, but in all cases a page referencing means that to some extent the chapters do not
number in bold type in the index indicates where the defi- have to be read in the order they are presented. In par-
nition of a term can be found. The mathematical compe- ticular, Chapters 6-14 can just as easily precede Chapters
tence of students taking courses in geomorphology varies 2-5 as follow them. Chapters 15-19, however, are to a large
enormously and this provides a major dilemma for the extent founded on material in earlier chapters. Readers with
textbook author. The view I have gained from my own a background in physical geology might want to skip much
experience and discussions with numerous colleagues both of Chapter 2 as it provides an introduction to global
in the UK and the USA is that while a fairly rigorous tectonics. For those without a geological training some
mathematical treatment might be desirable, such an approach useful background material is provided in the appendices.
would not be appropriate for the majority of students. These also contain information on the units of measurement
Moreover, a strong mathematical emphasis can in some used in the book and a brief discussion of dating techniques
cases create the impression that certain geomorphic pro- relevant to geomorphology.
cesses are more fully understood than they actually are. I
 Acknowledgements

Preparing a textbook is a major undertaking which can only I am especially grateful to those colleagues who put con-
be accomplished through the efforts of many individuals siderable time and effort into providing comments on the
besides the author. The diagrams are a crucial component of draft manuscript, or parts thereof. These individuals are
the book and I have been fortunate in having the services of listed separately. The responsibility for the final text is, of
Ray Harris and Anona Lyons of the cartographic staff at the course, solely my own.
Department of Geography of the University of Edinburgh. Finally, I would like to thank Sue Smith for her support
Initial assistance with the typing of some of the draft chapters and encouragement.
was provided by Mrs B. L. Summerfield and Mrs W. Rust and
most of the tables were typed by Miss S. Smith. A number The author and publishers are grateful to the following for
of individuals and organizations have been kind enough to permission to reproduce copyright figures and tables:
provide me with photographs and other imagery in areas
where my own resources have proved inadequate; these are Academic Press Inc. (London) Ltd. and the respective authors
acknowledged in the relevant figure captions. I would, for figs. 4.13 (Burke & Whiteman, 1973),6.28 (McFarlane,
however, particularly like to thank Nick Short who made 1983), 6.29 (Goudie, 1983), 13.10 (Hayes, 1976), 15.6
available to me several Landsat images originally published (Gansser, 1983); American Association of Petroleum Geo-
in his co-edited book Geomorphology from Space. I would logists for figs. 17.1 & 17.2 (Vail et al., 1977), 17.8 & 18.17
also like to thank those individuals and organizations who (Kominz, 1984) and table 13.4 (Wright & Coleman, 1973);
granted permission for copyright material to be reproduced American Association for the Advancement of Science and
or used as the basis for diagrams or tables. Specific the respective authors for figs. 16.6 (Burnett & Schumm,
acknowledgements are listed below. 1983), 17.4 (Hag et al., 1987) © 1983 & 1987 by the AAAS;
Some of the material presented in this book relates American Journal of Science and the respective authors for
directly, or indirectly, to my own research experience which figs. 3.22 (Le Fort, 1975), 15.15 (Ruxton & McDougall, 1967),
has included fieldwork in many parts of the world, and I am 15.18 (Ahnert, 1970), 18.17 (Berneretal., 1983); the editor,
grateful to those organizations that have provided financial American Scientist for fig. 19.21 (Head et al. , 1981); American
support for this research. These include the Natural Society of Civil Engineers for fig. 8.10 (Simons & Richardson,
Environment Research Council, the Royal Society, Texaco 1963) and table 9.3 (Benedict et al., 1971); Annual Reviews
Inc., British Aerospace p.l.c., the Carnegie Trust for the Inc. for fig. 9.8 (Schumm, 1985) © 1985 by Annual Reviews
Universities of Scotland, the University of Oxford and the Inc.; Edward Arnold (Publishers) Ltd. for figs. 5.7 (Clapper-
University of Edinburgh. I also owe a considerable debt to ton, 1977), 10.4, 10.23, 10.24, 10.25, 10.27 & 10.28 (Warren,
those numerous colleagues who have influenced my approach 1979),10.32 (Pye, 1984), 15.24 (Goudie, 1983), 19.5 (Baker,
to geomorphology, both through their published work and 1981); Edward Arnold (Publishers) Ltd. and the respective
informal discussions. Naturally this book draws extensively authors and figs. 11.18, 11.19, 11.20, 11.22 & 14.3 (Sugden &
on the work of a great number of earth scientists, only a small John, 1976), 12.3, 12.5, 12.7, 12.11 & 14.6 (Washburn, 1979)
proportion of whom are specifically mentioned in the main and tables 11.4 (Sugden & John, 1976), 12.3, 12.6 & 14.3
body of the text. The guides to further reading will, however, (Washburn, 1979), 12.2 (Embleton & King, 1975); A. A.
enable readers to follow-up work by individual researchers. Balkema Publishers (Netherlands) for fig. 17.13 (Dingle et al. ,
1983); the author, P. Bishop for fig. 15.17 (Bishop, 1985); burgh, 1972), 3.31 (Coney et al., 1980), 4.14 (Watts, 1982),
Basil Blackwell Ltd, for fig. 6.23 (Jennings, 1985); Blackwell 10.22 (Wasson & Hyde, 1983), 14.6 (Hovanetal., 1989), 19.22
Scientific Publications Ltd. for figs. 8.14 (Miller et al., 1977), (McGill, 1982) Copyright © 1972, 1980, 1982, 1983 & 1989
13.24 (Elliott, 1978); California Division of Mines & Geology Macmillan Magazines Ltd.; Macmillan Publishers Ltd. for fig.
for figs. 3.27 (Crowell, 1975), 16.7 (Jenkins, 1964), 16.11 18.4 (Thomas, 1974); Mineralogical Society of America for
(Wallace, 1975); Cambridge University Press for fig. B7.3 fig. 6.15 (Loughnan & Bayliss, 1961) Copyright © by Miner-
(Carson & Kirkby, 1972) and figs. 2.16 & 3.23 (Smith, Hurley alogical Society of America; Methuen & Co. for figs. 6.1
& Briden, 1981),7.4,7.5,7.27,7.28 & 7.29 (Carson & Kirkby, (Carson, 1969), 9.16 (Richards, 1982 & Dury, 1969), 9.17
1972); Cambridge University Press and the respective authors (Richards, 1982), 10.2 (Bagnold, 1954), 14.9 (Chorley et al.,
for figs. 13.25 (Stoddart, 1969), 14.5 (Harland et al., 1982); 1984), 14.18 (Schumm, 1969), 16.25 (Heinzelin, 1964), 18.1
Canadian Society of Petroleum Geologists for table 14.5 (Thomes & Brunsden, 1977); National Research Council of
(Baker, 1978); Carnegie Institution of Washington for fig. Canada for fig. 12.3 (Mackay, 1971); the author, A. Nur for
5.16 (Williams, 1942); Chapman & Hall Ltd. for fig. 16.21 fig. 3.32 (Nur & Ben-Avraham, 1983); the author, H. Ohmori
(Holmes, 1978); the authors, D. Chapman & H. Pollock for for figs. 15.19 &-15.20 (Ohmori, 1983); The Open University
fig. 2.8 (Chapman & Pollock, 1977); the author, J. Chappell for fig. 2.15 (Oxburgh, 1971) Copyright © 1971 The Open
for fig. 15.5 (Chappell, 1974); the editor, Clay Minerals for fig. University Press; the author, S. Ouchi for table 16.3 (Ouchi,
6.15 (Bayliss & Loughnan, 1964); Columbia University Press 1985); Oxford University Press for figs. E2 (Statham, 1977),
(New York) for fig. 18.13 (Johnson, 1931); Croom Helm Ltd. 15.26 (Selby, 1982) and tables 7.1 & 7:3 (Selby, 1982), 14.1
for figs. 6.10& 7.1 (Crozier, 1986}, 15.25 (Yoshikawa, 1985), (Douglas, 1978) Copyright © 1977, 1978 & 1982 Oxford
17.3 (Pillans, 1987); the author, J. Demek for fig. 12.17 (De- University Press; Pergamon Press pIc for figs. 2.12 (Heirtzler
mek, 1969); the author, D. Drewry for fig. 11.8 (Drewry, etal., 1966), 11.23 (Baker & Bunker, 1985) Copyright© 1966
1983); DSIR Publishing (New Zealand Depart. of Scientific & & 1985 Pergamon Press pIc; Pergamon Press Australia for
Industrial Research) for fig. 3.26 (Kingma, 1958) the editor, figs. 10.2 & 10.10 (Mabbutt, 1977); the author, W. Pitman for
Ecologae Geologicae Helvetiae for fig. 3.19 (Collet, 1935); figs. 17.6 & 17.7 (Pitman, 1978); Prentice-Hall (Englewood
Elsevier Science Publishers (Physical Sciences & Engineering Cliffs) for figs. 5.16 (MacDonald, 1972), 16.2 & 16.4 (Bloom,
Div.) for fig. 10.13 (Wilson, 1972); Elsevier Science Publishers 1978) and table 5.2 (MacDonald, 1972); the editor, Quater-
(Physical Sciences & Engineering Div.) and the respective nary Research for fig. 17.9 (Clark et al. , 1978); Routledge Ltd.
authors for figs. 9.20 (Allen, 1964), 13.16 (Ingle, 1966), 14.21 for figs. 7.18 & 17.20 (Parsons, 1988); The Royal Society and
(Lancaster, 1981), 18.17 (Barron, 1983) and table 6.4 (Lough- the author, C. Beaumont for fig. 4.17 (Beaumont et al. , 1982);
nan, 1969); Environment Canada and the author, M. Church the author, S. A. Schumm for fig. 9.13 (Schumm & Khan,
for fig. 12.6 (Church, 1974); W. H. Freeman & Co. for fig. 1.12 1972); Scientific American Inc. for fig. 19.28 (Soderblom &
(Leopold, Wolman & Miller, 1964) Copyright © 1964 W. H. Johnson, 1982) Copyright © 1982 by Scientific American Inc.
Freeman & Co; Gebriider Borntraeger for figs. 7.19 (Dalrym- All rights reserved; Society of .Economic Paleontologists &
ple et al., 1968), 10.29 (Tsoar, 1984), 18.8 (Brunsden, 1980) Mineralogists for figs. 3.26 (Crowell, 1974),8.11 (Middleton
and table 7.2 (S,elby, 1980); the editor, Geographica Helvetica & Southard, 1984),9.9 (Schumm, 1981) and table 13.5 (Mor-
for fig. 16.6 (Scheidegger, 1979); the editor, Geological Jour- gan, 1970); Seismological Society of America for fig. 2.10
nal for fig. 9.20 (Allen, 1970); Geological Society and the (Barazangi & Dorman, 1969); Springer-Verlag and the au-
respective authors for figs. 3.11 (Gansser, 1973), 16.12 & thor, G. Scott for fig. 17.20 (Scott & Rotondo, 1983); Minister
16.13 (Frostick & Reid, 1987); the editor and The Geologists of Supply and Services Canada for table 11.7 (Prest, 1968);
Association for fig. 3.20 (Oxburgh, 1974); Gulf Coast Asso- Syracuse University (Dept. of Geography) for fig. 16.19
ciation of Geological Societies for fig. 13.24 (Frazier, 1967); (Oberlander, 1965); Taylor & Francis Ltd. (Wykeham Pub-
the author, W. Hamilton for fig. 3.14 (Hamilton, 1977); Insti- lications) and the author, J. Guest for table 19.2 (Guest &
Itute of British, Geographers for fig. B17.1 (Chappell, 1987) Greeley, 1977); Transportation Research Board for figs. 7.6,
and fig. 7.24 (Kirkby, 1971); International Glaciological Soci- 7.9,7.10,7.11 and table 7.5 (Varnes, 1978); The Unesco Press
ety and the author, J. Nye for fig. 11.6 (Nye, 1952); the author, for fig. 8.3 (Unesco, 1978); United States Geological Survey
D. James for fig. 3.13 (James, 1971); Longman Group UK for figs. 3.34 (Wesson et al., 1975),5.19 (Hunt, 1953),9.22
Ltd. for figs. 1.15 (Tricart & Cailleux, 1972), 5.17 (Sparks, (Lustig, 1965),9.26 (Leopold & Miller, 1954), 10.17 (McKee,
1971), 6.13 (OIlier, 1984), 9.25 (Sparks, 1972), 12.6 & 12.7 1979), 19.5 (Scott & Carr, 1978); The University of Chicago
(French, 1976), 13.6, 13.9, 13.14, 13.19 & 13.25 (Davies, Press for figs. 1.11 (Brunsden & Kesel, 1973), 2.1 (Marner,
1980),14.14 (Lowe & Walker, 1984), 16.1 (Morisawa, 1985), 1976),13.1 & 13.2 (Inman & Nordstrom, 1971), 14.19 (Baker
16.9 (Ollier, 1981) and tables 15.3 (Morgan, 1986), 16.1 & & Penteado-Orellana, 1977), 18.15 (Bishop et al., 1985) and
16.2 (Morisawa, 1985); Longman Group UK Ltd. for table tables 13.1,13.2 & 13.3 (Inman & Nordstrom, 1971) © 1971,
11.8 (Price, 1973) published by Oliver & Boyd; Macmillan 1973, 1976, 1977 & 1985 by the University of Chicago. All
Magazines Ltd. and the respective authors for figs. 3.20 (Ox- rights reserved; University of Natal Press (Pietermaritzburg)
for fig. 17.16 (King,'1982); University of Texas (Texas Bureau has been made to trace the owners of copyright material, in a
of Economic Geology) for fig. 13.21 (Fisher et at., 1969); few cases this has proved impossible and we take this oppor-
University of Uppsala (Geological Institute) for fig. 8.13 tunity to offer our apologies to any copyright holders whose
(Hjulstr6m, 1935); Unwin Hyman Ltd. for fig. B11.3 (Boul- rights we may have unwittingly infringed.
ton, 1974) and figs. 7.17 (Young & Saunders, 1986), 8.9 &
9.11 (Allen, 1970), 15.23 (Middleton et at., 1986), 16.19 & We are grateful to the following for permission to reproduce
16.20 (Oberlander, 1985), 18.6 (Schumm, 1975), 18.14 (Sum- copyright photographs:
merfield, 1985), 19.13 (Greeley, 1985) Copyright © 1970,
1974, 1975, 1976, 1985 & 1986 Unwin Hyman Ltd.; Van Aerofilms Ltd. for figs. 3.18 & chapter heading photograph
Nostrand Reinhold for fig. 12.12 (Muller, 1968) Copyright © for Chapter 16; J. Chappell for chapter heading photograph
1968 Van Nostrand Reinhold; the author, P. Vogt for fig. 4.4 for Chapter 17; K. G. Cox for cover photograph. Geodetic
(Vogt, 1981); John Wiley & Sons Ltd. for figs. 6.2,6.4 & 6.16 Institute, Copenhagen for chapter heading photograph for
(Curtis, 1976), 7.14 (Atkinson, 1978), 7.21 (Abrahams & Chapter 11 & fig. 11.1; Geological Survey of Canada for fig.
Parsons, 1987), 7.23 (Carson, 1976), 10.13 (Wasson & Hyde, 12.13; Goddard Space Flight Center for figs. 1.1OA & 3.28; K.
1983), 15.17 (Brown et at., 1980), 15.9 (Walling & Webb, Mulligan for fig. 10.18; D. Munro for figs. 3.12 & 6.25; G. M.
1983),15.13 (Clayton & Megahan, 1986), 15.14 (Foster et at., Robinson for fig. 11.15; © Royal Geographical SocietylM. J.
1985), 15.21 (Walling & Webb, 1986), 15.22 (Schmidt, 1985) Day for fig. 6.24; Satellite Remote Sensing Centre for chapter
Copyright © 1976,1978, 1980, 1983,1985,1986 & 1987 John heading photographs for Chapters 1, 14 & fig. 18.12; S. J.
Wiley & Sons Ltd.; John Wiley & Sons Inc. for figs. 2.2,2.3 & Smith for fig. 9.6; D. E. Sugden for figs. 11.2 & 11.12; J. T.
2.5 (Wyllie, 1976),5.9 (Rittman, 1962), 7.13 (Elter & Trevi- Teller for fig. 10.19; USGS for fig. 1.1OC; C. Warren for cover
san, 1973), 9.19, 9.26 & 16.7 (Schumm, 1977), 9.25 (Thorn- photograph; A. L. Washburn for figs. 12.8, 12.10 & 12.16;
bury, 1969), 11.16 & 11.26 (Flint, 1971), 18.2 (Strahler, 1969) M. I. Whitney for fig. 10.8; World Data Center A for Rockets
and tables 5.3 (Rittman, 1962), 9.2 & 14.4 (Schumm, 1977) and Satellites for figs. 1.1, 1.2, chapter heading photograph for
Copyright © 1962, 1969, 1971, 1973, 1976 & 1977 John Wiley Chapter 3, Figs. 19.2, 19.3, 19.6, 19.7, 19.8, 19.10, 19.11,
& Sons Inc.; Yale University Press for fig. 19.18 (Carr, 1981) 19.12, 19.14, 19.15, 19.16, 19.17, 19.19, 19.23, 19.24 & cover
Copyright © 1981 Yale University Press. Whilst every effort photograph, 19.25 & 19.26.
List of reviewers

Athol D. Abrahams, Department of Geography, State

University of New York at Buffalo
Victor R. Baker, Department of Geosciences, University
of Arizona
Paul Bishop, Department of Geography and Environ-
mental Science, Monash University
Peter A. Bull, School of Geography, University of Oxford
Chalmers M. Clapperton, Department of Geography,
University of Aberdeen
Ian Douglas, School of Geography, University of
Manchester
Thomas Dunne, Department of Geological Sciences,
University of Washington
Andrew S. Goudie, School of Geography, University of
Oxford
Barbara A. Kennedy, School of Geography, University of
Oxford
Patrick D. Nunn, Department of Geography, University of
the South Pacific
Kenneth Pye, Postgraduate Research Institute for
Sedimentology, University of Reading
Stanley A. Schumm, Department of Earth Resources,
Colorado State University
Alan G. Smith, Department of Earth Sciences, University
of Cambridge
David E. Sugden, Department of Geography, University of
Edinburgh
Stanley W. Trimble, Department of Geography, Univer-
sity of California, Los Angeles
Tjeerd H. Van Andel, Department of Earth Sciences,
University of Cambridge
Andrew Warren, Department of Geography, University
College,London
 Chapter heading
plate captions

1. View of the Earth showing landforms at a global scale. 6. Rectangular joint system on Checkerboard Mesa, Zion
The image shows a nearly cloud-free Africa with National Park, Utah, USA. The fractures may have
southern Europe towards the top and the north-east of resulted from cyclic near-surface volume changes
South America on the left. Landscape features evident resulting from temperature fluctuations, wetting and
at this scale are primarily related to the major horizontal drying or freeze-thaw.
motions of the continents over the past 200 Ma. Such 7. Compound slope developed in flat-lying, alternating
movements have created the Atlantic Ocean and led to beds of shale and resistant, massive sandstone, Utah,
the more recent opening of the Red Sea (top right USA.
corner). (Meteostat image courtesy Satellite Remote 8. Highly turbulent flow in a boulder-bed channel, River
Sensing Centre, RSA.) Inn, Austria.
2. Apollo 7 image showing a view north-east across the Sinai 9. The finely dissected terrain of the Loess Plateau,
Peninsula and the northern end of the Red Sea with the Shanxi Province, China. The high drainage density in
Gulf of the Suez on the left and the Gulf of Aqaba on the combination with the ease with which the wind-borne
right. This region is seeing the early stages of the develop- silt (loess) covering the region can be eroded by runn-
ment of a divergent plate boundary separating the Afri- ing water causes extraordinarily high rates of erosion.
can and the Arabian Plates with a spreading ridge extend- The Huang He River visible flowing north to south on the
ing northward up the Red Sea. The linear feature in the right of the image is estimated to carry an annual sediment
top right comer of the image is the Dead Sea Rift, a load of between 1.5 and 1.9 x 109 t where it leaves the
tranform plate boundary which extends northward from loess plateau. The area covered is about 130 km across.
the Gulf of Aqaba. The eastern side of this rift is moving (Landsat image courtesy N. M. Short.)
northward relative to the western side. (Image courtesy of 10. Part of the Rub'al Khali, or Empty Quarter, of southern
NASA, Lyndon B. Johnson Space Center, Houston, Saudi Arabia, the largest sand sea on Earth. The large-
Texas.) scale regularity that can be achieved by wind-shaped
3. Looking west towards the Himalayas and Tibetan depositional forms is clearly evident in this image
Plateau. The dark area to the left is the lower forested which spans an area 140 km across. Three categories
slopes of the Nep~l Himalaya. Mount Everest is of dune form are visible; complex barchanoid ridges
located a little below the middle-centre of the view. which interconnect to form a dune network or akle
(Apollo 7 image courtesy R. 1. Allenby Jr and the World pattern cover the bottom and right half of the image,
Data Center A for Rockets and Satellites.) while complex linear dunes occupying the top left
4. The suInmit of the Drakensburg Escarpment in Natal, quarter pass into star dunes in the extreme top left
South Africa, capped at this location by basalt lavas more comer. Note that north is at the bottom of the image.
than 1 km thick. This part of the Great Escarpment of (Landsat image courtesy A.S. Walker.)
southern Africa lies about 160 km inland from the coast 11. Crevasse patterns and medial and lateral moraine visible
and has formed along a sheared margin. on a glacier in northern Milne Land, eastern Greenland.
5. The snow-covered summit of Mount Taranaki (Egmont), (Photo courtesy of the Geodetic Institute, Copenhagen,
an almost perfectly symmetric strato-volcano, North Denmark.)
Island, New Zealand. 12. Ice-wedge polygons near Barrow on the northern
 coastal plain of Alaska. The polygons range from 7 to Okavango. The area covered by the image which is
15 m across. (Photo courtesy United States Geological about 180 km across is located within the area of
Survey Photo Library, Denver, Colorado, T. Pewe Group A dunes in Figure 14.21. (Landsat image pro-
845.) cessed by the Satellite Remote Sensing Centre of the
13. Oblique, northward-looking view of the coastline of CSIR, RSA.)
North Carolina and Virginia, USA taken with a hand- 15. View from the Tasman Sea eastwards to the Southern
held camera on the Apollo 9 Earth-orbital mission. Alps, South Island, New Zealand. The Southern Alps
Cape Hatteras is the eastern extremity of the line of have one of the highest sustained rates of crustal uplift
barrier islands cleary visible running down the centre known, reaching a maximum of around 10 000 m Ma- 1 .
of the image. These barrier islands, which have a total These rapid rates are matched by equally high rates of
length of over 300 km in the area covered by the image, erosion.
form part of a nearly continous chain of barrier islands 16. Anticline in the Zagros Mountains, Iran, at an early
extending from Long Island, New York, to Florida. stage of fluvial dissection. (Photo © Aerofilms Ltd.)
Cape Hatteras is just one of a number of cuspate fore- 17. A flight of uplifted coral reef terraces recording sea-
lands along the eastern coast of the USA. Cape Look- level highstands during the past 120 ka, Huon Peninsula,
out is the promontory at the bottom of the image and Papua New Guinea. (Photo courtesy J. Chappell.)
Pamlico Sound is the large body of water to the west of 18. Structural control of the Great Escarpment in western
Cape Hatteras. (Image courtesy NASA, Lyndon B. Cape Province, South Africa, provided by a massive
Johnson Space Center, Houston, Texas.) sandstone unit (see Figure 18.12.)
14. Landsat image of the north-central Kalahari showing a 19. The south polar region (left part of image) of Triton,
range of landforms created under alternating humid one of Neptune's moons. The morphology of Triton
and arid conditions. East-west orientated relict dune seems to be a collage of landscapes observed on other
systems evident from vegetation contrasts on the image bodies in the Solar System - certainly Triton has the
and indicative of past aridity have been cut by river most enigmatic surface of any planetary body yet
channels which thus post-date them and demonstrate explored. On this exceedingly cold world the surface
later humidity. Flooded pans apparent from their dark appears to be formed of frozen methane and nitrogen.
tone are visible in the top right comer of the image. The dark streaks visible at the bototm left of the image
The dark channel in the extreme bottom left comer of may be formed by eruptions of nitrogen from volcanoes,
the image is the Okavango River which has a perennial some of which may still be active. The long features on
discharge fed from the Angolan Highlands far to the the right side of the image are probably down-faulted
north-west. The broad light-toned channel running blocks analogous to rift valleys on Earth. (Voyager 2
across the centre of the image is of uncertain origin but mosaic courtesy JPL and NASA.)
may represent a periodically active overflow from the
Part I
Introduction
 1
Approaches to
geomorphology

1.1 The science of landforms Table 1.1 Examples of relationship between geomorphology and allied
disciplines
Geomorphology is the science concerned with the form of
the landsurface and the processes which create it. It is DISCIPLINE EXAMPLE OF EXAMPLE OF
extended by some to include the study of submarine fea- CONTRIBUTION CONTRIBUTION
TO GEOMORPHOLOGY FROM GEOMORPHOLOGY
tures, and with the advent of planetary exploration must now
incorporate the landscapes of the major solid bodies of the Geophysics Mechanisms and rates of Erosional response of land-
Solar System. One focus for geomorphic research is the uplift surface to uplift
relationship between landforms and the processes currently Sedimentology Reconstruction of past Form of alluvial channels
erosional events from in interpretation of
acting on them. But many landforms cannot be fully ex- a sedimentary sequence fluvial sediments
plained by the nature and intensity of geomophic processes Geochemistry Rate and nature of Mobilization of elements
now operating so it is also nec~ssary to consider past events chemical reactions in in earth surface
that may have helped shape the landscape. To a significant rock weathering environments
extent, then, geomorphology is a historical science. Hydrology Frequency and intensity Sediment concentration
of flooding in stream
Since the landsurface is located at the interface of the
Climatology Effect of climatic elements Effect of surface deposits
Earth's lithosphere, atmosphere, hydrosphere and biosphere, on rate and nature of and morphology on
geomorphology is closely related to a wide range of other geomorphic processes climatic variable
disciplines (Table 1.1). While having a central interest in Pedology Effect of soil properties Topographic control over
landforms, geomorphologists must, none the less, be aware on slope stability soil-forming processes
of those aspects of allied disciplines that bear on their sub- Biology Role of vegetation cover in Topographic control over
affecting rates of erosion micro-environments of
ject. Equally, geomorphology has a potential, as yet only plant growth
partially realized, of making significant contributions to these Engineering Techniques for analysis of Identification of
other areas of knowledge. slope instability morphological features
indicative of slope
instability
1.2 The development of ideas Space science Context for understanding Interpretation of planetary
special characteristics landscapes by analogy with
The way geomorphologists ~pproach the study of landforms of landform-creating terrestrial landforms
at the present time can only be seen in a proper context if environment on the Earth
we appreciate how the central concepts of geomorphology
have been developed. Long before the term geomorphology
itself was introduced in the 1880s people had speculated on of the landsurface. The idea that streams have sufficient
the forces and mechanisms that had created the natural power to erode their valleys was appreciated to some extent
landscape around them. Aristotle, Herodotus, Seneca and by Seneca, and certainly by Leonardo da Vinci in the fif-
Strabo, among other Greek and Roman philosophers, wrote teenth century, but it was not until the late eighteenth century
on phenomena such as the origin of river valleys and deltas, that the implications of this fundamental concept began to
and the relationship between earthquakes and deformation be fully explored.
4 Introduction

1.2.1 The age of Hutton and Lyell proposition that changes on the Earth's surface are
usually slow, steady and gradual. Although Lyell did
In 1785 James Hutton presented a paper to the Royal Society
acknowledge that major events, such as floods and
of Edinburgh in which he argued that the landsurface had
earthquakes, do take place, he maintained that such
been shaped by the slow, unremitting erosive action of
phenomena are local in extent and that they occurred in
water rather than by the catastrophic events advocated by
the past with the same average frequency as they do
biblical scholars; to the history of the Earth Hutton saw 'no
vestige of a beginning, no prospect of an end'. His ideas today.
4. Uniformity of state: this is the idea that, although
disseminated only slowly until in 1802, five years after his
change occurs, it is directionless; that is the Earth
death, John Playfair, his friend and Professor of Mathematics
always looked and behaved much as it does at the
at the University of Edinburgh, restated and elaborated his
present time. This concept was a central pillar in
views with an elegance and clarity that has rarely been
Lyell's grand vision of earth history as an endless
matched in scientific writing. In his Illustrations of the
Huttonian Theory of the Earth Playfair provided the first
succession of cycles.
These multiple meanings of uniformity led to much con-
detailed and closely reasoned account of several important
fusion in the vigorous debate which Lyell provoked after
aspects of landform genesis, most notably the relationship
between rivers and their valleys: 1830 because it was possible to accept some of the pro-
positions embodied in uniformitarianism while at the same
Every river appears to consist of a main trunk, fed from time rejecting others. Opposition to uniformitarianism came
a variety of branches, each running in a valley pro- from geologists who subscribed to catastrophism - the idea
portioned to its size, and all of them together forming a
system of valleys, communicating with one another, and that many of the features of the landscape were to be explained
having such a nice adjustment of their declivities, that by rapidly occurring events, rather than by gradual change.
none of them join the principal valley, either on too high The great majority of catastrophists were not, as is often
or too Iowa level, a circumstance which would be in- portrayed, believers in a landscape created by acts of divine
finitely improbable if each of these valleys were not the intervention, since, by the 1830s, few serious geologists
work ofthe stream which flows in it.
accepted a 6000 year biblical time scale for Earth history.
Hutton's methodology, founded on the belief that the slow In fact they had no argument with the uniformity of law
but continuous operation of processes observable at the and uniformity of process, but, on the basis of their
present day provided a sufficient basis for explaining the interpretation of the available field evidence, they firmly
present configuration of the Earth's surface, was taken up and rejected Lyell's ideas on the uniformity of rate and
developed by Charles Lyell in his idea of uniformity. uniformity of state. One important type of evidence the
Through his highly influential work, Principles of Geology catastrophists pointed to was the so-called 'drift' deposits
(1830-33), Lyell became the 'great high priest' of what formed of a mixture of boulders, gravel and sand which
became known as the principle of uniformitarianism, a were known to blanket large areas of northern Europe.
concept frequently (but inadequately) summarized by the These materials, which were to be found on hilltops as well
phrase 'the present is the key to the past'. Lyell's notion of as in lowlands, had earlier been cited as evidence of the
uniformity was far more complex than is often appreciated biblical Flood but, scriptural arguments aside, they pro-
by many earth scientists and this has led to much confusion vided a powerful argument against Lyell's extreme notions
as writers have failed to distinguish between its various of gradualism and uniformity of state. The argument in
meanings. In fact four distinct meanings can be identified effect centred around the extent to which the intensity of
in Lyell's Principles. particular landscape-forming processes might change over
1. Uniformity of law: this is the assumption that natural time, and this debate has continued in various guises to the
laws are constant in time and space. present day where the primary concern is the relative
2. Uniformity of process: this is the proposition that if significance of rare, large magnitude (catastrophic) events
past events can be explained as the consequence of in landform genesis.
processes now known to be operating then additional The idea that these drift deposits had been laid down by
unknown causes should not be invoked. In essence, glaciers gradually emerged in the early nineteenth century;
this is the principle of simplicity adopted in all but a glacial theory did not become a widely accepted expla-
scientific explanation; if known processes are capable nation for drift deposits, and other landforms which were
of explaining natural phenomena additional 'exotic' apparently inexplicable in terms of normal fluvial erosion,
mechanisms should not be introduced. For Lyell, and until after the publication in 1840 of Louis Agassiz's Etudes
Hutton before him, this principle was in fundamental sur les Glacier.s. His notion of a Great Ice Age was soon
opposition to notions of divine intervention as an being applied by other workers to the landscapes of northern
explanation for the Earth's surface form. Britain, while Agassiz himself, who eventually moved to the
3. Uniformity of rate (gradualism): this is the USA from his native Switzerland, extended his glacial
 Approaches to geomorphology 5

theory to North America. Concurrent with the acceptance 1.2.3 The modern era
of the idea of continental glaciation was a continuing debate
A lack of empirical evidence as to the nature and rate of
over the relative importance of marine and fluvial erosion.
landscape change through time, coupled with a poor level
By the 1870s this had been resolved firmly in favour of the
of understanding of the processes responsible for landform
predominance of rivers in shaping the landscape.
genesis, led to increasing doubts among many geomor-
1.2.2 Developments in North America phologists as to the viability of historical explanation in
geomorphology. Foreshadowed by R. E. Horton's remarkable
If the foundations of the scientific study of landforms were
synthesis of drainage basin hydrology published in 1945, the
laid in Europe in the first half of the nineteenth century, following decades witnessed a growing emphasis, especially
much of the conceptual structure of the modem discipline in the UK and North America, on both the quantitative
was erected in the second half of that century by a re- analysis of landform morphology (landform morphometry
markably gifted group of ~merican geologists led by John or geomorphometry) and on the field measurement of ge-
Wesley Powell and Grove'~arl Gilbert. Exploring the mostly omorphic processes. These developments were not so evident
semi-arid terrain of the western USA ~here the detailed in Continental Europe where the earlier tradition of geomor-
relationships between rock structures and landforms were phology founded on the relationship between landform
largely unobscured by soil and vegetation, Powell was able
characteristics and climatic zones was strengthened after
to develop a structural and genetic classification of moun-
the Second World War.
tains, as well as classifications of valleys and drainage
The 1960s and 1970s saw a major reorientation of ge-
systems. His greatest conceptual contribution was the rec- omorphology in the UK and USA towards the development
ognition of the importance in landform development of base of predictive models of short-term landform change. These
level- the lower limit in the landscape, ultimately represented
were based on a much improved knowledge of geomorphic
by sea level, below which rivers cannot erode.
processes founded on a greater understanding of the basic
Of even greater significance to the later development of the
physical principles involved. Indeed, these models often
subject was the pioneering research of Gilbert. He created
reflected a significant input from research by engineers,
the first systematic analysis of the mutual interaction be-
particularly with respect to slope stability, the flow of water
tween the driving forces of erosion and the resisting forces
in river channels and the entrainment and transport of
represented by the rocks and superficial deposits of the
sediment. During this period there was also a rapid growth
.Earth's surface. The resulting series of laws of landscape
in applied geomorphology with predictive models being
.development, founded upon Gilbert's concept of dynamic
used to assess the likely response of the landscape to chang-
equilibrium (see Section 1.3.4), were brilliantly presented in
ing conditions brought about by human actiVities, such as land
his classic monograph, Report on the Geology of the Henry
use changes and dam construction.
Mountains, and developed further in a novel quantitative
treatment of fluvial processes published in 1914.
1.2.4 Future directions
Whereas Gilbert emphasized the adjustment between
present forms and present processes, his compatriot and Although greatly advancing our knowledge of surface pro-
contemporary William Morris Davis founded a school of cesses, much current research in geomorphology contributes
geomorphology based on the concept of a systematic pro- little to our understanding of how extensive areas of a land-
gression of landform change through time initiated by rapid scape change over long periods of time. But revolutionary
uplift of the landsurface. This evolutionary sequence, termed changes since the late 1960s in disciplines allied to geomor-
the cycle of erosion (see Section 1.3.4), was enthusiasti- phology, together with the application of new techniques,
cally extended and applied by Davis's students and other now provide the opportunity to look at this problem anew
researchers in the USA and the UK. But it was never and to develop a global perspective for geomorphology to
accepted by the majority of European geomorphologists accompany the existing disciplinary focus on small-scale,
who reacted against what they saw as the overly theoretical surface process studies. What new directions, then, can we
and idealized nature of the model, as well as the way it see for geomorphology within such a global perspective?
underplayed the importance of climate in influencing land- The late 1960s and early 1970s saw a dramatic change in
form development. A further European challenge came our understanding of the Earth as the notion of continental
from Walther Penck, who rejected the crucial assumption mobility embodied in the concept of plate tectonics was
inherent in the cycle of erosion that the Earth's surface can developed and refined (see Ch. 2). Davis's scheme of long-
be stable for a sufficient period of time after an episode of term landform development had no comprehensive tectonic
rapid uplift for an evolutionary sequence of landforms to be theory with which to work, and consequently his assumptions
developed. He argued instead that the overall form of the about uplift and landsurface stability were inadequately
landscape would depend primarily on whether the rate of grounded. We now have a model capable of making pre-
uplift was increasing, decreasing or constant through time. dictions of general patterns of uplift and stability for the
6 Introduction

Earth's surface through time and space and know that a alien landscapes of the Moon (Fig. 1.1), Mars (Fig. 1.2) and
mode of landscape development applicable to one tectonic the other planetary bodies have presented a major challenge
setting is not necessarily relevant to another. Moreover, to geomorphologists.
significant advances in dating techniques mean that in many The study of planetary surfaces is a 'two-way street'; on
cases long-term rates of uplift and denudation can be esti- the one hand terrestrial analogues have been invaluable in the
mated_ with some accuracy. One of the results of the appli- interpretation of the surface forms of other planetary bodies
cation of these dating techniques is the realization that in (Fig. 1.3), but on the other hand planetary landscapes provide
relatively stable regions landsurfaces may survive without an invaluable perspective with which to consider long-term
significant modification by erosional processes for several landform development on the Earth. Many planetary sur-
tens of millions of years. faces have experienced relatively little change for billions
A related advance in the earth sciences of great potential of years and so retain the effects of very rare, but cata-
value to geomorphologists has come from the exploration of
the oceans. Since the 1960s an active programme of drilling
in the deep ocean basins and in the shallower waters around
the margins of the continents has produced a wealth of data
on the quantities of sediment that have accumulated in these
environments over periods of 100 Ma or more (see Ch. 15).
These sedimentary sequences provide a unique repository
of information recording the erosion of adjacent land
masses. There have been numerous studies by geomor-
phologists which have used the volume of sediment accumu-
lating in lakes and reservoirs to quantify rates of denudation
over periods of up to a few thousand years, so the approach
is not novel; what is novel is the temporal and spatial scale
which can now be addressed.
A third development of importance to problems of long-
term landform evolution has been the revolution in our
understanding of climatic change (see Ch. 14). We now know
that climatic fluctuations, especially during the past 2 Ma or
so, have been both more frequent and more extensive in
their effects than was appreciated up to the 1960s. The
implications of these new ideas of global climatic change
for models of long-term landscape development have been
especially appreciated by geomorphologists investigating
humid tropical landscapes where earlier ideas of climatic
stability have had to be dramatically revised. Given the in-
creasing importance now being attached to the interactions
between the atmosphere, the oceans and the landsurface in
climatic modelling, we can anticipate that geomorphology
will come to play a more central role in the understanding
of climatic change itself. We now know, for instance, that
atmospheric carbon dioxide is a crucial factor in determining
global temperatures and that its concentration in the atmos-
phere is partly controlled by the rate of weathering reactions
on the landsurface; this, in tum, is influenced by a range of
geomorphic factors.
One further opportunity for developing a broader, global Fig. 1.1 Hadley Rille, a sinuous channel more than 115 km in
perspective for geomorphology lies in the comparative length cutting across a lava plain adjacent to the Montes
analysis of the planets and moons of the Solar System (see Apennines (mountains on the right of the image) on the edge of
Ch. 19). The period since the 1960s has seen a new research the Moon's Imbrium Basin. This rille, like other similar channels
frontier open for geomorphologists with the beginning of the on the Moon, is thought to be a lava channel on the basis of
analogous landforms occurring in volcanic regions on the Earth
exploration of the Solar System. As with the exploration of (although the lunar versions are many times larger than their
the then unfamiliar arid landscapes of the south-west USA terrestrial equivalents). (Apollo 15 image, World Data Center A
which led to the pioneering work of Gilbert and Powell, the for Rockets and Satellites.)
 Approaches to geomorphology 7

Fig.I.3 Oblique view from a height of about 10 km of the

canyon of the Little Colorado River in Arizona, USA, a tributary
· Fig. 1.2 Part ofNirgal Vallis, an 800 km long channel system of the Colorado River. This canyon, like many others in the
on Mars incised into old cratered terrain (see Section 19.3.8). The Colorado Plateau in Utah and Arizona, has a 'box' form very
tributary heads have an amphitheatre-like morphology and similar to some of the channels on Mars. Such landforms provide
through analogies with equivalent terrestrial forms this supports valuable analogies for interpreting equivalent Martian features
an origin involving runoff emerging from springs (a process (see Figure 1.2). It is thought that such box canyons on Earth
known as spring sapping or ground water sapping). Note the form as a result of water percolating through a thick permeable
crater located on the channel near the left-hand edge of the image sandstone unit and emerging where it encounters a less
which must have been formed after the channel. The area shown permeable underlying lithology. The resulting spring sapping
is about 55 km across. (Viking 1 image, World Data Center A for causes the canyon head to recede.
Rockets and Satellites.)

clysmic, landscape-forming events. On the Earth the effects 1.3 Some key concepts
of these events, such as the impact of large objects and
catastrophic floods, are usually rapidly obliterated by
erosional, tectonic and volcanic processes, but in some 1.3.1 Endogenic and exogenic processes
cases we can use the landscape history of other planets to The Earth's detailed form at any instant in time represents
assess their role in Earth history. For instance, although the net effect of surface, or exogenic processes, and internal,
only about 120 craters formed by impacts rather than or endogenic processes. Exogenic (also termed exogenetic)
volcanic activity have been identified on the Earth (Fig. processes, including the action of water, ice and wind,
1.4), we can estimate the approximate frequency of major predominantly involve denudation, that is, the removal of
impact events in the past caused by comets, meteorites and material, and thus generally lead to a reduction in elevation
asteroids by using the largely preserved history of bom- and relief (Fig. 1.5). (Note that the term relief refers to a
bardment experienced by the Moon and some other plane- difference in height and must be distinguished from
tary bodies. elevation and altitude which refer to height above some datum
8 Introduction

Fig.1.4 ExampLes of impact structures identified on Earth. (A) Manicougan (Quebec, Canada),formed around 210 Ma BP, is about
70 km across and has a broad centraL upwarp and a surrounding depression now filled with water after the construction of a dam. (B)
CLearwater Lakes (Quebec, Canada), a pair of depressions 20 and 30 km across which have been much eroded. (C) Serra da CangaLha
(BraziL), a 12 km diameter ring structure, and the smaller 4 km diameter Riachao Ring (inset). It has been estimated that a 10 km-
diameter object hits the Earth on average every 100 Ma at a veLocity of about 20 km S-l forming a 150 km wide crater. (Landsat images
courtesy of J. McHone and N. M. Short.)

Tectonic uplift Tectonic subsidence - normally mean sea level.) An exception is the localized
® e deposition of material, to form sand dunes for instance, which

.
causes an increase in relief. Denudation can involve the re-
moval of both solid particles and dissolved material. In this

• book we will generally refer to the former as erosion or

mechanical denudation, and to the latter as chemical denuda-
tion.
The two sources of energy which power the various exo-
Volcanism genic processes are solar radiation and the potential energy
® arising from the gravitational attraction of the Earth which,
t in the absence of sufficient resisting forces, causes the down-

.
slope movement of water, ice and particles of rock and soil.
>-
(9 Solar radiation acts in diverse ways, providing the energy for
cr:
L.U
Z
biological activity, the evaporation of water and the func-
L.U tioning of the Earth's atmospheric circulation.
.....J
« Endogenic (alternatively endogenetic) processes are gener-
Denudation
i= ally constructional in that they usually lead to an increase in
Z
L.U elevation and relief. Three major types of processes are in-
f-
0
a..
volved. Igneous activity consists of the movement of molten
............ rock, or magma, on to, or towards, the Earth's surface.
Z Orogenesis (orogeny) is the formation of mountain belts which
0
are typically arcuate or linear in plan form. Epeirogenesis
~
:::> (epeirogeny) is the uplift of usually large areas of the Earth's
L.U
.....J
L.U e @ surface without significant folding or fracture. The broad
structures of the Earth's crust and the processes of deforma-
tion and faulting which give rise to them are described by the
term tectonics, while the term morphotectonics is applied to
Sea level the interaction between tectonics and landform genesis.
rise
Neotectonics refers to the processes and effects of recent tecto-
nic activity and is usually applied to Late Cenozoic events.
Fig.1.5 Schematic illustration of the change in eLevation (height
reLative to a datum) associated with various exogenic and 1.3.2 Geomorphic systems
endogenic processes. The positive and negative signs indicate,
respectiveLy, increases and decreases in eLevation and potentiaL While the application of systems analysis to geomorphic
energy. The mechanism of isostasy is discussed in Section 2.2.4. phenomena has, arguably, not in itself led to any great
 Approaches to geomorphology 9

· advances in understanding, there has been a widespread use represented by process-response systems (Fig. 1.6).
of systems concepts in geomorphology and an extensive Before a system can be analyzed its boundaries must be
adoption of its terminology. A system can be defined as a defined. In an open system there is a movement of both
set of objects or characteristics which are related to one energy and matter across the system boundary, whereas in a
another and operate together as a complex entity. Systems closed system only energy is transferred. An input of mass or
analysis focuses on the relationships between these objects energy into a system is transmitted through it (throughput)
or characteristics. and leaves as an output. Changes in inputs of energy or
In geomorphology three kinds of system can be ident- mass usually produce changes in outputs, but may also give
ified. Statistical relationships between the morphological rise to adjustments in the structure of a part of the system
properties of landform elements are represented by morpho- (subsystem). Changes in the flow of energy and mass as
logical systems, while movements of mass and flows of well as adjustments to the structure of the system are
energy through the landscape are described by cascading controlled by the relationships between variables within the
systems. Interactions between these two types of system system. These variables, representing the form of the
resulting from adjustments between process and form are landscape, the rate of geomorphic processes acting upon it

A MORPHOLOGICAL SYSTEM B CASCADING SYSTEM

Atmospheric
Moisture

Soil
Water

Surface
Water

C PROCESS - RESPONSE o PROCESS - RESPONSE

SYSTEM SYSTEM
Negative Feedback Positive Feedback

Uplift
+ Storm event

Stream Rate of
Gradient Surface
Runoff

Rate of
Erosion Water
Rate Infiltration
into Soil

Stream Rate of
Flow Soil
Velocity Erosion

Fig.1.6 Examples of simple morphological (A), cascading (B), and process-response systems (C) and (D). Negative and positive
signs indicate the nature of the relationship between variables. In (B) rectangles represent storages and triangles transfers between
storages. Note that negative feedback systems (C) have an odd number of negative relationships, whereas positive feedback systems (D)
have either an even number or none.
10 Introduction

and the environmental factors influencing these geomorphic nantly dry conditions are occasionally punctuated by violent
processes, may be either independent (causal) or dependent storms, river discharges may vary by two or three orders of
(responding to causal variables). magnitude (by a factor of 100 or 1000).
A very common characteristic of geomorphic systems is Understanding the frequen~y with which geomorphic
negative feedback, a condition whereby the structure of events of different magnitudes occur is clearly a crucial
the system is capable of adjusting in a way that minimizes component of any explanation of landform genesis.
the effect of externally generated changes. Such an ability Measurements of the operation of various geomorphic pro-
of self-regulation or homeostasis, means that a system can cesses over a range of time scales show that extr~me, high-
maintain a state of balance or equilibrium. If faulting across intensity events are rare and that low- tt> medium-intensity
a river bed causes an instantaneous increase in channel events prevail for the great majority of the time. Accordingly
gradient, for instance, the resulting increase in river flow we find that plots of the frequency distribution of many
velocity will tend to promote a local increase in the rate of variables which influence the operation and intensity of
channel downcutting and, as a consequence, a reduction in geomorphic processes, such as wind speeds and river dis-
channel gradient (Fig. 1.6 (C». charges, have a positive (right)-skewed form (Fig. 1.7). The
In other cases an input change may engender a system rarity of many types of extreme event, such as major river
response which produces an output which reinforces the floods, arises from the fact that these generally require a
original input and eventually causes a shift in the system to specific combination of conditions which, while individually
a new equilibrium state. This 'snowball effect', as it is not uncommon, are very unlikely to coincide. A simple
colloquially called, occurs in systems exhibiting positive analogy is provided by the throwing of dice. The prob-
feedback and is precipitated by the breaching of a threshold ability of throwing a six with one die is 1 in 6 (16.67 per
in the system. When a severe storm leads to the erosion of cent), but the probability of obtaining three sixes from three
the uppermost, permeable layer of the soil the less permeable dice thrown simultaneously is only 1 in 216 (0.46 per cent).
subsoil is exposed (Fig. 1.6(D». As this horizon cannot
absorb the water running off the surface so effectively the
depth of runoff increases, the rate of erosion accelerates and
subsoil material with an even lower capacity for water Q)
(.)
absorption is exposed. In this way it is possible for an entire c::
soil profile to be removed in a single severe storm. The new Q)
l-
equilibrium is reached when all the loose, readily erodible I-

::J
soil has been removed and the much more resistant under- (.)
(.)
lying weathered bedrock is exposed.
A final important characteristic of systems is their hier-
archical property. A specific system may be composed of
numerous smaller systems, but itself form a part of a larger
-
o
o
>
(.)

system. Perhaps the clearest example of a system hierarchy c::

Q)
in geomorphology is provided by drainage basins; a ::J
C'
medium-sized catchment may contain numerous individual Q)
I-
streams, each with its own small basin, but at the same time U.
form part of a much larger drainage system. l<...--- ------'~

Magnitude of Event
1.3.3 Magnitude and frequency Fig. 1.7 Characteristic right-skewed frequency distribution of
There are great variations in the rates at which different variables (such as wind speed, river discharge and wave height)
affecting geomorphic processes. The highest frequency of events
geomorphic processes operate. The rate of some processes, occurs in the lower to middle magnitude range, whereas there is
such as the flow of ice within an ice sheet, may be fairly only a small proportion of high magnitude events.
constant over millennia, whereas other mechanisms, such
as landsliding, are inactive for long periods, although when
they do operate they act with great rapidity. Moreover, the The frequency of an event of a specific magnitude is
same process may exhibit a markedly different intensity and expressed as the average length of time between events of
degree of variability under contrasting climatic conditions. that magnitude and is known as its return period or
In a constantly humid environment lacking high-intensity recurrence interval (Fig. 1.8). If a particular maximum
storms, river discharges may not vary by more than an order annual discharge of a river has a recurrence interval of 20 a
of magnitude (by a factor of 10) over several years. By this means that there is a 1 in 20 (5 per cent) chance that a
contrast, under a semi-arid climatic regime where predomi- flood of this magnitude or greater will occur in anyone year
 Approaches to geomorphology 11

25 1.3.4 Equilibrium and evolution

Earlier in this chapter (Section 1.2) reference was made to

two contrasting approaches to the understanding of landform
genesis. In the fIrst of these, pioneered by Gilbert, the empha-
sis is placed on the mutual adjustment between present
20 forms and processes. This is sometimes described as the
functional approach to geomorphology, and the kinds of
I
en geomorphic systems with which it is usually concerned
M
commonly exhibit negative feedback and the tendency for
E
an equilibrium form to be restored after a disturbance by a
Q)
high magnitude - low frequency event. The second mode of
...
lJ)

explanatIon, which is indelibly linked with the work of Davis,

~(.) 15
en focuses on progressive changes in the landscape through
C time and is usually described as the evolutionary, or his-
torical approach to geomorphology. We clearly need to
examine further this apparent paradox that landforms can
be considered both to retain an equilibrium form and under-
go a progressive change in form through time.
10
If we imagine observing the boulder-strewn bed of a
mountain stream for a period of a few hours we would be
very unlikely to witness any measurable change in its form
(assuming, of course, that the river is not in flood). Water
1-01 2 5 10 20 50 100 flows in the channel transporting fine sediment and dis-
Recurrence Interval (a) solved material, but its gradient, form and elevation above
sea level remain essentially the same. We can describe this
Fig. 1.8 A typical relationship between stream discharge and
recurrence interval. situation as one of static equilibrium (Fig. 1.9(A)).
If we were to remain at our observation point for several
months or years we might well see a major flood which, by
and that, on average, there will be one such flood every causing a significant amount of erosion, lowers the bed of
20 a. The accurate calculation of such probabilities requires the channel by a small, but measurable, amount. Although
a record of measurements which is at least as long as the this temporarily causes a change in the form of the channel,
recurrence interval being estimated. Consequently the more over the succeeding years the slow deposition of sediment
extreme and rarer an event, in general the more difficult it will lead to a restoration of the original channel bed elevation;
is to estimate accurately its recurrence interval. A further in these circumstances the channel can be said to be in
problem is that over the long term the average state of a steady-state equilibrium since there is no change in its
variable may change significantly. Climatic changes, for in- mean elevation over the period in question (Fig. 1.9(B)).
stance, can involve marked shifts in average temperatures and Note that this is essentially the kind of adjustment envisag-
precipitation and these may in tum modify the magnitude ed by Gilbert in his concept of dynamic equilibrium (see
- frequency relationship of geomorphic processes. Section 1.2.2).
It is important to emphasize that for virtually all geomor- Stretching our powers of imagination further we could'
phic events such recurrence intervals are averages only and envisage continuing our observations over thousands or
do not indicate when we should expect an event of a particu- hundreds of thousands of years. Over such a period of time
lar magnitude to occur. We could, for instance, observe five there would be hundreds of floods including many events
separate floods with a recurrence interval of 20 a in a single of very high magnitude. Episodes of severe erosion would
year, although this would be highly unlikely. In the case of lead to a progressive lowering of the channel floor and poss-
some processes, however, a degree of regularity is evident ibly a gradual decrease in channel gradient. This situation is
in the occurrence of events of a similar magnitude. Earth- one which is now generally termed dynamic equilibrium
quakes, for example, represent the release of stresses which (Fig. 1.9(C)).
gradually build up within the Earth's crust. A large Finally, we can imagine what might happen over a
earthquake, with its related aftershocks, is therefore unlikely period of millions of years. As the altitude of the mountain
to be followed within a period of several years by another carrying our stream decreases we would see a progressive
large earthquake in the same locality since much of the stress reduction in the elevation of the channel bed. However, we
would have been released in the earlier event. would expect the rate of decrease in elevation to decline
12 Introduction

A planation surface is created. (Fig. 1.9(D)). This is essent-

ially the notion of landscape change which forms the basis
of Davis's cycle of erosion.

1.3.5 Scale in geomorphology

1.3.5.1 Temporal scale

From our discussion of equilibrium and evolution it will be
Static Time (10- 3 a) ~
apparent that the time scale over which the development of
B a landform is being considered is fundamental in determining
Static Time what kind of conception of equilibrium is relevant. The
time scales indicated in Figure 1.9 serve as only a very
general guide since the precise period of time required to
attain any particular type of equilibrium depends on a
number of factors. In environments in which rates of erosion
and deposition are high, landforms can be modified rapidly
and the effects of occasional high-magnitude - low-
frequency events can generally be obliterated fairly quickly.
Steady Time (10'a ) ~
In such environments steady-state equilibrium can be
attained in a short period of time. A second factor is the
c
0
degree of resistance of the material experiencing erosion.
The form of a channel in unconsolidated alluvium may
change in a few hours in response to a marked change in
co
>
river discharge, whereas a channel of similar size cut in
Q) densely cemented, resistant bedrock may only adjust after
hundreds or thousands of years (assuming the change in
UJ
discharge persists). A third factor is the size of the landform
being considered. The shape of a ripple on a sand dune will
Dynamic Time (10 3 a) ~
adjust much more rapidly to a change in wind direction
than the form of the sand dune on which it lies simply
Q)
because a far smaller amount of sediment has to be moved
c
c Dynamic Time by the wind to accomplish the adjustment. Similarly, we
co would expect a small mountain mass a few hundred metres

~
.r= high to reach decay equilibrium in a few million years,
U
whereas tens of millions of years would probably be
required for a major mountain range several thousands of
metres high.
In more general terms we can talk of the speed with which
Cyclic Time (10 7 a ) ~
a change in an input to a geomorphic system, such as an
Fig. 1.9 Different types of landform equilibrium illustrated with increase in rate of uplift or decrease in river discharge, is
reference to schematic changes in channel elevation over fully reflected in a change in form. This is referred to as the
different time spans: (A) static equilibrium; (B) steady-state relaxation time of the system and can range from a few
equilibrium; (C) dynamic equilibrium; (D) decay equilibrium.
minutes for changes in a small section of an alluvial channel
The time scales shown are merely suggestive and will vary by
orders of magnitude depending on the type of landform being to tens of millions of years for the uplift of a major mountain
considered and the nature and intensity of the prevailing range.
geomorphic processes.
1.3.5.2 Spatial scale
From our discussion of temporal scale it is clear that there
through time as the channel bed approaches base level is a close relationship between the temporal and spatial
and the channel gradient becomes so low that rates of scales at which landform change occurs. Returning to our
erosion reach a minimum. Over this longest time scale the illustration of the different types of equilibrium in Section
landscape evolves towards a decay equilibrium and a 1.3.4, we can see that while it is appropriate to consider a
landsurface of subdued relief, termed an erosion surface or small section of a stream channel in terms of static, steady-
 Approaches to geomorphology 13

state, and perhaps even dynamic equilibrium, it is inap- 1.3.5.3 Scale and causality
propriate to discuss long-term landform evolution and the As the temporal scale of analysis changes so the status and
attainment of decay equilibrium in terms of such a restricted interrelationships of variables in geomorphic systems alter
spatial scale. Not all parts of a landscape can be simulta- accordingly. If we examine landforms over dynamic time
neously in a steady-state equilibrium since sediment is con- then variables which only exhibit measurable change over a
tinuously being removed and relief progressively reduced. longer period can be regarded as fixed; that is, they are
Geomorphology is concerned with phenomena over an independent variables of the external environment. Over
enormous range of scales from the form of an individual dynamic time such variables include lithology and structure
boulder to the morphology of the Earth's major relief features and the initial relief of the region. Those factors which change
(Fig. 1.10). It is often useful to be able to categorize this over dynamic time, such as hillslope morphology and channel
range of scale and to talk, for instance, of microscale or gradient, are elements of the geomorphic system and are
macroscale forms. Such a classification is presented in therefore dependent variables at this scale. Variables such
Table 1.2 with some suggested ranges of linear and areal as water discharge and rate of sediment transport in a river
scale, although it must be emphasized that the divisions channel which change very rapidly over dynamic time can
between each scale are somewhat arbitrary. Examples of be regarded as having mean values which are part of the
landforms at each scale are indicated together with the system, but around which occur random fluctuations which
main endogenic and exogenic factors influencing landform are irrelevant to the system at this scale.
genesis at the different scales. A very approximate A further consideration in understanding the factors
equivalence between these spatial scales and the temporal controlling landform development at different temporal and
scales associated with the different types of equilibrium is spatial scales is the relative importance of endogenic and
also suggested; the exact nature of this relationship will, exogenic processes. If we are examining a drainage basin
however, depend on the characteristics of the particular of a few square kilometres in a recently uplifted mountain
landform being considered. range the tectonic history of the area will be essentially

Table 1.2 Hierarchy of spatial and temporal scales in geomorphology

DIMENSIONS EXAMPLES OF LANDFORMS MAJOR CONTROLLING

SPATIAL FACTORS TEMPORAL DURATION
SCALE SCALE
Linear Areal Endogenic Exogenic Endogenic Exogenic
(km) (km 2) Fluvial Glacial Aeolian

Micro <0.5 <0.25 Minor Pools and Small Sand Individual Microclimates; Steady lOla
fault riffles in moraine ripples earthquakes meteorological time
scarps a small ridges and volcanic events
river channel eruptions

Meso 0.5-10 0.25-10 2 Small Small Local and Local climates;

volcanoes Meanders glacial Dunes regional short-term Dynamic 103 a
valleys isostatic climatic time
uplift; change
localized
volcanism and
seismicity

Macro 10-103 102-106 Block- Floodplains Highland Sand Regional Regional

faulted of major ice caps seas uplift and climates;
terrain rivers subsidence long-term
climatic
change
(glacial-inter-
glacial cycles)
Cyclic 107a
time
Mega >10 3 >10 6 Major Major Continental Large Long-term Major climatic
mountain drainage ice sheets sand patterns of zones; very
ranges basins seas uplift, long-term
subsidence and climatic
continental change
motion (ice ages)
14 Introduction

Fig.1.10 Images illustrating the range ofspatial scales

considered in geomOlphology. (A) The megascale landscape
assemblages of the western part of the USA covering an area of
about 2.25 x 106 km 2 (linear dimension around 1200 km). At this
scale the overall tectonic framework of the region is clearly
visible. Among the most prominent features are the Sierra Nevada
(SN) and the Basin and Range Province of Nevada and southern
California (BR), a region characterized by a succession of
mountain ranges and intervening basins which includes Death
Valley (DV). (Part of mosaic of ERTS-I (Landsat) imagery
prepared for NASA Goddard Space Flight Center by the USDA
Soil Conservation Service.) (B) The macroscale landscape of
Death Valley (DV) and adjoining NW-SE-orientated basins
containing extensive lake deposits (light tones) laid down during
more humid periods in the past (boxed area on (A)). The western
side of Death Valley is flanked by a coalescing series of alluvial
fans (A) formed by sediment brought down from the adjacent
mountain range. At this scale both detailed tectonic controls and
the broad effects of exogenic processes are evident. The area
covered is about 34 x 103 km 2 (linear dimension 185 km).
(Landsat image courtesy A.S. Walker.) (C) Alluvialfans at the
northern end of Death Valley (boxed area on (B)). The area
covered is about 70 km 2 (linear dimension about 8 km). At this
mesoscale the effects of exogenic processes (in this case erosion
and the deposition ofsediment) can be clearly seen. (Air
photo by United States Geological Survey.) (D) A view across the
swface ofone of the fans shown in (C) (arrowed). At this scale
the detailed effects of exogenic processes are apparent.

A
 Approaches to geomorphology 15

c
o
16 Introduction

irrelevant to understanding the present-day morphology of a landscape we can either try to discover what processes are
landform elements in the basin. The prevalence of steep currently active and attempt to explain its present form with
slopes and rapid rates of erosion related to its history of reference to these processes, or we can endeavour to unravel
uplift would be regarded as elements of the external environ- the history of the landscape and understand its present form
ment outside the geomorphic system being studied. If the in terms of a sequence of landscapes through time (see
scale were to be expanded to a few thousands of square Section 1.3.4). The first of these (the functional approach)
kilometres a full understanding of the landscape would . emphasizes the immanent aspects of reality, the second (the
require some consideration of its history over the previous ,evolutionary or historical approach) emphasizes the con-
million years or so and an assessment of the interaction figurational aspects. Whereas relating present forms to
between endogenic and exogenic processes over that currently active processes may be a successful strategy if
period. At the largest scale we might be dealing with an we are working at the small scale, or where landforms are
entire mountain range covering tens of thousands of square adjusting very rapidly to the operation of geomorphic
kilometres; in this case it would be the endogenic processes processes, this is not an adequate approach where we are
that would have exerted the major control over the gross considering landscapes at the large scale or which have
morphology of the landscape with the operation of exo- long relaxation times (Table 1.3).
genic processes playing a role in shaping the details. An important question raised by the historical approach to
explanation in geomorphology concerns the assumptions we
make about the rate at which processes have operated in the
1.3.6 Explanation in geomorphology past. As we have already seen (see Section 1.2.1) Lyell em-
The vast range of temporal and spatial scales means that no phasized the uniformity of both the rates of geomorphic pro-
one methodological approach to explanation is appropriate cesses through time and the average relief of the Earth's
for all research in geomorphology. At very short time surface (he considered uplift at one place would be roughly
scales we may be concerned solely with the operation of balanced by subsidence at another). Such a view is, of course,
processes and their relationships with presently existing land- untenable in the light of our knowledge of major climatic
forms; at the other extreme we may be aiming to establish a changes during the Earth's recent history. Indeed, some ge-
historical sequence of landform development over a period omorphologists have argued that the great magnitude of these
of millions of years and relating this to long-term changes changes casts doubt on the whole idea of uniformitarianism
in endogenic processes. This distinction involves two aspects and have instead advocated neocatastrophism in its place.
of reality termed by the palaeontologist G.G. Simpson the This rejection of uniformitarianism arises from a confusion
immanent and the configurational. By immanent he meant over the diverse concepts encompassed by the term - a con-
those aspects of reality to do with the inherent properties of fusion which, as we have seen (Section 1.2.1), began im-
the Universe, that is, the physical laws that govern the mediately on the introduction of the concept in the 1830s.
behaviour of matter. By configurational he meant those A valuable attempt to clarify the term has been made by
forms (or configurations) which arise from the operation of S. J. Gould who distinguishes between two fundamentally
the physical laws of the Universe at a particular point in different types of uniformitarianism. The first, which he calls
time. methodological uniformitarianism and which encompasses
The importance of this distinction lies in the contrasting uniformity of law and of process, is the proposition that
approaches to explanation that it implies. When looking at natural laws are invariant - that is, they are constant in

Table 1.3 Relationship between spatial and temporal scale and approaches
to explanation in geomorphology

SPATIAL SCALE

Micro Meso Macro Mega

Predominant Exogenic Primarily Exogenic/ Primarily

genetic exogenic endogenic endogenic
mechanisms
Time required
for adjustment of Short Moderate Long Very long
form to process
Temporal scale Steady Dynamic Cyclic Cyclic
Appropriate
explanatory Immanent Immanent/ Configurational Configurational
basis configurational
 Approaches to geomorphology 17

space and time - and that those observable at the present unvegetated terrain was sufficiently high for a lowering of
time are sufficient to explain past events. The second, slopes of over 20 mm to be observed in this short period of
which he terms substantive uniformitarianism and which time.
incorporates uniformity of rate and of state, postulates rates For periods longer than a few years other methods have to
of natural processes and material conditions that are be employed to document landform change. Aerial photo-
essentially constant through time. This proposition involves graphy can be particularly valuable in areas where repeated
a claim about the world which can be tested and which we surveys are available. In the U K aerial surveys extend back
now know to be false (in any strict sense), whereas to the 1940s and a similar period of coverage is available
methodological uniformitarianism, or actualism as it is for parts of Europe and North America. This source of
also called, is a statement of scientific method and is a information can be especially valuable for tracking landscape
fundamental element of any attempt to provide scientific features such as those associated with landsliding where
explanations of how landscapes have changed through time. significant landform changes may take several decades but
in which the rate of modification varies enormously through
time. Satellite remote sensing is now beginning to provide
1.4 Methods of analysis an important additional means of monitoring landform
A battery of techniques and instruments are now available change, especially where extensive or remote areas are
to monitor the day-to-day operation of a wide range of being studied.
geomorphic processes, such as the gradual downslope move- Topographic maps provide another valuable source for
ment of debris on a slope, the transport of sediment in a documenting landscape change. In some regions accurate
river or the movement of ice at the bed of a glacier. These topographic surveying extends well back into the nine-
methods of data acquisition, which include laboratory ex- teenth century and sequences of maps have been used to
periments as well as field measurements, have generated a reconstruct various landform changes including those af-
wealth of data on short-term landform change, but we can fecting alluvial channels and coastal features. It is neces-
rarely apply these results directly to the problem of long- sary to use such sources with care, however, as some early
term landscape change. Primarily this is because there are surveys may not be sufficiently accurate. In rare instances a
almost invariably changes in the magnitude-frequency re- particular landscape-forming event may be anticipated and
lationships of geomorphic processes in the long term. valuable measurements recorded before a significant change
Although there are now techniques which can be applied to in form occurs. A notable example is provided by the fall of
estimate average denudation rates over millions of years, Threatening Rock in New Mexico, USA in 1941. The
indirect strategies have to be adopted if we are to determine progressive movement of this vertical column away from a
how the form of the landscape has changed over long periods cliff face had been monitored over several years prior to it
of time. We can do this in two ways; by space-time eventually toppling over.
substitution, where variations in form over space are inter-
preted in terms of changes through time, or by simulating
1.4.2 Space-time substitution
landform changes either mathematically or through the use
of hardware models. Space-time substitution in landform analysis was pioneered
by Charles Darwin in the testing of his hypothesis of coral
reef formation (see Section 17.6.3). Barrier reefs, fringing
1.4.1 Direct observations
reefs and atolls occurring at various locations in the world's
With a few rare exceptions direct observations of changes oceans were considered by Darwin to represent different
in form are confined to features of limited dimensions over evolutionary stages of island development applicable to any
periods of months or years. Significant changes in form can particular subsiding volcanic peak in tropical waters where
occur ~here readily mobilized unconsolidated sediments, coral growth could occur. Influenced by Darwin's metho-
such as beach sand or alluvium, are subject to frequent and dology, Davis found support for his evolutionary scheme of
intense geomorphic activity. Such changes can be instru- landform change through time in the form of the landscape
mented and monitored over periods of weeks, months or in different localities which he considered to represent
years. Occasionally it is possible to observe landform changes particular temporal stages of development.
which, although normally occurring very slowly, under Clearly this approach has its dangers. A researcher
certain conditions take place sufficiently rapidly to produce might, for instance, endeavour to fit landforms in different
measurable changes in a short period of time. In a classic places into an assumed temporal sequence simply to satisfy
study examining slope evolution in the clay badlands at a preconceived notion of how such landforms change
Perth Amboy, New Jersey, USA, S. A. Schumm measured through time, even though other sequences of change might
the depth of erosion on slope profiles over a period of ten be equally justified by the evidence. The essential guard·
weeks. The rate of erosion on this impermeable and largely against this kind of erroneous reasoning is a sufficiently
18 Introduction

specific model of landform change linked to a causal there is unambiguous evidence for the sequential develop-
mechanism which indicates particular changes in form. In ment of adjacent landform features. This might occur
Darwin's case his causal mechanism was the growth of coral where, for example, the gradual downstream migration of a
as the volcanic substrate subsided; no other explanation meander leads to the progressive elimination of active
could so adequately account for the different forms of channel erosion at the foot of a meander bluff (Fig. 1.11).
oceanic islands he observed. A second problem is the In this case it is possible to observe in the spatial sequence
danger of assuming that a temporal sequence exists simply the progressive change in slope form after the cessation of
because spatial variablity in form is evident. In reality active removal of basal debris by the river.
spatial variations in landforms may arise simply from ran- Some geomorphologists have attempted to set the pro-
dom fluctuations around an equilibrium form and this cedure of space-time substitution in an apparently more
possibility must be eliminated before temporal sequences rigorous framework by invoking the ergodic hypothesis.
are proposed. A third difficulty is that factors other than This notion was originally applied in the field of statistical
time may be responsible for systematic variations in form mechanics and proposes that sampling in space is the
over space. The form of a slope, for instance, may be equivalent of sampling through time. In order for this
significantly related to lithological controls which may vary assumption to be valid the statistical distribution of objects
spatially in a consistent fashion. or events over space and through time must be the same. If
Space-time substitution is most safely applied where this is the case then the probabilities of sampling a property

Fig.l.ll Sequence of slope profiles along Port Hudson bluff on the Mississippi River in Louisiana, USA (A), and map showing
location ofprofiles (B). The entire bluff segment was being undercut by the Mississippi River in 1722 since when the channel has
shifted downstream about 3 km, being at the location ofprofile 36 in 1849, 34 in 1883, 25 in 1909 and 23 in 1941. Profiles 2-7 in (A)
are being actively undercut by the river; profiles 10-14 are undercut during high flows only; profiles 20-36 are subject to basal
aggradation. The profiles, which show a decline in mean slope angle from 44 to 20°, represent a temporal sequence reflecting changes
in basal conditions. (From D. Brunsden and R. H. Kesel (1973) Journal of Geology 81, Figs. 4 and 6, pp. 581 and 584.)
 Approaches to geomorphology 19

MISSISSIPPI b

"'v.£/i

Based on 1963 U.S.G.S. topographic

map (Port Hudson Quadrangle)

-50- Contours in feet ~

2- Location of slope profiles
1- 9 High Intensity lone
10-18 Backwater at base }
Slope base covered Intermediate Zone
19 - 24 at flood stage only )
25-38 Low Intensity Zone
c:::. Depressions and Lakes

Mile
I
1 Kilometre

of the landscape through tIme and over space are inter- dictions of the morphological changes ansmg from the
.changeable. It is clear that there are few instances in operation of well-defined processes. This, of course, is a
geomorphology where such an assumption can be shown to familiar notion to earth scientists in general who are con-
be justified. Space-time substitution as an approach to cerned with the interpretation of geological structures or
understanding the way landforms change over periods of sedimentary bodies which they are often forced to interpret
time beyond those accessible to direct observation must, with reference to examples in different stages of develop-
.therefore, ~e based in most circumstances on specific pre- ment in different localities.
20 Introduction

1.4.3 Simulation x, X2 Uniform spacing at origin

An alternative, but also complementary, approach to deter-

mining the nature of landscape change through time is
through the use of simulation models of which there are
essentially three types: hardware models, analogue models
and mathematical models.
Hardware models, of which laboratory flumes are perhaps
the best known example, are frequently used to study the
operation of specific geomorphic processes, such as the
transport of sediment by wind or water; but they can also
be used to explore the sequence of landform adjustments that
arise as a result of changes in external variables influencing
these processes. An excellent example of this kind of
application of hardware models is the assessment of the
effects of a lowering of base level on an artificial channel
created in a flume.
The major problem with hardware models is that of
scale. Such models are almost invariably constructed at a
size much smaller than reality and usually the aim is to
accelerate specific geomorphic effects under controlled con-
ditions so that substantial changes in form can be observed
over a relatively short period of time. Unfortunately, though,
changes in scale affect the relationships between the various
properties of hardware models and the real world in different
ways. Scale ratios for length are different for those involv-
ing velocities and acceleration, while properties such as mass
and inertia which are critical in processes such as sediment
transport are also not simply a linear function of length. A
partial solution to this problem is to maintain the correct
scaling for a particular variable of interest by distorting the
scale relationships of other variables which are not of specific Fig. 1.12 The development of rills on a hillslope simulated
interest. using a stochastic (random) model. The model specified that in
each time increment the 'random walk' could proceed forward at
A second type of simulation is carried out using analogue any angle, but could not go backward. Note the equal spacing
models. These involve the replacement of materials which between the points of origin (X/-X n ) for each rill. (After L. B.
occur in the real world with other materials which enable the Leopold, M. G. Wolman and J. P. Miller (1964) Fluvial Processes
effects of processes to be more readily observed. Examples in Geomorphology. W. H. Freeman, San Francisco, Fig. 10-4,
p.416.)
include the use of clay to synthesize the effects of glacier
flow, or the development of fault patterns. Such models have
now been largely replaced by mathematical models which factors, the net effects of which can be adequately treated
represent geomorphic phenomena and the relationships as random (Fig. 1.12).
between them in terms of mathematical expressions. Given the complexity of most geomorphic processes, the
These mathematical models constitute the third kind of construction of mathematical models necessitates a good
simulation model. Deterministic mathematical models are deal of abstraction and selection, and the geomorphic
based on exact relationships between independent (causal) phenomena considered are almost invariably represented in
variables and dependent (response) variables, such as those a highly idealized fashion. Landform properties and pro-
that predict the way in which a slope profile is transformed cesses must be strictly defined if they are to be represented
over time from some initial form. By contrast stochastic mathematically. Mathematical models simplify the com-
mathematical models incorporate a random component plexity of form existing in the real world and in so doing
which allows different possible outcomes to arise from a enable us to predict how landforms will change through
given set of initial conditions. One of the most widely applied time. How successful such predictions are depends on how
types of stochastic model is that used to simulate the good a representation of the real world our model is. The
development of stream networks where the growth of indi- most accurate simulations of landform change are generally
vidual channel links in nature is affected by numerous minor those which treat relatively simple geomorphic systems, and
 Approaches to geomorphology 21

which are based on a detailed knowledge of the geomor- endogenic mechanisms comes primarily from geothermal
phic processes involved. heat, although small contributions are also made by tidal
There are two major problems in using mathematical energy generated by the gravitational attraction of the Sun
models to simulate landform change over time. One is and the Moon and by rotational energy derived from the
equifinality - different models may generate similar results momentum of the Earth's rotation. The ultimate sources of
and unless the geomorphic processes and landforms involved energy for exogenic processes are the potential energy
are very well understood it may not be possible to decide arising from the height of material above base level, and
which version most adequately describes the processes that that proportion of solar radiation received by the Earth.
are operating. The second problem is that we need to have a
detailed knowledge of the actual changes in form that occur
1.5.1.1 Internal energy
in the landscape in order that mathematical models pre-
Evidence of the Earth's internal energy is provided by the
dicting particular changes may be tested. This is an often
geothermal heat flow which can be measured, and in some
neglected but crucially important issue.
cases observed, at the surface. Volcanoes, of course, represent
local areas of higher than average heat flow, but the Earth's
1.5 Endogenic and exogenic factors internal heat is also the ultimate source of energy for
virtually all tectonic processes and the associated horizontal
1.5.1 Sources of energy
and vertical movements of the crust. Although the energy
The processes that shape the world's landscapes are powered involved in the various endogenic mechanisms is not
by two major sources of energy (Fig. 1.13). The energy for known, the present rate of heat flow to the surface is fairly

SOLAR RADIATION INTERCEPTED

BYTHEEARTH
17.8x 10" W

t
SOLAR ENERGY TO ATMOSPHERE
12.3 x 10"W ...
~
HYDROLOGICAL
CYCLE
... FLUVIAL AND
GLACIAL
4x10"W PROCESSES

t ""t"l
"'><
00
r'lC'J
KINETIC ENERGY (WIND)
7x10 14 W
...
~
AEOLIAN t"lt"l
~~

~
PROCESSES
~r'l
COASTAL WAVES
AND CURRENTS
5x10"W - COASTAL

OCEAN TIDES 2.2x 10"W / I PROCESSES

t TRANSFER OF
HEAT TO SURFACE ... VOLCANISM

ROTATIONAL ENERGY OF
THE SOLAR SYSTEM t ""t"l
",2

~
2.9x 10"W GEOTHERMAL HEAT 08
SEISMICITY QC'J

•
FLOW "'t"l
"'2
t"l_
"'r'l
Global36xl0"W
Continents 11 x 10" W
~

~~
SOLID EARTH TIDES Oceans 25 x 10" W
0.4xI0"W
TECTONICS

+
EARTH COOLING
AND RADIOACTIVE
DECAY

Fig.1.13 Estimated energy flows relevant to various geomorphic processes (data from various sources).
22 Introduction

well established. It can be measured from the increase in these latitudes. The resulting latitudinal temperature gradient
temperature with depth (averaging around 20-30 °C lan-I), gives rise to the general circulation of the atmosphere which
although allowances must be made for the thermal con- contributes, along with ocean currents, to the redistribution
ductivity of the rock. Heat flow rates within the continents of heat from the equator towards the poles and is the major
average around 56mWm-2 and range up to 200mWm-2 factor in the climatic zonation of the Earth. At the regional
in areas of active volcanic activity. In the ocean basins the scale marked temperature differences are generated by th€'
average is about 78 mW m-2 , but rates as high as 250 mW m-2 contrasting thermal properties of the continents and oceans,
have been recorded. These variations in heat flow provide with the higher heat capacity of the oceans helping to moder-
important clues as to the nature of the Earth's interior and ate in coastal areas the extremes of temperature charac-
the operation of endogenic processes. teristic of continental interiors. At a local scale altitude
The major source of geothermal heat is the radioactive becomes a primary factor affecting the heat budget. With
decay of the long-lived isotopes of uranium, thorium and increasing elevation there is a progressively greater heat loss
potassium. About 83 per cent of the heat flow to the Earth's through long-wave radiation from the Earth's surface. This
surface is attributable to this process, the remainder being leads to an overall decrease in mean temperature but an in-
provided by the continued cooling of the Earth which has crease in diurnal range.
been proceeding since its formation some 4.6 Ga ago.
Because the half-lives of the major heat-producing isotopes
are in the range 109-10 10 a the supply of heat from radio-
1.5.2 The hydrological cycle
active decay has been more or less constant over the past
several hundred million years. Eventually, though, this heat The hydrological cycle can be conceived as a system of
supply will gradually diminish and the endogenic processes storages between which water is transferred. The oceans
arising from it will become less active and ultimately cease. represent by a considerable margin the largest storage, but
Here, then, we have a system which, although in a' steady ice sheets and glaciers account for a significant proportion
state for a geologically significant period of time, is in fact, of fresh water (Fig. 1.14). Of more geomorphic signifi-
in the very long term, moving towards a decay equi- cance are the magnitudes of the transfers between storages.
librium. Approximately 517 x 103km3 of water is annually
evaporated and reprecipitated over the globe as a whole;
1.5.1.2 Solar radiation this is roughly equivalent to a layer of water 1 m thick over
Solar radiation provides an enormous source of energy, but the entire surface of the Earth. Water is transferred to the
only a very small proportion of this is utilized in the operation atmosphere as a vapour by evaporation from the oceans and
of geomorphic processes. The upper atmosphere intercepts a combination of evaporation and transpiration (evapo-
about 17.8 x 10 16 W of largely short-wave radiant energy transpiration) from the continents, and is returned to the
from the Sun of which about 30 per cent is immediately surface in a liquid or solid state as various forms of
reflected back into space (Fig. 1.13). The remainder heats precipitation (Fig. 1.14). The significant excess of pre-
the atmosphere and the surface and generates a mean global cipitation over evapotranspiration on the continents (around
surface temperature of about 15°C. Of the 12.3 x 10 16 W of 40 per cent) means that large amounts of water are returned
solar energy received by the atmosphere a significant to the oceans as surface runoff. Although only 0.0001 per
amount (about 33 per cent) drives the hydrological cycle cent of all water (and 0.004 per cent of fresh water) is to be
(see Section 1.5.2), the continuous movement of water in found in rivers at anyone time, runoff is by far the most
its gaseous, liquid and solid states between and within the important element of the hydrological cycle in terms of
atmosphere, oceans and landsurface. Latent heat is absorbed landform genesis.
by the vaporization (evaporation) of water from the surface In addition to transfers between major storages there are
of the oceans and continents and through transpiration (the movements within them (that is, transfers between sub-
loss of water vapour from plant cells) by vegetation. A storages). Most precipitation, for instance enters the soil
small proportion (about 1 per cent) is converted into kinetic water or ground water storage before returning to the oceans
energy and powers the circulation of the air in the atmos- as surface runoff. The rate at which water moves through
phere (winds) and water in the oceans (ocean currents). A the hydrological cycle also varies significantly. Water which
further minute, but highly significant, proportion (about 0.1 becomes part of the shallow ground water storage may
per cent) is consumed in photosynthesis (the fixation of remain there for several years while water entering deep
radiant energy by plants). ground water storage may be removed from the active
The receipt of solar radiation varies over the Earth's circulation of the hydrological cycle for hundreds of
surface both temporally and spatially. Taking a global view thousands of years. Similarly, water frozen into large ice
there is an excess of incoming over outgoing radiation be- sheets may travel hundreds of kilometres over tens of
tween about latitude 40° Nand S and a deficit polewards of thousands of years before eventually returning to the ocean.
 Approaches to geomorphology 23

HORIZONTAL ATMOSPHERIC
TRANSFER OF WATER
VAPOUR

r------.,--------l9

EVAPORATION PRECIPITATION
FROM ON
OCEANS CONTINENTS
PRECIPITATION
IN OCEANS EVAPOTRANSPIRAT10N
RUNOFF TO FROM
OCEANS CONTINENTS

PERCENTAGE OF WATER ON
PERCENTAGE OF ALL WATER THE CONTINENTS

RIVERS 0.004%

ICE SHEETS
OCEANS AND
ATMOSPHERE 0.386%
97.2% GLACIERS
76.71%
LAKESO.61%

GROUNDWATER
22.14%

CONTINENTS
2.8%

Fig. 1.14 The hydrological cycle. The figures representing transfers are percentages of the mean annual global precipitation of
approximately 1000 mm. The proportion of water in the various storages is also illustrated. (Data from various sources.)

1.5.3 Climatic controls assumption of climatic geomorphology which seeks to

establish the nature of landforms associated with distinctive
The wide range of climatic conditions over the Earth's sur- climatic regimes and to identify the specific combinations
face, together with the role of climatic factors in influencing of geomorphic processes which give rise to them. Areas
the nature and. rate of operation of geomorphic processes, has characterized by landforms associated with a particular
led some geomorphologists to suggest that different climates climatic environment are called morphoclimatic zones (or
are associated with characteristic landform assemblages. regions). The term morphogenetic region is also used, but
More specifically, it has been argued that particular cli- this is arguably less appropriate as it implies that climatic
matic conditions can have an effect on geomorphic processes factors alone control landform genesis.
sufficient to outweigh the influence on landform develop- The validity of the assumptions behind climatic geomor-
ment of differences in tectonic setting, rock type and relief, phology have been severely challenged by many geomor-
and over a period of time can generate a distinctive associ- phologists who see little evidence of a close relationship
ation of landforms of regional extent. This is the essential between climate and landform morphology except under
24 Introduction

Table 1.4 The Earth's major morphoclimatic zones (based on various sources)

MORPHOCLIMATIC MEAN ANNUAL MEAN ANNUAL RELATIVE IMPORTANCE OF GEOMORPHIC PROCESSESS

ZONE TEMPERATURE (OC) PRECIPITATION (mm)

Humid tropical 20-30 >1500 High potential rates of chemical weathering; mechanical weathering
limited; active, highly episodic mass movement; moderate to low rates of
stream corrasion but locally high rates of dissolved and suspended load transport
Tropical wet-dry 20-30 600-1500 Chemical weathering active during wet season; rates of mechanical weathering
low to moderate; mass movement fairly active; fluvial action high during wet
season with overland and channel flow; wind action generally minimal but
locally moderate in dry season.
Tropical semi-arid 10-30 300--600 Chemical weathering rates moderate to low; mechanical weathering locally
active especially on drier and cooler margins; mass movement locally active but
sporadic; fluvial action rates high but episodic; wind action moderate to high
Tropical arid 10-30 0-300 Mechanical weathering rates high (especially salt weathering); chemical
weathering minimal; mass movement minimal; rates of fluvial activity generally
very low but sporadically high; wind action at a maximum
Humid mid-latitude 0-20 400-1800 Chemical weathering rates moderate, increasing to high at lower latitudes;
mechanical weathering activity moderate with frost action important at higher
latitudes; mass movement activity moderate to high; moderate rates
of fluvial processes: wind action confined to coasts
Dry continental 0-10 100-400 Chemic'al weathering rates low to moderate; mechanical weathering, especially
frost action, seasonally active; mass movement moderate and episodic; fluvial
processes active in wet season; wind action locally moderate
Periglacial <.v 100-1000 Mechanical weathering very active with frost action at a maximum; chemical
weathering rates low to moderate; mass movement very active; fluvial processes
seasonally active; wind action rates locally high
Glacial <0 0-1000 Mechanical weathering rates (especially frost action) high; chemical weathering
rates low; mass movement rates low except locally; fluvial action confined to
seasonal melt; glacial action at a maximum; wind action significant
Azonal mountain Highly variable Highly variable Rates of all processes vary significantly with altitude; mechanical and glacial
zone action become significant at high elevations

the most extreme climatic contrasts. Suffice it to say here ground temperatures fall below freezing, while permanently
that with the exception of landforms which have very short frozen ground will only develop where the mean annual
relaxation times, most elements of the landscape are likely temperature is at least below O°C (Fig. 1. 16(A)). Wind
to be, to a greater or lesser extent, out of equilibrium with action is most active and widespread in arid regions within
prevailing climatic conditions because of the rapidity and a zone fairly well defined by the 200 mm mean annual
magnitude of global climatic changes that have occurred, isohyet (Fig. 1.16(B)), although wind can be an important
especially over the past 2-3 Ma. Indeed in some regions, process on coasts in all climatic zones. Global variations in
such as central Australia and central southern Africa, there precipitation undoubtedly have some influence on rates of
is generally a very low rate of geomorphic activity and the weathering and fluvial activity, but the frequency of high-
landscape is dominated by relict landforms developed under intensity rainfall events is probably far more significant in
climates quite different from those prevailing now. affecting rates of erosion than overall rainfall amounts (Fig.
Notwithstanding uncertainties about the precise relation- 1.l6(B)).
ship between climate and landform genesis, it is clear that
there are major contrasts in the kind of geomorphic processes
active under certain climatic regimes and that all climatically 1.5.4 Human agency
related processes vary in the intensity with which they Landforms, and the processes that create them, impinge on
operate from one climatic region to another (Table 1.4). human activity in many ways. High magnitude geomorphic
Indeed a number of attempts have been made to produce events, such as large landslides or major floods, become
global maps of morphoclimatic zones, and Fig. 1.15 shows natural hazards if they affect people. Whether this happens
a modified version of the widely cited map by the French depends on the distribution of people and geomorphic
geomorphologists 1. Tricart and A. Cailleux. events in time and space. By examining the landforms
The boundaries between morphoclimatic zones are some- themselves it is often possible to tell which areas are likely
what arbitrary, but some climatic parameters are rather to be at risk from high-magnitude-low-frequency geomorphic
specifically related to the operation of particular geomorphic events, while a knowledge of the conditions that occur
processes. Frost action can obviously occur only where immediately before a major event, such as high-intensity
 Approaches to geomorphology 25

6 2
r"""-<::::::-....;:""",:::::---"o.,o:;;:::----"~-~----T-~-__i:__-+_-_f_-_t'-----~-+-___7"_::....-_7"L-__,;;1"L....-____:::"...,c.-~600 10 km
1~
1000 0 1000
km

~ ITIIIIJ Tropical semi-arid

.,...

~
,

....•••• Humid tropical

.~~ Humid mid-latitude ~~\;(~ Periglacial

1\':';':>1 Tropical wet-dry ~ Dry continental ~-=-=3 Tropical arid • Glacial

• Mountains (altitudinal zonation important)

Fig.1.15 GLobaL distribution of morphocLimatic zones (see TabLe 1.4 for details of climatic parameters and geomorphic processes.)
(Modified from 1. Tricart and A, CaiLLeux (1972) Introduction to Climatic Geomorphology. Longman, London, Map III.)

rainfall preceding a flood, can provide the opportunity to which together constitute the subdiscipline of applied
issue warnings for evacuation. A knowledge of geomor- geomorphology. None the less, it is vital to take account of
phology can also contribute to environmental management the way human activity itself has affected the natural land-
in regions where special account has to be taken of particu- scape. Human agency as a factor in landform genesis began
lar geomorphic phenomena; examples include the impact of with the first fires lit by our ancestors which burnt the
pipeline and road construction on permanently frozen ground vegetation and temporarily left the soil more vulnerable to
in high-latitude environments, and the encroachment of erosion. The impact of human activity on the landscape has
sand dunes on human settlements in arid regions. In some been increasing at an accelerating rate, especially since the
cases an understanding of the factors controlling geomorphic Industrial Revolution and the introduction of modem farming
processes can be directly relevant to engineering problems, techniques.
an example being the application of geomorphic research on Human activity can affect the landscape in two ways; by
weathering to the problem of the deterioration of buildings the direct creation of new landforms, and by an alteration in
in arid regions due to the presence of salts. Finally, a the rate at which geomorphic processes operate. Examples
knowledge of landform genesis can assist in prospecting for of the creation of new features in the landscape include
certain kinds of mineral resources. reservoirs formed by the construction of dams (Fig. 1.17),
This book does not specifically consider these topics and the effects of large-scale strip mining. In the
26 Introduction

6 2
10km

1~
1000 0 1000
km

~ Precipitation intensity on average exeeds 50mm per hour

~ at least once every 2 years
j;.{(\~J Mean annual precipitation < 200 mm 6
10km
2

1~
1000 0 1000
km

Fig.l.16 Global distribution of some key climatic variables: (A) important temperature limits; (B) areas experiencing high
precipitation intensities and regions with low mean annual precipitation. (Based on data in R. Common (1966) in: G. H. Dury (ed.)
Essays in Geomorphology. Heinemann, London 53-81; and B. M. Reich, 1963, Journal of Hydrology 1, 3-28).
 Approaches to geomorphology 27

Fig.1.18 Terraced hillsides in southern Cyprus dating back to

at least the period ofRoman occupation.

are in danger of interpreting the past in terms of an atypical

present.

Further reading
A wide range of material is available for those interested in
the history of geomorphology. An excellent recent survey
is that by Tinkler (1985) while Chorley et ai. (1964)
provide a detailed assessment of developments up to the
turn of the century. The second volume of this work
(Chorley et ai., 1973) is devoted to W. M. Davis but two
further planned volumes have yet to appear. A valuable
study of the early years of geomorphology in Britain is that
Fig.1.17 Oblique aerial view of the western end ofLake Mead
on the Arizona-Nevada border, USA, a large reservoir formed
by Davies (1969) who has also recently presented a sober
through the impoundment of the Colorado River by the Hoover re-evaluation of Hutton's original contribution (Davies,
Dam (visible in the centre of the photograph). 1985). Gould (1987) contains brilliant and highly readable
essays exploring the conceptions of time embodied in the
models of Earth history presented by Hutton and Lyell; this
Appalachian region of the USA strip mining has produced is required reading for those who really wish to grasp the
benches on hillsides formed of rock waste which, it has philosophical framework of landform analysis established
been estimated, extend for a total distance of around in the late eighteenth and early nineteenth centuries. Rudwick
30000 km. The modification of hillslopes is not, however, (1985) is (in spite of its title) also worth consulting for its
only a recent phenomenon, as a visit to many areas in the discussion of the various meanings of uniformitarianism.
Mediterranean region would demonstrate (Fig. 1.18). G. K. Gilbert is the subject of an excellent biographical
Of greater global significance than the direct creation of study by Pyne (1980) emphasizing his methodological
landforms is the effect of human activities on the rates of approach; briefer assessments of Gilbert's work are
geomorphic processes. Agriculture, lumbering and the con- proviaed by Pyne (1975) and Yochelson (1980) while his
struction of settlements can all cause dramatic changes in key monographs are Gilbert (1877, 1914). Other original
land use which in turn may greatly change the susceptibility works of particular importance include: Hutton (1788)
of the landsurface to erosion. Not surprisingly, the extension (contained in Eyles, 1970); Playfair (1802), which is avail-
of activities such as agriculture into new regions over the able as a facsimile edition (White, 1956); Lyell (1830-33);
past 100 years or so has led to some dramatic changes in Agassiz (1840), available in an English tr.anslation
the quantity of sediment being transported by rivers in the (Carozzi, 1967); Davis (1909); and Penck (1924), which is
areas affected. Fluvial systems have, in many instances, also available in translation (penck, 1953). More recent
undergone dramatic and complex adjustments as a result. If developments, with a particular emphasis on fluvial geo-
we are to understand how the natural landscape behaves it morphology are reviewed by Gregory (1985), while Horton's
is vital that these effects are taken into account, otherwise we seminal paper is well worth consulting (Horton, 1945).
28 Introduction

Recommended reading on the various topics considered in forms are presented by Howard (1971) and King and
the discussion of future directions in geomorphology is McCullagh (1971). Mosley and Zimpfer (1978) and Schumm
indicated in the various chapters which develop these et al. (1987) provide examples of the use of hardware
themes, but it is worth mentioning here the thought- models and Anderson (1988) contains detailed presen-
provoking essays by Baker (1986) and Hayden et al. tations of mathematical models.
(1986), the assessment of impact cratering on the Earth by The Earth's internal energy and the present global heat
Grieve (1987), the review of comparative planetary geo- flow are treated at an introductory level by Pollack and
morphology by Sharp (1980) and the excellent case study Chapman (1977) while Williams (1982) outlines the surface
of interpreting planetary landforms in terms of terrestrial energy budget (see especially Chapters 1 and 3). Barry (1969)
analogues contained in Howard et al. (1988) provides a useful summary of the global hydrological
Thorn (1988) provides a useful introduction to concepts cycle, although some of the data quoted have subsequently
in geomorphology (although I feel he severely underplays been revised. The principles of climatic geomorphology are
the importance of endogenic factors in landscape critically summarized by Stoddart (1969) and presented in
development) while Baker (1988), Chorley (1978), Ritter detail by Biidel (1982) and Tricart and Cailleux (1972).
(1988) and Scheidegger (1987) discuss from various There is a large literature on applied geomorphology and
perspectives key principles and methodological approaches. the impact of human activities on the landscape with useful
On specific concepts in geomorphology, Scheidegger (1979) starting points being provided by Coates (1981), Cooke and
considers the' antagonism' between endogenic and exogenic Doornkamp (1974), Craig and Craft (1980), Gregory and
processes and Summerfield (1981) briefly discusses their Walling (1987) and Hails (1977).
relative significance as the spatial scale is altered. Finally, it is appropriate here to mention some of the
Magnitude-frequency relationships are treated in a classic more important general sources of research and reference
paper by Wolman and Miller (1960) and reconsidered by information in geomorphology. The four major journals
Wolman and Gerson (1978). The application of systems covering geomorphic research are Catena, Earth Surface
analysis to landform studies is discussed by Chorley (1962), Processes and Landforms, Geomorphology and Zeitschrift
Chorley an.cl Kennedy (1971) and Huggett (1985). The pro- fur Geomorphologie, although articles on geomorphology
blem of temporal scale in geomorphology is treated from are to be found in a wide range of geography and earth
various perspectives by Cullingford et al. (1980), science journals. The reviews of recent developments in
Montgomery (1989), Schumm and Lichty (1965), Thorn various fields within geomorphology contained in Progress
(1982) and Thomes and Brunsden (1977). The related issue in Physical Geography are particularly useful. Fairbridge
of appropriate conceptions of equilibrium in geomorphic (1968) is still a valuable reference work, but this has
research is considered by Chorley and Kennedy (1971) and recently been supplemented by the less detailed but broader
Schumm (1977) (especially Chapter 1), although it is import- Goudie (1985). Snead (1981) is a useful geomorphic atlas
ant to be aware of the different uses of the terms dynamic while Short and Blair (1986) is an invaluable source of
equilibrium and steady state in the latter in comparison with images of landforms taken from space with detailed accom-
some earlier work. The modem usage of the term uniformi- panying commentaries.
tarianism is critically evaluated by Gould (1965) and Simpson
(1970) provides a detailed discussion which introduces the References
notion of the immanent and configurational aspects of reality. Agassiz, L. J. (1840) Etudes sur les Glaciers. Neuchatel.
There are relatively few general treatments of the problem Anderson, M. G. (ed.) (1988) Modelling Geomorphological
of landform change through time, but Thomes and Brunsden Systems. Wiley, Chichester and New York.
(1977) cover a broad range of issues. Baker, V. R (1986) Introduction: Regional landform analysis. In:
N.M. Short and R W. Blair Jr (eds) Geomorphology From
On methods of analysis Gardiner and Dackombe (1981) Space: A Global Overview of Regional Landforms. NASA,
and Goudie (1981) between them comprehensively cover Washington D C, 1-26.
field and laboratory techniques in geomorphology. Schumm's Baker, V. R (1988) Goologic3.J. fluvial geomorphology. Geological
detailed investigation of the evolution of microscale badland Society ofAmerica Bulletin 100,1157-1167.
topography (Schumm, 1956) and the study by Schumm and Barry, R G. (1969) The world hydrological cycle. In: R J. Chorley
(ed.) Water, Earth and Man. Methuen, London; Barnes and
Chorley (1964) of the fall of Threatening Rock provide Noble, New York, 8-26.
classic examples of the value of direct observations. The Brunsden, D. and Kesel, RH. (1973) The evolution of a
general principles of space-time substitution are thoroughly Mississippi river bluff in historic time. Journal of Geology 81,
reviewed by Paine (1985) while Brunsden and Kesel (1973) 576-97.
provide an example of the application of this approach. A BUdel, J. (1982) Climatic Geomorphology (translated by L. Fischer
and D. Busche). Princeton University Press, Princeton and
useful introduction to simulation modelling is provided by Guildford.
Chorley and Kennedy (1971) and examples of the appli- Carozzi, A. V. (1967) Studies on Glaciers by Louis Agassiz.
cation of simulation modelling to fluvial and coastal land- Hafner, New York.
 Approaches to geomorphology 29

Chorley, R. J. (1962) Geomorphology and general systems theory. Global Overview of Regional Landforms. NASA, Washington
United States Geological Survey Professional Paper 500B. DC., 657-72.
Chorley, R. J. (1978) Bases for theory in geomorphology. In: C. Horton, R. E. (1945) Erosional development of streams and their
Embleton, D. Brunsden and D. K. C. Jones (eds) Geomorphology: drainage basins: hydrophysical approach to quantitative
Present Problems and Future Prospects. Oxford University morphology. Bulletin of the Geological Society of America 56,
Press, Oxford and New York, 1-24. 275-370.
Chorley, R. J., Beckinsale, R. P. and Dunn, A. J. (1973) The Howard, A. D. (1971) Simulation model of stream capture.
History of the Study of Landforms or the Development of Geological Society of America Bulletin 82,1355-76.
Geomorphology. Volume Two: The Life and Work of William Howard, A. D. and Kochel, R. C. (1988) Introduction to cuesta
Morris Davis. Methuen, London; Harper and Row, New York. landforms and sapping processes on the Colorado Plateau. In: A.
Chorley, R. J., Dunn, A. J. and Beckinsale, R. P. (1964) The D. Howard, R. C. Kochel and H. R. Holt (eds) Sapping Features
History of the Study of Landforms or the Development of of the Colorado Plateau NASA SP-491, NASA, Washington,
Geomorphology. Volume One: Geomorphology Before Davis. 6-56.
Methuen, London; Wiley, New York. Huggett, R. J. (1985) Earth Surface Systems. Springer-Verlag,
Chorley, R. J. and Kennedy, B. A. (1971) Physical Geography: A Berlin and New York.
Systems Approach. Prentice-Hall, London and Englewood Hutton, J. (1788) Theory of the earth; or an investigation of the
Cliffs. laws observable in the composition, dissolution, and restoration
Coates, D. R. (1981) Environmental Geology. Wiley, New York of land upon the globe. Transactions of the Royal Society of
and Chichester. Edinburgh 1, 209-304.
Cooke, R. U. and Doornkamp, 1. C. (1974) Geomorphology in King, C. A. M. and McCullagh, M. J. (1971) A simulation model of
Environmental Management: An Introduction. Clarendon a complex recurved spit. Journal ofGeology 79, 22-37.
Press, Oxford and New York. Lyell, C. (1830-33) Principles of Geology (3 vols) Murray,
Craig, R. G. and Craft, J. L. (eds) (1980) Applied Geomorphology. London.
Allen and Unwin, London and Boston. Montgomery, K. (1989) Concepts of equilibrium and evolution in
Cullingford, R. A., Davidson, D. A. and Lewin, J. (eds) (1980) geomorphology: the model of branch systems. Progress in
Timescales in Geomorphology. Wiley, Chichester and New Physical Geography 13, 47--66.
York. Mosley, M. P. and Zimpfer, G. L. (1978) Hardware models in
Davies, G. L. (1969) The Earth in Decay: A History of British geomorphology. Progress in Physical Geography 2, 461-83.
Geomorphology 1578-1878. MacDonald, London. Paine, A. D. M. (1985) 'Ergodic' reasoning in geomorphology:
Davies, G. L. H. (1985) James Hutton and the study of landforms. time for a review of the term? Progress in Physical Geography
Progress in Physical Geography 9, 382-9. 9,1-15.
Davis, W. M. (1909) Geographical Essays (edited by D. W. Penck, W. (1924) Die Morphologische Analyse: Ein Kapital der Phy-
Johnson). Ginn, Boston. sikalischen Geologie. Engelhorn, Stuttgart.
Eyles, V. A. (1970) James Hutton's 'System of the Earth, 1785; Penck, W. (1953) Morphological Analysis of Landforms (translated
Theory of the Earth, 1788; Observations on Granite, 1794'. by H. Czech and K. C. Boswell) Macmillan, London.
Hafner, Darien. Playfair, J. (1802) Illustrations of the Huttonian Theory of the
Fairbridge, R. W. (ed.) (1968) Encyclopedia of Geomorphology. Earth. William Creech, Edinburgh.
Reinhold, New York. Pollack, H. N. and Chapman, D. S. (1977) The flow of heat from
Gardiner, V. and Dackombe, R. V. (1981) Geomorphological the earth's interior. Scientific American 237(2), 60-76.
Field Manual. Allen and Unwin, London and Boston. Pyne, S. (1975) The mind of Grove Karl Gilbert. In: W. N. Melhorn
Gilbert, G. K. (1877) Report on the Geology of the Henry and R. C. Flemal (eds) Theories of Landform Development.
Mountains. United States Department of the Interior, Allen and Unwin, London and Boston, 277-98.
Washington DC. Pyne, S. J. (1980) Grove Karl Gilbert: A Great Engine of
Gilbert, G. K. (1914) The transportation of debris by running Research. University of Texas Press, Austin.
water. United States Geological Survey Professional Paper 86. Ritter, D. F. (1988) Landscape analysis and the search for
Goudie, A. (ed.) (1981) Geomorphological Techniques. Allen and geomorphic unity. Geological Society of America Bulletin 100,
Unwin, London and Boston, 160-71.
Goudie, A. (ed.) (1985) The Encyclopaedic Dictionary of Physical Rudwick, M. J. S. (1985) The Meaning of Fossils: Episodes in the
Geography. Blackwell, Oxford and New York. History of Palaeontology (2nd edn) University of Chicago
Gould, S. J. (1965) Is uniformitarianism necessary? American Press, Chicago and London.
Journal ofScience 263, 223-8. Scheidegger, A. E. (1979) The principle of antagonism in the
Gould, S. J. (1987) Time's Arrow, Time's Cycle: Myth and Earth's evolution. Tectonophysics 55, TI-TlO.
Metaphor in the Discovery of Geological Time. Harvard Scheidegger, A. E. (1987) The fundamental principles of
University Press, Cambridge and London. landscape evolution. Catena Supplement 10, 199-':210.
Gregory, K. J. (1985) The Nature of Physical Geography. Arnold, Schumm, S. A. (1956) Evolution of drainage systems and slopes
London. in badlands at Perth Amboy, New Jersey. Bulletin of the
Gregory, K. J. and Walling, D. E. (eds) (1987) Human Activity and Geological Society ofAmerica 67, 597-646.
Environmental Processes. Wiley, Chichester and New York. Schumm, S. A. (1977) The Fluvial System. Wiley, New York and
Grieve, R. A. F. (1987) Terrestrial impact structures. Annual London.
Review of Earth and Planetary Sciences 15, 245-70. Schumm, S. A. and Chorley, R. J. (1964) The fall of Threatening
Hails, J. R. (ed.) (1977) Applied Geomorphology. Elsevier, Rock. American Journal ofScience 262, 1041-54.
Amsterdam and Oxford. Schumm, S. A. and Lichty, R. W. (1965) Time, space and
Hayden, R. S., Blair, R. W. Jr, Garvin, J. and Short, N. M. (1986) causality in geomorphology. American Journal of Science 263,
Global geomorphology: outlook for the future. In: N. M. Short 110-119.
and R. W. Blair Jr (eds) Geomorphology From Space: A Schumm, S. A., Mosley, M. P. and Weaver, W. E. (1987)
30 Introduction

Experimental Fluvial Geomorphology. Wiley, New York and Thorn, C. E. (1988) An Introduction to Theoretical Geomor-
Chichester. phology. Unwin Hyman, Boston and London.
Sharp, R. P. (1980) Geomorphological processes on terrestrial Thomes, J. B. and Brunsden, D. (1977) Geomorphology and Tlme.
pfanetary surfaces. Annual Review of Earth and Planetary Methuen, London.
Sciences 8, 231-61. Tinkler, K. J. (1985) A Short History of Geomorphology. Croom
Short, N. M. and Blair, R. W. Jr (eds) (1986) Geomorphology From Helm, London.
Space: A Global Overview of Regional Landforms. NASA, Tricart, 1. and Cailleux, A. (1972) Introduction to Climatic
Washington D C. Geomorphology (translated by C. J. Kiewiet de Jonge).
Simpson, G. G. (1970) Uniformitarianism. An inquiry into Longman, London.
principle, theory, and method in geohistory and biohistory. In: White, G. W. (1956) 'Illustrations of the Huttonian Theory of the
M. K. Hecht and W. C. Steere (eds) Essays in Evolution and Earth', by John Playfair. University of Illinois Press, Urbana.
Genetics in Honor of Theodosius Dobzhansky. Appleton- Williams, P.1. (1982) The Surface of the Earth: An Introduction
Century-Crofts, New York, 43-96. to Geotechnical Science. Longman, London and New York.
Snead, R. E. (1981) World Atlas of Geomorphic Features. Wolman, M. G. and Gerson, R. (1978) Relative scales of time and
Reinhold, New York. effectiveness of climate in watershed geomorphology. Earth
Stoddart, D. (1969) Climatic geomorphology: review and re- Surface Processes 3, 189-208.
assessment. Progress in Geography 1, 159-222. Wolman, M. G. and Miller, J. P. (1960) Magnitude and frequency
Summerfield, M. A. (1981) Macroscale geomorphology. Area 13, of forces in geomorphic processes. Journal of Geology 68,
3-8. 54-74.
Thorn, C. E. (ed.) (1982) Space and Time in Geomorphology. Allen Yochelson, E. L. (ed.) (1980) The scientific ideas of G. K. Gilbert.
and Unwin, Boston and London. Geological Society ofAmerica Special Paper 183.
Part II
Endogenic processes and
landforms
 2
Global morphology and
tectonics

2.1 Global morphology

.lOOt
·SOm

2.1.1 The ge~id • SOm

.100m ?O'N
The Earth is only approximately a sphere. As a result of the
centrifugal force of rotation it bulges at the equator and its
polar radius (6378 kIn) is 21 kIn shorter than its equatorial
radius (6397 km); thus the Earth is more accurately
described as an oblate spheroid. Even this description, how-
ever, is not completely accurate as inhomogeneities in the
distribution of mass in the Earth's interior produce further
small but measurable irregularities on its surface. These
irregularities have been determined with great precision
over recent years through the very accurate measurement of 20'N
deviations in the orbits of artificial satellites.
The shape of the Earth determined in this way is known 10'N

as the geoid and is represented by the surface defined by

mean sea level over the oceans and the extension of sea
level along imaginary canals across the continents. As is
evident in Figure 2.1 this surface shows many irregularities 10'5

compared with a simple oblate spheroid; there is, for

instance, a bulge of 76 m near New Guinea and a
depression of some 104 m to the south of India. The sig-
nificance of the geoid for the operation of geomorphic Xl's

processes is that it defines the ultimate base level for

denudation; any change in the geoid would therefore cause
consequential changes in base levels. This is an important
issue in studies of sea-level change and is examined further SO'S

in Chapter 17.

l00m

$
0 • SOm
Fig.2.1 Geodetic sea level profiles at intervals of 10 latitude.
The horizontal latitudinal lines separate positive geoid values
(above)from negative geoid values (below). (From N. -A. Morner -so m
(1976), Journal of Geology 84, Fig. 1, p. 124,) -lOOm
34 Endogenic processes and landforms

30 10,----------------------,-10
Sea Highest mountain
level
~

~
0 Continental
~ platform
co Seamounts
~
co 20
Q)
u
r-

I
Mid-oceanic ridges
Continental slope
_ _5'::::-:-1'---------------.::S~ea~l=-'ev~el_+o
______ -=_':=_.",

-t :>
~
- Oceanic
'"
b
Continental
shelf
trench
Minimum Maxim um
>- 10 Ocean basin
5
u
c elevation r- elevati on
floor
Q)

~ n- ~
:>
cr
u..
1:

0
-10
~

-5 0
rh. 5 10 o 10 20 30 40 50 60
Greatest ocean depth

70 80 90
10

100
Elevation with respect to sea level (km)
Cumulative Area as a % of Total Surface
Area of the Earth Above Each Level
Fig. 2.2 Proportionate areal distribution of the solid sUiface of
the Earth between successive elevations. Note the two peaks Fig. 2.3 Global hypsometric curve indicating the main
representing the ocean basins and the continental platforms. morphological features characterizing the Earth's surface at
(Modified from P.l. Wyllie 1976, The Way the Earth Works. different elevations. (Modified from P. J. Wullie (1976) The Way
Wiley, New York, Fig. 3-11 (a), p. 38. the Earth Works. Wiley, New York, Fig. 3-11 (b), p. 38.).

necessary to identify the major components of global

2.1.2 Global hypsometry morphology that give rise to the curve (Fig. 2.4). Beginning
If the overall distribution of height of the Earth's solid with the more familiar form of the continents we can
surface above and below sea level is plotted as a frequency distinguish between the continental platforms formed by
distributiotl. its most obvious feature is the bimodal form of plateaus and lowlands, and the major linear mountain
the curve (Fig. 2.2). The two peaks in the distribution systems known as orogenic mountain belts, or simply
represent the continental platforms and the ocean basins. orogens. One major mountain system cuts across central
Between these two maxima there is a lower frequency of and southern Eurasia extending from the Alps in the west,
depths representing the more shallow regions of the oceans. through the Himalayas to the mountains of western China.
At either end of the histogram are thin tails indicating that Another, formed by the Andes and the North American
very little of the Earth's solid surface is found more than 5 Cordillera, runs along the entire western margin of the
km above, or 6 km below, mean sea level. The extremes of Americas. Other older and more subdued ranges include
the distribution mark the greatest elevation of 8.8 km above the Appalachians in eastern North America and the Urals in
sea level (Mount Everest) and the greatest depression at western Eurasia. Extensive areas within continental
more than 11 km below sea level (the Marianas Trench in platforms are formed of basement, a complex of meta-
the western Pacific Ocean). morphic and igneous rocks of Palaeozic or Precambrian
Other features are more clearly revealed if the data are age. In some localities, notably East Africa, continental
plotted as a cumulative frequency distribution indicating platforms are traversed by rift valleys consisting of linear
the proportion of the surface above any given height (Fig. troughs formed by the subsidence of crust between parallel
2.3). This produces a hypsometric curve which shows the systems of faults.
elevations of those regions lying between the continental The submarine extension of a continent is called the
platforms and the ocean basins and suggests that the true continental shelf. This is bordered by a continental slope
break between the continents and oceans lies not at present which inclines at an angle of around 3-6° towards the
sea level but some 200 m or so lower. As defined by ocean basin and which is separated from it by a continental
present mean sea level the continents account for 29 per rise (Fig. 2.5). The ocean basins themselves are traversed
cent of the Earth's surface area and the oceans 71 per cent, by a vast system of mid-oceanic ridges up to 1000 km in
but if the break of slope represented on the global width and tens of thousands of kilometres in length. On
hypsometric curve is used the figure for the continents is average they rise some 2 km above the surrounding ocean
nearer 37 per cent. floor but in a few localities, notably at Iceland, they reach
above sea level. Although the surface of the ocean basins is
relatively uniform it is punctuated in places by volcanoes.
2.1.3 Major morphological features Some of these rise above sea level, but the vast majority
Although the global hypsometric curve provides a valuable are submarine features known as seamounts, or if they
statistical summary of the Earth's relief as a whole it is also are flat-topped, guyots. The greatest ocean depths are
 Global morphology and tectonics 35

6 2
10km
1~
1000 0 1000
km

Mid-oceanic ridges 1~'Wf~1 Mountain belts of Cenozoic age

~ Partially erotled mountain belts of Mesozoic
' - Island arcs ~ and Late Palaeozoic age
Lowlands and plateaus
D (continental platform)
1m Highly eroded mountain belts of
Early Palaeozoic age
,,'\. Rift valleys

Fig.2.4 Major morphological features of the Earth.

