Modern and Past Glacial Environments Revised Student Edition

Modern and Past Glacial Environments Revised Student Edition

To my wife, Teresa and daughters, Erica, Fiona and Rebecca and To the memory of Hilt Johnson

Modern and Past Glacial Environments A Revised Student Edition

Editor: John Menzies

OXFORD

AUCKLAND

BOSTON

JOHANNESBURG

MELBOURNE

NEW DELHI

Butterworth-Heinemann Linacre House, Jordan Hill, Oxford OX2 8DP 225 Wildwood Avenue, Woburn, MA 01801-2041 A division of Reed Educational and Professional Publishing Ltd A member of the Reed Elsevier plc group First published 2002 © John Menzies 2002 All rights reserved. No part of this publication may be reproduced in any material form (including photocopying or storing in any medium by electronic means and whether or not transiently or incidentally to some other use of this publication) without the written permission of the copyright holder except in accordance with the provisions of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London, England W1P 0LP. Applications for the copyright holder’s written permission to reproduce any part of this publication should be addressed to the publishers

British Library Cataloguing in Publication Data Modern and past glacial environments. – Rev. student ed. 1. Glaciology 2. Glaciers 3. Glacial epoch I.Menzies, John 551.3'1 Library of Congress Cataloguing in Publication Data A catalogue record for this book is available from the Library of Congress ISBN 0 7506 4226 2 For information on all Butterworth-Heinemann publications visit our website at www.bh.com

Composition by Genesis Typesetting, Rochester, Kent Printed and bound in Great Britain

CONTENTS

Preface Acknowledgements List of Contributors List of Symbols

Chapter 1 J. Menzies 1.1. 1.2. 1.3. 1.4. 1.5.

xi xiii xv xvii

Glacial environments – modern and past

Introduction Impact of ice masses on global habitats and earth systems Research in modern glacial environments Research in past glacial environments Research issues in glacial environments

1

1 2 4 5 10

Chapter 2 Global glacial chronologies and causes of glaciation P. E. Calkin (with contributions by G. M. Young)

15

2.1. 2.2. 2.3. 2.4. 2.5. 2.6. 2.7. 2.8. 2.9. 2.10.

15 16 24 26 32 34 39 44 46 50

Introduction Pre-Cenozoic glaciation Late Cenozoic glaciation and the classic subdivisions Oxygen isotope stratigraphy and the marine record of glaciation Long terrestrial records of climate and glacial fluctuations for the Quaternary The Late Pleistocene climatic record revealed by deep ice cores Correlation of Late Cenozoic glaciations in the northern and southern hemispheres The last 25 000 years and synchroneity of northern and southern hemisphere glacial cycles Causes of climatic change for Late Cenozoic glaciation Future climate change and glaciation v

vi

CONTENTS

Chapter 3 Glaciers and ice sheets J. Menzies (with contributions by T. J. Hughes)

53

3.1. 3.2. 3.3. 3.4. 3.5. 3.6. 3.7. 3.8. 3.9.

53 54 54 55 60 63 65 74 77

Introduction Ice mass types Formation of glacier ice Mass balance and glacier sensitivity Ice sheet modelling strategies Ice mass changes and fluxes Structure and thermal characteristics of ice masses Ice structures Concluding remarks

Chapter 4 J. Menzies 4.1. 4.2. 4.3. 4.4. 4.5. 4.6. 4.7. 4.8. 4.9. 4.10. 4.11. 4.12. 4.13. 4.14.

Ice flow and hydrology

Introduction Ice mechanics and the thermo-mechanical problem The flow of ice The flow of ice sheets The flow of valley glaciers The flow of ice shelves Velocities of ice sheets and glaciers Glacier surges Hydrology of glaciers The nature of meltwater flow and routing at the ice–bed interface Hydraulic systems and deformable beds Subglacial lakes High-magnitude meltwater discharges Concluding remarks

79

79 79 82 83 87 88 90 94 102 114 123 125 125 130

Chapter 5 Processes of glacial erosion N. R. Iverson

131

5.1. 5.2. 5.3. 5.4. 5.5. 5.6.

131 132 139 142 143 145

Introduction Abrasion Quarrying Rates of erosion Large-scale processes Summary

CONTENTS

vii

Chapter 6 Processes of glacial transportation M. P. Kirkbride

147

6.1. 6.2. 6.3. 6.4. 6.5. 6.6.

Introduction Ice properties affecting sediment transport Debris sources Transport pathways through ice masses Modification of sediment during transport Glaciological effects of debris in transport

147 148 149 152 164 167

Chapter 7 Processes of terrestrial glacial deposition C. A. Whiteman

171

7.1. 7.2. 7.3. 7.4. 7.5.

171 171 172 178 179

Introduction Factors contributing to terrestrial glaciogenic deposition Mechanisms of till deposition Mechanisms of glaciofluvial deposition Tills of Eastern England – debates about mechanisms of till deposition – a case example

Chapter 8 Subglacial environments J. Menzies and W. W. Shilts

183

8.1. 8.2. 8.3. 8.4. 8.5. 8.6.

Introduction Definition The subglacial interface – bed types and conditions Spatial variations in subglacial polythermal-rheological bed conditions Subglacial sediments and landforms Sediment structures and related sedimentological and geotechnical characteristics within subglacial sediments 8.7. Subglacial–proglacial transition environments 8.8. Subglacial landforms and bedforms 8.9. Subglacial erosional forms under active ice 8.10. Sediments and landforms of passive ice flow 8.11. Repetitive event histories in subglacial environments 8.12. Summary

183 183 185 190 193

Chapter 9 J. Maizels

279

9.1. 9.2. 9.3. 9.4. 9.5.

Sediments and landforms of modern proglacial terrestrial environments

Distinctiveness of proglacial environments Morphology and landforms of proglacial environments Proglacial meltwater channels systems Characteristics of proglacial outwash sediments Issues and future prospects

210 210 218 237 263 277 278

279 288 294 295 315

viii

CONTENTS

Chapter 10 Supraglacial and ice-marginal deposits and landforms W. H. Johnson (deceased) and J. Menzies

317

10.1. 10.2. 10.3. 10.4.

317 318 325 333

Introduction Sediment and sediment associations Landforms Summary

Chapter 11 Glaciolacustrine environments G. M. Ashley

335

11.1. 11.2. 11.3. 11.4. 11.5. 11.6. 11.7.

335 340 344 352 356 357 359

Introduction Limnology of glacier-fed lakes Ice-contact glacier-fed lakes Distal glacier-fed lakes Stratigraphy and landforms Recognition of past glaciolacustrine environments Conclusion

Chapter 12 Modern glaciomarine environments R. Powell and E. Domack

361

12.1. 12.2. 12.3. 12.4. 12.5. 12.6. 12.7. 12.8.

361 363 363 374 378 383 385 387

Introduction Subglacial processes Marine-ending termini Types of grounding-lines Proglacial and paraglacial environments Other non-glacial and modifying processes Sedimentation rates and fluxes Advance and retreat of marine-ending glaciers

Chapter 13 Past glaciomarine environments A. Elverhøi and R. Henrich

391

13.1. 13.2. 13.3. 13.4.

391 391 400

Introduction Past glaciomarine sediments: classification and identification Examples of Pre-Cenozoic glaciomarine sequences Glaciomarine sediments and the sedimentary environment of a passive margin and its adjacent ocean. The Norwegian-Greenland Sea and the Norwegian/Barents Sea margin – a case example 13.5. Conclusion

402 414

CONTENTS

ix

Chapter 14 Processes of glaciotectonism F. M. Van der Wateren

417

14.1. 14.2. 14.3. 14.4. 14.5. 14.6. 14.7.

417 418 421 423 431 439 441

Introduction Soft sediment deformation Glaciotectonic regimes Subglacial shear zones Marginal compressive belts Passive deformations Conclusions

Chapter 15 Glacial stratigraphy J. Rose and J. Menzies

445

15.1. 15.2. 15.3. 15.4. 15.5. 15.6.

445 446 449 453 455 471

Introduction Rationale Stratigraphy within glacial environments Stratigraphic nomenclature Glacial stratigraphic procedures and methods Conclusion

Chapter 16 J. Menzies 16.1. 16.2. 16.3. 16.4.

Problems and perspectives

Introduction Paradigm shifts Questions illustrative of problems Epilogue

475

475 476 476 481

References

483

Subject Index

529

This Page Intentionally Left Blank

PREFACE

possible without the authors whose chapters appear in this new revised edition, namely Gail Ashley, Parker Calkin, Gene Domack, Anders Elverhøi, R¨udiger Henrich, Terry Hughes, Neal Iverson, Martin Kirkbride, Judith Maizels, Ross Powell, Jim Rose, Bill Shilts, Grant Young, Dick Van der Wateren and Colin Whiteman; to them I give a special thanks. It is with great regret that in the course of producing this revised edition I must record the loss of Hilt Johnson. This book is dedicated in part to his memory. I must thank the many individuals at Brock University who have aided in the production of this book, in particular my many students over the past several years, Loris Gasparotto and Mike Lozon for their special cartographic skills, to Gail Elliot, Virginia Wagg and Diane Gadoury for word processing and scanning skills, Shara Lee Foster for painstakingly reading through revised chapters and to Rebecca Menzies who, single-handedly, took on the onerous task of revising all the references. I would like to thank the staff at Butterworth-Heinemann for their support and efforts in making this book possible, especially Sue Thomas (freelance editor) and Sue Hamilton. Finally, I wish to thank my family, my wife, Teresa and daughters Erica, Fiona and Rebecca for their constant support and patience in this venture.

The purpose of this Revised Student Edition of the two-volume Glacial Environments (Modern, Menzies, 1995 and Past, Menzies, 1996) is to provide an affordable and accessible text principally for undergraduate students of Geology, Earth Sciences, Physical Geography, Environmental Sciences and related disciplines. Just as a small-scale map, compared with a large-scale intrinsically detailed map, must sacrifice some detail in order to provide direction and a broader ‘picture’, so this revised edition hopes to supply an entry and ‘pathway’ to an understanding of glacial environments. In producing this edited version emphasis has been placed on providing key elements of interest to students entering into the study of glacial environments. As in any such endeavour that distills a larger text it is incredibly difficult to decide on what stays and what must be left out of such a text. As noted above, the pivotal editorial arguments were in providing accessibility to the fundamental material and maintaining affordability. To that end certain chapters had to be modified, shortened and, sadly, as needs must, removed from this edition. It is testament to all the authors involved in the development of the earlier two-volume texts that a remarkably limited amount of editing was required other than up-dating and some amendments based on more recent discoveries and findings. I sincerely wish to thank Jim Rose and Jaap van der Meer for their unstinting sound advice and encouragement in producing this revised student volume. Of course, such a ‘third’ version would not have been

John Menzies Brock University Ontario Canada xi

This Page Intentionally Left Blank

ACKNOWLEDGEMENTS

3.9.(a,b) and 4.2.; J. Shaw for Plate 4.8.; H. Bjornsson for Plate 4.9.; D. E. Sugden for Plate 6.1.; L. Homer for Plate 6.2.; J. J. M. van der Meer for Plates 12.1. and 12.4.; W. Mitchell for Plate 9.1.; A. Russell for Plate 9.7.; T. Gustavson for Plate 11.1.(c); J. Hartshorn for Plate 11.2.(a); R. Aario for Plates 8.8., 8.9.(b), 8.27.(a,b) and 8.28.(a); B. C. Ferries for Plate 8.13.(a); P. Birkeland for Plate 10.4.; L. Clayton for Plate 10.6.; Geological Survey of Canda for Plate 8.25.(b); Geological Survey of Finland for 8.21.(a); P. L. Gibbard for Plate 8.3.(b); Greenland Geoscience (Meddelesler om Groenland) for Plate 3.3.(b); G. Hamelin for Plates 8.16.(a,b), 8.18.(a,b), 8.20.(d) and 8.22.(c); S. Hicock for Plate 8.7.; R. Lagerb¨ack for Plates 8.27.(c) and 8.28.(b); D. W. Moore for Plate 10.1.; National Air Photo Library of Canada for Plates 8.12. and 8.25.(a); R. Powell for Plate 3.9.; M. Rappol et al. for Plate 9.5.(c); J. C. Ridge for Plate 12.1.; J. Shaw for Plate 8.20.(a,b); Society of Economic Paleontologists and Mineralogists for Plate 3.7.; D. Stetz for Plate 8.9.(a); W. H. Wolfestan for Plate 8.13.(b,c); C. Zadanowicz for Plate 8.26.(a,b).

The editor, authors and publisher wish to thank the following for their kind permission to reproduce copyright material. Full citation can be found by consulting figure and plate captions and references. The authors wish to thank the following individuals and publishing companies for their kind permission to use figures and plates. Figures W. Ruddiman and H. E. Wright for Figs 2.6.(d) and 2.17.(a); J. Imbrie and K. P. Imbrie for Fig. 2.17.(b); R. Powell for Fig. 3.4.; C. Raymond for Fig. 4.5.; G. S. Boulton for Fig. 3.12.; B. Hallet for Fig. 5.4.; J. Harbor for Fig. 5.8.; N. Eyles for Fig. 6.2.; D. Lawson for Fig. 6.4.; R. J. Small for Figs 6.7., 6.8.; T. Gustavson and J. Boothroyd for Fig. 11.6.; J. Teller for Fig. 11.12.; R. Powell for Fig. 12.11.; and E. Domack and C. R. Williams for Fig. 12.12.; R. LeB Hooke for Fig. 8.9.; C. Kaszycki for Fig. 10.8.; C. H¨allestrand for Fig. 8.17.; R. R. Parizek for Fig. 10.11.; D. R. Sharpe and J. Shaw for Fig. 8.41.(a,b), 8.42.(a,b); J. Shaw for Figs 8.39., 8.40.; G. M. Young for Fig. 1.1.(b).

Publishers American Meteorological Society for Fig. 2.16.; American Association of Petroleum Geologists for Fig. 15.6.; American Association for the Advancement of Science for Fig. 2.10.(b); American Geophysical Union for Figs 2.6.(a,b,c), 2.18., 4.10.,

Plates C. C. Langway for Plate 2.1.; G. Clarke for Plate 3.1.; E. Evenson for Plates 3.2., 4.1.(a,b), 4.3., 4.5. and 4.6.(a,b); G. Holdsworth for Plates 3.4., 3.7., 3.8., xiii

xiv

ACKNOWLEDGEMENTS

4.12.(a,b), 4.13., 4.25., 8.12. and 12.3.; A. A. Balkema Publishers for Figs 4.20., 6.2., 7.2., 7.3., 8.9., 8.48., 14.9. and 14.15.; British Museum Press for Fig. 15.2.; C. R. C. Press for Fig. 2.2.; Cambridge University Press for Figs 3.12., 13.8. and 15.14.; Canadian Society of Petroleum Geologists for Figs 9.19. and 9.20.; Elsevier Science Publishers BV for Figs 2.9.(a,b,c), 4.6., 4.26., 8.2.(a), 8.3., 8.4., 8.6., 8.16., 8.18., 8.20., 8.28., 8.36., 12.4., 12.6., 13.2., 13.11., 13.12. and 13.16.; Elsevier Science Ltd for Figs 2.7.(b), 2.13.(a,b), 2.14., 2.15., 3.8., 3.11., 5.1., 5.2., 7.5., 7.8., 13.7. and 15.13.; Edward Arnold Publishers for Fig. 6.1.; Fennia (Geographical Society of Finland) for Fig. 8.22.; Geological Magazine for Fig. 8.10.; Geological Society of America for Figs 10.9. and 15.3.; Geological Society of Denmark for Figs 15.8., 15.9. and 15.10.; Geological Society of London for Figs 8.11. and 12.8.; Geological Survey of Canada for Figs 8.26. and 9.7.; Geological Survey of Norway for Fig. 8.32.; IAHS for Fig. 3.5.; Icelandic Glaciological Society for Figs 4.27. and 9.8.; International Glaciological Society for Figs 3.3.(b), 3.14., 4.8., 4.18., 4.19., 4.21.(a,b), 4.28., 5.7., 6.7.(a,b), 6.12.(a,b,c), 8.23., 8.24., 8.25., 8.30. and 8.38.; International Association of Sedimentologists for Figs 6.3.(b), 6.11.(b), 10.1. and 11.9.;

Kluwer Academic Publishers for Figs 2.3., 2.4., 2.7.(a), 2.10.(b), 2.19.(b), 4.1., 9.22., 13.13., 13.14., 13.15., 13.19. and 13.20.; Laval Universit´e Presse for Figs 3.1. and 4.2.; Lingua Terrae for Fig. 14.13.; Methuen Publishers for Fig. 8.29.; National Academy of Sciences for Fig. 2.1.; National Research Council of Canada for Figs. 8.39., 8.40. and 8.41.; Nature for Figs 2.18.(a,b,c), 2.10.(a), 2.11., 2.12., 4.9., 8.37. and 15.1.; Norsk Polarinstitutt for Fig. 13.10.; Oxford University Press for Fig. 4.11.; Precambrian Research for Fig. 13.8.; Royal Scottish Geographical Society for Fig. 8.43.(c); Royal Society of Edinburgh for Figs 9.16. and 15.11.; Scandinavian University Press for Figs 4.16., 7.4., 7.6., 8.34., 9.11., 9.21., 10.6., 10.10., 14.11. and 11.7.(b); Science Progress for Fig. 15.12.; Scottish Academic Press for Fig. 15.5.; Societas Upsaliensis pro Geologica Quaternia for Fig. 8.27.; Society of Economic Paleontologists and Mineralogists for Figs 9.14. and 11.6.(b); Springer-Verlag for Fig. 8.31.; State University of New York (SUNY) Binghampton for Fig. 10.10.; University of Colorado and Arctic and Alpine Research for Fig. 3.10.; University of Oulu for Figs 8.13. and 8.47.; and John Wiley and Sons for Figs 4.15., 4.17., 4.22., 4.23., 4.24., 8.5., 8.14., 9.2., 9.4., 9.9., 9.17. and 9.18.

CONTRIBUTORS

Gail M. Ashley Department of Geological Sciences, Rutgers University, New Brunswick NJ, USA

Judith Maizels 88 Morningside Drive, Edinburgh; formerly Department of Geography, University of Aberdeen, Aberdeen, UK

Parker E. Calkin formerly of Department of Geological Sciences, SUNY at Buffalo, Amherst NY, USA

John Menzies Departments of Earth Sciences and Geography, Brock University, St. Catharines, Ont., Canada

Gene W. Domack Department of Geology, Hamilton College, Clinton NY, USA

Ross Powell Department of Geology and Environmental Geosciences, Northern Illinois University, DeKalb IL, USA

Anders Elverhøi Department of Geology, University of Oslo, Oslo; fomerly Norsk Polarinstuut, Oslo, Norway

Jim Rose Department of Geography, Royal Holloway, University of London, Egham, UK

Rudiger ¨ Henrich Department of Geosciences, University of Bremen, Bremen; formerly GeoMar – Research Center for Marine Geosciences, Kiel, Germany

Bill W. Shilts Chief, Illinois State Geological Survey, Champaign IL; formerly Terrain Sciences Division, Geological Survey of Canada, Ottawa, Ont., Canada

Terry J. Hughes Institute of Quaternary Studies, University of Maine, Orono ME, USA

Dick M. van der Wateren Faculty of Earth Sciences, Vrije Universiteit, Amsterdam, The Netherlands

Neal R. Iverson Department of Geological and Atmospheric Sciences, Iowa State University, Ames IA, USA W. Hilt Johnson (deceased) Department of Geology, University of Illinois, Urbana-Champaign IL, USA

Colin A. Whiteman Earth and Environmental Science Research Unit, University of Brighton, Brighton, UK

Martin P. Kirkbride Department of Geography, University of Dundee, Dundee, UK

Grant M. Young Department of Earth Sciences, University of Western Ontario, London, Ont., Canada xv

This Page Intentionally Left Blank

SYMBOLS

a

Ablation

c˙

Accumulation rate

␧

Ablation rate

cr

Constant related to regelation process

ag

Geometric factor in clast ‘ploughing’

cs

Summer accumulation

ap

Sinusoidal bed amplitude

ct

Total accumulation

as

Summer ablation

cw

Winter accumulation

asp

Bed asperity roughness

C

Cohesion

asr

Measure of surface roughness

Cr

Areal concentration of clasts in contact with the bed

at

Total ablation

C3

aw

Winter ablation

A

Ice hardness coefficient

Empirical factor related to the mode of water flow in a conduit, the conduit geometry and ice flow conditions

b

Mass balance

d

Obstacle size

bL

Balance rate at the ice terminus

dc

Controlling obstacle size

bn

Net balance

do

Mean particle size diameter

bo

Stream channel width

dw

Depth of water film at the ice/bed interface

bs

Summer balance

D

Constant related to ice and water density

bw

Winter balance

Dc

Conduit diameter

B

Ice sliding coefficient

De

Longitudinal strain or extension

*

Be

Energy balance on ice mass surface

Bf

Buoyancy flux

Bp

Back pressure within an ice shelf

c

Accumulation

E

Activation energy of the creep of ice

f

Area of ice/bed interface occupied by meltwater

fsh

Shape factor affecting frictional drag in valley glaciers and ice streams

xvii

xviii

SYMBOLS

ft

Fraction of the bed that is thawed

ktc

Thermal conductivity

F

Froude number

kw

Wave number = 2␲/␭

Fc

Effective contact force between an indenting clast and the bed

K

Hydraulic conductivity

K*

Material constant

Fw

Longitudinal normal force within a sediment wedge

Kb

Constant related to deforming sediment

Fwd

Channel width/depth ratio

Kf

g

Acceleration due to gravity

Constant related to ice shelf ice fabric, temperature and ratio of components of strain rate

G

Hydraulic gradient

Kp

Apparent conductivity of a particle array in relation to ice

Geo

Geometric factor relating mean basal normal stress to basal shear stress and that fraction of the ice/bed interface covered in meltwater

l

Glacier length

L

Distance from ice divide to ice margin

h

Ice elevation above grounding line of floating ice margin

Lc

Length of main channel

Lh

Latent heat of fusion of ice

hbc

Height of cavity backwall

Lic

Length of ice cliff face

hd

Thickness of debris-rich ice layer

Lm

Latent heat of melting

hi

Ice thickness

Lv

Length of valley axis

hR

Present-day elevation of bedrock above or below the margin of the ice sheet

m

Constant of ice sliding related to ice creep

m ˙

Rate of melting of conduit walls

hs

Sediment thickness in a wedge

m ˙b

Rate of basal ice melting

hsw

Height of submerged ice-cliff wall

m*

hw

Water depth

H

Maximum surface elevation of an ice mass

Constant related to spectral density of the bed in terms of mean quadratic slope in the direction of the ice–bed slip

Ho

Channel depth at stream mouth

m ˙c

Rate of conduit closure by volume

i

Hydraulic gradient

M

Annual mass balance

ic

Critical hydraulic gradient

M1

Melting rate of ice cliffs

I*b

Ploughing index on a clast in a deforming sediment

M2

Mean melting rate of ice cliffs

Mf

Momentum flux of a meltwater jet

Js

Approximation of sediment flux along a channel

n

Visco-plastic deformation

k

Permeability of sediment

nK

km

Reciprocal parameter

Number of cavities/passageways across the width of a glacier

np

Porosity of sediment

P

Solid precipitation

ko

of

the

Critical wave number

Manning

roughness

component

for

ice

creep

SYMBOLS

xix

Pc

Cavity roof perimeter

T¥

Far-field ambient water temperature

Pcs

Channel sinuosity

TCL

Total length of all braid channels

Pcs(tot) Total channel sinuosity

u

Spreading ice velocity

Pc

Cavity roof perimeter

u˙

Average ice velocity in ice column

Po

Normal stress distribution across a glacier bed

u+

Sliding velocity of a surging ice mass

ub

Basal sliding velocity of ice

Q

Volumetric flux (discharge) of meltwater

uby

Vertical ice flow vector

ubz

Transverse ice flow vector

Qmax Peak discharge from a glacier during a j¨okulhlaup Qo

Stream discharge

uc

Velocity of a clast

Qs

Suspended load carried by stream

ucs

Creep velocity of sediment

Qx

Water influx along a section of channel x metres in length

u˙ c

Effective rate of cavity closure

ui

Overall ice mass velocity where ui = us + ub

r˙

Rate of closure of a conduit

uL

Terminus ice velocity

r

Radius of a conduit

um

rc

Radius of curvature of a clast

Sliding velocity of mobile sediment at its upper interface

rp

Particle radius

un

Component of ice velocity normal to the bed

rsr

Isostatic ratio variable

uo

R

Universal Gas constant

Component of un resulting from melting of ice caused by geothermal heat and sliding friction

RI

The Cailleaux-Tricart roundness index

us

Surface velocity of ice

Rn

Richardson number

ut

Ru

Bed roughness parameter

Component of ice velocity parallel to the bed interface

s

Areal fraction of the bed cavity/passageway uncoupled

uw

Dynamic viscosity of water

Um

sr

Bed roughness slope

Mean centreline velocity of stream at conduit mouth

S

Cross-sectional area of channel

Umo

Initial near-centreline stream velocity

Sg

Stream gradient

Uw

Water velocity

Sj

Area of clast imposed upon by ploughing pressure

vp

Radius of a particle

v¯ w

th

Time in years for an ice mass to equilibrate

Average water velocity within a meltwater film at the ice/bed interface

Tm

Response time of an ice mass to changes in equilibrium status

Vm

Mean flow velocity of sediment in a channel

Vmax

Ts

Surface water temperature

Total maximum volume of meltwater drained from an ice mass owing to a j␾kulhlaup

xx

SYMBOLS

Vw

Specific discharge along a channel where inflow and outflow are in equilibrium

w

Width of ice column

W

Width of deforming sediment layer, transverse to maximum ice flow

Wb

Width of channel bed

Ws

Stream power

z

Some vertical thickness of ice

zcs

Depth within a deforming sediment

zs(crit) Depth within a deforming sediment at which point basal shear stress = ␭s Zo

Constant related to sediment flux along a channel

␣

Ice surface slope

␣b

Slope at interface

␣c

Slope of a conduit

␣hc

Constant dependent on hardness of clasts, the bed and the geometry of the striator point

the

basal

ice/sediment

(bed)

⌬P

Difference in pressure in ice adjacent to a conduit and the water pressure within a conduit

⌬T

Temperature difference between ice and water

⌬␴p

Stress fluctuation

␦

Parameter in Fowler’s surging theory

␦f

Loss of water pressure in a conduit owing to friction

␦h

Incremental steps of ice thickness

␦p

Change in local water pressure in a conduit

␦s

Length of a straight conduit

␦u

Incremental lateral spreading of ice owing to internal deformation

␦z

Incremental vertical motion of ice owing to internal deformation

␥i

Incremental longitudinal strain

E˙ d

Strain rate for dilatant deforming sediment

E˙ sh

Strain rate within an ice shelf

␣i

Incremental ice surface slope

␩

Ice viscosity

␤

Slope of glacier bed

␩s

Newtonian viscosity

⌫

Constant related to ice regelation processes

␩sed

Sediment viscosity

⌫pg

Pressure gradient of meltwater in a channel in the direction of the mean slope

⌽

Angle of internal friction of sediment

⌽s

Slope of conduit

␥

Shear strain

␬

Thermal diffusivity of ice

␥f

Finite longitudinal strain

␬b

Bulk density of deforming sediment

␥g

Density of sediment

␬s

Fluid density

␥s

Shear strain rate of sediment in plane of shearing

␭

Bed wavelength

␥w

Specific weight of water

␭f

‘Frictional parameter’

⌬

Dilation within a sediment

␭rs

Removal particles

⌬h

Incremental ice thickness steps

⌬x

␭t

Incremental length steps from ice divide to margin

Transition wavelength at ice/bed interface where regelation and ice deformation are equally efficient

rate

constant

for

suspended

SYMBOLS

xxi

␭wr

Characteristic bed wavelength – ‘white roughness’

␴cj

Local effective stress upon a clast of jth size class

␭1

Principal direction of finite extension of sediment under shear

␴␧

Effective stress

␭3

␴i

Normal stress at a point within an ice mass

Shortest direction of finite extension of sediment under shear

␴zs

Effective stress within deforming sediment

␭R

Bed roughness term

␶

Shear stress

␮

Coefficient of rock friction

␶0

Shear stress of ice along flow line

␮cs

Coefficient of friction of sediment

␶b

Basal shear stress of ice

⌶

Melting-stability parameter

␶c

Yield strength of sediment

␰

Balance velocity

␶D

Shear stress for a dry bed

␰j

Controlling obstacle factor

␶g

Gravitational shear stress in sediment wedge

␳a

Average compressive stress

␶j

␳ad

Ambient fluid density

Basal shear stress applied to a clast of jth size class

␳i

Density of ice

␶s

Shear strength of deforming sediment

␳i

Average density of ice

␶w

Shear stress for a wet bed

␳o

Initial discharge density

␶xy

Internal ice shear stress

␳R

Rock density

␶xz

Applied shear stress in z-plane

␳s

Density of sediment

⌿

Drag coefficient applied to a clast at the ice/ bed interface

␳sw

Density of sea water

␳w

⌿c

Porewater pressure

Rate of channel closure resulting from ice creep

␳wt

Density of water

⌿e

Rate of erosion in evacuating a till channel

␴

Standard deviation of the Gaussian distribution for stream velocity across plume width

⍀

Conduit shape factor

␴b

⍀w

Ice overburden normal pressure at the base of an ice mass

Constant between 0 and 1 that depends on ␭ and the transition wavelength (␭t)

w ˙

␴b

Mean stress level at ice bed

Angle made by equipotential lines with the ice surface

This Page Intentionally Left Blank

1

GLACIAL ENVIRONMENTS – MODERN AND PAST J. Menzies Testimony to these glacial periods can be found in the geological stratigraphic record stretching from the Precambrian to the Quaternary (Deynoux et al., 1994) (Fig. 1.1a,b). Evidence, for example in the form of roche mouton´ees resulting from subglacial streamlining of bedrock, can be recognized from the Ordovician in Mauritania in Saharan Africa; glacial striations and floating ice mass scourings in the Sturtian of the Precambrian in South Australia; or dropstones and fissure patterns in the diamictites of the Precambrian Gowganda in Ontario. In mountainous areas of the world, today, extensive glacial sediments and associated landforms can be found, proof of past colder climatic conditions during the Pleistocene. More recent evidence of the cold climate periods in medieval times, known as the Little Ice Age, are well documented. For example, in Europe, Swiss parish priests, Icelandic saga writers and Flemish painters all recorded the very cold winters and resultant famines of the period (Grove, 1988). Glaciers and ice sheets have today, as in the past, had an enormous influence, both beneficial and detrimental, upon all aspects of earth systems. At present, the complexities of ice dynamics and the relevant variables and influences they have on ice mass balance, and on the specific reasons for ice front and meltwater discharge fluctuations, remain imprecisely understood. The processes of ice basal movement and the intricate

1.1. INTRODUCTION Glaciers and ice sheets are found in remote and wild terrains far from centres of population. Yet this physical and mental distance should not betray their fundamental significance and importance to all our lives. Glacial environments hold an essential key to our knowledge of the present, as well as past and future global environmental conditions. The Huttonian principle of understanding the present to permit a grasp of the past (and of the future) is never better exemplified than in our attempts to comprehend and explain contemporary dynamic processes in glacial environments. At first sight glacial environments are chaotic, complex and geologically ephemeral, yet beneath this apparent chaos exists a process-driven pattern. Perhaps no other sedimentological environment exhibits such rapid, dynamic and spatially variable changes of processes that, when viewed at the small scale appear chaotic, yet at other larger scales begin to reveal a precision in the spatial and temporal organization and execution of individual processes. Throughout Earth’s history, the interaction of oceans, continents and the atmosphere in response to solar forcing has resulted in repeated global glaciations. The underlying theme of this book, therefore, is to bring these environmental patterns into better focus. 1

GLACIAL ENVIRONMENTS – MODERN AND PAST

Relative Changes of Sea Level Rising

Falling

0

100

le

a Oce

ru nc

st

r acc

e

n tio

CO 2

Relative Changes of Sea Level Rising

QUATERNARY

Present Sea Level

Falling

0

CRETACEOUS

100

JURASSIC 200

TRIASSIC

200

PERMIAN 300

PENNSYLVANIAN

300

MISSISSIPPIAN DEVONIAN 400

I

TERTIARY

SILURIAN

400

GEOLOGIC TIME IN MILLIONS OF YEARS

GEOLOGIC TIME IN MILLIONS OF YEARS

S

ea

l ve

PERIODS

2ND ORDER CYCLES (SUPERCYCLES)

CLIMATE STATES

1ST ORDER CYCLES

GLACIATIONS

2

G

I

G

ORDOVICIAN 500

500

CAMBRIAN

See FIG. 1.1. (b)

I

PRECAMBRIAN

FIG. 1.1 (a) The Earth’s glacial record from the Cambrian to the Quaternary including sea level variations (after Vail et al., 1977), atmospheric CO2 (after Berner, 1990), rate of ocean crust accretion (after Gaffin, 1987) and climate ‘states’ (G – Greenhouse; I – Icehouse) (after Fischer, 1984). (Diagram modified from Deynoux et al., 1994; reprinted with permission from Cambridge University Press).

relationship between ice motion and subglacial hydrology continue to be poorly understood, as do thermal and sedimentological states within subglacial environments. Modern glaciers can be viewed as active analogues of past glaciers and ice sheets. Modelling of modern ice masses and mass balance studies advance our knowledge in explaining past global ice sheet development and expansion. The diverse subenvironments of modern glaciers, both on land and subaquatically, provide an active field laboratory for studying present glacial sedimentological processes that can be employed to understand and interpret

glacial sediments of the geological past. However, in utilizing modern ice masses as analogues of the past, care must be exercised, since past ice masses may have been significantly different in many critical aspects. 1.2. IMPACT OF ICE MASSES ON GLOBAL HABITATS AND EARTH SYSTEMS Rather than view global glaciation as a series of environmental aberrations, it is now apparent that the Earth is essentially a Glacial Planet punctuated by periods of ameliorative conditions similar to or

Relative Changes of Sea Level Rising

Falling

GLACIATIONS

GLACIAL ENVIRONMENTS – MODERN AND PAST

3

PERIODS CAMBRIAN

ea

-

l

CO 2

2000

3000

P R E C A M B R I A N

?