75°W 150 E
Pacific Ocean South America South Atlantic Ocean Africa
10 Mountain 10
bett .' - An..d,es
5 ~ Continental 5
~~~IC ./t~:j~l!l! ! :~: :,. . . Plat\rm..... ...

13,000 14,000 15,000 16,000 17,000 18,000 19,000 20,000 km

1,000 2,000
Vertical exaggeration X 100 °
, ,
km
,

Fig. 2.5 Schematic cross-section of major morphological features from western South America to the west coast ofAfrica at the
latitude of the Tropic of Cancer. Note that the curvature of the Earth is not shown. (Modifiedfrom P. J. Wyllie (1976) The Way the
Earth Works. Wiley, New York, Fig. 3-10, pp. 36-7.)
36 Endogenic processes and landforms

200S nOw
East Peru- Continental
Pacific Chile rise and Brazilian
Ocean basin

--------- ---- ---------

Trench slope Andes
---~-./'------
----------
Shield

A 3
1

7
Horizontal scale 1'5,000,000 Vertical scale 1, 2,000,000 Vertical exaggeration 25X

B F=======~---~=-=~---------;[~:
Horizontal and Vertical scale 1'5,000,000

Fig.2.6 Cross-section o/the Andes Mountains and the Peru-Chile Trench drawn with a vertical exaggeration 0/25 times (A), and
with the same horizontal and vertical scale (B).

achieved in oceanic trenches which are to be found both S-waves, are shear waves and they transmit energy by an
along some continental margins, notably western South 'up and down' motion of particles normal to the direction
America (Fig, 2.6), and in association with island arcs of propagation, The density and elastic properties of rock
which are usually located close to the edges of continents. are both important in affecting the passage of seismic
The relief exhibited by the highest mountains and the waves, P-waves can be transmitted through any material
deepest oceanic trenches, although apparently spectacular, (sound waves are of this type), but S-waves can only be
is in fact trivial when compared with the dimensions of the transmitted through solids. From the response of these
Earth as a whole, The maximum difference in altitude waves as they travel through the Earth it is possible to build
associated with the Earth's solid relief is about 20 km, or up a general picture of the properties of its interior.
only 0.31 per cent of the planet's mean diameter. This is Two major discontinuities in the variation of the velocity
why substantial vertical exaggeration must nearly always of earthquake waves with depth show that the Earth's
be employed in the depiction of extensive areas of relief in interior can be divided into three zones; the crust, the
cross-section (Fig. 2.6). Only with the most dramatic mantle and the core. The boundary between the core and
topography can large-scale relief forms be clearly drawn the mantle is very sharp and is located at a depth of about
with identical horizontal and vertical scales, 2900 km, The mantle-crust boundary is marked by the
Mohorovicic discontinuity (or the Moho as it is usually
.2.2 Earth structure abbreviated), Its depth below the continents averages
around 35 km, but below the ocean basins it is only
5-10 km. The crust, therefore, is an extremely thin surface
2.2.1 Seismic evidence
layer representing only about 0.5 per cent of the Earth's
Studies of the way in which earthquake waves are trans- radius.
mitted through the Earth have shown that its interior has a
concentric structure. When an earthquake occurs the energy
2.2.2 Mantle, asthenosphere, crust and lithosphere
released is transmitted in wave form in all directions. The
velocities of earthquake-(seismic) waves, together with the The mantle is mainly solid and appears to comprise
paths they follow, are determined by the properties of the minerals of high density, rich in magnesium and iron. The
material through which they pass. Two major types of wave boundary between the mantle and crust represents a
are produced and these travel at different speeds. Primary, significant change in rock density but there may also be
or P-waves, are the fastest. They are compressional waves, differences in chemical composition, Detailed studies of the
the energy being transmitted by an initial compression of velocity of seismic waves in the upper mantle show that
particles which is then passed on to adjacent particles. This after the marked increase corresponding to the Moho at the
produces a sequence of zones of compression and expan- crust-mantle boundary their velocity gradually becomes
sion which travel away from the source. Secondary, or greater to a depth of around 100 km, where a small, but
 Global morphology and tectonics 37

Table 2.1 Summary of the properties of the crust, mantle, lithosphere and asthenosphere

THICKNESS VELOCITIES OF MEAN

P-AND S-WAVES DENSITY BEHAVIOUR
(kms- I ) (kgm-3 )

Continental crust Increase with depth to Continental crust: 2700 Solid to top of low velocity zone.
Mean: 35km 6.6 P-waves Elastic deformation under vertical
-...'"= Min: <30km
Max: 70km
3.8 S-waves
Oceanic crust: 3000
crustal loading.
.........
U but considerable spatial variation ..c
Oceanic crust: c.
5-IOkm '"0
.:
I--- ~ Moho ~
Immediately below Moho: 3320 immediately below
Base of crust to depth 8.5 P-waves Moho to 5600 at base of
of 2900 km 4.8 S-waves mantle

Decrease to minimum of Very weak in low velocity zone .........

7.8 P-waves (depth range 100-300 km). Inelastic ..c
..2:i c..
.
=
~
4.2 S-waves
in low velocity zone between depth
deformation or 'flow'. '"0
...c
of 100 and 300 km .:"'.
<
Below 300 km increase to Below low velocity zone
14.0 P-waves progressively less weak, but
7.5 S-waves convection probably possible deep
at base of mantle into the mantle.

significant, decrease is found which continues to a depth of OCEAN CONTINENT

~/-------~-------
about 300 km (Table 2.1).
CRUST Depth
This layer of attenuated velocity is known as the low (km)
o
velocity zone and the reduction in seismic wave velocities
is considered to be due to partial melting in this region of 100

the mantle. It partly corresponds to the asthenosphere, the 200

ASTHENOSPHERE
plastic-like properties of which permits slow 'flow' of
material in response to forces applied over long periods of 300

time (Fig. 2.7). Lubrication may be provided by melting

~
400
between mineral boundaries, but molten material must only
form a small proportion of the asthenosphere since it is 500

capable of transmitting S-waves.

600
Compared with the mantle, the crust forms a very thin km
'?O 270 300 400
layer. Beneath the oceans it is thought to extend to a depth. To 2900 km 700

of around 10-15 km, that is to only 5-10 km below the

Fig.2.7 Schematic representation of the relationship between
ocean floor. Below the continents the crust is much thicker, crust, mantle, lithosphere and asthenosphere. Reincorporation of
averaging about 35 km but reaching up to 70 km beneath the lithosphere into the underlying mantle is also shown (see
some mountain ranges. Section 2.4.)
Of greatest significance to the creation of macroscale
landforms is the boundary of the asthenosphere with the flow measurements have been employed to map variations
solid mantle above. The crust and the immediately in lithospheric thickness on a global basis (Fig. 2.8). These
underlying mantle located above the asthenosphere in fact indicate generally thin lith?sphere beneath the oceans,
appear to behave as a coherent semi-rigid layer called the especially in the vicinity of mid-oceanic ridges, and a
lithosphere. Two types can be distinguished: oceanic marked thickening in certain continental areas, including
lithosphere is capped by thin oceanic crust, while Antarctica, North Africa, Brazil and north-eastern North
continental lithosphere is capped by much thicker con- America.
tinental crust. The thickness of the lithosphere itself shows
considerable variability and the boundary with the under-
lying asthenosphere is gradational and difficult to define 2.2.3 Gravity anomalies
precisely. In addition to the analysis of seismic data, heat Measurement of variations in the gravitational attraction at
38 Endogenic processes and landforms

Fig.2.8 Global variation in the thickness of the lithosphere estimated from heat flow measurements. (From D. S. Chapman and H. N.
Pollack (1977) Geology 5, Fig. 4, p. 268.)

the Earth's surface has provided much useful information density and thickness. The term isostasy (meaning 'equal
about the characteristics of the crust and the lithosphere. standing') was introduced by the American geologist
The force of gravity measured at the surface depends on the Dutton in 1889 to describe this state of equilibrium. Until
altitude of the location, the gravitational pull of the under- quite recently it was considered that it was the crust that
lying rock and the attraction exerted by any areas of attained equilibrium with respect to the mantle, but it is
higWand near by (this last effect being negligible except now known that isostatic adjustments also involve the rigid
close to large mountain masses). mantle forming the lower part of the lithosphere. Never-
The reference point for the calculation of gravity theless it is differences in the density and thickness of the
anomalies (which are measured in milligals) is an inter- crust that are largely responsible for variations in the
national standard for the gravitational attraction at the isostatic adjustment of the lithosphere as a whole.
Earth's surface which takes into account the effects of the If it is in isostatic equilibrium one section of lithosphere
rotation of the Earth. Correction for altitude alone gives the will stand higher than another because it is of lower density
free-air gravity anomaly. To allow for the mass present in (the crustal density or Pratt model), of the same density
a mountain range above the reference surface the theoreti- but thicker (the crustal thickness or Airy model), or
cal pull of the rock, based on an assumed average rock through a combination of both a lower density and greater
density, is subtracted from the free-air anomaly (or a thickness (Fig. 2.9). Continental lithosphere stands higher
correction is made for the presence of sea water if the point than oceanic lithosphere because continental crust is both
is over the ocean). This gives the Bouguer anomaly which of greater thickness and lower density than oceanic crust.
is the most widely used measure of gravitational deviations The great differences in elevation within the continents are,
over the Earth's surface. in most cases, related to variations in crustal thickness with
areas at high elevations generally being underlain by deep
roots of buoyant crustal rock (Box 2.1).
2.2.4 Isostasy Bouguer anomalies are generally negative on land and
As all but the uppermost part of the mantle appears capable positive over the oceans. This would be expected from
of viscous flow when subject to prolonged stress, the semi- differences in the density and thickness of continental and
rigid lithosphere, capped by continental or oceanic crust, oceanic crust. Bouguer anomalies include a correction for
can be viewed as 'floating' on the underlying astheno- altitude but only on the basis of an average crustal density.
sphere (the most easily deformable viscous part of the The continents, particularly where elevations are high,
mantle). To attain hydrostatic equilibrium the position of exhibit lower than predicted (negative) gravity values
the lithosphere adjusts vertically in accordance with its because of the greater thickness of underlying crust of
 Global morphology and tectonics 39

CRUSTAL THICKNESS MODEL (AIRY) CRUSTAL DENSITY MODEL (PRATT)

Mountain Mountain
Range Range
Continental Continental
Platform Platform
- - Ocean Basin - f- Ocean Basin
Low
I Med i u m
Hi g h

+ +
] J _Bouguer AnomaI Y_[

Mountain
Range
Continental
Platform
Ocean Basin
L w
I
0
Hig h
I
...........- - -.....
4 - - - . . , , / "'--.....- -......
4--..",#
Continental Crust Oceanic Crust
2700 kg m-3 3000,kg m-3

COMPOSITE MODEL
Fig. 2.9 A schematic representation of three models of isostatic equilibrium. Although differences in the density and thickness of the
crust are illustrated, the lithosphere as a whole is involved in isostatic compensation.

relatively low density. Higher than predicted (positive) covered much of North America and northern Europe in the
gravity values occur over the oceans because they are recent past had largely disappeared by 10000 a BP, the
underlain by thin and relatively dense crust. landsurface in these regions is still rising rapidly in
When corrections are made to take account of these response to the removal of this load.
differences in crustal density the magnitude of gravity A third reason for isostatic anomalies is that there are
anomalies is significantly reduced. Remaining discrepancies dynamic forces present in the sub-lithospheric mantle
are termed isostatic anomalies, but why should such which are capable of actively dragging down or pushing up
: anomalies exist when the lithosphere is apparently free to the lithosphere. These forces also play a key role in the
attain isostatic equilibrium with respect to the asthenosphere? , large horizontal movements experienced by the lithosphere
One reason is that the lithosphere is not divided into which we will be examining in detail later in this chapter.
small discrete blocks able to move freely up and down with
, respect to each other. The lithosphere possesses a certain
degree of rigidity so that the mass of any load (such as an
2.2.5 Crustal structure
ice sheet) placed on it is supported over a greater area than Seismic and gravity data, together with direct evidence
that covered by the load itself. In other words the litho- from the rocks themselves, allow us to identify the struc-
sphere experiences flexure, just as a springboard does when tural and compositional differences between oceanic and
a diver walks along it. This kind of behaviour, known as continental crust. Oceanic crust has a mean density of about
flexural is'ostasy, is especiaily, important for oceanic litho- 3000 kg m-3 and is composed of layers of basic rocks, broadly
sphere and for thick continental lithosphere with a high basaltic and gabbroic in composition, with a thin veneer of
rigidity (see Section 4.2.3). sediments. Over most of the ocean floor this sedimentary
A second reason for isostatic anomalies is that the cover is only 1-2 km thick and in the vicinity of mid-
lithosphere is not capable of adjusting instantaneously to a oceanic ridges it becomes very thin indeed. One of the most
change in load. Although the great ice sheets which remarkable discoveries arising from oceanographic research
40 Endogenic processes and landforms

Box 2.1 Surface elevation and crustal thickness depth of between 10 and 30 kIn and this divides the con-
tinental crust into upper and lower layers. The upper crust
Regions of high elevation on the continents that are in is composed of rocks of a broadly granitic composition,
approximate isostatic equilibrium are known to be 'sup- with a highly variable covering ranging from unmeta-
ported' by crust of greater than average thickness (although morphosed sedimentary and volcanic rocks to medium- and
compensation actually occurs at the base of the litho- high-grade metamorphosed sedimentary strata. The
sphere). The increase in crustal thickness required to sup-
composition of the lower continental crust is less well
port topography of a given elevation is determined by the
relative densities of the continental crust and the sub- known but probably consists mainly of granulite, an inter-
lithospheric mantle displaced at depth. Assuming isostatic mediate to basic rock formed in very high pressure and
equilibrium has been attained the relationship between temperature environments and containing mostly calcium-
crustal thickness (c) and elevation (h) is given by rich feldspars and pyroxene. Another possibility is the
_ hpm presence of amphibolite rocks with an intermediate to basic
c=c+ - - -
Pm - Pc composition since this would also give the seismic wave
velocities recorded at this depth. The earlier notion that the
where c' is the average thickness of continental crust, and Pc
and Pm are the densities of the crust and the sub-lithospheric lower continental crust is largely basaltic in composition
mantle respectively. (The average elevation of continental has received little support from more recent studies.
crust, including that part below sea level, is assumed to be In addition to differences in the thickness and density of
okm; the actual figure is around 120 m above sea level.) continental and oceanic crust identified from seismic
Taking the average thickness of continental crust to be
evidence and the nature of gravity anomalies, it is possible,
35 km, and crustal and mantle densities to be 2700 kg m- 3
and 3300 kg m-3 respectively, we can calculate values to through a more detailed analysis of such data together with
show the dependence of elevation on crustal thickness (Fig. h~at flow measurements and geological information indica-
B2.l). tive of the history of crustal stability, to identify several,
These results broadly accord with what we know about major structural regions within the continents and oceans
the thickness of the crust under regions of high elevation
(Table 2.2). Most of these regions are also topographically
from other lines of evidence. For instance, the Tibetan
Plateau, which has extensive areas over 6km in elevation, is distinct, but some, such as the continental shield, or craton,
thought to be underlain by crust around 70 km thick. In and mid-continent types, are primarily differentiated on the
some elevated regions, such as the Basin and Range Pro- basis of their detailed structure and the thickness of their
vince of the south-west USA, this relationship does not sedimentary cover. The validity of any model of global
seem to apply because the lithosphere is apparently much
tectonics must be judged by its ability to account for the
thinner than av'erage and' is underlain by unusually hot
asthenosphere of lower than normal density. distribution and characteristics of these major structural
regions.

2.3 Development of ideas on global tectonics

Serious scientific attempts to explain the major structural
and relief features of the Earth began only in the nineteenth
century. In 1829 Elie de Beaumont put forward the idea
that the Earth is contracting and argued that compressional
40 50 60 70
Thickness of Continental Crust (km I
80
stresses set up in the crust as a result of the cooling of the
Earth's interior would give rise to faulting, folding and
Fig. B2.1 Elevation of continental areas above sea level as a
function of crustal thickness. thickening of the crust, and eventually to the formation of
mountain ranges. This proposal found support in the work
of Lord Kelvin, the pioneer Victorian geophysicist, who
over the past two decades or so has been the youth of attempted to calculate the age of the Earth from its probable
oceanic sediments and the underlying basaltic crust. The rate of cooling on the assumption that it had formed as a
oldest known rocks from the ocean floor come from the molten offshoot of the Sun. However, as the role of radio- .
western Pacific, but these are only of Jurassic age. This can active decay in generating heat in the Earth's interior (and
be contrasted with the antiquity of rocks exposed over large thus drastically reducing the previously hypothezised rate
areas of the continents which in some cases are more than of cooling) became appreciated the notion of a contracting
3000 Ma old. Earth was rejected. Various global tectonic models were
Like the oceanic crust, continental crust has a layered proposed during the nineteenth and the first part of the
structure, but one which is much more complex and less twentieth century, with most attention being focused on the
clearly defined. A seismic boundary, known as the Conrad origin of the major mountain systems of the Earth's
discontinuity, underlies some continental regions at a surface. The majority of the models assumed that the
 Global morphology and tectonics 41

Table 2.2 Structural classification of crustal types

CRUSTAL TYPICAL HEAT FLOW BOUGUER DEGREE OF CHARACTERISTICS

TYPE CRUSTAL ANOMALY STABILITY
THICKNESS (mWm-2) (mgal)
(km)

Continental 35 29-38 -10 to -30 Very stable Low to moderate elevation. Composed of highly defonned
shield, or Precambrian metamorphic and plutonic rocks, unaffected by
craton post-Precambrian tectonism. No covering of post-
Precambrian sediments
Mid-continent 38 33-50 -10 to-40 Stable Generally similar to continental shields, the major difference
being the development since the Precambrian of broad
undulations which have led to the accumulation of thick
sedimentary sequences in extensive basins, particularly near
continental margins. The mid-continent structural type
commonly occurs adjacent to continental shields and
together they account for the majority of the area of
continental platfonns
Basin and range 30 71-105 -200 to -250 Very unstable Named after the type area of the Basin and Range Province
of the western USA, this structural type is characterized by
great instability associated with significant extension of the
crust giving rise to a series of basins and intervening,
usually significantly eroded, upland areas. Notable are the
very thin crust (for continental regions), the high mean
elevation, high heat flow, strongly negative gravity
anomalies and marked volcanic and seismic activity
Young mountain 55 29-84 -200 to -300 Very unstable Comparatively narrow, elongated regions attaining elevations
belt or active in excess of 3 km which have experienced relatively recent
orogen and often rapid uplift, in many cases preceded by intense
folding and thrusting of thick sedimentary sequences as a
result of crustal compression. Crustal thicknessess are highly
variable, ranging up to a maximum of about 70 km. There is
also a wide range of heat flow rates with higher rates being
characteristic of younger mountain belts
High plateau 35? 84? -150 to -250 Very unstable this crustal type lacks a distinct geophysical character, but
is typified by high elevations resulting from uplift lacking
associated folding or thrust faulting. The Colorado
Plateau provides an example
Island arc 30 29-167 -50 to +100 Very unstable This type includes a wide range of structural forms, including
significant 'continental' fragments such as Japan and New
Zealand, as well as arcs formed of numerous individual
volcanic peaks of predominantly andesitic composition.
Crustal thickness, heat flow rates and Bouguer anomalies are
all highly variable both between and within island arc systems.
Volcanic and seismic activity is intense
Oceanic trench Low Strongly Very unstable Closely associated with island arcs but are best considered
negative separately as they may be found adjacent to young mountain
belts. Earthquake activity is marked, particularly towards the
adjacent island arc or mountain belt
Ocean basin 11 54 +250 to +350 Very stable Covers extensive areas of the ocean. Ocean basins are broken
by long linear fractures or faults and are punctuated in places
by volcanoes of basaltic composition
Mid-oceanic 10 42-335 +200 to +250 Unstable Crustal type typified by shallow earthquakes and abundant
volcanic activity and composed of basaltic lava. Some ridges
have a central rift valley and at a limited number of points
they break the surface of the ocean to fonn islands

Source: Based mainly on J. N. Brune (1969) American Geophysical Union Monograph 13,230-42; and E. W. Spencer (1977) Introduction to the Structure
a/the Earth (2nd edn) McGraw-Hill, New York.

positions of the continents were fixed and that the ocean vassed on a number of occasions in the nineteenth century,
basins were ancient features. but its most influential advocate was Alfred Wegener who
intially presented his ideas in 1912. He outlined three main
lines of supporting evidence. First, he pointed out the 'fit'
2.3.1 Continental drift
of coastlines now separated by thousands of kilometres of
The notion of continental drift is not new. It was can- ocean, in particular those of South America and Africa. He
42 Endogenic processes and landforms

reconstructed the present land masses into a single super- 2.3.2 Palaeomagnetic evidence
continent called Pangaea and suggested that this had
initially split into two continents, Gondwanaland (Gondwana) During the mid-1950s, at a time when continental drift was
to the south and Laurasia to the north, before further not seriously considered by most earth scientists, new
rupture and drift resulted in the familiar shape and location evidence in the form of palaeomagnetic data from rocks
of the continents today. again began to bring into question the notion of stationary
Secondly, Wegener, a climatologist by training, referred continents. S.K. Runcorn and his associates, working in
to the global distribution of rocks characteristic of par- Britain, conducted an intensive programme of data col-
ticular climatic environments, particularly ancient glacial lection involving the measurement of remanent magnetism
deposits (tillites). He maintained that the distribution of in rocks of various ages from around the world. Earlier
tillites and patterns of glacial striations produced by ice studies of such palaeomagnetism in France and Japan had
sheets during the Late Palaeozoic glaciation (now termed shown that iron-rich volcanic rocks, such as basalt, record
the Gondwana Ice Age) found in the now widely dispersed the magnetic field prevailing at the time they are formed. As
continental areas of southern Africa, Australia, South basaltic lavas cool through the temperature interval 500 to
America, India and Antarctica indicated that these land 450°C (the Curie point) the atomic groups within the iron
masses were contiguous at that time and probably located minerals they contain become aligned parallel to the magnetic
fairly close to the South Pole. He also pointed out, though lines of force acting upon them. Once the temperature of
somewhat less convincingly, that some geological struc- such rocks falls below 450 °C this magnetic orientation
tures could be traced from one continent to another across becomes 'frozen' into the individual minerals, only being
what are now wide stretches of ocean. subsequently disturbed by marked heating. It was found
Wegener's third main category of evidence was palaeon- that the palaeomagnetism of young rocks tended to be close
tological. He argued that there were marked similarities to that expected from the present magnetic field, but older
between fossil terrestrial fauna and flora from Palaeozoic rocks showed marked deviations.
strata in the various southern continents and between North As a rough approximation the Earth's present magnetic
America and Europe, suggesting that during this era there field can be represented by a regular dipolar pattern (similar
was free movement over a single large land mass. Fossils to that produced when a bar magnet is held beneath iron
from more recent strata showed progressively less simi- filings scattered on a sheet of paper). If it is assumed that
larities with time in the different continents; this he the Earth's magnetic field was of this form in the past, then
attributed to the contrasting evolutionary paths of groups of it is possible to estimate the position of the magnetic pole
animals and plants separated by continental drift. for rocks of known ages (dated by radiometric methods) by
The greatest obstacle to the acceptance of the drift measuring their palaeomagnetism. On the basis of such
hypothesis was not so much the nature of this supporting measurements on rocks from Europe, Runcorn and his
evidence (although alternative explanations were readily colleagues demonstrated an apparent movement of the
proffered) but rather the failure to find a convincing magnetic pole over the past 500 Ma. This phenomenon was
mechanism by which it could occur. From an examination termed polar wandering and is now thought to arise from
of global hypsometry Wegener had noted the concentration the movements of the continents themselves rather than
of large areas of the Earth's surface around two levels from any significant shift in the location of the magnetic
representing the continental platforms and the ocean basins. pole itself.
He saw this as being compatible with the crust being made After extending palaeomagnetic investigations to rocks
up of two layers, the upper one of relatively low density, from North America it was found that while these too
the lower layer of higher density. He suggested that tidal showed an apparent movement of the magnetic pole the
forces resulted in the lighter continental crust 'ploughing' path of polar wandering appeared to differ systematically
tlirough the substratum of denser crust underlaying the from that determined for Europe. This discrepancy could be
oceans. This proposal was readily dismissed by geo- explained If North America and Europe had been moving
physicists who were easily able to demonstrate that the with respect to each other as well as with respect to the·
Earth is far too strong to be deformed by such tidal forces. magnetic pole. Further work began to suggest that
Suggestions made by Arthur Holmes and others that palaeomagnetic data supported the model of continental
continents could be moved by convection currents within drift proposed by Wegener although there were initial
the mantle were ignored rather than countered by opponents uncertainties over the accuracy of the technique. Palaeo-
of drift. While information about the Earth's interior and magnetic data could indicate north-south movements
the ocean floors was so sparse, most geologists preferred (palaeolatitudes) but they could not give past longitudinal
the safety of the established doctrine of s~ationary continents. positions. More,over, the weakness of the remanent
magnetism contained within volcanic rocks together with
 Global morphology and tectonics 43

60

30
,. ...
, . ....:
.,......
~
.0/'

...
",. 1
.'"
..1.·····_-' ..
:..~
-:- .
.' ....'""",,IA·: . . .. .:. "':'t. •
. -... :••"1:.0/... 60°
60 ..
A

30

o 0 Regions with deep focus earthquakes·

.. ::.
...
B
Fig. 2.10 Location ofapproximately 30 000 earthquakes recorded by the US Coast and Geodetic Survey between 1961 and 1967. In
(A) all the registered earthquakes are shown, whereas (B) records only those earthquakes generated at depths in excess of100 km, with
a separate indication ofthose regions experiencing deep focus seismic events (300-700 km). (Modified from M. Barazangi and J.
Dorman (1969) Bulletin of the Seismological Society of America 59,370-6.)
44 Endogenic processes and landforms

various sources of error meant that large numbers of are shallow, but they become progressively deeper further
individual measurements had to be averaged in order to away from the trench up to a maximum depth of about
provide a reasonably precise estimate of past continental 700 km. This relationship gives rise to a dipping planar
positions. Consequently the palaeomagnetic evidence did region of seismic activity, known as a Wadati-Beonioff
not lead to an immediate resurrection of the continental zone, inclined at an angle of between about 20 and 55 .
drift hypothesis. A third area of earthquake activity extends as a rather
diffuse belt from the Mediterranean region, through the
Himalayas and into Burma. The earthquakes here, which·
2.3.3 Global seismicity
are associated with a long and complex mountain belt, are .
Although the general distribution of earthquakes around the generally shallow and generated primarily by compressive
,world had been known for many years, it was only with the stresses. The final zone of seismic activity is confined'to
establishment of a comprehensive network of seismographic relatively small areas within the continents and is charac-
stations to monitor underground nuclear weapons testing in terized by shallow earthquakes which are not associated
the early 1960s that a detailed global picture emerged (Fig. with marked volcanic activity. The most notable example
2.10). Far from being randomly distributed, nearly all is the famous San Andreas Fault Zone of California, where
seismic activity is concentrated in relatively narrow zones. crustal movement involves large lateral displacements.
This pattern coincides with a number of structural.f.~atures,
including mid-oceanic ridges, mountain belts, oceanic
trenches and volcanic island arcs. 2.3.4 Sea-floor spreading
Looking in more detail, it is possible to classify seismic The idea that the polarity of the Earth's magnetic field has
activity into four zones. The first is located along the mid- reversed through geological time was proposed early this
oceanic ridges. Earthquakes here are generated at shallow century by Brunhes in France and Matuyama in Japan.
depths (less than 70 km), and they coincide with the high Measurements of remanent magnetism in a variety of
heat flows and volcanic activity of the ridges. The second volcanic rocks found on the continents had shown that
category is closely related to oceanic trenches and their some were magnetized in the opposite direction to the
associateq island arcs or mountain belts. Such a zone Earth's present magnetic field. Further data on these
extends more or less continuously around the western, geomagnetic reversals were collected and by the mid-
northern and eastern margins of the Pacific Ocean. The 1960s a chronology of reversals had been established
most notable feature of this seismic zone is the relationship through the radiometric dating of a large number of
between the depths of earthquake generation and their samples from all over the world. The resulting detailed
location with respect to the adjacent island arc or mountain palaeomagnetic record showed that during the past 4 Ma
belt (Fig. 2.11). On their oceanic margin earthquake foci the average duration of 'normal' fields (like the present
one) had been about 420 ka and that of 'reversed' fields a
little longer (about 480 ka). Periods of a particular polarity
A Im:~:::::::~{:::t:I:i:::~~~:I::i~:f:::Ii:~~o km are termed epochs, with shorter phases of opposite
1500 km: 0
magnetization within these being referred to as events.
B r::::~~~:It:fff::::::I:::::~:::~:::~~o km Geophysical surveys during the late 1950s involving the
1500 ' I , i , 'km ! 0
W : E measurement of the magnetic characteristics of the crust of
.--.-----.----a:::ciJ'iTl;-O 0
the ocean floor indicated the presence of anomalous linear
CD km patterns (Fig. 2.12). These magnetic anomalies were found
to be hundreds of kilometres long and typically 20-30 km
km
across, and it initially seemed that they were caused by
1500
alternating bands of weakly and strongly magnetized rocks.
o Earthquake epicentre
No convincing explanation was initially produced to
... Volcano
,,,,,•.•,.,. Line of volcanoes
explain these feat\lres.

~o .A Oceanic trench
In 1962 H. H. Hess, of Princeton University in the USA,
published a radical explanation for much of the geological
data that had been accumulating over the previous decade
Fig.2.11 Map of earthquake epicentres in the northern or so. This proposed that mid-oceanic ridges represent
Japan-Kurile-Kamchatka region of the north-western Pacific. regions where new oceanic crust is being generated by the
Earthquakes with a magnitude of 6 or more on the Richter scale upwelling of hot mantle material. He suggested that this
are shown. The composite profile plots the depth of all the
new crust spreads laterally away from these ridges until it
earthquakes on the map in relation to their distance from the
oceanic trench. (After H. Benioff, 1954, Bulletin of the Geological reaches an island arc or mountain belt where it descends
Society of America 65, Fig. 2, p. 388.) into the mantle along the adjacent oceanic trench. This
 Global morphology and tectonics 45

Geomagnetic
reversals Transform
fault across
mid-oceanic
ridge
A B

Ridge axis

Lithosphere
thickening as it ages
2000m and becomes cooler
Submanne
contour Fig.2.13 Schematic illustration of the sea-floor spreading
hypothesis of Hess as related to the record ofpalaeomagnetic
reversals by Vine and Matthews. A displacement across the ridge
representing a transform fault is also shown.

direction of magnetization. Positive anomalies represented

underlying rocks with normal polarity, negative anomalies
those with reversed polarity. As new crust emerged at mid-
oceanic ridges, they suggested, it became magnetized with
the polarity of the prevailing magnetic field (Fig. 2.13).
Subsequently the widths of the linear magnetic zones
were related to the duration of each successive epoch of
normal and reversed polarity on the assumption that the
ocean floor spreads at a constant rate away from mid-
oceanic ridges. The close relationship observed between the
width of each magnetic anomaly and the corresponding
time scale of magnetic reversals, which by now had been
accurately established by the radiometric dating ?f
terrestrial lavas, convinced many earth scientists of the
reality of sea-floor spreading. Further corroborative
evidence was provided by the palaeomagnetic record of
ocean sediments containing particles of iron-rich minerals
Fig. 2.12 Magnetic anomalies across the Reykjanes Ridge in the which lay on the basaltic oceanic crust.
North Atlantic. This is a summary diagram of several traverses When the patterns of magnetic anomalies across mid-
across the ridge, the axis o/which is indicated by the diagonal
line. Areas ofpositive anomalies are shown in black. (Magnetic oceanic ridges were mapped out it was noticed that in
anomaly map after 1. R. Heirtzler et al. (1966) Deep Sea Research several cases the parallel bands of similarly magnetized
13, Fig. 7, p. 435.) crust were broken by numerous offsets or displacements
which seemed to be aligned approximately at right angles
'conveyor belt' view of the oceanic crust, which owed to the axis of the ridge. Some of these displacements were
something to the earlier ideas of Holmes, was termed sea- very large, amounting to 1000 km or more in places. It
floor spreading by Dietz, another American geologist. appeared that these offsets marked some kind of lateral
In 1963 Vine and Matthews, two Cambridge geophysicists, movement between adjacent sections of oceanic crust. In
took up the idea of sea-floor spreading and linked it with 1965 J. T. Wilson of the University of Toronto proposed
data on the palaeomagnetic anomalies observed along mid- that these were not simple strike-slip faults, as had been
oceanic ridges (Fig. 2.12). They proposed that these previously suggested, but were an entirely new variety, which
anomalies, which seemed to be arranged symmetrically on he termed a transform fault (Fig. 2.13).
either side of, and roughly parallel with, the ridge crest, were During the late 1960s considerable effort was invested in
not a result of variations in the intensity of magnetization, establishing the history of the ocean floor, much of this
as earlier suggested, but rather a consequence of the involving the activities of the Deep Sea Drilling Project.
46 Endogenic processes and landforms

The chronology of geomagnetic reversals was extended the inclined zone 'of seismicity (the Wadati-Benioff zone)
back into the Mesozoic and this together with the dating of associated with island arcs and active continental margin
ocean floor sediments enabled the age of large areas of mountain belts could be explained by the existence of a
oceanic crust to be determined. Little was found to be more slab of lithosphere plunging down into the mantle. Analysis
than 150 Ma old. of motions associated with earthquakes generated in the
Wadati-Benioff zone suggested that lithosphere was indeed
moving downwards into the mantle along what became
2.4 Plate tectonics
termed subduction zones. This supported Hess's idea of
By the late 1960s a number of lines of evidence were being crustal formation along mid-oceanic ridges and its re-
brought together into a new, radical model of global absorption or subduction, into the mantle below oceanic
tectonics. Palaeomagnetic studies on the continents had trenches.
indicated that continental drift had, in fact, occurred, while All these ideas were incorporated into the theory of plate
more detailed information on the Earth's crustal and sub- tectonics which was more or less separately developed by a
crustal structure had led many to accept the existence of a number of workers at about the same time in the late 1960s.
relatively mobile zone within the mantle (the astheno- The outer layer of the Earth was seen as consisting of a
sphere). Detailed mapping of the mid-oceanic ridges, rigid lithosphere, composed of several 'plates', which
together with magnetic and seismic surveys, had indicated moved over the underlying more mobile asthenosphere.
that these were zones of crustal tension where new crust Sea-floor spreading and continental drift were seen as
was being extruded from the mantle below. The history of involving these lithospheric plates rather than just oceanic
polarity reversals of the Earth's magnetic field recorded in or continental crust.
this newly generated crust as it cooled indicated that, once
formed, it moved away from ridges towards oceanic trenches.
2.4.1 Outline of the plate tectonics model
,On the basis of the much more complete seismic data
available by the late 1960s it was being suggested that The plate tectonics model proposes that the Earth's surface

EURASIAN

IC

+
Absolute
Divergent
boundary •• •• Convergent
boundary
Transform
boundary
40
----t.~ plate
velocity
(mma-1 )
Fig.2.14 Map of the major lithospheric plates. The various types ofplate boundary are shown and the estimated current rates and
directions ofplate movement are indicated by arrows (rates in mm a-I.)
 Global morphology and tectonics 47

comprises seven major, and at least a dozen minor, litho- sufficient for it to subside spontaneously into the astheno-
spheric plates composed of the crust and the upper, more sphere. This may have occurred in the western Pacific
rigid, part of the mantle (Fig. 2.14). These plates are Ocean where the age of the lithosphere presently being
constantly in motion with respect to one another and to the subducted indicates that cooling over a period of about
EaJ;th's axis of rotation, and the motion of one plate 180-200 Ma is required for this process of spontaneous
influences the movement of others. Rates of movement subduction to occur. By contrast continental lithosphere has
range up to 100 mm a-I (l00 km Ma- 1) and average around a significantly lower mean density than oceanic lithosphere
70 mm a-I (70 km Ma- l ). Much, though not all, of the and its buoyancy with respect to the asthenosphere prevents
Earth's tectonic, volcanic and seismic activity is directly or all but limited subduction.
indirectly associated with movement between neighbouring The two fundamental assumptions underlying the plate
plates. The narrow zone marking the relative movement tectonics model are that the surface area of the Earth has
between two plates, wpich in most cases is fairly clearly not changed significantly with respect to the rate of
demarcated by seismic activity, is termed a plate boundary, generation of new oceanic crust, and that there is a lack of
with the peripheral region adjacent to this boundary being internal deformation within plates compared to the relative
referred to as a plate margin. motion between them. If the Earth's surface area has
There are three types of plate boundary. At a divergent increased significantly, sea-floor spreading could occur
boundary, such as the Mid-Atlantic Ridge, new crust is without associated plate subduction, the rate of crustal
formed and attached to the upper part of adjoining formation at mid-oceanic ridges matching the increase in
lithospheric plates while new upper mantle is accreted to surface area. There are, however, good reasons to believe
the lower part. Plate movement is laterally away from the that the Earth's radius has not increased significantly during
spreading ridge. At a convergent boundary, such as the the past 500 Ma or so.
western coast of South America, two plates are in motion Other evidence supports the idea of a lack of major
towards each other, with one plate slipping down below the deformation within plates, at least over plates composed of
other along a subduction zone. The surface area of a plate is oceanic lithosphere. Oceanic crust has a layered structure
reduced at a subduction zone whereas it is increased along which shows remarkably little disturbance and this would
a spreading ridge. Along a transform boundary, two not be the case if significant deformation had occurred. In
plates simply move laterally past each other along a trans- order for plates to behave in the manner proposed the
form fault without any major element of divergence or con- lithosphere of which they are composed must be suf-
vergence. The most famous example of this kind of boundary ficiently rigid compared with the underlying asthenosphere
is the San Andreas Fault System in California, USA. At for stress to be transmitted from one side to the other. If
some localities three plates may come into contact. Such this were not the case significant deformation would occur
a boundary is known as a triple junction, an example within plates rather than being concentrated along plate
being the junction of the Pacific, Nazca and Cocos Plates margins. The movements of plates would be greatly
(Fig. 2.14). complicated if they were to undergo significant internal
The upper layer of a plate is composed of either oceanic deformation rather than acting as thin but rigid caps in
or continental crust or both. There is a fundamental dif- motion around a sphere. It is quite clear, however, that
ference between the behaviour of lithosphere capped with marked deformation does in fact occur within some
oceanic crust (oceanic lithosphere) and that covered with continental regions which are thousands of kilometres away
continental crust (continental lithosphere) since only from the nearest plate boundary.
oceanic lithosphere can be generated at mid-oceanic ridges In the remainder of this book we will be using plate
and subducted at oceanic trenches. tectonics as a framework for the interpretation of the
Oceanic lithosphere has a mean density rather close to Earth; s large-scale topographic features. As wiil become
that of the immediately underlying asthenosphere. When apparent, however, there are several instances where it will
newly formed along mid-oceanic ridges it is hot and thin be necessary to modify the basic model outlined in this
and it probably has a slightly lower density. However, as chapter.
oceanic lithosphere ages, cools and thickens it becomes
2.4.2 Classification of plate boundaries
more dense than the asthenosphere and rests upon it in an
unstable state. Along zones of plate convergence, In Table 2.3 each form of plate boundary is classified
especially where these occur along the margins of according to the type of interacting lithosphere involved.
continents (such as along the west coast of South America), The most well-defined divergent boundaries are associated
buckling of the oceanic lithosphere causes it to founder and with mid-oceanic ridges, but they may also occur in an
sink into the underlying asthenosphere, thereby forming a incipient form within continents as rift valleys comprising
subduction zone. Alternatively, it seems possible that old, linear systems of faults associated with tension in the crust.
cold and thick oceanic lithosphere can attain a density Some rift valleys are analogous to mid-oceanic ridges in
48 Endogenic processes and landforms

Table 2.3 Classification and primary characteristics of plate boundary types

MORPHOLOGICAL AND STRUCTURAL FEATURES

Boundary Stress Oceanic-oceanic Oceanic-continental Continental-

type lithosphere lithosphere continental
lithosphere

Divergent Tensional Mid-oceanic

ridge Rift valley
Volcanic
activity Volcanoes
Oceanic trench Oceanic trench and
and volcanic continental-margin
island arc mountain belt and
volcanicity
Convergent Compressional
Complex Modified continental- Mountain belt,
island margin mountain limited volcanic
, arc collision belt activity
zone
Ridges and Fault zone,
Transform Shear valleys normal no volcanicity
to ridge axis

experiencing shallow earthquakes and volcanicity. surrounded by divergent or transform boundaries. This
Convergent boundaries between oceanic lithosphere are means that it must be growing in area and that, conse-
marked by an oceanic trench, a volcanic island arc and a quently, the total area of all the other plates must be
Wadati-Benioff zone. Subduction of oceanic lithosphere declining if the 'constant area' assumption of plate tectonics
beneath continental lithosphere is also associated with a is to be satisfied. Moreover, since sea-floor spreading is
trench and an inclined zone of seismicity, but in this case taking place from both the Mid-Atlantic Ridge and the
volcanic activity occurs within the active mountain belt Carlsberg Ridge (in the Indian Ocean), one or both of these
located on the continental lithosphere fringing the trench. ridges must be experiencing movement with respect to the
Where two continents collide neither experiences signi- Earth's rotational axis.
ficant subduction but some crustal thickening occurs and a If plates are regarded as rigid caps involved in relative
mountain belt is formed. motion around a sphere their movement can be analysed in
Transform boundaries, marked by transform faults, occur terms of specific geometric controls. In Figure 2.15A we can
both as fracture zones along mid-oceanic ridges and as imagine a segment of the Earth's surface (ABE) which has
strike-slip fault zones within continental lithosphere. Plate split into two, the two halves (ABC and ADE) remaining in
movements can be oblique as well as simply divergent, contact only at point A. Point A is the pole of rotation
convergent or transform. Oblique divergence is most about which movement occurs, and the separation of
commonly accommodated by transform offsets along a originally adjacent points on each segment occurs along
mid-oceanic ridge crest, while oblique convergence is small circles about this point. If we consider irregularly
resolved by the complex adjustment of lithospheric frag- shaped areas on the sphere (X and Y), rather than the
ments along the plate boundary. regular segments, these can be seen to move apart in such a
way that their motion can similarly be described with
respect to the same pole of rotation at A. The effect is
2.4.3 Plate motion
clearly illustrated in the alignment of transform faults
Plates vary greatly in area (Fig.2.14). The Pacific Plate, for which mark the off-sets along either side of divergent plate
instance, underlies almost the whole Pacific Ocean whereas boundaries (fig. 2.15B). These represent segments of small
a number of minor plates within the continents have an area circles concentric about the pole of rotation with respect to
of less than 1 x 1Q6 km 2 . Except where subduction zones lie which the diverging plates are moving. A consequence of
adjacent to mountain belts on continental margins, plate this is that the rate of spreading along a ridge will be a
boundaries do not coincide with continental coastlines. function of the sine of the angular distance from the pole of
Many continental margins are not separated by subduction rotation; the greater the distance from the pole of rotation:
zones from the divergent boundaries marked by mid- the greater the rate of spreading (up to an angle of 90\ At
oceanic ridges. For example, apart from its northern margin convergent boundaries the rate of plate subduction is
along the Mediterranean Sea, the African Plate is determined in a similar way.
 Global morphology and tectonics 49

.'

111
i
:111111 1:/111/':.·....Ii."I liil:ll:.:i:I::

:11111: :i·:::ii·:::I::::;.::IIII::1i:·:·iil:1::: ::::::::ii'::!:i1t'ft:,::f i:::..

Fig.2.15 The geometry ofplate motion: (A) the movement of two segments of the Earth's surface with respect to a common pole of
rotation; (B) the alignment of transform faults (dashed lines) in the equatorial Atlantic compared with small circles (solid lines)
concentric about a pole located at 58W, 36'W. (After E. R. Oxburgh (1971) in: I. G. Gass, P. J. Smith and R. C. L. Wilson (eds)
Understanding the Earth. Open University Press, Milton Keynes, Figs. 19.2 and 19.3, p. 266.)

By comparing the alignments of transform faults, the telescopes. The measurement is derived from the very
poles of rotation of the major plates have now been small difference in the time of arrival of signals received at
established. Moreover, by relating this information to data different telescopes from a distant radio source (usually a
on spreading rates, the directions of relative plate motions quasar). The accuracy of this technique has been steadily
derived from the analysis of earthquakes, and palaeo- . improving and measurements can now be made over
magnetic evidence of continental palaeolatitudes, it has thousands of kilometres to an accuracy of a few milli-
been possible to establish the direction and speed of· metres. Preliminary results have confirmed that plates are
present-day plate motions and what is probably a fairly moving in the directions, and at the rates, indicated by
accurate history of movements over the past 200 Ma or so already acquired geophysical data. These results are·
(Fig. 2.16). important for plate tectonic theory because they show that
Since the mid-1980s a direct means of measuring present the Earth is not currently expanding as has been suggested
rates and directions of plate motion has become available. by a small minority of earth scientists, at least not at a rate
This technique, known as very-long-baseline interferometry which is significant compared with rates of plate motion.
(VLBI), can determine with great accuracy the distance The 1990s will see further refinements to these measure-
between two points on the Earth's surface by using radio ments as the Global Positioning System becomes fully
50 Endogenic processes and landforms

........... : :
.0 .::• • •

200 Ma ago

160 Ma ago

Fig. 2.16 The pattern of continental drift over the past 200 Ma. The six maps show the estimated past positions of the continents at 40
Ma intervals. Note the assemblage of the continents into one major landmass (Pangaea) 200 Ma ago with the Tethys Sea as a large
embayment on its eastern flank (A). By 160 Ma BP this landmass had begun to rupture between North Africa and North America (B)
and 40 Ma later the separation ofGondwana (the southern continent) and Laurasia (the northern continent) was complete (C). Over
the past 120 Ma the Atlantic Ocean has opened and the various continents making up Gondwana have drift!d apart (D-F). Particularly
noticeable are the rapid northward movement ofIndia and Australia away from Antarctica. For purposes of comparison all the maps
have been produced on a cylindrical equidistant projection. Such a projection, however, grossly distorts areas and shapes in high .
latitudes. (Modified from A. G. Smith, A. M. Hurley and J. C. Briden (1981) Phanerozoic Paleocontinental World Maps. Cambridge
University Press, Cambridge, Maps 1 and 2, pp. 8...,.9, Maps 13 and 14, pp. 20-1, Maps 21 and 22, pp. 28-9, Maps 29 and 30, pp. 36-7,
Maps 37 and 38, pp~ 44-5 and Maps 45 and 46, pp. 52-3.)
 Global morphology and tectonics 51

c
120 Ma ago

80Ma ago

operational. This method of measuring distances between existence of plate tectonics the question of what causes
points on the Earth's surface is similar to VLBI but uses plates to move has yet to be fully resolved. One idea is that
signals from Earth-orbiting satellites rather than extra- lateral flow in the asthenosphere, arising from convection
terrestrial radio sources. in the mantle, drags along the overlying lithosphere (Fig.
2.17(A-C). Convection currents rise and diverge below
mid-oceanic ridges and converge and descend along
2.4.4 Mechanisms of plate movement subduction zones. Convection occurs when too much heat
In spite of the wide range of supporting evidence for the is present at depth to be conveyed upwards solely by
52 Endogenic processes and landforms

40 Ma ago

Present day

Fig.2.16 cont.

thermal conduction. Temperatures increase and material fact occur in the mantle; whether this convection is solely,
begins to rise since it is less dense than the surrounding or even primarily, responsible for plate movement is
medium; cooler, denser material moves laterally at depth to another matter.
take its place. If, however, the medium is too viscous Most of the dispute about the nature of convection within
convection will be inhibited, or even prevented altogether. the Earth "revolves around whether the circulation pattern
Although our knowledge of the physical properties of the involves all, or only part, of the mantle. On the basis of
Earth's interior is at" present far from complete, most certain assumptions about the chemical heterogeneity of the
geophysicists now support the idea that convection does in mantle and a postulated marked increase in viscosity with
 Global morphology and tectonics 53

Mid-oceanic Mid-oceanic
ridge ridge

! !

700km
--

A 8
Mid-oceanic Mid-oceanic
Subduction ridge Subduction ridge
zone ! zone ! Ridge-push---+
! ~~ -.~ !
~ I
~ ,~-...
Man tie ~
t
Lava
y.+\ \ ./ \ J jr injection

Ck~llilill;lt.I.~~
Lithospheric
0
Mid-oceanic
thickening ridge
!
---1 ~-... -...~
Gravity
sliding
--+ ---. - . - .
Dense
sinking
~

~
Iithospheri c
slab

E F
Fig.2.17 Possible mechanisms of plate movement: (A) shallow convection confined to the asthenosphere (maximum depth about 300
km); (B) convection in the upper mantle extending to a depth of about 700 km; (C) whole mantle convection; (D) ridge push; (E)
gravity sliding; (F) slab pull.

depth, it has been argued that convection would be confined convection models predict that convection cells have
to the asthenosphere, that is to a maximum depth of about similar horizontal and vertical dimensions, although argu-
300 Ian (Fig. 2.17(A». Some researchers, however, have ments have been presented against this idea. As some plates
suggested that 700 krn may represent the lower limit of have horizontal dimensions of thousands of kilometres it
convection (Fig. 2.17(B». has been argued that convection cells of equivalent hori-
Recently the idea of whole-mantle convection has zontal and vertical dimensions would be required to move
gained support, largely as a result of new estimates of the them. An alternative view is that convection involves a
viscosity of the mantle below the asthenosphere. It has been two-scale pattern consisting of large-scale convection in the
suggested that a hot layer at the core-mantle boundary, upper mantle responsible for plate movement superimposed
generated by convection in the mobile outer core, produces on smaller-scale convection cells confined to the topmost
a circulating system extending up through the entire mantle 650 Ian of the mantle.
(Fig. 2.17(C». An important point is that these large-scale Although convection in the mantle has attracted con-
convection cells fit in with the dimensions of plates. Most siderable support as a likely cause of plate motion, it has
54 Endogenic processes and landforms

also been argued that flow in the asthenosphere may simply (1967) (although its treatment of tectonic processes and
be a consequence of the drag exerted by the movement of landform development has been superseded by more recent
overlying plates. Such movement may be caused by at least research). Cogley (1984, 1985) and Harrison et ai. (1983)
three different mechanisms. The injection of lava along contain detailed analyses of continental hypsometry. An
mid-oceanic ridges, for instance, could push the adjacent accessible summary of the properties of the mantle is
plates apart and thereby contribute to plate motion (Fig. provided by McKenzie (1983), while Pollack and Chapman
2.17(D». Perhaps more significant is the thickening of (1977) discuss the use of heat flow data in the estimation of
oceanic lithosphere as it moves away from mid-oceanic global variations in the thickness of the lithosphere. A
ridges. Such a configuration would tend to make a plate much more advanced treatment of the Earth's interior with
slide under the force of gravity downwards from a mid- particular reference to isostasy is to be found in the
oceanic ridge towards a subduction zone (Fig. 2.17(E». conference proceedings edited by Morner (1980).
This would drag the underlying asthenosphere along and The developments which led to the formulation of the
promote a compensatory upward movement within the plate tectonics model are charted in the collection of
mantle under mid-oceanic ridges. original articles edited by Cox (1973). This incorporates all
Cool, thick, old oceanic lithosphere is gravitationally the classic papers including those by Hess, Vine and
unstable as it is generally denser than the asthenosphere Matthews, and Wilson. Clear introductions to each topic
over which it lies. Consequently a cold, dense lithospheric make this an ideal starting point for those interested in
slab descending up to 700 km into the mantle at a following the birth of plate tectonics. A briefer historical
subduction zone will tend to pull the remainder of the platesurvey is provided by Hallam (1973) who has also written
with it (Fig. 2.17(F». The low conductivity of lithosphere an assessment of Wegener's pioneering contribution
(Hallam, 1975). For those with a deeper interest in the
contributes to this effect, it being calculated that the centre
0
of a descending lithospheric slab may be 1000 C cooler history of the plate tectonics revolution Le Grand (1988) is
than the surrounding mantle at a depth of 400 km. Warming an important source. An excellent guide to the study of
occurs so slowly that the descending lithosphere retains itsplate motions, which has a strong practical bias, is that by
high density characteristics for a very long time. Cox and Hart (1986). There are a number of reconstruc-
Although there is still disagreement as to the relative tions of past continental mov~ments available but the series
significance of the various possible mechanisms of plate of maps produced by Smith et ai. (1981) is particularly
motion, recent assessments have indicated that the pull of useful.
descending lithospheric slabs is the predominant driving The application of the plate tectonics model to problems
force. This interpretation is supported by the relationship of Earth structure and topography are considered in the
observed between the rate of movement of individual plates following two chapters and guidance to further reading will
and the length of subduction zones along their margins. be found there. However, for those wanting a fuller account
Those attached to long subduction zones, such as the of the mechanisms of plate movement, McKenzie and
Pacific Plate, have been moving at comparatively high Richter (1976) discuss the possibility of two scales of
velocities (60-90 mm a-I), whereas plates lacking extensive convection while Boss (1983) reviews research on the
subducting boundaries have experienced slower rates of convection question. The book edited by Davies and
movement (below 40 mm a-I). Runcorn (1980) contains a number of papers on the various
suggested mechanisms. Finally, Van Andel (1984) and
Molnar (1988) present brief but thought-provoking assess-
Further reading ments of the current status of plate tectonics theory and
Since the intention of this chapter has been to provide an highlight the remaining problems of the model particularly
introduction to global tectonics and the major elements of when applied to the continents.
the Earth's morphology and structure, the suggested read-
ing is mainly limited to texts which will be useful in pro-
References
viding more details on these topics. For those readers with
little or no background in geology there are a number of Boss, A. P. (1983) Convection. Reviews of Geophysics and Space
Physics 21, 1511-20.
excellent introductory texts. Two of the best are those by Brown, G. C. and Mussett, A. E. (1981) The Inaccessible Earth.
Press and Siever (1986) and Skinner and Porter (1987). A Allen and Unwin, London and Boston.
more advanced coverage of the Earth's internal structure is Cogley, J. G. (1984) Continental margins and extent and number
provided by Brown and Mussett (1981) while Turcotte and of the continents. Reviews of Geophysics and Space Physics 22,
Schubert (1982) give a thorough quantitative grounding in 101-22.
Cogley, J.G. (1985) Hypsometry of the continents. Zeitschrift fur
tectonic processes. Geomorphologie Supplementband 53, 1-48.
A continent-by-continent survey of the Earth's major Cox, A. (ed.) (1973) Plate Tectonics and Geomagnetic Reversals.
morphological features is to be found in the book by King W. H. Freeman, San Francisco.
 Global morphology and tectonics 55

Cox, A. and Hart, R. B. H. (1986) Plate Tectonics: How It Works. McKenzie, D. P. and Richter, F. (1976) Convection currents in the
Blackwell, Palo Alto and Oxford. Earth's mantle. Scientific American 235(5), 60-76.
Davies, P. A. and Runcorn, S. K. (eds) (1980) Mechanisms of Molnar, P. (1988) Continental tectonics in the aftermath of plate
Continental Drift and Plate Tectonics. Academic Press, London tectonics. Nature 335,131-7.
and New York. Morner, N-A. (ed.) 1980. Earth Rheology, Isostasy and Eustasy.
Hallam, A. (1973) A Revolution in the Earth Sciences: From Wiley, Chichester and New York.
Continental Drift to Plate Tectonics. Oxford University Press, Pollack, H. N. and Chapman, D. S. (1977). The flow ofheat from
Oxford and New York. the earth's interior. Scientific American 237(2), 60-76.
Hallam, A. (1975) Alfred Wegener and the hypothesis of Press, F. and Siever, R. (1986) Earth (4th edn) W.H. Freeman,
continental drift. Scientific American 232(2), 88-97. San Francisco.
Harrison, C. G. A., Miskell, K. J., Brass, G. W., Saltzman, E. S. Skinner, B. J. and Porter, S. C. (1987). Physical Geology. Wiley,
and Sloan, J. L. II. (1983) Continental hypsography. Tectonics New York and Chichester.
2,357-77. . Smith, A. G., Hurley, A. M. and Briden, J. C. (1981) Phanerozoic
King, L. C. 1967. ·The Morphology of the Earth: A Study and Paleocontinental World Maps. Cambridge University Press,
Synthesis of World Scenery (2nd edn). Oliver and Boyd, Cambridge and New York.
Edinburgh. Turcotte, D. L. and Schubert, G. (1982). Geodynamics:
Le Grand, H. E. 1988. Drifting Continents and Shifting Theories: Applications of Continuum Physics to Geological Problems.
The Modern Revolution in Geology and Scientific Change. Wiley, New York and Chichester.
Cambridge University Press, Cambridge and New York. Van Andel, T. H. (1984) Plate tectonics at the threshold of middle
McKenzie, D. P. (1983) The Earth's mantle. Scientific American age. Geologie en Mijnbouw 63,337-41.
249(3),50-62.
 3
Landforms and
tectonics of plate margins

3.1 Convergent plate margins: general 3.1.1 Steady-state margins

characteristics
Figure 3.1 illustrates the two types of steady-state margin.
Where oceanic lithosphere is subducted beneath another
The fundamental types of landform created at convergent
oceanic part of a plate the associated volcanic activity and
plate margins are orogens and island arcs. Two distinct
other thermal effects produce an intra-oceanic island arc.
mechanisms are responsible for their development into
significant morphological features. One involves collision and Volcanic island arc
occurs where the motion of two plates towards each other
brings into contact two continents, a continent and an island
arc or two island arcs. The other mechanism involves the
effects of heating associated with plate subduction. Such
consumption of lithosphere can take the form of either sub-
duction of oceanic lithosphere beneath a plate capped by
continental crust or the subduction of oceanic lithosphere
beneath a second oceanic plate.
The subduction of oceanic lithosphere is a steady-state Intra·oceanic island arc
process in that it can continue until a continent or island arc
arrives at the subduction zone. The west coast of South
Continental· margin
America provides an example of oceanic lithosphere sub-
orogen
duction (Nazca Plate) beneath continental lithosphere. The
distinction between steady-state subduction and plate colli-
sion at convergent boundaries is due to the resistance of
continental crust and island arcs to subduction, although it
is possible that the latter can be subducted. The collision of
two continents leads to the eventual cessation of subduction,
whereas other types of collision may cause a change in the
location of the subduction zone. The main types of conver-
Continental· margin orogen
gent plate margin and their possible modes of development
are summarized in Table 3.1. They are classified into steady-
• Oceanic 'crust
state and collision margins and then further subdivided
according to the nature of the crust on the interacting plates, {;\~~I Continental crust
the mode of plate interaction and the type of orogen or Ivt-"'' I Volcanic rocks
island arc formed. Also listed are possible courses of subse-
quent development for each margin subtype. Fig. 3.1 Major elements of the two types of steady-state margin.
 01
OC

~
Q.
=
o
~
=
n'
'C
.,
g
Table 3.1 Classification of convergent plate margins
~
~
MARGIN TYPE STEADY STATE COLLISION Il)

Q.
=
TYPE OF Oceanic/oceanic Oceanic/continental Continental/ Oceanic (island arc type)/ Oceanic (island
;"
CRUST ON (though may contain continental continental arc type)/ Q.
=
INTERACTING sliver of continental oceanic (island 0-
.,
PLATES type crust) arc type) 3
'"
NATURE Oceanic ---> oceanic Oceanic ---> continent Continent ---> Island arc Continent ---> Island arc--->
OF continent ---> continent island arc island arc
INTERACTION

TYPE OF Intra-oceanic Continental- Inter- Modified Modified Compound

OROGEN island arc margin orogen continental continental- 'passive' intra-oceanic
FORMED collision margin continental-margin island arc
orogen orogen orogen (subduction orogen
halted)

Island arc---> Continent ---> No further Back-arc Subduction Subduction of

POSSIBLE continent continent development subduction halted 'underthrusting
collision OR collision as convergent halted but OR arc' continues
EVOLUTION continent ---> OR boundary forearc polarity OR
island arc island arc ---> since subduction subduction may reversal with polarity reversal
collision continent halted continue new subduction of subduction
collision zone in back arc behind 'overriding
arc'
 Landforms and tectonics of plate margins S9

These are particularly extensively developed in the western by the leading edge of the continent on the subducted plate
Pacific Ocean. Their typical arcuate plan form is probably are dramatic since there is a conversion from a 'passive' con-
due to the relationship between the angle of subduction and tinental margin (see Chapter 4) to an active plate boundary.
the Earth's curvature. Some idea of the effect can be gained Eventually the subduction zone is converted into a suture
by depressing the surface of a table tennis ball with the zone as the two continental blocks are welded together.
thumb. Subduction of oceanic lithosphere beneath a plate A second type of collision involves the migration of an
carrying continental crust gives rise to a continental-margin intra-oceanic island arc towards a subduction zone bounded
orogen. The Andes provide an excellent example and this by continental crust (Fig. 3.3). Unless the island arc is sub-
type of orogen is sometimes referred to as an Andean type ducted it will be accreted to the continental-margin orogen
or Cordilleran type. However, such an orogen may also form previously formed along the edge of the continent. Sub-
a continental-margin island arc if the continental crust duction along the continental margin would eventually cease,
behind the arc is below sea level. This is the case with the but it would probably continue along the outer edge of the
Sumatra-Java part of the Sunda Arc in the East Indies. island arc as oceanic lithosphere would still be available for
subduction there. The resulting orogen would be a modified
continental margin type.
3.1.2 Collision margins
A collision margin can take a number of forms, depending
on the characteristics of the interacting plate boundaries. The
first type involves the collision of two continents which gives
rise to an intercontinental collision orogen (Fig. 3.2), the
most notable example of which are the Himalayas. This type
must begin with a continental-margin orogen on the over-
riding plate and be transformed into an intercontinental colli- Precondition to island arc'" continent collision
sion orogen when continental crust on the underthrusting
plate reaches the subduction zone. The changes experienced

Modified continental-
margin orogen
'Passive' continental Continental-margin
margin orogen

,_ ..... \
'- 1,-\

Island arc"" continent collision

Figure 3.3 Possible consequence of an intra-oceanic island arc

Precondition to continental collision moving towards a subduction zone adjacent to a continent (key as
for Figure 3.1). .
Intercontinental
collision orogen
Where a continent moves towards a subduction zone as-

-/-:--;:,-~/::--,,-,:::-~\- . .~
_\,J,-;:' V \ -:;;\ -_
.....::--_--:,/
/'- '\ ;'\ . . . \<.. "\
</~ ~,;.,~
c ,\ i
sociated with an intra-oceanic island arc the consequences
are rather different because of the resistance of continental
-<V \-~ ~ :::- crust to significant subduction. The interaction can be accom-

-----_._~ ""
" ,~
modated in two ways (Fig. 3.4). A possible example of such
a continent - island arc collision is provided by northern
",", New Guinea. The collision of intra-oceanic arcs can also take

","
two forms (Fig. 3.5). Some of these plate boundary configur-
\ ations and stages of development are largely conjectural since
\
clear examples do not exist at the present day. There is, for
Continental collision example, uncertainty as to whether island arcs can be sub-
Fig. 3.2 Development ofan intercontinental collision orogen ducted and whether the direction of subduction can be re-
(key as for Figure 3.1.) versed, as illustrated in Figure 3.4.
60 Endogenic processes and landforms

A A

..

Precondition to island arc collision

Precond ition to continent"" island arc collision

Compound intra-
Modified 'passive oceanic island arc

B continental margin B

Island arc
SUbduction?7 / /
/ /
/ /
/ /
/ /

Island arc collision with closure of one subduction zone

Modified continental-
c
c margin orogen

Fig.3.5 Possible evolutionary sequences following the collision

of two intra-oceanic island arcs. When the island arcs collide,
Continent - island arc collision - subduction renewed
assuming that one is not subducted below the other, the
subduction zone at the collision site will be closed off (B). Part of
through polarity reversal
the original relative motion between the two island arcs could be
Fig. 3.4 Possible development of a convergent margin where maintained by continued subduction adjacent to one of them.
continental crust reaches a subduction zone adjacent to an intra- However, another possibility is a partial polarity reversal of the
oceanic island arc. One possibility is that the collision between system by the establishment of a new subduction zone behind the
the island arc (on the overriding plate) and the continental island arc on the overriding plate (C) (key as for Figure 3.1).
margin (on he subducting plate) will eventually stop subduction
(B). An alternative mode ofdevelopment which maintains the
relative opposing motion between the two plates is a reversal of such as the Hawaiian Islands, which are located far from
the polarity ofsubduction, a process sometimes called 'flipping' plate boundaries and have a different origin (see Section
(C). The original subduction zone is closed off but a new one
4.2.1).
forms behind the advancing island arc. Both processes would give
rise to a modified continental-margin orogen. In (B) a 'passive'
continental margin would gradually be created since subduction 3.2.1 General characteristics and formation
would eventually stop. In (C) an active modified continental-
margin orogen would be formed with an accreted island arc At a subduction zone the downgoing oceanic lithosphere
along its edge and a new subduction zone beyond the island arc finds its topographic expression in the trench formed on the
(key asfor Figure 3.1).
frontal side of the island arc (Fig. 3.7). Much of the veneer
of sediments resting on the oceanic crust of the downgoing
plate will not be subducted since it is not firmly attached to
3.2 Intra - oceanic island arcs the underlying crust and because it is not sufficiently dense
There are about 20 intra-oceanic island arcs located along to sink of its own accord into the asthenosphere. Conse-
subduction zones at the present day, the great majority being quently it either accumulates in the trench or is scraped off
found in the western Pacific Ocean (Fig. 3.6). Such volcanic the downgoing plate and forms wedges of sediment on the
island arcs must be distinguished from other volcanic chains, inner (arc side) of the trench.
 Landforms and tectonics of plate margins 61

BACK ARC ZONE VOLCANIC FORE ARC ZONE

CONTINENT
ARC ZONE

South-West ~
...---....,. r...-----' , r t

<101,H onshu Arc J ./ Volcanic Frotltal

Marginal basin arc (act.i.ve) arc(jnactive)
Outer arc
ridge Trench
Ocean
basin
~ .---'Kurile Arc Aleutian
Arc
North·East
Honshu Arc

<0·
SOUTH
CHINA
rv~~"",""" '" ) MaArianas
rc

SE~ ~o Philippine PAC I F I C

Sulu(J Arc . Heat Flow
(mWm-2 )
Aro OCEAN
/ J~
Celebes
A r c ( " Halmahera
_---~ ~120
o· ,

Band;'\
Arc
"--------1 40

Arc . / Fig.3.7 Schematic representation of the processes operating

••• ~ t:? during the formation of intra-oceanic island arcs and the major

~
New
Hebrides
morphological elements common to most arcs.
Arc Tonga
) Arc
~

AUSTRALIA
Kermadec of the downgoing lithosphere. This is related to the angle of
/ Arc subduction and the depth at which melting of the subducted

!J~
iithosphere becomes significant. With typical angles of
NEW subduction of 30-50° this means that there is a gap of some
ZEALAND
75-175 km between the trench and the arc of volcanoes.
170·E 170·W
The active volcanic belt may consist of a single chain of
volcanoes or form a broad belt up to 150 km in width. At
Fig. 3.6 Location of intra-oceanic island arcs in the western anyone time, however, volcanic activity is generally con-
Pacific Ocean. Other intra-oceanic arcs are the Lesser Antilles in fined to a zone no more than 50 km across.
the Caribbean, the South Sandwich Islands (Scotia Arc) in the
South Atlantic and the Andaman and Nicobar Islands between
Burma and Sumatra (see Figure 3.9). 3.2.2 Morphological and structural elements
If we look at intra-oceanic island arcs in more detail it is
As the plate descends frictional heat is generated between possible to identify several major morphological and struc-
its upper surface and the surrounding mantle. Heating pro- tural elements that are common to many (Fig. 3.7). The
bably begins at quite shallow depths but because the down- fundamental division is into the forearc, volcanic-arc and
going plate is cold virtually all of this heat is initially ab- back-arc zones. Beginning in the forearc zone and moving
sorbed. This is why low heat flows are recorded at the towards the volcanic arc the first feature encountered is the
surface immediately behind the trench (Fig. 3.7). At a depth trench which rises to the outer-arc ridge. This is located over
of about 100 km melting begins to occur along the edge of the subduction zone and consists largely of thrust sheets of
the subducted lithosphere and heat begins to be transferred sediment and slivers of oceanic crust, known as ophiolite,
upwards by rising pockets of magma. The dramatic surface emplaced by the subducted plate. The outer-arc ridge is usual-
expression of this marked increase in heat flow is the for- ly a submarine feature but in a few cases, such as Barbados
mation of volcanoes along the arc although relief is also in the West Indies and Middleton Island in the Aleutian Arc
generated by the emplacement of large igneous intrusions in Alaska, parts of it are. exposed above sea level. Between
(see Section 5.1) in the underlying crust. The most common the outer-arc ridge and the volcanic arc a forearc basin may
form of volcanic activity in this zone is the eruption of be formed. The volcanic-arc zone itself may consist of two
andesitic lavas, although there is a broad compositional elements: the active volcanic arc and an inactive frontal arc
range from basalt to rhyolite. located on its forearc side and composed of older volcanic
Fairly consistent patterns are evident in the location of rocks. The whole entity is sometimes referred to as a mag-
the volcanic activity with respect to the fringing trench. The matic arc. In most island arcs only a relatively small pro-
volcanoes tend to be located between 80 and 150 km verti- portion of the individual volcanoes actually rise above sea
cally 'above the Wadati-Benioff zone marking the position level.
62 Endogenic processes and landforms

A complicated pattern of basins and ridges is located be- Back arc

Fold- Thrust
hind the vo'lcanic arc. This back-arc zone may simply consist Forearc Basin Arc Belt

of a large marginal basin separating the intra-oceanic island

arc from an adjacent continent, or it may be a more complex
feature containing one or more remnant arcs, separated from
the volcanic arc and each other by interarc basins. All these /:~: ttJ :.:.:.:.:.: : : : : : : : : :.'.

~'I'
structures are submarine.
o 100
A number of processes have been proposed to explain the I
km
I

OCEANIC
formation of back-arc basins but no one mechanism seems LITHOSPHERE
to be applicable to all cases. One favoured process is known
Fig.3.8 Main structural and morphological features of a
as back-arc spreading and this has been applied to several
continental-margin orogen. The type illustrated is the
such basins. Japan, for example, is composed of relatively 'contracted' form (see text). (After W. R. Dickinson (1977) in:
thick continental-type crust, some of it quite old (Early M. Talwani and W. C. Pitman III (eds) Island Arcs, Deep Sea
Palaeozoic), rather than the more usual modified oceanic Trenches and Back-Arc Basins. Copyright by the American
crust of most other intra-oceanic island arcs. This is diffi- Geophysical Union, Washington, DC, Fig. 5, p. 38.)
cult to explain unless Japan has been detached from the
eastern margin of the Eurasian continent and has migrated sive vertical development and in some cases rise above sea
eastwards. The process thought to be responsible is broadly level. The Mentawai Islands off the west coast of Sumatra
analogous to sea-floor spreading along mid-oceanic ridges. represent the highest parts of the outer-arc ridge of the Sunda
Indeed, linear magnetic anomalies have been identified be- Arc and this structure can be traced northwards as far as the
hind a number of island arcs; those behind the Scotia Arc in Indoburman Ranges in Burma (Fig. 3.9). Parts of the coastal
the southern Atlantic Ocean, for instance, indicate that ranges of Mexico, Peru and Chile similarly represent outer-
spreading has occurred here over the past 7-8 Ma. Back-arc arc ridges. Volcanicity and magma emplacement, largely in
spreading may be initiated within the volcanic arc because
the high heat flux from below weakens the lithosphere. If the
lithosphere of the back-arc zone has a component of motion '_I
I<q (
away from the volcanic arc relative to the underlying asthe- /8, ~ Strike-slip
nosphere the arc may be split apart. This eventually leads to 'co,
Ii, .. " fault
..... .....
the creation of a remnant arc separated by an interarc basin I~\ \ " Outer-arc
\ 'Z- \ \ ridge
from an active volcanic arc. \:V~\ .. ,\\ SHAN .... Volcano
0\
~\ ..
,..........
\
,BURMA
FAULT
o I
km 500
I

3.3 Continental-margin orogens and continental-

margin island arcs
,
, I

Continental-margin orogens and continental-margin island

arcs are broadly similar to intra-oceanic island arcs in their
tectonic setting. The differences between them arise from the
presence of continental crust on the margin of the overrid-
ing plate. This influences not only the nature of volcanic ac-
tivity and landforms produced but also the processes tending
to initiate back-arc sprea<hng behind the volcanic arc. Con-
tinental-margin orogens are found along the western edge
of South America (the Andes) and along part of the west
coast of North America (the Cascade Mountains of Oregon
and Washington, USA.). The western part of the Sunda Arc
MENTAWAI
in the East Indies is an example of a continental-margin ISLANDS --*-~

island arc since its back-arc zone comprises drowned con-

tinental crust (the Sunda Shelf). 65mma-1

As with an intra-oceanic island arc the oceanward side of Rate and direction
of convergence of
a continental-margin orogen is marked by an oceanic trench the Indian Plate
associated with the downgoing oceanic lithosphere (Fig. 3.8).
An outer-arc ridge, formed of wedges of deformed oceanic Fig.3.9 Map of the major structural and morphological
sediment and slivers of oceanic crust in addition to sediment elements of the eastern margin of the Indian Ocean (based on
originating from the adjacent continent, may undergo exten- various sources).
 Landforms and tectonics of plate margins 63

the fonn of large granite intrusions, characterize the volcanic

arc. This is usually located between about 175 and 275 km
from the trench, a rather greater distance than is typically
the case with intra-oceanic island arcs.
The sequence of continental-margin orogen development
begins with the subduction of oceanic lithosphere at, or close
to, a continental margin. Frictional heat generated along the
upper edge of the underthrusting lithospheric slab at depths
exceeding 100 km leads to the upwelling of magma in an
expanding dome within the growing orogenic belt which
eventually rises above sea level. As the elevation of the oro-
gen increases, the presence of hot, mobile rock within its
core promotes gravitational spreading and sliding (see
Section 7.3 ) through upwelling and lateral flowage into the
more rigid outer crust. Eventually granite intrusions are em-
placed at high levels within the mountain mass and volcanic
activity develops. Extensive stretching of the crust above
these granite intrusions produces a faulted terrain with active
volcanoes. Crustal thickening, due in large measure to the
vast volume of granite emplaced above the subduction zone,
is considerable in some continental-margin orogens. For in-
stance, crust up to at least 70 km thick occurs below parts
of the Andes .
Some continental-margin orogens show evidence of defor-
mation attributable to contraction in the back-arc region.
They appear to be jammed against the adjacent cratonic
block and are known as contracted continental-margin
orogens. In other orogens, such as the coastal ranges of the.
north-west USA, this feature is absent and the tenn non-
contracted continental-margin orogen is applied. In con-
tracted continental-margin orogens the back-arc region, Fig.3.10 A snow-capped range separating sediment-filled
known as a retroarc basin, is separated from the rear of basins in thefau/ted Basin and Range Province, Nevada, south-
the volcanic arc by a system of down-thrusted and folded west USA.
masses of rock (fonning a back-arc fold-thrust belt)
(Fig. 3.8). These features develop if the back-arc litho-
sphere behind the orogen is moving towards it relative to arc spreading is analogous to that encountered behind some
the underlying asthenosphere, since this convergence will intra-oceanic island arcs. However, the Pacific-North
probably lead to partial subduction of the back-arc litho- American plate boundary has undergone significant changes
sphere. As the lithosphere carrying the rigid cratonic crust over the past few million years with the subduction of part
moves towards, and underthrusts, the continental-margin of the East Pacific Rise spreading ridge and the evolution of
orogen, the stripping of its sedimentary cover gives rise to a predominantly transform margin. It is possible that these
the back-arc fold-thrust belt. This part of continental-margin changes have been .critical in generating the considerable
orogens can therefore be viewed as a kind of subduction crustal extension (possibly up to 100 per cent) calculated for
zone, albeit one that involves orders of magnitude less con- the Basin and Range Province.
vergence that oceanic trench subduction zones.
The presence of anomalously thin crust, widespread exten-
3.3.1 The Andes
sional faulting and high rates of heat flow has led some
researchers to interpret the Basin and Range Province of the 3.3.1.1 General characteristics
western USA and northern Mexico as an area of back-arc The Andes constitute by far the most spectacular present-
spreading developed on continental crust behind the sub- day example of a continental-margin orogen. Stretching from
duction zone along the Pacific coast. The region, which as a the Toco Peninsula of Trinidad in the north to Tierra del
whole has been elevated by 1000-2000 m, has a fault- Fuego in the south, they fonn a varied and complex mountain
generated relief of up to 4000 m comprising uplifted moun- system some 9000 km long with elevations in excess of
tain blocks and intervening basins (Fig. 3.10). Such back- 5000m. Three major divisions can be recognized (Fig. 3.1-1):
64 Endogenic processes and landforms

In the range as a whole block faulting is significant and

.
Caribbean Arc vertical movements predominate over thrusting and other
effects involving laterally induced stress. Any plate tectonics
model of the Andes must in fact account for the uplift es-
sentially in terms of vertical tectonics. Another feature which
has to be explained is the emplacement of massive granite
intrusions of various ages which cover about 464 000 km 2
(about 15 per cent) of the surface of the range; this probably
represents a volume in excess of 2x106 km 3 - the greatest
mass of granite intrusions on Earth. They are concentrated
in the Coastal Cordillera of the central Andes but also occur
~i>r;.e
in the High Cordillera (Fig. 3.11). Volcanic activity is
~'<);.c.~" localized within the range; in the high plateaus of the central
.- Andes in the border area of Chile, Argentina and Bolivia
~ Coastal Cordillera
where the range reaches its greatest breadth, layers of lithified
Western Cordillera
{High Andes in Chile l volcanic ash from 500 to 1500 m thick produced by ex-
• Central Cordillera plosive rhyolitic eruptions, cover an area in excess of
ED Puna Block 200000 km2 . Quaternary lavas have erupted through these
• Eastern Cordillera
deposits to form over 900 volcanoes ranging in elevation
from 5000 to 7000m (Fig. 3.12).
~ Sub-Andean Cordillera

Chile Ridge
-------

------
Scotia Arc
2000
,
km

Fig.3.11 Major structural and morphological divisions o/the

Andes. (Modified/rom A. Gansser (1973) Journal of the
Geological Society London 129, Fig. 2, p. 96. Reproduced by
permission o/the Geological Society.)

(1) the Southern (Patagonian) Andes, extendIng from Tierra

del Fuego northwards to the Gulf of Penas at latitude 47 oS Fig.3.12 The volcanic peak o/Cotopaxi reaching an elevation
where the actively spreading Chile Ridge extends westwards 0/5896 m in the Ecuadorian Andes. (Photo courtesy D. Munro.)
into the Pacific; (2) the Central (Chilean-Peruvian) Andes
reaching from the Gulf of Penas northwards to the Amotape 3.3.1.2 Evolution of the central Andes
cross structure at the Peru-Ecuador border where the north- In the central Andes there has probably been more or less
western trend of the mountain belt changes to a north- continuous subduction since the Mesozoic. Consumption of
easterly orientation; and (3) the northern (Colombian- the Nazca Plate along the nascent Peru-Chile Trench had
Venezuelan) Andes, extending from the Amotape cross begun by at least the Late Triassic - Early Jurassic and
structure northwards and then eastwards to eventually link marked the conversion of the previously 'passive' continental
with the Caribbean Arc. margin (see Chapter 4) of western South America to an active
Since the development of the plate tectonics model, the convergent plate boundary (Fig. 3.13 (A) and (B)). The
Andes have sometimes been regarded as representing a start of subduction may have been associated with the
classic example of the orogenic consequences of the sub- initiation of spreading on the East Pacific Rise, but in any
duction of oceanic lithosphere beneath an overriding conti- case it certainly preceded the splitting of South America
nental plate. This, however, is an over-simplified view as the away from Africa 135-140 Ma ago. This early phase of plate
current phase of uplift and volcanicity represents only the interaction was probably marked by the development of a
most recent episode in a complex history stretching back to volcanic arc off the north Chile coast and was apparently
the Mesozoic; moreover, the history of this vast mountain contemporaneous with minor emplacement of granite in-
system differs greatly from one part to another. trusions in the Eastern Cordillera.
 Landforms and tectonics of plate margins 65

A two ranges a large structural basin was created which now

Coastline Sedimentary basin
forms the high plateau of the Altiplano. The modem phase
of orogeny began in the Miocene with volcanism and the
further emplacement of granite intrusions. A more active
period of igneous activity in the Pliocene - Quaternary,
involving massive emplacements of magma, caused intense
folding and the formation of large thrust sheets in the Eastern
About 300 Ma ago Cordillera. A series of high, narrow ranges was formed as
B the Eastern Cordillera were pressed against the resistant block
of the Brazilian Shield to the east (Fig. 3.13 (D». This folded
and thrusted belt probably represents the partiai' subduction
of the western fringe of the Brazilian Shield, and the major
change of the orientation of the Andes at the Amotape cross .
structure seems to be the result of the movement of separate
basement blocks in the back-arc zone.
There appears to be a clear distinction between the causes
of uplift on the western and eastern sides of the central
Andes. In the Western Cordillera evidence of tensional de-
c Formation of
formation indicates that crustal stretching has been associ-
Development of Altiplano Development of
Western Cordillera structural basin Eastern Cordillera ated with the emplacement of magma and volcanic activity
I related to subduction zone melting. In the Eastern Cordillera,
however, compressional stresses apparently generated by the
growth of the volcanic arc to the west have caused crustal
shortening, and uplift here is attributable to folding, thrusting
and crustal thickening. Nevertheless, for the Andes as a
whole, compressive stresses arising from plate convergence
have not been the major cause of uplift.

o 3.3.2 The Sunda Arc

The islands which make up the East Indies archipelago form
a region of fascinating complexity. For the present, however,
we will confine our attention to the Sunda Arc comprising
Sumatra, Java and t~e chain of smaller islands extending
eastwards as far as longitude 123°E (Fig. 3.14). To the
west of this point oceanic lithosphere of the Indian Plate is
currently being subducted below the Sunda Arc but to the
east the Arc is in collision with the continental shelf of
northern Australia (this is marked on Figure 3.14). There is
- M - Moho
EJ Pre-Palaeozoic rocks thus a transition from a steady-state boundary in the west to
Ej Volcanic rocks a collision boundary in the east. Since the western part of the
(j) Granite intrusions
~ Sedimentary rocks Sunda Arc is located on the edge of a block of continental
Fig.3.13 Model of the evolution of the central Andes. (Based crust (at present largely below sea level) it is a continental-
partly on D. E. James, (1971) Geological Society of America margin island arc rather than an intra-oceanic island arc.
Bulletin 82, Fig. 10, p. 3341.) Across the Sunda Arc there is a distinct sequence of
oceanic trench, outer-arc ridge, forearc basin, volcanic-arc
A major episode of mountain building occurred in the Late and back-arc zone (Fig. 3.9). The volcanic arc is located
Cretaceous-Early Cenozoic to the east of the earlier orogeny about 125 km above the Wadati-Benioff zone representing
(Fig. 3.13(C)). This was the major period of emplacement the seismic activity along the downgoing Indian Plate. This
of granite intrusions which resulted in the growth of the maintains a shallow angle of subduction beneath the outer
Western Cordillera. A further consequence of this large arc but steepens below the main volcanic arc to about 60°.
addition of granite to the core of the mountain mass was the Beneath Sumatra the subduction zone does not exceed a
transmission of compressive stresses to the east where the depth of 200 km, but north of Java it reaches 400 km and
Eastern Cordillera were folded and uplifted. Between these north of Flores, at the extreme eastern end of the zone of
66 Endogenic processes and landforms

South Philippine "'" fj

China Sea "f!
Sea 4,)'''

J
,;)~"'J.:~Q
.....", v
", <:'
Q. ...0
.....Caroline
..... Plate

Outer-arc Arc reversal Arc colliding Miocene arc collision

ridge exposed now underway with continental N.W. New Guinea
in Mentawai north of Timor shelf from here rotates away from
Islands and Sumba eastwards advancing arc

)(

\
~
Subduction
zone
'C Outer-arc
\ . ridge
-\'
A
A Volcanoes
\ Spreading
centre

Fig 314 The tectonic and morphological elements of the Sunda Arc and adjacent region. (Modifiedfrom W. Hamilton (1977! in: M .
Tal~a~i and W.C. Pitman III (eds) Island Arcs, Deep Sea Trenches and Back-Arc Basi.ns. PUblis~ed by the ~merican Geophyslcal Umon,
Washington, DC, Fig. 3, p. 18 and W. B. Hamilton (1988) Geological Society of Amencan Bulletm, 100, Flg. 2, p. 1511.

steady~state subduction, it attains a depth of 700 km. Sumatra 3.4 Intercontinental collision orogens
is composed of old, thick continental crust comprising
Although crustal collisions may involve intra-oceanic as
.volcanic rocks of Permian, Cretaceous and Cenozoic age.
well as continental-margin island arcs, it is the convergence
Recent volcanic rocks, including extensive sheets of volcanic
and' eventual collision of continental crust that gives rise to
ash, are predominantly aCidic and rnteimediate in compo-
intercontinental collision orogens (Fig. 3.15). Such mountain
sition since the magmas have risen through continental crust.
belts develop when the oceanic lithosphere originally lying
By contrast in Java, where the crust is relatively young and
between two continents is eventually consumed. The pre-
thin the volcanic rocks are mostly intermediate in compo-
viously existing subduction zone is converted into a suture
siti;n and the large ash sheets indicative of highly explosive
zone marking where the two continents are welded together.
eruptions are scarce (see Chapter 5). An embayment of oceanic crust not consumed before the
The outer-arc ridge of the Sunda Arc is of particular
boundary becomes 'locked' is called a remnant-ocean basin.
morphological significance. Although it is from 1 to 3 km
A peripheral foreland basin is developed on the margin
below sea level off Java, along most of the coast of Sumatra
of the underthrusting plate roughly parallel to the strike of
parts of it are exposed as the Mentawai and other islands the suture belt, either on a remnant-ocean basin covered by
(Figs 3.9 and 3.14). The height of the ridge is apparently
a thick layer of sediment, or on true continental crust. A
related to the amount of sediment being moved into the ad-
retroarc foreland basin occurs on the overriding plate behind
jacent subduction zone where lithosphere is being subducted
the volcanic arc previously developed in association with
~t a rate of about 65 mm a-I. Large quantities of sediment are
the earlier subduction of oceanic lithosphere.
present on the ocean floor off Sumatra because of its pos-
ition with respect to the Bengal Fan which represents the
3.4.1 General sequence of development
vast submarine extension of the Ganges delta on the Indian
subcontinent some 2000. km to the north (Fig. 3.9).. The initial impact occurs between the outer-arc ridge of the
 Landforms and tectonics of plate margins 67

Volcanic arc the subduction zone and uplift ensues (Fig. 3.16). Mountain
Peripheral foreland ranges formed at the points of initial continental collision
basin
Suture belt Retroarc foreland
provide the sediment which accumulates in adjacent remnant-
basin ocean basins (Fig. 3.16(B». On reaching the subduction zone
this sediment is scraped off and stacked into thrust sheets.
Final closing of a remnant-ocean basin results in the thrusting
of remaining sediments deposited along the continental
margin on to the overriding plate and the crustal thickening
produced leads to isostatic uplift (Fig. 3.16(C».
Large-scale crustal shortening is often present in inter-
continental collision orogens and this may be accommodated
by folding (Figs 3.17, 3.18), or by the limited subduction of
continental crust during the initial phase of collision follow-
ing the previous episode of steady-state subduction. There
Fig. 3.15 Major structural and morphological elements of an
intercontinental collision orogen. appear to be two possible mechanisms to account for this.
Since the mantle part of the descending lithosphere is cold
relative to the asthenosphere it may be sufficiently dense to
overriding continental margin and promontories on the pull the lower part of the crust down if the upper part is de-
leading edge of the continent entering the subduction zone. tached. Alternatively, subducted oceanic crust may be able
The first sections of the converging continental margins to to pull adjacent continental crust down into the astheno-
collide suffer the most intense deformation. Wedges of sphere. Depending on the assumptions applied this process
oceanic crust are thrust up on to the overlying sediments of could lead to the subduction of several hundreds of kilo-

Continental- margin
orogen deformation of sediments at
foot of inner wall of trench continental margin

II oceanic crust

Q continental crust lithosphere

OJ] mantle
lithospheric

Fig. 3.16 Schematic representation of the development ofan intercontinental collision orogen. (After J. F. Dewey and J. M. Bird, (1970)
Journal of Geophysical Research 75, Fig.l3,p. 2642. Copyright by The American Geophysical Union.)
68 Endogenic processes and landforms

metres of continental crust; more likely, however, is rather rock masses under the force of gravity along the flanks of
more limited subduction, but even this could involve the the uplifted orogen (see Section 7.3). Accumulation of sedi-
complete consumption of peninsulas and microcontinents. ment through the stacking of thrust sheets in the suture zone .
Partial subduction of the underthrusting plate continues is probably the major cause of subsidence of the underlying
until frictional forces and the buoyancy of the continental subducted plate. Plate descent itself does not seem to be a
crust exceed the driving force of plate motion (Fig. 3.16 major factor as isostatic rebound does not appear to occur
(D)). Cessation of subduction is eventually followed by a after subduction has stopped. A peripheral foreland basin
halt to volcanic activity on the overriding plate. Further begins to develop as the continental crust of the underthrust-
thrusting may develop at this stage through sliding of large ing plate is bent just before entering the subduction zone.

Fig.3.17 Landsat scene ofpart of the southern Zagros Mountains, Iran. Sediments ofMesozoic and Early Cenozoic age laid down on a
continental shelf have been folded into a succession ofplunging anticlines as a result of compression arising from intercontinental
collision. The sediments contain thick salt beds which in places have intruded overlying strata as a result of their lower density and
reached the surface to form salt fountains up to 1 km high (small, dark patches on image) The salt spreads under its own weight at rates of
a few m a-I forming salt flOWS. (Image courtesy N. M. Short.)
 Landforms and tectonics of plate margins 69

After the phase of partial subduction and the accumulation identified by various lines of evidence such as contrasting
of a considerable thickness of sediment, possibly up to pre-collision geological histories either side of the postulated
5000 m, the peripheral foreland basin may become a site of collision zone, the presence of deep ocean floor sediments
marked uplift. and associated volcanic rocks, and preserved remnants of'
Once the two continents have been welded together and andesitic lavas erupted from the continental-margin orogen
active volcanism and tectonic uplift have ceased, a new plate located on the originally overriding plate prior to collision.
is created in which the orogenic belt formed becomes an In inter-continental collision orogens major sustained hori-
intra-plate feature no longer associated with an active conver- zontal stresses can be generated which give rise to recumbent
gent plate boundary. A number of ancient 'suture zones rep- folds. Moreover, the back limbs of such folds may become
resenting previous plate collisions have been recognized but detached along major thrust faults to produce nappes. The
only a few of these, such as the Urals and the Appalachians classic region for these kinds of structures is the European
formed 300 to 250 Ma HP, are major relief features at the Alps. Here the convergence of the African and Eurasian
present day, most having been extensively eroded and cover- Plates has led to the development of successive thrusts and
ed by later sediments. Such ancient suture zones may be nappes carrying huge masses of rock northwards towards

Fig.3.18 Limb of the Saidmarreh anticline, southern Zagros Mountains, Iran. This plunging anticline is experiencing active uplift and
although breached by a major stream has retained much of its primary form. (Photo courtesy Aerofilms Ltd.)
 70 Endogenic processes and landforms

Dinaric
Pre-Alps High Calcareous Alps Alps

o 50 100 150 Horizontal distance (km)

Fig.3.19 Highly simplified cross-section from the piedmont plain of northern Italy north-westwards to Lake Geneva illustrating the
relationship of major nappes to Alpine topography. (Modified from L. W. Collet (1935) The Structure of the Alps (Edward Arnold,
London) Plate Xl after E. Argand, 1916, Ecologae Geologicae Helveticae 24,145-91).

the European foreland (Fig. 3.19). Horizontal movements distance of perhaps 100 kIn or more. Separation of the upper
along these major thrusts can exceed 100 kIn and rock thick- third of the continental crust on a plate has the effect of
nesses of several kilometres can be involved. The thrust plane reducing the overall buoyancy of the subducted crust by
itself represents a zone of detachment (or decollement) nearly one-half, and this in tum allows more subduction than
along a bed of salt or other incompetent lithology. However, would otherwise have occurred. The eastern Alps have been
detailed investigations of these structures suggest that much suggested as an example of such 'flake tectonics' since this
of this large-scale lateral movement is accomplished by the mechanism accounts for the existence of a 'platelet' in the
spreading or sliding of rock masses under gravity (see Alps which is much thinner than normal and dips in a
Section 7.3). southerly direction (Fig. 3.20(B».
Such a structure is particularly likely to form where a
promontory on the advancing edge of the continent on the
3.4.2 Pre-collision history and configuration of subducting plate collides with the opposing continent well
converging continental margins
before the rest of the continental margin (Fig. 3.20(E), (F».
Although it is possible to make some generalizations about Convergence associated with the continuing subduction of
the likely sequence of development of an intercontinental the remaining oceanic lithosphere along the plate margin
collision orogen, the pre-collision history of the active mar- would be maintained until the plate boundary had been fixed
, gin of the overriding plate and its subduction zone has an at a sufficient number of points to resist further subduction.
important influence on the subsequent evolution of an Salients of oceanic crust may be left along the suture zone
orogenic belt. The margin of the overriding plate is constantly and parts of the Mediterranean Sea represent remnant ocean
being modified by the subduction of oceanic lithosphere basins formed in this way.
beneath it; volcanic material is added, sediments accumulate
and the margin may be broken by strike-slip faults. The most
important of these variables is probably the heating of the 3.4.3 The Himalayas
plate margin since this will affect its thickness and strength.
Thick, strong continentai-margin crust will be associated with 3.4.3 .1 Morphological and structural elements
either slow or recently initiated subduction since in neither The Himalayan intercontinental collision orogen extends in
case would there be sufficient heating for melting of the a southward-bending arc, some 200-250 kIn across, for over
upper surface of the downgoing lithosphere. A marked con- 2500 kIn from the Indus River in the west to the Brahmaputra
trast in behaviour would be expected during continental River in the east (Fig. 3.21). It is separated from the
collision between the very thick crust of the Andes, where Trans-Himalayan zone and the Tibetan Plateau to the north
relatively rapid subduction has occurred throughout at least by the structurally controlled valleys of the Indus and the
the Cenozoic, and the Alpine region where the closure of Brahmaputra (in its upper reaches called the Tsangpo) and
the ocean separating North Africa and southern Europe has to the south it is bounded by the sediment-filled peripheral
been more gradual. foreland basin of the Indo-Gangetic plain. Distinct morpho-
The configuration of continental margins and the angle at tectonic zones can be recognized across the orogen along
which they converge also affects significantly the nature of its entire length, although their detailed interpretation is
the orogenic belt that develops (Fig. 3.20). One plate may disputed (Fig. 3.22). The two major ranges, the Lesser
simply override the other, but another possibility is that a Himalayas and the Higher Himalayas, both of which exceed
'flake' of continental crust from the upper surface of the 8000 m in height, are bounded by three major linear struc-
downgoing plate may be thrust over the adjacent plate for a tures: from south to north these are the Main Boundary
 Landforms and tectonics of plate margins 71

Lak~Baikal
Baikal
TT'I Rift

" ....

Fig. 3.21 The Himalayas and Tibetan Plateau in relation to the

major structural elements of central Asia (based on various
sources).

INDIAN PLATE: EURASIAN PLATE

ssw -:- , NNE
Himalayas:
,
Sub- Lesser- Higher-: Trans-
--~ ,
'---
A :/Indus-Tsangpo Suture Zone

• Oceanic crust D Continental crust

Fig.3.20 Various ways in which the collision between the

opposing margins of two continents may be accommodated. Main Boundary Main Central
Shown are: (A) simple overriding; (B) flake formation (e.g. Thrust \ ThJust IndUrTSangpO Suture Zone
eastern Alps); (C) a detachedflake; (D) a deep flake with B

~~7
separating fractures extending below the Moho (e.g. central
Alps); (E) a three-dimensional view offlake formation along an
irregular continental margin; (F) flake formation associated with M'"' - ;;; I

f_ ---- 7/
oblique subduction and consequential development of strike-slip
faulting. (Modified from E. R. Oxburgh (1972) Nature 239, Figs 2
and 3, p. 204 (E), and E. R. Oxburgh (1974) Proceedings of the
Geologists Association 85, Fig. 26, p. 344 (F) and Fig. 27, p. 345
Fig. 3.22 Two interpretations of the tectonics of the Himalayas
(A-D).)
according to Le Fort (1975) (A), and Powell and Conaghan
(1973) (B). (Modifiedfrom P. Le Fort (1975) American Journal of
Science 275A, Fig. 15, p. 38).
Thrust, the Main Central Thrust and the Indus-Tsangpo suture
zone. The belt of ophiolites following the Indus and Tsangpo
valleys marks the boundary between the Eurasian Plate and 180 mm a-I. Subsequently, movement has been much slower
the Indian Plate and to the north the presence of acidic at only about 50 mm a-I. This reduction in velocity seems
volcanic rocks indicates the subduction of oceanic litho- to be related to the timing of the contact between India and
sphere prior to the collision of the Indian and Eurasian land Eurasia. Collision along what is now the Indus-Tsangpo
masses. suture zone seems to have occurred from about the Late
Paleocene until the Early Eocene, or possibly a little later.
3.4.3.2 Models of development Only limited continental subduction appears to have occurred
From the pattern of ocean floor magnetic anomalies it is at this stage. Significant vertical uplift probably began in the
possible to trace the movement of India northwards over the Oligocene about 35 Ma ago and has continued to the present
past 80 Ma or so (Fig. 3.23). During the period 80 to time, but at varying rates.
50 Ma BP it drifted rapidly at a rate of between 100 and Although clearly an intercontinental collision orogen, the
72 Endogenic processes and landforms

The most significant development during this period was

the formation of the Central Gneiss Zone, some 10-20 km
broad and dipping at an angle of 3~0° to the north and
bounded to the south by the Main Central Thrust (Fig. 3.22).
The development of this zone appears to be related to the
major underthrusting and associated metamorphism of local
rocks that has apparently occurred along the Main Central
Thrust some lO0-200 km to the south of the Indus-Tsangpo
suture zone. This underthrusting began in the Early Miocene
and contributed to the doubling of crustal thickness (up to
80 km in places) and the impressive subsequent uplift of the
Higher Himalayas. Detailed seismic evidence indicates that
the underthrusting has been at an angle of about 15 0 over a
distance of about 300 km; perhaps significantly this roughly
corresponds to the average width of the Himalayas.
The precise relationship between the Main Central Thrust
and the intense metamorphism of the Central Gneiss Zone
of the Higher Himalayas remains somewhat u;ncertain.
Some researchers have suggested that the metamorphism
preceded thrusting and in fact actually promoted it by locally
weakening the crust. Another interpretation is that metamor-
phism occurred at the same time, or post-dated thrusting,
and that the frictional heating produced during thrusting
contributed to the partial melting of the crust and the intru-
sion of granitic rocks. A further explanation involves the
decoupling of the crust from the underlying lithosphere along
the northern margin of the Indian Plate and the exposure of
the base of the crust to hot asthenosphere. This could have
led to the upwelling of magma, granite emplacement and
isostatic uplift through the substitution of hot, low density
mantle from the asthenosphere for the cold, dense mantle of
the upper lithosphere. A further consequence of this heating
could have been a weakning of the upper layers of the crust
Fig. 3.23 The northward drift ofIndia over the past 80 Ma
and the formation of the northward dipping Main Central
reconstructed from palaeomagnetic data. The position of Eurasia Thrust along its southern edge.
is assumed to be fixed in its present location and the configuration The extent of subduction of continental crust below the
of the Indian sub-continent is drawn in its presentform (although Himalayas is also in dispute. At one extreme underthrusting
the shape of the northern boundary prior to collision is unknown). is seen as proceeding for up to 1500 km at the base of the
(Based on A. G. Smith, A. M. Hurley and 1. C. Briden (1981)
Phanerozoic Paleocontinental World Maps. Cambridge
continental crust through the peeling off of the lower part
University Press, Cambridge Map 2, p. 9, Map 10, p. 17, Map 14, of the lithosphere of the Eurasian Plate (Fig. 3.22(B». It is
p. 21, Map 18, p. 25 and Map 22, p. 29.) more generally accepted, however, that the buoyancy of the
continental crust has limited the extent of continental under-
present elevated form of the Himalayas is not a direct con- thrusting and that this is why the more recent Main Boundary
sequence of continental collision; rather it reflects develop- Thrust dipping under the Lesser Himalayas has now taken
ments that have occurred since the initial impact some over the role played earlier by the Main Central Thrust (Fig.
50 Ma BP. Subduction of the oceanic lithosphere of the lead- 3.22(A». It seems likely that yet another northward-dipping
ing edge of the Indian Plate prior to collision may have been thrust will develop even further to the south in a few million
associated with the emplacement of large granite intrusions years.
as well as volcanic activity in southern Tibet. Little relative Although the rate of movement between India and Eurasia
movement between the Indian and Eurasian Plates seems to fell markedly after their initial collision some 50 Ma ago,
have occurred between the Middle Eocene and Early Miocene, convergence of the two land masses since then may have
but isostatic readjustment of the presumably thickened crust amounted to as much as 2500 km. This represents a major
resulted in a withdrawal of the sea on the adjacent Indian problem because such an amount of continued convergence
continental lowland. between two continents after collision is not easily accom-
 Landforms and tectonics of plate margins 73

modated within a simple plate tectonics model. The extent that it implies (5° or less). This contrasts with the observed
of crustal shortening required implies a vast degree of under- northwards dip of the Main Central Thrust at about 15°.
thrusting by continental lithosphere, extensive crustal com- Moreover, there is little seismic evidence for the intermedi-
pression, crustal deformation over an extensive area behind ate and deep earthquakes below the Himalayas that would be
the suture zone, or a combination of two or more of these expected if active subduction is, in fact, occurring.
effects. Critical to an evaluation of these various hypotheses Of the estimated post-collision convergence of up to
is an understanding of the development of the high Tibetan 2500 km about 500 km could be accounted for by under-
Plateau lying to the north of the Himalayas. thrusting beneath the Himalayas and southern Tibet, another
200-300 km by thrusting and crustal thickening in the various,
3.4.4 The Tibetan Plateau ranges bordering the northern perimeter of the Tibetan
Plateau (such as the Pamir, Tien Shan, Altai and Nan Shan
3.4.4.1 Morphology and structure ranges) and perhaps a further 300-400 km by the crustal
The Tibetan Plateau, located between the Himalayan and shortening arising from gentle folding within the Tibetan
Karakorum ranges to the south, and the Kunlun and Altyn Plateau itself. This leaves around 1400 km of post-collision
Tagh ranges to the north, is a roughly triangular area some convergence to be accounted for, much of which might be
1000 km from north to south and 1700 km from east to explained by a doubling of the thickness of the Tibetan crust
west (Fig. 3.21). The local relief is relatively subdued but during the past 50 Ma from a 'normal' value of 35 km to the
the mean elevation is some 5000 m with isolated masses of present 70 km. But not all the remaining convergence can be
probable volcanic origin rising to between 6000 and 7000 m; explained by this process and it has been suggested that the
the Tibetan Plateau is thus by far the highest plateau of rest has been accommodated by the lateral movement of
significant extent on the Earth's surface. To the east the blocks of continental lithosphere along east-west trending
elevation falls and several major rivers such as the Mekong,
Salween, Chiang Jiang (Yangtze) and Huang-He (Yellow)
drain in an easterly and southerly direction from the
Tibetan Plateau eventually reaching the Yellow and South A
China Seas.
The Tibetan Plateau has been an unstable area since at
least the Late Palaeozoic. It seems to have formed through
the successive accretion of continental and island arc frag-
ments (see Section 3.6) from then until the middle Cretaceous
and at the time of its collision with 'the IndIan Plate it was
probably an area of relatively warm and weak lithosphere.
The presence of shallow marine sediments of Late Cretaceous '
age indicate that the region was below sea level at this time
and it apparently did not emerge until the Early Cenozoic.
Late Cenozoic volcanism, which is extensive over much of
the area, and low seismic wave velocities together suggest
that the lithosphere below the Tibetan Plateau is still
unusually hot and is therefore probably weaker than in ad-
jacent areas. The crust is now also abnormally thick (about
70 km) but gravity measurements indicate that it is in ap-
proximate isostatic equilibrium.

3.4.4.2 Models of development

B A
A number of explanations have been put forward to account - - - - -__ Tibetan ...- - - - - Tien Shan
Himalayas Plateau Tarim Basin Folding
for the thick crust and high mean elevation of the Tibetan
Plateau which also help to explain the great crustal shorten- ~_U_P~.l-lif_t / [ Stable ] --. :

ing indicated by the post-collision convergence of the Indian

and Eurasian Plates. One model proposes crustal thickening
through large-scale underthrusting by the leading edge of the
Indian continent (Fig. 3.22(B», an idea first put forward as
long ago as 1924 by the Swiss geologist, Emile Argand. A Fig. 3.24 Schematic representation of the possible effects of
major problem with the extent of underthrusting required post-collision convergence between the Indian and Eurasian land
(in excess of 1000 km) is the very low angle of subduction masses based on the model ofP. Molnar and P. Tapponnier.
74 Endogenic processes and landforms

strike-slip faults to the north of the Tibetan Plateau in strike-slip movements have recently been recorded.
Mongolia and western China (Fig. 3.24). Between 500 and The hypothesized eastward movement of China towards
1000 kIn of north-south crustal shortening could be ac- the Pacific would be expected because the continental litho-
counted for by an equivalent amount of lateral crustal move- sphere of Eurasia would provide more resistance to lateral
ment arising from the 'ploughing' motion of the Indian Plate movement than the subduction zones of the western Pacific
as it moved northwards and displaced lithospheric blocks in margin. Deformation of the weak Tibetan crust and flow in
the Eurasian Plate. Supporting evidence for extensive east- the upper mantle may have resulted from the north-south
west movement comes from the identification of a number squeezing generated by plate convergence. The Tibetan
of major faults clearly visible on satellite imagery, most Plateau can perhaps be viewed as the 'pressure gauge' of
notably the Altyn Tagh Fault (Fig. 3.25) which is com- Asia, with the pressure applied by the Indian continent
parable in length to the San Andreas Fault in California. against Eurasia maintaining its considerable elevation.
The amount of displacement along these faults is unknown, Moreover, the Tibetan Plateau transmits further pressure to
but several major earthquakes associated with horizontal the stable Tarim Basin to the north and this in turn pushes

Fig.3.25 Landsat image of the Altyn Tagh Fault north of the Kunlun Shan Mountains. Left-lateral movement along the fault is evident
from the offsetting of stream channels which cross it (arrowed). (Image courtesy N. M. Short.)
 Landforms and tectonics of plate margins 75

northwards to produce thrusting and uplift in the Tien Shan volcanic arc, both in the Sunda Arc and to the north in
range (Fig. 3.24). The association of thrust faulting on the Burma, movement is predominantly along transform faults.
margins of the Tibetan Plateau with normal faulting in its These extend into the back-arc basin of the Andaman Sea
highest regions has been interpreted as suggesting that it where there is evidence of some divergent motion. Where
has attained its maximum elevation, and that consequently the transform faults occur on land they are associated with
it is tending to grow outwards rather than increase in significant topographic features such as the valley of the
altitude. This may be because lower stresses are required for Irrawaddy River in Burma.
the formation of new thrust faults on its margin than are In order to emphasize the pattern of stress in oblique-slip
needed to elevate the entire Tibetan Plateau. margins, the terms transpression and transtension have
Whatever explanation is applied to the post-collision con- been applied to, respectively, the convergent and divergent
vergence of the Indian and Eurasian land masses and the varieties. Although these two types of margin are associated
formation of the Tibetan Plateau, it is clear that it cannot be with distinct types of landscape, they may occur in close
accommodated in any simple plate tectonics model. To con- proximity owing to the tendency of transform faults to
sider the hypothesized eastward moving segments of litho- contain offset segments or to have a sinuous form rather
sphere in Mongolia and China as involving 'micro-plates' than being purely straight (Fig. 3.26). Divergent oblique-slip
is to employ the term 'plate' in quite a different sense from
its original definition. The apparent extension of the effects
of the collision of the Indian and Eurasian Plates to much of A B
central and eastern Asia, and possibly as far as the Baikal Tra nstension Fault termination

\111:111II::tIIIIlII;I~
Rift (Fig. 3.21), in fact indicates that the consequences of
plate interaction need not necessarily be confined to -plate
margins.
Pull-apart basin

3.5 Oblique-slip margins

c
So far we have confined our attention to convergent margins Transpression
where two plates are in motion towards each other. We now
need to consider margins where the movement is predomi-
nantly transform and plates meet along transform faults rather
than at subduction zones or along collision boundaries.
Transverse- orogen
However, it will have become clear from the preceding dis-
cussion that purely convergent or transform movement is
unlikely to occur except along short sections of a plate E Anastomosing faults
boundary. So-called convergent boundaries almost invariab-
ly have some element of lateral movement (see, for instance,
Figure 3.20(F», while most transform boundaries possess
some convergent or divergent component. It is more realistic,
therefore, to think in terms of a continuum of plate boundary
types related to various degrees of convergence and diver-
Before movement After movement
gence. Where a convergent or divergent margin has a sig-
n!ficant transform component it is appropriate to refer to it Fig.3.26 Fault structures and major landformfeatures
associated with oblique-slip margins: (A) pull-apart basin; (B)
as an oblique-slip margin. graben and horst structures caused by fault termination; (C)
Some idea of the variation that may occur along an indi- folding and faulting on a side-stepping fault and creation of
vidual plate boundary can be gained by considering the east- transverse orogen; (D) development offault splays and wedges
em margin of the Indian Ocean. Here oceanic lithosphere through the progressive bending (2-4) of an originally straight
fault (1); and (E) movement along anastomosing faults leading to
of the Indian Plate is being subducted beneath oceanic and
upthrusting and the formation of down-sagged pods. (Based
continental lithosphere of the Eurasian Plate (Fig. 3.9). The largely on J.e. Crowell (1974) in: W. R. Dickinson (ed.) Society
Indian Plate is moving north-north-eastwards relative to the ofEconomic Paleontologists and Mineralogists Special
Eurasian Plate and the margin as a whole is one of low- Publication 22 Tectonics and Sedimentation. Figs. 5-7,
angle oblique convergence. However, when examined in p.194; H. G. Reading (1980) in: P. F. BallanceandH. G.
Reading (eds) Sedimentation in Oblique-Slip Mobile Zones.
more detail it is possible to identify significant variations
Special Publication of the International Association of
along the margin. For instance, although the forearc zone is Sedimentologists 4, Fig. 3, p. 12; and J. T. Kingma, (1958)
predominantly one of convergence and compression, there New Zealand Journal of Geology and Geophysics, 1, Fig. 1,
are also transform elements, such as north of Sumatra. In the p.270).
76 Endogenic processes and landforms

motion leads to extensional (nonnal) faulting and subsidence

and if this is of sufficient magnitude a pull-apart basin may
be fonned. Oblique convergence, on the other hand, generates
compression and results in folding, uplift and the formation
of a transverse orogen astride the fault.
Complexity is added to this simple relationship because
the predominant direction of stress along oblique-slip margins
may change through time as a consequence of adjustments
to the overall direction of plate movement. The San Andreas
Fault System in California, for instance, changed from main-
ly transtensile in the Miocene to predominantly transpressive
in the Pliocene. Very complex patterns of fracture are cha-
racteristic of major transfonn faults and these give rise to
local zones of compression and extension over comparatively
short distances. Because of the resulting close juxtaposition
of areas of uplift and subsidence, considerable local relief Fig. 3.28 Landsat mosaic showing the main structural and
morphological features along the San Andreas Fault System in
can be generated. southern California. The mosaic covers approxi;hately the same
area as Figure 3.27. (Part of mosaic ofERTS-i (Landsat)
imagery prepared for NASA Goddard Space Flight Center by the
3.5.1 The San Andreas Fault System USDA Soil Conservation Service.)
This situation is illustrated in classic fashion along the
southern section of the San Andreas Fault System (Fig. 3.27). The individual faults which branch, join, bend ~nd side-
Overall the fault system has a length of 1200 km, but the step each other, give rise to numerous separate areas of uplift
average breadth of the fault zone is some 500 km and it fonns and subsidence (Fig. 3.28). A major subsidence feature is the
a very diffuse boundary between the Pacific and North Salton Sea trough, an excellent example of a pull-apart basin.
American Plates. It is estimated that there has been about Although it contains a considerable depth of sediment shed
1000 km of lateral movement along the fault system over from the adjacent mountain ranges its floor still has a maxi-
the past 25 Ma. Of the large number of individual faults mum elevation of just 74 m below sea level. By contrast an
(Figure 3.27 just shows some of the major ones) only a few east-west trending Zone of compr.ession to the north-east of
are currently regarded as being active, although others have Los Angeles has led to the fonnation of a transverse orogen,
been active in the recent past. comprising the San Gabriel and San Bernardino Mountains
and collectively known as. the Transverse Ranges. These
attain elevations in excess of. 3000 m and much of this uplift
100
km I
has probably occurred as recently as the Late Quaternary.

3.5.2 The Southern Alps

The Southern Alps of New Zealand provide another example
of dramatic uplift and mountain construction along an
oblique-slip margin. They run parallel to the Alpine Fault, a
-remarkably straight transfonn fault representing the boundary
PACIFIC
between the Indian and Pacific Plates. It is estimated that
~.
th;e present net movement along the fault involves an aver-
OCEAN
age of 40 mm a-I of transform motion and about 22 mm a-I
of Qrthogonal convergence (Fig. 3.29).
~ ~ Major mountain Significant compres~ion has only developed comparatively
--...... Major faults ~ ranges' .
recently (the Alpine Fault was transtensile until the Pliocene)
;

Fig. 3.27 Simplified map of the main structural and but has generated rapid uplift with the highest peaks in the
morphological features associated with the San A,ndreas Fault Southern Alps now exceeding an altitude of 3000 m. The
System in southern California. The points marked by an X . compression has caused thrusting on an eastward-dipping
indicate areas of similar terrane considered to have been . :plane':along which the coptinental crust of the Pacific Plate
originally adjacent but subsequently displaced laterally along the
. is being heaved up (Fig. 3.30). The western slopes of the
fault system. (Based onJ. C. Crowell (1975) San Andreas Fault in
Southern California. California Division ofMines and Geology Southern Alps are, in effect, an enormous self-perpetuating
Special Report 118, Fig.i, p. 11.) fault scarp; as rapidly as material is eroded it is replaced by
 Landforms and tectonics of plate margins 77

Fig. 3.29 Schematic representation of the formation of the

Southern Alps through the oblique convergence of the Pacific
Plate along the Alpine Fault. The approximate direction of
movement of the Pacific Plate is indicated by the arrow labelled
A. (Based partly on J. Adams (985) in: M. Morisawa and J. T.
Hack (eds) Tectonic Geomorphology. Allen and Unwin, Boston,
Fig. 5, p.nD.)

further uplift and crustal shortening against the Alpine Fault

(see Section 18.2.4).

3.6 Displaced terranes

As our knowledge of both ancient and modem mountain
belts has developed, it has become increasingly evident that
not all the complexities observed can simply be explained
by the collision of large continental blocks. Many mountain Fig.3.30 Schist (S) thrust over Quaternary glacial deposits (Q)
belts situated along present or past plate boundaries contain on the western flank of the Southern Alps, New Zealand. Figure at
numerous individual slivers of crust some of which, on the the bottom of the photograph indicates scale. Fault plane is
basis of palaeomagnetic and structural evidence, together indicated by arrow.
with their assemblage of rock types, appear to have travelled
great distances before arriving in their present positions. arcs and slices detached from unknown continental margins.
Other crustal fragments, while appearing to be of more local Numerous displaced terranes have also been identified in
origin, have, none the less, experienced significant horizontal the Alpine-Himalayan chain (Fig. 3.32). Most of these are
displacement and rotation with respect to the plate of which thought to have rifted away from the margins of Gondwana
they form a part. These anomalous crustal fragments have and moved northward before colliding with the southern'
been given various names-suspect terrane, allochthonous' margin of the Eurasian land mass and causing extensive
terrane~but w~ will use the term displaced terrane (Note deformation and thrusting. Indeed, it has been suggested
that in this book the spelling 'terrane' is used to refer to struc- that much of the structural and morphological complexity
tural units whereas 'terrain' is applied in a topographic sense.) of the Alps may be due to the accretion or partial sub-
Several recent' studies have indicated the significant role duction of such crustal fragments rather than the effects of
played by displaced terranes in the morphology and structure simple collision of two large land masses. Similarly, the
of both intercontinental collision orogens and continental- morphology of the Himalayan orogen, in particular that of
margin orogens. For instance, it has been estimated that over the Tibetan Plateau, may be partly attributable to the accretion
70 per cent of the North American Cordillera is made up of of small crustal fragments prior to the main collision
displaced terranes (Fig. 3.31). Most of these seem to have' between India and Eurasia.
travelled thousands of kilometres and have been accreted to The composition of displaced terranes ranges from that
the North American cratonic margin during the Mesozoic of typical oceanic crust to significantly less dense granitic
and Cenozoic. The terranes are of several types and include rock with clear continental affinities., It has been suggested
slivers of oceanic crust, fragments of intra-oceanic island that such terranes are equivalent to oceanic plateaus, which
 78 Endogenic processes and landforms

are to be found in the present-day ocean basins and which

often rise several kilometres above the adjacent ocean floor
(Fig. 3.32). Many of these plateaus have anomalous crustal
thicknesses of between 20 and 40 km and an upper layer 10
to 15 km thick with P-wave velocities in the range 6.0-6.3
km S-I. These values are in the range of granitic rocks in
the continental crust and so it is inferred that many oceanic
plateaus (such as the Seychelles Bank) are drowned conti-
nental fragments originating from the edges of ancient land
masses and destined to be swept towards a subduction zone
in the future. As these plateaus are more buoyant than normal
oceanic crust they would tend to resist subduction and there-
fore be accreted to continents lying along plate margins.
The implications of displaced terranes for the investi-
gation of the relationships between tectonics and landform
development have yet to be assessed. Nevertheless, the jux-
taposition of terranes with often dramatically different litho-
logies and structural characteristics is likely to be a significant
factor in the long-term development of landscapes.

~ Oceanic
~ plateaus
o 2000

O Displaced
terranes
L--....--J
km

N.CHINA

S.CHINA

~
S.TIBET

K~NTUMC5~~~L
THAILAND D~

C~LON ~REO
SEYCHELLES ~
BANK SUMBA~
)

QAMGA'CAR
Fig. 3.31 Map of displaced (suspect) terranes in western North
America. The labelled blocks represent individual terranes which
~
BROKEN
AGULHAS lJ1 CROZET RIDGE
differ significantly in their geological characteristics. A PLATEAU tJjl1 PLATEAU
proportion have clear continental affinities and of these some Q.,

'
apparently originated in other continents at more southerly KE~GUELEN
RIDGE
latitudes. Other terranes have rock associations indicative of an
oceanic origin and represent island arcs and slivers of oceanic
crust. The shaded area indicates the extent ofNorth American Fig. 3.32 Map of the probable displaced terranes of the
cratonic basement and the arrowed line marks the eastern limit of Alpine-Himalayan chain. Several major oceanic plateaus in the
Mesozoic-Cenozoic deformation in the North American Indian Ocean are also shown. (Modified from A. Nur and Z. Ben-
Cordillera (From PI Coney et al. 1980 Nature 288, Fig. 1, Avraham (1982) Joumal of Geophysical Research 87, Fig. 1,
p.330.) p. 3649. Copyright by the American Geophysical Union.)
 Landforms and tectonics of plate margins 79

3.7 Mesoscale and microscale landforms

associated with faulting
Plate margins are invariably regions of active faulting,
and microscale and mesoscale landforms directly related to
. crustal movements along individual faults often provide im-
portant evidence of the nature of tectonic activity at a broader
scale. Although it must be emphasized that fault-related
landforms can be important in areas remote from plate
boundaries it will be useful to consider them briefly here.
Faulting takes place through sudden, rapid movements over
distances ranging from a few centimetres to several metres,
each individual movement being associated with an earth-
quake. Here we will focus on the primary tectonic landforms
created by faulting and leave the question of the mode of
subsequent landscape development of faulted terrains until
Chapter 16.
Fig.3.33 Block diagram showing various forms offault scarp
Both normal and reverse faulting can create a fault scarp, and their associated erosional features: (A) afault scarp dying
the initial angle of which will reflect the dip of the fault out laterally but merging into a monocline to the right; (B) en
plane. But weathering and slumping will very rapidly destroy echelon scarps to the right and the development ofgullies and
the original form and reduce the slope to an angle of stab- triangular facets to the left.
ility of typically between 20 and 40°. Fault scarps will only
be formed where a fault breaks the surface; they can either uplifted block, whereas if they are flowing towards the down-
die out laterally or merge into a monocline (Fig. 3.33(A)). thrown side they will rapidly cut down through the fault
Where faults die out laterally and are replaced by further scarp. This leads to the formation of gullies separated along
offset faults an en echelon scarp can be formed (Fig. the scarp by characteristic triangular facets (Fig. 3.33 (B)).
3.33(B)). Vertical movement along a fault may lead to the Eventually the scarp will retreat from the original position
impoundment of streams if they are flowing towards the of faulting and'it may be buried by sediment. A scarp formed

Linear ridge Offset drainage channel

Linear valley Sag pond
Shutter ridge

Fig.3.34 Typical micro- and mesoscale landforms associated with major strike-slip faults. (After R. L. Wesson et al. (1975) United States
Geological Survey Professional Paper 914A, Fig. 11, p. A21.)
80 Endogenic processes and landforms

Fig.3.35 Block diagram showing a shutter ridge formed along a left-lateral strike-slip fault.

at the site of a fault through the differential erosion of two introductory reviews, and Molnar (1988) considers the pro-
rock types juxtaposed along it or through the exposure by blem of reconciling the plate tectonics model with the
erosion of a previously buried fault scarp is termed a fault- problem of orogenesis. More details on the complexities of
line scarp. It is important to distinguish such erosional convergent margins are contained in articles by Cross and
features from true tectonic scarps since only the latter indi- Pilger (1982), Dewey (1980) and Uyeda (1982) while the
cate recent or current fault activity. possibility of the reversal of subduction polarity is assessed
Strike-slip faulting gives rise' to an intriguing range of by Johnson and Jaques (1980) and Mattson (1979). Various
landforms at the micro- and mesoscale in addition to the aspects of island arcs are discussed in the collection of papers
major topographic features associated with oblique-slip mar- edited by Talwani and Pitman (1977), with a more specifi-
gins (Fig. 3.34). Where there is no element of convergent cally geomorphic perspective being provided by Nunn
or divergent movement along a strike-slip fault it may gener- (1988), Woodroffe (1988) and Yoshikawa et ai. (1981).
ate no significant landforms. If, however, faulting occurs Continental-margin orogens are considered in several of
across a gullied terrain the resulting offsetting of truncated the references mentioned above. Useful introductions to the
gullies produces what are known as shutter ridges (Fig. tectonics and structure of the Andes are provided by Gansser
3.35). Localized zones of compressional and tensional stress (1973), James (1971) and Dalziel (1986). Gamer (1983) dis-
are common along strike-slip faults and they give rise to a cusses the tectonic geomorphology of the Ecuadorian Andes
number of distinctive landforms. Divergence across the fault and Jordan et., ai. (1983) consider the effects on Andean
leads to subsidence and the formation of shallow, elongated topography of variations in the angle of plate subduction.
troughs or sags, typically a few tens of metres across and a The North American Cordillera, which has a tectonic history
few hundred metres long, which, many contain a sag pond. of bewildering complexity, has not been examined in detail
In the zones of convergence, on the other hand, compression in this chapter but various aspects of its evolution are
generates positive topographic features through the formation considered by Dickinson (1976), Eaton (1987) and Smith
of ridges and linear and en echeion scarplets. Strike-slip and Eaton (1978). The Sunda Arc is discussed by Hamilton
faulting also usually leads to the displacement of stream (1979) in a comprehensive study of the whole Indonesian
channels (Fig. 3.25) and this feature is often useful in deter- region, while Hamilton (1977) provides a more concise
mining the magnitude and frequency of individual move- analysis.
ments along the fault (see Section 16.3.2). Articles considering specific aspects of intercontinental
collision include those by Burke et ai. (1977) on suture zones,
Molnar and Gray (1979) on the question of limited continen-
Further reading tal subduction and Sengor (1976) on the effects of the irre-
The literature on the structure and tectonic development of gular configuration of converging continental margins. There
plate margins is vast, but unfortunately very little of it takes has been a surge of field studies in the Himalayan-Tibetan
a specifically geomorphic perspective. An excellent summary region since the late 1970s and this has generated a large
is provided by Miall (1984, pp. 384-440) which, although body of published work. General overviews include those
focusing on sedimentological aspects, also has much rele- by Allegre et ai. (1984), Gupta and Delany (1981), Molnar
vant material on the morphology of plate margins. Dickinson (1984, 1986) and Tapponnier et ai. (1986), and Selby (1988)
and Seely (1979) give a more specific treatment of forearc provides a specifically morphological study of the Nepal
regions with excellent diagrams of their morphology, struc- Himalaya. Two contrasting models of the evolution of the
ture and evolution. A good coverage of the major processes Himalayas are presented by Le Fort (1975) and Powell and
of orogenesis, containing a number of detailed case studies, Conaghan (1973), while the tectonic development of the
can be found in Hsii (1983) and Schaer and Rodgers (1987), Tibetan Plateau is discussed by Bingham and Klootwijk
while Von Huene (1984) and Molnar (1986) provide useful (1980) and Ni and York (1978). The model of the Tibetan
 Landforms and tectonics of plate margins 81

Plateau uplift associated with the lateral displacement of Burke, K., Dewey, 1. F. and Kidd, W. S. F. (1977) World distribu-
lithospheric blocks is introduced by Molnar and Tapponnier tion of sutures; the sites of former oceans. Tectonophysics 40,
(1975), but there is also a more accessible and well-illustrated 69-99.
Coney, P. 1., Jones, D. L. and Monger, 1. W. H. (1980) Cordilleran
version (Molnar and Tapponnier, 1977). suspect terranes. Nature 288, 329-33.
Oblique-slip margins are considered in detail in various Cross, T.A and Pilger, R H. Jr (1982) Controls of subduction
papers in Ballance and Reading (1980); the introductory geometry, location of magmatic arcs, and tectonics of arc and
article by Reading (1980) is particularly useful as it pre- back-arc regions. Geological Society of America Bulletin 93,
sents a general review of sedimentation patterns in relation 545-62.
Crowell, J. C. (1974) Origin of Late Cenozoic basins in southern
to topography. The evolution of the San Andreas Fault California. In: W. R Dickinson (ed.) Tectonics and Sedimentation.
System is concisely covered by Crowell (1974, 1979). The Society of Economic Paleontologists and Mineralogists Special
Southern Alps have been the subject of a number of studies Publication 22 (Tulsa) 190-204.
examining the relationship between tectonics and landforms; Crowell, J. C. (1979) The San Andreas fault system through time.
these include those by Adams (1985), Basher et al. (1988), Journal of the Geological Society London 136,293-302.
Dalziel, I. W. D. (1986) Collision and cordilleran orogenesis: an
Whitehouse (1988) and articles included in Soons and Andean perspective. In: M. P. Coward and A. C. Ries (eds)
Selby (1982). Berryman (1988) and Kamp (1988) examine Collision Tectonics. Geological Society Special Publication 19,
landforms associated with oblique convergence in North 389-404.
Island, New Zealand. Dewey, J. F. (1980) Episodicity, sequence, and style at convergent
Displaced terranes are starting to generate a significant plate boundaries. In: D. W. Strangway (ed.) The Continental
Crust and Its Mineral Deposits. Geologists Association of
literature and this has been reviewed by Howell (1985) and Canada Special Paper 20,553-73.
Schermer et al. (1984). The large number of displaced Dickinson, W. R. (1976) Sedimentary basins developed during
terranes in the North American Cordillera are considered by evolution of Mesozoic-Cenozoic arc-trench system in western
Coney et al. (1980) and the possible role of oceanic plateaus North America. Canadian Journal of Earth Sciences 13, 1268-
in orogenesis is assessed by Nur and Ben-Avraham (1983). 87.
Dickinson, W. R and Seely, D. R (1979) Structure and stratigraphy
OIlier and Pain (1988) draw attention to the neglected pro- of forearc regions. American Association ofPetroleum Geologists
blem of how displaced terranes relate to landscape develop- Bulletin 63, 2-31.
ment in the context of Papua New Guinea. Examples of Eaton, G. P. (1987) Topography and origin of the southern Rocky
specific studies of micro-and mesoscale fault-related land- Mountains and Alvarado Ridge. In: M. P. Coward, J. F. Dewey
forms include those by Gerson et al. (1985), Nash (1981) and P. L. Hancock (eds) Continental Extensional Tectonics.
Geological Society Special Publication 28, 355-69.
and Wallace (1977), while Armijo et al. (1986) and Molnar Gansser, A (1973) Facts and theories on the Andes. Journal of
et al. (1987) provide detailed examples of how such the Geological Society London 129, 93-131.
landforms can assist in the interpretation of major tectonic Garner, H. F. (1983) Large-scale tectonic denudation and climatic
features. morphogenesis in the Andes Mountains of Ecuador. In: R.
.
"'"
Gardner and H. Scoging (eds) Mega-Geomorphology. Clarendon
Press, Oxford and Oxford University Press, New York., 1-17.
Gerson, R, Grossman, S. and Bowman, D. (1985) Stages in the
References creation of a large rift valley-geomorphic evolution along the
Adams, J. (1985) Large-scale tectonic geomorphology of the southern Dead Sea Rift. In: M. Morisawa and 1. T. Hack (eds)
Southern Alps. In: M. Morisawa and J. T. Hack (eds) Tectonic Tectonic Geomorphology. Allen and Unwin, Boston and London,
Geomorphology. Allen and Unwin, Boston, and London 105- 28. 53-73.
All~gre, c.j. et al. (1984) Structure and evolution of the Himalaya- Gupta, H. K. and Delany, F. M. (eds) (1981) Zagros, Hindu Kush,
Tibet orogenic belt. Nature 307,17-22. Himalaya-Geodynamic Evolution American Geophysical Union,
Armijo, R, Tapponnier, P., Mercier, J. L. and Tong-Lin, H. 1986. Washington, DC.
Quaternary extension in southern Tibet: field observations and Hamilton, W. (1977) Subduction in the Indonesian region. In: M.
tectonic implications. Journal of Geophysical Research 91, Talwani and W. C. Pitman III (Eds) Island Arcs, Deep Sea
13803-72. Trenches and Back-Arc Basins American Geophysical Union,
Ballance, P. F. and Reading, H. G. (eds) 1980. Sedimentation in Washington, DC, 15-32.
Oblique-Slip Mobile Zones. International Association of Hamilton, W. (1979) Tectonics of the Indonesian region. United
Sedimentologists Special Publication 4. States Geological Survey Professional Paper 1078.
Basher, L. R, Tonkin, P. J. and McSaveney, M. J. (1988) Howell, D. G. (1985) Terranes. Scientific American 253(5), 90-103.
Geomorphic history of a rapidly uplifting area on a compres- HsU, K. J. (ed.) (1983) Mountain Building Processes. Academic
sional plate boundary: Cropp River, New Zealand. Zeitschrift Press, London and New York.
fur Geomorphologie Supplementband 69, 117-31. James, D. E. (1971) Plate tectonic model for the evolution of the
Berryman, K. (1988) Tectonic geomorphology at a plate boundary: central Andes. Geological Society of America Bulletin 82,
a transect across Hawke Bay, New Zealand. Zeitschrift fur 3325-46.
Geomorphologie Supplementband 69, 69-86. Johnson, R W. and Jaques, A L. 1980. Continent-arc collision
Bingham, D. K. and Klootwijk, C. T. (1980) Palaeomagnetic con- and reversal of arc polarity: New interpretations from a critical
straints on Greater India's underthrusting of the Tibetan Plateau. area. Tectonophysics 63, 111-24.
Nature 284, 336-8. Jordan, T. E., !sacks, B. L., Allmendinger, R W., Brewer, J. A,
82 Endogenic processes and landforms

Ramos, V. A. and Ando, C. J. (1983) Andean tectonics related Reading, H. G. (1980) Characteristics and recognition of strike-
to geometry of subducted Nazca plate. Geological Society of slip fault systems. In: P. F. Ballance and H. G. Reading (eds)
America Bulletin 94, 341-61. Sedimentation in Oblique-Slip Mobile Zones. International
Kamp, P. J. J. (1988) Tectonic geomorphology of the Hikurangi Association of Sedimentologists Special Publication 4, 7-26.
Margin: surface manifestations of different modes of subduction. Schaer, J.-P. and Rodgers, J. (eds) (1987) The Anatomy of Mountain
Zeitschrift fur Geomorphologie Supplementband 69, 55-67. Ranges. Princeton University Press, Princeton.
Le Fort, P. (1975) Himalayas: the collided range; Present knowledge Schermer, E. R., Howell, D. G. and Jones, D. L. (1984) The origin
of the continental arc. American Journal ofScience 275A, 1-44. of allochthonous terranes:·perspectives on the growth and shaping
Mattson, P. H. (1979) Subduction, buoyant braking, flipping, and of continents. Annual Review of Earth and Planetary Sciences
strike-slip faulting in the northern Caribbean. Journal of Geology 12, 107-31.
87,293-304. Selby, M. J. (1988) Landforms and denudation of the High
Miall, A. D. (1984) Principles of Sedimentary Basin Analysis. Himalaya of Nepal: results of continental collision. Zeitschrift
Springer-Verlag, New York. fur Geomorphologie Supplementband 69, 133-52.
Molnar, P. (1984) Structure and tectonics of the Himalayas: Sengor, A. M. C (1976) Collision of irregular continental margins:
constraints and implications of geophysical data. Annual Review Implications for foreland deformation of Alpine-type orogens.
ofEarth and Planetary Sciences. 12,489-518. Geology 4, 779-82.
Molnar, P. (1986) The structure of mountain ranges. Scientific Smith, R. B. and Eaton, G. P. (eds) (1978) Cenozoic Tectonics
American 255(1), 70-9. and Regional Geophysics of the Western Cordillera. Geological
Molnar, P. (1988) Continental tectonics in the aftermath of plate Society of America Memoir 152.
tectonics. Nature 335, 131-137 Soons, J. M. and Selby, M. J. (eds) (1982) Landforms of New
Molnar, P., Burchfiel, B. C, Liang K'Uangyi and Zhao Ziyun Zealand. Longman Paul, Auckland.
(1987) Geomorphic evidence for active faulting in the Altyn Talwani, M. and Pitman, W. C. III (eds) (1977) Island Arcs, Deep
Tagh and northern Tibet and qualitative estimates of its con- Sea Trenches and Back-Arc Basins. American Geophysical
tribution to the convergence of India and Eurasia. Geology 15, Union, Washington, DC.
249-53. Tapponnier, P., Peltzer, G. and Armijo, R. (1986) On the mechanics
Molnar, P. and Gray, D. (1979) Subduction of continental of the collision between India and Asia. In: M. P. Coward and
lithosphere: Some constraints and uncertainties. Geology 7, A. C. Reis, (eds) Collision Tectonics. Geological Society Special
58-62. Publication 19, 115-57.
Molnar, P. and Tapponnier, P. (1975) Cenozoic tectonics of Asia: Uyeda, S. (1982) Subduction zones: an introduction to comparative
effects of a continental collision. Science 189,419-26. subductology. Tectonophysics 81, 133-59.
Molnar, P. and Tapponnier, P. (1977) The collision between India Von Huene, R. (1984) Tectonic processes along the front of modem
and Eurasia. Scientific American 236(4),30-41. convergent margins - Research of the past decade. Annual Review
Nash, D. B. (1981) Morphologic dating of degraded normal fault of Earth and Planetary Sciences 12, 359-81.
scarps. Journal ofGeology 88, 353-60. Wallace, R. E. (1977) Profiles and ages of young fault scarps,
Ni, J. and York, J. E. (1978) Late Cenozoic tectonics of the Tibetan north central Nevada. Geological Society of America Bulletin
Plateau. Journal ofGeophysical Research 83, 5377-84. 88, 1267-81.
Nunn, P. D. (1988) Plate boundary tectonics and oceanic island Whitehouse, I. E. (1988) Geomorphology of the central Southern
geomorphology. Zeitschrift fUr Geomorphologie Supplementband Alps, New Zealand: the interaction of plate collision and
69,39-53. atmospheric circulation. Zeitschrift fUr Geomorphologie
Nur, A. and Ben-Avraham, Z. (1983) Displaced terranes and moun- Supplementband 69, 105-16.
tain building. In: KJ. Hsti (ed.) Mountain Building Processes. Woodroffe, C. D. (1988) Vertical movement of isolated oceanic
Academic Press, London and New York, 73-84. islands at plate margins: evidence from emergent reefs in
allier, C. D. and Pain, C. F. (1988.) Morphotectonics of Papua Tonga (Pacific Ocean), Cayman Islands (Caribbean Sea) and
New Guinea. Zeitschrift fur Geomorphologie Supplementband Christmas' Island (Indian Ocean), Zeitschrift fur Geomorphologie'
69, 1-16. Supplementband 69, 17-37.
Powell, C. McA. and Conaghan, P. J. (1973) Plate tectonics and Yoshikawa, T., Kaizuka, S. and Ota, Y. (1981) The Landforms of
the Himalayas. Earth and Planetary Science Letters 20, 1-12. Japan. University of Tokyo Press, Tokyo.
This page intentionally left blank
 4
Landforms and
tectonics of plate interiors

.....
4.1 Landscapes of plate interiors Tibesti Plateau

We saw in Chapter 3 how a number of major morphological

features, notably orogenic mountain belts and island arcs,
coincide with convergent plate boundaries; a superficial glance
at the Earth's morphology might even suggest that all sig-
nificant sub-aerial topographic features are confined to such
boundaries. But the fact that the effects of plate interactions
can extend for thousands of kilometres away from plate
boundaries, as we have seen in the case of the Tibetan Plateau
and central Asia (see Section 3.4.4.2), shows that this is
certainly not the case. Moreover, a closer inspection of global
morphology reveals a range of major landform features which
are located in the interior of plates and which are umelated to
even distant plate convergence events (Fig. 4.1).
Some continents, most notably Africa, have extensive
plateaus which lie at elevations (2000 m or more) which far
exceed the average height of the continental platforms. Basins
represent another major type of landform of plate interiors.
Rifted_
These are commonly 1000 km or more across and are either margin
entirely enclosed and therefore drained internally, such as
the Lake Eyre Basin of Australia and the Chad and Kalahari
Basins of Africa, or they are breached by one or more major
river system, such as the region drained by the Zaire (Congo) o 500 1000
I I
river system (Fig. 4.1). Although volcanoes and associated
volcanic landforms are clearly concentrated along plate mar-
~ Land over 1000m ' , _ Sheared margin
gins, significant occurrences are also to be found in plate
Volcanic rocks with ~ Riftvalley
interiors. On the continents some of this intra-plate volcanic ages less than 25 Ma (- .... Boundary of hot spot
activity is closely related to rift valleys which themselves re- / " Major escarpment \ swell
present sites of continental rifting where the crust has been ,"', Boundary of basin
stretched and faulted. Fig. 4.1 Examples of landforms related to tectonic processes
The break-up of the supercontinent of Pangaea, which operating in plate interiors on the African continent. Note that for
began about 180 Ma ago (Fig. 2.16), created many new con- clarity only some of the basins and hot-spot swells are shown.
tinental margins. This event was of great significance for
the development of landscapes on a continental scale because activity than the active margins located along convergent
it led to the establishment of new base levels for continental plate boundaries and are consequently termed passive con·
erosion. These margins exhibit a much lower level of tectonic tinental margins (or simply passive margins). Where formed
86 Endogenic processes and landforms

~--------------

~~

Drakensburg
Mountains
Great Escarpment of
Great Escarpment of t::'
Southern Africa
Eastern Australia
6 2
10km

-cF~~....--"""""'=--~- --- - --- 1~

1000 0 1000
km

Fig.4.2 Great escarpments on passive margins. Although lacking major escarpments, a number of other passive margins possess low
amplitude marginal up warps flanked by a significant break ofslope. An example is the Fall Line of the eastern seaboard of North
America which marks a significant increase in stream gradients and in places this develops into a distinct escarpment.

by divergent plate movement they are described as rifted vertical movements of the crust, that is, the mechanisms of
margins, but where the motion between two adjacent con- epeirogeny, that are of most concern.
tinental blocks has been transform they are called sheared
margins (but note that the term rifted margin is often applied
rather loosely to passive margins of any type). Many, though
4.2 Mechanisms of epeirogeny
certainly not all, passive continental margins are characterized While we now have a reasonably good understanding of how
by major escarpments (sometimes called great escarpments) interactions at plate boundaries give rise to major morpho-
(Fig. 4.2). By any standards these are significant morpho- logical features on the continents our knowledge of the tec-
logical features - the Great Escarpment in southern Africa tonic processes operating in plate interiors is by compari-
has a relief exceeding 1000 m in places - and they typically sion rather poor. Numerous mechanisms have been suggested
separate distinct geomorphic environments. Below great to explain the often significant uplift that has occurred in
escarpments the topography is usually highly dissected, but plate interiors but none seems to be applicable to all cases.
inland on the plateau surface the relief is generally subdued. Although the various models defy any simple classification,
Along plate margins the horizontal motion of plates is for convenience we will consider them under the headings
the primary force driving the uplift which occurs in orogens of thermal, phase change and mechanical models.
(although as we saw in Section 3.3 vertical movement re-
sulting from thermal effects is of great importance in con-
4.2.1 Thermal models
tinental-margin ~rogens). In plate interiors horizontal plate
movements playa less direct role in creating landforms and There are many occurrences of volcanic activity in areas re-
it is the processes which generate the broad warping and mote from plate margins; the volcanoes of the East African
 Landforms and tectonics of plate interiors 87

to establish because of the problems of identifying specific

hot spots and incomplete data in some regions. Along mid-
oceanic ridges and subduction zones it is often a problem
distinguishing between 'normal' and anomalous volcanism.
Since it is also difficult in some cases to determine whether
neighbouring centres of active volcanism are associated
with one or more hot spots, estimates of the total number of
hot spots vary significantly from around 40 to well over 100.
A hot spot arises from the presence of unusually hot mantle
at the base of the lithosphere (Fig. 4.5(A)). Such a sub-
lithospheric thermal anomaly may be generated by a
mantle plume, an upwelling of hot mantle originating deep
in the Earth's interior, possibly at the core-mantle bound-
ary; alternatively, it may be caused by localized heating at
shallower depths in the mantle. In either case the presence
of hot mantle leads to melting at the base of the lithosphere
and the lithosphere as a whole becomes thinner (Fig. 4.5(B)).
Since part of the lithosphere is replaced by hot mantle of
lower density from the asthenosphere, the lithosphere as a
whole experiences isostatic uplift. As the change in density
is the result of a change in temperature this kind of isostatic
adjustment is termed thermal isostasy. Eventually uplift
may be sufficient to stretch the brittle overlying crust to the
point where it fractures and a rift valley forms (Fig. 4.5(C)).
The time scale over which uplift takes place depends upon
the way in which heat is transferred through the lithosphere.
Conduction of heat occurs only very slowly through rock,
and if this is the only mechanism operating uplift will
occur over a very long period - something of the order of
100 Ma. In contrast penetrative magmatism, whereby hot
Fig.4.3 Old Faithful geyser, Yellowstone National Park. and partially molten material from the asthenosphere forces
Wyoming, USA, representing hydrothermal activity associated
its way up towards the surface, is a much more rapid mechan-
with a hot spot beneath the NorthAmerican continent. The
surrounding region is characterized by active volcanism and ism of heat transfer and is capable of thinning 100 km thick
uplift. lithosphere and producing uplift at the surface in only about
20Ma.
As several million years are required for the transfer of a
Rift system, such as Kilimanjaro and Mount Kenya, the significant amount of heat to the surfac.tE:'9f continental Iitho-
volcanic and associated hydrothermal activity of the Yellow- sphere, it has been suggested that hot spots are most likely to
stone region in Wyoming, USA (Fig. 4.3), and the chain of develop where the lithosphere is more or less stationary with
volcanic peaks making up the Hawaiian Islands are examples respect to sub-lithospheric thermal anomalies as this would
from both continental and oceanic regions. There are also allow time for sustained heating to occur. This idea has been
anomalous areas of volcanism on, or near, mid-oceanic ridges; applied to the African Plate which, according to some analy-
Iceland, for instance, is formed of unusually thick oceanic ses of plate motions, appears to have been almost stationary
crust produced by very high localized rates of volcanic ac- with respect to the underlying asthenosphere for the past
tivity on the Mid-Atlantic Ridge. Such an area of anomalous 25-35 Ma. This period coincides with a well-established in-
volcanism, which is not related to processes of subduction crease in volcanic activity in Africa and, less certainly, with
or normal mid-oceanic ridge spreading, is termed a hot a phase of uplift.
spot. In addition to volcanic activity hot spots are in many' Although the idea of linking volcanism and uplift to sub-
cases associated with significant crustal uplift. The resulting lithospheric thermal anomalies is attractive there are problems
hot-spot swells are typically several hundred kilometres with this hypothesized relationship. One difficulty is that
across and may rise 1 km or more above the surrounding there is evidence indicating that hot spots can migrate with
terrain. respect to each other at velocities of up to 20 mm a-I. This
Hot spots appear to be rather irregularly distributed over means that both hot spots can move under 'stationary' litho-
the Earth's surface (Fig. 4.4) but the real pattern is difficult sphere and that the lack of motion of the African Plate is
88 Endogenic processes and landforms

~--------------

~.
• ,~

•
----------e----
____tP • _

• • • • •
• • •
• • •
DATA SPARSE DATA SPARSE 6 2
10km
1R:D
1000 0 1000
km

Fig.4.4 Global distribution of hot spots. (Based on P. R. Vogt, (1981) Journal of Geophysical Research 86, Fig. 1, p. 951. Published
by The American Geophysical Union.)

A Rift
formation

', ...
. . '

. '.'
.. "
',,: .

HOT SPOT

Fig. 4.5 Domal uplift generated by a sub-lithospheric thermal

anomaly. The thermal anomaly may give rise to a hot spot and the
,',: : : .. :',; .: .: . : crustal extension induced by uplift may eventually lead to the
.. ,', .". ...... development of a rift (see Section 4.3.2) .
 Landforms and tectonics of plate interiors 89

thrown into doubt because its 'past positions have been in Box 4.1 Lithospheric vulnerability
part determined with respect to a reference frame of suppos-
edly fixed hot spots. The lithospheric vulnerability index (v) is defined in terms
of lithospheric thickness (l) and plate velocity (u) as
4.2.1.1 Lithospheric vulnerability K
For a more realistic analysis of the occurrence of hot spots V= Ih"
and their associated uplifts we need to consider the thickness lu
of the lithosphere as well as its velocity with respect to sub- where K is a constant for a property of the lithosphere known
lithospheric thermal anomalies. This is because thick litho- -as thermal diffusivity which expresses the rate of propaga-
sphere will tend to be more resistant to the effects of heat tion of a temperature change through time. This expression
demonstrates that vulnerability is the same for 200 kIn thick
conduction and penetrative magmatism. By relating vari- lithosphere moving very slowly at 10 mm a-I as it is for
ations in the thickness of the lithosphere over the Earth's lithosphere which is only 63 km thick but which is moving
surface to variations in the velocity of plate movement it is rapidly at 100 mm a-I.
possible to determine global variations in lithospheric
vulnerability (Fig. 4.6; Box 4.1). Lithosphere which is thin
and slow moving is predicted to be the most vulnerable to African Plate the relationship is, in fact, the opposite of that
penetrative magmatism and therefore the most likely to ex- predicted with several hot spots located in areas ot" low
perience hot-spot volcanism and associated uplift. lithospheric vulnerability. There are several possible reasons
On a global basis there is a fairly strong relationship be- for this including inaccuracies in the estimates of lithospheric
tween lithospheric vulnerability and the distribution of hot thickness and the over-zealous identification of hot spots on
spots; few hot spots are to be found in areas of low litho- the African Plate. Nevertheless it is probable that factors other
spheric vulnerability whereas they are common in areas of than lithospheric vulnerability, such as the actual global
high vulnerability. When specific regions are examined, distribution of sub-lithospheric thermal anomalies, are im-
however, this relationship does not always seem to apply. By portant in influencing the occurrence of hot-spot volcanism
carefully examining Figure 4.6 you can see that on the and associated uplift.

Fig. 4.6 Global distribution of hot spots (triangles) and contours of the dimensionless lithospheric vulnerability index; lower values
of the index indicate areas more vulnerable to penetrative magmatism. Major plate boundaries are also shown. Note that the set of hot-
spot locations shown is slightly different from that illustrated in Figure 4.4. (After H. N. Pollack et aI., (1981) Journal of Geophysical
Research 86, Fig. 3, p. 965. Copyright by the American Geophysical Union.)
90 Endogenic processes and landforms

4.2.1.2 Effects of hot-spot migration our lack of detailed knowledge of the abundance of specific
One intriguing aspect of the effects of sub-lithospheric ther- minerals in the lower crust and upper mantle makes their
mal anomalies on continental topography arises from the evaluation rather uncertain.
consequences of plates moving over such regions of unusually A phase change, triggered by a small amount of heating at
hot asthenosphere. As yet there has been no detailed attempt to the base of the lithosphere, has been proposed to account for
assess the possible effects of such hot-spot 'migration' on the uplift evident in a number of regions in plate interiors,
landscape development, but there is some evidence from the including southern and eastern AfrIca, south-east Australia
stratigraphic record of the movement of topographic swells and the Transantarctic Mountains of Antarctica. Each of these
across continents which may be related to the passage of areas of continental lithosphere are thought to have over-
hot spots or similar thermal phenomena. For example, from ridden regions of hot asthenosphere associated with former
about 160 to 100 Ma BP, North America drifted over a hot mid-oceanic spreading ridges. It is thought that this would
spot which is now located near the Great Meteor Seamount have increased the flow of heat to the base of the lithosphere
in the eastern Atlantic Ocean. Patterns of erosion, sedimen- sufficiently to cause a phase change in the uppermost conti-
tation and igneous activity recorded over the area during this nental mantle after a delay of up to 70 Ma (the time taken
period seem to be consistent with a migrating region of uplift for heat to be conducted from the base of the lithosphere to .
related to the track of this hot spot. just below the Moho). The conversion of eclogite, a rock with
On the basis of an estimated 10 per cent of the Earth's a density of 3400 kg m-3, to basalt (density 3000 kg m-3) is
surface currently being part of a hot-spot swell it has been the kind of phase change envisaged. If this is so a 11.25 km
calculated that on average hot-spot epeirogeny should affect thick layer of eclogite would need to be converted to a thick-
a particular area of crust about every 600 Ma. We would ness of 12.75 km of basalt to produce an uplift of 1500 m.
expect the slow migration of hot-spot swells across con- The attractiveness of this model is that it requires only a
tinental interiors to have a significant impact on landscape modest increase of temperature of around 100°C at the base
development through, for instance, the disruption and diver- of the lithosphere. This seems much more reasonable than
sion of drainage systems. Geomorphologists have yet to the very large increases of up to 1000 °C which would be re-
assess this possibility, although the relationship between quired to produce an uplift of around 1500 m by purely ther-
migrating hot-spot swells and landscape development pro- mal effects. A major problem with the model, however, is
vides an intriguing subject for future research. that it predicts a rather uniform and extensive pattern of
uplift which does not accord with the typical morphology of
continental plate interiors. Africa, for instance, is characteriz-
4.2.2 Phase changes
ed by areas of domal uplift, from 500 to 2000 km across and
A major problem with hot spots as a general explanation by a prevalence of high elevations along its continental
for uplift in plate interiors is that many elevated continental margins.
regions in such areas show no evidence of volcanism. Large
parts of southern and eastern Africa stand at over 1500 m
4.2.3 Mechanical models
yet recent volcanism has been very localized. This might be
explained by sub-lithospheric thermal anomalies causing The most straightforward mechanical models of uplift are
uplift without penetrative magmatism developing to a point those that involve the isostatic adjustment that takes place
where volcanic activity occurs at the surface. Another dif- when a load is removed from the crust (see Section 2.2.4).
ficulty with the hot-spot model is that unrealistically large The term epeirogeny was in fact first coined by G. K. Gilbert
increases in heat flow are apparently required to explain the in 1890 to describe the isostatic uplift which occurred as a
magnitude of uplift recorded in some regions. result of the evaporation of Lake Bonneville in Utah in the
An alternative explanation of such uplifts involves the western USA (leaving the much smaller Great Salt Lake
effects of density changes in minerals in the upper mantle. of the present day). During a period of high lake levels
Such phase changes occur when minerals adopt different 15000--25000 a BP the load provided by the average water
atomic configurations as temperatures and pressures change. depth of about 145 m depressed the crust. After unloading
The temperatures experienced by a mineral in the crust or by evaporation the crust rose and shoreline features are
mantle can be affected by heat flow from below while the now elevated up to 64 m above the present water level.
pressure can be altered by loading and unloading of the crust Far more spectacular examples of isostatic uplift are to be
above. Consequently the depth at which particular phase found in the areas of the northern hemisphere covered by
changes (and therefore density changes) occur can migrate great ice sheets during the Pleistocene.· Detailed records of
up and down within the lithosphere. The surface above will post-glacial uplift have been established for such areas as
rise or fall in response to the changes in volume brought the Fenno-Scandian Shield and the Canadian Shield from
about by these alterations in mineral density. A number of both raised and warped shorelines and gravity data. Uplift of
possibly significant phase changes have been suggested but the central region of the Fenno-Scandian Shield began about
 Landforms and tectonics of plate interiors 91

A Local (Airy) Isostasy Change in actually increase in elevation as a result of the erosional un-

.><
<>
e
_"0
G)

<>-0
@
~------- -------.
surface
elevation loading along the escarpment (Fig. 4.7(B)). A similar effect
can lead to the summits of mountains in orogenic belts in-
creasing in elevation as the incision of deep valleys between
o Q) 'p c::
peaks causes unloading of the crust. The extent to which
_ >
.c 0 '" :l
-lJ).o
0
Q.E isostatic compensation occurs regionally through flexure,
oQ) Q)
~ ~~
rather than locally, depends on the flexural rigidity of the
lithosphere, a property which indicates its resistance to bend-
ing in response to a change in load.
Another proposed mechanism of epeirogenesis rests on the
B Flexural Isostasy
G)
observation that continental lithosphere is in an unstable
Landsurface .9.!:i.!!i~~~~C!s~:!~c~ _ mechanical equilibrium because its mantle is denser than the
,
,'-------
Rock removed underlying asthenosphere. Consequently if some process,
such as cracking or 'erosion' by mantle plumes, punctures the
Uplift with respect continental lithosphere as far as the crust the relatively dense
@ Landsurface ~originallandSUrfaCe mantle layer of the lithosphere could become detached and

t Area affected by isostatic compensation

t
sink (Fig. 4.8). This process has been termed delamination

A
Fig.4.7 Models of isostasy showing the contrasting modes of
vertical adjustment between local (Ail)') isostasy (A), and
flexural isostasy (B).
MANTLE PART
OF LITHOSPHERE
13 ()()() a BP and the total uplift since then has exceeded 100 m.
Though involving less spectacular rates of uplift, crustal
ASTHENOSPHERE
unloading by denudation is a far more pervasive process
which extends over long periods of time, in contrast to the
rather transient effects of deglaciation and evaporation of B
lakes. The latter two phenomena essentially restore the
crust to its approximate elevation before loading by ice or
water, but denudation can lead to continued uplift of the crust
for as long as continental crust is sub-aerially exposed. If
isostatic compensation is local, that is if vertical adjustments
of the lithosphere in response to the removal of a load are
confined just to the area covered by the original load, then
isostatic uplift simply reduces the rate at which a land-
surface is lowered by denudation (Fig. 4.7(A)). Such local
compensation, however, cannot occur at the small scale
because the lithosphere has a finite strength and so changes
in load result in regional isostatic adjustments over a greater
t
area than that actually affected by the change in load (Fig.
4.7(B)).
Such flexural isostasy (see Section 2.2.4) is a familiar
phenomenon to geomorphologists who study the uplift and
warping of shorelines as a result of crustal loading by nearby
ice sheets. In this case flexing of the lithosphere leads to the
development of a forebulge at some distance from the ice
margin which experiences an increase in surface elevation.
It is possible that a similar flexural effect is associated with Fig.4.8 Schematic representation ofdelamination showing: (A)
great escarpments along passive continental margins. In this initial cracking of the mantle part ofthe lithosphere by
case there is active erosion (and therefore unloading) along penetrative magmatism; (B) the sinking of this portion of the
lithosphere into the less dense asthenosphere; (C) crustal uplift
the escarpment edge but relatively little erosion on the plateau associated with the replacement of lithosphere by warmer and
surface immediately inland of the escarpment. Consequently less dense sub-lithospheric mantle. Plate motion is to the right in
it is possible that the slowing eroding plateau summit could relation to the underlying asthenosphere.
92 Endogenic processes and landforms

and may be capable of causing epeirogenic uplift through the A Fault escarpments
Symmetric rift
exposure of the cold base of the crust to hot mantle from the /valle y
asthenosphere and the replacement of relatively dense litho-
spheric mantle by less dense mantle from the asthenosphere.
Two such delamination 'events' may account for the episodes
of uplift of the Colorado Plateau in the south-west USA.
around 30 Ma and 5 Ma BP.

4.3 Continental rifts

4.3.1 Rift structure and location

Rift valleys are important landforms on continental crust in B
Fault escarpment
plate interiors where tensional stresses predominate in the Main listric Asymmetric rift valley
lithosphere. They typically have widths of 30-60 km, but boundary fault / (halfgraben)
some are up to 100 km across. Pioneers of rift valley geology
such as 1. W. Gregory were strongly influenced in their inter-
pretations of rift structure by the apparently simple rift
morphology that they encountered in areas such as East
Africa. The conventional image of a rift valley is of a broad,
flat-floored and symmetric trough flanked by steep escarp-
ments. The most obvious structural interpretation of this
morphology is that of a graben with the rift floor being formed Hangingwall
on a downthrown block bounded by normal faults which
create the steep escarpments (Fig. 4.9(A)). But a closer exam- Fig.4.9 Schematic representation of contrasting rift structures:
(A) classic symmetric graben structure with a downthrown block
ination of both the morphological and structural evidence bounded by normal faults; (B) asymmetric, half-graben structure.
shows that this model is inadequate. In particular the appli- In both cases the number offaults in real rifts and the complexity
cation of seismic stratigraphy to the study of rift valleys of their structure is much greater than is indicated here.
during the 1980s has led to a considerable revision of ideas
about their structure. patterns, for instance, reveals that major faults tend to occur
Seismic stratigraphy is a technique whereby seismic waves on only one side of a rift along anyone section of the valley.
generated by. small explosions set off at the surface are re- Moreover, these major faults tend to be discontinuous and
flected or refracted from discontinuities in the underlying curved in plan and it seems that main boundary listric faults
sediments which represent changes in sediment properties. alternate along rift valleys and are separated by transfer
Such changes thus enable discrete sedimentary units to be faults (Fig. 4.10). Some rift valley lakes, such as Lake
identified and subsurface faults to be mapped (see Section Tanganyika, Lake Malawi and Lake Turkana in East Africa
17.2.2.2. for a more detailed explanation). Seismic data which
show the deep structure of rift systems have revealed a much
more complex pattern of faulting than was suspected in the
classic rift model, with the presence of listric as well as
normal faults. Most significantly, the structure of many rifts
examined by seismic methods has been shown to be asym-
metric with most of the downthrow occurring along a major
boundary listric fault on one side of the rift. Such rifts are
said, therefore, to have a half-graben structure with the main
boundary listric fault forming a footwall, and the opposing
downwarped crust forming a hanging wall leading down from
a roll-over zone (Fig. 4.9(B)). The presence of an antithetic
fault on the hanging wall margin can give the impression of
a symmetric rift valley if it is exposed and forms an escarp-
ment, even though the overall structure is asymmetric.
Further evidence contradicting the traditional symmetric
rift valley model comes from observations of their mor- Fig.4.10 Highly schematic illustration of alternating half-
phology and surface .structure. Careful examination of fault graben along a rift valley.
 Landforms and tectonics of plate interiors 93

have a slightly sinuous plan form which may reflect the well as the position of zones of crustal weakness, is im-
structural control exerted by these alternating curved boun- portant in determining the location and orientation of conti-
dary faults. nental rifts.
Even a cursory examination of global structure and mor- The major differences in rift characteristics relate to their
phology shows the generally close relationship between con- position with respect to plate boundaries and the intensity of
tinental rifts and areas of mid-plate volcanism and crustal volcanic activity associated with them. Some rifts appear to
doming and uplift. The location of rift systems also seems be clearly related to divergent crustal movements with lateral
to be closely related to the distribution of ancient structures extension being normal to the rift axis. The East African Rift
representing zones of weakness between stable cratons. The System is of this type and its interpretation as an incipient
East African Rift System, for instance, parallels the trend of site of plate rupture is further supported by its continuity
old mobile belts (regions of highly deformed rocks formed with the divergent plate boundaries of the Red Sea and Gulf
during early phases of. folding and orogenesis) and tends to of Aden. In other cases, such as the Dead Sea Rift, there is
follow the margins of Precambrian cratons (Fig. 4.11). Such a significant transform component in the extensional move-
ancient lines of weakness are apparently more readily acti- ment and this feature can in fact be regarded as a large pull-
vated than adjacent crust and consequently are favoured apart basin (see Section 3.5). The other main rift type is to
sites for later faulting and rift development. Some recently be found orientated approximately at right angles to the
active rifts, however, do cut across ancient Precambrian struc- strike of intercontinental collision orogens. The development
tures, suggesting that the stress pattern initiating rifting, as of this type of rift appears to be related to the crustal exten-
sion that occurs, as a result of continental collision, beyond
the immediate collision zone itself. Examples of this type
are the Rhine Rift in the foreland zone of the Alps, and the
Baikal Rift System which lies some 3000 km to the north of
the Himalayas (see Section 3.4.4.2).
There are significant differences in the morphology and
degree of volcanic activity associated with these two types
of rift. True intra-plate rifts, such as the East African Rift
Afro- Arabian System, are characterized by prolonged volcanism and typi-
Rift System cally exhibit 1-2 km of down-faulting along the crest of
NUBIAN
substantial crustal upwarps. In contrast orogen-related rifts,
BLOCK / such as the Rhine and Baikal structures, experience generally
less intense volcanism and much more marked down-faulting,
, I~MALIA amounting to 5-6 km in the case of the Baikal Rift. In this
Lake Victoria Itl( . . .
~~OCK kind of rift, however, the marginal uplift is either narrow or

we~~;;n# I )1 Eastern lacking altogether.

1/ C9 )JRift
h ~ 4.3.2 Rift formation
TANZANIA
BLOCK ~ 'J L (/
INDIAN OCEAN The commonly close relationship between rifting and crustal

~t~c~rf
doming in rifts not associated with orogens suggests a causal
relationship. There are two possibilities: one involves active
rifting in which rifting develops as a response to the tensional

~)'l
ZIMBABWE
stresses induced ill the crust by uplift resulting from up-
welling of the asthenosphere (Fig. 4.5(C)); the other involves
passive rifting in which rifting is initiated by extensional
BLOCK (I stresses in the lithosphere, and this permits the subsequent
upwelling of hot mantle which in tum induces thermal
uplift (Fig. 4.12).
Clearly the timing of uplift, volcanism and rifting is crucial
o 1000
I ,
in the assessment of the validity of the active and passive
km
rifting models. The typical sequence for passive rifting is
rifting followed by volcanism, which may in tum be succeed-
Fig. 4.11 Simplified map of the main components of the East
African Rift System. Note the continuity with the spreading ridge ed by uplift. Active rifting, on the other hand, would be
leading from the Red Sea through the Gulf ofAden and into the expected to begin with volcanism which is then followed
Indian Ocean. (Based on various sources.) by uplift before rifting itself is finally initiated; but in some
94 Endogenic processes and landforms

A in which the tensional stresses within the lithosphere are most

Rift Valley
readily accommodated. Eleven such triple junctions, each
possibly associated with a separate hot spot, have been iden-
~ :::i::::::::::~~:~:~f~:::i::::::>::8>~r~{<::/:::::/:::::::::.:::.:j:j:j:::j:':. tified on the continental part of the African Plate and an
::c
0-
en intriguing problem is the way in which they may sub-
o Tensional Stress
::c sequently promote continental rupture and develop into
I-
• . : : .~~~---:-:--:-:--;-;-:-~
...J :-:-
spreading centres.
: : : : :AS:TI~ EI'.JO:S.PHE R E : Spreading can apparently only occur when continental sep-
aration can be accommodated into the motions of the global
plate system. Where this is not possible rift development is
halted, rifts become filled with sediment and any associated
B uplift subsides. Although the triple junctions identified in
Uplift
Africa, including those along the continental margin as well
~
as those in the interior, have variously evolved by spreading
along one, two or all three rift arms, the most common se-
quence has been for one arm to remain inactive and form
an aulacogen, with spreading occurring along the other two
(Fig. 4.13). Of the 29 rifts identified in Africa, five have led
to continental separation and a further three have been
associated with the movement of Madagascar away from
the African mainland. The two rifts in the Benue Trough in
Fig.4.12 Formation of a rift through extension of the Nigeria went through a phase of opening followed by later
lithosphere. In contrast to the active rifting model, in this passive closing, and of the remaining rifts ten became inactive before
rifting mechanism extension and thinning of the lithosphere (A) the spreading stage and nine are currently active and may
precedes uplift of the rift flanks (B).
spread in the future.

cases uplift seems to precede volcanism. In practice some TRIPLE JUNCTION

rifts seem to be of the active type whereas others seem to FORMATION
have been initiated by passive rifting, although in many cases Rungwe, Kilimanjaro
(Neogene. still active)
there is no consensus as to whether specific rifts are active
or passive.

4.3.3 Rifting and uplift in Africa

---------- ----------
THREE RIFTS SPREAD TWO RIFTS SPREAD

(One ceases to be active

before spreading)
Afar (Red Sea, Gulf
of Aden active.
The domal uplifts and rift structures of Africa are at various Ethiopian rift Nakuru (North and
beginning to spread)
stages of development and thus provide a useful basis for South Gregory Rifts
beginning to spread.
examining the evolution of rift systems. Two major periods Niger Delta (Atlantic
Kavirondo ceased)
of uplift and rift formation have been recognized in Africa; rifts active since Lower Limpopo (Lebombo
Cretaceous. Benue and Sabi spread
a Jurassic-Cretaceous rifting phase between 180 and 130 Ma Rift opened and closed Limpopo ceased in
in Cretaceous) Jurassic)
BP associated with the break-up of Gondwana and a more
recent period initiated some 35 to 25 Ma BP and continuing ..
up to the present. ONE RIFT SPREADS NO SPREADING

There is some evidence in Africa for the idea that uplift (One ceases to be active (Sediment fills one or more
before spreading and arms, activity ceases)
may precede volcanism since the Adamawa dome in the the third becomes
transform forming Luangwa (Jurassic)
Cameroon region, which rises to elevations in excess of sheared margin) Poli (Cretaceous)
2000 m, lacks evidence of recent volcanic activity. Volcanism Takatu (northern Rift
and uplift without rifting is exemplified in North Africa spread as Mid-Atlantic
Ridge. eastern became
where Neogene volcanic rocks are to be found on the Tibesti transform. Takatu
ceased before spreading
and Ahagger Plateaus (Fig. 4.1). Domal uplifts with rifts in Cretaceous)
cutting across them are particularly common in East Africa
where the East African Rift System comprises a whole series Fig.4.13 Mode of evolution of triple junctions illustrated from
rift structures in Africa. (Adapted from K. Burke and A. J.
of such structures (Fig. 4.11). The crests of the domal uplifts Whiteman (/973) in: D. H. Tarling and S. K. Runcorn (eds)
typically have three rifts meeting at angles of about 120 0 ; Implications of Continental Drift for the Earth Sciences, 2 vols.
this geometrical arrangement appears to arise from the way Academic Press, London, Table 3, p. 749.)
 Landforms and tectonics of plate interiors 95

4.4 Continental basins denudation that has occurred. In most cases, however, this
appears to be insufficient to account for the amount of sub-
4.4.1 Basin morphology sidence observed.
In addition to upward vertical movements of the crust which If denudation is insufficient as a thinning mechanism then
give rise to broad upwarps or swells, downward movements sub-crustal processes must be considered. Phase changes pro-
also occur creating basins. Basins located in continental in- vide a possible mechanism which could account for subsi-
teriors are sometimes referred to as intra-cratonic basins dence as well as uplift, depending on whether the crystalline
while those located along continental margins in intra-plate transformations involved give rise to more, or less, compact
settings are called passive margin basins. The latter are mineral structures. A further potential thinning mechanism
primarily submarine features and will not be discussed here which has received considerable support since the late 1970s
although they are certainly an important component of passive is the stretching of the lithosphere as a result of sustained
margin development (see Section 4.5). tensional stress (Fig. 4.12). Lithospheric extension will pro-
Areas of crustal subsidence usually form topographic fea- mote faulting and an initial phase of subsidence associated
tures with minimal local relief because they become partly with upwelling of hot mantle from the asthenosphere into
or wholly filled by sediment transported from surrounding the base of the lithosphere. This is followed by a slower phase
highlands; examples include the Lake Eyre Basin in southern of subsidence promoted by cooling of this mantle material
Australia, the Kalahari Basin in southern Africa and the Ob as heat is conducted up to the surface.
Basin in the USSR. The volume of sediment that accumulates Irrespective of the mechanism initiating basin formation
in a basin depends on the rates of subsidence, sediment sup- the amount of subsidence will be amplified by the weight of
ply and sediment removal. Evacuation of basin sediments is sediments which accumulate. This is undoubtedly a highly
primarily by rivers for basins with an outlet to the ocean significant process where sediment thicknesses of several
(such as the Zaire Basin) but wind action is important in en- kilometres occur. It is also a major reason for the thickening
closed basins (such as the Kalahari Basin) (Fig. 4.1). Basins of sedimentary units towards the centre of basins. The initial
in continental interiors are a few hundred to a thousand or depression will be the first to accumulate sediment and this
more kilometres across and the sedimentary units, with which will lead to a positive feedback cycle of further subsidence
they are partially or completely filled, generally thicken and sediment accumulation extending outwards from the
towards the centre. This sediment thickenmg indicates the initial focus of subsidence. Because of the rigidity of the
role of sediment loading in sustaining basin subsidence but lithosphere, down-flexure may extend for 150 km or more
it does not explain how basins originate. Continental basins beyond the area of sediment loading. While subsidence is
generally seem to be long-lived features with a subsidence maintained by sediment loading, the surrounding upland areas
history often extending over a period of 100 Ma or more. supplying sediment will rise isostatically through unloading.
Once initiated such a complementary system of uplift and
subsidence is probably self-sustaining for long periods of
4.4.2 Mechanisms of subsidence
time in the absence of major tectonic disturbance. In enclosed
Basins form as a result of downwarping of the crust, as a basins water-loading by lakes may further contribute to sub-
consequence of uplift of the surrounding region, or through sidence, though their effect will be far more ephemeral than
a combination of both of these effects. Epeirogenic uplift and sediment loading as they are likely to undergo repeated phases
subsidence are closely related phenomena and a number of of growth and desiccation due to changes in climate.
the mechanisms of uplift discussed in Section 4.2 may be
relevant to crustal subsidence. For example, while heating'
of the lithosphere may account for uplift, subsidence may
4.5 Passive continental margins
result from lithospheric cooling. In this case the connection The sequence of uplift, rifting and continental rupture which
between uplift and subsidence is especially close as there leads to the formation of passive continental margins is of
must first be a heating event for the subsequent cooling to considerable importance in any attempt to understand the
give rise to a basin. On its own heating and cooling of the large-scale geomorphology of the present-day continents.
lithosphere will not create a basin as the crust will simply Moreover, passive margins have assumed great economic
return to its original elevation. What is required is a mechan- significance since the 1960s as their hydrocarbon potential
ism for thinning the lithosphere during the heating phase so has begun to be extensively exploited by the oil industry.
that when it cools and subsides a region of negative relief is The break-up of Pangaea (Fig. 2.16) generated a considerable
formed. length of passive continental margins and analysis of the
One possible way of accomplishing this is through the palaeomagnetic record of the continents and the subsequently
erosional thinning of the crust when it is elevated during the created ocean floor has given us a fairly good idea of the
period of uplift. The amount of thinning and consequently the timing of most of these rifting episodes.
amount of eventual subsidence will depend on the depth of A feature common to many, but certainly not all, passive
96 Endogenic processes and landforms

continental margins is a broad upwarp which separates the termed synrift sediments while those laid down on the mar-
coast from interior basins. Such upwarps consist of broad gin once continental separation has occurred are termed
swells running parallel to the coast and are in most cases postrift sediments. Interpretations of these sedimentary se-
flanked on their oceanward side by a great escarpment. When quences in conjunction with morphological and structural
rifting leads to the rupture of a continent the original elevation evidence has led to the development of numerous models
of the landsurface above sea level prior to rifting inevitably of the tectonic evolution of passive margins.
creates some relief once a new base level is established. But
a number of passive margins appear to have experienced
4.5.1 Active rifting
uplift either during or soon after rifting which has led to the
development of continental-margin upwarps which stand In the active rifting model (see Section 4.3.2) rifts propagate
higher than landsurfaces further inland. Clearly models of between a series of domal uplifts formed above sub-litho-
passive margin evolution must be able to account for the spheric thermal anomalies (Fig. 4. 15(A)). This gives rise to a
development of these upwarps. sequence of domes and intervening saddles which, as the
Our understanding of the vertical (as opposed to the hori- break-up of the continent proceeds and a new ocean is created,
zontal) movements of the lithosphere during continental rup- are split to form a sequence of ruptured domes and troughs
ture is largely derived from the interpretation of the sedi- along the newly formed passive margins. At this early stage
ments laid down in passive margin basins. Knowledge of of development these new continental margins are termed
these sedimentary sequences, which are now almost entirely nascent passive margins (Fig. 4.15(B)).
located below sea level, has been greatly enhanced by the As sea-floor spreading continues the margin ages and
application of seismic stratigraphy (see Section 4.3.1). subsidence begins to predominate over uplift, at least along
Together with evidence from boreholes drilled through con- its oceanward flank (Fig. 4. 15(C)). Such subsidence eventual-
tinental-margin deposits, investigations using seismic strati- ly leads to the submergence of the oceanward part of the
graphy have shown that most passive margins support a margin and the formation of a continental shelf. Subsidence
wedge of sediment which becomes gradually thinner as it during this mature passive margin stage is driven by both
passes seawards from continental to oceanic lithosphere thermal subsidence arising from the cooling of the margin
(Fig. 4.14). Deposits laid down during the rifting phase are as it moves away from the region of mantle upwelling

Flexural Hinge
node zone
I I
I I
I :
i : Continental Continental
Bulge : Coastal plain : shelf slope Continental rise
--------~. :......t-------~. :......t--------I.~I
I I
...J.o..lI- ------------------
I
I I I
I : :
I I I
I I I
, I I Sea level
-------_--.:""""'IIIIIII:I::~~ =:------.!....---'-'- - - - - - - - - - - - - - - - - - -

I
I

- - - - - - - - - - - - - - - - - -.....I......I---------I.~I ......I - - - - - - - - - - - - - - - - - - - -
Unstretched continental lithosphere Stretched Oceanic lithosphere
lithosphere

Fig.4.14 The major structural and morphological elements of passive (rifted) continental margins. Synrift sediments are those
deposited in a rift prior to continental rupture, whereas postrift sediments date from the initiation ofsea-floor spreading. The hinge zone
marks the line along the margin dividing continental lithosphere of normal thickness from lithosphere which has been thinned and
stretched during rifting and subsequently subsided under the load of overlying sediments. The flexural node separates the (ocean ward)
portion of the margin subjectto subsidence from that experiencing uplift. (Modified from A. B. Watts (1982) Nature 297, Fig. 2, p. 470.)
 Landforms and tectonics of plate interiors 97

Some models of passive margin tectonics assume that the

isostatic compensation in response to these changes in load
is local rather than regional, in which case no net uplift of
the landsurface occurs (see Section 4.2.3). But if the litho-
sphere along the margin has sufficient rigidity the resulting
imbalance in load will be accommodated through flexural
isostasy. This will lead to a net uplift of landsurfaces along
the least eroded part of the margin, that is, inland of major
escarpments. The magnitude of this effect is difficult to esti-
mate, but it will depend in part on the flexural rigidity of the
lithosphere. Nevertheless, flexural isostasy provides a poss-
ible uplift mechanism for passive margins during their mature
stage of development when the effects of the initial thermal
uplift have diminished.
Unless the zones of sub-lithospheric mantle upwelling
B Escarpment which precipitate uplift and rifting are in a line, sheared mar-
gins will develop along zones of transform plate motion
Upwarped margin
linking offset domal upwarps (Fig. 4.16(A)). Modelling by
. geophysicists indicates that a 200 Ian diameter convective
upwelling could produce a domal uplift about 1000 kIn across
with a maximum elevation of around 1000 m. Rift arms at
triple junctions which fail to spread will form aulacogens
striking at a high angle into the continental margin (Fig.
4.16(B), (C)).
The extensive series of domal uplifts traversed by the East
African Rift System has been widely regarded as a clear
example of active rifting with uplift preceding the develop-
ment of rift structures. Some recent work, however, has cast
doubt on this interpretation and suggests instead that rifting
c may have been initiated before uplift. The evolution of the
Mature passive margin nascent rifted margins of the Red Sea has also been interpret-
Escarpment ed both in terms of active and passive rifting mechanisms.
Broad continental shelf
Deposition along
margin
4.5.2. Passive rifting
The idea that simple extension of the lithosphere can produce
the structural features observed along passive margins was
first presented by D. McKenzie in 1978. In this initial presen-
tation of the passive rifting model it was assumed that the
amount of extension was uniform throughout the lithosphere
(Fig. 4.17(A)). Calculations based on this assumption,
however, indicate~ that any uplift would be confined to the
Subsidence caused
region over which stretching occurs and that no uplift at all
by cooling of would occur if the thickness of the' crust were less than
margin and sediment
loading about 20 per cent of the thickness of the lithosphere. In fact
Fig.4.15 Schematic representation ojpassive margin evolution this uniform extension model predicted that in most geo-
through active rifting showing the initial uplift and rifting (A), logically reasonable situations subsidence caused by thinning
nascent passive margin (B) and mature passive margin (C) of the lithosphere through stretching would outweigh any
stages. thermal isostatic uplift arising from the partial replacement
of cool lithosphere by hot asthenosphere below the zone of
located at the site of rifting, and by the isostatic loading caused extension.
by the growing wedge of sediment accumulating offshore. Since this simple model was unable to explain the kind of
This sediment loading is coupled with isostatic unloading on uplift observed along rifted margins there were immediate
the landward part of the margin as material is transported attempts to modify it. One way the model was altered was to
offshore. assume that the the mantle part of the lithosphere stretches
98 Endogenic processes and landforms

A C !: ~'.:-:
\ I ..
J I
///~
1,-
t \., \ .:: :.~
I / \
I I \ ",
,-- .....
I ( "- I
/

\ I ' .... ;'

",---
I ",-

/ / /
,/ ", I I
,-"" / I
/ I
I
'" /
I
/
J

I /
" / Rifted margin I

~~t(
I
(
/ \ ..... \
/ '\ \ "\ \
x
I
,
/

I
I
/
.......
,
\
\
\ Y x ,
I
I
\
\
"-
'-

x Sea level
y
x y ------1c ------
---~~ "------ ~
Fig.4.16 Plan view of the sequence of continental break-up and
':'::.>.: :: ,::: passive margin formation under active rifting: (A) rifting stage;
B ) (B) nascent passive margin stage with the formation of a rifted
/ margin where the motion along the new plate boundary is largely
/
/ divergent, and the creation of a sheared margin where the motion
I Aulacogen
I is largely transform; (C) mature passive margin stage with failed
l rift arms (aulacogens) striking into the new continental margin.
I X-Y cross-sections indicate the general nature of associated
vertical movements across the margin and the dashed lines
",""
/ schematically indicate contours.
/
{
by a greater amount than the crust during extension (Fig.
I 4.17(B)). In this non-uniform extension (or depth-depen-
/
/ dent extension) model uplift will take place because
/
Rifted margin relatively dense lithosphere is replaced by warmer and less
I
Oceanic crust I dense asthenosphere below relatively less stretched and
\
\,.
, thinned crust. Uplifts of 1-2 Ian can apparently be produced
if the crust experiences very little or no extension and the
\
lithosphere is stretched by a large amount (Box 4.2).
Another modification of the model assumes that uniform
extension occurs but takes account of the fact that stretching
,of the lithosphere is likely to occur over a long period of
time, probably several million years. In his original model
McKenzie assumed, for ease of calculation, that the litho-
sphere 'is stretched instantaneously. If we make the more
realistic assumption that it takes a finite period of time then
we have to take into account the fact that replacement of the
lower lithosphere by hot asthenosphere during extension
will lead to heat being conducted laterally into unextended,
 Landforms and tectonics of plate interiors 99

A Box 4.2 Uplift by non-uniform extension

Sea
Level
We can calculate the maximum uplift to be expected during
non-unifonn, or depth-dependent, extension if we know
Crust ---------- how much the sub-crustal lithosphere has been stretched' in
./
----------_/ regions where the overlying crust has experienced little or no
Lithospheric mantle / extension. It is in fact often difficult to detennine these
stretching factors empirically, but we can make some
calculations using geologically reasonable values. The
maximum uplift (b.h) is given by
B
Sea b.h=
0.5Pma(Ta - Ts) I) (
I--
I)
Pc ~L
Crust !'Z@;;/ r <rL~~~I< where II is the initial thickness of the lithosphere, ~L the
proportionate amount of stretching experienced by the sub-
/-----------
crustal lithosphere, Pm and Pc the densities of the mantle
----------- ----------
Lithospheric m ~
(3300 kg m-3 ) and crust (2700 kg m-3 ) respectively, Ta - Ts
the difference in temperature between the base and the top
of the lithosphere (1300 0c) and a the volumetric coefficient
of thennal expansion (that is, the proportionate change in
volume accompanying a change in temperature) (a = 3.4 x
c 10-SOC-I).
Sea If we assume that the initial lithospheric thickness is 150 km
Level and the sub-crustal lithosphere is stretched by a factor (~L)
Crust '---'-""'t 'Zi1ii.2.. : >... ::.....: .. ::.:." .. of 1.1. (that is, its 'length' is increased by 10 per cent) then
the maximum uplift in unstretched overlying crust will be
~-------- about 0.37 km. But if we increase the stretching factor we
----------- obtain greater changes in surface elevation; thus for ~L =
~ithosphericmantl~ 1.2 we get 0.68 km and for ~L = 1.5 we get 1.35 km. These
higher figures are comparable to the amount of uplift observed
Fig. 4.17 Response of rifted margins to different modes of on a number of passive margins. The initial lithospheric thick-
passive rifting: (A) (instantaneous) uniform extension; (B) non- ness also affects the ultimate amount of uplift generated in
uniform (depth-dependent) extension in which the sub-crustal unstretched overlying crust. For example, if the initial litho-
lithosphere stretches by a greater amount than the crust and uplift spheric thickness is only 100 km where the stretching factor
occurs in the zone of relatively unthinned crust; (C) uniform (~L) is 1.5, the uplift is reduced to 0.9 kIn.
extension over a finite period of time with uplift promoted by
lateral heat flow into adjacent unthinned lithosphere. (Modified
from C. Beaumont et al. (1982) Philosophical Transactions of the active rifting is that the convective upwelling is a conse-
Royal Society London A305, Fig. 10, p. 310.) quence, not a cause, of lithospheric thinning and rifting. The
secondary convection model has been applied to the Gulf
and therefore unthinned, lithosphere, as well as vertically into of Suez at the northern end of the Red Sea. Here significant
the overlying upper lithosphere (Fig. 4.17 (C)). Depending uplift of rift flanks has occurred although there is convincing
on the amount of heat involved this mechanism could lead evidence (from the deposition of marine sediments at the
to a significant uplift through thermal isostasy due to the time of rifting) that rifting preceded uplift.
heating of this unthinned lithosphere. Since this heat transfer The sequence of rifted margin evolution to be expected
will occur only slowly the maximum uplift of the passive from passive rifting is illustrated schematically in Figure
margin would probably occur as long as 60 Ma after rifting; 4.18. This figure is non-committal as to the precise mechanism
this contrasts with the non-uniform extension model in which likely to lead to the generation of passive margin upwarps
uplift takes place concurrently with stretching and rifting. as it illustrates secondary convection as well as non-uniform
A further approach to explaining the development of (depth-dependent extension) and the effects of lateral heat
significant continental-margin upwarps as a consequence of flow into unthinned lithosphere. As is evident from comparing
passive rifting has been to invoke a process known as sec- Figures 4.15 and 4.18, by the mature stage continental mar-
ondary convection. In this model it is proposed that thin- gins formed by both active and passive rifting tend to assume
ning of the lithosphere and its partial replacement by hot a similar mode of evolution, with subsidence, at least in the
asthenosphere might allow convective upwelling to become offshore zone, predominating over uplift as the margin both
established below the zone of extension. It is envisaged that cools and is loaded by sediment shed from the adjacent land
this active mantle upwelling would lead to a significant mass.
heating of the overlying lithosphere and could generate uplifts A further component of passive rifting that needs to be
of up to around 1 km. The difference between this model and considered is whether a rifting event is accompanied by
100 Endogenic processes and landforms

Symmetrical rift rock to move towards the surface. The resulting reduction
in pressure, or decompression, experienced by this material
causes it to melt. Some of the magma generated is erupted
on to the surface to form enormous lava flows (see Chapter
5), but by far the largest proportion is accreted to the base
of the crust, a process known (somewhat misleadingly) as
underplating. According to the underplating model the
addition of volcanic rock thickens the crust and the result-
ing isostatic adjustment leads to the formation of a broad
Thinning ofstretched
hot-spot swell up to 2000 km across and an increase in
lithosphere surface elevation of up to 2000 m. It is indeed interesting to
B
note that the highest passive margins also -seem to be those
characterized by rift-related volcanism, just as the model
predicts. Although mantle plumes do not drive the rifting in
this model, the surface uplift caused by underplating would
assist the rifting process as diverging plates would tend to
slide down either side of a hot-spot swell.
All the models of passive rifting discussed so far assume
that rifting is more or less a symmetric process; that is, we
would expect the opposing passive margins formed through
continental break-up to have a similar structure and mor-
Heat flow from asthenosphere Uplift caused by lateral
heat flow into unthinned
phology because they have experienced a similar tectonic
lithosphere or depth-dependent
extension history. But does symmetric rifting accord with the evi-
c Mature passive margin Persistence of upwarp
Sediment movement
across margin
dence? We saw in our discussion of continental rifts (see
Section 4.3) that asymmetric rather than symmetric rifting
appears to be the norm, at least at the scale of individual rift
structures. Moreover, if we compare opposing passive mar-'
gins there seem to be a number of cases where their mor-
phology and structure are very different. For example the
western margin of southern Africa has a relatively narrow
continental shelf and a well-developed marginal upwarp,
whereas the opposing margin of the eastern coast of
southern South America to which it was joined prior to the
break-up of Gondwana has a very broad continental shelf
and no significant marginal upwarp. Another instance of
Fig.4.18 Schematic representation o/passive margin evolution
apparent passive margin asymmetry is provided by the
through symmetric passive rifting showing the initial rifting (A), eastern margin of Australia, which has a marked upwarp
nascent passive margin (B) and mature passive margin (C) which forms the Great Dividing Range, and the Lord Howe
stages. Rise which represents a now submerged fragment of
continental crust which rifted away from eastern Australia
abundant volcanism. It has been pointed out that enormous around 95 Ma BP.
thicknesses of basaltic lavas erupted during the early stages Studies of extensional terranes, such as the Basin and
of rifting characterize some passive margins, such as those Range Province of the south-western USA, using seismic
of eastern Brazil, western India and the south-eastern part methods have revealed shallow-dipping faults which appear
of southern Africa. A number of other passive margins, how- to extend through the entire lithosphere. These have been
ever, show no sign of such major volcanism. This observa- interpreted as shear zones representing surfaces of litho-
tion has led to the suggestion that whether a passive margin spheric detachment. If lithospheric extension occurs primar-
is volcanic or non-volcanic depends on whether the site of ily along such detachment faults rather than by thinning
rifting happens to coincide with the location of a mantle throughout the entire zone of lithospheric extension, then we
plume and its associated hot spot. would expect opposing passive margins to display a marked
Where rifting occurs over a mantle plume the sub- complementary asymmetry (Fig. 4.19).
lithospheric mantle will be up to 200°C hotter than normal Such detachment models predict that two types of passive
and large quantities of magma will be generated. This margin will be produced by continental rupture. An upper-
occurs because rifting of the lithosphere allows hot mantle plate margin is formed by crust lying above the detachment
 Landforms and tectonics of plate interiors 101

A Upper plate sidence due to the considerable thinning of the lower plate
Asymmetric rift
\ lithosphere. The resulting continental margin morphology is
likely to consist of a broad continental shelf and a rather
modest amount of uplift inland.
The behaviour of the upper plate is quite different. The
pulling away of the lower plate along the detachment fault
exposes the base of the lithosphere on the upper plate to hot,
rising asthenosphere (Fig. 4.19(B)). This substitution of less
dense asthenosphere for denser lithosphere will result in uplift
of the landsurface on the upper-plate margin. Oceanward of
this uplift zone, however, there will be subsidence since here
ASTHENOSPHERE movement along the detachment fault results in the substi-
tution of mantle for less dense lower crust. The coupling of
B this offshore subsidence and onshore uplift will induce a
flexure of the margin which will accentuate uplift inland.
Uplift caused by
emplacement of The resulting upper-plate margin morphology will comprise
hot asthenosphere
below crust and a broad marginal upwarp which drops abruptly down to a
under plating
relatively narrow continental shelf.
Superimposed on all these models of passive margin
rifting is the possible contribution to uplift and subsidence
made by flexural isostasy as material is eroded from the
margin and transported offshore. As with the active rifting
model this effect could generate uplift of the margin, even
during its mature phase when other mechanisms are largely
inoperative.

4.5.3 Passive margins: a key research focus

c Broad continental shelf The modelling of passive margin evolution is still in its
infancy. A wide range of mechanisms have been proposed
by geophysicists attempting to model the broad patterns of
uplift and subsidence observed but modelling has run well
aItead of empirical testing. Figure 4.20 summarizes some
of the key processes that appear to be involved. To date
relatively little attention has been paid to the detailed mor-
phological evidence that might indicate which, if any, of
the proposed models is applicable to any particular passive
margin. Nevertheless, it is impossible to underestimate the
Detachment faults

Fig.4.19 Schematic representation of the detachment model of ....- - - - - E

asymmetric passive margin evolution. Asymmetric rifting (A)
leads to the development of morphologically and structurally
distinct upper-plate (B) and lower-plate (C) margins.

fault while a lower-plate margin is developed on deeper

crustal rocks lying below the detachment zone and which
may be overlain by faulted fragments of the upper plate.
Upper- and lower-plate margins will differ significantly in Fig. 4.20 Summary of some of the major tectonic factors
their uplift and subsidence history. controlling the morphological evolution of rifted passive margins:
As displacement occurs along the detachment fault the V T = thermally-driven uplift; V/ = isostatic uplift associated with
denudational unloading; ST = thermally driven subsidence; S/ =
margin of the lower plate will be upwarped as the load pre- isostatic subsidence associated with sediment loading; r =
viously exerted by the upper plate is progressively removed rotation of the margin resulting from V, and Sf; E = erosion by
(Fig. 4.19(C)). The predominant effect, however, will be sub- escarpment retreat; S.l. = sea level.
102 Endogenic processes and landforms

importance of passive margins to an understanding of long-

Further reading
term landform development, a topic we will be returning to
in Chapter 18. It seems probable that many of the models The large-scale relationships between landforms and tectonics
outlined above will be found to be applicable to at least in plate interiors have received comparatively little attention
some passive margins, but much work remains to be done from geomorphologists so the literature from this perspective
in this field. This is especially true of sheared margins .is rather limited. King (1967) is still useful for its descrip-
which have been neglected by geophysicists. Their attention tions of the continental landscapes of plate interiors although
has been focused on rifted margins largely because of their his interpretations of the tectonic processes involved have
potential oil and gas reserves. now been superseded by more recent work. Ollier (1981)
provides an introduction to several topics discussed in this
chapter, and my own brief reviews in Progress in Physical
4.6 The break-up of supercontinents Geography (Summerfield, 1986, 1988, 1989) try to relate
Discussion about the way continental rifts form and how, in some of the work by geophysicists to problems of landscape
some ~ases, they develop into sites of continental rupture development. The morphology of passive margins is dis-
and sea-floor spreading raises the much broader question of cussed in general terms by Ollier (1985), and with respect to
why supercontinents break up. At least part of the answer specific regions by Ollier (1982) (eastern Australia), Summer-
to this question may lie in the contrasting properties of con- field (1985a) (Africa), Ollier and Marker (1985) (southern
tinental and oceanic crust. Continental crust is only about Africa) and Ollier and Powar (1985) (India).
one-half as efficient at conducting heat as oceanic crust, so There is an extensive and rapidly growing literature on the
if a supercontinent, such as Pangaea, covers a significant part tectonics of plate interiors, but unfortunately much of it is
of the Earth's surrace heat will build up in the mantle below written from a geophysical rather than a geomorphic per-
it. After perhaps 80 Ma or so the increase in the temperature spective and it is not readily accessible to readers lacking a
in the sub-litho'spheric mantle below the supercontinent will good grounding in physics and maths. Two general reviews
lead to uplift, rifting and continental fragmentation. New of the field presented at an introductory level are Bally et
ocean basins are created which become wider through time al. (1980) and Burchfiel et al. (1980). Both of these edited
but eventually subduction zones will become established, volumes contain a number of relevant and clearly written
either along previously passive margins, or within ocean papers on a variety of specific topics. A more detailed treat-
basins. This establishment of subduction zones results in a ment of epeirogenic uplift is to be found in McGetchin and
change from continental dispersal to continental aggre- Merrill (1979); in spite of its general title it is almost solely
gation which eventually leads to the reassembly of a large concerned with the Colorado Plateau, but it does contain a
land mass. brief but wide-ranging review of uplift mechanisms. An in-
This idea of a repetitive sequence of continental fragmen- teresting assessment of the differential uplift of the conti-
tation and aggregation is termed the Wil~on cycle after its nents through the Late Cenozoic is provided by Bond (1978).
originator J. T. Wilson. The average duration of each cycle A clear introduction to hot spots is provided by Burke and
seems, on the basis of geological evidence, to be very rougWy Wilson (1976), while a more recent assessment of their glo-
around 500 Ma. This is well beyond the time scale generally bal distribution is that by Stefanick and Jurdy (1984). Hot
relevant to landform development but it is still important to spots as a cause of uplift are discussed by Crough (1983) and
be aware of how present landscapes, especially in plate Gass et al. (1978), while Theissen et al. (1979), Summerfield
interiors, relate to the most recent episode of supercontinent (1985b) and Sahagian (1988) examine the relationship be-
fragmentation. For instance, dispersal patterns of the conti- tween epeirogeny and hot spot activity in Africa. The im-
nents following the break-up of Pangaea show that the now portant, but neglected, problem of the range of scales of
separate continents have been moving away from Africa, uplift associated with thermal mechanisms is addressed by
which itself appears to have remained relatively stationary. Le Bas (1980), and the idea of lithospheric vulnerability is
In other words Africa may have remained close to the region presented by Pollack et al. (1981). Crough (1979) explores
of major sub-lithospheric heating which preceded the break- the possible morphological consequences of continents
up of Pangaea and this may explain why it has a high mean migrating over hot spots and also cites a possible example
elevation today in spite of lacking significant orogenic belts. from North America (Crough, 1981).
Australia, on the other hand, has moved rapidly away from Turning to other possible uplift mechanisms, the phase
its original location on the edge of Pangaea and it has the change model is considered by Smith (1982) with particular
lowest mean elevation of the major continents. There is reference to Africa and Australia, and by Smith and Drewry
obviously much uncertainty in these broad generalizations, (1984) with respect to the Transantarctic Mountains. The me-
but geomorphic evidence has a key role to play in relating chanism of delamination is described by Bird (1979). Flexural
such global tectonic models to the geological history of the isostasy is discussed by Karner and Watts (1982) and Watts
continents. et al. (1982), and at a more introductory level by Watts
 Landforms and tectonics of plate interiors 103

(1981), but unfortunately these treatments focus on offshore References

subsidence rather than uplift inland. Anderson, D. L. (1982) Hotspots, polar wander, Mesozoic con-
Rift systems have been studied for many decades but have vection and the geoid. Nature 297,391-3.
received renewed attention since the early 1970s because of Bally, A. W., Bender, P. L., McGetchin, T. R and Walcott, R 1.
their association with crustal upwarps and their role in con- (eds) (1980) Dynamics of PLate Interiors. American Geophysical
Union, Washington DC; Geological Society of America,
tinental rupture. An accessible introduction is provided by
Boulder.
Palmason (1982); this is an edited compilation containing a Bally, A. W. and Snelson, S. (1980) Realms of subsidence. In: A.
brief but stimulating revie.w of continental rifts by Mohr D. Miall (ed.) Facts and PrincipLes of WorLd PetroLeum
(1982). More recent ideas about the half-graben structure of Occurrence. Canadian Society of Petroleum Geologists Memoir
many rifts are introduced by Frostick and Reid (1987) and 6,9-94.
Beaumont, C., Keen, C. E. and Boutilier, R (1982) A comparison
examined in depth by Rosendahl (1987) and in Frostick et
of foreland and rift margin sedimentary basins. PhilosophicaL
ai. (1986). The influence of pre-rift structures on eventual Transactions of the RoyaL Society London A305, 295-317.
rift form is considered by Versfelt and Rosendahl (1989) Bird, P. (1979) Continental delamination and the Colorado Plateau.
and the relationship between rifts and domal uplift in Africa JournaL of GeophysicaL Research 84, 7561-71.
is discussed by Burke and Whiteman (1973). Bishop, P. (1988) The eastern highlands of Australia: the evolution
of an intraplate highland belt. Progress in PhysicaL Geography
Much of the research into the tectonic evolution of basins
12, 159-82.
overlaps with work on passive margin development, and Bonatti, E. (1987) The rifting of continents. Scientific American
although written for students of sedimentology the book by 256(3), 74-81.
Miall (1984) provides a valuable introduction for geomor- Bond, G. (1978) Evidence for Late Tertiary uplift of Africa rela-
phologists on both of these themes. Bally and Snelson (1980) tive to North America, South America, Australia and Europe.
JournaL ofGeology 86, 47-65.
provide an extremely comprehensive review of basin types
Buck, W. R, Martinez, E, Steckler, M. S. and Cochran 1. R
and origins and Turcotte (1980) presents one of the more (1988) Thermal consequences of lithospheric extension: pure and
accessible reviews of the geophysics of basin subsidence. simple. Tectonics 7,213-34
A good introduction to the tectonics of passive continental Burchfiel, B. c., Oliver, 1. E. and Silver, L. T. (eds) (1980)
margins is provided by Scrutton (1982) while Courtillot and ContinentaL Tectonics. National Academy of Sciences,
Washington D C.
Vink (1983) and Bonatti (1987) look at the processes of
Burke, K. and Whiteman, A. 1. (1973) Uplift, rifting and the break-
continental rifting and break-up. The extent and magnitude up of Africa. In: D. H. Tarling and S.K. Runcorn (eds)
of uplift to be expected from active and passive rifting me- Implications of ContinentaL Drift for the Earth Sciences (2
chanisms are compared by Keen (1985), while Beaumont et vols). Academic Press, London 735-55.
ai. (1982) and Buck et ai. (1988) assess the patterns of Burke, K. C. and Wilson, J. T. (1976) Hot spots on the Earth's
surface. Scientific American 235(2), 46-57.
uplift likely to result from different models of passive rifting.
Courtillot, V. and Vink, G. E. (1983) How continents break up.
The uniform extension model is presented by McKenzie Scientific American 248(7), 43-9.
(1978) and the idea of non-uniform stretching is clearly Crough, S. T. (1979) Hotspot epeirogeny. Tectonophysics 61,
described by Rowley and Sahagian (1986). Steckler (1985) 321-33.
proposes secondary convection as an explanation of the uplift Crough, S. T. (1981) Mesozoic hotspot epeirogeny in eastern North
America. GeoLogy 9, 2-6.
in the Gulf of Suez region, White and McKenzie (1989)
Crough, S. T. (1983) Hotspot swells. AnnuaL Review of Earth and
outline the underplating model, and the asymmetric detach- Planetary Sciences 11, 165-93.
ment model of continental break-up is presented by Lister Frostick, L. and Reid, I. (1987) A new look at rifts. Geology Today
et ai. (1986). The eastern Australian passive margin is one 3,122-6.
region where there has been a fruitful interchange of ideas Frostick, L .E., Renaut, R W., Reid, I. and Tiercelin, J.1. (eds)
(1986) Sedimentation in the African Rifts Geological Society
and data between geomorphologists and geophysicists,
Special Publication 25.
although this has not resulted in a consensus as to either the Gass, I. G., Chapman, D. S., Pollack, H. N. and Thorpe, R S. (1978)
history of uplift or the mechanisms that have caused it. Geological and geophysical parameters of mid-plate volcanism.
Veevers (1984) provides a comprehensive background to PhiLosophicaL Transactions of the RoyaL Society London A288,
the history of the eastern Australian margin while Bishop 581-97.
Gurnis, M. (1988) Large-scale mantle convection and the aggre-
(1988) reviews the various models that have been proposed
gation and dispersal of supercontinents. Nature 332,695-9.
to account for uplift. Lambeck and Stephenson (1986) and Kamer, G. D. and Watts, A. B. (1982) On isostasy at Atlantic-type
Wellman (1987, 1988) provide contrasting interpretations continental margins. JournaL of GeophysicaL Research 87,
of the uplift mechanisms that have operated in the region. 2923-48.
Finally, the possible mechanisms underlying the cyclic Keen, C. E. (1985) The dynamics of rifting: deformation of the
lithosphere by active and passive driving forces. GeophysicaL
fragmentation and reassembly of continents are discussed
JournaL of the RoyaL AstronomicaL Society 80, 95-120.
at an introductory level by Nance et ai. (1988) and in detail King, L. C. (1967) The MorphoLogy of the Earth: A Study and
by Gurnis (1988). Both these discussions derive ideas from Synthesis of WorLd Scenery (2nd edn). Oliver and Boyd,
earlier work by Anderson (1982). Edinburgh.
104 Endogenic processes and landforms

Lambeck, K. and Stephenson, R. (1986) The post-Palaeozoic uplift Smith, A. G. (1982) Late Cenozoic uplift of stable continents in a
history of south-eastern Australia. Australian Journal of Earth reference frame fixed to South America. Nature 296, 400-4.
Sciences 33, 253-70. Smith, A. G. and Drewry, D. J. (1984) Delayed phase change due
Le Bas, M. J. (1980) Alkaline magmatism and uplift of continen- to hot asthenosphere causes Transantarctic uplift? Nature 309,
tal crust. Proceedings of the Geologists Association 91,33-8. 536-8.
Lister, G. S., Etheridge, M. A. and Symonds, P. A. (1986) Detach- Steckler, M. S. (1985) Uplift and extension at the Gulf of Suez:
ment faulting and the evolution of passive margins. Geology indications of induced mantle convection. Nature 317, 135-9.
14,246-50. Stefanick, M. and Jurdy, D. M. (1984) The distribution of hotspots.
McGetchin, T. R. and Merrill, R. B. (eds) (1979) Plateau Uplift: Journal of Geophysical Research 89, 9919-25.
Mode and Mechanism. Elsevier, Amsterdam. Summerfield, M. A. (1985a) Plate tectonics and landscape develop-
McKenzie, D. (1978) Some remarks on the development of ment on the African continent. In: M. Morisawa and J.T. Hack
sedimentary basins. Earth and Planetary Science Letters 40, (eds) Tectonic Geomorphology. Allen and Unwin, Boston and
25-32. London, 27-51.
Miall, A. D. (1984) Principles of Sedimentary Basin Analysis. Summerfield, M. A. (1985b) Tectonic background to long-term
Springer-Verlag, New York. landform development in tropical Africa. In: I. Douglas and T.
Mohr, P. (1982) Musings on continental rifts. In: G. PAlmason (ed.) Spencer (Eds) Environmental Change and Tropical Geo-
Continental and Oceanic Rifts. American Geophysical Union, morphology. Allen and Unwin, London and Boston, 281-94.
Washington DC; Geological Society of America, Boulder, Summerfield, M. A. (1986) Tectonic geomorphology: macroscale
293-309. perspectives. Progress in Physical Geography 10,227-38.
Nance, R. D., Worsley, T. R. and Moody, J. B. (1988) The super- Summerfield, M. A. (1988) Global tectonics and landform develop-
continent cycle. Scientific American 259(1), 72-9. ment. Progress in Physical Geography 12, 389-404.
allier, c. D. (1981) Tectonics and Landforms. Longman, London Summerfield, M. A. (1989) Tectonic geomorphology: convergent
and New York. plate boundaries, passive continental margins and superc~ntinent
allier, c. D. (1982) The Great Escarpment of eastern Australia: cycles. Progress in Physical Geography 13,431-41.
tectonic and geomorphic significance. Journal of the Geological Theissen, R., Burke, K. and Kidd, W. S. F. (1979) African hot-
Society ofAustralia 29, 13-23. spots and their relation to the underlying mantle. Geology 7,
allier, c. D. (1985) Morphotectonics of continental margins with 263-6.
great escarpments. In: M. Morisawa and J. T. Hack (eds) Tectonic Turcotte, D. L. (1980) Models for the evolution of sedimentary
Geomorphology. Allen and Unwin, Boston and London, 3-25. basins. In: A. W. Bally, P. L. Bender, T. R. McGetchin and R. I.
allier, c. D. and Marker, M. E. (1985). The Great Escarpment of Walcott (eds) Dynamics of Plate Interiors. American
southern Africa. Zeitschrift fur Geomorphologie Supplementband Geophysical Union, Washington DC; Geological Society of
54,37-56. America, Boulder, 21-6.
allier, c. D. and Powar, K. B. (1985) The Western Ghats and the Veevers, J. J. (ed.) (1984) Phanerozoic Earth History of Australia.
morphotectonics of Peninsular India. Zeitschrift fur Clarendon Press, Oxford and New York.
Geomorphologie Supplementband 54,57-69. Versfelt, J. and Rosendahl, B. R. (1989) Relationship between
PAlmason, G. (ed.) (1982) Continental and Oceanic Rifts. American pre-rift structure and rift architecture in Lakes Tanganyika and
Geophysical Union, Washington DC; Geological Society of Malawi, East Africa. Nature 337, 354-7.
America, Boulder. Watts, A. B. (1981) The U. S. Atlantic continental margin:
Pollack, H. N., Gass, I. G., Thorpe, R. S. and Chapman, D. S. Subsidence history, crustal structure and thermal evolution. In:
(1981) On the vulnerability of lithospheric plates to mid-plate A. W. Bally et al. (eds) Geology of Passive Continental
volcanism: Reply to comments by P. R. Vogt. Journal of Margins: History, Structure and Sedimentologic Record (With
Geophysical Research 86, 961-6. Special Emphasis on the Atlantic Margin). Education Course
Rosendahl, B. R. (1987) Architecture of continental rifts with Note Series #19. American Association of Petroleum Geologists,
special reference to East Africa. Annual Review of Earth and Tulsa, Section 2.
Planetary Sciences 15, 445-503. Wellman, P. (1987) Eastern Highlands of Australia; their uplift
Rowley, D. B. and Sahagian, D. (1986) Depth-dependent stretching: and erosion. BMR Journal of Australian Geology and
a different approach. Geology 14, 32-5. Geophysics 10, 277-86.
Sahagian, D. (1988) Epeirogenic motions of Africa as inferred from Wellman, P. (1988) Tectonic and denudational uplift of Australian
Cretaceous shoreline deposits. Tectonics 7, 125-38. and Antarctic highlands. Zeitschrift fur Geomorphologie 32,
Scrutton, R. A. (ed.) (1982) Dynamics of Passive Margins. 17-29.
American Geophysical Union, Washington DC; Geological White, R. S. and McKenzie, D. P. (1989) Volcanism at rifts.
Society of America, Boulder. Scientific American 261 (1),44-55.
This page intentionally left blank
 5
Landforms associated
with igneous activity

5.1 Extrusive and intrusive igneous activity Secondly, the rocks of intrusions are often more resistant to
Igneous activity involves the movement of magma (molten erosion than the surrounding country rock so they frequently
rock), or solid rock undergoing slow deformation, towards, form prominent landforms once exposed at the surface.
or on to, the Earth's surface. Extrusive igneous activity is The various types of igneous rock associated with differ-
more commonly referred to as volcanism and occurs where ent forms of extrusive and intrusive activity, and the differ-
magma erupts on to the surface either as flowing lava, or as ent types of magma from which they originate, play an im-
fragmental material thrown into the air by explosive volcanic portant role in influencing the ultimate landforms created. Ig-
activity. Volcanism directly gives rise to predominantly con- neous petrology, the subdiscipline concerned with the classi-
structional landforms, although not all volcanic activity fication and origin of igneous rocks, is a complex field. None
results in the development of volcanoes since some types of the less, for the purposes of understanding the development
eruption create extensive sheets of lava or fragmental of landforms associated with volcanism and intrusive ig-
material. neous activity, we need to appreciate a few key properties of
There are .five major types of volcanic activity. One very igneous rocks. Figure 5.1 presents a simple classification of
obvious form is the volcanism of island arcs and continental-
margin orogens associated with plate convergence. More ex-
tensive, however, are the vast sheets of basaltic lavas which ACID INTERMEDIATE BASIC ULTRABASIC I
have formed on the continents at various times in the past
EXTRUSIVE _---'-Rh....:..Yo::....lic.:..:te_ _.1--=.D::.:aC::.::ite:.........lI..:..A::.:.nd::.:e:::;.Sit:::e-;:'=B:::a:::sa:::lt=:;
and are referred to as continental flood basalts. Yet
,-I

volcanism is predominantly a submarine rather than a conti- INTRUSIVE \

/.
rl --;;-::::;::----,;:::--=---;:--::r~=~1
Granite
~D~o~ler=ite=:::;-I____=,.......,...,,.......,...,---,
pranodioritel Diorite I Gabbro I Peridotite
nental phenomenon since the outpouring of magma associ- 100 , - - - - - - - - - - : ; ; ; : ; ' - - , - - - - - - - - ,

ated with submarine eruptions far exceeds anything wit- I

I
nessed on the continents. This submarine volcanic activity is 80 I
I
of three types. In addition to the volcanic activity marking PLAGIOC,LASE

60 FELDSPAR
the creation of new lithosphere at mid-oceanic spreading I
I
centres, there are also volcanoes in the ocean basins I OLIVINE
40 Sodium-rich
representing hot spots, as well as flood basalts covering large I

parts of the ocean floor. 20

Intrusive igneous activity involves the movement of rock
bodies towards the Earth's surface. Such rock bodies may
50 40
be either molten, in the strict sense, or they may be solid and
SiO, %
move very slowly at elevated temperatures and pressures
through a form of plastic flow. A rock body emplaced in the Fig.5.1 Mineral composition of the common igneous rocks.
Note that dolerite is a medium-grained rock of basic composition
surround~ng country rock in this way is called an intrusion
whereas gabbro is coarse-grained. The terms 'acid' and 'basic'
or, alternatively, a pluton. The formation of intrusions can originate from the idea current in the nineteenth century that
affect the landsurface in two ways. During emplacement the silica is derived from silicic acid, although we now know that the
overlying strata may be significantly uplifted and deformed. silica content of a rock is not a measure of acidity.
108 Endogenic processes and landforms

igneous rocks based on their mineral and chemical compo- fluenced by the rate at which they cooled when formed, and
sition and on differences in grain size, or texture. not surprisingly we find that intrusive igneous rocks nearly
Igneous rocks can be classified on the basis of the pro- always have a coarser texture than volcanic rocks.
portion of silica (Si0 2) that they contain. Acid rocks have
more than 66 per cent Si02, intermediate rocks between 52
and 66 per cent, basic rocks between 45 and 52 per cent, and
ultrabasic rocks less than 45 per cent. The proportion of
5.2 Volcanism
silica is roughly related to the relative abundance of light- In addition to its importance in directly creating landforms,
coloured felsic minerals (quartz and feldspar) and dark- volcanism is also of geomorphic significance for other
coloured mafic minerals (silicates rich in magnesium reasons. The material erupted during episodes of volcanic ac-
and iron). The grain size of igneous rocks is strongly in- tivity can be dated by radiometric techniques and can

Fig.5.2 Volcanoes of central Java, Indonesia. Several large strato-volcanoes can be seen in this Landsat image which covers an area
185 km across. The peaks labelled are Muria (Mu) (1602 mY, Lawu (L) (3265 mY, Merapi (Me) (2911 m) and Sumbing (S) (3371 mY·
Murjo is now inactive and, as the image shows, is more deeply eroded than the other volcanoes. The volcanoes of this region are
among the most active and explosive in the world. Merapi,for instance, erupts frequently producing nuees ardentes and lahars. (Image
courtesy N. M. Short.)
 Landforms associated with igneous activity 109

therefore provide a minimum age for the land urface over generation of magma in the overlying crust, promotes a
which it lies. Since volcanoes often have a rather regular, sequence of volcanic activity at the surface (Fig. 5.3).
symmetric form when first created, their original form can Within volcanic arcs individual volcanoes tend to be
often be reconstructed with some confidence even after sig- aligned along essentially traight lines over distances of up
nificant erosion has occurred. Together with radiometric age to 1000 krn, the arcuate form not being evident at this
determinations of the most recent eruptive episode, such re-
constructions enable estimates to be made of the amount of
Sea level
material eroded over a known period of time. Volcanoes can
thus provide important test cases for estimates of rate of de-
nudation (see Section 15.4.1.2). Since volcanoes composed A
of similar rock types are also found in a range of climatic
environments they also provide the opportunity to compare LITHOSPHERE
the effects of climate on the rate and nature of denudation.

5.2.1 Distribution of volcanic activity

There are presently around 600 active volcanoes on the con- ASTHENOSPHERE
tinents or exposed above the sea as islands. A further several
thousand extinct volcanoes on land are known from their
form, structure or characteristic rock types. But these totals
are dwarfed by the immense numbers of submarine vol- Sea level
canoes in the world's ocean basins; at least 50000 are
known to exist on the floor of the Pacific Ocean. B
Confining our attention here to sub-aerial volcanic ac-
tivity, we have already considered some aspects of the distri-
bution of volcanoes at the global scale in the three previous
LITHOSPHERE
chapters. Most present-day volcanic activity is associated
with convergent plate boundaries and there is a considerable
concentration of volcanoes in the island arc systems (Fig.
3.6) and continental-margin orogens bordering the Pacific
Ocean. Around 60 per cent of active volcanoes are, in fact, to ASTHENOSPHERE
be found around the rim of the Pacific Ocean, with a third
of these concentrated in Indonesia (Fig. 5.2). But volcanoes
are also found in the interiors of plates where they are Sea level
as ociated with hot spots and tend to occur either as clusters
or lines.
Volcanic clusters occur both on continental and oceanic
c
lithosphere, and are particularly common on the African
Plate which apparently has a high concentration of hot spots LITHOSPHERE
(see Section 4.2.1). Clusters range in size from fairly restric-
ted features such as Tristan da Cunha in the South Atlantic,
to much more extensive volcanic groups covering a con-
siderable area such as the Azores in the North Atlantic and
the Galapagos Islands in the eastern Pacific. Active ASTHENOSPHERE
volcanism at anyone time is normally confined to a limited
number of centres within a particular cluster. Volcanic lines
are most abundant in the Pacific Ocean and comprise series
of volcanic islands (or seamounts in the case of submarine Fig.5.3 Formation of a volcanic line through the motion of
volcanoes) which typically become progressively younger lithosphere over a sub-lithospheric thermal anomaly. A volcano
along a line and often terminate in an active volcano. The marking the hot spot grows from the ocean floor as melting of the
Hawaiian I lands provide a classic example of a volcanic crust generates magma (A) and eventually emerges above sea
level (B). Plate motion finally moves the new volcano away from
line, and the generally accepted explanation of this type of the underlying sub-lithospheric thermal anomaly (C); the original
feature involves the passage of part of a plate over a sub- volcano is eroded and a new volcano develops above the zone of
lithospheric thermal anomaly which, through the periodic magma generation (see Section 17.6.3).
110 Endogenic processes and landforms

smaller scale. In fact, volcanic arcs appear to consist of series 75000 a BP an estimated 2000 krn 3 of material was dispersed
of slightly off-set linear elements which have been in- over a wide area of the East Indies.
terpreted. as r~presenting the subduction of discrete segments Interesting as these figures are, they do not give an accu-
of lithosphere. Volcanic lines and clusters also exhibit rate indication of the average rate of volcanic activity over
patterning at the regional and local scale with both radial and time. Estimates based on volumes of erupted material in fact
rectilinear arrangements being found. Such patterns appear show that the discharge from volcanoes forming island arcs
to reflect the way in which the lithosphere fractures as and continental-margin orogens averages a very modest
magma forces its way up to the surface. In anyone cluster 1 km 3 a -I. This rate has no doubt varied somewhat through
the spacing of volcanoes is usually consistent and is gener- geological time as the rate of lithospheric subduction has
ally close to the thickness of the lithosphere at that locality changed. It is, in fact, the much less spectacular, but more
(Table 5.1). The inference here is that the volcano spacing consistent, activity of volcanic centres related to hot spots
is controlled by fracture patterns in the lithosphere, with that have more impressive rates of production. The island
increasingly thick lithosphere exhibiting progressively more of Hawaii alone seems to have a long-term average rate of
widely spaced fracture patterns through which magma is ex- construction of over 0.4 km 3 a-I, while Iceland has sustained
truded. a rate of about 0.13 km 3 a-I during historic time, and a rather
lower average of 0.06 krn 3 a -lover the past 16 Ma.
Although the volume of continental flood basalts amounts
Table 5.1 Relationship between volcano spacing and estimated
to around 1x107 krn3, the long-term average rate of accumu-
lithospheric thickness in eastern Africa
lation is very modest. There have been periods in the Earth's
LOCATION AGE VOLCANO ESTIMATED history, however, when this rate has increased dramatically.
SPACING LITHOSPHERIC It has been estimated, for example, that around 600 000 krn 3
(km) THICKNESS
of basalt was erupted within just 1 Ma to form the Deccan
(km)
flows in India. On the 'basis of their vast volume, flood bas-
Erta-ali, Afar Quaternary 10 ± 3 16 alts on the ocean floor have erupted at a much faster long-
Dubbi, Afar ?Pliocene-Quaternary 19 ± 6 ?25 term average rate than their continental equivalents, although
Ethiopian rift PI iocene-Quaternary 43 ± 13 35-50
Gregory rift PI iocene-Quaternary 42± 11 35-50 the major contribution of volcanic material on the ocean
Eastern Uganda Miocene 72±9 75-80 floor comes from mid-oceanic ridge spreading centres. These
Addis Ababa Miocene-Pliocene 70± 10 80 create between 5 and 6 krn 3 a-I of new volcanic material (that
Ethiopian plateau Oligocene-Miocene 109 ± 22 ?80-120
is, new oceanic crust). This production is, however, episodic
Source: From C. M. Clapperton (1977) Progress in Physical Geography 1, as is illustrated by the volcanic record of Iceland, which is
Table 1b, p.390 based on data in P.A. Mohr and C. A. Wood (1976) Earth
both a hot spot and part of a mid-oceanic ridge. In the major
and Planetary Science Letters 33, 126-144.
eruption of Laki in Iceland in 1783 over 10 krn 3 of lava was
produced in just 50 days, and it spread to eventually cover an
area of 370 krn 2•
5.2.2 Volcanic activity through time
Rates of volcanic activity through time are especially cri-
The fact that volcanic activity is highly episodic in nature is tical in determining whether intra-plate volcanoes formed in
evident even from the elementary classification of volcanoes ocean basins (as opposed to volcanoes associated with sub-
into active, dormant and extinct. Although widely used, these duction zones) will reach sea level to form a volcanic island,
categories are often difficult to apply in practice. Extinct or remain submerged to form a seamount. Once it has been
volcanoes are those that have not erupted in historic time, formed at a mid-oceanic ridge new oceanic lithosphere sub-
whereas active volcanoes have been seen to erupt. The term sides as it moves away from a spreading centre and becomes
dormant volcano is applied during the period between erup- cooler, thicker and more dense (see Section 17.6.3). Con-
tions to those volcanoes thought to be potentially active. sequently, if a volcanic island is to be formed the rate of
The most obvious manifestation of volcanic activity growth of a volcano must exceed the rate of ocean floor sub-
comes from the infrequent but spectacular eruption of vol- sidence. Since intra-plate volcanicity is linked to effectively
canoes on land. Such individual explosive events can indeed stationary sub-lithospheric thermal anomalies, there is only a
involve impressive quantities of material. Around 10 krn 3 of limited amount of time available for a volcano to develop
volcanic debris was hurled into the atmosphere during the before plate motion removes it from its underlying region of
famous 1883 eruption of Krakatau (Krakatoa) in Indonesia, crustal melting and source of magma generation (Fig. 5.3).
but more than ten times this quantity was involved in the Dating of the volcanic rocks of the Hawaiian Islands has
1815 eruption of Tambora, another Indonesian volcano. shown that each of the main islands was formed in about
.Going back into prehistoric time even these volumes were 1 Ma. These volcanoes are large landforms which load the
far exceeded by the eruption of Toba, yet another volcano in underlying crust and are partly isostatically supported by
the very active Indonesian region. When it erupted around deep roots (Fig. 5.4). Taking into account the volume of
 Landforms associated with igneous activity 111

5.2.3 Products of volcanic activity

Sea level
Volcanic eruptions can involve both lava which flows over
the surface, and fragmented particles, or pyroclasts, thrown
into the air by violent explosions. All volcanic material which
falls through the air is described as tephra, while pyroclastic
flow deposits result from the rapid 'flow' of pyroclasts mixed
with abundant hot gas.

5.2.3.1 Lava forms

Lava is formed from magma by cooling and the loss of gas.
The behaviour of lava once erupted is significantly affected
by its viscosity, that is, its resistance to flow (see Section
8.2.1). This in tum is related to the composition and tempera-
ture of the lava. Acid lavas, such as those composed of
.rhyolite, which are derived from magma with a high silica
content, are the most viscous and are generally erupted at the
lowest temperatures (Fig. 5.5). They are consequently able
Fig. 5.4 Growth of a volcanic island associated with a hot spot to flow only very short distances before they solidify. Basic
showing the growth of the compensating root to maintain isostasy. magmas, however, which are predominantly of basaltic com-
This root is approximately twice as thick as the volcanic pile built position, have a much lower silica content and are erupted at
upon the ocean floor. Since the lithosphere has a certain degree high temperatures. Their low viscosity consequently allows
of rigidity some of the isostatic compensation will be achieved by
flexure which creates a 'moat' and peripheral 'bulge' around the them to flow over considerable distances. Low-viscosity ba-
volcano (see Sections 4.2.3 and 17.6.3). saltic lavas produce extensive flows, whereas acidic lavas
tend to produce flows which are convex in cross-section, the
degree of convexity increasing with lava viscosity. Once
erupted volcanic material lying both on and below the erupted the rate at which a lava flow cools and solidifies also
surface of the ocean floor, it can be estimated that a mini- depends on its depth. While aIm thick flow of basalt will
mum construction rate of between 0.005 and 0.01 km 3 a-I is
required for a volcanic island to be produced in 1 Ma on
lithosphere between 20 and 40 Ma old. The rate is lower on
younger lithosphere underlying shallow ocean, but greater RHYOLITEU

on older lithosphere underlying the deeper parts of ocean

DACIT~
basins. The Hawaiian Islands have attained elevations well
above sea level so their rates of growth have been much
higher, with 0.44 km3 a-I being estimated for the island of
Hawaii itself. Interestingly, this is about ten times the rate of a::
eruption of lava recorded over historic time. The reason for
this discrepancy is not known; it might be because there are .
short-term fluctuations in the rate of lava eruption and that

~E
recently the rate has been well below the long-term average.
An alternative explanation is that much of the growth of the
island has occurred through intrusive igneous activity rather
than the extrusion of lava on the surface.
A revealing comparison can be made between the growth
of Hawaii and Iceland. Hawaii has exhibited a much faster
rate of construction, but its volcanism will be short-lived as
it will be carried away from its magma source by plate mo-
tion. Iceland, on the other hand, is located on a mid-oceanic 10 3 +---L---..,...----,,...----r-----.----.-------,
45 50 55 60 65 70 75
ridge and consequently has not moved significantly with
respect to its hot spot since the opening of the North Atlantic
Ocean over 50 Ma BP. It has accumulated a pile of volcanic
Fig. 5.5 Relationship between Si0 content and viscosity for
material 10 km thick and it has five times the volume of various magma types. Note that there2 is a difference of 10 orders
Hawaii in spite of a long-term rate of volcanIC activity less of magnitude between the most viscous rhyolitic magma and the
than one-sixth as high. least viscous basaltic magma.
112 Endogenic processes and landforms

take around 12 days to cool from 1100 to 750 DC, the same lumps can become somewhat rounded through rolling and
amount of cooling will take 3 a for a 10m flow and 30 a for a they are piled up on top of each other to create pillow lava.
100m flow. The shrinkage of lava as it cools produces joints. These
Eruptions of lava give rise to a diverse range of surface are usually irregular, but in extensive flows of basalt a highly
forms depending on composition, viscosity and gas content. regular hexagonal jointing pattern is sometimes produced
Pahoehoe has a smooth, ropy surface as a result of the giving rise to the kind of vertical columnar basalt outcrops
partial solidification of a thin skin which is dragged into which form the Giant's Causeway in Northern Ireland, UK.
folds by continuing lava flow underneath. Pahoehoe is Such regular joint patterns appear to develop when the
formed when lava has a low viscosity, but when it cools, or centres of contraction are evenly spaced. The lines joining
loses some of its gas content, it is transformed into a much these centres represent the directions of greatest tensile stress
more viscous form. This is called aa and its surface is com- in the lava flow as it cools, and the hexagonal joint patterns
posed of jagged, angular fragments resembling clinkers (Fig. are generated by cracks forming at right angles to these
5.6). maximum stress lines.
Where there is abundant gas in erupting magma the re- A common feature of some flows, particularly of the
sulting lava contains numerous gas-filled vesicles, or cavities, pahoehoe type, are lava tunnels or lava tubes. These form
and has a sponge-like form known as scoria. If such vesicles as a result of the continuing movement of lava below a sur-
are extremely abundant pumice is formed. This has such a face crust. A main feeder lava tunnel often branches towards
low density that it can float in water and blocks of pumice the terminus of a flow rather like the distributaries of a river
may travel thousands of kilometres across the world's oceans. delta. A deep and often sinuous lava channel can form if the
Lava erupted underwater cools very rapidly with a plastic roof of a lava tunnel subsequently collapses. In low-viscosity
skin forming around lumps of still molten material. These lavas features known as lava levees may also develop.
These consist of solidified fragments of lava which are piled
up on either side of a flow by the hotter lava moving more
rapidly in the centre of the flow.

5.2.3.2 Tephra
Tephra can be categorized on the basis of the size of the
pyroclasts of which it is composed. The finest material «
2 mm across) is termed ash when freshly deposited, and
volcanic tuffwhen compacted, while material of intermedi-
ate size (2-64 mm across) is called lapilli. The largest calibre
pyroclasts (> 64 mm across), which usually show some de-
gree of rounding, are known as volcanic bombs. Volcanic
bombs become partially streamlined and rounded as they are
thrown through the air or pass rapidly through the magma
in the neck of a volcano when they are violently erupted.
Volcanic bombs can be composed of either lava, or of pre-
existing rock which has become incorporated into the magma
in the neck of a volcano. Most volcanic bombs are relative-
ly small, being about the size of a football, but much larger
bombs can be produced in very violent eruptions. For in-
stance, in one eruption of the Andean volcano Cotopaxi a
200 t volcanic bomb was hurled a distance of 14 km.
Volcanic eruptions invariably generate pyroclasts of a
range of sizes. The resulting tephra is deposited predomi-
nantly downwind and forms well-bedded deposits which be-
come progressively finer upwards. This size-grading occurs
because the coarser material generated by an eruption is
heavier and is therefore deposited first. For the same reason
large-calibre pyroclasts are concentrated close to the site of
eruption, whereas ash can be dispersed over distances of
Fig.5.6 Recentflows ofpahoehoe (left) and aa (right) on the hundreds, or even thousands, of kilometres. Such ash forms
flanks of Kilauea, Hawaii. a progressively thinner deposit as the distance from the site
 Landforms associated with igneous activity 113

of eruption increases. In regions of highly active volcanism tivity occurs beneath a glacier. This has been a relatively fre-
tephra can periodically blanket the landscape. quent event in Iceland and it can lead to the rapid melting
Since it can be dated radiometrically tephra provides a of large volumes of ice and the release of enormous quan-
valuable stratigraphic marker and a means both of deter- tities of water, creating a cataclysmic flood known as a
mining the minimum age of the landsurface on which it is jokulhlaup (although such floods can also be produced by
deposited, and of recording the activity of particular vol- other means (see Section 11.4.1)).
canoes. So valuable has the dating of tephra become that it
has been given a specific name - tephrochronology. In
Iceland each major volcanic eruption has resulted in a layer 5.3 Volcanoes
of tephra being deposited, and in many cases these have
covered much of the country. Tephra horizons are also wide-
5.3.1 Types of eruption
spread in New Zealand where ashes of Late Quaternary age
cover about half of the North Island. Volcanic eruptions involve the rapid release of enormous
quantities of energy. A typical violent eruption from an indi-
5.2 .3.3 Pyroclastic flow deposits vidual volcano has an energy of between 10 12 and 10 15 1. This
In some kinds of violent volcanic eruptions pyroclasts com- compares with the 10 16 J released by a one megaton hydro-
bine with large quantities of hot gases to form a fluid-like gen bomb. But major eruptions involve much larger amounts
flow which is capable of moving very rapidly even over very of energy. For instance, the 1980 eruption of Mount St Helens
low gradients. The great mobility of these pyroclastic flows was equivalent to about 30 one megaton hydrogen bombs,
is attributable to the volume of hot gas present and the heat- while the cataclysmic eruption of Laki in Iceland in 1783
ing of engulfed air which renders the solid material highly had an estimated energy of 1020 J (equivalent to 10 000 one
buoyant. An alternative term for such flows is nuees ar- megaton hydrogen bombs).
dentes (literally burning clouds). The initial form of a volcano is very closely related to
As a result of their great mobility pyroclastic flows can the nature of the eruptive activity that produces it. Eruptions
generate pyroclastic flow deposits, or ignimbrites, which can be of three primary types, each associated with a par-
cover a large area. The ignimbrites of the Lake Toba region in ticular form of ejecta - exhalative (gas), effusive (lava) and
Sumatra extend over 25 000 km 2, while those of the Taupo- explosive (tephra). All volcanic activity involves the ex-
Rotorua area in the North Island of New Zealand cover more pulsion of some gas but major landforms are only formed
than 26000 km 2. Although ignimbrites can be unconsolidated through the eruption of lava or tephra. None the less, gas is
deposits, the pyroclasts are frequently so hot when deposited crucial in the actual mechanisms which control the kind of
that the contained gas is expelled and the individual particles eruptive activity that occurs (Fig. 5.7). Various attempts have
are welded together to form a compact, impervious rock. been made to categorize types of volcanic activity on the
basis of the nature of the material erupted, and the violence
5.2.3.4 Volcanic products associated with water of eruptive events. A simplified version of such a classifi-
Magma contains up to 4 per cent water by weight, and some cation is presented in Table 5.2.
volcanic activity takes the form of eruptions of super-heated As magma rises towards the surface it exerts a consider-
water and steam from geysers (Fig. 4.3). Such hydrother- able pressure on the overlying rock. Of particular importance
mal activity is most common during the later stages of is the pressure exerted by the gas (especially CO2 , S02 and
volcanism. High concentrations of dissolved material, es- water vapour) which is formed through the "release of vol-
pecially silica and calcium carbonate, are often contained in atiles and by the conversion of ground water to steam. The
hydrothermal solutions, and on cooling this material is pre- first manifestation of volcanic activity at the surface is nor-
cipitated to form sinter. mally the exhalation of gas from tensional fractures. The
Where volcanic debris is mixed with water a rapidly mov- character of the ensuing volcanism is largely a function of
ing mud flow, or lahar, can be produced. Lahars can be a the viscosity of the magma which is, in tum, dependent on its
direct result of an eruption, where for instance freshly ejected chemical composition and temperature. As the magma rises
lava or pyroclasts melt ice or snow, or they can occur as a towards the surface the confining pressure drops and gas
result of the destabilization of the unconsolidated volcanic bubbles begin to form and move upwards. In fluid magmas,
debris on the slope of a volcano by an earthquake, or as a such as those of basaltic composition, these gas bubbles can
result of heavy rain. In some cases lahars travel large dis- expand freely. Consequently, on reaching the surface they
tances and form thick deposits up to tens of metres deep. The have a comparatively low residual pressure and explosive
1980 eruption of Mount St Helens, Washington, USA, led activity is very limited. This contrasts with the situation for
to a lahar which flowed some 27 km down the valley of the a viscous magma such as rhyolite in which gas expansion is
Toutle River and filled the valley bottom to a depth of 60 m. constrained in the lower levels of the magma column. On
Equally catastrophic results can arise when volcanic ac- reaching the surface the residual gas pressure is consequent-
114 Endogenic processes and landforms

Viscous rhyolitic

c lava plug

Fig.5.7 Schematic representation of three major forms of

eruptive activity related to magma type and associated mode of
gas expansion. In (A) the andesitic magma in zone 1 may be
either saturated or undersaturated with volatiles, but as it rises it
enters a zone of lower pressure (2) and bubbles of gas begin to
form (3). The viscosity of andesitic magma is sufficient to restrain
gas expansion up to the explosion zone (4) where the gas pressure
is able to overcome the resistance of the overburden. In (B)
magma of basaltic compositio'n is fluid enough to allow the
unrestricted expansion of gas bubbles and their relatively
quiescent exhalation at the surface. In (C) the rhyolitic magma is
so viscous that the bursting of gas bubbles does not occur until
very near the surface. Violent lateral explosions producing
ignimbritic eruptions consisting ofdense mixtures of gases,
liquids and pyroclasts are caused by the periodic blocking oj the
main vent by a viscous lava plug. (After C. M. Clapperton (1977)
Progress in Physical Geography Fig. 7, p. 394.)

ly very high and this results in the violent bursting of gas steam or gas are emitted are known as fumaroles or solfa-
bubbles in the magma and the generation of highly explosive taras, while the precipitation of minerals such as silica and
eruptions which generate large volumes of tephra. calcium carbonate around a vent can create a sinter mound.
The emission of hydrothermal solutions, steam and other Mud volcanoes are rather more impressive features which
gases without significant quantities of fragmental material or may attain elevations of up to'lOO m, eruptions of mud arising
lava can give rise to minor landforms. Vents through which, from the heating of shallow ground water. If the confining

Table 5.2 Classification of volcanic eruptions

TYPE OF ERUPTION TYPE OF MAGMA NATURE OF EFFUSIVE ACTIVITY NATURE OF EXPLOSIVE STRUCTURES FORMED AROUND VENT
ACTIVITY

Icelandic Basic, low viscosity Thick, extensive flows from Very weak Very broad lava cones; lava
fissures plains with construction of cones
along fissures in terminal phase

Hawaiian Basic, low viscosity Normally thin, extensive flows Very weak Very broad lava domes and shields
from central vents
Strombolian Moderate viscosity; Flows absent, or thick and Weak to violent Cinder cones and lava flows
partly basic, partly acid moderately extensive
Vulcanian Acid, viscous Flows frequently absent; thick Moderate Ash cones, explosion craters
if present
Vesuvian (strong Acid, viscous Flows frequently absent; thick Moderate to violent Ash cones, explosion craters
Vulcanian) if present
Plinian (extremely Acid, viscous Flows may be absent; variable Very violent Widespread pumice and lapilli;
strong Vulcanian in thickness where present generally no cone construction
Pelean Acid, viscous Domes and/or short, very thick Like Vulcanian but with Domes; cones of ash and pumice
flows; may be absent nuees ardentes
Krakatauan Acid, yiscous Absent Cataclysmic Large explosion caldera

Source: Based largely on G.A.·MacDonald (1972) Volcanoes. (Prentice-Hall, Englewood Cliffs) Table 10 -1, p. 211.
 Landforms associated with igneous activity 115

pressure at the surface is reduced by some mechanism, such also affect its form, and its size. The length of single eruptive
as the draining of a pool, steam may be generated and any episode, which can range from a few days to several years,
overlying fine-grained sediments liquefied. will depend on the eruptive energy available once the ground
The complexity of some volcanic eruptions is illustrated surface is broken. The total active life of a volcano, possibly
by the detailed observations made of Mount St Helens in involving hundreds of individual eruptions and spanning up
1980. After 123 a of quiescence, noticeable seismic activity to hundreds of thousands of years, will be related to much
began on 20 March, 1980. A week later the first steam and more deep-seated processes associated with the heat source
ash began to erupt from a newly formed crater on the summit. promoting crustal melting and magma generation.
During April the movement of magma within the volcano Volcanoes are highly variable in their morphology and
was indicated by a bulge which began to develop on its north- are therefore very difficult to classify satisfactorily. Moreover,
eastern flank; by 23 April this had grown to 100 m and was size itself is an important factor since greater quantities of
expanding at a daily rate of 1.5 m. On 18 Maya magnitude lava or tephra do not simply produce landforms of a similar
5 earthquake triggered a failure of the bulge and a huge form but larger size. Rather volcanoes can be said to have a
landslide crashed down the mountain. This caused the morphological capacity representing the maximum size at-
instantaneous release of gas and steam which had been tainable by a particular type of eruption and associated ejecta.
confined under high pressure and the whole northern flank .For instance, cinder cones, which are composed entirely of
of the volcano was removed by an enormous lateral blast. tephra, are structurally too weak to attain a large size (Fig.
Hot gas, steam and tephra devastated an area of 600 km 2 5.8). Consequently all large volcanoes are composed of lava,
and pyroclastic flow deposits covered several square kilo- or a mixture of lava and tephra. This notion of morphologi-
metres. A vertical eruption carried volcanic debris to an cal capacity is evident in Table 5.3 and Figure 5.9 in which
altitude of 25 km and overall 'the elevation of Mount St the main types of volcanic landforms are classified and
Helens was reduced by 350 m. illustrated.
Once created, denudational processes can rapidly modify
the original constructional form of a volcano. The rate of this
5.3.2 Volcano morphology
modification will be a function of the resistance to weathe-
The form of a volcano is controlled by several variables. The ring and erosion of the volcanic materials laid down, the ini-
most fundamental is the type of magma present since this tial relief created by the volcano and the prevailing climatic
influences both the type of eruption and the nature of the environment. The effect of erosion on volcano morphology is
erupted material. Eruptions can occur through either fissures beautifully illustrated in the Hawaiian Islands where the
or vents, again depending on magma type. Andesitic erup- degree of dissection of each island is proportional to the
tions typically occur from a single central vent, while both time elapsed since volcanic activity ceased. The island of
basaltic and, to a lesser extent, rhyolitic eruptions are more. Hawaii itself is currently active and shows little sign of
often associated with fissures, in part because the major phase erosion (Fjg. 5.10), whereas the island of Kauai, where most
of expansion in these forms of volcanism occurs near to the volcanic activity ceased around 4 Ma BP, is deeply eroded
surface. The most violent eruptions in fact leave some of the and cut by spectacular valleys (Fig. 5.11).
least impressive landforms; in extreme cases these consist
simply of a broad depression surrounded by a plateau of
5.3.3 Basalt domes and shield volcanoes
ignimbrite.
The eruption of highly viscous lava, such as that formed The low viscosity of basaltic lavas means that extensive flows
by the degassing of rhyolite or andesite, can create large are more common than distinct cones. Where lava erupts
domes which may reach a height of several hundred metres from a single vent a low exogenous dome may be formed
and a width of several kilometres. Small domes composed from a succession of flows, but basaltic lava flows so
of viscous lava may form within craters. These can be either readily that such features will only develop on nearly level
rounded tholoids, cylindrical plug domes or irregular surfaces. Exogenous domes grade upwards in size to shield
cumulo-domes. Highly explosive eruptions of gas-charged volcanoes which, together with plateaus formed by flood
magma originating in the upper mantle produce small, shal- basalts or ignimbrite sheets, rank as the largest volcanic land-
low craters calied maars. forms. The smaller versions range in height from 100 m to
A change in magma composition may occur during a 1000 m, but because their slopes typically have low gradients
single eruption as different parts of the magma chamber are (usually < 10°) they may have basal diameters of 2-20 km.
tapped, but more significant compositional changes, often The largest examples are to be found on oceanic lithosphere
leading to the production of more acidic ejecta, may arise marking the site of hot spots. The greatest group of shield
during the overall lifespan of a volcano as a result of chemical volcanoes on Earth forms the island of Hawaii. The two
changes in the magma reservoir. The duration of both an major volcanoes, Mauna Loa and Mauna Kea, have attained
individual eruption and the overall lifespan of a volcano will elevations of over 4000 m above sea level, but they rise
116 Endogenic processes and landforms

Fig.5.8 Sunset Crater, Arizona, USA., a small cinder cone formed by an eruption in 1066 AD. Note the fresh appearance of the lava
flow in the foreground which issued from the base of the cone during the eruption.

Table 5.3 Classification of volcanoes and related landforms

TYPE OF MAGMA TYPE OF QUANTITY OF MAGMA PRODUCED

ACTIVITY
Small ...
I
• Large

Fluid, very hot, basic Effusive Lava flows! Exogenous 2 domes Basalt domes and I Iceiandic 3A
in composition shield volcanoes

!
I
I Hawaiian 3B

Scoria cones with

Increasing viscosity, gas flows 4
content, and acidity (high Mixed Composite or strato-
proportion of silica) Loose tephra cones volcanoes 6 Volcanic fields with

I
with thick flows s multiple domes

Endogenous domes Ruptured endogenous domes

(plug domes, tholoids)7 with thick lava
r flows 8

Viscous, relatively cool, Explosive Maars of tephra 9 Maars with ramparts 10 Collapse . Ignimbrite sheets
acidic and explosion
Extremely viscous, Explosive, Gas maars ll Explosion craters I2 calderas I3
abundant crystals mostly gas

Note: The numbers refer to the diagrams in Fig. 5.9.

Source: Adapted from A. Rittmann (1962) Volcanoes and Their Activity, Wiley, New York, Table 4, p. 115.
 Landforms associated with igneous activity 117

Fig.5.9 Schematic block diagrams of simple central volcanoes.

The numbers refer to the volcano types classified in Table 5.3.
The forms of actual volcanoes are normally far more complex
since they usually result from a complicated history of numerous
eruptive episodes. (After A. Rittmann (1962) Volcanoes and Their
Activity. Wiley, New York, Fig. 57, p.114.)

more than 9000 m from the floor of the Pacific Ocean from
a base well in excess of 200km across (Fig. 5.12). Never-
theless, these dimensions are dwarfed by the enormous vol-
canoes now known to exist on Mars, the highest of which,
Olympus Mons, rises 26 km from the planet's surface (see
Section 19.3.2).

5.3.4 Strato-volcanoes
More common than purely basaltic lava volcanoes are those
. comprising a mixture of tephra and lava. Such cones are
known as strato-volcanoes or composite volcanoes. They
often assume a highly symmetric form with slopes at, or
close to, the angle of stability of the ejecta (Fig. 3.12). Large
strato-volcanoes invariably contain lava flows which help
to support the weight of overlying tephra and in fact most
contain a complex network of lava flows together with hori-

Fig.5.10 Gently sloping south-eastern flank ofMauna Loa, Hawaii, showing no evidence of significant erosion.
118 Endogenic processes and landforms

Fig.5.11 The deeply dissected north-west coast ofKauai, the oldest of the five major islands in the Hawaiian group (compare with
Figure 5.10).

zontal and vertical igneous intrusions known respectively logical capacity since craters below this size are normally a
as sills and dykes (dikes) (Fig. 5.13) (see Section 5.5). constructional component of strato-volcanoes, whereas those
Rather than having straight slopes, strato-volcanoes above are nearly always formed by subsidence and collapse
typically have concave profiles. Such a slope form has been following catastrophic eruptions. Such explosion calderas
variously attributed to the effects of lava flows emitted from formed by rapid collapse after the evacuation of underlying
the summit crater, the changing angle of slope stability in magma chambers are relatively common on the summits of
response to a change in the particle-size characteristics of the large strato-volcanoes and should be distinguished from the
tephra ejected, and the effects of slumping. But such a form subsidence calderas of large shield volcanoes which are
can also be anticipated from the pattern of deposition of associated with the non-explosive eruption of basaltic lavas.
tephra as an ash cloud spreads out from the summit crater Calderas created by violent explosions can be of enor-
during an eruption over a wedge-shaped sector of the cone, mous size. One on the Yellowstone Plateau in Wyoming,
the rate of ash fall-out being the reciprocal of the distance USA, covers an area of around 2500 km 2 and forms a large
from the vent. Although this effect may be instrumental in basin partly filled by massive sheets of ignimbrite. Smaller,
creating the characteristic concave profile, it is the subse- but still impressive, calderas are to be found at the summits
quent formation of lava flows and development of intrusions of strato-volcanoes. An example often cited is Crater Lake in
within the cone that support the structure of the volcano and Oregon, USA. (Fig. 5.14). Here a 600 m deep lake occupies
allow its form to persist. a 9 km diameter crater (Fig. 5.15). The frequent association
of massive quantities of ignimbrite with large calderas
suggests a genetic link. A favoured hypothesis is that the
5.3.5 Calderas persistence of a large body of magma allows it to become
Calderas are sometimes simply defined as large craters, the vertically stratified with the more viscous and gas-rich com-
lower size limit normally being a diameter of 2 km. Although ponent accumulating towards the top (Fig. 5.16). An un-
arbitrary, this size limit can be related to the idea of morpho- known mechanism which causes the magma to subside leads
 Landforms associated with igneous activity 119

Fig.5.12 The island of Hawaii, the largest shield volcano complex on Earth, seen in a Landsat mosaic. It reaches a height of 4135 m
above sea level and descends 4900 m to the ocean floor. The more recent eruptive activity is evident from the dark,fresh lava flows.
Most of those visible on the flanks ofMauna Loa, the prominent summit just to the south of the centre of the island, have beenformed in
the past 150 a. The area covered by the image is 185 km across. (Image courte~y N. M. Short.)

to a drop in pressure which promotes a rapid expansion of

5.4 Other extrusive igneous landforms
gas and tephra through fissures in the cone above. This initial We have already noted how certain kinds of highly explosive
eruption further lowers the pressure in the magma chamber eruptions can produce massive pyroclastic flows which create
and more eruption are generated. sheets of ignimbrite. In some cases these can completely
120 Endogenic processes and landforms

Fig.5.13 Schematic cross-section of a typical strato-volcano showing dykes (D), a dyke conduit feeding a lateral cone (L), lava flows
(F), a buried cinder cone (C) and a sill (S). Ash deposits are dotted, layers of coarser tephra are marked with small triangles and lava
flows are irregularly cross-hatched. (After G. A. MacDonald, (1972) Volcanoes. Prentice-Hall, Englewood Cliffs, Fig. 2.J, p. 23.)

cover the topography over an extensive area and form an in India and Early Jurassic in South America - suggest that
ignimbrite plateau. It is the eruption of low viscosity basal- they were associated with the break-up of Gondwana (see
tic lavas, however, that is responsible for the construction Section 4.5.2). The Early Jurassic Karoo basalts of southern
of the most impressive and extensive volcanic landforms. Africa have been reduced by erosion to an area of
Basaltic magmas are normally expelled at the surface in 50000 km 2 , but originally they were probably ten times as
a very hot and fluid state, and the resulting lavas are often extensive. This is suggested by the widespread occurrence
able to flow over considerable distances. Thus, in addi- in southern Africa of thick dolerite intrusions which are
tion to forming shield volcanoes and domes, basaltic lavas contemporaneous with the Karoo flood basalts and which
erupting from an extensive regional system of fissures can themselves have created significant thickening of the sedi-
generate flood basalts and create massive basalt plateaus. mentary sequences into which they are intruded.
Small-scale basalt eruptions can fill valleys and in some Since highly fluid flows have an initial slope of only
cases over-top interfluves: but large-scale flood basalts can around 10 they form monotonously flat constructional land-
completely bury a pre-existing topography as great numbers scapes. Nevertheless the erosion of such landscapes can ex-
of superimposed individual flows build up a basalt plateau. pose the successive lava flows which may then form low,
The Columbia Plateau, which covers 130000 km 2 of but laterally persistent, cliff-like faces in which the hexag-
the states of Washington and Oregon in the north-west onal joint patterns that are sometimes developed in basalt
USA, is composed of hundreds of separate basalt flows of are visible. More generally, the great weight of basalt erupted
Miocene age which in places reach a total thickness of at the surface may, in some cases, be sufficient to promote
2000 m and have buried a pre-existing relief of more than significant isostatic subsidence in the underlying bedrock.
1500 m. Even more extensive examples are the Deccan
Plateau of India which covers more than 500000 km 2 and the
vast Parana Plateau which extends over some 750000 km 2
5.5 Landforms associated with igneous intrusions
of Uruguay and southern Brazil. The location and age of Igneous intrusions form through the penetration of country
these massive outpourings of basalt - Cretaceous to Eocene rock by a mobile body of igneous rock which need not be
 Landforms associated with igneous activity 121

Fig.5.14 Landsat image covering an area about 180 km across of south-central Oregon, USA, showing Crater Lake (arrowed). The
left third of the image covers part of the Cascade Range which includes several volcanic peaks, while the right half consists of lava
plains which fringe the ,southern edge of the Columbia Plateau. Crater Lake itself was formed by a catastrophic eruption of Mount
Mazama, which originally had an elevation of at least 3600 m, around 6500 a BP. Sufficient ejecta were generated to cover a
surrounding area of nearly 13 000 km2 with tephra at least 0.15 m deep. (Image courtesy N. M. Short.)

molten. If subject to stress for a sufficiently long period of of thousands of years for rocks at the surface. Consequently,
time all 'solid' materials are capable of flow. The length of intrusions can be formed either by molten rock or by rock
time required for a solid to flow rather than fracture is known experiencing rheid flow.
as its rheidity and the type of flow produced is termed rheid Intrusions can be composed of a wide range of rock types,
flow. The rheidity of ice is a few weeks, that of alt about a although very large intrusions with a surface exposur p of
year, but igneous rocks will only exhibit rheid flow after a over 100 km 2 known as batholiths, often have an acidic
stress has been applied for hundreds of years, in the case of granitic composition (Fig. 5.17). Granite masses originate at
deeply buried rocks subject to elevated temperatures, or tens depth and rise very slowly to the surface as diapirs over
122 Endogenic processes and landforms

Fig.5.15 Crater Lake, Oregon, USA. Wizard Island, a small cinder cone, can be seen emerging from the lake (see Figure 5.14).

millions of years. The process involved is known as dia- In contrast to the irregular form characteristic of acidic
pirism, and diapirs rise because the hot rock of which they granite batholiths, large basic intrusions which typically have
are composed is les dense than the cooler surrounding a gabbro-type composition are normally layered and have an
country rock. Diapirism involves both the moving aside of . overall saucer shape (Fig. 5.17). Such intrusions, which rarely
overlying strata and its partial reincorporation into the rising attain the dimenSIons of batholiths, are known as lo'poliths
mass of rock. Pods of country rock surrounded by intrusive and their layered structure tends to lead to the formation of
material create roof pendants when subsequently exposed series of outward-facing scarps as they are exposed by
by erosion (Fig. 5.17). Processes along continental-margin erosion.
subduction zones often lead to the formation of massive Small intrusions occur both in association with larger
granite batholiths which play an important part in orogenesis. forms, and with extrusive igneous phenomena. They can be
They frequently underlie and support the most elevated conveniently categorized as concordant or discordant de-
sectors of continental-margin orogens, as is the case in the pending on whether or not they cut across the bedding of the
Andes (see Section 3.3.1). pre-existing strata (Fig. 5.17). The form they take is a func-
An important factor in the mode of landscape develop- tion of both the viscosity of the magma and the configur-
ment on granite batholiths once they are sub-aerially exposed ation of fractures and other lines of weakness in the country
is the joint structure that they form. Initially, three sets of rock. Where exposed by erosion even quite small intrusions
largely orthogonal joints develop, but as the overlying rock is can be significant features in the landscape, especially if they
unloaded by erosion the release of pres ure in the upper are composed of a rock type which is markedly more re-
100 m or so of the batholith generates a secondary et of sistant to weathering than the adjacent strata.
joints aligned roughly parallel with the surface. These joint Dykes are vertically discordant intrusions often composed
systems play an important role in the development of weath- of dolerite and they frequently exist in swarms. They are
ering forms (see Section 6.3.1) and of drainage patterns (see typically 1-20 m across although much greater dimensions
Section 16.2.2). occur; the Great Dyke of Zimbabwe averages 6-8 km across
 Landforms associated with igneous activity 123

Fig.5.16 The formation of Crater Lake caldera, Oregon, USA:

(A) before the eruption; (B) during an early stage of the caldera-
forming eruption, showing an ash cloud and a small ashflow
issuing from the central vent; (C) during the climax of the
eruption, with large ash flows erupting from the central vent and
from ring fractures on the volcano slopes while the summit starts
to sink as a series of blocks; (D) after the eruption; (E) in its
present state with new eruptions on the floor and the caldera
partly filled with water. The small volcano marked X and clearly
visible in Figure 5.15 is Wizard Island, a cinder cone formed
since the Pleistocene. (From G. A. MacDonald (1972) Volcanoes.
Prentice-Hall, Englewood Cliffs, Fig. 12-5, p.301, modified after
H. Williams (1942) Carnegie Institute, Washington, Publication
540.)

Stocks

MAJOR INTRUSIONS

Batholith

MINOR INTRUSIONS Dykes

Sill

Laccolith
Sill
Sill Laccolith
Sill Sill

Fig.5.17 Various types of major and minor intrusion. Lopoliths, dykes, stocks and cone sheets are discordant structures, whereas
batholiths, sills, laccoliths are concordant. Note that the diagrams are not to scale and that many other varieties of intrusion exist.
(Modifiedfrom B. W. Sparks (1971) Rocks and Relief. Longman, London, Fig. 3.1, p. 68, Fig. 3.2, p. 70, Fig. 3.7, p. 90 and Fig. 3.17, p.
101.)
124 Endogenic processes and landforms

Fig.5.18 A large contorted dolerite dyke cutting across the unconsolidated sediments of a shallow basin (pan), northern Cape
Province, South Africa.

and is over 500 km in length. In some cases curved concen- 5.19). These formed the topic of a classic study by G. K.
tric series of dykes radiate outward and upward from a com- Gilbert published in 1877. Subsequent investigations have
mon magma chamber at depth; such features are described led to some revisions of Gilbert's original findings and it is
as cone sheets. As with other intrusions, dykes form either now thought by some researchers that the laccoliths radiate
positive or negative features depending on whether they are from central discordant intrusions known as stocks (Fig.
more or less resistant than the surrounding rock (Fig. 5.18). 5.17). Uplift associated with the intrusion of these stocks has
The most areally extensive small intrusion is a sheet-like given rise to several peaks rising some 1500 m above the sur-
form called a sill. Sills can be hundreds of metres thick but rounding Colorado Plateau.
10 m-30 m is a more typical range. If formed from viscous
magma they are normally limited in lateral extent but where
composed of basic rocks they may extend for thousands of
Further reading
square kilometres. The vast series of dolerite sills which have There are a number of texts on the nature of volcanic acti-
intruded the Karoo sediments of southern Africa underlie an vity and the development of volcanic landforms. Excellent
area of over 500 000 km 2 and probably constitute a total vol- general introductions to volcanic processes are provided by
ume of in excess of 200 000 km 3. In this region, as in many Bullard (1976), Decker and Decker (1989), Francis (1976),
others, the intruded rock is more resistant than the surround- MacDonald (1972), Rittmann (1962) and Williams and
ing country rock and gives rise to prominent topographic McBimey (1979). Volcanic landforms are considered more
features (Fig. 7.26). specifically in the classic work by Cotton (1944) and in the
A laccolith is a type of sill which has thickened to prod- more recent books by OIlier (1988) and Green and Short
uce a dome which causes the overlying rock to be upwarped. (1971). Examples of more detailed analyses of volcanic land-
Perhaps the most noted examples of this kind of structure forms in particular regions are to be found in Soons and Selby
are provided by the Henry Mountains in Utah, USA (Fig. (1982) (New Zealand) and Yoshikawa et al. (1981) (Japan).
 Landforms associated with igneous activity 125

1500 metres 1500 metres

c=:J Tertiary diorite porphyry stock

~
Laccoliths and other intrusive bodies
of Tertiary diorite porphyry

E2J Rocks of Jurassic age

.m Rocks of Triassic age

~ Rocks of Permian age

Fig.5.19 Fence diagram showing the structural relations of the intrusions of Mount Ellsworth, Henry Mountains, Utah, USA. (from
C. B. Hunt (1953) United States Geological Survey Professional Paper 228, Plate 16.)

allier (1988) discusses the various types of volcanic prod- a comprehensive study of volcanism in Hawaii. The cata-
ucts and their associated landforms in some detail. Spry strophic collapse of volcanoes is discussed by Francis and
(1962) considers the development of columnar joints in ba- Self (1987) and the global occurrence of different volcano
salts while Walker (1973) examines the factors that deter- types is assessed by Suzuki (1977).
mine the length of lava flows. Self and Sparks (1981) con- Landforms associated with intrusions are considered by
tains papers on various aspects of tephra, and Fisher and Sparks (1971), while the classic work by Gilbert (1877) on
Schmincke (1984) and Sparks et al, (1973) deal with pyro- laccolith structures in the Henry Mountains, Utah has been
clastic flow deposits. reassessed by Hunt (1953). allier and Pain (1981) consider
Factors controlling the global occurrence of volcanism are an interesting case of an apparent diapir of gneiss that has
evaluated by Clapperton (1977) while more detailed analyses risen to create a major domallandform.
of volcano spacing are presented by Mohr and Wood (1976)
and Vogt (1974). Menard (1986) discusses volcanism with
particular reference to oceanic islands. The characteristics of References
the various types of volcanic eruption are considered in the Bullard, F. M. (1976) Volcanoes of the Earth. Texas University
general texts mentioned above. For examples of particular Press, Austin.
eruptions see Self and Rampino (1981) and Simkin and Fiske Chester, D. K., Duncan A. M., Guest, 1. E. and Kilburn, C. R. J.
(1983) on the cataclysmic Krakatau eruption of 1883 and the (1985) Mount Etna: The Anatomy of a Volcano. Stanford
University Press, Stanford.
extremely detailed and fully illustrated collection of papers
Clapperton, C. M. (1977) Volcanoes in space and time. Progress
on the 1980 Mount St Helens eruption edited by Lipman and in Physical Geography 1, 375-411.
Mullineaux (1981). Chester et al, (1985) provide a detailed Cotton, C. A. (1944) Volcanoes as Landscape Forms. Whitcombe
analysis of Mount Etna, while Decker et al, (1987) present and Tombs, Christchurch.
126 Endogenic processes and landforms

Decker, R. and Decker, B. (1989) Volcanoes. (2nd edn) W. H. Rittmann, A. (1962) Volcanoes and Their Activity (Translated by
Freeman, San Francisco. E. A. Vincent). Wiley, New York and London.
Decker, R.W., Wright, T. Land Stauffer, P. H. (eds) (1987) Vol- Self, S. and Rampino, M. R. (1981) The 1883 eruption of Krakatau.
canism in Hawaii. United States Geological Survey Professional Nature 294, 699-704.
Paper 1350. Self, S. and Sparks, R. S. J. (eds) (1981) Tephra Studies as a Tool
Fisher, R. V. and Schmincke, H. U. (1984) Pyroclastic Rocks. in Quaternary Research. Reidel, Dordrecht.
Springer-Verlag, New York. Simkin, T. and Fiske, R. S. (1983) Krakatau 1883 - The Volcanic
Francis, P. (1976) Volcanoes. Penguin Books, Harmondsworth. Eruption and its Effects. Smithsonian Institute Press,
Francis, P. and Self, S. (1987) Collapsing volcanoes. Scientific Washington., DC.
American 256 (6), 73-9. Soons, J. M. and Selby, M. J. (eds) (1982) Landforms of New
Gilbert, G. K. (1877) Report on the Geology of the Henry Moun- Zealand. Longman Paul, Auckland.
tains. United States Department of the Interior, Washington, D C. Sparks, B. W. (1971) Rocks and Relief Longman, London and
Green, J. and Short, N. M. (1971) Volcanic Landforms and Surface New York.
Features. Springer-Verlag, New York. Sparks, R .S. J., Self, S. and Walker, G. P. L. (1973) Products of
Hunt, C. H. (1953) Geology and geography of the Henry Moun- ignimbrite eruptions. Geology 1, 115 - 18.
tains region, Utah. United States Geological Survey Professional Spry, A. (1962) The origin of columnar jointing, particularly in
Paper 228. basalt flows. Journal of the Geological Society of Australia 8,
Lipman, P. W. and Mullineaux, D. R. (eds) (1981) The 1980 erup- 191-216.
tions of Mount St Helens, Washington. United States Geological Suzuki, T. (1977) Volcano types and their global population per-
Survey Professional Paper 1250. centages. Bulletin of the Volcanology Society of Japan 22,
MacDonald, G. A. (1972) Volcanoes. Prentice-Hall, Englewood 27-40.
Cliffs and London. Vogt, P. R. (1974) Volcano spacing, fractures, and thickness of
Menard, H.W. (1986) Islands. Scientific American Books, New lithosphere. Earth and Planetary Science Letters 21, 235-52.
York. Walker, G. P. L. (1973) Lengths of lava flows. Philosophical
Mohr, P. A. and Wood, C. A. (1976) Volcano spacings and litho- Transactions of the Royal Society London A274, 107-18.
spheric attenuation in the eastern rift of Africa. Earth and Williams, H. and McBirney, A. R. (1979) Volcanology. Freeman
Planetary Science Letters 33, 126-44. Cooper, San Francisco.
allier, c. D. (1988) Volcanoes. Blackwell, Oxford. Yoshikawa, T., Kaizuka, S. and Ota, Y. (1981) The Landforms of
allier, c. D. and Pain, C. F. (1981) Active gneiss domes in Papua Japan. University of Tokyo Press, Tokyo.
New Guinea: new tectonic landforms. Zeitschrift fur
Geomorphologie 25, 133-45.
Part III
Exogenic processes and
landforms
 6
Weathering and associated
landforms

6.1 The weathering system formed at great depths and at high temperatures and
Material carried to the sea by rivers, or transported by pressures this adjustment can involve a complete trans-
formation of their constituent minerals. Weathering occurs
glaciers or the wind, experiences some degree of chemical
through complex interactions between the lithosphere, the
decomposition or physical breakdown prior to being
eroded. Weathering is therefore an appropriate place to atmosphere, the hydrosphere and the biosphere, and gives
rise to three major types of product. Chemical processes
begin our look at the operation and effects of exogenic
geomorphic processes. Geomorphologists are concerned lead to the release of compounds in solution and the
creation of new mineral products, while physical processes
with the rates at which different weathering processes
cause the breakdown of the original rock into smaller
operate as a function of environmental conditions, and with
particles. Dissolved material may subsequently be re-
the nature of the weathered material that is produced. But
precipitated or be reincorporated into other minerals, but
they are also especially interested in how weathering gives
the great proportion is carried by rivers to the ocean.
rise to specific landforms. This contrasts with, for instance,
the soil scientist's interest in weathering which stems from
a concern for the way it contributes to differences in soil 6.1.2 Water in rocks and soils
characteristics and the release and movement of nutrients.
Water plays a vital role in nearly all mechanisms of
physical and chemical weathering. This arises in part from
the fact that it is a polar solvent; that is, the covalently
6.1.1 The nature of weathering
bonded H20 molecule has a positive charge at each end
Weathering can be divided into those processes involving balanced by a negative charge in the middle. The positive
chemical reactions and the formation of new minerals and negative parts of water molecules become· attached,
(chemical weathering) and those that involve only physical respectively, to the anions and cations of solids and free
changes (physical weathering). Although the differences them by neutralizing· their charge; water is thus a highly
between these two types of weathering are distinct in effective solvent. A second significant property of water is
theory, in practice they rarely operate separately; rather, the its ionized state. A proportion of H20 molecules are always
effects of one aid the operation of the other. For instance, a decomposed into hydrogen ions (H+) and hydroxyl ions
rock shattered through physical weathering will be more (OH-) (a state known as dissociation), the concentration of
liable to chemical weathering because of the increased H+ ions being expressed as pH. This is defined as the
surface area made available for chemical reactions. Con- negative logarithm to the base 10 of the H+ concentration in
versely, chemical weathering along microfractures in a rock grams per litre. At the standard temperature of 25 0 C there
will weaken it and help physical processes break it down are lo-7g of H+ ions per litre of pure water, giving a neutral
more rapidly. pH of 7. Lower pH values indicate acidity (higher concen-
Weathering can be defined as the adjustment of the tration of H+ ions), and higher values represent alkalinity
chemical, mineralogical and physical properties of rocks in (lower concentration of H+ ions). It is important to bear -in
response to environmental conditions prevailing at the mind that a change in pH of one unit represents a tenfold
Earth's surface. For some igneous and metamorphic rocks change in hydrogen ion concentration.
130 Exogenic processes and landforms

Water may enter the soil or bedrock simply through therefore influence both the movement and distribution of
percolation through interconnected voids between particles water in soil and rock, and the interaction of these two
under the force of gravity. But gravity is not the only factor factors gives rise to four zones of moisture storage (Fig.
controlling the distribution and movement of water in rock 6.1). Below the water table rock and soil pores are satu-
or soil, as is apparent from the ability of water to move rated; water movement is controlled by gravity and is
laterally and upward as well as downward. The most subject to hydrostatic pressure, that is pressure exerted by
important mechanism whereby horizontal and upward the weight of overlying water. Above the water table
movements of moisture can occur is the capillary suction capillary suction provides an additional control. A zone of
which affects the water films attached to soil and rock capillary saturation exists immediately above in which the
particles. This suction effect can be readily illustrated by moisture content, which does not vary with depth, is
imagining the rise of water in a thin (capillary) tube present in the form of continuous films of water around
through surface tension effects. If such a tube is placed particles with entrapped air between. Above this zone the
vertically into a tank of water the level of the water in the moisture content decreases upward to a point where the
tube will rise above that in the tank. The rise is a result of continuous film of moisture eventually separates into dis-
the capillary suction which acts against the force of gravity. crete droplets. The two zones characterized by continuous
The suction is relatively more effective in comparison with moisture films are referred to collectively as the capillary
the effects of gravity in a thin tube because the circum- fringe, the depth of which depends on pore size. In
ference of the tube (along which the suction effect materials composed of sand-sized particles this depth is just
operates) is large relative to its cross-sectional area (to a few millimetres, but in clay-sized materials in which the
which the effect of gravity is proportional). Interconnected voids are 1 11m or less across it can be tens of metres. The
pores in soils and rock possess a suction in a manner rate of capillary movement in very fine pores is, however,
analogous to a capillary tube and this can be measured by so slow that the theoretical equilibrium height over which
the upward movement of water against the force of gravity. capillary rise should occur is rarely attained in such cases.
Water in fine~grained materials with small' pores experi-
ences a greater capillary suction in just the same way that
the height to which water rises in a capillary tube increases as 6.2 Chemical weathering
the diameter of the tube decreases.
Capillary suction, together with the force of gravity,
6.2.1 Chemical characteristics of rock-forming
minerals
SOIL SURFACE

I
There are two main types of chemical bond eXlstmg
between atoms contained in the compounds constituting the
Discrete moisture Earth's rock-forming minerals - ionic bonds and covalent
fT1
films at grain contacts Z bonds. Atoms with eight electrons in their outermost
5z (valence) shell are chemically stable, but all the elements
with which we are concerned have either more or less than
c;tin~~ ~;istu~fil~; ~
i~ forming an interconnected
system of capillary films.
0
r=
this stable number. Those with one additional electron
readily lose it and become positively charged ions (cations)
J;! ~
~ Moisture content increasing ~
(for instance, K+, Na+) and those with one less tend to gain
~ with depth ~
an electron and become negatively charged ions (anions)
: r~:t~:: -m:l:u~e~I~S- ~ (for example, F-, Cl-). For other elements the loss or gain
~ ~1
T1
h With entrapped air pockets. of two electrons similarly creates ions with a double charge
MOisture content (for instance, Ca2+, 0 2-). Such ions are stable on their own
t at capillary saturation
but only form stable compounds when their electrostatic
WATER TABLE charge is neutralized by ionic bonding. In those elements
which gain or lose three or more electrons in becoming
ions, adjacent atoms can share electrons through covalent
bonding. Some elements, such as oxygen, are able to form
both covalent and ionic bonds.
This is a simplified picture because in reality most
h= height of eqUivalent rise in a capillary tube chemical bonds in the majority of rock-forming minerals
Fig.6.1 The zones of capillary moisture in regolith and soils. are intermediate between ionic and covalent (Fig. 6.2). As
(From M. A. Carson (1969) in R. J. Carson (ed.) Water, Earth the bonds with oxygen are particularly important in rock-
and Man. Methuen, London, Fig. 4.11.4, p. 100.) forming minerals Table 6.1 gives estimates of the pro-
 Weathering and associated landforms 131

Simple Ionic Structures

Bonding entirely
electrostatic,
Close packed spheres
with contact of
unlike charges
but separation
of like charges.
Resulting crystals
CRYSTAL FORM are simple in form.

Complex Structures (include most silicates)

Two components:
1 COVALENT BONDS
Linking Si (AI)
& 0 in a covalent
"skeleton" with
overall negative
CRYSTAL FORM charge.
2 ELECTROSTATIC BONDS
between metal ions
& the covalent
"skeleton':

Fig.6.2 Ionic and covalent bonding in minerals. Silicate minerals have a basic structure consisting of silicon atoms covalently
bonded between oxygen atoms to form tetrahedra with an overall negative charge. This is neutralized through electrostatic bonds with
various cations, most commonly potassium, calcium, sodium and magnesium (from C. D. Curtis, (/976) in: E. Derbyshire (ed.)
Geomorphology and Climate. Wiley, London, Fig. 2.2, p. 34.)

Table 6.1 Bonding characteristics and ionic properties of some common built from complicated silicate structures with complex
elements charge patterns (Fig. 6.3).
ELEMENT ION APPROXIMATE IONIC RADIUS (A)
IONIC CHARACTER (0 2-= 1.40)
OF BOND WITH
6.2.2 Chemical reactions: thermodynamics and
OXYGEN (%) kinetics

Predominantly ionic bonding Although chemical weathering reflects the tendency for
Potassium K+ 87 1.60 new minerals to be formed which are stable under con-
Sodium Na+ 83 1.16 ditions prevailing at the Earth's surface, the rate at which
Calcium Ca 2+ 79 1.12
Magnesium Mg2+ 71 0.72 these stable forms are produced is often very slow. This
Iron (ferrous) Fe 2+ 69 0.77 necessitates two complementary approaches to the study of
Intermediate ionic-covalent bonding chemical weathering: thermodynamics considers the ulti-
Aluminium AI3+ 60 0.53 mately stable forms by analysing the energy changes
Iron (ferric) Fe 3+ 54 0.65 involved in chemical reactions, while kinetics focuses on
Titanium Ti 4 + 51 0.61
rates and mechanisms of change.
Predominantly covalent bonding The thermodynamic approach can be illustrated using a
Silicon 48
Phosphorus 35 simple mechanical analogy (Fig. 6.4). All substances
Carbon 23 contain various amounts of energy within their chemical
Sulphur 20 structures in rather the same way that the ball in Figure 6.4
Source: Based on data in C.D. Curtis (1976) in: E. Derbyshire (ed.) possesses a potential energy proportional to its height
Geomorphology and Climate. Wiley, London, Table 2.2, p. 33. above some datum. When chemical reactions occur energy
is usually liberated as heat (although in some kinds of
reactions heat is absorbed) just as the ball loses potential
portion of bonding of an ionic type for a number of energy as it moves downslope. This liberation of heat
common elements. Although the relative dimensions of represents a net release of free energy (M?) (more strictly
ions (indicated in Table 6.1 by ionic radius) can be used to referred to as Gibbs free energy). Whether a particular
predict the general form of crystalline compounds where chemical reaction is likely to occur is related to the change
the bonds are largely ionic (such as the cubic structure of in free energy involved. This can be calculated by sub-'
fluorite (CaF2, Fig. 6.2», most rock-forming minerals are tracting the sum of all the free energies of the reacting
132 Exogenic processes and landforms

ARRANGEMENT OF NO. OF OXYGENS

SHARED WITH EXAMPLES Height
Si 0 4 TETRAHEDRA ADJACENT TETRAHEDRA
A
w~w
A
None Olivine
A
D
W.W
~ NESOSILICATES
D

B
8><8 SOROSILICATES Melilite

C Datum ' - - - - - - - - - - - - - - - - - - - - - -

2 Beryl
Fig; 6.4 Types of stability illustrated by a ball on an undulating
surface. The potential energy at each position is proportional to
its height; the vertical axis is therefore an energy scale. The ball
is stable at Band C since it has no spontaneous tendency to move,
D~~ Pyroxenes
~V···W 2 e.g. hypersthene but A and D are unstable positions since the ball will
diopside spontaneously move down the slope towards the nearest trough
INOSILICATES augite
losing potential energy (height) as it does so. Position B is not as

E:«.:
..'.
....•..; .....•..
:;
2 and 3
Amphibolite
e.g. hornblende
glaucophane
stable as C (it is metastable) since with the addition of a small
amount of energy (M) it could roll over the hump to the right to
position C where it would have a lower potential energy (height)
(M) than at its original position. In this model the most stable
INOSILICATES
state is the one with the least potential energy (lowest height).
(based on C. D. Curtis (1976) in: E. Derbyshire (ed.)
Geomorphology and Climate. Wiley, London, Fig. 2.3, p.38, and
Fig. 2.4, p.43.)
Micas
3 e.g. muscovite
biotite
tell us that they are unstable. We therefore have to consider
the kinetics of weathering by examining those factors
which affect the rate of chemical reactions. A fundamental
PHYLLOSILICATES
control is the degree of instability of the reacting system, or
in our mechanical analogy (Fig. 6.4), how far up the slope
Fig. 6.3 The primary structural forms of silicate minerals
(shared oxygen atoms shown by open circles): (A) discrete our ball is located. Another crucial factor is the concen-
tetrahedra; (B) a tetrahedra pair sharing one oxygen; (C) a ring tration of reactants; in the case of a solution (such as soil
of six tetrahedra each sharing two oxygens; (D) a single water) reacting with a solid (such as a rock) this will be
tetrahedra chain; (E) a double tetrahedra chain with the outward- related to the surface area of contact between the water and
facing tetrahedra sharing two oxygens, the inward-facing
the constituent minerals of the rock. A fine-grained,
tetrahedra sharing three oxygens and the creation of hexagonal
'holes' between chains sufficiently large to accommodate ions pe~meable rock, for instance, will provide a much greater
such as OH- (hydroxyl) or F- (fluorine); (F) a tetrahedra sheet surface area than a densely cemented, massive rock.
with a continuous network of hexagonal holes (this is the Temperature is another significant factor; this is expressed
structure shown in Figure 6.2). A final structural type of great in the Arrhenius equation which indicates that reaction
importance (tectosilicates) consists of a continuous three- o
rates approximately double for each 10 C rise in tem-
dimensional framework of tetrahedra in which all four oxygens
are shared; examples are quartz andfeldspar). (Based partly on perature. Catalysts can also greatly speed up chemical
A. Holmes, (1978) Principles of Physical Geology (Nelson, reactions, as is illustrated by the familiar oxidation of iron.
Sunbury-on-Thames) Fig. 4.10, p. 52.) In completely dry air this reaction proceeds very slowly,
but with the addition of water as a catalyst the rate of
oxidation increases dramatically. A further important
substances from the sum of the free energies of the reaction
control is the intensity of leaching, the downward move-
products. If the calculated free energy change is negative
ment of water through the weathering zone which leads to
this indicates that the reaction will occur spontaneously; the
the removal of soluble products.
larger the negative value the more readily the reaction will
occur and the more stable the reaction products will be in
comparison with the original reactants. 6.2.3 Chemical weathering processes
Unfortunately, weathering processes and products cannot
be understood simply in terms of the formation of stable 6.2.3.1 Solution
mineral forms since many minerals are abundant on the Solution (or dissolution) i the simplest process whereby
Earth's surface even though thermodynamic considerations minerals can be decomposed and involves water acting as a
 Weathering and associated landforms 133

solvent. The dissolution of quartz (a crystalline form of saturated water if dissolution is to be an effective
silica) provides an example: mechanism.
Some minerals have a marked ability to absorb water
Si02 + Si(OH)40 into their crystal structure through a reversible reaction
quartz silica in solution
(silicic acid) known as hydration. This can be illustrated by the
hydration of iron oxide:
The equilibrium solubility of a mineral represents the
extent to which it will dissolve in water; it i~ usually 2Fe203 + 2FeP3· 3H20
iron oxide hydrated iron oxide
expressed in ppm (parts per million by volume) or mg 1-1.
Some minerals, such as halite (NaCI), are highly soluble in This reaction is significant in chemical weathering because it
earth surface environments but others, such as quartz, have aids other chemical processes by introducing water molecules
a very low equilibrium olubility and dissolve in pure water deep into crystal structures.
at an exceedingly slow rate. Equilibrium solubility is
affected by the temperature and pH of the environment 6.2.3.2 Hydrolysis
(Fig. 6.5) while the rate of throughput of water is important In addition to acting as a solvent, water may react directly
in controlling the rate of dissolution. This last variable is with minerals through hydrolysis. This involves the replace-
particularly significant as the film of water in direct contact ment of metal cations (most commonly K+, Na+, Ca2+ and
with the mineral surface eventually becomes saturated with Mg2+) in a mineral lattice by H+ ions and the combining of
solutes, thereby hindering further dissolution. Consequently these released cations with hydroxyl (OH-) ions. The effects
the saturated zone must be constantly flushed by under- of this process can be illustrated by crushing different
minerals in pure water. The resulting abrasion pH
600 indicates the amount of exchange between H+ ions in the
I i:
I 'J water and cations in the mineral, the increase in pH being
I
!I
II., due to the abstraction of the H+ ions from the water (Table
500
I I, 6.2).
I CaC03 jl
The basic hydrolysis reaction can be illustrated by the
I
FeO
~
I I weathering of the silicate mineral albite (a sodium-rich
E 400 I i plagioclase feldspar) to the clay mineral kaolinite:
c. r
c. I AI 2 0 3 ,I
I 'i
/. 4NaAISi 30 s + 6H20 ~ AI4Si40 lO (OH)g + 8Si0 2 + 4Na+ + 40W
I ,I albite water kaolinite silica dissolved hydroxyl
~ 300
:0 I sodium ions
~
(5
I /
/" .I
en
I I I Note that some of the silicon is retained in kaolinite and
I .
200 I I I
that sodium is removed in solution. This reaction leads to
~
/ i
_ Si02 (amorphous
- r - - - __ sJ!ical ,/
/
I the production of hydroxyl ions and would make pore and
I --- - --':- /' A1 2 0 3 ! surface waters alkaline, yet in most environments these
'00
I I waters are neutral to slightly acid. Consequently, we can
I i
/ conclude that hydrolysis as such is not a realistic weather-
, Si0 2 (quartz)
o
".
_._o-""='. " ing reaction.
,', 6.2.3.3 Carbonation
pH
The bicarbonate ion (HC03-) is invariably present in
Fig. 6.5 Equilibrium solubility (expressed in parts per million weathering solutions and is easily the most abundant anion
(ppm)) in water of some components involved in chemical
weathering as afunction ofpH. Note that: (1) the equilibrium
in most surface waters. It is formed from the dissolution
solubility of quartz (Si0 2) (crystalline silica) is only about one- and dissociation of carbon dioxide in water in a reversible
tenth of that of amorphous silica; (2) ferrous iron (FeO) is soluble reaction:
below a pH of about 7, whereas ferric iron (Fe20 j) is insoluble
except under highly acidic conditions (pH <3.5); (3) alumina H 2 0 + CO 2 =; H 2 C0 3 =; H+ + HC0 3 -
(AI 20 j ) is insoluble in the pH range 4-10; and (4) the solubilities water carbon carbonic acid hydrogen bicarbonate
of calcium carbonate (CaCO j ) and silica (Si0 2) are inversely dioxide ion ion
related over the pH range 6.5-9. (Based on various sources
including H. Blatt et al. (1972) Origin of Sedimentary Rocks.
Carbon dioxide is fixed from the atmosphere by photo-
Prentice-Hall, Englewood Cliffs Fig. 10-12; p. 364; R. K.ller
(1979) The Chemistry of Silica. Wiley, New York, Fig. 1.6, p.42; synthesis and enters the weathering system through the
and F. C. Loughnan (1969) Chemical Weathering of the Silicate intermediary of respiration by plant roots and the break-
Minerals. Elsevier, New York, Fig. 15, p.32.) down of plant debris by bacteria. Carbon dioxide is
134 Exogenic processes and landforms

Table 6.2 Abrasion pH of some common rock-forming minerals

MINERAL GENERAL FORMULA ABRASION

pH

Olivine (Mg, Felz Si04 10-11

Augite (pyroxene) Ca(Mg, Fe, AI) (AISilz0 6 10
Hornblende (amphibole) (Ca, Nalz (Mg, Fe, Al)s (AlSi)s022 (OHlz 10
Dolomite (carbonate) CaMg (C0 3lz 9-10
Albite (feldspar) NaAISi 3 0 S 9-10
Biotite (mica) K(MgFeh (AlSi 3) 01O(OHlz 8-9
Anorthite (feldspar) CaAl 2Si 20 S 8
Hypersthene (pyroxene) (Mg, Felz Si 2 06 8
Orthoclase (feldspar) KAlSi 30 S 8
Calcite (carbonate) CaC0 3 8
Muscovite (mica) K 2Al 4 (Si 6 Al 2 0 20 (OH)4 7-8
Quartz Si0 2 7

consequently abundant in the atmosphere of soils, especially c Z

those characterized by high rates of organic activity; it may
reach a concentration of 10 per cent (in comparison with
:-A:-t_m:-o:-s-'-p_h_e_re _l_ _(9_)- - - - - - - - -
Liquid ,
0.035 per cent in the free atmosphere). The dissolution of
carbon dioxide in precipitation provides an additional
source of bicarbonate ions.
H20~02l"'"
Surface waters are in fact weak carbonic acid solutions
and we can write a more realistic reaction for the weather- HzC0 3

~
ing of albite in which the release of metal cations (in this
case sodium) is matched by the production of bicarbonate
ions, a reaction termed carbonation:
4NaAISi 30 s + 6H20 + 4C0 2 -> 8Si0 2 + AI4Si40 lO (OH)s
albite water carbon silica kaolinite
dioxide

+ 4Na+ + 4HC0 3 -
dissolved bicarbonate
sodium ions

In humid tropical environments, where organic activity is

very high and leaching intense, we would expect silica to
be removed in solution as silicic acid:
D Dissolution products

Fig. 6.6 General scheme of the reactions involved in the

4NaAISi 3 O s + 22H2O + 4C0 2 -> AI4Si 4O lO (OH)s chemical weathering of calcium carbonate. (After S. Trudgill,
albite water carbon kaolinite 1985, Limestone Geomorphology. Longman, London, Fig. 2.8,
dioxide p.17.)
+ 4Na+ + 4HC0 3 - + 8Si(OH)40
dissolved bicarbonate silicic
sodium ions acid
(Fig. 6.6). In the weathering of calcite, half of the bi-
carbonate is derived from the calcite itself:
In general it appears that the leaching of metal cations from
silicate minerals is controlled by the supply of acids; this CaC0 3 Ca2 + + 2HC0 3 -
calcite calcium bicarbonate
not only involves carbonic acid, although this is the most ion ions
pervasive, but also sulphuric acid and, more significantly, a
range of organic acids. The factors controlling the efficacy of limestone solution
Carbonation plays a particularly important role in the are complex (Fig. 6.7) but the role of temperature is of
weathering of calcareous rocks. Limestones contain a high particular interest. Whereas the rate of chemical reaction
proportion of calcium carbonate (nearly always in the form between carbonic acid and calcite increases, as we would
of the mineral calcite) and its weathering involves complex expect, with temperature, the equilibrium solubility of
o
reversible reactions with carbon dioxide in the soil or sub- carbon dioxide decreases with temperature (at 20 C it is
0
terranean atmosphere and carbonic acid in natural waters only half that at 0 C). This has the effect that high
 Weathering and associated landforms 135

"/
/
Saturated /
/
/
/
c:
o /C
.;:; /
co /
Permeability of
Soil and Rock
...c:
~

B/' /
/
Q)
(J
c:
o
u
+
N
co
U
Water Flow Rate Organic Acids
Undersaturated

CO 2 Concentration
Fig.6.8 Schematic representation of the effect on the potential
dissolution of calcite of mixing waters with different bicarbonate
concentrations arising from the non-linear relationship between
Dissolution CO 2 and Ca 2 +. The curve represents the equilibrium
Rate and concentration ofCa2+ in solution at given CO 2 levels. Mixing of
Solute Load waters at A and B occurs along the dashed line C in the zone of
undersaturation. (After S. Trudgill (1985) Limestone
Fig. 6.7 Key variables in the control of limestone solution. Geomorphology. Longman, London, Fig. 2.16, p. 21.)
(After S. Trudgill (1985) Limestone Geomorphology. Longman,
London, Fig. 3.1, p. 27.)
Oxidation acts as a weathering process in two distinct
concentrations of carbonic acid can be attained in cold ways. Various elements, such as iron, titanium, manganese
regions even though the rate of production of carbon dioxide and sulphur can be oxidized to form oxides or hydroxides.
by organic activity in such environments is relatively low. For iron this reaction is written:
The mixing of saturated waters with different equilibrium 4Fe2 + + 302 -> 2Fe203
solubility concentrations can also cause further decom- iron oxygen iron oxide
position of calcite (Fig. 6.8).
The concentration of dissolved CaC03 in surface waters Iron usually exists in the bivalent ferrous state (Fe 2+) in
varies considerably. Tropical waters are frequently found to most rock-forming minerals, but it can be converted by
be supersaturated with respect to the prevailing pH levels oxidation to the trivalent ferric form (Fe 3+) with the effect
and the equilibrium concentration of dissolved carbon that the neutral charge of the crystal structure is upset and
dioxide. This suggests that the solution of limestone in such can only be regained by the loss of other cations. This
areas must be controlled by components other than mechanism can lead to the collapse of the mineral lattice
carbonic acid and it seems likely that organic acids playa (as in the formation of the clay mineral vermiculite from
major role. biotite) and it also renders a mineral more vulnerable to the
operation of other weathering processes.
6.2.3.4 Oxidation and reduction The tendency for oxidation or reduction to occur is
Oxidation is the process whereby an atom or ion loses an indicated by the redox potential (Eh) of the environment.
electron and thus acquires an increase in its positive charge This is measured in units of millivolts (mV), with positive
or decrease in its negative charge. Oxygen dissolved in values registering an oxidizing potential and negative
water is by far the most common oxidizing agent. The values a reducing potential. Abundant oxygen is dissolved
reaction can be reversed by reduction which involves the in most surface waters and the Eh is predominantly positive
gaining of an electron. in weathering environments; that is, oxidation occurs spon-
136 Exogenic processes and landforms

+1000 Cation exchange capacity is affected by the pH of the

surrounding solution since under acidic conditions H+
+800
-..... - ----....
"::"
readily replaces metal cations. Where the environment is
", alkaline, however, metal cations are less easily detached
+600 and may even replace H+ adsorbed on to a mineral surface.
®I
+400 ACID I ALKALINE 6.2.3.6 Organic processes
OXIDIZING
I OXIDIZING
1
/
We have already referred to the role of organic acids in
+200
I 0 chemical weathering, the anions of which may be important
>
E
o ____ B I in the removal of metal cations from silicate minerals and
.c
w ---I®-- in the dissolution of carbonates. However, it is the forma-
-200
ACID @ I ALKALINE tion of chelating agents by organic processes that is of
REDUCING REDUCING
particular significance in weathering. These are able to
mobilize metal cations, such as Fe 3+ and A13+, which are
-400
virtually insoluble under normal Eh-pH conditions, a process
- 600 +------,------,-------,,----+-.,---~--~
known as chelation. Chelating agents can come from
o 2 4 6 8 10 12 various sources, but most are organic compounds either
pH
secreted directly by organisms such as lichen, or formed
Fig. 6.9 Range of Eh-pH values encountered in the through the decomposition of humus in the soil. Although
environment. The enclosed area represents the extreme limits of the effects of chelation are well known its detailed mechan-
natural Eh-pH values as indicated by the extensive collation of isms are not. It appears that a stable ring structure is formed
data by Baas Becking et al. (1960). The letters represent typical
between complexing agents and metal cations. The structure
values for various types of environments: (P) precipitation, (S)
stream water, (B) bog water, (W) waterlogged soils, (G) ground may be subsequently broken down by microbial activity
water and (A) aerated'alkaline environments. (Based on R. M. and the complexed cation precipitated.
Garrels and C. L. Christ (1965) 'Solutions, Minerals, and The dark staining characteristic of rock surfaces in arid
Equilibria. Freeman, Cooper, San Francisco, Fig. 11.1, p. 380 regions, which is also observed to a limited extent in more
and Fig. 11.2, p. 381 after L. G. M. Baas Becking et al. (1960)
humid environments, appears to be primarily a result of
Journal of Geology 68 Fig. 31,p. 276.)
biogenic processes. Desert varnish (alternatively rock
varnish), as it is termed, is composed of clay minerals,
taneously, though not necessarily rapidly. However, in oxides and hydroxides of manganese and/or iron, together
some localities, such as below the water table and in with detrital particles, which form a layer typically around
waterlogged soils, reducing conditions can prevail. Eh also 10-30 f.lm thick. It probably takes many thousands of years
varies with pH, becoming generally lower as alkalinity for a characteristic black to brown desert varnish to
increases. This relationship can be seen clearly by plotting develop, apparently through the action of bacteria which
pH and Eh values measured in various Earth surface preferentially oxidize and concentrate certain elements
environments (Fig. 6.9). Where a particular environment is (especially manganese) supplied from dust and surface
located on such an Eh-pH diagram provides a useful water.
indication of the nature of chemical reactions that are likely
to occur.
6.2.4 Products of chemical weathering
6.2.3.5 Cation exchange
Cation exchange is the substitution of one cation for 6.2.4.1 The weathering mantle
another of a different element in a mineral structure. This Chemical processes assisted by the physical disintegration
can occur with any mineral, but it is by far the most of the bedrock combine to produce a weathering mantle or
common in clay minerals which typically have rather regolith which, if differentiated into identifiable horizons,
loosely bonded cations on their surfaces which can be constitutes a weathering profile. The interface between
readily exchanged for cations in solution. Each type of clay weathered material and unweathered bedrock is known as
mineral has a different propensity for the exchange of the weathering front. The thickness of a weathering mantle
cations which can be measured by its cation exchange at a particular locality represents a balance between the rate
capacity (c.e.c.) in units of milliequivalents per 100 g (meq of bedrock weathering and the rate of removal of weathered
100 g-l) of clay. The mechanism is extremely important in material by denudational agents (Fig. 6.10). Depths of
the alteration of one clay mineral to another. The two weathering may exceed 100 m and exceptionally reach
cations most readily absorbed on to clay minerals are H+ 300 m or more.
and Ca2+, and the cation most easily released is Na2 +. Thick weathering mantles will only form in regions of
 Weathering and associated landforms 137

of weathering related to conditions at the weathering front.

A further complication in the interpretation of regolith
thickness is climatic change. In presently arid central
Bedrock weathering Australia, for instance, deep weathering profiles are to be
found which have clearly been inherited from a previously
Net more humid climatic regime. Their partial preservation has
erosion
I ~ been aided by the presence of siliceous crusts which are
Net \ } Steady
state themselves highly resistant to weathering and erosion (see
accumulation
Section 6.5).
The physical characteristics of weathering profiles depend
upon the rock type and its structural properties and miner-
.....o alogy as well as the intensity and nature of the chemical
.! Regolith erosion weathering processes. A number of geomorphic studies
CQ
a: have focused on weathering profiles developed on granite
under humid tropical and subtropical climates. Although
Time (Depth) considerable differences exist in profile characteristics on
Fig.6.10 Schematic representation of the achievement of a other rock types and under different climatic regimes, deep
steady-state weathering profile thickness dependent upon the rate weathering profiles on granite exhibit features which are
of weathering and rate of erosion. (After M. J. Crozier (1986) common to many (Fig. 6.11). There is typically a gradation
Landslides: causes consequences and Environment. Croom Helm, over tens of metres from fresh bedrock, through weathering
London, Fig. 4.8, p. 138.)
material incorporating corestones of relatively unweathered
rock, to clay-rich horizons capped by a soil..
minimal local relief where rates of erosion have been low The mineralogical composition of the various horizons
for a prolonged period. High rates of weathering are not a will be influenced by the mineralogy of the bedrock, the
prerequisite for the formation of deep weathering profiles if hydrological regime and, related to this, the intensity of
the rate of removal of weathered material is sufficiently leaching. For instance, ferrous iron taken into solution
slow. This observation is well supported by the presence of below the water table where reducing conditions prevail
very great depths of weathering on the Russian Platform. may be drawn upwards by capillary suction and precipi-
The general absence of deep weathering mantles in many tated as ferric iron in the upper part of the profile. Under a
regions of low relief in the mid-latitudes of the northern seasonally dry climate this leads to the formation of a
hemisphere can probably be attributed to the effects of characteristically red-coloured, iron-rich horizon overlying
glacial scouring during the repeated ice-sheet advances of a bleached pallid zone which has been effectively leached
the Pleistocene. of ferric iron.
Deep weathering profiles are widespread in the humid In some weathering mantles original bedrock structures
tropics, Where high temperatures and abundant rainfall are preserved. The term saprolite is given to this type of
create conditions in which the potential for chemical regolith and it reflects the operation of isovolumetric
weathering is great. However, when the weathering mantle weathering, that is, weathering accomplished without any
has developed to a depth of several metres the rate of water change in volume. This is probably a very common phenom-
movement at the weathering front becomes very slow. enon, but in many cases its occurrence cannot be proved
Consequently, under these conditions the rate of chemical either because the original bedrock lacked well-defined
decomposition at the weathering front is minimal, an effect structures or because any such structures have been
which is clearly evident in the typically very low concen- destroyed by organic activity.
trations of solutes recorded in streams draining highly
leached deep weathering profiles in lowland humid tropical 6.2.4.2 Mineral stability and the formation of secondary
regions (see Section 15.3). minerals
The situation on mountain slopes in the humid tropics is, The types and proportions of the various minerals in a
however, quite different as here high erosion rates lead to weathering profile are usually quite different from the
the rapid removal of weathered material. Deep weathering original bedrock. Some minerals seem to survive more Of
mantles cannot develop even though the actual rate of less unaltered even after being subject to prolonged wea-
weathering is probably very high as a result of the proxim- thering, whereas others decompose very rapidly. A classic
ity of the weathering front to the surface and the associated study by S. S. Goldich published in 1938 showed a con-
rapid downslope movement of soil water. We need, there- sistent variation in the stability in the weathering environ-
fore, to distinguish carefully between potential rates of ment of common silicate minerals occurring in a range of
weathering related to climatic parameters, and actual rates igneous and metamorphic rocks (F~g. 6.12). Goldich pointed
138 Exogenic processes and landforms

ZONE COMPOSITION AND % SOLID ROCK

Highly leached pallid zone

.....
composed predominantly of
<50
4 kaolinite with residual quartz
and small corestones towards
base

Abundant corestones still

3 50-90
largely interlocked

2 Minor weathering along joints >90

WEATHERING FRONT

Unweathered bedrock 100

Fig.6.11 Features occurring in afully developed weathering profile on granitic rocks. Based on B. P. Ruxton and L. Berry, 1957,
Bulletin of the Geological Society of America 68, Fig. 2, p. 1266.)

having few silicon-oxygen bonds (such as the neosilicates

IGoldich Weathering Series I IBowen Reaction Series I and inosilicates) generally being less stable than those with
Olivine Anorthite Olivine Anorthite more bonds (such as the phyllosilicates and tectosilicates)
I (ea-rich feldspar)
I I .
Augite
(pyroxene)
Pyroxene Intermediate (Fig. 6.3). The relationship can also be v'iewed thermo-
I AmJibole Co-No f1eldsp;rs dynamically, there being a larger change in free energy
Hornblende
(amphibole) accompanying the decomposition of an unstable mineral
I
Biotite Albite
I Biotite
;;
Albite i like olivine than a comparatively stable mineral like
I I ~/ ~g
(Na·rich feldspar)
I muscovite.
Orthoclase Orthoclase ~~
In addition to quartz, which can at least partially survive
(K-richfeldspar)
I I .~ 1)l
prolonged weathering in most environments, the mineral-
Muscovite Muscovite ~ il
I I ~'O ogical composition of the weathering mantle includes a
Quartz Quartz
range of secondary minerals (Table 6.3). These form by
Fig.6.12 The weathering series ofGoldich and the reaction the decomposition of primary minerals and the incorpor-
series of Bowen. ation of dissolved constituents into new mineral forms, by
far the most important of which are clay minerals. Clay
minerals occur in a bewildering variety of forms; only the
out that this weathering series closely matched Bowen's major groups are listed in Table 6.3. In most cases they are
reaction series which listed various silicate minerals in formed through the recombination of silica, alumina and
terms of their order of crystallization from a silicate melt. metal cations released during weathering into layered
Olivine, for instance, is the most unstable mineral in the phyllosilicate-type structures (Fig. 6.13). Other secondary
weathering series and crystallizes at the highest tem- minerals formed during weathering include those composed
perature in the reaction series. Quartz, on the other hand, is of oxides and hydroxides of iron and aluminium and, less
the most stable and crystallizes aJ the lowest temperature. significantly, of titanium oxide, as well as amorphous or
Part of the reason for this relationship between suscep- finel y crystalline silica.
tibility to weathering and temperature of crystallization The type of clay mineral produced during weathering
appears to lie in the relative strength of the bonds between depends on the composition of the circulating pore waters
oxygen and the cations in each mineral. But the structure of (specifically the concentration of dissolved silica and the
the bonding also seems to be significant, with minerals type and concentration of cations), the mineralogy of the
 Weathering and associated landforms 139

Table 6.3 Major secondary minerals formed in weathering environments

MINERAL GROUP COMPOSITION GENERAL FORMULA MAJOR VARIETIES COMMENTS

Oxides and hydroxides

Oxides of silicon Silicon dioxide { Si0 2 Quartz

Si0 2 Amorphous silica
Si0 2 • nH 20 Opaline silica Water content usually < 10%
Hydroxides of Hydrous alumina { AI 2 (OH)6 Gibbsite Has a clay-like layered structure
aluminium -AIO(OH) Boehmite
Oxides and hydroxides Iron oxide and Fe203 Hematite
of iron hydrous iron oxide FeO.OH Goethite
Oxide of titanium Titanium dioxide Ti0 2 Anatase

Clay minerals

Kaolinite group Hydrous aluminium Kaolinite is the most common

silicate variety
Illite group Hydrous potassium KAI 2 (AlSi 3)Oro (OH)2 Layered structure similar to mica.
silicate (K content varies and there is partial substitution Occurs in 1-, 2- and 3-layer forms
of Al by Mg and Fe)
Smectite group Complex hydrous AI 4Si s0 20 (OH)4 . nH 20 Readily hydrated with expansion
magnesium aluminium (variable replacement of Mg for Al and Al for Si of up to 10 times: montmorillonite
silicate with attachment of Ca and Na) is a common variety.
Chlorite group Hydrous silicate of (Mg, Fe)s AI(AISi 3)OIO (OH)9
aluminium iron and
magnesium
Vermiculite group Hydrous magnesium Mg 3[Si 40 ro l (OH)2 . nH 20 Structure intermediate between
silicate (some AI replacement of Si and introduction of smectite and chlorite
Mg, Fe and Al may occur)
Mixed-layered group Hydrous silicate Regular or irregular
interstratification of two clay
minerals. Illite-smectite and
chlorite-vermiculite mixed layers
are the most common
Palygorskite and Hydrous magnesium Modified amphibole-type double-
sepiolite group silicate chain structure rather than layered
phyllosilicate structure of other
clay minerals

KAOLINITE ILLITE SMECTITE (Hydrated) bedrock, the intensity of leaching and the prevailing Eh-pH
conditions (Table 6.4). Primary minerals with an existing

m ffi water
layers
layered structure, such as biotite, can be fairly readily
altered into a clay mineral (often illite which is similarly
potassium-rich), whereas most other silicate minerals re-

m ~
quire a more fundamental reorganization of their crystal
0 structure.
Thermodynamics can indicate which mineral species will

em] ;1
be in equilibrium with conditions prevailing in the weather-

~., 1
layers
about
ing environment; but the rate at which this equilibrium is
achieved is often extremely slow and consequently it is
15

ffir ~
Fig. 6.13 Basic structures of some major clay mineral groups.
All the clay minerals illustrated have a layered phyllosilicate

~' structure comprising successions of tetrahedral (5i-0) and

octahedral (Al-O or Al-OH) layers with, in some cases,
intervening layers ofabsorbed water. The various clay minerals
have a characteristic spacing of these layers (shown in the
0
0
e
OH
0

Si
8
Si-AI
•
AI
()

AI-Mg
0K diagram in angstroms (A) (10- 10 mY. (Modified from C. Ollier,
(1984) Weathering. (2nd edn) Longman, London, Fig. 6.5, p. 73.)
140 Exogenic processes and landforms

Table 6.4 Mineralogy of weathering products in relation to environmental controls

ENVIRONMENT pH Eh BEHAVIOUR OF MAJOR ELEMENTS MINERALOGY OF WEATHERING PRODUcrS

Non-leaching. Alkaline Oxidizing Some loss of Na+ and K+. Partly decomposed parent minerals.
Mean annual precipitation Iron present in ferric state Illite, chlorite, smectite and mixed-
0-300 mm. Hot layered clay minerals. Hematite,
carbonates, secondary silica and
salts. Organic matter absent or sparse
Non-leaching below water Alkaline to Reducing Some loss of Na+ and K+. Partly decomposed parent materials.
table neutral Iron present in ferrous state Illite, chlorite, smectite and mixed-
layered clay minerals. Siderite (iron
carbonate) and pyrite (iron sulphide).
Organic matter present
Moderate leaching Acid Oxidizing to Loss of Na+, K+, Ca z+ and MgZ+. Kaolinite with or without degraded
mean annual precipitation reducing Some loss of SiO z. Concentration (K-deficient) illite. Some hematite
600-1300 mm. Temperate of Al z0 3, FeZ03 and TiO z present. Organic matter generally
present
Intense leaching. Acid Oxidizing Loss of Na+, K+, Ca z+, MgZ+ Hematite, goethite, gibbsite and
Mean annual precipitation and SiO z. Concentration of boehmite with some kaolinite.
> 1300 inm. Hot Al z0 3, FeZ03 and TiOz Organic matter absent or sparse
Intense leaching. Very acid Reducing Loss of Na+, K+, Ca z+, MgZ+ Kaolinite, possibly with some
Mean annual precipitation and .some iron and Al z0 3. SiO z gibbsite or degraded illite. Organic
> 1300 mm. Cool and TiO z retained matter abundant

Source: Modified from F. C. Loughnan (1969) Chemical Weathering of the Silicate Minerals. Elsevier, Amsterdam, Table 21, p. 73.

important to know what factors control the rate at which most weathering environments (Fig. 6.14). Clay minerals
more stable weathering products are created. The intensity such as kaolinite, which have been stripped of metal cations
of leaching is probably the major such kinetic control in together with iron and aluminium oxides and hydroxides,
are prevalent in environments characterized by intense
leaching since here the large throughput of water removes
PRIMARY SEOONDARY MINERALS cations in solution and prevents their concentration in pore
WEA~l..Pc~r~JlfRATE I INTENSE LEACHING I vI~~t~1~ttrE waters within the regolith. Conditions of extreme leaching
MIN ERA L S • DEPTH IN PROFILE
TIME can even lead to a significant removal of iron and silicon to
1 . leave an aluminium-rich residue in the form of the mineral
,--_ _,Na,si "_.~ Smectite g;rg,Na'SI

Albite
-.L~ Mg, Na, Si, AI Kaolinite
gibbsite.
Si AI Where leaching is of only moderate intensity, cations
Na. Si. AI ~Na.Si I .

~Na,Si I ""Si
I AI released during weathering can build up in the solutions
I I moving through the weathering mantle and the formation of
K,Si ~ cation-bearing clays such as illite and smectite is favoured.
""O-,-tho-c-Ia-se'.-L

K,Si,AI
AI
~K,Si I ~Si I

J+ii,------.I
I K,Si I

r-----,~w Illite I Y' ~'; I

Fig.6.14 Schematic representation of the relationship between
leaching intensity and types of clay minerals formed. The main
Muscovite K,Si,AI ~ Kaolinite
Si, AI Gibbsite
constituent elements of each mineral are indicated together with
K,Si, AI ~K,Si I AI the progressive loss in solution of soluble components at each
~KSi I ~Si I
' I I stage of weathering. A range ofprimary rock-forming minerals
are shown in order to illustrate the role of bedrock mineralogy in
controlling the types of clay minerals produced under conditions
of weak leaching. Under intense leaching such differences are
gradually eliminated. Since the weathering ofa unit of bedrock
AI
~K,Mg I ~Si,Fe I and the leaching ofsoluble constituents will be progressive the
Fe sequences of clay minerals may, in some cases, represent a
temporal succession. However, irrespective of the duration of
weathering, gibbsite is only likely to form where there is free
drainage and intense weathering, Similarly, leaching intensity is
likely to vary down profile and this may be reflected in the
~Ca/N .. M~ ~ Si, Fe :
changing proportions of various secondary minerals
 Weathering and associated landforms 141

In arid and semi-arid regions leaching may be minimal and 0..--------r----r----------r--1

solutions in the weathering mantle can attain high concen- ® KAOLINITE
(f)
a:
L.U
trations of dissolved constituents. This condition commonly I

leads to the accumulation of calcium carbonate at, or near, 2 b

the surface, while under more extreme conditions of aridity
soluble salts such as gypsum (CaS04 . 2H20), halite (NaCI)
4
and thenardite (Na2S04) may be precipitated.
E
The intensity of leaching is never uniform throughout an
entire weathering profile since the throughput of water ILLITE QUARTZ
6
normally decreases with depth due to compaction and the
downward translocation of fine particles which reduces the
size of voids. The resulting reduction in the intensity of
leaching is usually reflected in mineralogical changes with
depth (Fig. 6.15). Clay minerals such as illite and smectite,
which are readily altered in the zone of active leaching
towards the top of a profile, may be relatively stable at
lower levels where leaching is less intense. Vertical dif-
ferentiation in mineralogy in weathering profiles may also
reflect the stage-by-stage alteration of primary rock 20 40 60 80 100
minerals. Smectite and illite may form preferentially near Mineralogy (Cumul. %)
the weathering front only to be eventually altered to
kaolinite, and perhaps gibbsite, as gradual lowering of the
weathering front and erosion at the top of the profile 0
effectively causes individual clay mineral particles 'to move ® BOEHMITE
(f)
up through the profile. a:
L.U
The relative abundance of dissolved constituents released 2 I

through weathering reactions into surface waters can be GIBBSITE

b
compared to their relative abundance in the bedrock in E 4
order to provide information on the types of chemical re-
~
actions that are occurring. A number of studies have ;;:::
0
attempted to rank the relative mobilities of the major et 6
constituents released during weathering, and although the c
precise order varies depending on the environmental con- £
~
0..
ditions the generally agreed sequence is OJ 8
0
Ca2+ > Mg2+ > Na+ > K+ > Fe2+ > Si4+ > Fe3+ > A}3+

10
6.2.5 Factors influencing chemical weathering
It has long been recognized that five factors control the o 20 40 60 80 100
nature of soil development: climate, parent material, topo-
Mineralogy (Cumul. %)
graphy, organic activity and time. The factors controlling
chemical weathering, and indeed weathering as a whole, Fig.6.15 Variation in mineralogy for weathering profiles
can be viewed in a similar way. Figure 6.16 illustrates the developed on clay slates at Goulburn, New South Wales,
Australia, under a humid, warm temperate climate (A) and on a
way in which the four environmental factors influence both
sandstone at Weipa, Queensland, Australia, under ahot,'
the thermodynamics and kinetics of weathering reactions monsoonal climate (B). The approximate transition between
by indicating the conditions required for maximum rates of bedrock and weathered material is indicated by a dashed line.
weathering to be attained. (Adaptedfrom P. S. Bayliss and F. C. Loughnan, (1964), Clay
These factors clearly operate at different scales. Apart Mineral Bulletin 5, Fig. 3, p. 358, and F. C. Loughnan and P.
from montane environments, where climatic gradients are Bayliss, 1961, American Mineralogist 46, Fig. 4, p. 212.)
typically sharp, the effects of climatic differences on
weathering tend to be most apparent at the continental and ably good correlation between the major climatic regions
global scale. Indeed attempts have been made to produce and particular weathering zones, the correspondence be-
world maps indicating the distribution of different comes very weak in more mountainous areas where
weathering products (Fig. 6.17). Although there is a reason- topographic factors can become predominant. This is
142 Exogenic processes and landforms

Fig.6.16 Schematic representation of the way in which

environmental factors can contrihute to high rates of chemical
weathering. (Modifiedfrom C. D. Curtis (1976) in: E. Derhyshire
(ed.) Geomorphology and Climate. Wiley, London, Fig. 2.5,
p.50.)

especially the case when the depth of weathering is

considered; although a fairly systematic relationship with
climate and vegetation is evident as a consequence of
variation in precipitation and temperature (Fig. 6.18), this
only applies in the absence of marked relief.

{j--------------
n
~
pl>ot
&\)

6 2
10km
1~
1000 0 1000
km

Zone of low temperatures

III Iii
Zone of very intense weathering
and low rates of chemical weathering.
characterized by kaolinite, gibbsite
Deep weathering possible in regions of
and aluminium oxides and hydroxides
minimal local relief
Zone of intense weathering
characterized by kaolinite and
iron oxides and hydroxides D Ice-covered areas

II
-----"
Zone of moderate weathering
characterized by illite and
smectite
Mountainous regions of highly
variable weathering with little
accumulation of weathering products

11
-------_.
--------- Arid regions with low rates
---------
of chemical weathering and little
-------_. accumulation of weathering products

Fig. 6.17 Global distribution of major weathering zones. Modified from N. M. Strakhov (1967) Principles of Lithogenesis Vol. 1.
Oliver and Boyd, Edinburgh, Fig. 1, p. 5.)
 Weathering and associated landforms 143

DRY CONTINENTAL TROPICAL WET-DRY TROPICAL WET-DRY

PERI-
GLACIAL HUMID MID-LATITUDE ~ SEMI-ARID &ARID ~ HUMID TROPICAL ~
_3000

.sE 2500
p'Oc;pLtion I
c:: 2000
o .. . . . . .. Litter.
production Q)
.~ 1500 Evapotranspiration I
_-~'--------~J I 25 5
20 «i
.5. 1000 T I I I
_:::"'-1" ·,·...:..-._.-._._._._.-.:/- ·-1 ::~-
-..........:.. . - ' - ' - ' - '
lempera~~_'-'I-'-'- 15 Q)
.~ 500 ___--".,..-e..:..::._+__. ------- .' 10 Co
• •• J 5 E
c: 0
O~

§
~~~1i1!r
Kaolinite
Unweathered

m.
_
Gibbsite
bedrock
Little chemical
decomposition
Iron and aluminium r:j:Q:«~l Predominantly
oxides and hydroxides ::0 p.o::,.: illite and smectite

Fig.6.18 Variation in weathering mantle depth and composition in relation to climatic and biotic variables. (Modifiedfrom N. M.
Strakhov, (1967) Principles of Lithogenesis Vol. 1. Oliver and Boyd, Edinburgh, Fig. 2, p. 6).