S

?

e lev

1000 MESOPROTEROZOIC

PALEOPROTEROZOIC

2000

3000

ARCHEAN

GEOLOGIC TIME IN MILLIONS OF YEARS

GEOLOGIC TIME IN MILLIONS OF YEARS

1000

PROTEROZOIC

NEOPROTEROZOIC

4000

4000

HADEAN

FIG. 1.1. (b) Record of Precambrian glacial record (diagram and interpretation by Young, 1994).

occasionally warmer than the period we live in today. Almost all aspects of life on Earth are influenced to a greater or lesser extent by the impact and persisting effects of glaciation. The distribution of plants, animals, early humans, soil types and coastal morphology are a few examples of the direct influence of global glaciation. Even in the tropics, where climatic conditions have remained almost unchanged for at least the past 15 million years, the northern and

southern boundaries on land and the repeated changes in ocean sea level have resulted in climatic and biogeographic changes all as a response, however imperceptible and subtle, to global glaciation. The effect of ice sheets and glaciers, particularly on global habitats and earth systems, can be viewed at two levels of impact: first, their influence upon humans and habitats within their immediate locality and second, on their much more pervasive influence

4

GLACIAL ENVIRONMENTS – MODERN AND PAST

on all global habitats owing to the effect of modern ice masses on global climate and sea level. The effect of ice masses in the immediate proximity to humans is well documented (Tufnell, 1984; Grove, 1988; Bell and Walker, 1992; Hambrey and Alean, 1992) in terms, for example, of meltwater outbursts and rapid ice advances resulting in the loss of pasture lands, property and, in some cases, human fatality. These detrimental aspects, of course, need to be balanced with the beneficial resources of water for hydroelectric projects, irrigation and fresh domestic water supplies. Less obvious, but actively researched today, is the more insidious and pervasive influence of present day ice masses on global climate and oceanic currents. As predictions of global warming increase, so knowledge of modern glacial conditions needs to be amplified if we are to cope with and predict sea level rise in the coming century when this knowledge will become acutely significant. Considering the enormity of this influence, the relevance today of past glaciations and their associated sediments and forms cannot be undervalued. It is more than likely that the Earth will experience further global glaciations. It is also possible that human activities, especially over the past 200 years, have exacerbated and possibly accelerated some of the complex oceanic/atmospheric and solar forcing interrelationships but to an extent that remains unknown. To be able to predict and be prepared for future global change, a profound knowledge of past glacial environments must be gleaned from the vast record that past glaciations have left behind. Related to potential sea level rise resulting from ice sheet melting in Greenland and Antarctica is the question of future ice sheet stability. If these ice sheets melt at an increased rate vast plumes of cold fresh water may be injected into the polar oceans affecting future oceanic habitats, currents, surface ocean water temperatures (e.g., El Ni˜no–Southern Oscillation events) and, consequently, global weather patterns (Sarmiento, 1993). However, whether ice sheet accelerated melting may or may not be occurring because of global warming remains the subject of considerable controversy (Kasting, 1993). A further aspect of glacial environments is in providing analogues to past glacial conditions as a key to understanding glacial sedimentation processes.

By studying present day glacial environments and sedimentological processes, considerable knowledge can be gleaned as to how sediments of past glacial events have been derived, transported and finally deposited both on land and in water. Pleistocene glacial sediments cover today at least 30 per cent of the Earth’s continental landmasses, and an even greater area must be included when Pre-Pleistocene sediments ranging over vast areas of India, Australia, Africa and South America are considered (Hambrey and Harland, 1981; Eyles, 1993). These sediments affect almost every aspect of human life from establishing foundations and footings for buildings, roads and runways; the nutrient content of soils; the nature of groundwater supplies; to the potential soil-routes for contaminant waste disposal and the location of landfill sites. These few examples illustrate the vital need to understand glacial processes ongoing in glacial environments (De Mulder and Hageman, 1989). Until relatively recently this figure of 30 per cent for Pleistocene sediments was accepted yet, today, perhaps approximately 60 per cent would be a more accurate figure if glaciomarine sediments are included. These thick sediments lie over the ocean floors covering enormous parts of the northern and southern areas of the Atlantic and Pacific oceans. The impact of glaciomarine sediments on landbased habitats is certainly limited, affecting only fisheries to a little-known degree, but if future utilization of oceanic basins occurs the influence of these vast areas of glacial sediments may become increasingly meaningful. 1.3. RESEARCH IN MODERN GLACIAL ENVIRONMENTS Research in a wide range of modern glacial environments has only become significant since the 1950s. Prior to then, work in often remote areas was restricted to occasional explorers and adventurers and to sporadic scientific studies. In Europe, especially in the Alps, inaccessibility was not a problem but scientific interest was fostered by only a few individuals. In the nineteenth century, the work of the Swiss from the time of Charpentier and Agassiz is well known, as are the travels and observations of James Forbes (Cunningham, 1990). Likewise, considerable work on both modern and past glacial environments

GLACIAL ENVIRONMENTS – MODERN AND PAST

began in Germany and Scandinavia. The tragic contretemps between Forbes and Agassiz, concerning a misunderstanding as to who had first publishing rights on certain scientific observations, may have increased what was an already growing interest in glaciers and ice mechanics in the scientific community by the mid-nineteenth century. By the early 1900s the observations of Nansen in Greenland, of Amudsen and Sverdrup in other parts of the Arctic, and many others, increased the scientific interest in modern glaciers and ice sheets; and without doubt the expeditions of Shackelton and ‘the doomed trek’ of Scott’s desperate expedition in Antarctica only heightened the curiosity of the general public to these environments. Other equally famous expeditions occurred in the Arctic to Nova Zemlya, Spitsbergen and the Canadian Arctic islands. In continental North America the famous expeditions, for example, of Harriman in the Canadian Rockies and the area of Glacier Bay in Alaska, brought forth the beginnings of scientific writings on modern glacier environments. Following this early period, a phase of scientific development began that is still very much in existence today. Many influences beyond natural curiosity have engendered and encouraged this research work. In particular, the conflicts of the Second World War and the potential invasion of Alaska and continental North America by Japanese forces brought a renewed focus on polar areas and research problems. The Cold War that followed spurred on research in Alaska and the Canadian and Soviet north to a degree never before (or since) witnessed. Perhaps in the interests of petty nationalism and territorial rights, Antarctica by the early 1950s was sectorized into areas where different nations strove to establish research bases, under United Nations auspices, to further scientific research. Whatever the motives may or may not have been, the establishment of so many research bases in Antarctica has led to a vast outpouring of scientific literature and an expanding appreciation of modern glacial environments. 1.4. RESEARCH IN PAST GLACIAL ENVIRONMENTS Research into past glacial environments has reached a new intensity over the past two decades. The reasons

5

for this acceleration are a growing awareness of the relevance of past glacial events and environments to modern habitats and life in those areas of the Earth glaciated in the past. 1.4.1. Recognition of ‘ice ages’ From medieval times there have been innumerable explanations of features that we now recognize as glacial, such as erratic boulders viewed as the putting stones of giants, potholes as devils’ punch bowls and other demonic interpretations for glacial phenomena. Where and when precisely a glacial explanation of many of these features was first enunciated is difficult to determine, but by the mid-eighteenth century in Scandinavia, Germany, Iceland and Switzerland, several individuals had begun to suggest that glaciers had been more extensive in the past (Nilsson, 1983). In the nineteenth century, Charpentier, Agassiz, Buckland and Esmark, for example, had begun to realize that large areas of Europe had been glaciated by vast ice sheets and thus the concept of the ‘ice age’ was proposed. This suggestion was not established, however, in many parts of Europe and North America until the late 1800s, and even up to the 1920s there were individuals who still questioned the very idea of an ice age (Table 1.1). As early as 1863, Archibald Geikie interpreted the unlithified sediments and landforms of Scotland as evidence of glaciation and, from that beginning in Britain, a rapid period of geological mapping and stratigraphic interpretation spread throughout most of the northern hemisphere. By the mid-1930s knowledge of the extent and details of multiple glaciations to have affected Europe and North America was well established (Flint, 1947, 1971; Charlesworth, 1957). 1.4.2. Development of Glacial Chronologies Penck and Br¨uckner (1909), in the European Alps, had established a four-fold sequence of major glaciations based upon interpretations of the extent and distribution of outwash fans and terraces in the northward-trending Bavarian, Alpine-foreland, river valleys of the G¨unz, Mindel, Riss and W¨urm (oldest to youngest glaciation; Chapter 2). The younger three glacials of the four-fold sequence of Alpine

6

GLACIAL ENVIRONMENTS – MODERN AND PAST

TABLE 1.1. Developments in glacial studies up to 1986 10th Century AD

Recognition in Icelandic sagas of the power and impact of glaciers upon the landscape

1740

D. Tilas

Finnish prospector recognized the concept of drift prospecting tracing ore bodies from dispersal trains

1745

James Hutton

Identified erratics in the Jura Mountains as being transported by glaciers

1802

John Playfair

Supports Hutton’s concepts and suggests glaciation had occurred in Scotland

1815

J. P. Perraudin

Suggests, from observations in Val de Bagnes, Switzerland, greater extension of glaciers

1821

I. Venetz

A Swiss highway engineer agrees with Perraudin and elaborates these ideas scientifically to the Society of Natural History, Luzern

1824

J. Esmark

Recognition of former extent of glaciation in Norway

1829

I. Venetz

Argues that most of Europe was glaciated

1830

C. Lyell

Puts forward his ‘drift theory’

1832

Berhardi

Recognizes former continental glaciation in Germany

1837

K. Schimper

A German botanist introduces the term Ice Age (Eiszeit)

L. Agassiz

Convinced by Venetz and J. de Charpentier of the validity of the ‘glacial theory’, announces his theory of the ‘Great Ice Age’ to the Swiss Society of Natural Sciences in Neuchatel ˆ

G. Martins

A French geologist suggests concept of vast continental ice sheets, based on work from Spitsbergen

W. Buckland

Renounces his belief in a biblical flood to explain drift sediments

1838

1839

C. Lyell

Introduces the term Pleistocene

Conrad

Accepts glacial theory in North America

1840

L. Agassiz

Publishes Etudes sur les glaciers, Neuchatel. ˆ He travels in Scotland with Buckland and Lyell convincing them that the surficial sediments are of glacial origin

1841

C. McLaren

Argues from evidence in Scotland for eustatic changes in sea level during the Ice Age.

J. de Charpentier

Publishes Essai sur le glaciers et sur le terrain erratique du bassin du Rhone, ˆ in Switzerland

1842–1843

J. Adhemar and U. Leverrier

In France the concept of an astronomical theory to explain the origin of glaciation is put forward

1843

J. D. Forbes

Following expeditions to the Swiss and French Alps and to Norway, publishes the first ‘glaciological’-style text on glacier movement, mass balance and erosional processes

1845

W. Hopkins

Recognition of mechanism of glacier sole sliding

1847

L. Agassiz

Recognition that North European and Alpine glaciations separate

1854

–

The term ‘Quaternary’ generally accepted in Europe and North America

1859

O. Torell

Postulation of a Fenno-Scandian Ice Sheet

1863

A. Geikie

Publishes first scientific paper on glacial deposits mapped in Scotland

1864

J. Croll

Publishes his theory on astronomical causes of glaciation

1865

T. Jamieson

Argues that isostatic depression of the land surface occurred owing to overlying weight of Pleistocene Ice Sheets, evidence from raised beaches

1870

G. K. Gilbert

Shows from mapping in Utah, USA, the past existence of large proglacial and ice-dammed lakes (Lake Bonnevile)

F. von Richtofen

Concludes that loess is of glacio-aeolian origin

1871

A. Worthen

Shows that more than one glaciation occurred in Illinois

1875

–

HMS Challenger returns from circum-global oceanographic expedition with extensive deep-sea deposit data

J. G. Goodchild

Recognition of melt-out tills

GLACIAL ENVIRONMENTS – MODERN AND PAST TABLE 1.1. Continued 1883

T. C. Chamberlin

Develops a ‘till’ classification

W. J. McGee

Recognizes the concept of U-shaped valleys and links to glacial erosion

1892

W. Upham

Introduces and describes ‘lodgement till’

1894

J. Geikie

Publishes new edition of the ‘Great Ice Age’ with glacial maps of North America, Europe and Asia

1897

G. De Geer

Recognition of use of varves in glacial chronology

1906

B. Brunhes

Discovery of paleomagnetic evidence of polar wanderings and reversals

G. K. Gilbert

Major scientific paper on glacial erosional forms and processes

1909

A. Penck and E. Bruckner ¨

Mapping of Alpine foreland terraces and reconstructing a Pleistocene age succession (Gunz, ¨ Mindel, Riss, Wurm) ¨

1914

R. S. Tarr and L. Martin

Recognition of surging behaviour in glaciers

1920

M. Milankovitch

Publishes astronomical theory of ice ages based upon solar forcing

1926

H. Mothes

Uses seismograph and dynamite charges to estimate ice thickness on the Hintereisferner, Switzerland

1929

M. Matuyama

Discovery of polar reversals

1933

E. Sorge

Estimation of Greenland Ice Sheet thickness

1935

W. Schott

German meteor expedition uncovers evidence of Pleistocene period from equatorial Atlantic ocean cores

1941

C. D. Holmes

Extensive discussion on use of till fabrics

1944

S. Thorarinsson

Introduction of use of tephrochronology

1945

C. M. Mannerfelt

Recognition of value of glacial meltwater channels in interpreting deglaciation

1947

H. Urey

Publishes work on oxygen isotope ratio dating method

H. Carol

Suggested glacial erosion by plucking may be due to regelation processes

R. F. Flint

Publication of Glacial Geology and the Pleistocene Epoch

1948

H. W. Ahlmann

Develops thermal classification of ice masses

1950

R. Beschel

Introduction of use of lichenometry

1951

W. Libby

Develops radiocarbon (14C) dating method

1952

J. Nye

Development of theories on ‘glacier flow’

1953

G. Hoppe and V. Schytt

Describe ‘flutings’ on subglacial sediment surfaces

1954

J. Glen and M. F. Perrutz

Experimental work on ice rheology

G. deQ. Robin

Estimation of Antarctic Ice Sheet thickness

1955

J. Glen

Development of law of ice deformation ‘Glen’s Law’

R. F. Sitler and C. A. Chapman

Early use of micromorphology in glacigenic sediments

C. Emilliani

Use of oxygen isotopes in determination of sea paleotemperatures

1957–1958

–

International Geophysical Year – major initiatives in glaciology

1957

J. Weertman

Discussion of ice basal motion by slippage on a film of water

J. K. Charlesworth

Publication of the Quaternary Era

L. Liboutry

Discussion of ice basal motion by basal cavity development

J. B. Sissons

Re-interpretation of glacial meltwater channels systems and their significance in deglaciation

J. H. Hartshorn

Introduces term ‘flowtill’

7

8

GLACIAL ENVIRONMENTS – MODERN AND PAST

TABLE 1.1. Continued 1961

D. B. Ericson et al.

Early compilation of deep sea sediment cores showing evidence of warm and cold phases in synchroneity with on-land warming and cooling phases

J. A. Elson

Recognizes ‘deformation’ till

1962

J. B. Sissons

Re-interpretation of raised shorelines and the significance of glacial isostasy in Britain

1963

A. Gow

Reports details of penetration of Antarctic Ice Sheet at Byrd Station

1964

A. T. Wilson

Suggestion that Antarctic Ice Sheet may surge leading to a trigger for global glaciation

1965

J. Gjessing

Suggest role of wet till in forming P-forms

1966

I. J. Smalley

Introduces idea of dilatant subglacial landforms

1967

J. B. Bird

Regional synthesis of glacial landforms

1969

IGCP

Initiation of the International Geological Correlation programme Project 38 on pre-Pleistocene tillites

J. Lunqvist

Recognizes regional patterns of subglacial landforms

J. A. T. Young

Recognition of extreme local variability of till fabrics

1970

W. F. Budd et al.

Early attempts at mathematical ice sheet modelling (Antarctica)

1971

A. Dreimanis and U. Vagners

Develop idea of ‘terminal’ grade in tills

1972

H. Rothlisberger ¨ R.L. Shreve

Major papers on subglacial hydrology

1973

D. Krinsley and J. Doornkamp

Publication of Atlas of Quartz grains using SEM

1974

G. S. Boulton

Develops an abrasion rate equation and links glacial erosion to lodgement process

1976

G. deQ. Robin

Advocates a ‘heat pump’ effect beneath ice masses – polythermal bed condition

1977

R. Aario

Recognizes possibility of subglacial bedform continuum – rogen moraine-drumlin-fluted moraine

1978

D. E. Sugden

Recognition of continental and regional patterns of glacial erosion related to bed thermal states

1979

G. S. Boulton and A. S. Jones

Theoretical discussion on model advocating deformable subglacial bed conditions

N. Eyles

Recognition of the use of lithofacies in glacigenic sediments

B. Hallet

Develops an abrasion rate model

J. Menzies

Recognizes role of porewater in subglacial sediment deposition and related drumlin formation

1981

G. H. Denton and T. J. Hughes

Compilation of Late Wisconsinan Ice Sheets Chronology and extents

1983

J. Shaw

From recognition of stratified drumlin core suggests ‘flood’ hypothesis for formation

1984

S. Manabe and A. J. Broccoli ˘ V. Sibrava et al.

Recognition of the role of continental plate positions and motion in timing of glacial periods Major correlation of Quarternary glaciations of the northern hemisphere

R. B. Alley et al.

Report of soft deformable bed beneath Ice Stream B, West Antarctica

1986

Pleistocene glaciations were adopted in northern Europe as the Elster, Saale and Weischsel Glaciations, and in Britain as the Lowestoft, Gipping and Devensian Glaciations. In North America, four glacial phases, similar to the European Alpine Model,

were primarily used – the Nebraskan, Kansan, Illinoian and Wisconsinan. Today this simple fourfold sequence has been shown to be rudimentary and a larger number of glacial and interglacial events have been shown to exist. At present, over 17 major

GLACIAL ENVIRONMENTS – MODERN AND PAST

continental glaciations appear to have occurred during the Pleistocene and this figure may increase as new data are uncovered (Andrews, 1997). 1.4.3. Multidisciplinary Nature of Glacial Studies The study of past glacial environments has generated a multidisciplinary strategy of scientific inquiry (Fig. 1.2). The underlying thesis of all these separate, yet connected, studies is to understand past glacial events and processes, global climatic and oceanic circulation patterns, and botanical and zoological adaptations and adjustments. With that knowledge, predictions can be made of possible future global events, patterns and responses. Central to the study of past glacial environments are glacial sediments. Their properties, characteristic structures, fossil content, age, stratigraphic position,

Dating Methods

landform association, morphology and location are characteristically the sole evidence from which reconstruction of past glacial environments can be made. To aid in reconstruction, surrogate and longdistance evidence must be gathered to augment what may, at times, be scanty data. In recent years, for example, radiometric dating techniques, and oxygen isotope records from deep ocean sediments and ice sheets, have supplied objective and precise information. Similarly, models that can be run repeatedly with ever-changing parameters such as those for weather patterns or oceanic circulation provide additional clues as to conditions during incipient, full and waning global glacial and interglacial phases. Such models reveal theoretical possibilities and feasibilities, providing scientific support for explanations of past environmental conditions and constraints.

Paleopedology Paleozoology Glaciology

Cryology

Paleoecology

Studies of Past Glacial Environments

Geology

9

Geomorphology

Paleoceanography

Anthropology

Sedimentology

Geographical Information System Paleobotany

Geotechnique Remote Sensing

Paleoclimatology

FIG. 1.2. Diagram of the multidisciplinary nature of studies pertinent to past glacial environments.

10

GLACIAL ENVIRONMENTS – MODERN AND PAST

1.4.4. Glacial Sediments and Glacial Geomorphology/Geology This textbook takes glacial sediments as central to any explanation of modern or past glacial environments. The study of these sediments comes under the heading of glacial geomorphology or geology. Glacial studies, in the past, have been of two types: those interested in sediments and landforms and those interested in the chronological sequence of glacial events. In many cases these two approaches were combined, thus developing a series of chronostratigraphic models of glacial events for a particular location (Chapter 15). Prior to the Second World War, research in glacial studies was strongly geological with an emphasis on processes and sedimentology but often within spatially limited areas. After 1945, a morphological approach was adopted in which the geographical distribution of landforms was used foremost in developing explanations of the glacial events for a specific site. Emphasis on sediment types and stratigraphic relationships was only used where stratigraphy was of interest and, too often, little attention was given to glacial sedimentology. By the 1970s, a search for explanations led research to again rely more heavily on glacial sedimentology (with a dependence on glaciological conditions) thereby de-emphasizing the once strongly geographical paradigm (Boulton, 1987b). This glaciosedimentological approach seeks to find answers by considering all glacial sediments and landforms within the framework of known sedimentological and glaciological conditions before, during and after specific events within a glacial system. The success of this approach therefore hinges upon knowledge of those ‘events’ and ‘conditions’. 1.5. RESEARCH ISSUES IN GLACIAL ENVIRONMENTS When considering the vast array of research in glacial environments some general trends can be observed that have had an enormous impact on various facets of glacier studies. Over the past several decades the techniques for studying glaciers have vastly improved, for example, in the accuracy of measurements of ice movement and mass balance through the

use of Landsat and other satellite imaging methods. From the first ice cores extracted from near Byrd Station, Antarctica, to the more recent discoveries of the Greenland Ice-Core Project (GRIP) (GRIP, 1993), the accuracy, detail and outpouring of data on ice cores, their geochemistry, geochronology and climatic implications, have been astounding. The precision of techniques in geochemical analyses of ice cores has reached a level whereby details of atmospheric chemistry can now be revealed that were hitherto unknown (Lorius, 1999). As understanding of modern ice sheets has grown, the intricate relationships of feedbacks and counter-feedbacks between the atmosphere, oceans and ice mass modifications have been translated into first-order ice sheet models. Recently, sophisticated computer-generated ice sheet models have increased our understanding of ice sheet growth and potential stability/instability, permitting future predictions of ice sheet behaviour (Jenson et al., 1996; Marshall and Clarke, 1997a). In tandem with these findings has been the increasing knowledge concerning Pre-Pleistocene glaciations and the overall causative mechanisms that may lead to global glaciation. In pondering how global glaciations began and ended, our knowledge of past global climates, carbon dioxide levels, biomass productivity and other habitat environmental indicators has dramatically improved, thereby permitting predictions of the possible effects of global greenhouse warming to be better understood, if not anticipated (Raynaud et al., 1993). What has often been ignored is an attendant and critical understanding of ice physics. A tacit recognition of this subdiscipline has always existed, but attempts to directly merge the findings of glaciological research into glacial geomorphology has been latently resisted. The result has been the divergence at times between glaciology and glacial geology. Today the value of combining both disciplines in a concerted effort to understand glaciers and their environments is gradually being accepted. It is no longer possible to try to understand the deposition of subglacial sediments without first considering the thermal and mechanical parameters of subglacial environments. What has emerged from the chaos of ideas and mis-developed theories concerning glacial sedimentological processes is that glacial environments

GLACIAL ENVIRONMENTS – MODERN AND PAST

are much more complex than perhaps previously considered. Just as glaciology has undergone profound changes and evolution, the wider field of glacial geology has undergone maturation. When the Europeans and North Americans first began studies of glaciated terrains they brought with them a profound geological appreciation of sedimentary stratigraphy as well as a spatial awareness of topography (physiography). Descriptions of sections in the field were often so detailed and accurate that even today when visiting these same sites, the accuracy of the precise details commented upon by early ‘glacialists’ is evident. Somewhere in the early twentieth century the science became diverted into a dominance of form over sediment. Landforms and physiographic studies became de rigueur and increasingly detailed studies of sediments and depositional processes were often relegated to a subordinate place or, in many studies, virtually ignored. From this developed what might be termed a ‘spatiomorphological school’ of thought that placed emphasis on spatial interrelationships and supposed geographical locational cause and effect. Likewise, an intense period of glacial chronological studies evolved. Glacial sediments and interrelationships were only lightly considered in pursuit of the establishment of temporal–spatial relationships between landforms and landform assemblages in order to develop scenarios of glacier fluctuations at local, regional and continental scales. This paradigm achieved enormous successes in explaining and developing global and regional glacial chronologies. However, in detailed studies of local glacial variations this paradigm failed to provide answers, if only for want of sound sedimentological and glaciological inputs. Although many workers held to a strong sedimentological methodology, only since the 1970s has a general swing toward a ‘glaciosedimentological’ school of thought begun to evolve. Emphasis was placed on all aspects of glacial processes and glacial ice dynamics in an attempt to establish glacial process-patterns within a time frame constrained not by potentially perilous spatial–topographical relationships but a broader, more complex and stochastic, appreciation of sediment processes, ice mechanics and transient environmental conditions. At the same

11

time a concomitant shift away from a ‘unique’ appreciation of landforms has emerged that places landforms as part of suites of associated forms developing in similar environmental conditions. As Figure 1.2 illustrates, the breadth of scientific interest in glacial environments encompasses many diverse fields of inquiry and each one has pertinent research issues. At the general level, however, there are several issues that transcend discipline boundaries, for example: 䊉 䊉

䊉 䊉

䊉 䊉 䊉

䊉

䊉

the problems of recognizing glacial from nonglacial sediments; discrimination of different lithofacies and facies associations within one or adjacent glacial environments; recognizing the contribution of water in the many subenvironments of glacial systems; identifying boundary interface processes and related bed- and landform initiation in subenvironments of the glacial system; identifying those sediment characteristics that are indicative of diagenesis; distinguishing and verifying the impact of freezing conditions on sediments; characterizing habitat changes close to ice masses in relation to climate change, plant colonization, and faunal and human migration; resolving, at the more local scale, the interplay between land and sea levels before, during and after global glaciations in relation to rapid changes in ice sheet volumes, ice marginal positions and oceanic circulation; developing objective multiple taxonomic criteria in glacial stratigraphy; acquiring dating techniques and refining dating resolution to permit more precise determination of events and process rates; determining modes of transport in glacial systems from better definition of transport signatures on individual grain surfaces and understanding of erosion, transport and deposition sequences as manifest in sediment placer bodies.