Temperature is an important factor in rates of chemical related to climatic controls, but vegetation type also varies
weathering, both through the direct effect it has on the rate at a local scale as a result of topographic factors and soil
of chemical reactions (see Section 6.2.2), and indirectly properites.
through its influence on rates of organic activity and hence Topography affects weathering largely through the way
the production of both soil carbon dioxide and organic it influences the movement of water through the regolith.
acids - both critical components in chemical weathering. Weathering rates and the intensity of leaching depend cri-
As long as there is an abundant supply of water high tically on the total throughput of water and this is likely to
temperatures tend to be associated with high rates of be far higher on a steep, well-drained slope than in flat terrain
weathering, although this does not necessarily result in with poor drainage. If drainage is too efficient, however,
deep weathering 'profiles for, as we have already noted, little water may be present in the regolith for much of the
where slope gradients are steep the products of weathering time and rates of weathering will consequently be reduced.
may be removed almost as soon as they are created. Parent material influences the nature and rate of weather-
Rates of weathering are potentially highest where tem- ing in two major ways. Its mineralogical composition,
peratures are high and rainfall is abundant. As indicated in partic'ularly in terms of the proportion of unstable mineral
Table 6.4 intensity of leaching, and hence the mineralogy phases present, will affect both the rate of chemical
of weathering deposits, is largely a function of temperature, decomposition and the kinds of secondary minerals that are
precipitation and drainage. Although the precise mineral- most likely to form. Although we might expect bedrock
ogical composition of weathering mantles is influenced by mineralogy to exert a pervasive control over the products of
bedrock mineralogy, the effects of variations in precipita- chemical weathering this is not the case, at least at the
tion are often evident with the relative abundance of broad scale, both because the rocks and minerals exposed at
kaolinite, gibbsite and iron and aluminium hydroxides the Earth's surface are dominated by just two or three types
showing a positive correlation with mean annual pre- (Table 6.5), and because prolonged weathering tends to lead
cipitation (Fig. 6.19). to a convergence of secondary mineral types irrespective of
Vegetation influences weathering, as we have already parent material mineralogy. The second factor is the
indicated, through the release of organic acids and in the physical nature of the parent material, especially its particle
supply of carbon dioxide to soil waters. Important to both size and permeability, which influences the rate of
these factors is the production of litter. This varies weathering though not the weathering products.
enormously not only between desert and forest ecosystems The final factor of time is important because of the slow-
but also between temperate forests with a typical range of ness of most chemical reactions at the Earth's surface.
0.1-0.3 x 106 kg km-2 a-I and tropical rain forests which There is always a significant lag between the establishment
produce 0.4-1.3 x 106 kg km-2 a-I. Organic activity is closely of a particular set of conditions in the weathering environ-
144 Exogenic processes and landforms

100 100

IRON
*'
Q)
80
*'
Q)
80
AND
"0 "0
0
x 'x0 ALUMINIUM
-0>- 60 -0>- 60
-
.c
iii
Q;
c 40
.c
"-
iii
Q;
c 40
HYDROXIDES

E 'E
>- >-
,'0
<1J
U U SMECTITE
Q)
,::: 20 Q)
> 20
';::;
co <1J
Qi Qi
a:: a::
0
500 1000 1500 2000 0 1000 2000 3000
A C
100 100

VERMICULITE
IRON
*'
Q)
80
*' 80
AND
Q)
"0
'x "0
'x0 ALUMINIUM
~

- -0>- 60
"0
>- 60

-
.c HYDROXIDES
KAOLINITE .c
iii iii
Q; AND SMECTITE
Q;
c 40 c 40
E HALLOYSITE 'E
>-
<1J
>-
<1J
U U
Q)
~ 20 > 20
.~ SMECTITE .~

Qi Qi
a:: a::
0 0
0 500 1000 1500 2000 0 1000 2000 3000
Mean annual precipitation (mm) Mean an'lual precipitation (mm)
B 0
Fig.6.19 Variations in the clay mineral and residual hydroxide composition of soils in relation to mean annual precipitation: surface
soil samples on feldspar and quartz rich igneous rocks (A) and an olivine, amphibole and pyroxene-rich igneous rocks (B) (bothfrom
California); soils developed on basalt under an alternating wet and dry climate (C) and a continuously humid climate (D) (both from
Hawaii). (Constructed from data and figures in I. Barshad (1966) International Clay Conference Proceedings 1, Figs 1 and 2, p. 167
and G. D. Sherman (1962) Problems of Clay and Laterite Genesis. American Institute of Mining and Metallurgical Engineering, New
York, Fig. 2, p. 158 and Fig. 3, p.159.)

ment and the adjustment of the mineralogical and physical 6.3 Physical weathering
properties of the regolith to these conditions. Consequently,
Physical weathering encompasses a range of mechanisms,
weathering profiles are rarely in full equilibrium with
the relative effectiveness of which are not accurately
environmental conditions; in most cases the weathering
known but clearly vary significantly as a function of en-
mantle adjusts to long-term average conditions rather than
vironmental conditions. The physical breakdown of rock is
to conditions at a specific time. We always have to be
always associated with some kind of volume change and it
aware of the kinetics of weathering reactions, when looking
is useful to categorize the various processes into those
at the mineralogical composition of weathering profiles. As
involving an overall volumetric change in the rock mass
we shall see in Section 6.5, some weathering products show
and those related to changes in volume of material
a remarkable resilience to changing environmental con-
introduced into voids or fissures in the rock.
ditions to the extent. that in some circumstances they can
persist in the landscape essentially unaltered for tens of
millions of years. 6.3.1 Volumetric changes of the rock mass
As noted in Chapter 5 in the context of granite intrusions
(Section 5.5), rocks formed at depth or located beneath a
 Weathering and associated landforms 145

Table 6.5 Relative abundance of rock types and minerals exposed to

weathering at the Earth's surface

Proportion of area covered by major rock types (%)

Shale 52
Sandstone 15
Granite and granodiorite 15
Limestone and dolomite 7
Basalt 3
Others 8

Total sedimentary -75

Total igneous and metamorphic - 25
Proportion of common minerals in bedrock exposed to weathering at the
Earth's sUlface (%)
Feldspars 30
Quartz 28
Clay minerals and micas 18
Calcite and dolomite 9
Iron oxide minerals 4
Pyroxenes and amphiboles 1
Others 10

Source: Data from L. B. Leopold et al. (1964) Fluvial Processes in

Geomorphology. W. H. Freeman, San Francisco, p. 100.

thick overburden are under considerable internal stress. As

the overlying strata are gradually removed by erosion and
these rocks reach the surface they undergo expansion or
dilation - a process known as pressure release - which
promotes the development of joints. At the smaller scale
micro-fissures and incipient joints related to the original
pattern of mineral crystallization in the rock provide lines
of weakness along which exfoliation (the spalling off of
thin sheets of rock) (Fig. 6.20) and granular dis- Fig. 6.21 An exfoliation dome formed in granite in the Yosemite
integration (the disaggregation of individual crystals or Valley, Sierra Nevada, USA.
particles) can occur. In some kinds of rock, such as granite,
the creation of steep, bare rock faces can lead to significant
lateral expansion into the valley side as well as vertical Insolation weathering, that is, the breakdown of rock as
dilation and this may give rise to exfoliation domes (Fig. a result of volume changes arising from thermal expansion
6.21). and contraction, is another potentially important physical
weathering process. Fluctuations in temperature experi-
enced by rocks at the Earth's surface due to day-time
heating and night-time cooling certainly cause them to
expand and contract. Nevertheless it has proved difficult to
establish whether the resulting volume changes are suf-
ficient to shatter rock, even in desert environments where
diurnal temperature ranges on rock surfaces can easily
o
exceed 30 C. Since rock is a poor conductor of heat a
thermal gradient is created when the surface is warmed.
The exterior of the rock consequently expands more than
the interior and stresses are set up. Data on bedrock sur-
faces in desert regions show that on dark rocks
o
temperatures of 80 C can be attained by solar radiation and
. 0
dIUrnal ranges of 50 C are not uncommon. Further
(intergranular) stress might be expected to arise from the
Fig.6.20 Exfoliation in granite-gneiss in north-western Cape differential thermal expansion of individual minerals in a
Province, South Africa. The hat towards the right-hand edge of rock related to differences in colour, specific heat and
the photograph indicates the scale. coefficient of thermal expansion.
146 Exogenic processes and landforms

In spite of the abundance of' shattered rock in hot desert 6.3.2 Volumetric changes within rock voids and
regions and elsewhere, experimental work in the 1930s cast fissures
considerable doubt on the efficacy of insolation weathering. Under this heading we need to consider two main sets of
Small rock samples subjected to repeated cycles of heating
processes: the effects of crystal growth and expansion,
and cooling over temperature ranges well in excess of those primarily of ice and salt, and the stresses induced by
produced naturally by solar radiation failed to shatter. The
biological activity. The latter are induced by both fauna and
presence of moisture was found to be necessary to promote
flora. Earthworms, for instance, ingest and excrete huge
chemical weathering and weaken the rock sufficiently for volumes of soil, as so graphically reported in a classic
disintegration to occur. Subsequently these conclusions
study by Charles Darwin in the last century. The resulting
have been challenged, it being pointed out that experiments mixing of the soil, or bioturbation, probably averages
involving the rapid heating and cooling of small individual about a 5 mm depth annually over the more humid parts of
rocks do not accurately re-create field conditions. In
the continents. Other soil fauna contribute to this effect,
particular, stresses set up in such samples can be relieved in
with termites being particularly significant in moving large
all directions whereas stresses in bedrock are much more
quantities of soil in the tropics and subtropics.
concentrated since the rock is confined both laterally and The growth of tree roots and their penetration and
vertically.
enlargement of incipient rock fractures can contribute
The occurrence of scrub and forest fires provides another significantly to the break-up of rock which may have
mechanism whereby rocks can be subjected to significant experienced relatively little weakening through chemical
thermal expansion and contraction. It is certainly important
alteration. Large tap roots can penetrate rock to a depth of
in semi-arid regions characterized by vegetation communi- several metres and the process commonly works in associ-
ties, such as chaparral (North America) and eucalyptus
ation with other mechanisms, particularly ice crystal
(Australia), which require fires for regeneration. Such areas
growth. It is also likely that chemical weathering is
are frequently seen to be characterized by spalling rock enhanced in the vicinity of roots where carbon dioxide
outcrops and shattered boulders.
concentrations are likely to be higher and where the
Chemical weathering frequently gives rise to minerals
extraction of metal cations is occurring. Although such
which are less dense than their precursors. The associated
organic activity can be significant, especially locally, it is the
volume increase can exert pressure on the surrounding rock
physical processes of crystal formation and expansion that are
and the weathering rinds produced peel off to expose fresh
of more fundamental importance in rock weathering.
rock to further weathering. The chemical processes involved
are usually associated with hydration (see Section 6.2.3.1) 6.3.2.1 Frost weathering
and the addition of water plays a significant role in the In arctic and alpine environments the surface is often seen
physical breakdown of rock. Such hydration weathering to be composed of a layer of angular rock fragments com-
can occur simply through wetting and drying which causes monly described by the term felsenmeer ,and attributed to
expansion and contraction. The process is most active in the operation of frost weathering. This process, which is
some clay minerals, notably those of the smectite group, also referred to as frost shattering and frost wedging,
which are capable of absorbing large quantities of water involves the breakdown of rock or other solid materials as a
into their crystal structures. result of stresses induced by the freezing of water.
A further hydration weathering mechanism, restricted to Early explanations of frost weathering alighted on the
environments in which freezing occurs, is hydration simple and seemingly obvious effect of the 9 per cent
shattering. This can occur in very fine-grained materials volume expansion which accompanies the phase change
such as clays which are capable of retaining significant from water to ice. Under ideal conditions it was estimated
quantities of unfrozen water at temperatures well below that ice formation could exert a maximum pressure of
o
o C. At these low temperatures the dipolar molecules around 200 MPa (at which point the freezing temperature is
of this supercooled water become ordered in such a way -22 0 C), but it was appreciated that this theoretical
that a repulsive force is set up across voids. It is thought maximum pressure could never be attained under natural
that while this force is insignificant in large voids it can be conditions, not the least because it far exceeded the tensile
sufficiently strong in the small pores in very fine-grained strength of most rocks (around 25 MPa). Even to build up
material to cause rock fracture. Chemical weathering is also moderately high pressures sufficient to shatter rock it was
probably associated with the mechanism which so far has realized that a closed system was required in order for the
only been investigated on clay':rich rocks, and it may be pressure not to be relieved by the expulsion of water into
restricted to such lithologies. adjacent voids. Such a closed system might be produced if
freezing proceeds rapidly from the surface downwards
since initial ice formation at the surface can then seal in
water contained within pores in the rock.
 Weathering and associated landforms 147

Subsequent experimental work has failed both to clarify contammg cations released during bedrock weathering
fully the exact mechanisms involved in frost weathering which have subsequently been concentrated by evapor-
and to define precisely the climatic conditions under which ation, precipitation with a high salt content (especially in
the process is likely be most effectiv~. It has been regions near coasts), saline ground waters and saline
suggested that a freezing rate of at least 0.1 °C per minute and deposits blown inland from the coast or originating in
temperatures of -5 to -10 °C are reqUIred . for rock shatter- inland saline basins. A number of salts are found in the
ing to occur. Most experimental studies have emphasized soils of arid regions but only a. few are generally
the importance of rapid freezing, a minimum temperature quantitatively significant - CaS04(.2H20), NaCI, Na2S04,
of --5 °C, the frequency of freeze-thaw cycles and the Na2C03, K2S04, MgS04 and KCl.
overriding significance of the moisture content of the rock. When saline solutions in rock pores become saturated as
This last factor is crucial since if a rock is not saturated ice . a result of a temperature. change or evaporation, salt
formation will simply cause the expulsion of water into air- crystals begin to form and considerable pressures are
filled cavities, thereby inhibiting a build-up of pressure. generated. Whether these lead to rock fracture will depend
Some theoretical work, however, has questioned the. on both the stresses produced by crystal growth and the
importance of water freezing in sealed cracks and the tensile strength of the rock. Salt crystal growth in unconso-
frequency of freezing events, suggesting instead that crack lidated sediments can cause considerable ground heave, and
growth will be greatest when the temperature ranges from where salt solutions enter desiccation cracks various types
approximately -4 to -15 °C. Other experimental studies of patterned ground may be formed (Fig. 6.22). A second
have supported the idea that frost weathering is likely to be cause of stress arises from the capacity of several common
more effective in alpine and marine arctic environments, salts to absorb significant quantities of water into their
where freeze-thaw may occur diurnally, than in polar crystal structure. Such hydration expansion commonly
regions where freezing and thawing are predominantly occurs in response to changes in relative humidity which,
seasonal phenomena. since they are closely related to temperature, may be
Uncertainties about the efficacy of volume expansion on diurnal. The importance of thermal expansion, the third
freezing in rock shattering have encouraged the examination mechanism of salt weathering, has yet to be fully evaluated.
of other possibilities. One of these is the pressure exerted Its possible significance derives from the observation that
by ice-crystal growth, the idea being that this is directional the coefficients ,of thermal expansion of many salts are
and that if the direction of growth is resisted by the wall of greater than that of most rocks. Sodium chloride (common
a void or crack then pressure will be exerted which may be salt), for instance, can expand by up to 1 per cent under the
sufficient to wedge the crack apart. As applied to the very large diurnal temperature ranges typical of continental
crystallization of ice this mechanism is largely conjectural deserts.
at present since little experimental work has been under- The rather specific environmental requirements of salt
taken to estimate the pressures likely to be generated. A weathering mean that its action is spatially highly variable.
further possibly important mechanism in rock breakdown It is likely to be intense in west coast deserts, such as the
associated with freezing, at least in very fine-grained Atacama (Chile/peru) and Namib (Namibia), where fre-
lithologies, is hydration shattering (see Section 6:3.1) and it quent fogs supply both salt and moisture. It is also active in
is likely that some of the effects of physical weathering areas adjacent to nearly enclosed seas in hot deserts, for
previously attributed to frost weathering in fact result from example the Red Sea, where both humidity and salinity are
this process. high. Other favourable environments include the margins of
salt-covered basins, along river channels and in other areas
6.3.2.2 Salt weathering of low-lying topography where shallow saline ground
Salt weathering involves three major processes: the waters can be drawn to the surface by capillary action.
precipitation of salt in voids and the expansion of salt Different salts also vary significantly in their efficacy as
crystals through either hydration or heating. It is most weathering agents, and several laboratory investigations
active in arid environments where rates of evaporation are have indicated the particular effectiveness of sodium
high relative to precipitation, and consequently surface and sulphate. This can be attributed' to its various physical
soil waters can become saturated with respect to a variety properties. First, it experiences a large increase in volume
of salts. Although frequently associated with hot desert when hydrated (from Na2S04 to Na2S04' lOH20). Secondly,
regions, the effects of salt weathering have also been it has a high equilibrium solubility so that large quantities
observed in high-latitude arid regions such as Antarctica. can be precipitated from a given volume of solution.
Moreover, the high salinity of sea water makes it likely that Thirdly, its equilibrium solubility is particularly sensitive to
salt is an active weathering agent in coastal environments temperature, reaching a maximum at 32.3 °C. This is within
(see Section 13.3.1.1). In inland areas salt is derived from a the range of diurnal temperature change in many hot
number of sources. These include surface and soil waters deserts and precipitation will occur as the temperature rises
148 Exogenic processes and landforms

Fig. 6.22 Polygonal patterned ground developed in the saline bed of the southern margin of Lake Eyre, South Australia.

or falls across this value. Finally, crystallization occurs discontinuous or thin soil cover, abundant enclosed
preferentially along a single crystal axis and large pressures depressions and a well-developed system of underground
can be generated within rock fractures. drainage including caves, all features attributable to the
solubility of limestone. Limestone is defined as a rock
which contains at least 50 per cent CaC0 3 , which nearly
6.4 Lithology and weathering forms always occurs as the mineral calcite. Other carbonate
Just as variations in the physical and mineralogical pro- minerals include aragonite (a rare form of CaC0 3 ) and
perties of bedrock can influence the mineralogical products dolomite (CaMg(C0 3)z). Rocks in which more than 50 per
of weathering so different lithologies give rise to a range of cent of the carbonate content occurs as the mineral
weathering forms. On most lithologies landforms related dolomite are given the rock name dolomite. The term
directly to weathering tend to be minor features, but on karstification is used to refer to the process of karst
rock types such as limestones, in which a large proportion landscape development. Karst landforms are classically
of the products of chemical weathering processes are re- exemplified by the landscape of northern Yugoslavia and a
moved in solution, major landforms can be produced. In number of terms used to describe karst landforms originate
this section we will focus on weathering forms in lime- from this region.
stone, although it should be pointed out that similar The term pseudokarst is applied to landforms in non-
weathering forms can develop on other lithologies. carbonate rocks which are morphologically similar to those
characteristic of limestone terrains. In most cases pseudo-
karst develops as a result of processes analogous to those
6.4.1 Karst weathering forms operating on true karst; such is the case on some pure
Karst is the German form of a Slovene word meaning siliceous rocks subject to prolonged weathering where the
'bare stony ground' and is used to describe limestone slow dissolution of quartz creates karst-like forms. In other
terrain characterized by a lack of surface drainage, a cases quite different processes can give rise to topography
 Weathering and associated landforms 149

superficially resembling true karst, an example being the and part of the area is more than 2000 m above sea level.
creation of 'underground drainage' through the formation Although chalk is a very pure limestone it is mechanically
of lava tunnels in volcanic terrains. too weak to allow the development of large underground
Karst scenery is not entirely, or in many cases even cavities. Similarly, thin beds of limestone do not permit the
primarily, a product of weathering since fluvial processes development of underground drainage while in arid regions
playa crucial role in its development. Nevertheless, it is there is insufficient water for significant limestone
appropriate to consider here those components of karst weathering to occur. There is an extensive vocabulary of
topography that owe much of their origin to weathering special terms to describe karst landforms which can be
processes. Karst landforms related more specifically to confusing to the uninitiated, and even to the expert.
fluvial activity are discussed in Chapter 9.
The fullest development of karst scenery is attained 6.4.1.1 Minor forms
where the limestone is relatively pure (in excess of 80 per The German term karren and the French term lapies refer
cent CaC0 3), very thick, mechanically strong and mass- to the small-scale solutional form developed on limestone.
ively jointed (so that water flow is concentrated along (Note that the terms 'solution' and 'dissolution' are used
joints rather than passing through the mass of the rock). In widely in discussions of karst landforms, but this is just a
addition a humid climate is required together with sufficient form of shorthand since, as was discussed in Section
relief for extensive vertical water flow down to a regional 6.2.3.3, limestone weathering involves complex chemical
water table deep below the surface. In the karst region of reactions.) There is no English equivalent for these terms
Yugoslavia there are pure limestones up to 4000 m thick although some specific forms have acquired local names. A

Table 6.6 Classification of solutional microfonns developed on limestone

FORM TYPICAL DIMENSIONS COMMENTS

Rainpit <30 mm across, Produced by rain falling on bare rock. Occurs in fields on gentle rather than steep
<20 mm deep slopes. Can coalesce to give irregular, carious appearance
.~ Solution ripples 20-30 mm high; Wave-like form transverse to downward water movement under gravity. Rhythmic
~ may· extend horizontally form implies that periodic flows or chemical reactions are important in their
~ ] for >100 mm development
~ ~ Solution flutes 20-40 mm across, Develop due to channelled flow down steep slopes. Cross-sectional form ranges from
j ~ (rillenkarren) 10-20 mm deep semi-circular to V-shaped but is constant along flute

j
~
I~
Solution bevels 0.2-1 m long,
30-50 mm high
Flat, smooth element u ually found below flutes. Flow over them occurs as a thin
sheet
"'" Q Solution runnels 400-500 mm across, Down runnel increase in water flow leads to increase in cros - ectional area. May
j (rinnenkarren) 300-400 mm deep, have meandering form. Ribs between runnels may be covered with oludon flutes
~
~ 10-20 m long
,~ ----------------------------------------------------
_ ~ Grikes (kluftkarren) 500 mm across, up to Formed through the solutional widening of joints or, if bedding is nearly vertical, of
~~ several metres deep bedding planes
~~
~ i5 Clints (flackkarren) Up to several metres Tabular blocks detached through the concentration of solution along near- urface
~.~ across bedding planes in horizontally bedded limestone
~ i
-
Solution spikes
(spitzkarren)
Up to several metres Sharply pointed projections between grikes

Solution pans 10-500 mm deep, Dish-shaped depressions usually floored by a thin layer of soil, vegetation or algal
0.03-3 m wide· remains. CO 2 contributed to water from organic decay enhances dissolution.

Undercut solution 400-500 mm across, Like runnel but become larger with depth. Rece sion at depth probably associated
runnels (hohlkarren) 300-400 mm deep with accumulation of humus or soil which keeps sides at ba e constantly wet
10-20 m long
Solution notches I m high and wide, Produced by active solution where soil abuts against projecting rock giving rise to
(korrosionkehlen) 10 m long curved incuts

Rounded solution 400-500 mm across, Runnels developed beneath a soil cover which become smoothed by the more active
runnels (rundkarren) 300-400 mm deep corrosion associated with acid soil waters.
10-20 m long
Solution pipes 1 m cross, Usually become narrower with depth. Found on soft limestones such as chalk as well
2-5 m deep as mechanically stronger and less penneable varieties

Note: The commonly encountered Gennan terms are given in parentheses

Source: Based largely on discussion in 1. N. Jennings (1985), Karst Geomorphology. Blackwell, Oxford, pp. 73-82.
150 Exogenic processes and landforms

bewildering variety of karren have been described and the

most important of these are classified and described in
Table 6.6
Several factors influence the form of karren. Probably of
most importance is the presence or absence of a cover of
soil or vegetation during its formation. This is probably
because the solution process operates differently on
covered and bare surfaces. On covered surfaces soil and
vegetation maintains a continuous presence of moisture on
the rock surface while organic decay increases the quantity
of organic acids and the abundance of carbon dioxide; both
increase the aggressiveness of the water percolating on to
the rock surface. (Aggressivity refers to the propensity of
water to dissolve calcium carbonate.) The resulting sol-
utional forms tend to be rounded and less sharp-edged than
those formed on bare rock.
A second factor is lithology. Karren are best developed Fig.6.23 Major types of doline: (A) collapse doline; (B)
on relatively massive and uniform, mechanically strong and solution doline; (C) subsidence doline; (D) subjacent karst
impermeable limestones in which there is a sharp contact collapse doline; (E) alluvial stream sink doline. (From J. N.
between soil and rock. Chalk, for instance, is too porous Jennings (1985) Karst Geomorphology. Blackwell, Oxford,
and mechanically weak for karren development. Climate is Fig. 37, p. 107).
also important since adequate moisture must be present for
effective solution. In arid climates there is generally in-
sufficient surface water, whereas in periglacial regimes development of further depressions near by.
frost action may be so effective that solutional forms have Various types of doline are recognized in terms of their
little time to develop. Snow patches, with their enhanced mode of formation (Fig. 6.23). A collapse doline is created
concentrations of CO 2, which experience seasonal melting through the fall of the roof of a cave to leave a steep-walled
may, nevertheless, be favourable environments for karren cavity. These steep walls are, however, likely to be subject
dev~lopment. Time is the final significant factor since to rapid degradation through dissolution and physical
karren developed under a relatively humid climate may weathering processes in all but arid climates, and their
survive for a period after a change to a more arid climatic gradients may be consequently reduced in a short period of
regime. Moreover, forms developed under a soil cover may time to the extent that a colIpase doline may be indis-
subsequently be exposed by erosion, in some cases pro- tinguishable in form from other types. Collapse into a
moted by human activities such as forest clearance. water-filled cave or a subsequent rise in the local water
The role of organisms in directly creating both erosional table can lead to the formation of a lake, such as in the case
and depositional microforms on limestone has probably of the cenotes which are common in the Yucatan Peninsula,
been underestimated. Micro-organisms may be instru- southern Mexico.
mental in many commonly observed microforms in lime- Solution dolines are formed when localized dissolution
stones, b.ut the precise nature of their role is as yet poorly usually concentrated at the intersection of major joints
known. The terms biokarst and phytokarst are used to lowers the bedrock surface. The initial development of the
describe forms attributable in large part to organic processes. depression leads to the capture of more surface drainage
and this in turn promotes further doline enlargement.
6.4.1.2 Major forms Solution dolines are one of the few karst landforms which
If one landform can be regarded as typifying karst scenery are capable of forming in relatively soft limestones such as
it is the doline. A doline is a closed depression which may chalk. Subsidence dolines form in a similar manner, but in
range in shape from bowl-shaped to cylindrical and in size this case there is a covering' of superficial deposits which
up to 100 m deep and 1 km across. In some limestone collapse either suddenly or progressively into the cavity
terrains the surface is pitted with large numbers of dolines developing in the underlying limestone. More dramatic
which each provide separate foci for surface drainage, in collapse characterizes subjacent karst collapse dolines in
contrast to the integrated valley systems of fluvially eroded which a covering non-calcareous rock unit collapses into a
landscapes. In fact it is likely that doline formation may be void in the underlying limestone. The final generally
a kind of contagious process whereby the establishment of recognized type is the alluvial stream sink doline which
an initial depression will tend, through its effect on lower- consists of a stream flowing into a doline before dis-
ing the water table in its immediate vicinity, to promote the appearing underground.
 Weathering and associated landforms 151

Dolines, especially of the solution type, may be so formation of poljes regarded them as a late stage in a
numerous that they merge into each other to form uvalas. developmental sequence of depression enlargement begin-
Where dissolution is significantly controlled by joint ning with dolines, but current ideas now focus on structural
systems uvala may be elongated forms, but in other cases controls. Many poljes appear to be aligned with structural
they are irregular depressions. In the humid tropics dolines trends represented by fold axes, faults or junctions between
characteristically have an irregular to star-shaped plan form limestone and non-calcareous rocks; some may indeed be
due to their mutual interference as they have grown and graben but in the majority of cases it appears that structure
completely consumed the original surface. This produces a controls the location of lateral planation necessary for polje
polygonal pattern of ridges surrounding individual dolines development. Most poljes are located at the junction of
which are sometimes described as cockpits following the limestone and impermeable beds, and runoff from the latter
local name given to them in Jamaica. Residual hills are will be undersaturated with calcium carbonate and will
often present on the divides separating depressions. They therefore be aggressive when entering the limestone terrain.
are usually rounded in form and are termed cones or kegel, In humid tropical and subtropical environments some
and the whole landscaRe assemblage is called cockpit limestone landscapes are dominated by spectacular tower-
karst (or alternatively cone karst or kegelkarst). Many like hills, or mogotes, up to 100 m or more high separated
cockpits contain stream sinks and the morphology of by broad, alluvial valley floors (Fig. 6.25). This type of
cockpits is probably attributable to the abundance of sur- terrain is described as tower karst and has been thought to
face runoff in humid tropical environments which leads to represent the effects of enhanced limestone dissolution in
the creation of minor valley-like solutional forms on their regions of high precipitation. Acid waters in the alluvial
slopes. In some localities in the humid tropics the residual valleys cause rapid lateral planation and undercutting of
hills on doline divides have a remarkable sharp-edged, rock faces while high temperatures promote the evapor-
pinnacle form and this has given rise to the term pinnacle ation of water flowing across the exposed rock leading to
karst (Fig. 6.24). the deposition of a protective layer of calcite. This climatic
Large flat-floored depressions ranging in size up to interpretation of tower karst, however, has been revised in
200 km 2 or more are found in some karst terrains and are the light of the recognition of tower karst in the far from
termed poljes. Polje floors are commonly covered by tropical environment of the Mackenzie Mountains in north-
alluvium and one or more sides of the depression has a west Canada. In this region the limestone is massive and
o
steep angle in excess of 30 . Water flows on to the polje very thick with widely spaced joints and the tower kar t
floor via springs along its edge or from impermeable strata appears to represent the last tage of karst landscape
which commonly form one side and leaves the polje either development. This apparently began with the formation of
through stream sinks called ponors or via gorges penetrat- deep, joint-controlled dolines and progressed through the
ing one of the polje walls. Early attempts to explain the creation of long, narrow gorges called karst streets and
then the development of a rectilinear network of deep gorges
forming labyrinth karst to a final phase of lateral planation
of the gorge rock walls to form towers.

6.4.2 Other weathering forms

Exposure of bare rock are found in many environments and
result from differential weathering of bedrock and the
removal of the weathered debris by slope processes. Rock
outcrops which stand out on all sides from surrounding
slopes are known as tors. They are particularly common on
crystalline rocks, but also occur on other resistant litho-
logies such as quartzites and some sandstones. Some
researchers have regarded deep weathering as a prerequisite
for the development of tors, with a phase of intense
Fig. 6.24 Pinnacle karst in the Mulu area of Sarawak. The chemical weathering preferentially focused along joints
development of this extraordinary topography is probably being followed by stripping of the weathered material in
attributable to preferential dissolution focused along joints or response to a change in conditions more favourable to
bedding planes in steeply dipping metamorphosed limestone. The erosion. Other workers, however, have argued that tors can
apparently rapid rate of limestone dissolution is related to high
rainfall totals (in excess of5000 mm a-I ),frequent mists and high develop in the absence of deep weathering where weather-
rates of organic activity. (Interpretation by M. 1. Day, and photo ing and stripping operate simultaneously on rocks of
courtesy of the Royal Geographical Society and M. 1. Day.) variable resistance.
152 Exogenic processes and landforms

Fig.6.25 Tower karst near Guilin, Guangxi Province, China. (Photo courtesy D. Munro.)

Nearly all exposed rock outcrops have irregular surfaces

which appear to be related to the effects of weathering.
These include rills, pits and cavernous forms. They are
found on essentially all rock types and in all climates, but
in view of the extent of exposed bedrock they are most
apparent in arid and semi-arid environments. The origin of
the wide range of forms that occurs is poorly understood;
suggested mechanisms include salt weathering, hydration
shattering, insolation weathering and frost action.
Two forms which have attracted particular attention are
honeycomb weathering and tafoni. Honeycomb weathering
consists of numerous small pits a few millimetres or
centimetres in width and depth and the honeycomb form is
produced when they coalesce to create a network structure.
Fig. 6.26 Tafoni developed on the side ofAyers Rock, Northern
Tafoni are larger features ranging up to several metres in Territory, Australia.
size and are cut into steep, bare rock faces (Fig. 6.26). In
some cases selective chemical disintegration of the rock is
apparent but in others evidence of chemical decomposition ant. Case-hardening appears to result from the localized
seems to be quite absent and mechanical processes such as mobilization and reprecipitation of minerals on the rock
salt weathering appear to be more probable processes. surface, thereby strengthening it and rendering it generally
Organic activity may also be instrumental through the less permeable. Areas of the rock surface less effectively
localized enhancement of acidity. case-hardened would then be susceptible to weathering
Case-hardening and core softening may also be import- processes which could then selectively penetrate into the
 Weathering and associated landforms 153

rock interior producing cavIties. Similarly, there is

evidence that the cores of boulders may be weathered more
rapidly than the surface and again this could lead to the
selective development of cavities.
Differential heave in unconsolidated deposits arising
from changes in volume can create a range of patterned
ground phenomena (Fig. 6.22). Such changes in volume
are most commonly associated with wetting and drying and
freeze-thaw cycles and patterned ground is most evident in
arid and periglacial regions (see Chapter 12).

6.5 Duricrusts
Duricrusts are hard layers formed in the weathering zone
Fig. 6.27 A calcrete profile more than 10m thick exposed in the
at, or near, the landsurface as a consequence of the absolute Molopo gorge in the southern Kalahari Basin on the horder
or relative accumulation of particular components hetween Botswana and South Africa. The calcrete has formed
through the replacement or cementation of pre-existing through the replacement of quartzites and tillites.
rock, soil, weathering materials or other unconsolidated
deposits. The most important components in duricrust
formation are iron and aluminium oxides and hydroxides, They cover a significant proportion of the world's semi-arid
silica, calcium carbonate and gypsum. They are of interest environments and they possibly underlie up to 13 per cent of
not only in terms of their origin but also because of their the global landsurface area. Gypsum crusts, or gypcretes,
role in landscape development and the evidence they by contrast have a much more limited distribution and
provide of past climates. Duricrusts may reach thicknesses appear to be largely confined to very arid regions where the
in excess of 50 m although I-10m is a more usual range. mean annual precipitation is below 250 mm. The gypsum
The ferruginous, calcareous and especially siliceous forms (CaS04 . 2H 20) content is highly variable but may excep-
can be exceedingly hard, and particularly where they cap tionally reach 95 per cent. Gypcrete reaches a maximum
less resistant materials they form prominent and highly thickness of about 5 m, somewhat less than other duricrust
resistant elements of the landscape. Duricrusts may form types.
several metres below the landsurface but they can be
subsequently exposed by erosion of less resistant overlying
6.5.1 Models of duricrust formation
materials.
Iron- and aluminium-rich duricrusts are known as, respect- The formation of the different types of duricrust can be
ively, ferricret~ and alcrete. Both occur in association with largely understood in terms of the relative mobility of dif-
deep weathering profiles in the humid to subhumid tropics ferent elements in the weathering environment. Ferricretes
although alcretes are more prevalent where rainfall totals and alcretes in most cases represent the relative accumu-
are particularly high. The term laterite has long been used lation of iron and aluminium oxides and hydroxides as
to describe iron- and aluminium-rich weathering deposits more mobile components are removed from the weathering
while the term bauxite refers to deposits containing mantle (Fig. 6.28). Such relative accumulation occurs under
economically extractable concentrations of aluminium. conditions of intense leaching under climates where rainfall
Many laterites and bauxites are, however, relatively weak is abundant for a significant part of the year. Locally absolute
materials and the terms ferricrete and alcrete are reserved accumulation may occur as iron and aluminium are mobilized
for the indurated forms. to a limited extent within the weathering mantle or are
Siliceous duricrusts, or silcretes, are commonly com- transported either mechanically or in solution from high to
posed of more than 95 per cent Si02 and are found in both low topographic positions.
humid and arid tropical environments. They are particularly Silcrete, calcrete and gypcrete form through absolute
prominent in central Australia and parts of southern and accumulation of silica, calcium carbonate and gypsum
northern Africa. In some cases they occur in weathering respectively. Their formation requires a source for these
profiles in close association with ferricretes, while in more constituents, a means of transferring them to the site of
arid regions they are found in conjunction with calcium formation and a mechanism for precipitation. Possible
carbonate crusts or calcretes. Calcretes have an average sources include the weathering of bedrock or sediments,
CaC0 3 content of around 80 per cent and their distribution inputs from rainfall or dust, plant residues and solutes in
in general broadly coincides with areas with a current mean ground water. Translocation of solid or dissolved source
annual precipitation of between 200 and 600 mm (Fig. 6.27). materials can be either lateral or vertical, and in the latter
154 Exogenic processes and landforms

Loss of resistant materials silcrete can develop within calcrete while under different

t\
Stage 3 Accumulation of very resistant materials conditions calcrete can be seen replacing silcrete. A con-
Stage 2
Loss of less mobile materials trasting situation exists in the highly acidic conditions

iI
Accumulation of resistant materials
......... associated with abundant organic activity in humid tropical
Loss of most mobile materials
Stage 1
Accumulation of less mobile materials
regions. At a pH of below 4 aluminium become more
1\1\1\1\1\/1. mobile than silica so clay minerals can be silicified as they
1\1\1\,,1\/\
/I /I /I /I /I /I
",\/\/,,1\/\
Parent material (p.m) lose aluminium, even though the concentration of silica in
T m e
solution is invariably very low.
There are few firm data on how quickly duricrusts can
1\ t 1\ 1\ /\
/1/1/1/1 /I p.m form although it is clear that gypcrete and calcrete
1\1\1\1\/\
A 1\1\1\/\
/\/\1\1\/\
generally develop much more quickly than the other types.
1\1\1\1\/\
1\1\1\1\1\ The rapid formation of some calcretes is demonstrated by
1\1\."\1\/\
1\1\1\1\/\
1\1\1\1\/\ 3
the cementation of human artefacts, such as gravestones
1\1\1\1\/\
1\1\1\1\/\ and even Coca-Cola tins! The slow development of silcretes,
1\1\1\1\/\
1\1\1\1\/\ 2 ferricretes and alcretes is probably in part explained by the
/1/1/1/1/1
1\1\1\1\/\
1\1\1\1\/\
I\I\{\.'\/\
very low concentrations of silica, iron and aluminium in
/1/1/1/1/1
/1/1/1/1/1 most surface and regolith waters.
1\1\1\1\/\
1\1\1\1\/\

Fig. 6.28 Stages of weathering leading to the relative 6.5.2 Environmental controls
accumulation of iron or aluminium to form ferricrete or alcrete.
The profile in which the iron and aluminium accumulates As with other weathering products the factors important in
experiences a continuous development whereby stage 1 the development of the different types of duricrust are scale
weathering effectively occurs at progressively lower levels as the
dependent. At the broadest scale climate (and vegetation in
weathering front moves downward. Similarly, the landswjace
experiences a progressive lowering and the iron and aluminium so far as it affects pH) is the predominant control; but the
gradually becomes concentrated in the zone of stage 3 overlap in the climatic conditions under which different
weathering. (From M. 1. McFarlane (1983) in A. S. Goudie and K. types of duricrust develop (Fig. 6.30) demonstrates that
Pye (eds) Chemical Sediments and Geomorphology: Precipitates other factors operate at the regional and local scale.
and Residua in the Near-Surface Environment. Academic Press,
Topography is of considerable significance. Low local
London, Fig. 2.3, p. 22.)
relief is an essential requirement for virtually all duricrust
formation since the rate of development must exceed the
case may be either upward as a result of capillary rise (per rate of denudation for a duricrust to be created. Secondly,
ascensum model), or downward through percolation (per topography is the key factor in determining local drainage
descensum) model (Fig. 6.29). Downward movements are conditions. Silica will not be retained in weathering profiles
probably more important in most cases since capillary rise to form silcrete unless the drainage is very poor whereas
will lead to a concentration of precipitation close to the alcrete develops preferentially where efficient drainage
surface and this will reduce the efficiency of surface encourages intense leaching. Bedrock also plays a role
evaporation and thus reduce the rate of subsequent capillary since it influences the availability of source constituents -
rise. silica, iron, aluminium, calcium carbonate, gypsum - but,
Precipitation can arise from a variety of factors in as with the development of other weathering products,
addition to the evaporation of weathering solutions: these lithological variations tend to become progressively masked
include pH changes, reactions with other cations in solution after prolonged duricrust development.
and organic activity. Changes in pH appear to be of Although the climatic parameters for calcrete and
particular importance in the case of silcrete and may gypcrete ,genesis are known fairly well, the same is not true
explain why this type of duricrust is found in both pre- for other types of duricrust. This is because ferricretes,
dominantly alkaline arid and semi-arid environments, and in' alcretes and silcretes are chemically and mechanically
acidic humid tropical weathering profiles (Fig. 6.30). The resistant so that once formed under a certain set of climatic
solubility of silica is more or less constant up to a pH of conditions they can persist in the landscape for long periods
around 9, but above this point its solubility increases .even though the prevailing climate may have changed
dramatically (Fig. 6.5). In alkaline environments, such as dramatically. Consequently, these types of duricrust may
those commonly found around ephemeral lake beds in frequently not be in equilibrium with prevailing environ-
semi-arid and arid regions, pH may vary between 8 and 10 mental conditions; indeed, they are likely to reflect in their
over short distances. Silica in solution moving from a chemical, mineralogical and physical characteristics a
locality where the pH is above 9 to one where it is below 9 history of changes in climate and other variables affecting
will therefore be precipitated. In such circumstances their development rather than a single equilibrium state.
 Weathering and associated landforms 155

A r Transpiration
of H2 0

Leached zone

-:].~'-I~-J-~I--J~:l"~·l·
'I. '. ~·I·-O£r1·· '~l~l·rr.
°00 .-0-''0-' J..-...::.O' - -
~1'~'-TH"d~':-'~I--::"
' ..
,r:,. " ·~to·,···
Zone.ofcapilliaryrise
and lime deposition
'00
-0 - o
"0
+'..
- - -><---.:..... ..:_~_9..· ~
.. o· ", , ,
--Q....-!.~ _~ .. _'_-'_' .:.
.....
tL.:.-J .__
Q..o~_ .. _ CaC0 3 .. rich
groundwater level

Dust and loess,

products of aeolian sand
grain abrasion etc
B
Volcanic dust and ash

Depth of rainfall
penetration

Upward flow of nutrients throU 9

root sy stem
- - - - - - - - -- 1
!
\1')-. - - - - - - - - - - - - Water-table

c
Cambering and undercutting

\
New massive
calcrete
Water-table

Bedrock 1

Fig. 6.29 Models of the formation of calcrete: (A) the capillary rise (per ascensum) model; (B) the percolation (per descensum) model
with inputs of carbonate from above; (C) the detrital model in which a secondary calcrete is formed through the solution and
disintegration of an existing calcrete horizon at a higher level in the landscape. Calcium carbonate as well as silica may also be
translocated in river waters, eventually being precipitated at some distance from the original source. (From A. S. Goudie (1983) in: A.
S. Goudie and Pye, K. (eds) Chemical Sediments and Geomorphology: Precipitates and Residua in the Near-Surface Environment.
Academic Press, London, Fig. 4.4, p. 113 and Figs. 4.5 and 4.6, p. 115.)
156 Exogenic processes and landforms

pH Fig. 6.30 A schematic representation of the relationship

w between the formation of different types of duricrust and
~ prevailing conditions of climate and pH. Note that the chemical
w.
a:: w - - and physical resistance of some kinds of duricrust mean that
u ~
Q.. w many occurrences may now exist outside the climatic and pH
>- a:: zones in which they were formed. In many instances local factors
c..9 u
-J such as topography and drainage outweigh broad climatic
«
u controls.

Calcretes and especially gypcretes are by contrast much

« w
~ more responsive to changing conditions; the relatively high
w
"0 a:: solubility of gypsum, for instance, would lead to the rapid
(,,) u
« -J destruction of a gypcrete should the climate become humid.
U5
w This contrasting 'survivability' in the face of environmental
w ~
~ w changes is largely responsible for the fact that few
w c:::
a:: u
-J gypcretes and calcretes which have been continuously
~ «
a:: exposed at the surface are older than Late Cenozoic while
a::
w some ferricretes and many silcretes date back to the Early
u..
Cenozoic, and even beyond.
Duricrusts are normally significantly harder and more
Rainfall resistant to erosion than the materials which they overlie.
Consequently they tend to armour, or protect, landsurfaces
from denudational processes (Fig. 6.31). Although some

Fig.6.31 Flat-topped hills (koppies) capped by silcrete near Riversdale, in southern Cape Province, South Africa. It is likely that the
silcrete originally formed at a low elevation in the landscape and that subsequently topographic inversion has occurred (see Figure
6.32).
 Weathering and associated landforms 157

A by Whalley and McGreevy (1985, 1987, 1988) are

excellent sources for references, especially on physical
weathering processes. The book by Gerrard (1988) is a
valuable summary of the role of lithology in weathering
and landscape development. The behaviour of water in the
soil and regolith which is so important to an understanding
of both physical and chemical weathering processes is
discussed in detail by Williams (1982).
The best introduction to chemical weathering for the
geomorphologist is probably still the short book by
Loughnan (1969) although it is confined to silicate minerals.
Drever (1985) provides a more detailed treatment of pro-
cesses while Colman and Dether (1986) consider various
aspects of chemical weathering rates. Curtis (1976a) and
Ross (1987) provide brief and accessible introductions to
energy considerations in chemical weathering while the
complex sequence of reactions involved in the weathering
of limestone is discussed by Gunn (1986) and Trudgill
(1985). The pH and Eh conditions in natural environments
are assessed by Baas Becking et al. (1960). The role of
earthworms in bioturbation is examined in the classic study
by Darwin (1881) and Goudie (1988) looks at the important
role played by both earthworms and termites in the tropics.
The probable role of organic processes in the formation of
desert varnish is discussed by Dorn and Oberlander (982)
and Whalley (1983).
Weathering products are covered in the book by OIlier
(1984) and in the volumes edited by Goudie and Pye (1983)
Fig. 6.32 Schematic representation of stages in the development and Wilson (1983). The deep weathering profiles of the
of inverted topography. A resistant duricrust, such as silcrete, is humid tropics are discussed by Thomas (1974) and in the
initially formed preferentially at low points in the landscape (A). classic paper by Ruxton and Berry (1957). Mineral stability
Subsequently, differential erosion lowers the surrounding, less is treated from a thermodynamic perspective by Curtis
resistant, terrain (B) to leave the duricrust capping residual hills
(C). (1976b), while Williams et al. (1986) look in detail at the
chemical weathering of granite in terms of mineral stability
and element mobility. The classic study of granite weather-
types of duricrust seem to preferentially develop at low ing by Goldich (1938) is still worth consulting. The
points in the landscape where ground water flows and formation of secondary minerals is examined in general
surface runoff converge, this armouring can subsequently terms by Loughnan (1969) and with specific reference to
lead to the more rapid lowering of surrounding areas and clays by Millot (1979). Loughnan (1969) also provides an
the creation of inverted topography (Fig. 6.32). Even excellent discussion of the environmental factors control-
where a duricrust horizon has been broken up by prolonged ling chemical weathering, while Pye et al. (1986) show
erosion the duricrust fragments produced may still remain how apparently minor differences in bedrock chemistry can
on the surface and perform a protective role for a long influence weathering rates.
period of time. This is the case in the gibber plains which Turning to physical processes Rice (1976) reassesses the
cover extensive areas of central Australia and which consist importance of insolation weathering while McGreevy (1985)
of silcrete boulders scattered on the surface (Fig. 6.33). looks at the implications of the thermal properties of rock
for physical weathering. Valuable field data on temperatures
attained by rocks in the field are presented by Kerr et al.
Further reading (1984), Peel (1974) and Smith (1977), while Ollier and Ash
There is an extensive liter.ature on weathering though not (1983) look at the role of fire in rock weathering. The
all of it is written from a specifically geomorphic perspec- importance of cracks within rocks for the effective oper-
tive. The books by Birkeland (1984) and OIlier (1984) ation of frost and salt weathering is emphasized by Whalley
provide good introductions to the processes and products of et al. (1982), while McGreevy and Whalley (1985) assess
weathering in the context of landforms, while the reviews the importance of rock moisture to frost weathering both in
158 Exogenic processes and landforms

Fig. 6.33 A gibber plain east ofLake Eyre, South Australia.

the laboratory and under field conditions. Fifteen years of morphometric techniques to karst terrains and Brook and
experimental work on frost weathering is summarized in Ford (1978) and Williams (1987) consider the long-
Lautridou and Ozouf (1982), and Thorn (1979) reassesses standing problem of the development of tower karst. A
the efficacy of freeze-thaw action in alpine environments. variety of cavernous weathering forms, including tafoni and
An important recent trend has been the greater application honeycomb weathering, occur on non-calcareous rocks
of basic physical principles to studies of frost weathering and the origin of these features is considered by Conca and
and this approach is exemplified by Walder-and Hallet Rossman (1985) and Mustoe (1982, 1983). Pseudokarst
(1985). Salt weathering has attracted increasing attention microforms on granite are recorded by Watson and Pye
from geomorphologists as its importance, especially in arid (1985) while Young (1986) reports karst-like tower forms
environments, has become appreciated. Sperling and Cooke developed in sandstone.
(1985) and Smith and McGreevy (1983) report exper- Good summaries of the characteristics and origin of the
imental and simulation studies of salt weathering, while various forms of duricrust are given in the volume edited
Goudie and Day (1980) and Goudie and Watson (1984) by Goudie and Pye (1983). Of particular relevance are the
provide evidence of its potency in the field. chapters by Goudie (1983), McFarlane (1983), Summerfield
There is a very large literature on weathering forms (1983a) and Watson (1983). Other useful articles on par-
developed on bedrock, especially in the case of limestone. ticular duricrust occurrences are contained in Wilson
Both minor and major karst forms are comprehensively (1983). Additional information can be found in Watson
treated in Jennings (1985), Sweeting (1972), Trudgill (1985) on gypsum crusts in Tunisia and Namibia, in
(1985) and White (1988), while Viles (1988) provides a Summerfield (1983b) and Young (1985) on contrasting
useful review of biokarst. Kemmerly (1986) presents a interpretations of the climatic conditions under which
general model for the development of closed depressions, different types of silcrete form and in Goudie (1985) on the
while Williams (1972) shows the value of applying role of duricrusts in landscape development.
 Weathering and associated landforms 159

References
Baas Becking, L. G. M., Kaplan, I. R. and Moore, D. (1960) du CNRS. Progress in Physical Geography 6, 215-32.
Limits of the natural environment in terms of pH and Loughnan, F. C. (1969) Chemical Weathering of the Silicate
oxidation-reduction potentials. Journal of Geology 68, 243-84. Minerals. Elsevier, New York.
Birkeland, P. W. (1984) Soils and Geomorphology (2nd edn.). McFarlane, M. 1. (1983) Laterites. In: A S. Goudie and K Pye
Oxford University Press, New York and London. (eds) Chemical Sediments and Geomorphology: Precipitates
Brook, G. A and Ford, D. C. (1978) The origin of labyrinth and' and Residua in the Near-Surface Environment. Academic
tower karst and the climatic conditions necessary for their Press, London and New York, 7-58.
development. Nature 275,493-6. McGreevy, J. P. (1985) Thermal properties as controls on rock
Colman, S.M. and Dether, D.P. (eds) (1986) Rates of Chemical surface temperature maxima and possible implications for rock
Weathering ofRocks and Minerals. Academic Press, Orlando. weathering. Earth Surface Processes and Landforms 10,
Conca, J. L. and Rossman, G. R. (1985) Core softening in 125-36.
cavernously weathered to"nalite. Journal of Geology 93, 59-73. McGreevy, 1. P. and Whalley, W. B. (1985) Rock moisture
Curtis, C. D. (1976a) Chemistry of rock weathering: fundamental content and frost weathering under natural and experimental
reactions and contr.ols. In: E. Derbyshire (ed.) Geomorphology conditions: a comparative discussion. Arctic and Alpine
and Climate. Wiley, London and New York, 25-57. Research 17, 337-46.
Curtis, C. D. (1976b) Stability of minerals in surface weathering Millot, G. (1979) Clay. Scientific American 240(4), 76-84.
reactions: a general thermo-chemical approach. Earth SUJjace Mustoe, G. E. (1982) The origin of honeycomb weathering.
Processes 1, 63-70. Geological Society ofAmerica Bulletin, 93, 108-15.
Darwin, C. (1881) The Formation of Vegetable Mould, through Mustoe, G. E. (1983) Cavernous weathering in the Capitol Reef
the Action of Worms, with Obse,:vations on their Habits. desert, Utah. Earth SUJjace Processes and Landforms 8,
Murray, London. 517-26.
Dorn, R. I. and Oberlander, T. M. (1982) Rock varnish. Progress Ollier, C. (1984) Weathering (2nd edn). Longman, London and
in Physical Geography 6, 317-67. New York.
Drever, J. I. (ed.) (1985) The Chemistry of Weathering. Reidel, OIlier, C. D. and Ash, J. E. (1983) Fire and rock breakdown.
Dordrecht. Zeitschriftfur Geomorphologie 27, 363-74.
Gerrard, A.1. (1988) Rocks and Landforms. Unwin Hyman, Peel, R. F. (1974) Insolation weathering: some measurements of
London and Boston. diurnal temperature changes in exposed rocks in the Tibesti
Goldich, S. S. (1938) A study of rock weathering. Journal of region, central Sahara. Zeitschrift fur Geomorphologie
Geology 46, 17-58. Supplementband 21, 19-28.
Goudie, AS. (1983) Calcrete. In: A S. Goudie and K Pye (eds) Pye, K, Goudie, A S. and Watson, A (1986) Petrological in-
Chemical Sediments and Geomorphology: Precipitates and fluence on differential weathering and inselberg development in
Residua in the Near-Surface Environment. Academic Press, the Kora area of central Kenya. Earth Surface Processes and
London and New York, 93-131. Landforms 11, 41-52.
Goudie, A S. (1985) Duricrusts and landforms. In: K S. Richards, Rice, A (1976) Insolation warmed over. Geology 4,61-2.
R. R. Arnett and S. Ellis (eds) Geomorphology and Soils. Allen Ross, S. M. (1987) Energetics of soil processes. In: K J. Gregory
and Unwin, London and Boston, 37-57. (ed.) Energetics of Physical Environment: Energetic Approaches
Goudie, AS. (1988) The geomorphological role of termites and to Physical Geography. Wiley, Chichester and New York,
earthworms in the tropics. In: H. A. Viles (ed.) Biogeomor- 119-43.'
phology. Blackwell, Oxford and New "Xork, 166-92. Ruxton, B.P. and Berry, L. (1957) Weathering of granite and
Goudie, A S. and Day, M.1. (1980) Disintegration of fan sedi- associated erosional features in Hong Kong. Bulletin of the
ments in Death Valley, California, by salt weathering. Physical Geological Society ofAmerica 68, 1263-92.
Geography 1, 126-37. Smith, B. J. (1977) Rock temperature measurements from the
Goudie, A S. and Pye, K (eds) (1983) Chemical Sediments and northwest Sahara and their implications for rock weathering.
Geomorphology: Precipitates and Residua in the Near-Surface Catena 4, 41-64.
Environment. Academic Press, London and New York. Smith, B. 1. and McGreevy, 1. P. (1983) A simulation study of salt
Goudie, A S. and Watson, A (1984) Rock block monitoring of weathering in hot deserts. Geografiska Annaler 65A, 127-33.
rapid salt weathering in southern Tunisia. Earth SUJjace Sperling, C. H. B. and Cooke, R. U. (1985) Laboratory simulation
Processes and Landforms 9, 95-8. of rock weathering by salt crystallization and hydration
Gunn, J. (1986) Solute processes and karst landforms. In: S. T. processes in hot, arid environments. Earth Surface Processes
Trudgill (ed.) Solute Processes. Wiley, Chichester and New and Landforms 10, 541-55.
York, 363-437. Summerfield, M. A (1983a) Silcrete. In: A S. Goudie and K. Pye
Jennings, 1. N. (1985) Karst Geomorphology. Blackwell, Oxford (eds) Chemical Sediments and Geomorphology: Precipitates
and New York. and Residua in the Near-Surface Environment. Academic
Kemmerly, P. R. (1986) Exploring a contagion model for karst- Press, London and New York, 59-91.
terrane evol~tion. Geological Society of America Bulletin, 97, Summerfield, M. A (1983b) Silcrete as a palaeoclimatic indicator:
619-25. ' Evidence from southern Africa. Palaeogeography,
Kerr, A, Smith, B. J., Whalley, W. B. and McGreevy, 1. P. (1984) Palaeoclimatology, Palaeoecology 41,65-79.
Rock temperatures from southeast Morocco and their Sweeting, M. M. (1972) Karst Landforms. Macmillan, London.
significance for experimental rock weathering studies. Geology Thomas, M. F. (1974) 'Tropical Geomorphology: A Study of
12,306-9. Weathering and Landform Development in Warm Climates.
Lautridou, 1. P. and Ozouf, J. C. (1982) Experimental frost Macmillan, London.
shattering: 15 years of research at the Centre de Geomorphologie Thorn, C. E. (1979) Bedrock freeze-thaw weathering regime in a~
160 Exogenic processes and landforms

Alpine environment, Colorado Front Range. Earth Surface Whalley, W. B. and McGreev'y, J. P. (1987) Weathering. Progress
Processes 4, 211-28. in Physical Geography 11, 357-69.
Trudgill, S. (1985) Limestone Geomorphology. Longman, London Whalley, W. B. and McGreevy, J. P. (1988) Weathering. Progress
and New York. in Physical Geography 12, 130-43.
Viles, H. A. (1988) Organisms and karst geomorphology. In: White, W. B. (1988) Geomorphology and Hydrology of Karst
H. A. Viles (ed.) Biogeomorphology. Blackwell, Oxford and Terrains. Oxford University Press, New York and Oxford.
New York, 319-50. Williams, A. G., Ternan" L. and Kent, M. (1986) Some observ-
Walder,1. and Hallet, B. 1985. A theoretical model of the fracture ations on the chemical weathering of the Dartmoor granite.
of rock during freezing. Geological Society ofAmerica Bulletin Earth Surface Processes and Landforms 11, 557-574.
96,336-46. Williams, P.1. (1982) The Surface of the Earth: An Introduction
Watson, A. (1983) Gypsum crusts. In: A. S.Goudie and K. Pye to Geotechnical Science. Longman, London and New York.
(eds) Chemical Sediments and Geomorphology: Precipitates Williams, P. W. (1972) Morphometric analysis of polygonal karst
and Residua in the Near-Surface Environment. Academic in New Guinea. Geological Society of America Bulletin 83,
Press, London and New York, 133-61. 761-96.
Watson, A. (1985) Structure, chemistry and origins of gypsum Williams, P. W. (1987) Geomorphic inheritance and the develop-
crusts in southern Tunisia and the central Namib Desert. ment of tower karst. Earth SUlface Processes and Landforms
Sedimentology 32, 855-75. 12,453-65.
Watson, A. and Pye, K. (1985) Pseudokarstic micro-relief and Wilson, R. C. L. (ed.) (1983) Residual Deposits: Surface Related
other weathering features on the Mowati Granite (Swaziland). . Weathering Processes and Materials. Geological Society
Zeitschrift fur Geomorphologie 29, 285-300. Special Publication 11. Blackwell Scientific Publications,
Whalley, W. B. (1983) Desert varnish. In: A. S. Goudie and K. Oxford and Boston.
Pye, (eds) Chemical Sediments and Geomorphology: Precipitates Young, R. W. (1985) Silcrete distribution in eastern Australia.
and Residua in the Near-Surface Environment. Academic Zeitschriftfur Geomorphologie 29, 21-36.
Press, London and New York, 197-226. Young, R. W. (1986) Tower karst in sandstone: Bungle Bungle
Whalley, W. B., Douglas, G. R. and McGreevy, 1. P. (1982) Crack massif, northwestern Australia. Zeitschrift fur Geomorphologie
propagation and associated weathering in igneous rocks. 30, 189-202.
Zeitschrift fur Geomorphologie 26, 33-54.
Whalley, W. B. and McGreevy, J. P. (1985) Weathering. Progress
in Physical Geography 9, 559-81.
This page intentionally left blank
 7
Slope processes and forms

7.1 Properties and behaviour of slope materials consolidated deposit which forms an essentially continuous
mass lacking significant joints or fissures, but which is
Slopes constitute the basic element of the landsurface and further weakened when saturated with water. This defi-
so it is not surprising that they have long formed a focal nition of soil includes unconsolidated weathering and
point for landform studies. The main emphasis in this sedimentary deposits as well as materials we would more
chapter is on the characteristics and behaviour of valley- normally regard as soils. Many slopes are, of course,
side slopes or hillslopes which extend from drainage composed of a mixture of rock and soil, but the distinction
divides to stream channels, although much of the dis- between the two is still important when analyzing the
cussion is also relevant to other kinds of slopes such as behaviour of slope materials.
those along coasts. Until the 1950s much of the research Unconsolidated material transported across, and deposited
carried out on slopes was focused on the question of how on, slopes is termed talus when composed of relatively
the history of a landscape might be reconstructed from the large rock fragments, and colluvium when composed
slope forms of which it is composed. Such studies rarely predominantly of finer material. The term talluvium is
considered in detail the specific processes operating on sometimes used for material that is a mixture of fine and
slopes and it has been only in the past two decade~ or so coarse material. Colluvium may contain fossil soil layers
that major advances have been made in this regard. These representing periods of relative slope inactivity character-
advances have been due in large part to the application of ized by low rates of erosion or deposition. Boulders
research carried out by engineers into the problems of slope dislodged from cliffs or free faces accumulate at the cliff foot
stability to slopes in the natural landscape. to form a talus slope, (or scree slope). Where the rock
At any point on the landsurface the form of the landscape fragments are funnelled down a notch or gully in an exposed
is dependent upon the nature, frequency and intensity of the rock face the material accumulates as a debris cone.
geomorphic processes acting upon it and the strength, or
resistance to deformation, of the surface materials of which
it is composed. The properties of these surface materials are 7.1.1 Factors determining the strength of slope
clearly important in understanding the form and mode of materials
development of the slopes which make up the landscape. The response of slope materials to stress is determined by
Two fundamental types of material can be distinguished - their strength which we can define as the ability to resist
rock and soil. Since our concern here is with mechanical deformation and fracture without significant failure. In the
properties which are relevant to the behaviour of slope context of slope materials experiencing stresses generated
materials, it is necessary to define rock and soil in this by gravity, it is strength in relation to shear stress that is
specific context. most important. The shear strength of slope materials as
Rock is a hard, coherent material comprising individual they occur in the field is usually rather variable over time
particles or crystals. It is discontinuous in the sense that it is and space because rocks and soils are generally complex
broken to a greater or lesser extent by joints and fractures mixtures of mineral particles, water and air; nevertheless,
but it is not significantly weakened when saturated with the factors determining shear strength are well known.
water. Soil, (or regolith), by contrast, is a weak, un- One controlling factor, which is particularly important
164 Exogenic processes and landforms

v stress. This accounts for the common observation that slope

failures often occur after heavy rain when pore-water
pressures in slope materials are high and effective normal
J
stresses are low. Where slope materials are completely dry
and voids are entirely occupied by air the pore-water
pressure is zero' (atmospheric pressure) and effective
a normal stress and total normal stress are identical.
The second factor affecting shear strength is cohesion
which can be defined as the inherent strength of a material
V is the vertical stress
which is present irrespective of any load imposed directly
" 'is the shear stress
on it normal to the surface along which movement tends to
a is the normal stress
occur. Cohesion includes the chemical bonding of rock and
{3 is the angle of the shear plane soil particles and the adhesion of clay-sized material as a
Fig. 7.1 Change in shear stress and normal stress with change result of electromagnetic and electrostatic forces. Most
in angle of the shear plane (After M. 1. Crozier (1986) Landslides: rocks are highly cohesive but the chemical bonding of soils
Causes, Consequences and Environment. Croom Helm, London, can also occur through the presence of cements commonly
Fig. 3.2, p. 41.) composed of silica, carbonates or iron oxides. In partially
moist soil, water is drawn over particle surfaces by capil-
for soils, is the frictional resistance between the constituent lary forces and pore-water pressure is negative. In these
particles of a material. This is related to the size of par- circumstances thin water films on particles contribute to
ticles, their shape and arrangement, their resistance to adhesion by creating capillary cohesion. This effect is
crushing and the number of contact points per unit volume. most apparent in soils with a high silt or clay content which
These inherent frictional properties collectively determine are able to retain some moisture permanently in very small
the angle of internal friction of the material. The greatest pores. It is also illustrated by the ability of sand to stand at
angles are achieved by materials composed of angular a much higher angle when it is damp than when it is
particles of a range of sizes which are capable of close .completely dry and cohesionless.
packing. We can now take into account the role of pore-water
The effects of gravity on a slope are expressed by the pressure in either moderating normal stress or contributing
weight per unit area of the slope materials. This load to cohesion by defining shear strength as effective normal
contributes to both the frictional resistance of the material stress plus cohesion, a relation often referred to as the
and to its propensity to move downslope and it can be . Coulomb-Terzaghi shear strength equation (Box 7.1).
resolved into two components (Fig. 7.1). One acts parallel The relative contribution of normal stress and cohesion to
to the shear plane, or surface of rupture along which shear strength will vary greatly depending on the state of
movement occurs, and constitutes the shear stress. The the material being considered (Fig. 7.2).
other is termed the normal stress because it acts at right The shear strength of soils can be measured by using
angles to the shear plane. It contributes to frictional either a shear box, in which the maximum force that can be
resistance because the load exerted forces the constituent transmitted through the sample before failure is established,
particles of the slope material together. Frictional resistance or by triaxial apparatus which has the advantage of being
is therefore a function of both the inherent frictional able to monitor and control the pore-water pressure in
properties of slope materials and the normal stress acting the sample. Such tests show that natural soils show rather
on them. Clearly, as the shear plane angle becomes steeper more complex responses to shear stress than would be
the shear stress becomes larger and the normal stress anticipated from the factors we have discussed so far.
smaller. The maximum angle attained when failure of the
slope materials occurs is known as the threshold angle of
stability. Box 7.1 The Coulomb-Terzaghi shear strength
This analysis is in fact over-simplified because normal equation
stress is exerted only at points of contact between indi-
vidual particles and not along the entire surface area of the The total shear strength of a slope material (s) is given as
shear plane. Below the water table the voids between s=c+o'.tan<j>
particles are filled by water and this gives rise to a positive
where c is cohesion, 0' the effective normal stress and tan <j>
(greater than atmospheric) pore-water pressure which has the coefficient of plane sliding friction which characterizes
a buoyancy effect on the overlying material and thus acts in the packing, surface roughness and hardness of the consti-
opposition to the normal stress. Consequently, in this case tuent particles.
the effective normal stress is less than the total normal
 Slope processes and forms 165

..c:
..... A Peak strength
C>
c:
...
Q)
.....
CJ)

...
co
Q)
..c:
CJ)

Effective Normal Stress Time

Fig. 7.3 Change in shear strength with time of densely and

loosely packed soils. The behaviour of most soils lies somewhere
between the two curves depicted.

Figure. 7.3 illustrates curves showing changes in shear

..c:
..... stress over time for both densely packed and loosely
C> packed soils. Although both ultimately attain the same
c:
...
Q)
..... residual shear strength their initial behaviour differs. On the
CJ) application of stress the densely packed material rapidly
...
co attains a peak she.ar strength which then declines to the
Q)
..c: residual value. This behaviour can apparently be explained
CJ)
by the rearrangement of soil particles as stress is applied.
Cohesion This rearrangement eventually leads to a decline in shear
strength to the residual value. During this rearrangement
under stress the material as a whole may expand, an effect
known as dilatancy, probably due in part to the riding of
Effective Normal Stress particles over each other.

Fig. 7.2 Relationship between shear stress and normal stress

7.1.2 Rock properties
for two types o/slope material: (A) a dry sand in which there is Having considered the general factors determining the
only a frictional component directly related to effective normal strength of slope materials we will now look at rock and
stress; (B) the usual situation where the slope material has some
initial strength from cohesion irrespective of effective normal soil more specifically. Different rock types vary
stress, but its shear strength increases as effective normal stress enormously in their intact strength, that is, the strength of
increases. the rock excluding the effects of fractures and joints (Table

Table 7.1 Intact strength of various types of rock

CHARACTERISTICS SCHMIDT HAMMER INTACT STRENGTH EXAMPLES

REBOUND VALUE CLASSIFICATION

Requires severe blows from hammer to break 100-60 Very strong Quartzite, dolerite, gabbro, basalt
intact sample
Hand-held sample can be broken with hammer 60-50 Strong Marble, granite, gneiss
Shallow indentations can be made by firm 50-40 Moderately strong Sandstone, shale, slate
hammer blow
Deep indentations can be made by firm hammer 40-35 Weak Coal, siltstone, schist
blow
Crumbles under sharp hammer blow - can be 35-10 Very weak Chalk, rock salt
cut with knife

Source: Modified from M. J. Selby (1982) Hillslope Materials and Processes. Oxford University Press, Oxford, Table 4.3, p. 65.
166 Exogenic processes and landforms

7.1). Intact strength can be rapidly assessed using a determining rock strength, it is the clay content that is the
Schmidt hammer, an instrument which measures the most critical factor in influencing the strength of the
rebound from the rock surface of a known impact. The majority of soils. With an increasing proportion of clay in
Coulomb-Terzaghi relation is capable of describing the the soil there is a greater potential for swelling and shrink-
shear strength of intact rock, but in a fractured and jointed
ing through hydration, for lowering of permeability due to
rock mass factors other than the cohesion and frictional increased water retention, for an increase in cohesion and
properites of the rock are largely responsible for its shear
for a lowering of the angle of friction.
strength. These factors include the degree of weathering, Soils are complex materials containing voids filled with
the width, continuity and spacing of joints and other air or water, in addition to solid particles of various sizes.
Whether a soil behaves as an elastic, plastic or viscous
fissures, the orientation of fissures (especially joints and
bedding planes) with respect to the ground slope, the material under stress depends on the proportions of these
presence of infilling material in partings, and the movement
components and especially the water content. A fine-
of water through the rock mass. Of these factors the grained cohesive soil with a low water content will exhibit
essentially elastic behaviour, failing by brittle fracture with
distance between individual fissures is in general the most
minimal plastic deformation. As the moisture content
important, there being no cohesive strength along fissures.
In some cases, however, the orientation of fissures can beincreases the behaviour of the soil changes from elastic to
critical, especially where the dip of bedding planes is close
plastic, the point of transition being known as the plastic
to the slope angle. limit. If the moisture content is increased further, a stage is
It is possible to assess the presence of these factors in the
eventually reached where the soil consists of a suspension
field and thereby provide a rating of rock strength in of particles in water and exhibits viscous behaviour. This
transition is referred to as the liquid limit. It has also been
relation to shear stress (Table 7.2). Each variable is given a
noted that when fine-grained cohesive soils are .dried
weighting in proportion to its estimated importance and its
value can be assessed for each variable for any particularbeyond the plastic limit the soil continues to decrease in
rock type. This enables an estimate of rock mass strength volume down to a certain moisture content, known as the
to be made. shrinkage limit. Beyond this limit no further volume
change occurs in spite of further moisture losses. These
transitions between different forms of soil behaviour are
7.1.3 Soil properties
known as Atterberg limits, after the Swedish soil scientist
Whereas joint spacing is the most significant variable in who first devised a means of quantifying the effects of

Table 7.2 Classification of rock mass strength

VARIABLE WEIGHTING % VERY STRONG STRONG MODERATE WEAK VERY WEAK

Intact rock strength 20 10<H>0 60-50 50-40 40-35 35-10

(Schmidt hammer r = 20 r= 18 r= 14 r = 10 r=S
rebound value)
Weathering 10 Unweathered Slightly weathered Moderately Highly weathered Completely
weathered weathered
r = 10 r=9 r=7 r=S r=3
Joint spacing 30 >3m 3-1 m I--D.3 m 300-S0mm <SOmm
r = 30 r=28 r= 21 r= IS r=8
Joint orientations 20 Very favourable. Favourable. Fair. Horizontal Unfavourable. Very unfavourable.
Steep dips into Moderate dips into dips or nearly Moderate dips Steep dips out of
slope, cross joints slope vertical dips out of slope slope
interlock (hard rocks only)
r = 20 r= 18 r= 14 r=9 r=S
Joint width 7 <0.1 mm 0.1-1 mm I-Smm S-20mm >20mm
r=7 r=6 r=S r=4 r=2
Joint continuity 7 None, continuous Few, continuous Continuous, no Continuous, thin Continuous, thick
andinfill infill infill infill
r= 7 r=6 r=S r=4 r=1
Ground-water 6 None Trace Slight Moderate Great
outflow <40 ml s-I m-2 40-200 m s-I m-2 >200 ml s-1 m -2
r=6 r=S r=4 r=3 r= I

Total rating 100-91 90-71 70-51 50-26 <26

Source: Modified from M. J. Selby (1980) Zeitschriftfur Geomorphologie 24, Table 6, pp. 44 and 45.
 Slope processes and forms 167

Table 7.3· Atterberg limits for various clay minerals

MINERAL SURFACE AREA PLASTIC LIMIT LIQUID LIMIT SHRINKAGE LIMIT VOLUME CHANGE
(m 2 kg-I) (%) (%) (%)

Smectite 800000 50--100 100--900 8.5-15 High

Illite 80000 35--{)0 60--120 15-17 Medium
Kaolinite IS 000 25-40 30--110 3-15 Low

Source: Modified from M. J. Selby (1982), Hillslope Materials and Processes. Oxford University Press,
Oxford, Table 4.8, p. 79; based on data in R. E. Grim, 1968, Clay Mineralogy. McGraw-Hili, New York.

water content on soil behaviour. They are expressed in 7.2 Mass movement
terms of the weight of contained water as a percentage of Mass movement is the downslope movement of slope
the weight of dry soil.
material under the influence of the gravitational force of the
When subject to the same stress different soils with an material itself and without the assistance of moving water,
identical moisture content may fail by brittle fracture, de- ice or air. The distinction between mass movement and the
form plastically or behave as a viscous fluid. This is transport of material by other denudational processes is,
because of the varying abilities of different clay minerals to however, not always clear-cut in practice since mass
absorb water (Table 7.3). Smectite generally has the highest movements involving material with a high water content
plastic and liquid limits because its crystal structure enables grade into fluvial transport where streams carry very large
it to provide an enormous surface area for water absorption. loads of fine sediment. In a very real sense glacier flow
Consequently, smectite-rich soils swell and shrink signi- itself is a form of mass movement, involving as it does the
ficantly when they are wetted and dried. Kaolinite, by con- downslope movement of coherent masses of ice; never-
trast, has a much more limited water absorption capacity. theless, glaciers have particular characteristics which merit
An important parameter derived from Atterberg limits is special consideration and they are examined separately in
the plasticity index. This is defined as the liquid limit the context of glacial landscapes (see Chapter 11). Very
minus the plastic limit and indicates the range of moisture
large-scale movements of rock which are transitional to
content over which a soil exhibits plastic behaviour. It is an
tectonic processes can also occur under gravity; we discuss
important indicator of the potential instability of the soil these briefly in Section 7.3. The term mass wasting is often
since the higher the plasticity index the less stable slope regarded as synonymous with mass movement, but it is also
materials will be. used in a broader sense to encompass all processes involved
Two further aspects of soil properties need to be ment- in the lowering of the landscape. Before we consider the
ioned briefly. Certain soils have an open 'honeycomb' various mechanisms of mass movement we will look
structure which allows them to retain water at proportions briefly at the conditions which give rise to them.
in excess of the liquid limit. The structure of these soils,
termed sensitive soils, is potentially unstable. If they are
subject to high shear stresses (such as those induced by an
7.2.1 Slope stability
earthquake), or high compressive stresses (such as those
arising from loading by burial), they can collapse catastro- The stability of a slope can be expressed in terms of the
phically as the water is squeezed out and the soil becomes a relationship between those stresses tending to disturb the
fluid. Such soils are sometimes called quick clays and are slope material and cause it to move and those forces tending
often associated with major and rapid flows of slope to resist these driving stresses. Clearly, movement will
materials. occur where driving forces exceed resisting forces and this
Sands can also act in a fashion similar to sensitive soils relationship is represented as the safety factor for a slope.
under certain conditions. In a saturated mass of sand most This is expressed as the ratio between shear strength and
of the strength arises from point to point contacts between shear stress (Box 7.2).
the solid sand grains. If the sand is shaken violently, by a Slopes can exist in one of three states. Where shear
seismic shock for instance, all the effective stresses can be strength is significantly larger than shear stress the slope is
transferred from grain-to-grain contacts to the pore water, described as stable (safety factor> 1.3). Where shear stress
an effict known as liquefaction. All the strength from exceeds shear strength (safety factor < 1) there will be
interparticle friction is thereby lost, the sand mass con- continuous or intermittent movement and the slope is de-
sequently has no resistance to shear stress and liquid scribed as actively unstable. Since shear strength can vary
deformation occurs. over time, especially in response to changes in the water
content of slope materials, the third stability category is the
168 Exogenic processes and landforms

Box 7.2 Safety factor for a slope the factors which promote movement, it has limited
applicability to specific situations. This is because both
The safety factor (F) is defined as cohesion and pore-water pressure are highly variable on
most natural slopes, even over short distances and brief
s
F=- periods of time. For instance, fissures may traverse slope
't
materials leading to drastic variations in pore-water pressure
where s is the total shear strength along a specific shear from place to place.
plane, and 't the total amount of shear stress developed along
this plane. For shallow, translational slides F is defined as
(yz cos2 ~ - u) tan <I>
7.2.2 Mass movement proces~es
F=c+ There have been numerous attempts to classify the diverse
yz sin ~ cos ~
where c is cohesion, y the uni~ weight of regolith, z the ver-
modes of mass movement, none of them univ~rsally
tical depth to the shear plane, ~ the angle of the shear plane, satisfactory. Here we identify six fundamental types of
u the pore-water pressure at the shear plane and <I> the angle movement - creep, flow, slide, heave, fall and subsidence.
of internal friction. Each of these can be subdivided into more specific forms of
mass movement (Table 7.5). Classifications of the various
processes of mass movement are valuable in indicating the
conditionally stable slope which has a safety factor of range of mechanisms and forms of motion, but it must be
1-1.3 and fails on occasion in response to transient changes appreciated that most movements in reality involve a
in shear strength. Numerous factors contribute to the combination of processes. Debris avalanches, for example,
occurrence of mass movements, and these are listed in may begin as slides consisting of large masses of rock but
Table 7.4. They can be categorized as either preparatory then rapidly break up to form flows as the material is
factors or triggering factors. Preparatory factors make the pulverized in transit. The compound nature of many forms
slope susceptible to movement without actually initiating of mass movement is illustrated in Figure 7.4, which also
failure by transforming it into a conditionally stable state. indicates how the different types of movement vary in their
Triggering factors transform the slope from a conditionally moisture content and velocity. Flows tend to be wet and
stable to an actively unstable state. slides dry, while heave can occur over a fairly broad range
Although the slope stability approach to analyzing mass of moisture conditions. Heave processes are invariably
movements provides a good theoretical understanding of slow, whereas both flows and slides tend to be rapid.

7.2.2.1 Creep
Table 7.4 Factors contributing to the occurrence of mass movement Creep is the slow, plastic deformation of rock or soil in
FACTOR EXAMPLES

Factors contributing to increased shear stress FLOW

Removal of lateral Erosion by rivers and glaciers, wave action,
support through faulting, previous rock falls or slides /
undercutting or slope
steepening /
Removal of underlying
support
Undercutting by rivers and waves, subsurface
solution, loss of strength by extrusion of
underlying sediments
/
Loading of slope

Lateral pressure
Weight of water, vegetation, accumulation of
debris
Water in cracks, freezing in cracks, swelling
/
(especially through hydration of clays).
pressure release
Transient stresses Earthquakes, movement of trees in wind

Factors contributing to reduced shear strength

Weathering effects Disintegration of granular rocks, hydration
of clay minerals, dissolution of cementing
minerals in rock or soil
Changes in pore-water Saturation, softening of material
pressure Fig.7.4 Classification of mass movements in terms ofpure flow,
Changes of structure Creation of fissures in shales and clays, slide and heave. Note the compound nature of some movements.
remoulding o·f sand and sensitive clays (After M. A. Carson and M. 1. Kirkby (1972) Hillslope Form and
Organic effects Burrowing of animals, decay of tree roots Process. Cambridge University Press, Cambridge, Fig. 5.2,
p.100.)
 Slope processes and forms 169

Table 7.5 Classification and characteristics of the major types of mass movement

PRIMARY MASS MOVEMENT MATERIALS IN MOISTURE TYPE OF STRAIN AND RATE OF

MECHANISM TYPE MOTION CONTENT NATURE OF MOVEMENT MOVEMENT

Rock creep Rock (especially readily Low Slow plastic deformation of Very slow to
deformable types such as rock, or soil producing a extremely slow
"'- shales and clays) variety of forms including cam-
"~ bering, valley bulging and out-
U crop bedding curvature
Continuous creep Soil Low
Dry flow Sand or silt Very low Funnelled flow down steep Rapid to
slopes of non-cohesive extremely rapid
sediments
Solifluction Soil High Widespread flow of saturated Very slow to
soil over low to moderate extremely slow
angle slopes
Gelifluction Soil High Widespread flow of seasonally Very slow to
saturated soil over permanently extremely slow
frozen subsoil
Mud flow > 80% clay-sized Extremely high Confined elongated flow Slow
Slow earthflow > 80% sand-sized Low Confined elongated flow Slow
Rapid earthflow Soil containing Very high Rapid collapse and lateral Very rapid
sensitive clays spreading of soil following
f-< ~
disturbance, often by an initial
Z 0 slide
«
Z
u:: Debris flow Mixture of fine and High Flow usually focused into Very rapid
:E0 coarse debris (20-80% of pre-existing drainage lines
Cl particles coarser than
u.l sand-sized)
~
"- Debris (rock) Rock debris, in some Low Catastrophic low friction Extremely rapid
f-< avalanche cases with movement of up to several kilo-
Z
u.l (sturzstrom) ice and snow metres, usually precipitated by a
Z major rock fall and capable of
0
"- overriding significant topo-
::E graphic features
0
u Snow avalanche Snow and ice, in some Low Catastrophic low friction Extremely rapid
....l
« cases with rock debris movement precipitated by fall or
~
u.l slide
f-<
«
....l
Slush avalanche Water-saturated snow Extremely high Flow along existing drainage Very rapid
lines
Rock slide Unfractured rock mass Low Shallow slide approximately Very slow to
parallel to ground surface of extremely rapid
coherent rock mass along
<:; single fracture
s::
.,g Rock block slide Fractured rock Low Slide approximatellr parallel to Moderate
-'l ground surface of ractured rock
'"s::
"
:9 ~'" Debris/earth slide Rock debris or soil Low to moderate Shallow slide of deformed
masses of soil
Very slow to
rapid
Vi
Debris/earth block Rock debris or soil Low to moderate Shallow slide of largely Slow
slide undeformed masses of soil
<:; Rock slump Rock Low Rotational movement along Extremely slow to
s:: concave failure plane moderate
·2
§ Debris/earth slump Rock debris or soil Moderate Rotational movement along Slow
c
<>:: concave failure plane
Soil creep Soil Low Widespread incremental Extremely slow
">
oj
downslope movement of soil or
rock particles
:r:" Talus creep Rock debris Low
Rock fall Detached rock joint Low Fall of individual blocks from Extremely rapid
blocks vertical faces
f-< 0;
Debris/earth fall Detached cohesive units Low Toppling of cohesive units of Very rapid
Z
u.l
2
'" (topple) of soil soil from near-vertical faces such
0 as river banks
"-
::Ef-< Cavity collapse Rock or soil Low Collapse of rock or soil into Very rapid
02 underground cavities such as
u«
....l~ limestone caves or lava tubes
«::E "u
S!o
f-<Cl
""
:9
Settlement Soil Low Lowering of surface due to
ground compaction usually
Slow
O::u.l
u.l~
15::l resulting from withdrawal of
>"- r/) ground water

Source: Based largely on D. J. Varnes (1978) in: R. L. Schuster and R. J. Krizek (eds) Landslide Analysis and Control, Transportation
Research Board Special Report 176. National Academy of Sciences, Washington, DC, 11-33.
170 Exogenic processes and landforms

response to stress generated by the weight of overburden. It A PURE FLOW

begins once the yield stress of the slope material is ex-
ceeded. In rock it can extend to hundreds of metres below
the surface. It occurs at very slow rates, typically
1 mm to 10 m a-I and is likely to be especially active where
weakly competent materials, such as clays, are overlain by
more competent beds. Creep is often a precursor of slide-
type movements, but it can also cause specific observable
effects. The bending of the lower parts of tree trunks is
often cited as evidence of creep, but this phenomenon can
also be· caused by other mechanisms. More substantial
evidence of creep is provided by the downslope curvature
of strata near the surface. Another consequence of creep is
cambering which involves the extrusion of weak rocks B PURE SLIDE
(usually clays), either lying below, or interbedded with,
more rigid strata which causes valley sides to bulge.
It is important to distinguish the type of creep described
here from soil creep and talus creep. The former acts solely
under gravity, whereas the latter involve heave and are con-
sequently considered in Section 7.2.2.4.

7.2.2.2 Flow
In a pure flow, shear occurs throughout the moving mass of
material and there is no well-defined shear plane (Fig.
7.5(A». Flow is distinguished from creep by having dis-
crete boundaries or narrow peripheral zones experiencing
shear. Shear is at a maximum at the base of the flow, but C PURE HEAVE
here the rate of flow is relatively slow and nearly all the
movement occurs as turbulent motion within the body of
the flowing mass. Dry flows can occur, but abundant water is
usually present. They are often initiated by falls or slides,
becoming flows when the moving soil or rock mass breaks
up. Flows are categorized as avalanches, debris flows,
earthflows or mudflows depending on whether they
consist of predominantly snow and ice, rock fragments,
sand-sized material or clay (Fig. 7.6).
Where the flow has a high water content it may extend as
a long, narrow tongue well beyond the base of the slope
from which it originated. Such flow are usually more or
less confined to existing drainage lines (Fig. 7.7) and there
is in fact a transition between mudflows and streams laden Fig. 7.5 Velocity profiles for ideal types of mass movement: (A)
with abundant fine sediment. Earthflows involve the ex- pure flow, (B) pure slide and (C) pure heave. (Modifiedfrom
trusion of lobes downslope and are usually slow moving. M. A. Carson and M. 1. Kirkby (1972) Hillslope Form and
Where the slope material is composed of sensitive soils, Process. Cambridge University Press, Cambridge, Fig. 5.1,
however, an initial disturbance can cause an instantaneous p.100).
loss of shear strength and promote a rapid earthflow. The
slowest type of flow is solifluction which involves the frequently occurs in close association with soil creep. Many
downslope movement of saturated soil. Solifluction can hillslopes show the combined effects of both processes
0
occur at slope angles as low as 1 and is particularly active (Fig. 7.8).
in periglacial environments. Here abundant moisture is In addition to the transport of material from higher to
made available by seasonal thawing of soil above a frozen lower el~vations, flows can be effective erosional agents,
subsurface, and this form of solifluction is termed especially the more energetic varieties. This is paiticularly
gelifluction (see Section 12.2.3). Although solifluction can true for debris flows and for debris and snow avalanches
be regarded as a distinct mass movement process it which travel over the ground, rather than predominantly
 Slope processes and forms 171

DRY FLOW SOLIFLUCTION

EARTH FLOW DEBRIS FLOW DEBRIS AVALANCHE

Fig.7.6 Varieties offlow-type mass movement. (Based on D. 1.

Varnes (1978) in R. L. Schuster and R. J. Krizek (eds) Landslides:
Analysis and Control. Transportation Research Board Special
Report 176. National Academy of Sciences, Washington, DC,
Fig. 2.1.)

through the air. The major geomorphic effects of ava-

lanches are the removal of debris from gullies and slope
faces, the excavation of rock to form avalanche chutes, the
erosion and redistribution of unconsolidated slope deposits,
Fig.7.7 Debris flow on western flank of the Southern Alps,
and the deposition of snow and/or rock debris.
South 1sland, New Zealand.
Major avalanches can be one of the most violent and
destructive forms of geomorphic activity. Two particularly
catastrophic avalanches, both triggered by earthquakes,
crashed from the the mountain peak of Hua caran in the
Peruvian Andes in 1962 and 1970. In the 1962 event 3 Mt
of ice and 9 Mt of rock were transported at speeds well in
excess of 100 km h- 1 over a horizontal distance of 20 km.
An estimated 3500 people were killed. In 1970 a second
avalanche occurred which initially moved along the same
route as the 1962 event, but after travelling some 16 km at
an average speed of around 300 km h- I part of the turbulent
flow of rock debris and ice jumped across a 300 m high
ridge and buried the town of Yungay beyond. Op this
occasion the death toll was possibly as high as 40000.
The extremely high velocities achieved by some ava-
lanches clearly require explanation. Some may ride on a Fig. 7.8 Valley cut into a chalk escarpment in Kent, UK. Such
layer of compressed air trapped between the avalanche debris valleys on the chalk are known as coombes, and solifluction
and the ground surface. This has the effect of greatly re- deposits, known as coombe rock, are often present on valley
ducing the frictional drag on the moving mass. Other possible floors (evident in photograph from the flat bottom of the valley).
These solifluction deposits prohahly accumulated at the end of the
explanations of high rates of movement include fluidiza-
last glacial 10000-15000 a BP under periglacial conditions (and
tion where fine particles are kept in suspension by a flow of are therefore more accurately descrihed as gelifluction deposits).
air, and cohesionless grain flow in which particle motion is Terracetles can be seen on the valley-side slopes and are
sustained by continuous collisions as in a fluid. prohahly related to soil creep.
172 Exogenic processes and landforms

Although spectacular, avalanches are not of widespread Minor scarp Main scarp

significance. They are largely confined to terrain character-

ized by high local relief and steep 'slopes and are thus most
common in active orogenic belts. On a global basis the
more ubiquitous but less spectacular slower forms of flow
are more significant denudational agents.

7.2.2.3 Slide
Slide is an extremely widespread form of mass movement,
and the term landslide is part of our everyday vocabulary.
This presents problems when using it in a specific technical
sense since landslide in general usage simply means the
rapid downslope movement of slope material. Applied in
Surface of
this sense many landslides also involve fall and flow. In a separation

pure slide failure occurs along a well-defined shear plane Fig.7.10 Major features of a rotational slide (Modified from
(Fig. 7.5(B)). Resistance to movement falls sharply immedi- D. J. Varnes (1978) in: R. L. Schuster and R. J. Krizek (eds)
ately the initial failure takes place, and downslope movement Landslides: Analysis and Control. Transportation Research
Board Special Report 176. National Academy ofSciences,
continues until there is a sufficient increase in resistance, Washington, DC, Fig. 2.1.)
often related to a decrease in slope angle, to halt it.
Slides are nearly always long in relation to their width
surface. Both rotational and translational slides are precipi-
and depth, their length-width ratio typically being 10: 1.
tated by a temporary excess of shear stress over shear
They can be subdivided into translational slides, which
strength within the slope (Box 7.3). The difference between
have predominantly planar shear surfaces, and rotational
the two types of movement is that in thick, relatively
slides in which the shear plane is concave-up (Fig. 7.9).
homogeneous material the depth at which the ratio between
Rotational slides are most common where slopes consist of
thick, homogeneous materials, such as clays. The rotational
movement can result in the upper part of the slumped mass Box 7.3 Analysis of rotational slides
being back-tilted towards the failure surface (Fig. 7.10).
The material can move as a single block, but usually it is The shear plane of a rotational slide is curved and therefore
broken into several discrete segments separated by transverse the stability analysis outlined in Box 7.2 must be modified.
fissures. Movement at the base of rotational slides in clay One way the stability of a rotational slide can be evaluated
is to divide the slide into a number of 'slices' of length L
or similar cohesive material is often transformed into that and aggregate the forces acting of each of these slices (Fig.
of an earthflow and this gives rise to a chaotic, humm'ocky B7.3). The weight (W) is taken as operating through the
centre of each slice. The angle of the shear plane (a) is
calculated for each slice from the centre of rotation (0).
TRANSLATIONAL TRANSLATIONAL ROCK SLIDE The effective normal stress (cr') at the base of each slice is
ROCK BLOCK SLIDE
W cos a and the shear strength (s) is W sin a. The safety
Scarp face-
c;:o,ntrol by Dip slope- factor (F) can then be defined as
JOInts .......... b~~td?~~Y

~
A

(J) F= L [cL+ (W cos (X- uL) tan <1>]

B A

~. L
B
Wsina

where c is cohesion, u the pore-water pressure at the base of

TRANSLATIONAL EARTH ROTATIONAL SLUMP the slice and tan <I> the angle of internal friction.
BLOCK SliDE

Fig. 7.9 Varieties of slide-type mass movements. (Based on D. J. Fig. B7.3 Stability analysis of a deep-seated slide using the
Varnes, 1978, in: R. L. Schuster and R. J. Krizek (eds) Landslides: method of slices. (After M. A. Carson and M. J. Kirkby (1972)
Analysis and Control. Transportation Research Board Special Hillslope Form and Process. Cambridge University Press,
Report 176, National Academy ofSciences, Washington, DC, Cambridge, Fig. 7.11,p.167.)
Fig. 2.1.)
 Slope processes and forms 173

shear strength and shear stress is at a minimum (that is, the ROCK FALL EA~TH FALL (TOPPLE)

potential shear plane) forms an arc rather than a straight line.

The great majority of slides are small and shallow with
lengths of a few tens of metres and depths of 2-3 m. Slides
in bedrock are less common, but may attain enormous
dimensions and involve the movement of millions of cubic
metres of material. Very large slides usually break up to
form debris avalanches but in some cases the rock travels a
significant distance as a coherent mass. This is most likely
to occur where competent beds slide over incompetent
strata (often clay or mudstone) dipping steeply roughly
Fig.7.ll Varieties offall-type mass movements. (Based on D. 1.
parallel with the ground slope. Such conditions contribute
Varnes (1978) in R. L. Schuster and R. J. Krizek (eds) Landslides:
to low shear strength and enhance the probability of slides Analysis and Control. Transportation Research Board Special
occurring. Report 176. National Academy ojSciences, Washington, DC,
Probably the largest slide on Earth is the Saidmarreh Fig. 2.1.)
slide located in south-western Iran. Although it occurred
more than 10 000 a BP it has suffered only ~uperficial the operation of soil creep because above an angle of
modification by subsequent erosion. The deposits are, around 25° the soil cover is thin or absent. On steep, grass-
crudely stratified, indicating that the movement was not covered slopes flights of narrow steps, called· terracettes,
predominantly one of a turbulent debris avalanche. A mass are commonly present and it is likely that these are related
of limestone some 15 km long, 5 km wide and at least 300 to soil creep, although shallow landslides may also be instru-
m thick slid off the underlying interbedded marl and lime- mental in their development (Fig. 7.8).
stone which dips at an angle of around 20° out of the slope.
The initial vertical component of movement was only about 7.2.2.5 Fall
1000 m, but the slide travelled a total distance of 18 km, Fall involves the downward motion of rock or, more rarely,
crossing an 800 m high ridge en route. soil through the air (Fig. 7.11). Soil IS not frequently in-
volved for the simple reason that free fall can only occur
7.2.2.4 Heave from very steep slopes or cliff faces which, of course, have
In pure heave the slope material experiences cycles of very little soil cover. An exception is the toppling of slabs
expansion and contraction (Fig. 7.5(C)). Downslope move- of earth along river banks, a process often referred to as
ment arises from the fact that while expansion occurs bank calving. This arises from the undercutting of banks
normal to the sloping ground surface contraction under by streams and is a very common phenomenon which con-
gravity tends to be more nearly vertical. Cohesion between tributes large quantities of sediment to river channels.
particles usually prevents a purely vertical return movement Topples are distinguished from other types of fall by the
under gravity. Two types of heave can be distinguished on rotation of the block of material as it falls away. Topple can
the basis of the size of the constituent particles - soil creep occur in rock especially where joints are vertically exten-
and talus creep; the latter involves coarser material than sive in relation to their width and where they dip out of a
the former. Expansion and contraction can be caused by slope.
wetting and drying, freezing and thawing (in which case the Rock can become detached as a result of various physi-
process is described as frost creep - see Section 12.2.3), cal weathering processes, including pressure release and
temperature changes and the burrowing activity of worms joint widening by frost action, and the fragments produced
and other organisms. are rapidly removed under gravity. Rock falls are common
The rate of soil or talus creep on a slope depends on a phenomena in terrain characterized by high, steep rock slopes
number of factors. It will become greater with increasing and cliffs (Fig. 7.12). Where the dislodged fragments are
slope angle since this increases the downslope component large they accumulate a significant amount of kinetic
of movement. It will also be high in soils containing energy by the time they impact on the slope below and they
abundant quantities of clays, such as smectite, which can therefore be an active erosive agent by detaching other
expand significantly on wetting, or in silt-sized material fragments. Large rock falls originating from a considerable
which is capable of substantial ice accumulation. Soil and height above the ground spread their debris over an extensive
talus creep will, however, decrease with depth below the area unless the dispersal of material is confined by topo-
slope surface, both because of the moderation of changes graphy. As already mentioned, large rock falls are often
associated with wetting and drying and freeze-thaw, and transformed into debris avalanches once they have made
the increase in the weight of the overburden. There is some their initial impact. In situations where deep valleys are cut
support from experimental and field observations for these into hard rocks, such as granites and some sandstones, by
expected relationships, although there is an upper limit to glacier.s or rapidly incising river channels, the release of
174 Exogenic processes and landforms

also occur naturally where the volume of poorly compacted

materials is decreased by the addition of water (hydro-
compaction) or by vibrations such as those generated by
earthquakes.

7.3 Gravity tectonics

Gravity tectonics is a useful term which covers a range of
processes extending from the very large-scale movements
of rock masses involved in the development of thrusts and
nappes to smaller scale downdip translocations of material
which are transitional to landslides. Many of the massive
nappes occurring in intercontinental collision orogens such
as the Alps are now considered to be vast gravity slides
moving away from their high, axial zones (see Section
3.4.1).
Gravity tectonics can involve both spreading and sliding,
the two processes being closely associated, but not ident-
ical. Rocks located high up in mountain masses and
bounded by steep slopes gradually yield and move
downslope under gravity. This kind of motion can be
accommodated by internal movements within the rock
(gravity spreading) or it may occur primarily through
gravity sliding of the rock mass over a few well-defined
planes composed of incompetent strata. Although the
movement is essentially downslope, internal deformation
and the rotation of blocks along small-scale faults can
create chaotic patterns at the local scale (Fig. 7.13).
Such structures clearly have a combined endogenic and
exogenic origin. The initial energy input is endogenic,
Fig. 7.12 RockfaLL in a recently glaciated vaLLey in the Southern involving as it does orogenic uplift. But the subsequent
o
Alps, South Island, New Zealand. The area offresh rockfacefrom sliding, which may take place over gradients as low as 0.5 ,
which the large blocks of rock have faLLen is more than 20 m high.

A
lateral confining pressure along the valley walls can give ENE

rise to pressure release and generate tension joints running

roughly parallel with the ground slope. These can promote
slab failure when the progressive widening of these joints Ophiolites Dolomite Upper Triassic
evaporites
eventually leads to the detachment of thick slices of rock
(Fig. 6.21). B

-- - ~
/
,\" \ " \ : \ \ \ \ \ \ \ ',\\
7.2.2.6 Subsidence
Subsidence can occur either as the more or less instan-
taneous collapse of material into a cave or other cavity c Surface of
(cavity collapse) or .as a progressive lowering of the decollement
ground surface (settlement). Cavity collapse is largely
confined to limestone terrains where the roofs of under-
ground cavities occasionally collapse. More rarely lava
tubes' within lava flows may experience a similar fate. Fig. 7.13 Schematic representation of gravity.tectonics in the
Cavity collapse can also occur as a result of human acti- northern Apennines, near Florence, Italy: (A) the original
vites such as mining. Settlement usually arises from the depositional basin; (B) thrusting from the west; (C) gravity
sliding over Upper Triassic evaporites (solid shading) acting as a
lowering of water tables and is most dramatically illustrated decoLLement surface. (After P. Elter and L. Trevisan (1973) in:
in areas where there has been oil drilling or large-scale K. A. De long and R. Scholten (eds) Gravity and Tectonics. Wiley,
abstraction of ground water for irrigation. Settlement can New York, Fig. 15,p.187.)
 Slope processes and forms 175

is dependent on stresses ansmg from the gravitational which it can absorb water. Many factors influence the
instability of the rock mass. Unloading of mountain ranges infiltration capacity of natural surfaces. The most important
by gravity spreading and sliding, a process sometimes are particle size and the abundance of organic matter and
described as tectonic denudation, occurs in continental- intensity of faunal activity. Each 'of these factors promotes
margin arc orogens such as the Andes as well as in inter- the development of an open soil structure which is capable
continental collision belts. Slide rates are slow, reaching a of absorbing water efficiently from a slope surface. Slope
maximum of about 100 m a-lover low-friction decollement surfaces with low infiltration capacities are most likely to
layers such as salt, but are orders of magnitude lower in be encountered in arid and semi-arid environments where
most cases. soils have a low organic content, but they are also common
in all environments where human activities lead to an
artificial removal of the vegetation cover or compaction of
7.4 Water erosion and solute transport on slopes the surface.
In addition to being affected by mass movement processes, Where the intensity of precipitation exceeds the infil-
slope material can be transported by water. Three mechan- tration capacity of the surface a proportion of the precipi-
isms are involved: rainsplash erosion, in which particles tation will flow over the surface as infiltratiorr-excess
on the soil surface are dislodged by the impact of raindrops; overland flow (or Hortonian overland flow as it is
slope wash, which is the process whereby sediment is sometimes termed after the pioneer in this aspect of
entrained and transported by a thin sheet of water flowing hydrological research, R. E. Horton). In other cases the
over the slope surface; and solute transport where soil infiltration capacity of the slope surface may not be
materials taken into solution during weathering reactions are exceeded by the rate of incoming precipitation but the
transported downslope as solutes. These mechanisms are combined intensity and duration of precipitation may be
sometimes collectively referred to as wash processes. sufficient locally to saturate the soil and raise the water
Rainsplash and slope wash are effective only on soil- table to the surface. These conditions will give rise to
covered slopes, whereas solution can operate on both soil saturation overland flow which is most likely to occur in
and rock slopes. In order to understand how materials are -areas of a drainage basin where the water table is already
transported downslope by water we first need to look at the relatively close to the surface, such as close to stream
way water moves over, and within, hillslopes. channels. This is the predominant form of overland flow in
humid environments.
7.4.1 HilIslope hydrology
7.4.2 Rainsplash erosion
Precipitation falling on a slope either runs off the surface, is
held in surface depressions (surface detention), or infiltrates Raindrops possess kinetic energy by virtue of 'their mass
through the slope surface. Except on impermeable rock and velocity. Although the impact velocity of raindrops
slopes some proportion of the precipitation reaching the varies depending on droplet size, wind speed and turbu-
slope almost invariably infiltrates the surface and either lence, raindrops of the maximum size under normal
percolates down to the water table to contribute to ground- conditions of around 6 mm diameter have an impact
water storage, or moves laterally down through the slope velocity of about 9 m S-I. At this speed rain drops can
more or less parallel to the surface (Fig. 7.14). This-lateral directly move particles more than 10 mm across and
movement occurs because compaction, the infilling of coarser material can be dislodged by the removal of
voids by fine particles washed down from above, and, in downslope support provided by finer sediment.
some cases, the precipitation of iron oxide, silica, calcium Rainsplash erosion can occur wherever vegetation does
carbonate and other compounds reduces the ability of the not entirely cover the ground, although it is a more potent
soil to transmit water at depth. Water that is diverted late- erosive agent in environments where there is little or no
rally is termed throughflow, although the term interflow is vegetation cover. Both slope gradient and surface charac-
also sometimes applied to water moving laterally below the teristics influence the effectiveness of rainsplash erosion.
soil proper but above the water table. The route taken by Experimental studies have shown that on low angle slopes
water that has infiltrated the slope surface and the rate at of 5° only about 60 per cent of the particles dislodged by
which it is transmitted through the slope are particularly raindrop impacts move downslope but this percentage
important in terms of the transport of solutes. increases with gradient reaching 95 per cent on 25 ° slopes.
The proportion of precipitation that flows down a slope It also appears that rainsplash erosion is more effective on
surface rather than infiltrating can depend on both the sandy surfaces than those containing a high proportion of
intensity and duration of precipitation as well as the pro- clay and silt-sized material, apparently because the pres-
perties of the slope surface. The latter determine the ence of finer particles contributes to cohesion.
infiltration capacity of the surface - that is, the rate at Although rainsplash alone can cause significant_ erosion,
176 Exogenic processes and landforms

/ I /
/ I I
/ / I
Precipitation / / /

/ / 1
Subsurface flow from matrix into

1 11 upstream end of pipes

w
=:!
u..
::::!o
o a::
(/)0..

pipe
Z
.0
u..t-
0«
....J
o
wu
Za::
Ow
No..

Throughflow seepage from

saturated part of soil

~ Turbulent flow of water

~Laminar flow of water

Fig.7.14 Routes of subsurface flow on hills/opes. (From T. C. Atkinson (1978) in M. 1. Kirkby (ed.) Hillslope Hydrology, Wiley,
Chichester, Fig. 3.1,p. 74.)

especially on very steep, highly permeable slopes where it trated. Finally, rainsplash can contribute to slope wash by
may be as important as slope wa h, it is most effective lowering infiltration capacity through the creation of a thin
when it is combined with slope wash. If slope wash occurs surface crust formed through the infilling of near-surface
without rainspla h the loose surface material is rapidly voids by fine particles displaced by raindrop impacts.
removed, but the rate of erosion quickly declines once the
more compacted material below is exposed. But where
7.4.3 Slope wash
rainsplash is combined with slope wash a high rate of
erosion can be more effectively sustained because the The movement of water across a slope surface, irrespective
surface is constantly being disrupted by raindrop impacts of how it is generated, is termed sheet flow, although this is
and particles are continuously being made available for a rather misleading description since the water flow is
entrainment and transport. Raindrop impacts also increase never of uniform depth because of the microtopography of
the turbulence of the water flow and this enhances its hillslope surfaces. Sheet flow can in fact grade into
ability to entrain and transport sediment. Probably the main channelled fluvial flow as the water movement becomes
role of rainsplash is the dislodgement of particles from progressively more concentrated into particular downslope
small ridges and bumps on hillslopes which are not covered routes, and the distinction between the two is sometimes
by overland flow. These particles move into adjacent difficult to make.
microtroughs where surface-water movement is concen- The mechanisms whereby particles are entrained and
 Slope processes and forms 177

transported by flowing water will be considered in Chapter channel network (see Section 8.1.3). Contour curvature has
8. Nevertheless, an important distinction which needs to be a significant effect on slope wash since where contours are
made here is that between true sheet flow, which can only convex in plan sheet flow will be dispersed downslope and
move very fine particles, and concentrated flow where the erosion will be minimized. Conversely, at valley heads and
greater depth of flow allows larger material to be trans- other locations where contours are concave in plan, the
ported. The erosional effectiveness of sheet flow is largely flow will be concentrated downslope and rill erosion will
controlled by the characteristics of the surface. These consequently be more effective (Fig. 7.16).
include particle size and degree of particle cohesion, the
extent and nature of the vegetation cover, and slope
gradient. By contrast, the rate of erosion for concentrated 7.4.4 Soil erosion
flow is determined more by the depth and velocity of flow. The rate at which material is eroded from a slope is a
In addition to material dislodged by rainsplash, slope wash function of both erodibility, or the resistance of slope
readily removes particles disturbed by the growth of ice materials to entrainment and transport, and erosivity, the
crystals or salts, or by animal activity. potential of slope processes to cause erosion. A multitude
Sustained concentrated flow can eventually produce of factors influence erodibility and erosivity, and it is
rills, microchannels a few centimetres in depth and width extremely difficult to quantify these in order to predict the
(Fig. 7.15). In humid environments the presence of veg- rate of erosion on a particular slope under a given set of
etation means that rills usually develop only on artificially conditions. None the less, there have been several attempts
disturbed surfaces, but in arid and semi-arid environments to express the rate of soil erosion as a function of specific
they can occur naturally. Although rills may be destroyed variables and these have been applied to the prediction of
between rainfall events by other slope processes, especially soil loss, especially where this is of importance to agriculture.
soil creep, those favourably located may eventually be The most comprehensive index of soil erosion is pro-
enlarged into gullies and form a permanent part of a vided by the Universal Soil-Loss Equation. This was

Fig. 7.15 Rills feeding into gullies on an unvegetated natural surface in the Painted Desert of northern Arizona, USA. The bag in the
bottom-left corner of the photograph indicates the small size of the rills.
178 Exogenic processes and landforms

greatest in those parts of a slope where subsurface flow is at

a maximum, such as at the base of slopes and where contour
curvature encourages the convergence of throughflow.

7.5 Rates of slope processes

Numerous estimates have been made of the rate of oper-
Fig.7.16 Effect of contour curvature on sheet flow and rill ation of individual slope processes and of the overall rate of
erosion: (A) contours convex in plan with sheet flow dispersed; slope denudation. A range of techniques have been
(B) contours concave in plan with sheet flow concentrated.
employed (Table 7.6), but virtually all measurements are
limited to a period of a few months or a few years, and it is
specifically developed for cropland and it incorporates a uncertain whether such short-term estimates are representa-
range of erodibility factors. These include the resistance of tive of average rates over the long term.
the soil to rainsplash and slope wash, slope length and
gradient, the proportion of the ground surface that is
Table 7.6 Techniques for estimating present rates of slope processes
covered by crops, and the presence or absence of soil
conservation measures such as contour ploughing. In ad- PROCESSES TECHNIQUES
dition it contains a measure of erosivity derived from the
kinetic energy of rainfall events. Although extensively used Soil creep 1. Young pits - rods or plates buried in a pit which is
re-excavated after a period of years.
to predict soil erosion on agricultural land, the Universal Soil- 2. Surface rods and buried and re-excavated cylinders,
Loss Equation is only of limited applicability in studies of deformation of plastic tubes
erosion on slopes under natural conditions. 3. Tilt of cylinders monitored from above by viewing
through cross-wires
Solifluction I. Buried plastic tubes with deformation measured by
7.4.5 Solute transport inclinometers or strain gauges
2. As for soil creep
The transport of slope materials ~n solution has been the Slope wash 1. Recording of ground loss using rods marked to
least studied aspect of slope denudation, and there are com- and rainsplash show position of the ground surface (erosion pins)
paratively few quantitative data by which direct comparisons 2. Collection of sediment moving over slope in a
downslope wash trap
can be made with the effectiveness of other slope pro-
Solute transport 1. Monitoring discharge and solute concentration in
cesses. The loss of weathered material in solution leads to and loss streams to estimate solute loss from slopes
the rearrangement and settling of the remaining particles, 2. Monitoring of solute concentration in throughflow
although it is difficult to calculate the effect of such loss on on individual slope sections.

the form of the slope itself. The rate of solute transport can Slides Estimation of volume of debris moved in individual
slides related to their estimated frequency of occurrence
certainly be estimated by measuring the discharge of sub-
Cliff and slope 1. Repeated surveying or comparison of large-scale
surface flow on a slope and relating this to its solute retreat maps constructed at different times over periods of
concentration. But this cannot be converted to a volume several decades
change unless we know what alterations in bulk density 2. Erosion pins and the direct recording of individual
rock falls and talus accumulation
have accompanied the weathering reactions which gave rise
to the release of the solutes. Indeed, in the extreme case of Source: Based on I. Saunders and A. Young (1983) Earth Surface
Processes and Landforms 8, 475-87.
isovolumetric weathering (see Section 6.2.4.1), the release
of material in solution is exactly compensated by a
decrease in the bulk density of the weathered material, so Few studies have attempted simultaneously to monitor
there is no change in volume and, therefore, no effect on various slope processes in a specific area. One notable
the configuration of the slope surface. example, however, is the classic investigation by A. Rapp
The clearest impact of solution is on slopes in limestone in the Karkevagge area of northern Sweden. Rapp moni-
terrain. Here the release of solutes occurs without the tored a range of processes over a period of 9 a on slopes in
accumulation of a significant weathering residue so solute mica-schist and amphibolite with gradients ranging from 15
transport is converted more directly into a change in surface to 45°. Rock fall was estimated from the accumulation of
form, even allowing for a significant proportion of the debris on seasonal snow patches, on areas of matting laid
solution occurring below the surface. Rates of solute release on the ground and in wire netting slung at the base of rock
are likely to be most rapid during the initial phase of faces. The rates of the various processes measured are
percolation or throughflow and decrease thereafter as the soil given in Table 7.7. Two comments need to be made on
solutions move towards equilibrium with the minerals these results. First, the considerable proportion of movement
present within the slope materials. Solute transport will be accomplished by earth slides was largely achieved in a
 Slope processes and forms 179

Table 7.7 Relative importance of slope A more useful basis for understanding the relationship
proces es at Kiirkevagge, between slope processes and slope form is the idea of
Sweden
detachment control. All material forming hillslopes can be
Rapid movements (%) regarded a having a range of detachability with re pect to
Rock falls 7 particular hillslope processes (Fig. 7.18). At one end of this
Debris avalanches 8 continuum, surface materials may be completely undetach-
Earth slides 34
able by a particular process, and erosion is limited by the
Total 49 rate at which weathering creates particles which are
Slow movements (%) detachable by the process. At the other end of the con-
Solifluction 2 tinuum surface materials may, in effect, be infinitely
Scree movement 1 detachable by a particular process and the rate of erosion on
Solute loss 48
the hillslope is controlled by the transporting capacity of
Total 51 this process.
The fashioning of slope form by slope processes can be
Source: Data from A. Rapp (1960)
Geograjiska Annaler 42, 73-200. influenced by both external conditions and by the nature of
the slope system itself. An example of the role of an
single major storm event; this underscores the problem of external factor is provided by the situation where the
assessing the true long-term significance of geomorphic moisture content of slope materials is significantly in-
events which have long recurrence intervals. Secondly, the creased by a period of intense rainfall. As the moisture
impressive figure for solution initially suggests the import- content rises and pore pressures increase, the effective
ance of chemical weathering in ground loss on slopes in the normal stress acting on the slope will fall until a point is
area. But a more appropriate assessment of its significance reached where shear stress exceeds shear strength and
would have to make allowance for any decrease in the bulk failure occurs. A similar situation occurs when surface
density of the regolith associated with the weathering runoff is established on a slope since the depth and
reactions leading to the release of solutes. velocity of flow must attain a critical level before the
The rates at which different slope processes operate on detachment and transport of slope materials can begin.
an individual slope vary greatly, but there is also a con- Both these cases provide examples of slope behaviour
siderable variation in the activity of particular processes being influenced by thresholds. And since in both cases it is
under different morphoclimatic regimes (Fig. 7.17). The external (climatic) factors that are promoting the changes in
latter arise from the influence of climate on erosivity and its slope conditions, these are instances of the operation of
indirect effects on the erodibility of slope materials, extrinsic thresholds.
especially in terms of the abundance of vegetation and A contrasting situation occurs where internal changes
moisture content. We consider the rates of overall land- within slope materials themselves lead to an adjustment of
scape denudation further in Chapter 15. slope form. Where a slope is being actively undercut and
thereby maintained close to its threshold angle of stability,
it will experience successive periods of progressive weather-
7.6 The slope system ing and weakening of slope materials interspersed with
The ability of processes on hillslopes to fashion slope form episodes of slope failure when the regolith becomes
is determined by the capacity of these processes to trans- incapable of sustaining the existing slope angle. In this case
port the available slope material. In 1877 G. K. Gilbert slope failure occurs as a result of the breaching of an
identified two kinds of situation for slope development. In intrinsic threshold since there are no changes in external
one, erosion is limited by the rate at which material is made variables. The slope failure itself exposes less-weathered
available through weathering. In the other, there is no material which is able to maintain a higher slope angle, and
effective limit to the availability of weathered material and failure will not occur again until the shear strength of the
slope erosion is therefore controlled by the capacity of the slope materials has been sufficiently reduced by further
transport processes. This distinction has subsequently weathering. It is, of course, possible that changes in
developed into the idea that slopes can be viewed either as external variables could also promote a slope failure, but
weathering-limited or as transport-limited. But how the important point is that this is not necessary for a
valid is this widely accepted categorization? It misleadingly discontinuous adjustment in slope form to occur.
implies that on slopes where there is a ready supply of
erodible particles transport processes will always operate at
7.6.1 Slope form
capacity. Moreover, it underplays the contribution of mass
movement processes in determining slope form and over Slope form is most often represented in terms of two-
emphasizes the role of water erosion on slopes. dimensional slope profiles. Slope profiles run from drain-
180 Exogenic processes and landforms
Soil Creep and Solifluction Slope Wash and Rainsplash
1000 -r---'---l....-----l.---'----'---l--....L...-----r
both up to
500 500 10000 mm ka- 1
Solifluction Key
steep and
gentle
slopes
200 200 gentle
slopes
100 100

I 50
<1l
'I 50
<1l
.:£ .:£
E E
E E
20 20

10 10

5 5

2 2

1 P/M TM TC M Ts-A SH TW-O HT 1 P/M TM TC M TS-A SH WO HT

Morphoclimatic Regime
Morphoclimatic Regime

Solute Transport and Loss Cliff and Slope Retreat

1000,----'---'-------''-----'----L----'-----'---r 106 -r---'---'----'-------''-----'----L----'------'------'--..,

500 Key
Marine

. o·
..
:.: -:: limestones cliffs

200
...~~. both

[
. ....
100
0.:
00
siliceous
rocks
"
' ....
o

'. '."

~ :OO:::S{:}:
I
<1l
50 .:£
E
E 20 ~·.o·~ '.~' ... ~ •• E
oo:~ 00000
~ (,.~ ~ 0 0 0 0 0
103 .,.......,....,......,
. . "<?' 0 0 0 0 0 0"""- 0
0 o o
~ .~
0 0 0
10°000° 0 0 0
o 0 0
0.
o 0 0 0 0 o ·0'
o 0 0 0 0 o • [ . .0 . 0
000000 • 0 .o. .:....~. _
0 0°00 00 • 0 0 000
5 000 0°°0 o • 0 000
o
.. 0 0 0
°• °
o 0
o

00. .°
o
°
0

• 0 0

.
o 0 0
o • 0
o 0 •
o 0 •
o 0 • o • 0
0 0 o 0 0
2 o 0
0
0
·00 0
• 0
......

o 0 0
o 0
o •
0 0
0
101 .L--.-----r--,---,.--__r----,-----,-~-~-...J
P/M TM TC M TS·A SH TWO HT P/M TM TC M TS·A TA SH TW·O HT
Morphoclimatic Regime Morphoclimatic Regime

Fig. 7.17 Rates of different slope processes in different morphoclimatic environments based on short-term field measurements. Key to
morphoclimatic regimes: PIM - periglacial and montane environments: TM - temperate maritime including eastern USA and western
Europe; TC - temperate continental including the humid interior of the USA and eastern Europe; M - Mediterranean and regions with
similar climate elsewhere such as southern California; TS-A - tropical semi-arid; TA - tropical arid; SH - subtropical humid
environments such as the south-east USA; TW-D - tropical wet-dry; HT - humid tropical. (Modified from A. Young and I. Saunders
(1986) in A. D. Abrahams (ed.) Hillslope Processes. Allen and Unwin, Boston, Fig. I.I, p. 7.)
 Slope processes and forms 181

WEATHERING - TRANSPORT-
LIMITED LIMITED
CASE CASE

TYPE OF
t Coarse fragments Water-stable Crusted
t
Cohesion less
SLOPE Bare on a rocky soil fine-grained fine-grained
MATERIAL rock substrate aggregates soil soil
I I
o DETACHABILITY co

Fig.7.18 The detachability continuum for slope processes. (After A. 1. Parsons (1988) Hillslope Form. Routledge, London,
Fig. 7.1, p. 108.) .

~ DIAGRAMMATIC REPRESENTAT IVE PROFILES

( a ) SOUTH AUCKLAND
GREYWACKE LANDSURFACE ~

(b)· IGNlt~BRJTE VALLEY WITH PUMICE ~

AGGRADATION SURFACE - VOLCANIC PLATEAU ~

(c) SOUTH AUCKLAND 1 2

ASH-MANTLED LANDSURFACE

INTERFLUVE

-i r- - _::--r
1 SEEPAGE SLOPE
2
C~~~~~
SLOPE
3
~ 2(2 )

0°_1° : 2°_4°
.
--1~ : "l
Modal slope ·angles
FALL FACE
(minimum angle 45°
normally over 65°)

TRANSPORTATIONAL MIDSLOPE
(frequently occurring ® INDICATES MOVEMENT IN
angles 26°_35°)
A DOWNVALLEY DI RECTION
" 5
"~ COLLUVIAL
ARROWS INDICATE DIRECTION AND + -~ FOOTSLOPE ALLUVIAL
RELATI VE INTENS ITY OF MOVEMENT OF :'OPRO../7........ ~_ 6 TOESLOPE
WEATHERED ROCK & SOIL MATERIALS so/[7-4,("I'- ~_
BY DOMINANT GEOMORPHIC PROCE.SSES rOt? lO";-1- - -=
~"'Iol'll.ii'iir-
: + 0,-
.... 0: ....
-o: .... z
....
....
I., u t- UJ
l.i.J_<t:2:
....
....u
V'H-3W
11>0: :> -0:
.... w 0 0:

~
u:> ::E:
o
0: _....
0..:':0
.... II>

U
i 0..
_ 0 .... W •
........ u .... z
........ -0: 0:0
z-o:,,- u-
w_o:
<.!lU::>
0011>
o II> en
.. i
_ 0:
.... 11>::> 00
0.. -0: II> VI"-

3 4
PREDOMINANT CONTEMPORARY GEOMO~PHIC PROCESSES

Fig. 7.19 A theoretical slppe profile containing a comprehensive range of slope units related to particular types of slope processes.
(From 1. B. Dalrymple et al. (1968) Zeitschrift fUr Geomorphologie 12, Fig. 1, p. 62).
182 Exogenic processes and landforms

age lines up the steepest slope to drainage divides. They are Profile Form

usually measured using simple surveying equipment, such RECTILINEAR CONVEX CONCAVE
as an Abney level or clinometer, which provides slope
angle measurements over a known distance on the ground.
Slope profiles can be subdivided into individual com-
ponents or units consisting of convex and concave
elements and straight, or rectilinear, slope segments.
Slope forms vary enormously, but in many cases they
comprise an upslope convexity leading down to a recti-
linear main slope which terminates in a basal concavity.
The main slope can consist of either a single segment or a
more complex sequence of segments at different angles. In
some cases bare rock is exposed, most often in the upper
part of the main slope, and this part of the slope is termed a
free face. Where active vertical or lateral undercutting is
present at the slope foot the basal concavity will be absent.
Figure 7.19 illustrates the complete range of slope units that
can be encountered, although very few individual slopes
contain all these components, and not all the units always
occur in the sequence illustrated. This diagram also indi-
cates how different slope processes tend to predominate on Fig. 7.20 The nine possible shapes of three-dimensional
different slope units. hillslope forms. (Modified from A. f. Parsons (1988) Hillslope
Although slope form is conventionally thought of in two- Form. Routledge, London, Fig. 2.5, p. 16.)
dimensional terms, it must be remembered that slopes are
components of the three-dimensional surface that consti-
100
tutes the landscape. The plan form of slopes is important
because contour curvature controls the routes taken by
water, sediment and solutes moving downslope (Fig. 7.16). Cfl
OJ 80
Since slope units can be straight, convex or concave in plan ~
Cl
as well as in profile there are nine possible three- OJ
Cl
dimensional slope forms (Fig. 7.20).
Numerous factors control slope form, but since there are
...c 60
OJ
important contrasts between those influencing rock slopes "'0

and those affecting soil-mantled slopes we will consider '"

~ 40
each separately. OJ
c-
o
7.6.1.1 Rock slopes CIl 20
.:.!
(,)
Rock slopes generally lie towards the weathering-limited o
cr:
end of the continuum of detachability, their form being
controlled by the weathering resistance and shear strength o
of the rock rather than by the activity of transport pro- 40 50 60 70 80 90 100
Rock Mass Strength (r)
cesses. The detailed form of rock slopes depends on varia-
tions in their rock mass strength as defined in Table 7.2 (see Fig. 7.21 Relationship between rock mass strength (r) and the
Section 7.1.2). A rock slope may therefore have several gradient of rock slopes based on over 250 measurements.
units but each unit will tend towards an angle which is in (Modified from A. D. Abrahams and A.f. Parsons (1987) Earth
Surface Processes and Landforms 12, Fig. 2, p. 634.)
equilibrium with rock strength on that part of the slope
(Fig. 7.21). We would expect these equilibrium angles to be
sustained through time for as long as the rock strength rock slopes are generally confined to coasts and regions of
characteristics remain constant. In rock types with a very high local relief where steep slopes associated with active
high rock mass strength vertical, or even overhanging, free basal undercutting allow the removal of weathered material
faces can develop (Fig. 7.22). as rapidly as it is produced.
Rock slopes are not common in humid tropical morpho- Talus will accumulate below rock slopes if basal removal
climatic environments because rock strength is usually occurs at a slower rate than debris is supplied. Moreover,
significantly reduced by weathering. In such environments the thickness of talus that accumulates reflects the balance
 Slope processes and forms 183

compound slopes comprising a series of free faces separ-

ated by talus slopes.

7.6.1.2 Soil-mantled slopes

Rectilinear units on soil-mantled slopes are developed by
those mass movement processes which transport debris
downslope to a relatively constant depth. Extensive straight
main slopes are common where stream incision is active.
The slope is steepened by basal undercutting until a
shallow debris or earth slide restores the slope to the angle
at which it is stable. As mentioned earlier (see Section
7.1.1) the gradient at which failure occurs is termed the
threshold angle of stability (alternatively, limiting angle of
stability) and is related to the shear strength of the slope
material. Numerous field measurements of the angle of
straight slope segments appear to show that some angles
are rather more frequent than others in the landscape (Table
7.8). These peak angle frequencies are most clearly evident
when the data for particular lithologies are examined (Fig.
7.23) and appear to be linked to the threshold angle of
stability of different kinds of slope materials.
Convex slope segments most frequently occur on slope
crests. The cause of this slope convexity was attributed to
soil creep by G. K. Gilbert in a paper published in 1909. He
argued that if the interfluves between drainage channels are
to be lowered there must be increasing rates of transport
of material at increasing distances from the divide. Since
rates of soil creep increase with gradient, but not with
distance from the divide, the increasing rate of transport
can only be accomplished by a downslope increase in slope
Fig. 7.22 Overhang (slope gradient> 90°) developed in a rock angle, thereby giving rise to slope convexity. This argu-
slope in Clarens Sandstone, a thick unit at the top of the Karoo
ment also applies to rainsplash erosion and solifluction.
sedimentary sequence in southern Africa.
Whereas soil creep is probably the main cause of the
convex divides so common in humid temperate environ-
ments, rainsplash erosion and solifluction may equally be
between the supply and removal of material through time. responsible for convex crests in, respectively, semi-arid and
In their initial stage of development talus slopes tend to be periglacial regions.
concave in form. This is due in part to the rolling and Concave slope segments are normally associated with
bouncing of debris falling from the rock face over the talus either slope wash or, as already mentioned, the deposition
surface, although slumping and sliding in the upper part of of talus. In the case of slope wash there is an increase in
the talus slope can also contribute to the development of a
concave profile. As the talus slope builds up and the rock Table 7.8 Typical threshold slope angles for various types of slope
face becomes progressively buried the distance that debris material
can fall from the rock face becomes gradually less. Falling
rock fragments consequently have, on average, less kinetic SLOPE ANGLE (0) SLOPE MATERIAL

energy and are less likely to move beyond the talus toe, so 43-45 Jointed and fractured rock that is virtually
the talus slope becomes gradually straighter. Talus slopes cohesionless but with a high packing density
formed of fine material can also be significantly modified 33-38 Same material as above but with looser packing
by overland flow. 25-28 Taluvial slopes in which high pore-water pressures
The form of rock slopes is influenced a great deal by the can be attained
lithological and structural properties of the rock. Tensional 19-21 Sandy material
joint systems generated in response to pressure release can 8-11 Clays
induce a degree of slope convexity (Fig. 6.21), while Source: Based on M. A. Carson and M. J. Kirkby (1972), Hills/ope Form
alternating beds of varying resistance can give rise to and Process. Cambridge University Press, Cambridge, pp. 183-184.
184 Exogenic processes and landforms

10 20 30 40 50 60 over a lower gradient. Slopes dominated by the effects of

slope wash are therefore concave since they become
A progressively less steep downslope.
Field measurements of rates of various slope processes
can be incorporated into mathematical models which
predict the approximate equilibrium form that should arise
given the operation of specific processes. One approach
B b~Q which has been widely applied uses a continuity' equation.
This expresses the transporting capacity of the slope as a
function of slope gradient and distance from the drainage
divide, and indicates the form which the profile will tend to
approach over time (Fig. 7.24). Since this equilibrium
c bill
profile depends on the processes operating and not the
initial slope geometry it is described as a characteristic
form (see Section 18.2.3).

o 1-0------

0·8

0·6

E
0·4

0·2

0·1 0·2 0,3 0·4 0·5 0·6 07 0·8 0·9 ,.0

Fig. 7.24 Characteristic slope forms for different slope

processes. (ModifiedfromM.J. Kirkby (1971) Institute of British
Geographers Special Publication 3, Fig. 5, p. 26.)

10 20 30 40 50 60 In the absence of basal undercutting many soil-mantled

slopes are convexo-concave in profile as the dominance of
Slope Angle (degrees)
rainsplash or soil creep in the upper part of the slope gives
Fig. 7.23 Frequency distribution of angles of straight slopes in way at some point downslope to the prevalence of slope
six areas illustrating the presence of threshold slopes: (A) slopes wash. Where active stream incision removes material from
in shale, Mesa Verde, Colorado, USA; (B) undercut (a) and
protected (b) slopes, Verdugo Hills, California, USA; (C) slopes the slope base as quickly as it is supplied the basal
in Queenstown Shale, Ontario, Canada, with rubble (a) or clay concavity is replaced by a rectilinear slope standing at, or
(b) regolith; (D) slopes in sandstone, Pennines and Exmoor, UK; close to, the threshold angle of stability of the slope
(E) slopes in Laramie Mountains, Wyoming, USA (unshaded material.
histogram denotes dry talus slopes); (F) slopes in clay-shale,
Derbyshire, UK. Thresh.old angles are divided into semi-frictional
7.6.2 Slope evolution
(arrowed) andfrictional. (After M. A. Carson (1976) in: E.
Derbyshire (ed.) Geomorphology and Climate, Wiley, London, The manner in which slope form changes through time
Fig. 4.11 , p. 123 r based on various sources.) provided a central focus for geomorphic research until the
1950s. Various models of slope evolution were proposed,
depth of flow downslope ·because the contributing area notably by W. M. Davis, W. Penck and L. C. King, which
increases away from the drainage divide. Consequently the formed a framework within which the development of the
velocjty of flow can be maintained at progressively lower landscape as a whole was considered. Davis presented a
slope" angles. In addition the continuous weathering and model of slope decline in which there is a progressive
sorting of particles as they move downslope means that decrease in overall slope angles through time as the rate of
the average grain size generally decreases away from the basal downcutting by streams decreases and slopes become
slope crest. Since a given load of fine material requires less mantled with weathered material of ever finer calibre which
power to transport it than an equivalent load of coarser can be transported across ever lower gradient slopes (Fig.
material the rate of sediment transport can be maintained 7.25(A».
 Slope processes and forms 185

A DAVIS Slope
decline

B PENCK AFTER DAVIS

C PENCK Slope
replacement

D KING Parallel

---~
retreat

-----~

Fig.7.25 Classic models of slope development through time: (A) Davis's model of slope decline; (B) Davis's misinterpretation of
Penck's model indicating parallel retreat; (C) Penck's model of slope replacement; (D) King's model ofparallel retreat.

Penck's model was originally presented in German, and §ince the 1960s have clearly demonstrated that the
it was introduced to the English-speaking world by Davis evolution of slopes is, in most cases, likely to be far more
who seriously misrepresented Penck's ideas. Rather than complex than implied in these classic models, and that the
advocating parallel retreat of the major part of a slope, as mode of development will depend on structural and
implied by Davis (Fig. 7.25(B», Penck in fact considered lithogical properties as well as the slope processes
that where rates of denudation are declining hillslopes evolve operating.
through a process of flattening from the base upwards (see A rock slope will experience parallel retreat if the
Section 18.1.2). In effect each part of the slope profile is strength of the rock mass remains constant and basal debris
replaced by a slope of lower gradient as it retreats, and is continuously removed, but lithological variations and
through this process of slope replacement a broadly con- climatic changes mean that over an extended period of time
cave profile is produced (Fig. 7.25 (C». there will inevitably be some change in slope profile form.
In rejecting Davis's assertioQ that slope gradients decline Parallel retreat is also likely to predominate in situations
through time, King pointed to the widespread occurrence of where a flat-lying resistant lithology overlies less resistant
escarpments, notably in southern Africa, which have appar- strata. Such situations are common where weakly resistant
ently experienced prolonged parallel retreat (Fig. 7.25 (D)). regolith or other types of unconsolidated deposits are
King argued that the free face retreats parallel to itself as overlain by duricrusts (Fig. 6.31), or where igneous
material is weathered and then removed by rill erosion, intrusions, such as dolerite sills, cap less resistant sedimen-
and a low angle slope, or pediment, grows at its base (see ta~y strata (Fig. 7.26). Once the resistant cap rock is finally
Section 14.1). Such slope retreat eventually creates isolated, removed by back-wearing, subsequent slope evolution may
residual hills, known as inselbergs. occur through slope decline.
Although they have been extremely important in in- Where rates of basal undercutting decline we might
fluencing previous views about landscape development (see expect a reduction in slope gradient as material can be more
Chapter 18) none of these classic models is based on detailed intensively weathered and the angle of threshold slopes
empirical observations. Moreover, they do not yield spe- is reduced. This is suggested by the presence of distinct
cific quantitative predictions as to how slope form may be groupings of threshold slopes, the number and angle of
expected to change over time. Studies of slope processes which apparently depends on the weathering sequence of
186 Exogenic processes and landforms

Fig.7.26 An example ofparallel slope retreat where a resistant dolerite sill caps less resistant Karoo sedimentary rocks, near
Steynsburg, South Africa.

the bedrock (Fig. 7.23). We can envisage slope develop- Since the 1970s much of the research generated on
ment in a well-jointed rock in which there is an initial hillslopes has reflected the growing links being developed
phase of scree formation, a second stage marked by the between engineers concerned primarily with questions of
production of taluvium stable at a lower angle, and finally slope stability, and geomorphologists attempting to apply
the development of a cover of colluvium with the lowest physical principles to slope behaviour in the natural land-
angle of threshold stability. Whether the associated changes scape. These links are evident in the pioneering book by
in slope gradient occur by an overall decline in slope angle Carson and Kirkby (1972) which remains a valuable guide
or by a replacement of individual slope elements by to the broad field of slope studies. Although it emphasizes
segments standing at a lower angle is probably dependent physical principles it is firmly focused on the problem of
on the nature of the predominant slope processes and the how real slopes develop. The excellent book by Parsons
way these interact with the slope materials. These factors (1988) provides a more up-to-date treatment of the factors
are summarized in Figures 7.27, 7.28 and 7.29 which controlling hillslope form. Finlayson and Statham (1980)
illustrate the probable course of profile development on also give a useful introduction to slope processes, while the
different lithologies and under contrasting climatic con- book by Selby (1982a) is particularly strong on rock slopes.
ditions. The profiles show the likely change in form for A clear appreciation of some of the maj~r issues currently
slopes initially subject to active basal downcutting, but being addressed can be gained from the ,,:olume edited by
which subsequently develop in the absence of active basal Abrahams (1986).
erosion. The books by Statham (1977) and Williams (1982)
provide an excellent background on the properties and
behaviour of slope materials, and Selby (1980) describes
Further reading the classification and application of rock mas~ 'strength to
Given the nodal position of slope studies in geomorphology slope analysis. The papers included in the two books edi!ed
it is not surprising that a large literature has been produced. by Brunsden and Prior (1984) and Anderson and Richards
 Slope processes and forms 187

Ar------..-----------:.........----------.

1.2 48 4b
./
./
./

B I - - - - -:------------=----.
I 1

Fig. 7.27 Slope profile development in massive sandstone. (A)

Semi-arid environment: (1) Slope development during initial
phase of active stream incision. The initial slope is assumed to be
vertical because slope processes act very slowly relative to the
rate of stream downcutting. Slope processes predominate Fig. 7.28 Slope profile development in strong, closely jointed
thereafter. There is no immediate replacement of the initial slope rock. (A) Semi-arid environment: (1) Slope development during
by another unit because loose sand grains which accumulate at the initial phase of active stream incision. Initial slopes are
the cliff base are easily removed by slope wash. (4) Modification assumed to be steep (45-75°) but not vertical since weakening of
of the stable main slope predominantly by retreat due to active the jointed rock mass accompanies stream downcutting and the
slope wash of loose material that is only poorly, if at all, bound rock mass effectively acts as a densely packed cohesionless
together by vegetation. Shortening of the main slope is due mostly system. (2) Slope at instant when active stream incision ceases.
to rainsplash. (5) Final elimination of rock wall leaves extensive Slope processes dominate thereafter. (3) The rock wall is
but gently inclined concave slope or pediment. (B) Humid transformed into a talus-mantled slope. It is assumed that the
temperate environment: (1) As for (A) but slope gradient is less removal offine debris from the talus slope by sUljace water
(20-30°) and the summit convexity is more fully developed. prevents replacement by a lower angle main slope. Consequently,
Lower main slope angle assumes rapid weathering of the bedrock the rock wall-talus slope retreats as one slope leaving behind a
under humid conditions and loss of cohesion of the surface soil concave basal slope or pediment. (4) Main slope progressively
cover. Development of convexity during stream incision is due to consumed by upslope extension of basal concavity predominantly
soil creep and rainsplash. (2) There is no immediate replacement through slope wash. (5) Further extension of basal concavity
of the initial slope by another unit (3) because the sandy soil- obliterates main slope. (B) Humid temperate environment: (1) As
mantled slope is assumed to be already at the angle of ultimate for (A). (2) As for (A). (3) Three phases of instability related to
stability, since once the rock has weathered to a sandy mantle the weathering sequence (bedrock-talus-taluvium) are indicated,
little further weathering and therefore change in shear strength is although there may be more. The rock wall (3a) is replaced by a
likely to occur. (4) Modification of main slope mainly through talus-mantled slope (3b) and this, in turn, is replaced by the
shortening as a result of both soil creep and rainsplash. Slope taluvial slope (3c). (4) Conversion of the taluvium tofiner
wash is much less effective than in semi-arid regions because of colluvium could result in further replacement or decline (3c-4)
the abundance of vegetation. Retreat is consequently much less, before the ultimate angle of stability of the regolith is reached.
although solution may be important in some instances. (5) (5) Upslope extension of basal concavity and downslope
Further shortening through downslope expansion of summit extension of summit convexity obliterates the main slope and
convexity produces a predominantly convex slope. (From M. A. produces a convexo-concave profile. (From M. A. Carson and
Carson and M. J. Kirkby, (1972) Hillslope From and Process. M.1. Kirkby (1972) Hillslope form and Process. Cambridge
Cambridge University Press, Cambridge, Fig. 15.7, p. 380, and University Press, Cambridge, Fig. 15.8, p. 380, and caption
caption based on same source, pp. 379-82.) based on same source, pp. 379-82.)

Browning (1973) assesses the 1970 Huascaran avalanche.

(1987) provide a very broad and detailed coverage of slope Slides are discussed by Crozier (1986), and Watson and
stability analysis, while Selby (1982a) and Crozier (1986) Wright (1969) describe the Saidmarreh landslide. Soil
present useful introductions to this topic. creep is considered by Kirkby (1967) and in the classic
The terminology of mass movement is a quagmire of paper by Gilbert (1909). .
confusion with different researchers using different terms Hillslope hydrology is comprehensively covered in
for the same phenomenon. Various classifications have Kirkby (1978) and reviewed concisely by Kirkby (1985a,
been proposed including those by Sharpe (1938) and 1988) while the movement of solutes on slopes is discussed
Skempton and Hutchinson (1969), but the categorization by Burt (1986). Morgan (1986) presents an accessible
used in this chapter is based on that of Varnes (1978). Innes treatment of soil erosion on slopes and Dunne and Aubry,
(1983) provides a detailed review of debris flows, while (1986) provide a detailed analysis of slope wash and rain-
Hsii (1975) looks at catastrophic debris avalanches and splash erosion. Data on slope process rates are presented by
188 Exogenic processes and landforms

A ,-.--. modelling of slope form are contained in Ahnert (1976,

!!
•I
I
2
1987), Armstrong (1987) and Kirkby (1985b), while
Douglas (1988) discusses the problems of applying such
models to slope development in real landscapes.

References
Abrahams, A. D. (ed.) (1986) Hillslope Processes. Allen and
I Unwin, Boston, and London.
I
·--1 I
Ahnert, .F. (1976) A brief description of a comprehensive three-
dimensional process-response model of landform development.
Zeitschrift fur Geomorphologie Supplementband 25, 29-49.
Ahnert, F. (1987) Process-response models of denudation at
different spatial scales. Catena Supplement 10, 31-50.
Anderson, M. G. and Richards, K. S. (eds) (1987) Slope Stability:
Geotechnical Engineering and Geomorphology. Wiley,
Fig.7.29 Slope profile development in a clay mass. (A) Semi- Chichester and New York.
arid environment: (1) Slope development during initial phase of Anderson, M. G., Richards, K. S. and Kneale, P. E. 1980. The role
active stream incision. The slope is steep as a result of true of stability analysis in the interpretation of the evolution of
cohesion together with capillary cohesion, and slightly concave threshold slopes. Transactions of the Institute of British
due to rill and gully erosion which accompanies sliding during Geographers NS 5, 100-12.
stream incision. (2) Slope at instant when active stream incision Armstrong, A. C. (1987) Slopes, boundary conditions, and the
ceases. Slope processes dominate thereafter. There is no development of convexo-concave forms - some numerical
immediate replacement of the initial slope by another unit as experiments. Earth Surface Processes and Landforms 12,
there is little change in shear strength or cohesion at this stage. 17-30.
(4) Continuing retreat of main slope with little change in shear Browning, J. M. (1973) Catastrophic rock slide, Mount
strength or cohesion. (5) Main slope is eventually eliminated to Huascaran, north-central Peru, May 31, 1970. American
leave a concave profile. (B) Humid temperate environment: (1) As Association of Petroleum Geologists Bulletin 57, 1335-41.
for (A) except that slope gradient is less (20-30°) and summit Brunsden, D. and Prior, D. B. (eds) (1984) Slope Instability.
convexity is more fully developed. The lower-angle main slope Wiley, Chichester and New York. ,
reflects the more rapid weathering of the clay under humid Burt, T. P. (1986) Runoff processes and solutional denudation
conditions and loss of cohesion. Slope convexity develops through rates on humid temperate hillslopes. In: S. T. Trudgill (ed.)
the action of soil creep and rainsplash. (2) As for (A). (3) 0 ne Solute Processes. Wiley, Chichester and New York, 193-249.
phase of slope replacement or decline may OCcur associated with Carson, M. A. and Kirkby, M.1. (1972) Hillslope Form and
a change from a temporary to an ultimate angle of stability. (4) Process. Cambridge University Press, Cambridge.
Continued reduction in angle of the main slope. (5) Downslope Carson, M. A. and Petley, D. 1. (1970) The existence of threshold
extension of summit convexity produces a predominantly convex hillslopes in the denudation of the landscape. Transactions of
profile. (From M. A. Carson and M. J. Kirkby (1972) Hillslope the Institute ofBritish Geographers 49, 71-95.
Form and Process. Cambridge University Press, Cambridge, Crozier, M. J. (1986) Landslides: Causes, Consequences and
Fig. 15.9, p. 381, and caption based on same source, pp. 379-82.) Environment. Croom Helm, London and Dover.
Dalrymple, 1. B., Blong, R. 1. and Conacher, A. J. (1968) A
hypothetical nine-unit landsurface model. Zeitschrift fur
Saunders and Young (1983) and Young and Saunders Geomorphologie 12,60-76.
(1986) while Leopold et al. (1966) and Rapp (1960) Douglas, 1. (1988) Restrictions on hillslope modelling. In: M. G.
provide examples of the detailed field monitoring of slope Anderson (ed.) Modelling Geomorphological Systems. Wiley,
processes, and Reneau et al..(1989) employ dated colluvial Chichester and New York, 401-20.
Dunne, T. and Aubry, B. F. (1986) Evaluation of Horton's theory
deposits.
of sheetwash and rill erosion on the basis of field experiments.
Turning to the slope system as a whole, the distinction In: A. D. Abrahams (ed.) Hillslope Processes. Allen and Unwin,
between weathering-limited and transport-limited move- Boston, and London, 31-53.
ment is clearly drawn by Carson and' Kirkby (1972), and Finlayson, B. and Statham, 1. (1980) Hillslope Analysis.
reassessed by Parsons (1988). Young (1972) provides a Butterworths, London.
useful survey of slope profile forms, while Dalrymple et al. Francis, S. C. (1987) Slope development through the threshold
concept. In: M. G. Anderson and K. S. Richards (eds) Slope
(1968) present a nine-unit landsurface model, and Parsons Stability: Geotechnical Engineering and Geomorphology.
(1979) demonstrates the importance of considering slopes Wiley, Chichester and New York, 601-24.
in three-dimensions as well as in profile. The development Gilbert, G. K. (1909) The convexity of hilltops. Journal of
of rock slopes is considered by Moon (1986) and Selby Geology 17, 344-50.
(1980, 1982b, 1987), while Carson and Petley (1970) apply Hsti, K. 1. (1975) Catastrophic debris, streams (sturzstroms)
generated by rock falls. Geological Society of America Bulletin,
the concept of threshold stability angles to slope develop- 86,129-40.
ment, an approach examined further by Anderson et al. Innes, J. L. (1983) Debris flows. Progress in Physical Geography
(1980) and Francis (1987). Examples of the mathematical 7,469-501.
 Slope processes and forms 189

Kirkby, M. J. (1967) Measurement and theory of soil creep. and Landforms 8, 473-501.
Journal of Geology 75,359-78. Selby, M. 1. (1980) A rock mass strength classification for
Kirkby, M. 1. (ed.) (1978) Hillslope Hydrology. Wiley, Chichester geomorphic purposes: with tests from Antarctica and New
and New York. Zealand. Zeitschrift fur Geomorphologie 24, 31-51.
Kirkby, M. J. (1985a) Hillslope hydrology. In: M. G. Anderson Selby, M.1. (1982a) Hillslope Materials and Processes. Oxford
and T. P. Burt (eds) Hydrological Forecasting. Wiley, University Press, Oxford and New York.
Chichester and New York, 37-75. Selby, M. J. (1982b) Rock mass strength and the form of some
Kirkby, M. J. (1985b) A model for the evolution of regolith inselbergs in the central Namib Desert. Earth Surface Pro-
mantled slopes. In: M. 1. Woldenberg (ed.) Models in Geomor- cesses and Landforms 7, 489-97.
phology. Allen and Unwin, Boston, and Chichester, 213-37. Selby, M.1. (1987) Rock slopes. In: M. G. Anderson and K. S.
Kirkby, M. (1988) Hillslope runoff processes and models. Richards (eds) Slope Stability: Geotechnical Engineering and
Journal of Hydrology 100, 315-39. Geomorphology. Wiley, Chichester and New York, 475-504.
Leopold, L. B., Emmett, W. W. and Myrick, R. M. (1966) Sharpe, C. F. S. (1938) Landslides and Related Phenomena.
Channel and hillslope processes in a semi-arid area, New Columbia University Press, New York.
Mexico. United States Geological Survey Professional Paper Skempton, A.W. and Hutchinson, 1.N. (1969) Stability of natural
352G, 153-253. slopes and embankment foundations. State-of-the-Art Report:
Moon, B. P. (1986) Controls on the form and development of rock Proceedings of the 7th International Conference on Soil
slopes in fold terrane. In: A. D. Abrahams (ed.) Hillslope Mechanics and Foundation Engineering, Mexico, 291-355.
Processes. Allen and Unwin, Boston and New York, 225-43. Statham, I. (1977) Earth Surface Sediment Transport. Clarendon
Morgan, R. P. C. (1986) Soil Erosion and Conservation. Press, Oxford and New York.
Longman, London and Wiley, New York. Varnes, D. J. (1978) Slope movement and types and processes. In:
Parsons, A. 1. (1979) .Plan' form and slope profiie form of hill- R. L. Schuster and R. 1. Krizek (eds) Landslides: Analysis and
slopes. Earth Surface Processes and Landforms 4, 395-402. Control. Transportation Research Board Special Report 176.
Parsons, A.1. (1988) Hillslope Form. Routledge, London and National Academy of Sciences, Washington DC, 11-33.
New York. Watson, R. A. and Wright, H. E. Jr (1969) The Saidmarreh
Rapp, A. (1960) Recent development of mountain slopes in landslide, Iran. Geological Society of America Special Paper
Kiirkevagge and surroundings, northern Scandinavia, Geogra- 123, 115-39.
fiska Annaler 42, 71-200. Williams, P.1. (1982) The Surface of the Earth: An Introduction
Reneau, S. L., Dietrich, W. E., Rubin, M., Donahue, D. J. and Jull, to Geotechnical Science. Longman, London and New York.
A. 1. T. (1989) Analysis of hillslope erosion rates using dated Young, A. (1972) Slopes. Longman, London and New York.
colluvial deposits. Journal of Geology 97, 45-63. Young, A. and Saunders, I. (1986) Rates of surface processes and
Saunders, I. and Young, A. (1983) Rate of surface processes on denudation. In: A. D. Abrahams (ed.) Hillslope Processes.
slopes, slope retreat and denudation. Earth Surface Processes Allen and Unwin, Boston and New York, 3-27.
 8
Fluvial processes

8.1 Drainage basin hydrology Runoff, or more strictly basin channel runoff, is the quan-
tity of water which enters stream channels in a drainage
In Chapter 7 we looked briefly at the movement of water basin over a specified period of time, and can be determined
on and within slopes, but it is now necessary to extend our by a water-balance equation. This expresses runoff in
discussion of water movements to the scale of entire drain- terms of precipitation, losses through evapotranspiration
age basins. All water in rivers ultimately originates as pre- and changes in the amount of soil moisture and ground
cipitation, although there can be a considerable lag before water storage. In environments where much of the preci-
this water enters the fluvial system. The main components pitation falls as snow the water-balance equation is com-
of drainage basin hydrology are illustrated in Figure 8.1, plicated by having to take into account water released
and are represented schematically in terms of a system of during melting. A further output which has to be considered,
inputs, outputs, storages and transfers in Figure 8.2. in ,addition to runoff and evapotranspiration, is deep out-

p e c p a a n

~
E v a p a a n s p a a n
CHANNEL A
SURFACE STORAGE INTERCEPTION STORAGE STORAGE :
I
I
I
I
I
I
Soil moisture storage I
I
I
Aeration zone storage

Ground water storage

r:::=>
BASIN CHANNEL
RUNOFF

WATE R

Fig. 8.1 Primary storages and transfers of water within drainage basins.
 192 Exogenic processes and landforms

represent discharge as an equivalent mean depth of water

over the basin. A common index of this kind is obtained by
dividing the mean annual discharge by the drainage basin
area. This mean annual depth-equivalent measure of dis-
charge ranges from around 1000 mm for the Amazon River
to 177 mm for the Mississippi and 31 mm for the Colorado.
Overland In terms of absolute discharge the Amazon has by far the
flow
highest mean runoff in the world estimated at 230 000 m3 s-1"
not a surprising figure when we consider that its channel at
Obidos, 700 km upstream from its mouth, is nearly 2.5 km
n~~~~~f----------{ThroU9hfIOWH ----,
across and attains a depth of 60 m. The Amazon is followed
by the Zaire River with a mean discharge of around 40 000
m3 S-I. These totals can be compared with the mean dis-
r::§~~~~f----------( Intertlow }-r----,
charge from the Mississippi Basin which is a mere
18 000 m3 S-I; even the greatest ever recorded flood of the
Mississippi achieved a discharge of only 57 000 m3 S-I.

8.1.1.1 Spatial variations

In the great majority of river systems discharge increases
downstream as tributaries progressively add more runoff to
the trunk channel. In arid regions, however, losses through
evaporation and seepage can lead to a downstream decrease
@RCAPILLARY
RISE in discharge. At the global scale river discharge is closely
related to the balance between precipitation and evapo-
Fig. 8.2 Major elements of the drainage basin hydrological transpiration, and the highest discharges are achieved under
cycle.
humid climatic regimes (Fig. 8.3). Since precipitation also
flow from ground water. In most cases this happens so tends to increase with elevation, the greatest depth-equiv-
slowly that it can be ignored in calculating the water balance alent discharges are found in the mountainous regions of
for a basin. In limestone terrains, however, movements of south-east Asia and along the mid-latitude mountain ranges,
water at depth may have little relation to surface patterns of such as those of Chile and Scandinavia, which intercept
water flow and a complex hydrological system can result. rain-bearing weather systems.

8 .1.1.2 Temporal variations

8.1.1 Channel discharge
In environments with a clear seasonal variation in precipi-
Water discharge in river channels varies over both time and tation or temperature, stream flow will tend to vary fairly
space at a variety of scales. Changes in discharge through systematically throughout the year, being highest during the
time are represented by a hydrograph. In a discharge wet season or during the period when lower temperatures
hydrograph changes in discharge are plotted against time, reduce water losses through evapotranspiration. The term
but the direct measurement of stream discharge is time- river regime is applied to these average annual variations
consuming and more often discharge is estimated from a in discharge. River systems in humid wet-dry and monsoonal
stage hydrograph. This shows variation in stage, or elev- environments experience dramatic annual fluctuations in
ation of the water surface, through time. These changes in discharge as do basins experiencing heavy winter snow-
stage can be converted to a record of discharge variations falls. But not all channel systems maintain a constant flow
using a stage-discharge rating curve which plots changes of water. In intermittent stream channels flow occurs for
in discharge against changes in water depth. The curve is at least one month per year in response to seasonally
initially calibrated by taking discharge measurements in a generated runoff. In arid environments many channels carry
well-defined channel cross-section, often a specially con- only occasional flow after storms and are consequently
structed weir. Thereafter a stage recorder can be used to described as ephemeral stream channels.
monitor subsequent variations in discharge. Much attention in fluvial geomorphology has been
Although discharge is usually expressed in units of m3 S-I, focused on the factors that determine variations in the
represented in this way it is a function of drainage basin discharge of streams draining comparatively small drainage
area. If we wish to relate discharge to the rate of operation basins. It is only over relatively small areas that it is
, of geomorphic processes, such as the transport of sediments realistic to monitor changes in soil moisture and ground-
or solutes, over a basin as a whole, it may be appropriate to water storage, and to track the movement of water from
 Fluvial processes 193

~--------------

~~

p ~-4
6 2
10km
1~
1000 0 1000
km

Mean annual runoff (mm) . . > 1000 1lIlII600- ~ 400- ~ 200- [QZJ 50- c=J <50
1000 600 400 200
Fig.8.3 Global variation in mean annual runoff expressed as depth-equivalent discharge. (Based on Unesco (1978) Atlas of the
World Water Balance, Unesco, Paris.)

where precipitation arrives at the surface to the point at

which it leaves the drainage basin. Since it is the peak
flows resulting from storms which are often of most
significance in terms of the geomorphic work that they
accomplish, particular attentionha's been paid to how run-
off is generated by periods of heavy rainfall.
Hydrographs for small basins in humid environments
typically show a relatively constant discharge, termed base r----_...-.::: =_==----
Base flow
flow, punctuated by flood events which are represented by
a rapidly rising limb, a short-lived peak and a slowly
declining falling limb (Fig. 8.4). The precise form of such a
storm hydrograph depends on the intensity, duration and
Time
areal extent of precIpitation, and on the hydrological pro-
perties of the drainage basin. Fig. 8.4 Typical form of a storm hydrograph.
The mean frequency, or recurrence interval, of floods of
a specified magnitude can be estimated if there is a record
of stream discharge extending over many years, preferably equal or greater in magnitude than one of a particular size
several decades. Recurrence intervals of floods can be in the annual series is then given as (n + l)/r, where n is the
expressed in two ways. The usual procedure is to rank the number of years of record, and r is the rank order of the
highest discharge attained in each year to produce an flood of the specified magnitude. For instance, if we have a
annual series. The recurrence interval of a flood which is 50 a annual series the recurrence interval of a flood of the
194 Exogenic processes and landforms

same or greater magnitude than the tenth largest annual colation of water through the soil. By contrast, water that
flood in the series is (50 + 1)/10 = 5.1 a. enters stream channels soon after a storm is termed quick
The alternative means of expressing the recurrence flow and is generated in different ways depending on the
interval of a flood is through a partial duration series. nature of the topography, vegetation and soils in a basin.
This is constructed by listing and ranking all discharges In drainage basins lacking a significant vegetation cover,
which exceed a given magnitude over a specified period of such as in semi-arid and arid regions, or where vegetation
time. The recurrence interval for a partial duration series is
has been cleared and the surface compacted, soil infiltration
calculated using the same formula as for the annual series capacities may be sufficiently low to be exceeded by rain-
except that n represents the total number of flood events fall intensities during storms. In such circumstances
exceeding the specified magnitude. In a partial duration infiltration-excess overland flow will be generated once the
series there may be more than one flood event recorded in storage capacity of the surface has been exceeded; this is
some years, and none in others. A recurrence interval known as the partial area model of stream flow gener-
calculated using this approach gives the mean interval of ation. Large volumes of water can be transmitted to river
time between discharges of a size which equal, or exceed, channels by this means because of the high velocities
the specified flood magnitude. achieved by overland flow - up to 500 m h- 1 - and its
A flood which is at least equal in magnitude to the mean simultaneous occurrence over large parts of an individual
flood in an annual series is termed the mean annual flood. basin.
This has a recurrence interval of about 2.33 a; that is, the In well-vegetated basins soils characteristically have in-
maximum discharge reached each year will equal, or exceed, filtration capacities far in excess of likely rainfall inten-
the mean annual flood on average once every 2.33 a. The sities, yet high peak discharges are still observed during
most probable annual flood in an annual series has a storms. In such basins peak flows are generated from rain
recurrence interval of about 1.58 a, this being equivalent infalling directly on stream channels and from saturation
magnitude to the flood in the partial duration series which overland flow contributed from zones adjacent to channels
has a recurrence interval of 1 a. In general, in alluvial where the soil has become completely saturated. Areas
channels the most probable annual flood seems to be contributing saturation overland flow change during a
roughly equivalent in magnitude to bankfull discharge, storm as the saturated zones initially expand with the onset
that is, the discharge at which the channel is filled to of rain, and then contract once the rain stops. This is known
capacity. Floods of this magnitude appear to be of special as the variable source area model of stream-flow generation.
significance because, in many cases, they show a close An important distinction between infiltration-excess and
relationship with certain channel form variables and are saturation overland flow is that whereas the former is
therefore presumably important in shaping channel influenced by soil infiltration capacity the latter is related to
morphology. antecedent soil-moisture conditions and location within a
Rare, 'instantaneous' increases in channel discharge can basin.
occur as a result of the catastrophic failure of natural dams. A second important route for quick flow in humid
Such dams can be formed by glacial ice or moraines (see environments where thick soils are present is the lateral
Chapter 11), but may also be created by landslides. Many movement of water through the soil itself. This lateral
landslide dams fail soon after they are formed, usually by movement is most pronounced immediately above soil
overtopping. On occasion, however, massive landslides can horizons with a lower hydraulic conductivity, that is, a
dam major rivers and lead to the formation of large lakes. lower permeability or capacity to transmit water. Under
In 1911 an earthquake in the USSR triggered a 2.0-2.5 x such conditions a higher proportion of the water percolating
109 m3 rock avalanche which dammed the Murgab River. through the soil will be deflected laterally downslope. Such
This natural dam is 550 m high, nearly twice the height of subsurface flow, or throughflow, is generally much slower
the world's largest artificial dam. The failure of such dams than overland flow, and usually only reaches a maximum
can, of course, create enormous, short-term discharges. velocity of 0.4 m h- 1• In some soils, however, much more
rapid subsurface flow, at velocities of up to 200 m h- 1, can
occur through natural pipes formed by the rotting of roots
or the burrowing of animals. Such pipes can develop to
8.1.2 Runoff generation
form well-integrated conduits enabling the rapid downslope
Water contributing to stream flow can be divided into two transmission of water. Where there is a downslope thinning
types in terms of the rapidity with which it enters the of permeable soil horizons subsurface flow may emerge at
stream channel after a rainfall event. In most streams there the surface as return flow towards the slope base.
is a relatively low level of base flow which is maintained Although most subsurface flow is too slow to contribute to
between major rainfall events and this is supplied by the main peak flows accompanying storm events, it is
delayed flow from the ground water or from slow per-' important in priming areas adjacent to the stream channel
 Fluvial processes 195

and thereby enhancing the area of saturation overland flow

in subsequent storms.
Various properties of the drainage system and morpho-
1 1 1 1 1 1 1 1 PRECI PITATION
logical characteristics of the basin also influence the runoff
response to storms. These include the density of channels,
and the relief and shape of the basin (see Section 9.2). Such /profile of overland flow
factors affect the arrival time of runoff from tributaries at
the main channel. Material in suspension

8.1.3 Channel initiation

An obvious question to ask about stream channels is how
do they originate? Channels may be created on a newly
exposed surface or develop through the expansion of an
existing channel network, but in order to understand how 7=R
they are initiated we must look at the conditions under
which water flowing on a slope becomes sufficiently con-
centrated for channel incision to occur. It is also necessary
to establish how, once they are established, channels are 1": eroding stress R: shear resistance of soil surface
maintained and enlarged to form 'permanent' features in Xc : critical distance of overland flow
the landscape.
In Chapter 7 we saw how both surface and subsurface Fig. 8.5 Horton's model of surface erosion by infiltration-excess
flows converge in areas of contour concavity, and such overlandflow. (After R. E. Horton (1945) Bulletin of the
Geological Society of America 56 Fig. 14, p. 316.)
convergence is an important factor in channel development.
Moreover, it was also pointed out how infiltration-excess
overland flow can lead to the development of rills, although likely to be especially favoured where subsurfice pipes are
the precise mechanism that brings this about is far from present. Pipe networks appear capable of promoting
clear. Microtopography on slopes tends to disrupt sheet channel development either through roof collapse or by the
flow and promote the concentration of water movement and concentration of runoff and erosion downslope of pipe
rill formation. But such rill development can be counter- outlets. Although infiltration-excess overland flow is usually
acted by the lateral shifting of flow lines, or by rainsplash regarded as dominant in semi-arid regions, piping is also
erosion which tends to even out the surface. frequently important in such environments.
In the model put forward by R. E. Horton, before erosion Channels may also be initiated through spring sapping
by overland flow can occur on a hillslope it must reach a where ground-water flow is concentrated in more per-
critical depth at which the eroding stress of the flow meable zones within the bedrock. This encourages chemi-
exceeds the shear resistance of the soil surface (Fig. 8.5). cal weathering or dissolution of the bedrock which, in tum,
Horton therefore thought that a 'belt of no erosion' is pre- leads to an increase in hydraulic conductivity and a further
sent on the upper part of slopes because here the flow depth increase in the rate of water movement. This positive
is insufficient to cause erosion. Subsequent work has shown feedback mechanism can lead to channel development by
that some surface wash is possible even on slope crests, headward erosion as the water flow emerges at the surface
although here it does not lead to rill development because along a zone of enhanced permeability. Spring sapping is
the rate of incision is slow and incipient rills are infilled as likely to be especially active where a permeable lithology,
a result of rainsplash. such as sandstone or limestone, overlies an impermeable
Although the infiltration-excess overland flow model lithology, such as clay.
provides a reasonable framework for understanding channel
initiation in semi-arid and arid environments, in humid
regions we have to look for an alternative mechanism. In 8.2 Open channel flow
such areas channel initiation is less related to a critical Two opposing forces act on water flowing in an open
distance of overland flow than to the location of surface channel. The driving force is gravity which acts in a
and subsurface flow convergence, usually in slope con- downslope direction and is determined by gravitational
cavities and adjacept to existing drainage lines. In one case acceleration and channel gradient. The resisting force arises
rills have been seen to develop as a result of the sudden from friction both within the water body and between the
outburst of subsurface flow at the surface close to the base flowing water and the channel surface. The ability of
of a slope. In humid environments channel development is flowing water to entrain and transport material, and hence
196 Exogenic processes and landforms

its capacity to do geomorphic work, is essentially deter- LAMINAR FLOW

mined by the relationship between these two forces. How-
ever, before looking at the processes operating in natural
river channels we need to consider the fundamental charac-
teristics of the behaviour of flowing water.

.s1------------r--'7'-J
a.
8.2.1 Resistance to flow Q)

o~-----
Water is a fluid - that is, its shape is changed continuously
by the smallest applied external stress. This change in
shape is sustained for as long as the force is applied.
Resistance of a fluid to a change in shape is represented by
its viscosity. One type of resistance to deformation pro-
vided by viscosity arises from internal friction caused by Flow Velocity
cohesion and collisions between molecules as they move
past each other, and consequently this type of viscosity is
termed molecular viscosity or dynamic viscosity. Around TURBULENT FLOW
97 per cent of the energy of rivers is expended as frictional
heat generated by molecular impacts, leaving only about
3 per cent for the transport of sediment. As with all liquids
the molecular viscosity of water increases with a decrease
in temperature. Moreover, in the natural environment we
are concerned not only with pure water but also with water
containing fine sediment or dissolved constituents which
has a greater dynamic viscosity.

8.2.2 Laminar and turbulent flow

A fluid moving over a flat solid surface can act as a series
of thin 'layers' sliding over one another, the resistance to Flow Velocity
movement resulting from molecular viscosity. This form of
motion is described as laminar flow and while it is
relatively common in highly viscous fluids, such as lava Fig. 8.6 Schematic representation of velocity profiles for
flows, it is extremely rare for water moving in natural laminar and turbulent flow in a river channel.
channels (Fig. 8.6). In stream channels water movement
nearly always occurs as turbulent flow, that is, the velocity
of flow fluctuates in all directions within the fluid. Water is
constantly interchanged in eddies between adjacent zones wetted perimeter, the latter being the length of the
of flow, and local changes in velocity occur which work boundary along which water is in contact with the channel
against the mean velocity gradient and lead to a loss of (Fig. 8.7). In broad, shallow channels the hydraulic radius
energy. The resulting additional resistance to shear is is closely approximated by the flow depth.
termed eddy viscosity. In most channels there is a thin Conditions under which laminar or turbulent flow occur
laminar sublayer within which the velocity of flow are defined by a Reynolds number (Re) which is essen-
initially increases in a roughly linear fashion. Above this tially a dimensionless measure of flow rate. A Reynolds
zone the rate of increase in flow velocity is approximately number is calculated by multiplying the mean flow velocity
logarithmic (Fig. 8.6). and hydraulic radius, and dividing by the kinematic
Whether flow is laminar or turbulent is determined by viscosity which represents the ratio between molecular
the mean flow velocity, the molecular viscosity and density viscosity and fluid density (Box 8.1). In stream channels
of the fluid and the dimensions of the flow section. For the maximum Reynolds number at which laminar flow is
stream channels the dimensions of the flow section used are sustained is around 500. Above a value of around 2000
either the depth of flow or the hydraulic radius. The flow is turbulent, and at transitional values of between 500
hydraulic radius of a stream channel is defined as the cross- and 2000 elements of both laminar and turbulent flow are
sectional area of the flow in a channel divided by the present.
 Fluvial processes 197

Mean flow velocity Hydraulic jump

Cross-sectional
area

Supercritical Subcritical
(Froude No. >1) (Froude No. < 1)

Hydraulic drop

Discharge = Cross - sectional area x Mean flow velocity -_ ........... ~---

Hydraulic radius = Cross-sectional area/wetted perimeter

Fig.8.7 Properties of a stream channel relevant to flow

~
characteristics. Subcritical Supercritical
(Froude No. < 1) (Froude NO.>1)
Box 8.1 Reynolds number
Fig. 8.8 Transitions between supercritical and subcritical flow
resulting in a hydraulic jump or hydraulic drop.
The Reynolds number (Re) is defined as
vR
Re=-
v changes in both the depth and velocity of flow; in other
where v is the velocity of flow, R the hydraulic radius and vwords a given discharge can be transmitted along a stream
the kinematic viscosity. channel either as a deep, slow-moving, subcritical flow or
as a shallow, rapid, supercritical flow. Transitions between
subcritical and supercritical flow are determined by the
velocity of flow. A sudden transition from supercritical to
8.2.3 Flow regimes
subcritical flow is called a hydraulic jump and it gives rise
In' natural channels local variations in the depth of flow to a stationary wave and an increase in water depth (Fig.
caused by irregularities on the channel bed create waves 8.8). When a flow changes from subcritical to supercritical
which exert a weight or gravity force. The ratio of the mean there is a decrease in the water depth, a situation described
flow velocity to the velocity of these gravity waves, or as a hydraulic drop. Sudden transitions in flow regime can
ripples, defines the Froude number (F) of the flow and occur where there is an abrupt change in channel bed form.
this index can be used to distinguish different flow states This is a common feature in mountain streams where there
(Box 8.2). When the Froude number is less than 1 the wave are large boulders or bedrock obstructions in the channel.
velocity is greater than the mean flow velocity and the flow In natural channels mean Froude numbers rarely exceed
is described as subcritical or tranquil. Under such flow 0.5 and supercritical flows are only temporary since the
conditions ripples propagated by a pebble dropped into a large energy losses associated with this kind of flow
stream can travel upstream. When the Froude number is 1 promote bank erosion and channel enlargement. This
the flow is termed critical and when it is greater than 1 it is erosion in tum leads to a lowering of flow velocity and a
described as supercritical or rapid. consequential reduction in the Froude number of the flow
The reason for the existence of these different types of (an excellent example of negative feedback). Froude and
flow is that changes in discharge can be accomplished by Reynolds numbers can be combined to specify four distinct
flow regimes (Fig. 8.9). As we will see in Section 8.2.4 the
Box 8.2 Froude number Froude number associated with a particular flow regime has
important consequences for bedforms in sandy, alluvial
channels.
The Froude number (F) is defined as
v
F= -
Vid 8.2.4 Velocity of flow
where v is the flow velocity, g the acceleration of gravity
and d the depth of flow. The velocity of stream flow is influenced by the gradient,
roughness and cross-sectional form of the channel. Natural
198 Exogenic processes and landforms

10 Box 8.3 Estimates of flow velocity

Subcritical
turbulent flow The Chezy equation defines the mean flow velocity (v) as
1-0
v=c{&
F < 1
E where R is the hydraulic radius, s the channel gradient and
0·1 C the Chezy coefficient which represents gravitational and
~
..... frictional forces .
a. F > 1
Q) The Manning equation defines the mean flow velocity (v)
£:)
as
~ lOO
0
u:::: V=
c n
co Subcritical
Q)
~ laminar flow where R is the hydraulic radius, S the channel gradient and
0·001
n the Manning roughness coefficient.

0·0001 L..-.-_ _. L . -_ _..L..-_ _-l...-_-----".....L.---"-_--l

which incorporates an index of channel bed roughness.

0·001 0·01 0·1 1-0 10 100
This Manning roughness coefficient (n) is usually esti-
Mean Flow Velocity (ms-') mated from tables (see Table 8.1), or by comparison with
Fig.8.9 Boundary values for Reynolds (Re) and Froude (F) photographs illustrating channels of known roughness.
numbers specifying distinctfiow regimes related to mean flow The use of the Manning equation is complicated by the
velocity and depth offlow. (Modified from 1. R. L. Allen (1970)
fact that resistance to flow varies with discharge and flow
Physical Processes of Sedimentation. Allen and Unwin, London,
Fig. 4.3, p. 126.) depth. The Manning roughness coefficient decreases as
flow depth increases up to bankfull discharge. Once chan-
nel capacity is exceeded the flow spreads over a much
channels almost invariably have rough surfaces which larger area and the Manning roughness coefficient increases.
induce significant frictional energy losses and cause a A further complication arises from the adjustments in bed
reduction in flow velocity especially close to the channel form, and hence channel roughness, that accompany
boundary. The direct measurement of stream flow velocity changes in flow regime in alluvial channels (Fig. 8.10).
is time-consuming and consequently empirical equations
have been developed in order to estimate mean flow
velocities (Box 8.3). In the Chezy equation velocity is 8.3 Fluvial erosion and sediment entrainment
estimated in terms of the hydraulic radius and gradient of
8.3.1 Erosion of bedrock channels
the stream channel, and a coefficient expressing the gravi-
tational and frictional forces acting on the water. The The reduction of the landscape through the action of fluvial
Manning equation is a more widely applied estimator processes can involve the incision of stream channels into

Table 8.1 Values of Manning's roughness coefficient for various types of natural channel

CHANNEL TYPE NORMAL VALUE RANGE

Small channels (width <30 m)

Low-gradient streams
Un vegetated straight channels at bankfull stage 0.030 0.025-4:l.033
Unvegetated winding channels with orne pools and
shallows 0.040 0.033-0.045
Winding vegetated channels with stones on bed 0.050 0.045-0.060
Sluggish vegetated channels with deep pools 0.070 0.050-0.080
Heavily vegetated channels with deep pools 0.100 0.075-0.150
Mountain streams (with steep unvegetated banks)
Few boulders on channel bed 0.040 0.030-0.050
Abundant cobbles and large boulders on channel bed 0.050 0.040-0.070
Large channels (width >30 m)
Regular channel lacking boulders or vegetation 0.025-0.060
Irregular channel 0.035-0.100

Source: Based on data in V. T. Chow (ed.) (1964) Handbook ofApplied Hydrology. McGraw-Hili, New York.
 Fluvial processes 199

BEDFORM FROUDE MANNING of corrosion are bedrock mineralogy, the solute concen-
NUMBER ROUGHNESS
COEFFICIENT tration of the stream water, the stream discharge and
A
Ripples 0·017 - 0·028
velocity of flow. Maximum rates of corrosion are achieved
where fast-flowing, undersaturated stream waters pass over
lithologies with a high proportion of reactive minerals; for
Weak boil
B Dunes with ~ instance, corrosion is an important process in bedrock
superimposed --, .....--. I I ---- F « 1
ripples
¥ ........ /I
channels in mountainous limestone terrains in humid
environments.
c
Dunes -~
--- BOil
-......J'") --..
'-.J F < 1 0·018- 0·035
A second mechanism is abrasion, or corrasion, and
consists of the wearing away or detachment of bedrock by
particles moved by the water flow. The particles involved
D
can be of any size that can be transported at prevailing flow
Washed-out velocities, and large boulders several metres across may be
F ~ 1
dunes
in motion in fast-flowing, deep river channels. The effec-
tiveness of abrasion depends on the concentration, hardness
and kinetic energy of the impacting particles and the
Plane bed
I I f =1 0·011 - 0·015 resistance of the bedrock surface. Since kinetic energy is
proportional to the square of velocity, rates of abrasion
increase rapidly as flow velocities increase.
Standing
wave
antidunes
~f>1 0·012- 0·016
A third mechanism involves hydraulic action, that is the
movement of water alone. One way this can occur is
through the detachment of loose rock fragments by the
G Incipient breaking
wave force or moving water. Another process which is almost
F ~ 1 0·012- 0·020
Antidunes
certainly more important, but which is poorly documented
in natural channels, is cavitation. This is a well-known
effect on ship propellers, dam spillways and other arti-
H
Chute and F > 1 ficial structures which are subject to rapid flow. An
pool acceleration of flow in a fluid causes a drop in pressure
which, if of sufficient magnitude, leads to the formation of
Fig.8.10 Bedforms in sandy alluvial channels in relation to air bubbles. Cavitation occurs when these bubbles implode
flow regimes expressed by Froude numbers. At low flow velocities and emit tiny jets of water at velocities as high as 130 m S-I.
ripples are formed (A), but as the flow velocity increases ripples
Such velocities can generate stresses sufficient to fracture
are transformed into larger forms called dunes (B and C), both
being out ofphase with waves on the water sUlface. With a solid rock. Although the high initial mean flow velocities of
further increase in velocity bed undulations are planed off, 10 m S-l' or so required for cavitation to operate are certainly
resistance to flow is lowered and sediment transport rates attained in some stream channels, especially in rapids and
increase (D and E). This is a transitional state between at the base of waterfalls, its effects usually appear to be
subcritical and supercritical flow. With a further increase in
concealed by features produced by abrasion.
velocity, supercritical flow gives rise to antidunes which because
they are in phase with standing waves at the water sUlface
present a low resistance to flow (F and G). Antidunes move
upstream since sediment is lost from their downstream side more 8.3.2 Sediment entrainment
rapidly than it is deposited. At the highest flow velocities fast - The majority of rivers do not cut directly into bedrock but
flowing shallow chutes alternate with deeper pools (H). (Based on
flow in alluvial channels formed in unconsolidated sedi-
D. B. Simons and E. V. Richardson (1963) Transactions of the
American Society of Civil Engineers, 128, Fig. 2, p. 289.) ments. These sediments may range in calibre from boulders
to clay-sized material. Alluvial channels are 'self-formed'
equilibrium or quasi-equilibrium landforms in that their
bedrock as well as the entrainment and downstream morphology arises from the mobilization, transportation
transportation of sediment. The erosion of bedrock chan- and deposition of sediment and represents an adjustment to
nels must be of considerable significance in mountainous prevailing hydrological and sedimentological conditions.
regions, and although the mechanisms involved are poorly Alluvial channels can generally adjust rapidly to changes in
understood three major processes appear to operate. the balance between the stresses generated by the flowing
Corrosion is the chemical weathering of minerals in water and the resistance of the channel bed sediments to
contact with stream water and the removal of soluble pro- movement. In this respect they differ significantly from
ducts downstream, and the weathering processes involved bedrock channels which can usually change only slowly
are discussed in Chapter 6. The key factors controlling rates and whose morphology is dominated by structural and
200 Exogenic processes and landforms

t-
u. Suspension
::i
Saltation

Rolling
~"="~

Fig.8.12 Schematic representation of the modes of sediment

/
transport in flowing water.
/,~­
Gravity Direction of easiest movement
Unless the flow is highly turbulent and energetic larger
particles move on a trajectory converging with the channel
bed at a low angle (Fig. 8.12). This kind of particle motion
Fig. 8.11 Schematic representation of the forces acting on a is termed saltation. When large numbers of grains are in
grain resting with others of similar size and shape on a channel motion under rapid flow conditions the ideal saltation
bed and subject to fluid flow. (After G. V. Middleton and 1. B.
trajectory is not attained because there are frequent
Southard (1984) Mechanics of Sediment Movement. Society of
Economic Paleontologists and Mineralogists Short Course No.3 collisions between particles. In this situation there is a
Fig. 6.1,p. 6-3.) concentrated dispersion of particles near the channel bed
dominated by interparticle collisions and deflections.
Larger particles which cannot be lifted from the channel
lithological controls. Sediment in alluvial channels includes bed may simply move across it by either rolling or sliding.
particles previously carried downstream as well as material Impacting saltating grain returning to the channel bed may
contributed directly from valley-side slopes and from bank help to precipitate this movement.
erosion. In addition to rolling, sliding and saltating, solid particles
The initial setting into motion of a solid particle in a may experience suspended motion in which their tra-
fluid is called entrainment and this occurs when the jectories are more irregular and more prolonged than for
stresses acting on a particle exceed the resisting forces. The saltation (Fig. 8.12). The weight of fine particles in true
resisting forces are the immersed weight of the particle in suspension is entirely supported by the upward pulses of
water acting normal to the channel bed and the constraining flow generated by eddies. Grain descending during
effects of neighbouring grains. The driving forces are the saltation may be temporarily buoyed up by upward move-
downslope component of the immersed weight of the ments in turbulent flows and this condition is more
particle and the fluid forces, the latter being of two types appropriately described as incipient suspension.
(Fig. 8.11). A particle on a channel bed is subject to a drag
or tractive force largely as a result of the difference in the
8.4 Fluvial transport and deposition
fluid pressure on its upstream and downstream sides. The
stress becomes larger as the velocity of flow increases, but
in turbulent flow temporary increases in velocity in eddies 8.4.1 Modes of fluvial transport
can generate local stresses three or four times the mean Material can be transported by rivers either as solid par-
value. Sll1all particles experience relatively low drag force ticles or in solution. Solute load, or dissolved load, which
since they are locate4 within the laminar sublayer. As the is derived largely from bedrock weathering (see Chapter 6),
flow velqcity increases this zone becomes thinner, and is dispersed throughout the flow. Solid load is of two main
particles are subject to markedly increased stresses as they types. Bed load, or traction load, encompasses all material
begin to protrude through the laminar sublayer and are rolling, sliding or saltating along the channel bed.
exposed to higher velocity turbulent flows and eddies. Suspended load is invariably of fine calibre and includes
Particles are al 0 subject to a lift force arising from the all particles prevented from falling to the channel bed by
acceleration in the flow of water over projecting grains. the upward momentum imparted by eddies within turbulent
This increase in velocity leads to a drop in pressure above flows. The finest fraction of suspended load, consisting of
the grain and is known as the Bernoulli effect. If this lift very small clay-sized particles, is termed wash load and is
force is sufficient to overcome the resistance to movement able to stay essentially in permanent suspension as long as
the grain will rise from the channel bed. As it does so the some flow is maintained.
lift force rapidly declines, the drag force quickly increases Solute and solid loads can be estimated by relating solute
and the particle is carried downstream in the water flow. and sediment concentrations to stream discharge (see
 Fluvial processes 201

Section 15.3.1). The concentration of solutes and sus- Box 8.4 Stream power
pended load can be measured from water samples collected
by inserting sampling devices into the stream at various Stream power (n), or the power per unit length of stream,
depths and at different stages of flow, although such is measured in W (watts) or J S-1 (joules per second) and is
devices also collect some saltating grains which are a com- defined as
ponent of bed load. These measurements can then be n= PwgQs
multiplied by discharge to provide an estimate of sus- where Pw is the density of water, g the acceleration of gravity, Q
pended sediment and solute transport rates. Bed load is discharge and s channel gradient.
much more difficult to estimate because most of the devices The energy available per unit area of the channel bed (w)
used interfere to some extent with the channel bed. can be related to stream power (n) to give the specific stream
Techniques employed have ranged from a variety of traps power (co):
to acoustic and pressure-difference recording apparatus. n _
co= - = TOV
The movement of very coarse bed material can be W
monitored by painting individual boulders and tracking where w is channel width, TO channel bed shear stress and Ii
their movement along the channel bed over a period of mean flow velocity. The specific power (co) of rivers in
time, while fluorescent dyes and radioactive tracers can be Britain varies over three orders of magnitude from ioo to
used to monitor the movement of finer material of a specific 1000 W m-2 in steep gradient, high runoff streams, to less
size or mineralogy. than 100 W m-2 in low gradient, low runoff streams.
Sediment in transit in river channels is subject to
abrasion and weathering. Although there is evidence for
significant rounding of pebbles as they are transported for transporting sediment. Stream power is at a maximum
downstream, it is less certain to what extent sand and finer, in streams with high discharge, high gradient and. large
material experiences attrition during fluvial transport. A hydraulic radius. Specific stream power is defined as stream
downstream reduction in mean' particle size is a widespread power per unit area of channel bed and equals bed shear
feature of river channels, but the great majority of this stress times mean flow velocity.
appears to be due to sorting (see Section 8.4.3). The concept of stream power can be applied to con-
ditions under which erosion or deposition will predominate
through the threshold of critical power. This is defined as
8.4.2 Sediment transport
the state in a specific channel reach when the power
Although the general principles whereby grains are set in available is exactly sufficient to transport the mean avail-
motion in a fluid are fairly well understood, the problem of able sediment load. If the available power is greater than
predicting the rate of particle movement in natural river that required to transport this mean sediment load then
channels under different flow conditions and with varying erosion and downcutting will occur; if it is less than that
sediment characteristics has proved extremely difficult. required, deposition will predominate.
Advances in this field have been achieved largely through In order to predict the overall rate of transport of sedi-
research employing laboratory flumes to measure rates of ment in a channel rather than the conditions which lead to
sediment movement under a range of flow conditions. In the entrainment of individual particles, it is convenient to
such experimental work the variables involved can be fairly use a readily measured average property of the water flow.
closely controlled, but the complexity present in natural One approach, pioneered by F. Hjulstrom, uses mean flow
channels means that it is not possible to extrapolate these velocity to define conditions of sediment entrainment and
laboratory findings directly to field situations. transport. Hjulstrom collated data from some 30 experi-
It is also possible to approach the problem of sediment mental studies relating flow velocity to sediment transport
transport from fundamental physical principles. A particle to construct the curves illustrated in Figure 8.13. These
lying on a channel bed possesses potential energy by virtue curves, which have been broadly confirmed by later
of its elevation above base level. Once set in motion a research, show the competence of different flow velocities
particle moves by expending kinetic energy and by over- - the competence of a flow being the maximum particle
coming frictional resistance. An important concept in size that can be entrained at a specific flow velocity. It is
relating fluvial energy to sediment transport is that of important to note that flow velocity in this case refers to the
stream power (Box 8.4). In order to transport sediment, mean velocity in the channel and not the much lower
work must be performed. Wark is defined as the product of velocity on the channel bed where the sediment is entrained. A
force and distance and power is the rate of doing work. major limitation of this approach is that there is no unique
. Stream power is therefore the power per unit length of mean flow velocity at which particles of a particular size
stream. In other words, it is the rate of energy supply at the are set in motion. The velocity on the channel bed is the
channel bed which is available for overcoming friction and more important controlling variable, and far a given
202 Exogenic processes and landforms

10 100

I"
<f)

E
' ....
........
1-0 ~/)""'"
Erosion
p
"";,,,,,,,,,,
,...-
-
:p
....1',.
. . .'!..!.I')........... V
~ ·,:!!?s........
'uo ..............l'1t ...... _ ~."..
.-' ",

Q) ............ ~:.!£.~;.'i- .-",'-

> 0·1
::
o
u:::
~0·01
Transport
/ Deposition Boundary Reynolds Number

Fig. 8.14 A Shields diagram showing the relationship between

/.11
Q) V
~ the dimensionless critical bed shear stress and the boundary
v tOCity Reynolds number. The scatter evident is a result of a wide range
0·001
of experimental conditions and the problem of deciding precisely
0·001 0·01 0·1 10 100 1,000 when a threshold is reached. Note that the dimensionless critical
bed shear stress is almost constant, at a mean value of around
Particle Size (mm)
0.06,jor a broad range ofparticle dimensions above a boundary
Fig. 8.13 Velocity thresholds for entrainment and deposition as Reynolds number of around 1.0. At lower boundary Reynolds
afunction ofparticle size. Once in motion a particle is not numbers the dimensionless critical bed shear stress increases
deposited until the flow velocity falls below the fall threshold progressively down to a value ofabout 0.3. This is thought to be
velocity; this is less than the velocity required to set it in motion due to the presence of a smooth boundary to flow with particles
because more force is required to entrain a particle than to keep lying entirely within the laminar sublayer. (Based on M. C. Miller
it in motion. Note that these curves relate to data for well-sorted et al. (1977) Sedimentology 24, Fig. 2, p. 511.)
material and do not indicate threshold velocities for poorly sorted
sediments. (After F. Hjulstrom (1935) Bulletin of the Geological
Institute, University of Uppsala 25.) difficult to apply to natural channels for a number of
reasons. First, they relate to well-sorted sediments, that is,
sediments of a relatively uniform grain size. In poorly
threshold bed velocity the mean flow velocity increases sorted, or heterogeneous sediments, small particles may be
with water depth. sheltered by larger grains from the drag stress exerted by the
An alternative approach uses the mean bed shear stress water flow. The extreme example of this is the production
necessary to initiate particle movement. This depends not of a channel armour where coarse grains cover the channel
only on particle size but also on the roughness of the bed and conceal the finer sediment below. Secondly, the
channel bed. The critical shear stress on the channel bed curves refer to smooth channels whereas most natural
can be represented in a dimensionless form and related to a channels are irregular. Thirdly, the curves assume flow to
boundary Reynolds number which defines the degree of be steady but in natural channels the velocity of flow is
channel bed roughness and is proportional to the ratio . often highly variable. Fourthly, the way particles are
between grain size and the thickness of the laminar sub- packed together can significantly affect the ease with which
layer. This relationship can be represented in a Shields they are entrained. Finally, experimentally or theoretically
diagram (Fig. 8.14) and shows a broadly similar form to derived sediment transport equations are especially difficult
the Hjustrom curve relating mean flow velocity to sediment to apply to gravel-bed channels. Sediment transport occurs
entrainment. at almost all flows in sandy, alluvial channels, but signifi-
The curves indicate that the particle size which can be cant transport in gravel-bed rivers only occurs at relatively
moved at the lowest velocity (about 0.2 m S-I) is roughly high flows, normally where bankfull discharge is approached.
0.25-0.5 mm (medium sand). Larger, heavier grains require Consequently, in channel beds composed of gravel or
higher threshold velocities to initiate movement, while silt boulders. there are relatively long periods of base flow with
and clay-sized material also requires higher v~locities little or no transport during which sediment can settle and
because such small particles are at least partially protected interlock thereby rendering it less susceptible to entrain-
by lying within the laminar sublayer. For clay-sized ment in subsequent high flows than would be expected
material cohesion between individual particles further from standard transport equations.
enhances their resistance to entrainment. Once set in
motion fine particles can continue to be transported in
8.4.3 Fluvial deposition
suspension even if the flow velocities fall significantly, but
coarser grains are more rapidly deposited when flow Just as there are threshold flow velocities for the entrain-
velocities are reduced. ment of particles of different sizes so there are thresholds
Although useful in indicating some of the fundamental for sediment deposition. The velocity at which a particle
controls on sediment transport rates, the relationships settles to a channel bed, known as its fall velocity, is a
evident in the Hjulstrom curves and the Shields diagram are function of both its density, size and shape and of the
 Fluvial processes 203

viscosity and density of the transporting fluid. Since vis- this topic include those by Graf (1971) and Yalin (1977),
cosity and density change with the concentration of while Bagnold (1966, 1977) demonstrates how general
sediment in a stream flow, deposition is not related simply physical principles can be applied to the problem of sedi-
to flow velocity. As flow velocity decreases the coarser ment transport. Bull (1979) introduces the idea of a
sediment begins to be deposited while the finer particles threshold of critical power for sediment transport and
remain in motion, and this differential settling of the deposition. Sundborg (1956) develops the approach pion-
material in transit gives rise to sediment sorting. eered by Hjulstrom of relating sediment entrainment to a
Conditions on a channel bed can change rapidly over threshold flow velocity, while the problems involved in
time and space, so whether sediment is entrained or de- defining a critical threshold parameter for particle
posited depends on local rather than average conditions. entrainment are reviewed by Miller et al. (1977).
Indeed it is possible for depositioY{ and entrainment to be Turning to more specific topics, sediment transport in
occurring simultaneously in the same channel reach and river bends, as opposed to straight channels, is examined by
coarse sediment can be deposited at the same time that finer Dietrich (1987), and the difficulties of predicting sediment
particles are being entrained. Nevertheless, as discharge transport rates in coarse-bed channels are discussed by
fluctuates between high and low stages, episodes of de- Baker and Ritter (1975), Carling (1983), Bathhurst (1987)
gradation during which erosion predominates will alternate and in numerous contributions in the book edited by
with periods of aggradation during which deposition Thome et al. (1987). Channel bed armouring is of particu-
prevails. lar significance in modifying sediment transport rates and
this phenomenon is discussed by Gomez (1983). Techniques
for monitoring sediment movement in natural channels are
Further reading described by Crickmore (1967) and Kennedy and Kouba
Fluvial processes are discussed in a number of texts on (1970) for sand, and by Sayre (1965) and Reid et al. (1985)
fluvial geomorphology including those by Knighton for bed load, while the paper by Ergenzinger and Conrady
(1984), Morisawa (1985) and Petts and Foster (1985), but (1982) illustrates an approach to the measurement of the
the pioneering work by Leopold et al. (1964) still remains overall rate of bed-load transport. Sediment deposition and
valuable as a clear introduction to the topic. A somewhat the complex relationships between erosion and deposition
more advanced treatment of processes in alluvial channels are examined by Colby (1964).
is provided by Richards (1982). Other useful material is to
be found in Richards (1987) and Schumm et al. (1987). References
The hydrology of drainage basins has generated an
Bagnold, R. A. (1966) An approach to the sediment transport
enormous literature, but the important concepts are sum-
problem from general physics. United States Geological Survey
marized in several general texts including those by Dunne Professional Paper 422-1.
and Leopold (1978) and Ward (1975) and in the book Bagnold, R. A. (1977) Bed load transport by natural rivers. Water
edited by Kirkby (1978). Numerous papers on specific Resources Research 13, 303-12.
aspects of drainage basin hydrology are to be found in the Baker, V. R. and Ritter, D. F. (197~) Competence of rivers to
periodicals Journal of Hydrology, Hydrological Processes transport coarse bedload material. Geological Society of
America Bulletin 86, 975-8.
and Water Resources Research. The initiation of erosion on Bathurst, J. C. (1987) Measuring and modelling bedload transport
slopes subject to infiltration-ex~ss overland flow is con- in channels with coarse bed materials. In: K. Richards (ed.)
sidered in the classic paper by Horton (1945), while more River Channels: Environment and Process. Institute of British
recent ideas on the role of both surface and subsurface flow Geographers Special Publication 18. Blackwell, Oxford,
in establishing networks of open channels are reviewed by pp.272-94.
Bull, W. B. (1979) Threshold of critical power in streams.
Jones (1987). Catastrophic discharges resulting from the Geological Society ofAmerica Bulletin 90, 453-64.
failure of natural dams are considered by Costa and Carling, P. A. (1983) Threshold of coarse sediment transport in
Schuster (1988). broad and narrow natural streams. Earth Surface Processes and
The book by Leeder (1982), although directed towards Landforms 8, 1-18.
sedimentologists, provides an excellent treatment of open Colby, B. R. (1964) Scour and fill in sand-bed streams. United
States Geological Survey Professional Paper 462D.
channel flow, a topic which is also well covered by Costa, J. E. and Schuster, R. L. (1988) The formation and failure
Knighton (1984) and Richards (1982). A detailed dis- of natural dams. Geological Society of America Bulletin 100,
cussion of bedform - flow interactions is to be found in 1054-68.
Simons and Richardson (1966). The erosion of bedrock Crickmore, M. J. (1967) Measurement of sand transport in rivers
channels is considered by Foley (1980). Leeder (1982) is with special reference to tracer methods. Sedimentology 8,
175-228.
also a good starting-point for acquiring a more detailed Dietrich, W. E. (1987) Mechanics of flow and sediment transport
understanding of the physical principles controlling sedi- in river bends. In: K. Richards (ed.) River Channels: Environ-
ment entrainment and transport. Advanced treatments of ment and Process. Institute of British Geographers Special
204 Exogenic processes and landforms

Publication 18. Blackwell, Oxford, pp. 179-227. Morisawa, M. (1985) Rivers: Form and Process. Longman,
Dunne, T. and Leopold, L. B. (1978) Water in Environmental London and New York.
Planning. W. H. Freeman, San Francisco. PeUs, G. and Foster, 1. (1985) Rivers and Landscape. Edward
Ergenzinger, P. and Conrady, 1. (1982) A new technique for Arnold, London and Baltimore.
measuring bedload in natural channels. Catena 9, 77-80. Reid, 1., Frostick, L. E. and Layman, 1. T. (1985) The incidence
Foley, M. G. (1980) Bed-rock incision by streams. Geological and nature of bedload transport during flood flows in coarse-
Society ofAmerica Bulletin 91 Part II, 2189-213. grained alluvial channels. Earth Surface Processes and
Gomez, B. (1983) Temporal variations in bedload transport rates: Landforms 10, 33--44.
the effect of progressive bed armouring. Earth Surface Pro- Richards, K. (1982) Rivers: Form and Process in Alluvial
cesses and Landforms 6, 235-50. Channels. Methuen, London and New York.
Graf, W. H. (1971) Hydraulics of Sediment Transport. McGraw- Richards, K. (ed.) (1987) River Channels: Environment and
Hill, New York. Process. Institute of British Geographers Special Publication
Horton, R E. (1945) Erosional development of streams and their 18. Blackwell, Oxford.
drainage basins: hydrophysical approach to quantitative mor- . Sayre, W. W. (1965) Transport and dispersion of labeled bed
phology. Bulletin of the Geological Society of America 56, material, North Loup River, Nebraska. United States
275-370. Geological Survey Professional Paper 433-C.
Jones, 1. A. A. (1987) The initiation of natural drainage networks. Schumm, S. A., Moseley, M. P. and Weaver, W. E. (1987)
Progress in Physical Geography 11, 207--45. Experimental Fluvial Geomorphology. Wiley, New York and
Kennedy, V. C. and Kouba, D. L. (1970) Fluorescent sand as a Chichester.
tracer of fluvial sediments. United States Geological Survey Simons, D. B. and Richardson, E. V. (1966) Resistance to flow in
Professional Paper 562-E. alluvial channels. United States Geological Survey Pro-
Kirkby, M. 1. (ed.) (1978) Hillslope Hydrology. Wiley, Chichester fessional Paper 422J.
and New York. Sundborg, A. (1956) The river Klaralven. A study in fluvial
Knighton, D. (1984) Fluvial Forms and Processes. Edward processes. Geografiska Annaler 38, 125-316.
Arnold, London and Baltimore. Thome, C. R, Bathurst, J:c. and Hey, RD. (eds) (1987)
Leeder, M: R. (1982) Sedimentology: Process and Product. Allen Sediment Transport in Gravel-Bed Rivers. Wiley, Chichester
and Unwin, London and Boston. and New York.
Leopold, L. B., Wolman, M. G. and Miller, J. P. (1964) Fluvial Ward, R C. (1975) Principles of Hydrology (2nd edn). McGraw-
Processes in Geomorphology. W. H. Freeman, San Francisco. Hill, London and New York.
Miller, M. C., McCave, 1. N. and Komar, P. D. (1977) Threshold YaHn, M. S. (1977) Mechanics of Sediment Transport (2nd edn).
of sediment motion under unidirectional currents. Sedi- Pergamon, Oxford and New York.
mentology 24,507-27.
This page intentionally left blank
 9
Fluvial landforms

9.1 The fluvial system predominates is usually a mountainous or upland region.

The zone dominated by deposition is generally located
Rivers and the landforms they create can be considered at along a coast and takes the form of a delta or lowland
an enormous range of scales. In Chapter 8 we focused on coastal plain. The depositional zone may, however, be
the microscale aspects of how sediment is entrained, located at the centre of an interior basin, and local sites of
transported and deposited by rivers. At this scale we can significant deposition may occur in the piedmont
examine the movement of individual particles on a river environment where rivers emerge at a mountain front. In
bed and the relationships between such factors as the the intervening zone in which transport predominates,
roughness of the channel floor, the velocity of water flow inputs and outputs of sediment may be roughly in balance,
and the rate of sediment transport. In this chapter we are at least over the medium term. Sediment may, however, be
primarily concerned with the way fluvial processes create stored in this zone for long periods of time before being
fluvial landforms. These range from the processes oper- removed downstream into the zone of final deposition.
ating in a single bend in a river, to the different channel
patterns arising from contrasting conditions of water flow,
sediment transport and channel gradient, and ultimately to
9.2 The drainage basin
the morphology of entire drainage basins . A drainage basin is an area within which water supplied
Rivers and streams can be simply defined as bodies of by precipitation is transferred to the ocean, a focus of
water flowing in an open channel. They have three import- internal drainage, such as a lake, or to a larger stream.
ant roles in landscape creation: they erode the channels in (Note that the term drainage basin is synonymous with
which they flow, they transport sediments and solutes catchment and, in the USA, with watershed. In the UK
provided by weathering and slope processes as well as by the term watershed refers to the drainage divide between
the other denudational agents of ice and wind, and they basins.)
produce a wide range of erosional and depositional land- For a number of reasons the drainage basin is the
forms. River systems are the primary agents of erosion, fundamental unit of fluvial geomorphology. Drainage basins
transportation and deposition in most landscapes, including are usually well-defined areas, clearly separated from each
many where surface water is not present for most of the other by drainage divides, within which surface or near-
time. Over much of the Earth fluvial processes determine surface flows of water and associated movements of
the overall form of the landsurface. sediment and solutes are contained. Since it is the transfer
Fluvial systems can for convenience be regarded as con- of material that causes changes in the elevation and form of
sisting of three main elements: a zone of sediment pro- the landsudace over time, drainage basins constitute the
duction, a zone of sediment transfer and a zone of sediment natural unit for the analysis of fluvially-eroded landscapes.
deposition. This categorization is, of course, over- The outlet of a drainage basin provides a very convenient
simplified because some erosion, transport and deposition point at which to monitor these movements of water,
occurs in all three zones; nevertheless, within each zone sediment and solutes. There are important exceptions to
one of these three processes is usually dominant. In large these generalizations - such as the partially subsurface flow
basins the upstream zone in which sediment production of water in limestone terrains which may be unrelated to
208 Exogenic processes and landforms

surface topography (see Section 9.6) - but they are valid

for the great majority of landscapes. Another important
property of drainage basins is their hierarchical nature; each
tributary in a drainage system has its own basin area con-
tributing runoff, and so larger basins consist of a hierarchy
of smaller ones.
A final feature of drainage basins which makes them
such important units of analysis in geomorphology is that,
following the pioneer work of R. E. Horton and A. N.
Strahler, many of their important properties can be
expressed quantitatively in a way which allows one basin to
be compared with another. Such quantitative description is
termed drainage basin morphometry and can be applied
to the areal and relief properties of basins as well as the
characteristics of their river channel systems.

9.2.1 Channel network characteristics

River systems are a type of network; that is, they consist of
a series of links which connect nodes. Networks can be
analyzed with respect to two main sets of properties: the
topological aspects of stream networks concern the inter-
connections of the system, whereas the geometrical aspects
involve length, area, shape, relief and orientation p~operties.
The basic element of stream networks is the stream Fig. 9.1 Schemes of stream ordering proposed by A. N. Strahler
segment, or link. This is a section of stream channel be- (stream order) and R. L. Shreve (stream magnitude) (see text for
tween two channel junctions or, for 'fingertip' tributaries, explanation).
between a junction and the upstream termination of a
channel. Stream order expresses the hierarchical relation-
ship between stream segments. It is a fundamental property the number of the next highest order. This number varies
of stream networks since it is related to the relative dis- slightly between different successive orders in a basin (that
charge of a channel segment. is, between first and second order, second and third order
Various systems of stream ordering have been proposed, and so on), so a mean bifurcation ratio for a whole basin is
but the two most frequently used are those of A. N. Strahler normally used. Where lithology is relatively homogeneous
and R. L. Shreve (Fig. 9.1). In the Strahler system a stream the bifuraction ratio is rarely more than 5 or less than 3.
segment with no tributaries is designated a first order Nevertheless, a value of 10 or more can be attained in
segment. A second order segment is formed by the joining highly elongated basins which can develop where there are
of two first order segments, a third order segment by the narrow, alternating outcrops of soft and resistant lithologies.
joining of two second order segments and so on. It is Various geometrical properties of stream networks are
important to note that with the Strahler ordering method defined in Table 9.1. Probably the most important of these
there is no increase in order when a segment of one order is is drainage density. This reflects a balance between
joined by another of a lower order. In contrast, the stream erosive forces and the resistance of the ground surface, and,
ordering system proposed. by Shreve defines the magnitude as a consequence, is closely related to climate and lith-
of a channel segment as the number of fingertip tributaries ology. Drainage densities range from less than 5 km km- 2 on
that feed it. As stream magnitude is closely related to the permeable sandstones, to extreme values of more than
proportion of the total basin area contributing runoff, it 500 km km- 2 on unvegetated clay 'badlands'. The role of
provides a good estimate of relative stream discharge for climate is indicated by the very high drainage densities in
small river systems. some semi-arid environments which appear to result from
Stream order as defined by Strahler has been appli~d to the prevalence of surface runoff and the relative ease with
numerous river systems and has been shown to be statisti- which new channels are initiated. Drainage density is
cally related to various elements of drainage basin mor- related to the length of overland flow (Table 9.1), the
phometry (Fig. 9.2). A widely used topological property of latter being approximately equal to the reciprocal of twice
stream networks is the bifurcation ratio. This is the ratio the drainage density.
between the number of stream segments of one order and Further aspects of channel network geometry include
 Fluvial landforms 209

'"~ drainage patterns and orientation. In the absence of

~
~
The number of streams of each order
structural effects tree-like patterns develop, but where
ci)
'0
approximates an inverse geometric series structural influences are strong a variety of drainage
in which the first term is the number of first
ci
z order streams. patterns occur. Drainage patterns are discussed further in
ci>
0
Chapter 16.
...J
Initial studies of stream networks indicated that fluvial
~
systems with topological properties similar to natural

/
0, systems could be generated by purely random processes.
C
ell
...J The mean lengths of streams of each order Although such random-model thinking has been extremely
E approximates a direct geometric series in
til
~ which the first term is the mean length of the influential in channel network studies, later research has
ci) first order streams.
ci> identified numerous regularities in stream network topo-
0
...J
logy. These systematic, as opposed to random, variations
c: appear to be attributable to various factors including the
ell
need for lower order basins to fit together, the sinuosity of

~
'ii
til
~ The mean gradients of streams of each order valleys and the migration of valley bends downstream, and
approximates an inverse geometric series in
~ which the first term is the mean gradient of the length and steepness of valley sides. Although they are
~
ci) first order streams.
most pronounced in large basins, such non-random
til
0
...J elements of stream network topology are also present in
3 4 small catchments.

/
til
ell
.{ The mean area of drainage basins of each
C
order approximates a direct geometric
.~ series in which thefirstterm is the mean area
CD of the first order basins.
9.2.2 Areal and relief characteristics
Cl
0
...J
Areal properties express the overall plan form and dimen-
sions of drainage basins, while relief properties express
Stream Order elevation differences (Table 9.1). The elongation ratio has
Fig. 9.2 Statistical relationships between stream order and important hydrological consequences because, in contrast
various drainage basin properties. These relationships are often to more circular catchments, precipitation delivered during
referred to as 'laws' of drainage basin morphometry, although it a storm in highly elongated basins has to travel a wide range
must be emphasized that they are only general trends around of distances to reach the basin outlet. The resulting delay in
which there may be significant variability.
the arrival of a' proportion of the storm flow consequently
leads to a flattening of the storm hydrograph (Fig. 9.3).
The' relief, or height differences, in a basin can be
expressed quite simply using maximum and minimum
Table 9.1 Some morphometric properties of drainage basins
elevation values. Local relief is the difference between
PROPERTY DEFINITION maximum and minimum elevations within a given area
(usually of limited extent). Relief is related to the slope and
Network properties stream gradients in a basin, and so indirectly has an influence
Drainage density Mean length of stream channels per unit area
on the rates of slope processes and sediment transport by
Stream frequency Number of stream segments of all orders per
unit area rivers. Applied to large basins, however, an index such as
Length of overland Mean distance from channels up maximum relief ratio (Table 9.1) will invariably conceal significant
flow valley-side slope to drainage divide variations in relief within the basin. A much more useful
Areal properties measure of relief can be produced by averaging the local
Circularity ratio Total drainage basin area divided by the area relief over cells of a given size across an entire basin, and
of a circle having the same perimeter as the such average relief estimates have been found to be closely
basin
correlated with fluvial denudation rates (see Section 15.5).
Elongation ratio The diameter of a circle of the same area as
the drainage basin divided by the maximum There is a close relationship between drainage density,
length of the basin measured from its mouth mean slope angle and relief (Fig. 9.4). If drainage density is
Reliefproperties constant and stream channels maintain a constant spacing
Basin relief Difference in elevation between the highest through time, an increase in local relief due to stream
and lowest point in a drainage basin incision must, of necessity, cause an increase in mean slope
Relief ratio Basin relief divided by the maximum length of angles in the basin. There is a limit to this effect, however,
the basin
as a progressive increase in slope angles cannot continue
Ruggedness number Basin relief multiplied by drainaged den ity
indefinitely. At a certain point rates of erosion will be so
210 Exogenic processes and landforms

RELIEF CONSTANT: DRAINAGE DENSITY VARIES

Bifurcation Time
ratio =2·08
DRAINAGE DENSITY CONSTANT: RELIEF VARIES

Fig. 9.4 Relationships between relief, drainage density and

Bifurcation
Time mean slope.
ratio = 12

can operate in conjunction with denudation to change the

areal distribution of elevation over time.

Fig.9.3 Schematic representation of the effect of basin shape on 9.3 River channels
the form of the storm hydrograph.
Three major types of river channel can be identified.
Bedrock channels are cut into rock. In general they experi-
ence gradual modification but retain their overall form for
high that interfluves will be lowered as rapidly as stream
long periods of time. Significant lateral shifting of channels
channels and local relief will attain a constant value.
may, however, occur where the bedrock is only weakly
Another approach to the description of drainage basin
resistant. In alluvial channels the bed and banks are
relief involves hypsometric analysis in which elevation is
composed of sediment being transported by the river. They
related to basin area (Fig. 9.5). Such analysis allows the
can undergo dramatic changes in form as weakly resistant
calculation of a hypsometric integral which summarizes
alluvium is eroded, transported and redeposited in response
the form of a drainage basin in a single value. Hypsometric
to changes in water discharge and sediment load, among
integrals lie between 25 and 75 per cent for most drainage
other factors. Semi-controlled channels are of intermediate
basins, high values indicating a relatively large proportion
type, being only locally controlled by bedrock or resistant
of land at high elevation within the basin and low values
alluvium. A semi-controlled channel will be stable where it
indicating a small proportion~ Thus basins in which stream
is cut into bedrock or resistant alluvium, but over time it
channels have incised deep valleys, leaving extensive areas
may migrate laterally into alluvium and be much more
of relatively high elevation, have high hypsometric inte-
responsive to changes in hydrological and sedimentological
grals. The idea that such basins might be modified by
variables.
denudation through time to a state where little uneroded
high land remains has led to the use of the hypsometric
9.3.1 Alluvial channels: plan form
integral as an index of the stage of development of a land-
scape. Although this may be true in a limited sense for Alluvial channels exhibit a great variety of plan form.
relatively small catchments, in large drainage basins Numerous channel patterns can be recognized but they all
tectonic uplift and subsidence as well as sea-level changes represent variations of just a few basic types. One key
 Fluvial landforms 211

A gradient to channel gradient provides an alternative defi-

nition. Sinuosity ranges from 1.0 for perfectly straight
-=-~I channels to around 3 for highly tortuous river courses.
I Channels with a sinuosity greater than 1.5 are usually
I
I described as meandering (although this is a purely arbitrary
A--_ _- _ ~ - Land surface
break-point). Even straight channels generally have a
I
I sinuous thalweg, that is, a line of maximum depth along
I
I the channel floor.
A second fundamental form that channels may assume
Base plane involves braiding. This represents the extent to which flow
in' the channel is divided by islands or bars, that is exposed
accumulations of sediment (Fig. 9.6). Islands are vegetated
B Contour and are relatively long-lived features, whereas bars are less
stable, being composed of unvegetated sands or gravels.
The degree of braiding is expressed quantitatively as the
percentage of channel length that contains islands or bars.
If a channel contains islands whose width is more than
Mouth three times the width of the water at mean discharge it is
J'\-----<::::<----.L-~

described as anabranching. The degree of anabranching is

expressed as the percentage of channel length occupied by
such islands.
Area a Area A A third type of channel pattern is termed anastomosing.
(entire basin) Anastomosing channels consist of distributaries which
branch and rejoin and have a superficial resemblance to
~ braided patterns. Braided channels, however, are single-
x = a x 100
c channel forms in which the flow is diverted around obstruc-
o
'';:;
A
co tions in the channel itself, whereas anastomosing patterns
>
Q) y = h x 100 consist of discrete, interconnected channels which are sep-
ijj
Ft Hypsometric integral
c
= % area under curve arated by bedrock or by stable alluvium. While braided
'iii
co channels are primarily depositional forms, anastomosing
aJ
co channels are essentially erosional in nature since material
~
o between channels is too resistant to be transported, except
.....o
by exceptional flows.
>!2.
o
Although many alluvial channels can be described as
stable in that they are not experiencing a dramatic change
% of Total Basin Area (x)
in form, some change is an inevitable element of the
Fig.9.5 The components of hypsometric analysis showing how behaviour of all alluvial channels since they are, at least
the area-elevation data are calculated (A) and plotted to produce partly, composed of material which is eroded or deposited
a hypsometric curve (B), The proportion of total basin elevation is as the stress exerted on the channel bed and banks by the
calculated for contours at appropriate intervals by dividing the
contour elevation by the maximum basin elevation (calculated flowing water changes over time. Such changes can occur
with respect to the basin outlet), The proportion of total basin in various ways (Fig. 9.8), including the downstream
area is calculated by dividing the area of the basin above each migration of bars, the gradual shifting of meanders and the
contour interval by the total basin area. The plot of these two rapid alteration of course through cut-offs or channel diver-
ratios (normally represented as percentages) defines the sion. A most dramatic example of the latter process, known
hypsometric curve for the basin. The hypsometric integral is the
area under the hypsometric curve as a percentage of the total as avulsion, occurred in China in 1851 when the mouth of
area defined by the two graph axes and can be simply estimated the Huang He shifted over 300 km to the north.
by superimposing it on graph paper and counting squares. Studies of alluvial channels from a wide range of
(Modified from A. N. Strahler (1952) Bulletin of the Geological environments have demonstrated that although the size of
Society of America 63, Figs 1 and 2, pp.1119 and 1120.) alluvial channels is controlled largely by the water dis-
charge flowing through them, the channel pattern and shape
property is sinuosity which represents the irregularity of are related primarily to the quantity and size of sediment
the channel course and is expressed as the ratio of channel being transported and the valley floor gradient. A genetic
length (measured along the centre of the channel) to valley classification proposed by S. A: Schumm identifies the three
length (measured along the valley axis). The ratio of valley fundamental channel patterns and rdates these to the nature
212 Exogenic processes and landforms

Fig.9.6 Braided channel of the Waimakariri,just north of Christchurch, South Island, New Zealand. (Photo courtesy S. 1. Smith.)

of the transported sediment among other factors (Fig. 9.9). Table 9.2 Classification of stable alluvial channels
Although the proportions of bed load and suspended load
TYPE OF BED LOAD AS CHANNEL CHARACTERISTICS
invariably change over time it is useful to distinguish be-
CHANNEL % OF TOTAL LOAD
tween suspended-load channels transporting less than 3
per cent of total sediment as bed load, bed-load channels Suspended <3 Wirlth-depth ratio <10; sinuosity
transporting more than 11 per cent as bed load, and mixed- load usually >2.0; rdatively gentle
gradient
load channels with 3-11 per cent of sediment being
Mixed load 3-11 Width-depth ratio 10-40; sinuosity
transported as bed load (Table 9.2). Suspended load usually 1.3-2.0; moderate gradient
channels are narrow and deep with a width-depth ratio of Bedload > 11 Width-depth ratio >40; sinuosity
less than 10. If the valley gradient is low the channel will usually <1.3; relatively steep
be straight (Fig. 9.9, pattern 1), but sinuosity becomes gradient
greater with an increase in gradient (Fig 9.9, pattern 3a). Source: Modified from S. A. Schumm (1977), The Fluvial System. Wiley,
Mixed-load channels have a lower width-depth ratio of New York, Table 5-4, p. 156.
10-40 and sinuosity ranges from 1.3 to 2.0 (Fig. 9.9,
pattern 3b). Even if the channel is relatively straight the
thalweg is nearly always sinuous (Fig. 9.9, pattern 2). Bed- 9.3.1.1 Meandering channels
load channels are straight and have a high width-depth Meandering channels vary in form, but a number of mor-
ratio of over 40. More than one thalweg tends to develop phological characteristics can be defined which are relatively
(Fig. 9.9, pattern 4) and where bed-load transport is very consistent for a large proportion of rivers (Fig. 9.10). These
high distinct bars form and a braided channel is created include the observation that meander wavelength is
(Fig. 9.9, pattern 5). Overall the mean flow velocity, the commonly about ten times channel width and about five
drag force on the channel bed and stream power increase, times the mean radius of curvature. Natural meanders
while channel stability decreases from patterns 1 to 5 (in rarely have a perfectly symmetric and regular form appar-
Figure 9.9). ently because of variations in channel bed material. In
 Fluvial landforms 213

Fig.9.7 Anastomosing channel of the Zaire (Congo) River in west-central Africa. The Zaire River is second only to the Amazon in
. annual discharge, and the anastomosing channel system in its lower reaches shown here is up to 15 km across. The area covered is
about 170 km across. (Landsat image courtesy N. M. Short.)

A B

CJ2!
Transverse Bar Shift Alternate Bar Shift
Neck Cutoff
Fig.9.8 Types of channel changes. (A) and (B) show the
downstream migration of bars or islands within channels; in (B)
o E alternate bars are moving in association with a shifting in position

~
--" of the deepest part of the channel, or thalweg, with the channel
"' ....7 -- -
'/ '/ ,,"
'\
"\ banks alternatively being protectedfrom, and exposed to, erosion
\ \ as the bars migrate; (C), (D) and (E) illustrate rapid changes in the
course ofmeandering channels,. (F) shows the establishment ofa
new course through avulsion. (After S. A. Schumm, (1985) Annual
Chute Cutoff Meander Shift Avulsion Review of Earth and Planetary Sciences 13, Fig. 4, p.ll.)
214 Exogenic processes and landforms

CHANNEL TYPE A A
Erosion
Suspended Load Mixed Load Bed Load
I
~ ~ t9
00
1~ :t .::
....J....J
i i
:c
i C/)

2 '2?-o- .::
:cC/) :t:
0
~ Qi '5
~
Z
a: .~ 3a
_"0 u
....J C

~
W a:: -"
ii5111 <.>Q)
~ a~~ ~~ z B A
~
....J

{i
~~
I ~
3b -~J
C/)

~~ -"'G
4,~
W
.&:;t9 :t:
Z I-C/)
B
z -t5 :)Qi ::I D c
« ~ U
I
U
w"o
a: c ..
Q)
Q)

'5
.&:;
\Q7~
~ u D
Channel Boundary
! ! Flow 5~-:§~~~ !~

U
.&:;.&:; :!:.&:;
OlOl c:D Bars oC/)
II c
....Jg' 0
~ 'iii
HIGH <-- RELATIVE STABILITY --. LOW (ij "3
(3%» Low'<-- Bed Load !Total Load Ratio --. high (>11%) .&:; > Decline in mean flow velocity,
Small Sediment Size Large
I- « bed shear stress and stream power
Small Sediment Load Large B
Low Flow Velocity High
Low Stream Power High

Fig.9.9 Classification of channel form based on pattern and

sediment load. (After S. A. Schumm (1981) Society of Economic
Paleontologists and Mineralogists Special Publication 31,
Fig. 4, p. 24.)

Length

Fig.9.11 Nature of water flow in a meandering channel: (A)

downstream changes in cross-sectional channel form and sites of
erosion and deposition; (B) water movement in a channel bend
and associated stresses exerted on channel bed material. ((B)
:t>
3 based on 1. R. L. Allen (1970) Physical Processes of
Point of ~
Sedimentation. Allen and Unwin, London, Fig. 4.5, p. 133(B).)
inflection ;;.'
C
0-
CD

Axis engineers has now shown that friction with the channel bed
of
bend and banks causes shear and turbulence in the water flow
and the development of instabilities which promote the
Concave formation of alternating bars along the channel. A helical
bank
flow is established, with the water surface being elevated on
the outer (concave) bank of each curve and return currents
Fig.9.10 Principal components of meander geometry.
at depth directing the flow towards the opposite bank
downstream (Fig. 9.11(A». The outer bank is eroded as a
rivers with coarse bed material meander forms are often result of the higher flow velocity, whereas deposition takes
highly distorted. place along the inner (convex) bank forming a point bar
The fact that straight channels are rare and meandering (Figs 9.ll(B), 9.12).
channels are common raises the question of what causes If meandering is an inherent feature of fluid flow we
meandering. The meandering behaviour of rivers may might ask why all channels do not meander. Experimental
initially seem anomalous since sinuous channels take a studies have shown that at a given discharge water intro-
longer, lower gradient course when a steeper, shorter duced at an angle into a flume can, as expected, generate
course is available. But we simply have to note that air- excellent meanders. If the gradient is reduced, however,
flows such as jet streams and ocean currents meander to there is a critical minimum gradient below which a straight
recognize that meandering may be the normal behaviour of channel will form and be stable even though lateral flow is
fluids in motion. still introduced into the flume (Fig. 9.13). The explanation
There is an extensive literature on the origin of meanders for this behaviour seems to be that, although secondary
suggesting a range of possible mechanisms for initiating hellcal flows are established at very low gradients under the
and sustaining meandering behaviour. Research by hydraulic prevailing low flow energy conditions they are not suf-
 Fluvial landforms 215

load is the first to be deposited in the channel. This material

forms the nucleus of bars which grow downstream as the
flow velocity is reduced and finer sediment accumulates.
With further decrease in discharge the water level pro-
gre sively falls and the bars are gradually exposed. During
subsequent floods some, or all, of the bars in a braided
channel may be submerged depending on the discharge
attained. During large floods braided channels can experi-
ence major diversions of flow.

9.3.1.3 Channel pattern change

Experimental studies, together with numerous field studies,
have shown how channel patterns change in response to
Fig. 9.12 Point bars clearly evident along a meandering changes in controlling variables. An important result from
channel, Louisiana, USA. these flume studies is the crucial control that appears to be
exerted on channel sinuosity by stream power. Since channel
gradient influences stream power it is not surprising that
gradient is also related to channel sinuosity. There are thus
1·3
a number of interrelated variables including discharge, flow
velocity, sediment concentration and size, and channel
~ 1·2 gradient that can change in response to tectonic, climatic
'Vi
o
...
~ and anthropogenic factors and promote a change in channel
E 1'1 Cl
pattern. The effects of climatic change are examined further
(f)

1·0
~ -Meandering----·I Braided
."'----+-....
in Chapter 14, while tectonic influences are discussed in
Chapter 16.
0'---------"'------'------'------'------'-----'
o 0,4 0·8 1·2 1·6 2'0
Flume Gradient (%) 9.3.2 Alluvial channels: hydraulic geometry
Fig. 9.13 Relationship between valley gradient and thalweg In addition to changes in plan form, alluvial channels may
sinuosity determined from experimental studies employing a also adjust their cross-sectional form in response to changes in
flume. (After S. A. Schumm and H. R. Khan (1972) Geological discharge. These cross-sectional components of channel form
Society of America Bulletin 83, Fig. 6, p.1761.)
were termed hydraulic geometry by L. B. Leopold and
T. Maddock in their innovative analysis of channel adjust-
ficiently powerful to form alternating bars and cause bank ments published in 1953. There are two components of
erosion. By contrast, in high-velocity flows in steep gradi- hydraulic geometry that need to be considered. One involves
ent channels the downstream momentum of flow prevents the relationships representing adjustments over time at-a-
significant cross-channel flows and the development of station, that is, at a particular point along a channel. The
alternating bars, and consequently straight, braided channels other concerns downstream changes evident at a specific
are formed. time.

9.3.1.2 Braided channels 9.3.2.1 At-a-station adjustments

The development of braided channels is favoured by several The analysis of a large number of rivers has demonstrated
factors (Fig. 9.9). In addition to a steep channel gradient the rather consistent relationships between changes in discharge
most important appear to be a large proportion of coarse at-a-st~tion and resulting adjustments in channel width,
material being transported as bed load, and readily erodible channel depth and mean flow velocity (Fig. 9.14). (Note that
bank material which enables channel shifts to occur with these relationships are plotted on logarithmic scales since
relative ease. Once formed', b'ars in braided channels can velocity, width and depth are power functions of discharge)
become rapidly vegetated and thereby stabilized as islands. (Box 9.1). As discharge increases these dependent variables
This points to the role of the highly variable discharge which increase at different rates for different river channels
is typical of many braided rivers. By promoting alternating depending on a number of controlling factors. Where bank
channel degradation and aggradation, large fluctuations in materials are fine-grained and cohesive, channel sides are
discharge help to suppress the establishment of vegetation on steep, and depth increases proportionately more rapidly
braided-channel bars. than width as discharge becomes greater. Where channel
As discharge declines after a flood peak the coarse bed materials are coarse and non-cohesive the banks slope
216 Exogenic processes and landforms

Box 9.1 Hydraulic geometry

At-a-station adjustments of mean flow velocity, channel

width and channel depth to changes in water discharge can
<I>
c:: be described by power function relations of the form:
c::
co
~
U
v= kQrn w = aQh d = cQf
C> where Q is the Qischarge, v the mean flow velocity, w the
o
....J channel width, d the mean channel depth, k, a and care
constants, and m, band f are exponents.
Since vdw = Q it follows that
k+a+c=l and m+b+f=l
~
Although the values of the exponents m, b andfvary, they
a.
<I> seem to approximate to 0.33 for channels in cohesionless
o
sediment. Channels in cohesive sediments have steep banks,
<I>
c::
c::
so channel width increases only slowly with increasing dis-
~
co charge and the exponent b is low (-0.05). The velocity
U exponent is high where channel bed roughness decreases
C> rapidly as discharge increases, such as in straight cobble-bed
o
....J
reaches.
The same power functions used to describe at-a-station
adjustments in hydraulic geometry can be used to express
downstream changes in flow velocity and channel cross-
.~
() section. In a study of rivers in the mid-west of the USA the
o
<I> exponents m, band f were found to be 0.16, 0.46 and 0.38
> respectively. These values show that, in general, mean flow
~ velocity increases downstream and that width increases
o
u.. more rapidly than depth. This indicates that large rivers are
c:: usually slightly wider relative to depth than small rivers.
co
<I>
~
C>
o
....J
proportionately more rapidly than channel depth as dis-
Log Discharge charge increases downstream, while in meandering channels
an increase in discharge is also accompanied by an increase
Fig. 9.14 Schematic representation of at-a-station relationships
between stream discharge and channel width, channel depth and
in meander wavelength.
mean flow velocity. Until the statistical analysis of downstream hydraulic
geometry in the 1950s it was widely believed that mean
gently towards the channel bed, and width increases rapidly flow velocity decreases downstream in response to the
in response to an increase in discharge. The influence of general decrease in channel gradient. The analysis by
channel bank material helps to explain the observation that Leopold and Maddock, however, suggested that this was
channels in semi-arid regions tend to adjust to an increase not the case and that velocity typically increases slightly
in discharge primarily by an increase in width rather than downstream. This might be explained by a decrease in
depth, whereas the converse is usually the case in humid frictional effects due to an increase in flow depth and a
environments where channels generally have more cohesive decrease in channel roughness (associated with a decrease
banks. in sediment size) which more than counteracts· the effects
of a reduction in channel gradient. Nevertheless, such a
9.3.2.2 Downstream changes decrease is not common to all rivers; in some large rivers,
For the majority of rivers, where seepage losses and for instance, velocity may be more or less constant down-
evaporation are not significant, discharge increases down- stream, while in small streams velocity may either increase
stream. As with at-a-station hydraulic geometry it is or decrease downstream depending on specific channel
possible to plot corresponding changes in channel width, conditions.
depth and flow velocity (Fig. 9.15). Because of the dif-
ficulty of acquiring data on bankfull discharge these
9.3.3 Alluvial channels: longitudinal form
relationships are normally expressed in terms of mean
annual discharge. As with at-a-station hydraulic geometry, A further component of downstream adjustments in alluvial
downstream adjustments are also represented as power channels is that of changes in channel gradient, and these
relationships (Box 9.1). In general, channel width increases can be considered at a range of scales. Looking at long
 Fluvial landforms 217

one since a progressive decrease in sediment calibre down-

stream could arise because coarser sediment, once de-
posited, cannot be transported over lower gradients. Neither
Cll
channel gradient nor sediment size are independent vari-
c:
c: ables and a feedback relationship operates between them.
co
.!:
U
The importance of increasing downstream discharge in
en
o producing concave profiles in alluvial channels is under-
...J
lined by observations from rivers in semi-arid environments
which experience a net loss of discharge downstream. Such
rivers typically have a convex profile as a result of down-
stream aggradation during rare flood events.
.!:
A progressive decline in channel gradient in alluvial
a.
o
Cll channels can be disturbed by either changes in base level
m downstream or by tectonic activity causing vertical move-
c:
c:
co ments of the channel bed itself. A fall in base level can give
.!:
U rise to a knickpoint, a discontinuity in the longitudinal
en
o profile of a river, and the downstream channel gradient is
...J
thereby increased. Alluvial channels will tend to adjust
rapidly to such a change through an increase in flow velocity
which leads to increased erosion in the steepened reach. The
>- role of active tectonics is discussed in Section 16.4.
.~
(.)
c: 0 In addition to these broad-scale changes in gradient,
~m
~> other much smaller variations in channel slope can be
d>~
0_ observed in many rivers. Particularly in channel beds
...J L.l.. L....- • _
composed of material of a wide range of sizes, shallow and
Log Mean Annual Discharge
deep reaches are seen to alternate downstream. The
.!:
shallows are formed by high points in the channel bed
....0> known as riffles. They are composed of coarser material
c:
Cll and are characteristically spaced at about five to seven
Cll
>
co times the channel width. The deep reaches are called pools
~ and have a bed of finer calibre than the riffles on either
side. At low stage the water surface is steeper over riffles
and the flow is shallower and more rapid, whereas over
pools the water surface has a lower slope and the flow is
0>
o deeper and slower.
...J L....- _
Theoretical work has shown that even in straight, uni-
Log Bankfull Discharge form channels, turbulence along the channel boundary will
Fig.9.15 Schematic representation of typical downstream generate large-scale roller eddies associated with alternat-
relationships between mean annual discharge and channel width, ing acceleration and deceleration of flow (Fig. 9.16(A)).
channel depth and mean flow velocity, and between bankfull The dimensions of these eddies are a function of channel
discharge and meander wavelength. size, and their spacing averages about six channel widths. It
seems likely that these downstream alternations of fast and
sections of a river course, channel gradient is generally slow flow lead to the development of zones of erosion
seen to decrease downstream giving a concave-up longi- (pools) and zones of sediment accretion (riffles) which
tudinal profile (long profile). This downstream decrease in eventually assume an equilibrium form. It is the riffles that
gradient is accompanied by an increase in discharge, and it are the sites of accretion and this may seem to contradict
can be explained by the ability of a river to transport the the faster flow observed over riffles. This faster flow, how-
same quantity and calibre of sediment load in a lower ever, occurs during low stage' when little sediment is in
gradient channel as discharge increases. Average sediment transit. At high stage when the flow velocity is sufficient
size does, of course, tend to decrease downstream as a for significant erosion and deposition to occur, it is the
result of sorting, attrition and weathering, and at a given pools that experience higher bed velocities and greater
flow velocity finer sediment can be transported in lower- shear stresses as a result of their lower roughness and the
gradient channels. None the less, the relationship between convergence of flow from the riffles upstream.
channel gradient and sediment size is not a simple causal In Section 9.3.1.1 it was mentioned that lateral com-
218 Exogenic processes and landforms

A fast slow fast slow pools and riffles and the wavelength of the meanders that
PLAN eventually develop, meander wavelength being approxi-
mately twice the pool spacing (Fig. 9.16(B». This situation
erosion
LONG may change, however, if the channel between each
PROFILE meander loop is lengthened in which case new pools and
initial
/ bed
riffles may develop (Fig. 9.16(C».
~ ~
\
final
CROSS-
SECTION V -~-- initial bed
bed

9.3.4 The alluvial channel system

Figure 9.17 summarizes the relationships between the
~ erosion of initial channel numerous components of the alluvial channel system dis-
cussed in the preceding sections. It is a speculative attempt
I:~.:':.::..:.J accretion on initial bed to illustrate the way in which alluvial channels can adjust to
changes in the independent variables of valley slope, dis-
B charge and bed material size and sorting.
Initial uniform channel -

~ ..... ~----:-:.:
:.::.,
.. ~

~pool
-:.:.; riffle

Fig. 9.17 The alluvial channel system. Independent variables

are indicated by a bold outline. Direct relationships are shown by
+ and inverse relationships by -. Indeterminate relationships are
indicated by double-headed arrows. (After K. Richards (1982)
Rivers: Form and Process in Alluvial Channels. Methuen, London,
Fig.1.8,p.26.)

Under a relatively consistent hydrological regime alluv-

~ Pool ·::·;·.Riffle ~ Erosion ~ Point bar ial channels experience constant adjustments in response to
changes in hydraulic and sedimentological variables which
Fig.9.16 Development ofpool and riffle sequences in channels
and their relationship to meander morphology: (A) erosion and maintain a more or less stable channel form. Such a state of
accretion of a channel bed corresponding to alternate zones of equilibrium is usually described as grade. The concept of a
fast and slow flow; (B) transformation ofa straight to a graded river was elegantly expressed by J. H. Mackin in
meandering channel in relation to pool spacing; (C) the 1948.
development of additional pools and riffles with the lengthening
of a meandering channel. (After K. Richards (1982) Rivers: Form A graded river is one in which, over a period of years, slope and
and Process in Alluvial Channels. (Methuen, London, Fig. 7.2C, channel characteristics are delicately adjusted to provide, with
p.184; and G. H. Dury, (1969) in: R. 1. Chorley (ed.) Water,
available discharge, just the velocity required for the
Earth and Man (Methuen, London) Fig. 9.11.4, p. 180.)
transportation of the load suppli~d from the drainage basin. The
graded stream is a system in equilibrium; its diagnostic
ponents of flow are present even in straight channels which
characteristic is that any change in any of the controlling factor
are capable of initiating channel sinuosity by selective bank
will cause a displacement of the equilibrium in a direction that
erosion under certain hydraulic and sedimentological con-
will tend to absorb the effect of the change.
ditions. As a result of the greater shear stress exerted on the
channel boundary in pools during channel-forming high The last sentence provides an excellent description of the
discharges, banks are eroded preferentially in pools. There proce s of negative feedback and, indeed, most alluvial
is consequently a correspondence between the spacing of channels for most of the time exhibit this form of ystem
 Fluvial landforms 219

behaviour. But changes in controlling variables in most channel or by material deposited in the channel by, for
rivers occur very frequently in comparison with the time instance, a landslide. The most obvious discontinuities in
taken for adjustments in channel properties to occur so a the longitudinal profile of bedrock channels are represented
graded channel is most realistically regarded as one in a by waterfalls and rapids. The term waterfall is applied
state of quasi-equilibrium rather than true steady-state where there is a vertical fall of water, a succession of
equilibrium. waterfalls giving rise to a cataract. Rapids are less marked
irregularities in the channel bed, consisting of short, steep
reaches which may be entirely submerged at high stage.
9.3.5 Bedrock channels
Given their importance in deciphering the history of river
Bedrock channels represent a dramatic contrast with alluv- systems, waterfalls have been remarkably little studied. It is
ial channels in that they are capable of only very slow widely considered that the great majority of waterfalls
adjustment of form in response to changes in discharge, develop as a result of the erosion of weak rock from
sediment load, gradient and other factors. Bedrock channels . beneath a resistant caprock. The classic example are held to
cannot, therefore, be realistically analyzed in terms of the be the Niagara Falls not only because of their popularity as
concept of grade because their response time is so long. a tourist venue but also because G. K. Gilbert carefully
It is generally held that rivers are most commonly cut into considered their development. He emphasized the import-
bedrock in the upper part of their courses where. channel ance of the undermining of the shale at the base of the falls
gradients are usually steeper and the coarseness of the load by the rolling of large limestone boulders detached from
they carry is greater and thus more effective as an agent of the caprock in the cascading water flow. The Niagara
abrasion. Although this is undoubtedly true in the majority of model is, however, far from universal since many water-
cases, there are a number of instances where major rivers falls possess no overhanging caprock but instead are
flow in alluvial channels for much of their upper course buttressed at the base (Fig. 16.24).
before plunging into bedrock channels as the channel I Contrary to Gilbert's widely accepted view thin falling
gradient steepens markedly downstream. A number of large water has little erosive power unless it carries a coarse
rivers in Africa exhibit thi characteristic, for example the sediment load, calculations of both the force exerted by a
Orange River in southern Africa (see Fig. ·16.24). large mass of falling water and the evidence of rapid scour
Little research has been carried out into bedrock at the base of dam overflows indicates that the impact of
channels in comparison with the abundance of studies of water is probably the predominant mechanism of fluvial
alluvial channels. From the research that has been under- erosion operating on waterfalls. Indeed plunge pools are
taken it appears that the longitudinal profile is normally commonly found at the base of waterfalls and the overall
much more irregular than in alluvial channels, while the process of waterfall retreat is also aided by cavitation and
cross-sectional form of bedrock channels is greatly influ- the weakening through weathering of rock surfaces exposed
enced by structural controls with fluvial erosion tending to to frequent wetting. The rate at which waterfalls retreat and
widen and deepen joints or other zones of weakness. whether the discontinuity in the longitudinal profile that
More attention has been paid to the smaller-scale erosional they represent is smoothed over time are important
features evident in bedrock channels. Plucking and cavi- questions which still remain to be satisfactorily answered.
tation gives· rise to irregular channel-bed surfaces, while
abrasion and corrosion tend to smooth scour marks. The 9.3.5.2 Incised meanders
most characteristic erosional form is the pothole. Three Although we noted above that meandering patterns are
types can be recognized: a narrow cleft in a rock face common in alluvial channels, meandering channels are also
caused by the cutting back of a waterfall; shallow de- found incised into bedrock. Incised meanders are particu-
pressions resulting from separated flows which promote a larly well developed in horizontally bedded strata and form
secondary circulation in the water flow on the channel bed; when a river cuts down through its alluvium and into the
and deep and often almost perfectly circular eddy holes underlying bedrock. Two types of incised meander can be
where a vortex-like circulation has become established. identified. Intrenched meanders are symmetric forms and
develop where downcutting is sufficiently rapid to mini-
9.3.5.1 Longitudinal profile mize the lateral migration of the meanders. This can occur
Irregularities in the longitudinal profile of bedrock channels where a significant fall in base level leads to the upstream
can arise from a downstream steepening of gradient as a migration of a knickpoint. Ingrown meanders, by contrast,
consequence of a fall in base level, but the knickpoint pro- are asymmetric as a result of lateral meander migration
duced may often be difficult to distinguish from a channel occurring concurrently with relatively slow incision. This
bed discontinuity related to the differential erosion of rocks slow downcutting is initiated by regional warping rather
of contrasting resistance. Discontinuities may also be pro- than rapid base level changes. Lateral migration of meanders
duced by vertical movements along a fault lying across a can proceed to the point where two meander loops cut
220 Exogenic processes and landforms

Fig. 9.18 Deformed, intrenched meander, part of the Goosenecks of the San Juan River, Utah, USA.

through a bedrock spur and form a natural arch. 9.4 Fluvial depositional landforms
The classic intrenched meanders of the San Juan River, a Deposition by rivers occurs predominantly in the bottom of
tributary of the Colorado River in the south-west USA, are valleys where gradients are low, at locations where there is
highly deformed and compressed and appear not to have a significant change in gradient, or where channelled flow
been inherited directly from a meandering pattern devel- diverges, and is consequently reduced in depth and vel-
oped in overlying alluvium (Fig. 9.18). Formation of the ocity. It must be emphasized, however, that deposition is
meanders seems to have been initiated through rapid not exclusive to any part of the fluvial system. Four major
downcutting of the Colorado River which lowered the base types of fluvial deposition can be distinguished: channel
level of the tributary San Juan. As headward erosion deposits and channel margin deposits which accumulate
progressed up the initially alluvial channel of the San Juan
within and along river channels; overbank deposits formed
the underlying bedrock would have been encountered first when bankfull discharge leads to the deposition of fine
on the lower downstream limb of each meander. Further sediments beyond the confines of the channel itself; and
downstream, migration of this part of the meander would valley margin deposits which accumulate at the base of
then be severely retarded, but the upstream limb could valley slopes (Table 9.3).
continue to shift downstream until it also encountered bed-
rock thereby compressing the meander loop (Fig. 9.19).
9.4.1 Floodplains
Interestingly, incised meanders on the Colorado Plateau
exist on the upstream but not the downstream flanks of Except in mountainous terrain most river channels are
folds. This seems to be because on the upstream flank flanked by an area of subdued relief termed a floodplain
intrenching meanders would encounter bedrock, but on the formed by deposits laid down when the river floods. Low
downstream flank the river flows down dip and the increase magnitude - high frequency floods cover only a part of the
in channel gradient destroys the meandering channel floodplain, and it is only during rare major floods that the
pattern before it can become incised. entire floodplain is inundated. Depths of flood water range
 Fluvial landforms 221

plains, this lies withIn a much more extensive area of older

floodplain deposits. Where closely spaced river channels
flow across a region of subdued topography adjacent flood-
plains may merge to form extensive alluvial plains.
Two modes of sediment deposition can be identified;
lateral accretion deposits are laid down as the river
channel migrates across its floodplain, while vertical
accretion deposits accumulate beyond the confines of the
channel during floods. Lateral accretion deposits are
formed by both meandering and braided channels, but in
the latter case a large part of the floodplain effectively
operates as part of the active river channel at anyone time.
Rapid changes in channel courses in braided rivers leave
abandoned channels floored by coarse bed-load material
3 which, if they survive reworking during subsequent channel
shifts, are eventually filled by organic material and fine
sediments laid down during floods. Lateral accretion by
meandering channels occurs through, the mbre gradual
growth of point-bar deposits on the convex side of meander
bends (Fig 9.12, 9.20). The overrunning of one meander
loop by another produces a meander cut-off, or chute,
which shortens the channel course (Fig. 9.8). The aban-
Fig.9.19 Development of a deformed, intrenched meander (J, 2, doned meander loop is gradually isolated by the deposition
3) when the downstream limb is fixed by bedrock. The letter B
indicates the location at which the channel enters bedrock. of bed material at each end by the main channel and
Upstream ofB the channel is in alluvium. (After S. A. Schumm becomes an oxbow lake. Eventually the oxbow lake ac-
(1977) The Fluvial System. Wiley, New York, Fig. 6-15, p. 200.) . cumulates channel-fill deposits laid down during floods.
Whereas lateral accretion deposits are constantly being
from a few centimetres up to several metres for large floods formed and reworked, vertical accretion occurs only during
in major rivers. As with the dimensions of river channels, floods when bankfull discharge is exceeded. Morpho-
the width of floodplains in most rivers is roughly pro- logically the most prominent depositional landforms in most
portional to discharge. The active floodplain of the lower floodplains are levees (Fig. 9.20). These are ridges which
~ississippi is about 15 kIn across, but, as with most flood- lie parallel to the river channel and may rise to a height

Table 9.3 Classification of valley sediments

PLACE OF DEPOSITION NAME CHARACTERISTICS

Channel Transitory channel deposits Primarily bed load temporarily at rest; part may be preserved in more durable
channel fills or lateral accretions
Lag deposits Segregations of larger or heavier particles, more persistent than transitory
channel deposits
Channel fills Accumulations in abandoned or aggrading channel segments; ranging from
relatively coarse bed load to fine-grained oxbow lake deposits
Channel margin Lateral accretion deposits Point and marginal bars that may be preserved by channel shifting and added
to overbank floodplain
Overbank floodplain Vertical accretion deposits Fine-grained sediment deposited from suspended load of overbank flood
water; including natural levee and backswamp deposits
Splays Local accumulations of bed-load materials spread from channels on to
adjacent floodplains
Valley margin Colluvium Deposits derived chiefly from unconcentrated slope wash and soil creep on
adjacent valley sides .
Mass movement deposits Earthflow, debris avalanche, and landslide deposits commonly intermixed
with marginal colluvium; mudflows usually follow channels but also spill
overbank

Source: After P. C. Benedict et al. 1971, J oumal of the Hydraulics Division, Proceedings of the American Society of Civil Engineers 97, Table 2 - Q.l,
p.44.
222 Exogenic processes and landforms

Floodplain
Channel-fill
downslope from a point on a mountain front, usually where
Point-bar
deposit a stream emerges (Fig. 1.10(C)). Most alluvial fans have a
Levee
radius of less than 8 km, but under certain conditions fan
Earlier
deposit radii may exceed 100 km. In fomi they have concave-up
long profiles, but convex-up cross-profile~. Th~ mean sur-
face slope of fans generally ranges from 1 to 5 , but at the
fan apex gradients can exceed 10°. Since the slope of
streams emerging at the mountain front is usually similar to
the gradient of the upper fan surface, it seems that the
Crevasse-splay deposition that causes fan building occurs primarily as a
EJ ~ lm ~ result of the sudden change from a confined to an unconfined
Decreasing Cross- Cross- Flat-
grain size- bedding lamination bedding condition as the stream leaves the mountain gorge rather
than as a result of reduced channel gradient. The calibre of
Fig.9.20 Depositionalforms developed along a meandering sediment deposited generally decreases down-fan, but fan
channel. Note how levees give rise to steep high banks where they
deposits are often poorly sorted as they are frequently laid
are undercut on the concave side of meander bends. After J. R. L.
Allen (1964) Sedimentology 3, Fig. 4, p. 168 and 1970, down by torrential floods. In catchments containing abun-
Geological Journal 7, Fig. 1, p.131.) dant, easily erodible debris, sediment concentrations may
reach the point where high-viscosity debris flows are
generated.
Alluvial fans develop where there is a relatively abun-
dant sediment supply, adequate relief for vertical fan growth
of several metres above the general level of the floodplain.
and a suitable location for sediment accumulation. Favour-
When bankfull discharge is exceeded water starts to spread
able environments include active faulted mountain fronts
rapidly across the floodplain. The decrease in depth in
where the rate of mountain uplift, or adjacent basin sub-
comparison to the river channel rapidly reduces the
sidence, exceeds the rate of downcutting of the trunk
competence of the flow and the coarser fraction of the
stream feeding the fan. Individual fans along a mountain
suspended load is promptly deposited to form levees. The
front may grow laterally to the extent that they coalesce to
presence of levees affects the movement of flood waters
form a continuous piedmont sedimentary apron. In arid
once bankfull discharge is exceeded. Initially the water is
environments such landforms are known as bajadas (Figs
ponded by levees, but as the level rises it breaks through
1.10(B), 9.21). In humid environments, where alluvial fans
forming crevasses. The accelerated flow through crevasses
tend to be larger, the coalescence of fans can produce
is able to transport a relatively coarse suspended load
immense, gently inclined alluvial slopes, such as those that
which is quickly deposited as the flow disperses to form
flank the southern margin of the Himalayas.
fan-shaped crevasse-splay deposits in the backswamp
Although some alluvial fans in arid environments are
zone beyond the levee barrier (Fig 9.20). Flow diverted
simple undissected forms, most are dissected - that is, the
through a crevasse may travel many kilometres down-
trunk channel is entrenched below the upper part of the fan
stream before it is able to regain access to the main
surface and deposition is focused towards the fan toe
channel. This may not occur until a tributary is reached, but
(Fig. 9.22). Fan-head trenching will occur if sediment
tributaries themselves may be diverted downstream by
yield decreases, or channel gradient and flow velocities
levee development along the main channel. Occasionally
increase. Such changes can be brought about by climatic
the main channel itself may shift its course into the
changes, increased tectonic activity or even human-induced
backswamp zone.
land use changes in the upstream basin. But we might also
Many of the world's floodplains have thick accumu-
anticipate that fan incision will occur in the long term
lations of vertical accretion deposits, but this is probably a
without any change in external factors simply as a result of
reflection of the particular conditions of the past 10 000 a or
the reduction in sediment yield as relief, and therefore rates
so when most river systems have been responding to a rise
of erosion in the basin, progressively declines through time
in base level in response to the post-glacial rise in sea level
as the moutain mass is gradually lowered.
(see Section 17.6.1). Under conditions of stable base levels
Not surprisingly, fan area has been found to be positively
we would expect lateral accretion deposits to be quan-
correlated with source basin area, although there may be an
titatively more significant.
order of magnitude difference in fan area for a given con-
tributing area. Some of this variation may be explained by
9.4.2 Alluvial fans
lithological factors. Basins in sandstone, for instance, have
An alluvial fan is a body of sediment whose surface form been found to have smaller associated alluvial fans than
approximates to the segment of a cone which radiates those underlain by shales or mudstone. This is presumably
 Fluvial landforms 223

Fig.9.21 Bajada/ormed along the/aulted, northern margin o/the Turfan Depression, Xinjiang Autonomous Region, China. The
area covered by the image is about 180 km across. (Landsat image courtesy A. S. Walker.)

because sandstone is less easily eroded and gives rise to ~~e fan to another. In large arid fans only a very small
lower sediment yields. Tectonic factors can also influence proportion (usually < 5 per cent) is active during anyone
fan area. In Death Valley progressive eastward tilting has flood event. In contrast humid fans are fed by perennial,
confined the east~side fans but enabled the west-side fans and often braided, channels which tend to migrate progress-
to extend several kilometres beyond the mountain front ively across the fan surface. Such channel migration is well
(Fig.l.IO(B)). illustrated by the giant Kosi Fan on the southern flanks of
Alluvial fans in humid environments are distinct in a the Himalayas (Fig. 9.23). The course of the Kosi River
number of respects from those in arid environments. Apart across the fan has moved more than 100 km to the west
from being generally smaller and steeper, fans in dry over the past 250 a. Coarse debris is transported only a few
regions are fed by ephemeral stream channels and active kilometres beyond the fan head and the greater part of
deposition tends to move unpredictably from one part of the fan is composed of fine- to medium-grained deposits.
224 Exogenic processes and landforms

lowest will be the youngest and may retain traces of

floodplain morphology, while the highest will be the oldest
and will usually be partly degraded. Paired terraces form
when vertical incision is rapid in comparison with the
lateral migration of the river channel (Fig. 9.25). Morpho-
logically similar features can, however, be produced by
resistant beds in flat-lying strata. These give rise to
structural benches and structural controls must be elimin-
ated before a river terrace interpretation is accepted.
Unpaired terraces form where lateral shifting of the channel
is relatively rapid; this results in the river cutting terraces
alternately on each side of the valley floor. Since valleys
experience phases of aggradation and degradation, once
formed terraces can subsequently be covered by sediment.
Thus, we would expect complex sequences of buried and
partially eroded terraces to lie beneath modern floodplains
(Fig. 9.26). Terraces can be formed under a wide range of
circumstances. Unpaired terraces do not, in fact, require
any change in external conditions since they can form
simply through the progressive incision of a river channel
migrating from one side of its floodplain to the other.
Many episodes of channel entrenchment and terrace
formation probably arise from changes in base level or
fluctuations in climate, although tectonic deformation along
river courses can also lead to local terrace development
(see Section 16.4). A fall in base level can lead to terrace
development if it generates a downstream increase in
channel gradient. This is then propagated upstream through
knickpoint retreat. The past 3 Ma have seen frequent
changes in sea level of up to 100 m or so, and this has
created conditions highly favourable to terrace formation.
Many other river terraces currently present in the landscape
are attributable to climatic fluctuations which lead to
changes in channel discharge and sediment load. As with
base level changes, climatic fluctuations have' also been
Fig.9.22 Building (A) and entrenchment (B) phases of alluvial frequent and rapid over the past 3 Ma or so. The response
fan development. (From R. U. Cooke and A. Warren (1973) of river systems to such environmental changes are
Geomorphology in Deserts. Batsford, London, Fig. 3.8, pp. 186
complex and can vary dramatically depending on local
and 187, after L. K. Lustig (1965) United States Geological
Survey Professional Paper 352-F, Fig. 137, p. 185.) channel conditions. For instance, during glacials the
Mississippi incised into its floodplain in its lower valley
producing a series of terraces (see Section 17.6.1), but in
9.4.3 River terraces
the upper part of its basin the great increase in sediment
Changes in channel gradient, discharge or sediment load supplied from the Laurentide ice sheet led to channel
can lead to a river channel incising into its floodplain. The aggradation.
original floodplain is thereby abandoned and is left as a The tendency of fluvial systems to aggrade or incise is
relatively flat bench, known as a river terrace, which is particularly sensitive to climatic fluctuations in semi-add
separated from the new floodplain below by a steep slope. regions. This is because in such environments modest
As well as forming in the alluvial fill of river valleys, river changes in annual precipitation can produce significant
terraces can also be cut into bedrock and in this case they changes in vegetation cover which are reflected in large
are covered by only a thin sediment veneer. The term changes in the rate of sediment supply to stream channels.
strath terrace is sometimes applied to these forms. In the south-west USA many ephemeral stream channels,
River terraces are inclined downstream but not always at known locally as arroyos, show evidence of phases of
the same inclination as the active floodplain. A valley side aggradation and entrenchment. Dramatic changes have
may contain a vertical sequence of terraces (Fig. 9.24). The occurred in a period of a few hundred years with the most
 Fluvial landforms 225

Fig.9.23 The Kosi Fan on the southern flank of the Himalayas. Thefan covers an area of 15 000 km 2 , but has an extremely low
gradient averaging 1 m km- 1 at its apex, and only 0.2 m km-1 at its toe near the Ganges River (which flows west to east across the
bottom of the image). The Kosi River drains an area of very high relief and monsoonal climate, and these factors combine to generate
a high sediment yield. Dates are given for the progressive westward migration of the course of the Kosi River since 1731. The image
covers an area about 180 km across. (Landsat image courtesy N. M. Short.)

recent period of entrenchment and terrace formation occur- 9.S Fluvial activity through time
ring between the 1860s and around 1915. The cause of this
episode of channel entrenchment is uncertain. In some
areas it is apparently due to subtle climatic fluctuations, but
9.5.1 Geomorphic thresholds, complex response and
in other localities significant changes in land use may also
episodic erosion
have played a role by influencing the supply of sediment to In Section 7.6 we noted how slope behaviour can be
channel systems. The possibility remains, however, that analyzed in terms of the exceeding of extrinsic or intrinsic
alternating phases of aggradation and degradation may thresholds. Similar reasoning can be applied to fluvial
represent the normal behaviour of fluvial systems even in systems, but in this case changes through time in the
the absence of changes in external controls. morphology of the landform itself can lead to the crossing
226 Exogenic processes and landforms

Fig.9.24 Terrace sequence formed in glacial outwash debris, Cave Stream, Canterbury High Country, South Island, New Zealand.

A Paired, poly-cyclic terraces B Unpaired, non-cyclic terraces

~~
A No terrace

~
2
~r
B One terrace

~
-:-- --:-:-- ~-:--
----_. ~ .
--:::=:::-:---:-:::::-- -'\. --- - --:=--..:: ;;~ --::_-. ," _:::--

C Two terraces

Fig.9.25 Paired and unpaired terraces. Note how the vertical Fig. 9.26 Valley cross-sections showing some possible terrace
spacing of terraces is more or less retained in paired terraces but and alluvial-fill combinations. (After L. B. Leopold and J. P.
converges downstream in unpaired terraces. (After B. W. Miller (1954) United States Geological Survey Water-Supply
Sparks (1972) Geomorphology (2nd edn). Longman, London, Paper 1261, Fig. 2,p. 5).
Figs 9.5,9.7 and 9.8, pp. 296 and 297; and W. D. Thornbury,
(/969) Principles of Geomorphology (2nd edn). Wiley, New York
and London, Fig. 6.9, p. 157 after Longwell et al.)

of a threshold. An example is provided by ephemeral erosion ensues. This is an example of a geomorp.hic

stream channels. These may experience periods of sediment threshold since the breaching of the threshold is inherent
accumulation on the valley floor until a point is reached in the way the morphology of the landform changes
where the channel gradient has been so increased that rapid through time. (Note, however, that the term geomorphic
 Fluvial landforms 227

threshold has been applied more broadly by some writers to

UPSTREAM
encompass both intrinsic changes and external factors - I
retain the more specific usage here.) The existence of
geomorphic thresholds means that many dramatic adjust- BASE LEVEL
ments in fluvial systems which have hitherto been ascribed FALL- NOCHANGE
to the effects of external factors may simply be the result of INCISION NO
CHANGE
the inherent behaviour of the landform itself. We must
clearly be very careful in our interpretations of river
channel adjustments, especially in climatically variable and INCISION
2 EXTENDS INCISION
tectonically active environments where we might tend to UPSTREAM
assume that all dramatic landform changes are the result of INCREASE
external factors.
The appreciation of the importance of thresholds in
fluvial systems has important implications for the way in 3 AGGRADATION INCISION
which we conceive how such systems change through time. INCREASE
The existence of thresholds indicates the inability of river
channels necessarily to adjust rapidly to new equilibrium
AGGRADATION
conditions, and we would expect different components of 4 EXTENDS AGGRADATION
the fluvial system to respond at different rates to external UPSTREAM
changes. Because of the intricate relationships between DECREASE

these different components - alluvial channels, tributary

streams, valley-side slopes and divides - we would expect
fluvially-eroded landscapes to exhibit a complex response 5 INCISION AGGRADATION

as the effects of any change in external factors, such as DECREASE

those arising from tectonic activity or climatic change, dif-
fuse through drainage networks.
INCISION
This behaviour has been demonstrated experimentally in 6 EXTENDS INCISION
laboratory drainage .systems where a base-level fall at the UPSTREAM
INCREASE
basin outlet leads to successive phases of incision and
aggradation as the system hunts for a new equilibrium (Fig.
9.27). The fall in base level initially promotes incision, but
as the increase in channel gradient diffuses throughout the 7 AGGRADATION ~
drainage system the sediment supply from the upper tri-
butaries increases to the point where the downstream
channel is incapable of transporting all the sediment being
delivered and aggradation begins. This aggradation then
extends upstream, reduces channel gradients and leads to a Fig. 9.27 Schematic representation of complex response in a
decrease in upstream erosion and sediment supply. This in drainage system as a result of a fall in base level.
tum allows channel incision to begin again downstream.
The response of the drainage basin is characterized by
episodic erosion punctuating periods of deposition. In this since the early 1960s has been at what relative magnitude
model of drainage basin development, features such as of discharge is most sediment transported in rivers and
flights of river terraces and alluvial-fill sequences may most geomorphic work accomplished. Although sediment
simply reflect the attempt of the system to re-establish transport increases with increasing flow velocity and dis-
equilibrium after an initial perturbation and need not be charge, high flows occur much less frequently than low
indicative of successive changes in external variables. The discharges. In a highly influential paper published in 1960,
concepts of geomorphic threshold, complex response and M. G. Wolman and 1. P. Miller argued that in most river
episodic erosion clearly have implications for the way we systems about 90 per cent of the total sediment load is
view the long-term evolution of fluvially-eroded landscapes, transported by relatively modest flows, that is, those that
and we examine this topic in Chapter 18. have a recurrence interval of up to 5-10 a. In most cases
large floods individually transport large sediment loads, but
they do not occur with sufficient frequency to have a major
9.5.2 Magnitude and frequency of fluvial activity
impact on overall sediment transport rates. Although there
A key question addressed by fluvial geomorphologists are some discrepancies between studies of different rivers
228 Exogenic processes and landforms

with contrasting proportions of high discharges, most Most alluvial channels have a low resi tance threshold
researchers support the view that bankfull discharge and are highly sensitive to floods of relatively small magni-
corresponds closely with the most effective sediment- tude and high frequency. When a major flood occurs there
transporting flows. This observation is of particular interest is a rapid response in terms of channel morphology and bed
because there i~ considerable evidence supporting the view roughness, but there is equally a relatively prompt readjust-
that in many rivers bankfull discharge is generally the most ment to lower flow conditions after a flood event. The
significant in determining channel morphology. situation in bedrock channels, or channels composed of
Althou$h the prevailing philosophy of fluvial geomor- very coarse alluvium, is very different. In highly resistant
phologists is that most rivers for most of the time are in a bedrock, for instance, erosion and channel modification
state of quasi-equilibrium in which channel morphology is will only occur during extremely high magnitude and very
broadly adjusted to prevailing discharge and sediment load, rare floods. Since such floods are usually short-lived, there
some researchers have highlighted the critical role of major is insufficient time for bed morphology to achieve
floods in certain kinds of fluvial systems. There are 'equilibrium' with flow conditions, and consequently the
numerous examples of high magnitude floods which have resulting 'disequilibrium' forms may persist for a long
been seen to produce either little, or very short-term, period_of time in certain environments (see Section 11.4.1).
change in fluvial channel form, but there are also instances
where spectacular changes have occurred. One such case
was associated with a storm over a part of central Texas, 9.6 Fluvial systems in limestone terrains
USA, on 2 August 1978. At least 790 "mm of rain fell in just
The susceptibility of rocks composed predominantly of
24 h in some localities, the most intense rainfall ever
calcium carbonate to the weathering process of carbonation
recorded in the USA up to that time. At a gauging station
gives rise to a distinctive form of fluvial activity on
on the Medina River the peak flow was 15 m deep and the
limestone terrains. We have already discussed the develop-
discharge reached 7920 m 3 S-1 from a catchment area of a
ment of solutional forms on limestone associated with
little over 1200 km 2 . The recurrence interval of such a flood
weathering and the development of karst scenery in general
cannot be accurately estimated from the brief, 38 a dis-
terms (see Section 6.4.1), but we have yet to consider
charge record available, but probably lies in the range
specifically the genesis of fluvial landforms in such
100-500 a. The geomorphic effects of this flood were
environments.
considerable. Trees lining the low-flow banks of the
The distinctive nature of fluvial activity on limestone
Medina River were ripped up and the underlying alluvial
arises from the high permeability of the bedrock and the
sediments which had accumulated over centuries suffered
diversion underground of a variable, but in many cases
extensive erosion. Boulders over 2 m in diameter were
significant, proportion of runoff. The permeability of rock
detached from jointed bedrock ledges and carried tens of
is in part related to its primary porosity, that is, the
metres downstream. In other channel sections giant ripples
proportion of the rock occupied by voids. (Strictly it is the
were formed; the largest of these were composed of gravel
effective porosity that is important - that is, the proportion
and boulders and their enormous wavelength of up to 80 m
of voids that are connected and thus able to transmit water.)
indicates a flow velocity of 3-4 m S-l and a flow depth of
In many types of limestones, however, it is their secondary
10m.
porosity, arising from the solutional enlargement of joints,
The role of flood events in shaping fluvial landforms
fissures and bedding planes, that makes by far the greatest
depends in part on the frequency distribution of high dis-
contribution to their permeability. Ultimately such enlarge-
charges. In semi-arid environments, for instance, rare flood
ment can give rise to caves. Since such secondary porosity
events may be relatively important in determining channel
increases over time as calcium carbonate is removed in
characteristics because the much smaller discharges
solution, the proportion of subsurface flow depends both on
occurring for the great majority of time will be ineffective
the lithological properties of the limestone and on the
in reshaping the channel between these major flood events.
length of time that it has been subject to active dissolution.
Another very important factor, however, is the resistance of
the materials experiencing the effects of flood discharges.
To playa significant geomorphic role the erosional force
9.6.1 Surface drainage
exerted by floods (expressed as specific stream power)
must exceed the resistance of materials, including veg- In addition to a topography characterized by numerous
etation, forming the channel bed and banks. The important closed depressions, karst landscapes may also contain
point here is that there is an enormous range in the fluvially eroded valleys. In limestone regions, such as those
resistance of different kinds of channel materials, and underlain by chalk, where karst has not developed the
consequently a wide range in the critical discharge at which landscape may be predominantly one of fluvially eroded
significant channel modification can occur. valleys. If recently exposed through the removal of over-
 Fluvial landforms 229

lying impermeable strata these valleys may contain the junction of an impermeable lithology and a thick over-
streams, but dry valleys with no active stream, or valleys lying limestone, and the marked difference in surface flow
which occasionally support short-term flow following a up-valley and down-valley of the spring causes a much
storm or longer period of heavy rainfall, are much more more rapid incision of the down-valley section forming a
typical of limestone terrains. pocket valley (Fig. 9.28(B». A steep headwall formed in
In some cases only the lower section of a valley is dryas the limestone may be produced which gradually retreats
a result of the surface flow in its upper part disappearing through spring sapping at its base. An excellent example is
underground via a sinkhole. If this situation persists the provided by Malham Cove in Yorkshire, UK.
floor of the upper part of the valley may be lowered Allogenic valleys are formed where river courses cross
significantly below the level in its lower part due to the from an area of strata capable of supporting permanent
drastically reduced potential for erosion down-valley of the surface flow on to limestone (Fig. 9.28(C». The length of
sinkhole. The two sections of the valley may eventually be such valleys on the limestone outcrop depends both on the
separated by a limestone 'cliff', in which case the upper discharge of the river and the permeability of the valley
section is referred to as a blind valley (Fig. 9.28(A». In floor. Small streams may disappear underground within a
other cases water may emerge at the surface on a valley few hundred metres, whereas large rivers may be able to
floor as a spring or resurgence. This frequently occurs at sustain flow for several tens of kilometres or more.
Allogenic valleys cut through massive limestone can form
spectacular gorges, such as the 300 m deep Tarn Gorge in
BLIND VALLEY A southern France.
Blind Although individual dry valleys can generally be ex-
7
Dry
11ey
Sinkhole I
Valley
plained by the progressive loss of subsurface flow over
time as the permeability of the underlying limestone

¥: i @@dIT8J3
increases, extensive integrated networks of dry valleys,
such as those of the chalklands of southern England and
northern France are somewhat more problematic. Numerous
hypotheses have been proposed to explain their develop-
POCKET VALLEY B
ment. One idea is that a drainage system was initially
Dry established on overlying impenneable strata which provided
Valley
sufficient surface runoff to cut a valley network in the
\ underlying chalk as the latter was gradually exposed.
Another possible explanation is that during the glacial
episodes of the past few million years the ground became
perennially frozen a short distance below the surface and
this rendered the chalk effectively impermeable and thus
allowed surface flow and valley formation to occur. Yet
ALLOGENIC VALLEYS c another possibility is that the water table has been
progressively lowered through time at a faster rate than the
.. JL~G:Ti!T:rges
..:.: ~.: .
upper tributary streams could lower their valley floors. This
is particularly likely to happen where the strata are gently
.. ~
. ' .. '. .
.... ... ....
'S
.,. dipping (Fig. 9.28(D», but it might also be a nonnal
. ., Dry / 0 product of the dissection of highly permeable lithologies
Valley ~J \ IV I
..
.
/
0
Sinkholes E: where lower lying trunk streams can maintain flow even
'\
Abandoned Sinkhole
during periods of low water table.

DIPSLOPE DRY VALLEYS o 9.6.2 Subsurface drainage

The size and frequency of joints and fissures in limestone
have an important influence on the type of subsurface water
flow that occurs. At one extreme limestones, such as chalk,
possess a high density of small fractures and consequently
surface infiltration rates are uniformly high. Chalk is
mechanically weak so fissures can only be enlarged to a
Fig.9.28 Formation 0/ different types o/valleys in limestone certain point before they collapse. The rock is essentially
terrain. permeable throughout (although water movement is still
230 Exogenic processes and landforms

concentrated in the larger fissures) and there is a distinct Box 9.2 Darcy's law
and continuous water table. At the other extreme are the
majority of limestones which are densely cemented and The flow of water flowing through a porous medium is given
through which the greater amount of water flow occurs along by
major joints and fi ures. This creates discrete systems of dh
subsurface flow with no single continuous water table, v=-K
dl
although in time a somewhat more integrated system of
subsurface drainage may develop as conduits are enlarged where v is the specific discharge (discharge per unit cross-
sectional area), K the hydraulic conductivity and dh/dl the
to the point where the rock is honeycombed with linked hydraulic gradient. The hydraulic gradient is determined by
cave systems and there is one general water table level. the length of flow (dl) and the difference in the elevation of
In limestone hydrology the term vadose zone is usually the inflow and outflow point (dh). The negative sign on
applied to the region lying above the water table in which the right-hand side of the equation indicates a loss in hy-
voids may contain either air or water. This is equivalent to draulic head in the direction of flow. Note that, strictly,
hydraulic conductivity is not synonymous with perme-
the soil moisture and aeration zones defined in Figure 8.1. ability; the former refers to the ability of a medium to tran-
Below the water table is the phreatic zone (equivalent to smit fluids in terms of the properties of both the medium
the ground water zone (Fig. 8.1)) in which all voids are and the fluid (specifically its density and dynamic viscosity),
completely filled with water. At any point below the water whereas the latter refers to material properties only (such as
table there is a pressure equal to atmospheric pressure plus void size and connectivity).
the pressure head (the product of the depth of water and
its unit weight). The hydraulic head at this point is the
sum of the pressure head and the eleva~ion above some assumes that flow is laminar, and while this assumption is
datum. Points of equal hydraulic head define equipotential valid for slow rates of flow it is not for high flow rates or in
surfaces within the phreatic zone. Water moves from high larger conduits (above about 10 mm diameter) where flow
to low potential along flow paths which are orthogonal to is turbulent and frictional effects become important.
these equipotential surfaces (Fig. 9.29). In lithologies with A major point of dispute in limestone hydrology is
relatively uniform permeabilities, including limestones whether subsurface drainage and the forms it produces are
such as chalk, water flow in the phreatic zone approximates best interpreted in terms of water moving through the mass
to Darcy's law (Box 9.2). This states that flow is of the rock as diffuse flow and establishing an extensive,
proportional to the hydraulic gradient, that is, the differ- continuous water table, or whether the concept of water
ence in elevation between the inflow and outflow points movement largely confined to discrete conduits (conduit
divided by the distance between them. Darcy's law flow) with no single water table is more realistic. On the
basis of the large quantity of field data that has now been
collected it is apparent that both diffuse flow and conduit
flow occur and that their relative importance varies
depending on topography and lithology. Diffuse flow is
more likely in low relief situations and where the limestone
is horizontally bedded and relatively porous throughout.
The conduit flow model is more valid, however, in region
of high relief with densely cemented limestones.
Concentrated subsurface flow is a prerequisite for the
development of caves, which can be arbitrarily defined as
cavities sufficiently large for a person to enter. Rates of
solution increase rapidly once conduits have been enlarged
to the size where turbulent flow begins to predominate.
Although the initial opening of fissures to form larger
.... conduits must occur primarily through solution, abrasion by
--.-. Flow line
sediment carried in subsurface flows can be important in
Water lable
contour the enlargement of caves especially where they are fed by
EQuipotential
surfaces
sediment-charged rivers. Caves vary from those that
permanently contain water, to those that are inactive or
experience only occasional flow. They are preferentially
Fig.9.29 Water tahle contours, equipotential sUlfaces andflow
lines. (After D. Ford and P. Williams (1989) Karst
developed in massive well-jointed limestone which has
Geomorphology and Hydrology Unwin Hyman, London. Fig. 5.7, sufficient mechanical trength to resist crushing and roof
p. 138.) collapse.
 Fluvial landforms 231

The most active debate about ·the formation of caves topics introduced here. To be recommended are Knighton
revolves around the level at which they develop relative to (1984), Morisawa (1985), Petts and Foster (1985) and
the water table. Most caves are presently located in the Richards (1982), although in spite of being published two
vadose zone and we would expect the most aggressive decades ago Leopold et at. (1964) still provide a stimu-
waters to be located in this zone. Nevertheless, such caves lating survey of some of the core concepts of fluvial
may initially have formed in the phreatic zone before being geomorphology. More recent treatments focusing on con-
drained following a fall in the water table. Many other cave cepts and experimental work are Schumm (1977) and
systems lie just below the present water table, and Schumm et at. (1987).
development here would be favoured by the high rates of The drainage basin as a basic geomorphic unit is
ground water flow at this level and the presence of rela- considered briefly by Chorley (1969) and in detail by
tively undersaturated waters. Formation deep within the Gregory and Walling (1973). An overview of the quanti-
phreatic zone is also possible since here large volumes of tative description of drainage basins is provided by Strahler
water can be transmitted through the rock at high rates of (1964), while Abrahams (1984a) presents a comprehensive
flow as a result of high pressures. Although we might review of channel network analysis. The assumption that
expect waters at depth within limestone to be saturated with stream network topology is random is challenged by
respect to calcium carbonate and thus incapable of Abrahams (1987). Other aspects of channel networks are
accomplishing further solution, the mixing of different discussed by Abrahams (1984b) and Shreve (1967), and
water bodies can produce aggressive, undersaturated Howard (1967) considers the classification of drainage
solutions (see Section 6.2.3.3). A fall in the water table patterns. The hypsometric analysis of drainage basin form
would elevate the passages produced into the vadose zone. is outlined by Strahler (1952), while Melton (1958) provides a
Phreatic caves typically have rather regular circular or pioneering attempt to relate basin morphometry to geomor-
elliptical cross-sections indicative of uniform rates of phic processes.
solution by flows completely filling the passage. In vadose The best introduction to alluvial channels is the book by
caves vertical incision of passages is important as the lower Richards (1982) and this is updated by papers in his later
part of the passage is subject to higher rates of solution than edited volume (Richards, 1987). Schumm (1985) provides
the partly air-filled upper part. If the water table is lowered an excellent introduction to channel patterns, while Hooke
and a phreatic passage is moved into the vadose zone a (1984) reviews ideas on the way meanders change
channel will begin to cut down into the base of the phreatic through time. The classic paper by Langbein and Leopold
passage. Most large caves are probably of multiple origin, (1966) is worth reading as a statement of the equilibrium
having experienced phases of both phreatic and vadose symmetrical meander model, although this has been
zone development. Structural factors, such as the orien- challenged by Carson and Lapointe (1983), Howard and
tation of bedding planes and fracture systems are certainly Knutson (1984) and Lapointe and Carson (1986) who argue
crucial in determining the overall architecture of cave that asymmetry is a natural element of meander
systems. development. Carson (1986) looks at gravel-bed meandering
Although the morphological evidence which would help rivers, while Callander (1978) represents a hydraulic
document stages of cave development is usually destroyed engineer's approach to understanding meander development.
during their formation, cave deposits can not only provide Thompson (1986) points to the role of secondary flows in
information on the most recent phases of cave history but initiating meander development.
also material which can yield palaeoclimatic data. Calcite Braided channels are considered by Miall (1977), Parker
deposits precipitated in the vadose zone are known as (1976) and Smith (1974), while Leopold and Wolman
speleothems and include the familiar vertically hanging (1957) attempt to define the conditions under which
stalactites and upward-growing stalagmites as well as meandering, braided and straight channels develop. Schumm
other more exotic forms. The dating of such speleothems and Khan (1972) report experimental work on this problem,
can provide detailed information on the more recent phases while Carson (1984a, 1984b) and Smith and Smith (1984)
of cave history by indicating periods of limestone solution discuss field examples. The complexity of controls on
and precipitation. Such data can be supplemented by sedi- channel form are emphasized by Ferguson (1987).
ments which may contain mineralogical and fossil evidence Hydraulic geometry is defined and introduced in the classic
of surface environments as well as hydrological conditions paper by Leopold and Maddock (1953), while Ferguson
within the cave itself. (1986) provides an excellent review of this topic. Other
aspects of hydraulic geometry are considered by Knighton
(1975, 1987) and Park (1977). The longitudinal form of
Further reading alluvial channels is discussed in detail by Richards (1982)
Several textbooks, each with a different emphasis, provide and riffle-pool sequences are considered by Lisle (1979)
an excellent coverage of fluvial landforms and develop the and Richards (1976). Mackin (1948), with his detailed
232 Exogenic processes and landforms

discussion of the concept of grade, and Richards (1982) Bull, W. B. (1977) The alluvial fan environment. Progress in
provide overviews of the alluvial channel system as a Physical Geography 1, 222-70.
whole. Callander, R A. (1978) River meandering. Annual Review of
Fluid Mechanics 10, 129-58.
Bedrock.channels are a neglected landform and there is Carson, M. A. (1984a) Observations on the meandering-braided
little further reading that can be recommended. Of par- river transition, the Canterbury Plains, New Zealand: Part One.
ticular value are the re-evaluation of waterfalls by Young New Zealand Geographer 40,12-17.
(1985), the discussion of incised meanders contained in the Carson, M. A. (1984b) Observations on the meandering-braided
river transition, the Canterbury Plains, New Zealand: Part Two.
book by Schumm (1977) and the discussion of the role of
New Zealand Geographer 40,89-99.
large floods in eroding bedrock channels by Baker (1988). Carson, M. A. (1986) Characteristics of high-energy 'meandering'
Turning finally to depositional landforms Lewin (1978) rivers: the Canterbury Plains, New Zealand. Geological Society
provides a useful review of floodplain forms and in a later ofAmerica Bulletin 97,886-95.'
paper he discusses the response of floodplains to changing Carson, M. A. and Lapointe, M. F. (1983) The inherent
asymmetry of river meander planform. Journal of Geology 91,
channel patterns (Lewin, 1983). The factors influencing the
41-55.
development of alluvial fans is discussed by Bull (1977), Chorley, R. J. (1969) The drainage basin as the fundamental
Harvey (1989) and Hooke and Rohrer (1979), and in terms geomorphic unit. In: R J. Chorley (ed.) Water, Earth and Man.
of specific examples in Rackocki and Church (1989). Green Methuen, London, 30-52.
and McGregor (1987) and Schumm (1977) assess the range Coates, D. R and Vitek, J. D. (eds) (1980) Thresholds in
Geomorphology. Allen and Unwin, Boston and London.
of factors that affect terrace formation, and Ritter (1982)
Cooke, R U. and Reeves, R W. (1976) Arroyos and Environ-
illustrates these with a case study. The entrenchment of mental Change in the American South-West. Clarendon Press,
ephemeral stream channels is considered in detail by Cooke Oxford.
and Reeves (1976). Ferguson, R I. (1986) Hydraulics and hydraulic geometry.
The importance of geomorphic thresholds, episodic Progress in Physical Geography 10, 1-31.
erosion and complex response in the functioning of fluvial Ferguson, R. (1987) Hydraulic and sedimentary controls of
channel pattern. In: K. Richards (ed.) River Channels: Environ-
systems, as well as other types of geomorphic systems, is ment and Process. Institute of British Geographers Special
highlighted by Schumm (1979) and illustrated in the Publication 18. Blackwell, Oxford, 129-58.
collection of papers edited by Coates and Vitek (1980). Ford, D. C. and Williams, P. W. (1989) Karst Geomorphology
Some of the original experimental work from which these and Hydrology. Unwin Hyman, London and Boston,
ideas were developed is presented by Schumm (1977) and Green, C. P. and McGregor, D. F. M. (1987) River terraces: a
stratigraphic record of environmental change. In: V. Gardiner
Schumm and Parker (1973). The magnitude and frequency et al. (eds) International Geomorphology 1986 Part I. Wiley,
of fluvial activity is discussed in the classic paper by Chichester and New York, 977-87.
Wolman and Miller (1960) and reassessed by Wolman and Gregory, K. 1. and Walling, D. E. (1973) Drainage Basin Form
Gerson (1978). The role of floods as channel-shaping and Process: A Geomorphological Approach. Edward Arnold,
events is discussed from various angles in the volumes London.
Harvey, A. M. (1989) The occurrence and role of arid zone
edited by Mayer and Nash (1987) and Baker et ai. (1988). alluvial fans. In: D. S. G. Thomas (ed.) Arid Zone Geomor-
Both surface and subsurface drainage in limestone phology. Belhaven Press, London; Halsted Press, New York,
terrains is covered in several excellent texts and in- 136-58.
numerable journal articles. Jennings (1985) provides a Hooke, 1. M. (1984) Changes in river meanders: a review of
more than adequate introduction, but for a comprehensive techniques and results of analyses. Progress in Physical
Geography 8, 473-508.
coverage of the topic see Ford and Williams (1989) or Hooke, R. Le B. and Rohrer, W. L. (1979) Geometry of alluvial
White (1988). fans: effect of discharge and sediment size. Earth Surface
Processes 4, 147-66.
Howard, A. D. (1967) Drainage analysis in geologic inter-
pretation: a summation. American Association of Petroleum Geo-
References logists Bulletin 51, 2246-59.
Abrahams, A. D. (1984a) Channel networks: a geomorphological Howard, A. D. and Knutson, T. R (1984) Sufficient conditions for
perspective. Water Resources Research 20, 161-88. river meandering: a simulation approach. Water Resources
Abrahams, A. D. (1984b) Tributary development along winding Research 20,1659-67.
streams and valleys. American Journal of Science 284, 863-92. Jennings, J. N. (1985) Karst Geomorphology. Blackwell, Oxford
Abrahams, A. D. (1987) Channel network topology: regular or and New York.
random? In: V. Gardiner et al. (eds) International Geomor- Knighton, A. D. (1975) Variations in at-a-station hydraulic
phology 1986 Part II. Wiley, Chichester and New York, geometry. American Journal ofScience 275,186-218.
145-58. Knighton, A. D. (1987) River channel adjustment - the down-
Baker, V. R. (1988) Flood erosion. In: V. R. Baker, R. C. Kochel, stream dimension. In: K. Richards (ed.) River Channels:
and P. C. Patton, (eds) Flood Geomorphology. Wiley, New Environment and Process. Institute of British Geographers
York and Chichester, 81-95. Special Publication 18. Blackwell, Oxford, 95-128.
Baker, V. R, Kochel, R. c. and Patton, P. C. (eds) (1988) Flood Knighton, D. (1984) Fluvial Forms and Processes. Edward
Geomorphology. Wiley, New York and Chichester. Arnold, London and Baltimore.
 Fluvial landforms 233

Langbein, W. B. and Leopold, L. B. (1966) River meanders - Ritter, D.F. 1982. Complex river terrace development in the
theory of minimum variance. United States Geological Survey Nenana valley near Healy. Alaska. Geological Society of
Professional Paper 422-H. America Bulletin 93, 346-56.
Lapointe, M. F. and Carson, M. A. (1986) Migration patterns of an Schumm, S. A. (1977) The Fluvial System. Wiley, New York and
asymmetric meandering river: the Rouge River, Quebec. Water London.
Resources Research 22, 731~3. Schumm, S. A. (1979) Geomorphic thresholds: the concept and its
Leopold, L. B. and Maddock, T. (1953) The hydraulic geometry applications. Transactions Institute of British Geographers NS
of stream channels and some physiographic implications. 4,485-515.
United States Geological Survey Professional Paper 252. Schumm, S. A. (1985) Patterns of alluvial rivers. Annual Review
Leopold, L. B. and Wolman, M. G. (1957) River channel patterns of Earth and Planetary Sciences 13, 5-27.
- braided, meandering and straight. United States Geological Schumm, S. A. and, Khan, H. R. (1972) Experimental study of
Survey Professional Paper 282B. channel patterns. Geological Society of America Bulletin 83,
Leopold, L. B., Wolman, M. G. and Miller, 1. P. (1964) Fluvial 1755-70.
Processes in Geomorphology. W. H. Freeman, San Francisco. Schumm, S. A., Mosley, M. P. and Weaver, W. E. (1987)
Lewin, 1. (1978) Floodplain geomorphology. Progress in Physical Experimental Fluvial Geomorphology. Wiley, New York and
Geography 2, 408-37. Chichester.
Lewin, J. (1983) Changes of channel patterns and flood-plains. In: Schumm, S. A. and Parker, R. S. (1973) Implications of complex
K. 1. Gregory (ed.) Background to Palaeohydrology. Wiley, response of drainage systems for Quaternary alluvial
Chichester and New York, 303-19. stratigraphy. Nature Physical Science 243, 99-100.
Lisle, T. (1979) A sorting mechanism for a riffle pool sequence. Shreve, R. L. (1967) Infinite topologically random networks.
Geological Society ofAmerica Bulletin 90, 1142-57. Journal of Geology 75, 178-86.
Mackin, J. H. (1948) Concept of the graded river. Bulletin of the Smith, N. D. (1974) Sedimentology and bar fonnation in the
Geological Society ofAmerica 59, 463-512. upper Kicking Horse River, a braided outwash stream. Journal
Mayer, L. and Nash, D. (eds) 1987. Catastrophic Flooding. Allen ofGeology 82, 205--23.
and Unwin, Boston and London. Smith, N. D. and Smith, D. G. (1984) William River: an out-
Melton, M. A. (1958) Correlation structure of morphometric pro- standing example of channel widening and braiding caused by
perties of drainage systems and their controlling agents. bed-load addition. Geology 12, 78-82.
Journal of Geology 66, 442-60. Strahler, A. N. (1952) Hypsometric (area-altitude) analysis of
Miall, A. D. (1977) A review of the braided-river depositional erosional topography. Bulletin of the Geological Society of
environment. Earth Science Reviews 13, 1-62. America 63, 1117~2.
Morisawa, M. (1985) Rivers: Form and Process. Longman; Strahler, A. N. (1964) Quantitative geomorphology of drainage
London and New York. basins and channel networks. In: V. T. Chow (ed.) Handbook
Park, C. C. (1977) World-wide variations in hydraulic geometry ofApplied Hydrology. McGraw-Hill, New York, Section 4-II.
exponents of stream channels: an analysis and some Thompson, A. (1986) Secondary flows and the pool-riffle unit: a
observations. Journal of Hydrology 33, 133~6. case study of the processes of meander development. Earth
Parker, G. (1976) On the cause and characteristic scale of Swjace Processes and Landforms 11, 631~1.
meandering and braiding in rivers. Journal of Fluid Mechanics White, W. B. (1988) Geomorphology and Hydrology of Karst
76,459-80. Terrains. Oxford University Press, New York and Oxford.
Petts, G. and Foster, 1. (1985) Rivers and Landscape. Edward Wolman, M. G. and Gerson, R. (1978) Relative scales of time and
Arnold, London and Baltimore. effectiveness of climate in watershed geomorphology. Earth
Rackocki, A. H. and Church, M. 1. (eds) (1989) Alluvial Fans: A Surface Processes 3, 189-208.
Field Approach. Wiley, New York and Chichester. Wolman, M. G. and Miller, J. P. (1960) Magnitude and frequency
Richards, K. (1976) The morphology of riffle-pool sequences. of forces in geomorphic proces es. Journal of Geology 68,
Earth Surface Processes 1, 71-88. 54-74.
Richards, K. (1982) Rivers: Form and Process in Alluvial Young, R. W. (1985) Waterfalls: fonn and process. Zeitschrift fur
Channels. Methuen, London and New York. Geomorphologie Supplementband 55,81-95.
Richards, K. (ed.) 1987. River Channels: Environment and
Process. Institute of British Geographers Special Publication
18. Blackwell, Oxford.
 10
Aeolian processes and
landforms

10.1 Aeolian activity glaciers and ice sheets. In such environments aeolian pro-
cesses may play an important role in landform development
Wind is a comparatively feeble geomorphic agent over and in areas of massive sand accumulations within the world's
much of the Earth's surface, but in areas free of vegetation major deserts wind is by far the predominant geomorph!c
agent.
it can have significant effects. In the arid regions of the sub-
tropics vast sand seas and extensive grooved bedrock
surfaces are a testament to the power of wind action. Deserts
10.1.1 Global distribution of aeolian landforms
are still some of the least known areas of the globe and
until comparatively recently our knowledge of desert land- The great proportion of wind-formed landscapes are con-
forms rested heavily on observations of explorers involved tained within the world's hot deserts and so we will largely
in pioneering expeditions to these inhospitable environ- confine our attention to these areas in this chapter. Aeolian
ments during the late nineteenth and early twentieth cen- features formed in periglacial and coastal environments are
turies. Increased accessibility associated with economic discussed in Chapters 12 and 13 respectively.
development, together with the global reconnaissance of The most striking aeolian landforms are dunes. These
deserts made possible by satellite imagery, is rapidly are accumulations of wind deposited particles and they can
increasing our knowledge of aeolian landforms, while assume a bewildering variety of forms. The great majority
experimental tudie and field instrumentation are begin- of dunes are composed of sand, although silt and clay-sized
ning to improve significantly our understanding of aeolian material may accumulate into dune-like features under
processes. certain conditions. Sand is not evenly spread throughout
The effectiveness of wind action is limited by a number desert regions but rather is concentrated into sand seas or
of factors, and on a global basis is a far less potent ero- ergs. It has been estimated that 85 per cent of the sand in
sional agent than fluvial activity. Compared with water, air active sand bodies is contained in ergs greater than
has a low density and viscosity so only very fine particles 32 000 km 2 in area. Large active ergs are more or less con-
can be carried in suspension, except at very high wind fined within the 150 mm mean annual isohyet, but fixed or
speeds. Moreover, vegetation greatly reduces wind speeds relict ergs (which have become inactive due to the stabili-
near the ground, and together with moisture it tends to bind zing effects of vegetation) are to be found in the subhumid
surface particles together and prevent them from being fringes of the world's arid regions (Fig. 10.1). Some active
entrained by the wind. Co.nsequently aeolian activity i ergs are o( vast extent; the largest is Rub' al Khali which
only effective in areas which lack a relatively complete covers 560 000 krn 2 of Saudi Arabia. Very extensive blankets
vegetation cover and where the surface material dries out at of silt-sized material, known as loess, are found in those
least occasionally. mid-latitude areas that were marginal to the Pleistocene ice
Most important among such areas are the world's arid sheets, and, to a more limited extent, at lower latitudes.
regions, but more localized wind action also occurs along Quartz is by far the predominant component of de ert
sandy coasts, over bare fields (especially where soil manage- sands, both because of its abundance as a rock-forming
ment techniques are poor) and in river plains containing mineral and because of its resistance to chemical decom-
migrating channels, particularly around the margins of position and abrasion. Ultimately aeolian quartz grains must
236 Exogenic processes and landforms

~--------------

~.
,~

• Active ergs

~ Relict ergs
2
,cfkm
,f(,~211
'~1000
1000 0
km
Fig.l0.l Global distribution of active and relict (fixed) ergs. The extent of relict ergs, thought to have been active during arid phases
during the Pleistocene, is approximate only. (Based partly on M. Sarnthein (1978) Nature 272, Fig. 1, p. 44 and G. Wells (1989) in: L.
Friday and R. Laskey (eds) The Fragile Environment. Cambridge University Press, Cambridge, Fig. 8.18, p. 171.)

have been derived from weathered or abraded rock, either standing of the development of aeolian landforms, yet there
as first-cycle quartz from igneous or metamorphic rocks or, is a frustrating paucity of data on the temporal and spatial
more frequently, as previously cycled grains in sedimentary variations in wind speed and direction in desert areas.
rocks. Most deserts contain extensive outcrops of quartzose Nevertheless, some generalizations are possible. The large-
sandstones, some of which themselves originated as aeolian scale properties of desert winds are determined by the
deposits. Present-day desert aeolian sands are mostly general circulation which, at the latitude of the world's hot
derived from alluvium rather than directly from bedrock deserts, is dominated by sub-tropical high-pressure systems.
weathering. This is suggested by the tendency for North Large areas are subject to fairly constant trade winds, but
African ergs to be located near the centre of basins towards regional and local factors, such as those arising from con-
which drainage systems (now largely inactive) are focused. trasts in the thermal properties of desert surfaces, can add
The distribution of erosional aeolian landforms is more to global effects. Differential heating of the surface associ-
difficult to specifiy in view of the uncertain origin of many ated with variations in albedo lead to pressure gradients
landforms ascribed to aeolian action, such as the de- which in tum generate local winds.
pressions of various sizes common in some desert areas. Of particular significance to the movement of particles
Satellite imagery is now making possible a new evaluation by the wind is the vertical variation of wind speed above
of the origin of some of the larger scale varieties of these the ground. Close to the surface wind speeds are reduced
possible aeolian features and many do indeed appear by friction, and the magnitude of this effect is largely a
explicable in terms of present or past phases of wind action. function of surface roughness. As wind speed and surface
roughness increase, the airflow becomes more turbulent
and the potential for erosion is consequently increased. Sur-
10.1.2 Wind characteristics face friction consumes energy, and as the frictional effect
A knowledge of wind characteristics is vital to an under- declines away from the surface wind speed increases with
 Aeolian processes and landforms 237

height. The rate of increase is rapid close to the surface but in the voids around the lower part of the grain and the
increases progressively more slowly at greater heights. The faster-moving air above.
actual rate of increase in any particular instance depends A much more important mechanism is the drag force
largely on surface roughness and wind speed. resulting from the difference in fluid pressure on the wind-
ward and leeward sides of grains in an airflow. This can
cause particles to move downwind by rolling or sliding, a
10.1.3 Effect of surface characteristics proce known as surface creep. Drag forces can probably
initiate movement at lower wind velocities than those
The character of desert surfaces merits consideration because
necessary for lift and they can also move particles too large
surface roughness and the size and cohesion of surface
to be lifted by the wind. The wind speed needed to promote
particles significantly affect the ability of the wind to erode.
movement through lift or drag is known as the fluid
Although sand dunes are perhaps thought of as the land-
threshold velocity.
forms most typical of arid regions, only 25-30 per cent of
This simple picture is greatly complicated by turbulence.
the total area of hot deserts is covered by sand. The re-
Airflows are rarely uniform but rather are characterized by
mainder is composed of bedrock, coarse weathered debris,
bursts of higher velocity flow which can momentarily pro-
fluvial sediments, weathering crusts, poorly developed soils,
duce a pocket of low pressure above the surface and create
and deposits of clay, silt or salt. The proportion of each of
a high potential for lift and drag. Such variations in wind
these types of urface in particular deserts reflects the com-
velocity are particularly important because the drag force
plex interaction of a variety of climatic and geological
varies with the second power of velocity; in other words a
factors.
doubling of wind velocity will produce a fourfold increase
Particle size and cohe ion, together with vegetation cha-
in drag.
racteristics, are the main properties of desert surfaces which
Whether a particle will move at a given wind speed
influence their susceptibility to wind erosion. In most cli-
depends on many factors in addition to its size; these in-
matic regions the surface is covered with soils rendered
clude its density and shape, the packing of grains around it
cohesive by the binding effects of moisture, humus and
and the amount of cohesion between particles. Neverthe-
clay-sized particles, but in hot deserts all of these elements
less, for dry particles of similar shape and density there is a
are, to a greater or lesser extent, lacking. The susceptibility
fairly direct relationship between size and fluid threshold
of soils to erosion by the wind is known to be related to
velocity since there is a strong correlation between drag and
their water content, desiccated soils being much more prone
average wind speed (Fig. 10.2). Grains larger than about
to erosion. Moisture retention and interparticle bonding are
1 mm across will rarely be moved by drag alone since the
more marked in clay- and silt-rich materials and these are
required wind speeds are hardly ever attained under natural
consequently more resistant to wind erosion.
conditions. For grains smaller than about 0.6 mm across,
the threshold velocity necessary for movement actually
increases as particle size decreases. One rea on for this is
10.1.4 Sediment entrainment and transport
the reduced surface roughness, and therefore lower level of
Both water and air are fluids and there are close similarities turbulence, generated by surfaces composed of fine material.
between the mechanism of sediment entrainment in air- Another reason is the greater interparticle cohesion charac-
flows and water flows (see Section 8.3.2). Movement is teristic of fine sediments which provides a greater resist-
resisted by the weight of particles, and by friction and inter- ance to entrainment. The sorting and packing of grains also
particle cohesion. Movement is induced by drag and lift affects their susceptibility to entrainment as small particles
forces, and the impact of grains already in motion. can be sheltered from the airflow by larger grains.
Lift forces can cause particles to jump up into the airflow Very small particles less than about 0.1 mm acr~ss are
in much the same way that the wings of aeroplanes normally capable of being carried away in suspension once
generate lift. In the case of a wing, lift occurs because its set in motjon by the wind. Somewhat larger grains up to
cross-sectional form is such that air flowing over its upper around 0.4 mm in diameter can be temporarily lifted into
surface is forced to take a longer path, and therefore moves the airflow before returning to the surface. As a result of
faster, than air flowing over its lower surface. This causes a this process of saltation, once lift and drag have initiated
pressure difference between the upper and lower wing particle movement grains downwind start to be bombarded
surfaces which induces lift. Imagine a sand grain pro- by those particles already in motion. This mechanism of
truding slightly above the average level of a sand-covered ballistic impact applies an additional forward momentum
surface. Depending on the exact form that the grain presents to surface particles and means that grains of a particular
to the wind it might also produce an acceleration of air over size can be set in motion at a wind velocity lower than that
it. In addition a pressure difference might result from the required to initiate movement. The difference between this
contrast in wind velocity between the almost stationary air impact threshold velocity and the fluid threshold velocity
238 Exogenic processes and landforms

8 return to the air. They may also cause other grains

0·6 -+--+----+----+-----+---r----if--_'---l
disturbed by their impact to saltate. This process gives rise
, ~
'0
o to a cloud of saltating grains extending a few centimetres
\ Qi-
I", \ ~T", above the surface. Most saltating grains travel within
.s 04+-\H\--t----+---jlC--'"' 7 - - - - - t - - - - - - i 6'§ .s Ql 10 mm of the surface with individual horizontal leaps of
~
'0 o~
" U
..c ... around 0.5-1.5 m. The height which saltating grains reach
o
~ '"
..c '"
:::l
~ Ql depends on particle size, wind speed and surface charac-
OJ
-Ql 0>
~ _.0 teristics. In general height increases with greater wind
o 0·2 +--+-->:--t--r"""7"''--t----t-----f------j 4 E 10
.- E
~ E
speeds and smaller particles, but as larger particles may
5.0
0.'- bounce better they may occasionally rise out of the salta-
~
ting cloud. Saltating grains also reach greater heights over
O+--+---t-----t----+----t-----i
hard, non-sandy surfaces apparently because they bounce
Creep more effectively on such surfaces.
The major factors influencing the rate of aeolian trans-
=~ormal
--Extreme
port are wind speed, air density (which varies slightly with
altitude and temperature), particle size and surface charac-
teristics. Experimental investigations have shown the rate
of sand transport to be proportional to the third power of
Dunes-+lCIZllS~~ wind speed. This relationship suggests that most sand is
Sand sheets and ripples
moved during periods of very high wind velocity. Very
strong winds, however, are much less frequent than those
of moderate speeds so it seems likely that most sand trans-
0·01 0·1 0·4 1·0 2·0 3·0 port in fact occurs during relatively frequent periods of
Grain Diameter (mm) moderately strong winds of around 12-18 m S-l (Fig. 10.4).
Fig.10.2 Relationships between grain size andfluid and impact
threshold wind velocities. Also shown are the relationships
between grain size and characteristic modes of wind transport
and associated depositional forms . (Modified from 1. A. Mabbutt
(1977) Desert Landforms. Australian National University Press,
Canberra, Fig. 65, p. 219, based on R. A. Bagnold (1954), The
Physics of Blown Sand and Desert Dunes. Methuen, London, Fig. 50
28, p. 88 and R. L. Folk (1971) Sedimentology 16, Fig. 16, p. 43.)
37.5

40
is small for fine particles but is quite significant for larger
Vi ...c
grains (Fig. 10.2). Once a gust of wind has caused an initial >-
co Q)

movement in some particles further motion can be tem-

"C
25 30 ~;::
c ~ 'co
porarily sustained by ballistic impact even after the wind .g E~
speed has dropped below the fluid threshold velocity. ~
:J
'g:Q
0 20 ~ ~
The proportion of total movement occurring by surface '+- E
12.5 o ....
creep becomes less with decreasing grain size, and even for ~-
co
relatively coarse sands it rarely exceeds 25 per cent. Salta- 10 a:
tion is the major mechanism of sand movement, and in view
of its importance we need to examine the process more
closely. On rising into the airflow, saltating grains en- 5 10 15 20 25
counter increasingly higher wind sp.eeds and they are carried Wind speed (ms-1 )
downwind (Fig. 10.3). They return to the surface along a
trajectory determined by a balance between gravity and Fig.10.4 Relationship between wind speed and sand movement.
drag and usually strike the surface at an angle of between 6 The frequency of winds of different speeds is based on a
and 12°. After impact saltating grains may bounce and hypothetical, but realistic, distribution through an annual cycle.
The estimated rate of sand movement (t m-width- I a-I) is based on
one of a number offormulae derived from experimental data.
(After A. Warren (1979) in: C. Embleton and 1. Thomes (eds)
Process in Geomorphology. Edward Arnold, London, Fig. 10.8,
p.335.)

Fig.10.3 Trajectory of a single saltating grain.

 Aeolian processes and landforms 239

10.2 Aeolian erosion

10.2.1 Deflation and abrasion
Aeolian erosion involves two processes; deflation, which is
the removal of loose particles by the wind, and aeolian abra-
sion, which is caused by the bombardment of rock and other
surfaces by particles carried in the airflow. Deflation
involving sand-sized particles is rather localized since sand
grains cannot be moved great distances except over long
periods of time. Silt, and especially clay-sized material, on
the other hand, can be lifted by turbulence and carried in
suspension in the atmosphere while the very finest material
can be carried great distances in dust storms.
The world's hot deserts are a major source of atmospheric Fig.10.5 Gravel lag deposits, Skeleton Coast, northern
dust (see Section 15.6). Accurate estimates are difficult to Namibia.
obtain but it has been estimated that between 130 and
800 Mt of material is deflated from the continents annually,
with the Sahara alone contributing between 60 and 200' Mt.
Deflation is a temporally and spatially concentrated pro-
cess. Even in arid regions it is largely restricted to areas
where surface conditions and wind speeds are particularly
favourable. Most of the material deflated each year is
removed in major dust storms, each usually lasting only a
few days, and satellite reconnaissance has provided a
potent technique for tracking these events. Some of this
material is redeposited elsewhere in desert areas forming
loess but much of the finer material is carried out over the
ocean and makes a significant contribution to sedimentation
in the world's ocean basins.
Rock- and boulder-covered surfaces can be abraded by
particles of a range of sizes. Although it has been widely
thought that sand-sized material is the most effective, some
research has pointed to the role of fine silt-sized particles
and dust, especially in producing fluted surfaces. The rate
of abrasion is maximized where high velocity winds cross
an abundant supply of hard sand grains upwind from soft,
friable lithologies. As sand particles are only carried close
to the ground, even at high wind speeds, abrasion caused by
this calibre of material is confined to within I or 2 m of the
surface.

10.2.2 Erosional landforms

Although aeolian erosion may be active over alluvial plains
and on beaches, wind-eroded landforms are rarely preserved
in such environments because of destruction by fluvial pro-
cesses or wave action. Only in arid areas, where other
denudational agents are lacking or weakly active, are Fig.10.6 Experimental plot in Arizona, USA, where the
aeolian erosional landforms abundant. A significant propor- surficial gravel lag deposit has been removed in order to
determine the rate at which if is re-established.
tion of many deserts is covered not by sand, but by lag
deposits comprising particles of gravel size or coarser
(Fig. 10.5). Lag deposits usually form a thin layer lying desert pavements, or stone pavements. It was long thought
over predominantly finer material (Fig. 10.6). Where the that they form by deflation, with the wind removing finer
stone cover is continuous such surfaces are known as particles from poorly sorted deposits such as alluvium.
240 Exogenic processes and landforms

Recently, however, it has been argued that the observed particles are carried in vortices in flow lines. Ventifacts
distribution offine and coarse particles is better explained by may also be disturbed or rolled into a new position during
the accretion over large areas of aeolian sand and silt which formation, thereby realigning them with respect to a single
progressively infiltrates into alluvial gravels. prevailing wind.
Bedrock surfaces may also exhibit evidence of wind
10.2.2.1 Small-scale forms abrasion in the form of small-scale pits, flutes, grooves and
A characteristic feature of stony desert surfaces is the polished surfaces, but such fOnTIS are often difficult to ascribe
presence of faceted cobbles and pebbles called ventifacts. with certainty to aeolian action as similar features can be
Their precise mode of development is disputed but abrasion produced by running water and chemical weathering.
by dust and silt (rather than sand-sized particles) is prob-
ably the main mechanism. Facets are produced at angles of 10.2.2.2 Intermediate-scale forms
between 30 and 60° facing the prevailing wind and are se- Much larger than these small-scale bedrock features are a
parated from the protected lee side by a sharp edge (Fig. variety of grooved forms and shallow depressions with
10.7). More than one side may be wind abraded and three- dimensions of tens to hundreds of metres. The most charac-
facet pyramid-shaped venifacts (called dreikanter) are teristic grooved form is the yardang, a streamlined parallel
particularly common. Abrasion of more than one side may ridge usually less than 10 m high and 100 m or more in
indicate winds prevailing from more than one direction, but length aligned with, and typically tapering away from, the
experimental studies demonstrate that ventifacts can be direction of the prevailing wind (Fig. 10.8). Yardangs are
formed by essentially undirectional winds. In this case most commonly developed in soft lithologies such as lacus-
erosion occurs simultaneously on all sides as dust and silt trine sediments and are numerous in some desert lake beds
(Fig. 10.9), but in areas such as western Egypt they occur in
extremely resistant granites and quartzites.
The extent to which yardangs are wind-formed features
is not universally agreed. It has been argued that some
merely represent a pre-existing bedrock relief only slightly
modified by wind action. Nevertheless their consistent
orientation, which can frequently be correlated with the
dominant wind direction, is strong evidence for a primary
aeolian origin. Their formation appears to depend essen-
tially on the excavation of material from the troughs lying
between the yardangs. They are frequently undercut at the
base and this would accord with a higher rate of wind

Fig.10.8 A mud yardang in the Kharga Depression, Egypt. The

unidirectional prevailing wind is from the right. Sand-hlasting is
confined to the hlunt, windward face, and the leeward tapering of
the yardang indica?es the importance of erosion hy fine particles
carried in suspension in secondary flows across the long leeside
tail. (Photo and interpretation courtesy M.I. WhitneyJram M.I.
Fig.10.7 Ventifaet on a gravel lag sUiface in the Namih Desert Whitney. (1985) Journal of Geological Education 33, Fig. J,
in a region of strong coastal winds. p.94.)
 Aeolian processes and landforms 241

Fig.10.9 Extensive yardang field in the Lut Desert, Iran. The grooving evident on the image, which covers an area about 140 km
across, is a result oflines of very large yardangs up to 80 m high and spaced at least 100 m apart. The yardangs are formed in horizontally
bedded lacustrine deposits and trend NNW parallel to the prevailing wind. (Landsat image courtesy A. S. Walker.)

abrasion near the ground. Weathering and surface runoff Similar in scale to yardangs are the shallow depressions
may be significant in modifying yardangs once they begin found across many desert. regions of low relief. These have
to form and recently one study has suggested that deflation been given a variety of local names but are generally refer-
of unconsolidated trough material may be of more import- red to as deflation hollows. They exist in a continuum of
ance than abrasion in their development. Whatever their sizes from small forms less than I m deep and only a few
exact mode of formation, yardangs seem to be confined to metres across, to large features which grade into the macro-
the very arid core regions of deserts where there is com- scale basins discussed below. The development of deflation
paratively little sand and where erosion by dust and silt hollows is clearly influenced by those factors which control
particles occurs in a unidirectional prevailing wind. the process of deflation and they tend to form where high
242 Exogenic processes and landforms

wind velocities are associated with bare surfaces covered

by relatively fine-grained, desiccated sediments. The problem
is to explain the differential erosion involved and this could
be due to a variety of local conditions, such as a lower sedi-
ment moisture content or disturbance of the vegetation
cover. In some areas, however, deflation hollows are regu-
larly spaced and this suggests some control by waves in the
airstream. Much research remains to be done before these
features are fully understood.

10.2.2.3 Large-scale forms

The occurrence of large enclosed basins in deserts has been
known since the first explorations of these regions, but the
extent and regularity of large-scale grooving in bedrock has
only been fully appreciated since the advent of satellite
imagery. The basins range from landforms a few metres
deep and upwards of 100 m across, such as the pans which
are so abundant in parts of southern Africa, to very large
features over 100 m deep and more than 100 km across.
The smaller, shallow basins probably represent localized
deflation; some are orientated along drainage lines whereas
others are located in troughs between dunes. But in both
...,.,.nT7Tn Escarpments
situations their long axes are aligned with the prevailing
-100m -- Watertable depth
wind.
Very large basins are much more complex landforms and ~ Dunefields o,
almost certainly owe their origin to the interplay of several
processes. In many cases their gross form appears to be Fig.10.10 Large enclosed basins in the western Egyptian
related to tectonic activity, and this can be established by Desert. Escarpments rim the northern margins of the basins but
there is no evidence of significant faulting in the area, only gentle
the presence of bounding faults. In such tectonic basins
warping. The southern basins are the oldest and were probably
aeolian activity may only slightly modify the existing form. initiated by fluvial erosion as far back as the Eocene, but the
But if surface deposits are susceptible to removal by wind, deepening of these basins once enclosed, and the development of
deflation may be a key process in maintaining their relief, the northern basins, is attributed largely to prolonged deflation.
or even increasing it, since this provides the only mechan- Judging by the relationship between basin depth (shown by spot
heights for low points) and depth of the water table (indicated by
ism capable of completely removing material from an
contours) deflation has been limited by the presence of ground
enclosed basin. Where faulting is absent there is frequently water. Retreat of the south-facing escarpments has occurred
much uncertainty about the origin of these features, some through spring sapping and weathering, the resulting sediment
of which represent the removal of enormous amounts of being deflated. This combination of denudational processes is
material. Perhaps the most remarkable concentration of clearly most effective since the Qattara Depression is cut into
Pliocene rocks and is therefore primarily of Quaternary age. This
large basins is to be found in Egypt where they cover more
implies d mean rate of lowering of around 90 mm ka- I , a rate
than 70000 km 2 and have an average maximum depth of comparable to fluvial denudation. (Modifiedfrom 1. A. Mabbutt
250 m (Fig. 10.10). The Qattara Depression, the deepest of (1977) Desert Landfonns, Australian National University Press,
the basins, reaches a depth of 134 m below sea level, and Canberra, Fig. 62, p. 212.)
has a volume of 3200 km 3 .
Large-scale bedrock grooving clearly visible on satellite wind abrasion and deflation. However, since they also fol-
imagery has been widely attributed to aeolian action. Such low the trend of bedrock lineations thought to have a tec-
grooves are the major landform over a 90000 km 2 area on tonic origin, it is possible that winds may have been in part
the south-eastern margin of the Tibesti Plateau in North channelled along pre-existing lines of structural weakness
Africa, and an aeolian origin is strongly suggested by the once these had been initially excavated.
close match between the regional atmospheric circulation
and their gently curving NE-SW trend. The grooves are 10.3 Depositional landforms
between 0.5 and 1 km wide, several tens of kilometres long
and are spaced 0.5-2 km apart. They are eroded into sand-
10.3.1 Basic depositional forms
stone, and are generally assumed to have formed in a similar
way to yardangs, that is, largely through a combination of Although capable of movement by surface creep and salta-
 Aeolian processes and landforms 243

research focused on Saharan dunes, that for a given grain

size the spacing of aeolian depositional forms falls into
a hierarchy of three discrete size classes consisting of
ripples, dunes and megadunes with no transitional forms
(Fig. 10.13). He suggested that the contrasting spacing of
dunes and draas was related to different scales of secondary
airflows. The dimensions of each of these depositional fonns
were held to be related to grain size, with coarse sands giving
rise to the largest forms in each category (Fig. 10.13). This
relationship was interpreted as being a result of the mini-
mum drag velocity required to move sand of a given size,
which, in turn, constrains the minimum spacing and dimen-
sions of depositional forms of each type.
For a time this idea was widely accepted and seemed to
provide a useful basis for classifying aeolian depositional
forms. More recent research, however, has led many geo-
morphologists to question whether this size hierarchy is
universally valid. Data from the Namib, Kalahari and
Australian Deserts, for instance, have failed to show a clear
discontinuity in the gradation from dunes to megadunes
(Fig. 10.13). Nevertheless, it is acknowledged that such a
size discontinUity may occur in dune types which develop
normal to a single prevailing wind direction, and in dune
fields where there is a large range in grain size.

10.3.1.1 Ripples
Ripples are asymmetric in cross-section with windward
slopes around 10° and lee-side gradients near the angle of
repose of dry sand of 30-35°. According to the model
presented by R. A. Bagnold they develop from very slight
Fig.lo.n Ripples on the sUlface of a dune, north of Walvis irregularities in the sand surface through a combination of
Bay, Namibia. The scale is indicated by the lens cap in the bottom
right corner. Regular ripples such as these form in well-sorted
surface creep and saltation. As saltating grains strike the
sand, whereas irregular ripples are produced in poorly sorted surface at a shallow angle they dislodge more grains on the
sediments. windward side of an irregularity than on the leeward side
which is protected from the flight path of the saltating
grains (Fig. 10.14). Ballistic impacts release grains from
tion, sand grains spend the vast majority of time in storage the windward side which either creep up the windward
in,sand accumulations which vary enormously in size and slope to the crest, or saltate a characteristic distance to add
form. The smallest depositional features are called ripples to the sand accumulation forming the next ripple down-
and consist of regular, wave-like undulations orientated at wind. The most common saltation distance depends on
right angles to the direction of the prevailing wind (Fig. particle size and wind speed and will influence ripple spacing.
10.11). The dimensions of ripples increase with particle As sand is constantly being eroded from the windward side
size, their heights ranging from 1 to 500 mm and wavelengths of ripples and accreted to the leeward side a series of ripples
from 0.01 to 5 m. Dunes are much larger depositional forms will migrate downwind while maintaining their charac-
having typical heights of 5 to 30 m and wavelengths of 50 teristic spacing. This process does not lead to an indefinite
to 300 m. Some dunes, however, attain even greater dimen- increase in ripple height because the ripple crest gradually
sions with heights of up to 400 m and wavelengths up to builds into a zone of higher wind speeds until eventually
4km (Fig. 10.12). The tenns draa, or megadune, are some- deposition is balanced by erosion.
times applied to these very large depositional fonns. Measure- This simple model works well for sand with a more or
ments from ergs around the world reveal a close relationship less uniform grain size, but most natural sands are only
between the height, width and spacing of sand dunes. This moderately well sorted. In such cases the large grains tend
suggests that dune forms are closely adjusted to controlling to move slowly by creep and they accumulate just on the
variables. leeward side of the ripple crest where they are protected
In the early 1970s I. G. Wilson argued, on the basis of from the impacts of smaller saltating grains. In this way
244 Exogenic processes and landforms

Fig. 10.12 Megadune scale forms of the southern Namib Sand Sea, Namibia. Three major dune types are represented. Compound
barchanoid ridges (B) predominate along the Atlantic coast, while compound and complex linear forms (L) predominate inland. Lines
of large star dunes (S) are visible in the centre of the image. The area covered is about 170 km across. (Landsat image courtesy A. S.
Walker.)

ripple height can increase substantially as much higher from the aerodynamic effects of the wind on sand, but such
wind speeds are required to move large grains. The presence forms are almost invariably obscured by ripples generated
of coarse particles on ripple crests provides a better bounc- by ballistic impact.
ing surface for saltating grains which consequently have
longer trajectories, and consequently ripple spacing increases. 10.3.1.2 Dunes
Where very coarse sand is subject to very strong winds For a dune to form a patch of sand must first begin to accu-
megaripples may form with wavelengths of 5 m and heights mulate. This occurs where the wind speed is reduced by an
of 0.5 m (Fig. 10.15). Low relief ripples may also originate increase in surface roughness, or by primary instabilities in
 Aeolian processes and landforms 245

Ripples Dunes Megadunes

A
40
ttl
Q)
30
~
'c:J 20
Qi
Q.
0 10
z
0
0'04 0'64 10 160 2560

B
£E 2·0
°E
~::: 1-0
Q) Q) 0'8

.~~ 0·6
.!: ~ 0'4
<0'-
'- Q)
<.::l Q. 0.2

0'04 0·64 10 160 2560

Wavelength (m I

Fig.10.13 Relationships proposed by I. G. Wilson between

a~'ealfrequency and wavelength (spacing) (A), and between grain
Slze and wavelength for aeolian depositional forms (B). Note that
ripples, dunes and megadunes form three discrete clusters, but
that Australian dunes plot between the dune and megadune
groups identified by Wilson. (Based on I. G. Wilson (1972)
Sedimentology 19, Figs. 2 and 3, p. 13 and R. 1. Wasson and R.
Hyde (1983), Earth Surface Processes and Landforms 8, Fig. 1, p.
302).)

Trajectory of impacting

------ - ---
saltating grains

- __~-- ....._ 0-0- 0-0/

-- - - ..... --....... :&3 S'"-
~. Fig.10.15 Large ripples with a wavelength of more than 1 m
developed in very coarse sediments on the flank of a linear
Surface creep of '\ --...=-- megadune, Skeleton Coast, northern Namibia (lens cap indicates
larger grains builds Protected leeside
up ripple crest of ripple scale).
Fig.lO.14 Mechanism of ripple formation according to the
model of R. A. Bagnold.
10.16). Where the airflow descends to the ground there is
more sand movement. Deposition occurs in the zones of
the airflow. Once formed the patch of sand grows by trap- converging, upward moving air forming dunes which grow
ping saltating grains which are unable to rebound on impact vertically until an equilibrium height is attained where the
as readily as they are on the surrounding stony surface. For wind velocity is sufficient to remove sand from the dune
this effect to work the sand body must be broader than the crest as rapidly as it is supplied from upwind. In addition
flight lengths of saltating grains. A width of around 1-5 m the wave-like form of the airflow itself may exert a control
seems to be a critical lower limit for the growth of sand
bodies and thus represents the limiting size for dunes. As a Prevailing wind

result of the separation and deceleration of the airflow on • crest

the lee side of the growing sand body, sand accumulates ----\-. .

more rapidly than it is removed and consequently the dune

increases in size.
~~
The basic shape of a dune formed by winds from a single bUried slip faces
prevailing direction is asymmetric in cross-section. Such represented by
bedding planes
dunes have a gently inclined windward slope at a typical
Fig.10.16 Basic components of the cross-profile of a transverse
angle of 10-15 0 separated by a sharp crest from a much
type dune with a single slip-face orientation. Linear dunes
steeper leeward slip-face sustained close to the threshold contain two opposing slip-face orientations and star dunes
angle of stability of dry sand of between 30 and 35° (Fig. multiple slip-face orientations.
246 Exogenic processes and landforms

on dune height. Sand eroded from the windward side is is considered in more detail below. Dunes with a single
deposited on the leeward slip-face so dunes move in the slip-face orientation are associated with unidirectional winds.
direction of the prevailing wind. Rates of dune movement Their axes are orientated normal to the prevailing wind, and
depend both on dune type and size and the frequency and for this reason they are often referred to collectively as
strength of winds necessary to cause sand movement, but transverse dunes. Simple, straight, parallel dunes called
10-20 m a-I is typical. transverse ridges (Fig. 10.l7(A» frequently grade into
The overall form of megadunes is similar in cross- more common barchanoid ridges (Fig. 1O.17(B», com-
section to that of dunes, but in detail is often complicated posed of coalesced crescentic forms, and into individual
by the presence of superimposed dunes. The formation of crescent-shaped barchans (Figs lO.l7(C), 10.18). Highly
megadunes appears to be more clearly related to primary elongated forms with two, more or less opposing, slip-
instabilities in the airflow than is the case for dune-sized faces, are called linear dunes although the term longitudi-
features. Since sand movement is confined to the surface of nal dune is used by some researchers for certain types
a dune, the area of which decreases as a proportion of volume (Figs. 10.17(D), 10.19). In contrast to transverse dunes net
as dune size increases, large dunes move much more slowly sand transport is parallel to the crest line. There is
than small dunes. Measurements on a medagune in Peru disagreement about the origin of linear dunes, but most
have indicated a rate of movement as low as 0.5 m a-I, researcher now consider that they develop where there are
whereas .smaller superimposed dunes were found to be mov- two obliquely converging prevailing winds. Linear dunes
ing at 9 m a-I. Indeed these slow rates of movement mean can attain lengths of tens of kilometres and they may
that very large megadunes must take hundreds of years to
attain an equilibrium form.

10.3.2 Classification of dune morphology

Our discussion so far has been confined to the general
characteristics of ripple and dune forms in cross-section.
Adding the third dimension of spatial organization we see
that, whereas ripples are comparatively simple features
forming a succession of more or less parallel ridges at right
angles to the prevailing wind, dunes occur in a great variety
of much more complex patterns. The extent and regularity
of these patterns at the megadune scale have only been
fully appreciated with the advent of satellite imagery and
the global overview that it makes possible (Fig. 10.12).
Any categorization of features as variable as dunes is to a
certain extent arbitrary, but a reasonably consistent classifi-
cation can be made on the basis of two key characteristics -
the overall shape of the dune and the position and number
of slip-faces. Plan shape is an obvious factor to consider,
but slip-face characteristics are also important since they
provide information on the nature of the formative winds.
Wind direction and velocity, sand supply and the presence E Reversing

of vegetation or topographic obstacles are the most important

factors influencing dune morphology. None the less, it is
convenient to distinguish between free dunes, whose form
is primarily a function of wind characteristics, and impeded
dunes whose morphology is influenced significantly by the
effects of vegetation, topographic barriers or highly localized
sediment sources.
Star G Dome

Fig.10.17 Morphology of the major types offree dune. Arrows

10.3.3 Classification of free dunes indicate the p'robableformative prevailing wind direction(s),
although note that the origin of linear dunes is controversial (see
The major types of free dune are classified in Table 10.1. Section 10.3.5.2). (After E. D. McKee (1979) United States
The primary wind regime thought to be associated with Geological Survey Professional Paper 1052, Figs. 3-5,7,10-12,
each type is given but the nature of the formative airflows pp.11-13.)
 Aeolian processes and landforms 247

Table 10.1 Classification of basic type of free dune

NUMBER AND GEOMETRY OF SLlP- INFERRED PRIMARY WIND DUNE TYPE MORPHOLOGY
FACE ORIENTATIONS REGIME

One; unidirectional Unidirectional § Transverse ridge Asymmetric ridge

<2 Barchanoid ridge Row of contiguous crescentic forms
~ Barchan
Cl.l
Crescentic form
>
Two; opposing Bidirectional; oppo ing ~ Reversing Asymmetric ridge
C<l
at -180°
t=:
Two; opposing Bidirectional; obliquely Linear Symmetric ridge; straight to sinuous in plan
converging?

Three or more; multidirectional Multidirectional Star Central peak with three or more arms

None Dome Circular or elliptical mound

coalesce downwind forming Y -shaped junctions (Fig. nately as wind direction changes periodically.
10.20). Dunes with several slip-faces orientated in different direc-
Reversing dunes (Fig. 1O.17(E)) are similar to linear tions have a roughly pyramidal shape but with elongated
dunes in that they have two slip-face orientations, but they and often irregular arms (Figs 10.17(F), 10.21). These forms,
are more appropriately regarded as a type of transverse dune. which are attributed to strong winds blowing from several
They are formed where two prevailing winds of similar different directions during an annual cycle, have a variety
strength and duration blow normal to the dune axis from of names, but will be referred to here as star dunes. Other
opposite directions. The opposing slip-faces develop alter- sand accumulations including sheets, streaks and dome dunes

Fig.IO.18 Barchan west of the Salton Sea, California, USA (Photo courtesy K. Mulligan.)
248 Exogenic processes and landforms

Fig.10.19 Complex linear megadune at the northern extremity

of the Namih Sand Sea. The vegetated area in the middle
foreground marks the site of the Kuiseh River which flows
occasionally and limits northward dune migration. (Photo
courtesy 1. T. Teller.)

exhibit no external slip-face. Sheets and streaks rarely give

rise to significant topographic features but domes (Fig.
10.17(G» are true dune forms, in some cases attaining
significant dimensions.
In addition to these basic types are compound dunes
comprising two or more of the same basic dune type which
have either coalesced or are superimposed. The latter can
occur when, for instance, a dune-sized barchan rides up the
back of a much larger megabarchan. Complex dunes can be
formed when different basic dune types are associated. For
example, star dunes are commonly seen surmounting linear
Fig.1O.20 Y-junction hetween linear dunes, South Australia.
dunes while barchans may occupy the troughs between
them.

dunes being much less common. Regional contrasts between

10.3.4 Global occurrence of free dune types
deserts are interesting; for instance linear dunes are par-
Only since the advent of the satellite reconnaissance of ticularly predominant in the Kalahari, whereas star dunes
dune fields has it been possible to gain any idea of the are numerous in the north-east Sahara.
relative abundance of the different types of dune on a
global basis. Previously our knowledge of the distribution
of dune types was provided only by rather patchy air photo- 10.3.5 Development of free dunes
graph coverage and by even more limited ground obser- The two most significant factors determining free dune mor-
vations. Although satellite imagery is extremely valuable it phology are wind regime and sand supply (Fig. 10.22). The
has limitations since some dune types cannot be readily importance of sand supply is illustrated by the occurrence
differentiated at the resolutions available. of barchans which appear only to develop on hard, rocky
Some idea of the relative abundance of the principal surfaces where sand is scarce. Under a similar undirectional
dune types in most of the world's major deserts, including wind regime abundant sand typically gives rise to series of
some fixed ergs, can be gained from Table 10.2, which is barchanoid ridges. Sand availability is also crucial in deter-
based on an extensive survey conducted by the United mining whether megadunes develop. Australian deserts are
States Geological Survey. It should be noted that the dune remarkable for their lack of megadunes, the Simpson Desert,
types listed are confined to those that can be distinguished for instance, being characterized by low linear dunes se-
on Landsat imagery and that major dune fields in Australia parated by relatively sand free inter-dune troughs. It has
and central Asia are omitted. Nevertheless the data indicate been estimated that if all the sand of the Simpson Desert
that sand sheets and streaks, transverse dunes and linear were spread out evenly it would be only 1 m thick. Similar
dunes are all quantitatively significant, with star and dome calculations for the Great Erg Oriental of Algeria, where
 Aeolian processes and landforms 249

Fig.l0.21 Star dune, United Arab Emirates. Note the orientation of the ripples and slip-face indicating a left-to-right prevailing
wind shaping the arm of the star dune in the foreground.

50
Fig.l0.22 Relationship of major dune types to measures of
sand availability (equivalent sand thickness) and variability in
wind direction. Equivalent sand thickness is the thickness that
would be attained by dune sand in an area if it were spread
40
evenly. Wind direction variability is measured by calculating the
E Star frequency and strength of winds blowing from different directions.
III
III
High values indicate a single predominant wind which accounts
(l) Transverse for a large proportion of sand movement, whereas low values
~ 30
()
indicate a highly variable wind regime. The majority of examples
:s of each dune type plot within the indicated ranges. (After R. 1.
"c:
~
III
Wasson and R. Hyde (1983) Nature, 304, Fig. 4, p. 339.)
C 20
(l)

~ The morphology of dunes represents a response to the

'S
W
0- complex interaction between secondary flows in the wind
10, and the form presented by the sand surface. Cause and
effect are so intimately related that it is frequently difficult
Linear Barchan to establish whether particular types of airflow around dunes
o +-__:.-:..::..:I!.:::..:==.__----.-_~~==!!!!.., are instrumental in their formation or merely a result of
o 0-2 0-4 0-8 ,.0 their existence. Empirical data on the configuration of air-
Multidirectional Unidirectional flows associated with dune forms are sparse, especially for
Wind direction variability larger-scale features, and there is still much to be learnt
about the detailed mechanisms of dune formation.
megadunes are abundant, give a much greater average depth Examination of the relationships between present-day
of 26m. wind regimes and dune patterns can provide valuable insights
 N
til
o
t!1
~
o
(JQ
t'D
='
;::;0
"C
""l
o(')
t'D
[Il
[Il
t'D
Table 10.2 Relative abundance of free dune types in some major world deserts [Il

SOUTH NORTH NORTH-EAST WEST AVERAGE

='
Cl-
THAR ALA SHAN SAHARA SAHARA SAHARA SAHARA ;'
- ='
Cl-
Linear type dunes (total) 13.96 1.44 24.08 22.84 17.01 35.49 30.54 ~
""l
Simple and compound 13.96 1.44 24.08 5.74 2.41 35.49 23.26 :3
[Il
Feathered 3.56 1.13 0.91
With crescentic
superimposed 4.02 7.32 1.46
With stars superimposed 9.52 6.15 4.92

Transverse type dunes (total) 25.61 27.01 28.37 33.34 14.53 19.17 24.09
Single barchanoid ridges 8.96 8.62 4.08 0.06 0.65 3.80
Megabarchans 7.18 1.98 0.92
Complex barchanoid ridges 16.65 18.39 24.29 26.10 12.55 18.52 16.45

Star dunes 2.87 7.92 23.92 5.00

Dome dunes 7.40 0.86 0.80 0.90

Sheets and streaks 31.75 33.56 45.44 13.56 23.24 67.82 47.54 35.92 39.25 45.34 38.34

Undifferentiated 6.71 4.50 1.12

Figures are for percentage area covered of sample Landsat images by the particular dune type.
Source: Modified from S. Fryberger and A. S. Goudie (1981) Progress in Physical Geography 5, Table 1, p. 423, based on analysis of maps in C. S. Breed et al. (1979) United States Geological
Survey Professional Paper 1052, 305-97 by A.S. Goudie.
 Aeolian processes and landforms 251

into the genesis of particular dune forms. Nevertheless, this

approach is not without difficulties. Some dune systems are
so large that they have been developing over periods of
thousands of years. It is unlikely that wind regimes have
remained constant over such long periods and today we
may just be witnessing the slow modification of patterns
formed by earlier wind regimes. This problem is most
relevant to megadune scale features since larger forms will .....

take longer to adjust to changing wind patterns that smaller

dune-sized forms. In some areas, such as the western Sahara,
dunes are orientated quite differently from the megadunes
on which they lie and it is difficult to explain this without
invoking a change in wind direction. Fig.10.23 A transverse ridge distorted by longitudinal vortices.
A second problem is the existence of relict dunes. Large These cause variations in wind speed along the ridge crest which
areas marginal to presently arid regions are covered by in turn lead to variations in the height of the ridge. Since lower
inactive dunes stabilized by vegetation. The patterns ex- dunes move more rapidly than higher dunes, the lower sections of
the undulating ridge crest move more quickly than the higher
hibited by these relict dune systems may be quite unrelated sections. This effect eventually produces a sinuous barchanoid
to present-day wind regimes and their interpretation may ridge composed ofprotruding linguoid and recessed barchanoid
have to rest on uncertain analogies with morphologically elements. (After A. Warren (1979) in: C. Embieton and 1.
similar, but active, forms elsewhere. Relict ergs, such as Thomes (eds) Process in Geomorphology. Edward Arnold,
that covering much of the Kalahari Desert of southern Africa, London, Fig. 10.11, p. 339.)
develop in response to major climatic changes and thus
provide important evidence of phases of greater aridity and
higher wind speeds in the past (see Section 14.5.2).

10.3.5.1 Transverse-type dunes

One possible explanation for the development of transverse
ridges involves wave-like motions in the wind. These could
be initiated by some pre-existing minor surface irregularity
or by temperature gradients in the atmosphere. Sand moves
most rapidly where such waves descend to the ground, and
consequently it will tend to accumulate in zones of slower
movement where airflows are ascending. An undulating
surface transverse to the wind direction will begin to form
which, once initiated, will enhance the wave-like airflow.
Fig.l0.24 A barchanoid ridge (A) propagating vortices which
Movement of sand up the windward slope of the develop-
shape a second ridge downwind. (FromA. Warren, 1979, Process
ing dune will eventually produce a slip-face and an equili- in Geomorphology. Edward Arnold, London, Fig. 10.12, p. 340.)
brium transverse ridge morphology.
Simple transverse ridges are comparatively rare apparently
because 'corkscrew'-like motions, or vortices, usually develop for producing the relationship observed between dune size
in the airflow probably as a result of surface irregularities. and spacing. Field data indicate that wind velocities return
These vortices distort the transverse ridge into the sinuous to their upwind value within a horizontal distance equivalent
form called a barchanoid ridge (Fig. 10.23). Ohce formed a to 12 to 15 times dune height.
barchanoid ridge will itself influence the airflow so that Where the supply is sparse, sand is entirely removed from
further vortices are propagated which shape the next ridge the zone represented by the linguoid elements of barcha-
downwind (Fig. 10.24). Linguoid elements of one ridge noid ridges and is swept downwind to form elongated arms
usually coincide with barchanoid elements of the next ridge thereby creating barchans. Indeed, transitions from trans-
downwind due to a complex interaction of transverse and vers~ ridges to barchans may be observed where there is a
longitudinal components in the airstream. This gives rise to marked diminution in sand supply downwind. Once formed
a series of nearly linked ridges trending obliquely at 10-20° a barchan generate its own secondary airflows which
to the prevailing wind direction which, with the transverse maintain its equilibrium form while it migrates downwind
elements, form what is termed a dune network or akle (Fig. 10.25). The elongation of barchan arms is due to the
pattern. The distance over which airflows are disturbed in higher rate of sand movement on the lower, outer parts of
the lee of dunes may provide another possible mechanism the dune. Eventually the arms extend sufficiently downwind
252 Exogenic processes and landforms

similarly reflect the effects of airflow patterns set up by

barchans upwind.

10.3.5.2 Linear dunes

Although all linear dunes consist of elongated ridges, two
subtypes can sometimes be distinguished. One type, some-
times referred to as a seif dune, is sinuous, relatively short
and has a sharp crest and a pointed downwind terminus.
The other subtype is long and straight and consists of a
narrow crest surmounting a broad plinth (Fig. 10.26). There
are widely divergent interpretations of the formation of
linear dunes. A minority view is that they are created by the
removal of sand from interdune troughs and that they are'
therefore erosional features. The great majority of researchers,
Fig.10.25 The pattern of airflow around an isolated barchan however, believe them to be, like other free dune types,
reconstructed from field data collected near In Salah in Algeria predominantly depositional forms. The main point of dis-
by P. Knott. (After A. Warren (1979) Process in pute is whether they develop in response to unidirectional
Geomorphology. Edward Arnold, London, Fig. 10.13, p.341.) winds or whether they are formed under bidirectional wind
regimes.
to receive some protection in the lee of the main body of Vortices in a unidirectional airflow might originate from
the barchan. Like transverse and barchanoid ridges, barchans convection cells in interdune troughs as a result of localized
often show a regularity in width and spacing and this may surface heating (Fig. 10.27). According to this model Y-

Fig.10.26 High altitude obLique aerial view of linear dunes in the Simpson Desert, centraL AustraLia. Y-junctions can be seen
between individuaL dunes, many of which extendfor tens of kilometres. Sand volumes here are low, as is evident from the broad,
sandless interdune troughs.
 Aeolian processes and landforms 253
A
CD B
".--
---@
m I
,_/
~)
A

_ _......A

C0Q::)C A
:JC
C ~CB

C.c)CC
Fig.10.27 Hypothetical reconstruction of roll-vortices that B
might shape linear dunes. Note the change in vortex diameter
downwind (cross-sections A, B and C); this provides a possible
mechanism ofY-junctionformation. As discussed in the text, this
model of linear dune development is not now widely favoured.
(From A. Warren (1979) Process in Geomorphology. Edward
Arnold, London, Fig.10.14,p.342.)

junctions would form where the axes of vortices rise from

the surface, possibly as a result of heating. The major short- A

coming of this explanation of linear dune formation is that Fig. 10.29 Two models for the development of self-type linear
the dimensions of vortices created by atmospheric instability dunes from barchans.In the model proposed by Bagnold (A) a
are much larger than average dune spacings. symmetric barchan is initially formed by unidirectional,
moderately strong prevailing winds (m) (1). High velocity storm
It has been noted that in many regions linear dunes are
winds (s) blowing obliquely across the barchan lead to a build-up
aligned obliquely to winds blowing from two prevailing of sand on one horn (2). This horn is elongated by storm winds
directions. The two principal winds may be either seasonal until it enters theflow of sand moving from the other horn (3-4) .
or diurnal, but in either case slip-faces develop in two . Subsequently the horn is elongated by winds from both directions
directions (Fig. 10.28). Under such conditions we would to form a sinuous linear, or seif, dune (5). By contrast, in the
model proposed by Tsoar (B) the orientation of a barchan is
expect a dune to extend in the resultant direction of sand
considered to be related to the prevailing direction of storm winds
movement. Seif dunes, for instance, can apparently develop (s) (1). The elongation of one of the barchan horns into a self
from barchans through the extension of one of the two arms dune is thought to result from sand transport by more common
by a second principal wind direction (Fig. 10.29). gentle winds blowing from a different direction (2-4). (After H.
Detailed field monitoring of a linear dune in the Namib Tsoar, (1984) Zeitschrift fUr Geomorphologie 28, Figs. 1 and 2,
pp. 100 and 101.)
Desert has demonstrated that its form changed in response
to a seasonal wind regime. In summer westerly winds were
found to transport sand from the western flank of the dune
and to deposit it on its easterly lee side. In winter the
situation was reversed. The crest of the dune was seen to
move 15 m back and forth over each annual cycle but the
base of the dune appeared to be fixed. Nevertheless, sand
was being transported across the dune which was extending
along a resultant of the easterly and westerly winds.
Although many researchers now believe that linear dunes
are a product of bidirectional wind regimes, problems remain.
For instance, it is not clear how compound linear dunes can
.be accounted for by this model. Moreover, it does not explain
the regular spacing of linear dunes.
The limiting case for dune extension is when the two
prevailing winds are each orientated normal to the dune
Fig.10.28 Internal structure of a linear dune formed by winds
blowing from two principal directions. Note that the slip faces dip crest. This type of wind regime produces a reversing dune
in opposing directions. (From A. Warren, (1979) Process in which grows by vertical accumulation rather than horizontal
Geomorphology. Edward Arnold, London, Fig. 10.15, p.342.) extension. Reversing dunes are therefore strictly speaking a
254 Exogenic processes and landforms

type of transverse dune. They have periodically opposing

slip-faces and typically develop under a strong seasonal
wind regime. Reversing dunes in Death Valley, California,
advance towards the north-east in spring and summer and
retreat to the south-east in the autumn.

10.3.5.3 Other types affree dune

Star dunes are most common in areas with complex wind
regimes and appear to form in response to strong winds
from several different directions. Some megadune-scale
examples, such as those in the Grand Erg Oriental of the
Sahara, reach an immense size and appear to be preferen-
tially located at the nodes of crossing trends formed by
linear dunes. It has been suggested that some isolated star
dunes form at the nodes of thermal convection cells, but it
seems unlikely that such airflows could be strong enough to
pile up large quantities of and. Star dunes along the eastern
and southern margin of the Rub'al Khali erg in Saudi
Arabia appear to fan out from the mouths of intermittent
streams and this suggests an association with localized sand
upplies.
Dome dunes probably form in a variety of ·ways. Some
reveal internal slip-faces dipping in one direction and this Fig.l0.30 Types of impeded dune. Prevailing wind direction
suggests that they may have evolved from barchans through indicated by arrows.
truncation of the crest and deflation of the arms by strong
winds. Other types of dome dune appear to have evolved from Table 10.3 and illustrated in Figure 10.30.
star dunes or to be controlled by vegetation and moisture In sand accumulations which have been stabilized by
distribution. vegetation, localized disturbance ot" the vegetation cover,
by for instance overgrazing, can give rise to blowouts.
These are circular or elliptical depressions, usually only
10.3.6 Development of impeded dunes
a few metres across, formed by the deflation of sand
A variety of dune forms are related to vegetation, topo- (Fig. 10.30(A». Higher wind speeds are attained over the
graphic barriers or localized sources of sediment. The major unvegetated surface and this dries the sand and makes it
varieties of these impeded dunes are listed and described in more susceptible to entrainment. The deflated sand is

Table 10.3 Classification of dune types controlled largely by vegetation, topographic feature or localized sediment sources

TYPE FORM AND POSITION MODE OF DEVELOPMENT

Blowout (A) Circular rim around depre sion Localized deflation

Parabolic dune (B) 'U' or 'Y' shape in plan view with arms opening Deposition of sand locally deflated upwind; arm are
upwind to enclose a blowout usually fixed by vegetation
Lunette (C) Crescent-shaped opening upwind Accumulation downwind of localized sediment
source such a desiccated lake basin or pan
Shrub-coppice dune (nebkha) Roughly elliptical to irregular in plan, streamlined Accumulation around and downwind of vegetation
(D) downwind clump
Lee dune (E) Elongated downwind from topographic obstruction Accumulation on protected lee side of obstacle
Fore dune (E) Roughly arcuate with arms extending downwind Accumulation in zone of disrupted airflow immediately
either side of obstruction windward of obstacle
Climbing dune (F) Irregular accumulation ri ing up windward side of Accumulation in zone of disrupted airflow on
large topographic obstruction windward side of obstacle
Falling dune (F) Irregular accumulation descending leeward ide of Accumulation in zone of disrupted airflow on upwind
large topographic obstruction side of obstacle
Echo dune (F) Elongated ridge roughly parallel to, and separated Accumulation in zone of rotating airflow upwind from
from, windward ide of topographic obstruction large ob tacle

Note: Letters A-F refer to illustrations in Figure 10.30

 Aeolian processes and landforms 255

carried downwind generally as far as the edge of the originally dried and carried as sand-sized aggregates downwind until
disturbed area. vegetation is encountered. The crescent shape may be
If deflation is intense the blowout becomes larger and its enhanced by wave action at times of lake or tidal flooding.
leeward rim migrates downwind leaving trailing arms on Lunettes are common in the semi-arid region of south-east
either side which are usually stabilized by vegetation. This Australia and around pans in the north-central Kalahari
forms a U or V-shaped parabolic dune (Fig. 1O.30(B)). Desert in Botswana. They appear to be confined to areas
High wind speeds may destroy the transverse component of with precipitation sufficiently high to introduce clays and
the parabolic form and leave only two elongated arms salts in runoff, but sufficiently low to permit desiccation
orientated downwind. Superficially, parabolic dunes may and deflation. An optimum mean annual precipitation of about
resemble barchans, but they can be readily distinguished 380 mm has been suggested for Australian lunettes, but it is
since their arms extend upwind whereas in barchans they likely that many are associated with fluctuating semi-arid
point downwind. Parabolic dunes are common where sand climates.
has been stabilized by vegetation, such as along coasts and Since vegetation reduces wind speeds near the ground
in semi-arid areas. Patchy reduction in vegetation cover can we would expect its presence to promote sand accretion.
eventually give rise to extensive fields of parabolic dunes. This is particularly likely to occur where vegetation lies
They cover nearly 30 per cent of the Thar Desert in north- downwind of a large sand supply. This situation is found
west India, a region which has experienced considerable not only in inland deserts, where localized vegetation sus-
climatic fluctuations and consequently changes in vegeta- tained by near-surface ground water may be present, but
tion density. also along coasts in most climatic zones where sand moving
Crescent-shaped lunettes are found on the downwind inland encounters vegetation. Isolated shrubs or vegetation
side of some saline ephemeral lakes and pans and along clumps in deserts generate localized sand accumulations
tidal lagoons in coastal areas (Fig. 1O.30(C)). In general, which taper downwind and are known as shrub-coppice
clay-sized particles constitute a much more important con- dunes or nebkhas (Figs 1O.30(D), 10.31). Especially along
stituent than sand. Periodically exposed salts and clays are temperate coasts, grasses may colonize and stabilize dunes.

Fig.l0.31 Nebkha, or shrub-coppice, dunes near Swakopmund, Namibia (hammer indicates scale).
256 Exogenic processes and !andforms

Such dunes are frequently sensitive to vegetation distur- sand may be funnelled up the slope to form a climbing
bance or short period of high intensity winds associated dune and if sand is carried over the crest of the barrier a
with storms, both of which can generate numerous blow- falling dune may develop on the lee side (Fig. 10.30(F)).
outs and initiate dune deflation.
Topographic features can significantly affect near-surface
10.3.7 Fine-grained deposits
airflows through the reduction of wind speeds and the
creation of eddies. If they lie astride sand-carrying winds, Silts and especially clays can be transported considerable
dunes will form where secondary airflows converge. This distances by the wind. Loess, consisting of well-sorted,
usually occurs close to the upwind side of the obstruction, very fine-grained deposits, cover large areas particularly in
where a fore dune develops, and on the downwind side, the mid-latitudes of the northern hemisphere, but they are
where a streamlined lee dune is formed (Fig. 1O.30(E)). A also found on the fringes of hot deserts. Loess blankets
single lee dune can develop behind small topographic fea- substantial areas of central Europe, the southern USSR and
tures, but a large obstruction characteristically generates China, and the USA (Fig. 10.32). Typically 80-90 per cent
two parallel lee dunes tapering downwind from its edges. of the particles are between 0.005 and 0.5 mm across. On a
Where the wind encounters a high, steep slope, such as global basis it has been estimated that some 10 per cent of
along an escarpment, a large roller vortex can develop which the Earth's total land area is covered by loess from 1 to
separates the fore dune from the topographic barrier. These 100 m thick. These deposits appear to have originated largely
may extend for several kilometres and because they reflect in the exten ive depositional plains formed on the margins
the form of the adjacent escarpment are often called echo of the Pleistocene ice sheets. Sediments were deflated,
dunes (Fig. 10.30(F)). They may develop into huge forms transported and redeposited during glacial episodes which
and the highest recorded dune in the world at over 400 m is were characterized by high wind speeds. This accounts for
an echo dune in Algeria. Where gradients along a topo- the concentration of loess in areas close to the equatorward
graphic feature are gently inclined, such as along gullies, limits of the Pleistocene ice sheets, since it is too coarse to

6 2
10km
1R1J
1000 0 1000
km

Fig.l0.32 Global distribution of major loess occurrences. (After K. Pye, (1984) Progress in Physical Geography 8, Fig. 1, p. 178).
 Aeolian processes and landforms 257

travel very large distances and in fact can be seen to thin eering treatment of aeolian processes by Bagnold (1941)
rapidly away from assumed source regions. These loess which is still worth consulting. Annual reviews of arid
deposits are considered further in Section 12.2.6. geomorphology in Progress in Physical Geography fre-
In most areas loess simply blankets the existing land- quently contain valuable updates of research on aeolian
scape and its main topographic effect is to smooth out minor processes and landforms.
urface irregularities since it accumulates more deeply in The literature on the mechani ms of particle movement
depressions. Rare occurrences of loess ridges, imilar in and deposition has grown rapidly over the past decade with
form to sand dune , are known, but it is not clear whether the increasing use of wind-tunnel experiments. An over-
these are primarily of depositional or erosional origin. Un- view of the mechanisms of sand entrainment and transport
usually thick deposits known as loess-lips, are found down- by the wind is provided by Sarre (1987), while Pye (1987)
wind of major river valleys, such as the Mississippi and considers the mobilization of finer particles. Specific fac-
Missouri, which provided major sources of silt during tors controlling sediment entrainment by the wind are
glacial phases. A high degree of sorting, angularity of con- examined by Logie (1982) and Willetts (1983). For more
stituent particles, and small particle size associated with detail on the effect of turbulence on grain entrainment see
relatively strong interparticle bonds and moisture-retention Lyles and Krauss (1971), and for quantitative estimates of
capability makes loess a relatively cohe ive deposit which critical threshold velocities for particle movement over a
in some cases forms steep cliffs. A high clay content or the wide variety of surface types see Gillette et al. (1980). A
precipitation of calcium carbonate in voids in the more general model of aeolian sediment transport by both salta-
calcareous forms of loess enhance this property. tion and suspension is presented by Anderson and Hallet
Surprisingly, dune forms in clay are more common that (1986), while saltation i considered specifically by Gerety
those developed in ilt. Clay dunes are mostly found near and Slingerland (1983), Greeley et al. (1983) and Iversen
coastal lagoons and downwind of exposed clay-rich saline (1983).
lake deposits and often take the form of lunettes. Clay dune Deflation is discussed in the context of the distribution
slopes are rarely above 15° and the, steeper gradient occurs and frequency of dust storms by Goudie (1978, 1983a) and
on the windward side. The origin of these features composed Middleton et al. (1986). Anderson (1986) considers the
of such fine material is more explicable when it is realized variation in rate of grain abrasion with height above the
that the original constituents were sand-sized aggregates of ground, and the formation of ventifacts is discussed by
clay particles which were entrained by the wind in the same Whitney and Dietrich (1973). The development of yardangs
way as ordinary sand grains. Such aggregates can form is discussed by McCauley et al. (1981) and Whitney (1985),
through the desiccation and cracking of clay-rich deposits, while Ward and Greeley (1984) provide a detailed field
perhaps aided by the binding effect of crystallized salts. study from Rogers Lake, California. Whitney (1978) exam-
Once deposited, rain leaches out the salt and the clay aggre- ines the role of vortices in generating wind-eroded linea-
gates decompose into fine particles to form a solid clay tions. The origin of pan is discussed by Lancaster (1978)
mass. with respect to the Kalahari and by Goudie and Thomas
(1985) in the context of southern Africa as a whole.
Turning to depositional forms, the development of ergs
Further reading is discussed by Fryberger and Ahlbrandt (1979) and Wilson
Aeolian processes and landforms are discussed in various (1971, 1973), while Wasson and Hyde (1983a) emphasize
books on desert geomorphology, while additional informa- the importance of wind regime and sand supply in con-
tion on littoral aeolian forms is to be found in texts on coastal trolling dune morphology. The existence of a well-defined
geomorphology (see Chapter 13). Cooke and Warren dimensional hierarchy of aeolian depositional forms is
(1973) and Mabbutt (1977) provide useful introductions to proposed by Wilson (1972, 1973) and challenged with
most of the topics covered in this chapter, but there have respect to particular localities and dune types by Lancaster
been major advances in the understanding of aeolian land- (1989a), Thomas (1988) and Wasson and Hyde (1983b).
forms since these books were written. This more recent The formation of ripples is considered by Brugmans (1983),
research is discussed in the volumes edited by Nickling Ellwood et al. (1975) and Sharp (1963). Useful discussions
(1986) and Thomas (1989). McKee (1979) is an important on the development of free dunes include those by Cooke
global survey of depositional forms drawing heavily on and Warren (1973), Fryberger (1979), Warren (1979) and
remote sensing data and EI-Baz (1984) similarly empha- Wilson (1973). Breed et al. (1979) provide a global survey
izes the value of satellite imagery in the interpretation of of major sand seas based largely on the interpretation of
aeolian features in desert landscapes. Other general sources satellite imagery, while studies of specific ergs include those
include the book on aeolian processes by Greeley and by Lancaster (1989a) and McKee (1982)- on the Namib,
Iverson (1985), the collection of paper edited by Brookfield Anton (1983) on the eastern Arabian Desert, Wasson et al.
and Ahlbrandt (1983) on aeolian sediments and the pion- (1983) on the Thar Desert, Wasson (1983) on the Strzelecki
258 Exogenic processes and landforms

and Simpson Deserts in Australia and Mainguet and Chemin Breed, C. S. and Grow, T (1979) Morphology and distribution of
(1983) on the Sahara. dunes in sand seas observed by remote sensing. United States
Important analy es of the origin of particular types of Geological Survey Professional Paper 1052, 253-302.
Brookfield, M. E. and Ahlbrandt, T S. (eds) (1983) Eolian Sedi-
dune include those by Hastenrath (1967), Howard et ai. ments and Processes. Elsevier, Amsterdam.
(1978) and Warren and Knott (1983) on barchans; Folk Brugmans, F. (1983) Wind ripples in an active drift sand area in
(1971), Lancas~er (1982) and Tsoar (1983) on linear dunes; the Netherlands: a preliminary report. Earth SUljace Processes
Lancaster (1980), Tsoar (1984) and Warren (1976) on the and Landforms 8, 527-34.
Campbell, E. M. (1968) Lunettes in southern South Australia.
relationship between barchans and seif dunes; and finally
Transactions of the Royal Society ofSouth Australia 92, 85-109.
Lancaster (1989b) on star dunes. In addition, the report of Cooke, R. U. and Warren, A. (1973) Geomorphology of Deserts.
the detailed monitoring of a linear dune in the Namib by Batsford, London.
Livingstone (1989) illustrates the value of field measure- EI-Baz, F. (ed.) (1984) Deserts and Arid Lands. Martinus Nijhoff,
ments in complementing laboratory and theoretical inves- The Hague.
tigations of aeolian processes. Ideas on the formation of Ellwood, J., Evans, P. and Wilson, I. G. (1975) Small-scale aeolian
bedform. Journal of Sedimentary Petrology 45,554-61.
obstruction and vegetation-related dunes are reviewed by Folk, R. L. (1971) Longitudinal dunes of the northwestern edge of
Cooke and Warren (1973), while Breed and Grow (1979) the Simpson Desert, Northern Territory, Australia. Part 1.
and Hack (1941) provide a more detailed examination of Geomorphology and grain size relationships. Sedimentology 16,
parabolic dunes, and Campbell (1968) discusses Australian 5-54.
lunettes. Relict dunes are briefly discussed by Goudie Fryberger, S. G. (1979) Dune forms and wind regime. United
States Geological Survey Professional Paper 1052, 137-69.
(1983b), and Bowler (1976), Grove and Warren (1968) and Fryberger, S. G. and Ahlbrandt, T S. (1979) Mechanisms for the
Talbot and Williams (1978) provide interesting regional formation of eolian sand seas. Zeitschrift fur Geomorphologie
studies. At the global scale Samthein (1978) presents evi- 23,440-60.
dence of the considerable extension of ergs during the Late Gerety, K. M. and Slingerland, R. (1983) Nature of the saltating
population in wind tunnel experiments with heterogeneous size-
Pleistocene.
density sands. In: M. E. Brookfield and T S. Ahlbrandt (eds)
A brief review of the voluminous literature on loess is Eolian Sediments and Processes. Elsevier, Amsterdam, 115-32.
provided by Pye (1984), but a much more detailed treat- Gillette, D. A., Adams, 1., Edno, A. and Smith, D. (1980) Thres-
ment is available in Pye (1987). The collection of papers hold velocities for input of soil particles into the air by desert
edited by Pewe (1981) contains much useful material on winds. Journal ofGeophysical Research 85, 5621-30.
Goudie, A. S. (1978) Dust storms and their geomorphological
the erosion, transport and deposition of desert dust. Smalley
implications. Journal ofArid Environments 1,291-310.
and Krinsley (1978) discuss the origin of desert loess and Goudie, A. S. (1983a) Dust storms in space and time. Progress in
Yaalon and Dan (1974) consider its distribution and ac- Physical Geography 7, 502--30.
cumulation on the desert fringe in Israel. Clay dunes remain Goudie, A. (1983b) The arid Earth. In: R. Gardner and H. Scoging
a little studied phenomenon, but Bowler (1973) provides a (eds) Mega-Geomorphology. Clarendon Press, Oxford and New
useful review of their occurrence and development. York,152-71.
Goudie, A. S. and Thomas, D. S. G. (1985) Pans in southern
Africa with particular reference to South Africa and Zimbabwe.
Zeitschriftfur Geomorphologie 29,1-19.
Greeley, R. and Iversen, J. D. (1985) Wind as a Geological.Pro-
cess. Cambridge University Press, Cambridge and New York.
References Greeley, R., Williams, S. H. and Marshall, 1. R. (1983) Velocities
Anderson, R. S. (1986) Erosion profiles due to particle entrained of windblown particles in saltation: preliminary laboratory and
by wind: application of an eolian sediment-transport model. field measurements. In: M. E. Brookfield and T. S. Ahlbrandt
Geological Society ofAmerica Bulletin 97, 1270-8. (eds) Eolian Sediments and Processes. Elsevier, Amsterdam,
Anderson, R. S. and Hallet, B. (1986) Sediment transport by wind: 133-48.
Toward a general model. Geological Society of America Grove, A. T. and Warren, A. (1968) Quaternary landforms and
Bulletin 97, 523-35. climate on the south side of the Sahara. Geographical Journal
Anton, D. (1983) Modern eolian deposits of the Eastern Province 134, 194-208.
of Saudi Arabia. In: M.E. Brookfield and TS. Ahlbrandt (eds) Hack, 1. T. (1941) Dunes of the western Navajo Country. Geo-
Eolian Sediments and Processes. Elsevier, Amsterdam, 365-78. graphical Review 31, 240-63.
Bagnold, R. A. (1941) The Physics of Blown Sand and Desert Hastenrath, S. L. (1967) The barchans of the Arequipa region,
Dunes. Chapman and Hall, London. southern Peru. Zeitschrift fur Geomorphologie 11, 300-31.
Bowler, J. M. (1973) Clay dunes: their occurrence, formation and Howard, A. D., 'Morton, 1. B., Gad-EI-Hak, M. and Pierce, D. B.
environmental significance. Earth Science Reviews 9,315-38. (1978) Sand transport model of barchan dune equilibrium.
Bowler, 1. M. (1976) Aridity in Australia: age, origins and ex- Sedimentology 25, 307-38.
pression in aeolian land forms and sediments. Earth Science Iversen, J. D. (1983) Saltation threshold and deposition rate model-
Reviews 12, 279-310. ling. In: M. E. Brookfield and T S. Ahlbrandt (ed ) Eolian
Breed, C. S., Fryberger, S. G., Andrews, S., McCauley, c., Sediments and Processes. Elsevier, Amsterdam, 103-14.
Lennartz, F., Gebel, D. and Horstman, K. (1979) Regional Lancaster, I. N. (1978) The pans of the southern Kalahari, Botswana.
tudies of sand seas using Landsat (ERTS) imagery. United GeographicaLJournaI144,81-98.
States Geological Survey Professional Paper 1052, 305-97. Lancaster, N. (1980) The formation of seif dunes from barchans -
 Aeolian processes and landforms 259

supporting evidence for Bagnold's model from the Namib Processes and Landforms 13, 545-53.
desert. Zeitschriftfur Geomorphologie 24,160-7. Thomas, D. S. G. (00.) (1989) Arid Zone Geomorphology. Belhaven,
Lancaster, N. (1982) Linear dunes. Progress in Physical Geography London.
6,475-504. Tsoar, H. (1983) Dynamic processes acting on a longitudinal
Lancaster, N. (1989a) The Namib Sand Sea: Dune Forms, Pro- (seif) dune. Sedimentology 30,567-78.
cesses and Sediments. Balkema, Rotterdam. Tsoar, H. (1984) The. formation of seif dunes from barchans - a
Lancaster, N. (1989b) Star dunes. Progress in Physical Geography discussion. Zeitschrift fur Geomorphologie 28, 99-103.
13,67-91. Ward, A. W. and Greeley, R. (1984) Evolution of the yardangs at
Livingstone, I. (1989) Monitoring surface change on a Namib Rogers Lake, California. Geological Society of America Bulletin
linear dune. Earth Surface Processes and Landforms 14, 95,829-37.
317-32. Warren, A. (1976) Morphology and sediments of the Nebraska
Logie, M. (1982) Influence of roughness elements and soil mois- Sand Hills in relation to Pleistocene winds and the development
ture on the resistance of sand to wind erosion. Catena Supple- of aeolian bedforms. Journal of Geology 84, 685-700.
ment 1, 161-73. Warren, A. (1979) Aeolian processes. In: C. Embleton and J.
Lyles, L. and Krauss, R. K. (1971) Threshold velocities and initial Thomes (eds) Process in Geomorphology. Edward Arnold,
particle motion as influenced by air turbulence. Transactions of London, 325-51.
the American Society ofAgricultural Engineers 14, 563-6. Warren, A. and Knott, P. (1983) Desert dunes: a short review of
Mabbutt, J. A. (1977) Desert Landforms Au tralian National needs in desert dune research and a recent study of micro-
University Press, Canberra. meteorological dune-initiation mechanisms. In: M. E. Brookfield
Mainguet, M. and Chemin, M. C. (1983) Sand seas of the Sahara and T. S. Ahlbrandt (eds) Eolian Sediments and Processes.
and Sahel: an explanation of their thickness and sand dune type Elsevier, Amsterdam, 343-52.
by the sand budget principle. In: M.E. Brookfield and T. S. Wasson, R. J. (1983) The Cainozoic history of the Strezelecki and
Ahlbrandt (eds) Eolian Sediments and Processes. Elsevier, Simpson dune fields (Australia), and the origin of the desert
Amsterdam, 353-63. dunes. Zeitschrift fur Geomorphologie Supplementband 45,
McCauley, J. F., Grolier, M. J. and Breed, C. S. (1981) Yardangs. 85-115.
In: D. O. Doerhing (ed.) Geomorphology of Arid Environments. Wasson, R. J. and Hyde, R. (1983a) Factors determining desert
Allen and Unwin, London, 233-69. dune type. Nature 304, 337-9.
McKee, E. D. (ed.) (1979) A study of global sand seas. United Wa son, R. J. and Hyde, R. (1983b) A test of granulometric con-
States Geological Survey Professional Paper 1052. trol of desert dune geometry. Earth Surface Processes and
McKee, E. D. (1982) Sedimentary structures in dunes of the Landforms 8, 301-12.
Namib Desert, South West Africa. Geological Society of Wasson, R. 1., Rajaguru, S. N., Misra, V. N., Agarwal, D. P., Dhir,
America Special Paper 188. R. P., Singhoi, A. K. and Kameswara Rao, K. (1983) Geomor-
Middleton, N.1., Goudie, A. S. and Wells, G. L. (1986) The fre- phology, late Quarternary stratigraphy and palaeoclimatology
quency and source areas of dust storms. In: W. G. Nickling of the Thar dune field. Zeitschrift fur Geomorphologie
(ed.) Aeolian Geomorphology. Allen and Unwin, Boston, and Supplementband 45, 117-51.
London, 237-59. Whitney, M. (1978) The role of vorticity in developing lineation
Nickling, W. G. (ed.) (1986) Aeolian Geomorphology. Allen and by wind erosion. Geological Society of American Bulletin 89,
Unwin, Boston, and London. 1-18.
Pewe, T. (ed.) (1981) Desert dust: origin, characteristics, and effect Whitney, M. I. (1985) Yardangs. Journal ofGeological Education
on man. Geological Society ofAmerica Special Paper 186. 33,93-6.
Pye, K. (1984) Loess. Progress in Physical Geography 8, Whitney, M. I. and Dietrich, R. V. (1973) Ventifact sculpture by
176-217. wind-blown dust. Geological Society of American Bulletin 84,
Pye, K. (1987) Aeolian Dust and Dust Deposits. Academic Press, 2561-82.
London and Orlando. Willetts, B. (1983) Transportation by wind of granular materials
Sarnthein, M. (1978) Sand deserts during glacial maximum and of different grain shapes and densities. Sedimentology 30,
climatic optimum. Nature 272, 43-6. 669-80.
Sarre, R. D. (1987) Aeolian sand transport. Progress in Physical Wilson, I. G. (1971) Desert sandflow basins and a model for the
Geography 11, 157-82. . development of ergs. Geographical Journal 137, 180-97.
Sharp, R. P. (1963) Wind rippies. Journal of Geology 71, 617-36. Wilson, I. G. (1972) Aeolian bedforms - their development and
Smalley, I. J. and Krinsley, D. H. (1978) Loess deposits associated origins. Sedimentology 19,173-210.
with deserts. Catena 5, 53-66. Wilson, I. G. (1973) Ergs. Sedimentary Geology 10, 77-106.
Talbot, M. R. and Williams, M. A. J. (1978) Erosion of fixed dunes Yaalon, D. H. and Dan, 1. (1974) Accumulation and distribution
in the Sahel, Central Niger. Earth Surface Processes 3, 107-13. of loess-derived deposits in tfle semi-desert fringe areas of
Thomas, D. S. G. (1988) Analysis of linear dune sediment-form Israel. Zeitschrift fur Geomorphologie Supplementband 20,
relationships in the Kalahari dune desert. Earth Surface 91-105.
 11
Glacial processes and
landforms

11.1 Glacier characteristics and dynamics Table 11.1 Present global distribution of glaciers

In some ways it is appropriate to describe the Earth as an REGION APPROXIMATE AREA

ice planet. At present some 10 per cent of the surface of the (km 2)
continents is covered by ice, but as little as 18 000 a BP the
North polar area
figure was nearer 30 per cent. This greater extent of ice Greenland ice sheet 1 726400
sheets has characterized much of the past 2-3 Ma and con- Queen Elizabeth Islands 106988
sequently landforms created by glacial action cover large Other Greenland glaciers 76200
Spitsbergen and Nordaustlandet 58016
areas that are now experiencing non-glacial climates. Since Other 114012
the majority of glacial landforms are formed beneath ice
Subtotal 2081616
sheets and glaciers it is only when these ice bodies retreat
that many types of glacial landform are revealed. This Continental North America
Alaska 51476
situation poses particular problems in understanding the Other 25404
processes whereby glaciers produce landforms since the
Subtotal 76880
mechanisms at work 'can only rarely be observed directly.
In this chapter we examine specific landforms created by South American Cordillera 26500
Europe 9276
glaciers, but as many glaciated landscapes are located in
Asian continent 115021
regions which are now experiencing quite different tem- African continent 12
perate conditions we consider the features of glacial land- Australasia 1015
scapes as a whole in Chapter 14. South polar area
Antarctic ice sheet (excluding ice shelves) 12535000
Other glaciers on Antarctic continent 50000
11.1.1 Glacier distribution and classification Sub-Antarctic islands 3000
Ice covers some 14.9 million km 2 at the present day. Most Subtotal 12588000
of this is accounted for by the two vast ice sheets ofAntarctica
and Greenland, the remaining 4 per cent comprising ice Total 14898320
•
caps and glaciers located mainly in high latitudes, and Source: Data from R. F. Flint, (1971) Glacial and Quaternary Geology,
innumerable glaciers found at high elevations in all lati- Wiley, New York, Table 4-B, pp. 76-7, based on various sources.
tudes (Table 11.1). Glaciers can be classified morphologi-
cally on the basis of their relationship to the underlying form where snow persists from year to year. Whether this
bedrock topography (Table 11.2; Figs 11.1 and 11.2). It is occurs is determined by the rate of snow accumulation,
difficult to generalize about the distribution of the different which is largely a function of the amount of precipitation
glacier types since, apart from ice sheets, most occur over a falling as snow, and the rate of melting, which is largely a
broad range of latitudes. function of temperature (Fig. 11.3). In the mid-latitudes the
The occurrence of glaciers is determined not only by cli- seasonal distribution of precipitation is more crucial than
mate but also by topography since there must be a suitable the total amount. Extensive glaciers are not found, for
surface on which ice can accumulate. Glaciers can only instance, in the high mountains along mid-latitude west
262 Exogenic processes and landforms

Table 11.2 Morphological classification of glaciers

BASIC TYPE COMPONENT OR SUBTYPES CHARACTERISTICS

Ice sheet and ice cap (unconstrained Ice dome A dome-like ice mass with a convex cross-profile fonned in re pon e to the
by topography)* basic flow characteristics of ice
Outlet glacier Glaciers which radiate out from an ice dome often occupying significant
depressions (Fig. 11.1). Within the ice dome they can be distinguished by a
zohe of rapidly moving ice tenned an ice tream
Ice shelf A floating ice cap or part of an ice sheet only partially constrained by the
coastal configuration and which defonns under it own weight
Glaciers constrained by topography Ice field A roughly level area of ice distinguished from an ice cap because of the
absence of a dome-like form and the control on ice flow exerted by the
underlying topography
Cirque glacier A small ice mass usually occupying an annchair-shaped bedrock hollow and
characteristically wide in relation to its length
Valley glacier A glacier which occupie a rock valley and is overlooked by rock cliffs
(Fig. 11.2). They may originate in an ice field or a cirque glacier into which
they may imperceptibly merge at their upper end. Large valley glaciers may
be joined by tributary glaciers, fonning a dendritic pattern of ice flow
Other small glacier Glaciers which occur in a wide variety of topographic positions, but all of
which are clo ely controlled by the underlying topography

*Ice sheets and ice caps are es entially distinguished on the basis of size, the former exceeding around 50 000 km 2 in area.
Source: Based on classification and discussion in D. E. Sugden and B. S. John (1976) Glaciers and Landscape. Edward Arnold, London.

Fig.11.2 Harker GLacier, a valley gLacier discharging into a

fjord, South Georgia. (Photo courtesy D. E. Sugden).

Fig.11.1 OutLet gLaciers near BarthoLins Brae, BLosseviLLe Kyst, CONTINENTALITY

GreenLand. Note the mediaL and LateraL moraine. (Photo courtesy
of the Geodetic Institute, Copenhagen, Denmark.)

coasts in spite of heavy precipitation because most falls as I TEMPERATURE I I PRECIPITATION

rain in summer and the relatively high summer tempera-

~ TOPOGRAPHY
tures lead to substantial melting of winter snows. By con-
trast the largest ice sheets receive very small amounts of
/'-
EXTENT AND THICKNESS
precipitation although this nearly all falls as snow. The OF SNOW COVER
mean annual precipitation over the northern part of the
Greenland ice sheet is only 150 mm, but rates of summer
melting are so low that ice accumulation takes place.
Although there are other sources of heat supplied to
glaciers, solar radiation is by far the most significant in Fig. 11.3 Major factors controlling the occurrence ofgLaciers.
 Glacial processes and landforms 263

accounting for their global distribution. Because melting the pressure melting point. In small glaciers the percolation
can occur only at temperatures above O°C it is the mean of surface meltwater may be the most important heat source
summer temperature rather than the mean annual value that since the refreezing of 1 g of water releases sufficient heat
influences the amount of snow-melt that will occur. Lati- to raise the temperature of 160 g of ice by 1 0c. The rate of
tude, elevation, aspect and continentality all exert an indirect fim accumulation can also be important since a high rate
control over the distribution of glaciers through their in- can introduce a significant quantity of cold material into the
fluence on precipitation and temperature (Fig. 11.3). The glacier. This helps to counteract the geothermal heating at
effect of continentality on precipitation is a particularly . the glacier base and so reduces the rate of increase in
significant factor in the development of ice sheets. This is temperature with depth. Once the base of a glacier reaches
illustrated by the pattern of snow accumulation over the pressure melting point all the heat from geothermal and
Antarctic ice sheet which is closely related to the distance frictional sources can be used in melting rather than simply
from the nearest area of ice free ocean in summer. raising the temperature of the ice.
It has been accepted practice to classify glaciers as tem-
perate or polar depending on whether they are supposed to
11.1.2 Characteristics of glacier ice
be composed of warm or cold ice. However, it is now re-
Glaciers are composed not only of ice but also of smaller cognized that not only do both types of ice occur in glaciers
amounts of air, water and rock debris. A property of ice of in both temperate and polar environments but that both
considerable geomorphic importance is its ability to deform types may be present in an individual glacier or ice sheet.
and flow under its own weight. Although glacier ice can The Antarctic ice sheet, for example, is composed pre-
fonn directly from the freezing of liquid water or water vapour dominantly of cold ice but is now known to contain basal
on the glacier surface, snowfall is the most important source. layers of warm ice in places. In the following discussion we
Snow with a density of only 50-70 kg m-3 is gradually con- shall refer to those glaciers or parts of glaciers with a basal
verted to ice by a complex process of compaction and layer at pressure melting point. as warm-based and those
recrystallization through an intermediate stage known as with a basal layer below pressure melting point as cold-
firn. This is composed of a loosely consolidated mass of based. A further important distinction is that between
ice crystals with a bulk density of around 400 kg m-3. The active ice which is moving downslope and being reple-
transformation into ice with a density of about 800 kg m-3 nished by fresh accumulations of snow in its source region
involves an increase in crystal size and the closing of voids, and stagnant ice which is decaying in situ and has ceased
but the final conversion to true glacier ice (density of about to experience any significant lateral movement.
900 kg m-3) only o'ccurs when the pressure of overlying ice
leads to the elimination of most of the remaining air
bubbles.
11.1.3 The glacier mass balance
Temperature is a key characteristic of glacier ice as it
exerts a profound influence on glacier behaviour. In addi- The gains and losses of ice experienced by a glacier
tion to solar radiation heat can be supplied to a glacier from constitute its mass balance or glacial budget. The ice, fim
the surface by the incorporation of fim which is warmer and snow added to the glacier constitute its accumulation.
than the existing glacier ice and by the release of latent heat The losses, or ablation, arise largely from melting in most
as water is refrozen. A glacier is also warmed from below glaciers, but evaporation, sublimation (the direct conver-
by geothermal heat and by frictional heat generated by sion of ice to water vapour), wind erosion and calving (the
sliding and the deformation of ice at its base. Variations in breaking away of blocks of ice into standing water) may
the supply of heat from these sources give rise to two types also be important in certain situations.
of glacier ice which have fundamentally different geomor- A mass balance is normally calculated over the balance
phic properties. year. This is the interval between two successive times at
Cold ice is at a temperature below melting point, but which ablation has reached a maximum value. Although
warm ice is so close to melting point that it contains liquid this normally occurs at the end of the summer the balance
water. Strictly we should use the term pressure melting year may not"be exactly 365 days. The total masses added
point here as the freezing point of water decreases as to and lost from a glacier during the balarice year are called,
pressure increases. For example, the melting point at a depth respectively, the gross annual accumulation and gross
of 2164 m at the base of the Antarctic ice sheet at Byrd annual ablation. The difference between these two amounts
Station is -1.6 0c. Cold ice occurs where the glacier surface is the net annual accumulation or net annual ablation,
experiences very low winter temperatures or where low depending on whether there has been a net gain or loss in
summer temperatures lead to negligible surface melting. mass. The net specific balance is the net annual accumu-
Warm ice occurs where geothermal or frictional heat are lation or ablation at a particular point on a glacier, and the
sufficient to raise the temperature of at least the basal ice to integration of a large number of such point measurements
264 Exogenic processes and landforms

Net
accumulation
Glacier surface at
end of following winter

Equilibrium
line

Glacier surface at end

of succeeding summer

Fig. 11.4 Variations in accumulation and ablation down an idealized glacier. The distribution of net annual accumulation and ablation
over a glacier is ofgreat significance to glacier movement. Net ablation generally increases down-glacier below the equilibrium line while
net accumulation tends to increase up-glacier above this point. The rate of increase in the net balance with height up a glacier is important
for the rate of ice movement because the higher the rate of increase the faster is the rate of ice movement needed to maintain the same
glacier profile.

can provide an estimate of the total mass balance. directions and is related to the mass of overlying ice, and
The taking of such measurements is a dIfficult and time- shear stress, which is related not only to the thickness of
consuming task and consequently there are few accurate overlying ice but also to the surface slope of the glacier.
mass balances for real glaciers. The data that do exist High shear stresses are generated towards the base of thick
clearly show that the annual net balance generally varies glaciers with steep slopes whereas lower basal shear stresses
over a glacier in a systematic manner with a positive balance are produced by thin, gently sloping glaciers. The variations
(net accumulation) in the upper part of the glacier and a in shear stress at the base of most glaciers are, however, not
negative balance (net ablation) in the lower part. These two large, usually lying between 0.05 and 0.15 MPa with a mean
zones meet at the equilibrium line where accumulation is of about 0.1 MPa.
exactly compensated by ablation and where the mass balance The main mode of internal deformation involves slip-
is therefore zero (Fig. 11.4). It is important to draw a page within and between ice crystals and is called creep.
distinction between the situation over an entire balance year Although the exact mechanism is not fully understood,
and seasonal variations since, depending on the season, various attempts have been made to model ice defomiation
much or all of a glacier may be experiencing accumulation by this process. In one such formulation, known as Glen's
or ablation. power flow law, the predicted rate of deformation depends
not only on the shear stress but also on the temperature of
the ice (Box 11.1). This model of creep appears to accord
with several observations about the behaviour of glaciers.
11.1.4 Glacier motion First, deformation rates are at a maximum at the base of the
glacier, where both stresses and, in the case of cold ice,
11.1.4.1 Mechanisms of ice movement temperatures are highest. Secondly, it explains the faster
The movement of glaciers can in many respects be treated rates of movement of warm ice. Thirdly, it accounts for the
in the same way as other forms of mass movement, al- way glaciers can regulate their discharge through negative
though ice masses have special properties which must be feedback since an increase in glacier thickness will increase
taken into consideration. Two, more or less distinct, types the basal shear stress which will in tum accelerate the rate
of ice movement can be distinguished - internal deforma- of ice flow and thereby reduce ice thickness. Where stresses
tion and basal sliding. The relative importance of these within the glacier cannot be accommodated sufficiently
two mechanisms varies significantly with basal sliding quickly the ice may move by fracturing. This is most likely
accounting for up to 90 per cent of the movement of warm- to occur on the margins of glaciers where thrusting can give
based glaciers, but being largely inoperative in cold-based rise to shear fractures and tensional stresses can produce
glaciers. Glacier ice deforms internally because it is subject crevasses (see Section 11.1.5.2).
to stress. At anyone point in a glacier this stress has two Basal sliding involves four major processes. One
components: hydrostatic pressure, which is exerted in all mechanism is the slippage of the glacier bed over a thin
 Glacial processes and landforms 265

Box 11.1 Glen's power flow law of subglacial sediment. By facilitating movement such bed
deformation would increase the rate of ice flow or allow
On the basis of laboratory experiments J. W. Glen estab- flow to occur at a similar rate over a lower bed gradient or
lished that the train rate in a block of ice soon attains a lead to a combination of these effects. In order for bed de-
steady value on being subject to a constant stress. The re- formation to occur the subglacial material must be saturated
lationship determined by Glen, known as the power flow with water and maintain a high pore-water pressure. Such
law, has been adapted by J. F. Nye to apply to glaciers. The
conditions will only be met if the glacier bed is at the
relationship is
pressure melting-point and if basal water accumulates in
e == At 'l
the sediment rather than being drained away.
where e is the strain rate, 't the effective shear stress, A a
constant related to temperature and n an exponent. 11.1 .4.2 Glacier flow
The values of n have been determined by several inves-
Average rates of movement vary enormously from one
tigators and vary from 1.7 to 4.5, the mean value being
around 3. This exponent makes the strain rate highly de- glacier to another but typically lie in the range 3-300 m a-I.
pendent on shear stress since with a value of n of 3 a doubl- The velocity of ice flow also varies spatially and tempo-
ing of the shear stress produces an eightfold increase in rally within an individual glacier. If we look at a vertical
strain rate. In Glen's experiments the value of A was found profile we find that the flow velocity is at a maximum at the
to be significantly affected by differences in temperature,
glacier surface and decreases downwards (Fig. II.S(A),
changing from 0.17 at aDC to only 0.00 17 at -13 DC, a
variation of two orders of magnitude. (B)). This at first sight appears paradoxical since, as we
have seen, most movement occur in the basal layer of a

A Increasing velocity

layer of water only a few millimetres thick. This can signi-

ficantly reduce the friction between the glacier and its bed
and considerably increase rates of glacier movement. A
second mechanism, known as regelation creep, involves
the movement of a warm-based glacier over minor irregula-
rities in its bed. The higher pressures on the up-glacier side
of an obstacle lead to melting and the water produced
migrates to the zone of lower pressure on the lee side of the
obstruction where it refreezes. The process is most effec- B Increasing velocity
tive where the obstacle is small enough (less than about 1 m
across) to enable the latent heat released by the refreezing
to be effectively conducted to the up-glacier side where it
promotes further melting. Where larger obstacles are present
on the glacier bed a third process known as enhanced
basal creep can become important. The increased stress on
the up-glacier side of such an obstruction allows the ice to
flow round it. This mechanism promotes the movement of
basal ice even where it is below pressure melting-point. C o
Traditionally, the analysis of basal sliding has been ~~~~~~~~~~~~
~
~

founded on the assumption that glaciers move over a rigid '" ~

\
rock bed. That this is frequently not the case is clear from
\
the sediment-covered surfaces found immediately in front /
of retreating glaciers. Furthermore, some 80 per cent of the /
/
area of Europe and North America covered by ice sheets /
during the glacial advances of the Pleistocene is mantled by ~
unconsolidated sediments rather than bedrock. Many in- Fig. 11.5 Schematic representation of vertical and horizontal
. stances of deformation structures attributable to ice sheet variations in the velocity of vaLLey gl':lcier movement: (A) cross-
movement are known from these sediments, and recent prOfile vertical variation in velocity in a glacier with no basal
experimental work has demonstrated that even quite coarse- sliding; (B) vertical cross-profile variations in velocity where
grained sediments can be easily deformed beneath a glacier. basal sliding is significant (isoLines intersect vaLLey floor); (C)
plan view of vaLLey glacier iLLustrating variations in velocity at
All this suggests that a fourth, and po sibly widespread, surface where there is no sliding along vaLLey waLLs. (D) plan
mechanism of glacier motion involves a significant part of view of velocity variation at surface where there· is significant
ice movement being accomplished through the deformation sliding along valley walls.
266 Exogenic processes and landforms

glacier where the shear stress is at a maximum. We have to Zone of

accumulation
remember, however, that each imaginary layer of ice is not
only moving as a result of the shear stress at that level but A
is also being carried along by the moving layer below it.
The velocities at each level in the glacier are thus cumu- Zone of ablation

lative from the base to the surface. There are also signifi-
cant variations in the velocity across confined glaciers since
the increased friction between the ice and rock wall retards
ice movement there (Fig. l1.5(C), (D)).
Changes in discharge longitudinally down a glacier are
strongly influenced by variations in total accumulation and B

~~
ablation. Discharge is generally highest around the equili-
brium line because the cumulative volume of ice increases
from the head of the glacier to reach a maximum at this
point. Below the equilibrium line discharge decreases pro-
COMPRESSIVE FLOW EXTENDING FLOW
gressively down-glacier with the increasing net loss of ice
through ablation (Fig. 11.4).
Because snow is being added from above in the accumu-
lation zone there must be some downward movement of ice
E E c d
,--------'----------'------1---------
c

1 I I I
from the surface. Similarly, in the ablation zone there must 1 I I
I I I
be some upward movement to maintain the surface form by I I I I
replacing ice lost through ablation. Evidence of these verti- I I I I
cal components of movement is provided by the burial of I I I I
surface debris in the accumulation zone and the emergence I I I
I I I
at the surface of previously buried debris in the ablation I I
zone. Such longitudinal variations in ice movement are
associated with two different types of flow regime within a
glacier (Fig. 11.6(A)). Above the equilibrium line extend-
ing flow predominates since the ice becomes more 'stretched Fig.11.6 Flow patterns within a glacier: (A) basic pattern of
out' down-glacier as the velocity of flow increases with the downward movement of ice in the accumulation zone and upward
steeper angle of movement. Conversely, there is a reduction movement in the ablation zone; (B) zones of extending and
in velocity in the zone of compressive flow below the compressive flow related to irregularities in valley floor
equilibrium line as the upward component of ice movement morphology (E = extending, C = compressive). The lines
indicating the direction of ice movement with respect to the
becomes more marked. In reality the longitudinal zonation
glacier sU/1ace are slip lines which indicate the direction in which
of extending and compressive flow is more complex than the ice has the maximum tendency to shear. (Part (B) modified
this since the flow regime is also affected by bedrock topo- from 1. F. Nye (1952) Journal of Glaciology 2, Fig. 7, p.88 and
graphy; extending flow is promoted over convex bedrock Fig. 8, p. 90.)
surfaces, especially in ice falls which occur where a glacier
flows down a steep slope, whereas compressive flow pre- Stream B, is moving over a bed composed of saturated,
vails over concave surfaces (Fig. l1.6(B)). The upward and deformable sediment.
downward component" of ice movement associated with Changes in ice movement over time occur in response to
zones of compressive and extending flow is accommodated alterations. in the mass balance arising from particular
along slip lines which meet the glacier surface at around meteorological conditions. Cooler weather promotes accu-
45° and represent trajectories of maximum shear stress. mulation and the glacier thickens and increases its velocity,
Movement along these slip lines can occur through creep as whereas in warmer conditions increasing ablation induces
well as by thrusting along 'faults' within the ice. thinning and a reduction in velocity, both adjustments pro-
In ice sheets and ice caps relatively narrow zones of ice gressing until a new equilibrium profile is attained. Changes
occur in which movement is much more rapid than the in mass balance can be transmitted down-glacier by kine-
adjacent parts of the ice mass. These ice streams are matic waves. These are complex phenomena, but we can
responsible for a significant proportion of ice movement in grasp a basic understanding of how they function by ima-
large ice masses and often feed outlet glaciers (Fig. 11.1). gining an increase in accumulation which causes a local
Several major ice streams have been identified in the raising of the glacier surface. The thic~er ice under the
Antarctic ice sheet and seismic measurements indicate that bulge will begin to move more rapidly th~m the surrounding
at least one of these in the west Antarctic, known as Ice ice because the basal shear stress will be higher. The zone of
 Glacial processes and landforms 267

thicker ice then moves down-glacier as a wave at between an increase in basal sliding. This could then promote a
two to five times the velocity of the ice itself. Adjustments cycle of positive feedback in which the initial increase in
to the profile of a glacier arising from changes in mass basal sliding generates more meltwater because of the greater
balance can consequently be transmitted to the glacier snout frictional heating associated with the higher rate of basal
much more rapidly by kinematic waves than by physical movement.
movement of the ice itself. Although kinematic waves have The concept of glacier movement assisted by the defor-
been observed moving down glaciers, in exceptional cases mation of subglacial sediments has also led to the idea that
involving a surface rise of some 100 m, they need not periodic changes in pore-water pressure could promote sur-
necessarily have any surface expression since the increase ging behaviour. Data collected from boreholes drilled into
in discharge represented by the wave can occur through Trapridge Glacier, Yukon Territory, Canada, indicate that it
adjustments in the ice flow other than those involving rests, at least in part, on a bed of potentially deformable
alterations in glacier thickness. sediment. It has been suggested that during non-surging
The rapidity with which glaciers respond to changes in interludes this substrate is efficiently drained through a
mass balance and attain a new equilibrium form varies system of channels and pipes; consequently, pore-water
enormously but it is generally related to the size of the ice pressures in the sediment are low and significant deform-
mass. Relaxation times vary from 3 to 30 a for valley ation does not occur. An increase in basal shear stress, per-
glaciers to thousands of years for ice sheets. Such long and haps related to an increase in glacier thickness, could
variable relaxation times make it difficult to relate changes promote an initial increase in subglacial sediment deform-
in glacier movement to particular climatic or meteorolo- ation leading to the disruption of the subglacial drainage
gical conditions. system. Such a reduction in permeability could then lead to
The most spectacular temporal variations in glacier a build-up of water within the sediment voids, and the
movement are glacier surges. These events, during which resulting increase in pore-water pressure could induce a
ice-flow velocities temporarily reach between 10 and 100 considerable increase in bed deformation and thereby
times their normal value, are experienced by some, but not glacier movement.
all, glaciers. Surging glaciers are fairly common, some 204
having been identified in North America alone, but they 11.1.4.3 Short-term glacier fluctuations
seem to have few characteristics in common. In some, surges Glaciers advance and retreat in response to changing
occur on a more or less regular cycle with a periodicity meteorological and climatic conditions over a range of
ranging from 15 to 100 a or more, but in others surging is time scales. Here were are concerned only with the shorter-
unpredictable. They also occur in a wide range of glacier term fluctuations of up to a few hundred years, since longer-
types, both warm and cold-based. Surging is initiated when term adjustments are caused by major climatic changes and
a threshold of instability is reached and ice in the upper are therefore more appropriately considered in Chapter 14.
ablation zone begins to move rapidly down-glacier. In some The relationships between changing meteorological con-
cases this may precipitate a rapid movement of the glacier ditions and short-term climatic fluctuations and glacier
snout, such as the 45 km advance at a rate of up to 5 m h- 1 behaviour are far more complex than was once assumed.
recorded during a surge by Bruarjokull, Iceland. In many Indeed, changes in weather and climate are so rapid in
North American glaciers, however, the effect of a surge is comparison with typical relaxation times of glacial systems
mainly confined to a thickening of ice towards the snout. that glaciers in true equilibrium with prevailing climatic
Surging is still one of the least understood aspects of conditions are rare. Only over the past 40 a or so have
glacier movement; it is necessary to explain both the detailed records of mass balance been kept and these for
mechanism that triggers a surge and the high flow veloci- only a few glaciers. Over a longer period less detailed
ties subsequently attained. The greater abundance of melt- information on glacier fluctuations can be gained from
water commonly associated with surges suggests that the maps, photographs, drawings, historical reports and, over
rapid movement may be related to greatly enhanced rates of longer periods, botanical evidence.
basal sliding through the effect of basal water reducing Variations in rates of glacier movement over weeks and
friction along the glacier bed. The trigger mechanism is months appear to be related to the abundance of meltwater
much more difficult to explain. It may be linked to external associated with seasonal changes in ablation rates. Year-to-
factors such as earthquakes or increased precipitation, but year fluctuations follow, with a lag, variations in winter
the regularity of surging in many glaciers and the rather snowfall and mean annual temperature, while the widespread
similar quantities of ice moved in each surge suggests that and significant phases of glacier advance and subsequent
it is more likely an intrinsic element of the behaviour of retreat documented over the past 300 a or so appear to be
some glaciers. Such glaciers may experience a gradual related to the cool climatic interval known as the 'Little Ice
build-up of basal water as a result of changes in the sub- Age'. Not all glaciers, however, respond in the same way to
glacial hydrological system which eventually precipitates a particular climatic fluctuation. Glaciers with a low
268 Exogenic processes and landforms

600
A
500

E
+-'
400
.!:
Ol
'Q)
I 300

200
0 2 4 6 8
Horizontal Distance (km)

600
B
Fig.11.8 Morphology of the Antarctic ice sheet illustrated by a
500 computer-generated three-dimensional reconstruction. (From
E D.1. Drewry (ed.) (1983) Antarctica: Glaciological and
+-' Geophysical Folio. (Scott Polar Research institute, Cambridge
.!: 400
Ol
'Q) Sheet 2, Fig. 2b.)
I
300
is largely controlled by the dynamics of ice flow. This can
be seen most clearly in the convex form of ice domes (Fig.
200 11.8). These have a maximum thickness which may exceed
a 2 4 6 8 10
4000 m in. Antarctica. Where there is sufficient accumula-
Horizontal Distance (km)
tion the ice builds up until the level of basal shear stress
600 promotes significant ice deformation. As we have already
c noted the amount of shear stress is related to glacier thick-
500 ness and surface slope so where the ice is thinner around
E the margins of the dome a steeper slope is required to main-
400 E.1.. tain the ice flow, and conversely where the ice is thicker

300
t E1
below the crest of the dome a less steep gradient is needed
to sustain movement. Theoretical models of ice deform-
ation generally accord quite well with the form of most ice
200 ;·;-;m~~~;·~~/;~·~\}f\\~:;.~~:?~~~(dI(~~f~JH\\;::}:~):\Q:/X~;f}~~;:W:\\/}/: domes (Box 11.2), but in reality a smoothly curved convex
a 20 40 60 80 100 120 shape is prevented by factors such as bedrock irregularities
Horizontal Distance (km) underlying the glacier and spatial variations in ice tempera-
Fig.11.7 Effects of a rise of iOO m in the altitude of the
equilibrium line (E, to E2) on a steeply inclined valley glacier (A), Box 11.2 Estimating the profile of an ice dome
a more gently inclined valley glacier (B) and on an ice cap (C).
Note that as the slope of the ice sUJiace decreases from A to C the
On the basis of a number of assumptions 1. F. Nye proposed
proportion of the glacier affected by the change in the altitude of
a simple formula which enables an estimate of the cross-
the equilibrium line increases.
profile of an ice dome to be made. The most important
assumptions are that the ice is actively flowing and the ice
dome is in equilibrium, but it is also assumed that the ice
surface gradient are more susceptible to such changes than flow is not affected by variations in temperature, by locali-
those with a steep slope since a vertical change in the equi- zed accumulation or by bed irregularities. The elevation (m)
librium line due to, for instance, a change in mean annual of the ice surface at any point (h) is given by
temperature, will affect a much larger proportion of the h = >J(2hos)
latter and thus produce a much more substantial change in where ho is 11 m, and s the horizontal distance from the
the areas of ablation and accumulation (Fig. 11.7). margin of the ice dome (m).
This formula provides a fairly good approximation of the
form of most ice domes, although it tends to overestimate
11.1.5 Glacier morphology the maximum elevation. Not only are such calculations use-
ful for examining existing ice domes, but they are also a
11.1.5.1 Large-scale forms useful way of estimating the form of the large ice sheets of
While the overall shape of confined glaciers is of course the Pleistocene.
closely related to topography, that of unconfined ice masses
 Glacial processes and landforms 269

ture and accumulation. Movement of ice in an ice dome A TRANSVERSE CREVASSES

occurs normal to its contours and it can operate through
basal sliding as well as internal deformation.

11.1.5.2 Small-scale surface features

Two sets of processes create forms on glacier surfaces -
those related to ice movement and those associated with
accumulation and ablation. Perhaps the most evident fea-
tures are crevasses formed through ice fracture. They are
vertical to subvertical cracks up to several metres across,
but do not generally extend to a depth of more than 30 m
since below this depth the ice is less rigid and is likely to B RADIAL CREVASSES
accommodate to tensional stress by creep rather than frac-
ture. Three types of crevasse can be identified and related
to contrasting stress patterns (Fig. 11.9).
A feature common in the ablation zone of many glaciers
and attributed to the metamorphism of ice during flow is a
banding of white ice containing abundant air bubbles be-
tween denser bluish ice. Another form of patterning is that
produced by ogives, or Forbes bands. These are alternat-
ing bands of light and dark ice lying across the glacier
surface and are developed in some glaciers below ice falls
(Fig. 11.10). Where movement is more rapid in the middle C CHEVRON CREVASSES
of the ice flow the bands are deflected down-glacier. The
total width of each pair of light and dark bands corresponds
to the distance moved by the glacier in a year. The dark
bands are generated during the passage of ice across the ice
fall in the summer when melting occurs. Conversely, light
bands are produced in winter when fresh snow is incor-
porated into the ice as it traverses the ice fall.
On the surface of the large ice sheets of Greenland and
Antarctica a variety of aeolian features have been recorded. Fig.11.9 Types of crevasse occurring in valley glaciers: (A)
transverse crevasses are associated with extending flow and
The best documented are dunes called sastrugi constructed result from expansion normal to the direction of ice flow; (B)
of hard-packed snow and aligned with the prevailing wind. radial (or splay) crevasses occur in zones of compressive flow,
They most commonly form in the lee of obstacles but are and although they also arise from expansion normal to the
also known in open situations. Wind is also funnelled around direction of ice flow they differ from transverse crevasses in being
rock outcrops protruding above an ice sheet, and snow is concave up-valley; (C) chevron (or en echelon) crevasses
associated with the tensional stresses generated by the drag of the
excavated from their lee side. On most glaciers the more valley walls on the margins of a glacier.
significant small-scale surface features associated with
ablation are produced by meltwater (see Section 11.4.1).
streams. The operation of abrasion has been further con-
11.2 Glacial erosion firmed by limited direct observations of glacier beds and
synthesized in laboratory studies. Its effectiveness is deter-
mined by glacier, bedrock and basal debris characteristics
11.2.1 Mechanisms of erosion and the role of these factors is summarized in Table 11.3.
Glacial erosion is accomplished by three major processes - Evidence for crushing' and fracturing comes from series
abrasion, crushing and fracturing, and joint·block removal. of crescentic cracks and grooves called chattermarks found
Abrasion involves the scratching, grooving and polishing of on some bedrock surfaces and thought to represent the
bedrock by debris carried in the base of a glacier (Box effects of pressure exerted by basal glacier debris, and from
11.3). Evidence of its efficacy is provided by smoothed the presence in glacial deposits of blocks which have
bedrock surfaces, often exhibiting parallel sets of fine grooves clearly been sheared off from fresh bedrock. The typical
known as striations, and the production of fine particles basal shear stress of around 0.1 MPa is quite inadequate on
(usually < 0.1 mm in diameter) known as rock flour and its own to shear off bedrock protrusions, but where large
occurring in high concentrations in most glacier meltwater boulders are being carried along the glacier bed a much more
270 Exogenic processes and landforms

Fig.ll.lO Ogives visible on the surface of AusterdaLsbreen, a valley glacier in southern Norway. Note the precipitous valley walls.
Box 11.3 Frictional shear at the bed of a glacier

Ice thickness determines the normal stress at the interface

between debris and substrate at the base of a glacier, but for 100ma"
warm-based glaciers this stress is counteracted to orne ;z
)( 1000
extent by basal water pressure. The basal frictional shear 'E
(F) is given by
F = Acl1.(vh - p)
.
g
II: 100
c:
0

where A c is the area of contact, J.L the coefficient of friction, )

«
v the unit weight of ice, h the ice thickness, and p the basal
water pressure. 4
Effective Normal PresaurelMPal
As the effective ba al pressure increases, the rate of abra-
sion, for any given sliding velocity and combination of basal Fig. Bll.3 Relationship between theoretical abrasion
debris/glacier bed hardness, will rise (Fig. B11.3). Even- rate and effective basal pressure for different rates of
tually, however, as the basal normal pressure increases, the glacier movement. The actual abrasion rate will depend on
friction between a particle and the glacier bed will reach a the relative hardness of the glacier bed and the
point where the ice begins to flow over the particle and the entrained subglacial debris. (Modified from G. S. Boulton
rate of abrasion begins to fall. Ultimately, with a further (1974), in D. R. Coates (ed.) Glacial Geomorphology. State
increase in effective basal pressure the particle will stop University of New York, Binghamton, Fig. 7, p. 52.)
moving and abra ion will cease.

substantial force can be exerted. The process is probably tonic processes, pressure release or frost wedging. Pres ure
most effective where the bedrock protrusion is much smaller relea e, however, can develop as a consequence of rapid
than the transported boulder, and where the shear strength glacial erosion since bedrock is replaced by much lower
of the protrusion is low. density ice. Moreover, joint widening by frost wedging
Joint-block remo'fal involves the 'plucking' or 'quarrying' may possibly occur under warm-based glaciers with abun-
of large joint-separated blocks by an overriding glacier. Such dant meltwater. Subglacial meltwater is, of course, another
bedrock joints may originate prior to glaciation through tec- potent erosional agent and is considered in Section 11.4.2.
 Glacial processes and landforms 271

Table 11.3 Factors affecting the mechanisms of glacial abrasion

FU DAMENTALFACTORS COMME TS

Presence of debri in basal ice Clean ice is unable to abrade olid rock. The rate of abrasion will increase with debris concentration up to
the point where effective basal sliding is retarded.
Sliding of basal ice Ice frozen to bedrock cannot erode. unless it already contains rock debris. The faster the rate of basal I iding
the more debris pa es a given point per unit time and the faster the rate of abrasion.
Movement of debris towards glacier Unless particle at the base of a glacier are constantly renewed they become poli hed and less effective
base abrasive agents. Thinning of the basal ice by melting or divergent flow around obstacles brings fresh
particles down to the rock-ice interface and increases abrasion.
Other factors affecting nature and rate of ahrasion
lee thickness The greater the thickness of overlying ice the greater the vertical pressure exerted on particles on the glacier
bed and the more effective is abrasion. This i the case up to a depth where friction between particles and
the bed becomes so high that movement is significantly retarded and abrasion decreases.
Basal water pressure The presence of water at the glacier base, especially when at high pressure, can reduce the effective pressure
on particles on the bed and thus abrasion rate by buoying up the glacier. However, sliding velocities may
tend to increase because of the reduced friction.
Relative hardness of debris particles The most effective abrasion occurs when hard rock particles in the glacier base pass over a oft bedrock. If
and bedrock the debris particles are soft in com pari on with the bedrock tHe former are abraded and little bedrock ero ion
is accompli hed.
Debri particle size and shape Since particles embedded in ice exert a downward pres ure proportional to their weight, large block hould
abrade more effectively than small particles. Moreover, angular debris will be a more efficient agent of
abra ion than rounded particles.
Efficient removal of fine debris To sustain high rates of abrasion fine particles need to be removed from the ice-rock interface since they
abrade less effectively than larger particles (assuming the latter are continually being supplied from above).
Meltwater appears to be the main mechanism for the removal of fine «0.2 mm) debris.