A persistent problem in all glacial studies is the recognition of glacial sediments. Although many characteristics have been suggested, there still remain concerns when interpreting lacustrine, marine

12

GLACIAL ENVIRONMENTS – MODERN AND PAST

and distal proglacial sediments. Once it has been determined that a particular sediment or form is glacial, controversy often surrounds the origin of a particular facies or sub-facies. Within glacial systems there are several environments that produce sediments, forms and internal structures that are virtually identical and indistinguishable (Menzies, 1996, chapter 9). Under these circumstances, facies associations may often aid in the resolution of a particular problem, for example, in the interpretation of the glacial sequence at Scarborough Bluffs near Toronto, Ontario (Eyles et al., 1983; Dreimanis, 1984; Karrow, 1984a,b; Sharpe and Barnett, 1985), or as to the nature of Precambrian Port Askaig sediments of Scotland (Chapter 13). However, since glacial sediments may progress through repeated cycles of erosion and deposition, an equifinality is commonly encountered in specific facies units. Only by using related diagnostic attributes possibly linked to stratigraphic position, location or other facies associations can a facies unit be designated in certain cases. Discrimination is especially problematic between sediments in adjacent facies environments such as proglacial proximal, and subaqueous or subglacial and subaqueous diamictons. It has become apparent that water, as meltwater and porewater, has played a much greater role in most glacial subenvironments than hitherto assumed. Processes of subglacial erosion, for example, are much more widespread than was recognized in the past both at the micro- and macro-scale levels (Kor et al., 1991). More controversial has been the concept of ice sheet stability and subglacial bed conditions controlled, to some extent, by massive subglacial floods (Rains et al., 1993; Shaw et al., 2000). The central role that meltwater and porewater plays in varying effective stress levels within glacial sediments has profound effects on sediment strength and mobility (Chapters 4 and 7). However, precise details of sediment geotechnical changes as controlled by porewater content remain rudimentary. The relative importance of porewater within glacial debris has been largely ignored. However, in many subenvironments where stress-sensitive sediments occur, such as flow tills, porewater content is the controlling variable in determining rates of deposition and transport, and effective stress levels. The pore-

water content of subglacial deformable beds controls the rate of debris mobilization and possible bedform development and/or survival (Menzies et al., 1997). From modern glacial environments it is known, for example, that massive j¨okulhlaups occur causing devastating effects in the proglacial zone, moving vast quantities of debris from the subglacial, terminal and proximal areas of ice masses, yet stratigraphic recognition remains imprecise and problematic. Much of geomorphology is concerned with the interaction of earth surface processes across the boundary interface between the atmosphere and the Earth’s surface, or the base of an ice mass and its sole, or the bed of a lake or sea or river and flowing water. It is at these interfaces that landforms and bedforms develop. Of intrinsic interest therefore in glacial environments is the reaction between ice, meltwater and the Earth’s surface and the formation of glacial landforms and bedforms. With insufficient knowledge of many glacial processes, explanations, although often remarkably accurate, were imprecise concerning processes and rates of landform development. As understanding of glacial processes increased, details of landform development have likewise become more complex. At the same time it has become apparent that many glacial landforms are not unique but are closely related in origin. In examining glacial sediments it must be remembered that some of these sediments have been deposited for many thousands of years resulting in subsequent changes in their properties. These changes may be the result of exposure at the Earth’s surface and the influence of subaerial processes, possibly exhumation owing to surface erosion, uplift from a subaqueous to subaerial position because of isostatic or eustatic changes, the impact of vegetation colonization and soil development, and the impact of climatic change. Any of these effects may act alone or in concert over various periods of time and to varying depths within the sediments. All of these influences are generally referred to as processes of diagenesis, but the degree and extent of impact is often poorly understood or even recognized at present. Diagenesis may manifest as geotechnical alterations, the result of consolidation, removal of fine sediments, particle to particle interrelationships of fabric or structure, and/or new fracture geometries. Geochemical changes may

GLACIAL ENVIRONMENTS – MODERN AND PAST

occur that cause authigenic mineralization, mineral weathering, pore cementation, and/or mineral precipitation. These changes to the original sediments may be visible and significant while in other sediments the changes are imperceptible and minor. Within the glacial environment transient freezing conditions are commonplace. Sediments are frozen on a seasonal basis or occasionally for longer. Freezing may occur owing to changes in stress levels, glacier ice thickness variations, meltwater surcharges or other temporary thermal fluctuations. As sediments transit the glacial system it is likely that they may undergo several freeze/thaw cycles. At present, recognition of the existence of past freezing conditions is difficult to substantiate. The affirmation of frozen conditions would aid in the recognition of particular environments within glacial environments. As ice masses creep forward, or slowly retreat, surge into a proglacial lake, float as a tidewater glacier or join an ice shelf, the margins of these ice bodies are in flux. Ice marginal conditions, therefore, are of immense importance since many sediments and landforms are associated with marginal environments and fluctuations. Recognition is often indirectly obtained from sediment facies sequences. Although land-based ice marginal fluctuations are understood, knowledge of land/water glacier margins are limited. This limitation is, first, due to a restricted knowledge of processes occurring immediately at the land/water ice grounding line positions – largely the result of past inaccessibility – and second, due to the inability to clearly discriminate in the glacial sediment record evidence of lithofacies indicative of the land/water margins. For example, it has been hypothesized that in some instances drumlin development may be linked to near marginal subglacial/subaqueous ice conditions (e.g., in Ireland (Dardis, 1987; Hanvey, 1987) and Canada (Menzies, 1986) drumlin formation may be linked to nearby grounding line positions in the sea or in a large proglacial lake).

13

Although present ice marginal environments are usually associated with tundra-like environments, the ice margins of the mid-latitude ice sheets during the Pleistocene were often less extreme than is witnessed today. Likewise, the margins of the vast Fennoscandian and Laurentide Ice Sheets varied enormously in terms of climate, flora and fauna along their southern and northern edges. Much remains to be learned of the margins of the past ice sheets to permit understanding of plant colonization, animal movement and migration, as well as plant and animal evolution and extinction. A persistent problem in studying glacial environments is in establishing the relative position of land and sea in local areas (Menzies, 1996, chapter 11). Often isostatic and eustatic changes have led to repeated inundations and re-emergence of land surfaces, generating complex stratigraphies and landform assemblages. Although the general framework of land/sea changes are known, much remains to be elucidated at the local scale owing to regional variations in mantle viscosity, ice mass volumes and marginal movements and local topography. As knowledge of glacial processes and depositional mechanics is refined, the ability to improve the stratigraphic resolution at different sites should be enhanced (Chapter 15). Allied to improvements in stratigraphic definition is the need to increase the resolution of dating techniques and the number of different dating methods that can be used on glacial materials (Menzies, 1996, chapter 14). Both stratigraphic definition and dating resolution demand a better understanding of glacial processes, process rates and the recognition of hiatus in erosional and depositional events before a greater comprehension of the chronology and sequence of glacial events can be resolved. The study of glacial environments remains a vibrant area of research in many disciplines with increasingly applied aspects related to human activities and global warming at the advent of the twenty-first century.

This Page Intentionally Left Blank

2

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION P. E. Calkin (with contributions by G. M. Young)

were separated by intervals when little or no terrestrial ice occurred and temperatures were often higher than those of today. This chapter provides a brief review of the glacial stratigraphic record with some consideration of causal factors and mechanisms responsible for driving climatic changes and the spectrum of glacial advances and retreats. Agassiz was not the first to discern the evidence of former extensive continental glaciations (Nilsson, 1983). Agassiz had been convinced of the idea by observations at the Swiss glaciers of Diablerets and Chamonix while in the company of Jean de Charpentier, Director of Mines for the Swiss Canton of Valais (Chorley et al., 1964). Charpentier (1841) supported a concept of a former much-enlarged Alpine glaciation. Similar innovative, but unaccepted, arguments had been presented even earlier, in 1821, to the Helvetic Society by the Swiss engineer, Venetz-Sitten. Both Charpentier and Venetz had acquired the idea of extensive glacier advances from the mountaineer Jean-Pierre Perrudin of Val de Bagnes. Venetz had also supported the idea presented by Hutton (1795) who suggested that erratic boulders were transported from the Alps to the Jura Mountains by a great mass of ice. Elsewhere, Esmarck (1824, as in Nilsson, 1983) proposed comparable erratic transport by

2.1. INTRODUCTION It was only 160 years ago that Louis Agassiz announced his theory of a ‘great ice period’ at a meeting of the Helvetic Society (Chorley et al., 1964). ´ In his publication, Etudes sur les Glaciers, Agassiz (1840) suggested that a vast sheet of ice had once extended from the North Pole to the Alps. He also presented a model involving a succession of ice ages driven by long-term trends of radiational cooling; however, he believed evidence was lacking for more than one glacial event. In its original form, Agassiz’s whole theory had several serious flaws, but his ideas were remarkably significant and far-sighted. Field observations and indirect probing via deep sea and ice sheet coring have now confirmed a long and complicated record of glaciations, or intervals of global cooling, and major ice sheet advances that extend over at least the last 2.4 million years (Ma) of Late Cenozoic time. Furthermore, geological studies on every continent have helped to document a record of at least four major intervals of geologic time preceding this last ‘ice age’ when fluctuating ice sheets blanketed a major portion of the Earth’s crust. The earliest of these pre-Late Cenozoic ice ages spanned tens to hundreds of millions of years. They 15

16

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

glaciers in Norway, as had Bernhardi (1832) for Scandinavia and north Germany. These ideas were disseminated rapidly by Buckland and Lyell in Britain and by Edward Hitchcock (1841) and eventually Agassiz in North America. By 1835 Lyell realized that many of the surface sediments in Britain contained marine shells and a novel theory to explain these sediments was required. Lyell’s ‘Drift Theory’, which was intended to combat the prevailing ideas of catastrophic flooding, linked these ‘diluvium’ or ‘drift’ deposits to transport by icebergs. The early glacialists accepted the concept of a single integrated glacial event forming all glacial deposits. However, this mono-glacial hypothesis was soon dispelled as two glaciations were inferred from distinct drift layers in the Vosges Mountains by Collomb (1847), in Wales by Ramsay (1852) and in Scotland by Chambers (1853). Interdrift strata were referred to as interglacial and the interval of time between two successive glaciations (glacials or glacial stages) were interglaciations (interglacials). The glacial stages were subsequently subdivided into stades (stadials), marking times of climatic deterioration and secondary glacial readvance, and interstades (interstadials). Interglacial times have been variously defined as times when the climate was essentially as warm as at present. Long before the extent of this last ice age had been delimited, the search for ice ages earlier in the geologic record had begun. In Britain, Ramsay (1855) suggested that the presence of angular, polished and striated rock fragments of boulders in Permian breccias of Shropshire and Worcestershire indicated the existence of glacier ice during the Permian. Although these were discounted as glaciogenic (glacial-formed), this find stimulated the search for evidence of a Permian glaciation elsewhere and led to the verified report of the glaciogenic ‘Talchir’ Boulder Beds in India (Hambrey and Harland, 1981). Agassiz’s work was barely published when Adhemar (1842), a French mathematician, suggested that the primary instigator of ice ages might be the Earth’s orbital path around the sun. Croll (1864) investigated these ideas and added the effect of orbital eccentricity or change in the Earth’s orbital shape. Croll suggested that long cold winters and short hot summers of high latitudes would favour glaciation and, in addition, the

precession effect caused ice ages to occur alternately in northern and southern hemispheres. Milankovitch (1930) suggested that the climatic effect of change in radiation owing to precession, eccentricity and the axial tilt (obliquity) would be sufficient to cause the glaciations and interglaciations that comprise ice ages. The basis of the astronomical theory of paleoclimates and, in particular, Milankovitch’s (1941) theory of variable solar insolation for climatic change has been intensely studied and has, since the late 1970s, been considered the major basis for pacing of Late Cenozoic glacials and interglacials. 2.2. PRE-CENOZOIC GLACIATION Study of ancient glaciations has been spurred primarily by concerns about possible anthropogenic modification of the Earth’s climate. The only long-term record of climatic change is the geologic record. Glaciations represent significant perturbations in the Earth’s climatic regime. The evidence of such glaciations is commonly preserved and relatively easy to recognize in the geologic record. 2.2.1. Astronomical Background It is considered that stars have a history involving a gradual increase in luminosity followed by a decrease (Gilliland, 1989; Kasting and Toon, 1989). According to this theory, the radiative power of the sun, during the early part of Earth’s history, was probably about 70 per cent of its present value. Kasting et al. (1984) estimated that the average temperature on Earth in Early Archean times (3–4 billion years ago (Ga)) would have been about 30°C lower than at present, given the present day concentrations of atmospheric greenhouse gases. The geological record is strongly at odds with such an interpretation, for the oldest rocks on Earth (~4 Ga) contain clear evidence of the presence of liquid water (not frozen). There is also a near-complete absence of evidence of glaciation in the stratigraphic record for the Archean. These contradictions led to the concept of the ‘faint young Sun paradox’ (Kasting, 1989). A solution to the problem is that the composition of Earth’s ancient atmosphere differed significantly from that of the present. It has

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

been proposed that the early atmosphere was rich in CO2 . There is a vast reservoir of CO2 (equivalent to about 60 bar) currently trapped in carbonate rocks at or near the surface of the Earth’s crust. Thus, in spite of the ‘faint young Sun’, geological evidence suggests that the climate on the early Earth was sufficiently warm to permit flowing water. 2.2.2. Possible Relevance of Plate Tectonics to Ancient Glaciations 2.2.2.1. The paleomagnetic record Since the development of plate tectonic theory there have been many attempts to apply these theories to ancient glacial deposits. Most involve attempts to correlate glaciation on a particular continent with periods when the continent was carried into high paleolatitudes. Glaciated areas, at present, are mostly in high latitudes. Therefore, it seems reasonable that ancient glaciations occurred in a similar paleogeographic setting. This line of reasoning has been successfully applied to glacial deposits of the combined southern continents (Gondwanaland) during the Paleozoic with glaciations ranging from the Devonian to Permo-Carboniferous times (Caputo and Crowell, 1985). Attempts to explain the very extensive Late Proterozoic glacial deposits by a similar mechanism have been frustrated by conflicting results from paleomagnetic studies. For example, work by Tarling (1974) on Late Proterozoic glacial deposits in Scotland suggested deposition in relatively low paleolatitudes. On the other hand, evidence has been provided favouring widespread glaciation during the Neoproterozoic (1.0 Ga–543 Ma), most of which appears to have occurred at low paleolatitudes (Embleton and Williams, 1986; Chumakov and Elston, 1989; Sohl et al., 1999). Recently, Williams and Schmidt (1997) also suggested that much older (Paleoproterozoic) glaciogenic rocks of the Huronian succession in Canada also formed at low paleolatitudes. These findings, among other lines of evidence, led Hoffman et al. (1998) to revive Harland’s (1965) idea of a completely glaciated planet – the ‘Snowball Earth’ hypothesis. Alternatively, it has been proposed that the obliquity of the Earth’s

17

ecliptic was much greater in Precambrian times, so that if it entered a glacial epoch, continents situated at a low latitude would be preferentially glaciated (Williams, 1975). 2.2.2.2. The supercontinental cycle It has been hypothesized that the Phanerozoic history of North America was characterized by a series of relative rises and falls in sea level leading to periods of flooding and periods of withdrawal of the sea. It has been proposed that Phanerozoic climatic history was largely controlled by a ‘supercycle’, involving flooding and exposure of the continents to produce what was termed alternating ‘greenhouse’ and ‘icehouse’ conditions, on a scale of about 400 Ma. Young (1991), among others, extended the concept of the supercycle to the Precambrian. The supercycle involves the periodic amalgamation of the continental crust into a supercontinent and subsequent fragmentation. During periods of ‘supercontinentality’ the amalgamated mass of continental lithosphere has a blanketing effect on the release of thermal energy from the Earth’s interior (Worsley and Nance, 1989), resulting in its elevation and consequent relative lowering of sea level. Subaerial exposure of such large areas of continental crust and concomitant elevation of mountain belts leads to enhanced weathering that, in turn, causes drawdown of large amounts of atmospheric CO2 . Up to 80 per cent of the present withdrawal of CO2 from the atmosphere is due to this mechanism. Thus, there is a possible connection between periods of supercontinentality (emergence of the continental crust) and reduction of atmospheric CO2 . The reduced greenhouse effect could result in significant cooling, leading to the onset of a glacial period. 2.2.3. The Archean Record An unusual aspect of the geological record is the near complete absence of evidence of glaciation during the first half of geologic history (Fig. 2.1). The Witwatersrand Succession of South Africa contains evidence of glaciation in Archean times (Young et al., 1998) as does the Stillwater complex in Montana (Page, 1981). One possible interpretation is that, despite of the

18

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

FIG. 2.1. Distribution in time of ice ages plotted linearly, at two scales. ␧, Cambrian; ␪, Ordovician; S, Silurian; D, Devonian; C, Carboniferous; P, Permian; TSR, Triassic; J, Jurassic; K, Cretaceous; T, Tertiary (from Crowell, 1982; reprinted with permission from Climate in Earth History. Copyright 1982 by the National Academy of Sciences. Courtesy of the National Academy Press, Washington DC).

inferred low solar luminosity, large amounts of atmospheric CO2 maintained relatively high surface temperatures. 2.2.4. The Paleoproterozoic Ice Age The earliest evidence of glaciation is recorded by terrestrial and marine diamictites and proglacial deposits. The oldest substantiated evidence of widespread glaciation is found in rocks of Paleoproterozoic age reported from North America, Finland, South Africa, Australia and India. Paleoproterozoic deposits of North America include the Huronian Supergroup of Ontario (Young, 1995, 1997) and a near-identical succession in southeastern Wyoming, the Snowy Pass Supergroup (Houston et al., 1981). Other deposits include the Hurwitz Group, west of Hudson Bay, the Chibougamau Formation, northern Quebec, tillites in Michigan’s Upper Peninsula and various locally developed diamictite-rich successions around the west end of Lake Superior.

In all of these areas glacial sediments may have been preserved by an episode of continental rifting (Young and Nesbitt, 1985). In some areas with thick stratigraphic successions there is evidence of several glacial episodes separated by rocks containing evidence of warm climate and relatively intense weathering (Nesbitt and Young, 1982). The alternation of cold and warm climatic conditions is attributed to a negative feedback mechanism. It has been inferred (Taylor and McLennan, 1985) that there was a period of significant addition to the continental crust at the end of the Archean. The resultant uplift and enhanced weathering of the continental crust would have caused a decrease in atmospheric CO2 . Once continental glaciation was initiated, weathering would be inhibited in ice-covered regions by lowered temperatures. Reduced weathering rates would have led eventually to a gradual build-up of atmospheric CO2 , subsequent destruction of the ice sheets and establishment of a warm climatic regime (Fig. 2.2). This long-term alternation of warm and cold climatic regimes would continue until it was interrupted by some other factor

28

I

III

19

II

Paleotemperature (ºC)

24

4000

20 16

12

1 3000

2 3

2000

4

8 1000 4 0 130

110

90

Cretaceous

70

50

P1

Mesozoic

30

P2

P3

10

N1

0

N2

CO2 Concentration in the atmosphere (ppmv)

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

0

Q

Cenozoic Time before present (Ma)

FIG. 2.2. Temperature and carbon dioxide change estimates from the Late Cretaceous until the present. Estimates of ocean surface temperatures in low latitudes (curve 1) and high latitudes (curve 2) of atmospheric CO2 concentration (curve 3) and of surface air temperature over the Russian plains (curve 4). Time periods are divided into I, non-glacial regime; II, glacial regime; III, transition regime; P, Paleogene: respectively Paleocene (P1 ), Eocene (P2 ), Oligocene (P3 ); N, Neogene: respectively, Miocene (N1 ), Pliocene (N2 ), and Q, Quaternary (from MacCracken et al., 1990; reprinted with permission from Prospects for Future Climate. Copyright 1990 Lewis Publishers, a subsidiary of CRC Press, Boca Raton FL).

such as fragmentation of the supercontinent. If the CO2 content of the early atmosphere was much higher than today, and if the ‘faint young sun’ theory is valid, then the onset of global glaciation would have been possible at much higher partial pressures of CO2 .

2.2.5. The Mesoproterozoic The Paleoproterozoic ice age was followed by an interval of 1300 Ma when evidence is lacking for glaciation. The Mesoproterozoic seems to have been characterized by the existence of one or more supercontinents (Hoffman, 1989).

2.2.6. The Neoproterozoic Ice Age In the Neoproterozoic, glaciation occurred on all continents. This widespread and long-ranging glacial episode may have had three peaks at about 940, 770 and 615 Ma (Crowell, 1981). Two of these are recorded in the same stratigraphic sequence around the North Atlantic Basin and in central and southwestern Africa, Brazil, western North America and Australia. Many Late Proterozoic glaciogenic successions display evidence of contemporaneous rift activity (Young and Gostin, 1989a, b). In some places, associated iron formations and manganese-rich sedimentary rocks have been interpreted as products of

20

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

hydrothermal activity related to rift-related volcanism (Breitkopf, 1988). Calcium carbonate sequences cap the glacial deposits globally. Paleomagnetic results suggest that many Neoproterozoic glaciogenic rocks formed in tropical latitudes (McWilliams and McElhinny, 1980). Neoproterozoic rocks are commonly associated with dolostones, and red beds generally considered indicative of warm climatic conditions. G. E. Williams (1975) attempted to resolve these enigmatic associations of rock types, proposing significant variations in the tilt of the Earth’s spin axis relative to the rotational plane of the solar system (variations in the obliquity of the ecliptic). According to this theory, during periods of greatly increased obliquity (>54°), annual insolation in polar regions would be greater than that in low latitudes. In the event of lowered global temperatures, the equatorial belt would have been preferentially glaciated. Crowell (1983) suggested that the Late Proterozoic was a period of rapid plate tectonic movement such that continents could have moved into high latitudes

where they were glaciated, but their magnetic signature was not imprinted on the sediments until much later when located in low paleolatitudes (Williams et al., 1998). If the concept of a supercontinent and its fragmentation (Young, 1995) in near-equatorial latitudes is adopted, then active chemical weathering and reduction in atmospheric CO2 could have led to glaciation. The first effects of such a cooling episode would have been freezing of the polar seas but eventually the continental region would have been glaciated (Fig. 2.3). Glaciation would have been initiated in regions situated at high altitudes, but ultimately would have extended to sea level. With the development of widespread glaciation and lowering of temperatures, negative feedback would have led to a build up of atmospheric CO2 again and introduction of a warm climatic period. Such a mechanism would be expected to produce a sedimentological record displaying evidence of alternation of warm and cold climatic episodes. Under high obliquity conditions, such glaciations would have taken place at low paleolatitudes.

Legend

?

Max. extent of glaciation in Europe

A

Anglian limit

S

Saale limit

?

Ural Ice Cap

Extent of last glaciation

W

Weichselian limit

?

?

DE

Devensian limit

VI

D W

E

DI

Younger Dryas

?

IC

?

Helsinki

Oslo

S

Stockholm Moscow

Edinburgh

? D

? A

?

Amsterdam Cologne

Warsaw

S

W

S Kiev

Paris

Prague

Cracow

Munich

Alpine Ice Cap

Caucasus Ice Cap

Pyrennees Ice Cap

FIG. 2.3. Extent of Pleistocene glaciation in Europe. Compiled from data in Nilsson (1983) (reprinted with permission of Kluwer Academic Publishers).

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

2.2.7. The Phanerozoic Ice Age There is evidence of glaciation in the Ordovician, Devonian and Permo-Carboniferous. Evidence is mostly found on the continents of the southern hemisphere that were then amalgamated into the Gondwana supercontinent. There appears to be a relationship between the sequential glaciations of the Permo-Carboniferous and the passage of different parts of the supercontinent across the south polar region. Two main factors may have contributed to Late Paleozoic glaciation. The emergent state of the Gondwana supercontinent could have led to atmospheric CO2 depletion (anti-greenhouse effect), and since part of the supercontinent was at high latitudes this would have contributed to lower average summer temperatures, the build up of snow and the development of continental glaciers. 2.2.7.1. The Ordovician–Silurian ice age The early Paleozoic paleogeography, like that of the Precambrian, was unlike that of the present. Most of the North American, European and Siberian cratons were located in the tropics or subtropics. However, the Gondwanaland supercontinent, encompassing South America, Africa, Arabia, Madagascar, India, Australia and Antarctica, as well as Florida, southern and central Europe, Turkey, Iran, Afghanistan, Tibet and New Zealand, extended between the equator and the South Pole throughout the Paleozoic. The Paleozoic glacial centres followed roughly the drift of the South Pole, with apparent paths of the pole crossing glacial sites in chronological order (Crowley et al., 1987). Continental reconstructions for the Late Ordovician suggest that the northern continents were scattered but the Gondwana continents were joined together. North Africa, where the best evidence of Late Ordovician glaciation is preserved, was close to the South Pole.

21

period of almost 100 Ma from about 360 to 255 Ma when Gondwanaland began to drift away from the South Pole. Devonian rocks of possible glacial origin have been reported from the Amazon basin, Brazil, West Africa and Antarctica 2.2.7.3 Permo-Carboniferous ice age and later periods Evidence for this widespread pre-Late Cenozoic ice age was first discovered in Africa and Australia in 1859. Subsequently, it has been reported at deeply eroded shield margins and within local basins on all Gondwanaland continents, including the Transantarctic Mountains, Antarctica. Paleolatitudes range generally from 48 to 80°S. However, Permian glaciogenic units are also known from northeastern Siberia, which lay near the North Pole at that time. Beginning 330 Ma in South America and southwest Africa, large-scale, lowland ice sheet glaciation covered much of Gondwanaland, including India, but particularly in Africa where, at ~280 Ma, a subpolar, marine-based ice sheet has been described in the Dwyka Formation (Visser, 1989). The glaciogenic Talchir Boulder Beds of India were deposited in Early Permian time. The termination of glaciation in Australia coincided with the formation of hot deserts in eastern South America and northern Africa, which were in the tropical latitudes at this time. Evidence for glaciation is extremely detailed, particularly, for example, from ice tongues that reached into Brazil and Uruguay from Africa and deposited the Itarare strata (Rocha-Campos and Dos Santos, 1981). In some areas of Gondwanaland, ten or more glacial episodes have been recorded with glacial units as thick as 1000 m, and multiple flow centres, many of which may have never merged. This was a time of fusing of the Gondwanaland and Laurasian supercontinents to form Pangaea and the creation of large outpourings of lava and construction of mountain chains.

2.2.7.2. The Devonian period Glaciation during the Devonian and Permo-Carboniferous periods would appear to have developed in a manner similar to that during the Ordovician–Silurian periods (Veevers and Powell, 1987), spanning a

2.2.8. The Pattern of Ice Ages and Events Leading to Late Cenozoic Glaciation Glaciation was apparently extensive and prolonged in the Precambrian relative to that of the Phanerozoic. It

22

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

is unlikely that all glaciations have a single cause, rather they are probably the result of a complex interplay of conditions within the Earth and at its surface. The near-complete absence of Archean glacial deposits in spite of the inferred ‘faint sun’, is attributed to high CO2 levels and a relatively small area of exposed continental crust. In the Paleoproterozoic the first widespread glaciations that occurred have been attributed to an increase in continental crust (Taylor and McLennan, 1985), and a reduction in CO2 . There is an absence of glaciation during the long period from about 2.0 to 1.0 Ga, despite evidence for the existence of a supercontinent. This is attributed to a long period of unusually widespread magmatic activity, postulated to have maintained high atmospheric CO2 levels and thereby precluding glaciation. The Neoproterozoic record reveals the greatest proliferation of ice sheets the world has known. Many of these glacial centres developed at low paleolatitudes. Widespread evidence of rifting and some paleomagnetic data support the concept of the existence of a supercontinent or supercontinents during this period. Such a supercontinental configuration at tropical latitudes could have led to an unprecedented reduction in atmospheric CO2 and extensive glaciation into low paleolatitudes. A negative feedback mechanism would eventually have induced a warming trend and a return to a warm climatic regime. By this means, an alternation of warm and cold climatic episodes could have been produced. The cycle of warm and cold climatic episodes was broken when continental fragmentation took place at or near to the end of the Precambrian, yet no simple cycle can be distinguished for these ancient ice ages (Crowell, 1982). Low latitude glaciation may thus be explained as the result of complete global glaciation (‘Snowball Earth’) or resulting from high obliquity of the Earth’s ecliptic (Hoffman et al., 1998). Data from the Phanerozoic are much more closely constrained, both paleontologically and paleomagnetically. These more recent glaciations appear to have taken place on continents in high paleolatitudes (Worsley and Kidder, 1991). The Cenozoic glaciation, and possibly also the Late Ordovician event, affected regions in high latitudes at periods when relative sea level was low.

No Triassic glaciogenic deposits are known and there are no tillites recognized from units of Jurassic and Cretaceous age, although the existence of significant glacier cover over Siberia and Antarctica at high paleolatitudes cannot be ruled out (Hambrey and Harland, 1981). The Mesozoic was an era of equitable global warmth as the progressive break-up of supercontinents, Laurasia and Gondwanaland, created high sea levels, extensive continental flooding and new ocean basins. In general, this and probably earlier non-glacial climate conditions were characterized by concentrations of CO2 five times greater than those of the pre-industrial world with temperatures 10–15°C above those of the present at high latitudes. Several major paleogeographic changes were taking place in the Mesozoic and in early Tertiary time. These were accompanied from the Late Cretaceous by a somewhat irregular decrease in global temperatures that was associated with a decrease in the atmospheric concentration of CO2 (MacCracken et al., 1990). Antarctica had moved into the South Pole area by the Tertiary Period and Australia had started its northward movement from Antarctica (Stump and Fitzgerald, 1992). A second major phase in Antarctica’s thermal isolation occurred between Late Oligocene to Miocene time (–24 Ma) (Table 2.1) with the opening and deepening of the Drake Passage between Antarctica and South America (Matthews, 1984). Simultaneously, the Arctic Basin was forming and northwest and northeast extensions of the North Atlantic were opened by the spreading of the North Atlantic Ridge. This allowed the separation of Greenland from Europe about 37 Ma, leading toward production of North Atlantic deep water flow. Strong cooling must have occurred in Antarctica during the Middle to Late Eocene time (Fig. 2.2) bringing winter snow cover and mountain ice buildup between 52 and 36 Ma. By at least the early Oligocene (–36 Ma), an extensive, full-scale ice sheet had formed (Barrett et al., 1989; Webb, 1990). Some interpretations of the Antarctic glacial record suggest that major fluctuations of the ice sheet(s) may have occurred at intervals of as little as one to two million years over the last 36 Ma (Clapperton and Sugden, 1990; Scherer, 1991) and that the ice cover was probably more extensive than today. Significant enlargement of the East Antarctic Ice

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

23

TABLE 2.1. Timing of some Cenozoic global events related to glaciation (modified from Herman et al., 1989) Time (Millions of Years)

Tectonic events

Climatic events

0

0.9–0 Himalayan, Alpine, etc. orogenies underway

0.9

Orogeny peak

1.6

2–1 Increased Tibetan, Himalayan and Sierra Nevadan orogenies

0.9–0 Major glacial–interglacial cycles; onset of large-amplitude climatic fluctuations

2.4–2.5

First major Northern Hemisphere ice sheets and onset of marked global cooling

2.5–2.6

Fluctuating temperature

3.5

Uplift of Panama Isthmus; opening of Bering Strait

5.5

Isolation of Mediterranean Sea

Gradual temperature decline. Southern Hemisphere lowland glaciation

6

Strong global cooling. 6–4 Expansion of Antarctic Ice Sheet

14

14–12 major Antarctic ice buildup?