Source: Based on discussion in D. E. Sugden and B. S. John (1976) Glaciers and Landscape. Edward Arnold, London, pp. 153-5.

The thermal regime of a glacier exerts a pervasive in- change, makes patterns of glacial erosion potentially very
fluence over processes of glacial erosion. For instance, the complex.
potential for abrasion in cold ice is negligible because of
the lack of basal slip and the absence of a significant load
of basal debris. However, the greater adhesion of cold ice
11.2.2 Debris entrainment and transport
to bedrock would be expected to increase the effectiveness Effective glacial erosion can only continue if the eroded
of joint-block removal. Abrasion is probably confined material is entrained and transported away by the ice flow
largely to warm-based glaciers since here the presence of (Fig. 11.11). The size of glacially transported debris ranges
basal water both promotes sliding and provides a means for from very small rock fragments up to huge boulders.
removing fine debris. The fact that a single glacier may Subglacial debris is transported along the base of a glacier
have both warm- and cold-based sections and that glaciers (Fig. 11.12), but glacier also acquire supraglacial debris
may alter from one type to the other in respon e to climatic through material falling on to the ice urface from rock

Fig.ll.ll Schematic illustration of travel paths of debris transported by valley glaciers.

272 Exogenic processes and landforms

subglacial into englacial debris, and especially in the lower

section of the ablation zone, into supraglacial debris.
In some glaciers englacial debris is distributed fairly
evenly with depth, but in others it is concentrated in bands
separated by relatively clean ice. Such bands run parallel
with flow lines and most are probably formed by the
freezing of subglacial debris prior to its upward movemet:lt
through the glacier. Under warm ice, basal melting produces a
high concentration of subglacial debris which can be attached
on to the base of the glacier by refreezing or regelation;
but as the regelation layer is only a few centimetres thick
this process can only account for thin bands. Creep provides
another mechanism for the incorporation of subglacial debris
into the base of a glacier, and in warm-based glaciers small
Fig. 11.12 Debris exposed at the base ofRusseLL Glacier, amounts of material can probably be squeezed up into sub-
Sf/Jndre Strf/Jmfjord, west Greenland. (Photo courtesy D. E. glacial cavities. Finally, where a glacier is overriding a
Sugden.) large obstruction on its bed, thrusts can develop in the ice
which provide a further means of emplacing englacial
walls or other ice-free areas (Fig. 11.1). Naturally, supra- debris.
glacial debris is likely to be most abundant in valley and
cirque glaciers and absent over large areas of ice sheets. It
can continue to be carried on the glacier surface within the 11.2.3 Erosional landforms
ablation zone, but above the equilibrium line it will become Glacial erosion produces landforms at a wide range of
progressively buried because of the accumulation of ice scales, but irrespective of scale the overriding control is the
from above. Once buried it becomes englacial debris and it attempt by glaciers to modify their beds in such a way as to
can travel as such to the glacier snout. Alternatively, it may improve the efficiency of ice flow. The extent of the ice
emerge at the surface in the ablation zone through melting cover, the thermal regime of the glacier, the amount of
of overlying ice or it can move down to the glacier bed and basal debris, the characteristics of the underlying bedrock,
be trapped by existing subglacial debris. Movement along the period of time during which glacial erosion has been
slip lines in zones of compressive flow tends to transform active and the form of the pre-existing topography all affect

Table 11.4 Landform and landscapes of glacial erosion

PREDOMINANT ASSOCIATED LINEAR DIMENSIONS

PROCESS MORPHOLOGY
0.01 m 0.1 m 1m 10m 100m Ikm 10km 100 km 1000 km 10000 km

Positive, ~ Whalebacks -----I.~ Streamlined ---.

streamlined spur
~ Rock drumlins ----II
••

Landscape
Unconfined Positive, partially Roches moutonnees -------Flyggbergs ~ of areal
ice flow streamlined scouring

Negative, streamlined ~ Striations ~ ~--- Grooves ---II.~

Negative, partially ....~ - - - Rock basins -------III.

streamlined

Channelled Negative, streamlined ~ Troughs - - - - Landscape of linear

ice flow ice-sheet erosion

Interaction of Negative ....~---- Cirques Valley

glacial and ------------------------------ glacier
periglacial Po itive . - Aretes --. landscape
proce es '-Horns~

. - Nunataks--' Nunatak landscape

Source: Modified from D. E. Sugden and B. S. John (1976) Glaciers and Landscape. Edward Arnold, London, Table 9.2, p. 169.
 Glacial processes and landforms 273

the nature of glacially eroded landscapes. A useful distinc- plays a role since it might be more effective where basal
tion can be made between those forms associated with shear stresses are highest under thicker ice occurring over
largely unconfined ice movement, those produced by chan- topographic depressions; in such a case an initially shallow
nelled glacier flow and those created through the interac- depression might then be prone to overdeepening.
tion of glacial and periglacial processes. This classification
forms the basis for our discussion of glacially eroded land- 11.2.3.2 Forms associated with channelled ice flow
forms (Table 11.4). Ice flow concentrated in channels gives rise to steep-sided,
glacial troughs (Figs 11.1, 11.2, 11.10). These may be
11.2.3.1 Forms associated with unconfined ice flow formed by valley or outlet glaciers, or by ice streams occur-
Unconfined ice flow is restricted to ice sheets and ice caps. ring within ice sheets and ice caps. The walls of glacial
Although the erosional activity below such ice masses is troughs, which may exceed 1000 m in height, frequently
probably for the most part limited to regions of warm basal truncate the tributary valleys and spurs of the pre~existing
ice a range of positive landform features have been attributed fluvially-eroded landscape thereby giving rise to hanging
to ice-sheet erosion. They include small, smoothly eroded valleys (Fig. 11.13). The cross-profile of many glacial troughs
forms a few hundred metres in length known as whale- approximates to a parabola and this form probably results
backs, larger streamlined hills sometimes called rock from a combination of higher basal flow rates, and there-
drumlins and tapered interfluves and spurs up to several fore more effective erosion, beneath the middle of the
kilometres long. Although there are doubts about the inter- glacier and fluctuations in the height of the surface over
pretation of some of the larger landforms their alignment time leading to less direct glacial erosion of the upper parts
with known directions of ice movement and their smoothed of the trough.
surfaces support a glacial origin. There is certainly no Three varieties of glacial trough can be distinguished.
doubt about the glacial nature of the asymmetric stream- Alpine troughs are eroded by valley glaciers whose accu-
lined form known as a roche moutonnee, and its larger mulation zones lie below a mountain mass. Icelandic
variant known as a flyggberg. Roches moutonnees range in troughs, by contrast, are formed by glaciers flowing from
size up to the dimensions of whalebacks. They have a ice sheets or ice caps over the trough head and some enor-
smoothed end facing the direction of ice flow but a craggy, mous example have been formed by outlet glaciers flow-
steeper lee side. Although it is apparent that they are ing from the Antarctic ice sheet. Where coastal Icelandic
formed by abrasion on their up-glacier side and joint-block troughs are partially drowned by rising sea level they form
removal on their lee side, it is not clear why their down- fjords. A third type, open troughs, are so called because
glacier side is not also smoothed by abrasion as is the case they are open at both ends. They frequently breach water-
in fully streamlined forms. This may be related to the lower sheds and most seem to have beep eroded by ice streams
ice pressure on the lee side of a bedrock obstruction in an within ice sheets or ice caps.
ice flow. The lower pressure would not only reduce abra- Alpine troughs characteristically have an overdeepened
sion but would also enable meltwater to migrate to the lee long profile; near the trough head the floor is steeply in-
side and possibly aid joint enlargement by frost wedging. clined, but down-valley there is a much lower or even
Low lee-side ice pressures would be favoured by thin or slightly reversed gradient. Icelandic troughs typically have
fast moving ice flowing around a prominent obstruction, an excessive steepening towards the trough head and flatter
since in these circumstances the ice would be less able to trough floors, whereas open troughs normally have a high
mould itself closely around the obstacle. point roughly in the middle of their long profile. The loca-
Negative relief forms attributed at least in part to the tion of overdeepening in alpine troughs is probably related
effects of unconfined ice flow include grooves aligned with to the position of the equilibrium line since here the ice is
a known direction of ice movement, and shallow rock thickest and we would expect this to coincide with the zone
basins often filled by a lake. Grooves up to 30 m deep, of maximum erosion. The overdeepening of Icelandic troughs
100 m across and 12 km long occur in the Mackenzie so near the trough head is·more difficult·to explain, but may
Valley in northern Canada· while rock basins, from several be associated with a localized zone of warm-based ice which
metres to a few hundred kilometres across, are more ubi- promotes more rapid erosion than the surrounding cold-
quitous and are to be found over much of northern Canada. based parts of the glacier.
A glacial origin for rock basins is suggested not only by Superimposed on these characteristic long profile forms
evidence of abrasion but also by their overdeepening, that are often smaller-scale irregularities consisting of alter-
is, their erosion to levels below regional base levels related nating rock bars and rock basins which form a series of
to fluvial systems. Some basins are aligned along faults and steps in the trough floor (Fig. 11.13). These features have
major joint systems and one explanation of their develop- generated much discussion and explanations have included
ment involves preferential joint-block removal where bed- more effective erosion where ice flow is constricted through a
rock joints are more closely spaced. Abrasion also probably local narrowing of the trough or where tributary glaciers
274 Exogenic processes and landforms

Fig.n.B Valley below the present location of the snout ofAusterdalsbreen, southern Norway. Note the hanging valleys and
truncated spurs, and the rock bar in the centre of the photograph.

enter the main trough, and differential erosion controlled by at the heads of deep valleys. A cirque consists of a bowl-
variations in the spacing of joints in the underlying bedrock shaped rock basin which extends from a steep headwall of
or the degree of pre-glacial weathering. If a trough floor shattered rock to a low rim. Fully developed cirques have a
contained small initial irregularities then any depressions fairly consistent ratio of height to length suggesting that they
would tend to be preferentially eroded, especially under are equilibrium forms which maintain their relative dimen-
warm ice, because of the higher basal shear stress associated sions as they grow in size. They range from modest de-
with the greater thickness of overlying ice. Deepening of pressions a few hundred metres across, to massive
depressions in this way would cause further ice thickening amphitheatre-like forms several kilometres wide and with
and so irregularities would tend to be accentuated by this headwalls several hundreds of metres in height. Wind-
positive feedback mechani m. This provides an interesting blown snow is important in the mass balance of cirque
contrast with fluvial systems in which adjustments between glaciers, so it is not surprising that they are best developed
channel bedform and changes in flow depth and velocity where prevailing winds can bring in large quantities of snow,
are characterized by negative feedback and therefore pro- or at least where accumulated snow is not readily removed
mote the reduction of irregularities. In glacier flow the by the wind.
development of a stepped trough-floor profile is further aided Cirque development tarts with the formation of a fim
by the existence of zones of compres ive and extending flow bank in a suitable depression. This i gradually enlarged by
since concavities in the glacier bed tend to be eroded to a combination of processes known as nivation ( ee Section
conform to slip lines. 12.2.4) involving active frost weathering and mass move-
ment processes promoted by the presence of meltwater
11.2.3.3 Forms associated with periglacial action around the fim bank. Deepening proceeds until the fim
Cirques (also called corries and cwms) are the major tums to ice which then begins to flow. Because cirque
landforms developed by a combination of periglacial action glaciers have a rather unusual bed configuration their flow
(see Chapter 12) and glacial erosion and are usually located characteristics are rather different from other glacier types.
 Glacial processes and landforms 275

cirque development focused along plateau margins create

Debris on
glacier deeply notched plateau edges giving rise to so-called
surface biscuit board topography. Cirque growth may eventually
progress to the point where two encroaching cirque head-
walls are only separated by a narrow ridge, known as an
arete, or where several headwalls converge, by a sharp
mountain peak or horn (Fig. 11.15). A summary of the
Debris sequence of features typically associated with the glaciation
path o'f mountain regions and created by the combined action of
channelled ice flow and the interaction of glacial and
periglacial processes is provided by Figure 11.16.
Periglacial processes also play an important role in the
shaping of nunataks. These are mountain peaks of limited
extent and entirely surrounded by glacial ice. They are most
common where an ice sheet or ice cap almost entirely
buries a mountain range with only the highest peaks re-
Fig.11.14 Schematic longitudinal section through a cirque maining above the ice urface. After wasting of the ice
glacier. Note theform of the flow lines and the path taken by there may be a significant contrast between the irregular,
debris which falls on to the glacier surface. sharp-edged topography characteristic of nunataks and the
somewhat smoother and more subdued forms of the sur-
rounding glacially modified terrain.

The regular concave long profile form induces rotational

sliding, and the flow lines are inclined away from the glacier
surface near the headwall and towards the surface near the
terminus. This is illustrated by the movement within the 11.3 Glacial deposition
glacier of rock debris which falls on to the surface of the Before we look at the processes of glacial deposition it is
ice at the base of the headwall (Fig. 11.14). As cirque glaciers useful to draw a distinction between the deposits laid down
become larger it is likely that rotational slip becomes less by ice, known as till or boulder clay, and the landforms
important and internal deformation more so. produced by such deposits, which are termed moraine. Al-
In order to maintain the rather con tant ratio between though tills are highly variable deposits they usually have
height and length, deepening of the cirque floor, largely by certain characteristics in cornmon. These are listed in Table
glacial abrasion, must be accompanied by retreat of the 11.5, but it must be emphasized that anyone till deposit is
headwall. This occurs mainly through the removal of large unlikely to exhibit all these features; moreover, deposits
blocks of bedrock, but uncertainty surrounds the mechanism originally deposited by ice may subsequently be reworked
whereby these blocks are produced. It has long been by glacial meltwater (see Section 11.4.1).
observed that the top layer of actively moving ice on a
cirque glacier is usually separated from the headwall by a
deep chasm known as a bergschrund. The bottom of a
Table 11.5 Characteristics of till
bergschrund would be a likely collecting point for melt-
water and, being exposed to the atmosphere, it might also 1. Poor sorting - there i a large range in grain size, with large clasts up
experience frequent temperature fluctuations across freez- to boulder size frequently contained in a finer, sometimes clayey,
ing point and therefore be a site of active enlargement of matrix

pressure-release joints by frost weathering. This superfi- 2. Lack of stratification - laminations and graded bedding (progressive
change in grain size with depth) are generally absent except in
cially plausible hypothesis unfortunately lacks supporting deposits modified by meltwater which may exhibit stratification
evidence since the little data collected do not indicate 3. Mixture of lithologies - particles may have been derived from
temperature changes which are either sufficiently frequent widely eparated sources, especially in the case of deposits laid
down by large ice sheets.
or rapid to promote intense frost weathering. A possible
alternative mechanism for rock break-up i hydration shat- 4. Frequent presence of particles with abraded facets and triations.

tering (see Section 6.3.1) although the effectiveness of this 5. Preferred orientation of particles
6. Compaction associated with pressure developed during depo ition
process in a wide range of rock types has yet to be firmly
established. 7. Overlies striated rock or sediment basement

Whatever the process of headwall erosion, cirque growth 8. Predominantly subangular particles due to a combination of
fracturing and rounding by abrasion
can consume large areas of upland. The early stages of
276 Exogenic processes and landforms

Fig.ll.15 The horn of Mount Cook, the highest mountain in the Southern Alps, South Island, New Zealand. (Photo courtesy G. M.
Robinson).

11.3.1 Mechanisms of deposition subsequently re-entrained and re-deposited elsewhere (Fig.

11.17).
Glacial deposition involves a range of processes and any At least three distinct mechanisms of subglacial deposi-
classification is somewhat arbitrary (Table 11.6). An added tion can be identified. Undermelt involves the deposition
complication is that researchers have used different terms of material through melting of the underlying ice. Melting
to describe the same process. Here we make a broad group- may arise from geothermal heat, through frictional heating
ing of processes on the basis of whether they operate at the or as a result of increased pressures around the up-glacier
base, on the surface, or around the margins of a glacier. side of obstructions. A second process, called basal lodge-
Nevertheless, material deposited in one situation may be ment, entails the smearing of predominantly- fine material

Table 11.6 Classification of mechanisms of glacial deposition

LOCATION RELATIVE TO PROCESS THERMAL AND DYNAMIC CO DITIONS

GLACIER

Subglacial Undennelt Wann-based only, active or stagnant

Basal lodgement Predominantly wann-based, active only
Basal flowage Wann-based only, active or stagnant

Supraglacial Meltout Active or stagnant

Flowage Active or stagnant

Marginal Dumping Active only

Pushing Wann- or cold-based, active only
 Glacial processes and landforms 277

of warm glaciers where up to 20 m of ablation may occur in

a single summer. Supraglacial flowage, involving the
movement down the ice surface of the debris-rich upper
layers of a glacier, has also been observed and is particular-
ly common near the snout where intense supraglacial melt-
out occurs. The movement ranges from slow creep to a
rapid liquid flow.
Deposition around a glacier margin can arise from a
variety of mechanisms. Water-soaked till can be squeezed
from under the ice, and if the glacier margin remains stable
for some time significant accumulations of supraglacial and
englacial debris can form through dumping of material by
meltout. Existing glacial deposits can also be pushed by an
actively advancing glacier and deposited when the advance
stops, or be overridden and recycled towards the snout as
englacial and supraglacial debri . Material deposited at a
glacier margin commonly suffers extreme disruption prior
to deposition so that any original orientation of particles in
the ice is usually lost, except possibly where a frozen mass
of till is bulldozed en masse.

11.3.2 Depositional forms

Most ice-laid deposit have a definite three-dimensional
shape, that is, they form moraines. Few moraines are com-
posed entirely of till - they also contain stratified sediments
deposited by meltwater and, in some cases, even a bedrock
core. Another characteristic of many moraines is their tran-
sience. Those occurring around the margins of a glacier
often have an ice core and they collapse into insignificant
forms when this melts. Moraines can also be destroyed or
Fig.11.16 Development of landforms associated with alpine drastically modified by meltwater and, in the case of those
glaciation: (A) a fluvially eroded mountain landscape prior to formed subglacially, can be modified by glacial activity at
glaciation; (B) the initial accumulation of snow and ice; (C) full the ice margin when the glacier retreats.
development of a network of valley glaciers; (D) after Perhaps the most helpful way to classify such complex
deglaciation showing landforms of mountain glaciation including
cirques, Q!'etes, horns, truncated spurs, hanging valleys and a landforms as moraines is to categorize them in terms of
glacial trough. (After R. F. Flint (l97i) Glacial and Quaternary their relationship to the direction of ice flow. On this basis
Geology, Wiley, New York, Fig. 6.3, p. 140.) there are three major types: those aligned largely parallel to
the direction of ice flow, those orientated roughly trans-
verse to the direction of ice flow and those lacking any con-
on the glacier bed. Lodgement of subglacial debris occurs sistent orientation (Table 11.7).
where the friction between a particle being transported by
the ice and the glacier bed becomes so great that its further 11.3.2 .1 Forms parallel to the direction of ice flow
movement is retarded. This situation is favoured by thick Moraines orientated roughly parallel with the direction of
ice promoting high basal pressures which increases friction ice flow can be formed subglacially, supraglacially, or at the
on the glacier bed and by low flow velocities which are ice margin. Most of the subglacially produced forms seem
associated with weak basal shear stresses. Basal flowage is to result from streamlining related to variations in stress
a third mechanism of subglacial deposition but it is also partly across the glacier bed. Often formless sheets of till, known
erosional. It involves both the squeezing of unconsolidated as ground moraine, can be modified subglacially by the
water soaked debris into basal ice concavities and the stream- ice flow into fluted forms up to 10 m high and 1 km long,
lining of till by overrriding ice (see Section 11.3.2.1). or exceptionally into megaflutes up to 25 m high and 20 km
Supraglacial debris deposition can occur by meltout or long. That they can be either predominantly depositional or
by flowage. Meltout involves deposition of sediment through erosional features is highlighted by the distinction made
melting of the glacier surface. It is most active in the snout between fluted ground moraine, in which the fluting is
278 Exogenic processes and landforms

DEBRIS
SUPPLY

TEMPORARY
DEPOSITION

Primary routes

Secondary
complex
routes

Fig. 11.17 Schematic representation of the possible routes that debris may follow in the glacial and jluvioglacial system prior to final
deposition.

Table 11.7 Classification of the major types of moraine

PARALLEL TO ICE FLOW TRANSVERSE TO ICE FLOW LACKI G CO SISTENT ORIENTATION

Subglacial forms with streamlining Subglacial forms Subglacial forms

Fluted and drumlinized ground moraine Rogen or ribbed moraine Low-relief ground moraine
Drumlins and drumlinoid ridges De Geer or washboard moraine Hummocky ground moraine
Crag-and-tail ridges Subglacial thrust moraine
Sublacustrine moraine
Ice-pressed forms Ice-pressed forms Ice-pressed forms
Longitudinal squeezed ridges Minor transverse squeezed ridges Random or rectilinear squeezed ridges

Ice marginal forms Ice front forms Ice sUiface forms

Lateral and medial moraines End moraines Disintegration moraines
Some interlobate and kame moraines Push moraines
Ice thrust/shear moraines
Some kame and delta moraines

Source: Modified from D. E. Sugden and B. S. John (1976) Glaciers and Landscape. Edward Arnold, London, Table 12.1, p. 236,
after V. K. Prest (1968) Geological Survey Papers Canada 67-57.

due to troughs within the till, and drumlinized ground in plan drumlins tend to be bluntly rounded on their up-
moraine where ridges rise above the general level of the glacier margin and sharply pointed on their lee side, while
till surface and more clearly represent constructional relief in profile they are usually highest towards their up-glacier
forms. There are numerous theories for the formation of end. Typical dimensions are 1-2 Ian in length, 5-50 m in
fluted moraine, but several point to the role of large boulders height and around 500 m in width. While some drumlins
on the glacier bed (Fig. 11.18). Constructional flutings appear to be composed entirely of clay-rich till, others have
could be formed by unfrozen till being forced into cavities a bedrock core. They rarely occur singly, much more com-
on the up- and down-glacier sides of an obstacle. monly forming drumlin fields in which individual drumlins
The most intensively studied subglacial moraine form is are randomly spaced.
the drumlin. Although morphologically somewhat variable, Drumlins remain difficult landforms to explain. The
 Glacial processes and landforms 279

Direction of Low pressure

ice movement zone
Direction of
ice movement

Fig.11.18 A probable model offluted moraine formation which Fig.11.19 Schematic representation of migration of debris on
may be applicable to most types offluted moraine ridges. (Based the flanks of a growing drumlin. (Based on D. E. Sugden and B. S.
on D. E. Sugden and B. S. John, (1976) Glaciers and Landscape. John (1976) Glaciers and Landscape. (Edward Arnold, London)
Edward Arnold, London, Fig. 12.2, p. 239.) Fig. 12.4, p.241.)

existence of a bedrock core in some examples suggests that a glacier bed will resist movement until the stress builds up
they may form around obstructions. But this is at best an to a sufficiently high level to induce the di1atant charac-
incomplete explanation since it does not account for those teristic. Once this level is reached the till will begin to
forms lacking a bedrock core. The concentration of drumlins deform readily and will continue to deform, even when the
in specific areas suggests that the initial accumulation of stress is decreased, until a lower critical threshold is reached
material is influenced by local conditions. Probably the and the dilatant property is lost. Imagine debris below thick
most significant are variations in bed roughness with ice. Here subglacial debris is moving freely in its dilatant
subglacial debris being preferentially deposited where the state since the basal stress is high. As the ice thins towards
bed is irregular. Bedrock protrusions or large boulders will the glacier margin the basal stress will gradually decrease
encourage debris to collect on their up-glacier side while until the critical lower limit for dilatancy is reached. The
deposition will also occur on the lee side where shear stresses basal debris will now suddenly revert to a compact form
are less. An initial stage of this kind of deposition is re- and resist further movement by the ice. As it is likely that
presented by crag-and-tail forms which consist of a tail of there will be variations in stress across the glacier bed,
glacial deposits in the lee of a rock obstruction. The zones of compact, stationary till will be surrounded by dila-
drumlin form appears to develop through variations in basal tant material which continues to be transported. The zones
pressure associated with ice flow over an initial irregularity of compact till will then start to be moulded by the still-
(Fig. 11.19). Drumlin crests would be zones of low pres- moving, debris-rich basal ice into a streamlined form offer-
sure while the troughs between would experience high ing least resistance to ice flow.
pressures. Particles being transported over the drumlin by Although widely regarded as ice-moulded forms, it has
the ice flow would tend to move across the gradient of also been suggested that some drumlins, at least, might be
decreasing basal ice pressure and be deposited on the crest. formed by subglacial meltwater. One idea is that drumlins
This kind of mechanism accords with the observation that can be formed when sedimentation occurs in cavities eroded
in many drumlin fields there is only a very sparse till cover by meltwater flow directed upwards towards the base of the
between drumlins. ice. Another possibility is that drumlins can be remnant
Variations in the mechanical properties of till may also erosional ridges created by meltwater flow. These kinds of
be important in drumlin formation, with subglacial deposi- drumlins are composed of pre-existing bedrock or sediment
tion occurring where till is more resistant to deformation. left protruding as the surrounding material is removed by
This might take place where subglacial meltwater is effi- meltwater erosion.
ciently drained from till which overlies a permeable bedrock. Moraines deposited on an ice sprface or along a glacier
Another explanation focuses on the role of dilatancy (see margin have a far lower chance of survival than forms of
Section 7.1.1). Most tills have the property of dilatancy; subglacial origin. Medial moraines are formed from supra-
that is, in very simplified terms, they expand when subject glacial debris concentrated in a thin ribbon in mid-glacier
to stress within certain limits. It is envisaged that debris on below the confluence of two tributary glaciers (Fig. 11.1).
280 Exogenic processes and landforms

Direction of
They can be significant features in the upper part of a glacier rce movement

but they usually only remain as a thin cover of till near the
snout. Lateral moraines are formed primarily from frost-
shattered debris which has fallen on to the edge of a glacier
from the adjacent rock walls and are more likely to survive
since they may lie partly on a bedrock substrate. Vertical
fluctuations in the level of the glacier can lead to lateral
moraines being perched high above the ice surface.

11.3.2.2 Forms transverse to the direction of ice flow

Perhaps the most obvious place to expect moraines aligned
perpendicular to the direction of ice movement is along the
front of a glacier. End moraines form in this position and
they mark the maximum extent of a glacier margin. Those
associated with continental ice sheets may be up to 100 m Fig.n.20 A model of the formation of ribbed (Ragen) moraine.
high and extend for tens of kilometres. A more or less Debris accumulates at the base of shear planes in zones of
compressive ice flow. In addition to particle movement along
parallel series of end moraines can often be found repre-
pressure gradients, the process offormation may also involve
senting stages of ice retreat. End moraines are usually pressure melting and lodgement where the shear planes 'take off
composed of both supraglacially and englacially transported from the glacier bed. (Based on D. E. Sugden and B. S. John
debris together with fluvioglacial and lacustrine sediments, (1976) Glaciers and Landscape. Edward Arnold, London,
and are formed by a combination of dumping, meltout and Fig. 12.8, p.245.)
flowage. Debris brought to the ice surface along flow lines
and shear planes developed near the glacier snout can be
dumped by sliding off ice-cored ridges on the glacier surface thrusting and therefore be allied to ribbed moraines or
or by collapse due to melting of underlying ice. The rate of Rogen moraines. These consist of fairly regularly spaced
melting can significantly affect the detailed form of the ridges, usually arcuate in plan and concave up-glacier, typi-
moraine. Where melting is rapid and meltwater abundant cally 10 m or more in height and up to 1 km long. Many,
supraglacial till can flow down even shallow slopes and though not all, have streamlined 'drumlinized' crests. One
rather featureless accumulations of till are produced. Where interpretation is that they form through preferential deposi-
melting occurs slowly debris can collect in troughs lying tion where shear stresses transverse to the direction of flow
between ice-cored ridges. As melting progresses these are relieved, and experience subsequent streamlining by the
ridges collapse and the originally debris-filled troughs ice (Fig. 11.20). Another explanation involves a genetic
become upstanding ridges as the topography is inverted. link between Rogen moraine and fluting, the development
Small end moraines, less than 10 m high, can also be form- of the former being seen as a response to variations in basal
ed by the squeezing of water-soaked debris from beneath stress, or the presence of topographic irregularities, across
the ice front or into subglacial crevasses. Moraines developed the glacier bed.
by this processes of basal flowage may be produced sea-
sonally when summer meltwater percolates to the bed of a 11.3.2.3 Forms lacking consistent alignment
warm-based glacier and saturates the underlying till. Some moraines have no obvious orientation or linear
Push moraines are another form constructed along an development and therefore fail to fit into either of the
ice front. They are produced by the bulldozing of glacial, morphological categories discussed above. Ground moraine
and even non-glacial, deposits by an advancing ice margin may occur as virtually featureless sheets of till, but it can
and can attain considerable dimensions - up to 100 m high also form a gently undulating topography or even a chaotic
and 30 km long. The pushing action of the ice can create hummocky terrain with a relief in excess of 100 m. This
complex thrust and fold structures in the till and typically may develop from debris which has accumulated in basal
results in a distal slope which is much steeper than the ice concavities or from the meltout of supraglacial debris
slope facing the glacier front. spread extensively over an area of wasting ice. On the mar-
Ice-margin processes are not the only mechanisms capable gins of a retreating ice sheet large masses of ice can become
of generating transverse moraines. Transverse ridges up to stagnant through being isolated from a source of active ice
a few metres in height, in many cases associated with lakes flow. Such stagnant ice will gradually be reduced by abla-
or former lakes, are known as washboard moraines, or De tion to discrete masses of decaying debris-rich ice and
Geer moraines. Some examples may be subaqueous push eventually a chaotic assemblage of disintegrating moraines
moraines, but others may owe their origin to subglacial will be created forming dead-ice topography.
 Glacial processes and landforms 281

11.4 Fluvioglacial erosion and deposition water and by the frictional heat generated by the flow of the
meltwater itself.
11.4.1 Glacial meltwater Depending on the thermal regime of the glacier, meltwater
Meltwater is an integral part of the glacial ystem (Fig. may travel over the glacier ~urface or move englacially and
11.21). It both significantly influences glacier behaviour subglacially (Fig. 11.22). Surface runoff may be concen-
through its effect on rates of basal sliding and has an ero- trated into stream channels up to several metres deep, in
sional and depositional role associated with, but distinct some cases sunk into 'valleys' on the ice surface. Extensive
from, that of glacier ice. On most glaciers surface melting drainage networks covering hundreds of square kilometres
is by far the most important source of meltwater, the can develop on ice sheet . Stream courses may be influenced
amount released increasing down through the ablation zone. by ice surface irregularities, but very regular meandering
In temperate regions this surface source may be supple- patterns have also been reported. Internal drainage within a
mented seasonally by rainfall and, in valley glaciers, by glacier evolves in a manner somewhat analogous. to the
runoff from valley-side slopes. The chief basal and internal formation of underground drainage in highly soluble rocks
sources of meltwater are basal melting through geothermal such as limestones. Initially, water permeates warm ice
and frictional heating, although ground water can be a along the boundaries separating individual ice crystals, but
significant additional source in temperate valley glaciers. eventually a secondary permeability develops with the crea-
Geothermal heat is capable of melting about a 6 mm layer, tion of surface sinkholes, known as moulins, and tunnel
and heat generated by basal sliding and internal deform- which tend to form preferentially along cracks and cavities
ation a 10-15 mm layer, of warm ice annually. Further basal in the ice. Glacier tunnels form dendritic networks both
melting can arise from the heat brought down by surface within, and at the base of, the ice, their direction and gradient
being controlled both by the general slope of the glacier
surface and the subglacial topography. In addition to basal
channels, water can collect to form large subglacial lakes. It
is suspected that these may be quite extensive below large
ice sheets.
Rates of meltwater flow from glaciers vary over a range
of time scales. Short-term fluctuations occur in response to
diurnal variations in ablation and changing meteorological
conditions. Seasonally the lowest flows usually occur in
late winter and the highest in early summer - how early
depending on the rapidity with which the internal drainage
channels are re-established after winter freezing and closure
by ice flow. Significant longer-term maxima in meltwater
discharge are associated with glacier surges, but the most
spectacular peak flows are caused by the catastrophic drain-
ing of ice-dammed lakes. Di charges several orders of
magnitude greater than normal are attained in a few hours
and diminish again even more quickly. Such jokulhlaups
are particularly common in Iceland where some are asso-
ciated with volcanism (see Section 5.2.3.4). One originating
from the western part of the Vatnajokull ice cap with a period-
icity of about IDa has reached an estimated peak discharge
of 40000-50000 m 3s- l , while another in Baffin Island,
Canada, involved the draining of a 5 x 106 m3 subglacial
lake in only 30 h. Such massive discharges could arise in
various ways, but one possibility is that a gradually extend-
ing internal drainage system eventually taps an ice-dammed,
subglacia11ake.
Even more cataclysmic jokulhlaups were apparently
associated with the draining of Late Pleistocene Lake
Missoula in the north-west USA (Fig. 11.23). This proglacial
Fig.ll.21 Meltwater emerging from the snout ofNigardsbreen, lake, which extended across a large area of western Montana,
a valley glacier in southern Norway. was the largest of a number in the region impounded by ice
advancing from the north. It is estimated that at its maxi-
282 Exogenic processes and landforms

mum Lake Missoula reached a volume of 2000 km 3 , and

much of this water was released during repeated failures of
its ice dam. The peak discharge attained during break-out
A
events is thought to have been a staggering 21.3 x 106 m 3 S-I,
or approximately 20 times the mean discharge of all the
world's rivers! Several separate flood events occurred be-
tween 16 and 12 ka BP which have been reconstructed from
the extraordinary landforms they produced. These include a
series of enormous streamless canyons, or coulees, which
form an anastomosing network and which contain giant
'ripples', plunge basins, deep rock basins and vast potholes
B (Fig. 11.24).
This whole region of anomalous topography is called
the Channeled Scabland and its origin was the subject of
intense debate after a catastrophic flooding hypothesis was
first put forward by J. Harlen Bretz in a paper published in
i[::::ij::::::::::::I::::! .
1923. This explanation was initially rejected by his
::::::::::.:::~:.::::::::::::::::::: ::::::;::::::::::::::::::::.::.::.::.::.:.".:: .
':::::::':':':':':::':::::::::::::::::::::.:'::::~.::: contemporaries, both because Bretz had no convincing
source for such enormous volumes of water (he was
unaware of literature documenting the existence of Lake
c Missoula published in the 1880s), and because they were
unwilling, on a priori grounds, to accept the notion of
catastrophic rather than gradual landscape modification.
Bretz's hypothesis was eventually substantiated by his own
persistent field work and by other workers who made the
link between catastrophic flooding and Lake Missoula.
Subsequent work has begun to relate the specific dimen-
sions of the landforms present to flow conditions. Of par-
ticular interest are the giant current ripples found in the
channels. These are composed of gravel and are typically
around 5 m high and have a wavelength of 100 m or so.
From these dimensions it is possible to infer mean flow
velocities, bed shear stresses and stream power values. It
seems that the floods that formed these ripples could
transport boulders as large as 10 m across.

11.4.2 Fluvioglacial denudation

The sediment-charged nature of glacial meltwater streams
D Cold ice
Water pockets
or subglacial
lakes
is testament to their transportational role and suggestive of
their erosional capabilities. Although data on sediment yields
are scarce it is known that the suspended loads of meltwater

..
Warm ice Meltwater flow.
Length approx. channels may attain concentrations in excess of 3000 mg 1-1.
proportional to Their characteristic greyish-white colour is due to an abun-
Ice below discharge dance of the fine sediment generated by glacial abrasion,
water table
Internal routes but up to 25 per cent or more of the total load may be
carried as bed load. Peak yields are usually reached in the
Fig.11.22 Possible melnvater routes in different types of
glaciers and glacial ice: (A) cold ice; (B) warm ice; (C) cold ice
early part of the summer presumably due to the flushing out
in the ablation zone and warm ice in the accumulation zone; (D) of sediment generated during the preceding winter. The
a high altitude glacier in equatorial latitudes. (After D. E. Sugden limited information available indicates that meltwater streams
and B. S. John (1976) Glaciers and Landscape. Edward Arnold, also probably carry a significant solute load. The high sur-
London, Fig. 14.18, p. 298.) face area of unweathered rock provided by fine suspended
sediment and the highly turbulent flows in which it is trans-
ported might be expected to promote active chemical
 Glacial processes and landforms 283

• Vancouver
CANADA

•
----------
Lakes MONTANA

D. '··..'.
" Flooding

Flood flow
direction

Margins of
ice sheet and
ice caps

'\... Political
\ boundaries

o 200
I I
km

t Lake
Bonneville

Fig.n.23 Area of the north-west USA affected by catastrophic flooding at the end of the last glacial. In addition to catastrophic
outbursts from Lake Missoula, flooding also occurred as a result of overspills from Lake Bonneville in Utah and from other ice-
dammed lakes in the Columbia Basin. (Modified from V. R. Baker and R. C. Bunker (1985) Quaternary Science Reviews 4, Fig. I, p.
2.)

weathering. Much chemical weathering undoubtedly occurs or simply as p-forms. Elongated varieties exhibiting stria-
subglacially and this may indirectly increase erosion rates tions are certainly likely to be due to glacial abrasion but the
by weakening the bedrock. During winter low flows the origin of another type, known as sichelwannen, which occur
solute load may even exceed the solid load. in resistant crystalline rocks, is less certain (Fig. 11.25).
High flow velocities and especially high peak discharges They consist of crescentic depressions up to 5 m or more
suggest that meltwater streams should be potent erosional across and a f1uvioglacial origin is suggested by the form-
agents. Although fine rock flour usually represents the ation by differential fluvial erosion of similar features in
major component of the suspended load it probably has a less resistant rock.
negligible erosional effect. The coarser suspended and sal- The major large-scale erosional f1uvioglacial landform is
tating load, however, is apparently a highly effective abra- the meltwater channel. These can be up to 100 m or more
sional agent; indeed there is observational evidence of the in depth and extend for tens of kilometres. They take various
cutting of meltwater streams into resistant bedrock in just a forms but can be broadly grouped into subglacial ice-directed
few years. Since high flow velocities in the range 8-15 m S-l forms, where active ice movement has exerted a major
are probably quite common in meltwater streams, and because control, and marginal and submarginal forms which tend to
rough channel beds will be the norm, cavitation may also run parallel to the glacier margin. Ice-directed channels are
be an important erosional process especially where stream generally aligned parallel to the direction of ice movement.
channels narrow. Some breach pre-glacial drainage divides while others run
Various minor erosional forms consisting of a range of downslope parallel with divide crests. They occur singly or
smooth rock depressions have been attributed to the action as bifurcating or anastomosing networks. A subglacial origin
of meltwater, although for some of these alternative ex- for such channels is supported by their up-and-down long
planations involving glacial erosion have been proposed. profiles, which are difficult to explain by anything other
Collectively they are known as plastically sculptured forms, than subglacial water flow under hydrostatic pressure, and
284 Exogenic processes and landforms

Fig.ll.24 Landsat image of the Channeled Scabland, Washington, USA. The vast anastomosing channel patterns are clearly visible
because they were cut through a cover of loess, thereby exposing the dark-coloured underlying basalt of the Columbia Plateau. The
area covered is about i80 km across. (image courtesy N. M. Short.)

by the presence of ice moulding on the upper slopes of some While submarginal channels are more lik~~y to form in
examples. association with warm ice, true marginal channels will
The distinction between subglacial and marginal channels develop along the margins of cold glaciers since here there
is not always clear since the latter can range from channels is little opportunity for meltwater to percolate along the
cut wholly in lateral moraine or bedrock to submarginal frozen ice-rock interface.
channels located subglacially at the extreme edge of the
glacier. The marginal and submarginal environment is cer-
tainly a likely focus for meltwater generated in the zone of 11.4.3 Fluvioglacial deposition
very active ablation of a valley glacier close to the trough Sediment transported by meltwater can be deposited either
wall, and for non-glacial runoff from adjacent valley slopes. in contact with the glacier or beyond the ice front in the
 Glacial processes and landforms 285

Proglacial features are much more likely to survive especially

if the ice front is receding.
The basic mechanisms of deposition from meltwater are
the same as those for ordinary fluvial systems, except that
deposition within a glacier may occur under high hydro-
static pressures if meltwater completely fills an englacial or
subglacial tunnel. There are some important differences,
however, in the environment of deposition; for instance,
tunnels within ice are not usually free to migrate laterally as
fluvial channels may do and the rapid temporal fluctuations
in meltwater discharge can produce equally rapid vertical
changes in the calibre of sediment deposited which may
range from large boulders to sand. The key characteristics
which distinguish water-laid fluvioglacial sediments from
Fig. 11.25 Crescentic scour marks, or sichelwannen, S¢ndre glacial deposits is the absence of fine particles due to sort-
Stromfjord, west Greenland. These features are probably formed ing and the presence of stratification. Meltwater deposits
as a result of erosion by high-velocity, sediment-charged may be less easy to differentiate from ordinary fluvial sedi-
subglacial meltwater flows. (Photo courtesy D. E. Sugden.)
ments, although the glacial sediment source and the typically
short distance of transport before deposition means that the
former are usually less rounded.
proglacial environment. This distinction provides a convenient Meltwater can deposit sediment on, in, under and along
basis for classifying fluvioglacial landforms (Table 11.8). the margins of glaciers (Fig. 11.26). Although most of the
Wide temporal variations in meltwater discharge and the ice-contact landforms to which this deposition gives rise
potential for reincorporation within a glacier of fluviogla- are formed in channels, some may also be constructed either
cially deposited debris by refreezing means that many ice- subglacially or marginally in lakes. The terminology used
contact forms are transient features, at least in active ice. to describe these features is extremely confused. Here we

Table 11.8 A classification of f1uvioglacial deposits

DOMINANT ENVIRONMENT GENERAL RELATIONSHIP GENETIC

SEDIMENT FORM TO ICE TERM

lIce-contact deposits Ridge Marginal, Esker

Sand subglacial,
and Fluvial englacial, Kame
gravel Mound supraglacial Kame complex

Spread with Marginal Kettled sandur

depressions
Proglacial deposits
Sand and Fluvial Proglacial Sandur
gravel
Spread
Silt and clay Lake plain

Sand and Lacustrine Terraces, ProglaciaV Beach

gravel ridges marginal

Clay sand Terrace Kame delta

and gravel

Silt and clay Spread Raised mud flat

Sand and Marine Terraces, Raised beach

gravel ridges

Clay, sand Terrace Raised delta

and gravel

Source: After R. J. Price (1973) Glacial and Fluvioglacial Landforms. Oliver and Boyd,
Edinburgh, Table 3, p. 138.
286 Exogenic processes and landforms

Braided
A stream
Subglacial Lake
channel Supraglacial
stream

B Kettle holes
Kame
terrace
Kame delta

Fig. 11.26 Development of ice-contact fluvioglaciallandforms and deposits: (A) late stage of deglaciation with extensive areas of
stagnant ice and abundant meltwater; (B) after deglaciation. (Based on R. F. Flint, (1971) Glacial and Quaternary Geology. Wiley,
New York, Fig. 8-4,p. 209.)

will simply distinguish between two main types of feature - Kame is a broad term describing a mound of sediment
eskers and kames. formed by the initial deposition of material within a cavity
Eskers are sinuous, sometimes discontinuous, ridges in the ice followed by slumping of this material as the sup-
formed of sand, gravel or boulders and range up to 200 m porting ice walls melt away. Often the term is used as an
in height, 3 km in width and 100 krn or more in length. adjective to describe landforms constructed in particular posi-
Some consist of mounds, lined by narrow ridges and they tion within, or along, a glacier. Kame terraces for instance,
may form single ridges or interconnected networks. Eskers are formed by the later~l or frontal accumulation of fluvio-
may initially be deposited in a variety of positions; they can glacial deposits along the ice margin, while a kame complex
be formed in ice tunnels either subglacially or, much more results from the letting down of numerous sed"iment-filled
rarely, englacially, or they may be deposited in channels on supraglacial depressions and cavities in stagnant ice on to
the glacier surface. Where initially deposited supraglacially the subglacial surface. All varieties of kame may be further
or englacially they can eventually be superimposed on to modified by slumping resulting from the melting of ice
the subglacial topography as the underlying ice melts. Eskers cores within the sediments thereby giving rise to kettle-
may also be deposited along the ice margin, either in asso- holes. Some landscapes described as kettle-and-kame
ciation with marginal and submarginal channels, or in ice- topography are dominated by mounds originally represent-
marginal lakes. ing sites of deposition in depressions and the secondary
 Glacial processes and landforms 287

effects of sediment collapse over melted ice masses. Within the general field of glacier characteristics and
Fluvioglacial deposition is dominant in the proglacial dynamics Paterson (1981) is the basic reference, while
zone both because of the decreased capacity for transport of Armstrong et al. (1973) provide illustrated definitions of
meltwater streams once they emerge from the ice front and ice and snow features. Chapter 5 of Sugden and John (1976)
the common presence of ice-margin lakes in this environ- contains a discussion of the factors controlling the global
ment. The rapid dumping of large quantities of coarse debris distribution of glaciers, and the temperature characteristics
immediately beyond the glacier margin and the associated of glacier ice are considered in Harrison (1975) and Paterson
shifting of stream channels within this zone produces an (1981). The complexity of the mass balances of real glaciers
extensive depositional plain known as a sandur (plural is examined in Mayo et al. (1972) and 0strem (1975) looks
sandar). Sandar are broadly analogous to alluvial fans with at the possibilities of the satellite monitoring of glacial
characteristic entrenchment by the main feeder channels, but budgets.
they experience much more drastic seasonal fluctuations in A general review of glacier movement is provided by
discharge. Valley sandar form in the laterally confined Kamb (1964), while a detailed treatment of internal de-
environment of glacial troughs, whereas plain sandar or formation of glacier ice is contained in the classic papers by
outwash plains can develop along the margins of ice Glen (1952, 1955) and Nye (1957, 1965a). For a similarly
sheets where braided rivers from numerous outlets along in-depth analysis of the mechanisms of basal sliding see
the ice front coalesce. Typically the proximal zone of a Kamb and La Chapelle (1964), Lliboutry (1968), Morris
sandur close to the glacier margin has only a few main (1976), Nye (1970) and Weertman (1964). More general
meltwater streams discharging from the ice front and a reviews can be found in Kamb (1970) and Weertman (1979).
characteristically pitted surface, known as a kettled-sandur, The role of bed deformation in glacier movement is assessed
produced by the melting of ice masses buried by the fluvio- by Boulton and Jones (1979) and Boulton and Hindmarsh
glacially deposited sediments. Channel braiding becomes (1987). Shreve (1984) challenges the conventional distinc-
accentuated further away from the ice front in the inter- tion between cold and warm-based ice and considers the
mediate zone and in the distal zone the channels are so mechanism of glacial sliding at subfreezing temperatures.
shallow that overflow commonly occurs especially during Ice flow in the Antarctic ice sheet is considered by Drewry
peak discharges. The distal zone commonly merges into an (1983) and Bentley (1987), while Alley et al. (1986) and
extensive delta system formed along a proglacial lake. The Blankenship et al. (1986) discuss the role of substrate
sandur may extend away from the glacier front, gradually deformation in the movement of Antarctic Ice Stream B. It
burying the delta as it advances. should be pointed out that a full understanding of most of
these articles on glacier dynamics requires a fairly sophis-
ticated level of mathematical expertise. The brief but very use-
Further reading ful review by Hutter (1982) is, however, readily accessible.
There are several comprehensive surveys of glacial geomor- Turning to the movement of glaciers as a whole, extend-
phology available including Embleton and King (1975), ing and compressive flow is treated by Nye (l952a) and
Flint (1971) and Price (1973), while Andrews (1975) pro- applied to a specific glacier by Meier and Tangborn (1965),
vides a brief but stimulating introduction to the subject. while the operation of kinematic waves is analysed by Nye
Drewry (1986) provides a detailed treatment of glacial pro- (1960, 1965b). Possible mechanisms of surging are con-
cesses but these are rather poorly related to landforms. My sidered by Budd (1975), Jarvis and Clarke (1975), Kamb et
own favoured treatment is that by Sugden and John (1976). al. (1985) and Robin and Weertman (1973), and Clarke et
This excellent text combines breadth with depth and has al. (1984) examine the role of bed deformation in the
provided much of the framework of discussion on which this· surging behaviour of Trapridge Glacier, Canada. Raymond
chapter is based. Useful supplementary material on a variety (1987) and Sharp (l988a) review ideas on the mechanisms
of topics is to be found in Coates (1974), while the volume of surging and Clarke et al. (1986) examine the glacier
edited by Gurnell and Clark (1987) contains various papers characteristics common to surging behaviour and Sharp
on glacial and fluvioglacial sediment transport with parti- (1988b) looks at its geomorphic consequences. Short-term
cular reference to Alpine terrains. glacier fluctuations are covered in Chapter 6 of Sugden
Articles on glacial geomorphology are published fairly and John (1976).
frequently in the general geomorphology journals, but the The large-scale form of ice masses is discussed by Nye
other important sources are the Journal of Glaciology and (l952b) and Weertman (1961), and Reeh (1982) presents a
Annals of Glaciology which emphasize the characteristics model which takes into account the effects of subglacial
of ice and mechanics of glacier flow. Some papers of geo- topography. The properties of the Antarctic ice sheet are
morphic interest are also published in Arctic and Alpine documented by Drewry et al. (1982) and Drewry (1983).
Research, Quaternary Research, Quaternary Science Reviews, Numerous ice-surface features are illustrated in Post and La
Journal of Quaternary Science and Boreas. Chapelle (1971) while Nye (l952a) considers the genera-
288 Exogenic processes and landforms

tion of crevasses and Hambrey (1976), Kamb (1964) and cussed in the classic paper by Mannerfelt (1949) and by
Lliboutry and Reynaud (1981) discuss the formation of Clapperton (1968) and Price (1973). Flint (1971), Parizek
various kinds of ice banding. (1969) and Price (1973) consider the complex range of
The processes of glacial erosion are reviewed by Boulton depositional fluvioglacial forms and Church (1972) and
(1974, 1979) and detailed discussions of particular ero- Krigstrom (1962) provide detailed treatments of the pre-
sional mechanisms are provided by Kamb and La Chapelle glacial environment.
(1964) and McCall (1960) (abrasion), Glen and Lewis
(1961) and McCall (1960) (crushing and fracturing) and
Lewis (1954), Linton (1963) and Trainer (1973) Uoint-
References
block removal). The formation of fluted surfaces at a variety
Alley, R. B., Blankenship, D. D., Bentley, C. R. and Rooney, S. T.
of scales is discussed by Boulton (1974, 1976), Flint
(1986) Deformation of till beneath ice stream B, West Antar-
(1971), Goldthwait (1979) and Linton (1963). The form of ctica. Nature 322,57-9.
glacial troughs is considered by Graf (1970), Harbor et ai. Andrews, J. T. (1975) Glacial Systems. Duxbury Press, North
(1988), Hirano and Aniya (1988) and Linton (1963), and Scituate.
their stepped long profiles are examined by Bakker (1965) Armstrong, T. E., Roberts, B. and Swithinbank, C. (1973) Illus-
and, in the broader context of contrasts between glacial and trated Glossary oj Snow and lee (2nd edn) Scott Polar Research
Institute, Cambridge.
fluvial systems, by King (1970). The morphology of cirques Baker, V. R. and Bunker, R. C. (1985) Catacylsmic Late Pleistocene
and factors controlling their development are treated by flooding from glacial Lake Missoula: a review. Quaternary
Derbyshire and Evans (1976), Haynes (1968) and Olyphant Science Reviews 4, 1--41.
(1981). Overviews of the creation of glacially eroded Bakker, J. P. (1965) A forgotten factor in the interpretation of
glacial stairways. Zeitschriftjur Geomorphologie 9, 18-34.
landscapes are provided by Linton (1963) and Sugden
Baranowski, S. (1970) The origin of fluted moraine at the fronts
(1974) while Sugden (1978) provides a more detailed of contemporary glaciers. Geografiska Annaler 52A, 68-75.
treatment of the morphology generated by the Pleistocene Bentley, C. R. (1987) Antarctic ice streams: a review. Journal oj
Laurentide ice sheet of North America. Geophysical Research 92, 8843-58.
There are several good reviews available on glacial Blankenship, D. D., Bentley, C. R., Rooney, S. T. and Alley, R. B.
(1986) Seismic measurements reveal a saturated porous layer
deposition and associated landforms, including Goldthwait
beneath an active Antarctic ice stream. Nature 322, 54-7.
(1971), Price (1973) and Schluchter (1979). Boulton (1974) Boulton, G. S. (1972) The role of thermal regime in glacial sedi-·
and Kamb and La Chapelle (1964) look at the processes of mentation. Institute oj British Geographers Special Publication
debris entrainment and transport and Boulton (1972, 1975) 4,1-19.
discusses mechanisms of deposition. Prest (1968) provides Boulton, G. S. (1974) Processes and patterns of glacial erosion. In:
D. R. Coates (ed.) Glacial Geomorphology. State University of
a comprehensive classification of depositional landforms.
New York, Binghamton, 41-87.
Fluted moraines are discussed by Baranowski (1970), while Boulton, G. S. 1975. Processes and patterns of sub-glacial sedi-
drumlins are reviewed by Menzies (1979) and considered mentation: a theoretical approach. In: A. E. Wright and F. Moseley
in detail in the symposium volume edited by Menzies and (eds) lee Ages: Ancient and Modern. Seel House Press, Liverpool,
Rose (1987). Specific processes of formation are proposed 7--42.
Boulton, G. S. (1976) The origin of glacially fluted surfaces -
by Evenson (1971), Shaw and Sharpe (1987) and Smalley
observation and theory. Journal ojGlaciology 17, 287-309.
and Unwin (1968). Embleton and King (1975) and Price Boulton, G. S. (1979) Processes of glacier erosion on different
(1973) contain extended treatments of end moraines, and substrata. Journal oj Glaciology 23, 15-38.
Cowan (1968) discusses the formation of ribbed (Rogen) Boulton, G. S. and Hindmarsh, R. C. A. (1987) Sediment deform-
moraine. ation beneath glaciers: rheology and geological consequences.
For excellent discussions of meltwater in glaciers see Journal oj Geophysical Research 92, 9059-82.
Boulton, G. S. and Jones, A. S. (1979) Stability of temperate ice
Shreve (1972), Stenborg (1969) and Weertman (1972). Nye caps and ice sheets resting on beds of deformable sediment.
(1976) provides an additional recent analysis with particular Journal oj Glaciology 24, 29--43.
reference to the generation of jokulhlaups, and Thorarinsson Bretz, J. H. (1923) The channeled scabland of the Columbia
(1953) contains a vivid description of the 1934 Grimsvotn Plateau. Journal oj Geology 31, 617--49.
Bretz, 1. H. (1969) The Lake Missoula floods and the channeled
jokulWaup. The Late Pleistocene Lake Missoula cataclysmic
scablands. Journal oj Geology 77,505--43.
flooding is discussed by Baker and Bunker (1985) and Budd, W. F. (1975) A first simple model for periodically self-
Waitt (1985), but the classic papers by Bretz (1923, 1969) surging glaciers. Journal oj Glaciology 14, 3-21.
are well worth consulting. The relationships between sedi- Church, M. (1972) Baffin Island sandurs. A study of arctic fluvial
ment yield and discharge in a meltwater stream are analysed processes. Canada Geological Survey Bulletin 216.
Clapperton, C. M. (1968) Channels formed by the superimposition
in 0strem et ai. (1967). Plastically sculptured forms (p-
of glacial meltwater streams, with special reference to the east
forms) are discussed by Dahl (1965), but Boulton (1974) Cheviot hills, northeast England. Geografiska Annaler 50,
and Gjessing (1965) provide alternative explanations for 207-20.
their development. The origin of meltwater channels is dis- Clarke, G. K. c., Collins, S. G. and Thompson, D. E. (1984) Flow,
 Glacial processes and landforms 289

thermal structure and subglacial conditions of a surge-type Hutter, K. (1982) Glacier flow. American Scientist 70,26-34.
glacier. Canadian Journal of Earth Sciences 21, 232-40. Jarvi, G. T. and Clarke, G. K. C. (1975) The thermal regime of
Clarke, G. K. c., Schmok, J. P., Simon, c., Ommanney, L. and Trapridge Glacier and its relevance to glacier surging. Journal
Collins, S. G. (1986) Characteristics of surge-type glaciers. of Glaciology 14,235-49.
Journal of Geophysical Research 91, 7165-80. Kamb, B. (1964) Glacier geophy ics. Science 146, 353-65.
Coates, D. R. (ed.) (1974) Glacial Geomorphology. State Univer- Kamb, B. (1970) Sliding motion of glaciers: theory and obser-
sity of New York, Binghamton. vation. Reviews of Geophysics and Space Physics 8, 673-728.
Cowan, W. R. (1968) Ribbed moraine: till fabric analysis and Kamb, B. and La Chapelle, E. (1964) Direct observation of the
origin. Canadian Journal of Earth Sciences 5, 1145-59. mechanism of glacier sliding over bedrock. Journal of Glacio-
Dahl, R. (1965) Plastically sculptured detail forms on rock sur- logy 5, 159-72.
faces in northern Nordland, Norway. Geografiska Annaler 47, Kamb, B., Raymond, C. F., Harrison, W. D., Engelhardt, H.,
83-140. Echelmeyer, K. A., Humphrey, N., Brugman, M. M. and Pfeffer,
Derbyshire,E. and' Evans, 1. S. (1976) The climatic factor in T. (1985) Glacier surge mechanism: 1982-1983 surge of
cirque variation. In: E. Derbyshire (ed.) Geomorphology and Variegated Glacier, Alaska. Science 227, 469-79.
. Climate. Wiley, London and New York, 447-94. King, C. A. M. (1970) Feedback relationships in geomorphology .
Drewry, D. J. (1983) Antarctic ice sheet: aspects of current con- Geografiska Annaler 52A, 147-59.
figuration and flow. In: R. Gardner and H. Scoging (eds) Mega- Krigstrom, A. (1962) Geomorphological studies of sandar plains
geomorphology. Clarendon Press, Oxford, 18-38. and their braided rivers in Iceland. Geografiska Annaler 44,
Drewry, D. (1986) Glacial Geologic Processes. Edward Arnold, 328-46.
London and Baltimore. Lewis, W. V. (1954) Pressure release and glacial erosion. Journal
Drewry, D. J., Jordan, S. R. and Jankowski, E. (1982) Measured of Glaciology 2, 417-22.
properties of the Antarctic ice sheet: surface configuration, ice Linton, D. L. (1963) The forms of glacial erosion. Transactions of
thrckness, volume and bedrock characteristics. Annals of the Institute of British Geographers 33, 1-28.
Glaciology 3, 83-91. Lliboutry, L. (1968) General theory of subglacial cavitation and
Embleton, C. and King, C. A. M. (1975) Glacial Geomorphology. sliding of temperate glaciers. Journal of Glaciology 7, 21-58.
Edward Arnold, London; Halstead, New York. Lliboutry, L. and Reynaud, L. (1981) 'Global dynamics' of a tem-
Evenson, E. B. (1971) The relationship of macro and micro-fabric perate valley glacier, Mer de Glace, and past velocities deduced
of till and the genesis of glacial landforms in Jefferson County, from Forbes' bands. Journal of Glaciology 27, 207-26.
Wisconsin. In: R. P. Goldthwait (ed.) Till: A Symposium. Ohio Mannerfelt, C. M. (1949) Marginal drainage channels as indi-
State University Press, Columbus, 345-64. cators of the gradients of Quaternary ice caps. Geografiska
Flint, R. F. (1971) Glacial and Quaternary Geology. Wiley, New Annaler 31, 194-9.
York. Mayo, L. R., Meier, M. F. and Tangborn, W. V. (1972) A system
Gjessing, J. (1965) On 'plastic scouring' and 'subglacial erosion'. to combine stratigraphic and annual mass-balance systems: a
Norsk Geografisk Tidsskrift 20, 1-37. contribution to the International Hydrological Decade. Journal
Glen, 1. W. (1952) Experiments on the deformation of ice. of Glaciology 11, 3-14.
Journal of Glaciology 2, 111-14. McCall, J. G. (1960) The flow characteristics of a cirque glacier
Glen, J. W. (1955) The creep of polycrystalline ice. Proceedings and their effect on glacial structure and cirque formation. In:
of the Royal Society London A228, 519-38. W. V. Lewis (ed.) Investigations of Norwegian Cirque Glaciers.
Glen, 1. W. and Lewis, W. V. (1961) Measurements of side-slip at Royal Geographical Society Research Series 4, 39-62.
Austerdalsbreen, 1959. Journal of Glaciology 3, 1109-22. Meier, M. F. and Tangborn, W. V. (1965) Net budget and flow of
Goldthwait, R. P. (ed.) (1971) Till: A Symposium. Ohio State South Cascade Glacier, Washington. Journal of Glaciology. 5,
University Press, Columbus. 547-66.
Goldthwait, R. P. (1979) Giant grooves made by concentrated Menzies, J. (1979) A review of the literature on the formation and
basal ice streams. Journal of Glaciology 23, 297-307. location of drumlins. Earth Science Reviews 14, 315-59.
Graf, W: L. (1970) The geomorphology of the glacial valley Menzies, 1. and Rose, 1. (eds) (1987) Drumlin Symposium: Pro-
cross-section. Arctic and Alpine Research 2, 303-12. ceedings of the Drumlin Symposium: First International Con-
Gurnell, A. M. and Clark, M. 1. (eds) (1987) Glacio-Fluvial Sediment ference on Geomorphology, Manchester 16-18 Sept. 1985.
Transfer: An Alpine Perspective. Wiley, Chichester and New Balkema, Rotterdam.
York. Morris, E. M. (1976) An experimental study of the motion of ice
Hambrey, M. J. (1976) Debris, bubble and crystal fabric charac- past obstacles by the process of regelation. Journal of Glacio-
teristics of foliated ice, Charles Rabots Bre, Okstindan, Norway. logy 17, 79-98.
Arctic and Alpine Research 8, 49-60. Nye, 1. F. (1952a) The mechanics of glacier flow. Journal of
Harbor, J. M., Hallet, B. and Raymond, C. F. (1988) A numerical Glaciology 2, 82-93.
model of landform development by glacial erosion. Nature 333, Nye, J. F. (1952b) A method of calculating the thickness of ice
347-9. sheets. Nature 169, 529-30.
Harrison, W. D. (1975) Temperature measurements in a temperate Nye, J. F. (1957) The distribution of stress and velocity in glaciers
glacier. Journal of Glaciology 14, 23-30. and ice sheets. Proceedings of the Royal Society London A239,
Haynes, V. M. (1968) The influence of glacial erosion and rock 113-33.
structure on corries in Scotland. Geografiska Annaler 50A, Nye, 1. F. (1960) The response of glaciers and ice sheets to sea-
221-34. sonal and climatic changes. Proceedings of the Royal Society
Hirano, M. and Aniya, M. (1988) A rational explanation of cross- London'A256,559-84.
profile morphology for glacial valleys and of glacial valley Nye, J. F. (1965a) The flow of a glacier in a channel of rectan-
development. Earth SUijace Processes and Landforms 13, gular, elliptic or parabolic cross-section. Journal of Glaciology
707-16. 5,661-90.
290 Exogenic processes and landforms

Nye, J. F. (1965b) The frequency response of glaciers. Journal of Sharp, M. (l988b) Surging glaciers: geomorphic effects. Progress
Glaciology 5,567-87. in Physical Geography 12, 533-59.
Nye, 1.' F. (1970) Glacier sliding without cavitation in a linear Shaw, J. and Sharpe, D. R. (1987) Drumlin formation by sub-
viscous approximation. Proceedings of the Royal Society glacial meltwater erosion. Canadian Journal of Earth Sciences
London. A315, 381-403. 24, 2316-22.
Nye, J. F. (1976) Water flow in glaciers: jokulhlaups, tunnels, and Shreve, R. L. (1972) Movement of water in glaciers. Journal of
veins. Journal of Glaciology 17, 181-207. Glaciology, 11,205-14.
Olyphant, G. A. (1981) Interaction among controls of cirque Shreve, R. L. (1984) Glacier sliding at subfreezing temperatures.
development: Sangre de Cristo Mountains. Colorado, U.S.A. Journal of Glaciology 30,341-47.
Journal of Glaciology 27, 449-58. Smalley, I. J. and Unwin, D. J. (1968) The formation and shape of
0strem, G. (1975) ERTS data in glaciology - an effort to monitor drumlins and their distribution and orientation in drumlim
glacier mass balance from satelli te imagery . Journal of Glacio- fields. Journal of Glaciology 7,377-90.
logy 15,403-14. Stenborg, T. 1969. Studies of the internal drainage of glaciers.
0strem, G., Bridge, C. W. and Rannie, W. F. 1967. Glacio- Geografiska Annaler 51A, 13-41.
hydrology, discharge and sediment transport in the Decade Sugden, D. E. (1974) Landscapes of glacial erosion in Greenland
glacier area, Baffin Island, N.W.T. Geografiska Annaler 49A, and their relationship to ice, topographic and bedrock con-
268-82. ditions. Institute of British Geographers Special Publication 7,
Parizek, R. R. (1969) Glacial ice-contact rings and ridges. Geo- 177-95.
logical Society of America Special Paper 123,49-102. Sugden, D. E. (1978) Glacial erosion by the Laurentide ice sheet.
Paterson, W. S. B. (1981) The Physics of Glaciers. (2nd edn) Journal of Glaciology 20,367-91.
Pergamon, Oxford. Sugden, D. E. and John, B. S. (1976) Glaciers and Landscape.
Post, A. S. and La Chapelle, E. R. (1971) Glacier Ice. University Edward Arnold, London.
of Washington Press, Seattle. Thorarinsson, S. (1953) Some new aspects of the Grfmsvotn
Prest, V. K. (1968) Nomenclature of moraines and ice-flow problem. Journal of Glaciology 2, 267-75.
features as applied to the glacial map of Canada. Geological Trainer, F. W. (1973) The formation o'f joints in bedrock by
Survey Papers Canada 67-57. moving glacial ice. United States Geological Survey Journal of
Price, R. J. (1 0 73) Glacial and Fluvioglacial Landforms. Oliver Research I, 229-36.
and Boyd, Edinburgh. Waitt, R. B. (1985) Case for periodic, colossal jokulhlaups from
Raymond, C. F. (1987) How do glaciers surge? A review. Journal Pleistocene glacial Lake Missoula. Geological Society of
ofGeophysical Research 92, 9121-34. American Bulletin 96, 1271-86.
Reeh, N. (1982) A plasticity theory approach to the steady-state Weertman, 1. (1961) Equilibrium profile of ice caps. Journal of
shape of a three-dimensional ice sheet. Journal of Glaciology Glaciology 3,953-64.
28,431-55. Weertman, 1. (1964) The theory of glacier sliding. Journal of
Robin, G. de Q. and Weertman, J. (1973) Cyclic surging of gla- Glaciology 5, 287-303.
ciers. Journal of Glaciology 12, 3-18. Weertman, J. (1972) General theory of water flow at the base of a
Schluchter, Ch. (ed.) (1979) Moraines and Varves: Origin. Genesis. glacier or ice sheet. Reviews of Geophysics and Space Physics
Classification. Balkema, Rotterdam. 10, 287-333.
Sharp, M. (1988a) Surging glaciers: behaviour and mechanisms. Weertman,1. (1979) The unsolved general glacier sliding problem.
Profvess in Physical Geography 12, 349-70. Journal of Glaciology 23,97-115.
This page intentionally left blank
 12
Periglacial processes and
landforms

12.1 The periglacial environment Here we will apply the term periglacial to those environ-
ments characterized by intense frost action and at least
In the high latitude regions of the northern and southern
seasonally snow-free ground. Such a qualitative definition
hemispheres, and in some areas at high elevation elsewhere,
makes it difficult to define precisely appropriate climatic
prevailing temperatures are so low that the ground remains
parameters in a way which satisfies all researchers. As a
frozen for much, or all, of the year. In such environments
general guide the upper limit for mean annual temperature
the effects of repeated freezing and thawing and the growth
is generally considered to lie between -1 and -3°C and the
of ice masses in the ground are so pervasive that they give
mean annual precipitation to be less than 1000 mm. Within
rise to a characteristic range of landfonns which merit special
this definition there are a great variety of climatic types,
consideration. This is the realm of periglacial processes and
ranging from the extremely cold and arid conditions exper-
landforms.
ienced in such areas as northern Greenland and parts of
Although their present extent is impressive; the Pleisto-
Antarctica, to the comparatively much warmer and humid
cene saw the extension of periglacial conditions well into mid-
maritime-influenced climates typified by Spitsbergen and
latitudes as the ice sheets of the northern hemisphere advanced
Iceland, and the alpine environments of high mountain
southwards. Thus large areas now experiencing compara-
ranges in all latitudes. Since an appreciation of this broad
tively mild climates retain evidence in the form of relict
climatic spectrum is important for an understanding of the
periglacial landforms of much colder conditions .as recently
global occurrence of the various periglacial processes and
as 15 ka BP. Many high latitude regions today have acquired
landfonns, it is helpful to subdivide the periglacial environ-
considerable economic and strategic significance and this
ment into a number of distinct categories (Table 12.1).
has served to focus attention on the special characteristics
of their landforms.
The term 'periglacial' was introduced in 1909 by the Table 12.1 Classification of periglacial climates
Polish scientist Walery von Lozinski to describe the land-
Polar lowlands Mean temperature of coldest month <-3°C.
forms and processes occurring around the margins of the Zone is characterized by ice' caps, bare rock
great Pleisto~ene ice sheets. Subsequently it was applied surfaces and tundra vegetation
more broadly to encompass those processes and landforms Subpolar lowlands Mean temperature of coldest month <-3 °C and
(regardless of age) associated with very cold climates in of warmest month> 10°C. Taiga type of
vegetation. The 10°C isotherm for warmest
areas not permanently covered with snow or ice (and in month roughly coincides with tree-line in
many cases located far from glaciers or ice sheets). Such northern hemisphere.
areas of extreme cold are commonly underlain by per- Mid-latitude lowlands Mean temperature of coldest month is <-3°C
manently frozen ground, or permafrost, and because many but mean temperature> 10°C for at least four
months per year.
landforms characteristic of these regions owe their exist-
Highlands Climate influenced by altitude as well as
ence to permafrost its presence is regarded by some as a latitude. Considerable variability over short
prerequisite for the action of periglacial processes. Never- distance depending on aspect. Diurnal
theless, since many other landforms characteristic of cold temperature ranges tend to be large

regions occur independently of pennafrost, most geomor- Source: Based on the classification presented by A. L. Wa hburn (1979)
phologists would not use the term in this restricted sense. Geocryology. Edward Arnold, London, pp. 7-8.
294 Exogenic processes and landforms

12.1.1 Characteristics of permafrost

ACTIVE LAYER
Although we have indicated that permanently (or more
OPEN
strictly perennially) frozen ground is not an element of all (SUPRA-PERMAFROST)
periglacial environments, it plays a key role in many TALIK

important periglacial processes. Permafrost was first defined

by S. W. Muller in 1947 as a thickness of soil, sediment or
bedrock at variable depth beneath the surface experiencing INTRA- PERMAFROST
TALIK
a temperature continuouly below freezing for a prolonged
period (Muller considered the minumum period to be 2 a).
According to this definition cold glaciers are a form of
SUB - PERMAFROST
permafrost, but because of their special characteristic of TALIK
movement they are usually considered separately. Various
other definitions have subsequently been offered, a number
of these emphasizing the importance of the presence of ice.
Although 'dry' permafrost can exist where there is insuffi-
cient moisture to permit ice formation, the role of ice is so Fig. 12.1 Nomenclature for features associated with permafrost.
crucial in the generation of landforms in periglacial en-
vironments that we will be almost exclusively concerned may be trapped between the permafrost table and the frozen
here with permafrost containing ground ice. The mean active layer. These water pockets, called taliks, develop in
surface temperature required to maintain, or develop, per- part because of the release of latent heat of fusion as water
mafrost may be less than O°C as pressure, salts in solution undergoes a change of state from a liquid to a solid. An
and the presence of clays can all affect the freezing point of additional factor is the build-up of cryostatic pressure; this
subsurface water. Consequently, most workers now cite a is pressure exerted as a result of the volume expansion
temperature a little below 0 °C in defining permafrost. associated with the freezing of water. Open, or supra-
Ground ice may be present in a variety of forms ranging permafrost, taliks may persist for months before finally
from a cement in soil pores to masses of almost pure ice in freezing, while intra-permafrost taliks, occurring within the
veins, lenses and wedges (Table 12.2). It is generally con- permafrost layer itself, may survive for much longer
fined to the upper part of a permafrost layer, the maximum periods.
depth recorded. being only about 45 m. In most areas of
permafrost, surface temperatures exceed 0 °C for at least a
12.1.2 Distribution of permafrost
brief period each year. Consequently there is usually a thin
active layer, up to 3 m in depth, overlying the permafrost The very considerable area covered by permafrost, even at
which freezes and thaws seasonally (Fig. 12.1). The bound- the present day, is evident from the data provided in Table
ary between the active layer and the permafrost is known as 12.3. If ground under ice sheets is included the total exceeds
the permafrost table. As a result of freezing penetrating one-quarter of the surface area of the continents. By far the
downwards at unequal rates in the winter, unfrozen water largest proportion occurs in the northern hemisphere, most-
ly in the USSR, Canada and Alaska (Fig. 12.2). In the very
Table 12.2 Classification of types of ground ice coldest regions permafrost is continuous, but as mean annual
temperatures approach 0 °C its occurrence becomes discon-
MAJOR TYPES SUBTYPES
tinuous and eventually sporadic. Its equatorward limit cor-
Soil ice Needle ice relates broadly with a mean annual air temperature of
Segregated ice -1°C. Climates associated with permafrost generally have
Ice-filling pore spaces
Vein ice Single veins Table 12.3 Global occurrence of permafrost
Ice wedges
Intrusive ice Pingo ice AREA COVERED (km 2 x 106)
Sheet ice
Extrusive ice (formed subaerially e.g. Continuous Discontinuous Total
floodplains)
Northern hemisphere 7.64 14.71 22.35
Sublimation, i.e. formed in cavities by Antarctica 13.21 13.21
crystallization from water vapour Mountains 2.59 2.59
Buried ice Buried icebergs
Buried glacial ice Totals 20.85 17.30 38.15

Source: After C. Embleton and C. A. M. King (1975) Periglacial Source: Modified from A. L. Washburn (1979) Geocryology. Edward
Geomorphology, Edward Arnold, London, Table 2.3, p. 34. Arnold, London, Table 3.1, p. 22, after S. R. Stearns (1966) US Army
Cold Regions Research and Engineering Laboratory, Cold Regions
Science and Engineering l-A2.
 Periglacial processes and landforms 295

Fig. 12.2 Distribution of continuous and discontinuous

permafrost in the northern hemisphere. Numbers indicate
localities in Table 12.4. (Map compiled from various sources.)

daily temperatures below 0 °C for at least nine months of

the year, and below -10 °C for at least six months of the
year. Temperatures rarely exceed 20°C at any time and
precipitation is typically low - less than 100 mm in winter
and less than 300 mm in summer. The large anticyclonic
continental polar air masses are largely responsible for
producing the required conditions of intense cold and
aridity. The distribution of permafrost in Eurasia clearly
illustrates the prominent southward depression of its equa-
torward limit in the continental interior of east-central
Siberia and its poleward deflection to very high latitudes in
maritime-influenced north-west Europe (Fig. 12.2). The
extent of permafrost under the Antarctic and Greenland ice
sheets is uncertain but there are strong theoretical grounds
for believing it to be present. The minimum elevation at
which permafrost is found tends to increase equatorwards
but small areas occur on the highest peaks even at very low
latitudes; limited occurrences, for instance, exist above an
elevation of 4170 m on Mauna Kea, Hawaii, at 19 ON. The
most extensive highland occurrences are in western China
• Continuous permafrost
on the Tibetan Plateau.
~ Discontinuous permafrost At the present day, depths of permafrost range up to at
least 1500 m (Table 12.4). Rates of permafrost aggradation

Table 12.4 Approximate depth to base of permafrost and prevailing temperature conditions at various localities in the
northern hemisphere.

LOCALITY DEPTH (m) MEAN ANNUAL AIR GROUND TEMPERATURE (0C)

TEMPERATURE (0C) AT DEPTH INDICATED (m)

USSR
1 Markha River, upper reaches 1450-1500
2 Udokan 900 -12.0 -3.3 500
3 Tiksi 630 -14.0 -3.3 500
4 Mirnyy 550 - 9.0 -1.8 300
5 Vilyuy River, mouth 420 -10.0 -1.0 240
6 Noril'sk 325 - 8.0 -7.5 50
7 Yakutsk 198-250 -10.13
8 Vorkuta 131
Canada
9 Winter Harbour 557 -16.0
10 Resolute 396 -16.2 -5.6 135
11 Mackenzie Delta 18-366 -9.1 to -11.3 -0.4 to -6.0 14
12 Port Radium 107 - 7.1
13 Inuvik >91 - 9.6 -3.3 8 to 30
14 Schefferville >76 - 4.5 -1.1 to -0.3 8 to 58
15 Churchill 30-61 - 7.2 -2.5 to-1.7 8 to 16
16 Dawson 61 - 4.7
Alaska
17 Prudhoe Bay 610
18 Barrow 204-405 -12.0 -6.3 179
19 Fairbanks 30-122 - 3.5 -7.0 25
20 Nome 37 - 3.5

Note: Locations are indicated by numbers on Figure 12.2.

Source: Data from A. L. Washburn, 1979, Geocryology. Edward Arnold, London, Ta\:>le 3.5, pp. 37-40.
296 Exogenic processes and landforms

can only be of the order of a few centimetres a year so thick 12.2 Periglacial processes
permafrost layers must have taken a minimum of several
Comparatively few geomorphic processes are solely con-
thousands of years to develop. In some areas present-day
fined to periglacial environments; rather, a number of pro-
climates are sufficiently cold for the development as well as cesses occurring in other morphoclimatic zones are
the maintenance of permafrost, but elsewhere permafrost
especially active in periglacial regimes and give rise to
depths seem to exceed those expected on the basis of cur-
some unique landforms. Most of the significant processes
rent temperatures. This suggests that permafrost depths may arise, directly or indirectly, from the freezing and thawing
be influenced by past as well as present climatic conditions.
of water. It is important, however, to emphasize that even
Other factors, including the presence and type of soil and though we will be focusing on these 'periglacial processes',
vegetation, the existence of water bodies, snow or ice on
the landscapes of periglacial regions as a whole are still
the surface and exposure to the Sun and wind can influence
dominated by the effects of 'normal' fluvial activity.
the depth of permafrost in a particular locality. These, and
other factors, can change over time to affect the intimate
12.2.1 Frost action
relationship between air and ground temperatures and there-
by lead to the degradation or aggradation of a permafrost Frost action encompasses those processes associated with
layer (Fig. 12.3). the freezing of water in rock, soil or other material, includ-

PERMAFROST AGGRADATION

Type of change ----""'4..

~ Processes
_ _ _ _ _ _ _ _..
~ Results
Responses Ground-Ice change

Active layer burial

from sedimentation
soil creep, slumping, Active layer
mudnows, etc. thlckenlDg lDduces
rise of permafrost
table bytloth down-
freezlnlland up-
freezing; ....turn to
Accumulation of thermal equilibrium
organkmatter

Ice at top of
permafrost Is trapped
active-layer ice
v <getatlon change
leading to ground
cooling

Cooling trend,
long or short term

PERMAFROST DEGRADATION

Type of change Processes Results ~

Responses - - - - - -.. Ground-Ice change

Active layer slope

erosion, slumping,
soil creep, mudnow,
blow-outs, etc.

Ice formerly at depth

DOW at top 01
Disturbances by permafrost
fire, construction,
vegetation, etc.
Changes leading to
warming of active
layer

Climatic warming
long or short term

Fig.12.3 Factors influencing the growth (aggradation) and decay (degradation) ofpemiafrost. (After A. L. Washburn (1979)
Geocryology. Edward Arnold, London, Fig. 5.3.31 and 3.32, p. 56, after 1. R. Mackay (1971) in: R. J. E. Brown (ed.) Proceedings of a
Seminar on the Permafrost Active Layer, 4 and 5 May 1971. Canada National Research Council Technical Memo 103 Fig. 1, p. 29
and Fig. 2, p. 30.)
 Periglacial processes and landforms 297

ing frost shattering, heaving and thrusting. Frost cracking a 2 a period at a depth of 50 mm in spite of 441 freeze-
(see Section 12.2.1.4) is also usually included although it is -thaw cycles at the surface; in many areas there may be
caused by the thermal contraction of materials at low only one annual cycle below a depth of a few centimetres.
temperatures rather than the formation of ice. Although Such findings have important implications for the efficacy
frost action occurs outside the periglacial zone, its signifi- of frost action in rock breakdown.
cance within it is the crucial element that differentiates
periglacial landscapes from those developed under other 12.2 .1.1 Frost weathering
morphoclimatic regimes. We have already outlined the mechanisms thought to be
Many periglacial landforms are associated with the freez- involved in frost weathering in Section 6.3.2.1, but it is
ing of water in soils and other unconsolidated deposits. At necessary here to reassess the role of this process under
first sight this appears a simple physical mechanism, but in present-day periglacial regimes. The important role tradi-
reality the large number of variables at play under natural tionally attached to frost shattering in producing the angular
conditions make it a complex process. Of particular import- rock debris so characteristic of periglacial environments
ance in terms of its geomorphic effects is whether water rests heavily on the assumed presence of suitable
freezes in situ in voids, or migrates to form discrete masses conditions of freeze-thaw and moisture availability, yet
of segregated ice. The key factors are the rate of freezing, field measurements of these crucial parameters are few in
the grain size (and hence pore size) of the material and the number. Moreover, some recent studies sugge t that suitable
potential for capillary suction (see Section 6.1.2) which conditions may be less common than previously thought.
expresses the ability of water to move towards the point in A detailed field investigation completed in the Colorado
the soil where it is freezing (often referred to as the . Rockies in the western USA, for instance, involved moni-
freezing front). toring temperatures beneath and adjacent to seasonal snow
Coarse deposits, such as gravel, are highly permeable but patches over 2 a. This is just the kind of alpine environment
have a low suction potential, whereas very fine-grained (abundant moisture from snow-melt coupled with frequent
materials with extremely small pores have a low per- air temperature fluctuations across the freezing point) in
meability but high suction potential. Consequently it is which we would expect frost shattering to be intense. None
sediments of intermediate grain size, such as silt, that are the less, this study demonstrates that even here the required
most susceptible to the formation of segregated ice since combination of freezing intensity and moisture availability
both suction potential and permeability are moderately high is absent. Because of the insulating effect of snow, bedrock
in such materials. Grain (pore) size also influences the temperatures at a depth of 10 mm rarely fell below -5°C,
temperature at which ice will form. As a result of the com- the apparently critical temperature for effective stresses to
plex interaction between water' molecules and the surfaces be generated by the expansion arising from ice formation.
of particles the proportion of water remaining unfrozen in On snow-free cliff faces temperatures dropped well below
sediment at a given temperature varies inversely with grain -5 °C but there was a lack of available moisture.
size since the surface area on to which absorbed water is Such studies question our ideas about conditions assumed
bound by a variety of physico-chemical effects is greater to occur in nature and in this case focus attention on other
for small particles. In clays only half of the total water mechanisms of rock breakdown, such as hydration shatter-
content may be frozen at a temperature of -2°C, whereas ing and, at least in arid and coastal locations, salt weather-
in sands virtually all the water freezes at 0 0c. In any ing. In the McMurdo Sound area of Antarctica, for instance,
material ice forms initially in the larger pores and only as it has been suggested that salts (mainly sodium chloride
temperature falls is it able to crystallize in smaller voids. transported from the sea or from surface accumulations) are
Crucial to an understanding of the role of frost action incorporated in snow. On thawing, salt solutions enter rock
processes are measurements of the frequency of freeze- pores and the salt crystallizes out. A variety of surface
thaw cycles experienced by surface materials or bedrock. weathering features occurring in periglacial environments,
The frequency of air temperature fluctuations across the including tafoni and honeycomb forms (see Section 6.4.2),
freezing point can be readily estimated from conventional have been attributed to this process. It may also play an
meterorological data. But unfortunately such data give little important role in rock breakdown. Alternatively, much of
indication of the effectiveness of frost action since many the angular rock debris found in some lower latitude peri-
other factors, such as the thermal properties of bedrock and glacial environments today may be a relic of Pleistocene
soil and the' insulating effects of vegetation and snow, must conditions more favourable to frost shattering.
also be considered. Discrepancies between the number of
freeze-thaw cycles in the air and those occurring in the 12.2.1.2 Frost heaving and thrusting
ground are especially marked in climates characterized by Frost heaving and frost thrusting refer respectively to the
many air cycles of small amplitude. In a study of Kerguelen vertical and horizontal movement of material due to the
Island in the sub-Antarctic· zone no frost was recorded over formation of ice. In nature a combination of vertical and
298 Exogenic processes and landforms

frost heaving is an important mechanism in the differential

Capillary suction increasing vertical movement of soil constituents of different sizes.
Permeability decreasing Numerous field observations confirm the upward move-
ment of stones in soils in periglacial areas, and various
mechanisms have been suggested to explain this phenom-
enon. They may be categorized for convenience into frost-
pull and frost-push hypotheses. The frost-pull hypothesis
proposes that stones together with the enclosing finer-
grained matrix are raised as the ground expands on freez-
...o ing. On thawing the fine material collapses while the bases
of the stones are still supported by ice. Subsequently, as
...
Q)

~ this ice melts, fine particles partly fill the space and support
c:
Q)
the overlying stone. The frost-push hypothesis 'is most
>
~ applicable to stones close to the surface. Soil water flowing
Q)
::I: around the stone will tend to collect beneath.it and on
Coarse Medium Fine freezing the stone is pushed upwards. On melting finer
Particle or Pore Size particles collapse into the void and support the stone. There
is experimental support for this process which is probably
Fig. 12.4 Relationship between heaving rate and pore size. most effective when freezing is rapid; nevertheless the
frost-pull mechanism is probably the more important under
natural conditions.

horizontal components is more common, although heaving 12.2.1.3 Mass displacement

probably predominates. This is because the pressure gene- Mass displacement is the local transfer of material within
rated by ice crystallization, which is at least as crucial as the soil as a result of frost action. Vertical movements are
the volume expansion due to freezing, acts predominantly most marked but horizontal movements also occur. A
parallel to the direction of the maximum temperature gradi- major cause of mass displacement is probably the cryostatic
ent, and this is approximately normal to the ground surface. pressures developed in pockets of unfrozen soil or other
For the same reason ice lenses, which lie parallel to the sediments trapped between the permafrost table and the
surface, are a common form of segregated ice, although freezing front migrating downward from the surface. Local-
their precise form is also influenced by the structure and ized pockets of unfrozen soil will be produced because the
stratification of soils and other deposits. freezing front will advance downward at different rates
Since heaving and thrusting are caused by the formation depending on the grain size and moisture content of the
of segregated ice, the factors which control the migration of soil. Some recent research, however, has questioned the
water to the freezing front also determine the susceptibility role of a cryostatic pressure in producing mass displacement
of surface materials to such processes (Fig. 12.4). As we since differential heave resulting from the annual freezing
have already noted, silts are most prone to the growth of and thawing of the ground may produce similar effects.
large ice masses. Variations in the particle size of surface Positive pore-water pressures towards the base of slopes
deposits can lead to localized ice formation and the devel- might also generate mass displacement through the injec-
opment of differential heaving as well as lateral pressures. tion of water-saturated soil upward through a thin frozen
The nature of the surface cover can also influence the pro- surface layer. This mechanism has been widely cited as a
cess as heaving has been observed to be most effective cause of significant displacements within sediments in the
where moisture is abundant and where there is little vege- active layer giving rise to structures consisting of interpene-
tation to insulate the ground. Localized upward movement trating layers of material which were originally flat-lying.
of surface stones can occur through the formation of needle Such forms are called periglacial involutions, the qualify-
ice which is composed of vertically elongated crystals up to ing adjective being necessary since non-periglacial processes,
30 mm in length. such as slumping and other types of mass movement, can
Frost heaving in the active layer appears to be attribut- produce similar features.
able to three main effects: the growth of ice lenses as
downward freezing progresses; the development of ice 12.2.1.4 Frost cracking
lenses near the base of the active layer by upward freezing Frost cracking is the fracturing of the ground by thermal
from the permafrost layer, and the progressive freezing of contraction at sub-freezing temperatures. It is probably the
unfrozen pore water as the active layer cools below 0 °e. In major cause of the polygonal fracture patterns so common
addition to displacing masses of soil and other materials, in periglacial environments, although similar crack polygons
 Periglacial processes and landforms 299

can be formed by desiccation cracking and by dilation crack- Rates of movement for gelifluction alone are strongly
ing resulting from differential heaving. Frost crack polygons dependent on moisture, and its importance in periglacial
are commonly 5-30 m across, crack spacing typically being environments can be ascribed to two primary factors:
two to three times crack depth. Although they are found saturation of the soil due to restricted drainage promoted by
outside the permafrost zone, they are best developed in a permafrost layer or seasonally frozen water table, and
areas of frozen ground rich in ice. The absence of a thick moisture provided by thawing snow and ice. The overriding
insulating snow cover appears to be a prerequisite for their importance of soil moisture compared with the influence of
development as crack frequency has been observed to vegetation and slope angle in controlling rates of gelifluc-
diminish markedly with an increase in snow depth. Rate of tion has been demonstrated by the observations of A. L.
temperature decrease appears to be more crucial in pro- Washburn at an experimental site at Mesters Vig in north-
moting cracking' than the actual minimum temperature east Greenland. Here the less steep but better vegetated
attained. Once cracking has been initiated ice films begin to slopes which tended to retain moisture experienced faster
fill the fracture and eventually ice wedges may form (see rates of movement than steeper but less well vegetated and
Section 12.3.2.1). drier slopes.
Free drainage in gravels and sands on the one hand, and
12.2.2 Chemical weathering the cohesive properties of clays on the other, means that
soils dominated by silt-sized particles are most susceptible
The absence of significant accumulations of chemically
to gelifluction. Under ideal conditions of soil type and mois-
weathered rock in most periglacial environments, together
ture content gelifluction can occur over gradients as low as
with the abundance of largely unweathered angular rock
1°. Maximum rates apparently occur at slope angles of
fragments and bare rock surfaces, even in areas of low
around 10-20° since efficient drainage on steeper slopes
relief, supports the long-held view that mechanical
limits the moisture content of soils. When estimated sep-
weathering predominates in such regions. We would expect
arately, rates of movement associated with gelifluction and
rates of chemical weathering to be low because of low
frost creep appear to be of the same order of magnitude.
temperatures, the retention of water as ice for a large
Maximum velocities are generally below 50-100 mm a-I,
proportion of the year and low levels of biotic activity.
but may exceptionally exceed 200 mm a-I. Rates decline
Nevertheless, there is a lack of detailed investigations
rapidly with depth and movement seems to be confined to
directly comparing rates of chemical and mechanical
within 1 m of the surface.
weathering in the periglacial zone and further research is
clearly required. One pioneer study carried out by A. Rapp
over a 9 a period in northern Sweden in a periglacial 12.2.4 Nivation
environment (but not within the permafrost zone) suggested
Nivation is localized denudation by a combination of frost
that material released by chemical weathering and removed
action, gelifluction, frost creep and meltwater flow in asso-
in solution by streams in fact accounted for about a half of
ciation with snow patches. It is most active in subarctic and
the total movement of material by all denudational processes
alpine environments and its major geomorphic effect is the
(see Section 7.5). Moreover, the absence of thick weathering
development of nivation hollows by the incision of snow
mantles in most periglacial environments may be partly
patches into hillsides. Once initiated, normally in some
accounted for by their removal from some areas by glacial
existing slight depression, a nivation hollow increases in
erosion during the Pleistocene.
size as it becomes a collecting site for snow in subsequent
years.
12.2.3 Mass movement
Snow provides both meltwater and acts as a ground
Most of the different types of mass movement are active in insulation. The erosional potential of nivation is controlled
periglacial environments to some extent, but two mechanisms primarily by snow thickness and the presence or absence of
- frost creep and solifluction - are of particular significance underlying permafrost. Where permafrost is absent an
(see Section 7.2.2). In periglacial environments solifluction extensive area of freeze-thaw can occur both beneath the
frequently occurs in association with permafrost or snow and around its margins. Where permafrost is present,
seasonally frozen ground, and under these circumstances it thawing, and hence frost action, gelifluction and frost
is more specifically described as gelifluction. In reality it creep, are confined to the periphery of the snow patch. As
often operates in conjunction with frost creep to the extent the snow patch gradually thaws the zone of maximum
that it is frequently difficult to distinguish between the action nivation will follow its contracting margin. A thick snow
of the two mechanisms. Downslope movements must rea- cover insulates the surface and reduces nivation by pre-
listically be assessed in terms of their combined effect and venting frequent freeze-thaw cycles. Data on rates of
they are often considered together in assessing rates of denudation by nivation are lacking, but it appears that fea-
movement (Fig. 12.5). tures extending 500 m along slopes with a width of 1 m
300 Exogenic processes and landforms

I,

Path of target during

frost heaving of ground

Path of target assuming m.aximum

possible frost creep (heaving and
vertical settling)

P 2->P3 Path of target during gelinuction

P3->P4 Path of target during settling of ground

Fig.12.5 Operation offrost creep and gelifluction illustrated by the movements of targets inserted in regolith on a slope. PFC
indicates the horizontal component of maximum possible frost creep, G is the horizontal component of downslope movement due to
gelifiuction and R is the apparent retrograde movement due to the tendency for the thawed layer to settle back against the slope rather
than purely vertically. (After A. L. Washburn (1979) Geocryology. Edward Arnold, London, Fig. 6.8, p. 201.)

may exceptionally be produced in one winter season in water region and meltwater can be temporarily dammed
areas of unconsolidated rock. Exceptionally large nivation behind vast ice accumulations further north before being
hollows recorded in Quebec, Canada, may have been released and causing widespread flooding.
excavated at a rate in excess of 1500 m3 a-I. Since a highly peaked river regime provides far more
erosional and transporting potential than a more evenly
distributed regime of equivalent total discharge, fluvial
12.2.5 Fluvial activity
action in periglacial environments is far greater than the low
Fluvial action has frequently been regarded as comparative- annual precipitation and modest annual discharges would
ly unimportant in periglacial environments, it being argued suggest. The limited field data available indicate that,
that its effectiveness is limited by the freezing of surface during the brief period of high snow-melt discharge, quite
water for a large proportion of the year and by the low small rivers can transport coarse debris and a high total
annual precipitation. Although the hydrological characteris- sediment load. The River Mechan in arctic Canada, which
tics of stream flow are certainly modified by the long lies well within the periglacial zone, provides a clear
period of winter freezing, the dominance of fluvial erosion illustration. Mean annual precipitation over the basin amounts
is reflected in the largely 'normal' morphometry of peri- to only around 135 mm, half of which falls as snow. Some
glacial environments. 80-90 per cent of the total annual flow is concentrated into
River regimes are highly seasonal with very large dis- a lO-day period during which peak flow velocities reach up
charges being sustained for short periods during the spring to 4 m S-I and the whole river bed may be in motion.
melt (Fig. 12.6). Major lowland rivers may continue to flow Although rates of downcutting may be rapid (a rate of
throughout the winter under an ice cover, but at greatly 1-3 m ka- I has been estimated for a river in Spitsbergen)
reduced rates of flow. In the northward flowing rivers of most rivers in the periglacial zone, other than mountain
arctic Canada and Siberia the thaw occurs first in the head- streams, flow on aggraded beds. This may be a result of
 Periglacial processes and landforms 301

6,000 the southern margin of the Pleistocene ice sheets. Loess is

4,000
Colville River, 1962 composed of angular to sub-angular particles and cohesion
between grains allows the development of steep, high cliffs
2,000 where thick blankets have been dissected by fluvial action.
Although loess can form in warm deserts, most is believed
May to have originated in past or present periglaci~l environ-
1·5 ments. Glacial grinding, frost cracking and salt weathering
Jasons Creek, Devon Island,1970 appear capable of producing sufficiently fine particles which
lO
'I can be picked up from sites such as glacial outwash plains
CJ) 0·5
(")
and carried considerable distances by the wind before being
E
deposited as extensive loess blankets, exceptionally up to
Ql
2co 100 m thick and covering tens of thousands of square
.r:.
u
CJ)
150 Lewis River, kilometres (see Section 10.3.7).
(5 100 Baffin Island, 1963 During the glacial phases of the Pleistocene, the steep
50 pressure gradient between the high pressure cells over the
Laurentide and Fenno-Scandian ice sheets and low pressure
to the south would have produced high mean wind velo-
cities and conditions conducive to significant aeolian trans-
Ogotoruk Creek
12 Alaska, 1960 port and deposition. Dating of discrete vertically stacked
layers of loess interbedded with soils indicative of more
8
temperate and humid conditions in China and Europe has
4
indeed indicated a close correlation between loess deposition
Aug. Oct. and phases of ice sheet extension (see Section 14.3.2).
Fig. 12.6 Runoff regimes of selected rivers in the periglacial
zone of North America. (Modified from H. M. French (1976) The 12.3. Periglacial landforms
Periglacial Environment. Longman, London, Fig. 8.1, p. 169,
hased on M. Church (1974) Permafrost Hydrology; Proceedings
of Workshop Seminar. Environment Canada, Ottawa 7-20.) 12.3.1 Patterned ground
First described more than a century ago, patterned ground
current rates of weathering and mass movement supplying is one of the most conspicuous features of periglacial environ-
more debris than can be transported by existing rivers. Some ments. It can vary widely in scale from patterns composed
of this material may, however, be a relic of higher rates of of elements only a few centimetres across to those with
rock breakdown and sediment supply during the Pleistocene. dimensions of 100 m or more. Although some forms of pat-
terned ground are found in other environments, especially
deserts, the extent of its development is much greater in the
12.2.6 Aeolian activity
periglacial zone than elsewhere since it is most charac-
Present-day periglacial areas characterized by extreme ari- teristic of surfaces subject to intense frost action.
dity are favourable environments for aeolian activity, and The most widely used descriptive classification of pat-
there is abundant evidence for much more widespread and terned ground has been developed by A. L. Washburn. This
intense wind action in areas marginal to the great northern emphasizes both shape and the degree of sorting of the
hemisphere ice sheets during the Pleistocene. In addition to constituent materials. Five basic patterns are recognized -
strong winds and the freeze-drying of ediments, low preci- circles, nets, polygons, steps and stripes - although one
pitation and low temperatures are associated with minimal form may grade into another. Each of these types is sub-
vegetation cover. Aeqlian erosion is evident in faceted and divided on the basis of whether the pattern results from the
grooved bedrock surfaces, deflation forms in unconsoli- sorting of surface materials into fractions of large and small
dated fluvial sediments, and in the formation of ventifacts size or whether it has developed without sorting. Slope
(see Section 10.2.2.1). Wind is also important in the move- angle is an additional factor. Circles, nets and polygons are
ment and deposition of snow which, together with topo- most common on flat surfaces, whereas steps and stripes
graphic features, determines areas of snow accumulation tend to form on gradients of between 5 and 30°, with transi-
and removal and thereby the distribution of an insulating tional forms developing on gentler slopes of between 2 and
snow cover. 5°. Mass movement is generally too active on slopes steeper
Of great geomorphic significance are various aeolian than 30° to allow patterns to form. A summary of the charac-
deposits, especially the loess which blankets much of those teristics and environmental associations of the various types
areas in North America and Eurasia which were located on of patterned ground is presented in Table 12.5.
302 Exogenic processes and landforms

Table 12.5 Characteristic of different forms of patterned ground

Circles Occur singly or in groups. Typical dimensions 0.5-3 m.

Nonsorted type characteristically rimmed by vegetation,
sorted type bordered by stones which tend to increa e in
size with size of circle. Found in both polar and alpine
environment but are not restricted to an~a of permafrost. SORTING
Unsorted circles also recorded from non-periglacial PROCESS
environments
Polygons Occur in groups. Nonsorted polygons range from small
features «I m across) to much larger forms up to 100 m
or more in diameter. Sorted polygons attain maximum
dimensions of only 10 m. Stones delimit polygon border
and surround finer material. Nonsorted·forms are
delineated by furrows or cracks. Some types of polygon
occur in hot desert environments but most are best
developed in areas subject to frost. Ice-wedge polygons
only form in the presence of permafrost Nonsorted
Nets A transitional form between circle and polygons. circles, nets and polygons
Usually fairly small «2 macro ). Earth hummocks
compri ing a core of mineral soil surmounted by
vegetation are a common form of unsorted net. PATIERNING PROCESS
Steps Found on relatively steep lopes. They develop either
parallel to slope contours or become elongated down lope
into lobate forms. In unsorted forms the rise of the step is
well vegetated and the tread is bare. In the unsorted type
Fig.12.7 Relationship between patterning, sorting and slope
the step i bordered by larger stones. Lobate forms are processes in the formation of patterned ground. (After A. L.
known as stone garlands. Neither type confined to Washburn (1979) Geocryology. Edward Arnold, London, Fig.
permafrost environments 5.41, p. 160.)
Stripes Tend to form on steeper slopes than step . Sorted stripes
composed of alternating stripes of coarse and fine
material elongated· downslope. Nonsorted variety
displacement contribute to the development of non-sorted
delineated by vegetation in slight troughs. Not confined to
periglacial environments circles while surface wash is probably a factor in the forma-
tion of sorted stripes.
Source: Based on discussion in A. L. Washburn (1979) Geocryology.
Edward Arnold, London, pp. 122-56. The regularity of many forms of patterned ground has
suggested to some researchers that convection occurring
when frozen ground thaws may be an important mechanism
Numerous factors contribute to the form of the great in its development. The idea is that water-saturated soil
variety of patterned ground including the interactions be- close to the surface warms up while the temperature of that
tween sorting, patterning and slope processes (Fig. 12.7). adjacent to the still frozen subsoil remains close to freezing
An important distinction is that between those forms of point. Convection in this situation is possible because water
patterned ground in which initial cracking of the surface is is densest at 4°C, so water at the thawing front is less dense
essential and those in which it is not (Table 12.6). Causes than the slightly warmer water close to the surface. When
of cracking include contraction on drying (desiccation crack- the active layer thaws the cooler water at the thawing front
ing), stretching of the surface as a result of heaving (dila- rises while descending plumes of warmer water lead to
tion cracking), and thermal contraction of seasonally or localized melting of the underlying ice. This produces a
perennially frozen ground (frost cracking). Only the latter regular undulating interface between frozen and unfrozen
is restricted solely to periglacial environments. Desiccation soil which, it is thought, can be reflected in the surface
and dilation cracking are important in the formation of small topography. The exact mechanism by which this occurs is
non-sorted polygons, whereas frost cracking is instrumental not clear, but it probably involves a range of processes
in the development of larger forms. The most important pat- including frost heave. Striped forms of patterned ground
terning processes which do not depend on initial cracking according to this model are explained by the downslope
include differential frost heaving and mass displacement. elongation of convection cells within the soil.
Frost heaving is apparently a primary cause of a number
of sorted forms of patterned ground; it helps to segregate
12.3.2 Ground-ice phenomena
larger stones, which tend to move upward and outward,
from the finer-grained matrix which forms a central core. 12.3.2.1 Ice wedges
Continued development at a number of discrete sites even- Ice wedges are downward-tapering bodies of ice (Fig.
tually leads to .coalescence and the formation of polygonal 12.8). They can be over 1 m across at the top and reach a
Eims formed primarily of stones. Sedimentation and mass depth of 10m or more. They have a polygonal form in plan,
 Periglacial processes and landforms 303

Table 12.6 Processes contributing to various types of patterned ground

PAlTER NED GROUND MORPHOLOGY

Circles Polygons Nets Steps Stripes

FORMATIVE PROCESSES N S N S N S N S N S

Cracking essential
Desiccation cracking * * *
Dilation cracking * *7
Salt cracking *
Seasonal frost cracking * * *7 *7
Permafrost cracking * * *7 *
Frost action along joints * * *7
Cracking not essential
Primary frost sorting * *7 *7 *7
Mass displacement * *7 *7 *
Differential frost heaving * *7 *7 *7 *
Salt heaving * *7 *7 *7 *
Differential thawing *7
and eluviation *7 *7
Differential mass movement * *7
Rillwork *7

Nand S indicate, respectively, nonsorted and sorted types of patterned ground.

Source: Modified from A. L. Washburn (1979) Geocryology. Edward Arnold, London, Table 5.1, p. 158.

Recent field studies have revealed that some ice-wedge

crack originate near the top of the permafrost and then
propagate both upwards to the surface and downwards into
the ice wedge. Only a proportion of wedges tend to crack in
anyone year, and the frequency of cracking seems to vary
inversely with the depth of snow cover. The average
amount of ice accreted each year varies enormously so the
total volume of ice present is not necessarily a good indi-
cator of how long an ice wedge has been growing. Obser-
vations made of the development of ice wedges after the
artificial draining of a lake in northern Canada have demon-
strated that under ideal conditions they can grow rapidly
and reach a significant size after just a few years.
Ice-wedge casts form when an ice wedge is replaced by
Fig. 12.8 Ice wedge in the right bank of the Aldan River, 170 km sediments as a result of the long-term thawing of the per-
upstream of its junction with the Lena River, Yakutia, USSR. mafrost. The form of the ice wedge is preserved by sedi-
(Photo courtesy A. L. Washburn.) ment and may survive long after the permafrost in which
the original ice wedge developed has completely thawed.
Ice-wedge casts are consequently an important indicator of
and ice-wedge polygons are associated with the most the previous existence of permafrost. Their palaeoenviron-
impressive form of patterned ground in periglacial environ- mental significance is also enhanced by the fact that ice
ments. Frost cracking appears to be the key mechanism wedges only seem to form in present-day permafrost en-
necessary for the development of ice wedges. Extreme cool- vironments where the mean annual temperature is below -6
ing of ice-rich permafrost in winter leads to thermal con- to -8°C.
traction and cracking of the surface (Fig. 12.9). As the
wedge grows it compresses the surrounding sediments 12.3.2.2 Pingos
which tend to bulge up around the perimeter of each polygon Pingos, named after the Eskimo word for a hill, are large
forming a low ridge. In other cases thawing and erosion of perennial ice-cored mounds. They ·are roughly circular to
the ice wedge produces a perimeter trough. In dry perma- elliptical in plan and range in size from 3 to 70 m high and
frost, where meltwater is scarce, the cracks may be filled from 30 to 600 m in diameter (Fig. 12.10). Large dilation
with loess or other coarser sediments to form sand wedges. cracks generated by the progressive growth of the ice core
304 Exogenic processes and landforms

~~Ijliii-- -----------------,
~F~ro~z~e~n~ Active layer T~a~~d_~~~.:.;~~~~

---r---
Permafrost

~ 1_~
A 1st Winter
B 1st Autumn

Fig.n.lO Pingo, Wollaston Peninsula, Northwest Territories,

Canada. (Photo courtesy A. L. Washburn.)

a lake in a permafrost zone is infilled by sediment and

vegetation, insulation of the underlying ground will be
reduced (Fig. 12.11) and freezing will encroach from the
base, side. and top. This will trap a body of water which on
freezing expands and domes the overlying sediments. Alter-
c 500th Winter
D 500th Autumn natively, diversion of a river or draining of a lake may have
a similar effect by reducing ground insulation. Support for
Fig. n.9 Hypothesized development of an ice wedge initiated this model of cryostatic pressure comes from the distribu-
by frost cracking due to thermal contraction. The initial crack, tion of pingos in the Mackenzie Delta area of the Canadian
which may be only afew millimetres wide, collects snow and
meltwater from thawing of the active layer (A). This water freezes Arctic where 98 per cent of the 1"380 pingos recorded are
below the permafrost table to form a thin ice vein (B). The initial located in, or closely adjacent to, lake basins.
fracture represents a zone of weakness so will be a favoured site The open system pingo model involves the freezing of
for frost cracking in subsequent years. However, cracking does ground water flowing under hydrostatic pressure as it forces
not necessarily occur each year and expansion during summer its way towards the surface from below a thin permafrost
heating will tend to close cracks developed during the preceding
winter. Further ingress of snow and meltwater adds new ice until layer (Fig. 12.12). It has been calculated that very high
an ice wedge is eventually formed (C and D). (After A. H. pressures are required to dome pingos and that these exceed
Lachenbruch (1962) Geological Society of America Special Paper those commonly attained within unconfined ground water.
70, Fig. 1, p. 5.) Consequently the open system type may in fact form under
temporary closed-system conditions as open taliks are
commonly mark the summit. These may widen sufficiently frozen in winter. Open system pingos extend into the zone
to expose the ice core and initiate thawing, thereby giving of discontinuous permafrost where mean annual tempera-
rise to a collapsed pingo. Various materials can envelop ture may be only a little below freezing since here water
the ice core, including clays, silts, gravels and even bed- circulation is freer. They tend to occur in cluster in favour-
rock. Drilling has shown that in some cases the core of able areas of drainage where the hydrostatic pressure is
clear ice extends to depths greater that the pingo height. more constant. In contrast, the closed system type usually
Vertical growth rates of pingos may reach 0.2 m a-I but occur as isolated features on flat surfaces related to lakes
they tend to decline rapidly as the pingo grows. Large and are confined to the zone of continuous permafrost where
pingos certainly take thousands of years to develop, as indi- minimum mean annual ground temperatures are around
cated by the radiocarbon dating of two pingos in the Canadian -5°C.
Arctic which gave approximate age~ of 4500 and 7000 a.
The ice core forming a pingo is thought to grow through 12.3.2.3 Palsas
the freezing of water forced upwards at pressure. Two Palsas are mounds or more elongated forms occurring in
mechanisms have been proposed to explain the generation bogs and containing perennial ice lenses. They differ from
of this pressure. In closed system pingos cryostatic pressure is pingos in containing peat as a major constituent and a core
involved. This may be brought about in two related ways. If composed of discrete superimposed ice lenses rather than a
 Periglacial processes and landforms 305

Aggrad ing permafrost Pond

-
-----_..::..:=~,~=.:..:-_~~---------'~---
it;

Pingo

Fig. 12.11 Schematic representation of the development of a closed system pingo following the infilling of a lake. (Modified from A. L.
Washburn, (1979) Geocryology. Edward Arnold, London, Fig. 5.49, p. 183.)

terrain. In addition to the collapse of material into the space

previously occupied by ice, thermokarst features may be
significantly affected by flowing water released by thawing.
High, short-term rates of erosion can occur, especially
where ice masses are exposed along cliffs or stream banks.
Although some thermokarst features can be attributed to
climatic change, many others form as a result of the
'normal' variability of the periglacial environment. Ther-
mokarst most commonly originates from a modification of
surface conditions and this can be brought about in a num-
ber of possible ways including disturbance of vegetation,
cliffretreai and changes in stream courses.
Thaw lakes are perhaps the most ubiquitous thermokarst

Fig. 12.12 Schematic representation of the development ofan

open system pingo. (Modified from F. Muller (1968) in: R. W.
Fairbridge (ed.) Encyclopedia of Geomorphology. Reinhold, New
York, Fig. 2, p. 846.)

single ice mass. The mode of ice formation is uncertain but

probably originates from differential frost heaving with
peat playing a crucial insulating role. Palsas are commonly
found in areas of discontinuous permafrost.

12.3.2.4 Thermokarst
Thermokarst is a term used to encompass a variety of
topographic depressions formed by the thawing of ground Fig.12.13 Small thaw lakes, Bathurst 1sland, Northwest
ice. It is so called because of the superficial resemblance of Territories, Canada. (Photo courtesy Geological Survey of
the landforms produced to those characteristic of true karst Canada.)
306 Exogenic processes and landforms

form, and arguably one of the most characteristic landforms specific direction. Their ongm is uncertain, but they are
of periglacial environments (Fig. 12.13). They are bodies of probably related to winds prevailing from a particular
water which fill small, shallow depressions. The majority direction which preferentially deposit sediments along lake
are less than 5 m deep, and they are rarely more than 2 km ,banks normal to the wind direction. These deposits insulate
across. The origin of thaw lakes seems to be related to the the banks from further thawing and protect them from wave
melting of permafrost which contains a volume of ice that erosion.
exceeds the normal pore space of the sediment. Subsidence Thaw slumps are arcuate embayments faciIV5 downslope
occurs and shallow depressions are formed which fill with and formed by the exposure and thawing of ground ice.
water. They are particularly common in poorly drained Basal undercutting of river banks or mass movement are
lowland regions where ground ice is abundant (Fig. 12.14). sufficient to expose ground ice and initiate the process.
They are usually filled quite rapidly by sediments and peat, Debris saturated by the meltwater produced moves down-
and have a limited life span of only a few thousand years. slope as a mud flow or by gelifluction. The headwalls of
In some localities, such as in the northern coastal plain of thaw slumps may reach 8 m in height and retreat at rates of
Alaska, thaw lakes are somewhat elongated with length- over 7 m a-I, making this a significant denudational process
width ratios of 2 : 1 or 3 : 1 and tend to be orientated in a in some periglacial environments. Thermocirques are

.1
,
·;·fi··
",lo • ;.

l.

:>'.. ~>.:

..

./l

Fig.12.14 Thaw lakes in the Tuktoyaktuk Peninsula which lies within the continuous permafrost zone of the Northwest Territories,
Canada. The area shown is about 180 km across. (Landsat image courtesy R. S. Williams fr)
 Periglacial processes and landforms 307

large-scale variants of thaw slumps formed when retreating occurred predominantly during a warm phase of the Holo-
slopes intersect ice wedges. The surface thawing of ice- cene between 9000 and 2500 a BP.
wedge polygons can produce linear and polygonal troughs
surrounding a central mound.
12.3.3 Depositional forms related to mass movement
Alases are major thermokarst depressions from 3 to 40 m
deep and 100 m to 15 km across. They are compound fea- The downslope movement of debris in association with
tures resulting from climatic change or a wide pread distur- periglacial mass movement processes gives rise to a range
bance of the surface such as that arising from a forest fire. of locally significant landforms. Deposits associated with
These environmental perturbations can lead to permafrost gelifluction assume a variety of forms, including sheets,
degradation which initiates a sequence of surface collapse, . benches and lobes. Gelifluction sheets have a smooth,
lake formation and pingo development (Fig. 12.15). Eventual gently sloping surfaces and a bench-like lower margin,
coalescence of individual alases gives rise to alas valleys whereas gelifluction benches are terrace forms. Particular-
which may be tens of kilometres long. Alas formation has ly characteristic are gelifluction lobes which consist of
affected large areas in central Yakutia in eastern Siberia tongue-shaped forms, 30-50 m across, which are elongated
where conditions are particularly favourable. It is estimated downslope (Fig. 12.16). They tend to occur on steeper
that in this area 40-50 per cent of the Plei tocene surface slopes (10-20 0 ) than benches. Where the downslope elon-
has been subjected to alas development, which apparently gation is very marked the term gelifluction stream is used.
Both benches and lobes have steep fronts, 1-6 m high,
Ice wedges which may be subject to erosion. They appear to form as
A~ stones, which have been pushed downslope, accumulate

r l r , f , and dam the movement of material above.

Some gelifluction deposits are crudely stratified and most
B
exhibit a preferred downslope orientation of the long axis
of their angular constituent particles, which may range in
size from boulders down to silt-sized material. Relict ge1i-
fluction deposits are difficult to distinguish from other soli-
c fluction debris, although they should be less weathered than
similar deposits from non-periglacial environments.
Block slopes are slopes mantled with angular boulders,
mostly between 1 and 3 m across, covering 50 per cent or
more of the ground surface. In the case of block streams
the boulders are concentrated in valley bottoms or occur as

Fig.12.15 Schematic representation of the sequence of

development and decay of alases: (A) original lowland surface
with ice wedges; (B) effect of increase in depth of active layer
promoted by climatic warming or ground disturbance; (C) initial
thermokarst stage; (D) young alas stage; (E) mature alas stage;
(F) old alas stage; (G) phase ofpossible pingoformation; (H)
development of collapsed pingos. (After P. J. Williams (1979)
Pipelines and Pennafrost. Longman, London, Fig. 2.2, p. 21,
modified from T. Czudek and J. Demek (1970) Quaternary Fig.12.16 Gelifluction lobe, Hesteskoen, Mesters Vig, north-
Research 1, Fig. 9, p.l 11.) east Greenland. (Photo courtesy A. L. Washburn.)
308 Exogenic processes and landforms

narrow downslope linear deposits on hillsides. Downslope Table 12.7 Nature of slope asymmetry in the periglacial zone of the
movement is indicated for both forms by the orientation of northern hemisphere

the long axes of the boulders which tends to be roughly

LOCALITY ORIENTATION (ASPECT) VALLEY
normal to the slope contours. Gelifluction and frost creep OF STEEPER SLOPE ALIGNMENT
are the favoured transporting mechanisms for most occur-
rences. Block fields, consisting of relatively flat boulder- Central Alaska N E-W
North-west Alaska N E-W
covered surfaces, are also common. They probably form Southampton Island, Canada N E-W
largely by in situ frost shattering and the heaving of jointed Banks Island, Canada SW NW-SE
bedrock and involve no significant downslope movement. Caribou Hills, Canada N, S E-W
Yakutia, Siberia (3 studies) N E-W
Rock glaciers are tongue-shaped masses of angular Andreeland, west Spitsbergen S E-W
boulders resembling in form a small glacier. They have a Conwayland, west Spitsbergen W N-S
steep front, exceptionally exceeding 100m in height, which Kaffioya-Ebene, west Spitsbergen S E-W
Wollaston-Vorl and, east Greenland N E-W
stands at the angle of repose of the constituent material.
Rock glaciers usually descend from cirques or cliff-faces Source: Modified from H. M. French (1976), The Periglacial
and may reach a length of I km or more. Large boulders Environment, Longman, London, Table 8.3 p. 179.
and smaller stones typically form the surface which com-
monly exhibits arcuate ridges and lobes. Active rock glaciers slope angle, while their colluvium will tend to divert streams
contain ice at depth, either filling voids (ice-cemented type) towards opposing north-facing slopes which will be subject
or forming a core (ice-cored type). Some of the ice-cored to undercutting and slope steepening. Moreover, as north-
variety are probably derived from debris-covered glaciers, facing slopes will be in shade longer during the summer
whereas others may be transitional to ice-cored moraines. thaw season, permafrost, where present, will be preferen-
In localities where rock glaciers and ice glaciers are both tially developed and will help to stabilize unconsolidated
currently active the distinction between them is essentially sediments at a steeper angle. Prevailing winds may play a
one of the quantity of debris carried. Where debris-rich glacier role since snow will tend to accumulate on lee slopes and
ice becomes stagnant sufficient ice may remain mixed with the greater quantity of meltwater produced on thawing will
fine sediment at depth to sustain slow movement, largely by augment mass movement on such slopes. Differences in
internal deformation and basal shear. Frost creep and geli- vegetation cover may also be important. In a study in Alaska,
fluction may also contribute to downslope movement. Irres- for instance, vegetation was found to be least developed on
pective of the actual mechanism of movement, rock glaciers the colder north-facing slopes which consequently exper-
are important transportational agents in highland periglacial ienced less active mass movement.
environments. Valley asymmetry has also been identified in areas sub-
ject to periglacial conditions during the Pleistocene. Although
these have commonly been attributed to some of the
12.3.4 Asymmetric valleys
mechanisms just discussed relevant to periglacial environ-
Asymmetric valleys are valleys with one slope significant- ments, it is difficult to separate any inherited periglacial
ly steeper than the other. Virtually all valleys are asymmetric effects from other processes capable of producing asymmetry.
to a certain extent, but some display a degree of asymmetry One important factor is the higher angle of inclination of
sufficiently marked to warrant some special explanation. the Sun in mid-latitude Pleistocene periglacial environ-
Asymmetric valleys are known from a wide range of en- ments in contrast with the present-day high latitude peri-
vironments and can clearly be caused by a number of factors, glacial zone.
but geomorphologists working in the periglacial zone have
long recognized the prevalence of asymmetric valleys
12.3.5 Cryoplanation terraces and cryopediments
there. Collection of field data has indicated no uniformity in
the preferred orientation of the steeper slope in periglacial Cryoplanation terraces (also known as altiplanation
asymmetric valleys (Table 12.7), although it has been sug- terraces) are level or gently sloping surfaces found in the
gested that steeper north-facing slopes predominate in the periglacial zone which are cut into bedrock on hill summits
northern hemisphere below a latitude of about 70 oN. or upper hillslopes. Cryopediments are a similar form
Various conditions existing in periglacial environments developed at the foot of valley sides or on marginal slopes.
may contribute to valley asymmetry. In the northern hemi- Unless they transect structure it may be difficult to dis-
sphere longer exposure to the Sun on south-facing slopes tinguish them from structurally controlled benches, and in
will result in a more rapid and prolonged thawing and a any case lithological and structural factors are important in
greater abundance of meltwater, promoting gelifluction and their formation. Cryoplanation terraces range from 10m to
other mass movement processes. Consequently, south- 2 km across and up to 10 km in length. The risers between
facing slopes will experience a more active reduction of terraces may reach a height in excess of 70 m and stand at
 Periglacial processes and landforms 309

an angle of 30° or more where debris-covered, and up to races and cryopediments are associated with present-day
90° where bedrock is exposed. The terraces are mantled by permafrost, frozen ground does not seem to be a prere-
gelifluction debris and may contain patterned ground. quisite for their formation.
Cryoplanation terraces are thought to be formed by
several processes working in conjunction. In essence they
appear to form by the combined effects of the break-up of
Further reading
bedrock by frost action and scarp recession. Various stages Students of periglacial geomorphology have no shortage of
are involved (Fig. 12.17). According to one model, devel- excellent texts which they can consult. Washburn (1979)
opment begins with the formation of a nivation hollow or provides a fully illustrated and referenced survey of pro-
bench associated with snow patches. Nivation then pro- cesses and landforms; a considerable strength of this text is
ceeds to erode a cliff which recedes as debris is transported the extensive reference made to the work of Russian and
away from the cliff base, largely by gelifluction and slope East European scientists who have made very important
wash. Continued cliff recession on either side of an inter- contributions to periglacial studies. Other general sources
fluve finally forms a summit terrace with residual rock masses. include Embleton and King (1975), French (1976) and
Some researchers, however, believe that most descriptions Williams and Smith (1989). The volumes edited by Church
of cryoplanation terraces are of relict forms, and that in and Slaymaker (1985) and Clark (1988) provide detailed
many cases the features described may be no more than up-to-date treatments of many aspects of periglacial geo-
benches related to the differential erosion of contrasting morphology, while a number of useful papers are contained
lithologies. in Zeitschrift fur Geomorphologie Supplementband, 71.
Cryopediments probably develop in a similar way Various regional studies are included in Boardman (1987),
except that slope wash may be more active than gelifluction although here the emphasis is on relict periglaCial forms in
in carrying debris away. Although many cryoplanation ter- the British Isles.
The problems involved in the human utilization of high
latitude environments have received much attention. Although
these are not specifically considered here, discussions of
various aspects of applied periglacial geomorphology are to
be found in Brown (1970), Harris (1986), Sugden (1982)
and Williams (1979). Journals containing articles on peri-
glacial geomorphology include Biuletyn Peryglacjalny (a
Polish publication but with many articles in English),
Canadian Journal of Earth Sciences, Quaternary Research
and Arctic and Alpine Research, in addition to other journals
which publish papers on geomorphology generally. Useful
annual reviews of periglacial geomorphology are to be found
in Progress in Physical Geography.
Harris (1986) provides a detailed discussion of the nature
and formation of permafrost, while the pioneer work by
Taber (1929, 1930, 1943) on frost action is still worth con-
sulting. Thorn (1979) presents some intriguing data on the
occurrence of freeze-thaw cycles' and the efficacy of frost
shattering in the alpine environment of the Colorado
Rockies, and White (1976) compares the role of frost shat-
tering and hydration shattering. Both field and laboratory
investigations of frost shattering are reviewed by McGreevy
(1981), and Pavlik (1980) provides a brief discussion of the
factors controlling ground ice formation. Salt weathering in
Fig. 12.17 Stages in the development of cryoplanation terraces
in resistant rock: (A) original sUlface; (B) formation ofnivation relation to tor formation is considered by Selby (1972).
hollows; (C) initial development of cryoplanation terrace; (D) Mass movement processes are covered by Harris (1987)
mature stage of cryoplanation terrace development; (E) initial and Rapp (1986), and also in the more general texts,
formation of cryoplanation surface; (F) development of especially Washburn (1979). Benedict (1976) gives a useful
cryoplanation summit sUlfaces. Arrows indicate the direction of review of frost creep and gelifluction and also provides a
surface modification. (Based on model of 1. Demek (1969)
Biuletyn Peryglacjalny 18,115-25, after H. M. French (1976) detailed discussion on mass movement mechanisms in alpine
The Periglacial Environment. Longman, London, Fig. 7.11, environments (Benedict, 1970). Fluvial processes in specific
p.160.) periglacial environments are discussed in Church (1972)
310 Exogenic processes and landforms

and McCann et al. (1972), while Scott (1979) provides a Embleton, C. (1979) Nival processes. In: C. Embleton and J.
comprehensive annotated bibliography on this topic. Embleton Thomes (eds) Process in Geomorphology. Edward Arnold,
and King (1975) and Thorn and Hall (1980) discuss niva- London, 307-24.
Embleton, C. and King, C. A. M. (1975) Periglacial Geomorpho-
tion, while Embleton (1979) and Thorn (1978) consider the logy Edward Arnold, London.
action of snow in the landscape. French, H. M. (1974) Active thermokar t processes, eastern Banks
The classification and origin of patterned ground is Island, Western Canadian Arctic. Canadian Journal of Earth
covered in the seminal papers by Washburn (1956, 1970), Sciences 11, 785-94.
but an extended discussion is also to be found in Washburn's French, H. M. (1976) The Periglacial Environment. Longman,
London.
general text. Nicholson (1976) provides some additional Giardino, 1. R., Shroder, J. F. Jr and Vitek, J. D. (eds) (1987) Rock
material on the characteristics and formation of patterned Glaciers Unwin Hyman, London and Boston.
ground, while Goldthwait (1976) concentrates on the sort- Gleason, K. J., Krantz, W. B., Caine, N., George, J. H. and Gunn,
ing processes. The possible role of convection in the forma- R. D. (1986) Geometrical aspects of sorted patterned ground in
tion of certain types of patterned ground is discussed by recurrently frozen soil. Science 232, 216-20.
Goldthwait, R. P. (1976) Frost sorted patterned ground: a review.
Gleason et al. (1986) and Krantz et al. (1988). The mechanism Quaternary Research 6, 27-35.
of frost cracking and the development of ice wedge poly- Hamilton, T. D., Ager, T. A. and Robinson, S. W. (1983) Late
gons is considered in the important paper by Lachenbruch Holocene ice wedges near Fairbanks, Alaska, U.S.A.: environ-
(1962), and more recent ideas are discussed by Hamilton et mental setting and history of growth. Arctic and Alpine Research
al. (1983), Mackay (1974, 1984, 1986) and Mackay and 15, 157-68.
Harris, C. (1987) Mechanisms of mass movement in periglacial
Matthews (1983). The formation of pingos is reviewed in environments. In: M. G. Anderson and K. S. Richards (eds)
Mackay (1978, 1987) and a detailed field investigation is Slope Stability: Geotechnical Engineering and Geomorphology.
reported by Mackay (1979). Wiley, Chichester and New York, 531-59.
Thermokarst is assessed in Czudek and Demek (1970) Harris, S. A. (1986) The Permafrost Environment. Croom Helm,
with an emphasis on Siberia, and French (1974) considers London.
Kennedy, B. A. (1976) Valley-side slopes and climate. In: E.
contemporary thermokarst development in Arctic Canada. Derbyshire (ed.) Geomorphology and Climate. Wiley, London
Martin and Whalley (1987) provide a useful review of rock and New York, 171-201.
glaciers, while a detailed treatment is to be found in the Krantz, W. B., Gleason, K.1. and Caine, N. (1988) Patterned
volume edited by Giardino et al. (1987). French (1976) ground. Scientific American 259(6), 44-50.
provides a useful discussion of valley asymmetry in both Lachenbruch, A. H. (1962) Mechanics of thermal contraction
cracks and ice-wedge polygons in permafrost. Geological
present and past periglacial environments, while Kennedy Society ofAmerica Special Paper 70.
(1976) takes a broader look at the general factors control- Mackay, 1. R. (1974) Ice-wedge cracks, Garry Island, Northwest
ling valley asymmetry. A model of cryoplanation terrace Territories. Canadian Journal of Earth Sciences 11, 1336-83.
development is presented in Demek (1969). Mackay, J. R. (1978) Contemporary pingos: a discussion. Biuletyn
Peryglacjalny 27, 133-54.
Mackay, 1. R. (1979) Pingos of the Tuktoyaktuk Peninsula area,
Northwest Territories. Geographie Physique et Quaternaire 33,
3-61.
References Mackay, 1. R. (1984) The direction of ice-wedge cracking in per-
Benedict, 1. B. (1970) Downslope soil movement in a Colorado mafrost: downward or upward? Canadian Journal of Earth
alpine region: rates, processes, and climatic significance. Arctic Sciences 21, 516-24.
and Alpine Research 2, 165-226. Mackay, 1. R. (1986) The first 7 years (1978-1985) of ice wedge
Benedict, J. B. (1976) Frost creep and gelifluction: a review. Qua- growth, Illisarvik experimental drained lake site, western Arctic
ternary Research 6,55-76. coast. Canadian Journal of Earth Sciences 23, 1782-95.
Boardman, J. (ed.) (1987) Periglacial Processes and Landforms Mackay, J. R. (1987) Some mechanical aspects of pingo growth
in Britain and Ireland. Cambridge University Press, Cambridge. and failure, western Arctic coast, Canada. Canadian Journal of
Brown, R. J. E. (1970) Permafrost in Canada. University of Earth Sciences 24, 1108-19.
Toronto Press, Toronto. Mackay, J. R. and Matthews, J. V. Jr (1983) Pleistocene ice and
Church, M. (1972) Baffin Island sandurs; a study of Arctic fluvial sand wedges, Hooper Island, Northwest Terrritories. Canadian
processes. Geological Survey of Canada Bulletin 216. Journal of Earth Sciences 20, 1087-97.
Church, M. and Slaymaker, O. (eds) (1985) Field and Theory: Martin, H. E. and Whalley, W. B. (1987) Rock glaciers part 1: rock
Lectures in Geocryology. University of British Columbia Press, glacier morphology: classification and distribution. Progress in
Vancouver. Physical Geography 11, 260-82. .
Clark, M. 1. (ed.) (1988) Advances in Periglacial Geomorphology. McCann, S. B., Howarth, P. J. and Cogley, J. G. (1972) Fluvial
Wiley, Chichester and New York. processes in a periglacial environment: Queen Elizabeth I lands,
Czudek, T. and Demek, J. (1970) Thermokarst in Siberia and its N.W.T., Canada. Transactions of the Institute of British Geo-
influence on the development of lowland relief. Quaternary graphers 55, 69-82.
Research 1, 103-20. McGreevy, 1. P. (1981) Some perspectives on frost shattering.
Demek, 1. (1969) Cryogene processes and the development of Progress in Physical Geography 5,56-75.
cryoplanation terraces. Biuletyn Peryglacjalny 18, 115-25. Nicholson, F. H. (1976) Patterned ground formation and descrip-
 Periglacial processes and landforms 311

tion as suggested by Low Arctic and Subarctic examples. Association ofAmerican Geographers 68, 414-25.
Arctic and Alpine Research 8, 329--42. Thorn, C. E. (1979) Bedrock freeze-thaw weathering regime in an
Pavlik, H. F. (1980) A physical framework for describing the alpine environment, Colorado Front Range. Earth Surface
genesis of ground ice. Progress in Physical Geography 4, Processes 4,211-28.
531--48. Thorn, C. E. and Hall, K. (1980) Nivation: an arctic-alpine com-
Rapp, A. (1986) Slope processes in high latitude mountains. parison and reappraisal. Journal of Glaciology 25, 109-24.
Progress in Physical Geography 10, 53-68. Washburn, A. L. (1956) Classification of patterned ground and
Scott, K. M. (1979) Arctic stream processes - an annotated biblio- review of suggested origins. Bulletin of the Geological Society
graphy. United States Geological Survey Water Supply Paper ofAmerica 67,823-65.
2065. Washburn, A. L. (1970) An approach to a genetic classification of
Selby, M. J. (1972) Antarctic tors. Zeitschriftfiir Geomorphologie patterned ground. Acta Geographica Lodziensia 24, 437--46.
Supplementband 13, 73-86. Washburn, A. L. (1979) Geocryology: A Survey of Periglacial
Sugden, D. (1982) Arctic and Antarctic: A Modern Geographical Processes and Environments. Edward Arnold, London.
Synthesis. Blackwell, Oxford. White, S. E. (1976) Is frost action really only hydration
Taber, S. (1929) Frost heaving. Journal of Geology 37, 428-61. shattering? Arctic and Alpine Research 8, 1-6.
Taber, S. (1930) The mechanics of frost heaving. Journal of Williams, P. J. (1979) Pipelines and Permafrost: Physical
Geology 38,303-17. Geography and Development in the Circumpolar North. Longman,
Taber, S. (1943) Perennially frozen ground in Alaska: its origin New York.
and history. Bulletin of the Geological Society of America 54, Williams, P. J. and Smith, M. W. (1989) The Frozen Earth:
1433-548. Fundamentals of Geocryology. Cambridge University Press,
Thorn, C. E. (1978) The geomorphic role of snow. Annals of the Cambridge and New York.
 13
Coastal processes and
landforms

13.1 The coastal environment Shore

A
Continental . Line
.... Slope ....-Continental Shelf------+l
So far in the preceding chapters we have discussed those l Coastal
- 1 - - Shore Zone--........' Plain"
processes active in the sub-aerial environment and the land- I Beach-Dune
forms to which they give rise. None the less, around 70 per
Sealevel
- Nearshore --I Ridge
cent of the Earth's surface is covered by water, and the ,.-------------••- .
: .."l./-:<:'::::;::••. ::::':::}/::.:?:.;.:....
coast, which forms the interface between land and water,
experiences a particular array of geomorphic processes and
a range of characteristic landforms. Waves and tides involve
the movement and dissipation of large amounts of energy
and this is capable of causing rapid and spectacular changes Shore
Line B
in landforms along coastlines.
Continental I
The coastal environment, or littoral zone, can be defined +- Slope ....-Continental Shelf~

as that area lying at the interface between the land and the ..Shore: Zone_Coastal Plain
sea (or other large water body). It includes both the zone +- Nearshore --I
of shallow water within which waves are able to move Sea level !..,
sediment, and the area landward of this zone, including
beaches, cliffs and coastal dunes, which is affected to some
degree by the direct or indirect effects of waves, tides and
currents. While the terms shoreline or littoral zone refer to
the area of frequent, or at least occasional, wave action along Fig. 13.1 Schematic representation of some of the important
a
the edge of the sea or lake, the coastal environment itself elements of coastal morphology: (A) low elevation coast, and (B)
may extend inland for many kilometres. Figure 13.1 illus- coast backed by cliffs. (After D. L. Inman and C. E. Nordstrom,
trates some of the major components of the coastal environ- (1971) Journal of Geology 79, Fig. 3, p. 6.)

ments for both sandy and rocky coasts.

As in other geomorphic environments, coasts are com- sheer size, are modified much more slowly. Great care must
posed of a wide range of landforms, only some of which are be taken in relating contemporary forms to current processes
in equilibrium, or quasi-equilibrium, with prevailing pro- especially since important agents of coastal modification,
cesses. The Holocene has seen great changes in the position such as storm waves, may change in their frequency of
of shorelines and in coastal morphology as a result of pre- attack over short spans of time.
dominantly rising sea levels, and there has been a con- Many attempts have been made to categorize coastal
tinuous succession of adjustments to these rapidly changing landforms, either in order to provide a descriptive frame-
conditions (see Chapter 17 for a detailed discussion of sea- work for subsequent investigations or to create a genetic
level change). While beaches composed of unconsolidated classification which can be applied to the interpretation of
and readily transportable sand can adjust quickly to changes coastal· morphology. Most such classifications have focused
in the nature and energy of coastal processes, other land- on the consequences of relative changes in sea level, cate-
forms, either because of their resistance to erosion, or just gorizing coasts as either emergent or submergent. Although
314 Exogenic processes and landforms

such classificatory schemes have their uses, the different as the west coast of South America, lie along convergent
rates of adjustment of different coastal landforms to changes plate boundaries, whereas passive margin coasts, as their
in sea level, as well as other factors affecting the develop- name implies, are located along passive continental margins.
ment of coastal features, make them difficult to apply in Several varieties of passive margin coast can be identified.
many cases. Nevertheless, a growing appreciation of the Young passive margins, such as those flanking the Red
nature of the global variation of a number of variables affect- Sea, give rise to nascent passive margin coasts, and these
ing coastal development does allow some valuable genera- evolve into mature passive margin coasts as the continen-
lizations to be made about the varying effectiveness of tal margin moves away from the spreading ridge, cools, and
different processes around the world's coasts. Some of this subsides. Passive margin coasts lying on the opposite side
variation is evident in the world maps included in this of a continent from a convergent margin can be described
chapter. Moreover, in keeping with the links between exo- as passive margin-distal orogen coasts. Finally, passive
genic and endogenic processes which form an important continental margins protected by island arc systems offshore
element of this book, it is appropriate here to anticipate the are termed marginal sea coasts.
theme of Part IV and illustrate the overall control that There is a fairly clear relationship between the tectonic
tectonics exerts on the broad nature of coastal topography. nature of coastal margins and the width of the continental shelf
It is possible to classify coasts on the basis of their tec- offshore with, for example, a strong predominance of narrow
tonic setting (Fig. 13.2). Convergent margin coasts, such shelves along convergent margin coasts (Table 13.1). These

6 2
10km
1~
1000 0 1000
km

~ ~ ~~~~~~~~:~~~ve I.:i:.: :.:· :~i: i: ~: ".: : ~!:.il

Convergent Mature passive
margin coasts margin coasts

~ ~
----
Marginal ----
----- Mature passive margin • Large deltas
sea coasts ----- coasts with distal orogen

Fig.13.2 Tectonic setting o/the world's coastline. (Based partly on 1. L. Davies (1980) Geographical Variation in Coastal
Development (2nd edn). Longman, LOY.ldon, Fig. 2, p. 9 and D. L.lnman and C. E. Nordstrom (1971) Journal of Geology 79, Fig. 4,
p.10.)
 Coastal processes and landforms 315

Table 13.1 Relationship of continental shelf width to tectonic setting

TECTONIC SETTING % OF WORLD

COASTLINE
CONVERGENT PASSIVE MARGIN COASTS
MARGIN COASTS
SHELF NASCENT MATURE MATURE MARGINAL
WIDTH DISTAL OROGEN SEA

Narrow «50 km) 23.9 6.5 11.2 12.6 3.3 57.5

Wide (>50 km) 1.4 0.1 3.6 29.1 8.3 42.5
Total 25.3 6.6 14.8 41.7 11.6 100.0

Note: All figures are percentages.

Source: Modified from D. L. Inman and C. E. Nordstrom (1971)Journal of Geology 79, Table 6, p. 14.

first order relationships are reflected in second order effects, gical units creates a characteristic large-scale coastal mor-
as is evident in the association between tectonic setting and phology (Fig. 13.3). In other cases, especially along passive
a range of specific types of coastal features (Table 13.2). margin coasts, the structural grain may be oblique or per-
Wave erosion, for instance, is important on convergent mar- pendicular to the coastline, and in this case a highly in-
gin coasts which receive comparatively meagre sediment dented coast can be produced, if fluvial valleys or glacial
supplies from continental drainage systems. Deposition by troughs are drowned by a rise in sea level. The resulting
marine action, however, is predominant on passive margin submerged valleys are known, respectively, as rias and
distal orogen coasts, such as those of eastern South America, fjords (Fig. 13.4). Finally, we can sec: in Table 13.3 to what
which are supplied with abundant sediment from an oro- extent it is possible to relate the overall morphology of coasts
genic belt on the other side of the continent. Large deltas in to their tectonic setting.
fact appear to be more or less confined to mature passive
margin and marginal sea coasts since it is only these that
provide the outlets for the world's major drainage systems 13.2 Waves, tides and currents
(Fig. 13.2). Further implications of the relationship between
tectonic setting and continental drainage systems are explored 13.2.1 Waves
in Chapter 16. Waves are characterized by their length, height (amplitude),
Structural and lithological controls are also evident at velocity (rate of forward motion of the wave peak) and
this second-order scale. Active convergent continental margins period (the interval of time between successive wave peaks
in many cases have fold structures running roughly parallel passing the same point) (Fig. 13.5). These properties, and
to the coastline, and differential erosion of adjacent litholo- the relationships between them, vary greatly depending on

Table 13.2 Relationship of coastal features to tectonic setting

TECTONIC SETTING TOTAL % OF WORLD

COASTLINE
CONVERGENT PASSIVE MARGIN COASTS
MARGIN COASTS
COASTAL NASCENT MATURE MATURE MARGINAL
FEATURES DISTAL OROGEN SEA

Wave erosion 47.9 5.6 3.8 4.9 37.8 100 44.7

Wave deposition 15.5 12.4 21.4 33.3 17.4 100 11.6
Fluvial deposition 5.7 6.4 9.9 62.4 15.6 100 3.2
Aeolian deposition 1.9 18.8 79.3 100 1.2
Biogenic activity 36.1 21.0 11.3 15.8 15.8 100 3.0
Glaciated coasts 6.4 7.2 86.4 100 30.7

% of world
coastline 39.1 4.3 6.8 35.4 8.8 94.4*

*Excluding the Antarctic coastline.

Note: All figures ar.e percentages.
Source: Modified from D. L. Inman and C. E. Nordstrom (1971) Journal of Geology 79, Table 7, p. 14.
 316 Exogenic processes and landforms

Fig. 13.3 The concordant convergent margin coast of western Pakistan which runs roughly parallel to the tectonically active Makran
Coast Range, a region of rapid uplift associated with the convergence of the Arabian and Eurasian Plates. The prominent headlands
visible in the image, which covers an area about 180 km across, are recently upthrustedfault blocks. Note the absence of major
indentations along the coast and the well-developed zetaform beach to the east of the prominent headland in the bottom left of the
image. (Landsat image courtesy N. M. Short.)

the nature of the mechanism generating the wave, the inten- fetch is the distance over which the wind blows and it has
sity of this generating mechanism and the environment in an important influence on wave height and period. The
which the wave exists. highest waves are produced by strong winds blowing in the
same direction over a long distance (Box 13.1). Excep-
13.2.1.1 Wind-generated waves tionally, wave heights of more than 15 m can be generated
Generation of waves by the wind involves a transfer of by the wind. It is possible broadly to identify shoreline en-
.energy from moving air to a water surface. Although a very vironments subject to different levels of wave energy on
familiar process, the way in which this occurs is still not the basis of prevailing wind speeds, fetch and coastal
fully understood. The amount of energy exchanged depends configuration (Fig. 13.6).
mainly on the velocity, duration and fetch of the wind. The Waves actively being generated by the wind are known
 Coastal processes and landforms 317

Fig. 13.4 The mature passive margin coast ofsouth-west Norway where the structural grain is oblique or perpendicular. The
discordance between structure and coastal alignment has been emphasized by glacial erosion which has led to the development of
fjords extending up to 150 km inland. Sognefjord, one of the largest, runs roughly east-west across the lower centre of the image. The
area covered is about 180 km across. (Landsat image courtesy N. M. Short.)

Box 13.1 Wave height as sea waves. Their periods vary, so as they disperse from
the zone of generation they become separated since longer
Relationships between wind velocity and fetch and wave period waves with a greater wave length travel more rapid-
height have been determined empirically from observa- ly (Box 13.2). This separation of waves of different periods
tional evidence. Wave height is proportional to the square produces swell waves. These may travel for thousands of
root of fetch and the square of wind velocity: kilometres across whole oceans, gradually losing height
H = 0.36 -VF' and H = 0.031 U2 and energy. Over the first 200 km of travel a 10m high wave
is
where H the wave height (m), F the fetch (lan) and U the would be reduced in height to around 2 m with an 80-90
wind velocity (m S-I). per cent loss in energy, but thereafter the rate of energy loss
becomes much less.
318 Exogenic processes and landforms

Table 13.3 Relationship of coastal morphology to tectonic setting

TECTONIC SETTING

CONVERGENT PASSIVE MARGIN COASTS

MARGIN COASTS
MORPHOLOGICAL NASCENT MATURE MATURE MARGINAL
CATEGORY DISTAL OROGEN SEA

Mountainous coa t 97.2 8.0 2.5

Narrow shelf, hilly coast 75.1 14.1 5.6
Narrow shelf, plains coast 15.9 46.2 1.5
Wide shelf, plains coast 4.0 89.3 3.1
Wide shelf, hilly coast 2.2 77.4
Deltaic coast 1.0 3.4 1.3 5.8
Reef coast 3.0 1.9 5.6
Glaciated coast 2.8 29.3 3.8
Total 100.0 100.0 100.0 100.0 100.0
% of world coastline* 39.0 4.6 7.5 35.2 8.1

*Excluding Antarctic coastline.

Note: All figures are percentages.
Source: Modified from D. L. Inman and C. E. Nordstrom (1971) Journal of Geology 79, Table 8, p. 19.

Wave length Box 13.2 Period and wave length for oscillatory
waves

Separation of oscillatory waves as they disperse from the

point of generation occurs because of the relationship be-
Still water tween wave length, wave velocity and wave period:
depth
g1'2 gT
Wave base
L= - -
2n
and u=-
2n

Fig.13.5 Major characteristics of waves. where L is the wave length (m), T the period (s), U the wave
velocity (m S-I) and g the acceleration of gravity (m S-2).
Thus waves generated by a storm with a period of 5 swill
In the deep water of the oceans there is little actual for- have a wave length of 39 m and a velocity of7.8 m s-I, while
ward motion of water in waves because it is the waves that those with a period of 10 s will have a wave length of 156 m
and a velocity of 15.6 m S-I.
move rather than the water. Such deep-water waves are
called oscillatory waves since the individual water particles
move with a virtually circular motion, oscillating around a
equal to water depth, but this relationship is also influenced
more or less fixed point. Motion declines rapidly down-
by wave length and shoreline gradient.
wards, and at a depth approximately one half that of wave
The type of breaking wave formed is related to wave
length, known as wave base, there is negligible movement
steepness and the gradient of the shore (Fig. 13.7). Once
(Fig. 13.5). As waves move towards shallower water their
the wave form has been destroyed the remaining water
mode of movement changes dramatically. Where the water
depth decreases to less than that of wave base the sea floor
starts to interfere with the oscillatory motion and the orbit Box 13.3 Velocity and wave length of translatory
of individual water particles become more elliptical. For- waves
ward movement of water now becomes important as the
oscillatory waves are transformed into translatory waves. As translatory waves enter shallower water their period re-
As the water depth becomes progressively more shallow, mains constant but their wave length and velocity changes:
wave length and velocity decrease, wave height increases U=>Jgd and L=Ngd
and consequently the wave steepens (Box 13.3). Eventually where U is the wave velocity (m S-I), L the wave length
the wave is over-steepened to the stage where it breaks as (m), d the water depth (m), T the period (s) and g the acce-
its crest crashes forward creating surf. This occurs when leration of gravity (m S-2).
wave height has built up to the point where it is roughly
 Coastal processes and landforms 319

P P P
P

West coast Storm wave Trade wind and

swell environments environments monsoon influences

~/
't East coast ~ Pro~ected sea ~ !ropical cyclone
~ swell environments ~ environments ~ Influences

Fig.13.6 Global distribution of major wave environments. (Modified from J. L. Davies, (1980) Geographical Variation in Coastal
Development. (2nd edn) Longman, London, Fig. 27, p. 43.)

SPILLING Fig.13.7 Major types of hreaking wave; in reality breaking

waves occur in a continuous spectrum offorms related to the
~~ gradient of the shore and wave period. Gentle gradients tend to
he associated with spilling breakers where water from the crest
cascades down the wave front. Plunging breakers, where the crest
curls over and crashes down on the advancing wave hase, are more
PLUNGING characteristic of steeply sloping shores. Where waves of low
steepness encounter high gradient shores they may fail to break

=
fully but form instead surging breakers in which the crest collapses
and the hase advances up the shore.

moves up the shore as swash and returns under the force of

SURGING gravity as backwash. Breakers, especially those associated
with large storm waves, expend large quantities of kinetic

]
energy and are capable of accomplishing considerable geo-
morphic work (Fig. 13.8). Because swell waves can travel
such large distances the energy released by breakers along
320 Exogenic processes and landforms

configuration is particularly favourable (as in the North

Sea) (Fig. 13.6). Storm surges are most destructive along
very low-lying coasts where their effects can extend many
kilometres inland, but their geomorphic significance arises
in large part from the way in which they lead to wave
attack at much higher levels along a shoreline than is
reached by normal waves.

13.2.1.3 Seismic sea waves

Seismic sea waves (or tsunami) are generated by displace- .
ment of the ocean floor. They can be produced by volcanic
eruptions and catastrophic submarine mass movements, but
by far the most significant cause is earthquakes. The dis-
placement of a large mass of ocean water, often at a great
Fig.13.8 Plunging breakers on a steeply sloping beach depth, generates waves of small amplitude (usually less
composed offine basalt gravel on the south-east coast of Hawaii. than 1 m), considerable length (up to 200 km or more) and
At the time the photograph was taken large waves were being high velocity. As with translatory waves the wave velocity
generated by a tropical storm less than 200 km away. is related to ocean depth and waves can travel at more than
600 km hr 1 over water 3000 m deep (Box 13.3). Seismic
a coastline may have originated in a storm thousands of sea waves can therefore cross the entire Pacific Ocean in a
kilometres away. matter of a few hours. As they near land the decrease in
The angle at which waves strike a shoreline is signifi- water depth leads to a marked reduction in velocity and an
cantly affected by wave refraction. This occurs when the increase in wave amplitude, so by the time the coast is
water depth as the shoreline is approached falls to less than reached wave heights may have reached 15 m. They seem
half the depth of wave base. Beyond this point further de- to be most effectively generated by shallow earthquakes
creases in water depth are accompanied by a reduction in associated with oceanic trenches and this probably accounts
wave velocity. Waves are refracted in this situation because for their predominance in the Pacific Ocean. They are
they are bent towards the part of the shoreline where they comparatively rare phenomena, major events being exper-
are moving most slowly. If we imagine a wave obliquely ienced on average only every 25-50 a in Hawaii, for example.
approaching a straight coast, those parts of the wave which Although capable of effecting catastrophic changes in'
reach shallow water first will slow down and will tend to be coastal landforms and causing severe flooding and loss of
caught by the other parts of the wave still in deep water. life, seismic sea waves are sufficiently infrequent to be
This refractive effect means that incoming waves tend to be outweighed in their geomorphic significance by less spec-
roughly parallel with the offshore submarine contours as tacular but more consistent wind-generated wave action.
they reach the shoreline, although they are rarely exactly
parallel. Wave refraction is important because it causes
wave energy to be concentrated at headlands where waves 13.2.2 Tides
converge and to be dissipated in bays where waves diverge. Tides result from the gravitational attraction exerted on
ocean water by the Moon and the Sun, the former having
13.2.1.2 Storm surges around twice the effect of the latter. The motions of the
Storm surges are generated by the combined effects of low Earth, Moon and Sun with respect to each other produce
atmospheric pressure and very high wind speeds. In the semi-diurnal tides along most coasts in which there are
centre of tropical storm systems atmospheric pressure may two lows and two highs approximately every 24 h. Tides
drop as far as 100 mb below normal, and this can 'suck up' higher than normal, known as spring tides, occur every
sea level below the centre of the cyclone by up to 1 m. The 14-75 days when the Sun and Moon are aligned. In be-
largest storm surges are produced when winds are blowing tween these periods lower than normal, or neap tides,
towards coastal embayments and are accentuated if they occur when the Sun and Moon are positioned at an angle of
coincide with high tides. By the time the waves generated 90° with respect to the Earth. Spring and neap tides involve
by the storm surge reach the coast they may have built up deviations of about 20 per cent above and below normal
to a height of several metres above normal high tides. tidal ranges.
Although comparatively uncommon on a global basis, Several factors complicate this simple picture. These
storm surges occur repeatedly along coasts experiencing include the size, depth and topography of ocean basins,
tropical cyclones, such as the Bay of Bengal, and in mid- shoreline configuration and meteorological conditions. Much
latitude areas subject to intense storms where the coastal of the Pacific coast, for instance, experiences a regime of
 Coastal processes and landforms 321

mixed tides in which the highs and lows of each 24 h Tidal range and type are of geomorphic importance for
period are of different magnitudes. Other coastlines, such several reasons (Fig. 13.10). Tidal type determines the
as those of much of Antarctica, have diurnal tides with only interval between tides and therefore the time available for
one high and one low per 24h. the shore to dry after high tide. This is important for shore-
Although contrasts between tidal types are important in line weathering processes and biological activity. More
some coastal processes, of much greater overall geomorphic significantly, it affects the intensity of tidal currents since,
significance is tidal range. This can vary from a minimal for a given tidal range, the velocity of water movement will
range in virtually enclosed water bodies ·such as the Medi- be greater in semi-diurna:I regimes than for mixed or diurnal
terranean Sea and Black Sea, to less than 2 m along some types because there is a shorter interval between high and
open coasts and up to 6 m or more where there is a major low tide. This effect is particularly important in narrow
semi-diurnal component to the tide. In fact it is useful to coastal embayments where tidal flows are concentrated.
consider three categories of tidal range; microtidal (less Tidal range is important because it controls the vertical
than 2 m), mesotidal (between 2 and 4 m) and macrotidal distance over which waves and currents are effective along
(above 4 m) (Fig. 13.9). The most extreme ranges occur the shoreline, and, in conjunction with the shoreline gra-
where the coastal configuration and submarine topography dient, it determines the extent of the intertidal zone, that
induce an oscillation of water in phase with the tidal period. is, the area between high and low tide.
This effect is particularly pronounced in the Bay of Fundy
in north-eastern Canada where the tidal range is nearly 13.2.3 Currents
16 m. In some estuaries with a large tidal range a single The geomorphic importance of currents lies in their role in
wave several metres high can be formed by the incoming transporting sediment and, less significantly, in erosion.
tidal flow as it experiences drag in entering shallower water. Currents associated with tides can transport and erode sedi-
Such waves, known as tidal bores, may reach velocities of ment where flow velocities are high. This is usually con-
30 km h- 1 and can be potent erosive agents. fined to estuaries or other enclosed sections of coast which

6 2
10 km
1~
1000 0 1000
km

Fig.13.9 Global variation in tidal range; the ranges indicated are/or spring tides. (Modified/rom 1. L. Davies (1980) Geographical
Variation in Coastal Development. (2nd end) Longman, London, Fig. 33, p. 51.)
322 Exogenic processes and landforms

IMICROTIDALI MESOTIDAL I MACROTIDAL I

/,,-i\ Rip Head rf
BARRIER
ISLANDS
.. t u

- I
Ripcurrent

'!.y
Breaker Zone
\\ (~)
(\~.~

/
f /.J
I f
I
Mass Transport
, _ _---"",],1:..-- _
\\
I~
\
I
I
TIDAL
DELTAS ..-F I u

-
........:....
----~\~
Longshore Currents - - -
-1
X\:~rr~~:~~·~::\}\.:~H.:.\~;·~·;·:{:::·:::·::·:!;~~:::}};::}:.!~:::::.:::::.:..::::':':~
TIDAL
SAND
RIDGES I I --E IIIIij Antinode
Incident ocean waves

Antinode Antinode

~---
TIDAL in phase out of phase in phase

INLETS I EQge~~ __

High breakers Low breakers High breakers

.. +

[~
TIDAL
I I I I
I I I Rip I I I
FLATS
... ) \
-----'"
J \
'---_/
/ \
...
I~
SALT
MARSHES

ESTUARIES

o
I~
!
2
I
4
!
6
!

Tidal Range (m)

Fig.13.10 Relative abundance of different types of coastal

landforms in relation to tidal range. (Modified from T. Elliott,
1986, in H. G. Reading (ed.) Sedimentary Environments and
Facies. (2nd edn) Blackwell Scientific Publications, Oxford,
Fig. 7.1, p. 156; and M. O. Hayes (1976) in: L. E. Gronin (ed.)
Estuarine Research: Vol. 2, Geology and Engineering, Academic
Press, London, Fig. 1, p. 5.)

experience semi-diurnal tides with a high range-. In mesotidal

and microtidal reg'imes fast tidal currents are confined to
zones of constriction such as around coastal promontories
and within narrow channels between islands. The highest
velocities usually occur midway between the high tide point
and the low tide point, but in long estuaries the timing of
maximum tidal current velocities in the tidal cycle may vary
with location. This is important since it determines the level
in the intertidal zone where tidal currents will be most
active. Fig.13.11 Currents in the nearshore zone: (A) feeder longshore
Currents are. also associated with wave action along currents, rip currents and mass transport returning water to the
shorelines (Fig. 13.11). Where waves move in a direction surf zone; (B) relationship between edge waves and rip currents;
perpendicular to a beach the backwash is normally confined (C) position of rip currents with edge waves and incoming waves
to fairly evenly spaced zones forming rip currents (Fig. which are 180 0 out ofphase; (D) longshore currents associated
with oblique wave approach. Based on 1. L. Davies (1980)
13.12) which are fed by water moving parallel to the beach Geographical Variation in Coastal Development (2nd edn)
in longshore currents (Fig. 13.11 (A)). This cellular circu- Longman, London Fig. 76, p. 103 and P. D. Komar and D. L.
lation pattern is caused by variations in wave height parallel Inman (1970) Journal of Geophysical Research 75, Fig. 9,p. 5926
Copyright by the American Geophysical Union.
 Coastal processes and landforms 323

but tidal type and meteorological factors are also important

since these affect, respectively, the time available for dry-
ing between tides and the rate of evaporation. The most
aggressive regime for water-layer weathering probably
occur along coasts characterized by high evaporation rates
and mixed or diurnal tides.
The most important process in shoreline weathering is
probably salt weathering, although its effectiveness depends
on the ability of shoreline rocks to absorb sea water and
spray. Chemical weathering also plays a role, the nature of
the chemical reactions being controlled by rock minera-
logy, and their rate being influenced by temperature and
variations in micro-environmental factors such as organic
activity and pH along the shore. Solution is apparently a
Fig. 13.12 Rip currents (arrowed) on a sand beach north of significant process on limestones and calcium carbonate-
Cairns, Queensland, Australia. cemented rocks, although there are difficulties in explaining
how it occurs since sea water is normally saturated or
with the shore. Water from the higher sections of breakers supersaturated with respect to CaC0 3 . The lower solubility
extends further up the beach and returns at the points where of CO 2 in warm water would suggest that solution should
the lower wave sections have broken. be a more effective proces along cold high latitude coasts,
Variations in the height of breaking waves may arise but apparently solutional forms are also found on low lati-
from the configuration of the horeline, but the presence of tude coasts. Rock pools on tropical shores may isolate poc-
cellular circulation along straight beaches suggests another kets of water which are capable of solution through pH
mechanism. Exactly what this mechanism is has been much changes associated with photosynthesis. There is still much
debated, but it may involve edge waves (Fig. 13.11(B)). uncertainty about the effectiveness of sea water in the dis-
These are formed by the resonance between waves arriving solution of carbonate rocks and recent attention has been
at the shore (incident waves) and those reflected from it. focused on the role of shoreline organisms, especially blue-
The resonance raises the height of breakers at regular inter- green algae, which may be capable of leaching carbonate
vals parallel to the shore. Irrespective of the precise mech- through complex biochemical reactions.
anism whereby they are initiated, once established rip cur- In high latitudes frost weathering is a potentially signifi-
rents and longshore currents are partially self-sustaining. cant shoreline weathering process because of the frequency
This is because offshore sediment transport in rip current of wetting of bare rock surfaces in the intertidal zone. Sea
zones, and onshore transport between them, modifies the water itself is not very effective because, on freezing, the
form of the beach so as to enhance the cellular circulation salts it contains become segregated and produce a rather
pattern (Fig. 13.11(C)). Where waves strike the shore obli- soft ice incapable of transmitting high stresses to the rock.
quely the long hore currents will be in predominantly one Frost weathering is apparently more effective where fresh
direction (Fig. 13.11(D)). As we will see this has important water is available from melting snow banks or permafrost.
implications for sediment movement along the shore.
13.3.1.2 Coastal erosion
Waves are the most important erosive agent along most
13.3 Coastal processes
coasts but their effect varies with wave energy and charac-
13.3.1 Destructional processes teristics, and with the nature of the material exposed to wave
attack. Where a coast is formed by steep cliffs which plunge
13.3.1.1 Shoreline weathering straight into deep water swell waves are not forced to
A shore zone may be subject to the same range of physical break before they impact. As there is virtually no forward
and chemical weathering processes that occur on land, but mass displacement of water in such waves they are reflec-
the presence of sea water and the cycle of wetting and dry- ted with little loss of energy and accomplish negligible
ing produced by tides introduces significant additional geomorphic work. Much more commonly coasts are ubject
factors. The tidal cycle of wetting and drying is instrumen- to breaking waves. These involve significant mass displace-
tal in a variety of we.a~hering processes which are appro- ment and a considerable loss of kinetic energy as they break
priately considered collectively as water-layer weathering. on a shoreline. This energy is dissipa~ed over a short dis-
The zone affected extends from iow water mark to the tance where the shore gradient is steep, but over a greater
furthest limit reached by waves and pray at high tide. Its horizontal distance where it is shallow. Of the main types
areal extent is therefore controlled largely by tidal range, of breaking wave, plunging breakers produce the highest
324 Exogenic processes and landforms

instantaneous pressures (up to 600 kPa or more) since air biological erosion is much greater along coasts characterized
can be trapped and compressed between the leading wave by low wave energies since here abrasion and quarrying
front and the shore. operate at only moderate or low intensities.
The combined effect of air compression and the impact
of a considerable mass of water is capable of dislodging
13.3.2 Constructional processes
fractured rock and other loose particles, a process known as
quarrying. Well-jointed rocks and unconsolidated or weakly 13.3.2.1 Sediment movement and deposition
consolidated sediments are particularly susceptible. Break- In considering the movement and deposition of sediment I
ing waves may also throw particles against the shore and along coasts it is necessary to identify the main sources of
this leads to the abrasion of shoreline materials. The effec- sediment, its mode of transportation and the zones in which
tiveness of abrasion is highly dependent on wave energy it accumulates or is removed from the littoral zone. If we
and on the availability of suitable material, such as pebbles, ignore, for the moment, the movement of sediment parallel"
along the shore. Large boulders can only be moved in the to the shore, there are three main sources of sediment: (1)
most intense storms, whereas small boulders and pebbles the coastal landforms themselves, ir).cluding cliffs and beaches;
can be moved much more frequently by waves of only (2) the land area inland from the littoral zone; and (3) the
moderate energy. offshore zone and beyond.
A variety of sea shore organisms, including some molluscs, The erosion of coastal landforms, especially cliffs, can
boring sponges and sea urchins, can destroy rocks by phy- locally provide abundant· sediment in environments with
sically boring into them. Their effectiveness is influenced high wave energies (especially where unconsolidated deposits
by rock type, most sedimentary rocks being much more are being eroded), but this is riot generally a significant
susceptible than igneous rocks. The relative importance of source in the tropics where environments with low wave

Fig. 13.13 Oblique aerial view showing the effects of longshore drift along the coastline of Tasman Bay, North Island, New Zealand.
Abundant sediment is carried down from the Richmond Range by the Motueka River and some of this is transported southward to form
small barrier islands. Note the recurved spit towards the bottom centre of the view which clearly indicates the prevailing direction of
sediment transport.
 Coastal processes and landforms 325

energies are common. This is supported by the relative lack of wave attack (Fig. 13.11 (D». As a result of gravity it
of coasts formed of bedrock in the tropics. Even where pre- returns down the beach in the backwash perpendicularly to
sent, cliffs formed of well-consolidated strata recede slowly the shoreline, and so moves a short distance along the beach
and supply little sediment. In contrast, pre-existing con- in each cycle of swash and backwash. The overall signifi-
structional features such as fossil beaches and dunes are much cance of longshore drift depends on whether it occurs along
more susceptible to wave attack and sediment mobilization. coasts of free or impeded transport. Impeded transport is
Land-derived ediment can be provided by mass move- characteristic of coasts with an irregular configuration, the
ment, especially where cliffs composed of material suscep- amount of longshore drift being limited by the trapping of
tible to such processes are being actively undercut. In sediment against headlands. This contrasts with unimpeded
periglacial environments, such as the Arctic coast of the longshore sediment transport along straight coasts. Given a
USSR, gelifluction is a particularly important means of particular coastal configuration the rate of longshore drift
transporting sediment into the littoral zone, while in other will be related to the predominant angle of wave approach
high latitude environments material can be supplied by and the constancy and velocity of longshore currents.
glaciers. Nevertheless, by far the most significant source of
supply overall is from rivers (Fig. 13.13). It has been 13.3.2.2 Organic activity
estimated that on a global basis river contribute about one A variety of organisms are either directly or indirectly res-
hundred 'time it much sediment as marine erosion. This ponsible for the construction of some coastal landforms.
ratio varies latitudinally, with fluvial sources being of even The most spectacular example of dirett construction is by
greater relative importance throughout most of the tropics the corals and other carbonate-secreting organisms which
where low wave energie make marine erosion less form coral reefs. These structures can attain immense
effective, but even most beaches in the mid-latitudes, where sizes, as in the ca e of the Great Barrier Reef which extends
wave energies are on average significantly higher, probably along much of 'the north-east coast of Australia. Because
contain less than 5 per cent of material derived from cliff they require a minimum sea temperature for growth, corals
erosion. are concentrated in the tropics, as are various calcareous algae
Onshore transport may return to the littoral zone pre- which form carbonate encrustations along many tropical
viously eroded beach material or fluvial sediments initially shores.
deposited well offshore. PaI1icuiarly severe storm waves, A range of plants are adapted to salt water and these
storm surges and seismic sea waves may occasionally bring form salt-marsh communities in the intertidal zone along
in sediment from beyond the offshore zone but much of the sheltered, muddy coasts. In the tropics mangroves are an
present-day onshore movement of sediment arises from the important element in coastal vegetation and, together with
post-glacial rise in sea level (see Section 17.6.2). Sediments other halophytic (salt-tolerant) plants, they may playa geo-
deposited on previously exposed continental shelves are morphic role by trapping sediment within their root systems
still being moved inland in response to the Holocene rise in and thereby aiding the process of deposition. Their precise
sea level and the adjustment of the littoral zone to present- role, however, is uncertain and we shall return to this in the
day tidal and wave regimes. In some localities this·sediment discussion of tidal flats. More certain is the role of plants in
supply seems to have been exhausted and some depositional stabilizing coastal dune systems and contributing to 'sand
landforms constructed during tIlls period of ea-Ievel rise deposition. Again we will return to this topic in the section
are now being eroded. on coastal dunes.
While sediment is constantly being moved more or less
perpendicularly towards and away from the shoreline by
tidal and wave action, the predominant net movement of
13.4 Coastal landforms
sediment along most coasts is parallel with the shore through 13.4.1 Destructional forms
the effects of longshore currents (Fig. 13.13). Thi,s move-
ment is termed longshore drift and its rate is dependent on 13.4.1.1 Cliffs
wave energy and the angle at which waves strike the coast Cliffs are steep, often vertical, slopes which rise abruptly
(an angle around 30° being the most effective). The effects from the sea or basal platform. Where they rise sheer from
of longshore drift are clearly illustrated by the accumula- deep water they are termed plunging cliffs. Such cliffs are
tion of sand against groynes along a beach and this in fact formed by fault scarps, volcanic masses and drowned gla-
provides a means of measuring the rate of longshore sedi- ciated valleys and because they are affected largely by
ment movement. swell waves, which are efficiently reflected, they suffer
Longshore sediment transport occurs below the breaker minimal erosion by wave action. Much more commonly the
zone where wave steepness is high, or by beach drift where water depth is shallow enough for waves to break and there-
wave steepness is low. Beach drift involves the movement fore most cliffs are subject to basal erosion by wave action
of sediment obliquely up the beach as a result of the angle· in addition to sUD-aerial weathering and mass movement.
326 Exogenic processes and landforms

B
Fig.13.15 Shore platform cut in horizontally bedded sandstone,
near Sydney, Australia.

Intermittent tected since wave energy is expended in crossing the plat-

undercutting
- - - Siow-attrfilon-- form. Wave action gradually becomes less significant in
undermining the cliff, and other processes such as salt wea-
thering, solution, frost weathering and the erosive effects of
shoreline organisms become more significant. Especially in
environments in the tropics where wave energy is low, the
combination of organic and inorganic weathering processes
is relatively more significant than quarrying and abrasion in
cliff development. Well-developed basal notches are charac-
teristic of cliffs along tropical coasts and these seem to
develop through water-layer weathering operating within a
narrow tidal range. Outside the tropics cliff notches are less
well developed, partly because the typically greater tidal
range means that wave attack is not concentrated in a
narrow vertical zone.
Estimates of the rates of cliff retreat show that most
Fig.13.14 Major processes of cliff retreat: (A) undercutting shore platforms cannot have formed in their entirety at
with rapid removal of collapsed material; (B) undercutting with
slow removal of collapsed material; (C) mass movement and present sea levels which have persisted for only about the
removal of material at various rates. (After 1. L. Davies (1980) last 4000 a. Since many shore platforms exist near to
Geographical Variation in Coastal Development. (2nd edn) present sea levels it can be deduced that they must have
Longman, London Fig. 49, p. 76.) developed when cliff retreat was occurring during earlier
phases of sea level in the Pleistocene close to that of today.
Cliff morphology is partly determined by the balance Since a constant sea level would eventually produce a shore
between the efficacy of sub-aerial processes and the rate of platform sufficiently extensive to absorb virtually all the
basal erosion (Fig. 13.14). Where wave action is capable of incoming wave energy, periodic rises in sea level are prob-
removing all the debris accumulating at the base of the ably necessary to maintain wave attack on the cliff base and
cliff, lithological and structural controls become s.ignificant thereby sustain cliff retreat and shore platform extension.
in influencing cliff form. The development of long free
faces, up to 500 m or more in height, is favoured in mech- 13.4.1.2 Shore platforms
anically strong, horizontally bedded sedimentary rocks and Three categories of shore platform can be identified related
coarsely jointed igneous rocks. On poorly consolidated to contrasing tidal stages and subject to rather different
materials with a low threshold angle of stability true cliffs combinations of processes. Intertidal platforms, extending
may fail to develop even if wave action is aggressive. between high and low water mark, are formed primarily by
A platform at the level of wave action, known as a shore quarrying and abrasion and are often referred to as wave-
platform, is developed as the cliff retreats under the com- cut platforms. They are most common where wave energy
bined effects of quarrying and abrasion (Fig. 13.15). As this is high and are best developed where easily eroded strata
platform is extended the base of the cliff starts to be pro- are exposed. Once initially formed by cliff retreat, intertidal
 Coastal processes and landforms 327

platforms are subject to modification by weathering and 'thering or by the modification of pre-existing platforms.
abrasion. 'Low-tide platforms are especially associated with lime-
As already mentioned, present-day ero ional activity is stones and other calcareous strata. Solution and biological
probably only extending features developed during periods erosion appear to be the key formative processes and they
in the recent past when sea level was near to that of the are typical of microtidal, low wave energy environments.
present. Each new episode of wave attack probably has to They may be partly constructional features where, for
begin with the removal of superficial material deposited on instance, carbonate-secreting calcareous algae are active.
the platform during the intervening periods of low sea The action of algae seems to be important in developing the
level. Unles there is a gradual rise in sea level, shore plat- basal cliff notches typically found on their landward mar-
forms can only extend themselves through cliff retreat by gins, but calcareous algae and other organisms contribute to
being progressively lowered seaward, otherwise sufficient their development through a combination of erosion and
wave action cannot be sustained at the cliff base. Indeed deposition. Outside the tropics biological activity is less
most intertidal shore platforms do exhibit a slight seaward significant in their development and high-tide platforms,
slope termed an abrasion ramp. As gradients are typically formed on hard limestones along mid-latitude and high
only around 10 , and as abrasion by waves is under most latitude coasts appear more directly attributable to solution.
conditions limited to a water depth of about 10 m, shore
platforms can only extend themselves in this way to a width
13.4.2 Constructional forms
of about 500 m in microtidal environments and perhaps up
to 1000 m in macrotidal environments. Any more extensive 13.4.2.1 Beaches
development than this would require a continuously rising Beaches constitute quantitatively the most significant accu-
sea level. Platform gradients tend to be less in microtidal mulations of sub-aerially exposed sediment along coasts.
than macrotidal environments, since in the former the ver- We have already considered some aspects of the supp~y and
tical range of wave action is much smaller. movement of sediment to beaches and will now look at
Shore platforms more extensive "than those just discussed beach form and process in more detail. Figure 13.16 pro-
must have formed primarily through mechanisms other vides a useful guide to the main features and processes to
than quarrying and abrasion. Those developed at approxi- be discussed.
mately high-tide level appear to be mainly attributable to Although beaches can be composed of material ranging
water-layer weathering. They seem to be best developed in from fine sand to boulders, most consist of sand or pebbles,
permeable rocks along coasts experiencing low to moderate intermediate particle sizes being rare. The characteristics of
wave energies, microtidal regimes and high evaporation local lithology can account for some of this variation in
rates. A low tidal range produces a well-defined upper zone sediment calibre, but it does not explain the prevalence of
experiencing regular saturation. Such high-tide platforms pebble beaches in the mid-latitudes and high latitudes, and
may develop through cliff retreat resulting from basal wea- of sand beaches along tropical coasts. Likely reasons for

WATER Oscillatory Wave Waves of Translation (bores): Longshore Collision Swash; Backwash
MOTION Waves Collapse Currents: Seaward Return Flow: Rip Currents Wind

DYNAMIC
Offshore Breaker Surf Transition Swash Berm Crest
ZONE

PROFILE

SEDIMENT Coarser ..... Coarsest ~ C AA ~lBi-modal """"-- C Wind-winnowed

SIZETRENDS - Grains oarser Lag deposi -..-- oarser - lag deposit
Accretion
PREDOMINANT
ACTION
Accretion Erosion
ITransportation
I Erosion and
Erosion

SORTING I+- Better - Poor Mixed Poor - Better ---+

ENERGY - Increase..... High - - Gradient---. High

Fig. 13.16 Major processes and characteristics of the nearshore zone along a sandy coast. Zones of maximum sand transport are
indicated by shading. (After 1. C. Ingle, 1966, The Movement of Beach Sand, Elsevier, Amsterdam, Fig. 116, p. 181).
328 Exogenic processes and landforms

Fig.13.17 Beachrock on the island ofAitutaki, Cook Islands, south-west Pacific.

this contrast include the greater predominance of fine mate- The profile form of beaches is determined by the size,
rial in fluvial sediments and the smaller relative contribution shape and composition of beach material, the tidal range
of material eroded from cliffs to littoral deposits in the and the type and characteristics of incoming waves. The
tropics. In contrast there is an abundance of coarse material upper section of many beaches consists of a horizontal to
in glacial and periglacial debris in coastal environments at slightly landward sloping surface called a berm (Fig. 13.16).
higher latitudes, especially in the northern hemisphere. This feature is not ubiquitous and is usually absent on
Beaches are composed of a variety of organic and inorganic gravel and pebble beaches. A berm is a zone of vertical
particles, but a major distinction can be made between car- accretion formed by backwash deposition and its elevation
bonate and non-carbonate beaches. Carbonate beaches were is therefore limited by the upper limit of swash.
long thought to be preferentially developed in the tropics On the seaward side of the berm, and separated from it
where there is a greater abundance of organic carbonate by the berm crest, is the beach face. Its gradient is largely
remains. However, it is now realized that many mid-lati- determined by the calibre of the beach material, and ranges
tude beaches have a comparatively high content of biogenic from around 2° in fine sand to as high as 20° in coarse
carbonate. The mineralogical composition of a particular pebbles. This can be accounted for by the greater permeabi-
beach will represent a complex balance between local sedi- lity of coarser sediments. Since swash percolates more
ment sources, material imported from inland by rivers and rapidly as it crosses a pebble beach than it does across a sand
material supplied by longshore drift or from offshore. beach, less water is available in the backwash to transport
Under certain conditions beach sediments can be lithified coarser material back down the beach face. A steeper slope
through precipitation of calcium carbonate to form beach- therefore develops to compensate for this. The importance
rock (Fig. 13.17). This is most commonly formed on tropical of percolation rates in influencing beach gradients is under-
beaches where calcium carbonate apparently upplied mainly lined by the observation that lower gradients are associated
from ground water is precipitated as a result of high rates of with poorly sorted beach sediments (which are more close-
evaporation. Micro-organi ms may also be involved in the ly packed and therefore less permeable), whereas high
precipitation of calcium carbonate. gradients characterize beaches with sediments of the same
 Coastal processes and landforms 329

mean size but better sorting. Wave height and steepness are slight irregularities in form along the beach. These are
also correlated with beach face slope angle, with teeper subsequently enlarged by swash with the eroded material
and higher waves being associated with gentler gradients as being deposited on either side by the backwash until an
they generate more backwash. equilibrium form is attained and the sediment moves in a
Where the beach gradient is shallow there is typically a closed cycle. Such movement has been documented with
submerged longshore bar running parallel with the beach the aid of dyed pebbles.
and separated from it by a trough. Several bars may be pre- Various other mechanisms contributing to cusp forma-
sent offshore from beaches with very shallow gradients. tion have been suggested including seasonal changes in wave
They seem to develop in response to the action of breaking energy and the breaching of dune ridges or berms, but more
waves and apparently migrate to and fro normal to the recently the role of standing edge waves with periods equal
shoreline as a result of chan"ges in wave characteristics. to, or twice that, of the incident waves has received atten-
The beach profile i not static but rather changes its form tion. The wide range of factors and mechanisms briefly
over a range of time scales, the most significant of which, at mentioned here indicate that we are still some way from a
least in the high energy storm wave environments of the comprehensive and detailed explanation of these intriguing
mid-latitudes, is a yearly cycle of erosion and deposition. features.
We know a lot about such annual changes because they have Significantly larger than cusps are sand waves and
been extensively monitored by repeated beach profile sur- crescentic bars. They are submerged features with a long-
veys. The predominance of swell waves in the summer is shore wave length usually around 200-300 m. They are
a sociated with deposition from swash and a phase of beach found off both straight beaches and along embayed coasts,
construction, often with the development of a berm just but are confined to microtidal environments. Standing edge
above high-water mark. During the winter, however, storm waves with a wavelength twice that of the crescentic bar
waves cut back or completely destroy the berm. The sedi- have been seen as being instrumental in their formation,
ment eroded is transported just offshore and deposited in with cell circulation and rip currents contributing to sediment
longshore bars, before being gradually returned to the beach transport. Longshore drift also seems to playa role in their
as the berm is constructed again the following summer. Wave development, and in this they differ from beach cusps.
steepness is apparently the critical factor in determining At a large scale many coasts display a more or less
whether a swell or a storm wave beach profile develops. regular succession of capes and bays. At some localities,
Since the berm is a major dissipator of wave energy, its along part of the east coast of Australia for instance, there
destruction by storm waves has a major impact on rates of exists an almost continuous series of asymmetrically curved
coastal erosion as it can lead to wave attack on backing bays linking each headland, with each beach section recessed
cliffs. behind its neighbour. They have been called headland bay
As well as characteristic profile forms, beaches also beaches, but the term zetaform beaches is also used
exhibit regularities in plan ranging in cale from 1 m or so because of their similarity in shape to the Greek letter (Fig.
to features over 100 km in length. These rhythmic features 13.3). The form of the Australian examples seems to be
involve a consistent repetition in form along the beach. The related to the incident angle of wave crests refracted around
smallest significant regularities are beach cusps. These are the headlands from the prevailing south-east swell. Beach
crescent-shaped indentations lying parallel with the shore steepness and sediment size increase northward in response
on the upper beach face and along the seaward margin of to increasing wave energy and the impeding of longshore
the berm. Although not confined to beaches with a particular transport of sediment by the headlands. Not all zetaform
calibre of sediment they are best developed in a mixture of beaches are, however, separated by rocky headland. Along
sand and gravel. coasts composed solely of unconsolidated sediments, such
In spite of their simple form there is little agreement a as the west coast of Sri Lanka and the eastern coast of the
to the origin of beach cusps, several factors seemingly con- Malay peninsula, river mouths may take the place of head-
tributing to their development. They form most readily lands in separating individual beaches.
where waves approach the coast perpendicularly and conse- Along some coasts low, linear beach ridges composed of
quently where longshore drift is minimal. Tides have also sand or shell debris surrounded by low-lying marshes are
been suggested as an instrumental factor; under high tidal found. These features are known as cheniers and were first
ranges they frequently form a serie of ridges down the described from the Louisiana coast of the USA. They can
beach but they have also been seen to develop along tide- be up to 1 km across, 100 km long and 4 m high. Coastal
less lake shores. It has been observed that cusp spacing is plains formed by alternating ridges and marshes are termed
positively correlated with wave height, but a stronger rela- chenier plains and have been recorded from a number of
tionship exists with swash distance - the distance between coasts including those of eastern China, Australia and New
the point at which waves break and the highest point on the Zealand. Their precise origin is uncertain although it is
beach reached by the swash. Cusps may be initiated by generally agreed that variations in sediment supply are
330 Exogenic processes and landforms

important in their development. Some workers consider

that the most significant factor controlling their formation is
rising sea level, but research in China has demonstrated that
the cheniers found there have developed since present sea
levels were more or less stabilized. Apart from a variable
supply of predominantly fine-grained sediment, cheniers
appear to develop preferentially where wave energies and
tidal ranges are low and longshore drift is active.

13.4.2.2 Barrier island Orlando

A barrier island (also called a barrier beach, barrier bar o
or offshore bar) is an elongated offshore ridge of sand
running parallel with a mainland coast and separated from
it for almost its entire length by a lagoon. Barrier islands
range in size from comparatively small features a few
metres in width and hundreds of metres long to massive
constructional forms up to 1 km or so wide, hundreds of
kilometres long and, where large dune systems are devel-
oped, up to 100 m high. It has been estimated that barrier
islands occur along 13 per cent of the world's coastline so
they are truly landforms of global significance. They are
most characteristic of very gentle sloping coasts with a
low tidal range and are developed particularly extensively
along the Atlantic and Gulf of Mexico coasts of the USA
(Fig. 13.18).
Tidal inlets are found at intervals along the barrier. They
are initially formed by storm breaching and subsequently
maintained by the ebb and flow of tidal currents and the
escape of river discharge entering the lagoon from its main-
land side. They migrate along the barrier as longshore drift
leads to removal of sediment from the updrift side and its
deposition on the downdrift side. After a certain amount of
longshore movement a new breach may form in the wake
of the migrating inlet and it too begins its longshore migra-
tion. Very large quantities of sediment are moved daily by
tidal currents and there are continuous fluctuations in sedi-
ment deposition and erosion on different sections of the
Fig.13.18 Barrier islands along the Atlantic coast of Florida.
barrier. USA.
The formation of barrier islands continues to be a contro-
versial topic, although data from boreholes have recently An alternative view is that barrier islands initially devel-
provided vital information on the-history of their develop- oped above sea level on beach ridges and coastal dunes,
ment and led to a revision of earlier ideas. Many barrier and were subsequently separated from the mainland by
islands were being formed prior to the Holocene rise in sea flooding and lagoon formation as sea level rose early in the
level, so clearly sea-level changes have played a role in Holocene. There is certainly a lack of evidence for beach
their construction. Some researchers envisage barrier islands processes in lagoonal sediments which would be expected
beginning as submerged offshore bars, or forming on earlier if barrier development had been initiated from submerged
foundations during the low sea levels of the Late Pleistocene. features. Yet another explanation sees barrier growth devel-
These were then built up by vertical accretion as sea level opment occurring during the past 4000 a of more or less
rose throughout the early part of the Holocene and large stable sea level. Each of the models of barrier island forma-
quantities of newly submerged sediment were transported tion may be relevant in particular cases, but the extent to
landwards. The present-day progressive erosion of some which barrier islands may be relict features out of equilibrium
barrier islands is seen as a consequence of the reduction in with prevailing environmental conditions adds greatly to
this sediment supply with the cessation of the major phase the difficulties of relating their development to specific
of sea level rise around 4000 a BP. processes.
 Coastal processes and landforms 331

13.4.2.3 Spits, baymouth bars and cuspate forelands I ...- - - - - T i d a l flat-----~~;

I I
Spits are elongated depositional forms attached at one end I
I Intertidal
I
I
to the mainland and usually developed where the coast :High tide slopes :High water
changes direction. Baymouth bars are similar features except : flat : level

that they extend into a bay rather than the open ocean.
I
I
I
z
Subtidal
one
: I

Low water
Longshore drift is the main process involved in their form- lev'el
ation, the transported sediment being deposited where it I
I
enters a zone of slack water. In addition to elongation they
may be widened by deposition on their offshore margin. If
baymouth bars continue to extend they may eventually form
a lagoon by totally enclosing a bay.
A feature of many spits as well as some barrier islands is
their landward curvature at their accreting terminus (Fig.
13.13). Two factors may be important in generating these Fig.13.19 Relationship of low- and high-tide positions to the
recurved forms. First, it has been seen simply as a normal morphological elements aftidalflats. (After J. L. Davies (1980)
consequence of spit development, resulting from the refrac- Geographical Variation in Coastal Development. Longman,
tion of constructional waves around the accreting terminus London (2nd edn) Fig. 123,p.170.)
of the spit and the consequent landward movement of sedi-
ment supplied by longshore drift at this point. This mech- between the high tide flat and the lower tidal limit; and (3)
. anism accords with the lack of recurvature in spits developed a subtidal zone which is submerged even at low tide.
in shallow, sheltered localities and its preferential develop- Clay and fine silt-sized sediments carried to the coast by
ment in macrotidal environments in which there is deep rivers tend to flocculate into larger aggregates when they
water in the high tide constructional phase. Indeed some encounter salt water. These particles will tend to settle in
spits have a concave plan form with respect to the mainland quiet coastal waters such as lagoons and sheltered estuaries.
at low tide and only assume their convex recurved form at This mud is brought in by the incoming tide and is deposited
high tide. A second possible cause of spit recurvature may before the tide reverses. If vertical accretion continues, part
arise from the effects of occasional periods of incident waves of the tidal flat will be exposed just above normal high tide
from a different direction than normal which modify the level. Once this occurs the mud is colonized by halophytic
form of the actively accreting terminus of the spit. This is plants and a salt marsh develops or, in the case of the tropics,
particularly likely to occur in exposed situations. Where commonly a mangrove swamp. Where both occur man-
there is a very constant undirectional swell wave regime, groves are characteristically concentrated on the offshore
concave rather than convex spits can develop, and this edge of the high tide flat where a rim with a slightly higher
situation is exemplified by the southern coast of Australia. elevation is sometimes developed.
Cuspate forelands are sediment accumulations which The role of root systems in aiding sediment deposition is
are roughly triangular in plan form and have their apex uncertain since colonization by vegetation may respond to
extending offshore. Some seem to be related to special con- spatial and temporal variations in sedimentation brought
ditions of funnelling of tidal currents, while others, such as about by channel shifting and other mechanisms. Continued
Dungeness in southern England, appear to have begun as a vertical accretion is limited by the reach of the highest
spit. In the case of Dungeness the spit extended through an tides, a nearly level equilibrium surface developing where
eastwards longshore drift into the path of storm waves from the addition of organic debris is more or less balanced by
the north. Its location in the narrow English Channel appears compaction and the decay of organic material. The high
to be crucial as constructional waves come from opposite tide flat is comparatively stable in comparison with the
ends of the Channel while cross-Channel waves have a very intertidal flat which is uswilly subject to successive episodes
small fetch and little power to blunt the apex. of erosion and deposition and provides an insufficiently
stable stratum for significant vegetation growth to occur.
13.4.2.4 Tidal flats
Tidal flats are largely depositional forms composed of 13.4.2.5 Deltas
muddy sediments and characteristically formed in lagoons Deltas are protuberances extending out from shorelines
and tidal estuaries. There is some confusion over the termi- formed where rivers enter the ocean, partially enclosed seas,
nology used to describe their various features but three barrier-sheltered lagoons or lakes, and supply sediment
basic units can be recognized (Fig. 13.19): (1) the high-tide more rapidly than it can be redistributed by coastal pro-
flat which has a very gently sloping surface, some of which cesses (Fig. 13.20). They are normally confined to regions
is partly submerged at high tide; (2) the intertidal slope, a with mature drainage systems where major river chaFmels
more steeply inclined though still gently sloping zone reach the coast and supply sediment to a limited area; in
332 Exogenic processes and landforms

Fig.13ol0 Landsat image of the Mississippi deLta, USA. The area shown is about 180 km across and covers not onLy the present deLta,
which can be seen surrounded by a Light-coLoured haLo of suspended sediment, but aLso some deLtas now abandoned (see Figure 13.24).
The deLicate tracery of distributary channeLs in the presentLy active deLta shows the overwhelming dominance offluviaL over tidaL and
wave processes. (Landsat image courtesy of N. M. Short.)

areas of less organized drainage many small river systems river water is more dense, and hypopycnal where it is less
supplying sediment to the littoral zone cause a more or less dense (Fig. 13.21).
uniform progradation of the entire coastal plain. Deltas form A number of attempts have been made to categorize the
because when a river enters a lake or the sea the outflow of broad range of delta types (Table 13.4) in terms of the major
river water expands and decelerates and its load is depo- processes controlling their development. A continuous
sited. The deposition of bed load occurs rapidly but the spectrum of delta forms is encountered in nature, and one
behaviour of the suspended load depends significantly on way of taking this variation into account is to categorize
the relative densities of the river, and sea or lake water. deltas using a ternary diagram in terms of fluvial, wave and
Three situations exist; homopycnal flow occurs where both tidal controls (Fig. 13.22).
waters are of equal density; hyperpycnal flow where the The location and configuration of the receiving water
 Coastal processes and landforms 333

Basin lead to the erosion of submarine channels on the delta front.

A Rapid CJ'J'--" '")"'
Finally, hypopycnalflow (C) occurs where the river waters are
deposition (\ J '--"' )
of bed . ~ less dense than those of the receiving basin. This is generally the
load '-...., ~ ---- case where river waters mix with sea water which is denser by
f - - - - -.........::.::
_ .. ,~ Thorough I
f--------'---, . ~ . '. ~ ..... mixi ng ./ virtue of its salt content. The buoyant river water spreads out
Density of
inflow and'
£)""
Axial jet
flow /
0J near the swface as a jet or plume and may travel a considerable
distance out to sea before losing its identity. (After W. L. Fisher et
basin are ~ J " t .J aI. (1969) Delta Systems in the Exploration for Oil and Gas.
Bureau ofEconomic Geology, University ofTexas, Austin, Figs.
equal Plan view
2-4, based partly on C. C. Bates (1953) American Association of
Petroleum Geologists Bulletin 37, Fig. 3, p. 2124.)

Axial cross-section

body influences the wave, tidal and associated current

processes. The morphology of deltas located on lakes or in
B
lagoons sheltered by barrier islands is dominated by fluvial
effects. In partially enclosed seas, such as the Gulf of Mexico
and the Mediterranean, wave energy is reduced because of
the limited fetch, and tidal effects are minimal. Consequent-
ly, delta located on the shores of such seas experience only
limited modification by coastal processes. By contrast, a
delta at the head of a narrowing gulf, such as the Ganges-
Brahmaputra delta at the head of the Bay of Bengal, may
experience a large tidal range which leads to the redistribu-
tion of great quantities of sediment. The intensity of fluvial
and coastal processes can vary through time, and if a sea-
Axial cross-section sonal fluctuation of sediment supply is out of phase with a
seasonal variation in wave and tidal energy then periods of
delta growth will alternate with periods of erosion.
c Basin
In essence, delta morphology and sedimentary structure
reflect the net 'effect of fluvial sediment supply and shore-
line sediment dispersal. The variables controlling delta struc-
ture and morphology relate to the nature of both the receiving
water body and the sediment supplied by the drainage
system (Table 13.5). Drainage basin characteristics influence
the fluvial regime and the nature of the sediment it trans-
ports to the delta, while the tidal and wave energy along the
shoreline affects the efficacy of sediment removal from the
delta area. Climate contributes to both the characteristics of
the drainage basin supplying sediment as well as the delta
environment itself. Tectonic controls are of particular signi-
Axial cross-section ficance, especially for marine deltas, because deltaic depo-
sition occurs at a land-water boundary which is very sensitive
Fig. 13.21 Various modes of mixing between sediment-laden to relative changes in sea level. A delta developing on a
river waters and receiving water bodies determined by relative stable platform, a common situation for small lake deltas,
water density. Homopycnalflow (A) occurs where both waters builds both upward and outward. But where the site of delta
are of equal density. There is an immediate vertical and
horizontal mixing which promotes rapid sediment deposition. This
formation is subsiding relative to sea level as a result of
type offlow is very common in lakes and results in a simple delta active tectonic subsidence, the rapid compaction of delta
structure consisting of bottomset beds formed from bed load, sediments or a sea level rise, a delta progrades by a process
fore set beds comprising both bed load and suspended load and of lateral displacement of sediment lobes. Because most
topset beds formed from overbank deposits and levees. coastlines have experienced a relative rise in sea level
Hyperpycnalflow (B) occurs where the river waters are more
during the Holocene, delta development in situations where
dense than the receiving water body and results in the creation of
high density basal currents which carry sediment beyond the relative sea level is falling is comparatively'flu-e.
shore zone and thereby restrict delta development. This can take Deltas consist of two fundamental components: a delta
place when very cold river water enters a warmer lake' and may front comprising the shoreline and the gently sloping sub-
334 Exogenic processes and landforms

Table 13.4 Characteristic delta morphologies and properties

DELTA COASTLINE AND RIVER MOUTH DELTA FRONT FEATURES DELTA PLAIN FEATURES
CONFIGURATION

Mississippi (1) Highly indented coastline, Indented marsh coastline, sand Marsh, open and closed bays
multiple extended digitate beaches scarce and poorly
distributaries - 'bird foot' developed
Danube (2) Slightly indented with Marsh coastline with sand beaches Marsh, lakes and abandoned
protruding river mouths adjacent to river mouths beach ridges
Ebro (3) Smooth shoreline with single Low sand beaches and extensive Salt marsh with a few beach
protruding river channel spits with some aeolian dunes ridges
Niger (4) Smooth, arcuate shoreline, Sand beaches nearly continuous Marsh, mangrove swamp and
multiple river mouths along shoreline beach ridges
slightly protruding
Nile (5) Gently arcuate, smooth Broad, high sand beaches and Floodplain with abandoned
shoreline with two slightly barrier formation with aeolian channels an"d a few beach ridges;
protruding distributary dunes; beach ridges at distributary hypersaline flats and barrier
mouths mouths lagoons near present shoreline
Sao Francisco (6) Straight, sandy shoreline High, broad sand beaches with Large linear beach ridges and
with single slightly constricted large aeolian dunes dunes
river mouth
Senegal (7) Straight coastline with extensive High, broad sand beaches with Large linear beach ridges and
barrier deflecting river mouth" large aeolian dunes swales, aeolian dunes
but no protrusion

Note: The numbers refer to delta types shown in Figure 13.22.

Source: After L. D. Wright and J. M: Coleman (1973) American Association of Petroleum Geologists Bulletin 57, Table 1, p. 376.

merged offshore zone, and a delta plain which forms an ances marking the location of channel mouths. In regions
extensive lowland area lying landward of the delta front characterized by mesotidal regimes tidal currents are active
and is made up of active and abandoned distributary channels. within the distributary mouths while wave processes
In some deltas, such as the Sao Francisco Brazil, there is operate along the remainder of the delta front. The delta
only a single channel, but it is much more common to find a front consequently comprises a series of cheniers separated
diverging set of channels fanning out across the delta plain. by tidal channels. In macrotidal settings tidal action is so
The area between these channels is occupied by flood- effective that the delta front is dominated by tidal current
plains, tidal flats, marshes and lakes. Delta plains are not ridges, channels and islands.
affected to any significant extent by wave action since they In deltas which form major protuberances, levees devel-
are protected by barrier island systems along wave-dominated oped along distributary channels may extend out into lakes
deltas. The detailed morphology of delta plains is signifi- or seas. These levees are generally lower than the water
cantly influenced by climate; mangrove swamps are common level attained in the channels during major floods and are
in humid tropical and sub-tropical deltas, dunes and shal- consequently fairly frequently breached. Once a breach has
low salt basins, or salinas, are frequently present in arid been made this may then provide a new route for the main
regions, while pingos, thaw lakes and other permafrost forms channel, especially if it is shorter and therefore has a steeper
are abundant in periglacial environments (Fig. 13.23). gradient than the existing main channel course. In the pre-
Delta front morphology is highly responsive to the sent Mississippi delta the Atchafalaya River is beginning to
relative effectiveness of fluvial, tidal and wave processes. divert a significant proportion of the flow from the
Delta front form is strongly influenced by fluvial action in Mississippi River, since from their point of bifurcation the
lakes because wave and tidal action is usually minimal. The course of the Mississippi to the Gulf of Mexico is well over
Mississippi is the only significant marine deita in which the twice as long (although this flow diversion is presently
delta front morphology is almost solely determined by the being limited by artificial channel controls). The steeper
mode of sediment deposition by the distributary channels gradient of the Atchafalaya is promoting a process common
with almost no subsequent modification by wave or tidal in deltas, that of channel switching, or avulsion, whereby
processes (Fig. 13.20). Where wave action is relatively the existing delta or delta lobe is abandoned and a new
more effective a relatively smooth arcuate or cuspate beach delta or lobe is initiated elsewhere. The sediment supply to
shoreline is developed as there is some redistribution of the previous delta gradually diminishes and eventually it
river-borne sediment. Dominant wave action is reflected in ceases to prograde while the diversion of sediment along
a very regular beach shoreline with only slight protuber- the new main distributary channel leads to the construction
 Coastal processes and landforms 335

Mississippi, USA Kutai, Borneo

o
I
40
I
Q
I
100I

km km

Ebro, Spain Fly, Papua New Guinea

km
DOMINATED

Wave Tidal
Processes Processes

Niger, Nigeria
o
L...I
100
~I
o
I
100
I

km km

Fig.13.22 Ternary diagram of delta types related to fluvial, wave and tidal controls. The numbers refer to the descriptions of delta
characteristics in Table 13.4. (Partly based on W. E. Galloway (1975) in M. L. Broussard (ed.) Deltas, Models for Exploration.
Houston Geological Society, Houston, Fig. 3, p. 92.)
336 Exogenic processes and landforms

Table 13.5 Factors influencing the morphology and edimentary characteristics of deltas

Quantity of suspended load and bed load

(stream capacity) increases during flood

Particle size of suspended load and bed load

(stream competence) increases during flood

Stream capacity diminishes during low

river stage

Stream competence diminishes during low

river stage

High wave energy with resulting turbulence

and currents erode, rework and winnow
deltaic sediments
High tidal range distributes wave energy
across an extended. littoral zone and creates
tidal currents
Strong littoral currents, generated by wave
and tide , transport sediment alongshore,
off hore, and inshore

Rigid basement precludes delta subsidence

and forces deltaic plain to build upward as
it progrades
Subsidence through downwarping coupled
with sediment compaction allows delta to
construct overlapping sedimentary lobes
as it progrades
Uplift of land (or lowering of sea level)
causes river distributaries to cut downward
and rework their sedimentary deposits

High temperature and humidity yield dense

vegetative cover, which aids in trapping
sediment transported by fluvial or tidal
currents
Seasonal character of vegetative growth is
less effective in sediment trapping; cool
winter temperature allows seasonal
.accumulation of plant debris to form delta
plain peat

Sparse vegetative cover plays minor role in

sediment trapping and allows significant
aeolian processes in deltaic plain
Sparse vegetative cover plays minor role in
ediment trapping; winter ice interrupts
fluvial processes: seasonal thaws and
aeolian processes influence sediment
transportation and deposition

Source: After J. P. Morgan (1970) in: J. P. Morgan (ed.) Deltaic Sedimentation: Modern and
Ancient, Society of Economic Paleontologists and Mineralogists Special Publication 15,
Table 1, p. 33.

of a new delta. This process is beautifully illustrated by the are known as crevasse-splay deposits. In well-vegetated deltas
sequence of abandoned deltas which lie to the east and west there is also a significant accumulation of organic debris in
of the modem Mississippi delta and which have formed these inter-distributary areas which may form thick sequences
over the past 7000 a (Fig. 13.24). of peat.
Breaches of delta distributary levees need not result in
the switching of main channels; more often they feed lakes 13.4.2.6 Coastal dunes
or bays on the delta plain known as crevasses and the deposits Coastal dunes are affected by the same basic aeolian pro-
they form which infill the area between extending levees cesses responsible for the formation of desert dunes (see
 Coastal processes and landforms 337

Fig.13.23 Part afthe delta of the Lena River on the arctic coast afthe USSR. River flow is restricted to a short summer season and
the delta plain is covered by landforms related to the presence ofpermafrost such as pingos (P) and thaw lakes (L). The image covers
an area about 180 km across. (Landsat image courtesy N. M. Short.)

Chapter 10) but additional factors must also be considered. and low humidity and preCIpItation and, of course, the
The key variable determining whether dunes will develop is availability of sediment capable of being moved by the
the rate of abstraction of sand from the beach. Other impor- prevailing winds. Tidal range will be crucial since a larger
tant variables are the effects of vegetation and the nature of area of beach sand will be exposed to aeolian action where
the backshore zone. there is a high tidal range, although this will also depend on
Whether sufficient quantities of sand will be blown land- the beach gradient. Tidal type is also important since semi-
ward from a beach to make dune formation possible will diurnal and mixed tides allow less time for beach sediments
depend on several factors. Inland aeolian transportation of to dry out and 'this renders them less susceptible to wind
beach sand will be favoured by high onshore wind speeds action (although tidal range outweighs this factor). The
338 Exogenic processes and landforms

rJ\\
ICJ
aquemines
Em Modern Delta
= St Bernard Delta
Complex
=:;I

ITIm Complex ~ Teche Delta Complex

Lafourche Delta
Complex EJ Maringouin Delta

o
I
100
!

km

@ I 16_ Plaquemines

2_
Modern
Mississippi River 13 Delta Complex

JBayou Lafourche 15_

Bayous Lafourche and Terrebonne 14
I Lafourche
Bayou Black Delta
I Complex
B~YoU Blue 10 I
6 Bayou Terrebonne

Jtii11
Bayou Sauvage 11

8 'SS'SS'PP' -~a Loutre StBernard

7 Bayou des Families Delta
Complex
5 _ Bayou Terre aux Boeufs

,-
31 Mississippi River 'and Bayou Lafourche

2_ ~~
I

Bayou Sale
I
Bayou Tec?e
I

Bayou Cypremort
Teche
Delta
Complex

Maringouin
Delta
8 7 6 5 4 3 2 o Complex
Age (ka B.P.)

Fig.13.24 Distribution of deltas and delta lobes at the outlet of the Mississippi River formed over the past 7000 a (A), and
chronology of individual phases of delta construction over this period (B). (After T. Elliott (1978) in H. G. Reading (ed.) Sedimentary
Environments and Facies. Blackwell, Oxford. Figs. 6.34A and 6.34B, pp.122 and 123, based on D. E. Frazier (1967) Transactions of
the Gulf-Coast Association of Geological Societies 17, Figs. 11 and 12, pp. 307 and 308.)

final significant control is sand supply. There is a clear poorly developed in the humid tropics. This is not surpris-
association between those coasts receiving an abundant ing since here fluvi,d sediments are typically poor in sand,
supply of sand from rivers and those with large coastal both humidity and precipitation are high and mean wind
dunes. Spectacular examples are the coasts of Oregon and velocities are low. Such a situation may not apply locally,
Washington in the USA which have massive dune systems as along the coast of Queensland, Australia, where there is
fed ultimately by fluvial sediments originating from high· an unusually abundant sand supply.
gradient, aggressively-eroding catchments. Primary dunes fed directly by beach sand can be cate-
A global perspective shows that dunes are preferentially gorized into free dunes and impeded dunes depending on
developed in the storm wave coastal environments of the whether vegetation has been important in initiating their
mid-latitudes, notably in north-west Europe and the north- formation (Table 13.6). Secondary dunes, which develop
west USA which are characterized by frequent high velocity through the erosion of impeded dunes, constitute a further
onshore winds. By comparison coastal dunes are generally major category. Free dunes are most abundant along desert
 Coastal processes and landforms 339

Table 13.6 Classification of coastal dune away from the beach (see Section 10.3.5.1).
More abundant globally are the impeded dunes charac-
TYPE FORMS ORIE TATION
teristic of humid and subhumid coasts. They take a variety
Primary dunes (sand derived from beach) of forms but most have a frontal dune running along the
Free dunes Transverse ridges Wind orientated and back of the beach. Their detailed form is significantly affect-
(vegetation barchans, oblique generally perpendicular ed by the type of colonizing plants instrumental in trapping
unimportant) ridges to direction of
constructional winds the sand. The eventual height reached by impeded dune i
Impeded dunes Frontal dunes, sand Nucleus orientated and essentially governed by wind velocity so that it is no
(vegetation beach ridges, dune generally parallel to rear surprise to find the highest dunes along the stormy west
important) platforms of source beach coasts of the mid-latitudes. Secondary dunes develop as a
Secondary dunes (sand derived from erosion of impeded dunes) consequence of dune erosion which can arise from a reduc-
Transgressive Blow-out dunes, Wind orientated and tion in sand supply or vegetation cover, or from an increase
dunes parabolic dunes, generally parallel to
longitudinal dunes, direction of con tructional in wind velocity. Some dunes are particularly sensitive to
transgressive sheets winds vegetation removal and blowouts and parabolic dunes can
Remnant dunes (eroded remnants of vegetated primary dunes) develop rapidly, initiating a new cycle of sand movement
Source: Adapted from J. L. Davies (1980) Geographical Variation in inland.
Coastal Development (2nd edn). Longman, London, p. 157. Dunes lithified by calcium carbonate are found on some
tropical coasts. They occur in rather similar environments
to beachrock, but their distribution is more extensive and
coasts, such as those of the Namib and Atacama Deserts reaches into sub-tropical latitudes. They are characteristic
and sections of Baja California, where vegetation is ex- of coasts experiencing a warm wet-dry climate in which
tremely sparse due to the hyper-aridity of these regions. there is sufficient moisture for CaC0 3 to be dissolved from
They also occur locally in more humid 'environments where the carbonate fraction of the sand, but sufficiently high eva-
species of sand-binding plants sufficiently vigorous to vege- poration during the dry season to promote its subsequent
tate and stabilize the dunes are absent. This is especially reprecipitation within the dune. Dunes subject to lithifica-
noticeable along the Washington and Oregon coast where tion do not have to be particularly carbonate-rich, lithifica-
native dune plants are in most localities incapable of estab- tion having been observed in dunes with as little as 8 per cent
lishing themselves in the midst of the very high rates of detrital carbonate. Lithified dunes are extensively devel-
sand movement. Free dunes are most commonly of the oped along the coasts of Australia, southern Africa and Brazil,
transverse type, but they may evolve into barchans as they among other areas, but direct climatic relationships are dif-
migrate inland as a result of the reduced sediment supply ficult to establish since many appear to be relict features.

... Coral reefs

Number of
coral genera
Equatorial Scale
o!
4000
I

km

Fig. 13.25 Global distribution of coral reefs and abundance of reef-building coral genera. The distribution of reefs is approximate
only as there are over 400 individual coral reefs. (Based largely on 1. L. Davies (1980) Geographical Variation in Coastal
Development, (2nd edn) Longman, London Fig. 45, pp. 66; and D. R. Stoddart (1969) Biological Reviews of the Cambridge
Philosophical Society 44, Fig. 16, p. 43.)
340 Exogenic processes and landforms

13.4.2.7 Reefs are to be found in Bird (1985), Clark (1984), Davis (1985)
Reefs are rocky, shallow-water, submarine ridges. Although and Komar (1983a) while Bird and Schwartz (1985)
they can be composed of inorganic rocks, biogenic forms provides a' global survey of coastal features and Fitzgerald
composed mainly of coral along with the remains of a wide and Rosen (1987) contains papers on glaciated coasts focus-
variety of other carbonate-secreting organisms are by far ing on those regions of North America affected by Quaternary
the most common. The environmental requirements of coral glaciation. Some recent ideas on aspects of coastal erosion
growth limit the global distribution of coral reefs broadly to are considered by Komar and Holman (1986), while Inman
the tropics (Fig. 13.25). A minimum sea temperature of and Nordstrom (1971) provide a stimulating tectonic frame-
about 18°C is required with optimum development occur- work within which to understand coastal morphology and
ring between 25 and 29 0c. Light requirements limit the Audley-Charles et ai. (1977) consider the location of deltas
depth at which corals grow to around 90 m, but the most in terms of tectonic setting.
vigorous development is confined to a depth of 20 m. A The generation of waves is discussed in Komar (1976),
normal ocean water salinity is required along with a mode- while Suhayda and Pettigrew (1977) look at the relation-
rate amount of water movement and a firm substrate on ship between wave height and velocity and Galvin (1968)
which the coral can establish itself. Coral growth is adver- examines the way in which beach gradient and wave height
sely affected by sediment deposition on the reef and this and wave length together influence the form of breaking
contributes to the comparative paucity of reef development waves. Davies (1980) considers the environmental controls
in the tropical Atlantic Ocean which receives abundant of tidal types and Komar (1976) provides a mathematical
sediment from several major river systems. treatment of the generation of tides' and a detailed discus-
Coral reefs occur in a variety of forms and positions with sion of currents.
respect to mainland coasts. Three major types can be identi- Turning to the processes operating in the coastal environ-
fied: fringing reefs which are separated from the mainland ment Robinson and Jerwood (1987) consider the role of
or island coast by a shallow channel; barrier reefs which frost and salt weathering while the idea of water-layer
are separated by a deep channel; and reefs forming islands. weathering is introduced by Wentworth (1938). Sediment
Fringing reefs are the most widespread type, but the other mobilization and movement in the swash and surf zone is
varieties are perhaps of more interest in view of their com- examined by Brenninkmeyer (1976) and Komar (1977), while
plex histories which reflect the interplay of sea-level longshore drift is discussed by Komar (1976). Various
changes, tectonic subsidence and coral growth (see Section aspects of the important role of organisms in coastal pro-
17.6.3). Indeed the present-day morphology of most coral cesses are assessed by Spencer (1988), Thom (1967) and.
reefs is a palimpsest of phases of construction and destruc- Stoddart (1969).
tion extending well back into the Pleistocene. The development of cliffs and shore platforms is con-
Most coral reefs now exposed above sea level have been sidered in a detailed treatment of rocky coasts by Trenhaille
subject to a range of geomorphic processes, with wave action (1987), while Emery and Kuhn (1982) look specifically at
and solution and other forms of weathering contributing to the nature and origin of cliffs and Bradley and Griggs (1976)
the formation of cliffs and shore platforms. Erosion provides and Trenhaille (1972, 1978) examine various features of
sediments which can be built into depositonal forms. Coral shore platform development. Komar (1976) examines the
reefs can also significantly influence coastal processes along relative importance of different sediment sources in the for-
the adjacent mainland, especially through the large amount mation of beaches, and Thorn (1968) points to the importance
of wave energy that they are capable of dissipating. Wave- of the Holocene sea-level rise in the translocation of sedi-
resistant species of coral are concentrated on the oceanward ment and the deveiopment of contemporary constructional
edge of reefs, while organic sediment eroded from this landforms. The formation of beachrock is discussed by
exposed margin is deposited in the calmer zone behind the Scoffin and Stoddart (1983). DuBois (1972) and McLean
reef. An intertidal reef flat or sandy lagoon may lie between and Kirk (1969) consider the relationship between sediment
the reef and island or mainland. Algal limestones are impor- size and sorting and beach gradient and Komar (1976)
tant in coating the very outer rim and contributing to reef discusses the role of wave height. General problems in the
stability. development of rhythmic topography are discussed by Komar
(1976, 1983b) while cusp formation is considered by Guza
and Inman (1975) and Dalrymple and Lanan (1976). The
Further reading characteristics and development of cheniers are examined
There are a number of texts which cover the field of coastal in general terms by Otvos and Price (1979) and with specific
geomorphology or large parts of it. Davies (1980) provides reference to China by Cangzi and Min (1987).
an excellent introduction with a global perspective, while There is a large literature on barrier islands; various views
Bird (1984), King (1972) and Pethick (1984) are also good concerning their development are given in Carter and Orford
starting points. More detailed treatments of specific topics (1984), Cooke (1968), Eddison (1983), Fisher (1968) and
 Coastal processes and landforms 341

Hoyt (1967, 1968), while the papers in Leatherman (1979) Dalrymple, R. A. and Lanan, G. A. (1976) Beach cusps formed by
and those in a special issue of the journal Marine Geology intersecting waves. Geological Society of America Bulletin 87,
(1985, vol. 63) give a good idea of the continuing points of 57-60.
Davies, J. L. (1980) Geographical Variation in Coastal Develop-
debate. The depo itional aspects of tidal flats are compre- ment (2nd edn). Longman, London and New York.
hensively discussed in Ginsburg (1975), while Thorn (1967) Davis, R. A. (ed.) (1985) Coastal Sedimentary Environments.
and Thorn et al. (1975) consider the relationships between Springer-Verlag, New York.
mangrove development and estuarine edimentation. Elliott DuBois, R. N. (1972) Inverse relation between foreshore slope
(1986) provides an excellent review of deltas with specific and mean grain size as a function of the heavy mineral content.
Geological Society ofAmerica Bulletin 79, 1421-6.
reference to their sedimentological characteristics, while Eddison, J. (1983) The evolution of the barrier beaches between
Morgan (1970a) provides a brief but useful introduction Fairlight and Hythe. Geographical Journal 149, 39-53.
and Morgan (1970b) a more detailed treatment of the topic. Elliott, T. (1986) Deltas. In: H. G. Reading (ed.) Sedimentary
The global variation in delta morphology in relation to Environments and Facies (2nd edn). Blackwell Scientific
environmental controls is examined by Coleman (1981), Publications, Oxford, 113-54.
Emery, K. O. and Kuhn, G. G. (1982) Sea cliffs: their processes,
Coleman and Wright (1975), Wright and Coleman (1973) profiles and classification. Geological Society of America Bulletin
and Galloway (1975), while Wright and Coleman (1974) 93,644-54.
look in detail at the development of the intensively studied Fisher, J. 1. (1968) Barrier island formation. Discussion. Geologi-
Mississippi delta system. cal Society of America Bulletin 79,1421-6.
Pye (1983) reviews the literature on coastal dunes, while Fitzgerald, D. M. and Rosen, P. S. (eds) (1987) Glaciated Coasts.
Academic Press, San Diego.
Pye and Bowman (1984) consider the impact of Holocene Galloway, W. E. (1975) Process framework for describing the
sea-level rise on their development. Guilcher (1988) pro- morphologic and stratigraphic evolution of the deltaic deposi-
vides a comprehensive coverage of the geomorphology of tional systems. In: M. L. Broussard (ed.) Deltas, Models for
coral reefs and Hopley (1982) presents a detailed study of Exploration. Houston Geological Society, Houston, 87-98.
Galvin, C. J. (1968) Breaker-type classification on three labora-
the Great Barrier Reef of Australia. Further references on
tory beaches. Journal of Geophysical Research 73,3651-9.
coral reef development are given at the end of Chapter 17. Ginsburg, R. N. (ed.) (1975) Tidal Deposits. Springer-Verlag,
Berlin.
Guilcher, A. (1988) Coral Reef Geomorphology. Wiley, Chichester
and New York.
References Guza, R. T. and Inman, D. L. (1975) Edge waves and beach cusps.
Audley-Charles, M. G., Curray, J. R. and Evans, G. (1977) Loca- Journal of Geophysical Research 80, 2997-3012.
tion of major deltas. Geology 5, 341-4. Hopley, D. (1982) The Geomorphology of the Great Barrier Reef:
Bird, E. C. F. (i984) Coasts: An Introduction to Coastal Geomorphol- Quaternary Development of Coral Reefs. Wiley, New York.
ogy. Blackwell, Oxford. Hoyt, J. H. (1967) Barrier island formation. Geological Society of
Bird, E. C. F. (1985) Coastline Changes: A Global Review. Wiley, America Bulletin 78, 1125-36.
Chichester and New York. Hoyt, 1. H. (1968) Barrier i land formation: reply. Geological
Bird, E. C. F. and Schwartz, M. L. (eds) (1985) The World's Society ofAmerica Bulletin 79, 947 and 1427-32.
Coastline. Van Nostrand Reinhold, New York. Inman, D. L. and Nordstrom, C. E. (1971) On the tectonic and
Bradley, W. C. and Griggs, G. B. (1976) Form, genesis, and morphologic classification of coasts. Journal of Geology 79,
deformation of central California wave-cut platforms. Geologi- 1-21.
cal Society of America Bulletin 87, 433-49. King, C. A. M. (1972) Beaches and Coasts. Edward Arnold,
Brenninkmeyer, B. (1976) Sand fountains in the surf zone. In: London.
R. A. Davis and R. L. Ethington (eds) Beach and Nearshore Komar, P. D. (1976) Beach Processes and Sedimentation.
Sedimentation. Society of Economic Paleontologists and Min- Prentice-Hall, Englewood Cliffs.
eralogists Special Publication 24, 69-81. Komar, P. D. (1977) Selective longshore transport rates of dif-
Cangzi, L. and Min, C. (1987) The chenier plains of China. In: V. ferent grain-size fractions within a beach. Journal of Sedimen-
Gardiner et al. (eds) International Geomorphology 1986: Pro- tary Petrology 47, 1444-53.
ceedings of the First International Conference on Geomorpho- Komar, P. D. (ed.) (1983a) CRC Handbook of Coastal Processes
logy Part I. Wiley, Chichester and New York, 1269-79. and Erosion. CRC Press, Boca Raton.
Carter, R. W. G. and Orford, J. D. (1984) Coarse clastic barrier Komar, P. D. (1983b) Rhythmic shoreline features and their
beaches: a discussion of their distinctive dynamic and morpho- origins. In: R. A. M. Gardner and H. Scoging (eds.) Mega-
sedimentary characteristics. Marine Geology 60, 377-84. Geomorphology. Clarendon Press, Oxford, 92-112.
Clark, M. W. (ed.) (1984) Coastal Research: UK Perspectives. Komar, P. D. and Holman R. A. (1986) Coastal processes and the
Geo-Books, Norwich. development of shoreline erosion. Annual Review of Earth and
Coleman, J. M. (1981) Deltas: Processes of Deposition and Models Planetary Sciences 14, 237-65.
for Exploration. Burgess, Minneapolis. Leatherman, S. P. (ed.) (1979) Barrier Islands: From the Gulf of
Coleman, 1. M. and Wright, L. D. (1975) Modern river deltas: St. Lawrence to the Gulf of Mexico. Academic Press, New
variability of processes and sand bodies. In: M. L. Broussard York.
(ed.) Deltas, Models for Exploration. Houston Geological Society, McLean, R. F. and Kirk, R. M. (1969) Relationship between grain-
Houston, 99-149. size, size-sorting and foreshore lope on mixed sand-shingle
Cooke, C. W. (1968) Barrier island formation: discussion. Geolo- beaches. New Zealand Journal of Geology and Geophysics 12,
gical Society ofAmerica Bulletin 79, 945. 138-55.
342 Exogenic processes and landforms

Morgan, J. P. (1970a) Deltas - a resume. Journal of Geological Suhayda, 1. N. and Pettigrew, N. R. (1977) Observations of wave
Education 18, 107-17. height and wave celerity in the surf zone. Journal of Geo-
Morgan, J. P. (ed.) (1970b) Deltaic Sedimentation; Modern and physical Research 82, 1419-24.
Ancient. Society of Economic Paleontologists and Mineralogists Thom, B. G. (1967) Mangrove ecology and deltaic geomorpho-
Special Publication 15. logy: Tabasco, Mexico. Journal of Ecology 55,301--43.
Otvos, E. G. Jr and Price, W. A. (1979) Problems of chenier genesis Thom, B. G. (1968) Coastal erosion in eastern Australia. Australian
and terminology - an overview. Marine Geology 31, 251-63. Geographical Studies 6,171-3.
Pethick, J. S. (1984) An Introduction to Coastal Geomorphology. Thorn, B. G., Wright, L. D. and Coleman, J. M. (1975) Mangrove
Edward Arnold, London and Baltimore. ecology and deltaic-estuarine geomorphology: Cambridge-Ord
Pye, K. (1983) Coastal dunes. Progress in Physical Geography 7, River region, Western Australia. Journal of Ecology 63,
531-57. 203-32.
Pye, K. and Bowman, G. M. (1984) The Holocene marine trans- Trenhaille, A. S. (1972) The shore platforms of the Vale of
gression as a forcing function in episodic dune activity on the Glamorgan, Wales. Transactions of the Institute of British
eastern Australian coast. In: B. G. Thorn (ed.) Australian Geographers 56, 127--44.
Coastal Geomorphology. Academic Press, Sydney, 179-96. Trenhaille, A. S. (1978) The shore platforms of Gaspe, Quebec.
Robinson, D. A. and Jerwood, L. C. (1987) Frost and salt weather- Annals of the Association of American Geographers 68,
ing of chalk shore platforms near Brighton, Sussex, U.K. 95-114.
Transactions, Institute of British Geographers NS 12, 217-26. Trenhaille, A. S. (1987) The Geomorphology of Rock Coasts.
Scoffin, T. P. and Stoddart, D. R. (1983) Beachrock and intertidal Clarendon Press, Oxford.
cements. In: A. S. Goudie and K. Pye (eds) Chemical Sediments Wentworth, C. K. (1938) Marine bench forming processes: Part 1,
and Geomorphology: Precipitates and Residua in the Near- Water-level weathering. Journal of Geomorphology 1, 6-32.
Swjace Environment. Academic Press, London and New York, Wright, L. D. and Coleman, J. M. (1973) Variations in morpho-
401-25. logy of major river deltas as functions of ocean wave and river
Spencer, T. (1988) Limestone coastal morphology: the biological discharge regimes. American Association of Petroleum Geolo-
contribution. Progress in Physical Geography 12, 66-101. gists Bulletin 57,370-98.
Stoddart, D. R. (1969) Ecology and morphology of Recent coral Wright, L. D. and Coleman, J. M. (1974) Mississippi River mouth
reefs. Biological Reviews of the Cambridge Philosophical Society processes: effluent dynamics and morphologic development.
44,433-98. Journal ofGeology 82, 751-78.
This page intentionally left blank
 14
Climate, climatic change
and landform development

14.1 Climate and landform development Disagreement comes with the finer distinctions between
degrees of humidity and temperature evident in attempts to
identify morphoclimatic zones (see Section 1.5.3), although
The relationship of landforms to climate is a topic which recently there have been some rather more sophisticated
has simultaneously been regarded as a focus for research in efforts to cast climatic controls in terms of the magnitude
geomorphology and an approach which has fruitlessly and frequency of meteorological events. Although variations
occupied the time of geomophologists over the past several in both temperature and, more especially, precipitation can,
decades. The former view has characterized research by in some instances, be sensitively reflected in processes such
many French and German geomorphologists, whereas the as those involved in weathering and fluvial activity, when
latter has prevailed among Anglo-American researchers. landscapes as a whole are considered other factors related
The processes and landforms associated with fluvial, aeolian, to tectonics and lithology often predominate. For instance,
glacial and periglacial environments reviewed in earlier if we examine a range of humid tropical environments around
chapters clearly demonstrate that major differences in cli- the world, diversity of landscape is more apparent th3;n
mate have a profound effect on landscape development. uniformity (Table 14.1). A further problem in trying to

Table 14.1 Comparison of landscape characteristics in a range of humid tropical environments

FEATURE NORTHERN AUSTRALIA SRI LANKA MALAYSIA BORNEO NEW GUINEA HAWAII

Dominant age of landscape Old Old Moderately old Young Very young Very young
Proportion of labile rocks Low Low Moderate Moderate High Moderate
Recent tectonism Practically none Practically none Negligible Some Intense Infrequent
Cenozoic volcani m Practically Practically Practically Some Abundant Highly
absent absent absent abundant
Tropical karst Practically absent Virtually absent Residuals only Abundant Abundant None
Relief Low to moderate Moderate Moderate Moderate (save Very high High
for Kinabulu)
Rate of current regolith
development Slow Slow Slow to moderate Moderate Rapid Rapid
Rate of current erosion and. Low to Low to
deposition moderate moderate Moderate High Very high Very high
Depth of colluvium Generally Generally Variable Variable Variable Variable but
shallow shallow usually high
Ferricrete Widespread Widespread Fragment Possible Absent Absent
Pallid zone of deep weathering Widespread Widespread Few Unknown Rare Rare

Source: After I. Douglas (1978) in: C. Embleton, D. Brunsden and D. K. C. Jones (eds) Geomorphology: Present Problems and Future Prospects.
Oxford University Pre ,Oxford, Table IV, p. 175.
346 Exogenic processes and landforms

correlate relatively small differences in climate with particular

A
landforms or landscape types is the diversity of conditions
under which apparently similar forms can develop.
Perhaps the most intractable problem in climatic geomor-
phology is the unambiguous linking of a particular landform
with a specific set of climatic conditions. No landform better
illustrates this dilemma than the pediment (see Section
7.6.2). Pediments are gently inclined slopes with gradients
of between 2 and 10° extending from slopes cut in bedrock B
down to an area of sediment deposition at a lower elev-
ation. The backing slope usually has a gradient of more
than 20° and joins the pediment at a marked change in
slope known as the piedmont angle. Some pediments have
a concave-up long profile, and some have a partial veneer
of sediment. Pediments can terminate upslope at a moun-
tain front, or at a residual hill, known as an inselberg if Fig.14.1 Two types ofpediment: (A) associated with a
large, and a tor if small. sedimentary caprock or duricrust; (B) rock pediment.
Part of the confusion about the origin of pediments and
their climatic affinities arises from the term being applied
to two morphologically similar, but genetically distinct,
forms. One type is associated with a resistant cap rock is undermined and collapses. Although such forms are most
(either a sedimentary unit or a duricrust) overlying more evident in semi-arid environments there is evidence from
readily erodible material (Figs 14.1 (A) and 14.2). The many regions that slope retreat and pediment growth is
pediment forms through slope retreat and the rate at which most active in more humid climatic interludes. This seems
this occurs is determined by the rate at which the cap rock to be especially the case where the cap rock is permeable

Fig.14.2 Pediments extending from silcrete-capped sUliaces, near Stuart Creek, South Australia.
 -Climate, climatic change and landform development 347

Fig. 14.3 Granite bornhardt (domed inselberg) standing above a gravel-mantled rock pediment in the hyper-arid Namib Desert east of
Walvis Bay, Namibia.

since under humid conditions ground water emerges at the ancient forms which are essentially 'frozen' in their develop-
cap rock base and promotes undermining. ment in the prevailing arid climate.
Much more controversy is associated with the other type Evidence of saprolite formation in association with rock
of pediment, the rock pediment. In this form the pediment pediment development comes from localities as diverse
and the backing mountain range or inselberg are composed as the central Sahara, Australia and southern Arabia. In the
of the same rock type, usually a coarse-grained intrusive Mojave Desert of southern California, USA, the presence
rock such as granite (Figs 14.1 (B) and 14.3). Rock pedi- of rock varnish on inselberg slopes indicates that they are
ments are most evident in, though not confined to, arid inactive at present although surrounded by pediments several
environments. One model suggests that they form through kilometres in length. Deep weathering under a humid climate
parallel slope retreat under present climatic conditions, and is supported by the presence of tors and core stones, rego-
that the piedmont angle is attributable to the way granitic lith stripping probably occurring when the climate became
rocks break down into sand-sized material, or grus, without somewhat drier (but not arid). The piedmont angle is attri-
intermediate pebble and cobble-sized debris. Such small buted to enhanced physical and chemical weathering related
calibre material can be readily removed by sheet flow and to concentrated water flow at the base of the back slope
therefore does not accumulate at the slope base. There are, during and after regolith stripping. This interpretation is
however, two difficulties with this model as a general supported by palaeoclimatic evidence which indicates that
explanation of pediment formation. One is that many rock desiccation of the Mojave region at the end of the Miocene
pediments are associated with pockets of deeply weathered was linked to the creation of a rain-shadow caused by the
regolith or are even cut in saprolite, and this clearly implies uplift of the Sierra Nevada and Transverse Ranges to the
a more humid environment at some stage during their west. Prior to about 6 Ma ago the vegetation cover ranged
development. A second difficulty is that the dating of some from woodland to semi-desert in type, but since that time
rock pediment surfaces has demonstrated that they are the climate has been distinctlyinore arid and pediment
348 Exogenic processes and landforms

modification has been extremely slow, apparently involving most processes operate in most environments at least to
an average of less than 100 m extension over the past 5 Ma. some degree. In many cases we do not yet know enough
In the eastern Mojave Desert 60 basalt flows up to 4 Ma old about the development of specific landforms to be able to
partly bury pediment surfaces, and comparison of the rela- say with any confidence which are being actively formed
tive positions of pediments covered by dated basalts and the in a padicular climatic environment. The problem of distin-
modern pediment surfaces indicates the degree of modifi- guishing between active and relict forms can be particularly
cation that has occurred over this time interval. The conclu- acute in those environments where even 'active' geomor-
sion is the amount of modification has been minimal, with phic processes operate sporadically. This is the case in
downwearing predominating over backwearing. many arid and semi-arid environments where the irregular
Although the exhumation model of pediment formation, patterns of precipitation mean that fluvial activity is very
along with its requirement for a previously humid climate, limited in duration, but, none the less, often very important
seems to have widespread applicability there appear to be in shaping the landscape. The question here is how long
situations where it cannot apply because there is compel- does a landform have to be inactive for it to be described as;
ling evidence of prolonged aridity. One such case is provided relict?
by the pediments flanking the domed inselbergs, or The past two decades have seen a quantum leap in our
bornhardts, of the Namib Desert in southern Africa (Fig. understanding of climatic change largely as a result of the
14.3). Here the cold Benguela Current offshore in conjunc- exploration of the ocean floor, which contains a relatively
tion with the dominant sub-tropical high pressure system complete and undisturbed climatic record, and of the devel-
~as created a hyper-arid coastal desert climate which has opment of new dating procedures and techniques of palaeo-
probably prevailed since at least the mid-Miocene. Al- environmental reconstruction. Our knowledge of climatic
though the present rate of slope retreat seems to be extremely fluctuations on the continents, nevertheless, is still severely
slow there are no remnants of an earlier phase of significant constrained by a lack of palaeoclimatic indicators capable
rock weathering to support the idea of exhumation. Such of yielding datable material. In fact landforms themselves
pediment and inselberg forms represent one of the major have been extensively used to reconstruct changes in tem-
enigmas of geomorphology. perature, precipitation and wind intensity and direction on
the continents. The use of this kind of evidence actually
raises a problem for geomorphologists concerned with inter-
14.2 The significance of climatic change preting the landscape in terms of climatic change because
The presence of landforms, such as certain kinds of pediments, in many cases the climate changes themselves have been
in climatic environments in which they could not have interpreted largely from landform evidence. Clearly there is
developed obviously raises the question of climatic change a danger of circular arguments here and, if possible, we must
in landform interpretation. The state of the atmosphere varies use independent information on climatic change.
from day to day and even hour to hour, and it is average The consequences of climatic change for landform genesis
weather conditions together with their variability that charac- can, very broadly, be divided into those effects related pri-
terize the climate of a region. But climate also fluctuates at marily to temperature changes, and those related primarily
all time scales from decades to tens of millions of years. to changes in precipitation. The growth and decay of ice
Whether a climatic change is of geomorphic significance sheets is clearly influenced significantly by long-term changes
depends on its magnitude and duration, and on the proper- in temperature, whereas surface runoff and fluvial processes,
ties of the landform concerned. The larger or more resistant along with the level of aeolian activity, are highly responsive
a landform, the longer, in general, it takes to adjust to a to precipitation changes. We must be careful, however, in
change in climate. The 'sensitivity' of different landscapes making broad generalizations. For instance, although fre-
to climatic fluctuations is an important but complex ques- quently resulting from a temperature fall, glacier growth
tion which affects the way we conceive long-term landscape can also occur as a result of an increase in precipitation
development and we examine it in detail in Chapter 18 (see when temperatures are stable, or even increase. Moreover,
Section 18.2.3). an increase in surface runoff can arise from a decrease in
An important distinction is between 'active' landforms temperature (and hence rates of evapotranspiration) as well
which seem to be in 'equilibrium' with the prevailing climatic as an increase in precipitation. Nevertheless, it is convenient
environment, and inactive or relict landforms which could to separate the discussion of the expansion and contraction
not have been formed under the current climatic regime. As of glacial and periglacial morphoclimatic regimes, which
will have been evident from our discussion of the pediment are largely a response to temperature fluctuations, from the
problem, this distinction is not always easy to make. Al- predominant impact of precipitation changes on fluvial and
though we saw in Section 1.5.3 that it is possible to relate aeolian systems. Before we do this, however, it is necessary
broad associations of geomorphic processes to major climatic to review briefly our present understanding of the history of
zones, the overlap between categories is significant since global climatic change.
 Climate, climatic change and landform development 349

14.3 The record of climatic change compared with the present. Polar regions probably exper-
ienced a climate rather similar to mid-latitude and sub-
A wide range of techniques are used to reconstruct past cli-
tropical climates of the present day, although the amount of
mates and they are employed on diverse types of evidence.
solar radiation received and the seasonality of climate
For long-term climatic change information is acquired from
would have been similar to equivalent latitudes today.
the changing distributions of plants and animals determined
H~wever this widely accepted ice-free Cretaceous interpre-
from fossils, and from rock types which are attributable to
tatIOn has been challenged by researchers who have inter-
particular climatic regimes. Evaporites and aeolian sand-
preted large exotic blocks found within Early Cretaceous
stones, for instance, are indicative of arid environments
whereas tillites demonstrate glacial conditions. Approache~
mudstones in Australia as dropstones, that is, debris
dropped from floating ice. This evidence indicates that ice
to reconstructing more recent climatic change include the
was present at sea level in high latitudes during the
analysis of pollen (palypology) and landforms themselves.
Cretaceous. Similar interpretations have been applied to
One technique, however, has done more than any other
exotic blocks in strata of a variety of ages and it has been
over the past two decades or so to revolutionize our under-
argued that these indicate, contrary to conventional wisdom,
standing of climatic change. In the late 1940s it was pointed
that the Earth may only rarely have been ice free in the
out that the ratio of the two stable isotopes of oxygen, 160
past. These interpretations have yet to be confirmed by
and 18 0, precipitated in carbonate would vary with water
other evidence, and meanwhile it is perhaps best to adopt
temperature. This idea was applied in 1955 by C. Emiliani
the generally accepted ice-free Cretaceous model. Towards
to the analysis of calcium carbonate (CaC03) shells secreted
the end of the Cretaceous there is good evidence for a slight
by various sea creatures. The technique of oxygen isotope
cooling, but this trend was reversed during the Paleocene
analysis, as it was termed, was most often applied to fora-
and Eocene when a warm phase saw tropical soil formation
minifera, single-celled organisms mostly between 0.1 and
extending to 45° latitude in both hemispheres and tropical
0.3 mm across. Emiliani found that the 18 0/ 160 ratio varied
species living in southern England and western Greenland.
in foraminifera shells recovered from ocean cores recording
The major point of dispute in the climatic record for the
sedimentation over the past few hundred thousand years. Both
past 50 Ma is when the Antarctic ice sheet became estab-
planktonic (surface and near-surface) and benthic (bottom-
lished. One interpretation of the oxygen isotope data suggests
dwelling) forms of foraminifera occur so it was possible to
that it first appeared about 15 Ma ago, but if 180/ 160 ratio
estimate temperature changes on both the sea surface and in
variations are caused predominantly by ice volume changes
the ocean depths. The changes in oxygen isotope ratios were
dated using the palaeomagnetic time scale in combination then a date of around 35 Ma is indicated (Fig. 14.4). This
with an assumed constant rate of sedimentation. earlier date is supported by geomorphic and sedimento-
18 16 logic~l evidence, including glacially abraded grains in off-
It was subsequently appreciated that the 0/ 0 ratio of
shore sediments, which indicate the first appearance of sea-
the sea water itself would also vary as the quantity of land
level glaciation in East Antarctica at this time. By about
ice changed with fluctuations in global temperatures. Such
26 Ma ago the East Antarctic ice sheet was apparently more
changes would be reflected in the foraminifera and would
extensive than it is at the present day. An ice sheet in the
arise because the lower atomic mass of 160 means that it is
Early Cenozoic also helps to account for the record of sea-
preferentially evaporated from the oceans. During glacials a
level change during this period which is difficult to explain
~r~portion of this 160-enriched water would be locked up
without the changes in ocean water volume generated by
In Ice sheets thus leaving the oceans relatively enriched in
18 0. It is difficult to separate these two effects - one indica- periodic fluctuations in ice volume (see Section 17.5.4).
The causes of the major climatic changes involved in the
ting changes in ocean temperatures, the other fluctuations
establishment of the Antarctic ice sheet are beyond the
in the volume of ice sheets - and this has led to divergent
scope of this book, but continental drift and the drastic
views of climatic history, especially for the past 40 Ma.
modifications of ocean currents that it generated were cer-
tainly important factors. Recall that the supercontinents of
14.3.1 The Cretaceolls to Neogene record Laurasia and Gondwana broke up progressively throughout
There are probably few areas in the world where landforms the Cretaceous (see Figure 2.16). In the northern hemisphere a
or weathering deposits can be regarded as having survived temperate climate persisted until the mid-Miocene but
~ore or less intact for more than the past 100 Ma, although significant cooling can be traced back to at least 10 M~ ago
It has been suggested that even more ancient forms exist in and the Arctic Ocean became ice-covered by 5 Ma BP.
northern Australia. It is appropriate, therefore, to begin our Although some uncertainty remains, it appears that ice sheets
brief survey of climatic change with the Cretaceous. became established in the northern hemisphere around
The climate of the Late Cretaceous - 100 to 65 Ma BP - 3.2 Ma BP. This date is indicated by the appearance of ice-
is generally thought to have been warm and equable with a rafted debris in the North Atlantic and North Pacific in
very much less marked temperature gradient with latitude sediments of this age.
350 Exogenic processes and landforms

18 AGE
EPOCH 0 OCEAN RECORD Ma B.P. ICE SHEET EXTENT MAJOR EVENTS
PLEIST
PLIOCENE
0
- Large N. hemisphere ice sheets

MIOCENE
10
-- Mountain glaciers in N. hemisphere
Sealevel glaciation in W. Antarctica

20

30
- Extensive ice sheet in E. Antarctica

---
OLIGOCENE

~
,, Coastal glaciation, E.Antarctica
Mountain glaciation in Antarctica?
40 Sharp cooling of Southern Ocean

EOCENE

50
- Antarctic-Australian passage opens

PALAEOCENE
..
Increase 60 Increase
•
Relative Change in 180 Relative Extent
Concentration in Oceans of lee Sheets
Worldwide

Fig.14.4 History of the development of glaciation during the Cenozoic. Note that the variations in ice sheet extent are speculative due
to lack of data, and that the dates at which glacial activity began at different localities is also subject to revision in the light of new
information. (Based partly on D. E. Sugden (1987) in M. J. Clark, K. J. Gregory and A. M. Gurnell (eds) Horizons in Physical
Geography. Macmillan, Basingstoke, Fig. 3.2.3, p. 217.)

14.3.2 The Quaternary record

directly, global temperatures (Fig. 14.5). Keyage markers
By 3 to 2.5 Ma BP large-scale oscillations in temperature re- for this ocean core oxygen isotope record have been pro-
flecting glacial-interglacial cycles had become established, vided by the palaeomagnetic time scale and there have also
with a regular 100 ka cycle occurring over the past 700 ka. been attempts to link: the ocean record with continental
Many researchers attribute these regular climatic fluctua- palaeoclimatic data. The most successful of these correlations
tions largely to the effects of slight variations in the position has been achieved with loess sequences (see Section 10.3.7).
and orientation of the Earth with respect to the Sun which The longest record is provided by the loess sequences of
affect the receipt of solar radiation; this is commonly known China which seem to give a continuous climatic record
as the Milankovitch mechanism for glacial-interglacial for the past 2.3 Ma. Enhanced rates of deflation and aeolian
cycles. Other research, however, has indicated that other dust transport occurred during periods of semi-arid climate
factors, such as the organic productivity of the oceans and coinciding with glacial advances in the northern hemi-
the concentration of CO 2 in the atmosphere, may also be sphere. A correlation has now been made between the past
important. The oxygen isotope data indicate that over the 500 ka of a Chinese loess sequence with the 180 record of a
past 3 Ma the Earth has experienced a glacial climate for deep-sea core in the north-west Pacific and this has greatly
a significant majority of the time, with glacial episodes of improved the dating of the loess sequence (Fig. 14.6).
around 80-100 ka duration alternating with 10 ka-20 ka A major contribution to the under tanding of the nature
interglacial interludes, at least over the past 700 ka. Signifi- and magnitude of Quaternary climatic change has come from
cant changes in the 180/ 160 ratio are now indicated by the CLIMAP (Climate/Long-Range Investigation Mapping
oxygen isotope stage numbers and these provide a good and Prediction) project which attempted to document the
guide to the major changes in ice sheet volume and, in- global environment at the last glacial maximum 18 ka BP.
 Climate, climatic change and landform development 351

Although concerned with just the most recent major glacial

advance, the environmental changes identified were probably
mirrored by earlier glacials of similar magnitude. Large ice
sheets blanketed a significant proportion of the land area of
the northern hemisphere and some 18 per cent of the Earth's
surface was ice-covered (Fig. 14.7). In eastern North
America the Laurentide ice sheet extended from the Rocky
Mountains to the Atlantic Ocean and from the Arctic Ocean
to the latitude of New York. In the ice free area to the south
large lakes formed, such as those of the Great Basin of the
western USA, in response to the supply of meltwater, and to
lower evaporation as a result of lower temperatures.
Until the 1960s it was widely thought that the humid
tropics had been little affected by the glacial-interglacial
fluctuations of the higher latitudes. New data from a range
of sources, including palynological studies of sediment cores
18
19 and work on changing lake levels, have now shown that,
during the latter part of the last glacial at least, much of the
? 22
humid tropics were probably cooler and drier than at
0.88 J';:ilA'M'iCl:o J IJ 23 present (Fig. 14.8). Although the evidence is sparse in some
0.94 ;.;.;.;.;.; .
? k1 regions, during the period around the last glacial maximum
at about 18 000 a BP the tropical rain forests seem to have
contracted to a smaller area than they presently cover. There
was a compensating increase in the extent of savanna vege-
tation, indicating a significant reduction in precipitation.
According to some interpretations of the sedimentological
evidence up to half of the land area within 30° of the
equator was covered by active ergs at this time, and lake
levels, especially in the northern hemisphere, were gene-
rally very low in the period 14 000-21 000 a BP. The evi-
dence for the magnitude of temperature depression is less
1.72 '""'~::,.,.,<~":":'}7."}""'}~m::: certain with a greater temperature decrease being inferred
~:&8U~A( from the lowering of the altitude of vegetation and glacier

1.88
r
~ ~ ~ ~ ~~ ~ ~ ~ ~ ~
limits than is indicated by information from deep-sea cores.
It is in fact not a surprise to find evidence of tropical
aridity coinciding with glacial episodes. We would expect a
reduction in precipitation to occur as a result of reduced
evaporation in response to lower ocean surface temperatures
and a decrease in the area of the oceans as consequence of a
global fall in sea level. Exceptions to this glacial aridity
trend did occur though, with some regions such as North
Africa being more humid (Fig. 14.8). This was due to the
equatorward deflection of rain-bearing westerly weather
systems by the large high pressure cells established over
the extensive northern hemisphere ice sheets.
Climatic changes have continued since the end of the last
Fig.14.S Oxygen isotope record for the about the past 2 Ma glacial around 10 000 a BP but their magnitude has been
from the Pacific deep-sea core V28-239 analyzed by N. J. much attenuated. Although mean global temperature does
Shackleton and N. D. Opdyke correlated with the palaeomagnetic not appear to have varied by more than about 2 °C over the
time scale of Berggren et al. The oxygen isotope stage numbers
past 8000 a or so, evidence from fluctuating lake levels indi-
are shown, the odd-numbered shaded stages indicating relatively
warm interludes. (Modified from WB. Harland et al. (1982) A cates that the tropics were generally much wetter in the
Geologic Time Scale. Cambridge University Press, Cambridge, period 12 000 to 5000 a BP than they were during the pre-
Chart 2.17, p. 42.) ceding 10 000 a, or have been over the past 5000 a. This
humid episode was, however, punctuated by much drier
352 Exogenic processes and landforms

Xifeng loss V21-146 Specmap

II Aeolian flux 0 18 0
a b c (mg cm- 2 kyr -1) d e
AQl
Ilyr l linit Unit a 200 400 600 800 6 5 4 3 2 2 o -I 2
a

100

200

300

100

500

600
KEY
o Loess m Loess inter-layer
D Weathered loess _ Palaeosol
~ Palaeosol inter-layer _ Black loam

Fig.14.6 Correlation of the loess sequence at Xifeng, China, with the oxygen isotope recordfrom west Pacific deep-sea core
V21-146. Dating of the loess sequence in column (a) is based on a technique known as magnetic susceptibility and in (b) on correlation
with the aeoljan dust flux. Column (c) shows the aeolian flux record of the deep-sea core while column (d) illustrates the oxygen isotope
record. Column (e) shows the global standard chronology for the oxygen isotope record (Specmap). (From S. A. Hovan et al. (1989)
Nature 340, Fig. 3, p. 298.)

periods in particular regions. In Africa, for instance, there

was a marked increase in aridity during the periods 11 000-
a
10 000 a BP, 8500-6500 a BP and 6200-5800 BP. It has
been the subhumid and semi-arid tropical regions that have
experienced the most significant post-glacial fluctuations in
precipitation, so it is not surprising that these areas pose
particular problems of landscape interpretation.

Fig.14.7 Probable extent of the northern hemisphere ice sheets

at the last glacial maximum 18000 a BP (Based largely on G. H.
Denton and T. 1. Hughes, (1981) The Last Great Ice Sheets.
Wiley, New York, p.viii.)
 Climate, climatic change and landform development 353

~--------------

~.
,~

Fig.14.8 World climate about 18000 a BP. Note that the palaeoclimatic reconstructions for some areas are uncertain. (Based partly
on T. H. Van Andel (1985) New Views on an Old Planet. Cambridge University Press, Fig. 4.4, p. 61 and various other sources.)

14.4 Effects arising predominantly from bedrock resistance influences glacial trough form with deeper
temperature changes and narrower valleys occurring in resistant lithologies. In
regions of modest elevation, landscape modification during
glacials may be confined to the formation of cirques. In the
14.4.1 Landscapes of deglaciation
mid-latitudes of the northern hemisphere cirques are prefer-
Virtually ail glacial landforms are in a sense relict since entially developed on north-east slopes. Snow accumulates
both erosional and depositional glacial landforms are only here because of the combination of shade from summer
fully exposed once the covering ice has retreated. Similarly, insolation with protection from the predominant westerly
the onset of a glacial morphoclimatic regime requires a winds. In the southern hemisphere. south-east slopes are
sustained cooling across the temperature threshold at which favoured for similar reasons.
snow can accumulate from one year to the next. In Chapter The nature of landscape modification by ice sheets is
11 we looked at the development of specific glacial land- controlled by ice sheet behaviour, the bedrock geology and
forms, but here we will confine our attention to the broad the pre-existing form of the landscape (Table 14.2 and Fig.
nature of landscape modification that resulted from the 14.9). The crucial factor seems to be the state of the basal
expansion of the Late Cenozoic ice sheets into the mid- ice. Active erosion can occur below warm-based ice and
latitudes (Fig. 14.7) and the extension of upland ice caps extensive areal scouring is accomplished where the pres-
and mountain glaciers to lower elevations. sure melting point is attained, normally beneath the centre
of ice sheets and around the margins in lower latitudes.
14.4.1.1 Relict landscapes of glacial erosion This broad pattern can be modified by the subglacial topo-
Mountain, or alpine, glaciation is characterized by a deep- graphy since upland regions will be more likely to be
ening of valleys by glacial erosion and active physical wea- covered by thinner, cold-based ice. Only very limited erosion
thering on valley sides. Glacial troughs develop generally is achieved beneath the cold-based polar flanks of continen-
along the course of pre-existing river valleys (Fig. 11.16); tal ice sheets.
354 Exogenic processes and landforms

Table 14.2 Main types of glacially eroded landscapes in relation to controlling variables

LANDSCAPE TYPE ICE CHARACTERISTICS LITHOLOGY TOPOGRAPHIC CHARACTERISTICS

Areal scouring Warm-based Crystalline, jointed rocks; Landscape of low relief with joint and fault control of deeper
impermeable depressions. Up-glacier sides of ridges smoothed; lee sides
rugged
Selective linear erosion Warm-based in troughs, Not relevant Troughs usually along pre-existing river valleys divide
cold-based over plateaus unmodified upland slopes and plateaus which may retain a
mantle of pre-glacial regolith
Little or no sign of glacial Cold-based, no basal Permeable? - reduces Minimal glacial modification of pre-glacial forms with
erosion debris water present at ice-rock regolith-covered slopes and interfluves
interface
Composite ice sheet/ Alternating ice sheet/ Hardrocks Cirques and other alpine forms well preserved where ice-sheet
alpine glacier types erosion predominantly linear, but highly modified where
upland erosion has occurred
Alpine landscapes Valley and cirque Affects glacial trough Dendritic network of deep troughs with cirques at higher
glaciers morphology elevations separated by horns and aretes

(Based largely on R. J. Chorley et al. (1984) Geomorphology. Methuen, London, Table 19.1, p. 515 and discussion in D. E. Sugden and B. S. John
(1976) Glaciers and Landscape. Edward Arnold, London, Chap. 10.)

A
ice flow

1\
basal thermal regimes

I warm- I / ~. I warm-
---I freezing ~warm-melting~I freezing
I I 1.....
----
.....- - - - - - - - - - . areal scour
bedrock obscured
1 by deposition
- - - - maximum excavation

ROLE OF TOPOGRAPHY

no erosion on
highest hills

c
ROLE OF LITHOLOGY

Rock type influences topography

in boundary zone between
contrasting basal thermal regimes
I I
1
I
1 1
1 I I
areal scour affects al!

~ permeable

Fig.14.9 Schematic representation ofglacial erosion by ice sheets. The left side represents polar or continental environments while
the right side illustrates maritime or mid-ladtude conditions. (A) The main basal regimes and the idealized erosional effects produced;
(B) the likely effects o/topography; (C) the effects o/variations in bedrock permeability on erosion and deposition. (From R. J. Chorley
et al. (1984) Geomorphology. Methuen, London, Fig. 19.8, p. 514.)
 Climate, climatic change and landform development 355

while active ice-retreat landscapes are characterized by sub-

glacial landforms traversed by marginal ridges.
Retreating valley glaciers leave a series of terminal
moraines which may be partially destroyed by meltwater
activity. Stagnation of the glacier snout during retreat allows
the development of kettle-holes and kames. Not surprising-
ly the patterns for ice sheets are more complex, but it is still "
possible to identify a sequence of depositional forms at the
continental scale left by retreating ice sheets in regions
such as North America (Fig. 14.13). This sequence can,
however, be modified by topography. On a relatively fea-
tureless plain the succession of subglacial and ice-marginal
depositional forms should be clearly developed, but where
a retreating ice front encounters an irregular topography
pockets of ice will eventually be isolated in depressions.
This stagnated ice will give rise to a chaotic depositional
topography characterized by moraine and modified by
fluvioglacial deposits.

14.4.1.3 Glacio-isostasy
Fig.14.10 Pattern of glacial erosion associated with the An important indirect effect of the fluctuations in ice sheet
Laurentide ice sheet of North America. (Based largely on B. 1. volume associated with glacial-interglacial oscillations is
Skinner and S. C. Porter (1987) Physical Geology. Wiley, New the changing load on the crust. As an ice sheet grows it
York, Fig. 19.27, p. 544 and D. E. Sugden (1982) Arctic and loads the crust which subsides; when the ice melts the load
Antarctic. Blackwell, Oxford, Fig. 4.16, p. 84.)
is removed and the crust rebounds. This effect is referred to
as glacio-isostasy.
" Since the lithosphere possesses some flexural rigidity (see
These anticipated patterns of erosion are apparent across Section 2.2.4) it bends in response to a load, and this gives
the large area of North America affected by the Laurentide rise to an accompanying peripheral forebulge (Fig. 14.14).
ice sheet (Fig. 14.10). Areal scouring is evident in those After the load is removed and the crust rebounds the
areas that lay beneath the centre of the ice sheet such as forebulge subsides. These adjustments do not, of course,
southern and western "Baffin Island (Fig. 14.11). In some occur instantaneously in response to changes in load since
areas the local direction of ice movement is clearly evident the rate of crustal displacement is controlled by the rate at
from the orientation of lakes, fluted ground moraine and which the highly viscous mantle moves at depth to com-
drumlin fields (Fig. 14.12). Selective linear erosion has pensate for load changes. Regions such as Scandinavia and
occurred in higher regions such as eastern Baffin Island and eastern Canada, which were covered by a great thickness of
north-east Greenland, while little erosion is evident in ice at the last glacial maximum, have so far only accom-
northern Greenland and over much of arctic Canada. Similar plished about half of the total vertical adjustment of between
patterns occur at a smaller scale in Scotland which lay beneath 500 and 800 m that will eventually occur in response to
the western margin of the Fenno-Scandian ice sheet. Areal deglaciation which was completed by 8000-10000 a BP.
scouring together with the erosion of deep troughs and The most obvious manifestation of such isostatic rebound
cirques (largely a result of high relief) prevails in the west, is the displacement of shorelines (Fig. 14.15; see Chapter
whereas selective linear erosion has been predominant in 17). Although the melting of ice sheets at the end of the last
the Cairngorm Mountains in eastern Scotland which retain glacial led to a rapid rise in sea level this was outpaced by
pre-glacial weathering forms on plateau summits. Preserva- shorelines located in the vicinity of the major centres of ice
tion of pre-glacial regolith is even more pronounced in the accumulation. This retreat of shorelines is still continuing,
Buchan area on the north-east coast of Scotland. and it is known that before the rebound of eastern Canada
has been completed Hudson Bay will have become dry
14.4.1.2 Relict landscapes of glacial deposition land.
The development of landscapes of glacial deposition depends
primarily on the duration of glaciation and the quantity of
debris carried by a glacier or' ice sheet. Two types of ice- 14.4.2 Relict periglacial landscapes
margin landscape can be distinguished. Stagnant-ice land- Relict periglacial features are widely known from those
scapes are dominated by irregular hummocky moraine, regions marginal to the northern hemisphere ice sheets and
356 Exogenic processes and landforms

Fig.14.11 AeriaL photo showing areaL scouring by the Laurentide ice sheet in southern Baffin IsLand, northern Canada. Numerous
small Lakes are characteristic of areaLLy scoured gLaciated Landscapes. The area covered is about 14 km across. (Photo courtesy The
Photographic Survey Corporation Ltd, Toronto, Canada.)
 Climate, climatic change and landform development 357

Fig.14.12 Orientated topography north ofAberdeen Lake, Keewatin District, Northwest Territories, Canada. The orientation is
evident largely from numerous lakes and indicates a NNW-SSE ice flow direction. Although within a zone of areal scouring the region
has a cover of till. The area covered is about 180 km across. (Landsat image courtesy R. S. Williams Jr.)

from other areas which experienced a marked cooling during glacial features have been recorded. The distribution of ice-
the Quaternary. Information for Europe is particularly wedge casts extends equatorwards about as far as the most
detailed. Remnants of pingos have been identified in southern southerly limit of the Laurentide ice sheet (Fig. 14.16).
England and Wales where mean annual temperatures are Relict periglacial features have played an important part
around 9°C today. Ice-wedge casts are widely known and in the reconstruction of climatic conditions prevailing during
some are associated with patterned ground; polygonal the Quaternary. Such palaeoclimatic reconstruction depends
patterns with meshes 8 m across have been identified near on a knowledge of the environmental parameters within
Evesham in southern England. Although the evidence is not which contemporary periglacial features are forming, but this
so abundant in North America, a number of fossil peri- procedure must be undertaken with caution. It is important,
358 Exogenic processes and landforms

. . - - Centre of ice mass and of

isostatic depression

Drumlinsand
eskers, and
End Disintegration possibly Drumlinized Fluted Scoured
moraines features localized Drumlins ridges ground bedrock
moraine
Rogen
moraines

WASTAGEZDNE' 'ACTIVEZONE' EROSION

(DEPOSITION FROM TRANSITION (STREAMLINING AND TRANSITION
ZONE
STAGNANT ICE) TILL LODGEMENT)

Isostatically depressed
Fig. 14.13 Schematic representation of the sequence of land surface
depositional landscapes to be expected in association with the
mid-latitude periphery of Late Cenozoic continental ice sheets.
(ModifiedJrom D. E. Sugden and B. S. John (1976) Glaciers and Fig.14.14 Loading of the crust by the growth of an ice sheet.
Landscape. Edward Arnold, London, Fig. 13.16, p. 275.) Note the flexure of the lithosphere and the formation of a
peripheralforebulge. (After 1. J. Lowe and M. 1. C. Walker
(1984) Reconstructing Quaternary Environments, Longman,
for instance, to appreciate the variablity of periglacial con- London, Fig. 2.30, p. 61.)
ditions in different locations in the Quaternary. The southern
margins of the ice sheet covering much of European Russia important point is the distinction that should be made
would have experienced an extremely cold and arid regime between present-day high latitude periglacial regions with
compared with the more maritime 'Icelandic' -type climate tundra or taiga vegetation, and those regions peripheral to
prevailing in the ice-free zone of southern Britain. A further the continental ice sheets of the Quaternary. Although some-

Fig. 14.15 Raised beach at about 8 m a.bove present mean sea level in the Isle ofArran, western Scotland. In this area of northern
Britain uplift associated with isostatic rebound following deglaciation has outpaced the post-glacial rise in sea level.
 Climate, climatic change and landform development 359

wedge polygons are confined to the zone of continuous

permafrost and can be preserved by infilling by sediment
on thawing. Pingos similarly require permafrost for their
development and can be identified as remnant collapsed
features, although there may be some confusion with other
thermokarst forms. Relict cryoplanation terraces and cry-
opediments have also been used, but these are less suitable
because they are not necessarily associated with permafrost
nor are they always readily distinguishable from structural
benches and other non-periglacial features. Deposits and
their associated structures may remain long after the mor-
phological form with which they were originally associated
has been degraded. Gelifluction deposits, involutions and
loess blankets have all been employed as evidence of peri-
glacial conditions although they do not provide very specific
guides to temperature.

Fig.14.16 Location of ice-wedge casts in North America. (After

A. L. Washburn (1979) Geocryology. Edward Arnold, London,
14.5 Effects arising predominantly from
Fig. 14.20, p. 307 and R. 1. E. Brown and T. L. Pewe (1973) in precipitation changes
North American Contribution, Permafrost Second International
Conference, National Academy of Sciences, Fig. 8, p. 90.) 14.5.1 Change in fluvial systems
times loosely regarded as analogous environments, they are Changes in climate during the Late Cenozoic have un-
in fact significantly different. The Quaternary ice sheets of doubtedly had a major impact on both global and regional
the northern hemisphere extended well into the mid-Iati- patterns of runoff and sediment transport, but in most cases
t~des (as far as 40 oN in eastern North America), and the it is only the 'changes of the relatively recent past that can be
climate of these areas would have been distinctly warmer, recognized clearly in the morphology of abandoned or highly
at least in summer, than present-day high latitude environ- modified river channels. The link between changing climate
ments. The higher angle of the solar radiation would have and river channel behaviour is not a simple one since the
promoted more rapid melting of snow and a greater depth fluvial system responds not just to a change in mean annual
of seasonal thaw. There is, indeed, no exact equivalent of precipitation but to changes in a complex range of other
such a climate today. A further 'problem is that some peri- factors which can affect the magnitude and temporal vari-
glacial forms assumed to be active today may in fact be ability of runoff and sediment transport. These factors include
relict features and therefore incapable of indicating the alterations in the density and type of vegetation cover,
critical temperature thresholds controlling their development. changes in rates of evapotranspiration related both to tem-
For instance, it has been argued that the present distribution perature and vegetation characteristics, changes in the inten-
of pingos does not necessarily indicate the temperature sity of rainfall and its distribution throughout the year and
limit required for their development since most examples in changes in the areal extent of permafrost and the contri-
North America and the USSR began to develop hundreds bution of snow-melt to runoff. An additional factor is the
of thousands of years ago under temperatures greatly differ- lags that may occur between a change in climate and the
ent from those of today. response of the fluvial system (Fig. 14.17). Several responses
In spite of these problems it is possible to indicate the are involved; first, the impact of climatic change on the
approximate climatic parameters of certain landforrps charac- vegetation cover; secondly, the effect of these external envi-
teristic of periglacial environments (Table 14.3). The most ronmental factors on the sediment and discharge charac-
useful features are: those confined to permafrost since-these teristics of streams; and finally, the morphological adjustments
provide a clear indication of the maximum mean annual of the channel and valley system to these sediment and
temperature; those that are not rapidly destroyed by pro- discharge variables.
cesses active in a subsequent morphoclimatic regime or are Some idea of the drastic changes that can occur to fluvial
preserved in some way; and those that cannot easily be systems as a consequence of climatic change can be gained
confused with non-periglacial forms. Relict ice wedges and, from research on the eastern Sahara Desert. Parts of this
in spite of reservations, pingos, most closely fulfil these region, especially the area known as the Selima Sand Sheet
requirements, being capable of giving minimum estimates on the Egypt-Sudan border, are quite appropriately described
of the temperature change to the present. They are conse- as hyper-arid at the present day with no trace of surface
quently widely used in palaeoclimatic reconstructions. Ice- water or fluvial activity. The Selima Sand Sheet itself is
360 Exogenic processes and landforms

Table 14.3 Environmental range of periglacial processes and landforms

LOWLANDS HIGHLANDS

Polar Subpolar Mid-latitude Polar Subpolar Mid-latitude Low latitude

Processes
Frost action
Frost wedging *** **** ** *** **** *** **
Frost heaving and thrusting **** *** ** *** **** ** *
Permafrost cracking **** *** *** **
Other processes
Frost creep *** **** ** ** **** *** *
Gelifluction **** *** ** *** **** ** *
Nivation **** *** *** **** *** **
Wind action **** *** ** **** *** ** **
Landforms
Patterned ground
Nonsorted circles **** *** *** *** **
Sorted circles **** ** *** ****
Small nonsorted polygons **** *** ** *** **** *** ***
Large nonsorted polygons
(ice wedge or sand
wedge) **** *** *** **
Small sorted polygons **** *** * *** **** ** ***
Large sorted polygons **** ** *** **
Small stripes **** *** *** **** ** ***
Large stripes **** ** *** ****
Other features
Permafrost **** *** **** *** **
Pingos **** ***
Palsas *** **** *** **
Thaw slumps **** **** ** **
Gelifluction lobes **** *** *** **** **
Block fields and rock
streams *** *** **** **** **
Rock glaciers *** **** ****
Nivation benches and
hollows **** *** *** **** **
Cryoplanation terraces *** *** **** ****
Asymmetric valleys related
to permafrost **** *** **** ***
Loess *** **** ** *
Notes: Asterisks indicate relative abundance of individual landforms and relative significance of a particular process in a range of
environments. **** Abundant, *** common, ** uncommon, * scarce, ? rare or absent.
Source: Modified from A. L. Washburn (1979) Geocryology. Edward Arnold, London, Table 13.1, pp. 277-8.

one of the most barren, featureless regions on Earth, con- dramatic climatic fluctuations can also lead to profound
sisting of an almost entirely flat plain with a cover of wind- changes in channel morphology. In Chapter 9 we described
blown sand and fine gravel. In 1981 aQ experimental radar the way in which channel morphololgy can change in res-
system installed on board the space shuttle Columbia acquired ponse to changes in controlling variables, but it is important
several radar images of this area. On inspecting these images it to emphasize that not all such adjustments are a consequence
became apparent that the radar pulses had penetrated the of natural processes. In examining changes that have occurred
relatively thin and completely dry sand cover to reveal in the past few thousand years, and especially the last 200 a, it
buried river valleys. Some of these are almost as broad as is important to keep in mind the dramatic impact that human
the present Nile Valley, and earlier work on the drainage activity can have on fluvial systems. This does not just
history of this region indicates that they may be as much as involve the direct control of rivers through dam construc-
25 Ma old. Superimposed on these large valley systems are tion and the modification of channels, but also encompasses
smaller incised channels which were probably periodically indirect effects, such as the increase in sediment supply to
active during the wetter phases of the Quaternary. channels which often follows the replacement of natural
This is obviously an extreme example of the affect that vegetation by agricultural land (see Section 15.3.1.1).
climatic change can have on fluvial systems, but far less
 Climate, climatic change and landform development 361

where underfit streams are widespread in an area.

I Relative Precipitation
I IRegime Of particular interest are misfit streams meandering within

HumId
.
~-I
I I I
, . . ----- valleys which themselves exhibit a regular meandering pattern
of much greater wavelength; such streams are described as
manifestly underfit. If the relationship between discharge
Arid I I I - and meander wavelength established from meandering
channels in alluvium is applied to these valley meanders,
Relative Vegetation Cover an estimate can be made of the discharge that would have
I been required to form them. Valley wavelengths are usually
Maximum
between 5 and 13 times the wavelengths of the sinuous
underfit stream channels contained within them, and since
Minimum
discharge is proportional to the square of the wavelength
the increased discharge required to form the valley meanders
I
Relative Hillslope Potential for Erosion is between 25 and 169 times. Even the low end of this range

Maximum
I is unrealistic if the change in discharge is brought about by
a simple alteration in precipitation, so other factors need to
Minimum
be considered. In some cases large discharges could be
produced by glacial meltwater, but it is also important to
I consider whether meanders at least partly cut in bedrock

f'dlm", rid
Relative Geomorphic Work
necessarily adjust in the same way to discharge as those
('.g I formed in alluvium and whether the lateral migration of a
Maximum
meandering channel could, of itself, create a meandering
valley pattern of longer wavelength.
Minimum
In some cases relatively straight channels are seen to
flow within wider meandering valleys, but more than the
Fig.14.17 Schematic representation of the effect of an abrupt
anticipated number of pools and riffles occur in the channel
change in climatic regime on fluvial systems. Maximum erosion within each meander bend. Such streams are termed Osage-
and sediment transport is associated .with the change from arid to type underfits and appear to represent a situation where
humid conditions when the increase in protective vegetation cover the channel is still adjusting to a decrease in discharge and
lags behind the increase in precipitation. Erosion rates has not yet decreased the amplitude of its meanders to
subsequently decline as vegetation density increases and surface
runoff is reduced. (After 1. C. Knox (1972) Annals of the
match the reduction in flow.
Association of American Geographers 62, Fig. 6, p. 408.)
14.5.1.2 Long-term river metamorphosis
Adjustments in river channel morphology which occur in
14.5.1.1 Underfit streams response to changes in discharge and sediment load are
As a broad generalization we would expect valleys to be of sometimes referred to as river metamorphosis. One of the
a size proportionate to the dimensions of the river channels most frequently cited examples of river metamorphosis is
that they contain. In a number of instances, however, streams the study of the Murrumbidgee River carried out by S. A.
appear to be far too small with respect to their valleys. Such Schumm. The Murrumbidgee is a tributary of the Murray
underfit streams or misfit streams could obviously arise River and drains westwards across southern New South
where there has been a reduction in precipitation and runoff; Wales on an alluvial plain with a mean slope of about
the channel can adjust its size rather rapidly to a decrease in 0.3 m km- I . The presently active channel is around 60 min
discharge but the response of the whole valley will be much width and it meanders across a floodplain which also con-
slower. (It is, of course, possible to have an increase in dis- tains many large oxbow lakes marking an earlier and much
charge and therefore a misfit in the sense of a stream that is larger meandering channel, and the trace of a much less
'too large' for its valley, but in this case the valley will sinuous channel (Fig. 14.18). Schumm measured various
experience rapid erosion until it reaches an equilibrium morphological and sedimentological properties of these
size.) Mechanisms other than a decrease in precipitation three distinct channels (Table 14.4) and interpreted them as
can also produce underfit streams; discharge can be reduced reflecting adjustments to the hydrological regime and sedi-
by one river system diverting flow from another through ment load characteristics prevailing when each channel was
river capture (see Section 16.2.3) and in permeable litho- active.
logies, such as limestone, flow can be diverted progressively Evidence from landforms and soils from the alluvial
underground thereby depleting surface discharge (see Section plain indicates that the older low-sinuosity palaeochannel
9.6). Climatic mechanisms, however, are the most likely was apparently active in the Late Pleistocene when the cli-
362 Exogenic processes and landforms

to changes in discharge but also to changes in the nature of

the sediment load which reflect indirect climatic influences
through changes in the vegetation cover in the catchment.
The active channel of the Murrumbidgee currently transports
small quantities of silt, sand and clay and when conditions
were wetter in the past a more dense vegetation cover would
have reduced the quantity of sediment being transported in
spite of the increase in runoff. In contrast, when conditions
were drier the resulting reduction in vegetation cover would
have allowed sediment supply to the channel to increase
significantly and more sediment would have been trans-
ported in the low-sinuosity. palaeochannel. Although the
total runoff would have been lower, peak discharges would
have been higher as a result of increased quick flow.
By assessing the morphological properties of the palaeo-
channels in the light of the relationships between channel
properties, discharge and sediment transport evident from
modem channels, Schumm showed how it is possible to
o 2 estimate the discharge and the nature of the sediment load
I I
km
of the Murrumbidgee palaeochannels. Calculations show
that the older low-sinuosity palaeochannel would have been
Fig.14.18 Active and palaeochannels of the Murrumbidgee capable of transporting large quantities of sand at bankfull
River, New South Wales, Australia. Flow is to the left. (After S. A. discharge, and this channel is, as anticipated, full of sand.
Schumm (1969) in: R. 1. Chorley (ed.) Water, Earth and Man. The large meandering palaeochannel would also have been
Methuen, London, Fig. 11.11.4, p. 531.)
competent to transport significant quantities of sand but it is
filled largely with clay and silt. This suggests that it was
Table 14.4 Morphology of the active and palaeochannels of the only in the drier phase when the low-sinuosity channel was
Murrumbidgee River.
active that large quantities of sand were being supplied to
MODERN YOUNGER OLDER
the channel from the poorly vegetated catchment. The main
MURRUMBIDGEE MEANDERING LOW-SINUOSITY response represented by the meandering palaeochannel to
PALAEOCHANNEL PALAEOCHANNEL an increase in discharge is the increase in channel dimen-
Channel width (m) 65 140 180
sions represented by a greater channel depth and width and
Channel depth (m) 6 II 3
longer meander wavelength.
Width/depth ratio 10 13 67
An important aspect of this analysis of the Murrumbidgee
Sinuosity 2.0 1.7 1.1
River is the manner in which the increased transport of sand
Gradient (m km- I ) 0.13 0.15 0.37
was accomplished during the dry episode, even though dis-
Meander
charge and the depth of flow decreased. Clearly an increase
wavelength (km) 0.85 2.14 5.49 in channel gradient would have been required, but this was
Channel not achieved by deposition and a steepening of the whole
Silt-clay (0/0) 25 16 1.6 alluvial plain but by a decrease in channel length through a
Bed load (0/0) 2.2 3.4 34 reduction in sinuosity. As shown in Table 14.4, the present
Bankfull discharge meandering channel has a sinuosity almost twice as great as
(m 3 s-l) 310 1430 640 the low-sinuosity palaeochannel and consequently the gra-
Estimated sand dient of the latter would have been nearly twice as high.
transport at bankfull
discharge (t day-I j 2000 21000 54000
Other stud'ies have broadly supported Schumm's conclu-
sions about the response of river systems to climatic changes,
Source: After S. A. Schumm (1977) The Fluvial System. Wiley, New York, at least in semi-arid environments. An investigation of the
Table 5-5, p. 169.
Colorado River in Texas, for instance, has shown a ten-
dency during dry phases for an increase in channel width,
meander wavelength and sediment size and a decrease in
mate in the region was drier than at present, whereas the channel sinuosity (Fig. 14.19). But other studies from tem-
younger meandering palaeochannel was functioning when perate environments suggest that these relationships are not
conditions became wetter around 3000 a BP. Nevertheless, necessarily universal. Indeed, complex sequences of change
changes in channel morphology occurred not only in response in river channel morphology may arise from different rates
 Climate, climatic change and landform development 363

Drier Humid
I I
-- Drier
I
Humid

.-
I
...
Drier

)~"""I~
Fine 10000 1000
1 100

Time (a B.P.)

Fig. 14.19 Changes in the Colorado River of central Texas in response to climatic fluctuations during the Late Quaternary. Six
phases are identified: (1) relatively arid climate associated with infrequent, high-magnitude floods, high sediment yields, aggradation
of coarse sediment and braiding, and low channel sinuosity; (2) cooler, wetter and more uniform precipitation associated with a
decreasing magnitude offormative discharges, transport offiner material and progressive change to high-sinuosity channels,foliowed
by more frequent but lower magnitude floods, vegetated slopes and less soil erosion, (3) warmer and more arid conditions with
increasing seasonality ofprecipitation associated with incisionfoliowed by aggradation and braiding, and wide, low sinuosity
channels; (4) more humid climate associated with lower formative discharges and more uniformflows, transport offiner material and
narrow, sinuous channels with a reduced meander wavelength; (5) a relatively dry climate associated with streams carrying a coarse
load, entrenchmentfoliowed by aggradation, and wider, less sinuous streams; (6) decrease informative discharges, coarse load and
progressive channel narrowing. (Modified from V. R. Baker and M. M. Penteado-Oreliana (1977) Journal of Geology 85, Fig. 11,
p.414.)

of change in discharge and sediment transport and from latter, exemplified by -the Jurua which rises at an elevation
changes in how frequently channel-forming floods occur. of only 600 m, display little variation in discharge through-
Clearly, more studies are needed from a wide variety of out the year, transport a negligible bed load and a modest
environments in order to assess the range of responses that suspended load of extremely fine calibre, and have narrow
can occur. floodplains and single meandering channels with variable
The necessity of expanding our understanding of fluvial but generally high sinuosity (Table 14.5). In comparison the
systems which have experienced different climatic histories rivers draining the Andes, such as the Solimoes (the Amazon
is underscored by research on the Amazon, the greatest river trunk river) (Fig. 14.20) and the Japura carry significantly
system in the world. At an estimated 230000m 3s- 1 (that is, more bed load and a higher suspended load and have broad
15 per cent of the total flow of fresh water to the oceans) its floodplains and anastomosing channels of low sinuosity.
mean annual discharge is over five times that of the next largest These contrasts have been interpreted as a consequence of
river, the Zaire (Congo). The gradient of the Amazon over the different abilities of these two types of river to trans-
a distance of more than 3000 km from the foothills of the port the vast quantities of coarse alluvium that were mobi-
eastern Andes in Peru to its mouth averages only 0.1 m km-I; lized during the much drier conditions which prevailed at
this is reduced to an almost imperceptible gradient of the height of the last glacial. The relationships established
0.03 m lan-lover the lower 1440 km of its course below the for river metamorphosis in semi-arid environments do not
city of Manaus. account for the kinds of channel adjustments evident in the
In the Amazon system there is a marked contrast in the Amazon sy tern where discharge characteristics and the
morphology of trunk rivers with high elevation sources in active role of vegetation in stabilizing channel banks seem
the Andes and rivers which drain the tropical lowlands. The to be critical factors.
364 Exogenic processes and landforms

Table 14.5 Properties of rivers in the Amazon Basin

JURUA JAPURA SOLIMOES

(JCT. SOLIMOES) (JCT. SOLIMOES) (NEAR FONTE BOA)

Environment Tropical Mixed Mixed

Drainage area
(103km 2) 217 289 1263
Discharge
(1012 m 3a-l) 0.197 0.351 1.161
Solid load
(% total load) 40 48 -30
Bed load
(% solid load) <1 2.3 -5
Sediment yield
(103kg km-2a- l ) 49.4 120.2 Not available
Pattern Moderate to Straight, Moderately sinuous,
tortuous moderately highly anastomosing
s'inuosity anastomosing
Meander form Compound Simple Simple asymmetric
asymmetric symmetric
Sinuosity 1.8-3.0 J.1 1.5

Source: After V. R. Baker (1978) in: A. D. Miall (ed.) Fluvial Sedimentology.

Canadian Society ofPetroluem Geologists Memoir 5, Table 2, p. 221.

14.5.2 Aeolian systems and intensity of the world's arid regions clearly must have
As in the case of rivers, aeolian systems are sensitive to had a profound effect on aeolian activity, and this is attested
changes in precipitation, although changes in temperature by the great expanse of relict ergs around the regions of
can also influence the level of aeolian activity through the currently active sand movement (see Chapter 10). It has
effect on evapotranspiration rates and thus soil moisture been suggested that mean wind speeds were higher during
and the density of vegetation cover. Similarly, as with fluvial glacials and that it is these strong winds that were respon-
systems, the change in precipitation will have most impact sible for creating the huge megadunes of the world's ergs.
if it occurs across a threshold range; in the case of aeolian This provides an interesting alternative explanation to put
systems this. will be the precipitation level at which vege- aside the generally accepted notion of megadunes being the
tation cover begins to restrict severely wind speeds at ground largest component of a hierarchy of aeolian bedforms (see
level and thereby curtail the entrainment of particles by the Chapter 10), and is supported by some recent research on
wind. The precise level of precipitation at which this occurs megadunes in a part of northern Saudi Arabia which inter-
will vary depending on local conditions, but 200 mm seems prets them as relict rather than active forms.
to be a reasonable approximation in many areas. In addition In the Kalahari Desert in southern Africa, three distinct
to changes in surface characteristics brought about by changes arid phases have been recognized during the Late Quaternary
in vegetation the variations in wind strength itself can also on the basis of relict linear dune systems with differing
affect the level of aeolian activity, and alterations in the morphologies and alignments (Fig. 14.21). The oldest (Group
directional characteristics of the wind regime also promote A) are located in north-western Botswana and adjacent
adjustments in the alignment of depositional, and ultimate- areas of Namibia and Angola. They are cut by fluvial and
ly, erosional forms (see Chapter 10). lacustrine features associated with a wet period which affected
A wide range of evidence, but especially pollen records the northern Kalahari from 30 to 20 ka BP and therefore
and dated fluctuations in lake levels in tropical latitudes, must pre-date these landforms. A second phase of aridity
have demonstrated major changes in climate in the world's during which dunes in north-eastern Botswana (Group B)
arid and semi-arid regions during the Quaternary, it now were active followed this humid period and coincides with
being well established that in general greater aridity at these the dry episode recorded from many localities in the tropics
latitudes coincided with glacials in the mid- and high lati- which corresponds with the last glacial maximum around
tudes. Indeed there may have been up to 25 arid-humid 18 ka BP. By about 12 ka BP the zone of aridity had shifted
cycles in this period and no doubt many more smaller ampli- south to the southern Kalahari and the dunes in this region
tude oscillations of the sort which have been evident in the were probably active at this time and remained so into the
past few thousand years. The major changes in the areal extent early Holocene.
 Climate, climatic change and landform development 365

Fig.14.20 Landsat image of an area about 300 km south of Manaus in the Amazon Basin, Brazil, covering an area about 180 km
across. The two major rivers visible are the Solimoes (one of the two primary tributaries of the Amazon system - extreme top of
image), and the smaller Purus. Both rivers are still transporting sediment mobilized during the last glacial and stored in the valleys of
the Andes upstream. The enormous accumulation of sediments in the vast floodplain, or varzea, of the Solimoes has diverted the course
of the Purus. The lakes visible in the centre of the image are known as rias fluviales and are formed by the damming of streams flowing
from interfluves by the floodplain sediments of the major channels. Post-glacial sea-level rise has also been important in the
development of the Amazon system and this is discussed in Section 17.6.1. (Landsat image courtesy N. M. Short.)

Further reading morphoclimatic categories is presented by Ahnert (1987).

The importance of climate in landform development is em- The problem of pediment formation has generated an enor-
phasized in the books by Blidel (1982) and Tricart and mous literature but the brief reviews by Oberlander (1989)
Cailleux (1972), but a critical appraisal of climatic geomor- and Twidale (1981) are useful; the development of the inten-
phology is provided by Stoddart (1969). A thoughtful analysis sively studied pediments of the Mojave Desert is discussed
of how the magnitude and frequency characteristics of by Dohrenwend et ai. (1987) and Oberlander (1972, 1974).
meteorological events might be used to identify meaningful There is no single comprehensive summary which covers
366 Exogenic processes and landforms

ice sheet, the extent of glacial modification of landscapes in

Scotland is considered by Hall and Sugden (1987) and
Andrews (1987) assesses the effects of glacial isostasy. A
comprehensive survey of relict periglacial landforms is
provided by Washburn (1979), and Black (1976) specifically
examines features indicative of permafrost. Reger and Pewe
(1976) and Hemal (1976) provide, respectively, details on
cryoplanation terraces and pingos as permafrost indicators,
while MacKay (1988) questions the palaeoclimatic value of
pingos. Boardman (1987) considers a number of reiict peri-
glacial forms in the British Isles.
Various aspects of the response of fluvial systems to
climatic change are covered in the volume edited by Gregory
(1983), and Knox (1972) provides a valuable analysis of
the effect of fluctuations in precipitation. The remarkable
buried channels of the eastern Sahara are described by
McCauley et ai. (1982). Dury (1977) presents a comprehen-
sive review of the problem of underfit streams and Schumm
(1969, 1977) considers various aspects of river channel
metamorphosis. Schumm (1968) presents a detailed study
of changes in channel morphology for the Murrumbidgee
River in Australia and Baker and Penteado-Orellana (1977)
Fig. 14.21 Distribution of major active and relict dune fields in present a similar investigation of the Colorado River of
southern Africa in relation to present-day mean annual Texas. Starkel (1983), however, emphasizes the complex
precipitation (mm). (From N. Lancaster (1981) Palaeogeography,
nature of the adjustments that can occur, while Baker (1978)
Palaeoclimatology, Palaeoecology 33, Fig. 8, p. 338.)
demonstrates the distinctive behaviour of some humid tropical
rivers. Finally, the impact of climatic change on aeolian sys-
the impact of climatic change on landform development, tems is considered in broad terms by Goudie (1983) and with
but the volumes edited by Derbyshire (1976) and Douglas respect to the Kalahari Desert by Lancaster (1981).
and Spencer (1985) cover a number of relevant topics.
On the long-term record of climatic change Frakes (1979)
provides a detailed background and Frakes and Francis (1988)
consider the evidence for Cretaceous glaciation. On the
Quaternary record Lowe and Walker (1984) is a good intro- References
duction, while Imbrie and Imbrie (1986) cover the oxygen Ahnert, F. (1987) An approach to the identification of morpho-
isotope record in detail. Denton arid Hughes (1981) is an climates. In: V. Gardiner et al. (eds) International Geomorphology
1986 Part II. Wiley, Chichester and New York, 159-88.
exhaustive treatment of the Quaternary ice sheets, while Andrews, 1. T. (1987) Glaciation and sea level: a case study. In:
Fink and Kukla (1977), Kukla (1987) and Pye (1987) dis- R. J. N. Devoy (ed.) Sea Surface Studies: A Global View.
cuss the related record of climatic change represented by Croom Helm, London and New York, 95-126.
loess deposits. The correlation of the Chinese loess sequences Baker, V. R. (1978) Adjustment of fluvial systems to climate and
with the deep-sea core 18 0 record is considered by Hovan source terrain in tropical and subtropical environments. In:
A. D. Miall (ed.) Fluvial Sedimentology. Canadian Society of
et ai. (1989). An overview of Late Quaternary lake fluctua-
Petroleum Geologists Memoir 5,211-30.
tions is presented by Street-Perrott and Harrison (1984), Baker, V. R. and Penteado-Orellana, M. M. (1977) Adjustment to
while Sarnthein (1978) considers the expansion of ergs dur- Quaternary climatic change by the Colorado River in central
ing glacial phases and Van der Hammen (1982) discusses Texas. Journal ofGeology 85, 395-422.
the palaeoecological evidence for greater aridity in the Bell, M. and Laine, E. P. (1985) Erosion of the Laurentide region
of North America by glacial and glaciofluvial processes. Qua-
humid tropics during the last glacial maximum. ternary Research 23, 154-74.
Turning to the landforms themselves, a broad overview Black, R. F. (1976) Features indicative of permafrost. Annual Review
of the impact of glaciation with respect to both landscapes of Earth and Planetary Sciences 4, 75-94.
of glacial erosion and glacial deposition is provided by Boardman, J. (ed.) (1987) Periglacial Processes and Landforms in
Sugden and John (1976), while Boulton et ai. (1985) relate Britain and Ireland. Cambridge University Press, Cambridge.
Boulton, G. S., Smith, G. D., Jones, A. S. and Newsome, J. (1985)
the dynamics of ice sheets to regional and continental-scale Glacial geology and glaciology' of the last mid-latitude ice
patterns of erosion and deposition. Bell and Laine (1985) sheets. Journal of the Geological Society London 142,447-74.
discuss the nature of erosion associated with the Laurentide Blidel, J. (1982) Climatic Geomorphology (translated by L.
 Climate, climatic change and landform development 367

Fischer and b. Busche). Princeton University Press, Princeton, MacKay, J. R. (1988) Pingo collapse and paleoclimatic reconstruc-
and Guildford. tion. Canadian Journal of Earth Sciences 25, 495-511.
Denton, G. H. and Hughes, T. 1. (eds) (1981) The Last Great Ice McCauley, J. F. et al. (1982) Subsurface valleys and geoarcheo-
Sheets. Wiley, New York. logy of the eastern Sahara revealed by shuttle radar. Science
Derbyshire, E. (ed.) (1976) Geomorphology and Climate. Wiley, 218, 1004-20.
London and New York. Oberlander, T. M. (1972) Morphogenesis of granitic boulder slopes
Dohrenwend, J. c., Wells, S. G., McFadden, L. D. and Turrin, in the Mojave Desert, California. Journal of Geology, 80, 1-20.
B. D. (1987) Pediment dome evolution in the eastern Mojave Oberlander, T. M. (1974) Landscape inheritance and the pediment
Desert, California. In: V. Gardiner et al. (eds) International problem in the Mojave Desert of southern California. American
Geomorphology 1986 Part II. Wiley, Chichester and New York, Journal ofScience 274, 849-75.
1047-62. Oberlander, T. M. (1989) Slope and pediment systems. In:
Douglas, 1. and Spencer, T. (eds) (1985) Environmental Change D. S. G. Thomas (ed.) Arid Zone Geomorphology. Belhaven
and Tropical Geomorphology. Allen and Unwin, London and Press, London; Halsted Press, New York.
Boston. Pye, K. (1987) Aeolian Dust and Dust Deposits. Academic Press,
Dury, G. H. (1977) Underfit streams: retrospect, perspect and pros- London and Orlando.
pect. In: K. J. Gregory (ed.) River Channel Changes. Wiley, Reger, R. D. and Pewe, T. L. (1976) Cryoplanation terraces:
Chichester and New York, 281-93. indicators of a permafrost environment. Quaternary Research
Fink, J. and Kukla, G. J. (1977) Pleistocene climates of central 6,99-109.
Europe: at least seventeen interglacials after the Olduvai event. Samthein, M. (1978) Sand deserts during glacial maximum and
Quaternary Research 7,363-71. climatic optimum. Nature 272, 43-6.
Flemal, R. C. (1976) Pingos and pingo scars: their characteristics, Schumm, S. A (1968) River adjustment to altered hydrologic
distribution, and utility in reconstructing former permafrost regimen - Murrumbidgee River and paleochannels, Australia.
environments. Quaternary Research 6, 37-53. United States Geological Survey Professional Paper 598.
Frakes, L. A (1979) Climate Throughout Geologic Time. Elsevier, Schumm, S. A (1969) River metamorphosis. American Society of
New York. Civil Engineers Journal of the Hydraulics Division 95, 255-73.
Frakes, L. A and Francis, J. E. (1988) A guide to Phanerozoic Schumm, S. A (1977) The Fluvial System. Wiley, New York and
cold polar climates from high-latitude ice-rafting in the Creta- London.
ceous. Nature 333, 547-9. Starkel, L. (1983) The reflection of hydrologic changes in the
Goudie, A (i983) The arid Earth. In: R. Gardner and H. Scoging fluvial environment of the temperate zone during the last 15,000
(eds) Mega-Geomorphology. Clarendon Press, Oxford, 152-71. years. In: K. J. Gregory (ed.) Background to Palaeohydrology.
Gregory, K. J. (ed.) (1983) Background to Palaeohydrology. Wiley, Wiley, Chichester and ·New York, 213-35.
Chichester and New York. Stoddart, D. (1969) Climatic geomorphology: review and re-
Hall, A M. and Sugden, D. E. (1987) Limited modification of assessment. Progress in Geography 1, 159-222.
mid-latitude landscapes by ice sheets: the case of Northeast Street-Perrott, F. A and Harrison, S. P. (1984) Temporal varia-
Scotland. Earth Sw1ace Processes and Landforms 12, 531--42. tions in lake levels since 30 000 BP: an index of the global
Hovan, S. A, Rea, D. K., Pisias, N. G. and Shackleton, N. 1. (1989) hydrological cycle. In: J.E. Hansen and T. Takahasi (eds) Climate
A direct link between the China loess and marine 18 0 records: Processes and Climate Sensitivity. Geophysical Monograph 29.
aeolian flux to the north Pacific. Nature 340,296-8. American Geophysical Union, Washington DC.
Imbrie, 1. and Imbrie, K. P. 1986. Ice Ages: Solving the Mystery. Sugden, D. E. and John, B. S. (1976) Glaciers and Landscape: A
(Harvard University Press, Cambridge). Geomorphological Approach. Edward Arnold, London.
Knox, J. C. 1972. Valley alluviation in south-western Wisconsin. Tricart,1. and Cailleux, A (1972) Introduction to Climatic Geo-
Annals of the Association of American Geographers 62, morphology (translated by C. J. Kiewiet de Jonge). Longman,
401-10. London.
Kukla, G. (1987) Loess stratigraphy in China. Quaternary Science Twidale, C. R. (1981) Origins and environments of pediments.
Reviews 6,191-219. Journal of the Geological Society ofAustralia 28, 423-34.
Lancaster, N. (1981) Paleoenvironmental implications of fixed Vander Hammen, T. (1982) Paleoecology of tropical South America.
dune systems in southern Africa. Palaeogeography, Palaeocli- In: G. T. Prance (ed.) Biological Diversification in the Tropics.
matology, Palaeoecology 33, 327--46. Columbia University Press, New York.
Lowe, J. J. and Walker, M. J. C. (1984) Reconstructing Quaternary Washburn, A L. (1979) Geocry9logy: A Survey of Periglacial
Environments. Longman, London and New York. Processes and Environments. Edward Arnold, London.
This page intentionally left blank
Part IV
Endogenic - exogenic
interactions
 15
Rates of uplift and
denudation

15.1 Tempo of geomorphic change This is simply equivalent to a change in ground elevation
Having examined the operation of the various endogenic expressed in mm ka- I or m Ma- I and seems to represent a
and exogenic geomorphic processes in Parts IT and ITI we rather needless proliferation of units.
can now look at the way these two sets of processes interact Since rates of denudation are frequently estimated from
in the formation of the landscape. It is appropriate to begin the mass of sediment transported by rivers, glaciers or the
with the overall rates at which endogenic and exogenic pro- wind, we need to be able to convert this mass to a volume
cesses operate since it is the balance between uplift and in order to estimate the equivalent rate of ground lowering.
denudation at any particular point on the Earth's surface This is done by relating the mass of material transported to
that determines whether that point becomes higher or lower its mean density, usually assumed to be the density of typical
with respect to sea level. Through our discussion of endo- crustal rocks (approximately 2700 kg m-3). Such a proce-
genic and exogenic progresses it will have become apparent dure, however, can raise difficulties when comparing the
that the rates at which they operate vary enormously; in fact removal of solid and dissolved constituents by rivers. This
a glance at Figure 15.1 shows that their rates range over problem is examined in Section 15.3.1.
some 16 orders of magnitude.
The rates of operation of geomorphic processes may be
15.2 Rates of uplift
expressed in various ways. One approach is to view the Before we discuss how rates of uplift can be,measured it is
transport of materials in terms of geomorphic work and to important to clarify what we actually mean when we use
represent this in units of power. This can be done by relating the term 'uplift'. This is necessary because it is used in two
the operation of a particular process to its duration. Although quite di'stinct senses. Here we will use the term surface
this approach is useful for comparing the relative impor- uplift to refer to the upward movement of the landsurface
tance of specific geomorphic processes, when considering with respect to a specific datum (normally sea level). What
the overall rates at which the landscape changes it is more we will term crustal uplift, on the other hand, refers to the
usual to express this in terms of an average lowering or upward movement of the rock column with respect to a
raising of the ground surface over a period of time. Since specific datum (also normally sea level).
we are concerned in this chapter primarily with the net effect As is apparent from Figure 15.2 the relationship between
of various geomorphic processes in modifying the land- surface uplift and crustal uplift depends on the prevailing
scape this is the approach we will use here. rate of denudation. Crustal uplift will only equal surface
Changes in the average elevation of the ground surface uplift if there is no denudation during the period of uplift.
are most commonly expressed in terms of mm ka- 1• For If, as is likely, denudation does occur then crustal uplift
long periods of time of the scale of millions of years it is will be greater than surface uplift. But if crustal uplift is
more appropriate to use the equivalent unit of m Ma- I (note exactly matched by denudation then there will be no sur-
that 1 mm ka- I equals 1 m Ma- 1). Very rapid processes may face uplift. In fact crustal uplift is frequently exceeded by
be better described in terms of mm a-lor even m S-I. It has denudation, and in this situation surface elevation is obvious-
become fashionable in recent years to express changes in ly reduced.
ground elevation in 'Bubnoff units', often represented as B. In most cases surface uplift is associated with active tec-
372 Endogenic-exogenic interactions

RATE
3
mMa-1/mmka-1 1 10 100 10
3
mma-1 1 10 100 10
ma-1 3
1 10 100 10
kma-1 3
1 10 100 10
ms-1
I 10'01
I
0·1
I
1 10 10o

Passive margin Isostatic rebound

subsidence ('instantaneous' unload ng)
I--------l I I

U
« Uplift in orogens
-Z Z
LLJ Hot sRot
Gravity
tectoniCS
migra ion f-----l
UJ ~
<..9 f-------i
0 Explosive
0 Horizontal volcanic
Z plate movement
0 eruptions
UJ I I I--l
Z J: Tectono-eustatic
sea level change
UJ a.. ~ Glacio-eustatic Lava flows
sea level change
I----i
Natural denudation
I'
Low local
I *~
relief relief
Aeolian
« Accelerated
dust
transport
U Z erosion
-Z UJ
I I
~ Dune migration Stream
UJ flow
0 Surface wash
<..9 Z Glacier flow
0 UJ
Debris
>< J: Free face retreat Marine slide Rockfall
cliff
UJ a.. retreat t------l I--------l
Solution I---t
I I Mud flow
Soil creep and
solifluction Debris flow
I I

Fig.15.1 Comparison of rates of various endogenic and exogenic geomorphic processes. Note that rates of specific slope processes
are discussed in Chapter 7 and are included here for comparative purposes.

tonic processes, but crustal uplift can occur simply as the Geodesy is concerned with the precise determination of the
inevitable isostatic response to denudation; that is, the re- Earth's surface form and geodetic methods can provide
moval of material through denudation reduces the load on information on short-term rates of surface uplift. One way
the crust and it moves upward to restore isostatic equili- this can be done is through. precise levelling surveys which
brium. In general such crustal uplift resulting from isostatic are repeated after a number of years or, more often, several
rebound does not exceed the rate of denudation so there is decades. Although systematic errors in levelling make it
no net surface uplift; nevertheless, where flexural effects difficult in some cases to distinguish a real vertical move-
are important (see Section 4.2.3), or where a landsurface is ment from an artefact of the measuring procedure, such
locally dissected by deep gorges, such isostatic recovery geodetic techniques provide an important basis for document-
can lead to surface uplift of parts of a landscape. ing surface uplift and subsidence over periods of a few
years to a few decades. The technology is now available
whereby satellites can determine with great accuracy the
15.2.1 Methods of measurement and estimation form of the Earth1s surface (satellite altimetry) and changes
Rates of uplift can be determined either directly or indirect- in the geoid over time have already been recorded by this
ly by a range of techniques appropriate to different time means.
scales (Fig. 15.3). Some of these methods measure or esti- Along coasts tide gauge records can provide high-quality
mate surface uplift, whereas others indicate rates of crustal information on vertical movements of the landsurface, as
uplift, and it is important to distinguish between them. long as allowance is made for the slow, continuing,. post-
 Rates of uplift and denudation 373

3
Initial D =Okm D =O.5km D =1km D =1km As the time period with which we are concerned becomes
condition Uc =1km U =1km U =1km longer, so the means of determining rates of vertical move-
~:
=2km
Us =1km U: =O.5km U: =Okm =1km
2 ment become more indirect and less precise. Historical and
archaeological evidence can be valuable in certain circum-
stances. Surface uplift in Iraq around AD 100-200 has been
E inferred from the breaching of an irrigation canal, and a
~
c
similar instance has been reported of the disruption of an
.g 0 Inca canal in Peru. In the coastal regions of Greece, a tec-
co
Q)
> tonically active area, rapid vertical movements have been
W
-1 documented from the displacement of buildings and other
constructions built. in antiquity. Discrete phases of surface
uplift and subsidence have been inferred from the displace-
-2
ment of the northern end of the ramp (didkos) used by the
Greeks to haul ships across the Isthmus of Corinth.
-3 Geomorphic evidence of vertical movements is represented
I by dated landform features which can be related to a pre-
Fig. 15.2 Schematic illustration of the distinction between
swface uplift and crustal uplift. D designates the depth of vious absolute or relative elevation. Such evidence might
denudation and U c and Us, respectively, the amount of crustal include dated basalt flows displaced by faulting or elevated
and swface uplift. Note that if local (Airy) isostasy is to be wave-cut piatforms or beaches which can be dated from
maintained then denudation of a given amount will be their associated deposits. One of the best-documented uplift
accompanied by crustal uplift of around 80 per cent of the depth
of denudation (see Figure 4.7.) histories in the world is to be found in the marine terraces
of the Huon Peninsula, Papua New Guinea. Here a series of
coral terraces has been cut by high sea levels during the
Late Pleistocene as the coastline has risen rapidly. The rates
10 9
of uplift can be established because the age of each coral
....J terrace can be determined by radiometric dating (see Sections
8 <X:
10
U
15.2.2 and 17.2.1).
c.::>
Geological evidence for uplift is usually relevant to
7
10
0 periods of millions of years and is established from the
....J elevation of rocks of known age formed at, or below, sea
6 0
10 LU level, and from a variety of dating techniques which enable
c.::> estimates to be made of the depth of burial of a rock at a
-'c" 105 known time in the past. The present-day height of marine
'"a. deposits above present sea level can be used to calculate a
(f)

Ql 10 4
long-term mean rate of surface uplift if the sea level at the
E time of the formation of a deposit is known. Such evidence,
r- U
3
....J however, only provides an estimate of net surface uplift since
10 I: <X:
a. ~
there could have been intervening periods of subsidence.
a:: a:: Increasing use is now. being made of radiometric dating
2
0 0
10
~ to estimate long-term rates of rock uplift relative to the

rn
2 (/)
0 -
I:
landsurface (Fig. 15.4(A)). The 'clocks' of radiometric
10' LU dating procedures, such as K-Ar and Rb-Sr dating, are only
(:J
set once the rock has cooled below a certain temperature,
10° known as the closure temperature; this is the point at
Fig. 15.3 Appropriate time scales for different types of evidence which the system becomes closed and the products of
used to determine rates of uplift. radioactive decay are retained (see Appendix B). If the
geothermal gradient (typically 20-30°C km- I in the upper
crust) is known, the depth below the surface at which the
glacial sea level rise (currently about 1 mm a-I). Tide gauge radiometric 'clock' started can be calculated. The radio-
stations are relatively numerous in some areas of the world. metric age of the rock then gives the time taken for denu-
In a few instances their records extend back into the last dation to remove the overlying strata and expose the dated
century, but the quality of some of these early records is too rock at the surface. This information provides us with a
poor for their use in precise calculations of surface uplift or mean denudation rate, and if we assume that there has been
subsidence rates. no change in surface elevation then this will be equal to the
374 Endogenic-exogenic interactions

A
Denudation Denudation Denudation

+ + + . . - Landsurface

I
0.>

....:5ro Mineral
exposed
~ at surface
0-

~
"-
E Rock
uplift j Rock
uplift j 0
Rock
upli~ j
.c
.... I
0- I

___1
0.>
0 I
C)
c::
.iii
I
ctl
---- -rL- ..-Critical

-
0.>
C3 I
Mineral at temperature/
critical depth atwhich
.£ Mineral ~~J temperature radiometric
to be 'clock' is set
dated

B
Depth Temp
IDenudation Denudation Denudation
(km) °C
o- 15 + + + . . - Landsurface
Apatite exposed
at surface

1 - 40 Rock
uplift j Rock
uplift j 10
Rock
uplift j
2 - 65

3 - 90
Typical closure
3-4 ----
Fission
---= Fission tracks
. . - temperature
for apatite
tracks start to be (~100C)
4 - 115 • not preserved preserved

Fig.15.4 Schematic illustration of the use of radiometric dating (A) and fission track dating (B) in the estimation of denudation and
uplift rates relative to the landsurface. In (A) the age of a mineral at the surface is determined by the time elapsed since it
passed through the critical isotherm at which its radiometric'clock' was set. This isotherm, and therefore depth, varies for different
minerals. The rate ofdenudation is calculated by dividing depth (D) by age. In (B) the estimation of denudation using fission track
dating of apatite is illustrated. If we assume that the geothermal gradient is 25°C km- I (as indicated) and the age of apatites exposed at
the surface is 10 Ma then the mean denudation rate is 3400 (m)110 = 340 m Ma- I .

mean crustal uplift rate relative to sea level. Note that, used denudation arid crustal uplift rates are likely to be approx-
alone, this technique does not provide any direct inform- imately equal for a significant period of time.
ation on the amount of absolute uplift that has occurred. An elaboration of this technique is to compare the radio-
The estimation of crustal uplift rates using radiometric metric ages of two different minerals which start their radio-
dating is most appropriately applied to orogenic belts since metric clocks at different temperatures. The radiometric
it is here that metamorphic events assQciated with signifi- clock for the mineral muscovite, for instance, starts about
cant heating of the crust are most likely to 'reset' the radio- 200°C higher than the radiometric clock for biotite. If the
metric clock by overprinting the original radiometric dates geothermal gradient is 20°C km- 1 this is eqivalent to a
established when the rocks were first formed. This enables difference in depth of 10 km, so the difference in age be-
a rock uplift rate relative to the landsurface to be esti- tween the two minerals when exposed together at the sur-
mated for a short period of a few million years since the re- face represents the time taken for 10 km of denudation to
setting event. It is also only in active orogenic belts that occur. Again, assuming a rough equilibrium between crustal
 Rates of uplift and denudation 375

uplift and denudation rates, this provides an estimate of the 15.2.2 Spatial and temporal variations
former.
From our discussion of the mechanisms of orogeny and
Another technique based on a similar principle is fission
epeirogeny in Chapters 3 and 4 we would expect consider-
track dating (Fig. 15.4(B)). The fission of uranium-238,
able differences in their associated rates of uplift. Although
which is relatively abundant in minerals such as apatite
this is clearly the case in the long term, evidence from rele-
found in granites and other basement rocks, involves the
velling surveys provides some equivocal data on rates of
break up of the original nucleus into two fragments of
uplift in regions affected by epeirogenic movements.
roughly equal mass. These charged particles recoil from
each other and move in opposite directions thereby creating
microscopic dislocations or 'tracks' in the crystal lattice. 15.2.2.1 Orogenic uplift
The number of tracks formed is a function of the concen- In active orogenic zones significant vertical movements of
tration of uranium present and the time elapsed since tracks the surface up to several metres may occur in a single earth-
began to be preserved. Determining the uranium concentra- quake. In the major Alaskan earthquake of 1964, for example,
tion and counting the number of microscopic tracks formed an area of some 400000 km 2 was raised by an average of
enables an age to be estimated. The particular advantage of 2 m and a maximum of 12 m. Although there may be rapid
the technique for estimating rock uplift relative to the land- variations in rates of vertical displacement in space and
surface and denudation rates is that for the mineral apatite time along convergent and transform plate margins, high
the fission tracks are destroyed, or annealed, at tempera- mean rates of uplift can be sustained. for periods of hun-
tures above about l00±20 0c. The fission track closure dreds of thousands of years (Table 15.1). An extreme
temperature for apatite is, therefore, located at a relatively example is the Ventura Anticline at the western margin of
shallow depth in the crust corresponding to this temperature. the Transverse Ranges in southern California which has
Depending on the geothermal gradient, this fission track experienced crustal uplift at an average rate of at least
annealing zone for apatite usually lies at a depth of be- 16000 m Ma -lover the past 1 Ma. Rather more modest rates
tween 3.5 and 5.5 km. of surface uplift of up to 3300 m Ma - 1 have been recorded
As with the radiometric methods already discussed, this from the elevated coral terraces of the Huon Peninsula (Fig.
technique can only provide estimates of crustal uplift rates 15.5).
indirectly through the calculation of denudation rates. Out- Minimum crustal uplift rates in major orogenic belts,
side active orogenic belts where these two rates will rarely such as the Alps and the Himalayas, located at convergent
be similar, the main use of both radiometric and fission track plate margins range from 300 to about 800 m Ma- I aver-
techniques is in the estimation of denudation rates (see aged over periods of several million years (Table 15.1). But
Section 15.4). rates may be much higher than this; in southern Tibet, for

Table 15.1 Long-tenn mean uplift rates in orogenic zones determined by various methods

LOCATION METHOD RATE PERIOD SOURCE

(mMa- l )

Central Alps Apatite fission 300-600* 6-IOMa BP Schaer et al. (1975)

track age Tectonophysics 29. 293-300
Central Alps Rb and K-Ar 400-1000* 10-35 MaBP Clark and Jager
apparent ages of (1969) American]oumal
biotite of Science 267. 1143-60
Kulu-Mandi Belt. Apparent Rb-Sr ages 700* 25 MaBP- Mehta( 1980)
Himalayas of coexisting Present Tectonophysics
biotites and muscovites 62.205-17
Southern Alps, Apparent K-Ar ages 10000* I MaBP- Adams (1981)
New Zealand of schists Present Geol. Soc. Lond.
Spec. Puhl.
9.211-22
Southern Alps. Estimated ages of 5000-8000t 140kaBP- Bull and Cooper (1986)
New Zealand elevated marine Present Science 234. 1225-8
terraces
Huon Peninsula. V-series and 14C 1000-3000 t 120 ka BP- Chappell (1974)
Papua New Guinea dating of elevated Present 1. Geophys Res.
marine terraces 79.456-64

* Rock uplift (relative to surface)

t Surface uplift.
376 Endogenic-exogenic interactions

Indian Plate

Subduction zone oL..--.....J

200
Thrust fau~s km

Strike-Slip faults
Spreading centre

400

E 0 5
"Q)300 I I

km
>
Q)
....J
CO
Q)
CJ)
....
a3 200
rJ)
Q)
L..
Q..
Q)
>
0
~ 100
rJ)
c
0
'';: Reef 1130 ka B.P
co
>
Q)
u::i 0 -----
2 3 4 5 6 7 8 9 10 11
Site Number

Fig.15.5 Tectonic setting and surface uplift history of the terraces of the Huon Peninsula, Papua New Guinea, determinedfrom the
14C and U-series dating ofwave7cut notches formed during Late Pleistocene sea level highstands. Uplift rates can be seen to have
varied along the coast with much higher rates towards the south-east. (Based on J. Chappell (1974) Journal of Geophysical Research 79,
Fig. 2, p. 392. Copyright by the American Geophysical Union.)
 Rates of uplift and denudation 377

instance, Late Pliocene - Early Pleistocene terrace deposits rate of crustal uplift averaging up to 10000 m Ma- 1 over
containing a fauna indicative of a lowland sub-tropical the past 1 Ma. This extraordinarily high uplift rate is, as in
climate have since been elevated to a height of 4000-5000 m the case of the Transverse Ranges of southern California,
indicating surface uplift averaging more than 2000 m Ma- 1• related to the oblique convergence occurring along the associ-
Overall, crustal uplift rates in the Himalayas appear to be ated plate boundaries (see Section 3.5.2).
currently averaging around 5000 mm ka- I . This is matched For some mountain ranges the variations in local rates
by long-term rates in some Andean ranges, such as the of surface uplift correspond well with their present-day
Cordillera Blanca in Peru which, on the basis of the expo- topography; that is, the regions of highest elevation are rising
sure of a granite batholith emplaced at a depth of 8 km about at the greatest rate (Fig. 15.6). In other orogenic belts there
10 Ma BP, has experienced crustal uplift of around 4000- is a mismatch between topography and present patterns of
5000 m Ma- I since that time. vertical movement. This is the case for the Alps where
Even these rates, though, are modest in comparison with detailed relevelling surveys documenting vertical movements
those estimated for sections of the Southern Alps in New over the past 60 a or so show that the current maximum
Zealand. This mountain range, located along the boundary rates of surface uplift are to be found in parts of the allu-
of the Pacific and Indian Plates, has apparently sustained a vium filled valleys ofthe Rhone and the Rhine. This apparent

CENTRAL HIMALAYAS
~
S ~ ~ N
km
10 g)
etl
u.
5
to ti
§ 0) c: c: :J "~ 2
";: :; 16"iii" co ..c ro.L:
~ c::> ~ ~~ ~~
11> I I
iJ:jOF======~=:::::::::::=-----------------------

mmka-1
+5000
.....
~
Q.
::J
'0 Of------~~----------------------------
2ro
a:::
-4000

CENTRAL ANDES
w E
km

[~t____ j J
mmka-1
+5000
~
Q.
::J
'0 Of-------------:::::o~---------------------==-----~~------
~
ro
a:::
-4000
Horizontal Scale
o km
100

Fig. 15.6 Relationship between topography and surface uplift rates along transects across the central Himalayas and central Andes.
(Modifiedfrom A. Gansser (1983) in: K. 1. Hsu (ed.) Mountain Building Processes. Academic Press, London, Fig. 2, p. 223.)
378 Endogenic-exogenic interactions

anomaly may be at least partly explained by the continuing Such data suggest remarkably high rates of surface uplift of
isostatic compensation for the removal of the ice loads up to 20 mm a-I at the present time in regions such as
occupying these valleys during the last glacial maximum. Florida. Thes'e rates exceed even those in the world's most
active orogenic belts so we are led to ask whether they are
15.2.2.2 Epeirogenic uplift credible or whether they are simply an artefact of system-
Perhaps because of the less obvious vertical movements in- atic errors in the levelling measurements. Although in most
volved, data on rates of epeirogenic uplift are rather sparse, cases the changes in elevation recorded exceed the random
at least for the long term. Rates for uplift averaged over error limits expected in such surveys, ~ystematic errors cer-
millions of years appear to lie in the range of 1~200 m Ma- 1• tainly cannot be ruled out. On the other hand, the patterns
The Colorado Plateau in the western USA, for instance, has of apparent movement observed along the eastern margin
risen about 500 m during the Late Cenozoic at a mean rate of the USA correspond to known tectonic structures and
of around 100 m Ma- I, while the Deccan Plateau of India various geomorphic features such as drainage divides and
has experienced uplift of 600 m or so extending over the stream gradient anomalies.
past 40 Ma at a mean rate of about 15 m Ma- I. If such movements are real the problem remains of find-
Although we would expect to see short-term rates that ing a plausible mechanism to explain them. It is obvious
Me similarly slow, this is apparently not the case. Detailed that such high rates in 'passive' tectonic settings of gene-
relevelling surveys carried out in the USA, for instance, rally subdued relief cannot be sustained for long periods of
demonstrate that over the past several decades there has time, otherwise they would generate a quite unrealistic
been little to distinguish apparent rates of vertical crustal topography (a sustained uplift of 20 mm a-I would give rise
movement in the tectonically active western USA from the to a 2000 m high mountain range in just 100000 a assum-
supposedly stable eastern part of the country (Fig. 15.7). ing no erosion). This implies that such epeirogenic movements

C1 mma-1
30

Ridge and km

8
2
(l)
0>
c
~ 20
u
App",,,,hi,,
Plateaus
j
Valley
Piedmont
2
c
0
'+=,
~
Q)
~,'- ......
j j
w
'0 10 '\\ " Coastal Plain
,.,/ \_.. . ./. , 1
*
II:

~
.;:;
co
\\
~v ' ......
...... ~
Cii 0
II:
0 500 1000 1500 2000

A Distance (km) A
1 2

mma- 1
30 San
Joaquim
- Basin and
.-
Range Province
km
mma- 1
~
Valley Cape Fear Cape
Sierra Arch Canaveral
co
.c Nevada km
U 20
•
2
Chesapeake •I / '.. ..-J{V\
V" 1)' 2
§
.;:;
co

w
iii
EmbaVr;.I/rv "J-
'0 I\..JV-\
~
(l)

II:
10 '-" -"
\r~
\J '\
*
0: ? ._-'-------/
,..
I'J-../ ' ....
I'

~ ~ 0 _,..--J o
~ ~~, .~

& ~=-_----I. ----'-- -'-_ _---'O Cii

0
o 500 1000 1500 2000 II: 500 1000 1500 2000 2500 3000 3500

8 Distance (km) 8 C1 Distance (km)

1 2
Fig.15.7 Apparent present rates of vertical movement for various transects in the USA based on precise levelling data. Solid lines
indicate topography, dashed lines apparent vertical movement. Note that rates of movement are expressed in mm a-I. (Based on L. D.
Brown et al. (1980) in N.-A. Morner (ed.) Earth Rheology, Isostasy and Eustasy. Wiley, Chichester, Fig. 1, p. 391 and Fig. 10, p. 399;
L. D. Brown (1978) Tectonophysics 44, Fig. 2, p. 211; and L. D. Brown and J. E. Oliver (1976) Reviews of Geophysics and Space
Physics 14, Fig. 11, p. 21.)
 Rates of uplift and denudation 379

are episodic or oscillatory, with periods of uplift alternating ling the rate at which landscapes are eroded. An assessment
with periods of subsidence. Although episodic or oscilla- of fluvial denudation rates is important because the output
tory movements must be related in some way to epeiro- of sediment and dissolved constituents from tJW usually
genic mechanisms (along passive continental margins) or well-defined boundaries of a ~rainage basin enables us to
plate interactions (along active plate margins), the specific monitor the net effects of weathering, slope processes and
processes involved are unknown. Nevertheless, the con- fluvial transport across a specific area. Defining source
siderable areal extent of the regions affected by these rapid regions for the removal of material by glaciers, and especially
epeirogenic movements (up to 1000 km across) implies that wind, is far more difficult. Perhaps the greatest limitation in
the cause cannot lie just in the crust but must be deep- using present-day denudation rates as a basis for under-
seated and probably involves the entire thickness of the standing long-term landform development is that we cannot
lithosphere and even possibly a part of the asthenosphere. be sure how typical they are. This is why" it is aiso vital to
Although there is great uncertainty about the cause of examine other evidence for long-term rates of denudation.
short-term rapid uplift in many regions of the world, far
more is known about the extremely high rates of uplift re-
corded from areas once covered by Pleistocene ice sheets. 15.3.1 Methods of Measurement and Estimation
In fact the relationship between the temporal and spatial The basis for determining fluvial denudation rates is the
variations in rates of isostatic rebound and the thickness estimation of the solid (sediment) and solute (dissolved)
and location of the Pleistocene ice sheets has provided vital load carried by rivers (see Section 804.1). A very approxi-
information about the nature of the sub-lithospheric mantle mate estimate can be achieved simply by multiplying the
and the way in which it responds to the removal of a load. mean sediment and solute concentration calculated from a
Detailed information on the decay of the great ice sheets of small number of samples by mean discharge. More accurate
the last glacial and the resulting isostatic rebound of the estimates, however, must take into account the way in
crust is provided by raised shorelines containing datable which the concentrations of sediment and dissolved
deposits. As can be seen in Table 15.2, such evidence constituents vary with discharge and, in particular, the way
demonstrates that the highest rates of surface uplift are in which they vary with flood events. This is accomplished
attained by crust lying below the thickest part of the ice by using sediment and solute rating curves constructed
sheet (where unloading generates the greatest degree of from equations which describe the best fit relationships
isostatic disequilibrium). It also shows that rates of isostatic between sediment and solute concentrations and discharge.
rebound decline exponentially through time, being extremely For greater accuracy separate rating curves can be used for
rapid immediately after deglaciation, with rates of up to rising or falling stage relationships and for seasonal flows.
100000 mm ka- 1. Somewhat slower rates of surface uplift, Solid and solute transport rates can then be calculated by
at around 5000-10 000 mm ka- 1, continue for at least a fur- relating the rating curve to either continuous stream-flow
ther 15000 a in the case of major continental ice sheets. data, or flow-duration curves based on hourly, daily or even
monthly data. Such a procedure is subject to errors,
particularly where the frequency of sampling is low. In small
Table 15.2 Post-glacial rates of vertical movement (mma- 1) in regions
covered by Late Pleistocene ice sheets basins transport rates can vary significantly from hour to
hour, and even in large basins a high degree of variability
LAURENTIDE FENNO-SCANDIAN SCOTrISH can occur between individual years.
- - - - - - - CENTRE CENTRE
Since data on sediment and solute transport rates are not
CENTRE EASTERN
PERIPHERY available for all drainage basins, estimates of denudation
rates on a continent-wide or global basis must be founded
10-8 ka BP 100-70 70 30 10 on some form of extrapolation. This is usually based on
4-3 kaBP 30 3-S 10-IS 4-S empirical relationships observed between measured solid
Recent decades >(S±2)
and solute transport rates and the factors thought to control
(20±S?) -(2-3) 9-10 3.8-S.8
these rates, especially those related to climate and relief.
Note: Post-glacial eustatic rise in sea level taken into account. Where such relationships are found to be strong it is then
Source: Data"from A. A.Nikonov (1980) in: N.~A. Momer (ed.) Earth
possible to estimate the sediment and solute transport rates
Rheology, Isostasy and Eustasy. Wiley, Chichester, Table 2, p. 347
compiled from various sources. on the basis of climate and relief, and, in some cases, other
variables.
This approach has been much favoured by some French
15.3 Present fluvial denudation rates geomorphologists. For instance, in an extensive survey
The analysis of contemporary rates of denudation raises major published in 1960, F. Fournier estimated the global pattern
difficulties, but it provides the only means we have of of denudation by relating sediment yield data on 78 basins
determining which factors are most significant in control- (ranging in area from 2460 to 1 060 000 km 2) in a variety of
380 Endogenic-exogenic interactions

climatic zones to a seasonality of rainfall index. This was Table 15.3 Comparison of sediment yields (t km-2 a-I)
expressed as p2/P, where p is the mean monthly maximum under natural and artificial conditions in various
countries
precipitation and P is the mean annual precipitation. The
idea of such a seasonality of rainfall index is that it indicates LOCATION NATURAL CULTIVATED BARE SOIL
the importance in sediment transport of peak discharges LAND

generated by intense seasonal rainfall. Fournier also incor-

UK 10-50 10-300 1000-4500
porated the effect of relief by using separate regression USA 3-300 500-17 000 400-9000
equations for different relief types. Because the relation- China <200 15000-20000 28000-36000
ships between sediment and solute transport rates and the India 50-100 30-2000 1000-2000
Nigeria 50-100 10-3500 300-15000
various climatic and relief variables used are not perfect, Ivory Coast 3-20 10-9000 1000-75000
such an approach can only give an approximate picture of
the variation in denudation rates. Source: Data from R. P. C. Morgan (1986) Soil Erosion and
Conservation. Longman, London, Table 1.1, p. 5, based on
Additional information on solid and solute loads for various sources.
major river basins is constantly accumulating, and most
estimates of contemporary global denudation are now in- the Amazon River. In mountainous terrains, though, bed-
creasingly based on these larger data sets rather than extra- load transport can be very significant and in such environ-
polation from results for a limited number of basins. Never- ments the widely applied assumption that bed load is about
theless, there remain significant difficulties in converting 10 per cent of suspended load may be a drastic underestimate.
solid and solute load data into meaningful estimates of Another problem in estimating natural rates of denuda-
denudation rates. A general problem is simply the unrelia- tion and in evaluating the factors that control them is that in
bility of measurements for many of the world's major many river basins sediment yields have been changed
rivers. The data on solid and solute load are inadequate for dramatically, at least at the local scale, through anthropo-
basins such as the Amazon, Ganges and Brahmaputra which genic effects. Dam construction promotes a downstream
contribute large quantities of both sediment and dissolved reduction in sediment yield, but cultivation, surface mining
constituents to the world's oceans. The data for Chinese and urban construction all lead to disturbance of the soil
rivers, which are also important sediment contributors, also and consequently can give rise to sediment yields far in
used to be poor but are now much improved following the excess of those generated from naturally vegetated surfaces
establishment of comprehensive monitoring programmes (Table 15.3). It has been estimated, for instance, that the
for several major basins. In some cases it is even difficult to conversion of forest to cropland in the states of the mid-
find out details of the method of measurement used. A Atlantic seaboard of the USA has led to a tenfold increase
widely cited estimate of the sediment yield for the Irrawady in sediment yield, but such increases within a basin are not
River in Burma is in fact based on a few measurements necessarily reflected immediately by an increase in sedi-
taken by unknown methods·in the 1870s. Even where modem ment load at the basin terminus. In the case just cited, for
techniques of measuring discharge and of sampling sedi- example, it has been calculated that over 90 per cent of the
ment and solute concentrations have been applied, there sediment stripped from the upland regions of the Piedmont
may still be significant errors in rating curves and insuffi- region of the south-east USA remains stored as colluvium on
cient information on the monthly and annual variability of hillslopes, and as alluvium In valley bottoms. In other words
discharge and the occurrence of rare peak discharges which there is not a steady state between sediment supply and
may be particularly important in affecting the overall rate sediment transport. A. similar situation exists in the upper
of sediment transport. Moreover, if there is no gauging Mississippi Basin, where a detailed sediment budget study
station at the mouth of a river the total load transported to has demonstrated that only 7 per cent of the sediment being
the ocean may be overestimated as there may be significant removed from the upper parts of a tributary valley is cur-
deposition or precipitation downstream of the sampling point. rently reaching the Mississippi River (Fig. 15.8).
Other important problems relate specifically to the accurate The figure of 7 per cent indicates the proportion of sedi-
estimation of solid and solute load, and we will now exam- ment being mobilized within the basin that actually reaches
ine these in some detail. the basin outlet and represents the sediment delivery ratio
for the basin. This is a central concept in understanding the
15.3.1.1 Solid load dynamics of drainage basin erosion through time. In the
In most measurements of solid load transport only the sus- long term the sediment delivery ratio must be approximate-
pended sediment is sampled and, because of the difficulties ly unity otherwise the river basin would become progress-
involved with large rivers, no attempt is made to measure ively choked with sediment. (In very large basins it may be
bed load. In lowland tropical rivers this may not be a major less than unity, since sediment may progressively accumu-
problem; for example it has been estimated that bed load late in the basin as the basin floor subsides under the load
accounts for only 1 per cent of the total sediment load of of overlying sediment.) In the short term, however, there
 Rates of uplift and denudation 381

SOURCES (100 %)

Upland sheet
and ri II erosion
77%

Upland
gullies
12 %

Tributaries
6%
Colluvium
55%
Upper main
valley 5%

Upland
valleys
7%

Upper
main valley
5%

Lower
Sediment yield to main valley
Mississippi River 26%
7%
SINKS (100 %)

Fig.IS.8 Sediment budget from 1938-75 for Coon Creek, a tributary of the Mississippi River in Wisconsin, USA. Thefigures show
the percentage of sediment mobilized from different sources within the basin and the percentage deposited in different temporary
storages (sinks). (Based on datafrom S. W. Trimble (1983) American Journal of Science 283, 454-74.)

may be transfers of material within a drainage basin but capacity in the form of extensive floodplains. An additional
little export of material from it. Debris removed from slopes factor will be basin morphology with mountainous catch-
·can be stored in fans, talus slopes and as river terraces and. ments generally having a smaller sediment storage capacity
floodplain sediment. than lowland basins. Eventually this stored sediment will
The disparity between the amount of erosion occurring be mobilized and transported out of the basin, but this may
on hillslopes within a basin and the total amount of sedi.- take decades or centuries, or, in the case of very large basins,
ment being removed from it will tend to become more many thousands of years or more. Glacial and talus deposits
marked as drainage basin area increases, since larger generated in mountain valleys during the Late Pleistocene
catchments are likely to have a greater channel storage have still not been removed from many mid-latitude basins,
382 Endogenic-exogenic interactions

while in the Amazon Basin only the major rivers are cur- HC03" which arises from the incorporation of atmospheric
rently able to rework the coarser alluvium that was depo- CO 2 during weathering reactions. Together these non-denu-
sited during the more arid phases of the Quaternary. dational components on average amount to about 40 per
cent of the solute load of rivers but there are major depar-
15.3.1.2 Solute load tures from this figure. In regions of rapid rock weathering
In view of the complexity of measuring separately each dis- in highly reactive lithologies the non-denudational compo-
solved constituent in stream water, the easily measured nent is relatively minor, but in some tropical lowland rivers
electrical conductivity, or specific conductance, of the water draining thoroughly leached and almost chemically inert
is often used to provide an estimate of the solute concen- weathering mantles the non-denudational component may
tration. Although there is a strong correlation between the be very significant. In the Rio Ucayali in the mountainous
concentration of ionic species in solution and electrical Andean region of the Amazon Basin it is estimated that
conductivity, the exact relationship varies depending on the only 4.8 per cent of Na+, K+, Ca2+ and Mg2+ comes from
concentrations present of particular dissolved constituents. precipitation but in the Rio Tefe, a lowland tributary of the
Moreover, Si0 2, which is a significant component of many Amazon, 81 per cent of these four constituents is contri-
tropical lowland rivers, is not recorded by this technique. buted by precipitation. Apart from their Si0 2 content the
The solute load of rivers can also be enhanced by anthro- chemical composition of many of ·the lowland Amazon
pogenic inputs. This is especially the case in basins con- tributaries is in fact very similar to extremely dilute sea
taining industrial activities, and also in agricultural regions water, a reflection of the predominance of precipitation-
where the fertilizers applied to crops can find their way into derived solutes. An additional consideration relevant to many
stream waters in significant quantities. Human activities can different environments and time scales is that the denuda-
also lead to an increase in rates of chemical weathering and tional component of the solute load' can also be under-
thereby enhanced rates of solute input into stream waters. estimated if solutes taken up in vegetation are removed
The most important problem in estimating the contribu- from the basin as litter in streams.
tion of solute transport to total denudation is the separation
of the denudational and non-denudational components. Allow- 15.3.1.3 Estimation of volumetric changes
ances must be made for dissolved constituents introduced A rather neglected problem concerning the estimation of
into a basin through precipitation, since these represent a denudation rates from solid and solute load data is the con-
non-denudational component additional to the solutes released version of a measurement expressed as a mass per unit area
through bedrock weathering. An adjustment can either be per unit time (usually t km-2 a-I) to a volumetric equiva-
made by using a knowledge of the chemical and minera- lent. Since we are concerned with the change in the form of
logical composition of the lithologies exposed in a basin to the ground surface through time, it is the volumetric change
predict the solutes likely to be derived from bedrock wea- which is of importance. At first sight this seems a trivial
thering or, for greater accuracy, the chemical composition task; we simply divide the mass of material by the density
of the precipitation can be measured directly. of the bedrock (say 2700 kg m-3 for rocks making up most
Table ·15.4 gives the average composition of world river of the upper continental crust) to give a volume (m 3 km-2
water and shows that two major non-denudational compo- a-I). We then convert this to an average rate of lowering
nents introduced by precipitation are Na+ and Cl-, the two for the whole basin, acknowledging, of course, that the
most abundant dissolved constituents in sea water. An even actual change in elevation will inevitably vary considerably
more important non-denudational component, however, is from place to place. Conveniently, a volume change express-

Table 15.4 Average composition of world river water and estimates of denudational and non-denudational contributions for different constituents

CI so~- HCO; Total

Average composition of 13.5 3.6 7.4 1.35 9.6 8.7 52.0 10.4 106.6
world river water
(concentration (mg I-I))

Provenance of major solute

components (0/0)
Non-denudational:
Precipitation (oceanic salts) 2.5 15 53 14 72 19 12
Atmospheric CO 2 57 28
Denudational:
Chemical weathering 97.5 85 47 86 28 81 43 100 60

Source: Based largely on data in M. Meybeck (1983) in: Dissolved Loads ofRivers and Surface Water Quantity/Quality Relationships International
Association of Hydrological Sciences Publication 141, 173-192.
 Rates of uplift and denudation 383

ed in m 3 km-2 a-I is equivalent to a mean rate of ground Where solutes corne from the direct solution of bedrock, as
lowering expressed as mm ka- I. The averaging over a occurs to a large extent in limestone terrains, this is a valid
period of 1000 a is largely conventional and has the advan- procedure. But in many cases bedrock weathering takes
tage of giving an amount of lowering that can be readily place without any change in volume. Where such iso-
appreciated. It does not imply that the rate of denudation volumetric weathering occurs (see Section 6.2.4.1) solutes
has been measured over this length of time, and this is, of are lost to stream waters, but there are no associated
course, not the case when using stream load data. volume changes as the transformation of fresh bedrock into
Unfortunately, this simple procedure does not necessarily saprolite composed largely of clay minerals is accompanied
yield an accurate estimate of the equivalent average volume by a compensating decrease in bulk density. Even where
change actually occurring in the landscape. If soil or wea- weathering is not strictly isovolumetric there is almost
thered material, which may have a bulk density between invariably a significant compensation through a decrease in
1100 and 2000 kg m-3 , i$ being eroded, is it appropriate to bulk density associated with the removal of dissolved con-
use the density of the underlying bedrock in calculating the stituents from the weathering mantle. Consequently, in
associated volume change? The answer is yes if the rate of lowland regions mantled by thick weathering profiles, it is
rock weathering and removal of weathered material and misleading to view the removal of weathering-derived
soil are more or less in a steady state, but our understanding solutes in the short term as necessarily being directly related
of the triode of landscape modification in many different to chemical denudation in. the sense of a reduction of the
geomorphic environments, and especially those of lowland average elevation of a basin. In considering the rates of
tropical regions of subdued relief, suggests that periods mechanical and chemical denudation discussed in the fol-
characterized by low denudation rates, during which the lowing section, it will be important to keep these points in
weathering mantle increases in depth, are punctuated by mind.
phases of active erosion. In such cases rock weathering and
weathering mantle erosion are not in equilibrium, and in
this situation it is not clear that converting sediment re- 15.3.2 Rates of mechanical and chemical denudation
moval into a rate of ground lowering on the basis of the ori- Table 15.5 lists a range of estimates of total solid and solute
ginal bedrock density is meaningful. load transport to the world's oceans and gives the equiva-
This problem is compounded when we consider chemical lent global mean denudation rates (assuming an average
denudation. Conventionally, the specific solute load of streams source rock density of 2700 kg m-3 ). It is important to point
measured in t km-2 a-I is converted into a rate of ground out here that these, and all other estimates of denudation
lowering in mm ka- I using an appropriate bedrock density. rates, are subject to variable, and often large, errors. Al-
though rates may be reported to a precision of one decimal
place, this certainly does not imply that they are accurate to
Table 15.5 Estimates of total transport by rivers of solids and solutes to that degree.
the oceans and equivalent estimated denudation rates The estimates in Table' 15.5 are based on the total land
area of 148 x 106 km 2 and would be increased by about 40
AUTHOR MEAN LOAD EQUIVALENT
DENUDATION RATEt per cent if only the global area of external drainage (about
(109ta- l ) (t km- 2 a-i) (rom ka- I ) 105 x 106 km 2) is considered. The various estimates in
Table 15.5 are not directly comparable as the basis for their
Solidload* Mechanical
Fournier (196Q) 58.1 392.6 145.4 calculation differs, but it appears that the early estimate of
Jansen and Painter 26.7 180.4 66.8 Fournier, based largely on his suggested relationship between
(1974) sediment yield and seasonality of precipitation, is too high.
Schumm (1963) 20.5 138.5 51.3
Holeman (1968) 18.3 123.6 45.8 The more recent low estimate of Milliman and Meade, on
Milliman and Meade 13.5 91.2 33.8 the other hand, is for actual solid load transport to the oceans
(1983) at the present day, and includes the effects of sediment
Lopatin (1952) 12.7 85.8 31.8
entrapment by darns on rivers such as the Colorado, Nile
Solute load Chemicat* and Zambezi.
Goldberg (1976) 3.9 26.4 5.9
Livingstone (1963) 3.8 25.7 5.7 What, then, is the best estimate we can make of current
Meybeck (1979) 3.7 25.0 5.6 global denudation rates excluding the effects of human
Meybeck (1976) 3.3 22.3 5.0 activities? Taking the Milliman and Meade estimate of
Alekin and Brazhnikova 3.2 21.6 4.8
(1960) 13 500 Mt a-I for suspended sediment transport, we can add
on 500 Mt a-I as a reasonable estimate of the material trapped
* Suspended load only. by darns and 1500 Mt a-I to allow for unrecorded bed-load
t Denudation rates based on a rock density of 2700 kg m- 3 .
:j: Rates for chemical denudation assume that 40% of total solute load is transport. In addition we perhaps should allow 500 Mt a-I
from non-denudational sources. as a deduction to take account of the increase in erosion
384 Endogenic-exogenic interactions

rates as a consequence of human activities. These are Such a global mean conceals great variations from area
admittedly rough approximations, but the total this gives of to area and basin to basin. Around 70 per cent of the total
15000 Mt a-I is probably a reasonable estimate of natural load transported to the oceans is provided from only 10 per
solid load transport to the oceans. cent of the land '!rea and just three rivers, the Ganges, the
Estimates for global solute travsport to the oceans are Brahmaputra and the Huang He (Yellow) carry 20 per cent
less variable, and we will take the recent figure established of the global fluvial sediment load. Figure 15.9 shows the
from a detailed global survey by M. Meybeck of 3700 Mt a-I results ot' a global survey of sediment yields based on mea-
as being the most accurate available. From this we need to surements from more than 1500 sites and indicates that
make a deduction to allow for the non-denudational com- specific sediment yield rates exceed 1000 t krn-2 a-I (rough-
ponent. As a global average 40 per cent is probably a rea- ly equivalent to a mechnical denudation rate of 370 rnrn ka- I)
sonable figure and this gives an estimate of 2200 Mt a-I for in the world's mountainous regions. By contrast, rates
'denudational' solute load transport. This gives a global are less than 50 t krn-2 a-I (equivalent to approximately
mean annual transport of 17 200 Mt of material to the oceans 19 mm ka- I) in many lowland regions.
which, averaged over the entire land area of the continents, If we examine the world's largest drainage basins we
gives a rate of 116 t km-2 a-I. Assuming a mean source rock find that present-day total denudation rates range from
density of 2700 kg m-3 , this converts to a mean global 3 mm ka- 1 for the interior Chari Basin in Africa and
denudation rate of 43 mm ka- I. If we exclude areas of 5 mm ka- 1 for the Kolyma Basin in eastern Siberia, up to
internal drainage where material is being transported but 529 mm ka- I for the Huang He Basin of China and
not removed from the continents, this figure rises to 677mm ka- 1 for the Brahmaputra Basin draining the eastern
61 mm ka- 1• Of this total about 85 per cent is accounted for Himalayas (Table 15.6, Fig. 15.10). These figures, how-
by solid load transport and 15 per cent by the transport of ever, by no means represent the extremes found in smaller
solute load. It is reasonable to assume that, over large areas, basins. Minimum sediment yields lie well below 2 t km-2 a-I
denudation will lead to a compensatory isostatic rebound (equivalent to <1 mmka- I) (1.7 t km-2 a-I, for instance, for
which reduces the 43 mm ka- I mean rate of lowering with the Queanbeyan Basin (172 km 2 ) in the southern tablelands
respect to sea level to only 8 mm ka- I. of the East Australian Highlands and less than 1 t km-2 a-I

Suspended
Sediment Yield
tkm- 2 a- 1

1000
750
500
250
100
50

~ Deserts and
~ permanent ice

Fig.15.9 Global pattern ofyields of suspended sediment. The values relate to intermediate-sized basins of 1OL 105 km 2 . (After D. E.
Walling and B. W. Webb (1983) in: K. J. Gregory (ed.) Background to Palaeohydrology. Wiley, Chichester, Fig. 4.2, p. 76.)
 Rates of uplift and denudation 385

Table 15.6 Estimated denudation rates for the world's thirty-five largest drainage basins based on solid and
solute transport rates

DRAINAGE RUNOFF TOTAL MECHANICAL CHEMICAL* CHEMICAL

AREA DENUDATION DENUDATION DENUDATION DENUDATION
(106 km 2) (mma- 1) (mm ka- 1) (mm ka- 1) (mm ka- I ) AS % OF TOTAL

Amazon 6.15 1024 70 57 13 18

Zaire (Congo) 3.82 324 7 4 3 42
Mississippi 3.27 177 44 35 9 20
Nile 2.96 30 15 13 2 10
Parana (La Plata) 2.83 166 19 14 5 28
Yenisei 2.58 217 9 2 7 80
Ob 2.50 154 7 2 5 70
Lena 2.43 206 11 2 9 81
Chiang Jiang
(Yangtze) 1.94 464 133 96 37 28
Amur 1.85 175 13 10 3 22
Mackenzie 1.81 169 30 20 10 33
Volga 1.35 196 20 7 13 64
Niger 1.21 159 24 13 II 47
Zambezi 1.20 186 31 28 3 11
Nelson 1.15 96
Murray 1.06 21 13 11 2 18
St Lawrence 1.03 434 13 I 12 89
Orange 1.02 89 58 55 3 5
Orinoco 0.99 1111 91 78 13 14
Ganges 0.98 373 271 249 22 8
Indus 0.97 245 124 108 16 13
Tocantins 0.90 385
Chari 0.88 69 3 2 I 29
Yukon 0.84 232 37 27 10 28
Danube 0.81 254 47 31 16 35
Mekong 0.79 595 95 75 20 21
Huang He
(Yellow) 0.77 63 529 518 11 2
Shatt-el-Arab 0.75 61 104 93 11 11
Rio Grande 0.67 5 9 6 3 38
Columbia 0.67 375 29 16 13 46
Kolyma 0.64 111 5 3 2 31
Colorado 0.64 31 84 78 6. 7
sao Francisco 0.60 151
Brahmaputra 0.58 1049 677 643 34 5
Dnepr 0.50 104 6 1 5 88
* Allowance made for non-denudational component of solute loads.
Source: Based primarily on data from M. Meybeck (1976) Hydrological Sciences Bulletin 21,265-89, and J. D.
Milliman and R. H. Meade (1963) Journal of Geology 91,1-21.

for several rivers in Poland). Maximum sediment yields I t km-2 a-I, whereas maximum yields of 6000 t km-2 a-I
exceed 10000 t km-2 a-I, with the Haast River draining occur in rare instances where rivers drain highly soluble
a region of high precipitation and very rugged relief in deposits such as halite. More usual maxima lying below
the Southern Alps in New Zealand having a yield of 1000tkm-2 a- 1 (equivalent to 370mmka- l ) occur in
12736tkm-2 a-I (equivalent to 4717mmka- I). But even limestone regions. High rates of chemical denudation are
this is exceeded by the 53500 t km-2 a-I of the Huangfu- invariably observed in humid mountainous regions. In such
chuan River, a tributary of the Huang He, which drains environments rates of chemical weathering are high as the
over 3000 km 2 of gullied loess-covered terrain in a region weathering front is at, or close to, the surface, but the
of sparse vegetation with a semi-arid climate characterized accompanying high rates of mechanical denudation prevent
by periodic intense storms. This is equivalent to a denu- the accumulation of thick weathering profiles (see Section
dation rate of 19814 mm ka- I, a figure which will only be 6.2.4.1). Conversely, minimum rates are recorded in semi-
sustained while supplies of readily erodible loess remain in arid regions where runoff is very low (although solute
the catchment. concentrations may be very high), in lowland humid tro-
Chemical denudation rates are in general less variable pical regions where solute concentrations are generally
than those for mechanical denudation, but none the less still extremely low and in high latitude lowland terrains where
exhibit a wide range. Minimum solute load yields lie below both runoff and solute concentrations are low.
386 Endogenic-exogenic interactions

106k rn 2

1~
1000 0 1000
km

Fig. 15.10 Denudation rates for the world's 35 largest drainage basins based on solid and solute load data. Allowance has been
made for the non-denudational component ofsolute loads. Source rock density is assumed to be 2700 kg m-3 (Based on data in Table
15.6.)

15.3.3 Relative importance of mechanical and unit area increases as total denudation increases, chemical
chemical denudation denudation becomes proportionally less significant in
drainage basins experiencing higher total denudation rates.
On the global scale we have already noted that the ratio of This is apparent in Fig. 15.11 which illustrates the relative
mechanical to chemical denudation is about 6 : 1, but what and absolute rates of solid and solute load transport for
are the variations about this world mean? It is evident from major basins. It is apparent that those rivers transporting the
Table 15.6 that for some basins, especially those in a greatest solid load are in general those also transporting the
predominantly humid lowland environment such as the highest solute load. Note also that the solute load tends to
great Siberian basins of the Ob, the Yenisei and the Lena, form a greater proportion of total load when the total load
chemical denudation can actually greatly exceed mechanical is small.
denudation (although we must keep in mind the point made A fairly clear pattern emerges from Figure 15.11 which
in Section 15.3.1.3). The other extreme is' reached with suggests the primary Influences on relative rates of mechanical
those basins with extremely high sediment yields such as and chemical denudation. The highest rates for both are
the Brahmaputra and the Huang He where chemical denu- found in basins draining major orogenic belts such as the
dation represents 5 per cent or less of total denudation. Over- Brahmaputra and the Ganges. The Chiang Jiang (Yangtze)
all chemical denudation rates are less variable than those Basin, on the other hand, has a particularly high rate of
for mechanical denudation, ranging over two orders of chemical denudation presumably because it occupies large
magnitude rather than three or more. Rates of mechanical areas of limestone terrain in Szechwan Province in western
and chemical denudation overall show a fairly strong posi- China. Rivers draining largely semi-arid regions such as the
tive relationship, but as the ratio of solid to solute load per Colorado, .Orange and Shatt-el-Arab (Tigris and Euphrates)
 Rates of uplift and denudation 387

Chari
Kolyma
Dnepr
Zaire
Ob
Rio Grande
Yenisei
Lena
Murray
St. Lawrence
Amur
Nile
Parana
Volga
Niger
Columbia
Mackenzie
Zambezi
Yukon
Mississippi
Danube

Colorado
Orinoco
Mekon
Shatt-el-Arab
Indus
Jian

2000 1500 1000 500 400 300 200 100

Sediment load (t km- 2 a- 1 )
a Solute load 100
(t km- 2 a- 1 )

Fig.15.11 Sediment and solute loads for the world's largest drainage basins. Solute loads represent the estimated denudational
component only. Data for the Nelson, Tocantins and Sao Francisco Basins are not available. (Based primarily on data in M. Meybeck
(1976) Hydrological Sciences Bulletin 21,265-89 and 1. D. Milliman and R. H. Meade (1983) Journal of Geology 91,1-21.)

can be seen to have very low rates of solute transport in

relation to total transport. This contrasts starkly with the
basins of humid and subarctic regions such as the Lena,
Yenisei, Db and Dnepr which have high rates of solute
transport in relation to total transport. The extremely high
relative rate of solute transport for the St Lawrence Basin
(some 87 per cent of total transport) is explained, at least in
part, by the large proportion of sediment currently being
trapped in the Great Lakes.
It is important to note that even with its predominantly
semi-arid climate the Colorado Basin still has a higher rate
of chemical denudation than either the humid tropical Zaire
Basin or the largely subarctic Db Basin. This is presumably
because of the high rates of chemical denudation occurring
in the high relief section of the Colorado Basin in the
Rocky Mountains and exemplifies the generalization that, o 1000
km
as for mechanical denudation, rates of chemical denudation
appear to be much more strongly related to relief than to
climate. This point is graphically illustrated in Figure 15.12 Fig.15.12 Rates of solid and solute load transport in sub-
which compares the solid and solute loads of major tribu- catchments of the Amazon Basin. Rates of solute load transport
taries of the Amazon. The Maran6n and Ucayali Basins in are for the estimated denudational component (Based on data in
M. Meybeck (1976) Hydrological Sciences Bulletin 21, 265-9.)
the mountainous Andean part of the Amazon Basin show
high rates of both solid and solute transport in comparison
with the Negro Basin which drains the plate~u country of the Amazon Basin originates from the Andean region of the
the Guiana Highland, and the Xingu Basin which occupies catchment with the Maran6n and Ucayali Basins alone
a region of very low relief on the northern margin of the jointly contributing 45 per cent.
Mato Grosso. Some 85 per cent of the total solute load of Although we expect that there will frequently not be a
388 Endogenic-exogenic interactions

125
sample. In particular, we have already highlighted the pro-

I
~ Chemical Denudation
blems of anthropogenic influences on present-day sediment
_ Mechanical Denudation
and solute load data, as well as the complications arising
from the lag between upstream sediment supply and sedi-
100
ment removal from basins. Fortunately, there are a number
of other indirect means which we can use to estimate long-
'I
'" 75 term rates of denudation.
E
.s 15.4.1.1 Estimates from sediment volumes
The variability of sediment yields over periods of a few
decades even in cool, humid temperate environments is
evident from studies of sedimentation rates in small lakes
and reservoirs. Denudation rates can be estimated on the
basis of a known volume of sediment deposited over a
known period of time and originating from a known source
a ~
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982
Year 1980 ....--r---r--~..........,
Fig.15.13 Annual range and variability in rates of mechanical
and chemical denudation for four catchments in Idaho, USA.
(Based on 1. L. Clayton and W. F. Megahan (1986) Earth Surface
Processes and Landforms 11, Fig. 3, p.394.)
1960
steady state between sediment supply and sediment transport
rates, in general this is unlikely to be the case for the release
of solutes and their transport out of the basin. This point
needs to be emphasized since it must be taken into account 1940
in any comparison of rates of mechanical and chemical
denudation. Another factor which has to be considered,
especially when comparisons are based on short-tenn records,
is the different ways in which sediment and solute transport
rates vary through time. This is well illustrated by a detail- 1920
ed study of four small forested catchments on a coarse-
grained granitic lithology in Idaho, USA (Fig. 15.13). Here
the mean total denudation rate over an eleven year period
was calculated to be 8.9 mm ka- 1• The 11 a mean mechanical 1900
denudation rate was found to exceed that for chemical
denudation for three out of the four catchments. Neverthe-
less, because of the greater temporal variability of mechanical
denudation it was more probable that in anyone year
chemical denudation would exceed mechanical denudation
in three out of the four basins. In this case episodic high
rates of mechanical denudation were considered to be related
to high peak flows generated by snow-melt runoff after
winters with heavy snowfalls.
5 10 15 20
t km- 2 a- 1
15.4 Long-term fluvial denudation rates o 5 10
I I I
mmka-1
15.4.1 Methods of estimation Fig.15.14 Temporal variations in the rate of mechanical
The major problem with using sediment and solute load denudation in the Merevale catchment, Warwickshire, UK, based
on rates of lake sedimentation over nine time periods since 1861.
data as a basis for estimating long-term denudation rates is The sedimentation rates have been adjusted to allow for non-
that we cannot be certain that we are not sampling an atypical catchment derived sediment. (Modified from I. D. L. Foster et al.
period and therefore extrapolating from an unrepresentative (1985) Earth Surface Processes and Landforms 10, Fig. 9, p.5 9.)
 Rates of uplift and denudation 389

area. In a study of a small catchment in Warwickshire, UK, can be applied at a range of temporal scales. For instance, it
changes in the volumes of sediment deposited in a reservoir is possible to estimate rates of soil loss from around the
in nine periods since 1861 were calculated on the basis of roots of trees dated by dendrochronology. Since some spe-
54 sediment cores dated radiometrically and correlated cies of tree may have long life spans (the 2500 a or more
from magnetic measurements (Fig. 15.14). In this case the recorded by some bristlecone pines, for example) this
source area for the sediment was accurately known (being technique provides a useful time range. In one study carried
the present catchment area) and there was also good dating out in the Piceance Basin, Colorado, USA, the date when
control. the tree roots were first exposed by erosion was determined
In order to find out about rates of denudation over much by a number of factors including the interpretation of the
longer periods of time we need to apply this approach at annual ring growth pattern and the earliest occurrence of
much greater temporal and spatial scales. We can, for reaction wood. Little variation in denudation rates was
example, attempt to equate denudation over large areas of found between two sites, although a significant difference
continental drainage with the volume of marine sediment was found between north-facing slopes with a mean rate of
deposited offshore along the adjacent continental margin. 560 mm ka- 1 and south-facing slopes with a mean rate of
In effect this is like treating offshore sedimentary basins as 1180 mm ka- 1• The use of the exposure of datable materials
enormous, long-lived reservoirs. can also be applied to human artefacts. Where archaeologi-
The procedure involves four steps. First, the volume and cal remains of known age are present it is sometimes
age of sediments offshore must be determined from borehole possible to estimate rates of ground lowering, or deposition,
data and seismic stratigraphy. The major errors here arise since their construction, or abandonment.
from an insufficient coverage of boreholes to give good For longer time periods it is necessary to use dated
dating control and the possibility of the addition and removal surfaces. A classic study using this approach is the investi-
of sediment by ocean-bottom currents. Secondly, a deduction gation by B. P. Ruxton and I. McDougall of the erosion of
must be made to allow for the porosity of the sediments and the Hydrographers Range, a dissected volcanic peak in
their content of biogenic carbonate deposits (remains of north.-east Papua New Guinea. On the basis of radiometric
marine organisms). Next, an estimate has to be made of the dating it is known that this composite andesitic volcano last
continental area from which the sediment was derived. This erupted about 650 ka BP and, although the original form of
can be a particularly difficult task, especially if we are deal- the now eroded upper part of the cone is not known, its
ing with drainage areas in existence many tens of millions likely elevation was about 2000 m. Below about 1000 m the
of years ago. Finally, the mean rate of mechanical denuda- presence of young lavas on interfluves means that the
tion is calculated by relating the volume of land-derived original surface here can be reconstructed fairly accurately.
sediments over a particular time period to the area of the By measuring the cross-sectional area of valleys cut into
assumed source region. this surface at a range of altitudes, Ruxton and McDougall
In one such investigation of the eastern seaboard of North were able to estimate the volume of material removed over
America it has been estimated that during the Cenozoic the past 650 ka. The denudation rate estimated in this case
about 1000 km 3 of sediment has been produced for each represented the combined effects of mechanical and chemical
kilometre of the 3000 km of coast from Georgia in the USA denudation and was found to increase with increasing local
northwards to Newfoundland. Deducting 600 km 3 km- 1 to relief and slope gradient (Fig. 15.15).
allow for the contribution of porosity and biogenic carbon-
ate gives 400 km 3 km- 1 for the volume of land-derived sedi- 15.4.1.3 Fission track and radiometric techniques
ment. Assuming that the present crest of the Appalachian The final approach we need to consider is one which is
Mountains represented the western limit of the sediment likely to become increasingly important in studies of long-
source this gives an estimate of 2 km of mechanical denu- term denudation rates, namely the exposure of rocks known,
dation over the past 65 Ma for the eastern seaboard of on the basis of fission track or radiometric evidence, to
North America at a mean rate of just over 30 m Ma- I . Such have been at a specific depth below the surface at a particular
estimates are, of course, approximate and do not include time in the past. We have already outlined the principles of
chemical denudation for the simple reason that solutes are this approach in Section 15.2.1; as pointed out there,
dispersed throughout the ocean (although they are even- although these techniques have been primarily used to infer
tually incorporated either into marine organisms (especially crustal uplift rates, the direct evidence they provide in fact
Ca 2+) or into deposits on the ocean floor). relates to denudation rates.
It is interesting to see how estimates of denudation rates
15.4.1.2 Erosion of dated surfaces based on this approach compare with those determined
A more direct and location-specific estimation of denuda- from the volumes of sediment deposited offshore. In a study
tion rates can be made on the basis of the amount of lower- of the exposure of granite intrusions in northern New England
ing experienced by a surface of known age. This approach on the northern seaboard of the eastern USA, differences in
390 Endogenic-exogenic interactions

800 metric and sediment volume estimates, the range of long-

Estimate of maximum_x term denudation rates is broadly comparable to that observed
700 amount of lowering
at centre of volcano for present-day rates of fluvial denudation. We now need
!co 600 to look in more detail at how long-term rates vary over time

. ....
~
and space and assess the degree of equivalence between
E
E 500
. modem and past rates.

....
$ One basis for the estimation of long-term global trends in
~ 400
c
0
.. ... continental denudation is provided by the sediment cores
recovered from the ocean basins during the various legs of
~ 300
~
cQ) ... \
the Deep Sea Drilling Project (DSDP). This evidence sug-
0200
c ......:.. gests that there have been globally synchronous fluctua-
tions in sedimentation rates (and therefore presumably rates
..
1'0

~ 100
. . :..
.. of continental erosion) (Fig. 15.16). High rates during the
Middle Eocene (49 to 45 Ma BP) and the Middle Miocene
100 200 300 400 500 600 700 to the present (from 14 BP) greatly exceed those for
Local Relief ( m ) intervening periods during the Cenozoic. These fluctuations
have been interpreted as the result of lower global preci-
Fig.15.15 Mean rate of denudation as afunction of local relief
(difference in elevation between major ridge crests and adjacent
valley bottoms) for the Hydrographers Range, Papua New
30 ATLANTIC OCEAN
Guinea based on the estimated dissection of dated swfaces.
(Modifiedfrom R. P. Ruxton and I. McDougall (1967) American
Journal of Science 265, Fig. 5(A) p. 557.)
20

fission track ages between apatite and zircon (the latter

10
having a significantly higher annealing temperature) have
been used to estimate their depth and temperature of em-
Ill!"
placement. It was calculated that a presently exposed o 10 20 30 40 50 60
Jurassic intrusion dated at 180 Ma would have solidified at Ma B.P.
a depth of 5.3-7.6 km and a Cretaceous intrusion (115 Ma)
at a depth of 3-3.6km. These values convert to long-term
30 PACIFIC OCEAN
rates of denudation of 42 to 29 m Ma- I since 180 Ma BP and c
31 to 27 m Ma- 1 since 115 Ma BP. It gives one a sense of ·B.~
cQ)
confidence that these estimates are broadly in line with the E 20
rate of 30 m Ma- I (for mechanical denudation only) for the i5
Q)
Vl
past 65 Ma estimated on the basis of offshore sediment u
.~
10
volumes for the eastern seaboard of North America. co
U
Similarly, comparable results have been established for
the Alps where differences in the radiometric ages of o 30 40 50 60
biotite and muscovite indicate a mean denudation rate of Ma B.P.
around 1000 m Ma- 1• This contrasts with an estimated mean
mechanical denudation rate of 100 m Ma- 1 for the Rhone
30 INDIAN OCEAN
Basin in the western Alps based on the sediment volume of
the offshore submarine Rhone fan. These results become
comparable if it is assumed, as seems likely, that the bulk 20
of the sediment transported by the River Rhone originates
from the mountainous Alpine section of its catchment. The
10
highest long-term denudation rate yet recorded appears to
be the 10000 m Ma- 1 over the past 1 Ma determined from
the K-Ar ages of exposed schists in the Southern Alps of o 10 20 30 40 50 60
New Zealand. Ma B.P.

Fig.15.16 Cenozoic rates of clastic (non-biogenic)

15.4.2 Variations in rates sedimentation in the world's ocean basins based on 110 Atlantic,
170 Pacific and 54 Indian Ocean DSDP cores. (Based on data in
We have just seen that, on the basis of fission track, radio- T. A. Davies et al. (1977) Science 197, Fig. 1, p. 54.)
 Rates of uplift and denudation 391

pitation during these periods, although there is evidence 100

that any increased continental denudation during arid phases
is likely to be a result of enhanced aeolian erosion (see
Section 15.6) rather than greater rates of fluvial sediment
transport. However, it is also necessary to consider changes
'I

5
~ 50
,...
-
A

!ro B
in rates of sediment deposition along continental margins a:
since much of the coarser material removed from the con- c
.g 10
e_
tinents is deposited there and fails to reach the deep ocean. ro
'0
:::J
An overriding climatic control has also been invoked by C n
~
another study examining variations in the rates of deposi-
tion of Al 20 3 recorded in DSDP cores; Al 20 3 is a good
5
~ A
E
- -

indicator of continental erosion because it is the major ele- f-

ment most characteristic of non-biogenic sedimentation. A

sixfold increase in the rate of Al 20 3 accumulation in the 1 I I I

northern and central Atlantic and Pacific Oceans over the

250 200 150 100 50 o
Ma B.P
past 15 Ma indicates a greatly increased rate of denudation
during this period, although the increase in the southern Fig.IS.I7 Rates of incision and denudation from the Murray
Basin, south-east Australia, based on various methods: (A) basin
Atlantic and Pacific is a more modest 100 per cent. Again, sedimentation; (B) fission track ages of exposed granitic rocks;
this may reflect primarily increased rates of aeolian denu- (C) river incision into dated basalts, (D) fan sedimentation; (E)
dation since Al2 0 3 could be derived from clay particles in river incision between basalts of two ages. (Modified from P.
soils deflated from regions experiencing phases of aridity. Bishop (1985) Geology 13, Fig. 3,p. 481.)
In a more recent analysis of DSDP cores it has been esti-
mated that the total of denudationally derived sediment and trends apparently present in DSDP records and especially
dissolved load delivered to the world's oceans over their the climatic interpretation placed upon them.
present lifetimes of 100-200 Ma is 1.65 km 3 a-I. This sug- What, then, can we say about the present-day relation-
gests a rather low mean rate of continental denudation of ship between long-term and present-day denudation rates?
11 m Ma- 1 (calculated on the basis of the present land area Although we are living in a rather unusual period of geo-
of 148 x 106 km 2). logical. time (the Quaternary) characterized by rapid changes
Several estimates of long-term denudation rates in south- in eustatic sea level and major oscillations in global
eastern Australia based on a variety of methods indicate climate, contemporary fluvial denudation rates, perhaps
low rates of landscape modification since the Late Mesozoic rather surprisingly, do not appear to be significantly
(Fig. 15.17). River incision in the upper Lachlan valley of different from the long-term average. The estimated
the Murray Basin into radiometrically dated basaltic lavas present-day global mean of 43 m Ma- I is somewhat higher
filling the valley indicates minimum rates of downcutting than the long-term estimate of 11 m Ma- I based on DSDP
of 8 m Ma- 1 for the past 20 Ma and 3-4 m Ma- 1 for the data, but the latter is very uncertain and is probably a
previous 40 Ma. The rate for the past 20 Ma is broadly significant underestimate of total continental denudation
confirmed by the rate of sedimentation in the fan at the because it does not adequately incorporate continental mar-
confluence of the Lachlan with the Murrumbidgee which gin sedimentation. Certainly, fission track estimates of mean
indicates a rate of 4 m Ma- 1. A mean rate of 1-3 m Ma- 1 is denudation rates of around 30 m Ma- 1 during the Cenozoic
indicated for the whole of the Murray Basin during the in regions of subdued relief are not incompatible with a
Cenozoic, while fission track ages from apatites suggest a global average in excess of 40 m Ma- 1• Moreover, the range
denudation rate for the southern part of the East Australian of present-day denudation rates appears to be very similar
Highlands of 9 m Ma- 1 over the past 230 Ma or so. Simi- to the range evident for the long term, with minima of
larly modest rates of denudation of 15-30 m Ma- 1 are indi- around 1 m Ma- 1 and maxima in excess of 5000 m Ma- I •
cated for the past 80-100 Ma following the uplift of the There is evidence that rates in the Late Cenozoic~ may have
east Australian margin. This evidence is of interest both increased significantly in some regions as a result of active
because of the low rates of denudation that it suggests for tectonism or climatic oscillations, and changes in areas of
the tablelands of south-eastern Australia and for the lack of internal and external drainage must also have affected sedi-
any evidence of a significant increase in erosion in the Late ment input on to the continental margins.
Cenozoic in parallel with the global trends indicated by the Although at a global scale there is a broad correspondence
DSDP data (with the possible exception of the Quaternary). between present-day and long-term rates of denudation,
This is in spite of evidence indicating increasing aridity and anthropogenic effects have clearly become overwhelming
climatic seasonality during this period. The Australian evi- in some areas. In Natal in South Africa, for instance, a
dence clearly brings into question the global validity of the comparison of modem mechanical denudation rates with
392 Endogenic-exogenic interactions

long-term rates based on the volume of offshore sediments 500

suggests that the former are between 12 and 22 times the
latter. Very high rates of mechanical denudation have arisen
over the past few decades as a result of rapid soil erosion • •
400
arising from poor land management practices associated
with excessive population pressures in an environment
characterized by steep slopes and periodic intense rainfall. •
• •
15.5 Factors controlling fluvial denudation rates
Some of the factors that influence rates of fluvial denuda- c
·g200
tion have already been mentioned in passing and we now Cll

consider these factors in more detail. Variations in denuda-

"C
::J
C
•
<IJ
tion rates can ultimately be explained by the way erosivity o • •
and erodibility interact. As we noted in Section 7.4.4, ~100 ••
<IJ
erosivity is the energy available at the surface to detach and ~ •
transport regolith; it represents the potential of denuda-
tional systems to remove material from drainage basins. As • ••
well as the effectiveness of the export of material by streams o 1000 2000 3000
it also includes the efficacy of other processes involved in Mean Local Relief (m)
sediment transport including rainsplash, sheet flow and rill
Fig.iS.18 Relationship between mean local relief and mean
erosion, and the various mechanisms of mass movement denudation rate for 20 mid-latitude drainage basins. (Modified
and other denudational processes. In terms of fluvial denu- from F. Ahnert (1970) American Journal of Science 268, Fig. 3,
dation the important factors affecting erosivity include p.251.)
runoff and the gravitational energy available for transport-
ing material downslope, the latter being related to slope and Extremely low rates of denudation are recorded for many
stream gradients within the basin. of the basins of tropical Africa, northern Eurasia and sub-
Erodibility is the susceptibility of materials at the surface arctic North America, in spite of substantial runoff, largely
to transport by denudational processes. This is related to a because the local relief in these areas is generally very low.
complex set of factors including mechanical strength, hard- In the Zaire Basin, for instance, gradients are so minimal in
ness, cohesion and particle size. In addition the rate of solute the lower part of the catchment that natural lakes have
transport is influenced by the susceptibility of rocks and formed.
sediments at, or near, the surface to mechanisms of chemical As we noted in Section 15.3.3 local relief also affects
weathering. The highest rates of denudation will be observed absolute rates of chemical denudation. This is presumably
where both erosivity and erodibility are high and, conver- because in regions of high local relief, with steep slopes
sely, minimal rates of denudation will occur where both are covered by thin soils or with bare rock surfaces, the potential
low. We now look at these factors further. for chemical weathering is greatest, assuming there is ade-
The strong relationship between relief and denudatioI). quate precipitation. This contrasts with areas of low local
rates is evident from the global pattern of denudation (Figs relief in most climatic. environments where the presence of
15.9 and 15.10). It is important here to distinguish between a thick weathering mantle inhibits the movement of water
elevation and local relief. The former itself has no direct at the weathering front where most of the weathering
influence on denudation rates, but the latter is closely reactions occur.
related to local slope and thereby to erosivity. As noted in Climate has been widely held to have a predominant
Section 9.2.2, local relief is the difference between the influence on rates of denudation, although in the light of
maximum and minimum elevation in an area of specific but more recent data it is doubtful whether this view can now
limited size. A very high correlation has been established be sustained. Several attempts have been made to analyze
between denudation rate and local relief measured within the relationship between denudation rate and mean annual
20 x 20 km (400 km 2) areas across a range of mid-latitude precipitation (Fig. 15.19). Although there is little agreement
drainage basins (Fig. 15.18). Although it is true that regions in detail between these postulated relationships certain
at high elevations also tend to have high local relief, espe- common elements emerge - an initial peak of erosion in
cially in orogenic belts, this is far from always the case. semi-arid environments and a progressive increase in denu-
Take, for instance, the high plateau of southern Africa, dation above a mean annual precipitation of around 1000 mm.
where rather subdued local relief (with equally modest rates The curve by Ohmori is based on the largest data set and
of denudation) is found up to 2000 m above sea level. indicates an initial maximum at 35Q-400 mm mean annual
 Rates of uplift and denudation 393

1000

500

200

'I
n:J
~
El00
..§ "" Douglas
....
~
Q)

50 ~.
c 0----0
o
.;:;
n:J "'" Langbein and Schumm
"C
E
Q)
20
o

500 1000 1500 2000

Mean Annual Precipitation (mm)

Fig. 1S.19 Various estimates of the relationship between (mechanical) denudation rate and mean annual precipitation. (Based on H.
Ohmori (1983) Bulletin, Department of Geography, University of Tokyo No. IS, Fig. 2, p. 84.)

precipitation, followed by a minimum at 600 mm and then 50

a progressive increase exceeding the initial maximum at
around 1000 mm. This two-maxima curve can be explained
by the way the amount of precipitation influences both 10
erosivity and, through the mediation of vegetation, erodi- 5
bility. The biomass, or weight of plant material per unit N
I
area, increases with mean annual precipitation up to about E
0)
2500 mm a-I in the transition from desert through desert =. lO
scrub to grassland and forest (Fig. 15.20). The presence of
~ 0·5
vegetation greatly reduces the erodibility of surface materials, n:J

and the abrupt retardation of mechanical denudation rate E

o
above a mean annual precipitation of around 400 mm i co
0·1
probably attributable to a marked increase in vegetation
cover, as it begins to more than outweigh the effects of 0·05
increasing runoff. Eventually, however, the maximum pro-
tective effect of vegetation is reached and beyond this point OL...--_ _----I. --L ---L... -I- ...L-

increasing precipitation, and therefore runoff, tends to lead o 1000 2000

to increasing rates of mechanical denudation. The importance Mean Annual Precipitation
of the protective role of vegetation in retarding erosion is
dramatically illustrated by the effect of different land uses Fig. 1S.20 Generalized relationship between mean annual
on denudation rates (Table 15.7). precipitation and biomass based on various studies. (Based on H.
Ohmori (1983) Bulletin, Department of Geography, University of
The relationship between runoff and chemical denuda- Tokyo No. IS, Fig. 3, p. 86.)
tion is rather different as there is a consistent positive
correlation between solute load and annual runoff (Fig.
15.21). This is apparently because as precipitation increases dation and temperature is very weak, however, apparently
there is more water available for chemical reactions in the because the effect of temperature on weathering rates and
regolith and solute release, and also greater runoff to trans- solute release is overwhelmed by other variables, especially
port these solutes. The relationship between chemical denu- precipitation and local relief.
394 Endogenic~xogenic interactions

Table 15.7 Typical rates of erosion for various types of land use in the mechanical denudation have been found to be negatively
USA related to mean annual runoff but positively related to
runoff variability (that is, the frequency of high magnitude
LAND USE TYPE SEDIMENT YIELD DENUDATION RATE
(tkm- 2 a-I) (mmka- 1) flows; Fig. 15.22). The relationship between runoff variabi-
lity and chemical denudation is, however, negative because
Forest 8.5 4.2 as total runoff and, therefore, chemical denudation increases,
Grassland 85.0 42.5
Cropland 1700 850 river discharges become less variable.
Felled forest 4250 2125 A further factor which indirectly influences fluvial denu-
Active open-cast mines 17 000 8500 dation rates is drainage basin size. Some studies have indi-
Construction sites 17 000 8500
cated that mechanical denudation rates are greater for
Source: Based on data in Environmental Protection Agency (1973) smaller catchments than larger basins, at least up to a basin
Methods for Identifying and Evaluating the Nature and Extent of Nonpoint size of 2000 km 2 or so. There seem to be a number of reasons
sources of Pollutants, Washington, DC.
for this relationship. Small catchments are commonly in the
upper parts of larger basins and they typically have steeper
It has long been recognized that the seasonality of valley-side slopes and channel gradients. Such catchments
precipitation has a crucial effect on rates of mechanical also have smaller floodplains and less potential for
denudation. This results from both the dramatic increase in sediment storage. Finally, it is possible for high-intensity
sediment-carrying capacity in peak flows and the reduction storms generating major floods to cover a small basin
in the protective role of vegetation during the dry season. entirely, whereas such storms will only affect part of a
The high sediment yields in semi-arid environments may larger basin. Contrasting findings to this inverse relation-
also be influenced by the frequency of high intensity storms, ship between basin area and mechanical denudation rate
which characterize such regions, as well as the lack of a have come, however, from a study of sediment and solute
continuous vegetation cover,. In a detailed study in the yields in western Canada. Here the lowest denudation rates
upper Colorado Basin in the western USA, where the mean occur in the smallest basins « 100 km 2) and the highest in
annual precipitation ranges from 150 to 1500 mm, rates of medium-sized catchments (1000-100000 km 2). The largest

500

100

"7
co
N
I
E
~

'::10
"0
co
.3
"0
~
(5
l/)
l/)

Ci
co
:J
c:
c:
«

10 100 1000 5000

Mean Annual Runoff (mm)

Fig.15.21 Generalized relationship between mean annual runoff and mean annual solute load for a sample of 496 rivers. The scatter in
the relationship is probably due largely to the effects of varying lithology. (Based on D. E. Walling and B. W. Webb (1986) in: S. T.
Trudgill (ed.) Solute Processes, Wiley, Chichester, Fig. 7.3, p. 260.)
 Rates of uplift and denudation 395

500 15.6 Rates of aeolian denudation

A
• • • ••
• Few surveys of global denudation have adequately eva-
•
luated the relative importance of aeolian denudation, but it
100
• is clear from our discussion of aeolian processes and dust
• •• fill
transportation (Chapter 10) that it is potentially significant,
I"
~

i
E
t1l

•
•
•••
• .. at least in arid regions. One reason for this relative neglect
is that rates of aeolian denudation are extremely difficult to
quantify. Present rates of dust transport from the continents
c:
0
'';:;
• • to the oceans can be estimated from samples collected from
"0
t1l
:::l
10 • • ships, while individual dust storms can now be monitored
c:
0
Ql
•• by remote sensing techniques either from satellites or by
cou photography from orbiting vehicles such as the space
'2 shuttle. Atmospheric transport can also be inferred from the
t1l
J::
U
Ql presence of dust derived from remote land areas in the soils
::E of oceanic islands. Such dust may be transported great
1
10 100 1000 5000 distances - up to 6500 km from the Sahara to Barbados,
Runoff Variability (Q Q,01,Q 9'99) 8000 km from the Sahara to Miami, 10 000 km from central
Asia to Alaska and 11 000 km from central Asia to the
north Pacific islands of Hawaii and Eniwetok (Fig. 15.23).
100 • B Estimates of rates of atmospheric dust transport vary
enormously from 100 up to 5000 Mt a-I. Assuming all this
I" dust is deposited in the oceans (and therefore lost from the
t1l
~

E continents) and that the source material density is 2700 kg m-3 ,

S this indicates a global mean aeolian denudation rate of
c:
0
'';:;
between 0.25 and 13 mrn ka- I. Any deposition on land would
t1l
10
"0
:::l
lower these rates, but we must also remember that deflation
c:
~ of dust is confined to a rather small proportion of the total
cou land area, so in these regions the aeolian denudation rate
'E would be much higher than these global average figures.
Ql

U
J::
• Recent data suggest that the high estimates of several
thousand million tonnes per year are, in fact, overestimates,
10 100 1000 5000
Runoff Variability (Q Q,01,Q 9'99)
but present-day aeolian denudation rates may still be impres-
sive at the regional scale. It is estimated, for' instance, that
Fig.15.22 Relationship between runoff variability and the rate of dust transport from the Sahara to the Atlantic
mechanical (A) and chemical (B) denudaiion rates in the upper averages 146 Mt a-I. This implies a sediment yield aver-
Colorado Basin, western USA. Runoffvariability is represented
as the ratio between the discharge that is equalled or exceeded
aged over the area of the Sahara as a whole of around
0.01 per cent of the time and the discharge that is equalled or 16 t km-2 a-I which converts to a denudation rate of
exceeded 99.9 per cent of the time (QO.oJ1Q99,9)' (Modified from 6 mm ka- I. However, it is clear that the source of this dust
K-H Schmidt (1985) Earth Surface Processes and Landforms 10, is largely confined to particular regions within the Sahara,
Fig. 6, pp. 504-5.) espedally the Bodele Depression, the alluvial plains of Niger
arid Chad, southern Mauritania, northern Mali and central
basins (>100 000 km 2) were found to have intermediate southern Algeria, southern Morocco and western Algeria,
rates. the southern fringes of the Mediterranean Sea in Libya and
There are considerable uncertainties as to the dominant Egypt and northern Sudan. If the dust source is limited to
factors controlling rates of fluvial denudation, and much these areas then they are probably experiencing a rate of
research is still needed to test some widely accepted, but aeolian denudation in excess of 20 mm ka- I , just under half
not necessarily well supported, generalizations. None the the mean global fluvial denudation rate. Similar estimates
less, it appears that at the global scale, relief and, to a lesser of aeolian denudation appear to be valid for east-central
extent, climate are the main determinants of denudation rates Asia, the source of around 20 Mt a-I of fine soil material
(Table 15.8). At the regional and local scale, lithology and carried to the north-west Pacific.
the specific factors of erodibility, which determine the supply Sediment in ocean cores enables past rates of dust depo-
of sediment and solutes, become more significant. sition to be monitored and both in the north-west Pacific
396 Endogenic-exogenic interactions

Table 15.8 Solid and denudational solute load of major rivers and total denudation in relation to climate and
relief

CLIMATE AND SOLID LOAD DENUDATIONAL TOTAL LOAD TOTAL TYPICAL SOLUTE
RELIEF ZONE (t km- 2 a-I) SOLUTE LOAD (t km- 2 a-I) DENUDATION LOAD AS % OF TOTAL
(t km- 2 a-I) (mmka- I)

Mountainous, 200-1500 70-350 250-2000 95-740 10

high precipitation
Mountainous, 100-1000 10-60 120-1000 45-370 10
low precipitation
Moderate relief, 40-200 25-60 80-300 30-110 35
temperate or
tropical climate
Low relief, 10-100 3-10 15-100 5-35 10
dry climate
Low relief, 20-50 12-50 40-80 15-30 65
temperate climate
Low relief, 1.5-15 5-35 5-40 5-15 80
subarctic climate
Low relief, 1-10 2-15 4-30 1.5-10 50
tropical climate

Source: Based partly on data in M. Meybeck (1976) Hydrological Sciences Bulletin 21, Table 2, p. 279.

Tropical arid

Tropical semi-arid - - - . ~~jn~~or~~~tof

Dry continental Regions of frequent 6 2

dust haz~ 10km

1~
1000 0 1000
km

Fig.15.23 Global distribution of dust storm activity and major dust trajectories. (Modified from N. J. Middleton et al. (1986) in W. G.
Nickling (ed.) Aeolian Geomorphology. Allen and Unwin, Boston, Fig. 4, p. 247.)
 Rates of uplift and denudation 397

and Atlantic Oceans this shows that current rates are areas, by contrast, are relatively 'blown out' with a surface
representative of the past few thousand years. Taking the cover of rocky plains or dune fields which yield little of the
ocean core record further back, however, we find some fine particles necessary for long distance aeolian transport.
interesting fluctuations in rates of deposition. Around 18 000 a
BP rates of accumulation off the coast of West Africa were
about twice those of the present, and similar increases at
15.7 Rates of glacial denudation
this time have been recorded off coasts adjacent to the As with aeolian denudation, rates of glacial denudation
Australian, Arabian and Thar Deserts. It is now possible to have received relatively little attention and are similarly
identify changing rates of aeolian denudation on the basis difficult to estimate. The methods used include the measure-
of atmospheric dust deposition in ocean cores as far back as ment of the sediment and solute load of glacial meltwater
the Cretaceous. Rates of deposition were apparently low streams, reconstructions of pre-glacial or interglacial land
during the Early Cenozoic, but increased after about 25 Ma surfaces, the relating of the volume of glacial drift in a
BP and then accelerated markedly from 7 to 3 Ma BP The specific region to an assumed glacial source area and the
most dramatic increase occurred, however, around 2.5 Ma estimation of the contribution of glacial debris to marine
BP accompanying the onset of major glaciation in the sediments. There are comparatively few estimates of rates
northern hemisphere. For much of the Cenozoic the Sahara of mechanical denudation for active valley glaciers and these
seems to have remained a major dust source. show an extremely broad range from 100 to 5000 mm ka- 1•
Variations in rates of dust deposition in the geological The limited evidence available suggests that rates of solute
past highlight some of the factors which control rates of transport in glacial meltwater streams are relatively high,
aeolian denudation. The dramatic increase in rates at the with a chemical denudation rate of about 200 mm ka - 1
beginning of the Pleistocene appears to be linked to the being calculated for the South Cascade Glacier in the
establishment of arid - humid climatic cycles at low lati- Cascade Range of the north-west USA. This is presumably
tudes at this time roughly in phase with the glacial - inter- related to the high surface area for chemical reactions
glacial cycles at higher latitudes. This accords with evidence provided by the fine particle size of rock flour, and may
from current dust storm activity which suggests that the help to explain the high rates of chemical denudation
most abundant sources. of dust are in areas that are pre- estimated for partially glaciated mountain belts.
sently changing to more arid climatic regimes. There was, There is considerable uncertainty as to the rates of denu-
for instance, a threefold increase in rates of dust transport dation associated with continental ice sheets. This is illus-
out of North Africa during the droughts of the early 1970s trated by the debate over the depth of erosion accomplished
and early 1980s. The observation that dust-storm frequency by the Pleistocene Laurentide ice sheet in eastern North
reaches a maximum where mean annual precipitation is America. One view is that the glacial scouring of this region
around 100-200 mm (Fig. 15.24) is consistent with evi- has been largely confined to the exhumation of a weather-
dence that water is essential for the weathering processes ing front formed under pre-glacial conditions and has
capable of producing large quantities of the fine sedimen- been limited to an average of a few tens of metres at most.
tary particles susceptible to deflation. Many hyper-arid In support of this idea is the absence of the quantities of
glacial till to be expected if deep erosion had occurred. On
the other hand, it has been pointed out that much of the
Vl
§ glacial debris produced could have been transported offshore
£
(f)
10 by meltwater streams. Recent estimates of marine deposi-
+-'
Vl tion rates along the eastern margin of North America do
::J
0 indeed suggest a significant input of glacially derived mate-
'5 rial during the Pleistocene. Although the precise source
>
()
cQ) region for this sediment is impossible to define, it has been
::J
0- estimated that the volume deposited offshore implies a
~ minimum mean depth of erosion of 120 m for the Laurentide
ro::J region with a possible maximum of 200 m. This gives a
c
c mean denudation rate of between 48 and 80 m Ma- 1 through-
«
c
co
out the Pleistocene, and this is clearly rather higher than
Q)
~ would be. expected of fluvial denudation for such an area of
subdued relief.
Mean Annual Precipitation (mm) Rates of sediment production from 'permanent' ice masses,
Fig. 15.24 Global mean annual frequency of dust storms in
such as the Antarctic ice sheet, are very low and it appears
relation to mean annual precipitation. (From A. S. Goudie (1983) that significant erosion is only accomplished by periodic ice
Progress in Physical Geography 7, Fig. 11, p. 615.) sheets such as those of the northern hemisphere in the
398 Endogenic-exogenic interactions

Pleistocene. Whatever the general picture it is clear that The evidence now available on rates of crustal uplift and
erosion under even periodic ice sheets can be highly selec- denudation in orogenic belts indicate that the interaction
tive. This is illustrated in the Buchan area of north-east between uplift and denudation has an inherent tendency
Scotland where, in spite of successive ice sheet advances towards a steady state where both are roughly equal in rate.
across the region, are still preserved deep weathering mantlesDuring the uplift of a mountain range the denudation rate
of pre-Pleistocene age. along its crest is initially slower than the rate of crustal
uplift. This is because as we have seen, the denudation rate
is largely a function of local relief, and this in tum is
15.8 Comparison of rates of uplift and denudation related to the degree of fluvial dissection. As this begins at
Few topics in geomorphology have received a more con- the margins of an uplifted mountain range and progresses
fused analysis than the relationship between rates of uplift towards its crest by headward erosion, the summit region is
and denudation and the implication this has for the way relatively unaffected by denudation in these early stages of
landscapes evolve. We will leave the question of landscape uplift. Consequently the elevation of the crest of the range
evolution until Chapter 18, but we will conclude this chapter increases, but as fluvial dissection works towards the crest
by considering relative rates of uplift and denudation. the rate of denudation increases until eventually rates of
In an influential paper published in 1963, S. A. Schumm uplift and denudation are roughly equal. In spite of further
asserted that modem rates of orogenic uplift at around crustal uplift, in part promoted by isostatic rebound as a
7500 mm ka- 1 are some eight times greater than average consequence of continuing denudational unloading, the moun-
maximum denudation rates. This conclusion has subse- tain range cannot become any higher as long as the long-
quently been widely accepted and has been interpreted as term rate of crustal uplift remains roughly constant and
meaning that the assumption that episodic phases of rapid there are no significant changes in exogenic conditions.
uplift punctuate the continuous but more leisurely progress This equivalence of rates of crustal uplift and denudation in
of erosion is broadly correct. Such an assumption has young mountain ranges is not fortuitous but simply reflects
important implications for the understanding of landscape the achievement of a steady state, and the time taken to
evolution so we must ask, is it really true? Schumm was attain the steady state elevation is largely a function of the
working with a much smaller database than that now avail- rate of ~plift.
able and his estimates of maximum rates of denudation are But how high can a mountain range become before this
unduly modest. We have seen, in fact, that maximum rates steady state is achieved? This seems to depend largely on
of both orogenic uplift and denudation lie between 5000 and the width of the mountain range, since the broader the range
10 000 mm ka- 1 • Rates of glacio-isostatic rebound can, of the longer it takes for fluvial dissection to reach the crestal
course, be even higher than this but such vertical displace- regions, and therefore the longer the time during which the
ments can only amount to a total of a few hundreds of crustal uplift rate on the divide exceeds the rate of denu-
metres. dation. Consequently, it is no surprise to observe that the

10000
5000

2000

.:rlOOO
ctl NE Japan
.::.::. 500
E
E SW Japan
....ctl 200
Q) Outer Zone
c:: "100
c NE Japan
o
.~ 50 SW Japan
'0
:J Inner Zone
a:; 20
Cl
10L.------'------'---------<--------'
o 250 500 1500 2000 30
Amount of Quaternary Uplift (m)

Fig. 15.25 Relationship between rates of Quaternary uplift and rates ofpresent denudation in Japan. (Based on T. Yoshikawa (1985)
in A. Pitty (ed.) Themes in Geomorphology. Croom Helm, London, Fig. 12.3, p. 203, and Fig. 12.4, p. 204.)
 Rates of uplift and denudation 399

crest of the relatively narrow Southern Alps of New Zealand, rapidly towards the west as the zone of oblique conver-
which are only some 80 km across, has an average height of gence marked by the Southern Alps is approached (see
about 3000 m whereas the Himalayas, which have a width Section 3.5.2; Fig. 15.26). The crest of the Southern Alps
of around 350 km, have an average crestal elevation of around seems to be in a steady state with very high rates of both
7500m. crustal uplift and denudation in excess of 5000 mm ka- 1• On
The relationships between rates of crustal uplift and the eastern flanks of the Southern Alps, however, in areas
denudation in orogenic belts can be instructively examined such as eastern Otago, rates of crustal uplift are consider-
by looking at New Zealand and Japan, both regions for ably lower 000-300 mm ka- 1), but still well in excess of
which relatively good data are available. In Japan the rates of denudation at around 70 mm ka- I . Here, then, there
central zone with crustal uplift rates of about 2200 mm ka- 1 has apparently been insufficient time for a steady state
has attained a steady state and the outer zone nearly so, between crustal uplift and denudation to be established.
whereas in the north-east and south-west zone crustal uplift So far we have limited our discussion of the relationship
still exceeds denudation (Fig. 15.25). It appears that in the between crustal uplift and denudation rates to the situation
central and south-west zones the rapid rates of crustal uplift occuring in orogenic belts. 'What of those regions exper-
have prompted a rapid response in the rate of denudation, iencing epeirogenic uplift? Excluding areas of glacio-isostatic
whereas in the other two zones the slower rates of crustal rebound, long-term rates of epeirogenic crustal uplift appear
uplift have not yet led to a fully compensating increase in to be of the order of 100 mm ka- 1 (in spite of possible
denudation rates. A similar picture is revealed in South short-term rates an order of magnitude higher than this).
Island, New Zealand, where rates of crustal uplift increase However, the low local relief of many areas experiencing

>7000

2000-7000

0-2000

Estimated
Denudation Rate
(mmka-1)

>5000

111500-1500

o
km

Fig.15.26 Comparison of rates of crustal uplift and denudation in New Zealand. (Modifiedfrom M. 1. Selby, (1982) Hillslope
Materials and Processes. Oxford University Press, Oxford, Fig. 11.11, p. 237.)
400 Endogenic-exogenic interactions

epeirogenic uplift, such as the regions of broad continental important effect that the disequilibrium between sediment
warping in Africa and Australia, mean that denudation rates supply and transport can have on estimates of denudation
in such areas are often even lower. Consequently, there is rates. Saunders and Young (1983) and Young and Saunders
little prospect of a steady state being achieved in these cases, (1986) provide useful overviews of the relative importance
a point which has interesting implications for models of of different denudational processes under various climates,
landscape development as we will see in Chapter 18. and Meybeck (1976) and Clayton and Megaham (1986)
provide, respectively, a general and specific assessment of
the relative importance of mechanical and chemical
Further reading denudation.
At the longer time scale, Carrara and Carroll (1979)
There is an enormous range of literature on rates of uplift provide a case study estimating erosion rates from tree root
and denudation, but fortunately much of this is summarized exposure, while Foster et al. (1985) demonstrate the vari-
in reviews. Most of the material on uplift rates is not set ability in sediment transport evident in lake sediment records
specifically in a geomorphic context, but the importance of over periods of a few decades. Assessments of long-term
such information for understanding landform development global rates of sediment supply to the oceans include those
cannot be overemphasized. It is, however, important to be by Davies et al. (1977) (updated by Worsley and Davies,
aware of the confusion between surface uplift and crustal 1979), Donnelly (1982) and Howell and Murray (1986).
uplift evident in much of the literature on uplift rates. Matthews (1975) uses the offshore sedimentary record to
The most useful general assessment of crustal move- examine Cenozoic rates of erosion in eastern North America,
ments and the techniques used to monitor them, especially while Ruxton and McDougall (1967) document the rate of
over the short to medium term, is the book by Vita-Finzi fluvial dissection of the Hydrographers Range in Papua
(1986). The ambiguities inherent in the measurement of New Guinea. The use of fi sion track and radiometric ages
present-day deformation using relevelling surveys are to determine long-term average denudation rates is illus-
discussed by Brown et al. (1980), while Adams (1984) trated in the papers already mentioned on the estimation of
looks at a range of evidence for recent crustal movements crustal uplift rates, and by Doherty and Lyons (1980).
along the northern Pacific coast of the USA. Schaer et al. Bishop (1985) convincingly demonstrates the predominance
(1975) compare modem and past rates of uplift in the Swiss of low rates of denudation throughout the Cenozoic in
Alps, while various aspects of the literature on rates of south-eastern Australia, while Young (1983) briefly examines
tectonic uplift are briefly reviewed in my short discussion the evidence for, and implications of, low denudation rates
of neotectonics and landform genesis (Summerfield, 1987). on a broader basis. At a more local scale Martin (1987)
Rates of glacio-isostatic uplift are considered by Nikonov compares long-term sedimentation rates off the Natal coast
(1980). Clark and Jager (1969) discuss the application of with contemporary sediment yields. The factors determining
radiometric dating to estimating crustal uplift rates, while denudation rates are examined briefly in the reviews of
Gleadow and Fitzgerald (1987), Moore et al. (1986) and solid and solute load data mentioned above, but more de-
Parrish (1983) apply fission track dating techniques. Both tailed evaluations are provided by Wilson (1973) on climatic
present-day and long-term rates of uplift in orogenic belts controls, Ahnert (1970) on the local relief factor, Schmidt
are considered by Gansser (1983). (1985) on the importance of runoff variability, Schumm
There have been several recent assessments of estimates (1963) and Slaymaker (1987) on drainage basin area and
of rates of mechanical and chemical denudation based on Ohmori (1983) and Trimble (1988) on the role of vegeta-
sediment and solute load data for rivers. Sediment yields tion. The influential paper by Langbein and Schumm (1958)
are dealt with by Milliman and Meade (1983) and Walling is still worth a look as long as its limited applicability is
and Webb (1983a), while the solute loads of rivers are appreciated.
considered at a predominantly local and regional scale by The role of dust transportation is considered by Goudie
Walling and Webb (1986) and at a largely global scale by (1983) and Middleton et al. (1986), while Lever and McCave
Walling and Webb (1983b). These wide-ranging reviews (1983) use evidence from deep sea cores to trace the varia-
also consider the problems of estimating sediment and solute tions in aeolian deposition in the Atlantic since the Cretaceous.
loads and of inferring patterns of denudation from them. On Rates of sediment transport in glacial streams are examined
this theme Meybeck (1983) and Cryer (1986) deal in detail by Gurnell (1987), and the high solute concentrations typical
with the question of atmospheric inputs in river solutes, of glacial meltwater are considered by Collins (1983). The
while Dunne (1978) provides an excellent investigation of debate concerning rates of ice sheet erosion is illustrated
chemical denudation rates carefully excluding non-denudational through the contributions of White (1972), Sugden (1976)
components. Turning to sediment loads, Douglas (1967) and Bell and Laine (1985), while Hall and Sugden (1987)
assesses the importance of anthropogenic factors, while provide specific evidence of the inefficiency, at least locally,
Trimble (1977, 1983) and Meade (1982) address the of ice sheets as denudational agents.
 Rates of uplift and denudation 401

On the question of the relative rates of uplift and denuda- Lake catchment based studies of erosion and denudation in the
tion it is appropriate to begin with S€humm' s influential Merevale catchment, Warwickshire, U.K. Earth Surface Processes
assessment (Schumm, 1963) and then to look at the excellent and Landforms 10,45-68.
Gansser, A (1983) The morphogenic phase of mountain building.
discussions by Adams (1985) on the Southern Alps and In: K. J. Hsii (ed.) Mountain Building Processes. Academic
Yo hikawa (1985) on Japan. The factors determining the Press, London and New York, 221-8.
elevation of a mountain range are considered by Ahnert Gleadow, A. J. W. and Fitzgerald, P. G. (1987) Uplift history and
(1984). structure of the Transantarctic Mountains: new evidence from
fission track dating of basement apatites in the Dry Valleys
area, southern Victoria Land. Earth and Planetary Science
Letters 82, 1-14.
Goudie, A. S. (1983) Dust storms in space and time. Progress in
References Physical Geography 7,502-30.
Adams, J. (1984) Active deformation of the Pacific northwest Gurnell, AM. (1987) Suspended sediment. In: A M. Gurnell and
continental margin. Tectonics 3, 449-72. M. J. Clark (eds) Glacio-Fluvial Sediment Transfer, Wiley,
Adams, J. (1985) Large-scale tectonic geomorphology of the Chichester and New York, 305-54.
Southern Alps, New Zealand. In: M. Morisawa and J. T. Hack Hall, A M. and Sugden, D. E. (1987) Limited modification of
(eds) Tectonic Geomorphology. Allen and Unwin, Boston and mid-latitude landscapes by ice sheets: the case of northeast
London, 105-28. Scotland. Earth Surface Processes and Landforms 12, 531-42.
Ahnert, F. (1970) Functional relationships between denudation, Howell, D. G. and Murray, R. W. (1986) A budget for continental
relief and uplift in large mid-latitude drainage basins. American growth and denudation. Science 233, 446-9.
Journal of Science 268,243-63. Langbein, W. B. and Schumm, S. A. (1958) Yield of sediment in
Ahnert, F. (1984) Local relief and the height limits of moun.tain relation to mean annual precipitation. American Geophysical
ranges. American Journal ofScience 284, 1035-55. Union Transactions 39, 1076-84.
Bell, M. and Laine, E. P. (1985) Erosion of the Laurentide region Lever, A and McCave, I. N. (1983) Eolian components in Creta-
of North America by glacial and fluvioglacial processes. Qua- ceous and Tertiary North Atlantic sediments. Journal of Sedi-
ternary Research 23, 154-74. mentary Petrology 53, 811-32.
Bishop, P. (1985) Southeast Australian late Mesozoic and Martin, A. K. (1987) A comparison of sedimentation rates in the
Cenozoic denudation rates: a test for late Tertiary increases in Natal Valley, SW Indian Ocean with modem sediment yields in
continental denudation. Geology 13, 479-82. east coast rivers, southern Africa. South African Journal of
Brown, L. D., Reilinger, R. E. and Citron, G. P. (1980) Recent Science 83, 716-24.
vertical crustal movements in the US: evidence from precise Matthews, W. H. (1975) Cenozoic erosion and erosion surfaces of
levelling. In: N.-A Marner (ed.) Earth Rheology, Isostasy and eastern North America. American Journal of Science 275,
Eustasy. Wiley, Chichester and New York, 389-405. 818-24.
Carrara, P. E. and Carroll, T. R. (1979) The determination of erosion Meade, R. H. (1982) Sources, sinks, and storage of river sediment
rates from exposed tree roots in the Piceance Basin, Colorado. in the Atlantic drainage of the United States. Journal of Geo-
Earth Surface Processes 4,307-17. logy 90, 235-52.
Clark, S. P. and Jager, E. (1969) Denudation rate in the Alps from Meybeck, M. (1976) Total annual dissolved transport by world
geochronologic and heat flow data. American Journal of major rivers. Hydrological Sciences Bulletin 21, 265-89.
Science 267, 1143-60. Meybeck, M. (1983) Atmospheric inputs and river transport of
Clayton, J. L. and Megahan, W. F. (1986) Erosional and chemical dissolved substances. In: Dissolved Loads of Rivers and Surface
denudation rates in the southwestern Idaho batholith. Earth Water Quantity/Quality Relationships. International Association
Surface Processes and Landforms 11, 389-400. of Hydrological Sciences Publication 141, 173-92.
Collins, D. N. (1983) Solute yield from a glacierized high moun- Middleton, N. J., Goudie, A. S. and Wells, G. L. (1986) The fre-
tain basin. In: B. W. Webb (ed.) Dissolved Loads of Rivers and quency and source areas of dust storms. In: W.G. Nickling (ed.)
SUiface Water Quantity/Quality Relationships. International Aeolian Geomorphology. Allen and Unwin, Boston and London,
Association of Hydrological Sciences Publication 141, 237-59.
41-9. Milliman, J. D. and Meade, R. H. (1983) World-wide delivery of
Cryer, R. (1986) Atmospheric solute inputs. In: S. T. Trudgill (ed.) river sediment to the oceans. Journal ofGeology 91, 1-21.
Solute Processes. Wiley, Chichester and New York, 15-84. Moore, M. E., Gleadow, A J. W. and Lovering, J. F. (1986) Thermal
Davies, T. A, Hay, W. W., Southam, J. R. and Worsley, T. R. (1977) evolution of rifted continental margins: new evidence from
Estimates of Cenozoic oceanic sedimentation rates. Science fission tracks in basement apatites from southeastern Australia.
197,53-5. Earth and Planetary Science Letters 78, 255-70.
Doherty, J. T. and Lyons, J. B. (1980) Mesozoic erosion rates in Nikonov, A A (1980) Manifestations of glacio-isostatic pro-
northern New England. Geological Society of America Bulletin cesses in northern countries during the Holocene and at present.
91, 16-20. In: N.-A Marner (ed.) Earth Rheology, Isostasy and Eustasy.
Donnelly, T. W. (1982) Worldwide continental denudation and Wiley, Chichester and New York, 341-54.
climatic deterioration during the late Tertiary: evidence from Ohmori, H. (1983) Erosion rates and their relation to vegetation
deep-sea sediments. Geology 10,451-4. from the viewpoint of world-wide distribution. Bulletin of the
Douglas, I. (1967) Man, vegetation, and the sediment yield of Department of Geography University of Tokyo 15, 77-91.
rivers. Nature 215, 925-8. Parrish, R. R. (1983) Cenozoic thermal evolution and tectonics of
Dunne, T. (1978) Rates of chemical denudation of silicate rocks in the Coast Mountains of British Columbia 1, fission-track dating,
tropical catchments. Nature 274, 244-6. apparent uplift rates, and patterns of uplift. Tectonics 2, 601-
Fo ter, I. D. L., Dearing, J. A, Simpson, A and Carter, AD. (1985) 32.
402 Endogenic-exogenic interactions

Ruxton, B. P. and McDougall, I. (1967) Denudation rates in north- Trimble, S. W. (1988) The impact of organisms on overall erosion
east Papua from potassium-argon dating of lavas. American rates within catchments in temperate regions. In: H. A. Viles
Journal of Science 265, 545-61. (ed.) Biogeomorphology. Blackwell, Oxford, 83-142.
Saunders, I. and Young, A. (1983) Rates of surface processes on Vita-Finzi, C. (1986) Recent Earth Movements: An Introduction
slopes, slope retreat and denudation. Earth Surface Processes to Neotectonics. Academic Press, London and New York.
and Landforms 8, 473-501. Walling, D. E. and Webb, B. W. (1983a) Patterns of sediment
Schaer, J. P., Reimer, G. M. and Wagner, G. A. (1975) Actual and yield. In: K. J. Gregory (eds) Background to Palaeohydrology.
ancient uplift rate in the Gotthard region, Swiss Alps: a com- Wiley, Chichester and New York, 69-100.
parison between precise leveling and fission-track apatite age. Walling, D.E. and.Webb, B. W. (1983b) The dissolved loads of
Tectonophysics 29, 293-300. rivers: a global overview. In: Dissolved Loads of Rivers and
Schmidt, K.-H. (1985) Regional variation of mechanical and Surface Water Quantity/Quality Relationships. International
chemical denudation, upper Colorado River Basin, U.S.A. Earth Association of Hydrological Sciences Publication 141, 3-20.
Surface Processes and Landforms 10,497-508. Walling, D. E. and Webb, B. W. (1986) Solutes in river systems.
Schumm, S. A. (1963) The disparity between present rates of In: S. T. Trudgill (ed.) Solute Processes. Wiley, Chichester and
denudation and orogeny. United States Geological Survey Pro- New York, 251-327.
fessional Paper 454-H. White, W. A. (1972) Deep erosion by continental ice sheets.
Slaymaker, O. (1987) Sediment and solute yields in British Columbia Geological Society ofAmerica Bulletin 83, 1037-56.
and Yukon: their geomorphic significance reexamined. In: V. Wilson/ L. (1973) Variations in mean annual sediment yield as a
Gardiner et at. (eds) International Geomorphology 1986 Part I. function of mean annual precipitation. American Journal of
Wiley, Chichester, 925-45. Science 273, 335-49.
Souchez, R. A. and Lemmens, M. M. (1987) Solutes. In: A. M. Worsley, T. R. and Davies, T. A. (1979) Sea-level fluctuations
Gurnell and M.1. Clark (eds) Glacio-Fluvial Sediment Transfer. and deep-sea sedimentation rates. Scienc'e 203, 455-6.
Wiley, Chichester and New York, 285-303. Yoshikawa, T. (1985) Landform development by tectonics and
Sugden, D. E. (1976) A case against deep erosion of shields by ice denudation. In: A. Pitty (ed.) Themes in Geomorphology. Croom
sheets. Geology 4, 580-2. Helm, London, 194-210.
Summerfield, M. A. (1987) Neotectonics and landform genesis. Young, A. and Saunders, I. (1986) Rates of surface processes and
Progress in Physical Geography 11, 384-97. denudation. In: A. D. Abrahams (ed.) Hillslope Processes. Allen
Trimble, S. W. (1977) The fallacy of stream equilibrium in con- and Unwin, Boston, and London, 3-27.
temporary denudation studies. American Journal of Science Young, R. W. (1983) The tempo of geomorphological change:
277,876-87. evidence from southeastern Australia. Journal of Geology 91,
Trimble, S. W. (1983) A sediment budget for Coon Creek basin in 221-30.
the Driftless Area, Wisconsin, 1853-1977. American Journal
ofScience 283, 454-74.
This page intentionally left blank
 16
Tectonics and drainage
development

16.1 Active and passive tectonic controls on structural controls) operate through the influence exerted
drainage by previous tectonic activity on subsequent drainage devel-
Tectonic controls can be exerted on. the development of opment. These controls are usually more subtle than those
drainage systems in two ways (Table 16.1). Active tectonic arising from active tectonism and are more difficult to
controls involve the response of fluvial systems to ongoing unravel because later erosion may have removed much of
tectonic activity. Such controls are most evident in tectonic- the evidence of pre-existing structures. Passive controls most
ally active environments, but even in apparently tectonically commonly operate through the effect that previous tectonic·
quiescent regions drainage systems can be significantly activity has on the structural organization (disposition and
affected by active tectonic controls. Such controls include arrangement) of lithologies of varying degrees of resist-
faulting as well as the tilting and deformation of the ground ance. Some of the world's most spectacular regional scale
surface. landforms are fold terranes which have been partIally
Passive tectonic controls (also sometimes referred to as eroded by drainage systems (Fig. 3.17).
Although the distinction between active and passive con-
trols is a useful one, it is important to emphasize that both
Table 16.1 Mesoscale tectonic controls of drainage systems kinds frequently operate together in influencing the devel-
opment of drainage systems. A drainage network in a sedi-
CONTROLS EFFECTS
mentary terrane which is being actively folded will both
Active controls respond directly to the ongoing ground deformation and to
Active faulting Linear, hanging and wineglass valleys; channel the arrangement of alternating weak and resistant beds
offsets, ponding and diversion; terraces and exposed as the folds are progressively eroded.
knickpoints
Active folding Antecedent and dip drainage; water gaps; channel
and tilting incision, aggradation and lateral shifting
16.2 Passive tectonic controls
Passive (structural)
controls
Fault traces
16.2.1 Drainage patterns
Linear, hanging and wineglass valleys; channel
offsets, ponding and diversion; terraces and Drainage patterns are influenced by many factors, including
knickpoints
climatic and lithological variables, but geological structure
Tilting Trellis drainage pattern; parallel, long dip and short
anti-dip streams is ovenill the most important factor. Indeed, drainage
Domes Radial and annular drainage patterns; superimposed patterns are often so dramatically affected by passive tec-
rivers tonic controls that the patterns themselves are frequently
Anticlines and Trellis drainage pattern; superimposed rivers; water used as a basis for unravelling geological structures from
synclines gaps data sources such as aerial photographs and satellite imagery.
Lineation Asymmetric valleys; linear channels The major types of drainage pattern associated with tec-
Joints Rectangular drainage patterns tonic controls are illustrated in Figure 16.1 and are listed,
Source: Modified from M. Morisawa (1985) Rivers. Longman, London, with their primary characteristics, in Table 16.2. The patterns
Table 10.1, p. 157. illustrated are those that occur most commonly at the meso-
406 Endogenic~xogenic interactions

B c

Radial

fa Annular'

G
~\/!%
;(r{ '-.;:
!
Centripetal

o 5 o~I 500
L.,'--'-km--....J' ( B- F) -....J' (A and G)
m

Fig.16.1 Major types of drainage pattern related to structural controls. These are actual examples taken from various localities in
the USA: (A) dendritic -Big Flat, Arkansas; (B) parallel-East Wind River Range, Sage Creek; (C) radial-Mount Ellsworth, Utah;
(D) trellis - Tuscarora Creek, Pennsylvania; (E) rectangular - Schroon and Bouquet Rivers, New York; (F) annular - Maverick Spring
Dome, Wyoming; (G) centripetal- Menan Buttes, Idaho. (Modifiedfrom M. Morisawa (1985) Rivers. Longman, London, Fig. 10.3,
p.161.)

Table 16.2 Classification of drainage patterns related to structural controls

TYPE DESCRIPTION OF PATTERN STRUCTURAL CONTROL

A. Dendritic Spreading tree-like arrangement. No evident orientation Horizontal sediments or homogeneous crystalline rocks.
of channels Lack of structural control on rocks of uniform resistance.
B. Parallel Main channels regularly spaced and parallel, or subparallel, Closely spaced faults, monoclines or isoclinal folds.
to each other. Tributaries join at very acute angles
C. Radial Streams flow outward from centre Volcanic cones, domes
D. Trellis A dominant drainage direction with a secondary direction Tilted or folded alternately resistant/weak sedimentary units
perpendicular to it. Primary tributaries join main stream at
right angles, secondary tributaries are parallel to main stem
E. Rectangular Drainage forms a perpendicular net with the two directions Joints or faults
equally developed.
F. Annular Main rivers have circular pattern with subsidiary channels Eroded dome in alternate resistant/weak sediments
at right angles
G. Centripetal Streams flow inward to centre Calderas, craters, tectonic basins

Source: Modified from M. Morisawa (1985) Rivers. Longman, London, Table lOA, p. 160.
 Tectonics and drainage development 407

scale and macroscale. Continental drainage systems, how- as is evident from the anti-dip tributaries being shorter than
ever, are usually more complex and may include a variety of the dip streams.
specific drainage pattern types. The angle of dip of alternating weak and resistant beds
The basis for comparison for structurally related drainage can obviously change, and the associated topographic forms
patterns is the dendritic type. This consists of a system of vary as a consequence (Fig. 16.4). Fluvial erosion of resis-
irregular branching tributaries forming a tree-like pattern tant beds in horizontal strata leads to the development of
with junctions at a variety of angles but usually well below steep-sided, flat-topped plateaus called mesas, and to similar
90°. Dendritic patterns develop in areas where there are no but smaller buttes (Fig. 16.5). With modestly dipping beds
marked lithological or structural controls affecting drainage. an asymmetric cuesta is produced, while steeply dipping
They are frequently associated with horizontal, or nearly strata at an angle of around 45° lead to the formation of an
horizontal, sedimentary strata, or with massive igneous approximately symmetric ridge known as a hogback
rocks. They may also be found on complex folded strata, or (Fig. 16.3).
on contorted metamorphosed rocks where lithological At its simplest, folded strata consist of pairs of limbs of
variations (in terms of resistance to weathering and erosion) roughly equal inclination. Fluvial dissection of a series of
are °insufficient to modify the dendritic pattern. folds will initially produce a symmetric trellis drainage
Where sedimentary rocks are tilted there may be a suc- pattern with trunk streams following the synclinal axis. But
cession of relatively weak and relatively resistant lithologies major valleys are frequently seen to be developed along
exposed. River channel incision will tend to be more active anticlinal axes with the synclinal axes forming ridges. A
on the less resistant lithology leading to the development of possible explanation for this inversion of topography is that
a strike valley flanked on the up-dip side by a dip slope the extension of strata across anticlines promotes joint
and on the down-dip side by an escarpment (Fig. 16.2). widening, and that series of joints along anticlinal axes then
This creates a trellis drainage pattern, with roughly parallel become favoured sites for weathering and erosion. In many
strike streams being joined at high angles by short dip cases, however, it can be demonstrated that several kilo-
streams and anti-dip streams (Fig. 16.3). As the fluvial metres of rock have been remo~ed by fluvial erosion prior
system incises downwards it will also migrate laterally, a to the exhumation of a particular anticlinal structure and
process known as homoclinal shifting. This arises from the that many phases of topographic inversion and 'normal'
following of a weak downward-dipping lithology by the relief development are likely to have occurred during the
strike stream and enhances the asymmetry of the strike valley, prolonged phase of landscape lowering. Whether an anti-

Strike
Present'--_ _"<--'........ stream
surface

Future
surface-------->,..,

I Resistant
I ./ beds
-----t---- I
I
I
~
Homoclinal
shift

Fig.16.2 Trellis drainage pattern developed in dipping strata with alternating weak and resistant bedso Homoclinal shifting of the
strike valley system down-dip is also shown. (Modified from A. L. Bloom (1978) Geomorphology: A Systematic Analysis of Late
Cenozoic Landfonns. Prentice-Hall, Englewood Cliffs, Fig. 11-10, p. 263.)
408 Endogenic-exogenic interactions

Fig.16.3 Trellis drainage pattern developed on steeply dipping Precambrian strata, South Australia.