18

18–16 Subsidence of Iceland-Faeroes Ridge 18–14 Renewed Himalayan Tibetan orogeny

24

Drake Passage open

24–14 Increased glaciation of Antarctica; intensification of global climatic gradients

36.6

Greenland separates from Eurasia; Tasman Seaway opens

Antarctic continental glaciation; slow cooling

37

37–35 major Himalayan and Alpine orogenies

Cooling at high and low latitudes; mountain glaciation in Antarctica

Sheet did occur in Middle Miocene time following opening of the Drake Passage. Continued cooling allowed development of ice shelves in West Antarctica and finally the grounding of the West Antarctic Ice Sheet below sea level by 10 Ma. Appreciable retreat of Antarctic maximum ice cover occurred after several million years ~10 Ma. A warm Early Pliocene may have caused further deglaciation; however, a period of rapid uplift in the Transantarctic Mountains followed that may correspond with the start of Late Pliocene glacier readvances in the northern hemisphere about 2.5–2.4 Ma (Behrendt and Cooper, 1991). Elsewhere in the southern hemisphere, Miocene uplift and cooling by 4.7 Ma caused mountain glaciers or ice caps in the mid-latitude

Patagonian Andes to be more extensive than the present (Rabassa and Clapperton, 1990; Clapperton, 1993) (Table 2.1). In the northern hemisphere, direct evidence of growth of the Greenland Ice Sheet is unknown, but mountain glaciers were established by 10–2.7 Ma in the mountains of East and West Greenland. An ice sheet at least as big as today probably formed over the continent by 2.4 Ma (Funder, 1989). Small-scale glaciation is recorded by about 3 Ma in Iceland (McDougall and Wensink, 1966). Glaciation in Alaska may have developed during the same period in response to uplift that accompanied convergence of the Pacific and North American Plates ~6 Ma (Eyles and Eyles, 1989).

24

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

2.3. LATE CENOZOIC GLACIATION AND THE CLASSIC SUBDIVISIONS 2.3.1. Scope and Classification Discussion will now focus upon the global terrestrial and marine records of the present ice age. These records delineate glacial and interglacial intervals of 10 000 to 100 000 years, stadial and interstadial events within the glaciations that span 1000 to 10 000 years, and major fluctuations of mountain glaciers in postglacial and historic times involving decades to centuries to millennia. The time of onset of the present ice age may be fixed in a number of different ways. The continental glaciation in Antarctica began –36 Ma, whereas initiation of glaciation in middle latitude lowland areas of the northern hemisphere began around 2.5 Ma. It is now accepted that initiation of glacial conditions on a global scale apparently does not correspond with most definitions of any of the widely accepted Cenozoic chronostratigraphic terms (Van Couvering, 1997). Desnoyers (1829) was the first to apply the term Quaternary to post-Tertiary terrestrial strata and Reboul (1833) modified the application to include flora and fauna still living. Later, Lyell (1839) introduced Pleistocene (most recent) to marine units that contained more than 70 per cent of mollusca still

living. Forbes (1846) who, following acceptance of the Glacial Theory, modified Lyell’s definition to make ‘Pleistocene’ equivalent to the Glacial Epoch and suggested the term Recent for post-glacial time. ‘Recent’ was replaced by Holocene (wholly recent) by the International Geological Congress of 1885. At present, the internationally proposed Pliocene–Pleistocene boundary stratotype is a highly fossiliferous marine strata of the Vrica section in Calabria, southern Italy. This boundary is given an age of ~1.8 Ma (Van Couvering, 1997; Pasini and Colalongo, 1997). The Pleistocene–Holocene boundary is usually taken arbitrarily at 10 000 BP (Before Present) in North America and elsewhere (Watson and Wright, 1980; Menzies, 1995, chapter 8). 2.3.2. The Classic Subdivisions of the Pleistocene With the recognition of multiple sets of Pleistocene glacial deposits several regional schemes of classification emerged (Table 2.2). The Pleistocene classifications of the European Alps, northwestern Europe and central North America are well-known, longestablished and often still used (Chapter 15). However, each, although much modified, has been unsatisfactory and misleading (Bowen, 1978). Regional classification schemes were well established before development of the marine oxygen isotope records or the long terrestrial sequences. The best known

TABLE 2.2. Major glaciations and interglaciations of western Europe and North America following the classical systems. The Alpine sequence and the pre-Illinoian sequence of North America are not considered useful for stratigraphic correlations. Major glaciation began following the Cromerian temperate stage in Europe European Alps

Northwest Europe

Britain

North America

Wurm ¨

Weichsel

Devensian

Wisconsinan

R/W

Eem

Ipswichian

Sangamon

Wolstonian

Illinoian

Warthe Riss

Saale Drenthe

M/R

Holstein

Hoxnian

Yarmouth

Mindel

Elster

Anglian

Kansan

G/M

Cromerian

Cromerian

Aftonian

Gunz Interglacials in italics; glacials in bold.

Nebraskan

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

Pleistocene succession of glaciations and interglaciations was the four-part Alpine classification (Penck and Br¨uckner, 1909). This attained the status of a continental to worldwide framework for glacial sequences despite its lack of stratigraphic basis (e.g., local identification, description and correlation of rock units in sequence). 2.3.2.1. European Alpine terrace model of Penck and Br¨uckner The Alpine ice mass (Fig. 2.3) at its maximum, covered about 150 000 km2 flowing as a network among mountain peaks and divides, but with coalescing valley and piedmont glaciers that reached down toward sea level on the north and south flanks. In north-draining valleys of the German Alpine foreland south of Munich the G¨unz (G), Mindel (M), Riss (R) and W¨urm (W) glaciations, and the G/M, M/R and R/W interglaciations were delimited (Table 2.2). Named after four Bavarian tributaries of the Danube (Donau) and Isar, all have their type localities at the Iller-Lech Platte; however, no true stratotypes were designated (Kukla, 1977). Subsequently, the four-part sequence was extended by the discovery of two older glacial stages: the Donau and Biber. The four principal glacial stages were represented by a sequence of terraced glaciofluvial outwash plains (sch¨otter); each younger (and generally lower) plain was tied successively farther upslope to end moraines representing contemporaneous glaciation (Fig. 2.4). Interglacials were represented only by terracing of the previous outwash gravels to produce steps. A chronosequence was established based on the depth of step cutting between one terrace and the next, the M/R interglaciation was estimated to be about four

25

times longer than the other interglaciations. Furthermore, based on the interval of postglacial time, the whole Pleistocene was estimated to have lasted about 650 000 years. The timing of the four glaciations corresponded well with estimates of radiation minima at high to mid-latitudes calculated by Milankovitch. Problems with this classic four-part sequence are that the terrace sequence is more complicated than originally envisioned and contains both interglacial and postglacial deposits; the erosion intervals (unconformities between terraces) are largely glacial rather than interglacial; and the deposits may represent only a few millennia (Kukla, 1977). This classical sequence of glaciations must be abandoned for external correlations. 2.3.2.2. Northern European glacial chronology At its maximum, the Scandinavian Ice Sheet extended eastward to the Ural Mountains, southeast to beyond Kiev, south into central Germany and westward to the British Isles (Fig. 2.3). The fundamental origin of the stratigraphic subdivision of this ice sheet is morphostratigraphic based on a progressively younging end moraine system marking glaciation limits from southern central Europe northward into Scandinavia. The classic stages (Table 2.2) were named (from oldest) Elster, Saale and Weichsel after rivers on the northwestern European plain. Later, the Warthe stage was added, originally with the Weichsel, then the Saale, and now singled out as a distinct glacial stage between Saale and Weichsel. All stages are now known to fall into the Brunhes paleomagnetic chron (–0.9 Ma). However, some evidence of earlier Pleistocene glacial advances is reported

End moraines Sandur (outwash) field Marginal basin Drumlin FIG. 2.4. Alpine terraces and moraines of Penck and Br¨uckner (from Nilsson, 1983; reprinted with permission of Kluwer Academic Publishers).

26

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

˘ locally (e.g., Britain and Poland) (Sibrava, 1986; Bowen et al., 1988). The system of moraine-based units in northern Europe is supported by classical interglacial stages (from oldest): the pre-Elster, Cromer, the Holstein (Hoxnian in Britain); and the Eem (Ipswichian in Britain). The interglacial units are represented by deposits of marine transgressions in the lowland areas, and by peat bogs with pollen documenting temperate, hardwood forests in northwestern Europe. Significant problems relative to the northern European scheme relate to correlating the end moraines with interglacial deposits, repetitive glacial and interglacial deposits, the fragmentary nature of individual Pleistocene deposits, and the scarcity of distinguishing fossils in Pleistocene strata (Cepek, 1986; ˘ Sibrava, 1986; Zagwijn, 1986). Finally, reorganization is needed to account for deposits including those of pre-Elster age (Bowen et al., 1988). In the British Isles, several centres of glaciation existed during the Quaternary, with frequent linkage with the Scandinavian Ice Sheet as it invaded from across the North Sea. Consequently, the moraine sequence is less clear than in continental Europe. One of the classic areas of Quaternary study in the British Isles has been on the North Sea coast of East Anglia (Bowen et al., 1986a). 2.3.2.3. Central North American sequence The Laurentide Ice Sheet extended from the Arctic Ocean in the Canadian Arctic to the mid-west of the USA in the south and from the Canadian Rocky Mountains in the west to the Grand Banks off Newfoundland (Fig. 2.5a), possibly incorporating 35 per cent of the world’s ice volume during the Late Wisconsinan maximum and 60–70 per cent of the ice wasted at its conclusion (Fulton, 1989). Named after the states where they were best characterized or easily studied, the Nebraskan, Kansan, Illinoian and Wisconsinan glaciations represented the glacial sequence of the interior lowlands between the Appalachian and the Rocky Mountains. In many areas, their drift boundaries were thought nearly coincident (Fig. 2.5b); however, the Kansan is considered the most extensive in the northern mid-

USA obscuring the older Nebraskan drift in all but a few areas. The original bases of the earliest three glaciations were till sheets, while the Wisconsinan was once based on the little-modified ubiquitous end moraines. The interglaciations (Yarmouth and Sangamon) were represented by paleosols (buried soils) developed in tills or other deposits; and the Aftonian interglaciation by Kansan outwash gravel before a major soil horizon was later identified. The Aftonian, Yarmouth and Sangamon interglacial paleosols were believed to have formed during discrete soil-forming intervals corresponding to the respective interglacial. Estimates of their duration based on soil thicknesses and development were later shown to be unreliable. Data gathered from precise numerical dating and correlation methods, together with information from deep coring through the drift sheets suggested that a complete reorganization was necessary in Nebraska, Iowa and Missouri. In general, glacial and non-glacial deposits formerly assigned to the Kansan and Nebraskan glaciations, and Aftonian and Yarmouth interglacials, have often been miscorrelated and are of such diverse age as to make the terms almost meaningless. Therefore, the terms Nebraskan, Aftonian, Kansan and Yarmouth have been abandoned in stratigraphic nomenclature since the 1980s. Stratigraphic relations are relatively well known for the Illinoian (Late Middle Pleistocene) and in considerable detail for Wisconsinan Stage deposits. These latter sediments occurred following a major, and locally well-documented, shrinkage of the Laurentide Ice Sheet during the Sangamon interglacial. 2.4. OXYGEN ISOTOPE STRATIGRAPHY AND THE MARINE RECORD OF GLACIATION Knowledge of the character, magnitude and timing of Late Cenozoic glaciation has arisen through the development of new, detailed records throughout both glaciated and unglaciated areas, as well as from major modifications of the classical glacial–interglacial chronologies. For example, radiocarbon dating (Libby, 1955) provided an opportunity to focus on details of the latest Quaternary glacial fluctuations

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

(a)

27

North Pole A C

O

R T

C

IC N A

E

GREENLAND ICE SHEET

INNUITIAN ICE SHEET

CORDILLERAN ICE SHEET P O

A

C

C

E

A

IF N

LAURENTIDE

IC

ICE

SHEET

Calgary

Toronto

CORDILLERAN MOUNTAIN GLACIATIONS 0

New York

St. Louis

500 km

Cincinnati

Washington

C TI N A AN TL E A C O

(b) 0

400 km

C A N A D A

Paci fic Ocea n

U N I T E D

Atlantic Ocean

S T A T E S

Legend Area glaciated during Wisconsin Glacial Age

End moraines of earlier glacial ages

Additional area glaciated during earlier glacial ages

Glaciated area in Cordilleran region is not differentiated and is only approximate

Conspicuous end moraines of Wisconsin age

FIG. 2.5. (a) Approximate maximum extent of glaciation and main ice sheets in North America. Inner dashed lines at the southern boundary show generalized limits of Late Wisconsinan glaciation (modified from Flint, 1971 and Fulton, 1989). (b) Extent of glaciation in the northern USA and southern Canada (modified from Flint, 1971).

28

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

and to infer changes in continental ice volume and climates. At about the same time, similar global glacial and climatic inferences were made by taking advantage of the development in 1947 of deep sea piston cores. These allowed long undisturbed records of ocean bottom sediments, deposited at relatively constant rates over long time periods to be obtained (Menzies, 1995, chapter 14). Perhaps the main revolution in thinking of ice ages comes from the stratigraphy and resulting chronologies obtained from deep sea cores that involve the acquisition of the oxygen-isotope ratios (18O/16O) of micro-organisms, principally the calcareous planktonic and benthonic foraminifera. Cores provided a continuous time series of the worldwide variations in continental ice volume for the Late Cenozoic that generally cannot be duplicated by the more fragmentary and shorter continental records. Because the mixing time of oceans is relatively short (on the order of millennia) the trends in isotopic ratio variations expressed in the deep sea cores are, globally, nearly synchronous; in addition, they are easily correlated on the basis of unique stratigraphic horizons (Prell et al., 1986). Therefore, they have been used as a standard against which both continental and other marine chronologies are measured (Fig. 2.6). Instead of the four classical major glaciations differentiated in central North America or Europe through the mid to late twentieth century (Table 2.2), these deep sea isotopic records indicate that as many as 16 or more major ice advances and retreats occurred during the Pleistocene and an even greater number when the record is extended into the Pliocene (Andrews, 1997). The patterns of isotopic change have also provided convincing support for the astronomical theories of Croll and Milankovitch as describing the principal ‘pacemaker’ in global climatic changes. This implies that the orbital factors set the phase and frequency of climatic changes, if not also driving them. The technique of ␦18O with its adaptation to defining worldwide glaciations was pioneered by Emiliani (1955) on Caribbean cores and based on the isotopic fractionation during crystallization of marine microfaunal tests. The size of ice sheets on surrounding continents may also be estimated from deep sea cores by

measurement of the input of ice-rafted continental debris, the estimation of sea surface temperature (SST) determined from cores using biological transformations and the CaCO3 per cent controlled by the depth of dissolution (Rea and Leinen, 1989). 2.4.1. Principles of Oxygen-Isotope Analysis (OIS) Oxygen occurs in three isotopic forms in nature: 16O (99.763 per cent), 17O (0.037 per cent) and 18O, the heavy element (0.200 per cent). The stratigraphic and correlation technique of the oxygen isotopes in deep sea cores is based on the fact that the proportions of these isotopes in sea water have changed through time. It also rests on the premise that the calcareous microfauna (largely foraminifera), whose tests collect on the sea floor, have incorporated the two major isotopes of oxygen (18O and 16O), in proportion to the isotopic abundance in the surrounding water. This isotopic abundance is dependent largely on the amount of water stored in glaciers on land; however, it is also dependent on sea water temperature and to a minor degree on the water’s salinity. For a detailed discussion on the principles of oxygen isotopes see Mix (1987), Ruddiman, (1987) and Wilson et al. (2000). 2.4.2. Timescale, Milankovitch Controls and Phase Lags of the Deep Sea ␦18O Marine Chronology Stages in the cores from the world’s oceans are numbered from 1, for the present (Holocene) warm stage, backward in time, with colder stages (glacials or stadials) given even numbers and warm stages (interglacials or interstadials) odd numbers. Boundaries are placed at the mid-point between temperature maxima and minima (Fig. 2.6). The timescale for marine ␦18O records was initially adjusted to numerical time, based on dated paleomagnetic reversal boundaries and assumed constant sedimentation rates as well as correlations with coral terraces dated with uranium-series isotopes (Broecker and van Donk, 1970). However, the close correspondence of orbital periods in ␦18O data has

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

29

(a) Timescale Chron Sub-Chron 5.5 Ma 0.0

Site 607 18O (‰) 4.5

3.5

Site 607 (–)

(b)

2.5

18O 5.5

4.0

2.5

0.0

Brunhes Normal

0.1

(c)

Age (Myr)

0.2

20

Pleistocene

0.5

3.0 0.4

0.5

1.5

0.6

40

Jaramillo

0.0 0.00

0.7

1.0

Matuyama Reversed

1.5

80

Matuyama 0.734 - 1.6 Myr

Composite Depth (m)

60

0.06

Brunhes 0.0 - 0.734 Myr

SPECMAP

Olduvai

Site 607 18O

0.3

Site 552A

(d)

18O (‰)

%CaCO3 0

2.0

50

100

5.5

4.0

2.5

1.4

Reunion 1.8

2.2

2.5

Olduvai

2.0

100 Age (Ma)

Pliocene

1.6

MATUYAMA

2.4 2.6

GAUSS

Gauss Normal

2.8

120

3.0

Keena

3.2 3.4

GILBERT

3.0

3.6

FIG. 2.6. Oxygen isotope record for the North Atlantic over the last 3.5 Ma. Deep Sea Drilling Project (DSDP) composite cores for site 607 (41° 00⬘N, 32° 58⬘W) with: (a) isotope stages against magnetic timescale (modified after Raymo et al., 1989 and Ruddiman et al., 1989, reprinted with permission of the American Geophysical Union). (b) Brunhes Chron portion overlaid onto the SPECMAP ␦18O stack (dashed line of Imbrie et al., 1984) (from Ruddiman et al., 1989; reprinted with permission of the American Geophysical Union). (c) Spectral analysis of Brunhes and Matuyama portions of ␦18O record of site 607 (from Ruddiman et al., 1989; reprinted with permission of the American Geophysical Union). Note strong obliquity (41) cycle for early record compared with eccentricity (100) and precession (23 and 19) cycles for Brunhes. (d) North Atlantic DSDP site 552 records show major (about 30 per cent) increase in northern hemisphere ice sheets at 2.4 Ma (about Stage 100) (data of Shackleton et al., 1984; reproduced from Ruddiman and Wright, 1987; reprinted with permission from the authors). Abrupt decrease of CaCO3 per cent values mark initiation of deposition of ice-rafted sand.

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

July Insolation Isolation (65ºN) July (65°N) (1033 cal/cm cal/cm22day) day) (10

d18O - d16OTOO (‰)

0.8

0

0

0.9

1.0

1

24

I

100

II

200

III

300

IV

400

Terminations

Age (x103 yr B.P.)

2

Scaled Variance

30

100

16

8 41

23

V VI

500

19 0

80

40

20

Period (103 yr/cycle) 600

VII

(a)

(b)

(c)

FIG. 2.7. Comparison of Milankovitch July insolation record for 65°N (a) with composite oxygen-isotope record (SPECMAP) from Imbrie et al. (1984) (reprinted with permission of Kluwer Academic Publishers), (b) showing ␦18O spectral analysis (from Broecker and Denton, 1989, reprinted from Geochimica et Cosmochimica Acta, 53, 2465–2501, 1989, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK). Horizontal dashed lines mark the termination while diagonal dashed lines suggest the periods of gradual ice buildup. (c) Scaled variance per cycles.

meant that timescales have now been adjusted by matching to orbital variations. This correspondence was achieved for particular latitudes and seasons using insolation curves and averaging of multiple core records from the world’s oceans that are matched to curves showing variations in obliquity and precession (Rea and Leinen, 1989) (Figs 2.6b, 2.7, 2.8a). Not all fluctuations and amplitudes detected in core signals can be resolved by orbital parameters. Large climatic changes are commonly detected in the records and frequently occur too rapidly to be explained by orbital forcing. Typical frequencies noted within cores (Fig. 2.7c) are: (1) a 100 000-year rhythm similar to the periodicity of orbital eccentricity that increases in importance from the Pliocene and dominates the Brunhes Magnetic Chron; (2) a 41 000-year period of obliquity (tilt) that dominates in the Late Pliocene

through the Matuyama Chron; and (3) weaker 23 000- and 19 000-year signals corresponding to precessional periodicity that influences mid-latitude insolation levels. The ␦18O, ␦13C, CaCO3 and the sea surface temperature (SST) all appear to exhibit a coherent, in-phase correlative response to these frequencies in cores. Ice sheets appear to have responded to orbital forcing (Ruddiman et al., 1989). In correlating data from deep sea cores, phase lags play an important role. The ␦18O signal, for example, lags behind terrestrial ice volume change by ~500 years during ice sheet shrinkage and by ~3000 years during ice sheet growth. In the marine cores the ␦18O signal lags insolation, forcing a change at the 41 000-year obliquity by ~8000–10 000 years; and at the 19 000 and 23 000 precessional periods a lag of ~5000–6000 years (Ruddiman, 1987).

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

31

FIG. 2.8. (a) Sea level curve for deglaciation from the last Pleistocene maximum (Late Wisconsinan) when the level was 121±5 m below present level. Curve is scaled to ␦18O using the calibration of 0.011 per cent per metre change in sea level to provide an upper limit for mean ice volume for global deglaciation (from Fairbanks, 1989; reprinted with permission from Nature, 342, 637–642, 1989, Macmillan Magazines Limited). (b) Rate of glacial meltwater discharge (solid curve) calculated from the Barbados sea level curve (a) compared with summer insolation (dashed and dotted) for two latitudes. The faint curve shows correction for atmospheric 14C changes (from Fairbanks, 1989; reprinted with permission from Nature, 342, 637–642, 1989, Macmillan Magazines Limited). (c) AMS 14C-dated oxygen isotope record for the deglacial interval starting 15 000 BP (from Fairbanks, 1989; reprinted with permission from Nature, 342, 637–642, 1989, Macmillan Magazines Ltd.). Meltwater pulses spanning the Younger Dryas interval are indicated.

32

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

2.4.3. Interpretation of Late Cenozoic Glaciation from the Deep Sea Records 2.4.3.1. The overall record Initiation of continental ice sheets may have occurred in North America and Europe by ~3.0–3.1 Ma (Fig. 2.6). At about 2.4 Ma (OIS 100), ␦18O values increased abruptly and CaCO3 suddenly decreased, indicating that ice sheets may have increased from one-quarter to one-half as large as those of the Late Pleistocene. Also sufficient ice build-up had occurred to initiate the first major episode of marine ice rafting of terrigenous debris into areas south of 40°N (Raymo et al., 1989). The ice sheets existing after 2.4 Ma (Matuyama Chron) fluctuated at least 40 full climatic cycles with a periodicity of ~41 000 years. By 0.9–0.6 Ma, signals indicate a doubling effect, such that both the North American and European ice sheets had probably doubled in size (Ruddiman et al., 1989). 2.4.3.2. Details of the last glacial–interglacial cycle and comparison with data from uplifted coral reef terraces The ␦18O record for the last 130 000 years has been divided into five stages and, in turn, into substages 5e to 5a or 5.5 to 5.1. OIS 5e has been correlated with the Eemian interglacial in northern Europe and the Sangamon in North America. Terrace studies and deep sea data show a temperature change at OIS 5e with sea level 6 m above present (Shackelton, 1969, 1987). Subsequent isotopic events (OIS 4 to 1) mainly represent ice volume changes on the continents. The last global glacial maximum occurred during OIS 2 and was followed by rapid deglaciation (Termination I) to the Holocene. A detailed examination of those cores with high sedimentation rates or with stacked (averaged) sets of well-dated cores over the last 20 000 years (Mix, 1987) reveals ice volume maxima during the last glaciation from 20 000 to 16 000 BP followed by a major deglaciation and a rapid decrease in ␦18O at about 14 000 to 12 000 BP. A second termination occurred between about 12 000 and 9000 BP (Fig. 2.8). These latter periods straddle an interval of

cooling in the North Atlantic and are perhaps elsewhere known in European pollen zone terminology as the Younger Dryas, named after the arctic flower Dryas octapetala in deposits near Aller¨od, Denmark. The first, almost continuous sea level curve constructed for deglaciation has been obtained from borings in uplifted interglacial coral reef terraces off Barbados (Fig. 2.8) (Fairbanks, 1989). This curve indicates a sea level 121±5 m below present during the last glacial maximum. Furthermore, the steep parts of this curve, which represent unusually rapid melting and sea level rise, appear to correspond with the ␦18O steps considered above and, in turn, the peak of Milankovitch summer insolation for 60°N at 11 000 BP. The data of Hanebuth et al. (2000) support this work. 2.4.3.3. Limitations of the Deep Sea Chronology The resolution of the oceanic record is limited by the normal slow pelagic sedimentation rates (1–4 cm ka–1 ), the bioturbation of the bottom sediments and the rate of ocean water circulation. It may take 1000–1500 years for complete ocean mixing. The lags inherent in the orbital, ice sheet and oceanic records lead to inaccuracies and, in addition, the isotopic record has its deficiencies, the most important of which is the unknown contribution of temperature change versus ice volume to down-core ␦18O records.

2.5. LONG TERRESTRIAL RECORDS OF CLIMATE AND GLACIAL FLUCTUATIONS FOR THE QUATERNARY 2.5.1. Loess (L¨oss), Pollen and Lacustrine Records Loess deposition is largely a Quaternary phenomenon (Menzies, 1995, chapter 6): sequences of alternating loess and dark soils that occur, for example, in central Europe are particularly well known and closely correlated with glacial fluctuations. Very extensive central Asian loess deposits, however, furnish more complete terrestrial paleoclimatic records for the Late

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

Climate Index

PMG 0 0

100

200

300

33

Units 400 HOLOCENE

ELEVTHEROUPOLIS

100

200

DRAMA PANGAION

SYMVOLON STRYMON

300

KAVALA KRIMENES

Time 103Y

LITHOCHORIS

400

LEKANIS

500

ALISTRATI I

ALISTRATI II

600

FALAKRON

NIKI

700

POLISTILOS

PROSOLSANI

800 KALAMONAS

900

FIG. 2.9. Climate changes expressed by an index based on oak, pine and total tree pollen records obtained by Wijmstra et al. from the Tenaghi-Phillippan peat bog in Greece (reproduced from Kukla, 1989, with permission of Elsevier Science Publishers, Amsterdam). PMG is paleomagnetic chronology with normal polarity in black.

34

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

Cenozoic. In addition, several long lacustrine sequences with their associated sediments and pollen records extend well back into the Pleistocene and show promise for the future. These types of records are dated using methods such as faunal changes, carbon– 14, fission track-tephrochronology, thermoluminescence and paleomagnetic measurements; in addition, many of the records can be correlated directly with the glacial–interglacial ␦18O changes of deep sea records (Kukla, 1989) providing details of the Middle Pleistocene. Loess is a silt, transported and deposited by wind in unconsolidated and unstratified accumulations that are usually loosely cemented by syngenetic carbonate. At sites in Luochuan, Xifeng and Xian, central China, a record of worldwide climatic oscillations, comparable with that of deep sea cores, is displayed in a sequence of alternating soil horizons and loess totalling over 200 m. This stratigraphic sequence extends back over the last 2.5 Ma. Pollen from a continuous 280-m-long peat core from Tenaghi-Philippan, Greece, reveals cyclical vegetation and climatic changes from open steppe to closed hardwood forest over the last 900 000 years (Kukla, 1989). Variations in the ratios or abundance of arboreal pollen to that of grass and herb pollen, as well as that in the oak to pine pollen (Fig. 2.9), allowed differentiation of cycles comparable with the last 25 oxygen isotope stages, from dry cool intervals of open steppe to relatively warmer times under hardwood forests. This sequence shows frequent large amplitude climate oscillations lasting a few millennia throughout the core. Long lacustrine cores come from the intermontane basins of the tropical Andes and southern Japan, and from the subsiding rift basins of the Dead Sea or northern California.

these similarities, the Devils Hole record is thought to reflect global climate. However, locally, the pattern of changes may be driven by winter–spring land surface temperatures in the southern Great Basin that yield isotopic variations in the local precipitation. 2.6. THE LATE PLEISTOCENE CLIMATIC RECORD REVEALED BY DEEP ICE CORES 2.6.1. General A valuable and continuous record of Late Pleistocene paleoclimate and paleoenvironmental changes, especially during the last glacial cycle, can be obtained

Devil’s Hole SPECMAP DSDP-609 0 0 100 1 Time (kyr B.P.) 2

0

A continuous 500 000 year ␦18O climate record (Fig. 2.10A) has been obtained from a 36-cm-long core of calcite precipitated along an open fault zone at Devils Hole, Nevada (Winograd et al., 1992, 1997). The record ‘mimics’ major features displayed in marine ␦18O, ice-volume records and ice cores from Greenland and particularly the four glacial–interglacials of the long Vostok core from Antarctica. Because of

50 %

4 5a 5b

100

5c 5d

150

6

200

7a 7b

5e1 5e3 5e5

5e2 5e4

5e

c2

7c 7d

a

2.5.2. Vein Calcite Record at Devils Hole, Nevada, USA

V27-116

3

50

Summit

b

c1

7e

d

250 14

15

0.5 0 -0.5

-40

-35

18O (‰)

(a) FIG. 2.10. (A) Climate records plotted to a common linear timescale (modified from Dansgaard et al., 1993). (a) ␦18O variation in vein calcite, Devils Hole, Nevada (Winograd et al., 1992). (b) Orbitally tuned SPECMAP ␦18O curve (Martinson et al., 1987). (c) Part 1, grey-scale measurements along marine core at site DSDP 609; part 2, per cent CaCO3 in core V27–116 from WGW+W of Ireland. (d) ␦18O along upper 2982 m of the GRIP ice core, Greenland. (Reprinted with permission from Dansgaard et al., 1993, Nature, 364, 218–220, 1993, Macmillan Magazines Ltd.).

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

Dye 3 1694 - 2036 m

35

Camp Century 1081 - 1387 m

Dome C 0 - 905 m

Byrd 1238 - 2163 m

10 8230 ± 50 years B.P.