• --="'-.;r~-------':'-.:---- Butte

Mesa

Fig. 16.4 Relationship between topographic forms and angle of dip ofstrata with alternating weak and resistant beds. (Modified from
A. L. Bloom (1978) Geomorphology: A Systematic Analysis of Late Cenozoic Landfonns. Prentice-Hall, Englewood Cliffs, Fig. 11-9,
p.262.)
 Tectonics and drainage development 409

Fig.16.5 Mesas and buttes in Monument Valley, Utah, USA, in afluvially eroded landscape developed in massive, flat-lying
sandstone formations.

cline is associated with a positive or negative topographic ciently well developed to produce linear 'zones which are
form probably reflects the relative resistance of the litho- more susceptible to weathering and erosion. In such cases
logy exposed along its axis at a given time. downcutting rivers will tend to exploit these lines of weak-
Obviously, folded strata are not infinitely continuous ness and the drainage pattern will be modified as a con-
along strike. In some cases folds plunge along strike over sequence. Such joint control is often evident on granitic
short distances in relation to the length of their axes. In rocks where an orthogonal jointing system is reflected in a
these circumstances a simple trellis pattern is modified by rectangular drainage pattern characterized by numerous
the bending of trunk streams around the limbs of folds. channel junctions with angles around 90°.
Except where folding is relatively recent, fluvial erosion of Joint and fracture orientations may correspond to either
folded strata will give rise to a landscape of successive ancient stress fields or to recent patterns of stress in the
cuestas and hogbacks facing inward towards the fold axes. lithosphere (the neotectonic stress field) which, at the
Where folds have very large dimensions and low ampli- broadest scale, is related to plate tectonic processes. There
tudes extensive cuesta and strike valley landscapes are has long been controversy, for instance, over whether the
developed; the Paris Basin of northern France and the trends of valleys in Switzerland are related to active tec-
Weald of southern England are well-studied examples. The tonic controls or whether they simply represent the remnant
Paris Basin is a large synclinal structure consisting of a of a drainage system created initially in the Miocene on a
succession of outward facing escarpments. The Weald, on surface sloping to the north towards an ancestral River
the other hand, is a complex eroded anticlinal structure with Rhine. Statistical analysis of valley orientations indicates a
inward-facing escarpments. non-random distribution corresponding broadly to major
joint directions and thus a strong tectonic control over
drainage patterns (Fig. 16.6). Studies in many other parts of
16.2.2 Fracture and joint control
the world have demonstrated a similar correspondence be-
Even on homogeneous lithologies, jointing may be suffi- tween drainage orientations and stress fields, and such a
410 Endogenic-exogenic interactions

tectonic control also appears to extend in some cases to

glacial valleys.
A further interesting example of the relationship between
fracture and joint systems and drainage is to be found in
north-west Europe. A north-west-trending system of large-
scale fractures in the Late Cretaceous and Paleogene strata
of southern England and northern France appears to have
been created by the stress field associated with the Late
Neogene and Quaternary phase of lithospheric compression
in the Alps. In northern France the system of fractures and
joints is coincident with a rectangular drainage pattern. In
s.outhern England, however, no such correspondence is evi-
dent, probably because channel orientations there are more
influenced by additional local fracture systems related to
flexures which trend obliquely to the major north-west
orientated fracture system.
Relationships between neotectonic stress fields and drain-
age patterns are most evident at the macroscale and mega-
scale since at smaller scales local controls over joint patterns
become predominant. The important point is that any signi-
ficant deviation from a random (dendritic) drainage pattern
at the megascale or macroscale should lead us at least to
consider the possibility of tectonic control (although depo-
sitional landforms such as drumlins and dune ridges can also
exert an important non-random influence over drainage
patterns).

16.2.3 River capture

An important process in the development of drainage
subject to structural controls is river capture, or stream
piracy (although it is important to emphasize that the pro-
c N cess can also operate in the absence of such controls). River
capture occurs when one stream erodes more aggressively
than an adjacent stream and captures its discharge by inter-
secting its channel. The faster rate of erosion of the 'captur-
ing' stream may be due to it having a steeper gradient or
discharge, or to the fact that it is eroding less resistant strata
or to a combination of these factors (Fig. 16.7). At the macro-
scale an additional factor could be contrasts in precipitation
either side of a mountain chain, with high discharge streams
on its more humid side cutting back to capture rivers on its
less humid side.
The .point at which the capture occurs is frequently
indicated by a sharp change in channel direction, often of
the order of 90°, known as an elbow of capture. The removal
of discharge from the captured stream (known as behead-
ing) means that its source will move down the original
Fig.16.6 Correspondence between joint direction and valley; the old river course upstream will be marked by
orientation of river channels in Switzerland: (A) drainage fluvial deposits and the abandoned river valley will form a
network; (B) rectified drainage network constructed by joining wind gap.
tributary termini and junctions with straight lines from which
river channel orientations are measured; (C) polar histogram of River capture is only possible through the agency of
river channel orientations. (Modified from A. E. Scheidegger headward erosion whereby valleys are extended towards
(1979) Geographica Helvetica 34, Figs. 2,3 and 4, p. 10.) and beyond their drainage basin divide. This process is well
 Tec~onics and drainage development 411

ment to unravel is that provided by transverse drainage.

Such a drainage pattern, which is also known as discordant
drainage, occurs when river channels cut across geological
structures (Fig. 16.8). Such drainage, which is common in
fold mountain belts, is often regarded as anomalous, al-
though this is not really an appropriate description. Much
of the early and influential research on transverse drainage
focused on the Appalachian Mountains of the eastern USA.
This was unfortunate in that these are ancient (Palaeozoic)
mountains which have 'experienced prolonged erosion and
no significant folding since the Permian. Consequently, it is
difficult to determine their early history in detail. This is
perhaps why several competing hypotheses were presented

® to explain their drainage characteristics and why they re-

mained current for so long - the evidence to reject them
was just not available.
The simplest explanation for transverse drainage is that
streams cut back by headward erosion along transverse
faults .. This is a varia~t of the river capture mechanism
since in cutting across a mountain belt such streams would
intersect the accordant strike drainage. This hypothesis is
testable with reference to existing field evidence. It has
specifically been rejected for the central Appalachians, but
may play an important role in drainage development in
evolving orogens (see Section 16.6).
Another explanation for transverse drainage is antece-
dence. The idea here is that a river which exists prior to the
uplift of a mountain belt and maintains sufficient erosive
power to keep pace with uplift will eventually be trans-
verse. Antecedence has been favoured as an explanation for
transverse drainage in many of the world's Cenozoic orogens,
for example the Alps, the Himalayas and the Cascade
Mountains of the north-west USA. Evidence favouring this
mechanism includes the general direction of drainage away
from the oldest structures in an expanding zone of defor-
mation. A variant of this mechanism is drainage inheri-
tance. This occurs where a drainage system developed on
Fig.16.7 Three stages of stream capture (A-C). The stream at an erosion surface lacking a sedimentary cover is rejuve-
the lower elevation cuts back through less resistant strata to nated and erodes down on to underlying structures which
capture the stream flowing at a higher level on more resistant are discordant with respect to the inherited drainage
formations. (From S. A. Schumm (1977) The Fluvial System
.rWiley, New York and London) Fig. 6.36, p.235 after O. P. orientation.
Jenkins (1964) California Division of Mines Bulletin 135, Fig. 72, A third mechanism is superimposition (superposition).
p.196.) Drainage developed on a sedimentary cover could be let
down on to an underlying structure, and such superimposed
documented for microscale gully systems but its operation drainage would, initially at least, be discordant with respect
at larger scales is less clear. The key problem is how runoff to this structure. Similarly, there might be marked structural
sufficient to promote the channel erosion and valley down- differences between vertically adjacent crustal units,
cutting necessary for headward erosion can be generated especially where major thrusting and sliding has occurred.
from the small area between existing valley heads and the Although a plausible mechanism, superimposition is
drainage divide. extremely difficult to verify except in the case of very young
orogens where vestiges of the original sedimentary cover
remain. In ancient mountain belts, denudation will have
16.2.4 Transverse drainage
removed all the evidence of any pre-existing sedimentary
One of the most perplexing problems of drainage develop- cover.
412 Endogenic--exogenic interactions

Fig.16.8 Glen Helen Gorge, in the western Macdonnell Ranges of central Australia, cut by a transverse southward-flowing
headwater tributary of the Finke River through a resistant east-west-orientated ridge. This transverse drainage appears to have
evolved through a combination of drainage inheritance and limited superimposition.

16.3 Tectonics and drainage adjustment between tectonic deformation, erosion and deposition. The
" point is well summarized by the dictum that drainage
The traditional approach to the analysis of drainage history systems have a heritage rather than an origin.
assumed that drainage systems are originally formed on an
initial slope before gradually becoming adjusted to struc-
16.3.1 Effects of warping
ture. It is now widely acknowledged, however, that this
kind of model of drainage development is unrealistic. Except The interplay between tectonics and drainage is clearly
in the rare circumstances where drainage is initiated almost shown in the effects of warping. Drainage disruption will
instantaneously on a virgin surface (such as that provided occur if the rate of uplift along the upwarp axis exceeds the
by a newly formed volcanic island) there is likely to be an rate of downcutting. The effects can be quite complex (Fig.
existing topography which will influence any subsequent 16.9) but the specific drainage patterns generated are, in
drainage development. This will invariably be the case with some cases, sufficiently diagnostic to enable important ele-
drainage developed on a landsurface newly exposed by a ments of a region's tectonic history to be reconstructed
fall in sea level. In these circumstances there is exposure of from drainage pattern evidence. In areas of minimal relief,
a pre-existing submarine topography which may have been with very low stream gradients, quite modest tectonic move-
partly formed during an earlier phase of sub-aerial expo- ments can have profound effects on drainage patterns. This
sure. There will also be an existing sub-aerial drainage is partly because only a very slight tilt of the landsurface,
system on the adjacent littoral zone which will extend over perhaps by only a fraction of a degree, is necessary to cause
the newly emerged coastal margin. Rather than considering drainage reversal. Another factor is that a river with a very
hypothetical starting points for drainage development, it is low gradient, especially when it also has a small discharge,
more fruitful to consider the progressive changes and modi- will be incapable of a rate of downcutting sufficient to
fications that can take place. There is a constant interplay maintain its original course.
 Tectonics and drainage development 413

Fig.16.9 Effects of warping on drainage systems. Downstream

a/the upwarp the channel gradient will be increased along part
of the channel (A), but the discharge from upstream of the upwarp
axis will be cut off (beheading) (B). Along the axis of the upwarp
channel gradients will be reduced and an area of disorganized
drainage, possibly including shallow lakes, may become
established (C). If upwarps are of low amplitude and long a 100
l.--.-...J
wavelength then regions ofpoorly integrated, low gradient km

drainage may be very extensive. On the upstream flank of the

upwarp the drainage will be reversed (D) as far as the point
where the initial downstream slope is re-established (E).
Tributary streams may have sufficient erosive power to maintain
their original direction offlow although their gradients will be
reduced (F). This will produce a barbed drainage pattern (usually
a diagnostic feature of drainage reversal). Lakes may also form
where reversed drainage joins the upstream drainage which has
maintained its original flow direction (G), but the development of
strike drainage parallel to the upwarp axis (H) may prevent lakes
from forming. (Based on C. Ollier (/981) Tectonics and
Landforms. Longman, London, Fig. /2.7, p./68.)

The effect on drainage systems of recent, and quite rapid,

warping are to be seen in the East African Rift System (Fig. Fig.16.10 Effects of warping on drainage development in the
16.10(A». A high rate of surface uplift occurred along a East African Rift System (A). Detail of Lake Kyoga in Uganda
NNE-trending axis to the east of the Western Rift and de- formed by drainage reversal caused by the Late Pleistocene back-
tilting of the plateau east of the Western Rift is also shown (B).
formed an Early Pleistocene landsurface. Rivers originally
draining into the rift were defeated and drainage reversal
occurred, giving rise to a barbed drainage pattern with
junction angles greater than 900 • Lakes Victoria and Kyoga the relative movement is downward on the downstream
developed due to ponding of this reversed drainage until side then a waterfall will be formed.
breaching by the present Victoria Nile re-established a Strike-slip faulting also tends to involve some vertical
connection with the Western Rift into Lake Albert. The movement and so may give rise to similar adjustments to
peculiar shape of Lake Kyoga (Fig. 16.10(B» is a result of normal and reverse faulting. The main effect of strike-slip
the drowning of a dendritic drainage system as a conse- faulting, however, is the creation of offset drainage along
quence of the warping of Early Pleistocene surface. This the line of the fault (Fig. 3.25). This is displayed in classic
particular example not only illustrates the effect of crustal fashion along sections of the San Andreas Fault System
warping on drainage but also indicates how the development where a series of crustal movements has led to repeated off-
of drainage patterns through time can be used to elucidate setting of streams traversing the fault. Various drainage pat-
the sequence and location of tectonic activity. terns have been produced in consequence, including right-
angle offsets, 'Z' patterns and trellis forms (Fig. 16.11).
Such drainage adjustments can provide valuable evidence on
16.3.2 Effects of faulting
the magnitude of strike-slip movement, but care must be
Although in comparison with warping its effects are nor- taken in their interpretation since modest differential ver-
mally more localized, faulting can also cause significant tical movements of only a few centimetres across a fault
drainage adjustments. A relative upward movement on its can be sufficient to deflect small streams and produce false
downstream side forming a fault scarp will, if of sufficient offsets, while vertical displacements of just a few metres
magnitude, behead a stream and create a fault-line lake. If are capable of offsetting even large river channels.
414 Endogenic-exogenic interactions

ABC

·.,BEHEArJED
I
I

~
t.1. .. . ..1,1·1·
t
, "

If
(!
t
tl.>;
BEHEADEDlt
.••• \III
I' ,
~ ~
<:
:.~.~~~~
. .'. .~~ ~I .. ff'~~~ •.
~ ••• _ ..

D E F
"

Fig.16.11 Diagrammatic representation ofpatterns offault-related stream channels occurring along the San Andreas Fault in the
Carrizo Plain area, California. Offset channels are shown in (A) and (B): (A) the misalignment ofsingle channels directly related to the
amount offault displacement and the age of the channel - there is no ridge on the downslope side of the fault and beheading is
common; (B) misaligned paired stream channels. Patterns resulting from a combination of offset and deflection are represented in
(C)-(F): (C) compound offsets of ridge spurs and both the left and right offsetting and deflection of channels; (D) trellis drainage
produced by multiple fault strands and shutter ridges; (E) indicates how a combination of offset and deflection by shutter ridges can
give rise to exaggerated or reversed apparent offset; (F) shows how capture by an adjacent channel followed by right-lateral slip may
produce a '2' pattern. Two types offalse offset are represented by (G) and (H): (G) shows how differential uplift may deflect streams to
produce false offsets, and (H) illustrates en echelon fractures over a fault zone followed by streams producing false offsets. (After R. E.
Wallace (1975) in J. C. Crowell (ed.) San Andreas Fault in Southern California. California Division ofMines and Geology, Special
Report 118, Fig. 2, p. 243.)

16.3.3 Drainage evolution in rift valleys or other fault systems which accommodate the movem.ent
between half-graben units of opposite polarity (Fig. 16.l2(B)).
Traditional views of drainage evolution in rift valleys were Large deltas, such as the Turkwel delta of the Turkana Rift
based on the classical symmetric model of rift structure. in Kenya, can form where these diverted rivers enter a rift
These emphasized the importance of axial drainage along valley.
the rift and the role of marginal streams crossing the oppo- The drainage of Lake Tanganyika in East Africa pro-
sing fault scarps in constructing alluvial fans on either side vides an extreme case where listric boundary faults have
of the rift. As we saw in Chapter 4 (see Section 4.3 and acted as a barrier to drainage (Fig. 16.13). In other parts of
Figure 4.9) the symmetric model of rift development has the East African Rift System both diverted rivers and axial
been challenged on the basis of seismic data which indicate drainage are more common. Nevertheless, irrespective of
that some rift valleys, at least, have an asymmetric half- their detailed structure, most of the sediment supplied to
graben structure. rifts with a half-graben structure generally comes not from
An important consequence of the half-graben model of streams draining the steep footwall scarp but from axial
rift structure is the back-tilting of drainage behind the drainage or the much larger catchments of the opposing
footwall scarp (Fig. 16.12(A)). This occurs through a com- roll-over section of the rift. Consequently, the alluvial fans
bination of the regional doming that frequently accom- located at the footwall scarp are normally small in com-
panies rifting and displacements associated with movements parison with the extensive alluvial cones which develop on
along the major listric boundary fault. If movement occurs the roll-over zone.
along more than one major boundary fault, then small back-
tilted blocks are formed which prevent direct drainage to
16.3.4 History of the Colorado river
the rift and divert rivers along strike. These rivers may
eventually gain access to the rift at the site of transfer faults The Colorado River of the western USA provides an excel-
 Tectonics and drainage development 415

A
Back- Strike
tilted drainage

~~ Major Fault
/
Footwall
scarp

Malagarasi
Listric River
boundary
faults
Catchment

B Strike drainage between

back-tilted blocks
Back-
tilted
drainage
/
_ _ _ Transferfault or
~ accommodation
'\ fault zone

Listric
boundary
fault
\
Lake Rukwa
Rift Basin

\
Roll-over
drainage

~
km
o 100
Axial
J drainage

Fig.16.12 Schematic representation of the relationship between

drainage patterns and half-graben rift structure showing cross-
sectional (A) and plan view (B). (Modified from L. E. Frostick and
I. Reid (1987) in L. Frostick and I. Reid (eds) Desert Sediments:
Ancient and Modern. Geological Society Special Publication 35, Fig.16.13 Lake Tanganyika illustrating relationships between
Fig. 3, p. 56. Reproduced by permission of the Geological major listric boundary faults (dashed lines with ticks indicating
Society.) downthrow side), drainage basin size and primary direction of
drainage systems (drainage vectors). Notice how the largest
lent illustration of the way in which .large-scale drainage catchments occur predominantly on the roll-over side of each
half-graben unit. (From L. E. Frostick and J. Reid (1987) in: L.
evolution is intimately linked to interpretations of tectonic Frostick and I. Reid (eds) Desert Sediments: Ancient and Modem.
history (Fig. 16.14). The composition and structure of Ter- Geological Society Special Publication 35, Fig. 5, p. 57.
tiary gravel deposits in ancient stream channels along the Reproduced by the permission of the Geological Society.)
southern margin of the Colorado Plateau indicate a source
416 Endogenic-exogenic interactions

16.4 Active tectonics and channel adjustment

As we saw in Chapter 9, alluvial channels are very sensi-

tive to variations in discharge and sediment load charac-
teristics, and many of the changes in channel morphology
that occur over time can be attributed to these variables. It
is important, none the less, to consider the possible role of
active tectonics in affecting channel properties. As we have
already noted, active faulting can have a dramatic impact
on fluvial systems (see Section 16.3.2); yet slower, vertical
deformations can lead to more subtle, but still significant,
adjustments in channel morphology. Anomalies in stream
channels which may be indicative of active tectonics include
the local development of braiding or meanders, the creation
of ponds or marshes, the accumulation of alluvial fills and
the widening or narrowing of channel reaches. The defor-
mation of fluvial channels can take many forms, but the
primary effect is usually either a local reduction or increase
in channel gradient, or lateral tilting of the channel leading
to channel aggradation or incision.

16.4.1 Field examples

The Mississippi Basin in the USA provides several well-
documented examples of the effects of slow, ongoing
deformation on river channel characteristics. On the borders
of the states of Louisiana, Arkansas and Mississippi an area
known as the Monroe uplift has been active since the
Fig.16.14 Course of the present-day Colorado River in relation Tertiary, and repeated levelling surveys during' the period
to the uplifted Colorado Plateau. 1934-1966 have shown that surface uplift is still continuing,
at least on the southern flank of the uplift. Quaternary uplift
to the south and south-west. At this time central Arizona is demonstrated by convexities in the longitudinal profiles
was more elevated than the region now occuped by the of river terraces of this age, and continuing uplift is indi-
Grand Canyon. Radiometric dating of basalt pebbles in cated by convex-up sections in the present-day valley floor
these deposits show that this situation lasted until at least profiles (Fig. 16.16). The response of an alluvial channel to
10 Ma BP . The cutting of the Grand Canyon by the Colorado this kind of slow crustal deformation can be illustrated by
River was promoted by the uplift of the Colorado Plateau Big Colewa Creek, one of the streams shown in Figure
(Fig. 16.15). This apparently occurred between 10 and 16.16. Several effects are apparent, including an abrupt
5 Ma BP since the age of valley-fill basalts indicates that the threefold downstream increase in channel gradient at the'
south-flowing drainage of the Verde River away from the axis of uplift which has led to channel incision and a
uplift had developed by 5 Ma BP. corresponding increase in average bank height from 2 to
Setting these changes in a broader perspective, the 6 m. Channel sinuosity increases from 1.2 upstream of the
Colorado Plateau was low relative to the elevated Basin uplift axis to 1.7 in the reach immediately downstream of
and Range Province to the west until at least 18 Ma BP. the zone of maximum uplift, before gradually falling to 1.5.
Between 18 and 10 Ma BP faulting to the west of the Active tectonics may also have important indirect effects
Colorado Plateau led to the differentiation of the Basin and on river channel morphology. A notable example is pro-
Range Province and the development of interior drainage vided by the New Madrid earthquake which occurred in the
within it. The ancestral Colorado only began to drain to the centre of the Mississippi Basin in 1812. It appears that this
west between 10 and about 3 Ma BP, as the Colorado Plateau major earthquake promoted widespread collapse of un-
experienced uplift relative to the 'collapsing' Basin and stable banks along the lower Mississippi which led to a
Range terrain. By about 3 Ma BP a well-developed Grand large injection of sediment into the river. This in tum caused
Canyon had been formed and the Colorado River had almost major downstream adjustments in channel morphology,
attained its present depth at its downstream end. including a reduction in meandering and the formation of
 Tectonics and drainage development 417

Fig.16.15 Landsat image showing the Grand Canyon carved into the uplifted Colorado Plateau by the Colorado River. The area
covered by the image is shown on Figure 16.14. (1mage courtesy N. M. Short.)

islands and point bars. These changes were so significant useful means of isolating these effects, and thereby reveal-
that they persisted for several decades after the earthquake. ing changes in channel characteristics which are influenced
by tectonic deformation of the channel bed.
One of the most sophisticated of such studies carried out
. 16.4.2 Experimental studies
to date has been that by S. Ouchi working at the Colorado
In the field it is often very difficult to separate the effects on State University. Using a large laboratory flume to investi-
channel morphology of active tectonics from variations in gate the effects of vertical deformation across alluvial
discharge and sediment characteristics, since there is often channels, he found that changes produced depended not
a complex interaction between these groups of variables. only on the rate and amount of displacement but also on the
Experimental studies in which the hydrological and morphology and sediment properties of the channel (Table
sedimentological variables can be controlled thus provide a 16.3). Braided channels were found to be unable to change
418 Endogenic-exogenic interactions

tion of terraces. Meandering channels, by contrast, adjusted

by an increase in sinuosity (resulting in greater channel
length), bank erosion and point bar growth in the steepened
reach downstream, and by flooding over channel bars in the
flattened reach upstream of the uplift axis. 'Subsidence was
c: found to lead to deposItion in the reach astride the synclinal
...
.2
Cll
axis and the downstream development of transverse bars in
>
0) braided channels but in meandering channels the flooding
ijj
of bars downstream of the zone of maximum subsidence
and an increase in sinuosity upstream was observed.
These experimental studies show that although active
deformation commonly occurs at rates which are slow in
comparison with changes in the other variables which
90 60 30 0 30 60 affect channel morphology, if such deformation persists for
Valley Distance from Uplift Axis (km)
only'a few decades the displacements achieved will be suf-
ficient to generate significant changes in channel morpho-
... 6 UPPER B
..r: ZONE logy. River channel morphology, therefore, appears to provide
Cl LOWER
'Q) ZONE a very sensitive means of detecting ongoing crustal defor-
I
~ "0 mation, assuming non-tectonic factors can be clearly isolated.
c:_
Cll
al_
E 4
0)
..r:
u
...c: In identifying the presence of active tectonics a wide range
0)
c: "0
a>c: E 0) of responses of fluvial systems can be considered (Table
c:
...c:
~,
..r:
O)U
..r:
U
c: 16.4).
0) > c:
Cll
..r:
U 0c: ......
._ 0)

u ...
Cll c:
...~c:
-0) ~
2
16.5 Global tectonics and continental drainage
"0
C c The geomorphic significance of the world's major rivers is
illustrated by the fact that, excluding the largely ice-covered
~~ 0·023
c: Channel gradient
0)0)
.- a. land masses of Greenland and Antarctica, some 47 per cent
~.2 of the area of the continents is drained by only 50 river
e;(J) 0·015
- > basins. Indeed the five largest basins - the Amazon, Zaire
0) 0)
c:= (Congo), Mississippi, Nile and Yenisei - alone account for
lij~ 0·008
..r: 10 per cent, with the Amazon itself contributing 5 per cent.
u
On examining drainage basins of this size it becomes
0·0 immediately apparent that tectonics exerts a strong in-
fluence over their large-scale morphology and development
2·0 o through time.
At the global scale it is possible to identify five major
types of drainage system in relation to tectonic setting
...
>
(Table 16.5). These are illustrated in Figure 16.17 where
.~ 1-5
:J
c the relationship between tectonic setting and drainage system
U5 characteristics is readily apparent. A passive margin basin
1-0 ............. .------.----r--..------.----r--,.-----r-...-------r--, drains to a passive continental margin from a cratonic
100 80 60 40 20 0 interior which lacks a Mesozoic or Cenozoic mountain belt.
Valley Distance from Mouth (km) This type of drainage system is particularly common on the
African continent and for this reason is sometimes referred
Fig.16.16 Affects of uplift on channel characteristics for streams
crossing the Monroe uplift in the lower Mississippi Basin: (A) the to as Afrotype drainage. Passive margin - distal orogen
10ngitL!dinal profiles of various stream channels,. (B), (C) and (D) basins are so called because, whilst draining to a passive
changes in channel bank height, channel gradient and channel margin, a significant proportion of their tributaries have their
sinuosity for Big Colewa Creek. (After A. W. Burnett and S. A. source in a mountain belt along the distal margin of the
Schumm (1983) Science 222, Figs 3(b) and 4, p. 50.) basin. The classic example of this kind of drainage basin is
the Amazon, but such drainage systems exist along most of
their sinuosity in response to anticlinal uplift but rather the eastern margin of the American continent as a con-
adjusted by channel incision which in tum led to the forma- sequence of the the American Cordillera which marks the
 Tectonics and drainage development 419

Table 16.3 Re ponse of experimental fluvial channels to uplift and subsidence

UPSTREAM FROM ZONE UPSTREAM FLANK AXIS OF DOWNSTREAM FLANK DOWNSTREAM FROM
OF DEFORMATION OF DEFORMATION DEFORMATION OF DEFORMATION ZONE OF DEFORMATION

Aggradation Degradation Aggradation

sQ.
l
13
"0
;:)
Thalweg shift
Submerged bars
Terrace formation
Single bars Braided

"§
' c:
"u Degradation Aggradation Degradation
l:Q
~
:>"0
en 'v.;
" Single thalweg Braided Flooding Single thalweg

Aggradation Degradation Aggradation

S Flooding Sinuosity increase
I'5 Q.
,;:) Multiple channels Bank erosion
(anastomosing)
~
'5
"0
;j "uc: Degradation Aggradation Local scour
~ "
"0
'u; Sinuosity increase Flooding, cutoffs
-g Bank erosion Multiple channels
en

Source: Modified from D. I. Gregory and S. A. Schumm (1987) in: K. Richards (ed.) River Channels: Environment and Process, (Blackwell, Oxford)
Fig. 3.3, p. 46, after S. Ouchi (1985) Geological Society ofAmerica Bulletin, 96, Tablc 1, p. 508.

Table 16.4 Types of adjustments of river channel characteristics relevant Table 16.5 Classification of major types of drainage system in relation to
to the identification of active tectonics tectonic setting

Deformation of valley-floor longitudinal profile TYPE CHARACTERISTICS EXAMPLE

Deformation of channel projected profile
Passive margin Drain to passive continental Zaire
Change of channel and valley gradient margin from cratonic hinterland
Change of channel width and depth lacking Mesozoic or Cenozoic
Conversion of floodplain to low terrace mountain belts
Active floodplain development Passive margin - Drain from Mesozoic or Cenozoic Amazon
distal orogen mountain belt to a passive margin
Variations in channel stability including
reaches of active channel incision, deposition, channel shifting and Intra-orogen Drain along strike within Irrawaddy
formation of knickpoints mountain belt
Indirect effects both upstream and downstream of Extra-orogen Drain along strike marginal to Ganges
the zone of deformation including deposition, channel incision, bank mountain belt
erosion and flooding Trans-orogen Drain across strike of a Columbia
mountain belt

complex zone of plate interaction along its western peri- Basin southwards towards the Gulf of Mexico.
phery; such drainage basins are indeed sometimes referred The three other types of drainage basin are associated
to as Amerotype. The presence of a high mountain belt in with mountain belts. An intra-orogen basin is located pri-
the distal part of the basin is probably the reason why the marily within a mountain belt and the drainage network
drainage pattern of this type of basin is relatively unidirec- runs parallel to its strike. Drainage parallel to strike also
tional in comparison with the rather more centripetal drain- characterizes an extra-orogen basin, but in this case the
age which characterizes Afrotype basins. basin is located along the margins of a mountain belt.
The often large drainage area of passive margin basins of Trans-orogen basins differ from the other two in that the
both types is one of the reasons why deltas are preferen- drainage traverses the strike (that is, runs across the grain)
tially located along passive margin and marginal sea coasts of the mountain belt.
(Fig. 13.2). Both types of passive margin basin may also be While useful, this classification scheme presents an over-
modified by the presence of a degraded ancient orogen simplified view of the relationship between global tectonics
along the continental margin (formed during an episode of and continental drainage. Some basins may contain elements
continental collision prior to the subsequent continental rift- of more than one type - the upper part of the Brahmaputra
ing). These may present a topographic barrier sufficient to River, for instance, runs along the strike of the Himalayan
prevent major continental drainage outlets developing along orogen before finally turning at right angles across the
extensive sections of passive margins. An example of this eastern Himalayas to join the Ganges at the head of the Bay
effect is the Appalachian Mountains of the eastern USA of Bengal. Moreover, other tectonic settings need to be
which appear to 'deflect' the drainage of the Mississippi considered. The cratonic interior basins of Africa, such as
420 Endogenic--exogenic interactions

IH~ffil Mountain belts of Cenozoic age

~ Partially eroded mountain belts of Mesozoic
~ and Late Palaeozoic age
6 2
Illll!lllllli Highly eroded mountain belts of 10km

1~
IlIlllllIl Early Palaeozoic age
~=.-:=====t

1000 0 1000
km

Fig.16.17 Examples of major types of continental drainage system in relation to tectonic setting.

the Chad and Okavango Basins, the river systems of block- because they provide the foci for extensive drainage basins
faulted terrains, such as those of the south-west USA, and entering the alluvial plain from the adjacent mountain
the river systems of western China dominated by the effects mass. They supply much more sediment than the smaller
of active strike-slip faulting all reflect specific tectonic streams draining the intervening salients. This is exempli-
controls. fied by the drainage of the Himalayas where the Punjab re-
entrant is occupied by the Indus, and the upper Assam re-
entrant is occupied by the Brahmaputra (Fig. 16.18). Be-
16.6 Drainage development in orogens tween these two major river systems only comparatively
Three components of structural control are typically evident small rivers drain to the Ganges aligned longitudinally
in the development of drainage in orogens. One is the dis- along the Indo-Gangetic plain.
tinct separation of the erosional zone of the mountain mass
from the adjacent aggradational basin by a zone of reverse
faulting. Secondly, longitudinal fault systems within the 16.6.1 Drainage of the Zagros Mountains
orogen predispose the drainage to be similarly orientated Modem orogens, which are actively growing and where
along strike; moreover, drainage in the aggradational basin primary structures have not yet been removed by erosion,
adjacent to the orogen also tends to be longitudinal except provide the optimal conditions for studying the relationships
for comparatively short, but erosionally vigorous, streams between structure and tectonics in mountain belts. The
draining directly from the mountain front. A third key Zagros Mountains of Iran, which represent an interconti-
structural element is the existence of salients and re- nental collision orogen, provide an excellent example. Parts
entrants along the boundary between the mountain mass of this range have experienced major folding and uplift as
and the adjacent basin. recently as the Pliocene. Rivers rising near the north-eastern
Along the Himalayas such irregularities probably date margin of the highland flow completely through the orogen
from the collision of the Indian and Eurasian Plates. Re- on their way to the Mesopotamian-Persian Gulf trough
entrants are structurally low regions and are significant (Fig. l6.l9(A». A detailed analysis of the drainage systems
 Tectonics and drainage development 421

Land over Axis of highest oI

500
I
2000m peaks of the
Himalayas km

Fig.16.18 Drainage systems of the Himalayas and Tibetan Plateau.

5p
Hamadano km
Reach developed in transverse subsequent valley;
• later superimposed upon resistant formations at
depth.

'i1 Reach antecedent to local structure.

o Reach superimposed from disharmonically folded

overburden.

X Reach extended headward in resistant rock.

Origin oftransverse reach indeterminate; principally

in lone of thrust faults.

500
km

Fig.16.19 Drainage of the Zagros Mountains, Iran: (A) general tectonic setting of the major drainage systems; (B) most probable
origins of stream reaches transverse to anticlinal structures of the central Zagros. (Modified from T. M. Oberlander (1965) The Zagros
Streams Syracuse University Geographical Series 1, Fig. 1, p. 2 (A) and T. M. Oberlander (1985) in: M. Morisawa and 1. T. Hack (eds)
Tectonic Geomorphology. Allen and Unwin, Boston Fig. 3, p.165 (B).)
422 Endogenic-exogenic interactions

of the Zagros Mountains by T. Oberlander has demon- of relatively less resistant strata which were successively
strated that no single hypothesis of transverse drainage can exposed as the folds were eroded. Far from being anoma-
explain all the observed drainage patterns. None the less, lous, the polygenetic development of transverse drainage in
every drainage anomaly in the region can be classified as the Zagros Mountains producing spectacularly discordant
some variant of one or other of the general models of drainage patterns is viewed by Oberlander as an inevitable
transverse drainage development (Fig. 16.19(B». consequence of normal stream development in the fold
Superimposed drainage was found to be confined to structures typical of intercontinental collision orogens (Fig.
specific structural associations and no evidence was found 16.20).
of a regional cover mass or erosion surface from which ex-
tensive superimposition or drainage inheritance could have
16.6.2 Drainage of the Himalayas
occurred. Similarly, antecedence was diagnosed as being
localized rather than regional. The most widespread me- The traditional view of the transverse rivers of the Himalayas,
chanism leading to the development of transverse drainage which include the Indus and the Brahmaputra, is that they
has apparently been headward stream extension into zones are antecedent, having originated on the high Tibetan

Fig.16.20 Model of drainage evolution in open folds, such as those which characterize much of the Zagros Mountains, in which the
depth of readily erodible beds between resistant formations exceeds fold amplitudes: (A) the initial erosion of anticlines exposes
erodible formations and leads to the expansion of axial basins; (B) the expanding axial basins begin to merge and so create outcrops of
erodible strata which extend across fold axes; (e) through-flowing transverse streams occupy continuous outcrops of erodible beds and
are superimposed on to lower resistant beds emerging in anticlinal cores; (D) anticlinal mountains which are resurrected on the lower
resistant beds are breached by the superimposed transverse streams (After T. M. Oberlander (1985) in M. Morisawa and J. T. Hack
(eds) Tectonic Geomorphology. Allen and Unwin, Boston, Fig. 8, p. 172.)
 Tectonics and drainage development 423

Plateau prior to the major phase of uplift of the Himalayan rivers, many of these are seen to coincide closely with
ranges. The upper courses of both the Indus and the transverse anticlines, that is, anticlinal structures which are
Brahmaputra (Tsangpo) closely follow the Indus-Tsangpo transverse to the main axis of the Himalayas. The River
suture zone marking the initial line of collision between the Arun, which has its headwaters just a few kilometres from
Indian and Eurasian continents. These rivers traverse the the Tsangpo and which has widely been regarded as antece-
axis of the Himalayas through spectacular gorges. Where dent, follows one of these structures (Fig. 16.21). The Arun
the Indus leaves Kashmir it is only 1000 m above sea level, transverse anticline, located east of Mount Everest, has an
but is overlooked by the world's fourth highest mountain, amplitude of 10000 m and the River Arun has incised itself
Nanga Parbat, towering to over 8000 m. A number of other to a depth of 15 000 m into this structure as isostatic re-
large rivers draining south from the Himalayas presently bound has caused significant crustal uplift in response to
have source tributaries on the Tibetan Plateau and on the denudational unloading. The transverse warps appear to
northern slopes of the Higher Himalayas. The antecedence have developed late in the history of the Himalayas, so if
model of Himalayan drainage requires that the river systems they have been instrumental in controlling the location of
are ancient since they must have originated before the transverse drainage systems then these must also be recent.
faulting, thrusting and folding that created the Himalayas. Rather than being antecedent it seems likely, then, that
A closer look at the relationships between river courses these transverse Himalayan fluvial systems represent a struc-
and geological structures in the Himalayas, however, indi- turally controlled form of superimposed drainage. Given a
cates that there is not complete discordance, as would be sufficient thickness of readily erodible strata within the
anticipated from a simple antecedence model. The Indus transverse anticlines of the Himalayas the unroofing of
and Brahmaputra, for example, cross the axis of the these structures could have enabled stream systems to locate
Himalayas exactly at the points where the trend of the preferentially within them. Suitable weak lithologies exist
range changes (Fig. 16.18). As for the other transverse within the Himalayas, but extensive erosion means that it is
not certain that they were present in the hypothesized loca-
lities. None the less, the correspondence between transverse
structures and river courses argues strongly against any
simple model of antecedence.

16.7 Drainage development associated with

passive margins
We have already noted how the major drainage basins of
Africa drain to passive continental margins, but the large-
scale morphology of these basins as well as the location of
their outlet to the ocean also appears to be subject to tec-
tonic control. The overall shape and size of the major drain-
age basins of Africa appear to be related quite closely to the
. basin and swell topography characteristic of the continent.
Breaches of swells (crustal upwarps) by rivers are rare
except where major river channels reach the coast. Here
. drainage outlets appear to be located either at structurally
low points along the margin between domal upw~s (as in
the case of the Orange and Zaire Rivers) or at aulacogens.
The classic instance of the latter situation is provided by the
Niger River which is located in the Benue Trough, an
aulacogen created during the formation of the southern
Atlantic Ocean. Another, perhaps less obvious, example is
~ Axes of major ranges
the Zambezi whose mouth is located within a complex rift
~ Major gorges
structure associated with the splitting of Madagascar away
from the African mainland.
Fig.16.21 Drainage in the vicinity of Mount Everest in the Somewhat similar structural controls have also been
eastern Himalayas. Note the discordant courses of the south- suggested for the Amazon. Its mouth is also located in an
flowing rivers which traverse the axis of the High Himalayas.
(Modified from A. Holmes (1978) Principles of Physical Geology
aulacogen which apparently continues inland as an elon-
(3rd edn, revised D. L. Holmes). Chapman and Hall, London, gated structural low as far west as the Andes. Such
Fig. 19.36, p. 387.) continental-scale structures originated during the break-up
424 Endogenic-exogenic interactions

of Gondwana and may have also influenced the courses of by continental break-up? This is an important question
other South American rivers (such as the Parana and Sao because many of the world's major drainage basins drain to
Francisco) which drain towards the Atlantic. passive margins formed by the break-up of Pangaea which
The relationship between tectonics and drainage patterns began about 200 Ma BP.
in Africa and South America are illustrated in Figure 16.22.
The two continents have been replaced into their approxi-
16.7.1 Drainage history of the African continent
mate relative positions prior to rifting around 140 Ma BP.
This reconstruction serves to emphasize the similarities Africa provides perhaps the best starting point for an
between drainage patterns and the tectonic settings of the investigation 'of the effects of continental break-up on drain-
passive margins on either side of the present-day South age development. This is both because it has no significant
Atlantic (although the pre-rift drainage would, of course, relict orogenic belts blocking drainage outlets across its
have been very different from that of the present). But how margins, and because, unlike South America, it is almost
do drainage systems adjust to the major disruption caused entirely rimmed by passive margins.

A Guinea - Guiana dome

B Nigeria - N.E. Brazil dome
C Angola - S. Brazil dome
D S.W. Africa - Uruguay dome

Fig.16.22 Relationship between present-day drainage and domal rift structures on the Atlcmtic margins ofAfrica and South America.
 Tectonics and drainage development 425

Even a cursory examination of a topographic map of The long profiles of the major river channels also raise
Africa shows that the drainage of the African continent has interesting questions. The gradient of the Zaire River along
a number of apparently anomalous feature . For instance, 2000 krn of its middle course averages only 0.05 m krn- 1,
the drainage pattern of several of the major basins is to a whereas from Kinshasa to Matadi near its mo~th the mean
large extent centripetal. This is particularly evident in the gradient increases by over 15 times to 0.78 m km- 1 as the
Zaire Basin where most of the major tributaries are orien- channel plunges down the knickpoint formed by the rapids
tated towards a point in the centre of the basin rather than at Stanley Pool. Other African rivers also have significant
towards its outlet. Other major rivers, such as the Niger, knickpoints such as theVictoria Falls on the Zambezi and
have long sections which flow directly away from a closely the Augrabies Falis on the Orange (Fig. 16.24).
adjacent coastline. Africa, along with Australia, is also
remarkable for the proportion of its total area occupied by
internal drainage systems.
Undoubtedly climatic factors, especially oscillations be-
tween arid and humid climatic regimes, have played an
important role in'the development of these drainage charac-
teristics. But tectonic controls also seem to have been crucial;
not only has the tectonically induced basin and swell topo-
graphy of Africa provided an overall constraint to drainage
basin development, but the presence of a crustal upwarp
extending around most of the margins of the continent has
probably been of great importance to the post-Gondwana
drainage evolution of the continent.
The effect of this marginal upwarp becomes very evident
when we examine the hypsometric curves for Africa's
major drainage basins (Fig. 16.23). Most obviou is the
very small proportion of the area of each basin at low
elevations. The form of these hypsometric curves is in stark
contrast to the form we might expect for mature drainage
basins which would have a significant proportion of their
total basin area at low elevations. Remember that the
drainage systems of Africa have had more than 100 Ma to
evolve since the final emergence of Africa as a discrete
continent after the break-up of Gondwana.

100

90

80 1 Zambezi
2 Orange
70 3 Zaire
4 Niger
60 5 Nile
c:
.2
.....
(tJ 50
>
~
w 40
Cf2.
30

20

10

0
0 10 20 30 40 50 60 70 80 90 100
Fig.16.24 The Augrabies Falls, a major knickpoint on the
% Area Orange River, southern Africa (top photo), which separates its
, Fig. 16.23 Percentage hypsometric curves for the five largest low-gradient braided middle course (middle photo) from its
drainage basins in Africa. deeply incised lower course (bottom photo).
426 Endogenic-exogenic interactions

One possible explanation for these anomalous hypsometric incorporation of much of the African continent into a fully
curves and longitudinal profiles is that the margins of developed external drainage system. The present-day pattem
Africa have experienced significant uplift in the relatively of drainage can best be explained by the large-scale capture
recent past, perhaps in the Late Pliocene and Pleistocene. of internal basins by the active headward erosion of streams
This view has certainly been put forward by several re- draining the oceanward flanks of the marginal upwarps.
searchers in one form or another, but it is not supported un- This idea is not entirely new since various interpretations
equivocaily by the geological and geomorphic evidence along these lines have been suggested previously, including
and, moreover, it is difficult to reconcile with what we the view that there was internal drainage over much of
know about the mechanisms producing uplift along passive Africa until the Miocene (Fig. 16.25). Other versions of this
margins. As we saw in Chapter 4, significant passive mar- drainage capture model have advocated rather earlier
gin uplift is likely to occur either immediately before con- capture events. Certainly, the relatively recent development
tinental break-up (in the active rifting model where domal of the Zaire Basin into an external drainage system is sug-
uplift precedes rifting) or soon after (in the passive rifting gested by the major knickpoint represented by the rapids
model where uplift arises as a consequence of the rifting below Stanley Pool and the extremely low gradients found
mechanism). The possibility remains, therefore, that the in the central part of the basin. Further south there is some
marginal upwarps that we see today influencing drainage evidence of major shifts in the location of the mouth of the
were present in some form at the time of passive margin ancestral Orange River, and of dramatic changes in the
creation, or within a few,million years thereafter. In most organization of fluvial systems draining to the eastern mar-
cases, any subsequent surface uplift has probably been gin of southern Africa (Fig. 16.26).
largely confined to flexural effects associated with passive
isostatic compensation as the coastal periphery of these
16.7.2 A model of passive margin drainage
passive margins has been gradually eroded.
development
This model of passive margin tectonics immediately sug-
gests a radical alternative model for the drainage evolution As yet the case for large-scale drainage capture is not proven,
of much of Africa. This model proposes that the upwarps but it does fit in well with current ideas about the tectonic
around the margins of Africa significantly delayed the evolution of passive margins. Much more research is needed
on this topic, but even now it is possible to put forward a
plausible model of how megascale drainage capture might
work.
Figure 16.27 provides a schematic representation of a
possible sequence of drainage adjustments during continen-
tal rifting and subsequent passive margin subsidence. The
model assumes that uplift precedes rifting and does not take
into account the possible development of an aulacogen (see
Section 4.5 for details of the tectonic evolution of passive
margins). The model is also specifically confined to the
evolution of rifted rather than sheared margins.
An additional factor which must be considered in the
drainage evolution of passive margins is the role of aulaco-
gens since they can be important foci for continental drain-
age. Of particular importance is the fact that they can be
established relatively soon after rifting (the Niger delta, for
instance, dates back at least 60 Ma) and thereby provide an
early outlet for continental drainage to a newly developing
ocean. The likely effects on drainage development of aula-
---... Hypothesised
Miocene
coastline cogens and other components of structural variability along
Present coastline a passive margin are illustrated in Figure 16.28 which is
----------
•"
km 1000
Ancient highlands
I
based on the 'plan view' of the passive margin development
discussed in Chapter 4 (see Figure 4.16).
,,~~ External drainage

Fig.16.25 Speculative map ofAfrican drainage as it might have

appeared immediately before the major uplift and rifting of the 16.7.3 Variation in passive margin drainage evolution
East African Rift System. (Modified from J. de Heinzelin (1964)
in F. Clark Howell and F. Bouliere (eds) African Ecology and Although it is possible to predict the broad outline of drain-
Human Evolution. Methuen, London, Fig. 7, p. 650.) age history from the likely tectonic -evolution of passive
 Tectonics and drainage development 427

Fig. 16.26 Highly conjectural reconstruction of the pre-Miocene drainage ofsouthern Africa (A) compared with the modern drainage
system (B). It has been suggested (c. 1. H. Hartnady per.comm.) that the protrusion on the east coast in southern Mozambique may
represent an enormous ancient delta which was fed by a very large drainage system possibly extending into central Africa. (Pre-
Miocene drainage reconstruction based on various sources.)

margins, careful examination of actual passive margin drain- challenge to reconcile this variability with tectonic models
age reveals considerable variation. While major river systems of passive margin development.
penetrate the marginal upwarps along the African passive
margin, on the subcontinent of India the drainage of the
steep western flank of the Western Ghats is completely se- Further reading
parated from river systems flowing eastwards. The drainage The topics discussed in this chapter are rather patchily cover-
divide between the two systems coincides with a major ed in the literature. Of the more general texts Ollier (1981)
escarpment and lies within 100 km of the west coast along discusses aspects of both active and passive (structural)
much of the margin. Since the east coast of India is also a tectonic controls on drainage and Morisawa (1985), (Chapter
passive margin the dramatic asymmetry of the drainage of 10) emphasizes the passive controls over drainage develop-
the subcontinent may be due to uplift having occurred ment. Structural controls are covered in detail in the
earlier in the west. books by Tricart (1974) and Twidale (1971), but these
The eastern margin of Australia provides a further example treatments do not include adequate discussions of recent
of passive margin drainage evolution. Here the drainage ideas on the development of transverse drainage in orogens
divide - separating eastward-flowing streams draining to or of the evolution of passive margin basins.
the Tasman Sea and westward-flowing streams draining Drainage patterns are examined in depth by Howard
towards the Australian interior - does not coincide with the (1967) who illustrates how they can be used in the inter-
generally well-developed great escarpment but lies at a pretation of geological structures. Joint control of drainage
variable distance to the west of it. Moreover, there is evi- patterns in homogeneous lithologies is discussed by Schick
dence to suggest that the continental drainage divide has (1965), and Scheidegger and Ai (1986) provide a useful
not shifted laterally to any significant extent since at least introduction to the relationship between valley orientations
the Late Oligocene - Early Miocene. Clearly, there is and the joint and fracture patterns generated by local neo-
considerable variation in the history of drainage evolution tectonic stress fields. Useful regional studies of these rela-
along passive margins, and it will be a significant research tionships are provided by Bevan and Hancock (1986) on
428 Endogenic~xogenic interactions

Fig.16.27 A speculative model of drainage development in passive margin basins. Note that this model assumes that uplift precedes
rifting and does not incorporate the effects of aulacogens striking into the margin. An initial drainage across a low relief landscape is
assumed (A). Initial epeirogenic warping disrupts the drainage system and causes drainage reversal on the 'upstream' (left) side of the
upwarp (B). Such drainage disruption could be achieved by quite slow rates of uplift where channel gradients are low and the erosive
power of rivers is minimal. Where stream power is greater (either due to higher channel gradients or higher (peak) discharges related
to climatic conditions in the catchment) rivers might initially be able to maintain their courses across the upwarp. The deveLopment of
significant down-faulting associated with rift development will, however, eventually bring about a complete disruption ofpre-rift
drainage (C). During this phase of rifting three drainage eLements are created: (1) low gradient networks draining the flanks of the rift
upwarp, most probabLy to newly formed internal basins; (2) steep gradient streams which cut deep canyons into the rift walls and
achieve rapid headward erosion towards the drainage divide formed along the line of maximum uplift on the rift shoulder; (3)
longitudinal drainage on the rift floor which redistributes sediments supplied from the rift walls. Continuing extension across the rift
andfurther subsidence of the riftfloor leads to the formation of a nascent ocean (D). The rift wall streams now form the new coastal
 Tectonics and drainage development 429

drainage which is initially confined to a narrow belt along the newly created rift margin. Thermal subsidence and sedimentary loading
will together induce progressive drowning of the margin but, perhaps more significantly, uplift will also be promoted inland by rotation
and flexure of the margin (D-F). The zone of maximum uplift along the margin is also likely to migrate inland as the lithosphere cools
and becomes more rigid. Uplift of the marginal upwarp will also be enhanced by denudational unloading as the steep gradient coastal
streams continue to erode inland following the migrating zone of maximum uplift. Isostatic uplift could result in a net increase in the
height of interfluves along the upwarp crest because erosion will probably be localized in canyons. Eventually the most aggressively
headward-eroding coastal streams will begin to capture the inland drainage by breaching the crest of the marginal up warp (G).
Capture would initially be confined to headwaters, but eventually entire internal fluvial systems would be incorporated into coastal
systems and a major knickpoint will be formed at the junction of the high gradient coastal channels and the much lower gradient trunk
channels of the captured internal drainage network (H).

southern England and northern France, and by Scheidegger controls and the long-term development of fluvial systems.
(1980) on Ontario, Canada. The development of struc- The effects of warping and faulting on drainage have
turally controlled drainage through time is discussed at been inadequately investigated, but the book by OIlier
length in the books by Tricart and Twidale mentioned above. (1981) (Chapter 12) provides some useful material on this
An excellent case study of transverse drainage is that by topic. Frostick and Reid (1987) consider drainage develop-
Mabbutt (1966) who provides a detailed analysis of the ment associated with rifts and specifically examine the
drainage history of central Australia and illustrates the com- impact of half-graben structures on drainage routeing. The
plex relationship that can exist between passive tectonic history of the Colorado River provides an interesting case

A c

B
\
J
/
I
I /
I I
J \

Fig. 16.28 Schematic model of the effects on drainage

development of structural and morphological variability along
passive margins. A-C represent stages of development. Note the
role of aulacogens as likely foci for external drainage early in the
life of the margin. The dashed lines schematically indicate
contours.
430 Endogenic-exogenic interactions

study of the effects of warping and faulting on drainage Frostick, L. E. and Reid, 1. (1987) Te'ctonic control of desert
development, and this example can be followed up in the ~ediments in rift basins ancient and modem. In: L. Frostick and
papers by Lucchitta (1972, 1979) and McKee and McKee t Reid (eds) Desert Sediments: Ancient and Modern. Geo-
logical Society Special Publication 35, 53-68.
(1972). Until the 1980s little research had been undertaken Gregory, D.1. and Schumm, S. A. (1987) The effect of active
on the adjustment of river channels to ongoing crustal tectonics on ~lluvial river morphology. In: Richards, K. (ed.)
deformation; now there are a number of both field and River Channels: Environment and Process. Institute of British
experimental studies on this topic. Gregory and Schumm Geographers Special Publication 18. Blackwell, Oxford and
(1987) and Schumm (1986) provide excellent introductions New York, 41-68.
Howard, A. D. (1967) Drainage analysis in geologic interpre-
to the existing literature, while Burnett and Schumm tation: a summation. American Association of Petroleum Geo-
(1983), Ouchi (1985) and Russ (1982) report specific field logists Bulletin 51, 2246-59.
investigations and experimental studies. Kale, V. S. and Rajaguru, S. N. (1988) Morphology and denuda-
To date the relationships between tectonic setting and tion chronology of the Coastal and Upland river basins of western
Deccan Trappean landscape (India): a collation. Zeitschrift fur
drainage systems at the macro- and megascale have been
Geomorphologie 32,311-27.
assessed largely in terms of the sedimentary characteristics Lucchitta, 1. (1972) Early history of the Colorado River in the
of the associated depositional basins. Miail (1981, 1984) Basin and Range Province. Geological Society of America
presents an excellent overview of this approach which con- Bulletin 83, 1933-48.
tains much of geomorphic interest; Audley-Charles et al. Lucchitta, 1. (1979) Late Cenozoic uplift of the southwestern
Colorado Plateau and adjacent lower Colorado River region.
(1977) and Potter (1978), on the other hand, give greater
Tectonophysics 61, 63-95.
emphasis to the fluvial systems themselves. The best analysis Mabbutt, 1. A. (1966) Landforms of the western Macdonnell
of drainage development in active orogens is that of Ranges: A study of inheritance and periodicity in the geomor-
Oberlander (1985) who also produced an earlier classic phology of arid central Australia. In: G. H. Dury (ed.) Essays in
study of the drainage of the Zagros Mountains (Oberlander, Geomorphology. Heinemann, London, 83-119.
McKee, E. D. and McKee, E. H. (1972) Pliocene uplift of the
1965). Aspects of the drainage history of Africa are in-
Grand Canyon Region - time of drainage adjustment. Geo-
cluded in the papers by De Swardt and Bennet (1974), logical Society ofAmerica Bulletin 83, 1923-32.
Dingle and Hendey (1984), Summerfield (1985) and Thomas Miall, A. D. (1981) Alluvial sedimentary basins: tectonic setting
and Shaw (1988). Kale and Rajaguru (1988) look at drain- and basin architecture. In: A. D. Miall (ed.) Sedimentation and
age development in western India while Bishop (1982, Tectonics in Alluvial Basins. Geological Association of Canada
1986) and Bishop et al. (1985) provide a detailed overview Special Paper 23, 1-33.
Miall, A. D. (1984) Principles of Sedimentary Basin Analysis.
of the drainage history of the south-eastern margin of Springer-Verlag, New York.
Australia. Morisawa, M. (1985) Rivers. Longman, London and New York.
Oberlander, T. M. (1965) The Zagros Streams. Syracuse Univer-
sity Geographical Series No. I.
Oberlander, T. M. (1985) Origin of drainage transverse to struc-
References .tures in orogens. In: M. Morisawa and J. T. Hack (eds) Tectonic
Audley-Charles, M. G., Curray, J. R. and Evans, G. (1977) Loca- Geomorphology. Allen and Unwin, Boston and London,
tion of major deltas. Geology 5, 341-4. 155-82.
Bevan, T. G. and Hancock, P. L. (1986) A late Cenozoic regional Oilier, C. (1981) Tectonics and Landforms. Longman, London
mesofracture system in southern England and northern France. and New York.
Journal of the Geological Society, London 143, 355-62. Ouchi, S. (1985) Response of alluvial rivers to slow active tec-
Bishop, P. (1982) Stability or change: a 'review of ideas on ancient tonic movement. Geological Society of America Bulletin 96,
drainage in eastern New South Wales. Australian Geographer 504-15.
15,219-30. Potter, P. E. (1978) Significance and origin of big rivers. Journal
Bishop, P. (1986) Horizontal stability of the Australian continen- ofGeology 86, 13-33.
tal drainage divide in south central New South Wales during the Russ, D. P. (1982) Style and significance of surface deformation
Cainozoic. Australian Journal ofEarth Sciences 33, 295-307. in the vicinity of New Madrid, Missouri. United States Geo-
Bishop, P., Young, R. W. and McDougall, 1. (1985) Stream profile logical Survey Professional Paper 1236, 45-114.
change and longterm landscape evolution - Early Miocene and Scheidegger, A. E. (1980) The orientation of valley trends in
modem rivers of the east Australian highland crest, New South Ontario. Zeitschrift fur Geomorphologie 24, 19-30.
Wales, Australia. Journal of Geology 93, 455-74. Scheidegger, A. E. and Ai, N. S. (1986) Tectonic processes and
Burnett, A. W. and Schumm, S. A. (1983) Active tectonics and river geomorphological design. Tectonophysics 126, 285-300.
response in Louisiana and Mississippi. Science 222, 49-50. Schick, A. P. (1965) The effects of lineative factors on stream
De Swardt, A. M. J. and Bennet, G. (1974) Structural and physio- courses in homogenous bedrock. Hydrological Sciences Bulletin
graphic development of Natal since the late Jurassic. Transac- 10,5-11.
tions of the Geological Society of South Africa 77, 309-22. Schumm, S. A. (1986) Alluvial response to active tectonics. In:
Dingle, R. V. and Hendey, Q. B. (1984) Late Mesozoic and Tertiary Active Tectonics. National Academy Press, Washington DC,
sediment supply to the eastern Cape Basin (SE Atlantic) and 80-94.
palaeo-drainage systems in southwestern Africa. Marine Geo- Summerfield, M. A. (1985) Plate tectonics and landscape develop-
logy 56, 13-26. ment on the African continent. In: M. Morisawa and 1. T. Hack
 Tectonics and drainage development 431

(eds) Tectonic Geomorphology. Allen and Unwin, Boston, and Tricart, J. (1974) Structural Geomorphology, (translated S. H.
London, 27-51. Beaver and E. Derbyshire). Longman, London and New York.
Thomas, D. S. G. and Shaw, P. A. (1988) Late Cainozoic drainage Twidale, C.R. (1971) Structural Landforms. MIT Press, Cambridge
evolution in the Zambezi Basin: geomorphological evidence and London.
from the Kalahari rim. Journal of African Earth Sciences 7,
611-18.
 17
Sea-Level Change

17.1 Global and regional sea-level change 1. Sea level rising while the landsurface subsides, remains
stationary or rises at a slower rate;
Sea-level change is a central concern of geomorphologists.
2. Sea level remaining stationary while the landsurface
Not only does the sea surface determine the base level for
subsides; or
erosion, but relative vertical movements of land and sea can
Sea level falling while the land surface subsides at a
greatly alter the area of land exposed to exogenic sub-aerial 3.
more rapid rate.
geomorphic processes. Indeed, over the past 100 Ma the
relative proportion of land and sea has changed dramati- A relative fall of sea level is an apparent fall in the mean
cally. The study of sea-level change is also important because, level of the sea surface with respect to a landsurface and
as we have seen in previous chapters, it can provide key can result from:
evidence of climatic change and can also give us a bench-
1. Sea level falling while the landsurface rises, remains
mark for estimating rates of tectonic uplift.
stationary or subsides at a slower rate;
Satellite technology now enables us to measure the form
2. Sea level remaining stationary while the landsurface
of the sea surface (the ocean geoid - see Section 2.1.1) to a
rises;
precision of a few millimetres. These accurate altimetric
Sea level rising while the landsurface rises at a more
measurements can be used to monitor changes in sea sur- 3.
rapid rate.
face elevation due to tides and currents which occur over
short periods (hours, days or months). Sea-level change in a Relative sea-level changes can occur on a regional or global
geological sense, on the other hand, occurs on a much scale, and one of the primary tasks of research into sea-level
longer time scale of thousands and millions of years. These fluctuations is to separate the regional components from
changes in sea level through time can only be determined in global changes. Global sea-level changes result from mech-
terms of movements of the sea surface with respect to ad- anisms which affect the level of the sea surface world-wide.
jacent land areas. Consequently, it is shorelines, marking Such global changes in sea level are often described as
the interface between land and ocean, which are the focus eustatic. By contrast, regional changes are (with one import-
of sea-level change studies. In many cases sea-level changes, ant exception - see Section 17.5.3) the result of uplift or
which of course involve vertical movements of the sea sur- subsidence of the land. The importance of eustatic sea-level
face, have to be inferred, often indirectly, from horizontal change was first highlighted in the late nineteenth century
shoreline movements. Transgressions occur when the shore- by the Austrian geologist Eduard Suess who remarked on
line advances landward, whereas regressions represent the the synchronous episodes of deposition of marine sedi-
retreat of shorelines from the land. ments around the world evident in the geological record.
A relative change in sea level is a rise or fall in the mean He interpreted these as being a result of world-wide changes
level of the sea surface with respect to the landsurface (or in sea level, inferring that outside the world's orogenic
the sea bed). Either the sea surface, or the landsurface, or belts the continents had exhibited long-term stability. This
the two in combination, may rise or fall during a relative eustatic model soon attracted criticism from those who saw
sea-level change. A relative rise of sea level is an apparent abundant evidence for widespread local and regional defor-
rise in the mean level of the sea surface with respect to a mation of the continents, but it served to emphasize the
landsurface and can result from: distinction between global and regional causes of relative
sea-level change.
434 Endogenic-exogenic interactions

17.2 Evidence for sea-level change morphological features. There is abundant evidence of
Quaternary sea-level change along the well-studied coast-
In order to unravel the complex history of sea-level changes
lines in the mid-latitudes of the northern hemisphere but
we have to make use of a wide range of techniques; these
much of this region has been dominated by local tectonic
involve both fixing the location of a past shoreline and
effects as land areas have subsided and risen isostatically in
dating it as accurately as possible. Different methods are
response to the growth and decay of ice sheets.
appropriate to various time scales and, clearly, as we go
Shorelines remote from the direct effects of glacial isos-
further back in time the task of reconstructing the record of
tasy are, however, to be found in the tropics and over much
sea-level change, either locally or globally, becomes in-
of the southern hemisphere and if they occur in regions of
creasingly difficult
active tectonic uplift they may provide a detailed record of
global sea-level fluctuations. One notable example of such
17.2.1 Quaternary sea level
a tectonically active coast is the Huon Peninsula of Papua
It is only over the past 100 a or so that sea-level changes New Guinea (see Section 15.2.1 and Figure 15.5). Here
have been determined from accurate measurements of the there is the ideal combination of an actively rising shoreline
height of the sea s~rface at tide gauge stations. Averaged (rates of 700-3300 mm ka- I) and datable deposits (coral
across the world these data indicate a current sea-level rise reef terraces) which range in age up to 120 ~a old. As uplift
of around 1 mm a-I. In contrast to the direct information has occurred along the coast each high sea-level stand has
from tide gauge records, sea-level changes over longer cut a new terrace on the shoreline. Since the age of each
periods of time must be based on indirect information. terrace can be determined radiometrically, the difference in
Relative falls in sea level are demonstrated by elevated height between present global sea level and these earlier
shoreline features such as raised beaches which provide eustatic highstands can be estimated if it is assumed that the
deposits which may yield material suitable for dating by rate of uplift has been constant at any given point along the
radiometric techniques. Vertical sequences of erosional shoreline (Box 17.1).
notches can often provide sufficient information to provide
a relative chronology of sea-level fall, even if absolute dating
17.2.2 Pre-Quaternary sea levels
is not possible. A relative rise in sea level is indicated by
submerged coastal topography and by drowned river valleys As we go back in time beyond the beginning of the Quater-
or glacial troughs which, in some instances, can be traced nary, morphological evidence becomes progressively more
seawards as submarine canyons and channel networks. Dating difficult to relate to particular changes in sea level, since
of such submarine features, however, must be based on landforms formed at such times have usually been severely
associated deposits; these may include submerged weather- modified by erosion; consequently, we have to rely largely
ing profiles which may be datable if they contain organic on the record preserved by marine sediments.
material. In regions affected by Quaternary glaciation it
may be possible to identify glacial sediments which, al- 17.2.2.1 Continental evidence
though now below sea level, can be shown to have been It is often possible to estimate approximately the depth at
deposited sub-aerially. which a body of marine sediment was deposited by deter-
It has long been appreciated that the fluctuations in ice- mining its facies - that is, those lithological and organic
sheet volume during the Quaternary have been the primary characteristics imparted by its environment of deposition.
control of global sea level for at least the past 2 Ma or so. Where marine sediments· are now exposed sub-aerially the
As we saw in Chapter 14, changes in global sea level during change in relative sea level since deposition can be esti-
the Quaternary have been reflected in the oxygen isotope mated by utilizing fossil and sedimentological evidence.
record preserved in the calcareous shells of microfossils Ma~ine sediment~. containing rapidly evolving faunal as-
which have accumulated in the sediments deposited on the semblages can be dated to a resolution of around 1 Ma or
ocean floor. The recovery of cores drilled at various loca- less, so a fairly detailed record of sea-level change can be
tions across the world's oceans have thus made it possible reconstructed by this means. As such, this technique only
to monitor global sea-level fluctuations over the last several provides information on regional sea-level changes, but
hundred thousand years. The oxygen isotope variations pro- where marine sediments overlie a cratonic terrane which is
vide a fi!st order approximation of global ice volumes and thought to have been tectonically stable, then any sea-level
hence global sea-level changes (a 0.01 per cent increase in changes inferred would be of global extent. Our confidence
the 18 0 concentration indicating an approximately 10 m in world-wide sea-level changes is greatly strengthened if it
decrease in global sea level). is found that sea-level changes identified in different con-
The oxygen isotope record can be more accurately cali- tinents are found to have been synchronous.
brated by correlating it with detailed local studies of sea- An alternative approach to inferring the depth of depo-
level change based on accurately dated sedimentological or sition from the facies of marine sediments is simply to plot
 Sea-level change 435

Box 17.1 Estimating global sea-level change from Quaternary marine terrace sequences

Estimating global sea-level change from the record provided Vila

by the marine terrace sequence of the Huon Peninsula first 200
requires that we separate the effect of surface uplift along
the coast. This is done by using the 120 ka terrace formed I 160
during the highstand of the last inter-glacial as a reference
~
120
point (labelled terrace VII on Figure B17.1 (A». The height
of this terrace is determined at various points along the
~
80
coast and the difference between this height and global sea
level at 120 ka BP gives the total amount of uplift at each t:Q) 40 0 200
point during this period. It is known from studies of stable VI ~
shorelines around Australia that sea level at 120 ka BP was £ 0

between 2 and 8 m higher than at present. We can therefore ~

tj 200
'I VII b
calculate the mean rate of uplift for each point along the B
coast (U) from: ~
~'- 160
(HT-H s) .r:. VI
U= ------'-
120 ~ 120
V
c:
where H T is the elevation (m) of the 120 ka terrace above 0
.~
80 IV
present sea level and Hs the height (m) of the 120 ka BP Q)

global sea level above the present. W 40 lila ......

..... lIlb
This gives the mean uplift rate in m ka- 1 which varies 0 ..... 11
from 3.3 to 0.7 along the coast (see Figure 15.5). If we 120 100 80 60 40 20 0
assume that the rate of uplift has been constant over the ka B.P.
past 120 ka, then we can calculate the fluctuations in global
sea level by relating the actual elevation of each dated ter-
race to the mean uplift rate of the terrace sequence (Fig. Fig. BI7.I Typical cross-section of raised coral reef
B17.1(B». Thus the-relative sea level (S) indicated by each terrace on the Huon Peninsula, Papua New Guinea, with
terrace of known age and present elevation is given by the same reef codes as used in Figure 15.5(A); illustration
of the uniform uplift assumption used as the basis for the
S'= h - Ut
estimation of the pattern of sea-level change (B). (After 1.
where h is the elevation above present sea level (m), t the Chappell (1987) in: M. 1. Tooley and I. Shennan (eds) Sea-
age (ka) of the terrace, and U the mean uplift rate (m ka- I ) Level Changes. Institute ofBritish Geographers Special
at that point. Publication 20, Blackwell, Oxford, Fig. IO.II ,p. 323.)

17.2.2.2 Seismic stratigraphy

the area of the continents covered by marine strata for par- Marine sediments deposited along continental margins con-
ticular periods of geological time-in effect tracing changes tain a wealth of evidence which could help us determine the
in shorelines through time. If sea level rises relative to an record of long-term sea-level changes, but how do we gain
adjacent land area there will be a transgression as the sea access to this information locked up in offshore sedimen-
encroaches on to the land. Conversely, if sea level falls tary sequences? One approach is to drill boreholes. These
there will be a regression as the sea withdraws across the give us a very good idea of the changes in age and charac-
continental margin. Again, if such regressions and trans- teristics of sediments with depth, but the information they
gr.essions are synchronous across several continents. we provide is specific to the particular drill site and the cost of
might feel confident in claiming that sea-level change had drilling offshore is very high. Although boreholes have
been of global extent. been used extensively by oil companies in their search for
One problem with this approach is that the hypsometry oil along the world's continental margins, it is the applica-
of the continents may change through time, and such changes tion of seismic techniques that has revolutionized our
in hypsometry can alter the areal extent of a regression or knowledge of offshore stratigraphy by providing indirect
transgression arising from a given change in sea level. A information on submarine sedimentary sequences.
second problem is that transgressions and regressions are If sound waves are transmitted downwards from a ship
not determined solely by the relative vertical movements of and the time taken for them to bounce off the sea bed and
land and sea. Erosion can cause a coastline to recede and return is measured it is possible to use the variations in
thus promote a transgression, while deposition of sediment, 'echo' time to map the morphology of the sea floor. This is
especially at the deltas of major rivers, can cause a re- the well-known technique of echo-sounding. Using more
gression. Such effects must be taken into account in assess- powerful but lower frequency seismic waves, which can
ing evidence for relative sea-ievel change and are discussed partially penetrate solid rock, it is possible to locate the
further in Section 17.5.4. boundaries, or discontinuities, between different sedimen-
436 Endogenic-exogenic interactions

1-

II~il~III-I,,:~%~~ COASTAL DEPOSITS

~ MARINE DEPOSITS
- - SEQUENCE BOUNDARIES
- REFLECTION TERMINATKINS
o 15 30 MILES I
I I
I I
o 25 50 KM
Fig.I7.I Offshore seismic sectionfrom the continental margin of north-west Africa. The section shows strata ofprobable Jurassic
age laid down at an early stage of continental rifting overlain by thick marine and coastal sequences of Cretaceous sediments. The
vertical scale is expressed in terms of the transmission time (in seconds)for the seismic waves to travel through the strata. The
equivalent total depth of the section is around 12 km. (From P. R. Vail et al. (1977) in C. E. Payton (ed.) Seismic Stratigraphy -
Applications to Hydrocarbon Exploration. American Association of Petroleum Geologists Memoir 26 Fig. 5, p. 69.)

tary units from differences in the velocity with which the- reflected in the lateral and vertical displacement of sedi-
seismic waves travel through them. As a ship tracks across mentary units (continental, littoral and marine) across the
the ocean, seismic profiles are generated which, after com- continental shelf and continental slope (Fig. 17.2). Since, as
puter processing, enable the underlying sedimentary struc- we have already noted, most seismic reflectors separating
tures and -sequences to be identified, a technique known as these different sedimentary units represent ancient sea
seismic stratigraphy (Fig. 17.1; see Section 4.3.1). floors, the highest point on each reflector indicates the sea
The echo times from these discontinuities between sedi-
mentary units, termed seismic reflectors, can be converted A
to a depth if we know (or can reasonably estimate) the vel- E!1i3HIGHSTAND DEPOSITS
ocity of seismic waves through the rock. Most seismic re- ~LOWSTAND DEPOSITS

flectors recorded in sedimentary sequences appear to be ~ OLDER ROCKS

surfaces which, prior to being buried, formed the sea bed or
an erosion surface on land. Such surfaces are therefore un- !4~~~~~iIIICOASTAL ONLAP

conformities, representing depositional breaks caused by

MARINE 200 KM B
erosion or lack of deposition. Since most unconformities OFFLAP

can be assumed to represent surfaces of uniform age they

can be regarded as chronostratigraphic markers. The ~SUBAERIAL
~EXPOSURE
actual age of surfaces represented by seismic reflectors can UNCONFORMITIES
~ SUBMARINE
only be determined by correlating them with borehole se- -
- - SUBAERIAL
quences which have been dated.
'How, then, can seismic stratigraphy contribute to our
knowledge of sea-level change? The most active researchers
Fig.I7.2 Typical depositional patterns (highly simplified)
involved in the interpretation of data from seismic stratigra- observed on seismic sections ofpassive continental margins: (A)
phy in terms of sea-level change have been based in the oil the prograding sedimentary sequences associated with high
industry, and foremost among these has been the group at relative sea level; (B) the onlap of marine deposits on to earlier
the Exxon Corporation originally led by P. R. Vail. This sedimentary sequences during a period of low sea level when
much sediment is carried beyond the edge of the continental shelf
group has argued that processes of sub-aerial erosion and
(From P. R. Vail et al. (1977) in C. E. Payton (ed.) Seismic
submarine deposition operate relative to a base level which Stratigraphy - Applications to Hydrocarbon Exploration.
is determined by the position o(the sea surface. A relative American Association of Petroleum Geologists Memoir 26
change in sea level leads to a change in base level which is Fig. 9, p. 73.)
 Sea-level change 437

level at the time it was formed. By correlating records of

relative sea-level change recorded on different continental
17.3 The record of Quaternary sea-level change
margins, the Exxon group have constructed what they The general trend of Holocene sea-level change is well
believe to· be a record of world-wide sea-level change established and is clearly related to melting of the northern
throughout the Cenozoic and Mesozoic (see Section 17.4). hemisphere ice sheets which began about 15000 a BP. This
The detailed application of seismic stratigraphy is far led to an initially rapid rise in global sea level at rates of up
more complex than has been indicated here because several to 12000 mm ka -1, which declined significantly from about
factors - including variations in rates of sediment supply 8000 a BP. The details of Holocene sea-level change have,
and removal and rates of relative sea-level change - can however, been more difficult to resolve, especially for the
greatly complicate the relationship between the position of past 8000 a. It has been variously argued that during this
seismic reflectors and sea-level change. It also needs to be period global sea level has risen continuously but slowly,
emphasized that many researchers have been extremely that it rose up to about 3600 a BP. but has subsequently
sceptical about certain aspects of the Exxon global sea-level been more or less stable, and, finally, that it rose to about
curve. Apart from alternative interpretations of the data, the 2 m above present sea level around 6000 a BP before declin-
curve has some remarkable features which challenge the ing. A significant increase in data on sea-level changes in
conventional wisdom about the mechanisms of sea-level the past few years, particularly in the tropics, has made it
change and some of the original data from which it has clear that there have been real differences in the Holocene
been constructed have been withheld for commercial reasons. 'eustatic' sea level record in different parts of the world and
None the less, the Exxon curve has been widely adopted as that these discrepancies cannot be entirely attributed to
the best available representation of the record of global sea- errors in dating or local tectonic effects. Moreover, as we
level change, especially since some of the criticisms of its shall see in Section 17.53, we now have an elegant ex-
original presentation have been met by the publication in planation of how these variations are produced.
1987 of a revised version in which the interpretation from For the Quaternary as a whole the deep-sea ocean core
seismic stratigraphy has been integrated with the record of record now provides the framework for studies of global
sea-level change determined from marine sediments now sea-level change (Fig. 17.3; see Section 14.3.2). Major deg-
exposed on land. laciation events associated with sea-level highstands corre-

o
Brunhes - Matuyama
~
c
o
P,'"~''''''' BOOOd"'\
.~ 1
C
Q)
u
c
o
uo
o
~
.S
~1
c
'"
L:
U
.~ 2 1

'"
Qi
a:
o 100 200
7

300 400
-
~xygen ~sotope sta ges correlate~ with sea \~Vel highs 13

-- - - 500 600 700

ka B.P.

South Taranaki,
New Zealand
V>
'C
IIII II I I I
Atauro,
Indonesia ~
Cl
I III I I I I I I
Huon Peninsula,
Papua New Guinea
il
.,
co
IIIII III I III I I I I I I
I II I I I I I II II III
South Australia Ul

0 1 0 200 3 0 400 500 6 0 760

ka B.P.

Fig.17.3 A composite record of variations in 180 concentrations from ocean floor sediments for the past 700 ka compared with the
chronology of highstands at various locations unaffected by glacio-isostatic rebound. (Based on 1. Imbrie et al. (1984) in: A. Berger et
al. (eds) Milankovitch and Climate. Reidel, Dordrecht, Vol. 1.. and B. Pillans (1987) in: R. 1. N. Devoy (ed.) Sea Surface Studies: A
Global View. Croom Helm, London, Fig. 9.10, p. 281.)
438 Endogenic~xogenic interactions

late with oxygen isotope stages 1,5, 7,9, 11, 13, 15, 17 and more variable. Much of this discrepancy is probably due to
19, while stage 3 represents a partial deglaciation ac- errors in dating (especially in the ocean floor record). Dating
companied by moderately high sea levels. The Brunhes- errors 'for shoreline features can be as great as 50 ka at
-Matuyama palaeomagnetic boundary, which is identifiable 370 ka BP, so correlations are fraught with difficulty once
in these cores, has been accurately located by K-Ar dating we go back beyond the past 200 ka or so.
at 730 ka B P and provides an important reference point in
the mid-Pleistocene. Since the major deglaciation events of
the Pleistocene have a broadly similar isotopic signal, we
can infer that the changes in ocean water volumes asso-
17.4 The record of pre-Quaternary sea-level
ciated with earlier glacial-interglacial cycles were similar
to the most recent transition from a glacial to interglacial
change
climatic regime. Consequently, it is likely that the magnitude Most discussion of pre-Quaternary sea-level change is cur-
of sea-level change was also similar. rently focused on the global record presented by the Exxon
The correlation between the ocean-core record of global group. Figure 17.4 illus~rates the most recent version of
sea-level change and evidence of sea-level change from well- their global sea-level curve for the Cretaceous to the present
qated coral terrace sequences' in the tropics is fairly good and shows fluctuations at a variety of scales. The long-term
back to about 400 ka BP. In addition to the Huon Peninsula rise evident for the Early Cretaceous was preceded by a
sequence, Barbados in the Caribbean provides an excellent prolonged period of generally low sea levels extending from
location for'recontructing Late PleIstocene global sea-level the Late Palaeozoic some 320 Ma BP until the Late Jurassic
change. It 'is tectonically active and has been rising rapidly about 150 Ma BP. Since the mid-Cretaceous there has been a
at up to 500 mm ka- 1 over the past several hundred thou- fairly persistent trend of falling sea level, although esti-
sand years. Fifteen separate coral terraces have been dated mates of the amount of this fall vary. The original analysis
up to an elevation of 200 m above present sea level with a by the Exxon group used an estimate of a Late Cretaceous
maximum age of 650 ka BP. The terrace sequence appears maximum some 350 m above present sea level, but their
to indicate six or seven complex interglacial sea-level high- revised curve uses a more modest estimate of 250 m. Perhaps
stands which correlate fairly well with oxygen isotope stages the most remarkable features of the Exxon curve are the
5-17. Moreover, the 120 ka terrace from the Huon Peninsula dramatic sea-level falls evident throughout the Mesozoic
has an equivalent of the same age in Barbados. Beyond about and Cenozoic, most notably the very rapid Late Oligocene
400 ka BP. the correspondence between the ocean floor record fall of around 150 m.
and sea-level records from coastlines around the world is Most researchers concerned with global sea-level change

EARLY CRETACEOUS
+250
] 200
150
ai 100
>
.3 + 50
0
t1l
Q)
(/)
130 120 110 100 90 80 70 65
t1l
.0 Ma B.P.
.2
<.9
c:
.- +250
Q) 200
OJ
c: 150
t1l +100

<5 50
0
- 50
-100
65 60 50 40 30 20 10 0
Ma B.P

Fig. 17.4 Global sea-level curves from the beginning of the Cretaceous produced by the Exxon group largely on the basis of seismic
stratigraphy. The long-term trend is shown by a dashed line with shorter-term fluctuations being indicated by a solid line. (Modified
from B. U. Haq et al. (1987) Science 235, Figs 2 and 3, pp. 1159 and 1161.)
 Sea-level change 439

have had little difficulty in accepting the general validity of ice sheets (ice shelves, pack ice and icebergs do not count
the long-term trends depicted on the Exxon curve because as they qisplace their own mass of sea water). We have
they accord with other evidence of high global sea level in already referred to this mechanism, known as glacio-eustasy,
the Late Cretaceous (primarily the wide extent of marine in relation to Quaternary sea-level fluctuations. Using the
strata of this age which outcrop across North America and total area of the oceans (3.6 x 108 km 2), the density of sea
Europe). It is the short-term fluctuations that have attracted water (l030 kg m-3) and the density of glacier ice
criticism, it being argued, for instance, that the rate and (920 kg m-3), changes in global sea-level can be calculated
magnitude of sea-level changes exhibited by the extremely if the volume of terrestrial ice that melts or accumulates is
rapid falls in the Exxon curve can only be explained by known. The volumes of present-day ice sheets are difficult
extensive glaciations. The major problem here is that there to estimate accurately, and of course estimates for the
is little evidence for glacial episodes over the past 200 Ma extensive ice bodies of the Late Cenozoic are even less
until the mid-Cenozoic (but see Section 14.3.1). It has also certain. Nevertheless, rough calculations can be made. A
been suggested that some of the short-term fluctuations rise of 60-75 m would be produced by the melting of the
may' not be of worldwide extent, but may simply be an present Antarctic ice sheet, and a further few metres would
artefact of the bias arising from the large proportion 'of be added by the dissipation of the Greenland ice sheet,
seismic data collected from passive margins with a similar giving a total rise of bewteen 65 and 80 m. Melting of other
history of subsidence. These criticisms are crucial to an land-based ice bodies would contribute only a trivial
understanding of long-term global sea-level change, and both additional rise in sea level. This does not give the ultimate
can only be assessed with reference to possible mechanisms change in global sea level, however, because we have to
of sea-level change. take into account the effects of isostasy. The additional
water added to the oceans will depress the crust of the
ocean floor and the newly flooded continental margins, a
17.5 Mechanisms of global sea-level change
process known as hydro-isostasy.This would reduce the
Our knowledge of global sea-level fluctuations over the past overall rise in sea level to between 40 and 50 m for the
100 Ma or so indicates that they have involved total move- melting of present-day ice sheets.
ments in excess of 200 m and have occurred over a wide Various attempts 'have been made to calculate the equi-
range of time scales. What, then, causes these global sea-level valent changes for the growth and decay of the Pleistocene
changes? There are several possible mechanisms, but they ice sheets. The limits of Pleistocene glaciations are fairly
can be grouped into three basic types - those that alter the well known, and the form of ice sheets, and hence their
volume of water in the world's oceans, those that change approximate volumes, may be inferred from these limits. It
the volume of the ocean basins and those that produce is not certain, however, whether the major ice sheets reached
changes in the geoid (Fig. 17.5). their maximum size simultaneously. If we assume, for
simplicity, that they did, then the maximum sea-level fall
for the most extensive well-documented glacial (correlated
with oxygen isotope stage 6 - about 120-190ka BP) was
CLIMATE
probably of the order of 150 m, or about 100 m if an allow-
ance is made for isostatic compensation. The rather less
extensive last glacial probably produced a more limited
maximum fall of around 115 m, or 75 m assuming isostatic
Isostasy
compensation. These figures, it must be emphasized, are
t Isostasy t only estimates and some of the evidence suggests some-
what more extensive ice sheets and hence rather larger sea-
GRAVITY
level falls of perhaps as much as 175 m. None the less, an
estimate of Quaternary glacio-eustatic changes of a maximum
TECTONICS of 150 m (isostatically uncompensated) is probably not far
from the mark.
Fig.17.5 Schematic representation of the major types of
controls of global sea-level change: (A) and (B) represent two The rates of Quaternary sea-level change are in accor-
sea levels which are associated with different amounts of change dance with likely rates of ice sheet growth and decay. The
in different locations due to geoidal effects (see Section 17.5.3). rapid sea-level rise between 15000 and 6000 a BP attained a
maximum rate of 12000 mm ka- I and would have required a
sustained melting of the predominantly northern hemi-
17.5.1 Changes in ocean water volume
sphere ice sheets at a rate of around 5000 km 3 a-I. Ice sheet
By far the most significant factor influencing the amount of growth at a similar rate is indicated by the rapid sea-level
water in the oceans is the waxing and waning of land-based fall prior to 18000 a BP. These rates of ice sheet growth and
440 Endogenic-exogenic interactions

decay may seem unrealistic until we remember that some ridge system world-wide as a result of the periodic reor-
1100 kIn 3 a-I of water equivalent is presently added to the ganization of plate boundaries, and from changes in the rate
Antarctic ice sheet. This annual accumulation is currently of sea-floor spreading. Of these two mechanisms it is pro-
more or less balanced by ablation but the potential for rapid bably changes in sea-floor spreading rates that have been
volume changes with fluctuations in accumulation and most significant, at least since the Late Cretaceous.
ablation rates is clear. Consequently, we have no difficulty . The effect of fluctuations in sea-floor spreading rates can
in accounting for very rapid global sea-level changes of a be estimated on the basis of the age-depth relationship
magnitude of up to 150 m during periods when the Earth is exhibited by oceanic lithosphere (Fig. 17.6). After it is
experiencing major glacial-interglacial cycles, but the form~d new oceanic lithosphere cools, thickens, increases
problem remains of explaining the very rapid falls evident in density and gradually subsides. Where spreading rates
in the Exxon sea-level curve during the Mesozoic and Early are slow, cool lithosphere and therefore deep ocean occurs
Cenozoic. close to the ridge axis (Fig. 17.7); conversely, where spread-
Other factors influencing ocean-water volume involve ing rates are high, the lithosphere is still relatively warm
changes of only 10m or less. It is well known that the tem- and buoyant some distance from the ridge axis and ocean
perature of the oceans has changed frequently throughout depths here are consequently fairly shallow. If the spread-
the Phanerozoic, and such changes are fairly well docu- ing rate changes there will be a consequential change in
mented for the Quaternary. Through its influence on the ridge volume with an increase in spreading rate producing
density of ocean water, and hence its volume, temperature an increase in ridge volume and a rise in global sea level,
fluctuations can be reflected in changes in sea level. Such and a decrease in spreading rate giving rise to a decrease in
changes, however, will be comparatively small; even an ridge volume and a sea-level fall. It is important to note,
increase of 10°C throughout the entire depth of the oceans however, that the rate at which sea level can change as a
would only produce a 10 m rise in global sea level. result of variations in spreading rate is limited by how
quickly oceanic lithosphere cools.
Evidence for this mechanism is provided by the corre-
17.5.2 Changes in ocean basin volume lation between periods of accelerated spreading rates, as
Changes in the volume of the ocean basins were almost cer- indicated by the age of the ocean crust astr~de mid-oceanic
tainly the most significant factor in influencing the broad ridges, and global sea-level rise. In particular, the major
trend of global sea level throughout the Mesozoic and Early Late Cretaceous sea-level high coincides with a well-docu-
Cenozoic. Such changes can arise from a variety of causes, mented phase of rapid sea-floor spreading. Although there
including orogenesis, the isolation and desiccation of small are uncertainties as to the exact magnitude and timing of
ocean basins, sedimentation and variations through time in changes in spreading rates, quantitative estimates can be
the volume of the world-wide mid-oceanic ridge system. made of the effect on global sea level. Initial calculations
Estimates of the likely effects on global sea level of other
mechanisms capable of altering the volume of the ocean
2·5
basins indicate that these are probably of less significance.
Extensive upwarps of the ocean floor related to concen-
3·0
trations of hot spots have probably had some effect on E
~
global sea level as they develop and subsequently decay. 3·5
.l:
One suggestion is that massive mid-plate volcanism in the 0-
Q)
western Pacific between about 110 and 70 Ma BP may have 0
40
led to a global sea-level rise of between 40 and 100 m. c:
III
Q)
Changes in the hypsometry of the continents as a conse- ()
0 4·5
quence of the break-up of Pangaea may also have caused
major changes in global sea level with a possible rise of 5·0
130 m between about 140 and 50 Ma BP and a small fall of
10m since then. 5·5

o 10 20 30 40 50 60 70 80
17.5.2.1 Variations in the volume of mid-oceanic ridges
Age of Lithosphere (Ma)
As the volume of the world mid-oceanic ridge system is
equivalent to about 12 per cent of the total volume of ocean Fig. 17.6 Age-depth curve for oceanic lithosphere. The process
water, changes in the overall ridge system volume through of subsidence with increasing 'lithospheric age and density is
relatively simple, and for the first 70 Ma or so the increase in
time are clearly a potentially important mechanism of global depth is proportional to the square root of the age of the
sea-level change. Variations in mid-oceanic ridge volume lithosphere. (After W. C. Pitman, (1978) Geological Society of
can arise both from alterations in the total length of the America Bulletin 89, Fig. 1, p. 1390.)
 Sea-level change 441

::~I
Time
Elapsed

OMa
+300

20Ma
:H /\ I :gt ~ I
]:
~Q)
+200

A
...J

40 Ma
:~~ I :~~ A <0
Q)
(/)

CO
.c
G
0

::~~I :H A
.!:
Q)
Cl
+100
60 Ma c:
<0
~
U

70 Ma
~~ :H 2000 0 2000 2000
Distance From Ridge Axis (km)
/\ 0
!

2000
o

80 60 40 20 o
Ma (BP)

Fig. 17.8 Change in global sea level since 80 Ma BP due to

changes in mid-oceanic ridge volume and corrected for post-
Miocene ice-volume change. The thicker line shows the best-
estimate curve and the shaded area indicates the probable range
of error in the estimates. (Modifiedfrom M. A. Kominz (1984) in:
1. S. Schlee (ed.) Interregional Unconformities and Hydrocarbon
Fig.17.7 Relationship between ridge volume and sea-floor Accumulation. American Association of Petroleum Geologists
spreading rate . (A) The cross-section of a mid-oceanic ridge that Mem'oir 36, Fig. 7, p. 121.)
has been spreading at 20 mm a-I for 70 Ma. At time 0 Ma the
spreading rate is increased to 60 mm a-I. The sequence ofprofiles American craton (which appears to have been essentially
shows the increasing cross-sectional area at 20,40,60 and 70
Ma after the increase in spreading rate. By 70 Ma the ridge has
stable since at least the mid-Cretaceous).
reached its new equilibrium profile which has three times the At certain times in the past, variations in the length of the
cross-sectional area of the ridge 70 Ma earlier. (B) The converse world mid-oceanic ridge system may have been more im-
situation with a ridge that has been spreading at 60 mm a-I for 70 portant than changes in spreading rates in altering the vol-
Ma. The spreading rate is then decreased at time 0 Ma to 20 mm ume of the ocean basins. Some of the rise through the Early
a-I. The sequence ofprofiles shows the decreasing cross-sectional
area at 20, 40, 60 and 70 Ma after the decrease i~ spreading rate. Cretaceous may have been related to the extension of the
By 70 Ma the new equilibrium profile of the ridge has attained a world ridge system arising from the break-up of Gond-
cross-sectional area only one third that of the ridge at time 0 Ma. wana. The modest sea-level rise in the Eocene, which
(C) Cross-sections of ridges which have spread at 20 mm a-I and occurred during a period of reduced rates of sea-floor
60 mm a-I superimposed. (Adaptedfrom W. C. Pitman (1978) spreading, may similarly have been a result of the develop-
Geological Society of America Bulletin 89, Fig. 2, p. 1391.)
ment of ridge systems in the North Atlantic.

by W. C. Pitman indicated a fall of global sea level from 17.5.2.2 Sediment accumulation on the ocean floor
about 350 to 50 m above present sea level from 85 to 15 Ma Sediment supplied to the oceans from the continents would,
BP and these data were used to calibrate the initial Exxon on its own, decrease ocean volume and therefore raise sea
curve. Subsequently, more detailed calculations have high- level. With current rates of sediment supply this would pro-
lighted the large errors inherent in estimating the effects of duce a rise of between 15 and 40 mm ka- 1, but we must take
changing ridge volume and have indicated a rather smaller into account two compensating factors. These are the iso-
best-estimate for global sea-level fall since the Late Cretaceous static depression of the ocean floor as a result of the sedi-
of about 230m (Fig. 17.8) (although the revised Exxon curve ment load and the removal of sediments from the sea floor,
uses a slightly higher figure of about 250 m). These estimates either through uplift or subduction. Short-term changes in
are in fairly close agreement with the 270 m fall recently global sea level of a few tens of metres could possibly occur
calculated independently from the elevation of sediments if these two factors do not exactly compensate for sedimen-
known to have been deposited near sea level on the North tation.
442 Endogenic-exogenic interactions

Changes in global sea level over greater time scales Death Valley, would have been about 2800 km in length,
could result from long-term variations in the amount of 850 km wide and up to 3 km deep. The water evaporated
sediment supplied to the oceans from the continents as a during such a desiccation event would have eventually been
consequence of climatic change or tectonic factors. Another returned to the world's oceans and would have initially
factor affecting global sea level has been the apparently produced a global sea-level rise of about 15 m, which would
sharp increase in rates of carbonate accumulation in the have been reduced to about 10m after the ocean floor had
ocean basins since the mid-Cretaceous (as a result of the adjusted isostatically to the increased water load.
more active growth of carbonate-secreting marine organisms). The initial isolation of the Mediterranean may have been
This may have resulted in a 300 m decrease in average caused by localized uplift around the Straits of Gibraltar
ocean depth which would produce a global sea-level rise of where it is connected to the Atlantic Ocean, or it may have
about 55 m after isostatic compensation. been due to a global sea-level fall, perhaps related to the ac-
A further effect of sedimentation rates is related to sea- celerated growth of the Antarctic ice sheet in the Miocene.
floor spreading. Since sea-floor spreading rates over the Although there are alternative possible explanations for the
past few million years have been slow in comparison with occurrence of evaporites on the floor of the Mediterranean,
the Cretaceous, the average age of the oceanic lithosphere the additional presence of aeolian sediments seems to be
is probably much greater now than around 100 Ma BP. As- conclusive evidence in favour of the desiccation hypo-
suming constant rates of sedimentation this means that there thesis. Subsequent research has indicated that the Mediter-
would be proportionately less sediment in the ocean basins ranean may have dried out on several separate occasions
during the Cretaceous than today. It has been estimated that between 5.5 and 5 Ma BP, with the Straits of Gibraltar
this extra sediment could produce a sea-level rise of 15 m periodically acting as a giant cataract as the Mediterranean
and that the combination of all three factors gives a net rise Basin was refilled after each desiccation event. While the
of global sea level of 77 m since the mid-Cretaceous. As connection with the Atlantic Ocean was completely severed
the volume of the mid-oceanic ridges has been decreasing the likely rate of evaporation of 4000 km 3 a-I could have
during this period this extra sediment would have had the reduced the Mediterranean to an arid basin in about 1000 a.
effect of reducing the expected fall in sea level by about 30 Given these enormous evaporative losses it is likely that
per cent. some 40000 km 3 a-I of water (about 1.25 x 106 m3 S-1 or
around seven times the mean annual discharge of the Amazon
17.5.2.3 Effects of orogeny River) must have cascaded through the Straits of Gibraltar
As orogeny is frequently associated with crustal shortening, during the refloocting episodes.
which gives rise to both a thickening of the continental Similar desiccation events may have occurred in the past,
crust and a reduction in continental area, it can potentially particularly during the early stages of continental break-up.
have a role in global sea-level change. The collision of It has been suggested that periodic drying out of the very
India and Eurasia leading to the formation of the Himalayas large isolated ocean basins of the nascent South Atlantic in
and the Tibetan Plateau represents the most significant oro- the Early Cretaceous, which is indicated by the presence of
genic event since the Palaeozoic. Assuming that the Tibetan thick evaporite deposits, may have produced rapid global
Plateau is formed of continental crust of twice the average sea-level rises of up to 60 m.
thickness, and that the reduction in continental area arising
from the crustal shortening is replaced by ocean of average
depth, then the resulting global sea-level fall would be 17.5.3 Geoidal effects
about 26 m, although isostatic compensation would reduce
this to 18 m. Other Mesozoic and Cenozoic orogenic events We noted in Section 17.5.1 how changes in the volume of
have probably had effects an order of magnitude smaller water in the oceans cause the crust to deform as it adjusts
than this. isostatically to the alterations in load. These crustal defor-
mations, however, are not confined to simple vertical move-
17.5.2.4 Desiccation of small ocean basins ments; they also involve a continuous redistribution of mass
The idea that the drying out of small ocean basins could between and within the ocean basins. The record of relative
occur and lead to a rapid, albeit rather modest, change in sea-level change observed at any point on the Earth's
global sea level is a rather recent one. In the early 1970s surface consequently represents the net effects of a series of
K. 1. Hsii cited the existence of thick evaporite deposits in complex interactions between ocean water, ice masses and
the sediments of the Mediterranean, and the presence of deep both lateral and vertical adjustments in the solid Earth. In
submarine gorges extending from the mouths of major the 1970s geophysicists working in conjunction with geo-
rivers, such as the Rhone and the Nile, as evidence for the morphologists developed a mathematical model which
evaporation of the entire Mediterranean Sea around 5 Ma predicted how these interactions would manifest themselves
BP. The resulting basin, looking something like a giant in different patterns of relative sea-level change around the
 Sea-level change 443

A B C

~
E-==:::J
o

G L(~~l~"'~
EARTH I~I OCEAN I
~
Attraction

Fig. 17.9 Schematic representation of the sequential adjustments to the geoid arising from ice sheet growth and decay and the
associated changes in the volume of water in the world's oceans. The weight of ice depresses the crust (A) and the ice mass attracts the
water (B); the resulting transfer of matter within the Earthfurther distorts the geoid (C). When the ice mass melts, the weight of
meltwater depresses the Earth differentially (D) and more water flows into the depression, thereby increasing the water load (E); these
changes lead to further deformation of the ocean floor and further transfers of matter within the Earth with additional adjustments to
the geoid (F). The interactions between ice loads, water loads and deformation of the Earth are also illustrated (G). (Based on 1. A.
Clark et al. (1978) Quaternary Research 9, Fig. 1, p. 266.)

world during the Holocene (Fig. 17.9). This model predicts Geoidal changes may generate sea-level fluctuations on
six zones across the world's oceans, each characterized by -their own but may also superimpose themselves on sea-
a different history of Holocene sea-level change as a result level movements produced by other mechanisms. If major
of both isostatic and geoidal effects (Fig. 17.10). changes in the geoid have occurred over geological time
Prior to the development of this model it was widely then this would greatly complicate the interpretation of
thought that a single eustatic curve existed which described fluctuations in global sea level. However, this mechanism
the changes in sea level brought about by the waxing and is extremely difficult to verify with respect to the pre-
waning of ice sheets in all areas remote from the main ice Quaternary history of global sea-level change, and although
sheets themselves (and consequently free from the effects it may be of significance its actual effects on sea-level
of glacio-isostasy - see Section 14.4.1.3). The presence of change are unproven.
raised beaches of Holecene age in regions remote from the
sites of Pleistocene ice sheets were attributed to occasional
intense storms, localized tectonic uplift or simply erroneous 17.5.4 Explaining the long-term record of g'lobal sea-
dating, rather than being viewed as evidence of real differ- level change
ences in sea-level history. It now appears that the geoidal
model is able to explain such emergences and account for Any proposed mechanism of global sea-level change must
the slightly different patterns of Holocene sea-level change obviously accord with the rates and magnitudes of sea-level
recorded around the world. fluctuations suggested by the evidence. The only mechan-
More extreme effects on sea level have also been attributed isms which it is generally accepted can cause global sea-
to changes in the geoid through time than tp-ose associated level changes in excess of 100 m are the growth and decay
with the changes in ice and water loads occurring during of major ice sheets and changes in the volume of the world
glacial-interglacial cycles. The present total relief of the mid-oceanic ridge system (Table 17.1). We might add long-
geoid is around 180 m (Fig. 2.1), and presumably the geoid term changes in the geoid or global hypsometry, but the im-
must change, at least to some extent, through time, although portance of these mechanisms and the magnitude of sea-
the likely magnitude and rate of this change is unknown. level changes that might be associated with them is very
444 Endogenic--exogenic interactions

~ ®
--~----------------------
®

6 2
10km

-~~~~~~-------- 1~
1000 0 1000
km

Fig. 17.10 The world's oceans and shorelines divided into zones in which the predicted history of relative sea-level change since the
beginning of the last deglaciation has a different form. Zone I comprises regions originally covered by ice in which isostatic uplift of
the initially ice-covered land outpaces the global sea-level rise but at a declining rate. This produces a continuous fall in relative sea
level which becomes progressively less rapid with time. Zone 2 is characterized by a continuous and rapid rise in relative sea level due
to the progressive rise in global sea level and crustal down warping as a result ofthe collapse of the ice marginal crustal upwarp
(forebulge). In the transitional region between zones I and 2 there is an initial relative sea-level fall followed by a later rise. For zone 3
the predicted pattern is of an initial rise in relative sea level on deglaciation followed by a slight fall, giving maximum sea levels 0.5 m
above the present. This pattern again results from the interaction of the gradually subsiding ice marginal forebulge and the
redistribution of the water load on the Earth's surface following deglaciation. In this zone the model does not exactly fit the field
evidence since many areas show a continuous sea-level rise. Zone 4 is characterized by a continuous sea-level rise while in zone 5 a
small relative fall in sea level occurs immediately after ice sheet melting has ceased. Finally, in zone 6, which comprises all continental
shorelines removed from ice sheet influences. a small relative sea-level fall is predicted as deglaciation is completed. This is a
consequence of the depression of the ocean basins and submerged parts of the continental shelf, resulting from the increased load
contributed by meltwater which causes a crustal tilt across the continental margins and a slight emergence of shorelines. (Based on
W. R. Peltier (1987) in R.i. N. Devoy (ed.) Sea Surface Studies: A Global View. Croom Helm. London, Fig. 3.12, p. 77.)

uncertain. Other mechanisms of fairly well known but more age Earth. The problem arises when we try to account for
limited significance include orogenesis, the desiccation of the large, rapid sea-level changes which characterize the
small ocean basins, and the effects of sedimentation. whole of the Exxon sea-level curve. Even in its more recent
What, then, can we say about the rate of sea-level modified form it contains numerous dramatic sea-level falls
change? Glacio-eustasy operates so rapidly that it is effect- throughout the apparently ice-free Mesozoic and Early
ively instantaneous on geological time scales, but of the Cenozoic. Perhaps the most remarkable of these is the Late
other mechanisms of global sea-level change only ocean Oligocene fall of around 150 m about 30 Ma BP. This seems
basin dessication generates changes at a comparable rate. to have occurred in less than 1 Ma at an average rate of
All the other mechanisms cause sea-level changes which 150 mm ka- 1. This is slow by the standards of glacio-
are slower by at least three orders of magnitude. eustasy, but far too rapid to be accounted for by changes in
We seem to have no difficulty in explaining both the mid-oceanic ridge volume.
rates and magnitudes of changes in global sea level which One possibility is that ice sheets have existed for a much
have occurred during the past few million years of an ice- greater proportion of geological time than previously thought
 Sea-level change 445

Table 17.1 Approximate magnitude and rate of global sea-level change associated with various eustatic mechanisms

Mechanism Magnitude Rate Confidence in Comments Data sources for* *

(m) (mmka- I ) estimate magnitude estimates

Changes in ocean water volume

Glacio-eustasy <150* <12500 High Donovan and Jones (1979)
Change in mean ocean temperature of 10 °C 10 Slow High Donovan and Jones (1979)

Changes in ocean basin volume

Change in volume of mid-oceanic ridges <230 <10 Moderate Kominz (1984)
Imbalance between inorganic sediment <7 <1.5 Moderate Over past 5 Ma Harrison et al. (1981)
deposition and removal
Relationship between sea-floor spreading <15 <0.15 Moderate Over past 100 Ma Harrison et al. (1981)
rates and mean sediment thickness
Variations in rate of organic carbonate 55 0.55 Moderate Over past 100 Ma Harrison et al. (1981)
accumulation
Orogenesis-formation of Himalayas and 18 1.6 Moderate Harrison et al. (1981)
Tibetan Plateau
Desiccation of Mediterranean Sea 10* <15000 High Donovan and Jones (1979)
Hot-spot-related ocean-floor deformation <100 Very slow Low 110-70 Ma BP Schlanger et al. (1981)
Changes in continental hypsometry <130 Very slow Low 140-50 Ma BP Wyatt (1986)

* Isostatic compensation for change in water load included in estimate.

** Full source details can be found in the list of references at the end of the chapter.

and so can explain rapid global sea level changes before the We have been talking here about sea-level rise, although
Late Cenozoic. This may be true for the Early Cenozoic, it is clear that the trend over the past 80 Ma or so has been
but the palaeoclimatic evidence strongly indicates a rather for global sea level to fall. Nevertheless, if we look at many
uniformly warm ice-free Earth in the Mesozoic. The key of the world's continental margins we find that a significant
problem is to determine when ice sheets first became estab- proportion of them are passive margins formed by the
ished on Antarctica as we are fairly certain that major ice break-up of Gondwana and Laurasia. As they have cooled
sheets did not become established in the northern hemi- and become loaded by sediment offshore, these margins
sphere until the Late Pliocene. The beginning of glaciation have been progressively subsiding (see Sections 4.4 and
on Antarctica is very uncertain, but the weight of evidence 4.5). Consequently most passive continental margins have
is currently in favour of it beginning between 45 and 20 Ma experienced a relative sea-level rise over the past 100 Ma or
BP; the most recent evidence from the Ocean Drilling so even though global sea level has been falling. Signi-
Program indicates that there was glacial activity at sea level in ficantly, because of their potential for oil exploration, a
East Antarctica by 35Ma BP (see Section 14.3.1). large proportion of the data used to construct the Exxon sea
A possible alternative explanation is that most of the level curve has come from these subsiding passive margins.
rapid sea-level falls recorded in the Exxon curve are in fact It'is possible, then, that all we need to explain the rapid
regional regressions and do not indicate changes in global regressions in the Exxon sea level curve are variations in
sea level. In order to understand how this might occur we the rate of global sea-level fall. This in tum would give rise
have to look at the relationship between sedimentation, to variations in the rate of relative sea-level rise along sub-
continental margin subsidence and relative sea-level change. siding passive margins and consequently a succession of
As we noted in Section 17.2.2.1, whether a shoreline moves regressions and transgressions.
landward (transgression) or seaward (regression) depends
not just on the relative movement of land and sea but also
on the rate of deposition. Along a continental margin which 17.6 Sea-level change and landscape development
is gradually subsiding and is receiving a more or less con-
stant supplyQf sediment, the shoreline will be located where
the rate of sea-level rise is balanced by the rate of depo- 17.6.1 Drainage basin response
sition. If the rate of sea-level rise slows down then the Relative changes in sea level have both direct and indirect
shoreline will move seaward, causing a regression, since effects on drainage basins. The direct effects are the adjust-
sedimentation will outpace the now slower rate of sea-level ments of drainage systems to the changes in base level and
rise. Conversely, if the rate of sea-level rise increases the the increase or decrease in drainage basin area brought
shoreline will move landward, producing a transgression as about by a relative rise or fall in sea level. The indirect effects
the rate of sea-level rise will exceed the rate of deposition. arise from the climatic changes associated with alterations
446 Endogenic-exogenic interaction

stream part of the drainage basin will be submerged, and if

the sediment supply is sufficient there will be deposition
and a delta may be constructed. Further deposition may
then occur above the delta as a new profile becomes estab-
lished. This regrading of the river long profile will not

B 51
!
j .... ~--

Fig.17.11 The geography of the Sunda 5heLfin south-east Asia

during the sea-levellowstands of the Pleistocene. The coastline is
drawn in the position it would occupy if sea level were 100 m
below that of the present. Much of the 5unda Shelf is dry land and
is drained by rivers originating in present-day Sumatra and
Borneo. (The reconstruction of the drainage is partly based on
x 51
I. H. F. Umbgrove (1942) The Pulse of the Earth. Martinus C
Nijhoff, The Hague, Fig. 83, p. 131.)

in the relative area of land and sea. A relative rise in sea

level reduces the drainage basin area, whereas a relative fall
can produce a marked downstream enlargement, especially
where there is a shallow continental shelf offshore. In such
cases the significant increase in land area is likely to lead to Fig. 17.12 Schematic representation of drainage system
response to afall in sea level along continental margins with
greater aridity in the continental hinterland and a conse- different configurations. In (A) there is afall in sea level (L,-L 2)
quential reduction in runoff. The sea-level fluctuations of at a margin with a constant gradient and the shoreline migrates
the Late Cenozoic caused major shifts in drainage area and from position 5, to position 52' Although deposition rates may
the periodic submergence of extensive drainage systems in consequently fall in the newly exposed zone between 5, and 52 we
the relatively brief interludes of high interglacial sea level would not expect any significant incision here as the gradient of
the margin is no greater than inland of the original shoreline.
such as we are experiencing at the present day (Fig. 17.11). Certainly it is unlikely that channel incision is likely to be
Although the most widespread effects of sea-level fluctu- sufficient to cause the rejuvenation of river systems inland of the
ations on drainage systems are associated with changes in original shoreline (5,). In (B) there is an increase in gradient
base level, these effects are not always simple or self- between the initial (5,) and subsequent (52) shoreline positions. In
evident. A relative fall in sea level will, of course, increase this case we would expect the greater erosional energy of rivers
seaward of the point at which the gradient increases to lead to
drainage basin relief and thus the potential energy available channel incision and a regrading of river long profiles (indicated
in the landscape (Fig. 1.5). But the response of fluvial by dashed line). Ultimately such regrading may extend inland
systems to a relative change in sea level depends on the beyond the original shoreline (5,). In (C) the fall in sea level
way in which the change is generated and on the morphology eventually exposes an area of lower gradient margin. In the upper
of the coastal zone across which the shoreline migrates. To section above the point at which the gradient decreases (x) we
would expect a simple extension of river systems with little
understand these factors we need to examine two situations. incision since there is no change in gradient above and below the
The first is where the land is stationary, and the relative initial shoreline (5,). Below this slope break, however, the
change in sea level is caused entirely by a change in the decrease in gradient would probably promote deposition. This
position of the sea surface; the second is where sea level is could eventually lead to the regrading of the whole of the newly
stationary, and the land rises or subsides. (In reality, of exposed coastal plain between 5, and 52 and deposition might
eventually extend inland of the original shoreline. (From M. A.
course, both effects may operate simultaneously.) Summerfield, (1985) in: M. Morisawa and I. T. Hack (eds)
Where the sea surface rises along a stable continental Tectonic Geomorphology. Allen and Unwin, Boston, Fig. 3,
margin the effects are relatively easy to predict. The down- p.39.)
 Sea-level change 447

necessarily extend far up the river system and consequently fall of 500 m exposing the present submarine topography of
the effects of such a base level rise will not necessarily the continental margin. The gradients of rivers extending
extend throughout an entire drainage basin. over the newly exposed broad coastal plain of the we tern
A fall in sea level might be expected to have the opposite and southern margins would in general be less than the
effect, that is, lead to erosion and the incision of drainage average gradients of the pre ent-day channels at the coast.
ystems. Whether this occurs, however, depends on the Only along a section of the eastern coast, where there is a
gradient of the surfaces over which a newly extended drain- very steep offshore slope, would there be a significant increase
age channel flows towards the now more distant shoreline. in seaward channel gradients and consequently active river
In particular, it depends on whether channel gradients incision into the exposed continental shelf. A situation similar
across the newly exposed continental shelf are greater, less, to this (although with a rather less dramatic sea-level fall
or the same as those where the channel approached the than we have assumed above for illustrative purposes) may
original shoreline (Fig. 17.12). Only if gradients across the have occurred during a possible major regression in the
newly exposed shelf are greater than those immediately late Oligocene-Early Miocene. Interestingly, the numerous
inland will significant river incision occur. This may in tum submarine canyons of the eastern margin of southern Africa
lead to rejuvenation of river systems inland of the original appear to have been cut during this period.
high sea level shoreline. Active incision of the continental The sea-level fluctuations of the Quaternary which have
shelf by rivers during periods of low sea level may cut deep caused base levels to be lowered by up to 100 m for sub-
valleys which, after a subsequent sea-level rise, become stantial periods of time during the past 1Ma or so have
submarine canyons. promoted widespread valley incision along continental mar-
The possible effects of different channel gradients across gins. Such valley downcutting is not normally evident today
a continental margin can be illustrated by looking at the since the Holocene sea-level rise has led to valley aggra~
consequences of a ignificant lowering of base level in dation and the burial of these incised valleys by sediment.
southern Africa (Fig. 17.13). Imagine a relative sea-level Perhaps one of the best documented instances of valley

I
Oliphants
- Berg
Estuary
o 500
I I

km

....- - - - - - . Sections of continental margin where gradient from sea level to

500m > gradient from 500 to 1000m

Fig. 17.13 Hypothetical effect of a major sea-level fall on the drainage systems ofsouthern Africa. The labelled contours indicate
elevations above sea level in metres after a relative sea-level fall of500 m. This means that the 500 m contour coincides with the
present-day coastline. The dashed lines schematically represent the courses of major present day rivers across the exposed continental
margin. Sections of the margin where a major relative fall in sea level would lead to a downstream increase in channel gradient, and
hence the potential for channel incision, tend to coincide with actual areas characterized by incised valleys and submarine canyons.
(Based on R. V. Dingle et al. (1983) Mesozoic and Tertiary Geology of Southern Africa. Balkema, Rotterdam, Fig. 190, p.313.)
448 Endogenic-exogenic interactions

Williana
Terrace
Bentley
Terrace

Montgomery
Terrace
Valley - cutting episodes
Prairie 5 4 3 2 1
Terrace
Present

_______ _~FIO~~~:_
--
...... : J
-- -- ----
..- - - ----
....... - - - -

Fig.17.15 An interpretation of the Quaternary histOly of the

lower Mississippi Valley based on an idealized representation of
terrace relationships. Five successive stages of channel incision
during low glacial sea levels and terrace formation during high
interglacial sea-levels are shown. (After H. N. Fisk (1944)
Geological Investigations of the Alluvial Valley of the Lower
Mississippi River. Mississippi River Commission, Vicksburg,
Fig. 77.)

downcutting and subsequent aggradation is provided by the

lower Mississippi Valley (Fig. 17.14). In this case there
was an extremely high rate of deposition because of the enor-
mous amounts of glacially-derived sediment being trans-
ported from the north by meltwater as the Laurentide ice
sheet retreated. According to some interpretations the terrace
deposits of the lower Mississippi Valley indicate several
distinct phases of sedimentation interspersed with episodes
of valley incision which can be correlated with alternating
high and low Pleistocene sea levels (Fig. 17.15).
Another dramatic example of the effects of Quaternary
sea-level change on drainage systems is provided by the
Amazon system. The channel of the Amazon River itself
has a remarkably gentle gradient of only 0.1 m km- I from
the foothills of the Andes to the Atlantic Ocean. Downstream
Fig. 17.14 Late Quaternary history of the lower Mississippi
Valley showing the transformation from an entrenched river of Manaus, which is almost 1500 km from the ocean, the
developed during the low sea levels of the last glacial to the gradient is a mere 0.03 m km-I. In this lower reach the
present-day meandering alluvial channel. (A) The situation Amazon has a very broad floodplain up to 100 km across.
around 15 000 a Sf with the Mississippi entrenched by around This extraordinary situation is lc:rrgely an artefact of post-
120 m in response to the lowering of base level and an increase in
glacial sea-level rise. During the low sea levels of the last
channel gradient to about 0.155 m km- I . From 14 000 to 4000 a
Sf the post-glacial rise in sea levelled to the deposition of mainly
(and previous) glacials the Amazon was deeply entrenched,
coarse sediments, supplied in prodigious quantities by meltwater but the rise in post-glacial sea level initially transformed
from the decaying Laurentide ice sheet to the north, and the the lower Amazon into an enormous ria. Once global sea
establishment of a braided channel with a gradient of around level began to stabilize in the Holocene the sediment-charged
0.140 m km- J (B). A progressive reduction in both the rate ofsea-
western Amazon rapidly infilled this embayment, thereby
level rise and the size of sediment being carried by the river
gradually led to the deposition of predominantly fine sediments creating the wide, low-gradient floodplain of the present-
and the reduction of the average channel gradient to about day lower Amazon. The rate of sediment' accumulation was,
0.127 m km- I (C). Finally, the stabilization of sea level and in fact, so rapid that floodplain deposits blocked the mouths
reduction in sediment load promoted the creation of a single of major lower Amazon tributaries, such as the Xingu and
meandering channel with an average gradient of approximately
Tapajos.
0.112 m km- I (D). (After H. N. Fisk (1944) Geological
Investigation of the Alluvial Valley of the Lower Mississippi If the relative change in sea level arises from vertical
River. Mississippi River Commission, Vicksburg.) movements of the land rather than the sea, the effects on
drainage systems will depend on the nature of the crustal
 Sea-level change 449

c cla ified on the basis of whether they were emergent

rrr
B (associated with a relative fall in sea level) or submergent
A (associated with a relative rise in sea level). Unfortunately,
uch a classification is not as relevant as it might fir t appear

t~LJ
for two main rea ons. First, since most coastlines have
A
B experienced a relative rise in sea level during the Holocene,
c
c
0
·iii erosional landforms diagnostic of submergence, such as
0 CIl

C
·iii
CIl
~ rias and fjords, are very widespread. Secondly, because
Ql Cl
OJ
CIl
C
Cll
many coastal landforms do not adjust quickly to changes in
Ql
a: ~ sea level, the rapidly fluctuating sea levels of the Quaternary
----. have produced coasts which exhibit characteristics asso-
Fig.17.16 Effects of continental-margin flexure on coastai
gradients and relative sea level. A-A shows the initial ciated with both submergence and emergence. This is clearly
configuration of the margin with the shoreline positioned at A . the case with many erosional forms, but even applies to
~-B shows the situation after the flexure of the margin has S depositional forms which would be expected to respond
mcreased and the axis of rotation has moved seaward. As a result more rapidly to changing sea levels. This has resulted in
the gradient across the margin increases and there is uplift along present depositional landforms related to high Holocene sea
the coast and a regression as the shoreline moves seaward to Bs .
C-.C shows afurther increase inflexure across the margin, but in levels frequently located on sediments laid down during
thls case the axis of rotation has moved inland. This also leads to previous interglacial sea level highstands.
a steepening of the gradient across the margin, but the landward One feature of present-day coastlines around the world is
migration of the axis of rotation means that there is a the widespread erosion of beache and other types of coastal .
transgression as the coast subsides and the shoreline migrates depositional landforms. This, too, appears to be related to
lan~war~ to Cs' (Based on L. King (1982) The Natal Monocline.
Umverslty ofNatal Press, Pietermaritzburg, Fig. 12, p. 34.)
sea-level change. During the rise in sea level at the end of
the last glacial, the transgression of shorelines across what
had previously been coastal plains meant that new sources
deformation. Where the entire land area rises or subsides
of sediment were constantly being encountered. The coarser
uniformly then the effects will be similar to a fall or rise in
fraction of this sediment was transported onshore to form
the sea surface. However, along passive margins, which are
beaches, or blown inland to create coastal dunes. As sea levels
characterized by crustal flexure with uplift inland and
began to stabilize at roughly their present position around
subsidence offshore (see Section 4.5.2), the picture is more
4000 a BP new sources of sediment no longer became avail-
complex. An increase in flexure of the margin will always
able and consequently many beaches have been experienc-
lead to a steepening of channel gradients in the coastal zone,
ing a net loss of sediment since that time.
but this can be associated with either a regression and rela-
One of the most critical factors determining the response
tive sea-level rise, or a transgression and relative sea-level
of coasts to a change in sea level is the morphology of the
fall, depending on whether the axis of rotation of the margin
continental shelf and the coastal plain. Three distinct effects
moves seaward or landward (Fig. 17.16).
can be identified. The most obvious is the rate of horeline
One further effect worthy of mention is hydro-isostasy.
migration. The lower the gradient of the continental shelf
A fall in sea level will reduce the water load on a con-
and coastal plain over which regressions and transgressions
tinental margin thus cause it to flex upwards. The vertical
occur the greater will be the amount of horizontal move-
movement involved is probably insufficient in mo t cases
ment of the shoreline for a given vertical change in sea
to disrupt river ystems draining across the margin, but in
level. Thus shoreline movements generated by global sea-
other instances the effects may be dramatic. It has been
level changes will be much more extensive and rapid along
estimated, for example, that the desiccation of the Mediter-
generally gently sloping rifted margins that along the
ranean Sea during the Late Miocene caused a flexural
steeper and narrower continental shelves of active margins.
upwarp of its north-western and south-eastern margins of
On continental shelves of very low gradient rates of shore-
up to 450 m along the shelf edge. This would probably
line migration during periods of interglacial sea-level rise
have been sufficient to have caused the landward reversal
have been prodigious. For instance, during the period 18 000 a
of rivers with small discharges and low coastal gradients.
BP to 6000 a BP the shoreline moved inland an average dis-
The Rhone and Nile, on the other hand, apparently had
tance of over 300 km across the eastern margin of the USA
enough erosive energy to maintain their courses across the
at rates of up to 8 m a-I in New Jersey and 13 m a-I in
margin.
Louisiana. It is along such gently loping coasts, charac-
terized by rapid rates of shoreline movement, that it is mo t
17.6.2 Coastal effects likely that there will be insufficient time for littoral pro-
Geomorphologist have long been aware of the importance cesses to cons~ct equilibrium depositional landforms. This
of changes in sea level on coa tal landforms, and this aware- indeed appears to be the case along the eastern seaboard of
ne s led some to propose that all coasts could be usefully the USA where the continental shelf has a range of mor-
450 Endogenic-exogenic interactions

phological zones which have not yet been fully modified by on the voyage Darwin had been reading the first two vol-
present-day shelf processes, and which appear to be related to umes of Lyell's Principles of Geology which argued for a
previous shoreline positions. version of uniformitarianism incorporating the notion that
Two further factors related to the interaction of sea-level the Earth's average morphology remained more or less the
change and the configuration of the continental shelf are same over geological time - that is the Earth's surface as a
wave energy budget and tidal range. The tidal range along a whole existed in a 'steady state' in which uplift in one area
continental margin can be greatly amplified if a low- gradi- would be compensated by subsidence in another. Darwin
ent shelf is submerged by a relative rise in sea level. Simi- applied this idea to the oceans, postulating that regions of
larly, the rate of change in water depth as the shoreline is uplift would be matched by regions of subsidence, and then
approached partly controls the rate at which wave energy is linked this proposition with the observation that areas of
dissipated and hence affects the amount of geomorphic the ocean containing atolls tend to be quite separate from
work that waves are able to accomplish (see Section 13.2). those with active volcanoes. Darwin then saw that the three
A change in sea level which alters the gradient of the shore- types of oceanic islands could be explained as stages in a
line, especially in the critical zone where water depth is less single sequence of development (Fig. 17.17).
than half the wavelength of waves, can therefore indirectly The first stage is the growth of a volcano from the ocean
affect the nature of wave activity along coasts. floor and its emergence above sea level. Volcanic activity
eventually ceases, and in warm tropical oceans coral can grow
in the shallow water around the edge of the volcano and
17.6.3 Oceanic islands
construct a fringing reef. As the volcano subsides the coral
Oceanic islands provide unparalleled 'laboratories' for ex- is able to build the reef upward sufficiently quickly to main-
ploring relative sea-level change, and, incidentally, for the tain its upper surface close to sea level. A barrier reef is
investigation of other important geomorphic problems. Al- consequently formed which is separated from the partially
though they are distributed across the world's oceans they
are far more numerous in the Pacific than elsewhere. In the
mid-nineteenth century, Charles Darwin, who began his
scientific career more as geologist than a biologist, recog-
nized that oceanic islands were of just three types; volcanic
islands, coral reefs and combinations of the two. (We now
know this is not strictly correct as a small proportion of
oceanic islands, such as the Seychelles in the Indian Ocean, A A-Outer edge of the reef at the level of the sell..
B B-Shores of the island.
rise from fragments of continental crust - see Figure 3.32 - A'A'-Outer edge of the reef, after its upward growth during a period
but the classification is true for the oceanic islands we are of subsidence.
C C·-The lagoon-channel between the reef and the shores of the now
concerned with here which rest on oceanic lithosphere.) encircled land.
Although coral reefs are highly variable in morphology, the B' B'-The shores of the encircled island.
N.B.-In this, and the following woodcut, the subsidence of the
ultimate form is the atoll, a ring of coral encircling a land could only be represented by an apparent rise in the level of
shallow lagoon. the sell..

The existence of coral atolls rising from the depths of the

ocean provided a puzzle for nineteenth-century geologists.
Charles Lyell, Darwin's geological mentor, regarded them
as merely the craters of submarine volcanoes which had
been overgrown by coral, but such an explanation did not
account for many of the morphological characteristics of
atolls. Although many hypotheses of atoll formation were
proposed in the nineteenth and early twentieth centuries, it
was Darwin who, in 1837, first put forward the theory link-
A'A'-Outer edges of the barrier-reef at the level of the sell.. The
ing volcanic islands and coral reefs that we know today to coco -nut trees represent coral-islets formed on the reef.
be essentially correct. C C-The lalioon-channel.
B' B'-The shores of the island, generally formed of low.alluvialland
Darwin's coral reef theory was one of the most remark- and of coral detritus from the lagoon-channel.
able examples of scientific deduction in the history of the A"A"--·The outer edges of the reef, now forming an atoll.
C'-The lagoon of the newly-formed atoll. According to the scale
earth sciences in that all its essentials were worked out before the depth of the lagoon and of the lagoon-channel is exaggerated.
he had ever seen an atoll. During his voyage around the
world in the surveying ship H M S Beagle (1831-36) Darwin Fig.17.17 Diagrams from Darwin's book on coral reefs
illustrating his theory of the development of atolls from volcanic
observed graphic evidence of recent significant uplift on the islands. (From C. Darwin (1874) The Structure and Distribution
western coast of South America which he considered to be of Coral Reefs (2nd edn) Smith Elder, London, Figs. 5 and 6,
closely linked to the active volcanism of the Andes. While pp. 131 and 133.)
 Sea-level change 451

submerged volcano by a lagoon. Eventually the volcano Sea Level

subsides completely below sea level to leave an atoll which
continues to build upwards at the rate of subsidence. Darwin ai
>
~ Ocean
was well aware that his entire t.heory depended on the abi- co
Q)
lity of coral to grow upwards at a rate at least as rapid as (J)

subsidence, and he was careful on his return to England to ~

o
verify this assumption from empirical observations. ai
co
Darwin's coral reef theory is remarkable both for its sim- ..c
plicity and explanatory power, but it was only with the
a.Q)
Q
advent of plate tectonics that a full understanding of the SUb-lithospheric
thermal anomaly .J\.
process of volcano subsidence (which is such an important '--------"-----'----------;'11
element of the theory) was achieved. Although drilling into Lithospheric Age
atolls has demonstrated, as Darwin predicted, great thick-
Fig.17.18 Changes in the elevation of the ocean floor as a
nesses of carbonate of up to 1200 m resting on a volcanic result of heating associated with a sub-lithospheric thermal
base, there are a number of potential causes of subsidence. anomaly. Such a heating anomaly may give rise to a volcano
One is global sea-level rise which could cause apparent which grows sufficiently to emerge above sea level. (After R. S.
subsidence, but the vertical movements involved and the Detrick and S. T. Crough (1978) Journal of Geophysical Research
subsidence histories reconstructed for different atolls rule 83, Fig. 4(C), p. 1239. Copyright by the American Geophysical
Union.)
this out as the most common mechanism. Isostatic subsi-
dence as a result of the load of the coral and volcanic edi- than the amount of subsidence experienced by many atolls.
fice on the oceanic lithosphere is another possibility but again The only feasible mechanism is therefore the subsidence of
the vertical movement that this could produce is far smaller the oceanic lithosphere as it ages and cools. As can be seen

Kamchatka 4 3 2 STAGES OF DEVELOPMEN

Pen ins u Ia ,.....,...,..,..,."".,.!T1\"T1:>,,·,,~,·f--------+-------+------+-----+----------A--l
Coral growth Sub-aerial

W
AI' Seamount ceases in cooler Fringing denudation
T~eu~~~n subduction water. Atolls reefs become and fringing
o submerged to atolls reef growth
70 Ma become seamounts

~Kurile Trench

m Sea level
3 o
"0 o EAST PACIFIC -
..,
Cll RISE
o.., o
o

(fl
Cll
OJ
'3
o
<;,... 050 Ma
IJl
40Ma Midway

I
o
Island
o
00 / ~ Volcanic and coral
o 00 ,30 Ma Gardner islands
.J '• .1. Pinnacles :) Shallow banks
It '.' (1 '-' •• 20 Ma I Oahu

'.O;~0/0 '" ,. 10,~:.

Pacific Seamounts
o"'Oii on / fHawaii
Ocean nd s ....
..~ <2 Ma
~ a
I
1000
I

km
B

Fig.17.19 The development of the volcanoes, atolls and seamounts of the Hawaiian-Emperor chain (A), and map showing the chain
extending from the island of Hawaii to the Kurile Trench off the Kamchatka Peninsula (B). The change in the orientation of the chain
about halfway along is due to an alteration in the direction of movement of the Pacific Plate about 40 Ma BP. The dates along the chain
indicate the age of volcanic activity. (Based partly on G. A. 1. Scott and G. M. Rotondo (1983) Coral Reefs 1, Fig. 5, p. 146 (A), and
D. A. Clague et al. (1975) Geological Society of America Bulletin 86, p. 991 (volcano ages in (B)).)
452 Endogenic-exogenic interactions

from Figure 17.6, this can involve subsidence of nearly Such a sequence is beautifully exemplified by the
3000 mover 80 Ma. Hawaiian-Emperor seamount chain which can be traced as
We can now construct a model of volcanic island and a line from the present centre of volcanic activity on the
atoll development by considering the effects of a sub-litho- island of Hawaii north-westwards towards the Kamchatka
spheric thermal anomaly (see Section 4.2.1) located beneath Peninsula (Fig. 17.19). In tropical waters coral growth can
moving oceanic lithosphere (Fig. 17.18). Heating at the keep pace with subsidence of this kind, but if atolls move
base of the lithosphere makes it less dense and causes it to into cooler waters (as in the case of the Hawaiian-Emperor
rise, thereby interrupting its 'normal' cooling and subsidence seamount chain) corals cannot grow and atolls subside
behaviour. If the heating anomaly is sufficiently vigorous to below sea level to form flat-topped seamounts, or guyots.
create a hot spot, magma will penetrate the lithosphere and A survey of the world's coral reefs shows that there is a
a volcano will form. The continuing motion of the litho- wide range of forms (Fig. 17.20). These include atolls that
sphere will, however, soon carry the volcano away from the have apparently been raised by 50 m or more above sea
heating anomaly and eruptive activity will cease. This vol- level, so we must ask whether the simple model that we
cano will then be carried gradually below sea level on the have just outlined provides a comprehensive explanation of
now recooling oceanic lithosphere. coral reef evolution. The answer is no, but it does appear

a) Volcanic island with no fringing reef b) Volcanic island with fringing reef c) Raised volcanic island with fringing reef

.......
>< t".. " ..
"" ........ < ~ )-
fringing reef Y.> ~ JI\ ..; A " ~
" ... ~~r----
.... \II' 1" A'" A ""'ii
/ <t'~ .. Jt>~~,,..\tI
\' '" ~ ,. ~ \'t"".," t""':" '"
'4 ...
t" ...... \,~ .. '4 <)
~t""'.,"''".) "1~A~"
'4 '" ........ \' ~

ego Kusaie, Caroline Is. ego Oahu, Hawaiian Is.

d) Almost-atoll e) Raised almost-atoll f) Atoll

raised lagoon sediments

lagoon
I

ego Truk, Caroline Is. ego Atiu, Cook Is. ego Arno, Marshall Is.
g) Inundated atoll (dry only at low tide) h) Part-raised atoll (lagoon open) Part-raised atoll (lagoon enclosed)

ego Pearl and Hermes Reef, Hawaiian Is. ego Gardner Atoll, Hawaiian Is. ego Swains I, north of Samoa
j) Raised atoll (dried out lagoon) k) Raised atoll (atoll form lost)
Basaltic core that was
initially a volcanic island

Reef facies produced during stable

or submergence phases

Reef developed during

or following emergence phase
ego Nauru, Gilbert Is. ego Niue, east of Tonga Is.

Fig. 17.20 Types of oceanic island in the Pacific Ocean. Differences between the various types reflect the net effects of volcanism,
oceanic lithosphere subsidence and flexure, sub-lithosperic heating and consequential uplift, eustatic sea-level change, coral growth
and sub-aerial denudation. (From G. A. J. Scott and G. M. Rotondo (1983) Coral Reefs 1, Fig. 2, p. 141, based largely on H. 1. Wiens
(1962) Atoll Environment and Ecology. Yale University Press, New Haven and O. K. Leont' yev et al. (1975) USSR Oceanology 14,
840-846.)
 Sea-level change 453

A 1Ma B.P that most of the deviations from the model can be explained
Coral reef by known erosional, tectonic and eustatic effects. For in-

AJ\
Sea level "-
stance, there are two ways in which atolls can be elevated.
Volcanic One is the possibility that an existing atoll may be carried
substrate
over another sub-lithospheric thermal anomaly. It will then
_Cooling and subsiding oceanic lithosphere
be elevated as the lithosphere on which it rests expcriences
300ka B.P uplift. A second possible mechanism is probably only
BQ
Co Volcano En;Jergence
~ relevant to rather modest amounts of uplift. This can arise
if a new volcano appears amidst a group of atolls as they
pass over a heating anomaly beneath the lithosphere (Fig.
17.2 I). The new volcano loads the lithosphere and causes it
to flex, producing a moat immediately around it and a more
distant peripheral arch - this is analogous to the crustal
forebulge around an ice sheet. Islands within the moat wilJ
consequently subside whereas those on the arch wiJ1 be
elevated. In some cases several volcanoes can be seen to be
simultaneously loading the lithosphere and this, not unex-
--Moat
pectedly, produces very complex patterns of island uplift
Further flexure of oceanic lithosphere
and subsidence (Fig. 17.21). We therefore have to be very
D careful when using oceanic islands as 'dipsticks' with which
AITUTAKI to monitor global sea-level change since closely adjacent

~®
'8 Ma MANUAE
• Mitiaro
islands may exhibit significantly different sea-level histories
Predict. 27m over time spans as ShOit as the Holocene.
Obs.27m
SOUTHERN a 0 r:'-l
COOK ~ ~
20·5 ISLANDS £J .11--"'// J,~ Further reading
RAROTONGA Until the mid-1980s anyone with. an Interest in sea-level
change had to search through a vast journal literature in
order to get an up-to-date perspective on the subject. Fortu-
nately, this is no longer the case, as there are now some
excellent reviews of various aspects of the history and causes
of sea-level change. The most recent and wide-ranging sur-
22"S 0 Obs 70m
vey is that edited by Devoy (1987a), while the collections
o
edited by Smith and Dawson (1983) and Tooley and Shennan
© o 100 (1987) provide more specialized treatments focusing on
~ , km !
Quaternary sea-level change.
160·W 158·W On the evidence for Quaternary sea-level change, Suther-
land (1983, 1987) provides a comprehensive discussion of
Fig. 17.21 Relative sea-level change associated with the flexure
of the lithosphere as a result of loading by a volcano. (A) The the problems of dating, while Chappell and Shackleton (1986)
initial situation with coral reeff growing on volcanic substrate attempt to reconcile the coral terrace sequence of the Huon
subsiding on coolinl( oceanic lithosphere. (8) The initial effects of Peninsula first presented by Chappell (1974) with the oceanic
lithospheric flexure brought about by the development ofa oxygen isotope record. Turning to the pre-Quaternary record
volcano near by with submergence in the 'moat' and emergence
Bond (1979) and HaITison et al. (1981) demonstrate that
on the 'arch'. (C) Further submergence and emergence occurs as
the mass of the volcano increases and the flexure of the lithosphere sea~level reconstructions based on the amount of present-
becomes more marked. Note the short time scale of such day land flooded at various times in the past must take into
deformation. (D) The application of a flexural model to the account changes in continental hypsometry, while Sahagian
southern Cook Islands in the southern Pacific. Loading has been (1987) shows how the central cratonic region of North
caused by three volcanoes - Rarotonga, Aitutaki and Manuae -
America can be used as a 'stable' frame of reference for the
and the predicted uplift for the surrounding islands is compared
with the observed uplift. The match is close exceptIor Mangaia estimation of the position of global sea level in the Cretaceous.
which may be within the zone of deformation ofa fourth volcano The procedures for calculating global sea-level fluctuations
to the south not included in the calculations. (Based on M. from seismic stratigraphy are described in detail in the semi-
McNutt and H. W. Menard (/978) Journal of Geophysi<.:al nal memoir edited by Payton (1977) (see especially Vail et
Research 83, Fig. 2, p. 1207 and Fig. 4, p. /210; and General
ai, 1977a, 1977b), and these are updated in Schlee (1984);
Bathymetric Chart of the Oceans, 1984. Canadian Hydrographic
Service (hathymetric contours).) Devoy (1987b) provides an accessible introduction to this
topic. Detailed assessments of the problems of reconstruc-
454 Endogenic-exogenic interactions

ting global sea-level change from seismic stratigraphy are evident in the detailed review by Orford (1987). Oceanic
made by Burton et al. (1987), Hubbard (1988) and Miall islands have recently attracted much attention, from geo-
(1984, 1986), while Haq et al. (1987) provides the most physicists as well as geomorphologists. The beautifully
recent version of the 'Exxon curve' which incorporates evi- illustrated book by Menard (1986) gives a superb introduc-
dence from continental outcrops of marine sediments. I tion to the topic, while Guilcher (1988) provides a more
The record of Quaternary sea-level change is covered in conventional geomorphic treatment. On more specific topics
the general reviews by Morner (1987a) for the northern Scott and Rotondo (1983) discuss the life history of oceanic
hemisphere and Pillans (1987) for the southern hemisphere islands as they subside on ageing lithosphere, and McNutt
(although the former should be read with due regard for the and Menard (1978) and Spencer et al. (1987) examine the
author's enthusiasm for geoidal effects on global sea level). uplift of atolls through the loading effects of new volcanic
More detailed discussions of particular areas are to be found islands. The reading of Darwin's original work on coral
in Tooley and Shennan (1987) which includes an excellent reefs, which is available as a reprint (Darwin, 1984), is a
evaluation by Chappell (1987) of the Late Quaternary record rewarding experience for any geomorphologist not just for
in the important Australian region. For the pre-Quaternary its detailed observations on atolls but also as an exemplar
record the Exxon group summary by Haq et al. (1987) and of the application of a scientific methodology appropriate to
the more wide-ranging review by Hallam (1984) are par- the study of long-term landform development. Finally, the
ticularly useful. potential of oceanic islands as a 'test-bed' for geomorphic
Mechanisms of sea-level change are discussed in a brief theory is advocated by Nunn (1987) who points out the
but useful review by Donovan and Jones (1979), but for simplicity of controls on their landform development in com-
more detail it is necessary to refer to papers on specific parison with continental areas.
mechanisms. The link between glacial-interglacial cycles
and changes in ocean water volume as established by the
oxygen isotope record is described by Shackleton and
Opdyke (1973), while Matthews (1984) considers the rel- References
evance of glacio-eustasy to Early Cenozoic global sea-level Andrews, 1. T. (1987) Glaciation and sea level: a case study. In:
change. Pitman (1978) provides an estimate of the effect on R. J. N. Devoy (ed.) Sea Surface Studies: A Global View.
Croom Helm, London and New York, 95-126.
global sea level of changes in mid-oceanic ridge volume
Bond, G. C. (1979) Evidence for some uplifts of large magnitude
which is refined by Kominz (1984). Changes in sea level in continental platform~. Tectonophysics 61,285-305.
related to sediment accumulation in the oceans are examined Burton, R., Kendall, C. G. St. C. and Lerche, 1. (1987) Out of our
by Harrison et al. (1981), while Schlanger et al. (1981) con- depth: on the impossibility of fathoming eustasy from the strati-
siders the effects of thermally induced uplift of the ocean graphic record. Earth-Science Reviews 24, 237-77.
Chappell, J. (1974) Geology of coral terraces, Huon Peninsula,
floor. Wyatt (1986) speculates on the effects on global sea
New Guinea: a study of Quaternary tectonic movements and sea-
level of changes in continental hypsometry and the Medi- level changes. Geological Society of America Bulletin 85,
terranean desiccation hypothesis is vividly presented by 553-70.
Hsti (1972). The geoidal model of sea-level fluctuations Chappell, J. (1983) Aspects of sea levels, tectonics, and isostasy
related to changes in ice and water masses around the globe since the Cretaceous. In: R. A. M. Gardner and H. Scoging
(eds) Mega-Geomorphology. Clarendon Press, Oxford and New
is covered in an accessible way by Peltier (1987) who has
York, 56-72.
also provided an overview of the complex mathematical Chappell, j. (1987) Late Quaternary sea-level changes in the
analysis on which the model is based (Peltier, 1982). Assess- Australian region. In: M. J. Tooley and 1. Shennan (eds) Sea-
ments of the correspondence between the model and degla- Level Changes. Institute of British Geographers Special Public-
ciation sea-level histories for particular areas of the world ation 20. Blackwell, Oxford and New York, 296-331.
Chappell,1. and Shackleton, N. 1. (1986) Oxygen isotopes and sea
include those by Andrews (1987) and Devoy (1987c),
level. Nature 324, 137-40.
while the more speculative consequences of possible large- Darwin, C. (1984) The Structure and Distribution of Coral Reefs.
scale changes in the geoid for long-temi sea-level change University of Arizona Press, Tucson.
are discussed by Morner (1981, 1987b). Finally, the pro- Devoy, R. J. N. (ed.) (1987a) Sea Surface Studies: A Global View
blematic relationship 'between changing rates of passive mar- Croom Helm, London and New York.
Devoy, R. J. N. (1987b) Hydrocarbon exploration and biostrati-
gin subsidence and the transgressive and regressive sedi-
graphy: the application of sea-level studies. In: R.1. N. Devoy
mentary sequences which form the basis for sea-level cur- (ed.) Sea Surface Studies: A Global View. Croom Helm, London
ves derived from seismic stratigraphy is considered by Watts and New York, 531-68.
(1982) and Pitman and Golovchenko (1983). Devoy, R. 1. N. (1987c). Sea-level changes during the Holocene:
The response of drainage basins to sea-level change has the North Atlantic and Arctic Ocean. In: R.J.N. Devoy (ed.)
Sea Surface Studies: A Global View. Croom Helm, London and
not yet received sufficient attention, but brief discussions of
New York, 294-347.
some of the relevant factors are to be found in Chappell Donovan, D. T. and Jones, E. 1. W. (1979) Causes of world-wide
(1983), Norman and Chase (1986) and Summerfield (1985). changes in sea level. Journal of the Geological Society London
Coastal effects have been more fully considered, as is 136, 187-92.
 Sea-level change 455

Guilcher, A. (1988) Coral Reef Geomorphology. Wiley, Chichester Global View. Croom Helm, London and New York, 264--93.
and New York. Pitman, W.e. III, (1978) Relationship between eustacy and strati-
Hallam, A. (1984) Pre-Quaternary sea level changes. Annual Review graphic sequences of passive margins. Geological Society of
of Earth and Planetary Sciences 12, 205--43. America Bulletin 89, 1389--403.
Haq, B. U., Hardenbol, J. and Vail, P. R. (1987) Chronology of Pitman,W.e. III, and Golovchenko, X. (1983) The effect of sea-
fluctuating sea-levels since the Triassic. Science 235, 1156-66. level change on the shelfedge and slope of passive margins.
Harrison, C. G. A., Brass, G. W., Saltzman, E., Sloan, 1. II, Southam, Society of Economic Paleontologists and Mineralogists Special
1. and Whitman, 1. M. (1981) Sea level variations, global sedi- Publication 33, 41-58.
mentation rates and the hypsographic curve. Earth and Planetary Sahagian, D. (1987) Epeirogeny and eustatic sea-level changes as
Science Letters 54, 1-16. inferred from Cretaceous shoreline deposits: applications to the
Hsii, K. J. (1972) When the Mediterranean dried up. Scientific central and western United States. Journal of Geophysical
American 277(6), 27-36. Research 92, 4895-904.
Hubbard, R J. (1988) Age and significance of sequence boundaries Schlanger, S. 0., Jenkyns, H. e. and Premoli-Silva, I. (1981) Vol-
on Jurassic and Early Cretaceous rifted continental margins. canism and vertical tectonics in the Pacific Basin related to
American Association of Petroleum Geologists Bulletin 72, 49- global Cretaceous transgressions. Earth and Planetary Science
72. Letters 52, 435--49.
Kominz, M. A. (1984) Oceanic ridge volumes and sea-level change Schlee, 1. S. (ed.) (1984) Interregional Unconformities and Hydro-
- an error analysis. In: J. S. Schlee (ed.) Interregional Uncon- carbon Accumulation. American Association of Petroleum
formities and Hydrocarbon Accumulation. American Association Geologists Memoir 36. Tulsa.
of Petroleum Geologists Memoir 36. Tulsa, 109-27. Scott, G. A. J. and Rotondo, G. M. (1983) A model to explain the
Matthews, R. K. (1984) Oxygen isotope record of ice-volume his- differences between Pacific Plate island-atoll types. Coral Reefs
tory: 100 million years of glacio-eustatic sea-level fluctuation. 1,139-50.
In: 1. S. Schlee (ed.) Interregional Unconformities and Hydro- Shackleton, N. J Opdyke, N. D (1973) Oxygen isotope and paleo-
carbon Accumulation. American Association of Petroleum magnetic stratigraphy of Equatorial Pacific core V28-238: Oxygen
Geologists Memoir 36. Tulsa, 97-107. isotope temperatures and ice volumes on a 105 to 106 year
McNutt, M. and Menard, H. W. (1978) Lithospheric flexure and scale. Quaternary Research 3, 39-55.
uplifted atolls. Journal of Geophysical Research 83,1206-12. Smith, D. E. and Dawson, A. G. (eds) (1983) Shorelines and
Menard, H. W. (1986) Islands. Scientific American Books, Isostasy. Institute of British Geographers Special Publication
New York. 16. Academic Press, London and New York.
Miall, A. D. (1984) Principles of Sedimentary Basin Analysis. Spencer, T., Stoddart, D. Rand Woodroffe, e. D. (1987). Island
Springer-Verlag, New York. uplift and lithospheric flexure: observations and cautions from
Miall, A. D. (1986) Eustatic sea level changes interpreted from the South Pacific. Zeitschrift fur Geomorphologie Supplement-
seismic stratigraphy: a critique of the methodology with partic- band 63,87-102.
ular reference to the North Sea Jurassic record. American Asso- Summerfield, M. A. (1985) Plate tectonics and landscape devel-
ciation of Petroleum Geologists Bulletin 70, 131-7. opment on the African continent. In: M. Morisawa and J. T.
Marner, N.-A. (1981). Revolution in Cretaceous sea level Hack (eds) Tectonic Geomorphology. Allen and Unwin, Boston
analysis. Geology 9, 344--6. and London, 27-51.
Marner, N.-A. (1987a). Quaternary sea-level changes: northern Sutherland, D. G. (1983) The dating of former shorelines. In:
hemisphere data. In: R 1. N. Devoy (ed.) Sea Surface Studies: D. E. Smith and A .G. Dawson (eds) Shorelines and Isostasy.
A Global View. Croom Helm, London and New York, 242-63. Institute of British Geographers Special Publication 16.
Marner, N.-A. (1987b) Models of global sea-level changes. In: Academic Press, London and New York, 129-57.
M. J. Tooley and I. Shennan (eds) Sea-Level Changes. Institute Sutherland, D. G. (1987) Dating and associated methodological
of British Geographers Special Publication 20. Blackwell, problems in the study of Quaternary sea-level changes. In:
Oxford and New York, 332-55. R J. N. Devoy (ed.) Sea Surface Studies: A Global View.
Norman, S. E. and Chase, e. G. (1986) Uplift of the shores of the Croom Helm, London and New York, 165-97.
western Mediterranean due to Messinian desiccation and flexural Tooley, M. J. and Shennan, I. (eds) (1987) Sea-Level Changes.
isostasy. Nature 322, 450-1. Institute of British Geographers Special Publication 20. Blackwell,
Nunn, P. D. (1987) Small islands and geomorphology: review and Oxford and New York.
prospect in the context of historical geomorphology. Institute of Vail, P. R, Mitchum, R M., Jr and Thompson, S. III (1977a)
British Geographers Transactions, N S 12, 227-39. Seismic stratigraphy and global changes of sea-level, part 3:
Orford, J. (1987) Coastal processes: the coastal response to sea- relative changes of sea level from coastal onlap. In: e. E. Payton
level variation. In: R J. N. Devoy (ed.) Sea Surface Studies: A (ed.) Seismic Stratigraphy - Applications to Hydrocarbon Ex-
Global View. Croom Helm, London and New York, 415-63. ploration. American Association of Petroleum Geologists Memoir
Payton, e. E. (ed.) (1977) Seismic Stratigraphy - Applications to 26, Tulsa, 63-81.
Hydrocarbon Exploration. American Association of Petroleum Vail, P. R, Mitchum, R M. Jr and Thompson, S. III (1977b)
Geologists Memoir 26. Tulsa. Seismic stratigraphy and global changes of sea-level, part 4:
Peltier, W. R (1982) Dynamics of the ice age Earth. Advances in global cycles of relative changes of sea-level. In: e. E. Payton
Geophysics 24, 1-146. (ed.) Seismic Stratigraphy - Applications to Hydrocarbon
Peltier, W. R (1987) Mechanisms of relative sea-level change and Exploration. American Association of Petroleum Geologists
the geophysical responses to ice-water loading. In: R J. N. Memoir 26. Tulsa, 83-97.
Devoy (ed.) Sea Surface Studies: A Global View, Croom Helm, Watts, A. B. (1982) Tectonic subsidence, flexure and global changes
London and New York. 57-94. of sea-level. Nature 297, 469-74.
Pillans, B. (1987) Quaternary sea-level changes: southern hemi- Wyatt, A. R. (1986) Post-Triassic continental hypsometry and sea-
sphere data. In: R. J. N. Devoy (ed.) Sea Surface Studies: A level. Journal of the Geological Society London 143, 907-10.
 18
Long-term landscape
development

18.1 Models of landscape evolution 18.1.1 The Davisian cycle of erosion: peneplanation

One of the most obvious questions we can ask about land- The model of landscape evolution usually known as the
scapes is how they came to be as they are. Indeed, the cycle of erosion was developed by W. M. Davis between
historical approach to landform analysis was the dominant 1884 and 1899 and owed much to the evolutionary thinking
perspective until the 1960s. Over the past two decades, that had permeated both the natural and social sciences in
however, the other obvious question - what are the processes Britain and North America during the latter half of the
operating in the landscape today and how do they relate to nineteenth century. Davis considered that in a similar way
the landforms we see - has become pre-eminent to the extent to life forms, landforms could be effectively analyzed in terms
that studies of landscape development through time have of their evolution. He regarded landscapes as evolving through
been rather neglected. The detailed work on surface pro- a progressive sequence of stages, each exhibiting charac-
'cesses over the past two decades has significantly increased teristic landforms. In his view these sequential changes in
our understanding of the relationships between process and form through time made it possible to infer the temporal
form at the small scale and over short periods of time. But the stage of development of a landscape from its form alone.
gap between' our understanding of landform genesis at this A second key concept implicit in the cycle of erosion
scale and our knowledge of how whole landscapes function model (although not explicitly referred to by Davis) is that
at long time scales has been widely acknowledged. A much of thermodynamics. The development of the principles of
better appreciation of the role of tectonic and climatic con- thermodynamics had been a major achievement of nineteenth-
trols over landscape development, coupled with the applica- century science with repercussions just as profound as
tion of new dating techniques and major theoretical advances, those of evolution. This aspect of the cycle of erosion model
is beginning, once again, to bring the problem of long-term has been highlighted only relatively recently and relates
landscape development centre stage. This chapter draws closely to systems analysis in geomorphology. The second
extensively on topics introduced earlier in the book and law of thermodynamics states that in an isolated system
tries to show how new concepts and data are beginning to (that is, one which cannot give off or receive either mass or
shed a fresh light on some long-standing problems in energy) entropy can never increase. The concept of entropy
geomorphology. has been applied in many different contexts, but a general
Geomorphology has seen various attempts to systematize definition is that it is a measure of the energy in a system
the development of landscapes through time by isolating which is unable to perform work. In a system in a state of
the key factors which determine the way in which landforms low entropy there are large differences in the distribution of
evolve. These models of landscape evolution, the most in- energy and the flows from areas of high to areas of low
fluential of which have been those proposed by W. M. ,energy allow work to be performed. Conversely, in a high
Davis, W. Penck, L. C. King and J. Biidel, have had a pro- entropy system energy is much more evenly distributed and
found effect on the kinds of problems that geomorpholo- the flow of energy and the performance of work is corres-
gists have considered and the ways they have attempted to pondingly reduced. At the theoretical point of maximum
tackle them.
458 Endogenic-exogenic interactions

entropy the distribution of energy is entirely uniform and tion of the anticipated sequence of forms (Fig. 18.2) was
no work can be done in the system. based on several important assumptions: that denudation
In strict terms, landscapes are neither isolated nor even occurs under a humid temperate climate (which Davi
closed systems since they are constantly importing and regarded as 'normal') on a uniform lithology, and that the
exporting mass and energy. Nevertheless, it has been argued cycle is initiated by the relatively brief and rapid uplift of a
that as the potential energy created by the uplift of a land- landsurface of minimal local relief which does not exper-
surface is the major source of energy in the landscape system ience significant erosion during the uplift phase. Given
it is a justifiable simplification to regard landscapes as if these conditions, he described a series of stages in the cycle
they are isolated systems. Given this assumption, the cycle of erosion categorized by way of analogy to the stages of
of erosion can be seen as representing a progressive decline human life as youth, maturity and old age. Davis argued
in potential energy and increase in entropy as the landscape that there would be a progressive decline in slope angles
is eroded. Indeed Davis saw the slope angles and stream (Fig. 7.25(A)) and stream gradients through time which
gradients at any particular point in the landscape as reflect- would ultimately result in the production of a landsurface
ing the distribution of potential energy expressed as local close to base level with very subdued relief. Such a surface
differences in elevation. The total potential energy of the he termed a peneplain and consequently Davis's model of
landscape, and hence its stage of evolution at any given time, landscape development characterized by declining surface
could be expressed in terms of its mean elevation above gradients through time is often referred to as peneplana-
base level (Fig. l8.l(A)). tion. It is important to note, however, that the term pene-
Although Davis acknowledged numerous complications plain is used by some writers much more broadly to refer to
that could affect the cycle of erosion, his detailed descrip- any low relief surface however formed. The alternative

UPLIFT YOUTH MATURITY OLD AGE A

Q) -----+
'0 Altitude of main Altitude of highest
;:, valley floors
+" divides
.~

<i: PENEPLAIN

Time
-- -- ------ - - - - - - - - - --.

WAXING WANING
DEVELOPMENT DEVELOPMENT
. . Altitude of B
~
Altitude of main highest divides
valley floors /

- - - - - __
--- ---- ENDRUMPF

~~~~I'~:::::::::::::::::t.d!..=.::::::"'=-
--- -- --
-=-':=-=::::::::-=-=--=-:":-~-~-

Time

UPLIFT YOUTH MATURITY OLD AGE

c
Altitude of main Altitude of
valley floors highest divides

---------~----------

Time
Fig.18.1 Schematic representation of the key elements of the models of landscape evolution proposed by Davis (A), Penck (B) and
King (C). Note that for simplicity base level is assumed to be fixed through time and that the temporal scale is not necessarily
comparable between diagrams. In the Davisian scheme the stage of old age should be regarded as many times longer than youth and
maturity. (Modifiedfrom J. B. Thomes and D. Brunsden, (1977) Geomorphology and Time. Methuen, London, Fig. 6.2,
p.122.)
 Long-term landscape development 459

Fig.18.2 The Davisian cycle of erosion under a humid climate. The assumed starting-point is a landsurface with little local relief,
either a peneplain developed during the previous cycle as shown in (A) or an emerged submarine surface. Uplift leads to rapid incision
of the landsurface by rivers. In early youth (B) narrow river valleys separate broad areas of largely uneroded uplands and river
gradients are irregular with waterfalls, rapids and lakes formed in response to lithological variations. These channel gradient
irregularities have been eliminated by the end of middle youth, and by the end of late youth (C) major rivers are graded and lateral
erosion enables the development of narrow floodplains in their lower courses. The flat uplands which have been steadily reduced in
area during youth as the drainage network has grown are eliminated altogether by the beginning of maturity (D). This is the stage of
maximum local relief and the drainage network becomes fully integrated and closely adjusted to structure. Hereafter local relief begins
to decline as the graded river channels, which by this stage have spread far up tributary valleys, are lowered progressively less rapidly
than interfluves. Associated with this change is the reduction in average slope angles, as the steep slopes ofyouth which are close to the
stability angle of the partially weathered debris are transformed into lower gradient slopes as the active basal removal ofdebris ceases.
Throughout maturity, floodplains become gradually wider and major rivers develop meandering channels. By late maturity (E) local
relief has been significantly reduced and the landscape comprises gentle valley-side slopes and extensive floodplains. As old age is
reached (F) the entire landscape is graded and floodplains are several times broader than the active meandering belts within them. The
mean elevation of the landsurface, already close to base level, is lowered further only very gradually. Note, however, that in regions
remote from the coastline to which rivers are flowing the developing peneplain will remain well above base level since river channels
must have a certain minimum gradient in order to transmit water. Low rates of erosion allow the accumulation of thick weathering
mantles which, in progressively masking the underlying bedrock, gradually free river channels from structural controls. None the less,
particularly resistant lithologies may allow erosional residuals, known as monadnocks, to survive into late old age. Finally, renewed
uplift will initiate a new cycle of erosion (G). (From A. N. Strahler (1969) Physical Geography (3rd edn.) Wiley, New York, Fig. 27.1, p.
466, drawn by E. Raisz.)

spelling 'peneplane' which is used occasionally is certainly ment of youthful forms which would coexist with older
misleading as Davis in no sense envisaged the development .forms and thereby create a polycyclic landscape. The simplest
of a planar surface as the ultimate product of the cycle of assumption was that such uplift would only manifest itself
erosion. as a fall in base level at the downstream extremity of drain-
As we have mentioned, Davis acknowledged the presence age basins (normally at the coastline), and would lead to
of factors that might complicate the stately progression of the gradual encroachment of steeper river gradients and
landscapes illustrated in Figure 18.2. The cycle might be slopes upstream through the drainage systems of the up-
interrupted by renewed uplift at any stage which would lifted landscape (Fig. 18.2(G)).
cause rejuvenation of the landscape through the develop- A second complication Davis noted was climate. Davis
460 Endogenic-exogenic interactions

effectively represented land cape development under the movements characteristic of plate interiors usually involve
humid temperate morphoclimatic regime of his home area slow, but prolonged surface uplift, while the high crustal
of the north-east USA as 'normal'. But he accepted th'at uplift rates characteristic of convergent and oblique- lip plate
the detailed nature of landform evolution under different margins persist for as long as the plate interaction giving
prevailing climate would not be identical because of varia- rise to them are sustained.
tions in the intensity of geomorphic processes under differ- Another major criticism of the Davisian model arises from
ent morphoclimatic regimes. Consequently, he developed its inability to accommodate the frequent and rapid climatic
'arid' and 'glacial' versions of the cycle of erosion while changes that have characterized the Quaternary. These have
later disciples of his evolutionary approach added further been of world-wide extent and, in conjunction with the
variants. frequent major changes in base level with which they have
A third complication was provided by lithology and struc- been associated through their effect on global sea level,
ture which Davis saw as exerting specific controls on land- they make it extremely unlikely that landscapes anywhere
scape evolution largely through their influence on drainage can be realistically viewed as representing a simple uni-
patterns. He maintained, however, that such controls would directional sequence of forms.
become progressively less significant as the cycle of ero-
ion proceeded. In the case of limestone terrains, later workers
18.1.2 The Penck model: uplift and denudation
found it necessary to develop a specific karst cycle of
erosion. Nevertheless, in spite of these complications, Davis related
maintained the value of regarding landscapes primarily in As has already been pointed out, the Davisian model never
terms of their evolutionary stage in a unidirectional tem- gained universal support, and geomorphologists on the
poral sequence. continent of Europe found its assumptions - especially those
Although his cyclic scheme never gained wide acceptance concerning the nature of uplift - drastically over-simplified.
on the continent of Europe, it dominated Anglo-American In spite of these criticisms the only coherent alternative
geomorphology for several decades. Since the 1950s, how- scheme of landform development to emerge prior to the
ever, both the theoretical utility and the empirical validity Second World War was that of W. Penck. Penck's ideas
of the cycle of erosion have been increasingly challenged. have never been popular among English-speaking geomor-
What, then, are the major criticisms of the model? Al- phologists both because of his rather obscure writing style
though contemporary critics have tended to focus on the and terminology, and because the majority of geomorpho-
rather vague understanding of surface processes evident in logists unable to read German had to rely for several decades
Davis's formulation of landform development in general, on misleading representations of his views by Davis and
and slope development in particular, perhaps the most other writers.
fundamental problem with the cycle of erosion arises from Penck's ideas on uplift differed significantly from those
the assumptions concerning the rates and occurrence of of Davis (Fig. 18.1(B». Whereas the latter assumed brief
uplift. episodes of rapid uplift punctuating prolonged periods of
Presumably due to the lack of quantitative data when he stability, Penck argued that, in orogenic belts at least, active
was writing, Davis was never very specific about actual uplift could continue for a considerable time and in such
rates of uplift and denudation. The estimates he did give, situations Davis's notion of evolutionary stages of land-
such as the 20-200 Ma for the peneplanation of the fault- scape development would be of dubious value. On the basis
block mountains of Utah, indicates that he envi aged exten- of the evidence from sedimentary sequences flanking the
sive time scales. Our current knowledge of uplift rates (see Alps, Penck considered that rates of active uplift initially
Chapter 15) suggests that few areas of the world remain increased slowly before reaching a maximum and then
stable for periods of tens of millions of years or more, and declining gradually.
therefore it seems that polycyclic landscapes are likely to In certain circumstances Penck thought that periods of
be the norm rather than the exception. Furthermore, iso- increasing and decreasing rates of uplift might be reflected
static uplift is an inevitable consequence of denudational in slope forms. This link could arise from the effect
unloading as the cycle runs its course. As a result, con- changing rates of crustal uplift could have on rates of river
tinuous crustal uplift, albeit at a declining rate through time, incision. High rates of crustal uplift, Penck argued, would raise
will affect the entire duration of a cycle of ero ion and river channels further above base level and thus increase their
greatly delay the attainment of full peneplanation. As we gradients. This would lead to an. acceleration in river down-
have seen in Chapters 3 and 4, inter-plate and intra-plate cutting until the rate of incision matched the rate of crustal
tectonic mechanisms give rise to quite different temporal uplift. The converse situation would apply during a decline in
and spatial patterns of uplift, and in neither case does the the rate of crustal uplift, with rates of river incision decreasing
elevation of the landsurface take the form of geologically as downcutting reduced channel gradients. Penck considered
brief, discrete episodes of rapid surface uplift. Epeirogenic that a uniform rate of river incision would give rise to
 Long-term landscape development 461

straight slopes which would retreat at a constant angle. If weathering, both of which can significantly affect relation-
the rate of downcutting were to increase, however, a phase ships between stream activity and slope form.
of waxing development would ensue and slopes would
steepen progressively from the base upwards to produce a
18.1.3 The King model: pediplanation
convex profile. Conversely, a decrease in river downcutting
could create a phase of waning development and slopes L. C. King's model of landscape evolution resembles Davis's
could become progressively less steep from the base in assuming that uplift is episodic and rapid in comparison
upwards, creating concave profiles. with rates of denudation, and that the overall morphology
Penck's model of landscape evolution can thus be sum- of a landscape at any point in time is diagnostic of its evo-
marized as follows. An initial gradual increase in the rate of lutionary stage of development (Fig. 18.I(C». The essen-
crustal uplift of a primary surface (Primarrumpf) leads to tial, and significant, difference in King's scheme lies in the
the widespread development of convex slopes. Further mode of slope development he proposed. King initially
acceleration in the rate of uplift results in the formation of a developed his model to account for the landscapes of
series of benches (Piedmottreppen) around the margins of southern Africa. These are characterized by extensive, gently
the primary uplifted surface. As the rate of uplift begins to inclined surfaces dotted with inselbergs and separated by
decline there is a transition from waxing development, escarpments, and have developed under predominantly arid
characterized by rapid downcutting, to waning development to tropical wet-dry morphoclimatic regimes. King's notion
where the rate of stream incision is reduced and valley of slope development appears to owe much to Davis's
widening through the parallel retreat of individual slope misrepresentation of Penck's ideas (compare Figure 7.25(B),
elements gradually becomes dominant (Fig. 7.25(C». As (D». Rather than the sequential replacement of parallel
noted in Chapter 7 (see Section 7.6.2), this form of slope retreating slope segments by lower angle elements, King
evolution is perhaps best described as slope replacement to envisaged the parallel retreat of a single free-:face slope unit
distinguish it from the version of whole-slope parallel retreat leaving a broad, concave pediment sloping at an angle of
advocated by King and misattributed by Davis to Penck 6-7 0 or less at its base. Gradually over time, pediments
(Fig. 7.25(B». The steepest slop~ elements forming free coalesce to form pediplains and this mode of landscape
faces retreat most rapidly leaving behind basal series of development is therefore called pediplanation.
lower angle debris slope segments. The retreat of free faces King considered that once pediment surfaces have been
eventually leads to the formation at drainage divides of formed they persist with little change until the next phase
large residual hills, or iriselbergs, which are flanked by of surface uplift promotes a new cycle of river incision and
pediments. The eventual elimination of inselbergs leaves a escarpment retreat which consumes existing pediplains and
landscape termed by Penck an endrumpf consisting entire- creates new ones. As in the Davisian model, the dating of
ly of slowly retreating, low angle concave slopes. such denudational episodes can be described in terms of the
Although Penck's emphasis on the response of drainage timing of the fall in base level initiating each new land-
systems to changing rates of uplift provides useful pointers scape cycle. None the less, the landsurface itself is diachro-
as to how we might attempt to integrate tectonics into nous because in King's model landscapes essentially develop
models of long-term landform development, his scheme as .through backwearing as escarpments experience parallel
a whole is untenable as it pays insufficient attention to other retreat; landsurfaces, therefore, are progressively older away
factors affecting landform development. In particular it from escarpments (Fig. 18.3). Consequently, it is possible
fails to acknowledge the importance of changes in river to talk of the local age of a landsurface, and even to refer to a
discharge which might arise as a result of climatic change, terminal age determined by the final removal of a pediplain
and it also underplays the role of lithology and the nature of remnant.

Terminal age
of 0
Local age I
of C ,

~--
sea level
Initial age
of A

\.,/ - ,--

B
- ~ c D c
A +
Local age
of A

Fig.18.3 Different criteria for defining the ages of erosion surfaces according to the model of landscape development proposed by
L. C. King. The surfaces labelled A-D were initiated during three episodes of base level fall. Each is diachronous and deposits on the
surface may be capable ofyielding a minimum local age at that point. The final elimination of the last remnant of a particular surface
(D) gives its terminal age.
462 Endogenic-exogenic interactions

King originally envisaged episodes of uplift occurring lATERITIZED ETCHPLAIN A

predominantly along continental margins as the result of a
delayed isostatic response to denudational unloading. He
considered that such isostatic uplift would only take effect
once an escarpment had retreated over a critical threshold
distance. This mechanism of discontinuous isostatic uplift
is based on a misunderstanding by King (and other geomor-
phologists since) of how the crust responds to changes in
load. The response is, of course, continuous on the geo-
logical time scale of denudational cycles (see Section 2.2.4),
although flexural isostasy does provide a possible means of
generating surface uplift along passive margins experien-
cing escarpment retreat (see Section 4.2.3). Subsequently,
King advocated the somewhat ill-defined mechanism of
cymatogeny, involving the upwarping and flexure of con-
PARTIALLY STRIPPED ETCH PLAIN C
tinental margins as a result of active subcrustal processes,
as the means by which new cycles of pediplanation could
be initiated. Although modem concepts of passive margin
tectonics have now replaced King's ideas about uplift
mechanisms, the notion of widespread and long-term escarp-
ment retreat has gained new impetus from recent attempts DOMINANTLY STRIPPED ETCHPLAIN / ETCHSURFACE 0
to understand the evolving morphology of passive margins
following continental break-up.

- .... -_.
18.1.4 The Budel model: etchplanation ---.....-..- -.;;_-- .-.::I

Although having little impact on the development of Anglo-

American geomorpholog)', the ideas of 1. Rudel have exerted INCISED ETCHSURFACE E
a considerable influence on workers in the continent of
Europe, especially in Germany. Rudel's key notion con-
cerning landscape development is that of a 'double surface
of levelling'. In regions covered with thick weathering
deposits, especially the relatively stable shield areas of the
humid tropics, denudation of the landscape occurs simul-
taneou,sly through the removal of material from the surface

mDl
lNDURATED
- largely, it is thought, by sheet wash - and by ongoing BEDROCK HORIZONS
chemical decomposition at the weathering front. This com- (especially ferricrete)
bination of deep weathering and surface removal produces
an etchplain (or an etchsurface where an uneven basal I: : : : : :: : : :1
WEATHERING
surface has been exposed) and the overall process is termed : : : : :: : : :: MANTLE
etch planation (Fig. 18.4). Elements of this model of land-
scape evolution can be traced back to the British geologist, Fig. 18.4 The development of different types of etchplains and
E. J. Wayland, who worked in Uganda in East Africa in the etchsUljaces. The diagrams do not necessarily represent an
1930s, but it is Budel who has developed the concept of evolutionary sequence as repeated episodes of accelerated
etchplanation most fully. erosion may only succeed in partially removing the weathering
mantle. The types of etchplains and etchsUljaces illustrated are:
In the humid tropics an important element in landscape
(A) lateritized etchplains comprising a sUljace of low local relief
development is the spatial variability of the factors which underlain by a thick weathering mantle, including indurated
determine weathering rates, especially lithology and drain- lateritic horizons (ferricretes), which has been subject to only
age. As a result of variations in these controls the form of limited stream incision; (B) dissected etchplains in which
the weathering front is highly irregular, and the depth of the accelerated stream downcutting promoted by climatic change
weathering mantle does not necessariiy bear any relation- or uplift leads to the development of well-defined valleys,
fringed in places by duricrust breakaways, and the very
ship to the form of the ground surface. During periods of localized exposure of bedrock and the formation of tors; (C)
tectonic and climatic stability rates of weathering and denu- partially stripped etchplains characterized by widespread stream
dation are roughly in balance and the depth of the weather- dissection and the extensive stripping of the weathering mantle,
 Long-term landscape development 463

including resistant duricrust layers, to reveal numerous rock out that this could not occur in any strict sense over large
outcrops in the form of tors; (D) dominantly stripped etchplains horizontal distances because of inevitable variations in rock
representing a very advanced stage of stripping in which the
strength associated with what might be quite subtle changes
weathering mantle is retained only in deep pockets along the
weathering front and where some of the exposed bedrock has also in lithology. While clearly accepting this point, it is useful
been subject to erosion (forming an etchsUlface where significant to retain the notion of parallel retreat in a broad sense in
relief is present); (E) incised etchsUlfaces in which the basal order that it can be contrasted with the idea of a progressive
bedrock sUiface has been extensively modified by fluvial erosion, decline in slope gradients in the landscape through time.
almost certainly as a"result of a significant change in base level The distinction between backwearing and downwearing
rather than climatic change. (Diagrams and descriptions of
etchplain and etchsurface types based on M. F. Thomas (1974)
is important because of the different isostatic responses to
Tropical Geomorphology, Macmillan, London, Fig. 41, which we would expect them to give rise. Isostatic compen-
pp.236-8.) sation of a landscape experiencing extensive downwearing
would not, in general, lead to any surface uplift, whereas
ing mantle varies little. But a change in climate, or increase flexural effects along the kind of sharp topographic dis-
in the rate of crustal uplift, can disrupt this steady state by continuity formed by a major escarpment could lead to
generating an increase in the rate of river incision or, in the localized surface uplift (see Section 4.2.3). It is, of course,
case of climatic change alone, through the disturbance of possible, and indeed likely, that both downwearing and
the vegetation cover. During such a perturbation of the backwearing occur simultaneously, although the latter may
geomorphic system the weathering mantle may be partially, be slow with respect to the former. Slow downwearing is,
or even wholly stripped. As the landscape is lowered in in fact, just what is implied by Budel's notion of etch-
response to the more vigorous erosional activity, water planation, and it certainly seems that we cannot assume that
tables will fall. This will tend to increase rates of water once pediments are created by escarpment retreat they
throughput at the weathering front and in tum lead to an necessarily remain immune from the effects of weathering
increase in the rate of weathering. and erosion. Indeed, if we accept the evidence for generally
Etchplains can assume a range of forms depending on a warmer (and probably also wetter) climates in the Cretaceous
number of factors including the lithology and structure of and Early Cenozoic (see Section 18.4.3) then very ancient
the local rocks (which influences the depth of the weather- landsurfaces are unlikely to have remained untouched by
ing mantle), the intensity and duration of erosional episodes episodes of deep weathering even in regions which are now
and the morphology of the basal weathering surface. Various predominantly arid.
types of etchplain can be produced as a result (Fig. 18.4), Although the subject of intense debate during the first
and this has led to some confusion over the application of half of this century, over the past two or three decades there
the term. has been relatively little discussion among geomorpholo-
gists about the relative merits of the classic schemes of
landscape evolution discussed here. Many have regarded
18.1.5 Classic models of landscape evolution:
them as being so deficient in their treatment of exogenic
summary and assessment
geomorphic processes that they are barely worth serious
At this point it is probably useful to summarize very briefly consideration. While accepting the value of the idea of pro-
the essential elements of what might be described as the gressive landscape change through time, others have rejected
classic models of landscape evolution, before seeing in the the specific models of landform change proposed as over-
following sections how the problem of l~~g-term landform simplified and inadequate. Yet other geomorphologists have
development is currently being tackled. The cycles of erosion pointed to the important effects that lithology or changing
envisaged by Davis and King are similar in that they both morphoclimatic regimes have on the way landforms evolve
assume that surface uplift occurs as more or less discrete through time, and have argued that these factors render the
pulses which punctuate the progressive erosional develop- search for an all-embracing model of landscape evolution
ment of the landscape. Penck, on the other hand, explicitly futile. Finally, there has been the idea that in reality land-
incorporated the idea of surface uplift occurring for much scapes do not in fact evolve in any systematic manner but
longer periods of landscape history and playing an integral simply oscillate around an equilibrium form.
role in how the landscape evolves rather than simply pro- Irrespective of the merits of these views, their effect has
viding an initial input of potential energy. But in terms of been to direct attention away from problems of long-term
the changes in form that the landscape experiences through landscape development to the apparently more tractable
time, King is much closer to Penck in proposing that back- questions posed by the nature of shorter-term, and smaller
wearing generally predominates over downwearing (although, scale, geomorphic change. None the less, this situation is
as we have pointed out, Penck's and King's conceptions of now changing as a result of both conceptual and technical
exactly how slopes retreat were different). With particular developments since the mid-1970s. One has been the attempt
respect to King's form of parallel retreat, it has been pointed to integrate a modified version of the concept of dynamic
464 Endogenic-exogenic interactions

equilibrium into the notion of progressive landscape change 18.2.2 The dynamic metastable equilibrium model
embodied in the principle of evolution. Another has been
In Chapter 1 we discussed how the idea of equilibrium in a
the revolution in our knowledge of tectonic processes that
landscape was linked to the temporal and spatial scale being
has occurred over the past two decades and in particular the
considered (see Section 1.3.4). However, we have yet to
way that this has immeasurably improve~ our understand-
consider exactly how a landscape compos~d of individual
ing of the nature and causes of uplift. Finally, there have components in a steady-state equilibrium can experience
been major advances in the dating of geomorphic events progr'essive lowering in the longer term. This problem has
and the ability to estimate long-term rates of denudation
been addressed by S. A. Schumm who has proposed that
(see Section 15.4). It is to these themes that we tum next.
these ideas of landscape stability and landscape change can
be reconciled by incorporating the concept of episodic ero-
sion (see Section 9.5.1) into a decay equilibrium model of
18.2 Landscape stability and change landscape evolution. The key element of Schumm's model
is that valley floors are lowered episodically rather than
18.2.1 The Hack dynamic equilibrium model continuously. This could occur through the accumulation of
sediment from the upper parts of a basin covering the bed-
One reaction to the evolutionary thinking embodied in Davis's
rock of the valley floor. Periodically this sediment is removed
notion of a cycle of erosion was the proposal by 1. T. Hack
and the bedrock of the valley floor is lowered. Such valley
that landscapes could be better understood in terms of
floor incision may be promoted by external factors, such as
'dynamic equilibrium'. In rejecting the idea of progressive
a major flood (see Section 9.5.2), but it might also arise as
change in the form of the landscape through time, Hack
a result of the breaching of a geomorphic threshold due to
resurrected the approach of G. K. Gilbert focusing on the
sediment deposition, causing the channel gradient to reach
continuous adjustment between force and resistance. He
a threshold of instability.
argued that in landscapes that have experienced a long
This situation can be described as one of dynamic meta-
period of denudation there will be a mutual adjustment
stable equilibrium, and differs from the concept of dynamic
between lithological controls and prevailing surface pro-
equilibrium in that the reduction in channel bed elevation
cesses. In the ideal case where base level, surface processes
takes place discontinuously not progressively. Under condi-
and lithology remain constant through time, the form of the
tions of dynamic metastable equilibrium, phases of steady-
landsurface remains unchanged since the whole landscape
state equilibrium are punctuated by adjustments involving
is lowered at a constant rate. Relief, slope angles and stream
the breaching of either extrinsic or geomorphic thresholds
gradients are adjusted in such a way that each unit area
which shift the steady-state equilibrium to a new level (Fig.
yields the same sediment load; regions of resistant rock
18.5). Schumm argues, therefore, that it is possible .to re-
have steep, rugged relief, whereas areas of less resistant
place the Davisian notion of progressive, gradual change
lithologies have subdued relief and gentle slopes. (Note that
through time by a model of landscape evolution in which
in the sense we have already defined the terms (see Section
change occurs in a step-like manner (Fig. 18.6). Interesting-
1.3.4) this model is essentially one of steady-state equili-
brium.) The major shortcoming of this approach as a general
landscape model is that, while this condition of uniform
lowering might apply to particular areas of limited extent, it
cannot apply to entire drainage basins in the long term.
This is because the lowering of the surface of a drainage
basin towards base level necessarily involves a reduction in 1
the gradient of trunk streams, and this will eventually affect
tributary basins. A further problem with Hack's dynamic
c:
equilibrium concept is that climatic change and tectonic o
activity are likely to lead to changes in the nature and rates .....
l'O
of processes through time, while progressive surface lower- > Q.l
ing will expose different lithologies. The maintenance of a UJ
'dynamic equilibrium; assumes a rapid adjustment to such
changes but there is abundant evidence in some landscapes
of the survival of relict landforms. It appears that the con-
cept of 'dynamic equilibrium' is likely to be most applicable
to parts of slowly eroding landscapes which have not Time
experienced major climatic shifts and which are effectively Fig. 18.5 Schematic representation of dynamic metastable
isolated from base level changes. equilibrium. Compare with Figure 1.9(C).
 Long-term landscape development 465

(/)

C1l

"
>
0
Fig.18.7 Davis's representation of the cycle of erosion
"c: incorporating the idea of valley aggradation (shown by the
CQ dashed line C-E-G) during drainage basin evolution (From
W. M. Davis, (1899) Geographical Joumal14, Fig.], p. 486.)
(/)

0
0 magnitude and frequency of external changes that they
LL
experience and in their ability to adjust to such changes. A
core concept in geomorphology is that of characteristic form;
>- distinct landform environments exist which are determined
C1l
by a set of external variables including rock type and struc-
CQ ture, climate, biological activity, and rates of tectonic activity
> and base level change . These variables interact over time
and space, giving rise to a characteristic magnitude and fre-
0 quency of landforming events and producing a uniform and
repetitive assemblage of landforms reflecting a strong inter-
c:
0
dependence between process and form which persists for as
c long as the controlling conditions remain constant.
CQ These characteristic forms are, in reality, subject to signi-
> ficant changes in external conditions, either in the form of
C1l (f)(f)
-i~
0
m
environmental change or more transient high magnitude -
(f)
-0
LU ~l> 0
(f) -i
m low frequency events, as well as inherent internal instabili-
m~ (f) ~ l>
=i -i ties at a variety of time scales. In fact it has been argued
0
m ~
Z that the conditions controlling landscape development rarely
remain constant sufficiently long for characteristic forms to
Time
develop and that in such cases transient forms dominate
Fig.18.6 Modifledform of the cycle of erosion incorporating the landscape. This is particularly likely to be true in environ-
the idea of episodic erosion and dynamic metastable equilibrium: ments where rare, large magnitude events can generate
(A) cycle of erosion as envisaged by Davis; (B) discontinuous substantial long-lived changes in the landscape. Viewed in
reduction in elevation of valley floors as a result of episodic
erosion; (C) close-up of (B) showing periods ofsteady-state this way we can divide geomorphic time into three cate-
equilibrium interrupted by episodes of rapid erosion followed by gories: lag time, which represents the time taken for the
deposition. (Based on S.A. Schumm (1975) in W. N. Melhorn and system to begin reacting to change; relaxation time, which
R. C. Flemal (eds) Theories of Landfonn Development. Allen and designates the time taken to achieve a characteristic form;
Unwin, London, Fig. 5, p. 77.) and characteristic form time which indicates the time
over which characteristic forms persist (Fig. 18.8). The cen-
ly, Davis himself appears to have been aware that the reduc- tral, and as yet largely unanswered, question is what is the
tion of elevation of valley floors was not necessarily a pro- relative importance of characteristic and transient forms in
gressive and continuous process. In his 1899 presentation the landscape?
of the cycle of erosion he pointed out that at some stage Clearly, a crucial factor in deciding this question is the
after the initiation of a cycle trunk rivers might not be able sensitivity of the landscape to change. Landscape sensiti-
to cope with the amount of sediment being generated by vity is defined as the propensity for the landscape to under-
actively downcutting upper tributary streams and would go a recognizable change in response to changes in the
adjust by aggrading. He even provided an illustration of external variables controlling the geomorphic system. Whether
this effect (Fig. 18.7) which in some respects anticipates a landscape experiences a change will depend on both its
the concept of episodic erosion. sensitivity (an inherent property of the landscape sy tern)
and the magnitude, frequency and duration of changes in
the controlling variables. Traditionally, the resistance of a
18.2.3 Landscape sensitivity
landscape to change has been seen largely in terms of either
Another necessary modification of our understanding of rock resistance (as reflected in rock strength, resistance to
landscape development is that landscapes vary in both the ~eathering and erodibility) or morphological resistance (in
466 Endogenic-exogenic interactions

CHARACTERISTIC
Table 18.1 Characteristics associated with sensitive and insensitive
~RELAXATION TIME....... FORM TIME - . landscapes

CHARACTERISTIC SENSITIVE INSENSITIVE

LANDSCAPES LANDSCAPES
Q)
I I
OJ Nature of geomorphic system Fast-responding Slow-responding
c I I
ltl
I I Local relief High Low
~
u I I Denudation rates High Low
I I
Response to climatic change Rapid Slow
() I I
~ I I I I I Relationship of landforms to Strong Weak
Co
l- +LAG+- RELAXATION .. +LAG+ contemporary morphoclimatic
e regime
E Landform persistence Short Long
e
Q)
Magnitude of geomorphic Small Large
<.::l
1
IMPULSE
1
IMPULSE
event required to overcome
barriers to change
OF CHANGE OF CHANGE Relative importance of high Low High
magnitude/low frequency
Time events
Abundance of characteristic High Low
Fig.18.8 Schematic illustration of the concepts of lag time,
forms
relaxation time and characteristic form time. (After D. Brunsden,
(1980) Zeitschrift fUr Geomorphologie Supplementband 36, Abundance of transient forms Low High
Fig. 1,p. 18.)
their long-term development. Insensitive landscapes will be
the sense that low gradient slopes and regions of low local
characterized by a persistence of landform patterns and of
relief will be less responsive to external changes).
structural and lithological controls. The landscape will not
Although these elements of landscape sensitivity are
reflect landform development under anyone morphoclimatic
important it is now realized that other factors are also
regime; rather it will be a palimpsest of geomorphic systems
significant. These include channel gradient, drainage den-
consisting of a series of forms partially developed or modi-
sity and the strength of the linkages between the various
fied under previous morphoclimatic conditions. Characteristic
components of the landscape system. Where channel gradients
forms will be rare or absent; transient forms will predomi-
are steep and drainage density is high, changes (in, for
nate. Sensitive landscape systems, on the other hand, will
instance, base level) can propagate rapidly throughout a
display a much closer correspondence with a single mor-
drainage basin. In headwater areas of drainage basins there
phoclimatic regime and lithological and structural controls
is a strong coupling between channel systems and slope
will be less marked. Characteristic forms are much more
systems, and consequently a high landscape sensitivity,
likely to be present but transient forms may also be abun-
whereas downstream floodplain development means that
dant, especially where changes in external controlling vari-
river channels are only weakly linked with valley-side slopes
ables are frequent.
and interfluves, thus making these landscape components
less sensitive to change. Landscape sensitivity can also be
18.2.4 Time-dependent and time-independent
affected by the sequence of morphoclimatic regimes pro-
landforms
duced by climatic change. Characteristic forms generated
under one morphoclimatic regime may be of too low a The concept of landscape sensitivity can be linked to ideas
gradient to be significantly altered under the following about the role of temporal and spatial scale in understanding
regime. landform genesis to provide us with a very promising
We can now identify two end-members of landscape sen- framework within which to address the problem of whether
sitivity: highly sensitive, fast-responding, mobile geomorphic landscapes should be viewed as evolving or in a steady
systems which adjust rapidly to change, and insensitive, state. At one extreme the form of a landscape can be viewed
slow-responding systems which are highly resistant to change. solely in terms of its stage of development and thus regard-
The characteristics of these two extreme types of geomor- ed as time-dependent; at the other a landscape can be
phic system are outlined in Table 18.1. Within any land- analyzed solely in terms of the interaction between the
scape, sensitivity will inevitably vary since certain landscape materials of which it is composed and the processes acting
elements, such as river channels and soil systems, will always upon them - the time-independent approach. Some re-
be more responsive than others, such as plateaus and inter- searchers regard Hack's dynamic equilibrium model to be
fluves. Nevertheless, the gross contrasts between different time-independent, but this is not the case for landscapes as
landscapes are of critical importance in the interpretation of a whole. Hack suggested that landforms would retain a
 Long-term landscape development 467

constant form over time, but since he also envisaged Mountain development
downwearing through time the landscape overall must
gradually be reduced in elevation and is therefore ultimately
time-dependent.
..in space and time

Whether a time-dependent or t,ime-independent approach SUi'face

=tZl,-----
is an appropriate basis for landscape analysis will depend
TIME DEPENDENT
very much on the temporal and spatial scale of our study.
LANDFORM EVOLUTION
But it also depends on the sensitivity of the landscape since,
for a given magnitude and frequency of perturbations in
external controls (such as climatic change and tectonic acti-
vity), a sensitive landscape will adjust much more rapidly
than an insensitive landscape and will therefore spend a
larger proportion of its time in equilibrium with prevailing
environmental conditions.
We would expect the most extensive development of time-
independent landscapes where local relief and denudation
rates are high and where there is a strong linkage between
landscape components, so that adjustments propagating
through the drainage system can extend rapidly through
slope systems and up to interfluves. Active orogenic belts
characterized by extremely high rates of crustal uplift and
denudation seem to be the only regions where these require-
ments are likely to be attained over a landscape as a whole.
One of the best examples is probably the Southern Alps
in New Zealand. In Section 15.8 we noted the relative rates
of uplift and denudation in these mountains and found them

--
to be comparable. The topography of South Island, New
Zealand, shows a transition from an incised erosion surface
in the east to sharp-peaked mountains in the west, reflecting
an increasing rate of crustal uplift and denudation as the
plate boundary represented by the Alpine Fault is approached
(Fig. 3.29). As the erosion surface has moved towards the Fig.18.9 Schematic representation of the transition from time-
plate boundary it has been uplifted, and the resulting increase dependent to time-independent landforms in the Southern Alps,
in channel gradients has led to river incision and the wide- New Zealand. The actual transition occurs between stages 4 and
5 when the flat-topped interfluves are finally eliminated as a
ning of river valleys with a corresponding narrowing of result of valley deepening and widening. (Based partly on 1.
plateau summits (Fig. 18.9). At an altitude of around 2000 m Adams, (1985) in: M. Morisawa and 1. T. Hack (eds) Tectonic
(at which point the crustal uplift rate is around 500 mm ka- 1) Geomorphology. Allen and Unwin, Boston, Fig. 4, p. 117.)
the original erosion surface has been completely removed and
above this height interfluves consist of sharp ridges. Up to
this point the morphology of the landscape is time-dependent the river channels. The highest parts of the Southern Alps
and we can plot an evolutionary sequence of forms. Surface can therefore be regarded as being in a true steady state
uplift is occurring because crustal uplift is increasing more where crustal uplift and denudation are in equilibrium and
rapidly than denudation as a result of the relatively poor elevation oscillates around a constant value. As such it is a
linkage between river channels and interfluves; that is, the time-independent landscape which will retain a constant
landscape is relatively insensitive to changes in external form for as long as external factors remain roughly
conditions (in this case the accelerating rate of crustal constant.
uplift). We can extend this steady-state model applied to the
Once valley deepening and widening has produced sharp- , Southern Alps to other active orogenic belts. The rugged
edged interfluves the morphology of the landscape changes mountains which form the backbone of the island of Taiwan
comparatively little, although it continues to experience appear to provide another example of an extensive steady-
surface uplift up to the point where the rate of crustal uplift state landscape. Taiwan lies along the active boundary be-
is balanced by the rate of denudation. Valley-side slopes tween the Philippine and Eurasian Plates and their oblique
are straight and at their threshold angle of stability, and convergence has led to the southward propagation of a
therefore respond very rapidly to changes in conditions in mountain belt, the Central Mountains, over the past 4 Ma.
468 Endogenic-exogenic interactions

channel gradient brought about by uplift will take longer to

propagate through large than through small drainage basins.
In intra-plate setting landscapes are very unlikely to
EURASIAN contain extensive areas of truly steady-state, time-independent
PLATE landforms, even though the very slow rates of landscape
lowering may give the impression of elevation stability.
The predominance of low channel gradients and poor link-
ages between fluvial and slope systems render such land-
scapes predominantly insensitive. This landscape insensitivity
Transition/
to steady reaches an extreme form in landscapes in parts of Africa
state ( and Australia which retain a landform record of tectonic
mountains and climatic events extending back over tens of millions of
of constant years. More generally, the major escarpments which charac-
width '" terize many passive margins effectively decouple the coastal
and interior geomorphic systems so that changes in base level
PHILIPPINE
are transmitted very inefficiently across these landscapes.
PLATE

Widening 18.3 Denudation chronology

Central Mountains
18.3.1 Principles of denudation chronology
Implicit in the Davisian model of landform evolution was
the idea that a landscape should retain some evidence of
earlier landscape cycles, or at least of separate phases of
o 100
, development associated with episodes of rejuvenation with-
km in a single cycle. It was thus possible, it was argued, to
construct a landscape history, or denudation chronology,
documenting the sequence of erosional events. Early attempts
Fig.18.10 Topography of Taiwan in its plate tectonic setting at such landscape histories were based largely on morpho-
showing the northward increase in the breadth of the Central logical evidence which could only yield a crude relative
Mountains to a constant width of90 km. chronology.
In theory a denudation chronology would be expected to
This is evident in the northward increase in width and record the sequence ?f erosional and depositional events
elevation of the mountain range over a distance of around resulting from climatic as well as tectonic and sea-level
120 km in the southern part of the island (Fig. 18.10). North controls. Although climatic oscillations clearly have a pro-
of this point the mountain belt has a constant width of found impact on landscape development, most denudation
about 90 km and is in a steady state with crustal uplift and chronologies constructed within a Davisian framework
denudation both at rates of around 5000 mm ka- 1• The pro- were concerned with the reflection of major base-level
gressive northward widening of the Central Mountains changes in the formation of sequences of erosion surfaces
provides an excellent case for the application of space-time over millions or tens of millions of years. Since the 1960s
substitution (see Section 1.4.2), since from the known rates there has been a severe reaction to the production of such
of plate motion it can be demonstrated that the growth in denudational schemes which have been widely regarded as
width of the mountain belt over a distance of 120 km takes going far beyond the limitations of the available data.
1.3Ma. Criticisms have been based on a number of grounds.
The elevation attained by steady-state topography will A major problem is the extrapolation of surfaces assumed
depend both on the rate of crustal uplift and lithological to be of the same age over large distances between often
controls. If rates of crustal uplift are slow then an equi- small erosional remnants. Accordance of summit elevation
librium with denudation is likely to be achieved before high as a basis for the correlation between, and relative dating
mountains have been created. In weakly resistant rocks, on of, erosion surfaces is complicated by the possibility that
the other hand, even rapid rates of crustal uplift will not surfaces may be warped after they are formed. This may be
generate very high topography because higher rates of denu- obvious where extensive remnants remain, but can cause
dation can be generated on lower gradient slopes. Of great considerable difficulties in correlation where there are only
significance, however, is the width of the mountain belt small, isolated erosion surface residuals. Erosion surfaces,
(see Section 15.8) because changes in base level and by their very nature, are often bereft of deposits which
 Long-term landscape development 469

might provide a basis for dating. In some cases weathering preservation of landforms related to earlier denudational
deposits have been cited as an indicator of relative age, it events.
being assumed that degree of weathering can be regarded It now seems that the greatest obstacle to reconstructing
as a measure of relative age. But it is now appreciated that landscape histories, that of attaching dates to denudational
local and regional factors of drainage and climate can have an events, can be tackled, at least to a certain extent. One means
overriding influence on the nature of weathering deposits. of doing this is by applying geochronometric dating to
Another problem is the· isolating of structural controls materials on the iandsurface (see Section 15.4.1.3). An
from base-level controls in the formation of erosion sur- example of this approach is the estimation of the ages of
faces. Although an erosion surface may locally truncate dif- valleys and their subsequent lowering by the radiometric
ferent lithologies, and thus appear to be related to purely dating of basalt flows that they contain (see Section 18.3.3.4).
erosional controls, this does not necessarily exclude the Another valuable source of information applicable to the
possibility of local structural controls on drainage channels timing and depth of erosion over large continental areas is
which in tum influence landscape development upstream. provided by the sedimentary deposits produced, particularly
Separating structural from drainage basin-wide base-level those that accumulate offshore.
controls is a particular problem in regions, such as parts of
eastern Australia (Fig. 18.11) and southern Africa (Fig.
18.3.2 Continental denudation and the marine
18.12), which are dominated by flat-lying lithologies of
varying resistance.
stratigraphic record
A further difficulty is that landsurfaces previously buried One of the most exciting potential developments in the
by sediments may be exhumed by subsequent erosion and understanding of landform development through time over
significantly influence the course of later landscape evolu- the past decade has been the vast improvement in our
tion. In southern Africa and Western Australia, for instance, knowledge of the offshore sedimentary sequences of the
there is evidence of the exhumation of pre-Permian land- continental margins. Such sequences have to analysed with
surfaces, and it may be difficult to differentiate such surfaces care in view of the possibilities of sediment redistribution
from those that have undergone continuous sub-aerial by submarine currents, and to allow for the effects of com-
development. A final possible problem is that of the preser- paction and the contribution of biogenic material; none the
vation of erosion surfaces over long periods of time. Given less they contain a wealth of information which could pro-
the recent evidence of how extraordinarily low rates of vide insights into the variations in rates of denudation
denudation can be on the plateau surfaces of areas such as through time in the adjacent continental hinterland. Such
south-eastern Australia, this is less of a difficulty than it sedimentary sequences have already been widely us~d to
once appeared. Moreover, it is evident that many low relief reconstruct sea-level changes (see Section 17.2.2.2) but
landsurfaces are armoured by duricrusts and are thereby their application to estimating long-term denudation rates is
very effectively protected from erosion. Clearly, the con- still in its infancy (see Section 15.4.1.1).
struction of denudation chronologies is going to be most
appropriate where landscape sensitivity is low and where,
therefore, there is going to be the greatest potential for the 18.3.3 Regional case studies
By their very nature, studies attempting to reconstruct land-
scape histories involve detailed morphological analysis and
the integration of numerous types of data. Such studies can-
not, therefore, be adequately summarized in a brief form.
The case studies outlined here merely indicate the approach
taken and the major conclusions reached. They can be
pursued, however, through the references provided at the
end of the chapter.

18.3.3.1 South-east England

In 1939 S. W. Wooldridge and D. L. Linton published a
detailed reconstruction of the denudation chronology of
south-east England which was revised in 1955. Their land-
scape history covered a relatively small area consisting of
Cenozoic sediments overlying chalk of Late Cretaceous age,
Fig. 18.11 Structural control of topography by a massive and drew on a wide range of data, including morphological,
sandstone unit in the valley of the Shoalhaven River, East lithological and structural evidence, drainage anomalies and
Australian Highlands, east of Canberra, Australia. datable deposits. They identified three distinct erosion sur-
470 Endogenic~xogenic interactions

Fig.18.12 Landsat image showing the structurally controlled Great Escarpment of western Cape Province, South Africa. The
Location of the photograph at the beginning of the chapter is arrowed. (Landsat image from Satellite Remote Sensing Centre, R. S.A.)

faces: a warped, sub-Cenozoic unconformity trimmed by higher, sub-aerially eroded central part of the region (the
marine action ('Sub-Paleogene surface'); an unwarped Weald) and the discordant drainage of the periphery.
Neogene erosion surface; and an unwarped Plio-Pleistocene Wooldridge and Linton's interpretation of landscape his-
marine platform. These provided the basis for a complex tory, involving distinct episodes of uplift interspersed with
denudation chronology (Table 18.2) in which the Sub- the formation of low relief erosion surfaces, was strongly
Paleogene surface was uplifted and deformed prior to the cyclic in nature and followed clearly in the tradition of Davis.
formation of the Neogene erosion surface, the margins of Subsequent criticisms have focused on the continuity of
which were in tum trimmed by a Late Pliocene-Early tectonic activity and the degree of landscape modification
Pleistocene marine transgression (the Calabrian transgression). achieved by the Calabrian transgression (Table 18.2). It is
This incursion of the sea also laid down marine deposits now thought that tectonic instability throughout the Cenozoic
from which drainage was superimposed. This led to the was much more complex and prolonged, and the Calabrian
contrast between the structurally adjusted drainage of the transgression probably less extensive, and certainly much
 'Long-term landscape development 471

Table 18.2 Comparison of two version of the denudation chronology of south-east England

EPOCH/PERIOD WOOLDRIDGE AND LINTON (1955) JONES (1980)

Holocene Impact of human activities. Post-glacial transgression

leads to valley sedimentation and coastline development

Pleistocene Cyclic fall in sea level producing erosion surfaces Widespread incision in response to glacially lowered sea levels.
and terraces. Periglacial phases. Slow retreat of chalk escarpments and lowering of clay vales leads
to increased local relief.

Pliocene Calabrian transgression-development of discordant Continued uparching of Weald but at a slower rate. Rivers cut down
drainage pattern to within 100-150 m of sea level in the Weald by end of Pliocene.
Widespread removal of Paleogene cover and Upper Cretaceous chalk
Development of Mio-Pliocene peneplain leads to exposure of various Lower Cretaceous strata. Lithological
control of drainage development becomes important. A Mio-Pliocene
Miocene
marine transgression leaves little impact on landscape

Alpine folding phase. Drainage disruption Low relief Paleogene surface with duricrusts. Further growth of
Oligocene Weald anticline. Progressive withdrawal of sea from structural
Sub-aerial denudation. Development of sarsens (silcrete) basins promotes further sub-aerial erosion. Integration of drainage.

Re-establishment of down-dip drainage pattern Continuing flexure leads to emergence of Weald and initiation of
Eocene concordant radial drainage pattern on mantle of Paleocene deposits

Paleogene transgression-development of Continuing tectonic deformation leading to predominantly eastward-

'Sub-Paleogene' surface flowing drainage. Transgression drowns structurally low areas. Up to
Paleocene 200 m of denudation.
Sub-aerial denudation and initiation of down-dip
drainage pattern

Uplift. Development of major structural pattern 300-450 m of uplift around 70 Ma BP along a NW-SE axis.
Upwarping leads to sub-aerial exposure
Cretaceous
Aptian-Cenomanian transgression. Deposition of
300 m of chalk

Source: Based largely on D. K. C. Jones (1980), in D. K. C. Jones (ed.) The Shaping of Southern England. Institute of British Geographers Special
Publication 11,13-47.

less erosionally significant that Wooldridge and Linton were extrapolated entirely from unconformities in sedimen-
envisaged. tary sequences along the adjacent coastal plain, and extensive
sedimentary covers were postulated (notably a marine Cre-
18.3.3.2 The Appalachian region taceous cover) for areas in which no remaining outcrops
The Appalachian region of the eastern USA has been a focus existed (Fig. 18.13).
of historical landscape studies for over a century since the In stark contrast to the cyclic interpretations of Davis and
publication of Davis's classic Rivers and valleys of his followers, Hack used the Appalachians as an example
Pennsylvania in 1889. In spite of the pervasive relationship of the operation of 'dynamic equilibrium' in his initial pre-
between morphology and lithology over much of the region, sentation of the concept. Where the evolutionists saw cycliC
Davis and his disciples considered that there were sufficient erosion surfaces Hack saw accordant summits at different
departures from lithological controls to warrant a cyclic inter- levels on different lithologies, and attributed them to the vari-
pretation involving successive episodes of uplift and erosion able rates of erosion on rocks of contrasting resistance. More
surface formation. The major anomalies identified included recently the long-standing problem of the geomorphic
the fact that although the tectonic structure of the Appala- development of the Appalachian region has been looked at
chians is broadly symmetric, the drainage is decidedly asym- afresh with the benefit of new techniques and data able to
metric with rivers crossing from west to east. This transverse yield estimates of actual rates of long-term denudation, and
drainage in tum suggested a history of antecedence, super- new tectonic concepts which place the landscape history in
imposition and headward erosion along less resistant litho- the context of early orogenic events and subsequent conti-
logical units. In addition, the apparent accordance of summit nental break-up and passive margin tectonics. Detailed
elevations in parts of the Appalachian region suggested the research on deep subsurface structures has demonstrated
existence of erosion surfaces. The lack of any datable deposits that some correspond in a complex fashion to surface mor-
between the Early Permian and the Pleistocene meant that phology, while new data on ~he tectonic and sedimentary
there were effectively no critical tests available for the history of the region have led to the abandonment by some
various schemes proposed. In some cases erosion surfaces researchers of the idea of an extensive cover of Cretaceous
472 Endogenic-exogenic interactions

It I ASS I C low LAN 0 :COASTAL PLAIN'

........-..... ,..

Fig.18.13 Scheme of landscape development proposed by D. W. Johnsonfor the eastern margin of the USA in the vicinity ofNew
Jersey involving the formation of a pre-Cretaceous Fall Zone peneplain which was covered by a Cretaceous marine transgression
before being upwarped and eroded to form the widespread Schooley peneplain. (From D. W. Johnson (1931) Stream Sculpture on the
Atlantic Slope. Columbia University Press, New York, Fig.1S, p. 77.)

strata over the region. Fission track dating inland in con- Table 18.3 Denudation chronology of southern Africa according to
junction with data on sediment volumes offshore has shown T. C. Partridge and R. R. Maud (compare with Fig. 18.14)

that denudation over the region of eastward draining rivers

EVENT AGE
has averaged around 2-2.5 kIn over the past 100 Ma, at rates
ranging up to 60 m Ma- 1 . Although rates of denudation have Post-African II cycle of erosion Late Pliocene to Holocene
apparently varied over both time and space, there is no evi- Major uplift of up to 900 m Late Pliocene
dence for the succession of brief episodes of uplift and pro- (- 2.5 Ma BP)

longed periods of erosion required by conventional cyclic Post-African I cycle of erosion Early mid-Miocene to
Late Pliocene
models of landscape evolution. (- 2.5 MaBP)
The growing consensus is that, although there have been Moderate uplift of 150-300 m End of Early Miocene
discontinuities in rates of crustal uplift, at least during the (- 18 Ma BP)
Late Cenozoic, landscape history since the opening of the African cycle of erosion with Late Jurassic/Early Cretaceous
North Atlantic and associated uplift of the new continental minor tectonic interludes to end of Early Miocene
(-18 MaBP)
margin in the Late Cretaceous has probably been dominated
Break up of Gondwana and Late Jurassic/Early Cretaceous
by ongoing denudation and compensatory isostatic uplift. initiation of Great Escarpment
While a mean thickness of about 2.5 km of rock has been
removed, isostatic rebound has reduced surface lowering to Source; Based on T. C. Partridge and R. R. Maud (1987) South African
Journal of Geology 90, Table 1, p. 187.
around 500 m. This situation has been described by some
researchers as representing a steady state between weather-
ing, erosion and crustal uplift, but this is not strictly true nology put forward by King has been subject to major
because local compensatory isostatic uplift alone leads to a revision (Table 18.3); it will be interesting to see to what
progressive, albeit in many cases very slow, surface lower- extent this revised chronology can be related not only to
ing. Flexural effects could allow a localized steady state to detailed offshore depositional evidence but also to estimates
be maintained for a short period, but these have yet to be of long-term denudation rates employing fission track dating
considered in relation to landscape development in this and contemporary ideas on the tectonic history of passive
area. margins.
Some idea of the likely value of offshore sedimentary
18.3.3.3 Southern Africa sequences to reconstructing landscape histories can be gained
Perhaps the grandest schemes of denudation chronology from Figure 18.14 in which the denudation chronology for
have been those put forward by L. C. King. On the basis of Africa proposed by King is compared with the record of
detailed investigations of the morphology and sedimentary offshore sedimentation. The most important point to note is
sequences of the coastal belt of Natal in South Africa, King the lack of any consistent correspondence between hiatuses
proposed a chronological scheme which, he argued could (gaps) in marine deposition and either phases of uplift or
be applied on a global basis (Fig. 18.14). Few would now erosion proposed by King. This lack of correspondence
accept the possibility of world-wide denudational episodes may be due to a number of factors. The denudation chrono-
in the light of our present understanding of uplift patterns in logy outlined by King may simply be wrong; alternatively,
plate interiors and the probable dismemberment of drainage the relationship between denudation inland and sedimen-
by continental break-up (see Section 16.7.1). Nevertheless, tation offshore may be complex, in which case unconformi-
the attempt by King to link onshore and offshore data anti- ties offshore cannot be simply equated with erosion surfaces
cipated what is likely to become an important approach in inland. This is almost certainly the case for, as we saw in
the documenting of landscape histories. Recently the chro- Section 17.6.1, river systems can respond in different ways
 Long-term landscape development 473

TIME SCALE CHRONOLOGY Of UPliFT AND PlANATION CONTINENTAl MARGIN

SEDIMENTATION
Me RIFTED 9EAIIED aMI(
0
OJATERNARY
Congo cycle of <IMp
PLIOCENE
~ _.,"",~,"".~ 12·4 3 158
10 '"Z S W coastal margin
I Younges, Cyelel
U
'"
~ MIOCENE

- '"0 Strong cymatogeny

}M~"_"
20
Z U1te &x:.w-
0 Second phase of

~ ~.~~.,~--
Early M~
30 N lW,despreod Landsalpel hiatus
OLIGOCENE

0 '"Z • Moderate uplift

40 I Active Episode 0 I
Z '"
~ Broad valley-Hoor pe<:hplalns

1
0 • deVf)loped In "Afne.n" surface
EOCENE
('Roiling Landsurtace')
50
'" '" ...J 17 JO 91
«
60
u 0-
PALEOCENE :::=:T"""","
'Afncan landscape' of U1te Maastrichtian /
extreme planation Early PlIeoalne hiatus
('Moorland planatlOf1"
70

U I
80 I lO7 97 46
LATE
<J) 7 Conllnent.wlde
uplift
- ::J ------.. I Act,ye Ep,sode 81

90 0
0 '"
u Cenomanian
« regression
f--
100
N '"
ex: Post- Gondwana MOSlon
117 224 64
u concentrated In upwarps
120 EARLY (Kretaclc planallon"I South Atlantic
0
unconfcwmity

40 <J)
Fragmentation of
Gondwanaland
I Act,ve Ep,sode AI
•
I
147 73 84

60
'" u
iii
LATE
'Gondwana landscape
of el(treme planatton
I ~

<J)
MIDDLE ('Gondwana plan.llon',
~ «
80 ex:
~
EARLY
- 56+
200

Fig.18.14 Comparison ofdenudation chronology and record of offshore sedimentation for Africa. The denudation chronology is
according to L. C. King and refers primarily to central and southern Africa. King's later modified terminology for the various
landsurfaces is shown in parentheses. Vertical arrows indicate approximate timing and duration of uplift according to King.
Sedimentation rates are for the western, southern and eastern margins ofAfrica as estimated by Dingle (1982); figures refer to mean
sedimentation rates in m Ma- I . Sunk margins are those on the east coast which have experienced persistent subsidence since rifting.
Hiatuses in deposition are also indicated. The time scale is based on Harland et aI. (1982). (From M.A. Summerfield (1985) in M.
Morisawa and 1. T. Hack (1985) Tectonic Geomorphology. Allen and Unwin, Boston, Table 1, pp. 34-5. Primary sources: L.C. King
(1967) The Morphology of the Earth (2nd edn.) Oliver and Boyd, Edinburgh; L. C. King (1976) Zeitschrift fiir Geomorphologie 20,
133-48; L. C. King (1983) Wandering Continents and Spreading Sea Floors on an Expanding Earth. Wiley, Chichester and New York;
R. V. Dingle (1982) inR.A. Scrutton (ed.) Dynamics of Passive Margins. American Geophysical Union, Washington: Geological
Society ofAmerica, Boulder 59-71; W. B. Harland et al. (1982) A Geologic Time Scale. Cambridge University Press, Cambridge.)

to a fall in base level depending on the gradient across gradual descent to the west, but a generally sharp drop in
continental margins. Although seaward warping of a con- elevation eastwards across a major escarpment to a coastal
tinental margin could increase channel gradients and promote plain. The Highlands lie along a rifted margin, albeit one
faster rates of offshore sedimentation, a reduction in base with a complex history.
level due to a eustatic sea-level fall might not generate a phase Attempts to interpret the history of the East Australian
of rejuvenation in drainage basins along the margin if Highlands in cyclic terms inferred a major Plio-Pleistocene
channel gradients decreased offshore. surface uplift on the basis of the present elevation of a
duricrusted erosion surface, assumed to have formed close
18.3.3.4 The East Australian Highlands to sea level in the Miocene. Deeply inci ed valleys cutting
The East Australian Highlands extend along almost the back from the coast, such as that of the Shoalhaven River
entire eastern periphery of the continent, although they (Fig. 18.11), added to the idea that uplift had been recent.
attain their greatest elevation in the south where heights in The challenge to this cyclic model was based on both the
excess of 2000m occur. Over most of its length the highland possibility of significant lithological control and dating of
belt has the form of a broad, asymmetric arch with a very Cenozoic lavas erupted on to the Highlands. Direct litho-
474 Endogenic-exogenic interactions

logical control of the topography of the Highlands is locally of the westward-draining Lachlan River, basaltic lavas
provided by essentially horizontally bedded sediments which erupted about 20 Ma BP and flowed into river valleys
(Fig. 18.11). Elsewhere, erosion surfaces bevelling steeply have been preserved as sinuous hill-top cappings. These
dipping strata may in some cases be graded to outcrops of basalt outcrops in places overlie fluvial sediments which
resistant beds downstream. The presence of such litho- show the flow direction of the pre-basalt streams to be
logical controls implies that the low relief surfaces of the similar to the present river channels. Together this evidence
Highlands need not have been formed close to sea level. enables the longitudinal profiles of the pre-basalt and
Radiometric dating of Cenozoic lavas on the summit of the modem river valleys to be compared (Fig. 18.15). These
Highlands in New South Wales also demonstrate that the show that there has been little change in longitudinal pro-
landsurfaces there are much older than the earlier cyclic files over 20 Ma, during which time valley floors have been
models assumed. Recent research by both geologists and lowered by about 100 m. This extremely slow rate of inci-
geophysicists concerning the tectonic history of the East sion is broadly in accord with other evidence of low denu-
Australian Highlands has not achieved a consensus, with at dation rates in the Murray Basin of which the Lachlan
least three distinct models being suggested: (1) isostatic catchment forms a part (see Section 15.4.2). Such rates
uplift of Mesozoic highlands in reponse to prolonged argue strongly against cyclic notions of landscape develop-
erosion; (2) active Late Mesozoic and/or Cenozoic uplift; ment and point to the predominantly slow rates of land-
and (3) passive margin subsidence of Mesozoic highlands scape modification in passive margin environments inland
punctuated by episodes of active uplift. of actively eroding escarpment environments. Assuming
A particularly detailed study has been undertaken by P. local isostatic compensation, the mean denudation rates esti-
Bishop in an area of the East Australian Highlands in central mated for the Lachlan headwaters area indicate a surface
New South Wales. In this region, located in the headwaters lowering of no more than 2 m Ma- 1 over the past 20 Ma.

(0) 650 Bevendole Basalt

600 . Lachlan
River
-J
Volley
vi
"l: 500
EO as
400 70 80
0 10 40 50 60

Wheeo Basalt
(b) 900

800

700
...j
vi
"l:
600
~

500

400
0 10 20 30 40 50 60 70 80
Dis 'oncII upvollllY from conflUflnclI of CrookwIIII EJ Lachlan RiJlflr s
( Va lilly Kilomllfru)

Fig. 18.15 Longitudinal profiles of the Early Miocene and present Lachlan River valley (a), and tributary Crookwell River - Wheeo
Creek valleys (b). Note that valley-floor incision from the upper (Early Miocene) profile has been relatively consistent downvalley and
that lithological control of the form of the longitudinal profiles has been sustained. (From P. Bishop et al. (1985) Journal of Geology
93, Fig. 7, p. 466.)
 Long-term landscape development 475

18.4 Tectonics, climate and landform genesis. Such influences can operate in various ways. Uplift
development occurring in an orogenic belt may cause the displacement
of landforms created in a particular climatic environment
Various points made in this chapter underscore the impor-
into a new morphoclimatic regime. Such transitions can
tance of tectonics in controlling the way landscapes evolve;
occur rather rapidly if the rate of surface uplift is high, and
but the role of tectonics in long-term landform development
climatic landform relicts formed at lower elevations may be
is not confined to these essentially direct effects. Tectonic
temporarily preserved as they pass into higher altitudes. In
activity can also play an important part in influencing local,
general, the high rates of denudation in most mountain belts
regional and global climates and thereby exert an additional
would make either significant vertical displacement or
indirect control over landscape 'evolution.
prolonged survival of climatic relicts unlikely, but this is
not necessarily so for montane plateaus such as the Tibetan
18.4.1 Continental drift and changing morphoclimatic Plateau and the Altiplano of the central Andes. In these
regimes . cases the much more subdued local relief - and conse-
One obvious way in which we might expect tectonics to quently more modest rates of denudation on the plateau
influence climate is through continental drift. At the global summits - could permit the long-term preservation of relict
scale, changes in the distribution of the continents over the landforms. Certainly, there is clear evidence from the
past few hundred million years have certainly had a pro- reduction in the size of lakes for a progressively more arid
found impact on climates world-wide, although the precise climate on the Tibetan Plateau as it has apparently experi-
effects have yet to be ascertained. It seems clear that the encedrapid surface uplift during the Quaternary.
movements of the southern continents played an important A gradual decrease in temperature with elevation is the
part in the transition from the apparently warm, equable most obvious effect of surface uplift. In extreme cases this
climates of the Late Cretaceous to the distinctly cooler, and can promote the growth of glaciers and the initiation of a
globally more diverse, climates of the Late Cenozoic (see glacial morphoclimatic regime. In most situations precipita-
Section 14.3.1). tion increases markedly with elevation, but in very high
An additional climatic consequence of continental drift mountains there can be a gradual reduction above a height
could arise from changes in the palaeolatitudes of indivi- of 3000-3500 m where moisture-depleted air begins to be
dual continents. In most cases, climatic changes associated encountered. At the equator the mean annual air tempera-
with drift have been slow in comparison to changes related ture is below 0 °C at an altitude of around 4500 m, but at
to major adjustments in global circulation patterns. But it higher latitudes this critical boundary approaches gradually
has been suggested that this has not been the case for closer to sea level. The Andes in Ecuador illustrate a full
Australia, which has been moving rapidly equatorwards for range of morphoclimatic regimes from lowland humid .tro-
about the past 40 Ma from a mean latitude of around 45°S pical to glacial. On the western slopes the upper, northern
to about 25°S at the present day. This has enabled it to part below the snow line is characterized by intensely
maintain an essentially tropical or sub-tropical climate in dissected topography created under prolonged humid con-
spite of the marked global cooling of the Late Cenozoic. It ditions. This contrasts with the lower and more southerly
is tempting to suggest that one consequence of this equator- section of the Ecuadorian Andes where landform develop-
ward movement was Australia's entry into the latitudinal ment has occurred under a distinctly more arid climatic
zone of aridity related to the sub-tropical high pressure cells regime. As illustrated by the Andes, climatic transitions
during the Late Cenozoic. That this is not an adequate ex- with altitude may not be uniform on either flank of a moun-
planation of changing climatic conditions in Australia is tain range. Where an orogen is transverse to the prevailing
indicated by evidence for a humid environment along the rain-bearing winds there will tend to be a marked asymmetry
southern coast of Australia in the Miocene which extended in climate, with a drier altitudinal sequence of morphocli-
deep into the interior. An alternative interpretation is that matic environments on the leeward side.
any climatic changes linked to drift have been subordinate
to major changes in global circulation patterns, with Australia 18.4.3 Plate tectonics, climate and weathering
being overtaken by intensified sub-tropical high pressure
cells migrating equatorwards which promoted aridity start- Recent research has begun to highlight a possible relation-
ing at about the Miocene-Pliocene boundary. ship between tectonics and climate that is even more
pervasive than those relating to continental drift and oro-
genic uplift. This arises from variations in rates of sea-floor
18.4.2 Orogenesis and changing morphoclimatic spreading which lead to changes in the amount of atmo-
regimes spheric CO 2 ; these, in tum, promote adjustments in global
Another way in which tectonics can influence climate (and temperatures (Fig. 18.16). In the short term the concen-
thereby indirectly landform development) is through oro- tration of atmospheric CO 2 is closely related to the life
476 Endogenic-exogenic interactions

25
A
computer
model A
a> 23
(,)

....co
:;
en ~
e Palaeobotanical
/ ' evidence
::l 21
c: .....
co
a>
co L-
a>
-, ........
~ c.
19

Fig.18.16 Schematic representation of the likely causal

(ij
.0
0
E
~ C02comput~ I'
<:5 model B
relationships between sea-floor spreading rates, atmospheric CO 2 17
concentration, climate and weathering rates.

processes of organisms, but in the long term it appears that

geological, or more specifically tectonic, controls are domi-
B
nant. Over geological time the amount of CO 2 in the 150
atmosphere represents a balance between its release by
co
volcanic activity and rock metamorphism, and uptake in Q)
L-

weathering reactions (see Section 6.2.3.3). The tectonic «

process which ultimately seems to control the release of '0 "'E 140
c: ~
co
CO 2 is sea-floor spreading. High rates of sea-floor ....J
cD

g
spreading will be associated with both active volcanism (ij
..... 130
along mid-oceanic ridges and, because of correspondingly 0
I-
high rates of subduction, considerable volcanic activity and
metamorphism along steady-state convergent plate margins.
120
In both cases increased rates of sea-floor spreading should
be associated with increased rates of CO 2 outgassing.
There is fairly convincing evidence that rates of volcanic
activity have varied significantly over periods of tens of 240
a>
millions of years, and there is certainly abundant data indi- E
::l
cating that rates of sea-floor spreading have changed over
the past 100 Ma or so (Fig. 18.17(C». Although weathering ~a> ;;- 200
E
reactions tend to absorb CO 2, the surface area of the con- C)~
'0 cD0
tinents available for these processes would be reduced ._
0::.=
during periods of high sea-floor spreading rates because of
c: 160
the corresponding rise in global sea levels as mid-oceanic co
Q)

ridge volumes grow (Fig. 18.17(B); see Section 17.5.2.1). ~

Atmospheric CO2 is transparent to incoming short-wave 120+----,-------.-----r----,-------,
radiation from the Sun, but efficiently absorbs outgoing
100 80 60 40 20 o
long-wave radiation emitted from the Earth. This is the so-
called 'greenhouse effect'. If the Cretaceous atmosphere Ma B. P.
had a significantly higher CO 2 content then the climate at Fig.18.17 Trends in mid-oceanic ridge volume (indicative of
that time should have been warmer world-wide, and a range sea-floor spreading rate), total land area and global mean
of palaeontological, geochemical and palaeogeographical surface temperature over the last 80-100 Ma. In (A) CO 2
data certainly supports this interpretation. We would expect computer model A is the predicted temperature fluctuation as a
consequence of changes in CO2 related to a variable rate of sea-
warmer temperatures to promote faster rates of chemical floor spreading since 100 Ma BP whereas CO 2 computer model B
weathering, not primarily through an acceleration in weather- is the temperature prediction based on a linear decrease in sea-
ing reaction rates, but rather as a result of enhanced rates of floor spreading rate. (Based on R. A. Berner et al. (1983)
organic activity leading to a greatef' production of soil CO2 American Journal of Science 283, Fig. 12, p. 679 (A);
and a consequential increase in soil acidity. Because the E. 1. Barron (1983) Earth-Science Reviews 19, Fig. 4, p. 317 (B);
and M. A. Kominz (1984) in J. S. Schlee (ed.) Interregional
concentration of CO 2 in soil is generally so much higher Unconformities and Hydrocarbon Accumulation. American
than that in the atmosphere, changes in the overall concen- Association of Petroleum Geologists Memoir 36 Fig. 5,
tration of CO 2 in the atmosphere probably have little direct p. 119 (C).)
 Long-term landscape development 477

effect on weathering rates; rather 'the effect is indirect 18.4.4 Tectonics, climate and landscape development:
through the influence of temperature on organic activity a summary
and hence soil acidity.
It is useful to conclude this discussion of long-term land-
Estimates of the concentration of atmospheric CO2 during
scape development by presenting an overview of the way in
the Cretaceous vary enormously from three times to 100
which tectonic processes are related to long-term fluctua-
times present values. Even assuming an increased concen-
tions in global climate, which in tum generate changes i~
tration at the lower end of this range, the resulting likely
morphoclimatic regimes. A schematic representation of
higher global temperatures in the Cretaceous (Fig. 18.17(A))-
these relationships is provided in Figure 18.18; this indi-
and the possibly significantly enhanced rates of chemical
cates climatic changes arising from sea-floor spreading,
weathering - need to be borne in mind when we try and
continental drift, orogenesis and volcanicity. It is important
intepret landscapes whose evolution extends back to this
.to note that the relationships suggested in this diagram refer
period.
to periods of tens of millions of years. Within this temporal
It is also worth noting that chemical weathering appa-
scale glacial-interglacial oscillations represent rapid fluc-
rently provides a vital natural negative feedback mechanism
tuations in morphoclimatic regimes within a framework of
limiting increases in the' concentration of CO2 in the atmos-
longer-term changes in climate. In the high and mid-latitudes
phere, and thereby helping to limit long-term fluctuations
the effect of these short-term climatic cycles dominate the
in global temperature. Since rates of chemical denudation,
landscape, but landform relicts are found from morpho-
and therefore chemical weathering, are highest in regions
climatic regimes operative several tens of million of years
of high local relief (see Section 15.5) we might also anti-
ago in the most insensitive landscapes of the tropics and
cipate that long-term changes in contin~ntal morphology, in
sub-tropics.
particular the rifting of continents and the growth and ero-
sion of orogenic belts, might also influence the concentra-
tion of atmospheric CO 2 , These possible relationships Further reading
represent a fruitful field for future research linking climatic Models of long-term landscape evolution are discussed by
change to geomorphic processes and long-term landscape Thomes and Brunsden (1977) and Thorn (1988), and in
history.

GLOBAL THERMALLY
CHANGEIN ALBEDO ISOLATED POLES
PALAEOLATITUDE

OROGENESIS

CHANGE IN ATMOSPHERIC-
OCEANIC CIRCULATION
AND CLIMATIC PATIERNS

CONTINENTALITY

FULL ICE AGE

CLIMATE (glacial
-interglacial and
Short-term arid-humid
glacio-eustasy oscillations
VOLCANICITY

Long-term
tectono-eustasy

VOLUME OF GLOBAL
t------il~MID-OCEANIC RIDGE SYSTEM

. Fig. 18.18 A schematic representation of some ?f the ways in wh~ch tectonic processes generate chan~es in morphoclimatic regimes
over time. The relationships shown are to a certam extent speculatlve and are by no means comprehenslve.
478 Endogenic-exogenic interaction

various papers in the volume edited by Melhorn and Flemal as case studies can be found in the following: Jones (1980),
(1975). Other contributions to the revival of interest in the Small (1980) and Wooldridge and Linton (1955) on south-
theoretical problems posed by long-term landscape devel- east England; Costa and Cleaves (1984), Davis (1889),
opment include those by Chorley (1962), Howard (1965), Johnson (1931), Hack (1960, 1979, 1982) and Pavich (1985)
Schumm and Lichty (1965), Thomas and Summerfield on the Appalachians; De Swardt and Bennet (1974), Partridge
(1987) and Thomes (1983). As with all reading on this topic, and Maud (1987) and Summerfield (1985) on southern
great care has to be taken when using these sources in Africa; and Bishop (1988), Bishop et al. (1985), Ollier
noting the sense in which the terms dynamic equilibrium, (1982), Wellman (1987, 1988) and Young (1983) on the
steady state and time-independent form are being applied East Australian Highlands.
since their precise usage varies. Much of the rather limited literature on the relationships
The cyclic model of W. M. Davis is given exhaustive between tectonics, climate and landform development has
treatment in the comprehensive survey of his work by not been written from a specifically geomorphic perspec-
Chorley et al. (1973) while key writings by Davis himself tive, but none the less is still thought-provoking for the
include his general statement of the cycle (Davis, 1899) and geomorphologist. The relationship of global climate to chang-
his later collection of essays (Davis, 1954). The Davisian ing patterns in the distribution of the continents is discussed
approach to landform analysis is assessed in the context of by Barron et al. (1984) and Cogley (1979), and the impact
systems analysis by Chorley (1962). The ideas of Penck are of orogenic uplift on montane climates is analyzed in terms
less well recorded in the Anglo-American literature; the of landform development by Garner (1974, 1983). The link
basic source is the translation of Penck's book originally between sea-floor spreading rates, atmospheric CO 2 , global
published posthumously in German (Penck, 1953). It is also temperatures and rates of chemical weathering is examined
worth consulting the paper by Simons (1962) which tries to by Berner and Lasaga (1989), while Barron (1983) reviews
clear up the confusion which arose from Davis's misrepre- the evidence for a warm, equable Cretaceous climate. Finally,
sentation of Penck's ideas concerning slope evolution. the idea of alternating geomorphic and sedimentological
King's model of landscape evolution is covered in a regimes arising from contrasting modes of tectonic activity
series of papers and books of which the most important are through geological time is presented by Mackenzie and
King (1953, 1957, 1967, 1976). Readers should be aware, Pigott (1981).
however, of the various revisions that King made to his
model, especially in terms of the timing and mechanisms of
uplift, and should consult the papers by De Swardt and
Bennet (1974) and Partridge and Maud (1987) which care- References
fully document these changes. The concept of etchplanation
Adams, J. (1985) Large-scale tectonic geomorphology of the
is comprehensively covered by Biidel (1982) and Thomas Southern Alps, New Zealand. In: M. Morisawa and 1. T. Hack
(1974), while Kiewietdejonge (1984a, b) provides a wide- (eds) Tectonic Geomorphology. Allen and Unwin, Boston, and
ranging review of Biidel's approach to geomorphology. London, 105-28.
Thomas (1989a, b) assesses the role of etching in landscape Ahnert, F. (1984) Local relief and the height limits of mountain
ranges. American Journal of Science 284, 1035-55.
development and Thomas and Thorp (1985) apply the etch-
Barron, E. J. (1983) A warm, equable Cretaceous: the nature of
planation model to the landscape of West Africa. the problem. Earth-Science Reviews 19, 305-38.
Hack's 'dynamic equilibrium' model is covered in Hack Barron, E. 1., Thompson, S. L. and Hay, W. W. (1984) Continental
(1960, 1975), and the dynamic metastable equilibrium model distribution as a forcing factor for global-scale temperature.
and the concept of episodic erosion is developed by Schumm Nature 310, 574-5.
Berner, R. A. and Lasaga, A. C. (1989) Modeling the geochemical
(1975, 1979). The concept of landscape sensitivity is dis-
carbon cycle. Scientific American 260 (3), 54-61.
cussed in detail by Brunsden and Thomes (1979), while Bishop, P. (1988) The eastern highlands of Australia: the evolu-
Brunsden (1980) provides a briefer and more accessible treat- tion of an intraplate highland belt. Progress in Physical Geo-
ment. The importance of relative rates of uplift and denu- graphy 12, 159-82.
dation in determining whether landscapes are essentially Bishop, P., Young, R. W. and McDougall, 1. (1985) Stream profile
change and longterm landscape evolution: Early Miocene and
time-dependent or time-independent is emphasized by Adams
modern rivers of the East Australian Highland crest, central
(1985) and Ahnert (1984). Adams (1985), Whitehouse (1988) New South Wales, Australia. Journal ofGeology 93, 455-74.
and Suppe (1987) provide case studies of steady-state land- Brunsden, D. (1980) Applicable models of long term landform
scapes along convergent plate boundaries in New Zealand evolution. Zeitschrift fur Geomorphologie Supplementband 36,
and Taiwan. 16-26.
Various issues relevant to denudation chronology are Brunsden, D. and Thomes, J. B. (1979) Landscape sensitivity and
change. Transactions of the Institute ofBritish Geographers NS
considered by Fairbridge and FinkI (1980), Mabbutt (1988), 4,463-84.
Ollier et al. (1988), Summerfield (1985), Twidale (1976) BUdel, J. (1982) Climatic Geomorphology (translated by L.
and Young (1983). Further details on the regions discussed Fischer and D. Busche). Princeton University Press, Princeton.
 Long-term landscape development 479

Chorley, R. 1. (1962) Geomorphology and general systems theory. Mackenzie, F. T. and Pigott, 1. D. (1981) Tectonic controls of
United States Geological Survey Professional Paper 500B. Phanerozoic sedimentary rock cycling. Journal of the Geological
- Chorley, R. 1., Beckinsale, R. P. and Dunn, A. J. (1973) The His- Society London 138, 183-96.
tory of the Study of Landforms. Volume Two: The Life and Melhorn, W. N. and Flemal, R. C. (eds) (1975) Theories of Land-
Work of William Morris Davis. Methuen, London. form Development. Allen and Unwin, London and Boston.
Cogley, J. G. (1979) Albedo contrast and glaciation due to Oilier, C. D. (1982) The Great Escarpment of eastern Australia:
continental drift. Nature 279, 712-13. tectonic and geomorphic significance. Journal of the Geological
Costa, J. E. and Cleaves, E. T. (1984) The Piedmont landscape of Society of Australia 29, 13-23.
Maryland: a new look at an old problem. Earth SUliace Processes OIlier, C. D., Gaunt, G. F. M. and Jurkowski, 1. (1988) The
and Landforms 9, 59-74. Kimberley Plateau, Western Australia: a Precambrian erosion
Davis, W. M." (1889) The rivers and valleys of Pennsylvania. surface. Zeitschriftfiir Geomorphologie 32, 239-46.
National Geographic Magazine 1, 183-253. Partridge, T. C. and Maud, R. R. (1987) Geomorphic evolution of
Davis, W. M. (1899) The geographical cycle. Geographical Journal southern Africa since the Mesozoic. South African Journal of
14,481-504. Geclogy 90, 179-208.
Davis, W. M. (1954) Geographical Essays (ed. by D. W. Johnson). Pavich, M. J. (1985) Appalachian piedmont morphogenesis: wea-
Dover, New York. thering, erosion, and Cenozoic uplift. In: M. Morisawa and J. T.
De Swardt, A. M. J. and Bennet, G. (1974) Structural and phy- Hack (eds) Tectonic Geomorphology. Allen and Unwin, Boston
siographic development of Natal since the late Jurassic. Tran- and London, 299-319.
sactions of the Geological Society of South Africa 77, 309-22. Penck, W. (1953) Morphological Analysis of Landforms (trans-
Fairbridge, R. W. and Finkl, C. W. (1980) Cratonic erosional lated by H. Czech and K. C. Boswell). Macmillan, London.
unconformities and peneplains. Journal of Geology 88,69-86. Schumm, S. A. (1975) Episodic erosion: a modification of the
Gamer, H. F. (1974) The Origin of Landscapes: A Synthesis of geomorphic cycle. In: W. N. Melhorn and R. C. Flemal (eds)
Geomorphology. Oxford University Press, New York and Theories of Landform Development. Allen and Unwin, London
London. and Boston, 69-85.
Garner, H. F. (1983) Large-scale tectonic denudation and climatic Schumm, S. A. (1979) Geomorphic thresholds: the concept and its
morphogenesis in the Andes mountains of Ecuador. In: R. applications. Transactions of the Institute of British Geographers
Gardner and H. Scoging (eds) Mega-Geomorphology. Clarendon NS 4, 485-515.
Press, Oxford and New York, 1-17. Schumm, S. A. and Lichty, R. W. (1965) Time, space and causality
Hack, 1. T. (1960) Interpretation of erosional topography in humid in geomorphology. American Journal of Science 263, 110-19.
temperate regions. American Journal of Science 258A, 80-97. Simons, M. (1962) The Morphological Analysis of Landforms: a
Hack, J. T. (1975) Dynamic equilibrium and landscape evolution. new review of the work of Walther Penck (1888-1923). Tran-
In: W. N. Melhorn and R. C. Flemal (eds) Theories of Land- sactions of the Institute of British Geographers 31, 1-14.
form Development. Allen and Unwin, Boston and London, Small, R. 1. (1980) The Tertiary geomorphological evolution of
87-102. south-east England: an alternative interpretation. In: D. K. C.
Hack, 1. T. (1979) Rock control and tectonism - their importance Jones (ed.) The Shaping of Southern England. Institute of
in shaping the Appalachian Highlands. United States Geologi- British Geographers 11. Academic Press, London and New
cal Survey Professional Paper 1126-B, BI-BI7. York,49-70.
Hack, J. T. (1982) Physiographic divisions and differential uplift Summerfield, M. A. (1985) Plate tectonics and landscape develop-
in the Piedmont and Blue Ridge. United States Geological ment on the African continent. In: M. Morisawa and 1. T. Hack
Survey Professional Paper 1265. (eds). Tectonic Geomorphology Allen and Unwin, Boston and
Howard, A. D. (1965) Geomorphological sy tems - equilibrium London, 27-51.
and dynamics. American Journal of Science 263, 436-45. Suppe, J. (1987) The active Taiwan mountain belt. In: J.-P. Schaer
Johnson, D. W. (1931) Stream Sculpture on the Atlantic Slope. and J. Rodgers (eds) The Anatomy of Mountain Ranges.
Columbia University Press, New York. Princeton University Press, Princeton, 277-93.
Jones, D. K. C. (1980) The Tertiary evolution of south-east Thomas, M. F. (1974) Tropical Geomorphology. Macmillan,
England with particular reference to the Weald. In: D. K. C. London and Halstead Press, New York.
Jones (ed.) The Shaping of Southern England. Institute of Thomas, M.F. (1989a) The role of etch processes in landform
British Geographers Special Publication 11. Academic Press, development 1. Etching concepts and their applications. Zeitschrift
London and New York, 13-47. fur Geomorphologie, 33, 129-42.
Kiewietdejonge, C. J. (1984a) BUde! 's geomorphology 1. Progress Thomas M. F. (1989b) The role of etch processes in landform
in Physical Geography 8, 218-48. development II. Etching and the formation of relief. Zeitschrift
Kiewietdejonge, C. 1. (1984b) BUde!'s geomorphology II. Progress fur Geomorphologie 33, 257-74.
in Physical Geography 8,365-97. Thomas, M. F. and Summerfield, M. A. (1987) Long-term land-
King, L. C. (1953) Canens of landscape evolution. Bulletin of the form development: key themes and research problems. In: V.
Geological Society of America 64, 721-52. Gardiner (ed.) International Geomorphology 1986: Proceed-
King, L. C. (1957) The uniformitarian nature of hillslopes. Tran- ings of the First International Conference on Geomorphology
sactions of the Edinburgh Geological Society 17, 81-102. Part II. Wiley, Chichester and New York, 935-56.
King, L. C. (1967) The Morphology of the Earth (2nd edn). Oliver Thomas, M. F. and Thorp, M. B. (1985) Environmental change and
and Boyd, Edinburgh. episodic etchplanation in the humid tropics of Sierra Leone: the
King, L. C. (1976) Planation remnants upon high lands. Zeitschrift Koidu etchplain.In: 1. Douglas and T. Spencer (eds) Environmen-
fiir Geomorphologie 20, 133-48. tal Change and Tropical Geomorphology. Allen and Unwin, Lon-
Mabbutt, 1. A. (1988) Land-surface evolution at the continental don and Boston, 239-67.
time- cale: an example from interior Western Australia. Earth- Thorn, C. E. (ed.) (1982) Space and Time in Geomorphology.
Science Reviews 25, 457-66. Allen and Unwin, London and Boston.
480 Endogenic-exogenic interactions

Thorn, C. E. (1988) Introduction to Theoretical Geomorphology. and Antarctic highlands. Zeitschrift fur Geomorphologie 32,
Unwin Hyman, Boston and London. 17-29.
Thomes, J. B. (1983) Evolutionary geomorphology. Geography Whitehouse, I.E. (1988) Geomorphology of the central southern
68,225-35. Alps, New Zealand: the interaction of plate collision and atmo-
Thomes, 1. B. and Brunsden, b. (1977) Geomorphology and Time. spheric circulation. Zeitschrift fur Geomorphologie Supplement-
Methuen, London. band 69, 105-16.
Twidale, C. R. (1976) On the survival of paleoforrns. American Wooldridge, S. W. and Linton, D. L. (1955) Structure, Swjace
Journal of Science 276, 77-95. and Drainage in South-East England (2nd edn). George Philip,
Wellman, P. (1987) Eastern Highlands of Australia; their uplift London.
and erosion. BMR Journal of Australian Geology and Geophysics Young, R. W. (1983) The tempo of geomorphological change:
10,277-86. evidence from southeastern Australia. Journal of Geology 91,
Wellman, P. (1988) Tectonic and denudational uplift of Australian 221-30.
Part V
Extraterrestrial landforms
 19
Planetary geomorphology

19.1 Approaches to planetary geomorphology understanding how our Earth alone among the planets and
moons of the Solar System came to develop its apparently
We conclude this survey of global geomorphology by ex- uniquely complex surface form.
panding our horizons to the earth-like planetary bodies of It is possible in this chapter to give only the very briefest
the Solar System. By earth-like we mean those planets and outline of the exciting findings that have come from the
large moons composed of a solid crust on which landforms various planetary orbiter and lander missions of the USA
can develop and be preserved. The term 'terrestrial' is re- and the USSR. This is a fast-moving field of research, and
served to designate those features actually occurring on the although spectacular landforms have already been revealed
Earth. The interpretation of surface planetary features is by the initial examination of the mass of imagery returned
carried out in part by analogy with terrestrial landforms, to Earth, further more detailed analyses will no doubt pro-
although some planetary landforms appear to have no ter- vide a much fuller picture of the development of planetary
restrial analogues. Many planetary bodies, such as the Earth's surfaces.
moon, retain landscape features formed very early in their Some of the global properties of the inner planets which
history, and their study provides a context in which to inter- directly or indirectly influence the operation of geomorphic
pret the long-term history of the Earth's surface. Moreover, processes are listed in Table 19.1. The mean distance of the
comparative planetary geomorphology gives a basis for Sun influences the amount of solar energy available at the

Table 19.1 Properties of the inner planets

EARTH MOON MARS MERCURY VENUS

Mean distance from the Sun (km x 106) 149.6 0.3844* 227.9 57.9 108.2
Period of revolution (Earth days) 365.26 27.32* 687 88 224.7
Rotation period (Earth days) 0.9983 27.32 1.026 59 243 retrograde
Equatorial diameter (km) 12756 3476 6787 4880 12104
Mass (Earth = 1) I 0.01226 0.108 0.055 0.815
Density (kg m- 3) 5':')0 3340 3900 5400 5200
Atmosphere (main components) Nitrogen, oxygen None Carbon dioxide None Carbon dioxide
Atmosphere (minor components) Carbon dioxide, None Noble gases, None Noble gases; hydrochloric
noble gases nitrogen hydrofluoric and sulphuric
acids
Mean temperature at surface (oq 15 107 (day) -23 305 (day) 480
-153 (night) -170 (night)
Atmospheric pressure at surface (millibars) 1000 0 6 <10- 9 90000
Surface gravity (Earth = 1) 1 0.16 0.38 0.37 0.88

* Relative to the Earth.

Source: Modified from B. Murray el al. (1981) Earthlike Planets. W. H. Freeman, San Francisco, Table 2.1 p. 28.
484 Extraterrestrial landforms

surface, and together with the rotational period and the nature in the seventeenth century two major types of terrain were
of the atmosphere this largely controls the average and identified - the high albedo, heavily cratered uplands and
extreme temperatures experienced on planetary surfaces. the low albedo, smooth, lightly cratered lowlands, called
Atmospheric pressure and temperature are two crucial vari- maria (singular mare) because of their supposed resem-
ables which determine whether water can exist in its liquid blance to seas). The Ranger, Lunar Orbiter and Apollo
state on a planet's surface, and consequently they deter- programmes of the USA, together with a number of Soviet
mine the nature of chemical and physical weathering and missions, have now provided an abundance of data on the
the presence or absence of fluvial activity. Surface gravity Moon. Although many questions remain, we now have a
is a significant factor in influencing atmospheric pressure fairly clear idea of the processes that have shaped the lunar
and indirectly in controlling the existence and nature of surface and an approximate chronology for its develop-
aeolian activity; moreover, it directly affects the operation ment, based on the radiometric dating of rock samples
of mass movement processes. There are in fact complex returned to Earth.
interactions between several of these factors, and we will Two major processes have fashioned the lunar landscape.
consider these further as we discuss the geomorphology of One is impact cratering, which involves the transforma-
individual planetary bodies. tion of the kinetic energy of an impacting object, or bolide,
and the consequent formation of impact craters and asso-
ciated features. Bolides can include comets and asteroids as
19.2 The Moon well as meteorites. The second significant process is volcanic
Knowledge of the broad outlines of the Moon's surface fea- flooding. The maria are plains formed by extensive sheets
tures extends back to the time when human beings first of basaltic lava, erupted mainly between 3.9 and 3.1 Ga BP.
looked heavenwards. Contrasts between light and dark areas Some 17 per cent of the surface is cover~d by lava plains,
are clearly visible with the naked eye and, following the but by far the greater proportion of these occur on the near
observations by telescope pioneered by Galileo and others side of the Moon. The far side, which is always turned away

NEAR SIDE FAR SIDE

N N

•
S

•
S

D MARE IMBRIAN NECTARIAN CRATERED HEAVILY

BASINS BASINS CRATERED

Fig.19.1 Maps of the near side andfar side of the Moon showing the major geological and morphological provinces. The location of
the landing sites for Apollo missions 14-17 are also indicated. (After K. A. Howard et al. (1974) Reviews of Geophysics and Space
Physics 12, Fig. 14 p. 322. Copyright by the American Geophysical Union.)
 Planetary geomorphology 485

from the Earth because of gravitational interactions arising

from irregularities in the form of the two bodies, is almost
completely covered by moderately to heavily cratered terrain
(Fig. 19.1).

19.2.1 Impact cratering

The Moon has experienced a history of bombardment ex-
tending back to its formation 4.6 Ga BP. Craters formed by
bolide impacts range from microscopic forms up to huge
circular basins, such as the Mare Imbrium, which have sub-
sequently been flooded by lava. A small number of craters
of probable non-impact origin have also been identified on
the Moon and are thought to be associated with volcanic
activity.
Impacts occur at tremendous velocities of up to 150 000
km h- I . The instantaneous conversion of the enormous
quantities of kinetic energy released into heat as the object
penetrates the surface, explodes and vaporizes means that
small objects can produce surprisingly large craters. For
example a 50 t meteorite about 3 m across can form a crater
150 m in diameter. As the kinetic energy of a projectile is Fig.19.2 View of the relatively young, 27 km diameter lunar
crater Euler located in the south-west of the Mare Imbrium. A
proportional to the square of its velocity, bolide size is less
central peak can be seen in the centre of the crater, probably
important than speed in determining the size of the crater formed as a result of rebound of the crater floor after impact. The
formed. effects of impact are seen to extend well beyond the primary
At the simplest level, lunar craters can be classified crater since in the Moon's low gravity field ejecta may travel for
according to their size and characteristic morphology into hundreds of kilometres before returning to the surface to form
large numbers of much smaller secondary craters. The bright
four categories. Microcraters « 10 mm in diameter) are bands that can be seen radiating away from the crater
found on rocks and boulders, and bombardment by very (particularly towards the top left) are known as crater rays. They
small meteorites is probably important in fracturing rocks are typical features of relatively young craters and are probably
and producing the fine-grained lunar regolith. Small craters composed offine material ejected on impact, including highly
(10 mm to 15 km across) have a simple form lacking major reflective fused silica (glass) beads formed by the shock waves
generated by the primary impact explosion. With age these bright
structural features, whereas large craters (15 to 300 km rays become darker and less conspicuous probably as a
across) are complex features and typically show evidence consequence of burial by ejecta from later impact events or
of modification by molten rock, either related to the changes in the crystal structure of the silica beads brought about
original impact or associated with subsequent volcanic by the intense ultra-violet radiation which reaches the unshielded
activity. Some contain one or more central peaks probably lunar surface. (Apollo 17 image, World Data Center Afar
Rockets and Satellites.)
formed through the rebound of the crater floor after impact
(Fig. 19.2).
In craters more than about 120 km across, a series of
on the Moon in the past. Most was associated with the
rings formed by uplifted masses of rock tends to replace the
formation of large impact basins, all of which were created
central peak. In very large craters (> 300 km across), usually
early in lunar history. The Orientale Basin is the most recent
termed basins, up to five rings may be found. Most lunar
and this was formed about 3.9 Ga BP. Puncturing of the
mountain chains are formed from these large circular struc-
early-formed crust by large bolides exposed the partially
tures and most large fracture systems are related to them.
molten interior and led to the eruption of lava flows which
The Montes Apennines, which reach some 5 km above
proceeded to fill the basins. Eruption of basalts appears to
the lunar surface, in fact represent part of the outermost
have continued for a considerable period after basin formation.
ring of the vast Imbrium Basin. The origin of these multi-
During the Apollo 15 mission to Hadley Rille, a sinuous
ringed basins is uncertain, but probably involved adjustments
channel cutting across the Mare Imbrium, individual lava
of the lunar crust after the impact of very large bolides
flows up to 60 m thick were observed. The Hadley Rille it-
during the early stages of the Moon's history.
self is thought to be a massive collapsed lava tunnel, smaller
versions of which are known from lava fields on Earth (Fig.
19.2.2 Volcanism and tectonics
1.1). Some idea of lava thicknesses can be gained by mea-
Although now quiescent, volcanic activity has been active suring the rim heights of craters partially buried by lava and
486 Extraterrestrial landforms

comparing these with their likely original heights. Such from bedrock, ejecta and molten rock from impact events
analyses indicate that basin lava thicknesses of 300 mare (the latter consisting largely of fused silica in the form of
typical. In addition to rilles, other landform a ociated with glass beads) and minor amounts of bolide fragments. The
volcanic activity include lobate flow fronts on lava flows, most effective bolide size for generating regolith is the range
low elevation shield volcanoes up to 15 km or so across, and which forms craters - about 10-1500 m in diameter. Larger
domed-shaped hills, which although of less certain volcanic bolides are too infrequent to be significant, while smaller
origin, may have been formed from lavas more viscous than objects are rarely able to penetrate the existing regolith cover
those which cover the mare basins. and thus create any additional regolith material. None the
Data from seismographs on the lunar surface show that less, small meteorites, down to those of microscopic dimen-
the current level of seismic activity on the Moon is minimal, sions, which can reach the lunar surface in the absence of
being only about one-millionth of that on the Earth. The an atmosphere, do playa role in pitting the rock surface and
Moon has in fact experienced little tectonic activity over smoothing features at the small scale. The effectiveness of
the past 3 Ga apart from crustal dislocations associated with this mechanism, however, is minimal in comparison with
impact events. Nevertheless, some tectonic structures have processes of terrestrial denudation. This is apparent from
been recorded; for instance, graben are locally abundant the preservation of 10m high features on lava flow over
an~ some mare ridges may not be of purely volcanic origin 3 Ga old observed by Apollo astronauts. The consequences
but rather true tectonic features associated with vertical of bolide impacts are not confined to fragmentation of the
movements within the sub-basin crust.

19.2.3 Surface materials and processes

Apart from forming craters, bolide impacts also create the
lunar regolith. This is made up of rock particles detached

Fig. 19.4 Boulder track on the slopes of the North Massif of the
Fig.19.3 The 18 km diameter lunar crater Dawes showing Moon near the Apollo 17 landing site. The track is nearly 1 km
clear evidence of mass movement in the form of extensive long and the largest boulders visible are about 5 m across.
accumulations of debris on the crater floor. (Apollo 17 image, (Apollo 17 image, World Data Center A for Rockets and
World Data Center A for Rockets and Satellites.) Satellites.)
 Planetary geomorphology 487

surface since the pressure and heat generated on impact is surface, and relating these variables to the radiometric dates
capable of fusing unconsolidated sediments together to form provided from the Apollo lunar samples (Table 19.2). There
breccia. are some uncertainties as to the accuracy of these dates
In the absence of the transporting agents of water, wind (different radiometric techniques' giving slightly different
and ice, movement of surface materials on the lunar surface results) and there are also discrepancies between the various
(excluding the effects of bolide impacts and lava flows) is approaches taken to relate the density of cratering to the
confined to mass movements. There is abundant evidence age of the surface. Nevertheless, the very low rate of modi-
for such activity; slumping associated with impact events is fication of the lunar surface compared with the Earth means
indicated by the morphology of the terraced inner walls of that we have a much clearer idea of its long-term geomor-
large craters (Fig. 19.3) and tracks left by large boulders phic history.
sliding down crater slopes have also been observed (Fig. After the formation of the Moon 4.6 Ga BP a process of
19.4). Landslides of unstable ejecta mantling crater wall differentiat'ion of its constituents led to the formation of a
slopes could be triggered by the seismic shocks generated crust which became sufficiently solidified by about 4.3 Ga
by large bolide impacts, but downslope movement may BP to record impact events. Most of this very early Pre-
also be precipitated by the subsequent impact of ejecta from Nectarian terrane (Table 19.2) is preserved as very subdued
such events. The scarcity of post-mare landslides, however, basins and craters on the far side of the Moon. Formation of '
suggests that the occurrence of mass movement may have the large lunar basins was concentrated between about 4.3
diminished after the early period of high cratering rates. and 3.9 Ga BP and ended with the creation of the vast
Imbrium Basin at the termination of the Nectarian System.
Prior to the Apollo missions the lunar highlands were
19.2.4 History of landscape development
thought to be remnants of the earliest accretionary phase of
The history of the lunar landscape has now been estab- the Moon's history when it was subject to continuous bom-
lished in broad outline by calculating the density of crater- bardment. Radiometrically dated Apollo samples, however,
ing and the freshness of craters over different parts of the suggest a predominant age of around 4 Ga which indicates

Table 19.2 Chronology of landscape development on the Moon

TIME-STRATIGRAPHIC UNITS AGE (Ga) ROCK UNITS EVENTS NOTES

Few large craters Craters with bright rays and sharp features at
Copernican System all resolutions (e.g. Tycho, Aristarchus)
Tycho
Aristarchus
Few large craters Craters with bright rays and sharp features
but now subdued at metre resolutions
(e.g. Copernicus)
Copernicus
Few large craters Craters with Copernican form but rays barely
Eratosthenes visible or absent
Eratosthenian System 3.2 Apollo 12 lavas
3.3 Apollo 15 lavas Imbrium lavas Few lavas with relatively fresh surfaces

3.42 Luna 16 lavas

Mare lavas Eruption of widespread iava Extensive piles of basaltic lava sheets with
Imbrian System 3.6 Apollo II lavas sheets on near side: some intercalated impact crater ejects sheets
few eruptions on far side
3.8 Apollo 17 lavas
3.9 Orientale Basin

Imbrium Ba in
Crisium Numerous overlapping large impaot craters
Nectarian System Muscoviense and associated ejecta sheets together with
Humorum large basin ejecta
Nectaris Basins
4.1 Serenitatis Any igneous activity at surface obscured by
Smythii impact craters
Pre-Nectarian Tranquillitatis
Nubium
4.6 Formation of Moon Crystalline rocks formed by early igneous
activity

Source: Modified from J. E. Guest and R. Greeley (1977) Geology on the Moon. Wykeham Publications, London, Fig. 1.5, p. 8.
488 Extraterrestrial landforms

that there was a period of frequent impacts by large bolides 19.3 Mars
well after the initial accretionary phase.
This episode of intensive impact cratering is known as The first close view of the Martian surface was provided by
the Late Heavy Bombardment. The dating uncertainties the Mariner 4 fly-by of 1964. Two more fly-by missions in
already mentioned have h~d to' a number of rather different 1969 by Mariners 6 and 7 provided further information, but
interpretations of the details of this event. One suggestion is detailed views of the planet's fascinating landforms had to
that the Late Heavy Bombardment represents the end of a await the arrival of the Mariner 9 orbiter towards the end
500 Ma period of continuous but declining impact activity. of 1971. At first the surface was tantalizingly obscured by a
Another view is that there was a short-lived cataclysmic huge dust storm, but after several months the atmosphere
increase in the rate of bombardment around 4 Ga. Whatever cleared to reveal some of the most spectacular landscapes
the correct interpretation, evidence of the Late Heavy in the Solar System.
Bombardment on the Moon is of crucial importance as it During a period of a single year the Mariner 9 orbiter
provides a time scale for similar events which appear to imaged almost the entire surface of Mars at resolutions of
have affected all the inner planets, and also possibly some 1-10 kIn, with selected areas being imaged at higher re-
of the more remote planetary bodies of the Solar System. solutions of up to 100 m. This mission provided the basis
Infilling of the large lunar basins by lava flows occurred for the initial mapping of the planet's surface and was
from about 3.8 to 3.2 Ga (Imbrian and Eratosthenian Systems). complemented by two Soviet orbiters in 1971 and 1973.
Geomorphic activity since then (Copernican System) has The Viking 1 and Viking 2 combined orbiter and lander
been confined to the formation of a few large craters and missions in 1976 provided more extensive high resolution
the continuing minor modification of the lunar surface through imagery totalling nearly 60 000 individual frames. In addi-
the impaCt of comparatively small bolides. . tion to monitoring changes on the surface during the slowly

NORTH
0 0 0 0 0 0 0 0 0 0 0
150 120 90 60 30 0 330 300 270 240 210
650 N r - - - - - , - - - - - - - , - - - - - - r - - - - - - , - - - - - , - - - - - - - r - - - - . - - - - - - r - - - - - , - - - - - , - - - - r - - - - - , 65 0 N

--+----+-(">;;;~~---_+_---+=_--__+--J2l--_1 50 0 N

EXPLANATION
o1....
I~!~! !:!:! ~ ~!:! ~:!:!~!I ~;~V~~s Cratered
1000 2000 km
-J...
I -----ll Scale at 650 _AureOle
• Rolling Plains

o
1
1000
I
2000 km
1 Scale at 50° D Northern Plains

o 1000 2000 km

~
_"-
I 1 1 Scale at 30° ~ ,'_,__ I..::: Knobby Terrain Ridge Plains
\r ..... I ; ' / , - , .

o 1000 2000 km Volcanic Plains

1
I 1 Scale at 0 0 of Tharsis I::::·:,::::·::::::j·:.::::i Volcanic Const ructs

Fig.19.5 Geomorphic map of Mars between latitudes 65°N and 65°5 constructed from Viking orbiter imagery. (Based on V. R. Baker,
1981, Progress in Physical Geography 5, Fig. 1, p. 474, modifiedfrom the Geologic Map of Mars by D. H. Scott andM. H. Carr (1978)
United States Geological Survey, Miscellaneous Geologic Investigations, Map /-1083.)
 Planetary geomorphology 489

changing Martian seasons the two Viking lander probes

provided close-up views of landforms around the landing
sites and the first analyses of the Martian regolith.
Maps constructed from Mariner and Viking orbiter imagery
indicate a wide variety of terrain types on Mars and the
operation of a range of geomorphic processes, including
fluvial and aeolian erosion and deposition, weathering, mass
movement and periglacial activity, as well as impact crater-
ing. At the very broad scale the planet may be divided into
two hemispheres separated by a great circle inclined at about
35° to the equator. The southern hemisphere is heavily
cratered, whereas the northern hemisphere is on average
lower and consists mostly of plains. The reason for this
contrast is not known. At a more detailed scale nearly a
dozen major morphological provinces have been identified
(Fig. 19.5; Table 19.3).

19·.3.1 Impact craters

Although impact craters are a very common element of
most planetary landscapes, many of those on Mars have
very special characteristics which make them unlike any-

Fig.19.6 The 18 km diameter Martian crater futy located at

Table 19.3 Principal morphological provinces of Mars
latitude 22°N, longitude 34°. Several sheets of ejecta can be
observed, each with well-developed lobes. These types of craters
MORPHOLOGICAL PROVINCE MORPHOLOGICAL CHARACTERISTICS
are common in a belt lying within 30-40° of the equator. Since
the small crater adjacent to futy has experienced only partial
Heavily cratered uplands Widespread hilly and cratered terrains of the burial the layers of ejecta must be relatively thin. (Viking 1 image,
equatorial and southern highlands
World Data Center A for Rockets and Satellites.)
Mountainous terrain Ejecta and uplifted blocks of ancient terrain
caused by large impacts
Knobby terrain Fretted uplands and isolated flat-topped thing observed elsewhere in the Solar System (Fig. 19.6).
residuals formed as erosional remnants The form of the ejecta, which occurs as overlapping lobes,
along the boundary between the heavily
cratered uplands and northern plains implies some form of fluid flow; indeed, these features have
Volcanic plains of Relatively young lava flows originating been termed 'splosh' craters. The most probable explana-
Tharsis from the Tharsis Montes volcanic province tion for this unusual morphology is that the material distur-
Rolling plains Intermediate age lava flows originating from bed on impact included subsurface ice (permafrost) which,
the Elysium volcanoes or other sources either as a gas or liquid, became incorporated into the ejecta
Northern plains Complex lowland exhibiting evidence of to form a debris flow. Secondary craters are also found
ice-contact volcanism, permafrost features
and aeolian modification which indicate that ballistic processes do operate to some
Ridged plains Relatively old lava plains that show extent. An intriguing finding is that the distribution of the
prominent ridges debris lobe type craters appears to be related to latitude and
Volcanic constructs Individual large volcanoes including altitude; this may be indirect evidence of a variation in the
Olympus Mons, Alba Patera and Elysium depth of permafrost over the planet, with ordinary ballistic
Mons
impact type craters only forming where the permafrost is
Aureole Complex terrain of elongate hills and ridges
occurring around the margins of certain confined to great depths.
volcanoes
Chaotic terrain Blocky, fractured terrain apparently
developed through the collapse and 19.3.2 Volcanic and tectonic features
subsidence of the heavily cratered terrain
On the basis of our present knowledge Mars possesses the
Valleys and channels Includes the highly modified landscape of
Valles Marineris and various valley and largest volcanoes in the Solar System. Olympus Mons
channel features attributed to fluid flow reaches 26 km above the surrounding plain from which it is
erosion separated by a scarp up to 8 km in height. A number of
Source: Based on terrain types identified in Fig. 19.5 and described in collapse structures are found around the summit, and the
V. R. Baker (1981) Progress in Physical Geography, 5, pp. 474-6. caldera complex alone is some 80 km across (Fig. 19.7).
490 Extraterrestiallandforms

Fig. 19.7 Oblique view of Olympus Mons, one of the largest volcanoes known in the Solar System. Based on the density of
cratering it has been estimated that Olympus Mons is some 400 Ma old, although this may simply be the age of the youngest flows, the
whole edifice probably having been built over billions ofyears. By comparison, volcanoes on Earth rarely survive more than afew
million years before being destroyed by erosion. (Viking image, World Data Center Afar Rockets and Satellites.)

The morphology of Olympus Mons is broadly comparable

to the shield volcanoes of Hawaii, but these two landforms
differ dramatically in size with Hawaii (the largest shield
volcano complex on Earth) rising only a mere 10 kIn from
the ocean floor! (Fig. 19.8). The great height of some
Martian volcanoes should not surprise us becau e the lower
surface gravity on Mars (only 38 per cent of that of the
Earth) means that large volcanic edifices there weigh less
than their terrestrial counterparts and consequently do not
load the underlying crust to the same extent.
The larger Martian volcanoes are all of the shield type
and are associated with massive lava flows. Alba Petera is
1500 km across and dwarfs even Olympus Mons. It has a
modest elevation of only 6 km so its flanks are inclined at a
very shallow angle. It was thought initially to be an ancient
degraded tructure, but Viking imagery has revealed nu- Fig.19.8 Olympus Mons (left) compared with the shield
merous juvenile feature on its slopes including sheet flows volcano of Hawaii (right)." Olympus Mons is 650 km in diameter
and lobe fronts. These large shield volcanoes are concen- and is composed of thousands ofindividuallavaflows./ts caldera
complex alone is over half the size of the whole subaerial part of
trated in an area of probable 'mantle' upwelling known as Hawaii shown here. (Olympus Mons - Viking 1 image, World
the Tharsis Montes volcanic province (Fig. 19.5). This is a Data Center Afar Rockets and Satellites; Hawaii - Landsat
massive crustal bulge which formed about 3.9 Ga BP and it mosaic courtesy N. M. Short.)
 Planetary geomorphology 491

MARS - Tharsis Montes Volcanic Province sequently, the amount of near-surface water declined and
later volcanism was confined to lava eruptions and the for-
E-100
mation of dome and shield volcanoes.
.>I.-zoo A range of tectonic features have been identified on the
-300 Martian surface, many in association with volcanic struc-
5,000 10,000
km
tures. Far and away the most impressive of these is the
system of canyons collectively called the Valles Marineris.

-
Increasing mantle density This complex graben-like feature is some 5000 km long,
averages around 200 km wide and is up to 7 km in depth.
Some idea of the size of this landform can be gained by
imagining a canyon four times as deep and about fifteen
times as broad as the Grand Canyon, stretching coast to
coast across the USA. The Valles Marineris has clearly
formed through the effects of large-scale tensional stress in
the Martian crust and its possible association with the Tharsis

-
Increasing mantle
density
bulge suggests that it may be of a similar age.

Fig.19.9 Combined crustal density (Pratt) and crustal thickness

(Airy) isostatic model for the Tharsis Montes volcanic province of 19.3.3 Weathering
Mars in comparison with a similar model for the Basin and Range Although contrasting starkly with the Earth, the atmosphere
Province in the western USA. (Modifiedfrom N. H. Sleep and R. 1.
Phillips (1979) Geophysical Research Letters 6, Fig. 2, p. 805. and surface conditions on Mars are closer to our own planet
Copyright by the American Geophysical Union.) than any of the other bodies of the Solar System. In spite of
marked differences in atmospheric pressure, atmospheric
origin may be similar to that of the Basin and Range Province composition and surface temperatures (Table 19.1) there are
of the western USA (Fig. 19.9). some apparent similarities in weathering processes. Analyses
In addition to extensive lava flows there is evidence in carried out by the Viking landers, although failing to con-
the form of blankets of tephra that many Martian volcanoes firm the presence of life, provided valuable data on the
have experienced explosive episodes early in their history. physical and chemical properties of surface materials. Some
This could be explained by the explosive consequences of of the boulders observed in Viking lander imagery (Fig.
magma penetrating a frozen, water-saturated regolith. Sub- 19.10) may have been formed by in situ weathering, but it

Fig.19.10 Ground view across Chryse Planitafrom the Viking 1 landing site on Mars. Large numbers of boulders, up to 1 m or so
across, can be seen in addition to drifts offine-grained material. (Viking 1 mosaic, World Data Center A for Rockets and Satellites.)
492 Extraterrestrial landforms

is more likely that they are simply ejecta from impact events.
A possible explanation for the formation of the fine-
grained material which mantles much of the surface at both
the Viking landing sites is salt weathering since the Martian
soil appears to be rich in sulphur, probably in the form of
sulphate salts. The presence of a thin salt crust (sometimes
misleadingly described as 'duricrust' by planetary geolo-
gists) is also indicated from the observation that surface
material broke into small cohesive fragments where it was
disturbed by the sampling arm attached to the Viking landers.
Frost weathering may also be significant on Mars since
both water and carbon dioxide ice have been identified on
the surface, but as with other Martian geomorphic processes
there is far more speculation than conclusive data. The closest
terrestrial analogues are the cold, arid valleys of Antarctica.
There is certainly a seasonal cycle of frost deposition on
Mars and a thin, bright surface covering was seen to form
during winter at the Viking 2 lander site.

19.3.4 Slopes and mass movement

The abundance of high escarpments on Mars.provides con-
siderable scope for the operation of mass movement mech- Fig.19.11 Large landslides on the south wall of Gangis
anisms; certainly the scale of such processes exceeds any- Chasma in the Valles Marineris region of Mars. The escarpment
thing occurring on Earth. In the Valles ~arineris concave running left to right across the centre of the image is about 2 km
high and separates a plateau region (top)from the floor of
chutes separate the comparatively flat upland surface from
Gangis Chasma (bottom). Some of the landslides show backward
the canyon itself (Fig. 19.11). These chutes lead down rotation (arrowed), afeature characterizing rotational slumps on
either into talus slopes or into a 'spur-and-gully topography' Earth. The area shown is about 60 km across. (Viking I mosaic,
made up of alternating ridges and swales. The talus slopes World Data Center A for Rockets and Satellites.)
typically have a mean angle of around 30° and in some cases
have a vertical drop of several kilometres. In a study em-
19.3.5 Aeolian processes and landforms
ploying Viking imagery, 25 large landslides ranging in size
from 40 to 7500 km2 were identified in the Valles Marineris; With a lack of surface water, the absence of any vegetation
one of these was seen to be recessed up to 30 km into the to stabilize the surface, and the presence of high winds and
escarpment wall. fine sediment it is not surprising that aeolian landforms are
One explanation for these massive landslide features abundant on Mars (Fig. 19.12). There is a vast erg sur-
involves dry avalanching of unconsolidated material. This rounding the north polar region, wind-laid dust and ice
implies that, at least in some areas, the surface is covered deposits at the poles themselves and numerous depositional
.by a thick, non-cohesive regolith possibly resulting from and erosional forms in the equatorial zone. Although now
the shattering of bedrock through bolide impacts. Another, recorded in detail by the imaging systems orbiting Mars,
now more favoured, hypothesis is that the undermining of the great dust storms that periodically sweep the planet had
free faces by sapping, involving the evaporation of ground long been observed from the Earth through high-powered
water or permafrost ice, is the primary landslide-triggering telescopes. Such storms may shroud the whole planet, as
mechanism. This process may also have been instrumental occurred when Mariner 9. arrived in 1971, and can last for
in the formation of the fretted terrain located along the up to four months. Wind speeds can exceed 140 m S-I (around
boundary between the ancient heavily cratered highlands 500 km h- I ) compared with a typical maximum of 30-40 m S-I
and the sparsely cratered lowland plains. This transition (about 120-160 km h- I ) on Earth; dust may rise up to 70 km
zone consists of vast debris aprons with a characteristic above the Martian surface.
convex lobate form and appears to represent an area of out- The much lower atmospheric pressure on Mars compared
ward flow of debris from highlands to lowlands. The fairly with Earth (Table 19.1) has important implications for the
uniform height of the eroding escarpments within the fretted operation of aeolian activity. Threshold drag velocities, for
terrain at between I and 2 km has been interpreted as indi- instance, are much larger on Mars (Fig. 19.13); consequent-
cating that scarp retreat may be controlled by the depth to ly saltating grains have about twenty times the velocity of
which ice-rich permafrost extends. similar-sized particles on Earth and the potential for aeolian
 Planetary geomorphology 493

10

'I Mars
UJ 0.115mm
E
-; 10
'0
Q
~
Ol
~
0
"0
0
.r:
UJ Earth
~ 0.075mm
.r:
f-
01

0.075mm
Fig.19.12 Dunes in the ancient cratered terrain ofMars
located at latitude 47°5, longitude 340°. Most of the dunes are of 0.01 +- ,.-_ _.-- .--_--,
the transverse ridge and barchanoid ridge type, but isolated 001 0.05 0.1 0.5 1.0
barchans are also visible (arrowed). The area shown is about Particle Size (mm)
60 km across. (Viking 2 image, World Data Center A for Rockets
and Satellites.) Fig.19.13 Threshold drag velocity in air as afunction of
particle size on the Earth, Mars and Venus. Note that for 0.2 mm
sized particles this velocity may be ten times greater on Mars than
on Earth. Saltating grains on Mars have about ten times the
erosion is therefore considerable. Silt to clay-sized particles momentum and one hundred times the kinetic energy of those on
seem to predominate on the Martian surface, but sand-sized Earth. The reasons for the higher threshold drag velocity for
aggregates can apparently form through the electrostatic particle sizes less than about 0.1 mm are not known for certain
bonding of this finer material. The dunes observed on Mars but are probably associated with interparticle cohesion and
aerodynamic effects. (Modifiedfrom R. Greeley (1985) Planetary
are probably formed largely from these aggregates and are Landscapes, Allen and Unwin, London, Fig. 3.36, p. 66.)
therefore in some respects analogous to clay dunes on
Earth. Wind-tunnel experiments suggest, however, that such
sand-sized aggregates may have a short life span since at
slightly above Martian threshold drag velocities they rapid- Aeolian erosional landforms include modified crater rims,
ly break up into fine fragments around 20!-lm in diameter. linear grooves, streamlined ridges and fluted cliffs. An
The image of the destruction of these aggregates as they aeolian origin is implied where parallel series of linear
smash into rock surfaces at high velocities has given rise to ridges and grooves extend over large distances, some of the
the vividly descriptive term kamikaze grain. forms seen on Mars being morphologically comparable to
Inspection of Mariner and Viking orbiter imagery has terrestrial yardangs. The distribution of aeolian erosional
revealed a wide variety of both erosional and depositional landforms is localized and this suggests that they may only
aeolian forms on Mars. Morphologically, many of these are form where there are friable surface materials. Estimates of
very similar to landforms recorded from terrestrial deserts, rates of aeolian erosion on Mars range from a negligible
but the great size of some features exceeds any analogous 0.001 mm ka- 1 for the lowland plains to a maximum of
forms occurring on Earth. Various dune patterns have been around 0.1 mm ka- 1 in parts of the heavily cratered high-
identified; those of the Hellespontus area, for instance, have lands. Although the apparent lower density of craters at high
a crescentic ridge form reminiscent of terrestrial dunes in latitudes may be due in part to the blanketing effect of
central Asia. Massive barchanoid and transverse forms have aeolian deposits, and thus imply the active deflation of
also been recorded as well as the coalescence of individual material from elsewhere, measurements from the Viking
barchans, a feature well known from terrestrial deserts (Fig. landers have demonstrated that current rates of aeolian ero-
19.14). sion and deposition are extremely low.
494 Extraterrestrial landforms

polygon formation on the Earth. Alternative explanations

for these phenomena exist, however, including the contrac-
tion of cooling lava and extension associated with tectonic
mechanisms.
The present water content of the Martian permafrost is
not known but there is abundant morphological evidence
for a high ice content in the past. Likely thermokarst fea-
tures include alases (irregular depressions) and thermocir-
que topography (embayed escarpments; see Section 12.3.2.4).
Such landforms, arising from the melting of ground ice, are
best developed in terrestrial periglacial environments where
the ice content is high. The chaotic terrain occurring in
parts of the Martian equatorial zone has been widely inter-
preted as the result of large-scale ground collapse but the
Fig.19.14 The arm of a migrating barchan (A) on Mars size of these features on Mars is much greater than ter-
coalescing with another barchan downwind (arrowed) restrial thermokarst forms (Fig. 19.15).
The labelled barchan is about 500 m across and 200 m long and
is situated near the southern edge of the erg located at latitude There is a clear latitudinal control over the operation of
73°N, longitude 39°50'. All the dunes visible are migrating permafrost-related processes on Mars. Estimates of the
towards the bottom of the image, that is, in an approximately likely heat flow from the planet's interior indicate that the
easterly direction. (Viking 2 image, World Data Center A for maximum permafrost depth ranges from about 1 km at the
Rockets and Satellites.) equator to around 3 km at the poles. Under the present
climatic regime permanent ice can probably exist to within
a few centimetres of the Martian surface at latitudes above
19.3.6 Periglacial processes and landforms
about 40°. Local melting of ice-rich permafrost to produce
It could be argued that 'periglacial' is an inappropriate term thermokarst and related landforms could result from a
to apply to geomorphic processes on Mars since there is no variety of mechanisms including volcanism and bolide
firm evidence that glaciers have ever existed on the planet. impacts. More widespread melting might be precipitated by
None the less, in its broader sense, as applied in Chapter climatic changes, while scarp retreat could lead to the release
12, the term is appropriate as it refers to environments of liquid water confined at depth within the permafrost
characterized by low temperatures with fluctuations above layer.
and below freezing. Although water in its liquid state is
known not to exist in significant quantities on the present-
19.3.7 Polar terrains
day Martian surface, locally it may occur at shallow depths
in the regolith in the equatorial zone. But freeze-thaw acti- The frost caps of Mars expand to cover up to 30 per cent of
vity and associated frost creep may have been widespread the planet's surface area during the Martian winter but re-
in the past when mean temperatures were probably rather treat to small residual 'ice caps' covering only 1 per cent of
higher and surface water may have been more abundant. If its area in the summer. The remnant north polar ice cap is
salt is an important constituent of the present-day surface, largely composed of water ice, but the nature of the perma-
freeze-thaw may be currently active since liquid saline nent southern cap is less certain and it may consist of both
solutions might be able to exist during the day at low frozen water and carbon dioxide. The seasonal frost blanket
latitudes. is predominantly formed by the freezing of atmospheric
The possible extent of periglacial activity on Mars is carbon dioxide, although it probably also includes dust par-
suggested by the presence of patterned ground recorded in ticles coated by water ice. ,
high resolution Viking imagery. Although the polygonal The perennial ice caps contain l~yered deposits which
patterns observed are certainly similar to terrestrial patterned have been partially exposed by the formation of deep can-
ground, orne occur at a much larger scale than analogous yons. This is the layered terrain in which alternating
forms on Earth. Ice-wedge polygons on Earth are typically bright and dark bands up to 30 m thick have been observed
1-100 m across, but some of the Martian polygons are exposed along canyon walls. This banding may represent
5-10 km in diameter, the cracks themselves being hundreds ~hanges in the ratio of dust to ice deposited on the ice caps,
of metres across. If they are indeed periglacial phenomena and it has been suggested that such fluctuations ultimately
it is likely that they have been generated by temperature have a climatic origin. Indeed, it has been speculated that
changes of considerable duration extending to great depth Mars may have experienced cycles of climatic change similar
within the Martian permafrost, rather than by the seasonal to those on Earth during the Late Cenozoic Ice Age and that
temperature fluctuations largely responsible for ice-wedge there may even be a common cause in the Milankovitch
 Planetary geomorphology 495

Fig.19.15 The head of a Martian outflow channel containing an extensive area of chaotic terrain (right centre oj image). Detailed
examination of this region has revealed slumped and collapsed blocks up to 10 km long lying at the base of steep escarpments rimmed
by arcuate features. Vast quantities of debris have clearly been removed and this may have been accomplished by the catastrophic
release of water from the melting of ice-rich permafrost. The area shown is about 300 km across. (Viking 1 mosaic, World Data Center
A for Rockets and Satellites.)

mechanism (see Section 14.3.2). The lack of fresh impact with their concentration within 30° of the Martian equator.
craters in the layered terrain certainly indicates that in As we shall now see, there is no shortage of alternative
Martian terms the surface features in this region are rela- hypotheses to explain these and other channel forms on
tively young and that the layered deposits must be accumu- Mars.
lating rapidly compared with the recent cratering rate.
As already mentioned, there is no unequivocal evidence
for glacial erosion on Mars and it certainly seems that the 19.3.8 Channels
existence of flowing ice is ruled out in the present Martian No landforms on the Martian surface have generated more
environment. Nevertheless, glacial processes have been controversy than the range of channel features first seen on
proposed as an explanation for the so-called outflow chan- Mariner 9 imagery and subsequently investigated in detail
nels on Mars (see Section 19.3.8) because their forms are, using the higher-resolution Viking orbiter data. In nearly all
in some respects, similar to terrestrial glacially eroded cases these forms are most accurately described as valleys,
valleys. An argument in favour of the glacial theory is the but as the term 'channel' has gained wide currency among
comparable scale of the Martian and terrestrial forms, but a planetary geomorphologists we will retain it here. Three.
glacial origin for outflow channels is difficult to reconcile main types of channel have been recognized.
496 Extraterrestrial landforms

terns; they are found throughout the older cratered terrain

(Fig. 19.16). When fir t observed they were simply attri-
buted to urface runoff processe and were thus considered
to be ancient features related to a past pluvial epoch on
Mars. The junction angle of these networks are, however,
much less acute than in most terrestrial fluvial systems. By
Fig.19.16 Nirgal Vallis, an 800 km long runoff channel on analogy with morphologically similar forms which occur in
Mars located at latitude 28°S, longitude 40° and incised into old certain environments on the Earth, such as on the Colorado
cratered terrain. See Figure 1.2 for detail from part of the Plateau in the south-west USA (Fig. 1.3), headward erosion
channel network towards the left-hand edge of the area shown by spring sapping is now thought to be a more likely mode
here. (Viking 1 mosaic, World Data Center A for Rockets and
Satellites.) of formation (Fig. 1.2). Meltwater from ice-rich permafrost
is the likely source of the required water.
Outflow channels are much larger features tens of kilo-
Fretted channels have wide, l1)ooth floors and steep metres wide and hundred of kilometres long (Fig. 19.15).
walls and are found in the fretted terrain marking the tran- They generally lack tributaries and most originate within
ition between the heavily cratered uplands and the lightly chaotic terrain or large concave depressions. Of particular
cratered lowland plains. The presence of large debris flows interest is the variety of bedforms seen on the floors of
where these channels extend into the lowlands indicates that outflow channels including scour marks, longitudinal grooves
they have formed through escarpment retreat by mass and streamlined upland remnants (Fig. 19.17). Scour marks
movement. and other erosional and depositional forms extend into the
Runoff channels, which are connected to form dendritic plains on to which the mouths of outflow channels open.
networks, are in many ways similar to terrestrial fluvial sys- Although they are widely regarded as having been formed

Fig.19.17 A streamlined 'island' about 40 km in length representing an eroded remnant of a former plateau located at the mouth of
Ares Vallis, an outflow channel on Mars located at latitude 20 0 N, longitude 31°. The flow from the lower left to the upper right
diverged around the crater rim forming a tapering tail downstream. (Viking 1 mosaic, World Data Center A for Rockets and Satellites.)
 Planetary geomorphology 497

by some kind of fluid flow there is little agreement as to the Most surviving
large (>SOkm)
particular fluid agent involved. The erosive agents pro-
-----
impact craters and basins
p6sed include water, ice, wind, debris flows, liquefied crustal
material and low-viscosity lava. Alba Patera ~
small Tharsis volcanoes \:
Detailed comparisons between the form of outflow Hadriaca Patera _ _ _ _ _ _ _ _O_ly'---m-'-p_us_M_o_ns~ ~
Tyrrhena Patera ..
channels and the morphology of the Channeled Scabland t-ormatlon of
Tharsis b~lge Canyon formation ~
of eastern Washington and Oregon in the USA (see Section --------"=.!...::..:.:..cc"-"'-'-'-'--------.J ~
z
II A.! ) have led to the suggestion that the Martian channels Circum·Tharsis fractures
i;j
similarly owe their form to the action of catastrophic floods.
Although the outflow channels are generally much larger
than their terrestrial equivalents, this may be explained by
Runoffchannels Fretted channels (mass movement?)
the significantly different conditions of surface gravity and
atmospheric pressure on Mars. The lower surface gravity, Outflow channels

for instance, could allow the transportation of cobble-sized

3000 20bO lObo
material in suspension, and it is likely that catastrophic Density of Craters > , km per 106 km 2
floods could attain a sediment concentration of up to 60-70
per cent by volume. Formation of the outflow channels could,
Approximate Age (Gal
therefore, have been accomplished very quickly, perhaps in
a matter of a few days or weeks. Even accepting the domi- Fig.19.18 A possible chronology for the Martian landscape.
nant role of some kind of water flow in sculpturing their (Modified from M. H. Carr (1981) The Surface of Mars. Yale
overall form, outflow channels have clearly also been exten- University Press, New Haven, Fig. 16.1, p. 202.)
sively modified by a range of other processes, including
thermokarst development, impact cratering, aeolian erosion, of channel formation from 4 to 3.7 Ga. Another chronology
spring sapping and a variety of mass movement processes. suggests more recent activity with highland lava emplace-
If outflow channels are accepted as essentially water- ment between 3.9 and 0.3 Ga BP and Tharsis region volcanism
eroded landforms, a crucial problem remains the source of starting about 0.5 Ga BP. More detailed investigations of
such water which clearly must have been present in very cratering rates in the outflow channels suggest that they
large quantities. An initial proposal was that volcanic erup- formed from 0.5 to 2 Ga BP, while the runoff channels are
tions below glaciers led to rapid melting and the cata- considered to be rather older. The most modem features of
strophic release of water, but this hypothesis requires the the Martian landscape are some of the dune formations
existence of significant ice sheets outside the polar regions. which, as repeated Viking orbiter observations show, are
A related idea is the melting of ice-rich permafrost asso- currently active. Figure 19.18 gives a very tentative chrono-
ciated with volcanic activity, although this would be a far logy for the main landforming episodes on Mars which will
less efficient process for rapidly generating large quantities no doubt be considerably revised when rock samples even-
of meltwater. tually become available for radiometric dating.
Another hypothesis involves the catastrophic release of
artesian water under high pressure. This possibility arises
from the postulated existence of an aquifer system extend- 19.4 Mercury
ing from about 1 krn to 5 or 10 krn below the Martian surface Before the fly-by Mariner 10 mission in 1974 Mercury was
and confined by a thick, impermeable, overlying permafrost the least known of the inner planets. Because of its com-
layer. This mechanism has the advantage of simultaneously paratively small size and proximity to the Sun no surface
accounting for catastrophic water release and the formation details could be discerned from the Earth. The Mariner 10
of chaotic terrain at the head of outflow channels through fly-by provided adequate coverage of nearly half of the
the collapse of the surface on the removal of the underlying planet at resolutions from 4 krn to 100 m. First impressions
water. are that the surface is very like that of the Moon (Fig.
19.19). Impact craters are the predominant landform and
most of the features associated with lunar craters are also
19.3.9 History of landscape development found on Mercury. Minor differences can be attributed largely
In the absence of dated surface materials from samples to the higher surface gravity of Mercury; ballistic ejecta, for
returned to Earth, the chronology of landscape development instance, are found closer to the primary impact crater.
on Mars rests on estimates of cratering rates. A range of A major contrast with the morphology of the Moon,
ti"me scales have been proposed; one of these gives the age however, is provided by the occurrence of lobate scarps.
of the oldest crust as 4.2-4.3 Ga, volcanic activity in the These appear to be associated with thrust faults resulting
Tharsis region at between 3 and 1 Ga BP, and the major phase from crustal compression of global extent. It has been sug-
498 Extraterrestrial landforms

been identified on Mariner 10 imagery, indirect evidence

suggests that, like their lunar counterparts, these plains are at
. least primarily of volcanic origin. There are, for instance,
many examples of material from the smooth plains filling
craters clearly formed at an earlier stage.
In the absence of the kind of radiometrically based land-
scape chronology established for the Moon, any dating
scheme for Mercury must rest on the assumption that it has
experienced a lunar-type history of bombardment. Although
there are similarities between the size frequencies and areal
densities of impact craters on Mercury and the Moon, it has
been questioned whether this necessarily indicates a tem.
poral equivalence in bombardment history. It is perhaps
safer, therefore, to talk in terms of a relative rather than an
absolute landscape chronology for Mercury (Fig. 19.20).

19.5 Venus
Similarities in size, density and distance from the Sun make
Venus in these respects the closest equivalent of Earth in
the Solar System. There is a stark contrast, however, in the
atmospheric properties of the two planets. Venus is blanketed
by a thick atmosphere with a surface pressure some 90 times
that of the Earth, while surface temperatures are around
480°C. The dense atmosphere has prevented direct visual
observation of the planet's surface from the Earth or by
conventional orbiting satellite imaging systems. Instead our
limited knowledge of the surface has so far come from the
Soviet Venera landers and both Earth-based and orbiting
satellite-based radar altimeters which are able to penetrate
the atmosphere and provide data on variations in elevation
Fig.19.19 The surface of Mercury as seen on approach during
over the planet.
the Mariner 10 fly-by mission. (Mosaic from the Atlas of
Mercury, Fig. 18, World Data Center A for Rockets and The presently available radar data are of a resolution
Satellites.) sufficient to pick out major morphological features, but in-
adequate to provide the detailed altitudinal information
needed to determine unequivocally whether Venus has an
gested that an episode of global contraction may have
resulted from the cooling and shrinking of Mercury's large
core and the stresses imparted by the slowing of the planet's

~
high
Impact
rate of rotation due to its gravitational interaction with the Rate
Sun.

------_
low

~2~
Formation of
Impact features on Mercury range in size from the Crater 5-3 1
Degradational Classes
smallest craters resolvable in Mariner 10 imagery up to the
Plains
Caloris Basin, a complex feature more than 1300 km across I~ ___ ~.....
•..• >
Formation
encircled by mountains up to 2 km high. Antipodal to the Caloris
Caloris t Impact.
Caloris Basin is an anomalous region of hilly and lineated Related
Hilly and Lineated Terrain

terrain which appears to disrupt earlier landforms. The crea- Scarp

.. .... .........
Formation
tion of this weird terrain, as it has been termed, may have
~
Thermal Melting Cooling
occurred as a consequence of the focusing of seismic waves History
&
Expansion
&
Contraction
generated by the massive impact event which formed the Time-_
Caloris Basin itself.
Fig.19.20 A relative chronology for the development of the
The smooth plains of Mercury, which cover about 15 per surface features of Mercury. Note that other interpretations are
cent of its surface, are similar in appearance to the lunar possible. (After R. G. Strom (1979) Space Science Reviews 24,
maria. Although no unequiv'ocal volcanic features have Fig. 38, p. 65.)
 Planetary geomorphology 499

>-
....
12 latter has arisen largely from a combination of density con-
~
E 10
A trasts between continental and oceanic crust, and the effects
« of plate tectonic mechanisms in continually generating new
~
§ dense oceanic crust and sweeping lighter material on to the
Cl: 8 continental margins. We can surmise, therefore, that if plate
~ tectonics does occur on Venus then it must be of a rather
"0
~
Q)
6 different form from that active on the Earth.
>
u
0 Examination of the gross topography of Venus does in
co
~ 4 fact suggest a number of features with at least a superficial
« resemblance to tectonic structures on the Earth (Fig. 19.22).

l
'0
Q) For instance, a plateau feature known as Beta Regio appears
OJ
co 2
c: to be a large domal uplift traversed by a rift with associated
Q)

~ 0
) volcanoes, and indeed it has been likened to the East African
6057 6062
Rift System. There are also extensive arcuate and linear
6042 6047 6052
Radius (km) troughs concentrated in the upland massifs, but these seem
to be more analogous to Martian canyons. The extensive
9 elevated area named Terra Ishtar reaches up to 11 km above
ttl
~
«Q)
B the median elevation of Venus, and around this massif are a
number of depressions which appear to be of volcanic ori-
~:J 6 gin. Another significant feature is a huge cone about 600 km
(f)
r/)
:.c: across with an apparent summit caldera and this has been
t:
compared with the Martian shield volcano Olympus Mons.
-
ttl
W
0 3 Numerous crater-like forms have also been recorded in the
Q)
OJ
ttl
lowland areas; some of these may be volcanic, others impact
C
Q) craters, but most are of uncertain origin.
~
8:. Further information bearing on the possibility of plate
0
-8 -4 0 4 8 tectonics on Venus has been provided by the Soviet Venera
Elevation Above Sea Level (km) 13 and Venera 14 landers. Chemical analyses of surface
9
materials by these probes indicate the existence of tholeiitic
co
~ basalts at the Venera 14 landing site, rocks typical of mid-
«
Q)
<.l
C oceanic spreading ridges on the Earth, and the presence of
ttl
t:
:J 6
high-alkali basalts characteristic of terrestrial continental hot-
(f)
.r/)
spot volcanism at the Venera 13 site some 1000 km away.
"€co Although the presently available topographic data do not pro-
w vide unequivocal evidence of the existence of plate tectonic
'0 3
Q)
OJ
structures on Venus, estimates of the likely rate of heat flow

\J~
ttl
to the Venusian surface imply that a very large number of
~
~ volcanic hot-spot centres must be present if spreading ridge
8:. J
0 systems are not available to dissipate heat.
-8 -4 0 4 8 Images of the surface returned by the Venera probes
Elevation Above Sea Level (km)
show a variety of terrains. The Venera 14 site is unusual in
Fig.19.21 Hypsometric curves for Venus (A) and the Earth (B lacking unconsolidated material, but the other sites display
and C). Curve (B) represents the Earth's actual topography, fine to coarse debris resting on bedrock. Evidence of ero-
whereas curve (C) represents the hypsometry with the loading sion of rock surfaces and the presence of angular rock frag-
effect of the oceans on the crust removed, the latter curve being
the more appropriate for comparisons with Venus. (After 1. W. ments in Venera lander images indicate that the Venusian
Head et al. (1981) American Scientist 69, Fig. 3, p. 617.) surface continues to experience geomorphic activity. Data
on the extraordinary atmosphere provided by the Venera
probes certainly suggest that chemical weathering should
Earth-like topography dominated by the effects of plate tec- be extremely active. Since the partial pressure of H20 in-
tonics, a Mars-like surface characterized by impact craters creases with elevation the most vigorous weathering prob-
and volcanic landforms or a combination of the two. Never- ably occurs in the upland regions.
theless, one major morphological distinction is clear; the In the absence of liquid water on the surface, at least in
hypsometric curve for Venus displays a single peak in con- recent geological time, wind is presumably the main denu-
trast to the Earth's bimodal distribution (Fig. 19.21). The dational agent. Theoretical calculations, in conjunction with
500 Extraterrestrial landforms

o· 20· 40· 60· 80· 100· 120· 140· 160· 180· 200· 220· 240· 260· 280· 300· 320· 340· 360·

70· 70·

60· 60·

50· 50·
40·
UJ
o 30·
::J
f-
20·
~
....J 10·
O·
-10·
-20·
-30· -30·
-40· -40·

-50· -50·

-60·

o· 20· 40· 60· 80·

------------
100· 120· 140· 160· 180· 200· 220· 240· 260· 280· 300· 320· 340" 360·
LONGITUDE
-60·

• >1 km above median elevation

rU!i[:[iI[i!il :!: 1km of median elevatk)n

D >1km below median elevation

Fig. 19.22 Major topographic features of Venus based on radar altimetry. Note that the Mercator projection used greatly exaggerates
the area covered by features at high latitudes. (After G. E. McGill (1982) Nature, 296, Fig. 2, p. 14.)

satellite data, indicate low wind speeds on the surface of which are of planet-sized dimensions. Until the Voyager
Venus of around 1 m S-I, but because the atmosphere is so programme of planetary exploration which began in the late
dense particles up to 10 mm across can probably be trans- 1970 little was known of the surface features of the rocky
ported (Fig. 19.13). Earlier in its history Venus may have and icy satellites of the outer planet , and the first close-up
been capable of retaining liquid water on its surface and so images of the moons of Neptune were only received from
ancient fluvial features may yet be found. Of crucial impor- Voyager 2 in 1989. The images that have so far been returned
tance to an understanding of the evolution of the Venusian from the hugely successful Voyager probes have revealed
surface is whether large impact basins remain from the Late landscapes with extraordinary characteristics, and which, in
Heavy Bombardment. This would indicate the extent to many cases, have no terrestrial analogues.
which endogenic and exogenic processes have subsequent-
ly modified the surface, but answers to these and other
intriguing questions about the smaller-scale morphological
19.6.1 The Galilean moons of Jupiter
features of the planet will have to await the return of high
resolution radar imaging data from the Magellan probe The four large (Galilean) satellites of Jupiter are compar-
launched from the space shuttle Atlantis in May 1989. able in size to Mercury and our own Moon and represent
extremes of planetary processes and morphology (Table
19.4). Each has feature preserved from different periods in
19.6 The moons of the outer planets its history, and each exhibits the effects of quite distinct
Beyond the inner Solar System lie the giant planets Jupiter, geomorphic processes. 10, the innermost of the four, is one
Saturn, Uranus and Neptune. Although the e bodies lack a of the most bizarre planetary bodies in the Solar Sy tern
solid surface, they all possess orbiting satellites, orne of (Fig. 19.23). It has a high albedo and a striking reddish-
 Planetary geomorphology 501

certainly comes from radioactive decay within the planet's

rocky interior, additional energy is also supplied from its
strong tidal interaction with the giant mass of Jupiter close
by. The surface is thus constantly being renewed by fresh
eruptions of both silicate-rich and sulphur-rich lavas; any
impact craters formed earlier in its history have conse-
quently been obliterated.
Europa, the next satellite out, is, like 10, composed pri-
marily of rock. Its very high albedo and slightly lower density
in comparison with 10 suggest a 50-100 per cent covering
of frost and ice, probably a few kilometres thick, forming
its remarkably smooth surface (Fig. 19.25). There are very
few impact craters indicating that resurfacing, possibly also
promoted by tidal energy derived from the moon's relative
proximity to Jupiter, has occurred since the Late Heavy
Bombardment (assuming that this event affected this part of
the Solar System).
Callisto, the outermost of the Galilean moons, has a much
more familiar looking surface dominated by impact craters
and multi-ringed basins, presumably formed during the
Late Heavy Bombardment and since preserved (Fig. 19.26).
Its low density indicates a high ice content, but its low
albedo suggests that any surface ice must be heavily con-
taminated with rock.
Ganymede, lying between the orbits of Callisto and
Fig. 19.23 View of /0 taken at a range of 862200 km. The lack Europa, has the most complex surface of the four Galilean
of impact craters on the sUljace, together with evidence of satellites representing an amalgam of heavily cratered Callisto-
continuing volcanism, show /0 to be an extremely active body. like terrain and the linear fracture patterns reminiscent of
(Voyager 1 image, World Data Center A for Rockets and
Satellites.) Europa. Its low density indicates a high ice content and its
high albedo suggests a 20-60 per cent surface covering of
fresh ice. It is conjectured that during the Late Heavy
Table 19.4 Properties of the major moons of Jupiter and Saturn
Bombardment the crustal layer of Ganymede was very thin
DIAMETER (km) DISTANCE (krn) DENSITY(kg m- 3 ) but it subsequently thickened, giving rise to a range of
tectonic structures including strike-slip faults and rift zones
Jovian Moons (from Jupiter)
into which fresh, clean ice was injected.
10 3640 421600 3530
Europa 3130 670900 3030
Ganymede 5280 1070000 1930
Callisto 4840 1880000 1790 19.6.2 The moons of Saturn
Saturnian moons (from Saturn) Of the moons of Saturn, Titan is by far the largest and is
Tethys 1060 294700 1210
1120
exceptional in having a significant atmosphere. It is for this
Dione 377 400 1430
Rhea 1530 527100 1330 reason that the surface is obscured and consequently its
Titan 5150 1 221 900 1880 surface morphology is unknown. In spite of their broadly
Iapetus 1460 3560800 1160
similar density and composition there is no lack of variety
in the geomorphology of the other Saturnian satellites imaged
by the Voyager fly-bys.
yellow colour, but quite the most remarkable observation Tethys is the closest to Saturn of the planet's five large
made by the Voyager I probe on its transit through the moons (diameter greater than 1000 km) (Table 19.4). It has a
Jovian system was of an active volcano hurling incan- heavily cratered surface and its most remarkable structure
descent plumes hundreds of kilometres above its surface is an enormous impact scar over 400 km across, that is, some
(Fig. 19.24). The lack of impact craters on its surface, toge- 40 per cent of the diameter of the satellite itself. As this
ther with the abundant morphological evidence of wide- feature now shows little vertical relief, it is hypothesized
spread volcanism, show 10 to be the most active planetary that the interior of Tethys was sufficiently warm and mobile
body in the Solar System observed so far. Although some early in its history to accommodate what was obviously an
of the energy for this high level of endogenic activity enormous impact without shattering. The burial of some
502 Extraterrestrial landforms

Fig. 19.24 A volcanic eruption in progress on 10. The brightness of the volcanic 'plume that can be seen on the horizon rising 100 km
from the surface has been greatly increased by computer enhancement. (Voyager 1 image, World Data Center A for Rockets and
Satellites.)

other large craters also indicates a degree of resurfacing on

the satellite.
Almost identical in size to Tethys is Dione, orbiting
some 80 000 Ian further out from Saturn. It is the densest
moon after Titan and detailed evaluation of its heavily cratered
surface suggests that it has been bombarded by two distinct
populations of projectiles. The trailing hemisphere, that is,
the side facing backward with respect to the direction of its
orbital motion around Saturn, has impact craters similar to
the lunar highlands and with all the hallmarks of the Late
Heavy Bombardment. A second phase of impact cratering
with a greater proportion of small projectiles is attributed to

Fig. 19.25 View of Europa showing a region 600 x 800 km.

Particularly noticeable is the network of lineaments on the
surface representing low relief ridges and grooves, and contrasts
in albedo. These patterns appear to have been caused by stress in
the surface crust generated by various possible mechanisms
including tidal deformation, global expansion due to dehydration
of the interior and freezing of an early ocean. (Voyager 2 image,
World Data Center A for Rockets and Satellites.)
 Planetary geomorphology 503

Fig. 19.26 View of Callisto from a range of 202000 km. The bright circular region to the upper left of the image is an impact basin
about 300 km across. (Voyager 1 mosaic, World Data Center A for Rockets and Satellites.)

debris generated through collision within the Saturnian 19.6.3 The moons of Uranus
system of moons. This has given rise to a moderate crater-
ing of the plains on Dione formed by resurfacing after the Although the images returned from the' Voyager probes as
Late Heavy Bombardment. A similar history appears to they passed through the Jovian and Saturnian systems
have been experienced by Rhea which also shows evidence provided plenty of surprises, the moons of Uranus examined
in its distribution of crater sizes for two populations of ,at close quarters by Voyager 2 in January 1986 yielded yet
impacting projectiles. As on Dione, part of the trailing more bizarre surface forms. The 1500 km diameter moon
hemisphere shows evidence of resurfacing. Oberon was seen to have large impact craters with bright
Closely equivalent in size to Rhea is Iapetus, the most rays, and one mountain peak observed was estimated to be
distant of Saturn's large moons. It has an incredibly low some 6 km high. Its comparatively low density of about
density of about 1100 kg m-3, and a startling contrast in 1600 kg m-3 indicates that it has an icy composition. The
albedo from around 50 per cent on its trailing hemisphere slightly smaller moon Ariel with a diameter of about 1200 km
to an extraordinarily low 3-5 per cent on its leading face; has a similarly low density. One of its most remarkable
this is possibly a result of dusting by particles originating features are canyons which appear to have been flooded by
from Phoebe, the most remote of the Saturnian satellites. eruptions of ice. These and similar ice-filled valley forms
An alternative explanation is that the dark material, which may owe their origin to the extrusion of ice on to the
has been observed in circular features similar to craters filled surface as a result of tectonic activity.
by lava flows on the Moon and Mars, was extruded from The strangest world in the Uranian system is Miranda.
the interior of the satellite. Just 480 km in diameter and with a density of only
504 Extraterrestrial landforms

1200 kg m-3 , it is the nearest of the large satellites to anticipation that re earchers awaited the arrival of Voyager
Uranus. Its surface i composed of two contrasting types of 2 at Neptune. In late August 1989 their patience was re-
terrain: one is bright and heavily cratered, whereas the other warded by views of a world stranger than anything that had
is darker and lower in elevation and is made up of aeries been previously encountered in the epic 12-year journey of
of discrete and apparently unrelated regions of grooved Voyager 2. Although initial attention was focused on Neptune
terrain. At present the most widely accepted explanation of itself, interest rapidly shifted to Triton, its largest moon
this topography is that Miranda suffered a cataclysmic with a diameter of 2700 km. Its fascinatingly complex land-
impact early in its history which caused it to break up and scape, which seems to be an assemblage of terrain types
then subsequently reaggregate. The now juxtaposed distinc- previously encountered on other planetary bodies, imme-
tive terrains are thought to have formed as a result of the diately precipitated a wide range of hypotheses attempting
late accretion of dense fragments which sank slowly through to explain its history.
Miranda's outer icy 'mantle' towards its rocky core. Predictions prior to the arrival of Voyager 2 were that
Triton should be a cold, dead world retaining the scars of
the Late Heavy Bombardment. Yet the surface showed
19.6.4 The moons of Neptune
much evidence of resurfacing with far less than the expected
After the startling images acquired of the surfaces of the density of impact craters (Fig. 19.27). Triton, it was thought,
moons of Jupiter, Saturn and Uranus, it was with great had neither the tidal energy of moons like 10, nor the radio-

Fig. 19.27 A close-up view about 500 km across of the surface of Triton. The lack of large craters indicates active resurfacing after
the Late Heavy Bombardment, but the presence of numerous small craters suggests that generally there has been only a low level of
activity over the past 2 Ga. (Voyager 2 mosaic courtesy IPL and NASA.)
 Planetary geomorphology 505

active energy of large, rocky planetary bodies to generate 19.7 Comparative planetary geomorphology
sufficient internal heat to have sustained 'volcanic' activity
long after its formation. The key to this paradox may lie in The quantum leap in our knowledge of planetary bodies
Triton's peculiar orbit - it moves opposite to the direction over the past two decades presents us for the first time with
of Neptune's rotation and at an inclination of 21 0 to the the possibility of making interplanetary comparisons of land-
planet's equator. Such an orbit suggests that Triton was forms and the processes that shape them. As we have al-
captured by Neptune and was pulled into a highly eccentric ready emphasized, much of the interpretation of the surface
orbit which alternately took it close to the planet anq far forms of the earth-like planets has been accomplished
away. This type of orbit would have indeed created the through analogies with terrestrial landforms. This has
kind of tidal energy that drives lo's intense volcanism at the served to highlight certain common elements in landform
present day. After a few hundred million years the orbit genesis, but for the most part has revealed the considerable
would have become less eccentric, and the reduction in contrasts in both the nature and relative intensity of the
gravitational energy would have caused Triton to cool operation of geomorphic processes throughout the Solar
rapidly. Consequently, the effects of the massive bombard- System. The comparative approach to the study of the evo-
ment experienced by all planetary bodies early in the history lution of planetary surfaces is in its infancy, but it is already
of the Solar System have been erased from Triton's surface possible to suggest some broad generalizations concerning
which now only retains evidence of less catastrophic impact the primary controls of the morphology of the earth-like
events. More recent signs of resurfacing post-dating the planetary bodies (Table 19.5).
Late Heavy Bombardment may be a result of the flowage Until the exploration of the moons of the outer planets, a
of methane and nitrogen 'ice' precipitated out of Triton's key variable in the geomorphic evolution of planetary bodies
thin atmosphere. was considered to be size. The Earth has sufficient mass to

Table 19.5 Comparative geomorphology of some earth-like planetary bodies

PLANET/MOON EXOGENIC PROCESSES IMPACT CRATERS VOLCANISM/RESURFACING TECTONICS

Earth Active chemical and physical weathering. Very rare-largely Locally active Active plate tectonics
Mass movement and fluvial, glacial and obliterated by
aeolian erosion and deposition denudation, volcanism
and tectonics
Moon Regolith production by micrometeorite Abundant Locally active early Limited faulting early in
impact and mass movement in history history
Venus Active chemical weathering and probably Probably present Probably active Uncertain-po ibly a form
aeolian action and mass movement of plate tectonics
Mars Physical and (limited) chemical Fairly abundant Locally abundant Major structural lineaments
weathering. Mass movement and aeolian
activity. Phase of fluvial erosion probably
early in history.
Mercury Probably regolith production by micro- Abundant Possibly active early Thrust faulting associated
meteorite impact and mass movement in history with crustal compression
early in history
10 None? pre umably Very active Related to volcani m
obliterated by volcanic
resurfacing
Europa Uncommon-presumably Resurfacing active Linear patterns related to
obliterated by resurfacing early in history? crustal stress'
Ganymede Relatively abundant None Lineaments with.evidence of
rifts and strike-slip faulting-
form of plate tectonics?
Callisto Mixing of rock with surface ice Abundant including None ?
(associated with impact cratering?) multi-ringed basins
Tehtys ? Abundant Limited resurfacing
Dione Abundant Limited resurfacing
Rhea Abundant Limited resurfacing
Titan Presence of atmosphere suggests some ?
surface weathering
Iapetus ? Abundant Resurfacing by low
albedo material
506 Extraterrestrial landforms

retain a significant atmosphere and thereby allow the oper- represent size, composition and state of thermal evolution
ation of a wide range of denudational processes. The rela- (that is, the nature and history of tectonic and volcanic
tively low surface/volume ratio of the Earth also means that activity). The inner planets all have a similar rocky com-
the rate of heat loss from its surface to space has been position, but several of the moons of Saturn are low-density
sufficiently slow to ensure continuing volcanic and tectonic bodies composed predominantly of ice. Even though they
activity promoted by the high temperatures of a molten are very small in comparison with the Moon and Mercury,
interior. Smaller bodies, such as the Moon and Mercury, these satellites (Tethys, Dione, Rhea and Iapetus) show
have a higher surface/volume ratio and consequently have some evidence of resurfacing after the Late Heavy Bombard-
experienced a much. more rapid rate of cooling than the ment. It seems likely that this activity, in spite of a lack of
Earth. A rigid outer crust formed early in their history mass, is a consequence of their icy composition since ice
which was thus able to retain the scars of the Late Heavy has a far lower melting-point than rock and is therefore
Bombardment. Moreover, both the Moon and Mercury are capable of producing tectonic activity at much lower tem-
too small, and therefore have insufficient surface gravity, to peratures. Hence on our diagram these satellites plot as about
have prevented the escape of gas molecules and thus retain- as tectonically evolved as the Moon, even though they are
ed an atmosphere. Mars is roughly intermediate in size be- much smaller.
tween the Earth and the Moon and Mercury and, as would Turning to the Galilean moons of Jupiter we appear to
be expected, it has both an intermediate atmospheric pres- have more exceptions to the size rule. These satellites are
sure and level of tectonic and volcanic activity. On this of the same order of size as the Moon, but they range in
basis we would expect Venus to have a level of endogenic their degree of thermal evolution from the highly evolved
activity not very dissimilar to the Earth, since it is of 10 to the poorly evolved Callisto. If we look more carefully,
comparative mass. though, we see that the level of thermal evolution of each
In broad terms, then, size (or more strictly mass) appears of the Galilean moons decreases with their increasing dis-
to be a significant factor in controlling the mode of geomor- tance from Jupiter. This is apparently because 10, the inner-
phic development of planetary bodies. This simple relation- most satellite, derives a significant proportion of its internal
ship, however, breaks down when we tum to the larger energy from its tidal interaction with its giant mother
satellites of Jupiter and Saturn. In Figure 19.28 most of the planet. This may also apply, albeit to much smaller extent,
larger earth-like planets and moons of the Solar System are to Europa.
plotted in a three-dimensional diagram, the axes of which Perhaps the most remarkable satellite in the Saturnian
and Jovian systems is Enceladus, a small moon orbiting
ICY~ ---,
close to Saturn. Although it has a diameter of only 500 km
and has only 0.1 per cent of the mass of the Moon, it has a
complex surface suggesting comparatively recent tectonic
activity. Again this is probably associated with the genera-
tion of tidal energy due to the close proximity of the large
mass of Saturn. Further examples of small planetary bodies
with active resurfacing processes are provided by the moons
of Uranus.

19.8 Global geomorphology

It is appropriate to conclude this chapter on planetary geo-
morphology by briefly considering what impact the study
of planetary landscapes has had on geomorphology as a
whole and, perhaps more importantly, what developments
it is likely to precipitate in the future. One obvious con-
tribution has been in the development, and more particularly
the use, of remote sensing techniques in landform analysis.
Planetary exploration has been founded, the manned Apollo
missions notwithstanding, on data collection by unmanned
THERMAL EVOLUTIONI--------- satellites carrying remote sensing equipment. This has
prompted the greater application of satellite imagery to the
Fig.19.28 A comparison of major characteristics ofsome
earth-like planets. (Adaptedfrom L. A. Soderblom and T. V. investigation of terrestrial landforms, as exemplified by the
Johnson, (1982) Scientific American 246(1), p. 86. Copyright © production of a landform atlas by NASA researchers and
(1982) by Scientific American, Inc. All rights reserved.) others based on satellite imagery, although there is much
 Planetary geomorphology 507

scope for further developments in this field. Global surveys appeared in the proceedings of the various lunar science
of terrestrial dune sy tems have now been undertaken, conferences published as supplements of Geochimica et
based largely on Landsat imagery and prompted by the Cosmochimica Acta. Burnett (1975) and EI-Baz (1975) also
need for a comparative database for the interpretation of provide assessments of the major results of the Apollo pro-
images of Martian dune systems. A further development gramme while Wilhelms (1987) has presented a superbly
has been the space shuttle imaging radar programme which, illustrated survey of lunar geomorphology. Contrasting
although so far attaining only a limited geographical cover- estimates of cratering rates throughout the history of the
age, has demonstrated the value of radar techniques in pro- Moon are to be found in Guiness and Arvidson (1977) and
viding images of landforms in regions such as the humid Neukum and Konig (1976), while crater morphology is
tropics where persistent cloud cover limits the value of considered by Hale and Grieve (1982), Head (1976a) and
conventional Landsat imagery. Pike (1977). Lunar volcanism is reviewed by Head (1976b),
A second contribution provided by the exploration of the and the origin of lunar rilles and mare ridges is discussed
inner planets has been the stimulation of research on par- by Lucchitta (1976) and Solomon and Head (1979).
ticular geomorphic processes. The realization of the impor- The discussion of Martian geomorphology in this chapter
tant role played by aeolian processes on Mars, for instance, draws extensively from the excellent review by Baker
arising from the identification of vast dune systems and (1981a); other general surveys include the introductory article
wind-eroded terrain on Pioneer and Viking imagery, has by Arvidson et ai. (1978) and the detailed and compre-
led to a significant growth in laboratory and field studies of hensively illustrated book by Carr (1981). A large number
terrestrial aeolian activity and landforms. of papers on various aspects of the geomorphology of Mars
A third contribution of planetary geomorphology to the are to be found in two supplements of the Journal of Geo-
study of terrestrial landscapes has been the emphasis placed physical Research (84, 7909-8519 (1979), and 87, 9717-
in the former on catastrophic events. Not only meteorite 10305 (1982)).
impact craters but also the formation of features such as the Impact craters on Mars and their associated layered debris
Martian outflow channels seem to be explicable only in terms are discussed by Carr et ai. (1977a) and Mouginis-Mark
of the operation of rapidly acting, highly energetic and (1979), while Carr (1976) provides an introduction to the
devastating geomorphic processes. Such an emphasis has volcanic features on the planet, a treatment which is supple-
led to a re-evaluation of the role of catastrophic processes in mented by a more detailed discussion in Carr et al. (1977b)
shaping the Earth's relief. and a comparison with terrestrial volcanism in Malin (1977).
Perhaps the most significant stimulus provided by the The Valles Marineris structure is considered by Blasius et
exploration of planetary landscapes, however, is one which al. (1977), Sleep and Phillips (1979) suggest an isostatic
has yet to be fully realized; it is an enlargement of the con- model for the formation of the Tharsis Montes volcanic
ceptions of the spatial and temporal scales with which province, and the broader question of Martian tectonics is
geomorphologists should be concerned. Coupled with major tackled by Wise et ai. (1979). Clark and Van Hart (1981)
advances in other fields within the earth sciences, such as and Berkley and Drake (1981) present ideas on weathering,
those represented by plate tectonics and contemporary and Sharp and MaIm (1984) provide an overview of our
theories of long-term climatic change, the investigation of present knowledge of deposits on the Martian surface based
planetary landscapes provides us with an exciting new on Viking lander data. Mass movement on Mars is
global perspective for geomorphology. discussed in papers by Lucchitta (1979) on the spectacular
landslides of the Valles Marineris and Kochel and Peake
(1984) on the morphometry of the debris flows of the fretted
Further reading terrain.
Before the space missions of the 1960s and 1970s little was Aeolian features on Mars have been extensively investi-
known of the surface features of the earth-like planets, but gated through analogies with terrestrial forms by Breed
since 1970 there has been a flood of data from a variety of (1977), Breed et al. (1979) and Ward (1979). Ward et al.
planetary orbiter and lander missions. There are now hun- (1985) provide a comprehensive assessment of the occurrence
dreds of articles on aspects of planetary geomorphology and of aeolian features on the planet, and the processes of wind
only a very small selection is mentioned here. Virtually all abrasion are considered in the context of terrestrial aeolian
of these have been published outside the normal outlets for action by Greeley et al. (1985). The best overall coverage
geomorphic research, the publications Icarus and Journal of planetary aeolian processes, not only for Mars and the
of Geophysical Research being particularly useful sources. Earth but also for Venus, is the book by Greeley and
For a general introduction to planetary geomorphology there Iversen (1985).
is no better tarting point than the extensively illustrated Permafrost on Mars is reviewed by Carr and Schaber
book by Greeley (1986). (1977) and Lucchitta (1981) draws comparisons between
The results of the Apollo missions to the Moon have I periglacial features on Mars and Earth. The layered depo-
508 Extraterrestrial landforms

sits of the poles and other problems of the polar terrain are ly convincing view of the potential value of thi infant
discussed in Cutts et ai. (1979), while Lucchitta et ai. research field for geomorphology as a whole.
(1981) speculate on the role of glacial erosion on Mars and
Clifford (1987) assesses the role of melting in generating
References
chaotic terrain and outflow channels.
Arvidson, R. E., Binder, A B. and Jones K. L. (1978) The surface
The enigmatic Martian channels have generated a large
of Mars. Scientific American 238(3), 76-89.
literature. A useful starting point is Baker (1981 b), while a Baker, V. R. (1981a) The geomorphology of Mars. Progress in
selection of views concerning the origin of outflow channels Physical Geography 5, 473-513.
include those in Ma ursky et ai. (1977), Carr (1979), Baker, V. R. (1981b) The Channels of Mars. University of Texas
Komar (1980), Nummedal and Prior (1981) and Cutts and Press, Austin.
Baker, V. R. (1985a) Models of fluvial activity on Mars. In: M. J.
Blasius (1981). The origin and modification of small
Woldenberg (ed.) Models in Geomorphology. Allen and Unwin,
Martian valleys are discussed by Baker and Partridge (1986) Boston and London, 287-312.
and Brakenridge et ai. (1985). In view of the wide range of Baker, V. R. (1985b) Relief forms on planets. In: A. Pitty (ed.)
proposed mechanisms of channel formation it is fortunate Themes in Geomorphology, Croom Helm, London, 245-59.
that there is also a comprehensive overview available (Mars Baker, V. R. and Partridge, J. B. (1986) Small Martian valleys:
Pristine and degraded morphology. Journal of Geophysical
Channel Working Group, 1983). Baker (1985a) looks at the
Research 91, 3561-72.
role of fluvial activity on Mars, while Carr (1987) takes a Barsukov, V. L. et al. (1986) The geology and geomorphology of
broad look at the role of water in the planet's history. the Venus surface as revealed by the radar images obtained by
A useful assessment of the results of the Mariner 10 mis- Venera 15 and 16. Journal of Geophysical Research 91,
sion to Mercury is provided by Strom (1979), while the 0378-0398.
Basi1evsky, A T. and Head, 1. W. III (1988) The geology of Venus.
initial results of this project are contained in Murray (1975)
Annual Review of Earth and Planetary Sciences 16, 295-317.
and a special issue of Icarus (28, 429-609 (1976». The Berkley, 1. L. and Drake, M. J. (1981) Weathering on Mars:
possible gravitational effects of the Sun on the tectonics of Antarctic analog studies. Icarus 45,231-49.
Mercury is examined in Bums (1976), while McCauley et Blasius, K. R., Cutts, J. A, Guest, J. E. and Masursky, H. (1977)
ai. (1981) review the Caloris Basin and Dzurisin (1978) Geology of the Valles Marineris, first analysis of imaging from
the Viking 1 orbiter primary mission. Journal of Geophysical
presents a general discussion of the tectonic features of the
Research 82, 4067-91.
planet. Brakenridge, G. R., Newsom, H. E. and Baker, V. R. (1985) Ancient
The morphology of the Venusian surface as revealed by hot springs on Mars: origins and paleoenvironmental signifi-
radar altimetry is discussed by Barsukov et ai. (1986), cance of small Martian valleys. Geology 13, 859-62.
Basilevsky and Head (1988) and Head et al. (1985). An Brass, G. W. and Harrison, C. G. A. (1982) On the possibility of
plate tectonics on Venus. Icarus 20, 326-40.
evaluation of the data collected by the various Soviet Venera
Breed, C. S. (1977) Terrestrial analogs of the Hellespontu dunes,
landers is provided by Garvin et ai. (1984), while White Mars. Icarus 20, 326-40.
(1981) and Nozette and Lewis (1982) speculate on the Breed, C. S., Grolier, M. J. and McCauley, J. F. (1979) Morpho-
nature of weathering and erosion. The intriguing question logy and distribution of common 'sand' dunes on Mars: com-
of whether there is a form of plate tectonics operative on parison with Earth. Journal of Geophysical Research 84,
8183-204.
Venus is addressed by Brass and Harri on (1982), Head et
Burnett, D. S. (1975) Lunar science: The Lunar legacy. Reviews of
ai. (1981) and Solomon and Head (1982), while Head and Geophysics and Space Physics 13, 13-34.
Wilson (1986) examine the possibilities of volcanic land- Burns, 1. A. (1976) Consequences of the tidal slowing of Mercury.
forms. Icarus 28, 453-8.
Results from the Voyager missions to Jupiter and Saturn Carr, M. H. (1976) The volcanoes of Mar. Scientific American
have rapidly generated a large literature. Useful surveys of 234(4),32-43.
Carr, M. H. (1979) Formation of Martian flood features by release
the initial data on the larger satellites are to be found in of water from confined aquifers. Journal of Geophysical Research
Soderblom (1980) (Jovian moons), Soderblom and Johnson 84,2995-3007.
(1982) (Saturnian moons) and Johnson et ai. (1987) (Uranian Carr, M. H. (1981) The SUlface of Mars. Yale University Press,
moons), but for a more detailed coverage the book by Greeley New Haven, and London.
(1986) should be consulted. Carr, M. H. (1987) Water on Mars. Nature 326, 30-5.
Carr, M. H., Crumpler, L. S., Cutts, J. A, Greeley, R., Guest, 1. E.
The broader topic of comparative planetary geomorpho- and Masursky, H. (1977a) Martian 'impact craters and emplace-
logy, as well as being addressed by Greeley (1986), is con- ment of ejecta by surface flow. Journal of Geophysical Research
sidered from a tectonic perspective by Head and Solomon 82,4055-65.
(1981). Rossbacher and Rhodes (1987) provide an excellent Carr, M. H., Greeley, R., Blasius, K. R., Guest, J. E. and Murray,
example of the analogies that can be drawn between ter- J. B. (1977b) Some Martian volcanic features as viewed from
the Viking orbiters. Journal of Geophysical Research 82,
restrial and planetary landforms attributable to catastrophic 3985-4015.
flooding. It is to the essay by Baker (1985b), however, that Carr, M. H. and Schaber, G. G. (1977) Martian permafrost fea-
readers should tum for a creative, optimistic and thorough- tures. Journal of Geophysical Research 82, 4039-54.
 Planetary geomorphology 509

Clark, B. C. and Van Hart, D. C. (1981) The salts of Mars. Icarus Lucchitta, B. K. (1981) Mars and Earth: Comparison of cold-
45,370-8. climate features. Icarus 45,264-303.
Clifford, S. M. (1987) Polar basal melting on Mars. Journal of Lucchitta, B. K, Anderson, D. M. and Shoji, H. (1981) Did ice
Geophysical Research 92, 9135-52. streams carve Martian outflow channels? Nature 290, 759-63.
Cutts, 1.A and Blasius, KR (1981) Origin of Martian outflow Malin, M. C. (1977) Comparison of volcanic features of Elysium
channels: The eolian hypothesis. Journal of Geophysical (Mars) and Tibesti (Earth) Geological Society of America
Research 86, 5075-102. Bulletin 88,908-19.
Cutts, 1. A., Blasius, K R and Roberts, W. 1. (1979) Evolution of Mars Channel Working Group. (1983) Channels and valleys on
Martian polar landscapes: interplay of long-term variations in Mars. Geological Society ofAmerica Bulletin 94, 1035-54.
perennial ice cover and dust storm intensity. Journal of Geo- Masursky, H., Boyce, J. M., Dial, A. L., Schaber, G. G. and Strobell,
physical Research 84, 2975-94. M. E. (1977) Classification and time of formation of Martian
Dzurisin, D. (1978) The tectonic and volcanic history of Mercury channels based on Viking data. Journal of Geophysical Research
as inferred from studies of scarps, troughs and other lineaments. 82,4016-38.
Journal of Geophysical Research 83, 4883-906. McCauley, 1. F., Guest, 1. E., Schaber, G. G., Trask, N. J. and
El-Baz, F. (1975) The Moon after Apollo. Icarus 25, 495-537. Greeley, R (1981) Stratigraphy of the Caloris basin, Mercury.
Garvin, J. B., Head, J. W., Zuber, M. T. and Helfenstein, P. (1984) Icarus 47, 184-202.
Venus: The nature of the surface from Venera panoramas. Mouginis-Mark, P. 1. (1979) Martian fluidized crater morphology:
Journal of Geophysical Research 89,3381-99. variations with crater size, latitude, altitude and target material.
Greeley, R (1986) Planetary Landscapes (revised edn). Allen and Journal of Geophysical Research 84,8011-22.
Unwin, London and Boston. Murray, B. C. (1975) Mercury. Scientific American 233(3), 58-68.
Greeley, R and Iversen, J. D. (1985) Wind as a Geological Neukum, G. and Konig, B. (1976) Dating of individual lunar
Process. Cambridge University Press, Cambridge and New craters. Geochimica Cosmochimica Acta Supplement 7, 2867-
York. 81.
Greeley, R, Williams, S. H., White, B. R, Pollack, J. B. and Nozette, S. and Lewis, 1. S. (1982) Venus: Chemical weathering of
Marshall, J. R (1985) Wind abrasion on Earth and Mars. In: igneous rocks and buffering of atmospheric composition. Science
M.1. Woldenberg (ed.) Models in Geomorphology. Allen and 216,181-3.
Unwin, Boston and London, 373-422. Nummedal, D. and Prior, D. B. (1981) Generation of martian
Guiness, E. A. and Arvidson, R. E. (1977) On the constancy of the chaos and channels by debris flow. Icarus 45, 77-86.
lunar cratering flux over the past 3.3 x 109 yr. Geochimica et Pike, R J. (1977) Apparent depth/apparent diameter relation for
Cosmochimica Acta Supplement 8,3475-94. lunar craters. Geochimica et Cosmochimica Acta Supplement 8,
Hale, W. S. and Grieve, R A. F. (1982) Volumetric analysis of 3427-36.
complex lunar craters: implications for basin ring formation. Rossbacher, L. A. and Rhodes, D. D. (1987) Planetary analogs for
Journal of Geophysical Research 87, A65-A76. geomorphic features produced by catastrophic flooding. In: L.
Head, J. W. III (1976a) The significance of substrate charac- Mayer and D. Nash (eds) Catastrophic Flooding. Allen and
teristics in determining the morphology and morphometry of Unwin, Boston and London, 289-304.
lunar craters. Geochimica et Cosmochimica Acta Supplement 7, Sharp, R P. and Malin, M. C. (1984) Surface geology from
2913-30. Viking landers on Mars: A second look. Geological Society of
Head, 1. W. III (1976b) Lunar volcanism in space and time. America Bulletin 95, 1398-412.
Reviews ofGeophysics and Space Physics 14,265-300. Sleep, N. H. and Phillips, R.1. (1979) An isostatic model for the
Head, 1. W. III, Peterfreund, A. R., Garvin, J. B. and Zisk, S. H. Tharsis province, Mars. Geophysical Research Letters 6, 803-6.
(1985) Surface characteristics of Venus derived from Pioneer Soderblom, L. A. (1980) The Galilean moons of Jupiter. Scientific
Venus altimetry, roughness and reflectivity measurements. American 242(1),68-83.
Journal of Geophysical Research 90, 6873-85. Soderblom, L. A. and Johnson, T. V. (1982) The moons of Saturn.
Head, 1. W. III and Solomon, S. C. (1981) Tectonic evolution of Scientific American 246(1), 73-86.
the terrestrial planets. Science 213, 62-76. Solomon, S. C. and Head, J. W. III (1979) Vertical movements in
Head, J. W. III and Wilson, L. (1986) Volcanic processes and land- mare basins: relation to mare emplacement, basin tectonics and
forms on Venus: Theory, predictions and observations. Journal lunar thermal history. Journal of Geophysical Research 84,
ofGeophysical Research 91, 9407-46. 1667-82.
Head, J. W. III, Yuter, S. E. and Solomon, S. C. (1981) Topography Solomon, S. C. and Head, J. W. III (1982) Mechanisms for litho-
of Venus and Earth: A test for the presence of plate tectonics. spheric heat transport on Venus: Implications for tectonic style
American Scientist 69,614-23. and volcanism. Journal of Geophysical Research 87, 9236-46.
Johnson, T. V., Brown, R H. and Soderblom, L. A. (1987) The Strom, R G. (1979) Mercury: A Post-Mariner assessment. Space
moons of Uranus. Scientific American 256(4), 40-52. Science Reviews 24, 3-70.
Kochel, R C. and Peake, R T. (1984) Quantification of waste Ward, A. W. (1979) Yardangs on Mars: evidence of recent wind
morphology in Martian fretted terrain. Journal of Geophysical erosion. Journal ofGeophysical Research 84,8147-66.
Research 89, C336-C350. Ward, A. W., Doyle, K B., Helm, P.1., Weisman, M. K and
Komar, P. D. (1980) Modes of sediment transport in channelized Witbeck, N. E. (1985) Global map of eolian features on Mars.
water flows with ramifications to the erosion of the martian Journal of Geophysical Research 90, 2038-56.
outflow channels. Icarus 42, 317-29. White, B. R (1981) Venusian saltation. Icarus 46,226-32.
Lucchitta, B. K. (1976) Mare ridges and related highland scarps - Wilhelms, D.E. (1987) The geologic history of the Moon. United
result of vertical tectonics. Geochimica et C osmochimica Acta States Geological Survey Professional Paper 1348.
Supplement 7, 2761-82. Wise, D. u., Golombek, M. P. and McGill, G. E. (1979) Tectonic
Lucchitta, B. K (1979) Landslides in Valles Marineris, Mars. evolution of Mars. Journal ofGeophysical Research 84, 7934-9.
Journal of Geophysical Research 84, 8097-113.
This page intentionally left blank
 Appendix A
Units of measurement

The units used in this book are based on the Systeme Inter- Acceleration
national d'Unites (SI) which has now become the standard m S-2 metres per second per second
for scientific literature world-wide. The only major deviation
from SI units is the recording of temperature in °C (degrees Mass
celsius - sometimes expressed as centigrade) rather than K g gram
(degrees kelvin). kg kilogram
Multiples of SI units are expressed as follows: t tonne (= 103 kg)
10-6 micro ~
10-3 milli m Density
10-2 centi c* kgm-3 kilograms per cubic metre (to convert to g cm-3
103 kilo k multiply by 0.001)
106 mega M
109 giga G Force
N newton (the force required to give a mass of 1 kg an
Time acceleration of 1 m-2 .
s second
h hour Pressure
a year Pa pascal (= I Nm-2 ) (to convert to mb (millibars)
ka 103 years multiply by 100)
Ma 106 years
Ga 109 years Work/energy
J joule (the work done when the point of application
Length of a force of 1 N is displaced through a distance of
m metre 1 m in the direction of the force; since changes in
cm centimetre energy result in work being done, energy and work
mm millimetre have the same units)
~m micrometre (micron)
km kilometre Power
W watt (= 1 J S-I)
Velocity
m S-1 metres per second Dynamic viscosity
kmh- I kilometres per hour (to convert to m S-I multiply N sm-2 newton second per metre squared
by 3.6)
Temperature
°C celsius (to convert to K (kelvin) add 273.15)
* Not strictly an SI multiple.
This page intentionally left blank
 Appendix B
Dating techniques

A wide range of techniques is used' in the dating of events in Geochronometric dating

Earth history. Although some of these techniques are more
directly relevant to landform analysis than others, it is
Radiometric dating
important for geomorphologists to have a broad understand-
ing of dating methods. The establishment of specific ages for events in Earth his-
. tory had to await the development of radiometric dating
methods involving the measurement of radioactive isotopes
Relative dating (although various attempts to estimate the age of the Earth
Relative dating methods indicate the order in which events had been made by various indirect methods before radio-
have occurred, but not the time elapsed between them. The metric techniques were developed earlier this century). Most
basis for the relative dating of rocks and sediments are the atomic nuclei are stable, but several elements have inhe-
principles of stratigraphic succession. The principle of ori- rently unstable, or radioactive, nuclei which undergo spon-
ginal horizontality states that sediments deposited from taneous transformation into other types of atomic nuclei.
fluids under the influence of gravity are laid down in hori- There are various ways in which this can occur, but the
zontal, or nearly horizontal, layers. The principle of strati- processes responsible for all such transformations are called
graphic superposition states that in a vertical profile through radioactive decay. The decay of a radioactive isotope can
sedimentary units which have not been subsequently over- involve the acquisition or emission of certain kinds of atomic
turned as a result of tectonic or other mechanisms, their age particles, the emission of electromagnetic radiation, or both.
increases from the top to the base. The recognition that fossil The atom experiencing radioactive decay is termed the
faunas in sedimentary strata succeed one another through parent isotope, and the product is called the daughter
time in a recognizable order enabled geologists in the early isotope.
nineteenth century to extend relative dating in the horizontal An example of radioactive decay is the transformation of
dimension by tracing sedimentary units characterized by 40K (potassium-40) into 40Ar (argon-40). Most of this trans-
specific fossil assemblages across country. By the 1830s this formation of 40K occurs through the emission of single
technique was being extended to the correlation of sedi- electrons (called Bparticles) to form 40Ca (calcium-40), but
mentary units originally defined on the basis of the fossil 12 per cent of the 40K atoms are transformed into 40 Ar by
content of British strata to places as far afield as Russia and electron capture. Such processes of radioactive decay
the USA. can be used as a basis for dating because the transforma-
To a limited extent the relative dating of geomorphic tion of certain types of atomic nuclei into other types occurs
events can be achieved with reference to morphological . at a predictable rate which is unaffected by any chemical
evidence, supplemented, where possible, by evidence pro- changes, or any normal changes in temperature and
vided from associated weathering materials. Such relative pressure. The rate of decay of a given mass of a radio-
dating was the primary basis for the reconstruction of land- active isotope declines exponentially; the time taken for the
scape histories, known as denudation chronology, prior to number of atoms of the parent isotope to be reduced by one
the advent of modern geochronometric dating techniques half is known as the half-life of that particular isotope. In
(see Section 18.3). each successive half-life period the remaining number of
514

atoms of the parent isotope is further reduced by half. The Table B1 Some important isotopes used in radiometric dating
half-life of 40K is 1.3 Ga, so after 2.6 Ga only one-quarter
ISOTOPE HALF-LIFE DATING DATABLE MATERIJ\LslMINERALS
of the number of atoms of the parent isotope remain. Radio-
RANGE
metric dating techniques are sometimes said to provide
absolute ages, but this implies a greater degree of accuracy Carbon-14 5730 a <50 ka Wood, charcol, peat, bone,
than is possible with present techniques. Geochronometric animal tissue, shells,
speleothems, ground water,
dating is a more appropriate term for such methods. ocean water, glacier ice
A number of factors have to be taken into consideration Thorium-230 75 ka <200 ka Organic carbonate
when radiometric dating is applied to geological materials. Uranium-234 250 ka 50-100 ka Coral
All radiometric dates have a margin of error. This can range Potassium·40 1.3 Ga >100 ka Muscovite, biotite,
from ±5 per cent for some methods to ±20 per cent for hornblende, whole volcanic
others. Another factor is the stage at which the radioactive rock

decay system becomes closed; that is, the point where all the Uranium-238 4.5 Ga >IOMa Zircon, uraninite

daughter isotopes being produced are trapped in the Rubidium-87 47Ga >IOMa Muscovite, biotite,
potassium-feldspar, whole
material and cannot escape. For example, at high tempera- igneous or metamorphic
tures argon diffuses out of minerals, so the amount of 40Ar rock
measured in a sample will only reflect that produced since
the sample cooled below the critical closure temperature
(this varies for different minerals). A radiometric age, there- counted under a microscope. The number of tracks per unit
fore, indicates the length of time that the products of radio- area is determined by the amount of parent isotope and the
active decay have been retained within the material being time elapsed since tracks were first preserved. Above a
dated. K-Ar dating is often applied to basaltic lavas so the certain temperature fission tracks are annealed, that is, they
K-Ar age gives the time since the lava cooled below the are destroyed, so the fission track dating clock only begins
critical temperature. If the lava is subsequently reheated once a sample cools below a critical temperature. These
then some of the argon may escape and the radiometric critical temperatures vary for different minerals, but the
clock will be reset. Since rocks cool as erosion strips away mineral apatite has a low annealing temperature of around
the overburden and they are brought up to the surface, 100°C which makes it especially valuable for estimating
K-Ar dating provides a means of estimatimg long-term long-term rates of denudation and rock uplift (see Sections
rates of denudation and rock uplift (see Sections 15.2.1 15.2.1, and 15.4.1.3). Depending on the uranium content of
and 15.4.1.3). the rock, fission track dating can be used to date rocks over a
Radiocarbon (14C) dating, which is widely applied to very broad time span. In certain types of volcanic rocks,
organic remains, is a particularly useful dating method dating down to a few thousand years is possible, but it can
because of the very short half-life of 14C (5370 a). During also be widely applied to rocks hundreds of millions of years
life, organisms contain the same proportion of 14C as the old.
atmosphere through the processes of photosynthesis, res-
piration and feeding. Once they die organisms acquire no
Palaeomagnetic dating
more 14C and the amount present at death declines through
the radioactive decay of 14C into 14N. Although invaluable Palaeomagnetic dating is based on the fact that the
for dating organic remains up to 50 ka old (and with new, properties of the Earth's magnetic field vary over time.
more sensitive techniques up to 100 ka) samples have to be Changes occur in its declination (the angle between mag-
carefully selected to overcome the problem of contamination. netic and geographical north), its dip (inclination to the
The more important isotope transformations used in horizontal) and its intensity. The most dramatic changes are
radiometric dating and the types of material to which they represented by complete reversals in the polarity of the
can be applied are listed in Table B 1. magnetic field (see Section 2.3.2), but other more subtle
changes also occur. These variations in the magnetic field
can be used to date rocks or sediments where magnetically
Fission track dating
susceptible minerals or particles containing iron have become
Fission track dating is based on the accumulation of radia- orientated with the magnetic field as they are formed or
tion damage from the 'spontaneous nuClear fission of 238U in deposited. These rocks and minerals 'freeze in' the pre-
uranium-bearing minerals, such as apatite, zircon and vailing magnetic field, and since they can be dated by inde-
sphene. These fission events lead to the emission of the two pendent means (such as the radiometric dating of basalt
fragments of the 238U nucleus into the surrounding crystal flows) it is possible to construct a palaeomagnetic time
lattice which is thereby damaged. The damaged areas can scale which can then be used as a basis for future dating on
be etched out by acid and the resulting fission tracks palaeomagnetic evidence alone.
 515

Varve dating of new wood in the spring and summer is affected by pre-
Varves are annually deposited layers of sediment. Each vailing weather conditions, and in extreme cases of drought a
varve consists of a layer of coarse sediment which becomes growth ring may be eliminated, dendrochronology is a use-
finer upwards. The most common type used in dating are ful dating technique for historic time and an invaluable
glacial varves. These are formed where glacial meltwater means of cross-checking 14C dates within this time range.
brings sediment into a body of still water. Each summer
Amino acid racemization dating
meltwater rivers carry a mixed load of sand and clay, and
the sand is deposited first. Such annual layers can often be The recently developed technique of amino-acid racemi-
counted back over several thousand years and they clearly zation dating is based on certain time-dependent chemical
indicate changes in sediment load over time. Other types of changes (termed racemization) experienced by proteins
varves may form as result of the production of organic mat- preserved in the remains of organisms. The main problem
ter over an annual cycle. with the method is that the rate of chemical change is also
temperature dependent, so environments with a uniform
temperature, such as deep caves, are required. Used alone it
Dendrochronology
can yield only a relative age, but if calibrated by 14C it can
The dating technique of dendrochronology is based on the provide useful geochronometric ages. Dating as far back as
annual growth rings of trees. Although the seasonal growth 200 ka is possible.
This page intentionally left blank
 517

Appendix C Geological time scale

Era Sub-era/period/ Age (A) Age (B)
sub-periodlEpoch (Ma BP) (MaBP)

Holocene
Quaternary 0.01 0.01
Pleistocene
1.64 1.6
Pliocene
5.2 5.3
u Neogene Late
0N
0 Miocene 14.2 15.8
Z
U.l :>-. Early
U .~
t:: 23.3 23.7
~
E-< Oligocene
35.4 36.6
Paleogene Eocene
56.5 57.8
Paleocene
65.0 66.4
Late
Cretaceous 97.0 97.5
Early
145.6 144
U
Late
0
N
0 157.1 163
en

~ Jurassic Middle
178.0 187
Early
208.0 208
Triassic
245.0 245
Permian
290.0 286
Carboni ferous
362.5 360
U
0N Devonian
0 408.5 408
U.l
<I::
:~ Silurian
0.. 439.0 438
Ordovician
510.0 505
Cambrian
570.0 570
PRECAMBRIAN

Source: Ages according to W. B. Harland et at. (1990) A Geologic Time Scale 1989, Cambridge University
Press, Cambridge (A); and Geological Society of America (1983) Decade of North American Geology
Geologic Time Scale. Geological Society of America, Boulder (B).
This page intentionally left blank
 Appendix D
Particle size

Table Dl indicates the main size classes used in the descrip- Table Dl Major class inter...als used in description of sediment sizes
tion of sediments. More detailed information can be obtained
from any textbook on sedimentology. The phi (<p) scale is mm 4' SIZE CLASSES

widely used in the description of sediments. It is an inverse Boulder

logarithmic scale where a particle size of 1 mm is given a <p 256 -8
value of zero. The convenience of the <p scale arises fr~m 64 -6 Cobble
32 -5
its small numerical range in comparison with the metric 16 -4 Pebble
equivalent. Grain sizes have traditionally been determined 4 -2
using square-meshed sieves or, for fine particles, pipette and 2.83 -1.5 Granule
2.00 -1.0
hydrometer techniques. Automated sediment size analyzers 1.41 -0.5 Very coarse sand
employing laser technology are, however, now becoming 1.00 0.0
available. 0.71 -0.5 Coarse sand
0.50 1.0
0.35 1.5 Medium sand
0.25 2.0
0.177 2.5 Fine sand
0.125 3.0
0.088 3.5 Very fine sand
0.0625 4.0
0.031 5.0 Coarse ilt
0.0156 6.0
0.0078 7.0 Fine silt
0.0039 8.0
Clay

Source: Based on F. J. Pettijohn et al. (1972) Sand and Sandstone.

Springer-Verlag, New York, Table 3-2, p. 71.
This page intentionally left blank
 Appendix E
Behaviour of materials

Stress is defined as force per unit area and is measured in TENSILE STRESS
pascals (Pa) (l Pa = I Nm-2 ). Strain is a measure of the
deformation experienced by a material as the result of an
applied stress and i expre sed in terms of the relative change
in shape or volume. Three fundamental types of tress can be
distinguished depending on the relative direction of the force
being applied (Fig. EI). Tensile stress is an extensional
force which tends to stretch or pull material apart. Such a
stress is capable of changing the shape of the material. It may
-00-
also increase its volume if the stretching leads to an open- COMPRESSIVE STRESS
ing of voids and hence a reduction of bulk density. Com-
pressive stress is a force which tends to compress material
and thereby change its shape. There can also be a reduction
in volume if pores and fractures are closed up, or if mineral
grains are reorganized to pack more closely. Material at
depth i subject to compressive stress proportional to the
weight of the overburden of rock. Finally, shear stress is a
force which acts so as to tend to deform a mass of material
by one part sliding over another along one or more failure
planes. Although shear tress causes a change in the shape SHEAR
of the material, it generally has no significant effect on
volume.
Materials respond in different ways to .an applied stress
depending on their inherent physical properties and the
•
CZJ . ..... .---

magnitude and duration of the stress (Fig. E2). A material SHEAR STRESS
exhibits elastic behaviour if it entirely recovers its original
form when the deforming stress is removed. Over a large
range of applied stresses, strain is proportional to the applied
stress, a relationship known as Hooke's law. Beyond a
certain point, known as the elastic limit, the deformation is
no longer recovered when the stress is removed and the 1
- .......1_ _.._-
material exhibits plastic behaviour. In brittle rocks plastic
deformation is very limited and failure occurs just beyond
the elastic limit. +-4=
In a perfectly pIa tic material no deformation occurs ==C=====~===:::J
:'=::::'2'-::::1
until a threshold level of stress (known as the yield stress) Fig. El Schematic illustration of tensile, compressive and shear
is attained; thereafter strain occurs at a constant rate as long stress.
522

as the stress is constant. In purely plastic behaviour a material

!L=
STRESS-STRAIN STRAIN - TIME STRESS-STRAIN RATE

Elastic limit
j
completely retains the form assumed during deformation
ELASTIC ~ when the applied stress is relaxed. Viscous behaviour
BEHAVIOUR li; occurs when the rate of deformation, or strain rate, is pro-
portional to the applied stress. This occurs in an ideal
STRAIN TIME Newtonian fluid where deformation occurs in response to
extremely small stresses. We normally associate viscous be-
At stresses
haviour with liquids, but a form of non-Newtonian viscous
greater than
PLASTIC ~ yieldstress behaviour can also apply to solids if the stress is applied
BEHAVIOUR ~ \ over a sufficient period of time; note, for instance, that
glaciers composed of solid ice can flow.
STRAIN TIME STRAIN RATE In reality materials making up lithosphere do not exhibit
a single type of behaviour, but rather respond to stresses in

!~
a complex manner. Nevertheless, the fundamental distinctions
VISCOUS
\ between elastic, plastic and viscous behaviour are import-
BEHAVIOUR Ideal
Newtonianfluid ant in analyzing the response of lithospheric materials to
stress.
TIME STAAINAATE

Fig. £2 Response ofmaterials experiencing elastic, plastic and

viscous behaviour. (Modified from 1. Statham (1977) Earth Surface
Sediment Transport. Clarendon Press, Oxford. Fig. 2.13, p. 53.)
 Appendix F
Geological structures

Faults creation of faults, representing fractures along which major

movements have occurred. These range from a few metres
Stress in the crust can be accommodated either by the flow in total, affecting only a shallow depth of rock or sediment,
and deformation of rocks over long periods of time or by to large movements of tens or hundreds of kilometres along
fracture, the latter being common in the brittle rocks of the major faults which extend to the base of the lithosphere.
upper crust. One manifestation of fracture is the develop- The surface trace of the fault is known as the fault line and
ment of joints; these are structures of small dimensions its horizontal orientation is described as its strike (Fig.
lacking any significant movement. The other involves the FI(A».

B
A Minimum ~ NORMAL
compressive
FAULT NOMENCLATURE ;s~tr.:e=ss~ ---.., , FAULT

c o STRIKE-SLIP
Maximum FAULT
REVERSE
compressive """
stress ~ FAULT

Fig. Fl Schematic representation of the major types offault: (A) illustration of dip, strike,fault line andfault plane; (B) normalfault;
(C) reverse fault; (D) strike-slip fault.
524

Normal faults develop under a pattern of predominantly sive stress is vertical, strike-slip faults are formed (equi-
tensional stress. Movement along the fault involves exten- valent terms are wrench fault and transcurrent fault)
sion of the crust as the dip of the fault plane (Fig. FI(A)) (Fig. FI (D)). The fault plane is essentially vertical, and
is less than 90° (Fig. FI(B)). A down-faulted block between a motion along the fault is horizontal. Sinistral strike-slip
pair of more or less parallel normal faults is known as a faults exhibit left-lateral movement across the fault (as
graben, and an up-faulted block is termed a horst. Reverse illustrated in Figure FI(D)), while motion across dextral
faults are normally associated with zones of compression, strike-slip faults is right-lateral.
and some crustal shortening occurs as one block rides over While these represent the basic types of fault, other varieties
the other along the fault (Fig. FI (C)). Where the angle of exist. Where both horizontal and vertical movements are
dip is low the term thrust fault (Fig. F2) is often applied, significant the term oblique-slip fault is applied; these can
and this kind of faulting is frequently associated with intense be either normal or reverse. Seismic studies of some block-
crustal folding in orogenic belts. Where the mean compres- faulted terrains and sedimentary basins have indicated that
the dips of their 'normal' faults often decrease with depth
until they become more or less horizontal. These are des-
cribed as listric faults and their identification has provided
~ine

(
important evidence bearing on the problem of crustal exten-
sion during continental rifting and break-up. Normal listric

~f faults are associated with crustal thinning, but a change

from a tensional to a compressive regional stress pattern
can lead to a reversal of movement and the development of
Anticline reverse listric faults.

Folding
While fracture is a common response to stress at the sur-
face, sustained stress under high confining pressures can
cause bedrock deformation and folding. Most folds originate
at some depth, the simplest forms produced being the
monocline, anticline and syncline (Fig. F2). In a recum-
bent fold the strata are overturned and both limbs of the
fold are nearly horizontal. The horizontal compression that
creates recumbent folds may eventually lead to the shearing
of the upper part of the fold along a thrust fault. The strata
Fig. F2 Basic types offoid structures. moved forward over the thrust fault is called a nappe.
 Index

Note that a page number in bold type indicates where the main definition of the
indexed term is to be found.

Aa 112 also passive margin basins Anabranching channels 211

Ablation of glacial ice 263, 266 (drainage)) Analogue models 20
Abney level 182 Agassiz, L. 4 Anastomosing channels 211,213,284
Abrasion (fluvial) 199, 283 (see also Aggressivity 150 Andes 34-6, 63-5, 122, 377,475
aeolian abrasion, glacial abrasion) Ahagger Plateau 85, 94 Angle of internal friction 164
Abrasion pH 133 Airy isostasy (see local isostasy) Anions 130
Abrasion ramps 327 Airy model 38, 39 Annealed fission tracks 375, 514
Accumulation of glacial ice 263, 266 Aitutaki, Cook Islands 328, 453 Annual series of river channel
Acidity 129 Akle pattern 234,251 discharge 193
Acid rocks 108 Alases 307 Annular drainage pattern 406
Active ice 263 Alaska, USA 375 Antarctica 346, 350, 445
Active layer 294 Alas valleys 307 Antarctic ice sheet 261, 263, 266, 268,
Actively unstable slope 167 Alba Petera, Mars 490 349,350,439,440
Active rifting 93, 96, 97 Alcrete 153 Antecedence of drainage 411,421,422,
Active tectonic controls of drainage Alkalinity 129 423
405,412-18 Allochthonous terranes (see displaced Anticline 69,404,524
Active volcano 109, 110 terranes) Anti-dip streams 407
Actualism 17 Allogenic valleys 229 Antidunes 199
Aeolian abrasion 239,240 (see also Alluvial channels 199,210-22, 228, Antithetic fault 92
abrasion) 416-18 Apennines, Italy 174
Aeolian activity, in periglacial Alluvial fans 15, 222-4, 225 Apollo space programme 484,485,
environments 301 Alluvial plains 221 486,487
Aeolian activity, on Mars 492-4 Alluvial stream sink dolines 150 Appalachian Mountains, USA 69,411,
Aeolian activity, on Venus 493, 499, Alpine Fault, New Zealand 76-7, 467 419,471,472
500 Alpine troughs 273 Applied geomorphology 24,25,27
Aeolian denudation rates 391, 395-7 Alps 69,70,71, 77, 375 Ares Vallis, Mars 496
Aeolian deposition 242-57,336-9,364, Altiplanation terraces 308 (see also Arete 272,275
366 cryoplanation terraces) Argand, E. 73
Aeolian erosion 239-42 Altiplano, Bolivia-Peru 65, 475 Ariel 503
Aeolian landforms, distribution 235, Altyn Tagh Fault, China 73, 74 Arno, Marshall Islands 452
236,248,250,256 Amazon Basin 365, 382, 387, 418 Arrhenius equation 132
Aeolian systems, response to climatic Amazon River 192, 363, 364, 365,380, Arroyos 224,225
change 364, 366 423,448 Arun River, Nepal 423
Aerial photography in monitoring Amerotype drainage basins 419 (see Ash 112
landform change 17 also passive margin-distal orogen Asthenosphere 37
Africa 85,87,89,90,94,102,426 basins) Asymmetric valleys 308
Afrotype drainage basins 418, 419 (see Amino acid racemization dating 515 At-a-station hydraulic geometry 215, 216
526 Index

Atchafalaya River, USA 334 Bathurst Island, Canada 305 Cairngorm Mountains, UK 355
Atiu, Cook Islands 453 Bauxite 153 Calabrian transgression 470, 471
Atolls 450-3 Baymouth bars 331 Calcrete 153-6
Atterberg limits 166, 167 Bay of Fundy, Canada 321 Calderas 118, 119, 123
Augrabies Falls, South Africa 425 Beach cusps 329 California, USA 76
Aulacogens 94, 97, 98, 423, 424, 426, Beach drift 325 Callisto 501, 503, 505
429 Beaches 324, 325, 327-30 Caloris Basin, Mercury 498
Austerdalsbreen, Norway 270, 274 Beach face 328 Calving 263
Australia 102, 329, 331, 475 Beachrock 328 Cambering 170
Australia, eastern 100,427,469,473, Bed deformation 265,267 Cape Hatteras, USA 312
474 Bed load 200,212,282,332,363, 364 Capillary cohesion 164
Avalanche chutes 171 Bed-load channels 212, 214 Capillary fringe 130
Avalanches 169,170,171 Bed load measurement 201, 380 Capillary rise 130
Avulsion 211, 213, 334 Bedrock channels 198, 199,210,219, Capillary saturation 130
Ayers Rock, Australia 152 220,228 Capillary suction 130
Beheading 410,413,414 Carbonation 133, 134, 135
Back arc 61 Benioff zone (see Wadati-Benioff zone) Carbon dioxide 133-5, 150,475-7
Back-arc fold-thrust belt 62, 63 Benue Trough 94, 423, 424 Carrizo Plain, USA 414
Back-arc spreading 62, 63 Bergschrund 275 Cascade Range 62, 121
Backswamp zone 222 Berm 328, 329 Cascading systems 9
Backwash 319, 322, 328, 329 Berm crest 327, 328 Case-hardening 152, 153
Baffin Island, Canada 281,355, 356 Bernoulli effect 200, 237 Catalysts 132
Bagnold, R.A. 243,245 Beta Regio, Venus 499 Cataract 219
Baikal Rift 71, 75, 93 Bicarbonate ion 133 Catastrophism 4
Bajadas 14, 222, 223 Bifurcation ratio 208, 210 Catchment 207 (see also drainage
Balance year for glacial budget 263 Big Colewa Creek 416,418 basin)
Ballistic impact 237, 238, 243, 244 Biological weathering (see organic Cation exchange 136
Bank calving 173 weathering) Cation exchange capacity 136
Bankfull discharge 194, 228 Biokarst 150 Cations 130
Barbados 438 Biomass 393 Caves 230, 231
Barbed drainage 413 Bioturbation 146 Cave Stream, New Zealand 226
Barchanoid ridges 234, 244, 246, 247, Biscuit board topography 275 Cavitation 199
248,251,493 Bishop, P. 474 Cavity collapse 169, 174
Barchans 246,247,248,249,251,252, Blind valleys 229 c.e.c. (see cation exchange capacity)
253,339,494 Block fields 308 Cenotes 150
Barrier bars 330 (see also barrier Block slopes 307, 308 Central Gneiss zone, Himalayas 72
islands) Block streams 307, 308 Central Mountains, Taiwan 467, 468
Barrier beaches 330 (see also barrier Blowouts 254, 255, 256, 339 Centripetal drainage pattern 406
islands) Bolides 484, 486 Chad Basin, central Africa 85
Barrier islands 312, 324, 330, 331 Borneo 446 Channel armouring 202
Barrier reefs 325, 340, 450 Bornhardts 347, 348 Channel deposits 220
Bars 211,213,215 Bottomset beds 333 Channel discharge 192--4
Basal concavity 182 Bouguer anomaly 38, 39 Channeled Scabland, USA 282, 284,497
Basal flowage 276, 277 Boulder clay 275 (see also till) Channel initiation 195
Basal lodgement 276, 277 Bowen, N.L. 138 Channel margin deposits 220
Basal sliding 264, 265, 267 Bowen's reaction series 138 Channel networks 20,208,209
Base flow 193 Box canyons 7 Channels on Mars 7, 495-7
Base level 5, 217, 222, 224, 227, 445, Brahmaputra Basin, central Asia 384, Channels, river (see river channels)
446-9,458-61,463,464,468,469 386 Chaotic terrain, Mars 494,495,497
Basement 34 Brahmaputra River, central Asia 419, Characteristic form 184, 465, 466
Basic rocks 108 420,421,422,423 Characteristic form time 465,466
Basin and range (structural type) 41, 63 Braided channels 211, 212, 214, 215, Chari Basin, central Africa 384 (see
Basin and Range Province 14, 63, 100, 221 also Chad Basin)
416,491 Braiding 211 Chattermarks 269
Basin channel runoff 191, 192,359 Brazilian Shield 64, 65 Checkerboard Mesa, USA 128
Basin relief 209 Bretz, J.H. 282 Chelating agents 136
Basins (fluvial) (see drainage basins) Brunhes, B. 44 Chelation 136
Basins (lunar) 485, 487 BOdel, J. 457,462,463 Chemical bonding in minerals 130, 131
Basins (tectonic/sedimentary) 85, 95, Buttes 407, 408, 409 Chemical denudation 8,299,477
242 Chemical denudation rates 379,380,
Batholiths 121, 122, 123 Cailleux, A. 24 382-8,392-5,397
 Index 527

Chemical weathering 129, 130-44, 146, Colorado Rockies, USA 297 Crater rays 485
151,299,323,475-7 Columbia Plateau, USA 120,121,284 Craters (impact) (see impact craters)
Chenier plains 329, 330 Competence 201 Craton 41
Cheniers 329, 330 Complex dunes 248 Creep (of ice) 264
Chevron crevasses 269 Complex response 227 Creep (of rock) 168-70
Chezy equation 198 Composite volcanoes 117 (see also Crescentic bars 329
Chiang Jiang (Yangtze) Basin, China strato-volcanoes) Crevasses (on deltas and floodplains)
386 Compound dunes 248 222,336
Chronostratigraphic markers 436 Compound slopes 162, 183 Crevasses (on glaciers) 260, 269
Cinder cones 115, 116, 120 Compressive flow 266, 269 Crevasse-splay deposits 222, 336
Circularity ratio 209 Compressive stress 521 Critical flow 197
Cirque glaciers 262 Concordant intrusions 122 Crookwell River, Australia 474
Cirques 272,274, 275, 353 Conditionally stable slope 168 Crust 36, 37,39,40,41
Clay dunes 257 Conduit flow 230 Crustal uplift 371-5,377
Clay minerals 138, 139, 140, 141 Cone karst 151 Cryopediments 308, 309, 359
Cliffs 180, 325, 326 Cones 151 Cryoplanation terraces 308, 309, 359
CLIMAP 350,351 Cone sheets 123, 124 (see also altiplanation terraces)
Climate 22,293,294,295,358,359 Configurational aspects of reality 16 Cryostatic pressure 294,298, 304
Climate and landform genesis 23,24, Conrad discontinuity 40 Cuesta 407, 408
25,26,345-8,459,460,465,475-7 Continental drift 41,42,44,50-2,475, Cumulo-domes 115
Climatic change and landform genesis 477 Curie point 42
6,156,224,225,251,255,261,267, Continental flood basalts 100, 107, 110, Currents 321-3
268,293,297,301,307,344,347, 120 Cuspate forelands 331
348,353-65,425,439,440,442,443, Continental lithosphere 37 Cut-offs 211, 213
444,445,449,450,459,460,463, Continental-margin island arcs 59, 62, Cwms 274 (see also cirques)
466,475-7 63,65, 66 Cycle of erosion 5,457-60,463,464,
Climatic change, history of 349-52,353 Continental-margin orogens 59, 60, 465,468
Climbing dunes 254,256 62-5 Cymatogeny 462
Clinometer 182 Continental-margin upwarps (see
Clints 149 marginal upwarps) Danube delta, Rumania 334
Closed system pingos 304, 305 Continental platform 34, 35 Darcy's Law 230
Closed systems 9 Continental rifting 85,87,88,92-102 Darwin, C. 17, 18,450,451
Closure temperature 373, 514 Continental rifts (see rifts) Dating techniques 373-5, 513-15
Coastal deposition 324, 325, 327-40, Continental rise 34, 35, 96 Daughter isotope 513
449 Continental shelf 34, ~5, 96, 101, 449 Davis, W.M. 5, 17, 184, 185,457-60,
Coastal dunes 33fr9 Continental shield 41, 65 463,464,465,470,471
Coastal environments 313-15 Continental slope 34, 35, 96 Dawes crater, Moon 486
Coastal erosion 323, 324,325-7,449 Continuity equation 184 Dead-ice topography 280
Coastal weathering (see shoreline Contour curvature 177, 178, 182, 195 Dead Sea Rift 93
weathering) Contracted continental-margin orogens Death Valley, USA 14, 15,223,254
Coasts, classification of 313, 314, 449 63,65 Debris avalanche 168, 171, 179
Coasts, organic activity 325, 327, 328, Convection (mantle) 51, 52-4 Debris cones 163
331,339,340 Convergent boundary 47, 48, 75 Debris flow 169,170,171,222
Coasts, response to sea-level change Convergent margin coasts 314,316 Decay equilibrium 12
449-53 Convergent margins 57-75 Deccan basalt flows, India 110
Coasts, tectonic controls 314, 315, 316, Cook Islands 453 Deccan Plateau, India 120, 378
317,318 Coon Creek, USA 381 Decollement 70, 174
Cockpit karst 151 Coral 339, 340 Deep outflow 191, 192
Cockpits 151 Coral reefs 17,18,325,339,450-3 Deep Sea Drilling Project 45,390,391
Cohesion 164, 165 Coral terraces 373, 432 Deep weathering profile 137, 138, 141
Cold-based glaciers 263, 353, 354 Core 36 Defeated drainage 413
Cold ice 263 Corestones 137 Deflation 239,240,242,254,255,395,
Collapse dolines 150 Corrasion 199 397
Collapsed pingos 304 Corries 274 (see also cirques) Deflation hollows 241, 242
Collision margins 59, 60 Corrosion 199 De Geer moraine 278,280 (see also
Colluvium 163 Cotopaxi, Ecuador 64, 112 washboard moraine)
Colorado Basin, USA 387,394,395 Coulees 282 Delamination 91, 92
Colorado Plateau, USA 7,92, 124, Coulomb-Terzaghi shear strength Delayed flow 194
220,378,415,416,417,496 equation 164, 166 Delta front 333, 334
Colorado River, Texas, USA 362,363 Covalent bonds 130 Delta plain 334
Colorado River, USA 220,386,414, Crag-and-tail 279 Deltas 314, 315,331-6,337,338
415,416,417 Crater Lake, USA 118, 121, 122, 123 Dendritic drainage pattern 406,407
528 Index

Dendrochronology 389, 515 Dreikanter 240 Enhanced basal creep 265

Denudation 7, 8 (see also denudation Drift deposits 4 Entrainment 200
rates) Dropstones 349 Entropy 457, 458
Denudation chronology 468-74, 513 Drumlin fields 278, 279 Epeirogenesis 8 (see also epeirogeny)
Denudation rates 109, 178, 179, 180, Drumlinized ground moraine 278 Epeirogeny 8,86-102, 378, 379
371,372,373,374,379-400,467, Drumlins 278,279 Ephemeral stream channels 192,224,
468 Dry flow 169,170, 171 225,226
Denudation rates, long-term 388-92, Dry valleys 229 Episodic erosion 227, 464, 465
395,397,398,467,468 DSDP (see Deep Sea Drilling Project) Epoch (palaeomagnetic) 44
Depth-dependent extension 98,99 Dune network 234, 251 Equifinality 21
Desert pavements 239,240 Dunes (aeolian) 235, 238, 243,244-57, Equilibrium concepts in
Desert sands 235, 236 336-9,344 geomorphology 11, 12, 218,219,
Desert varnish 136 (see also rock Dunes (aeolian) on Mars 493,494 464,466-8
varnish) Dunes (channel bedforms) 199 Equilibrium line 264, 266, 268
Desiccation cracking 299, 302 Dungeness, UK 331 Equilibrium solubility 133
Detachability 179, 181 Duricrusts 153-7, 462, 463, 469 Equipotential surfaces 230
Detachment control 179 Dust storms 239,395, 396, 397 Ergodic hypothesis 18
Detachment faults 100, 101 Dust storms on Mars 492 Ergs 235,236,351, 364,366
Deterministic mathematical models 20 Dykes 118, 120, 122, 123, 124 Ergs on Mars 492
Dextral strike-slip faults 524 Dynamic equilibrium 5, 11,464,466 Erodibility 177, 392, 393
Diapirism 122 Dynamic metastable equilibrium 464, Erosion 8
Diapirs 121, 122 465 Erosion rates (see denudation rates,
Dietz, R. 45 Dynamic viscosity 196 mechanical denudation rates)
Diffuse flow 230 Erosion surface 12
Dilatancy 165, 279 Earthflow 169, 170, 171 Erosivity 177, 392, 393
Dilation cracking 299, 302 Earth, fundamental properties 483, Eskers 285, 286
Dione 501,502,503,505 505,506 Etchplain 462, 463
Dip 523,524 Earthquakes (see seismic activity) Etchplanation 462, 463
Dip slope 407 Earth slide 169, 178, 179 Etchsurface 462,463
Dip streams 407 East African Rift System 93,94,97, Euler crater, Moon 485
Discharge hydrograph 192 413,414 Europa 501, 502, 505
Discordant drainage 411 (see also East Antarctic ice sheet 349 Eustatic sea-level change 433
transverse drainage) East Australian Highlands 473,474 Evapotranspiration 22
Discordant intrusions 122 Eastern Cordillera, Andes 64,65 Event (palaeomagnetic) 44
Displaced terranes 77, 78 Ebro delta, Spain 334, 335 Evolutionary approach to
Dissociation 129 Echo dunes 254,256 geomorphology 11 (see also
Dissolution 132, 133 (see also solution) Eddy viscosity 196 historical approach to
Dissolved load 200 (see also solute Edge waves 322, 323, 329 geomorphology)
load) Effective normal stress 164, 165 Exfoliation 145
Diurnal tides 321 Effective porosity 228 Exfoliation domes 145
Divergent boundary 47, 48 Effusive eruption 113 Exhalative eruption 113
Dolines 150, 151 Eh 135, 136 (see also redox potential) Exogeneous dome 115
Dome dunes 246,247,248,250,254 Elastic behaviour 521 Exogenetic processes (see exogenic
Domes (volcanic) 115 Elastic limit 521 processes)
Dormant volcano 110 Elbow of capture 410 Exogenic processes 7, 8
Draa 243 (see also megadune) Elements (slope) 182 Exogenic processes, energy sources 21,
Drag force 200,237 Elie de Beaumont, L. 40 22
Drainage basin hydrology 191-5 Elongation ratio 209, 210 Explosion calderas 118
Drainage basin morphometry 208-11 Emiliani, C. 349 Explosive eruption 113
Drainage basins 207-10 Enceladus 506 Extending flow 266,269
Drainage basins, effects of sea-level End moraines 278,280 Extinct volcano 109, 110
change 445-9 Endogenetic processes (see endogenic Extra-orogen basins (drainage) 419,
Drainage density 208,209 processes) 420
Drainage divides 207 Endogenic processes 7, 8 Extrinsic threshold 179, 464
Drainage evolution 405,407,410-16, Endogenic processes, energy sources Extrusive igneous activity 107
420-7 21,22 Exxon global sea-level curve 437, 438,
Drainage inheritance 411 Endrumpf 458,461 439,441,444,445
Drainage in limestone terrains 228-31 En echelon crevasses 269
Drainage patterns 405-10 En echelon scarp 79 Facies 434
Drainage reversal 412 Englacial debris 271, 272 Fall 169, 173
Drakensberg Escarpment 84 England, UK 229,410,469-71 Falling dunes 254, 256
 Index 529

Fall velocity 202 Frontal arc 61 Glacier distribution 261-3

Fall Zone peneplain 472 Frost action 29~9 Glacier ice 263
False off-sets 413 Frost cracking 297, 298, 299, 302, 303 Glacier morphology 268,269
Fan-head trenching 223 Frost crack polygons 299 Glacier motion 264-8
Fault line 523 Frost creep 173, 299, 300 Glacier surges 267
Fault-line lakes 413 Frost heaving 297, 298, 302 Glacio-eustasy 437,438,439,440
Fault-line scarps 80 Frost shattering 146 (see also frost Glacio-isostasy 90, 91, 355, 358, 378,
Fault plane 523, 524 weathering) 379,434,443,444
Faults 45,48,49, 73-7, 79,80,92,523, Frost thrusting 297, 298 Glen Helen Gorge, Australia 412
524 Frost weathering 146, 147,275,297, Glen, J.W. 265
Fault scarps 79 323 Glen's power flow law 264, 265
Felsenmeer 146 Frost weathering on Mars 492 Global Positioning System 49, 51
Felsic minerals 108 Frost wedging 146 (see also frost Goldich, S.S. 137
Ferricrete 153-6, 462 weathering) Goldich's weathering'series 137, 138
Fetch 316 Froude number 197, 198, 199 Gondwana 42,50,51, 120,349, 424,
Firn 263 Fumaroles 114 425,441
Fission track annealing zone 375 Functional approach 11, 16 Gondwanaland (see Gondwana)
Fission track dating 374,375,389, 390, Goosenecks of the San Juan River,
391,472,514 Galilean moons (see moons of Jupiter) USA 220
Fixed ergs 235 Ganges Basin, India-Nepal 386 Gould, S.J. 16
Fjords 262,273,315,317 Ganges-Brahmaputra delta, Graben 92, 524 (see also half-graben)
Flexural isostasy 39, 91, 97, 101, 111 Bangladesh-India 62,66,333 Grade 218
Flexural node 96 Ganges River, India-Nepal 225, 420, Graded river 218,459
Flexural rigidity 91, 355 421 Gradualism 4
Floodplains 220-2, 224 Gangis Chasma, Mars 492 Grand Canyon, USA 416,417
Florida, USA 330 Ganymede 501, 505 Grand Erg Oriental, Algeria 248,249,
Flow 168, 169, 170, 171 Gardner Atoll, Hawaiian Islands 452 254
Flow regimes 197 Gelifluction 169, 170, 171,299,300 Granite intrusions 64, 65, 72, 121, 122,
Flow velocity 197-8 Gelifluction benches 307 390
Fluidization 171 Gelifluction lobes 307 Granular disintegration 145
Fluid threshold velocity 237, 238 Gelifluction sheets 307 Gravel-bed channels 202
Fluted ground moraine 277, 278, 279 Gelifluction streams 307 Gravity anomalies 37-9
Fluvial activity in periglacial Geochronometric dating 513,514 Gravity slides 174
environments 300,301 Geodesy 372 Gravity sliding 174
Fluvial denudation rates 371, 372, Geodetic techniques 372, 373 Gravity spreading 174
379-95,396 Geoid 33,433,439,442,443 Gravity tectonics 63,68, 70,174,175
Fluvial deposits 220--2 Geomagnetic reversals 44 Great Barrier Reef, Australia 325
Fluvial erosion 198, 199 Geomorphic threshold 226, 227,464 Great Dyke, Zimbabwe 122
Fluvial systems, response to climatic Geomorphology 3 Great escarpments 84, 86, 96, 97,101,
change 359-64 Geomorphology, history of 3-5 456,470,473
Fluvioglacial denudation 282-4,285 Geomorphology, relationship to other Great Escarpment, South Africa 84,
Fluvioglacial deposition 284-7 disciplines 3 456,470
Fly delta, Papua New Guinea 335 Geomorphometry 5 Greece 373
Flyggbergs 272,273 Geothermal heat flow 21,263,281 (see 'Greenhouse effect' 476
Footwall 92, 415 also heat flow) Greenland 260, 262, 355
Foraminifera 349 Geysers 87, 113 Greenland ice sheet 261,262
Forbes bands 269 (see also ogives) Gibber plains 157 Gregory, J.W. 92
Forearc 61 Gibbs free energy 131, 132 Grikes 149
Forearc basin 61, 62 Gilbert, G.K. 5, 11, 90, 124, 179, 183, Gross annual ablation (glacial) 263
Forebulge 91,355,358,444,453 219,464 Gross annual accumulation (glacial)
Fore dunes 254,256 Glacial abrasion 269,270,271,273 263
Foreset beds 333 Glacial activity on Mars 494, 495 Ground ice 294, 297, 298, 302-7
Fournier, F. 379, 380, 383 Glacial budget 263 (see also mass Ground moraine 277, 280
France 229, 410 balance of glaciers) Ground water sapping (see spring
Free-air gravity anomaly 38 Glacial denudation rates 397, 398 sapping)
Free dunes 24~54, 338, 339 Glacial deposition 275-80,355,357, Grus 347
Free face 182 358 Gulf of Suez, Egypt 32,99
Freezing front 297, 298 Glacial erosion 269-75,276,277, Gullies 177
Fretted channels, Mars 496 353-5,356,357 Guyots 34, 452 (see also seamounts)
Fretted terrain, Mars 492, 496 Glacial meltwater (see meltwater) Gypcrete 153, 154, 156
Fringing reefs 340, 450, 452 Glacial troughs 262, 270, 273, 274, 353
530 Index

Haast River, New Zealand 385 Hydraulic conductivity 194 Incised meanders 219, 220, 221
Hack, J.T. 464,466,471 Hydraulic drop 197 India 70, 71, 72, 73, 427
Hadley Rille, Moon 6, 485 Hydraulic geometry 215,216 Indo-Gangetic plain, India-Pakistan 70,
Half-graben 92 Hydraulic gradient 230 71,420421
Half-life 513 Hydraulic head 230 Indonesia 110
Hanging valleys 273, 274 Hydraulic jump 197 Indus River, Pakistan 420,421,422,
Hanging wall 92 Hydraulic radius 196, 197 423
Hardware models 20 Hydrocompaction 174 Indus-Tsangpo suture zone, central
Harker Glacier, South Georgia 262 Hydrogen ions 129 Asia 71,423
Hawaii 110, 111, 112, 115, 119,320, Hydrograph 192 Infiltration capacity 175, 194
451,452,490 Hydrographers Range, Papua New Infiltration-excess overland flow 175,
Hawaiian-Emperor chain 451,452 Guinea 389, 390 194,195
Hawaiian Islands 109, 110, 111, 115, Hydro-isostasy 439, 449 Ingrown meanders 219
451 Hydrological cycle 22, 23 Input 9
Headland bay beaches 329 Hydrolysis 133 Inselbergs 185, 346, 347, 461
Headwall 274, 275 Hydrostatic pressure 130 Insolation weathering 145, 146
Headward erosion 410, 411 Hydrothermal activity 113 Intact strength 165, 166
Heat flow 21,22,61 (see also Hydroxyl ions 129 Interarc basins 62
geothermal heat flow) Hyperpycnal flow 332, 333 Intercontinental collision orogens 59,
Heave 168, 169, 170, 173 Hypopycnal flow 332,333 66-75
Henry Mountains, USA 124,125 Hypsometric analysis 34, 210, 211, 435 Interflow 175, 191, 192
Hess, H.H. 44 Hypsometric curves 34,211,425,426, Intermediate rocks 108
High-tide flats 331 440,499 Intermittent stream channels 192
High-tide platforms 327 Hypsometric integral 210, 211 Internal deformation 264
Hillslope hydrology 175,176 Hypsometry, global 34, 42, 499 Intertidal platforms 326, 327
Himalayas 56, 70-3, 223, 225, 375, 377, Intertidal slope 331
399,420,421,422,423,442 Iapetus 503,505 Intertidal zone 321
Hinge zone 96 Ice age, Gondwana 42 Intra-cratonic basins 95
Historical approach to geomorphology Ice age, Late Cenozoic 349,350,351, Intra-oceanic island arcs 57, 58, 59-62
11, 16, 457 (see also evolutionary 352,353-9 Intra-orogen basins (drainage) 419, 420
approach to geomorphology) Ice age, Pleistocene (see ice age, Late Intrenched meanders 219,220,221
Hjulstrom, F. 201,202 Cenozoic) Intrinsic threshold 179
Hogback 407, 408 Ice caps 262 Intrusions 107, 121-5
Hoggar Plateau (see Ahagger Plateau) Ice domes 262 Intrusive igneous activity 107, 120-5
Holmes, A. 42, 45 Ice falls 266 Inverted topography 157, 407
Homeostasis 10 Ice fields 262 10500,501,502,505
Homoclinal shifting 407 Iceland 87,110, 111, 113,281 Ionic bonds 130
Homopycnal flow 332, 333 Icelandic troughs 273 Iraq 373
Honeycomb weathering 152, 297 Ice lenses 298 Irrawady River, Burma 380
Hooke's Law 521 Ice sheets 262 Island arcs 36, 57-63
Horn 272, 275, 276 Ice shelf 262 Isle of Arran, UK 358
Horst 524 . Ice streams 266 Isolated system 457
Hortonian overland flow 175 (see also Ice-wedge casts 303,357, 359 Isostasy 8,38,39,40,91,95,97,100,
infiltration-excess overland flow) Ice-wedge polygons 292,303,359,494 111, 120, 355, 358, 439, 442 (see also
Horton, R.E. 5, 175, 195,208 Ice wedges 302, 303, 304, 359 flexural isostasy, glacio-isostasy,
Hot spots 87-90, 109, 110, 111,440, Idaho, USA 388 hydro-isostasy, local isostasy,
452,499 Igneous activity 8, 61 (see also thermal isostasy)
Hsii, K.J. 442 extrusive igneous activity, intrusive Isostatic anomalies 39
Huangfuchuan River, China 385 igneous activity, volcanism) Isostatic rebound (see isostatic uplift)
Huang He (Yellow) Basin, China 384, Ignimbrite 113, 115, 118, 120 Isostatic uplift 67,68,72,87,88,90,
386 Imbrium Basin, Moon 6,485,487 91,101,355,358,372,379,460,
Huang He (Yellow) River, China 206, Immanent aspects of reality 16 462,463,472
211 Impact cratering 8, 484, 485, 486 Isovolumetric weathering 137,178,383
Hudson Bay, Canada 355 Impact craters 7, 484-7, 489, 497, 498 Isthmus of Corinth, Greece 373
Huon Peninsula, Papua New Guinea Impact craters, lunar 484-7
373,375,376,432,434,435,437, Impact craters, Martian 489
438 Impact structures on Earth 8 Japan 62,399
Hutton, J. 4 Impact threshold velocity 237, 238 Japura (Rio), Brazil 363, 364
Hydration 133, 146 Impeded dunes 246, 254-6, 338, 339 Java 66,108
Hydration shattering 146,147,275,297 Incident waves 323 Jokulhlaups 113, 281, 282
Hydration weathering 146 Incipient suspension 200 Joint-block removal 269, 270, 271
 Index 531

Joints 112, 122, 145, 149, 150, 151, 166, Landscape sensitivity 465, 466 Lower-plate margins 101
183,229,230,273,405,407,409, Land use and denudation rates 380, Low-tide platforms 327
410,523 392,393,394 Low velocity zone 37
Jurua (Rio), Brazil 363, 364 Lapies 149 Lunettes 254,255,257
Lapilli 112 Lut Desert, Iran 241
Kalahari Basin, southern Africa 85,95 Late Heavy Bombardment 488,500, Lyell, C. 4, 16, 450
Kalahari Desert, southern Africa 255, 501,502,504,506
344,364,366 Lateral accretion deposits 221, 222 Maars 115
Kame complex 285, 286 Lateral moraine 260,262,280 Mackenzie delta, Canada 304
Kame delta 285, 286 Laterite 153 Mackin, J.H. 218
Kames 285, 286 Laurasia 42, 50, 51, 349 Macrotidal coasts 321
Kame terraces 286 Laurentide ice sheet 351, 352, 355, 356, Maddock, T. 215,216
Kamikaze grains 493 397,448 Mafic minerals 108
K-Ar dating 373, 375, 513, 514 Lava 107,111, 112, 115,485,486 Magma 107, 111
Karkevagge, Sweden 178 Lava channels 6, 112 Magmatic arc 61
Karoo flood basalts 120 Lava flows 107, 110-12, 114-18, 120 Magnetic anomalies 44,45,46
Karren 149, 150 Lava levees 112 Magnetic reversals (see magnetic
Karst 148-51, 228-31 Lava tubes 112 anomalies)
Karstification 148 Lava tunnels 112, 485 Magnitude and frequency of
Karst streets 151 Layered terrain, Mars 494 geomorphic events 10,11, 12,227,
Kauai, Hawaiian Islands 115, 118 Leaching 132,137, 140, 141, 143 228
Keewatin District, Canada 357 Lee dunes 254,256 Magnitude-frequency relationships 11,
Kegel 151 Lena delta, USSR 337 227, 228, 465
Kegelkarst 151 Length of overland flow 208, 209 Main Boundary Thrust, Himalayas 70,
Kelvin, Lord 40 Leopold, L.B. 215,216 71,72
Kerguelen Island 297 Levees 221, 222, 334 Main Central Thrust, Himalayas 71,
Kettle-and-kame topography 286 Lift force 200,237 72, 73
Kettled-sandur 287 Limestone weathering 148-51 Main slope 182
Kettle-holes 286 Limiting angle of stability 183 (see also Makran Coast Range, Pakistan 316
Kinematic viscosity 196 threshold angle of stability) Managaia, Cook Islands 453
Kinematic waves 266,267 Linear dunes 234, 244, 246, 247, 248, Mangroves 325,331
Kinetics of weathering reactions 131, 249,250,252,253,364,366 Mangrove swamps 331
132, 138, 139, 141, 142, 144 Linton, D.L. 469-71 Manifestly underfit streams 361
King, L.c. 184, 185, 457, 458, 461, Liquefaction 167 Manning equation 198
462,463,472 Liquid limit 166 Manning roughness coefficient 198, 199
Knickpoint 217,219,224,425 Listric faults 92, 414, 415, 524 Mantle 36, 37
Kolyma Basin, USSR 384 Lithosphere 37, 38, 87, 89, 90, 95, 440, Mantle plumes 87, 100
Kosi Fan, India 223,225 441,451 Manuae, Cook Islands 453
Kuiseb River, Namibia 248 Lithospheric extension 93-101 Marafi6n Basin, Peru 387
Kusaie, Caroline Islands 452 Lithospheric plates (see plates) Mare 484 (see also maria)
Kutai delta, Indonesia 335 Lithospheric stretching (see Mare Imbrium, Moon 485
lithospheric extension) Marginal sea coasts 314
Labyrinth karst 151 Lithospheric vulnerability 89 Marginal upwarps 9fr.101
Laccolith 123, 124 Little Colorado River, USA 7 Maria 484 (see also mare)
Lachlan Basin, Australia 391, 474 'Little Ice Age' 267 Mariner space programme 488,492,
Lag deposits 239, 240 Littoral zone 313 493, 495, 497
Lagoon 330 Local isostasy 91,474 Mars 7, 483, 488-97, 505, 506
Lag time 465,466 Local relief 209, 392 Mars, landscape history 497
Lahar 113 Loess 235, 238, 239,256,257,301, 350, Mass balance of glaciers 263, 264, 266,
Lake Bonneville, USA 283 352 267 (see also glacial budget)
Lake Eyre Basin, Australia 85, 95 Loess-lips 257 Mass displacement 298, 302
Lake Kyoga, Uganda 413 Loess Plateau, China 206 Mass movement 167-74, 299, 307, 308
Lake levels, changes in 351, 352 Longitudinal dunes 246 Mass movement, on the Moon 486,487
Lake Missoula, USA 281,282,283 Longitudinal profile 216,217,218,219, Mass movement, on Mars 492
Lake Tanganyika, East Africa 414, 415 425,426,446,474 Mass wasting 167
Lake Victoria, East Africa 413 Long profile 217 (see longitudinal Mathematical models 20, 21
Laminar flow 196 profile) Matthews, D.H. 45
Laminar sublayer 196, 200, 202 Longshore bar 329 Mature passive margins 96, 97
Landscape evolution 5,12,16,17, 101, Longshore currents 322, 323 Mature passive margin coasts 314, 317
102,275,277, 450-3, 457-77 (see Longshore drift 324,325,329,331 Maturity (landscape evolution) 458
also slope evolution) Lopoliths 122, 123 Matuyama, M. 44
532 Index

Mauke, Cook Islands 453 Monument Valley, USA 409 Nigardsbreen, Norway 281
Mauna Kea, Hawaii 115 Moon 6, 7, 483, 484-8, 505, 506 Niger Basin, West Africa 425
Mauna Loa, Hawaii 115,117,119 Moon, landscape history 487,488 Niger delta, Nigeria 334, 335, 426
McDougall, 1. 389 Moons of Jupiter 500,501,502, 503, Niger River, West Africa 423, 425
McKenzie, D. 97 506 Nile Basin 425
McMurdo Sound, Antarctica 297 Moons of Neptune 504,505 Nile delta 334
Meade, R.H. 383 Moons of Saturn 501-3,506 Nile (River) 442, 449
Mean annual flood 194 Moons of Uranus 503,504,506 Nirgal Vallis, Mars 7,496
Meander cut-off 213,221 Moraine 275,277-80 Niue 452
Meandering channels 211-15, 218, 219, Morphoclimatic zones 23, 24, 25, 345 Nivation 274,299,300
220,221,361,362 Morphogenetic region 23 (see also Nivation hollows 299
Mechanical denudation 8 morphoclimatic zones) Noncontracted continental-margin
Mechanical denudation rates 361, Morphological capacity 115, 118 orogens 63
379-88,392-8 Morphological systems 9 Non-uniform extension 98, 99
Mechan River, Canada 300 Morphometry 5 Normal faults 523, 524
Medial moraines 260,262,279 Morphotectonics 8 Normal listric faults 524
Medina River, USA 228 Most probable annual flood 194 Normal stress 164
Mediterranean Sea 70,442,449 Motueka River, New Zealand 324 North Africa 395,397
Megadunes243, 244, 245,246,248, Moulins 281 North America, eastern seaboard 389,
251,364 Mount Cook, New Zealand 276 390
Meltout 276, 277 Mount Ellsworth, USA 125 North American Cordillera 77, 78
Meltwater 267,269,277,279,281-5, Mount Everest, China-Nepal 56, 423 Nuees ardentes 113
361,397 Mount Mazama, USA 121 Nunataks 272,275
Meltwater channels 283, 284 Mount St Helens 113,115 Nye, I.F. 265,268
Mentawai Islands, Indonesia 62, 66 Mount Taranaki (Egmont), New
Mercury 483,497,498,505 Zealand 106 Oahu, Hawaiian Islands 452
Merevale Basin, UK 388,389 Mudflows 169, 170 Oberlander, T. 422
Mesas 407, 408, 409 Mud volcanoes 114 Oberon 503
Mesotidal coasts 321 Muller, S.W. 294 Oblique-slip faults 524
Mesters Vig, Greenland 299,307 Murgab River, USSR 194 Oblique-slip margins 75-7
Methodological uniformitarianism 16 Murray Basin, Australia 391, 474 Ocean basins 34, 35
Meybeck, M. 384 Murrumbidgee River, Australia 361, Oceanic islands 450-3
Microtidal coasts 321 362 Oceanic lithosphere 37,440,441,451
Mid-Atlantic Ridge 45, 87 Oceanic trenches 35, 36
Mid-oceanic ridges 34, 35, 440, 441 Namib Desert, Namibia 244,248,348 Offset drainage 74,413,414
Milankovitch mechanism 350 Namib Sand Sea, Namibia 244,248 Offshore bar 330 (see also barrier
Miller, J.P. 227 Nappe 69,524 islands)
Milliman, I.D. 383 Nascent passive margins 96,97 Ogives 269, 270 (see also Forbes bands)
Miranda 503, 504 Nascent passive margin coasts 314 Ohmori, H. 392, 393
Misfit streams 361 (see also underfit Natal, South Africa 391,392 Okavango River, southern Africa 344
streams) Natural arch 220 Old age (landscape evolution) 458
Mississippi Basin, USA 380,381,416, Nauru 452 Old Faithful geyser, USA 87
417,418,419 Neap tides 320 Olympus Mons, Mars 117, 489, 490
Mississippi delta, USA 332,334, 335, Nebkhas 254, 255 (see also shrub- Open channel flow 195-8
336,338 coppice dunes) Open system pingo 304, 305
Mississippi River, USA 224,332,334, Needle ice 294,298 Open systems 9
338,380,381,416 Negative feedback 10, 197, 218, 264, Open troughs 273
Mississippi Valley, USA 448 477 Ophiolite 61, 71
Mitiaro, Cook Islands 453 Negro Basin, Brazil 387 Orange Basin, southern Africa 425
Mixed-load channels 212,214 Neocatastrophism "16 Orange River, southern Africa 219,
Mixed tides 321 Neotectonics 8 386,425
Mogotes 151 Neotectonic stress field 409,410 Organic weathering 136, 146, 150
Moho 36,37 Net annual ablation of glacial ice 263, Orientale Basin, Moon 485
Mohorovicic discontinuity 36 (see also 264 Orogenesis 8 (see also orogeny)
Moho) Net annual accumulation of glacial ice Orogenic mountain belts 34 (see also
Mojave Desert, USA 347,348 263,264 orogens)
Molecular viscosity 196 Net specific balance of glaciers 263 Orogens 34, 35, 56-60,62-78,375-8,
Monadnock 459 Network 208 398,399,420-3
Monocline 524 New Madrid earthquake, USA 416 Orogeny 8,56-60,62-78,375-8,398,
Monroe uplift, USA 416,418 Newtonian fluid 522 399,442,475,477
Montes Apennines, Moon 6, 485 New Zealand 76,77,113,399 Osage-type underfits 361
 Index 533

Oscillatory waves 318 Pediplanation 461 Poljes 151

Ouchi, S. 417 Penck, VV. 5, 184, 185,457, 458,460, Polycyclic landscapes 459, 460
Outer-arc ridge 61, 62, 66 461,463 Ponors 151
Outflow channels, Mars 495, 496, 497 Peneplain 458, 459 Pools 217, 218
Outlet glaciers 262 Peneplanation 458-60 Pore-water pressure 164, 168,267
Output 9 Peneplane (see peneplain) Positive feedback 10, 267, 274
Outwash plains 287 (see also plain Penetrative magmatism 87 Postrift sediments 96
sandar) Periglacial activity 274, 275 Potholes 219
Overbank deposits 220 Periglacial activity on Mars 494 Powell, J.VV. 5
Oxbow lake 221 Periglacial environments 293-6, 357-9 Pratt model 38, 39
Oxidation 135, 136 Periglacial involutions 298 Pre-Nectarian terrane, Moon 487
Oxygen isotope analysis 349,350,351, Periglacial landforms , relict 355, 357, Pressure head 230
434 358,359,360 Pressure melting point 263
Oxygen isotope stage numbers 350, Period of waves 315, 318 Pressure release 145, 174, 270
351,437,438 Peripheral foreland basins 66, 67-9 Primarrumpf 461
Permafrost 293-6 Primary dunes 338,339
Pahoehoe 112 Permafrost on Mars 489, 492 Primary porosity 228
Paired terraces 224, 226 Permafrost table 294 Principle of original horizontality 513
Palaeomagnetic anomalies (see Peru 373 Principle of stratigraphic superposition
magnetic anomalies) P-forms 283 513
Palaeomagnetic dating 514 pH 129 Process-response systems 9
Palaeomagnetic time scale 349,350, Phase changes 90, 95 Proglaciallakes 281,282,287
351 Phoebe 503 Pseudokarst 148
Palaeomagnetism 42,44,45,46,514 Phreatic caves 231 Pull-apart basins 75, 76
Pallid zone 137 Phreatic zone 230 Pumice 112
Palsas 304, 305 Physical weathering 129, 144-8,296-9 Purus (Rio), Brazil 365
Palynology 349 Phytokarst 150 Push moraines 278, 280
Pangaea 42, 50,85, 95, 102,424,440 Piceance Basin, USA 389 P-waves 36
Pans 242,255,344 Piedmont angle 346, 347 Pyroclastic flow deposits 111, 113, 115
Parabolic dunes 254, 255, 339 Piedmottreppen 461 Pyroclastic flows 113
Parallel drainage pattern 406 Pillow lava 112 Pyroclasts 111-15
Parallel retreat 185, 186, 458, 461, 463 Pingos 303, 304, 337, 357, 359
Parana Plateau, Brazil 120 Pinnacle karst 151 Quarrying 324
Parent isotope 513 Pipes 194, 195 Qattara Depression, Egypt 242
Paris Basin, France 409 Pitman, VV.c. III 441 Queanbeyan Basin, Australia 384
Partial area model 194 Plain sandar 287 (see also outwash Quick clays 167
Partial duration series 194 plains) Quick flow 194
Passive continental margins 85 (see also Planation surface 12 (see also erosion
passive margins) surface) Radial crevasses 269 (see also splay
Passive margin basins (drainage) 418, Planetary geomorphology 6, 7, 482-507 crevasses)
419,420 (see also Afrotype drainage Plastic behaviour 521 Radial drainage pattern 406
basins) Plasticity index 167 Radioactive decay 513
Passive'inargin basins Plastic limit 166 Radiocarbon dating 376, 514
(tectonic/sedimentary) 95 Plate boundaries 46, 47, 48 Radiometric dating 373-5,376,389,
Pafsive margin coasts 314 Plate margins 47, 58 390,487,513,514
Passive margin distal-orogen basins Plate movement 46, 47, 48, 49,51-4, Rainpits 149
(drainage) 418,419,420 (see also 71 Rainsplash erosion 175, 176, 178, 183
Amerotype drainage basins) Plates 46 Raised beaches 358, 434, 443
Passive margin distal-orogen coasts 314 Plate tectonics 5,6,46,47,75,475-7, Rapids 219
Passive margins 85,86,91,95-102, 499 Rapp, A. 178,299
423-7,436,445,449,468,471-4 Playfair, J. 4 Rarotonga, Cook Islands 453
Passive rifting 93, 94, 97-101 Plug domes 115 Rates of aeolian denudation (see
Passive tectonic controls of drainage Plunge pools 219 aeolian denudation rates)
405-12 (see also structural controls Plunging breakers 319,320,323,324 Rates of chemical denudation (see
of drainage) Plunging cliffs 325 chemical denudation rates)
Patterned ground 147,148, 153,301, Pluton 107 Rates of denudation (see denudation
302,303,357 Pocket valleys 229 rates)
Pearl and Hermes Reef, Hawaiian Point bars 214, 215 Rates of fluvial denudation (see fluvial
Islands 452 Point-bar deposits 221, 222 denudation rates)
Pediments 185, 346-8, 461 Polar wandering 42 Rates of geomorphic processes, units
Pediplains 458,461 Pole of rotation 48, 49 371,372,382,383
534 Index

Rates of glacial denudation (see glacial Rift valleys 34, 85, 88, 92-4, 414, 415 Satellite altimetry 372
denudation rates) Rills 177, 195 Saturation overland flow 175, 194, 195
Rates of mechanical denudation (see Rip currents 322, 323 Scale in geomorphology 12-16
mechanical denudation rates) Ripples 243, 244, 245 Scale in hardware models 20
Rates of slope processes (see slope River capture 410, 411 (see also stream Schmidt hammer 165,166
processes, rates of) piracy) Schooley peneplain 472
Rates of uplift (see uplift rates) River channel patterns 210-5,361-4, Schumm, S.A. 17,211,361,398,464
Rates of uplift and denudation 365,416-8, 419 Scoria 112
compared 372, 398-400 River channels 210-20,360,361-4, Scotia Arc 62
Rectangular drainage pattern 406,409, 365,416-8,419 Scotland, UK 355
410 River regimes 192, 300, 301 Scree slopes 163
Recumbent fold 69, 524 River terraces 224, 226, 448 Sea-floor spreading 44, 45, 46, 476, 477
Recurrence interval 10, 11, 179, 193, Roches moutonnees 272, 273 Sea-level change 325, 326, 327, 330,
194,227,228 Rock (as slope material) 163, 165, 166 333,358,372,373,432-53
Redox potential 135, 136 (see also Eh) Rock bars 273, 274 Sea-level change, evidence for 434-7
Red Sea 32, 97, 99 Rock basins 272, 273 Sea-level change, impact on landforms
Reduction 135, 136 Rock drumlins 272, 273 325,326,327,330,333,445-53,476
Reefs 340 Rock fall 169, 173, 174, 178, 179 Sea-level change, mechanisms of
Regelation 272 Rock flour 269, 397 439-45
Regelation creep 265 Rock glaciers 308 Sea-level change, record of 437-9,476
Regolith 136 Rock mass strength 166, 182 Sea-level rise, post-glacial (Holocene)
Regolith, lunar 486 Rock pediments 346, 347 325,330,358,372,373,437,442,
Regolith, Martian 491,492 Rock varnish 136, 347 (see also desert 443,444,447,448,449
Regression 433 varnish) Seamounts 34, 109, 110,451,452
Rejuvenation 459 Rogen moraines 278, 280 (see also Sea waves 317
Relative dating 513 ribbed moraines) Secondary convection 99
Relaxation time 12,267,465,466 Roll-over zone 92, 414, 415 Secondary dunes 338, 339
Relict dunes 251, 364, 366 Roof pendants 122, 123 Secondary minerals 138, 139, 140
Relict ergs 235, 364, 366 Rotational slide 169, 172, 492 Secondary porosity 228
Relict landforms 24, 348, 353-66, 475 Rounded solution runnels 149 Sediment delivery ratio 380, 381
Relief 7, 209, 210 Rub'al Khali, Saudi Arabia 234,235, Sediment deposition in river channels
Relief ratio 209 254 202,203
Remanent magnetism 42 Ruggedness number 209 Sediment entrainment by wind 237, 238
Remnant arcs 62 Runcorn, S.K. 42 Sediment entrainment in river channels
Remnant dunes 339 Runoff (see basin channel runoff) 199-202
Remnant-ocean basins 66,67, 70 Runoff channels 496 Sediment load (see solid load)
Remote sensing techniques in Runoff generation 194, 195 Sediment movement on coasts 324, 325
geomorphology 17, 235, 236, 248, Russell Glacier, Greenland 272 Sediment rating curve 379
250,395,506-7 Ruxton, B.P. 389 Sediment transport 200-2, 227
Resurgence 229 Sediment transport by wind 237, 238
Retroarc basins 62, 63 Safety factor 167 Segments (slope) 182
Retroarc foreland basins 66, 67 Sag ponds 79, 80 Segregated ice 294,297, 298
Return flow 194 Sahara Desert 359, 360, 395, 397 Seif dunes 252, 253
Return period 10 (see also recurrence Saidmarreh anticline, Iran 69 Seismic activity 11, 43, 44, 375
interval) Saidmarrey slide, Iran 173 Seismic reflectors 436
Reverse faults 523, 524 Salinas 334 Seismic sea waves 320
Reverse listric faults 524 Saltation 200,237,238,243,244 Seismic stratigraphy 92,96,435-7
Reversing dunes 246,247,253,254 Salt intrusions 68 Selima Sand Sheet, North Africa 359,
Reynolds number 196,197,198,202 Salt marshes 331 360
Rhea 501, 503, 505 Salt weathering 147, 148,297,323 Semi-controlled channels 210
Rheid flow 121 Salt weathering on Mars 492 Semi-diurnal tides 320
Rheidity 121 San Andreas Fault System, USA 76, Senegal delta, Senegal 334, 335
Rhone Basin, France 390 413,414 Sensitive soils 167, 170
Rhone (River), France 442,449 Sandar 285, 287 Settlement 169, 174
Rias 315 Sandur (see sandar) Seychelles 450
Ribbed moraines 278, 280 (see also Sand waves 329 Shatt-el-Arab (River), Iran-Iraq 386
Rogen moraines) Sand wedges 303 Shear box 164
Riffles 217,218 San Juan River, USA 220 Sheared margins 86, 97, 98
Rifted margins 86, 95-102 Sao Francisco delta, Brazil 334 Shear plane 164
Rifts 88, 92-4, 96-101 (see also rift Saprolite 137, 347, 383 Shear strength 163-8, 179
valleys) Sastrugi 269 Shear stress 163-8, 200-2, 264, 521
 Index 535

Sheet flow 176 Solution notches 149 Stream segment 208

Sheets 247, 248, 250 Solution pans 149 Stress 521
Shield volcanoes 115,117,118,486, Solution pipes 149 Striations 269
490 Solution ripples 149 Strike 523
Shields diagram 202 Solution runnels 149 Strike drainage 411
Shoalhaven Valley, Australia 469, 473 Solution spikes 149 Strike-slip faults 71, 73, 74, 79, 80, 413,
Shoreline weathering 323 South Australia 158, 346, 408 414,523,524
Shore platforms 326, 327 South Cascade Glacier, USA 397 Strike valley 407
Shreve, R.L. 208 Southern Africa 100, 124, 156, 183, Structural benches 224
Shrinkage limit 166 186,392,447,469,472,473 Structural controls of drainage 405-12
Shrub-coppice dunes 254, 255 (see also Southern Alps, New Zealand 76,77, (see also passive tectonic controls of
. nebkhas) 171,370,375,390,399,467 drainage)
Shutter ridges 79, 80 South Island, New Zealand 467 Subcritical flow 197
Sichelwannen 283,285 Space-time substitution 17,18, 19,450, Subduction 46,47,67,68,69, 70, 72,
Silcrete 153-7, 158 451,468 73
Sills 118, 120, 123, 124, 185, 186 Specific conductance 382 Subduction zones 46,54,57-61,65,
Simpson Desert, Australia 248,252 Specific stream power 201 102
Simpson, G.G. 16 Speleothems 231 Subglacial debris 271, 272
Simulation models 20 Spilling breakers 319 Subjacent karst collapse dolines 150
Sinistral strike-slip faults 524 Spits 324, 331 Sublimation 263
Sinkhole 229 Splay crevasses 269 (see also radial Sub-lithospheric thermal anomaly
Sinter 113 crevasses) 87-9,96,109,110,451,452
Sinter mound 114 Spreading ridges 47 Submarine canyons 447
Sinuosity 211 Spring sapping 7, 195, 229 Subsidence (localized, of ground
Slide 168, 169, 170, 172, 178 . Spring tides 320 surface) 169, 174
Slip-face 245, 247 Stable slope 167 Subsidence (tectonic) 95-101, 372, 373,
Slip lines 266 Stage 192 379
Slope decline 184, 185 Stage-discharge rating curve 192 Subsidence calderas 118
Slope evolution 17,18, 184-8,459-61, Stage hydrograph 192 Subsidence dolines 150
463 Stage recorder 192 Substantive uniformitarianism 17
Slope form 179, 181-4 Stagnant ice 263 Subsurface drainage 229-31
Slope material behaviour 163-8 Stalactites 231 Subsystem 9
Slope material properties 163-7 Stalagmites 231 Subtidal zone 331
Slope processes 163-81 Star dunes 234, 244, 246, 247, 249, 250, Suess, E. 433
Slope processes, rates of 178-80 254 Sumatra 113, 446
Slope profiles 179, 181, 182 Static equilibrium 11 Sunda Arc 62,65,66
Slope replacement 185 Steady-state equilibrium 11, 464 Sunda Shelf 62, 446
Slope stability 167, 168 Steady-state margins 57 Supercontinent 102
Slope wash 175-8, 183, 184 St Lawrence Basin, North America 387 Supercritical flow 197
Snow avalanche 169, 170 Stochastic mathematical models 20 Superimposition 411,421,422,423 (see
Sogne fjord, Norway 317 Stocks 123, 124 also superposition)
Soil (as slope material) 163, 166, 167 Stone pavements 239 (see also desert Superposition 411 (see also
Soil creep 169, 173, 178, 183 pavements) superimposition)
Soil erosion 177, 178 Storm hydrograph 193 Supraglacial debris 271, 272
Solar radiation 8, 21, 22 Storm surges 320 Supraglacial flowage 277
Solfataras 114 Strahler, A.N. 208 Surf 318
Solid load 379-82, 387, 396 (see also Strain 521 Surface creep 237, 238, 243
bedload, suspended load) Strain rate 522 Surface detention 175
Solifluction 169, 170, 171, 178, 179, Strath terrace 224 Surface uplift 371,372, 373, 375, 376,
183,299 Strato-volcanoes 108, 117, 120 377,378, 379
Solimoes (Rio), Brazil 363, 364, 365 Streaks 247, 248, 250 Surging breakers 319
Solute load 200, 210, 282, 283, 379, Stream discharge (see channel Suspect terranes (see displaced
380,382,387,394,396 discharge) terranes)
Solute rating curve 379 Stream frequency 209 Suspended load 200,212,282,283,
Solute transport 175, 178, 179,200, Stream magnitude 208 332,363
201,382 Stream networks (see channel Suspended-load channels 212, 214
Solution 132, 133, 148-51,323 (see also networks) Suspended load measurement 201, 380
dissolution) Stream order 208,209 Suspension (in air) 237, 238, 239
Solution bevels 149 Stream piracy 410, 411 (see also river Suspension (in water) 200
Solution dolines 150 capture) Suture zone 59,66,67,69
Solution flutes 149 Stream power 201 Swains Island, American Samoa 452
536 Index

Swash 319, 328, 329 Tidal flats 331 Undermelt 276

S-waves 36 Tidal inlets 330 Underplating 100, 101
Swell waves 317, 323 Tides 320, 321 Uniform extension 97, 99
Switzerland 409, 410 Till 275 Uniformitarianism 4, 16, 17, 450
Syncline 524 Tillites 42 Uniformity of law 4
Synrift sediments 96 Time-dependent landscapes 466-8 Uniformity of process 4
Systems 9 Time-independent landscapes 466-8 Uniformity of rate 4
Systems, geomorphic 9, 10 Titan 501, 505 Uniformity of state 4
Systems analysis 8, 9, 457 Topographic maps, use in monitoring United Arab Emirates 248, 249
landform change 17 Universal Soil-Loss Equation 177
Tafoni 152, 297 Topological properties of channel Unpaired terraces 224,226
Taiwan 467, 468 networks 208 Uplift 8, 63, 64, 65, 67-9, 71, 73, 76,
Taliks 294 Topple 169, 173 77,86-102,124,371-9,398-400,
Talluvium 163 Tors 151, 346, 462, 463 458-75 (see also crustal uplift,
Talus 163 Tower karst 151, 152 isostatic uplift, surface uplift)
Talus creep 169, 173 Traction load 200 Uplift rates 371-9,398-400,467,468
Talus slopes 163, 183 Tractive force 200 Uplift rates, measurement 372-5
Tarn Gorge, France 229 Transcurrent faults 524 Upper-plate margins 100, 101
Tasman Bay, New Zealand 324 Transfer faults 92 Upslope convexity 182, 183
Tectonic denudation 175 Transform boundary 47, 48, 75 Ural Mountains, USSR 69
Tectonics 8 Transform faults 45,48,49 USA 14, 378
Tectonics, lunar 486 Transgression 433 Uvalas 151
Tectonics, Martian 491 Transgressive dunes 339
Tefe (Rio), Brazil 382 Transient forms 465, 466 Vadose caves 231
Tensile stress 521 Translational slide 169, 172 Vadose zone 230
Tephra 111-15, 120, 121 Translatory waves 318 Vail, P.R. 436
Tephrochronology 113 Trans-orogen basins (drainage) 419, Valles Marineris, Mars 491,492
Terraces (see river terraces) 420 Valley glaciers 262
Terracettes 171, 173 Transport-limited slopes 179, 181 Valley margin deposits 220
Terra Ishtar, Venus 499,500 Transpression 75 Valley meanders 361
Tethys 501,502, 505 Transtension 75 Valley sandar 287
Tethys Sea 50 Transverse crevasses 269 Variable source area model 194
Thalweg 211 Transverse drainage 411,412, 421-3 Varves 515
Thar Desert, India 255 (see also discordant drainage) Vatnajokull ice cap, Iceland
Tharsis Montes volcanic province 490, Transverse dunes 246, 249, 250, 251, Venera space programme 498,499
491 252,339 Ventifacts 240
Thaw lakes 305, 306, 337 Transverse orogens 75, 76 Ventura anticline, USA 375
Thaw slumps 306 Transverse Ranges, USA 76,375,377 Venus 483, 493, 498-500, 505,506
Thermal isostasy 87, 99 Transverse ridges 246,247,251,493 Vertical accretion deposits 221, 222
Thermal subsidence 96, 101 Trapridge Glacier, Canada 267 Very-long-baseline interferometry 49,
Thermocirques 306, 307 Trellis drainage pattern 406, 407, 408 51
Thermodynamics, application of Triaxial apparatus 164 Viking space programme 488, 491, 492,
concept to landform evolution 457 Tricart, J. 24 493,495
Thermodynamics of weathering Triple junctions 47, 94 Vine, F.J. 45
reactions 131, 132, 138, 139, 141, Triton 482, 504, 505 Victoria Falls, Zambia-Zimbabwe 425
142 Truk, Caroline Islands 452 Viscosity 196
Thermokarst 305,306,494 Tsunami 320 Viscous behaviour 522
Tholoids 115 Tuktoyaktuk Peninsula, Canada 306 VLBI (see very-long-baseline
Threshold 10 (see also extrinsic Turbulent flow 190, 196 interferometry)
threshold, geomorphic threshold, Turfan Depression, China 223 Volcanic activity (see volcanism)
intrinsic threshold) Turkwel delta, Kenya 414 Volcanic arcs 61, 62, 65-7, 109, 110
Threshold angle of stability 164, 183, Volcanic bombs 112
184 Ucayali Basin, Peru 387 Volcanic clusters 109, 110
Threshold of critical power 201 Ucayali (Rio), Peru 382 Volcanic eruptions 110--15
Throughflow 175, 191, 192 Ultrabasic rocks 108 Volcanicity (see volcanism)
Throughput 9 Unconformities 436 Volcanic lines 109, 110
Thrust faults 69, 524 Undercut solution runnels 149 Volcanic tuff 112
Tibesti Plateau, Chad 85,94,242 Underfit streams 361 (see also misfit Volcanism 61,62,63,66,87,88,89,
Tibetan Plateau, China 70, 71, 73-5, streams) 90,100,107,108-20,476,477
77,375,377,421,423,442,475 Underground drainage (see subsurface Volcanism, lunar 485,486
Tidal bores 321 drainage) Volcanism, Martian L!90, 491
 Index 537

Volcanism, on 10501, 502 Wave refraction 320 processes 236, 237, 247, 249, 252,
Volcanism, on Venus 499 Waves 315-20, 322, 323, 324,450 253
Volcanoes 108-11, 113-23 Waxing development of slopes 458, 461 Wind gap 410
von Lozinski, W. 293 Wayland, E.J. 462 Wizard Island, USA 122, 123
Voyager space programme 500,501, Weald, UK 409,470, 471 Wolman, M.G. 227
503,504 Weathering 129-57, 296-9, 323, 383, Wooldridge, S.W. 469-71
462.463, 475-7 Wrench faults 524
Wadati-Benioff zone 44,46, 61 Weathering on Mars 491, 492
Waimakariri River, New Zealand 212 Weathering front 136, 137, 141,402, Xingu Basin, Brazil 387
Waning development of slopes 458,461 463
Warm-based glaciers 263, 353, 354 Weathering-limited slopes 179, 181 Yakutia, USSR 307
Warm ice 263 Weathering mantle 136, 137, 143, 144, Yardangs 240, 241
Washboard moraine 278, 280 (see also 153,462, 463 Yield stress 521
De Geer moraine) Weathering profiles 136,137,138,141, Yosemite Valley, USA 145
Washburn, A.L. 299, 301 143, 144, 153 Youth (landscape evolution) 458
Wash load 200 Wegener,A.41,42 Y-shaped junctions 247, 248, 252, 253
Wash processes 175 Weird terrain, Mercury 498 Yugoslavia 148, 149
Water-balance equation 191 Western Australia 469 Yuty crater, Mars 489
Waterfalls 219 Western Cordillera, Andes 65
Water, in rock and soil 129, 130, 132, Western Ghats, India 427 Zagros Mountains, Iran 68, 69, 404,
143, 146, 147, 149, 164-7 Wetted perimeter 196, 197 420,421,422
Water-layer weathering 323, 327 Whalebacks 272,273 Zaire (Congo) Basin 85,95,387,392,
Water, physical properties 129, 146 Wheeo Creek, Australia 474 425,426
Watershed 207 (see also drainage Whole-mantle convection 53 Zaire (Congo) River 192, 213, 425
basin) Wilson cycle 102 Zambezi Basin, southern Africa 425
Water table 130, 230, 231 Wilson, I.G. 243, 245 Zambezi River, southern Africa 423
Wave base 318 Wilson, J.T. 45, 102 Zetaform beaches 316, 329
Wave-cut platforms 326 Wind characteristics and aeolian