Indian Ocean

Age

( Hays, Imbrie & Shackelton )

0

9 8

Emiliani Stages

20 2

6

40

5 3

4

60 3

-55

4

2

-50 ‰

80 5a

1

5b

-40

-35 ‰

100 5c 5d

-35

-30

120

-25 ‰ 5e

 18O)

x 103 years B.P. -40

-35

-30

3

2‰  O 18

(a)

(c)

(b)

(d)

(b) FIG. 2.10. (B) ␦ O profiles along the five deep ice cores from Greenland (Dye 3 (a) and Camp Century (b)) and Antarctica (Byrd (c) and Dome C (d)) plotted on linear depth scales for intervals indicated (modified from Dansgaard, 1987; reprinted with permission of Kluwer Academic Publishers). Ages on Greeland core arrows and Byrd and Dome C, Antarctica, logs are added from Paterson and Hammer (1987). The Late Wisconsinan maximum is estimated to be at a depth of 1854 m in the Dye 3 core where the ␦18O is at a minimum and the 10Be and dust reach their highest concentrations. Planktonic foraminiferal ␦18O profile from Hays et al. (1976) has been adjusted to Camp Century profile (now believed to end within Stage 5e) as plotted by Dansgaard et al. (1982) (reprinted with permission from the authors and Science, 218, 1273–1277, 1982, copyright 1982 by the American Association for the Advancement of Science). 18

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

12 10 8 6

50

100

150

200

250

a 0 -2 -4 -6

Insolation (W m-2)

b 50

c

0

d

d18 OSW (‰)

-50

PLATE 2.1. Thin section of an ice core sample taken from 160 m below snow surface at Camp Milcent, Greenland. Section is photographed under polarized light, showing the irregular shape of individual ice crystals after nearly 300 years of transformation from snow flakes to a polycrystalline aggregate. The diameters of the large crystals are approximately 1 cm. The colours depict the different orientations of individual crystals (photo courtesy of Chester C. Langway Jr).

0

0

7.1

e

5.5

6

?

1

7.3 7.5

7.2 7.4

1

300

f

250

200

CH4 (Vostok) (p.p.b.v.)

from ice cores (Dansgaard et al., 1993) taken through both polar (high latitude) ice sheets and mountain glaciers (Plate 2.1). The longest and most useful cores are obtained from the polar ice sheets where there is no surface melting and ice temperatures are usually below the pressure melting point (Fig. 2.10A, B). A stratigraphic record encompassing ~150 000 years through the penultimate interglacial has also been extracted from the surface ablation zone along the outer edge of the ice sheet (where older ice emerges) in central west Greenland near Pakitsaq. Long cores have been obtained from Devon Island Ice Cap, Canada; Camp Century, Renland, Dye 3, Greenland Ice Sheet Project 2 (GISP2) and GRIP sites in Greenland; and Dome C, Byrd, Vostok and Taylor Dome sites in Antarctica (Figs. 2.10A, B; 2.11; 2.12). All of these extend back chronologically into the Late Glacial Pleistocene maximum and beyond (Jouzel et al., 1993). The larger diameter core (3053 m in depth) of the GISP2 Project at Summit reached 1.5 m into bedrock below the ice sheet (Grootes et al., personal communication, 1993). At the summit divide of the central Greenland Ice Sheet 20 km away is the 3028-m GRIP core (Fig. 2.12). The undisturbed part of each core covers the past ~105 000 years. Drilling at the North GRIP site may extend the Greenland record at least beyond 135 000 years (Stauffer, 1999).

DT (°C)

2

␦d1818OO (‰) (‰)

SST (°C)

0

CO (Vostok) (p.p.m.v.)

36

700

g

600 500 400 300 0

50

100 Age (kyr

150

200

250

BP)

FIG. 2.11. Vostock ice core records and correlative marine and orbital variations (modified after Jouzel et al., 1993, reprinted with permission from Nature, 364, 407–412, 1993, Macmillan Magazines Ltd.). (a) Summer sea surface temperature at Indian Ocean site; (b) atmospheric temperature change at Vostock derived from the deuterium data of cores 3G and 4G; (c) summer insolation at 20°N; (d) Vostock ␦18O profile; (e) isotope change in sea water from marine V19–30 ␦18O record; (f,g) Vostock CO2 and CH4 profiles, respectively.

The last of three cores obtained at Vostok reaches 3623 m in depth (Fig. 2.11). The upper 3311 m of this core appears undisturbed and represents the last 420 000 years (Petit et al., 1999). The extraction of paleoclimatic data from ice cores has problems;

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION -36

0

-34 ‰ 10

Depth (m)

11 12

1500 Time (kyr BP)

14

1

1 Bølling

16

IS number

18 20

2 500 3

2 3 4

30

5 6 7 Denekamp 8 9 10 11 12 Hengelo 13

40

4

50

5 1000

14 15 16 17 18

60 6 7

1500

9

Glinde Oerel

2500

70

19 20

80

21 Odderade 22

100

(a)

2000

25

35

8

Depth (m)

a

120

b

23 Brørup 24

Eem

150 Saale

(b)

200 Holstein 250

-45

-40

-35

3000 -30

d18O (‰)

FIG. 2.12. Continuous GRIP ␦18O record from Summit, Greenland, plotted with 2.2-m increments in two sections on a linear depth scale (modified after Dansgaard et al., 1993). (a) From surface to 1500 m; (b) from 1500 m to 300 m depth. IS, interstadials with longer ones assigned European pollen horizons (reprinted with permission from Nature, 364, 218–220, 1993, Macmillan Magazines Ltd.).

however, these records provide better resolution of short-term change over the last glacial cycle than do deep sea cores (Paterson and Hammer, 1987). 2.6.2. Dating of Ice Core Profiles The most precise method of obtaining age–depth relationships is by counting seasonal changes distinguished by ␦18O, dust particles or acidity (Dansgaard et al., 1993). This is undertaken where accumulation rates are greater than 25 cm a–1 and at some height above the basal layers where diffusion may

37

make counting impossible. The chronology of the GRIP core (Fig. 2.12) is determined by counting annual layers back 14 500 years and, beyond this date, by ice flow modelling. Annual layers have been counted back almost 80 000 years in the GISP2 core (P. Mayewski, personal communication, 1993). Beyond the Holocene or latest Pleistocene in the case of GISP2, less precise methods are used such as: (1) matching reference horizons of known age, for example, in the Byrd core by volcanic acidity maxima (Hammer, 1989); (2) radioactive dating of carbon–14 from CO2 bubbles; (3) by 36Cl/10Be ratios or, for a shorter term, 32Si or 210Pb; (4) matching of the ␦s record with another dated climatic record (principally deep sea records); and (5) by the less precise method of ice-flow models. Ice-flow models are used in the lower regions of most polar ice cores and for the entire length of the two last Vostok cores, where accumulation rates are too low for annual counting (Jouzel et al., 1993; Petit et al., 1999). 2.6.3. Summary of Major Environmental Changes Revealed by Ice Cores and Possible Causes 2.6.3.1. The overall record The ice core ␦ records show qualitatively the same general features as do the deep sea ␦ records over the last glacial cycle and some important differences during earlier intervals (Fig. 2.12). The ice core timescales are independent of orbital tuning used in the marine cores; thus they may provide a valuable check on changes of environmental parameters obtained from cores in adjoining seas. These environmental parameters (e.g., atmospheric gases, dust, temperature proxy) appear to have frequently shifted rather simultaneously. The long Vostok record shows similar, but slightly different climatic and atmospheric trends across the four cycles (Petit et al., 1999). Temperature variations estimated from ␦D for the last two glacial periods were similar, but the cycles of the earlier two are of shorter duration. This is also true for marine cores. Nevertheless, minimum temperatures are very similar, as are interglacial maxima in the ice record; data for interglacial OIS 11.3 are uncertain, while deep sea cores suggest a particularly warm

38

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

climate for this interval (Howard, 1997). The Holocene, 11 000 years in duration, is shown to be, by a notable margin, the longest stable warm period in Antarctica over the past 420 000 years. For example, interglacials of OIS 5.5 and 9.3 show ~4000 years of warmth followed by relatively rapid cooling and then slower cooling. Interglacial OIS 7.5 is even more ‘spiky’ in shape. These interglacial span ~17 000 to 20 000 to 14 000 years, respectively. The greater length of interglacials shown by the Vostok data (compared with those of marine cores) is also suggested by the Devils Hole terrestrial record where these warm intervals average ~22 000 years in length (Winograd et al., 1997). The record of Vostok atmospheric greenhouse gases shows that present levels of CO2 and CH4 (~360 ppmv and ~1700 ppbv, respectively) are unprecedented over the past 420 000 years; preindustrial levels average ~280 ppmv and 650 ppbv, respectively. The close correlation of these gas concentrations with Antarctic temperature fluctuations is probably related to their contribution to glacial–interglacial changes (Petit et al., 1999). Greenland ice cores, including GRIP and GISP2 (Alley et al., 1993) and previous long Greenland records show abrupt mode switches in the isotope, dust and gas records (Figs. 2.10A, B). The switches are only weakly shown in Antarctic cores (Jouzel et al., 1987). The switches, called ‘interstadials’ (Johnsen et al., 1992) or ‘Dansgaard–Oeschger’ cycles, take about 50 years representing ␦18O changes of 2‰ and –3‰ or temperature shifts of ~5–7°C (Dansgaard and Oeschger, 1989). These changes occur every 500 to 2000 years, at least between 40 000 and 20 000 years ago. The Younger Dryas cooling beginning ~13 000 years ago represents the most recent of these major cycles. Furthermore, electrical conductivity studies of the GISP2 records show fluctuations (‘flickers’) (Taylor et al., 1993) of dust on scales of less than 5–20 years within and between these longer term, warm–cold Dansgaard–Oeschger cycles. 2.6.3.2. The last deglaciation The ␦18O minimum (Late Wisconsinan maximum), dated in Greenland ice cores at approximately 18 000 BP and in Antarctica at about 20 000 BP (Fig.

2.10) was followed by a general warming trend until about 17 000 BP when Greenland cores show major temperature drops. These represent the last Dansgaard–Oeschger ␦s cycles and are closely correlative with northern European pollen zones. The Older Dryas, the first of these coolings, ended with a sudden increase ␦s and temperature at about 14 700 BP that correlates with the Bølling–Aller¨od warming (Fig. 2.12). The Younger Dryas interrupted this warming about 13 000 calendar years ago with a temperature drop to full glacial conditions that lasted until 11 500 BP. The shifts to warm intervals following the Older and Younger Dryas took place in about 50 years according to isotope data, but as fast as 3–5 years based upon the dust profiles. The end of the Younger Dryas was marked by a temperature rise of about 7°C and a doubling of snow accumulation in Greenland (Alley et al., 1993). Antarctic cores show more muted reversals during these and earlier intervals (Jouzel et al., 1987). The Younger Dryas and other ‘coolings’, extending back through at least several hundred thousand years appear to correspond with one of many aperiodic intervals of prolific rafting of sediment iceberg across the North Atlantic Ocean. Called ‘Heinrich events’ (Heinrich, 1988; Andrews, 1998), these occurred at intervals of about 12 000 years over at least the last 250 000 years. Most ‘events’ are assigned to episodes of surging, or collapse of the Laurentide Ice Sheet in Hudson Strait (Andrews, 1997, 1998). The instability of the ice sheet may, in turn, be triggered by Dansgaard–Oescheger cooling events shown by the Greenland cores (Alley et al., 1999). Whether the Pleistocene–Holocene transition occurred as a result of rapid climatic jumps, or more gradually and uniformly, typical Holocene temperatures were established by about 9000 BP simultaneously in Greenland and Antarctica. The total shift in ␦s during the transition in Greenland cores varied from 7‰ to 11‰ representing a local warming of 10.5–16.4°C (Dansgaard and Oeschger, 1989), somewhat higher than in Antarctica. Coincident with the drop in ␦18O was a general increase in CO2 from Pleistocene values. By itself this fluctuation in CO2 would have caused a significant change in the global radiation balance and temperature rise (Oeschger et al., 1984).

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

Shorter-term climatic fluctuations are more difficult to extract from ␦18O ice core records than are the large amplitude fluctuations partly as a result of large variations between drill sites. Nevertheless, ice-core data from some Greenland cores suggest that Holocene ice surface temperatures reached a long-term maximum 1–2°C above present temperatures between about 6000 and 3000 BP (Dansgaard and Oeschger, 1989). The elevated snow accumulation shown in the GISP2 core between 9200 and 7300 BP suggests that this may have been the Holocene ‘climatic optimum’ or Hypsithermal warm interval. General cooling since the Hypsithermal culminated between about 1200 and 1900 AD in the most pronounced Holocene advance of mountain glaciers known as the ‘Little Ice Age’ (Grove, 1988). Ice cores from the tropical Quelccaya Ice Cap suggest that the Little Ice Age occurred in South America between 1490 and 1880 AD with onset and termination within a few decades (Thompson et al., 1986). 2.7. CORRELATION OF LATE CENOZOIC GLACIATIONS IN THE NORTHERN AND SOUTHERN HEMISPHERES Figure 2.13 represents an attempted correlation and dating of Quaternary ice sheet and mountain glaciations across Eurasia, Europe and North America ˘ (Bowen et al., 1986b; Sibrava et al., 1986). The primary basis for correlation is the even-numbered oxygen isotope stages of deep sea cores ascribed to individual glaciations. The primary indicators used for glacial advance and retreat in northern hemisphere correlations were tills or proglacial deposits. Secondary proxy indicators of ‘cold’ periods were not used in the chart. Glacial advances in the European Alps were based on the river terraces since they are tied to glacial deposits. Each glaciation is indicated by a wedge for which the bottom line indicates advance and the top line retreat. Wedges, generally, are not proportional in size to glaciation magnitude, except for latest Pleistocene with Holocene maxima shown for Canada. Chronological control from dating of lava and tephra by potassium–argon and fission-track methods allowed bracketing or limiting ages and reliable independent chronometric control in parts of the

39

USA. The last 50 000 years of the Pleistocene may be the most reliable, especially where 14C ages are applicable in the central Great Lakes area. In Europe, some chronometric methods such as 14C and particularly thermal luminescence have been applied in correlation. In correlation, synchroneity of glacial events is assumed, although this is problematic (Richmond and Fullerton, 1986a). 2.7.1. The North American Record In North America, several pre-Illinoian glaciations of the Laurentide and Cordilleran Ice Sheets, as well as of mountain glaciers, have been inferred (Richmond and Fullerton, 1986b). Major mountain glacier advances occurred by Miocene time in Alaska (Hamilton, 1994) and probable Late Pliocene mountain glaciers are recorded in the Rocky Mountains of Wyoming and the Cascades of Washington, USA (Fig. 2.13, glaciations L, J and/or K). The Cordilleran Continental Ice Sheet, a mountain ice complex that spread over the north–south trending mountain ranges west of the Canadian plains, formed the Puget Sound lobe (Washington) by Late Pliocene or Early Pleistocene time. Evidence based on tills indicates that the Laurentide Ice Sheet advanced by the Late Pliocene into Iowa. The till displays a paleomagnetic reversal that can be assigned to Glaciation K (Fig. 2.13). A possible correlative till occurs in western Wisconsin (Matsch and Schneider, 1986). Glacio-fluvial deposits in the Canadian prairies are considered Late Pliocene or Early Pleistocene in age (Fulton et al., 1986). During the interval 1.5–0.9 Ma, no persuasive evidence exists of glaciation in the USA, but mountain glaciers in the southern hemisphere may have reached their greatest extent during this period. Beginning at OIS 22, glaciations have been inferred to have occurred in the mountainous areas of western North America and in areas influenced by the Laurentide and Cordilleran Ice Sheets. Late Middle Pleistocene (Illinoian) glaciation is recorded by an early, as well as a two-pronged, late glaciation; a relatively well-developed soil marks the non-glacial interval in the type area of Illinois. The Sangamon paleosol, which is markedly diachronous, developed locally on these Illinoian units in the mid-western

40

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

FIG. 2.13. Correlation of Late Cenozoic glaciation in the northern hemisphere (from Bowen et al., 1986; reprinted from Quaternary Science Reviews, 5, Chart 1, 1986, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

FIG. 2.13. Continued

41

42

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

USA. An extensive terrestrial sediment record illustrating the Sangamon interglacial is found at sites along 1600 km of the southwest margin of Hudson Bay, Canada (Dredge and Cowan, 1989). These sediments may represent an early glacial lake formed by the receding Laurentide Ice Sheet margin, a highlevel sea that later regressed to levels similar to or below present, or a subaerial interval of deposition and a possible succeeding proglacial lake that preceded ice margin readvance over the area. Early Wisconsinan advances are generally considered to be less extensive than those of the Late Wisconsinan, but a considerable increase in global ice volume did occur during OIS 4. Advances along the southern margin of the Laurentide Ice Sheet are recorded from Michigan eastward into New England; however, some of these advances may be of older or younger age because dating is insecure. During the Middle Wisconsinan, the Cordilleran Ice Sheet may have disappeared; the southern margin of the Laurentide Ice Sheet was restricted generally to Quebec, eastern Ontario and possibly north-central New York State (Dreimanis, 1991). At least partial deglaciation and marine invasion of Hudson Strait and Hudson Bay had also occurred (Andrews, 1997). The Wisconsinan maximum of the Laurentide Ice Sheet at its southern margin in the USA, occurred between 24 000 and 14 000 BP (OIS 2); retreat from the northern and eastern margins in Canada occurred between 12 000 and 8000 BP. The mountain glacier maximum in the USA may have occurred before ca. 22 000 BP and before its record in deep sea cores by ca. 18 000 BP. The maximum extent of the Cordilleran Ice Sheet occurred 15 000–14 000 BP (Andrews, 1997). 2.7.2. The European Record Pliocene continental glaciation is generally unknown in Europe or Eurasia except for the Alps (Ehlers, 1996). Pollen data record the distinctive cooling at or near the Gauss–Matuyama paleomagnetic boundary in many areas of Europe (Zagwijn, 1986; Velichko and Faustova, 1986). The oldest loess sheets of the Alpine foreland have ages up to 2.5 Ma. These events correspond generally with major northern hemisphere ice rafting and glaciation indicated by deep sea cores.

Evidence of an Early Pleistocene advance by the Scandinavian Ice Sheet is sparse except possibly for two advances recorded in west-central Poland (Rzechowski, 1986) and adjacent Commonwealth of Independent States (Velichko and Faustova, 1986) (Fig. 2.13). Early Middle Pleistocene glaciations are reported from Britain, Poland, Belarus, Russia and the Ukraine. A lobe of the Scandinavian Ice Sheet is reported to have advanced into the Don River basin in the Central Russian Plain to at least 52°N at about this time (Arkhipov et al., 1986). Widespread advances across northern Europe and Eurasia may not have occurred until Elster time, correlated with OIS 14 and/or 12 (Fig. 2.13). Middle and Late Pleistocene glaciations show strong correlations with respective events in North America. For example, they display (1) weak mid-Holstein, OIS 10, advances; (2) a tri-partite Saale as in the Illinoian; (3) a weak early Weichsel as in the Wisconsinan; and (4) a strong Late Weichsel glacial maxima at about 20 000 BP. 2.7.3. The Southern Hemisphere Record Large glaciers have existed in Antarctica since the Oligocene as confirmed by the diamicts of the poorly dated Sirius Formation. The interpretation of this formation has important consequences for the stability of the Antarctic Ice Sheet through the Cenozoic (Clapperton and Sugden, 1990). It has been suggested that the Sirius Formation is associated with a very large Mid-Miocene ice sheet that subsequently reduced in size to its present form (Prentice et al., 1986; Denton et al., 1989). In contrast, Webb (1990) suggested that the Sirius Formation indicates that warm marine waters penetrated beneath the ice sheet with subsequent partial collapse during a Late Pliocene warm interval (Webb, 1990). The Southern (Patagonian) Andes were repeatedly glaciated during the Late Cenozoic and may provide one of the best preserved and well-dated terrestrial glacial records in the world (Clapperton, 1993). Glaciers, formed as early as in Late Miocene time and before 4.7 Ma, left multiple drifts on the mountain flanks throughout the Pliocene that may correlate with less well-dated deposits in the Peruvian– Bolivian Andes and New Zealand. The most

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION Northern Atlantic ocean V23-81

43

Insolation

FIG. 2.14. (a–c) Comparison of Late Pleistocene proxy climate records of ice sheet and mountain glacier advances, deep sea ␦18O record V23–81 and Dome C (Circe) ice core profiles with insolation curves from northern and southern hemispheres (from Broecker and Denton, 1989; reprinted from Geochimica et Cosmochimica Acta, 53, Broecker, W.S. and Denton, G.H., The role of ocean–atmosphere reorganization in glacial cycles, 2465–2501, 1989, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

FIG. 2.14. (d, e) (d) Shows oxygen-isotope and dust records from Dome Circe ice core on the East Antarctic plateau; (e) shows Southern Hemisphere alpine fluctutations. The grey bar marks the time of an abrupt global climatic event that ended full-global conditions (from Broecker and Denton, 1989; reprinted from Geochimica et Cosmochimica Acta, 53, The role of ocean–atmosphere reorganizations in glacial cycles, 2465–2501, 1989, with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

44

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

extensive Patagonian glaciation occurred during the Early Pleistocene at a time of limited northern hemisphere glaciation. At this time, 1.2–1.0 Ma, southern Patagonian glaciers reached 200 km east of the mountains to the Atlantic Continental Shelf (Clapperton, 1993). Glacial advances are inferred to have occurred throughout the southern hemisphere mountain region during the Late Middle Pleistocene (Fig. 2.13). Drifts and morainic limits of the Late Quaternary have been assigned to OIS4 in South America, New Zealand, Tasmania and in the sub-Antarctic. The last glacial maximum probably peaked in mountainous areas of the southern hemisphere between 24 000 and 18 000 BP, being well-dated in all but the African, New Guinea and Southern Ocean–sub-Antarctic (Clapperton, 1993) (Figs 2.13 and 2.14). 2.8. THE LAST 25 000 YEARS AND SYNCHRONEITY OF NORTHERN AND SOUTHERN HEMISPHERE GLACIAL CYCLES 2.8.1. Last Glacial Maximum 2.8.1.1. Greenland and Antarctica data It is well documented that glacial cycles of northern and southern hemisphere ice sheets are in phase, as shown by the ␦18O or ␦D records of the Greenland and Antarctic ice cores. Warming events in central Greenland cores lagged behind corresponding interior Antarctic (Vostok and Byrd) events by more than 1000 years. However, the ␦18O ice core data from Taylor Dome near coastal Antarctica did indicate synchroneity of global glacial climatic fluctuations with those in Greenland (Lowell et al., 1995; Steig et al., 1998; Grootes et al., in press). Data from the Ross Sea and Transantarctic Mountains, Antarctica, indicate ice advance through the Ross Sea coincident with the maximum of the southern margins of the northern hemisphere ice sheets. Antarctic Ice Sheet fluctuations are directly influenced by the growth and disintegration of the largely terrestrial, northern hemisphere ice sheets. This direct effect relates to the interplay of fluctuating global sea level change and continental ice sheet growth and decay. As much of

the Antarctic Ice Sheet is grounded below sea level, sea level recession (northern hemisphere ice growth) results in Antarctic Ice Sheet expansion and, conversely, Antarctic ice retreat ensues with increased calving and disintegration in relation to sea level rise (northern hemisphere ice wastage). 2.8.1.2. The mountain snowline One of the clearest demonstrations of interhemispheric coupling of climate change is shown by the uniform depression of mountain snowlines during the last glacial maximum. The transect in Fig. 2.15, along the western Cordillera of North and South America, indicates that a maximum depression of 900–950 m occurred in tropical and subtropical areas. The depression appears related mostly to cooling (~4–6°C) (Rind and Peteet, 1985; Broecker and Denton, 1989) in high mountain areas rather than to increased precipitation. Other data indicate lowland terrestrial temperatures as well as sea surface temperatures were uniformly ~3–4°C lower in glacials than during interglacials in the tropics (van Campo et al., 1990). However, these temperatures conflict with the findings of CLIMAP Project Members (1981) and some deep sea ␦18O data that indicate tropical ocean surface temperatures remained within ±2°C of their present values (Broecker, 1986). 2.8.1.3. Terrestrial ice advances in middle to high latitudes Figure 2.14 suggests a correlation of climatic data for both hemispheres across the last Pleistocene glacial maximum. The time–distance diagrams for the northern hemisphere (Fig. 2.14a) represent the southern margins of the Laurentide and Cordilleran Ice Sheets and mountain glaciers of the St. Elias Mountains of the Alaskan–Yukon border area. The southern hemisphere diagrams (Fig. 2.14e) show generalized indications of mountain glacier advance in the Southern Andes and Southern Alps of New Zealand. Two major maxima of the latest Pleistocene glaciation in both hemispheres occur at ~20 000 BP and ~15 000 BP; the latter more sharply defined, although of slightly lesser extent in many areas (Clapperton, 1990, 1993; Lowell et al., 1995). General mid-latitude ice recession, with

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

N

Alaska Range

Mexican Volcanoes Sierra Central Nevada America Cascade Range

4000

2000

Arctic Ocean

Elevation (m)

6000

0 90º

S

A n d e s

Antarctica Patagonia

Brooks Range

Scotia Sea

8000

45

60º

30º

0º

30º

60º

90º

Latitude FIG. 2.15. Average mountain snowlines of present day (heavy solid line) and last glacial maximum (heavy dashed line) on a meridional topographic profile along the cordillera of North and South America (from Broecker and Denton, 1989; reprinted from Geochimica et Cosmochimica Acta, 53, Broecker, W.S. and Denton, G.H., The role of ocean–atmosphere reorganizations in glacial cycles, 2465–2501, 1989, with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

minor readvances, began in both hemispheres by about 14 000 BP. 2.8.1.4. Sudden glacial readvances of the Younger Dryas in Northwestern Europe The glacial recession beginning ~14 000 BP was interrupted in northwestern Europe at ~13 000 BP by a rather sudden cooling to almost glacial conditions terminating ~10 000–10 200 BP. Ice core data indicate that this probably represents only one perturbation of a quasi-cycle during the Late Quaternary. During this cooling interval, the Scandinavian Ice Sheet built ice margin deposits nearly continuously around Scandinavia, and small mountain glaciers reformed in the marginal highlands (Ehlers et al., 1991) or readvanced to distinct moraines on the coast of Norway. There are widespread climatic events through both hemispheres that correspond in time with the Younger Dryas (LaSalle and Shilts, 1993). 2.8.2. Holocene Glaciation and the Little Ice Age Late Pleistocene retreat of ice sheets and major global warming by ~9000 BP led to average Holocene

climates being similar to today. Increased summer insolation brought on by a northern hemisphere summer insolation peak after ~10 000 BP caused temperatures (Fig. 2.16) in Eurasia and western North America to reach 2–4°C higher than present. However, south of the remaining ice sheets, in continental interiors, temperatures did not reach these Holocene maxima until –6000 BP (COHMAP Members, 1988; ~9200–7300 years ago according to GISP2 accumulation record). This period is the so-called ‘Hypsithermal’ (Altithermal or Climatic Optimum). Subsequent deterioration of temperatures owing to decreased summer insolation is usually considered to have culminated in the Little Ice Age. Superimposed on these broad Holocene climatic changes have been shorter-term temperature variations with periods of mountain glacier advance and retreat. The Little Ice Age was a global phenomenon of temperature variations, it appears to have begun with minor glacial expansions in the thirteenth and fourteenth centuries. Following a brief interval of warmer weather, it reached a major phase of repeated glacial expansions and retreats at ~1600 AD, culminating with global warming and retreat of temperate and subpolar mountain glaciers in the last 200 years. The avail-

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

18

15

12

9

6

3

0

AEROSOL

AEROSOL

LAND ICE (%)

100

SST (K)

CO2 (ppmv)

265

SJJA 4

50

0

-1 -2

-4

-3 -4

200

8

ICE

0 0

330

Ka BP

N. H. SOLAR RADIATION (%)

46

SST

SDJF

CO2

18

-8

15

12

9

6

3

0

Ka BP

FIG. 2.16. Changes in external climate-forcing elements (from Kutzbach and Guetter, 1986). Northern hemisphere solar radiation in June through August (SJJA ) and December through February (SDJF ) as per cent difference from the radiation at present; land ice (ICE) as per cent of 18 000 BP ice volume; global mean annual sea surface temperature (SST), including calculated surface temperature over sea ice, as departure from present, K; excess glacial-age dust (AEROSOL); and atmospheric CO2 concentration (reproduced with permission of the American Meteorological Society).

ability of precise dating by tree rings and, particularly, the many historical observations in Europe, suggests that glacial fluctuations were nearly synchronous within separate regions spanning several hundred kilometres (Grove, 1988).

2.9. CAUSES OF CLIMATIC CHANGE FOR LATE CENOZOIC GLACIATION Suggested causes of glaciation are much more numerous than the glaciations themselves. Both extraterrestrial and terrestrial causes have been invoked. Extra-terrestrial mechanisms include proposed changes in solar radiation, dust clouds interposed between

Earth and Sun, changes related to the galactic year (Steiner and Grillmair, 1973), shading of Earth’s equatorial zone by an icy ring similar to that of Saturn, and changes in Earth’s orbital parameters (Milankovitch effect). Terrestrial causes include explosive vulcanism producing a globe-encircling dust cloud (offset by increased CO2 levels?), antigreenhouse effects related to decreases in CO2 content caused by precipitation of large amounts of carbonate rock or decreased volcanic activity. Plate tectonic positioning of the continents can have a major effect on oceanic circulation and heat distribution across the surface of the Earth. Climatic fluctuations may also be explained in terms of changes in oceanic circulation.

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

2.9.1. A Range of Causes There is some general agreement as to the most important factors affecting climatic change over the Late Cenozoic as well as distant geologic past. All the causes listed below may have played some part in triggering, forcing, sustaining or otherwise controlling Late Cenozoic glaciation: 1 changes in solar radiation related to solar output; 2 changes in land–ocean distribution associated with plate tectonic movements with consequent changes in mountain elevation, ocean circulation, sea level and atmospheric composition; 3 changes in atmospheric composition, particularly of greenhouse gases carbon dioxide and methane. These gases may be linked to major reorganization of the ocean–atmospheric circulation system as well as to tectonic events; 4 changes in the albedo of the Earth’s surface; 5 changes in the Earth’s orbital parameters (Milankovitch hypothesis); 6 other changes including catastrophic events such as extended periods of volcanic eruption. 2.9.2. Initiation of Late Cenozoic Glaciation Since the Permo-Carboniferous Ice Age (~250 Ma), the solar flux increased by ~1 per cent (MacCracken et al., 1990), the continents joined and split apart again and moved to their present positions, and plateaux and mountain elevations have increased. Changes in ocean geometry, atmospheric vapour transport and current pathways must have had profound effects on ocean temperatures and salinity, on rainfall, temperature patterns and atmospheric composition. Evidence suggests that CO2 concentrations were 5–10 times greater during the Mesozoic compared with today. The apparent decrease in atmospheric CO2 from the Late Cretaceous through the Pleistocene and the consequent decreased greenhouse warming help explain the general, stepped, cooling over this interval. Nevertheless, the causes of CO2 changes and rapid cooling are uncertain. The decline in atmospheric CO2 levels has been ascribed to decreases in global sea-floor spreading rates and volcanic outgassing. Decreased global chemical weathering (reducing CO2 ) related to

47

elevated sea levels and hence a reduction in land area, may also have been a major cause of climatic change (Raymo, 1991). The movement of land masses into polar latitudes has often been considered critical in bringing on continental glaciation (Crowley et al., 1987). Paleogeographic reconstructions indicate, however, that Antarctica and the northern continents have been near their present latitudes for the past 100 Ma. Local geographic factors were once considered to be critical to the growth of Late Cenozoic glaciation at high latitudes. These included: the uplift of ice-sheet nucleation centres in arctic Canada; the opening of the Drake Passage isolating Antarctica from South America; and the formation of the Isthmus of Panama and the possible diversion of vital ocean currents to more northerly paths. In addition, increased Cenozoic volcanism and solar shielding have not proven to have provided adequate climatic forcing at appropriate times for Antarctic and mid-latitude northern hemisphere glaciations. Two complementary hypotheses for intensification of cooling and triggering of Late Cenozoic glaciation are related to the significant, post-Eocene uplift of the Himalayas, Tibetan Plateau, Andes and highlands of western North America. Uplift appears to have culminated by ~5 Ma and in the Himalayas and Tibet between 3 and 1 Ma (Table 2.1). Models of uplift in the northern hemisphere indicate that the modification of the path of the upper atmosphere jet stream would have altered the Rossby wave pattern, enhancing cold polar airmass penetration over more southern areas of the northern hemisphere continents. A related hypothesis argued that uplift of the Tibetan Plateau, the Himalayas and probably the Andes took place during the Late Neogene resulting in a dramatic increase in subaerial weathering that would have significantly decreased atmospheric CO, resulting in global cooling (Raymo, 1991). 2.9.3. The Milankovitch Astronomical Theory 2.9.3.1. General principles Strong and continuous global cooling since the MidPliocene, and an increase in intensity of glacial/ interglacial oscillations over the past 1–2 Ma, have

48

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

apparently been paced and to some extent driven by changes in the Earth’s orbital parameters. However, the long-term pattern of this forcing is not known to have changed sufficiently to explain selective and sudden glacial initiation or intensification. The Milankovitch (1941) theory holds that these orbital changes, resulting from gravitational perturbations by other planets, cause cyclic changes in latitudinal and seasonal distribution of solar radiation that, in turn, induce climatic fluctuations. Summer insolation received in middle latitudes is considered to control the volume of ice stored on continents. The seasonal insolation cycle arises mainly from the effects of (1) changes in eccentricity of the Earth’s orbit around the Sun; (2) change in tilt (obliquity of the ecliptic) of the Earth’s (spin) axis relative to the orbital plane (ecliptic); and (3) changes in the relative position of the equinoxes owing to precession of the Earth’s axis of rotation about the axis of the orbital ellipse (Fig. 2.17). Orbital mechanics indicate quasi-harmonic variations with periods of 100 000, 41 000 and 23 000 and 19 000 years, respectively (Figs 2.6c and 2.7c). These cyclical variations affect the Earth’s distance from the Sun, changes in the latitude of the tropics, the position of the polar circles by a few degrees, and the angle of incidence of the solar beam at different seasons. Orbital solar forcing is virtually constant on a global-mean annually averaged basis; therefore a feedback mechanism is needed to convert seasonal and latitude-specific atmospheric changes to a global climatic response. The Milankovitch orbital theory maintains that northern high-latitude summers receiving minimum radiation may be cooled sufficiently to allow winter snows and sea ice to persist, thereby permitting a positive annual accumulation budget to develop. This effect may eventually initiate a positive feedback cooling response over the Earth leading to a further latitudinal extension of snow and sea ice and a consequent increase in surface albedo (Oerlemans, 1991). Evidence, however, contradicts the effectiveness of this mechanism in initiating and terminating glaciations (Broecker and Denton, 1989).

tion of the North Atlantic Ocean. Under present conditions, warm waters of relatively high salinity in the equatorial areas of the Atlantic move northward at surface and intermediate depths. In transit north, the waters evaporate and release considerable heat into the Greenland–Norwegian and Labrador Seas. The resulting cooler, denser water sinks to the ocean bottom forming the North Atlantic Deep Water (NADW) current. An annual release of ~5 × 1021 calories of heat into the atmosphere occurs over the North Atlantic as a result of this process. The cool NADW then exports cool salty water southward around the tip of Africa into the southern oceans (Broecker et al., 1990). If some mechanism should ‘shut off’ or decrease the flow of warm salty waters to the north and hence saltladen NADW to the southern oceans, the result would be a cooling of North Atlantic surface waters, a decrease in salinity and possible growth of ice sheets in the surrounding area. In the reverse process, deglaciation may be brought on by a ‘turning on’ or increase of NADW. Such major changes would have a pronounced effect on atmospheric CO2 content reinforcing climatic changes. The rapidity of the CO2 exchange between ocean and atmosphere remains a major uncertainty (Sarnthein et al., 1987). Such jumps in ocean circulation between modern and glacial modes may be a consequence of Dansgaard–Oeschger oscillations as observed in Arctic ice cores (Alley et al., 1999).

2.9.4. Changes in North Atlantic Circulation

2.9.5.1. Fluctuations in solar radiation

An important hypothesis for rapid global cooling and glaciation revolves around the thermohaline circula-

Climatic imprints left by factors external to the climate system, such as variations in solar radiation or

2.9.5. Causes of Shorter-term Climate Change and Holocene Glacier Fluctuations Short-term changes in climate such as those recorded during the Pleistocene–Holocene transition or during the Holocene itself, cannot rely for explanation on the longer-term temporal cycles of Milankovitch-type orbital perturbations. That ocean–atmosphere or CO2 mechanisms could have played a part, at least in the early Holocene, seems likely. Two mechanisms are considered below for climatic change that may occur at any time, but are most evident in the late Holocene mountain glaciations.

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

(a)

N

N Maximum in N. Hemisphere winter insolation

Minimum distance

Maximum in S. Hemisphere summer insolation

Maximum in N. Hemisphere summer insolation

Maximum in S. Hemisphere winter insolation S

Minimum in S. Hemisphere winter insolation

Maximum distance Small tilt

Small tilt

S

N

N 21 June

21 December

Minimum distance

Maximum distance

Minimum in N. Hemisphere winter insolation

Large tilt

Large tilt

Minimum in S. Hemisphere summer insolation S

Percent

past

future

5

Eccentricity

Earth-Sun Distance in June less more

(b)

S

Minimum in N. Hemisphere summer insolation

21 June

21 December

49

3 1

Precession

Degrees

24.00 23.50 23.00

Tilt

22.50 250

200

150

100

50

0

-50

-100

Thousands of Years Ago

FIG. 2.17. Diagrammatic representation of the Earth’s orbital elements (eccentricity, tilt and precession) (from Ruddiman and Wright, 1987; reprinted with permission from the authors). Eccentricity is the distance from the ellipse centre to one focus divided by the length of the semi-major axis (i.e., halfway between perihelion and aphelion). It varies from a maximum where the receipt of radiation outside the atmosphere varied by 30 per cent between perihelion and aphelion to a minimum when the orbit is nearly circular and precession has little climatic effect. (b) Changes in eccentricity, tilt and precession of the Earth’s orbit (from Imbrie and Imbrie, 1979, reproduced with permission of the authors and Enslow Publishers).

50

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

changes owing to volcanic blocking insolation, are difficult to distinguish from internal climatic variability. Dark sunspots on the visible disk, which have cycles of 11 and 22 years as well as longer ones, correlate directly with intervals of increased internal activity of the Sun, increased solar brightness, surface activity and with a slight increase in the solar radiation constant (Pecker and Runcorn, 1990). An analysis of historic records of sunspot activity has revealed, for example, a remarkable correlation of sunspot minima with some of the historic and proxy data of cooling (e.g., 10Be concentration in ice cores) over the last millennia. This includes a solar maximum about 1200 AD (recognized as the medieval warm interval) followed by three sunspot minima, the Wolf (1280–1340 AD), Sporer (1420–1530 AD) and the Maunder Minimum (1645–1715 AD) with no recorded sunspots. The latter two minima correspond with what may have been the stormiest and coldest three centuries in the mid-latitudes of the northern hemisphere (Little Ice Age). Furthermore, the data support the existence of a 200-year solar cycle. Attempts to correlate climatic events based upon a variety of paleoclimatic proxy data, including glaciers with 14C solar proxy data, have yielded mixed results (Magny, 1993). Some of the Little Ice Age solar minima can be correlated with periods of decreased CO2 in ice cores (Stuiver and Braziunas, 1989); however, in general, the CO2 concentration is rather constant in ice cores. It is likely that the worldwide Holocene glacial correlations are not sufficiently well-dated to portray details of global change. Furthermore, the global increase in temperature over the past 150 years coincides with mountain glacier retreats, a 14C decrease and, apparently, with increased solar activity. Therefore, some component of the observed temperature increase may be due to elevated solar energy output (Pecker and Runcorn, 1990). 2.9.5.2. Changes in atmospheric transparency and volcanism Changes in atmospheric transparency and incident radiation caused by volcanic eruptions emitting micro-particles and gases, particularly SO2 and H2S, may compliment changes in solar activity. Most of the

sulfate gases are quickly converted to sulfate (H2SO4 ) aerosols that reflect solar radiation and cool the troposphere. Convincing correlations of volcanic eruptions with historic climatic change are available, but often offer conflicting evidence (Broecker, 1992). Estimates of the effects of volcanism on surface temperatures over the last century suggest that cooling may have been limited to a few tenths of a degree for a one- to two-year period after each event. However, long-term correlations seem to substantiate volcanism as a major, if not primary, factor in climate change (Zielinski et al., 1995). 2.9.6. Which Cause? The search for specific causes of long- or short-term climatic fluctuations within glacial or interglacial intervals, or for the triggers of longer-term, orbitally regulated coolings or warmings, may remain an enigma for some time to come. It has been suggested, when considering the Younger Dryas cooling phase, that at the start of major orbital changes small disturbances may produce large climatic changes by positive feedback amplification (Berger, 1990). However, the ability to distinguish between cause and effect in such feedback loops remains problematic. Furthermore, causes external to the climate system may greatly complicate the search for mechanisms to explain glaciations. 2.10. FUTURE CLIMATE CHANGE AND GLACIATION As understanding of past global climates and mechanisms of climatic change improves, the need to be able to predict future climatic change and the growth of ice sheets becomes increasingly pertinent. Such predictions must be couched in terms of the geological, climatic and astronomical factors and the impact of human civilizations on past and present global climates. From the review of Late Cenozoic climatic change and glaciation, the Earth appears to be experiencing an interglacial phase. Short-term climatic trends give only mixed clues to any impending glaciation. For example, mean global annual surface air temperatures have been rising since the end of the Little Ice Age approximately 100 years

0

Barbados III TERMINATOR II

Barbados I

Barbados II

-30

Late Wisconsin

0

Present High TERMINATION I

30

Early Wisconsin (?)

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

Period of Glacial Growth

50

100

Thousands of Years Before Present

150

0

-30 150

NEXT TERMINATION ?

30

Period of Glacial Growth ?

100

50

Thousands of Years in Future

0

Mean - Global Temperature (ºF)

FIG. 2.18. Insolation curves for 45°N covering the past 150 000 and future 150 000 years calculated from orbital data of Vernekar (1968) (after Broecker and van Donk, 1970; reprinted with permission of the American Geophysical Union).

65

51

ago, although there was a distinct cooling phase during the period from the 1940s to the 1960s. Nevertheless, paleoclimate data demonstrate that mean annual temperatures of the northern hemisphere have fallen ~1°C from the Mid-Holocene Climatic Optimum to present values. Since an apparent correlation of past climatic change with astronomical control of peak summer radiation for the northern hemisphere can be shown, it can be argued that long-term orbital curves may provide the single best available longterm climatic extrapolations for the future. The solar insolation curve calculated for 45°N over the next 150 000 years (Fig. 2.18) suggests that glaciers may start to build up rapidly over the next 50 000 years, reaching a maximum stage of development before the next interglacial some 100 000 years after present (AP). Perhaps a more realistic projection of the orbital data comes from models that use the insolation curve of Fig. 2.18 or similar data as input, but take into account the history of climate response to changes in the solar radiation patterns (e.g., from deep sea sediment curves of ␦18O). These irradiation simulation models indicate that temperatures will continue

Carbon Dioxide Induced “Super-Interglacial” Last Interglacial

Present Interglacial

60

Beginnings of Agriculture Last Glacial

55

LAST MAJOR CLIMATIC CYCLE

50 -150

-125

-100

-75

-50

-25

0

25

Thousands of Years (B.P. - A.P.) FIG. 2.19. Temperature curves based on deep sea ␦18O data and projection into the future. (a) Solid curve suggests rise in temperature owing to greenhouse effect before temperatures eventually fall again to join cooling trend projected by astronomical forcing (from Imbrie and Imbrie, 1979).

52

GLOBAL GLACIAL CHRONOLOGIES AND CAUSES OF GLACIATION

the general cooling trend begun in the Mid-Holocene (~6000 BP), leading toward global cooling and probably glaciation by ~23 000 AP (Fig. 2.19). However, an advance more comparable with the last glacial cycle (Late Wisconsinan), is proposed by ~60 000 AP (Berger, 1981). These orbital effects must be superimposed over and above climate fluctuations that have frequencies higher than the 19 000-year precession cycle and also over climate changes forced by anthropogenic effects. Human-induced climate change largely results from increases in the concentrations of greenhouse gases in the atmosphere. These gases include carbon dioxide, methane, nitrous oxide and carbon monoxide that absorb infrared (long-wave) radiation. Atmospheric CO2 , now at levels far above any in the last 160 000 years, is by far the most abundant of these gases. It has increased by 25 per cent over the last 150 years and from 315 ppmv to ~351 ppmv during the last 30 years alone, as a result of industry and agricultural expansion (Houghton et al., 1990; MacCracken et al., 1990). Projection of atmospheric gases, based on fossil fuel use, suggest that CO2 will double its preindustrial levels by the middle to late decades of the twenty-first century. Studies indicate that the increases in greenhouse gases have already led to global temperature rise, however, it is difficult to separate these from the possible increases in solar irradiation. By the middle of the twenty-first century, mean global surface air temperatures will have reached 1°C above pre-industrial levels and perhaps a few degrees higher in middle to high latitudes with increases at a rate of 0.75°C every 25–50 years (MacCracken et al., 1990). Extrapolation of the greenhouse effect beyond the next century is very speculative; however, unrestricted burning of fossil fuels without counteracting climatic fluctuations may involve a ‘super-interglacial’ unlike any interval experienced by Earth

during the past million years (Manabe and Stouffer, 1993) (Fig. 2.19). Furthermore, because of the slow absorption of CO2 by the oceans, the effects of this carbon dioxide may endure for a thousand years after fossil fuels are exhausted. A greater understanding of past natural climatic cycles and of atmosphere–ocean CO2 interactions may help to predict whether temperatures may peak in decades or millennia before there is a return to the cooling of the orbitally or otherwise-driven cooling trends. What are the possibilities of ice sheets and major mountain glaciation occurring in the immediate future? It has been demonstrated that initial ice sheet growth at the start of the last glacial cycle (~120 000 BP) occurred in high northern latitudes (65–80°N) sustaining climatic conditions quite similar to those of today. The geologic record indicates that these optimal ice-growth conditions include a warm northern ocean, strong meridional atmospheric–oceanic circulation and low summer temperatures as well as depressed winter temperatures. Since greenhouse warming is expected to be most pronounced in the Arctic and during winter months, together with decreasing summer insolation (Fig. 2.16), this anthropogenic warming could lead to snowline depression and ice sheet growth in the high northern latitudes. Both geological and historic evidence exists demonstrating glacier growth and expansion in the Arctic during periods of global warming (Miller and de Vernal, 1992). Increased precipitation during the Holocene in Antarctica is countered by increased calving at the ice margin contemporaneous with sea level rise (Clapperton and Sugden, 1990). Certainly more collection of paleoclimatic data, as well as information on chemical, physical and biological interactions of the contemporary atmosphere and oceans, will be needed before even our most sophisticated computer models can make climate predictions of any detail.

3

GLACIERS AND ICE SHEETS J. Menzies (with contributions by T. J. Hughes)

ice mass. Both examples depict complex and interrelated facets of the glacier system. The impact such glaciodynamic variations may have upon sedimentological processes and sediments is exhibited in the sediment/landform associations formed within specific glacial environments. In understanding the mechanics of glacial erosion, debris entrainment, transportation and subsequent deposition, the relationships between ice mechanics, underlying topography and sediment source and transport pathways must be considered. Each landform or bedform produced within a glacial environment is a product of, or a reaction to, glaciological factors that were either of ephemeral or long-term effect. Our past understanding of glaciers, of how they move over their beds, erode and transport debris, has been partly hampered by their general inaccessibility. Early pioneering work was largely based upon an understanding of alpine valley glaciers, a trend that dominated at least until the 1950s. By the International Geophysical Year (1957), studies of the Antarctic and Greenland Ice Sheets had begun and have since continued apace. Since then, studies of glaciers can be roughly partitioned into: (a) theoretical and applied glaciological studies; and (b) glacial geological studies of Pre-Quaternary sediments, Quaternary sediments and present-day glaciated areas.

3.1. INTRODUCTION In studying any glacial environment a fundamental appreciation of the glaciological factors that influence the glacial processes and sediments must be acquired. It is the purpose of this chapter to consider those aspects of glacial physics pertinent to the understanding of sedimentological processes within glacial environments. For recent studies of the wider field of glacial physics and glaciology, see Paterson (1994), Hooke (1998), Hughes (1998) and Van der Veen, (1999). The mechanics of ice movement, basal ice stresses, pressure melting, basal ice thermal conditions, glacial hydrology and the response of ice masses to climatic changes are examples of the influence exerted by the glaciological component upon the glacial system. When periodic increases of snow within the accumulation area lead to the development of kinematic waves, increased but localized basal slip velocity may occur. In time, the kinematic wave may induce an ice advance, perhaps localized deposition of subglacial debris melt-out beneath the area of the wave or increased mobilization of a deforming basal debris layer. Likewise, reduced accumulation or increasing ablation may lead to ice retreat and increased debris melt-out in the near marginal supraglacial areas of an 53

54

GLACIERS AND ICE SHEETS

3.2. ICE MASS TYPES Classifications based upon morphology, thermal structure, geographical location and velocity have all been used to various degrees of success (Paterson, 1994; Van der Veen, 1999). As new information has been acquired, ice mass classifications have been revised or abandoned. At present, no universally accepted classification exists. Ahlmann’s early thermal classification (1948) had been revised and later altered to accommodate increasing knowledge of ice mass thermal structures and, in particular, subglacial thermal states. Recently, classifications of the subglacial environment with a perspective on sedimentology have been proposed (Shoemaker, 1986c; Menzies, 1987). The value of any ice mass classification is in its ability to provide clues as to likely glaciologic conditions and sedimentologic processes. For example, where an ice mass is expected to have been a grounded tidewater-type glacier, as was the case in certain instances along the edges of the Great Lakes during the Late Wisconsinan, then specific environments are likely to be encountered when understanding the stratigraphic record from such a site. Where the record does not seem to correspond with the expected lithofacies associations, either the classification is faulty or inappropriate, or a greater understanding of a particular environment is necessary. 3.3. FORMATION OF GLACIER ICE As snow accumulates on the surface of a glacier in a loose and highly porous state, it begins a series of transformations termed diagenesis or metamorphosis. Following burial by the new year’s snowfall, a process of densification or consolidation ensues. The transformation of snow to ice occurs, under dry conditions, when increased grain packing, rounding (sintering), reducing porosity and permeability, and recrystallization takes place (Plate 2.1). Under these conditions, densification can be expressed by Sorge’s Law where density becomes an exponential function of depth below the original accumulation surface. Under these dry conditions, the process of transformation of snow to firn can take over 100 years. Where wet conditions of surface melting and percolation

prevail, the above process occurs more rapidly since meltwater from surface snow melt percolates into the snow and refreezes at depth thus reducing porosity, reducing snow grain size and accelerating transformation (Lock, 1990). Glacier ice is, by definition, a polycrystalline substance impermeable to air at the macroscopic level, but containing air bubbles and inclusions of other chemical species. Polycrystalline ice is strongly anisotropic owing to the variety of crystal shapes, sizes and orientations (Plate 2.1). These properties account for the remarkable ductility and strength exhibited by glacier ice. 3.3.1. Glacier Ice Density The process of snow transformation passes through two phases before becoming glacier ice, viz: firn – a compacted snow usually of the previous year’s accumulation (density: 0.4–0.8 Mg m–3 ) and n´ev´e – densified snow of reduced air permeability (density: 0.7–0.8 Mg m–3 ). Glacier ice, normally, has a density of ~0.9 Mg m–3; however, the input of impurities and particles can alter this figure slightly. Air bubbles are often found within glacier ice owing to the relative rapidity of the densification process. Where surface melting and refreezing occurs fewer air bubbles are present. Highly attenuated, sheared sets of air bubbles are commonly encountered in the lower basal zones of temperate ice masses where repeated regelation processes have resulted in the incorporation of air. 3.3.2. Glacier Ice Crystals Glacial ice polycrystals have two unique qualities. First, ice crystals are relatively weak along the basal plane, allowing comparatively easy adjacent crystalline slip (dislocation climb). Internal deformation of glacier ice thus occurs under very small stress applications resulting in glacier movement. Secondly, glacier ice is less dense than water and thus is buoyant. Ice crystal size, generally, increases from the surface downwards in a step-like fashion until, at ~100 m below the surface, a relative constant size (typically ~10–30 mm va) persists until close to the glacier sole where very large crystals (~5–10 cm va) are commonly found (Fig. 3.1) (Michel, 1978; Van

GLACIERS AND ICE SHEETS

a 0 20

Depth (metres)

40

1

2

55

(millimetres)

3 4

6 8 10

20 30

S I

60 80 100 120 140 160 180

FIG. 3.1. Average crystal size vs. depth in the Massif du Mont Blanc where S is in firn and I is glacier ice (after Michel, 1978; reproduced with permission of Les Presses de L’Universit´e Laval).

der Veen, 1999). Crystals may vary in size from fine to coarse depending upon the age and debris content of the ice and past changing climatic temperatures. Small crystal sizes are likely to be due to consolidation and/or crystallization around microparticles. With increasing stress application, ice crystals develop a stress-anisotropic fabric.

3.4. MASS BALANCE AND GLACIER SENSITIVITY The surface morphology, overall dimensions and geographical location of all ice masses are largely functions of the factors controlling mass balance (Fig. 3.2). Where low snowfall increments accumulate in

high-latitude, polar dry continental sites, low surface temperatures and low levels of solar radiation combine to preserve almost the complete annual snowfall. In contrast, in subarctic and mountainous areas in middle latitudes, snowfall increments are often extremely large, such that higher summer solar radiation values and high ablation rates result in large meltwater production. The preservation of snow is restricted to the higher parts of the glacier (accumulation area) where the influence of lower surface temperatures with increasing elevation is effective. Finally, in regions of high altitude, low latitude, high solar radiation and summer temperatures, snow accumulation may only survive if sufficiently large accumulations occur (Østrem and Brugman, 1991; Van der Veen, 1999).

56

GLACIERS AND ICE SHEETS

Topographic slope Bed relief Precipitation

Geothermal heat

Wind-blown snow

Extent of surface debris

Air temperature

Ice velocity

Glacier surface temperature

Basal ice temperature Temperature of glacier bed

Altitude Latitude Maritime influences Solar aspect

Distribution & Thickness of Snow Cover

Relief Solar radiation Cloud cover Dust content of atmosphere Avalanche frequency & magnitude

Distribution & Thickness of Glacier Ice

Presence/absence down-ice of topographic bars/rises, ice rises or ice shelf Basal debris extent & thickness geotechnical/rheological properties Extent & depth of basal meltwater Surface meltwater, seasonal distribution Addition of avalanched glacier ice Presence/absence of tributary Glaciers/ice streams

FIG. 3.2. A generalized linkage diagram exhibiting the relationships between external geographical and atmospheric variables, glacier mass balance and ice mass thickness and extent.

3.4.1. Factors Influencing Mass Balance Figure 3.3 shows, in simple form, the linkages between regional climate and ice mass response to mass balance change. A lack of universal correlation of mass balance to specific influencing factors can be illustrated over relatively short distances where comparable climatic conditions prevail. As an example, the mass balance relationships in the Canadian Rockies reveal that the Peyto Glacier, Alberta, is dependant upon summer temperatures; the Sentinel Glacier, British Columbia, upon winter precipitation;

while the Place Glacier, British Columbia, is influenced by both elements (Letr´eguilly, 1988; Sturm et al., 1991). Likewise, Powell (1991) has illustrated the complex relationships that exist within any tidewater/ floating glacier system and mass balance controls (Fig. 3.4). The influence and effect of mass balance upon any ice mass is difficult to ‘isolate’ but often is illustrated by the position and movement of the equilibrium line (equilibrium-line altitude – ELA). The equilibrium line, the boundary line between the upper accumulation and lower ablation sections of an ice mass, fluctuates, over time and space, as a

GLACIERS AND ICE SHEETS

57

(a)

1 REGIONAL CLIMATE

LOCAL MASS INPUT

TOPOGRAPHY 2

LOCAL ENERGY BALANCE

GLACIER MASS BALANCE

(b)

GLACIER MASS BALANCE

3

GLACIER GEOMETRY GEOLOGIC RECORD

TOPOGRAPHY 4

GLACIER FLOW DYNAMICS

FIG. 3.3. (a) Simplified linkages between regional climate and glacier mass balance where local topographic control exerts a local specific influence. (b) Reverse linkage diagram showing mass balance and glacier response as reflected in the geologic record (after Furbish and Andrews, 1984; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 30 (105), 200, fig. 1, 1984).

function of climatic variables over time (Oerlemans, 1989). Regression analyses have been used to correlate mass balance to climatic variables using, for example, annual specific mass balance, ELA, summer temperatures and precipitation, and annual temperatures and precipitation. The gradient of mass balance with altitude, or ‘mass balance gradient’, has become a common comparative value for mass balance and ice mass sensitivity analyses (Mayo, 1984). Gradients may have a steep ablation and a shallower accumulation section and vice versa. Two assumptions are implicitly accepted in evaluating mass balance change and ice mass response: (a) mass balance perturbations are independent of altitude; and (b) the mass balance profile is constant. Neither of these assumptions can be generally substantiated either from theory or actual observation. Mass balance gradient is a function of several contributing factors: (a) accumulation from precipitation is influenced orographically; (b) wind drift snow accumulation tends to be higher at high altitudes; (c) surface

albedo increases with altitude as a function of shortwave absorption; (d) air temperature lapse rate varies as a function of altitude; (e) the fraction of precipitation falling as snow increases with altitude; and (f) snow cover increases with altitude thus heat gain from side-wall areas in valley glaciers decreases. 3.4.2. Equilibrium Line: Sensitivity to Change In periods of higher snowfall and/or reduced ablation, the equilibrium line will move down-ice as the accumulation area enlarges, depending upon factors constraining flow and glacier dimensions. Glacier retreat, after a time lag, may follow periods of higher surface temperatures and/or reduced snowfall accumulation causing enlargement of the ablation zone. The time lag (relaxation or response time) following a change in input or output components to the glacier is the time delay necessary for a system change to pass through the ice mass and thus be evident as a visible expression of that change at the

58

GLACIERS AND ICE SHEETS SNOW ACCUMULATION CLOUD COVER

MELTING + EVAPORATION/SUBLIMATION

ELA

Atmosphere Temperature (ELA Slope/Surface Slope)

Seawater Rainfall

Cliff Orientation

Marine Currents

FLOW VELOCITY

Ice Regimen

Bed Condition

Bedrock

Water

Drawdown

Deforming Sediment

CALVING Ice Regimen

Tectonism

Isostasy

Sediment Yield Drainage Basin Area Subglacial Meltout

Erosion Rates

Cliff Meltout

Water Depth

Flow Velocity

Sedimentation

Eustasy

Grounding-line System Volume

Transportation Rates

Calve-dumping

Crevassing

Release Rates

Conveyor Belt

Squeeze/Push

Length

Fluvial

Critical Depth

Sediment Dispersal Patterns Sediment Repose Angle

High Density Gravity Flows

Slides/Slumps

Diffusion

Icebergs

Direct Deposition

Plumes

Low Density Gravity Flows

By-Pass

Landslides

Lodgement

FIG. 3.4. Factors controlling mass balance of ice masses, including sedimentological controls for tidewater termini. Many factors are dependent variables, feedback loops are not illustrated for simplicity (after Powell, 1991; reproduced with permission of the author).

terminus. This input/output response relationship, however, is too simple and rarely can be detected in reality. An increase in surface mass flux may often lead to readjustments in local ice thickness, surface gradient and flow rate (thus basal shear stress field). As these variables alter, a feedback response in terms of ice mass geometry and flow dynamics usually occurs. Major constraints affecting ice mass response to input/output fluctuations are extra-glacial factors, such as topography and the relation of any part of the glacier system to major and minor tributary ice masses. For instance, a long, narrowly constrained valley glacier entering into a large dendritic glacier system may experience a very slow response: so slow that such a response may be difficult, if not impossible, to detect. In contrast, a small, cirque glacier will respond very rapidly with a change at the ice terminus being readily noticeable. The position of the ELA is a reflection of climatic variability as it relates to altitude, latitude, con-

tinentality, maritime influences and hemispheric atmospheric circulation. The Global ELA, today, varies from being very high in the equatorial regions of the world, where the line may be at 6000 m (as on Mount Kenya), to 3000 m in the Sierra Nevada of California, to 200 m in coastal northern British Columbia, to sea level at around 66°N latitude in Glacier Bay, Alaska. In comparison with the presentday, during the Pleistocene the ELA at sea level was at least 30° closer to the Equator.

3.4.3. Net Mass Balance A glacier’s net mass balance is said to be in a steady state when the inputs of fresh snow, blown snow and avalanched snow are in equilibrium with outputs resulting from ablation, and loss by blowing wind and calving (Van der Veen, 1999). In mathematical terms, the mass balance (b) can be expressed as follows:

GLACIERS AND ICE SHEETS

b = c+a =

冕

t

(˙c + a˙ ) dt

(3.1)

t1

bn = bw + bs = ct + at

(3.2)

bn = cw + aw + cs + as =

冕

tm

t1

(˙c + a˙ ) dt +

冕

t2

(˙c + a˙ ) dt

59

the atmosphere to the surface of the ice mass; (3) change in the geothermal heat flux; and (4) change in the length of the ablation season. Kuhn (1984) has attempted to categorize glacier types with certain balance gradient characteristics based upon imbalances between positive and negative periods:

(3.3)

tm

where a is ablation (as , summer ablation, aw winter ablation), a˙ is ablation rate, b is balance or mass balance, c is accumulation, c˙ is accumulation rate, at is total ablation, ct is total accumulation, bn is net balance, bs is summer balance, bw is winter balance, cs is summer accumulation, and cw is winter accumulation within time frames for the winter (t1 to tm ) and summer (tm to t2 ) seasons. Thus a glacier’s accumulation area has a bn > 0, and the ablation area a value of bn < 0, while at the equilibrium line bn = 0. When and if bn = 0, for the total ice mass, a steady state is reached. In reality, this state is rarely, if ever, attained because of the changing spatial geometry of an ice mass over time and the impact of past variations in bw and bs . This latter problem increases when mass balance calculations are made for large ice sheets. Values obtained of total accumulation for the Antarctic and Greenland Ice Sheets are, within a reasonable margin of error, accurate, but total ablation values remain elusive. Whether these major ice sheets are stable or whether they are gaining or losing mass at present remains unknown. 3.4.4. Mass Balance Gradients The calculation of mass balance can be made by considering an energy balance equation that is altitude dependent. This approach utilizing the entire balance for a year (over a time step-wise function of 30 min) involves the atmospheric temperature, snowfall and atmospheric transmissivity for solar radiation related to altitude allowing a balance gradient to be calculated (see Menzies, 1995, eq. 4.4, p. 109). The response of mass balance gradients of differing glacier types to variations in climate and thus mass balance parameters is complex. The variations may consist of one or a combination of the following effects: (1) change in accumulation to the ice mass at the surface or base; (2) change in the energy flux from

(a) Polar–Extrapolar: exhibit greatest variation owing to length of ablation period and have the lowest balance gradients; (b) Dry Continental: are affected by albedo changes close to the ELA; (c) Maritime: are most affected by changes in the accumulation area and have steep balance gradients; (d) Alpine: show least variation from the linear balance model and imbalances appear independent of altitude. This ice mass taxonomy introduces a process-related relationship between the factors that affect mass balance, ice mass growth and decay and glaciodynamics. No other classification has attempted to relate changing ice mass geometry resulting from climatic and other responses to the effect such changes have upon subglacial conditions of temperature, basal shear stress and sliding velocity. This approach to ice mass differentiation introduces a process-oriented logic that has wide implications for glacial sedimentology integrated to climatic change and individual (ice mass) bed geometry over the long-term period of ice sheet growth and decay. 3.4.5. Ice Sheet Growth Models and Mass Balance In attempting to reconstruct past ice sheet dimensions and processes of growth and decay, the following scenario has been suggested: as the large Pleistocene ice sheets of the northern hemisphere developed, in the early stages of expansion high accumulation rates could be expected to occur with steep mass balance gradients, high ice flux rates and high basal shear stress values similar to maritime conditions. However, as ocean surface temperatures and total ocean surface area decreased, an associated diminution of precipitation onto the developing ice sheets must have

60

GLACIERS AND ICE SHEETS

occurred leading to a reduction in mass balance and a more dry continental state being established with lower balance gradients and basal shear stress values. With ice sheet outward expansion, moisture-laden air masses along the southern edge of the ice sheets were prevented from penetrating the dry elevated continental centres of the ice sheets resulting in most precipitation being deposited within the ablation areas which, with southern expansion, became areas of very high ablation. The effect of these two competing processes was to balance each other out. This latter effect, at near maximum ice sheet spread, was to ‘flatten-out’ the balance gradient even to the point of producing a reverse gradient for a short interval of time. An additional effect to be considered is the impact of isostatic depression upon the geometry of the overall ice sheet (Andrews, 1997). It is possible that where the rate of ELA adjustment to ice sheet advance is in disequilibrium with the rate of crustal depression, an unstable state will develop in which ice retreat or, at least, an ice front standstill will take place. Whether such an unstable state could lead to catastrophic collapse of an ice sheet is unlikely since in retreat a new equilibrium between the ELA and the rate of crustal adjustment should re-establish and a new advance could then occur. Both modern and Pleistocene ice sheets exhibit(ed) asymmetry across their ice divides. The possibility that this form was due to differential mass balance seems unlikely. The cause would appear to be the result of differences in basal topography, thermal regimens and margin sensitivity to glaciodynamic and extra-glacial influences such as floating ice fronts, grounding lines and proglacial terrains (Menzies, 1995, chapter 3). It can be expected that glaciers in the same mountain region, even neighbouring glaciers, will exhibit differing responses at least over decades. The response time (Tm ) over which a valley glacier reacts to prior climatic changes manifested as changes of mass balance can be quantified using a linearized theory of kinematic wave propagation travelling down a glacier (Section 3.6.1)(Paterson, 1994):

predicates responses of 102 –103 years. However, observations today indicate shorter responses. In the case of major ice sheets, response variations, unless of global magnitude, may be of the order of thousands of years. For example, the present Antarctic Ice Sheet exhibits wide variations in glacier activity between those glaciers flowing into the Ross Ice Shelf in West Antarctica and those outlet glaciers flowing to the coast in the Enderby Land and Kemp Coast areas of East Antarctica (Drewry, 1983). The response of the Laurentide Ice Sheet, during the Pleistocene, was not synchronous with the Fennoscandian Ice Sheet. Within the same massive ice sheet variations in response can be expected along its edges. The response of the Laurentide Ice Sheet along its front in southern Ontario and New York State, for example, can be expected to have been widely different and asynchronous with the ice termini in LabradorUngava and the western portions of the Beaufort Sea. Along the edges of any ice sheet, separate ice streams and lobes will react at differing rates, frequencies and magnitudes. Unless synchroneity is considered in millennia, smaller asynchronous temporal variations can be expected. If marginal areas of past ice sheets have been characterized by thin ice thicknesses, rates of advance and retreat over considerable distances of terrain may have been on a scale of centuries rather than millennia. Chronological correlation, therefore, along the margin of any ice sheet becomes extremely complicated (Chapter 15). Interglacial and interstadial phases of the Laurentide Ice Sheet during the Quaternary, when higher temperatures followed an ice retreat or when short ice-free periods of tundra-like conditions prevailed, should not necessarily correlate with similar periods in Europe. For example, the Aller¨od of Northwest Europe need not be expected to be synchronous with a similar phase in North America.

Tm ~ fl/uL

Broadly speaking, two strategies for ice sheet modelling have emerged in the past two decades (Menzies, 1995, chapter 3). These strategies are now converging but were originally quite distinct.

(3.4)

where l is glacier length, uL is terminus velocity and f is a constant of approximately 12 . This equation

3.5. ICE SHEET MODELLING STRATEGIES (contribution by T. J. Hughes)

GLACIERS AND ICE SHEETS

3.5.1. The Melbourne School The first strategy disregards glacial geology as input to the model, and uses it only as a control on model output. This strategy was pioneered at the University of Melbourne (Australia) in the early 1970s (the ‘Melbourne school’). The Melbourne school, originally, held the view that if all the essential physics of ice dynamics and boundary conditions were specified correctly, the model would compute the basal thermal conditions (whether the bed is thawed or frozen and, if thawed, the rates of freezing and melting) and their distribution over the subglacial landscape. This information would then permit an understanding to be achieved of the glacial processes that modify the landscape. Hence, glacial geology is relegated to an ancillary role at the tail-end of model output (Fig. 3.5). In principle, the Melbourne school is correct in producing a one-dimensional flowline model so sophisticated in its internal ice dynamics and so precise in accounting for all external boundary conditions that the topographical expressions of glacial geology have only one explanation, the one

Reality (complex, interconnected) + concepts

61

provided by the mechanical (Newtonian) model. For example, Sugden (1977, 1978) used this model to compute the physical characteristics of the Laurentide Ice Sheet at its Late Wisconsinan maximum extent, and used the model output to explain the macro-scale aspects of North American glacial landscapes. The primary assumption made was that, at the glacial maximum, a glaciodynamic steady state was achieved and thus steady-state processes of glacial erosion and deposition made their dominant imprint on the subglacial landscape. The Melbourne school models were based upon the surface interaction between ice sheets and the atmosphere, particularly the surface mass balance between ice accumulation and ablation. Later a two-dimensional model was applied as a time-dependent model to simulate a Quaternary cycle of North American glaciation (Budd and Smith, 1981). It was further applied as a steady-state model to compute physical characteristics of the present-day Greenland and West Antarctic Ice Sheets (Radok et al., 1982; Budd et al., 1987). In all cases care was devoted to the ice– atmosphere boundary. However, the ice–water and ice–bed boundaries were ignored entirely, since only flow over a frozen bed was permitted and glacial geology was not used (Van der Veen, 1999). Most three-dimensional time-dependent models today trace their origin to the Melbourne model. 3.5.2. The Maine School

something missed or oversimplified

modelling

measurements

Model (simple, closed) with adjustable parameters

Field Data for tuning

for testing

calculations

Tuned Model

test

yes

"Explanation" Predictions

no

FIG. 3.5. Flow diagram of ice mass modelling process (reproduced with permission of the International Association of Hydrological Sciences (IAHS)).

A second strategy for model construction used glacial sediments and landforms as primary data input, based on the premise that they are direct records of the existence of former ice sheets, and indirect reflections of ice sheet internal dynamics and the external boundary conditions that constrained them. This strategy, pioneered at the University of Maine (USA) in the late 1970s (the ‘Maine school’), had as its guiding philosophy that both internal ice dynamics and external boundary conditions for former ice sheets must have been compatible with processes of glacial erosion and deposition that produced the glacial landscapes. Glacial sediments and landforms, then, became tools for understanding ice dynamics and boundary conditions of former ice sheets. This approach assists in understanding the ice dynamics of ice streams and their

62

GLACIERS AND ICE SHEETS

subglacial and marine grounding line boundary conditions, especially during deglaciation. In modelling former ice sheets, both schools distinguish between first- and second-order glacial landscapes. Any glaciated landscape is a mosaic of both macro-scale and meso-/micro-scale temporal and spatial components. The latter group tends to indicate the last glacial advance and retreat from an area, whereas the former reflects forms and terrains that have survived possibly repeated glaciations. The meso-/micro-scale forms include moraines, striations, grooves, kames, eskers and drumlins. Most of these are time-transgressive forms that indicate the ice margin movement during retreat or close to maximum ice sheet extension and are sufficiently small and delicate to be obliterated or overprinted by subsequent glacial advance(s). Because of their small and transitory nature, these features constitute second-order glacial landscapes. The first-order (macro-scale) forms, products of long-term and repeated glaciation, include fjords, glaciated troughs and basins, and scoured and fluted terrains. In the Maine school, glacial sediments, landforms and landscapes were employed in the primary development of ice sheet models. These models attempt to simulate both advance and retreat of ice sheets, during which time steady-state processes of erosion and deposition imprint a first-order glacial landscape. This landscape consists of large-scale glacial landforms. In considering the Laurentide Ice Sheet, first-order glacial landscapes consist of: (1) the continuing glacio-isostatically depressed Hudson Bay and surrounding lowlands, where ice load was greatest and most prolonged; (2) an exposed Precambrian Shield surrounding Hudson Bay, which locates a region of sustained sheet-flow of ice sliding on a thawed bed and regions of basal freezing and thawing; (3) linear troughs along the outer edge of the Precambrian Shield generally radiating from Hudson Bay, locating the region where interior ice sheet flow characteristics began to change into ice stream flow characteristic of ice sheet margins. This zone identifies the outer limit of the relatively steady-state ice sheet’s core; and (4) the outermost terminal moraines on land, and sills or sediment fans at marine margins, which locate the maximum ice sheet extent not glacially altered during retreat (Bouchard, 1989; Andrews, 1997).

A two-dimensional, time-dependent version of the Maine school model was developed to study the role of ice streams in ice sheet dynamics and the role of changing sea level on the stability of marine ice sheets. It has also been used to simulate a full glaciation cycle for ice sheets in North America, Eurasia and Antarctica (Fastook and Hughes, 1990). Glacial geomorphology is used to specify whether the bed is frozen or thawed, with thawed beds ranging from rough bedrock providing high traction to soft till providing low traction. Ice stream flow exists as channelized low-traction basal sliding. The distribution of melting, freezing and frozen conditions beneath an ice sheet is determined in large part by the distribution of accumulation, ablation and temperature conditions on the ice sheet surface. Fisher et al. (1985) used this glaciated terrain–ice sheet linkage to postulate surface conditions that would be compatible with basal conditions deduced from glacial geology in their two-dimensional reconstruction of the Laurentide Ice Sheet at the last glacial maximum. Timedependent glaciated terrain–ice sheet models (GT–IS model) use second-order glacial geology to control ice sheet dynamics during retreat. 3.5.3. The Merger of Both Schools Ultimately, a merger of the mechanical and geomorphic approaches to modelling ice sheets during a glaciation cycle will produce a glaciological synthesis comparable with that in hardrock geology produced by plate tectonics. This merger will require incorporating the equations of heat transfer with the equations of mass transfer in the ice dynamics so that surface temperature and mass balance can be quantitatively related to the basal thermal conditions that produce glacial terrains, and in relating changes in sea level caused by ice-volume changes to glacio-isostatic adjustments along marine ice sheet margins (Greve and MacAyeal, 1996; Hooke, 1998; Hughes, 1998; Van der Veen, 1999). 3.5.4. The Glaciated Terrain–Ice Sheet Model (GT–IS) The GT–IS model relies on glacial geology revealing changing patterns of flow lines for an ice sheet to

GLACIERS AND ICE SHEETS

determine whether ice sheet flow was over a melting, freezing or frozen bed. All of these boundary conditions cannot be presented here, but all cause bed traction to change in space and time. In this model, knowledge of the surface mass balance is not necessary, but surface accumulation and ablation rates can be inferred from basal traction, since these rates produce changes in ice velocity and traction in response to velocity changes. Glacial geological input to the GT–IS model (Figs 3.6 and 3.7) shows boundary conditions for a generalized ice sheet. FirstC

B D ice

shelf

A

A'

ice lobes

ice tongues

D'

C' B'

63

order glacial terrains, created when an ice sheet closely approaches true steady-state conditions, are reinforced by each successive glaciation. First-order features are found in: (1) the central regions of postglacial isostatic rebound, whether this region was frozen continental highlands or a thawed marine basin; (2) the heavily pitted erosion zones of basal melting and freezing that surround this central region, and where eroded debris is incorporated into a basal layer of regelation ice; (3) an outer zone of deposition where the regelation ice melts, either during the glacial maximum or during transient conditions that accompany glacial retreat; and (4) in fore-deepened channels cutting across the outer zone occupied by either terrestrial or marine ice streams. Ice streams tend to form in sedimentary marine troughs or river valleys and depressions, where soft sediments become mobilized. The resulting deformable bed persists until erosion removes the sediments or ice–bed interface conditions alter (Chapter 8). Ice sheet surface profiles can be constructed knowing only the topography and glacial geology of the deglaciated landscape (Thorp, 1991); no knowledge is required of mean annual temperatures and accumulation/ablation rates of the former ice sheet surface. 3.6. ICE MASS CHANGES AND FLUXES 3.6.1. Kinematic Waves

FIG. 3.6. Idealized ice sheet flow regime. Flow in the plan-view for surface ice (top) and basal water (bottom). Shown at top are surface flowlines (solid lines) radiating from a terrestrial dome (inside hachured line) and a marine dome (outside hachured line), a surface equilibrium line (dashed line) that is highest on the equatorward flank of the ice sheet and ice shelf grounding line (dotted line) on the poleward flank of the ice sheet. Shown at the bottom are thawed patches (isolated black areas) where quarrying creates lakes, frozen patches (isolated white areas) where lodgement till creates drumlins, the arc of exhumation (black areas beneath the surface equilibrium line) where quarrying and regelation occur along a basal equilibrium line that separates an inner melting regime from an outer freezing regime, the arc of deposition (between the arc of exhumation and the ice margin) where regelation ice melts, selective linear erosion in ice stream channels (broad black bands radiating from ice domes) and selective linear deposition in eskers (narrow black lines between broad black bands).

In response to positive changes in mass balance, surface bulges called kinematic waves develop that travel down-ice several times (3–5 times) faster than the average velocity of a valley glacier or ice stream. These waves occur as local increases in the balance gradient and have been linked to the triggering of glacial surges (Clarke et al., 1984; J´ohannesson et al., 1989) (Plate 3.1). Waves are typically 10 m high and approximately 1 km in length. Some waves form down-ice of seracs and other fast-moving sections of ice masses. Since the velocity gradient is high through such fast-moving regions, ice rapidly accumulates in the downstream area where surface bulges on the ice pile up. Such localized ice thickenings, on thin valley glaciers, may influence subglacial shear stresses, temperature and later processes of erosion, transport and deposition.

64

GLACIERS AND ICE SHEETS marine

ice

terrestrial

ice

E

A

A'

E

F M

F

W

M

D

M

E

sea

level

B'

B M

F

M F

W

F

F

M

E

C'

C M

F

W

F

E

sea

level

D'

D M F

M

D

M

F

FIG. 3.7. Idealized ice-sheet flow regime. Flow of ice in vertical cross-section is represented by thin lines along dotted surface flowlines A–A⬘ through D–D⬘ shown in Fig. 3.2. Shown above are longitudinal flowline profiles along the crest of the ice divide (section A–A⬘), along opposite flanks of a saddle on the ice divide section B–B⬘), along opposite flanks of the marine ice dome (section C–C⬘), and along opposite flanks of the terrestrial ice dome (section D–D⬘). At the base (hachured line), a dry bed (D) is frozen everywhere, a wet bed (W) is thawed everywhere, melting beds (M) and freezing beds (F) have a mix of frozen and thawed areas, regelation ice (dotted areas) forms at ice-stream headwalls, over beds (F) have a mix of frozen and thawed areas, regelation ice (dotted areas) forms at ice-stream headwalls, and over the thawed parts of a freezing bed.

3.6.2. Ice Mass Flux or ‘Activity’ Since ice flow is caused by unbalanced stresses within the ice mass, mass balance (accumulation/ablation) is a major component in understanding ice mass dynamics (Van der Veen, 1999). Ice flow is a thermomechan-

ical process involving gravity, mass balance, surface and basal temperatures, and the geothermal heat flux. The contribution of mass balance changes to total ice mass (mass balance gradient) varies by several magnitudes in comparison with the relative constancy of all other factors and is therefore the driving force in

GLACIERS AND ICE SHEETS

PLATE 3.1. A large wave-like bulge growing at the boundary between warm-based and cold-based ice on the Trapridge Glacier, Yukon Territory, prior to the next surge. The bulge is approximately 40 m high and the glacier width at the bulge is 0.75 km. Photograph was taken June 1980 (photograph courtesy of Garry Clarke).

65

ice mass sole, (2) near-surface and atmospheric temperatures, (3) vertical and horizontal components of velocity at any point within the ice mass, and (4) basal ice friction at the glacier bed (Hooke, 1998). Thermal changes within ice masses affect ice mass dynamics and sedimentological environments. At the surface of a glacier within the snow/firn/n´ev´e zone temperature fluctuations occur in response to seasonal and diurnal changes in air temperature. Temperature changes reduce in amplitude with depth such that below 7–16 m fluctuations become negligible. Typically, the 10 m temperature is taken as the standard criterion by which comparison between ice masses can be made. Meltwater infiltration into the percolation zone influences englacial temperatures at depth, introducing higher temperatures into the upper layers and reducing temperatures owing to the latent heat of freezing as meltwater refreezes at depth (Sharp et al., 1998). In winter, ice mass upper-layer

Temperature (°C)

3.7. STRUCTURE AND THERMAL CHARACTERISTICS OF ICE MASSES 3.7.1. Thermal Structure of Ice Masses Temperature distribution within an ice mass is a function of (1) the rate of geothermal heat flux at the

0

-6

-5

-4

-3

-2

-1

0

1 2

Depth (metres)

ice movement. The ‘activity index’ is a measure of the net mass balance gradient calculated as a function of the net accumulation with altitude (mm–1 ) up-ice. The greater the gradient, the faster glacier flow will tend to be. In evaluating the activity index geographically, the ‘gradient’ tends to decrease with increasing latitude and continentality. In reviewing indices of activity, Andrews (1975) introduced the concept of the ‘energy of glacierization’, relating ice mass balance changes to changes in mass transfer (ice flux) and thus ice velocity, A significant relationship exists between mass balance changes and glacial processes; a relationship that is exceedingly complex and far from fully understood.

-7

3 4 5 6

June 25

July 5

July 15

July 25

7 8 9 10

FIG. 3.8. Removal of winter ‘cold wave’ from firn (adapted from Sverdrup, 1935) (after Paterson, 1981; reprinted from Paterson, W. S. B., Physics of Glaciers, 1981, p. 189, fig. 10.1, with kind permission of Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

66

GLACIERS AND ICE SHEETS

temperatures are typically higher than the ambient surface air temperature, resulting in heat conduction toward the ice surface. The effect of these processes is to create or propagate a cold wave into the upper 15 m of an ice mass (Fig. 3.8). Thermal anomalies such as cold waves and percolation-zone layers are preserved within the ice, resulting in the development of distinct paleo-thermal layers. In time these layers fade as the ice mass equilibrates to steadystate temperatures.

3.7.1.1. Ice facies Ice masses can be subdivided into surface facies zones (Fig. 3.9). At a depth of approximately 10 m, a further ice facies classification based on ice crystalline structures and fabrics, and debris content can be used to separate glacier ice into facies types (Sugden et al., 1987; Hubbard, 1991).

3.7.2. Basal Temperatures and Thermal Boundary Conditions The temperature at the ice–bed interface is a critical factor in the processes of erosion, entrainment and deposition. In temperate, wet-based glaciers basal temperatures are found at ~–1 to –3°C. In polar, coldbased glaciers the ice mass is frozen to its bed with basal temperatures of –13 to –18°C (Paterson, 1994; Van der Veen, 1999). Thermal conditions at the ice– bed interface are more complex than implied by these two thermal states. Basal ice temperatures may vary both temporally and spatially producing, in temperate ice masses, polythermal bed conditions (Fowler, 1979; Hutter, 1983). In investigating thermal regimes beneath active temperate ice masses, it is apparent that local, shortterm fluctuations of temperature very likely occur at the ice–bed interface. Since temperatures are so close to the zero-degree isotherm and are closely

GREENLAND ICE SHEET MOST OF THE ANTARCTIC ICE SHEET

MOST MOUNTAIN GLACIERS equilibrium line accumulation area

dry snow zone

percolation zone

wet-snow zone

max. height of surface in current year

previous summer surface

no runoff

ablation area abalation area

superimposed ice zone

surface in summer

runoff possible

runoff occurs

REGELATION GLACIER ICE

DEBRIS CLOTTED BED

FIG. 3.9. Ice facies zonation on an idealized ice mass (modified after Benson, 1962 and Knight, 1992).

GLACIERS AND ICE SHEETS

coupled to short-term pressure variations and widely fluctuating discharges of meltwater, spatially and temporally transient patches of frozen and melting sections develop across the bed interface. The presence of ‘cold patches’ may explain the ‘stickslip’ motion of ice observed in subglacial tunnels (Fig. 3.10). However, the significance of polythermal conditions is one of the cornerstones of modern subglacial sedimentology. Where, in the past, erosional and depositional processes were spatially separate, polythermal conditions allow these processes to occur penecontemporaneously in local juxtaposition, and to overlap and affect the same bed area at different times. Thus, for example, the fact that subglacial diamictons are of dominantly local origin undergoing only short distance transport can be explained.

67

When considering the temperature fluctuations at an ice–bed interface the following parameters converge to establish the specific thermal conditions: (a) rate of snow accumulation (densification/transformation) and snow temperature on deposition; (b) geothermal heat flux; (c) mean annual surface air temperature; (d) ice surface velocity and vectors; (e) basal ice velocity and vectors; (f) subglacial meltwater discharge flux and temperature; and (g) the imprinted ‘memory’ of previous thermal conditions. In appraising polythermal bed models, it is assumed that steady-state conditions of dynamic equilibrium (net mass balance zero) prevail. In transposing these concepts to the Laurentide Ice Sheet, for example, it is assumed that the ice sheet was in a steady-state phase during its maximum

A C

O

R T

C

T A

N A

E

O

C

L

A

N

T

IC

PA C

IFIC

OCEAN

IC N A

E

Legend Warm-melting Warm-freezing Cold-based

0

500 km

FIG. 3.10. Spatial zonation of basal ice thermal regime beneath the Laurentide Ice Sheet at its maximum (~18 000 BP) (afer Sugden, 1977; reproduced by permission of the University of Colorado, Institute of Arctic and Alpine Research, 1977, courtesy of the Regents of the University of Colorado).

68

GLACIERS AND ICE SHEETS

3.7.3. Subglacial Thermal Conditions of Existing Ice Masses

expansion –18 000 BP. Several models of polythermal conditions have been developed for the Laurentide Ice Sheet that illustrate the potential variation in bed conditions across the ice sheet from centre to margin (Fig. 3.10) (Sugden, 1977; Fisher et al., 1985; MacAyeal et al., 1995; Greve and MacAyeal, 1996; Jenson et al., 1996; Marshall et al., 1996; Marshall and Clarke, 1997a,b). The question of steady state, however, needs to be redefined. In the past, it has been assumed that steady state implies an unchanging ice sheet centre that directly responds to external climatic forcing without significant dynamic feedback loops developing. However, recent evidence based upon patterns of cross subglacial lineations suggests that, in the case of the Laurentide Ice Sheet, the ice sheet centre shifted over time probably reflecting a coupled ice sheet/ global climatic temporal response in terms of spatial fluctuating accumulation and ablation areas on the ice sheet surface and feedback effects causing fluctuating atmospheric conditions (Boulton and Clark, 1990a,b; Jenson et al., 1995; Clark et al., 1996; Maher and Mickelson, 1997).

ELEVATION KM

Investigations within subglacial tunnels and from ice cliffs and boreholes have supplied much needed information on the basal thermal regimes of valley glaciers and ice sheets. The information can be subdivided into data on: (a) ice temperatures and subglacial cavity air temperatures; (b) rates of net loss from the glacier sole by melting (liquid and/or sublimate); (c) rates of net gain to the glacier sole by freezing (liquid and/or sublimate); and (d) basal, englacial debris content and concentrations. The data can be further summarized under: (a) wet-based ice conditions; (b) cold-based ice conditions; (c) polythermal bed conditions; and (d) theoretical and observed rates of basal melting and freezing. The evidence that wet-based and cold-based glaciers exist seems fairly well established but whether such a simple two-fold classification accurately portrays conditions beneath most ice masses today or in the past is questionable. The relationship between such mono-thermal basal conditions and

4 3 2 1

EXTENT OF WET BASED SLIDING FG M PROBABLE LOCATION OF FORMATIOM

F

MG

DRUMLINS/FLUTED MORAINE

FG

FG

M

MG

F

DRUMLINS/FLUTED MORAINE

MG

M

DRUMLINS/FLUTED MORAINE

FLOW LINE

F

FREEZING

SEDIMENTARY ROCKS

THICK SEQUENCE OF DIRTY ICE

M

MELTED

SHIELD ROCKS

FROZEN

MG

MELTING

FIG. 3.11. Idealized basal thermal regimes beneath a large ice sheet and their relationship to the spatial distribution of zones of erosion and deposition (modified from Denton and Hughes, 1981, after Eyles and Menzies, 1983). Reprinted from Eyles, N. (ed.), Glacial Geology, 1983, p. 20, fig. 2.1, with kind permission of Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, UK.

GLACIERS AND ICE SHEETS

polythermal conditions is complex and as yet little understood. In Figure 3.11 a suggested relationship is drawn between these markedly different basal thermal conditions. 3.7.3.1. Wet-based ice conditions Air temperatures recorded beneath Østerdalisen, Norway, within a subglacial cavity some 65 m from the ice front and beneath 22 m of ice, indicate an approximate constant value of –1°C in summer (Theakstone, 1979). Winter temperatures are rather lower and, unless a connection exists to the ice surface or ice margin, the air temperatures remain

PLATE 3.2. Portal of subglacial meltwater tunnel at the front of the Canwell Glacier, Alaska. Note debris banding and fissuring within the glacier ice. Person for scale standing on supraglacial debris (photograph courtesy of Ed Evenson).

69

relatively constant, often passing below 0°C, thus causing the glacier to freeze to its bed (Anderson et al., 1982). Under wet-based or temperate ice conditions, ice is at the pressure melting point throughout the ice mass, water is observed occasionally being squeezed from ice in the ceilings of cavities and immediately refreezing on the ice surface as fragile ice crystals or spicules. The base of most cavities has a refrozen meltwater ice glaze and patches of wet or trickling meltwater (Plates 3.2 and 3.3). Debris (Ice Facies 4) is found in basal ice as striped or strata-form englacial inclusions. The debris exists as discreet bands typically 0.2–0.9 m in thickness (Plate 3.4). The ice encasing the debris is termed amber ice having a fine-textured crystal size and a low air or gas content (0.2 cm3/100 g for Byrd Station, Antarctica). Debris-rich ice rarely exceeds 0.4 m in thickness at the base of the ice. The relative thinness of basal debris-laden ice under these thermal conditions poses several intriguing questions when compared with the much thicker debris-ice zones observed beneath coldbased ice masses. The influence of basal ice tectonization, in particular fold structures, complicates basal debris sequences (Boulton and Spring, 1986). The debris provenance usually indicates increasing exotic sources with elevation above the bed (Fig. 3.12). The lowest basal ice and debris bands appear to have been formed and entrained by two possible processes. First, a process of ‘Weertman regelation’ of meltwater in the lee of bed obstacles to form ‘laminated ice’ (Weertman, 1966; Souchez and Lorrain, 1991) or by the formation of dispersed or ‘clotted’ ice typical of subpolar ice masses (Lawson, 1979b; Hubbard, 1991); and second, by ‘net basal adfreezing’ in which meltwater penetrating into the surface of saturated sediment of rock materials refreezes at the freezing front and thus envelops both debris and ice as part of the mobile basal ice. What is apparent from any study of wet-based conditions is the delicate balance between melting and freezing coupled with the pressure melting point of ice. Where deviations in stress conditions occur owing either to local increases or decreases in stress, ice temperatures fluctuate across the freezing line. In a phase of freezing conditions, meltwater and debris mixtures will become frozen to the sole of the glacier and entrained within the ice. Once this

70

GLACIERS AND ICE SHEETS

PLATE 3.3. Subglacial cavity beneath Nigardsbreen, Norway. Note the debris lying in the leeside of a bedrock rise. Ice moving from left to right.

process is repeated a series of layers or laminae of debris-rich ice accretes to the base of the ice. The thickness of these layers must, however, remain relatively thin since net melting of the ice occurs over time. The concentration of debris within basal

ice exhibits marked variations from layer to layer and place to place in the same ice mass and between differing ice masses. Values of debris to ice concentrations range from 4 to almost 50 per cent debris content by volume.

Mean crystal size on 25 cm2 vertical surfaces

d18 O‰

Air bubbles

Volume concentration in 0.5m, 0.1m, & 0.01m thick units

11

12

BUBBLY ICE

8

11 10 9 8 7

6

6

SILT & FINE SAND-RICH DEBRIS STRIATED & FACETTED BOULDERS

7

5 4 BUBBLE-FREE ICE

3 2 1 0

-14

-12

-10

0.1

1.0

10.0 mm

0.001 0.01 0.1

1.0

10%

4 3 2 1

Derived Sediment Mass

-16

5

Height in metres above base of ice cliff

Height in metres above base of ice cliff

10 9

(BOLTON 1970)

0 - 1.5 1.5 - 2.6 2.6 - 3.3 3.3 - 5.3 > 5.3 km km km km km SILT-POOR GENERALLY ANGULAR BOULDERS WITH OCCASIONAL FACETS & STRIAE

12

Travel distance & no. of clasts in 2m3 of ice local marine sediment

0 5 0 5 0 5 0 5 0 5 10

0

FIG. 3.12. Debris content and other characteristics of basal ice exposed in an ice cliff at the margin of Aavatsmarkbreen, Spitsbergen. Note the basal 6 m of ice are thought to be regelation ice (after Boulton, 1983; reprinted with permission of the author from Climate Record in Polar Ice Sheets, edited by G. deQ Robin, 1983, Cambridge University Press).

72

GLACIERS AND ICE SHEETS

PLATE 3.4. Ice cliff of the southern lobe of the Barnes Ice Cap, Baffin Island, revealing debris bands and folds (far left centre) as it terminates in Generator Lake (photograph taken 1969, courtesy of Gerry Holdsworth).

3.7.3.2. Cold-based ice conditions Cold-based ice masses are of two types: (a) those completely frozen (polar) and below pressure melting point throughout, and (b) those with frozen terminal zones but temperate wet-based interiors (subpolar). In the latter case frozen conditions tend to be climatically controlled. Typically, under coldbased ice conditions a sharp bed contact exists without cavities forming. Basal interface temperatures are well below the pressure melting point. The ice at the base is often white, debris free and composed of fine-textured ice crystals. The ice crystal size may be the result of either basal regelation in the past or from n´ev´e field snow burial down the ice flow line. Debris that is detected is thought to result from the burial of atmospheric dust particles (Koerner and Fisher, 1979; Thompson and Mosley-Thompson, 1982). From a Devon Island Ice Cap core, an amber ice zone containing clay and pebbles was found 10–60 cm above the bed. The emplacement of this debris, as pockets or ‘clots’ within a polar glacier (Ice Facies 5 and 6), points to past temperate-based thermal conditions and to the probability of subglacial ‘freeze-on’ as the dominant means of debris incorporation. A related phenomenon is the inversion observed between younger and older ice at depth within the Greenland and Antarctic Ice Sheets. Boulton and Spring (1986) demon-

strated that polar and subpolar ice masses have occasionally undergone basal ice tectonism that has led to over-, recumbent and thrust folding of ice layers and enclosed debris (Fig. 3.12). Basal ice tectonism explains inverted ice stratigraphies and the presence of debris-rich ice at high levels above the bed within polar ice masses.

3.7.3.3. Polythermal-based ice conditions A growing body of data reveals that many ice masses have or have had varying thermal conditions at their base. The Barnes Ice Cap on Baffin Island, Canada, is an example of such a type, exhibiting today a wetbased interior and a frozen cold-based frontal zone. Ice masses with past or presently fluctuating thermal basal conditions can be classified into three major subgroups under the general heading of polythermalbased ice masses (Fig. 3.13): (a) ice masses that are now polar cold-based but at some time in the past had warm-based bed conditions (e.g., Camp Century, Greenland Ice Sheet (Herron and Langway, 1979)); (b) ice masses that are now temperate wet-based but in the past had cold-based bed conditions (e.g., Byrd Station, West Antarctica (Gow et al., 1979));

GLACIERS AND ICE SHEETS

73

WET-BASED GLACIERS Dominant Ice Mass Form Ice Sheets

Surgin., Ice Masses Ice Streams

+ positive

100

Type B

Type C

THERMAL FLUCTUATIONS

10

Type C

P.M.P.

0

10

- negative 100

Type A

Cold-based

COLD-BASED GLACIERS 0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

POSSIBLE DISTANCE SEPARATION OF AREAS OF BED (km2)

Valley Glaciers

Wet-based

150

POSSIBLE TIME (years)

FIG. 3.13. Classification of polythermal subglacial bed types and their relationship to differing ice mass types.

(c) ice masses that have at present polythermal-based bed conditions. (e.g., Glacier d’Argenti´ere, French Alps (Goodman et al., 1979)). This classification explains the position and presence of debris within ice masses resulting from past or present thermal regimens and subsequent transport pathways. Ultimately, these critical transformations in thermal regimen are ubiquitous in affecting glacial sedimentological processes and glaciolithofacies environments. Polythermal Type A: from ice cores, retrieved from beneath the polar-based Greenland Ice Sheet at Camp Century, evidence reveals debris encased within basal ice to a height of 15.7 m above the bed (Fig. 3.14) (Herron and Langway, 1979). The debris appears to have been glacially abraded and has a distribution of pitted angular to smoothed surrounded particles (Whalley and Langway, 1980). The average volume

concentration of debris was 11.9 per cent. To explain this debris incorporation, it is thought likely that wetbased temperate ice conditions prevailed in the past. It is thought this debris is of basal origin and ‘frozen on’ to the ice sole during the Pleistocene. As a consequence, these ice masses have considerable volumes of englacial debris found at high levels within the ice. Polythermal Type B: on penetrating the West Antarctic Ice Sheet at Byrd Station, Gow et al. (1968, 1979) found temperate wet-based conditions. Debris was found up to 4.83 m above the bed (Plate 3.5) in volumetric concentrations of <7 per cent. The debris was well stratified, ranging from silt to cobble-sized particles. Typically, the finer debris occurred as discrete ‘mud clots’, not exhibiting boudinage as might be expected if shearing had been involved either in the entrainment mechanism or englacial deformation. The layered nature of the debris may

74

GLACIERS AND ICE SHEETS

of less than 1–10 km and/or over periods of less than a decade. Conversely, where thermal fluctuations range over tens of square kilometres of bed and/or persist for periods of decades polythermal beds of Type A or B may be found (Fig. 3.13). The attributes of Type C basal ice conditions conform to our present understanding of wet- and cold-based ice masses. However, in a Type C ice mass, it can be expected that debris content should be within a few centimetres of the glacier sole, likely dominated by locally derived materials with an englacial transport ‘duration’ of a few years rather than decades, compared with Types A and B.

Clay-Size Fraction

40 20

Per cent of of Total Total Debris Percent

0 Silt-Size Fraction

60 40 20 Sand-Size Fraction

80 60 40

3.8. ICE STRUCTURES

20

3.8.1. Primary

0

0

2

4

6

8

10

12

14

16

18

Distance from Bottom (m) FIG. 3.14. (a) Graphical representation of Camp Century ice core, Greenland, D–D⬘ represents the ice sheet divide; C–C⬘ represents the Camp Century core; and B–B⬘ represents the debris-regelation ice facies. (b) Debris particle size distribution in section B–B⬘ (after Herron and Langway, 1979; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 23(89), 194, 198, figs 1 and 3, 1979).

In wide and relatively flat accumulation areas the surface transformation, remelting and refreezing of snow and ice produce a distinctive stratification. This form of quasi-horizontal banding can be distinguished by layers of coarse bubbly ice (winter snow) alternating with coarse clear ice (refrozen meltwater) (Plate 3.6). The influence of ice flow deformation is virtually absent from such structures. 3.8.2. Secondary

indicate past episodic periods of freezing and melting, or alternate phases of freezing of debris-rich and debris-poor mixtures of meltwater. In Type B beds debris is being gradually released through melt-out from the glacier bed while in contrast, in Type A beds, debris incorporation, over wide areas of the glacier bed, is virtually zero and englacial debris at height is rare. Polythermal Type C: as the number of observations and our understanding of subglacial environments increase, it seems likely that almost all ice masses are of the Type C form, with Types A and B being only end-members of a fluctuating thermal system. Type C polythermal beds exist where spatial and/or temporal fluctuations at the bed occur either over adjacent areas

Where internal ice deformation occurs glacier ice exhibits several characteristic structures. Deformation may be further induced by changes in bottom topography, valley wall constrictions, ice stream bifurcation or tributary addition. The major secondary structures are surface and bottom crevasses, folds, faults and ogives. 3.8.2.1. Crevasses Crevasses form due to tensile stresses developing within the ice, overcoming the internal yield strength of the ice that, in essence, is the result of ice moving at a velocity in excess of its normal strain rate. Crevasses may range from minute cracks a few

GLACIERS AND ICE SHEETS

75

PLATE 3.5. Surface debris bands revealed along the edge of a large crevasse on Omsbreen, Norway.

millimetres across to major fractures several metres wide (Fig. 3.15; Plate 3.7). It can be shown from Glen’s Flow Law (Chapter 4, Equation 4.5) and assuming dry (no meltwater) conditions that crevasse depth is limited to between 25 and 35 m; below 35 m the pressure of internal deformation creep acts to seal the fracture. Where meltwater is present and flows into a crevasse (moulin) the influence of water turbidity, meltwater heat, frictional heat and erosive effects cause the crevasse to remain open to considerable depths. In winter,

without meltwater, these crevasses may seal up by in situ freezing of meltwater or internal deformation on becoming dry (Plate 3.8). 3.8.3. Non-laminar Flow in Basal Ice 3.8.3.1. Folds, fractures and thrust planes Basal ice flow is often neither parallel to its bed nor ‘laminar’. Owing to inhomogeneities and rheological anisotropy within basal ice resulting from debris

76

GLACIERS AND ICE SHEETS

PLATE 3.6. Example of regelation banding within an englacial tunnel (Saskatchewan Glacier, Alberta/British Columbia). Wooden plank on floor of tunnel is approximately 15 cm. "COMPRESSIVE" FLOW

NO LONGITUDINAL EXTENSION OR COMPRESSION

PLATE 3.7. Crevasses on an outlet glacier from Highland Ice Field, east-central Baffin Island (photograph courtesy of Gerry Holdsworth).

"EXTENDING" FLOW

Flow crevasses

principal strain directions

foliation

strain ellipses

FIG. 3.15. Types of crevasses and crevasse patterns.

GLACIERS AND ICE SHEETS

77

PLATE 3.8. Seracs and wave ogives on an outlet glacier from the Highland Ice Field approximately 50 km northeast of the Barnes Ice Cap, Baffin Island (photograph courtesy of Gerry Holdsworth).

content, ice crystal fabric and temperatures, and bottom topographic irregularity, basal ice flow often exhibits non-laminar flow. Observations made in subglacial tunnels or in natural ice cliffs show that glacier ice is often folded (Plate 3.9). Folds develop, in general, because of local stress concentrations or inhomogeneities within the ice.

confined glaciers or on outlet glaciers of ice sheets. Several explanatory hypotheses have been suggested: (a) seasonal pressure waves; (b) longitudinal stretching of ice through passage via ice falls; and (c) buckling.

3.9. CONCLUDING REMARKS 3.8.3.2. Ogives On many glacier surfaces transverse bands of debrisrich ice (ogives) can often be observed. Ogives are of two types: (1) Forbes-bands formed of alternating light and dark ice; and (2) wave ogives formed of trough and ridge construction (Plate 3.9). Ogives tend to form at the rate of one per year at ice falls on some

It is imperative that the theoretical and field observations of glaciology be tied closely to glacial sedimentological processes if a more accurate understanding of the latter is to be finally achieved. Until relatively recently glaciological processes have been largely left separate from any attempts to comprehend the basic sedimentological processes of the glacial

78

GLACIERS AND ICE SHEETS

(b)

PLATE 3.9. (a) Folds on the surface of an unnamed glacier G¨orner region of Switzerland. (b) Surface foliations on an alpine glacier east central Baffin Island (photographs courtesy of Gerry Holdsworth). (a)

environmental system. It is evident, for example, that the response of an ice mass to a series of kinematic waves and their rate of translation through an ice mass, down the flowline, has fundamental effects

upon the whole glacial system from meltwater activity to basal ice stresses to the mobility or otherwise of a potentially mobile bed and, ultimately, to the resultant subglacial bedforms produced.

4

ICE FLOW AND HYDROLOGY J. Menzies dimensional stress/strain conditions within ice masses are symptomatic of ice conditions.

4.1. INTRODUCTION Glacier movement is invariably imperceptible yet repeated visits to the same glacier indicate that an almost relentless process of motion occurs. Motion is apparent in how the glacier surface changes, crevasses appear while others close up, new debris arrives at the surface, moulins enlarge and fresh moraines are builtup. Without question it is an environment of enormous change driven by the dynamics of ice flow. This chapter will review ice motion and the hydrology of the glacier systems and the characteristics and hydrodynamics of meltwater and its pathways within glacial systems.

4.2.1. Rheological Behaviour of Ice Natural glacier ice is crystalline and may be either mono- or polycrystalline, usually the latter. The major difference between the two types is that polycrystalline ice is composed of a range of differently oriented

4.2. ICE MECHANICS AND THE THERMO-MECHANICAL PROBLEM In an endeavour to integrate the field of ice rheology into ice physics, Hutter (1983) pointed out the inadequacy of past approaches to ice mechanics and the thermo-mechanical nature of ice flow. Previously, attention had focused upon either stress/velocity or temperature fluctuations and their interrelationships. It is apparent, however, that ice flow and temperature distribution problems require an integrated thermomechanical solution. Although two-dimensional thermo-mechanical solutions are available for polar or temperate ice masses, no adequate solution(s) exist(s) yet for polythermal ice masses. Complex three-

FIG. 4.1. Stress–strain curves for non-basal glide of single ice crystals with orientations of 45° and 60°. The tensile axis lies in the basal plane (after Hutter, 1983; reprinted by permission of Kluwer Academic Publishers). 79

80

ICE FLOW AND HYDROLOGY

u

au ku

lu

u ea ek

eu

P

S

T

t

FIG. 4.2. Creep curves in polycrystalline ice from primary, secondary, tertiary creep. Various forms of deformation are shown where ␧ is pure elastic deformation, v is pure plastic deformation, ␣ is deformation caused by an increase in the number of dislocations in the ice and a corresponding decrease in ice viscosity (strain softening), ␬ is deformation accompanied by micro-cracking activity in the ice, which also decreases the apparent viscosity of the ice, and ␦ is deformation associated with decreased viscosity as a result of syn-tectonic recrystallization (after Michel, 1978; reproduced with permission of Les Presses de L’Universit´e Laval).

single crystals (Plate 2.1) and therefore any rheological understanding of polycrystalline ice is an average computed value based upon the relative amounts of single crystal types. The stress–strain curve of ice exhibits three creep ‘states’ (Fig. 4.1) and has been described, in the first instance, by the equation known as Glen’s Flow Law (Glen, 1955)(Eq. 4.5). Polycrystalline ice under stress exhibits linear viscoelastic behaviour. Generally, polycrystalline ice is regarded as isotropic but, under the effects of recrystallization, especially at the base of ice masses, it becomes strongly anisotropic. Polycrystalline ice under stress reveals that, as ice deforms, several interrelated processes occur: dislocation climb; grain boundary slip; cavity formation at grain boundaries; polygonization and recrystallization (Fig. 4.2) (Michel, 1978). 4.2.2. Stress Conditions Within Ice Masses Stress is transmitted within any ice mass through intergranular ice crystal contacts, and through bottom and side drag effects. In the vertical plane, normal

stress (σi ) at a point within the ice is a function of ice density and ice thickness above that point (Fig. 4.3). Where water-filled intercrystalline or basal cavities, joints or veins exist, an effective stress (σi –ρw ) can be envisaged, where ␳w is porewater pressure. In the central vertical plane of an ice sheet it is assumed that there is no shear stress (τ). Shear stresses develop as the ice deforms toward its margins, in simple shear, under its own weight. In a two-dimensional model, where consideration is given only to a parallel-sided ice slab, it can be shown that simple shear stresses vary with depth and ice velocity from a value of zero at the upper surface to a value at base computed as: ␶xy = ␳i ghi␣

(4.1)

where ␳i is the density of glacier ice, g the acceleration due to gravity, hi the thickness of the ice mass, and ␣ the surface angle of slope to the ice mass. Ice flow is assumed to be laminar and therefore parallel to the boundary surfaces of the ice slab. Increasingly, observation and research reveal this latter assumption to be untenable. However simplistic this two-dimensional model is, several important implications for ice flow are revealed. First, it is assumed that the value τ can be calculated from known values of ␣ and h, provided ice exhibits perfect plasticity and a ‘standard’ ice yield strength of approximately 100 kPa (100 kPa = l bar) applies. Secondly, it can be seen from Eq. (4.1) that basal shear stresses (where ␶xy = ␶b ) are assumed to be a function of the surface profile of an ice mass and not bottom topography. The value of α is generally taken to be very small and approximately constant over large areas of an ice mass. However, at the large scale, over distances >20h, it can be assumed that ␶b is relatively constant; whereas at intermediate horizontal distances of 1–4h, the influence of bottom topography and gradient becomes significant and the relationship shown in Eq. (4.1) breaks down. At the local scale, with horizontal distances of 艋h, a derivative equation of Eq. (4.1) is necessary where the bottom topographic gradient is integrated into the equation. Since many glaciological problems are of a ‘local’ nature, a more sophisticated three-dimensional approach is demanded. Thirdly, in the idealized case, deformation

ICE FLOW AND HYDROLOGY

81

Debris bands in active ice

True surface

Smooth surface

us angle

a

hi

Smooth bed

ub True bed angle

b Smoothed surface

Sandur plain Subglacial Diamicton

Flow till

Debris band in stagnant glacial ice

Deformable Bed

sb sin b

True bed

b

sb

Buried stagnant ice

sb cos b

FIG. 4.3. Model of ideal ice sheet-glacier. Dashed lines illustrate real top and bottom interface surfaces (adapted from Boulton, 1972 and Drewry, 1986).

within the ice is assumed to occur in only its lowest layers. In reality, ice flows three-dimensionally, thus shear stress conditions are not as simply defined. Finally, ice flow within confined valleys is inhibited owing to side-wall friction and the shape of the conduit. Basal shear stress calculations in these conditions are incomplete. Assuming continuity of flow discharge, it can be expected that over short distances of a few kilometres where valleys become constricted and narrow, ice velocities and basal and side-wall stresses will increase, and vice versa. A similar state will also occur as a boundary effect in ice streams. Stresses within ice shelves at basal boundary surfaces are zero. Only where the ice shelf impinges on the coast, ice rises and up-ice of the grounding-line are stress fields of significance. 4.2.3. Glacier Bed Stress Conditions Our understanding of stress level fluctuations and stress fields at glacier beds remains imprecise.

Glacier beds are extremely complex and exhibit transient temporal and spatial fluctuations of thermal, sedimentological and glaciological conditions. Over an area of bed the impact of meltwater, varying thermal regimens, ice pressures, bed topography and debris all collectively, singly or in variable association alter the local stress fields. The development of local, site-specific stress fields, in turn, affect everradiating stress conditions across the glacier bed. Nevertheless, normal stress levels acting in the vertical plane can be described by the following expression: ␴i = ␳i ghi

(4.2)

Significant stress levels are only likely to develop where no water is present. In general, this is a rare state beneath an ice mass except where ice is frozen directly in contact with minute asperities on the bed. Under these circumstances very high local stress levels capable of overcoming intact yield strengths of

82

ICE FLOW AND HYDROLOGY

bedrock can be achieved. Typically, normal stresses are effective stresses defined as: ␴␧ = (␳i ghi – ␳w )

(4.3)

Normal effective stress levels develop as a function of the presence of free meltwater at the glacier bed, or porewater within subglacial debris, or a combination of both. Values of normal effective stress lie within a broad range from <20 kPa to >101 kPa. Values of ␴␧ beneath Ice Stream B, West Antarctica have been reported in the range 50 ± 40 kPa (Blankenship et al., 1987); while beneath the Blue Glacier values of ~1100 kPa have been described (Englehardt et al., 1978). When the bed becomes locally ‘drowned’ (i.e., no contact points exist between the ice and its bed owing to the presence of an incompressible water layer or film) effective stresses approach zero. Under these circumstances a state of potential ice mass instability is reached (Shoemaker, 1992a). With the ice mass locally decoupled from its bed, shear stresses will also disappear and basal ice velocity should rapidly increase. However, as Weertman (1979) has demonstrated, this unstable condition is self-limiting if the source of water is totally derived from geothermal and frictional melting at the base. As the water layer increases in thickness, production of continued meltwater as a result of melting owing to ice–bed frictional energy will reduce. In the longer term, a consequent thinning of the water layer will occur and a return to stable conditions. Clearly, where water is obtained from other sources, especially the supraglacial environment or up-ice impounded meltwater, a state of instability resulting from bed asperity ‘drowning’ may lead to some form of catastrophic ice advance, enormous subglacial flood discharges (Shaw, 1988a) or vast and sudden short-lived subglacial cavity enlargement (Walder, 1986). The simple basal shear stress developed at the base of an ice mass can be defined as follows: ␶b = ␳i ghi␣

(4.4)

Basal shear stresses are found to range from 50 to 150 kPa and an average value of 100 kPa is assumed in calculations. However, under ice streams, outlet glaciers and ice masses with deformable beds, values

as low as 4–20 kPa have been suggested (Boulton and Hindmarsh, 1987; Ridky and Bindschadler, 1990). The impact of shear stresses, over a small area of a few square metres or across minute asperities on the bed, can be considerable. Generally, both tensile and compressive stresses much higher than the average are involved in any one given instant or location. Where the shear strength of materials is overcome, failure and erosion of the material results. Material may, as a consequence, become either entrained within the ice (englacial), as a tractive load (subglacial), squeezed (intruded or extruded) or pushed into cavities, depressions and cracks of underlying materials. 4.3. THE FLOW OF ICE Ice flows as a viscoplastic material under the influence of gravity, and is opposed by boundary interface friction. Ice sheet flow is unconfined by topography or competing ice mass units; whereas valley glaciers are confined and side-wall friction is significant. Ice shelves are confined within coastal embayments but overall have zero basal friction. There are exceptions to the above differentiation; for example, within ice sheets, fast-moving ice streams are influenced by side-wall friction owing to slowermoving adjacent ice. During surges of an ice sheet or valley glacier basal friction values, at least locally, may approach zero and, where piedmont glaciers emanate from confining mountainous terrain, a valley glacier may have little side-wall friction. Further distinctions that can discriminate all three types of ice mass are variations in mass balance, response time, and the location and possible spatial fluctuations of the equilibrium line. The physical principles of ice flow for all three are identical; only the overall behaviour, growth, surface morphological expression, hydrological characteristics, sediment entrainment mechanisms, subglacial processes and proglacial processes and forms vary in frequency and magnitude depending upon type and location of ice mass. 4.3.1. Longitudinal Strain Rate (Extending and Compressive Flow) As glacier ice moves under the influence of gravity and the internal yield strength of the ice is over-

ICE FLOW AND HYDROLOGY

come, a longitudinal strain rate develops. The strain rate gradually increases through the accumulation area to a maximum at the equilibrium line beyond which, in the ablation area, the effects of ablation cause the overall ice body to lose mass and the strain rate to decelerate toward zero at the glacier snout. This characteristic longitudinal attenuation and reduction is termed extending and compressing flow, respectively (Hutter, 1983). The appearance at the margins of valley glaciers of the corpses of mountain goats, and, infrequently, ill-fated mountaineers and, recently, a lost medieval traveller, all of whom have gone astray and/or fallen by accident into crevasses high in the accumulation areas of Alpine glaciers in Europe, attest to the slow efficacy of ice flow pathways (Ambach et al., 1991). Typical longitudinal strain rates are of the order 10–1 a–1 in temperate valley glaciers to 10–5 a–1 for the central parts of Antarctica (Paterson, 1994). One of the striking features of surging glaciers is the sudden increase in longitudinal strain rates immediately prior to surge initiation. The importance of these flow types and stress states is two-fold. First, under extending flow conditions, ice movement is generally toward the glacier bed while under compressive conditions ice flow is away from the bed. At the macroscale these directions of ice movement have important ramifications for subglacial depositional and erosional processes. Where ice is moving toward the bed, conditions suitable for melting and debris release from the ice occur. When ice flow is away from the bed, processes of regelation and debris entrainment can potentially result. These flow processes, when considered in a local context, allow for sediment distribution and re-distribution at the ice–bed interface. Secondly, these flow conditions permit the differentiation of long-distance transport pathways of glacially entrained sediment (Menzies, 1995, chapters 15 and 16). 4.4. THE FLOW OF ICE SHEETS Ice rheological behaviour can be subdivided into thermal and mechanical components. At present this artificial division is utilized for simplicity’s sake but, as our understanding of ice masses and their polythermal nature increases, there is a growing need for

83

the development of an overall thermo-mechanical approach (Blatter and Hutter, 1991). Ice masses flow across the Earth’s surface as a result of three processes, all of which may occur either independently or in concert. These mechanisms are: (a) internal deformation; (b) basal slip or sliding; and (c) ice mass movement owing to a deforming bed. The first two processes are well established and discussed in the literature; the third process is of recent derivation and its pervasiveness or otherwise remains to be established. In any one ice mass all types of flow are possible either in different parts of the ice mass or at different times depending upon basal thermo-mechanical conditions, topographic constraints and seasonal variations. 4.4.1. Ice Sheets and Bed States Ice masses can be subdivided into those that have Hard or Rigid Beds and those with Soft, Mobile or Deformable Beds (Hutter and Engelhardt, 1988; Menzies, 1989a,b). In the former case, it is assumed that the glacier bed beneath the ice is rigid, smooth with zero permeability and is undeformable. This case is the one most often modelled in the past. These beds are likely composed of unfractured bedrock of high intact shear strength or, if sediment-based, frozen and again impenetrable to meltwater (H-beds). Under these conditions, if the ice is temperate, meltwater will move along the single basal interface and thus motion of the ice mass is likely to be by basal slip. In the case of polar ice masses, the ice will adhere to the bed and internal deformation is likely to account for glacier motion. Where soft Mobile beds occur (M-beds), meltwater can penetrate the underlying sediment to a lesser or greater degree and the debris may be mobilized such that ice motion is largely accounted for by this underlying mobile bed. In this case both temperate and polar ice conditions seem possible with limited upper interface meltwater present. Unlike the H-bed situation, several interfaces may temporally exist beneath the ice mass, acting as shear zones to carry the pervasive or non-pervasive shear stress imparted to the mobile sediment by the overlying ice mass. It is likely beneath any ice mass both bed types may exist at differing times but in the same location. Under these common conditions,

84

ICE FLOW AND HYDROLOGY

a Quasi-hard/Mobile Bed state is likely to prevail (Q-beds) where all three forms of glacier motion may periodically occur. 4.4.2. Internal Deformation As snow transforms to glacier ice, the ice begins to deform internally under the influence of gravity. This gravitation motion is described typically as a downward concave flowline within the accumulation area and upward concave beyond the equilibrium line in the ablation area (Fig. 4.4). In order for ice to move, stresses must be generated that first overcome the shear strength of its own internal crystalline structure (yield strength) (Fig. 4.2). From that point, beyond the yield strength of the ice, the ice deforms under stress as a viscoplastic material (Van der Veen, 1999). The process of deformation within glacier ice consists of a series of intercrystalline dislocations. Since glacier ice crystals are anisotropic, irregular, with lattice defects, slippage of planes of atoms over each other is possible at even very low applied stresses. This process of slippage is termed creep. Creep occurs in two forms: (a) creep relatively insensitive to confining pressure, and (b) creep that is prevented by high confining pressures. Initially, dislocation movements appear to be few but gradually increase by a positive feedback

R1 =

dm dx

x

R1

R2 =

-ve

dm dx

mechanism leading eventually to a ‘piling-up’ of dislocations possibly at ‘point/edge defects’ and a hardening of the ice. Within secondary deformation (Fig. 4.2), glacier ice deforms approximately to the following expression now commonly termed Glen’s Flow Law:

␧˙

= A␶ n

(4.5)

where ␧ ˙ is the average uniaxial strain or creep rate, ␶ is the internal shear strength of polycrystalline ice, both n and A are constants dependent upon temperature, ice crystal size, orientation and debris/ impurity content, and confining pressure. Considerable debate exists as to the values of A and n, the latter parameter varies from 1.5 to 4.2 to a mean ~3 (Paterson, 1994) (Fig. 4.5). Two aspects of ice flow related to ice dynamics must be considered. First, ice movement is temperature dependent. Therefore, at the base of an ice mass where most vigorous internal flow occurs, the lower the temperature the slower the deformation and vice versa. Secondly, internal ice movement is influenced by confining tangential stresses. Where tangential shear stresses increase because of increasing ice thickness and/or flow constriction, an exponential increase in ice mobility occurs with the lower

+ve

ELA ELA

m

Bed Bed

(a)

-ve

x

ELA +ve

+ve

m R2

-ve

Bed Bed

(b)

FIG. 4.4. Models of a valley glacier (a) and ice sheet (b) showing flowlines and longitudinal strain both positive (extending flow) in accumulation areas, and negative (compressive flow) in ablation areas. R1 and R2 are longitudinal strain in accumulation and ablation areas, respectively; d␮ and dx are flow vectors in vertical and horizontal directions, respectively; and ELA is the equilibrium line altitude.

M JeBud athew ns d s se & n

cr

im

in

(m

ns

& Sa

ot d

har

rson Pate

an

&S

rd ra er

eve

d

Kamb

& Sh

mon Ray

-2

reve

Ny

e

Shr

-1

vage

he

rs

&

lb

Co

p

k

ec

a Ev

G

. Log ( / yr -1)

Glen

p)

ee

um

(And

Gle

0

rade

n (m

inim

's La

w)

um

cre

ep)

ICE FLOW AND HYDROLOGY

-0.5

0 Log ( / bars)

0.5

FIG. 4.5. Relationship of shear strain rate and effective shear stress for ice from laboratory experiments (Glen, 1955; Colbeck and Evans, 1973); analysis of borehole deformation (Mathews, 1959; Paterson and Savage, 1963; Shreve and Sharp, 1970; Kamb and Shreve, 1966; Kamb 1970; Raymond, 1973); surface velocity (Budd and Jenssen, 1975) and tunnel closure (Nye, 1953) (after Raymond, 1980; reproduced with permission of the author and Academic Press).

zones of any ice mass becoming increasingly mobile. This basal elevated mobility may translate into increased basal friction leading to high rates of melting at the basal interface 4.4.3. Basal Slip In ice masses where ice is at pressure melting point (temperate, warm-based), ice not only moves by internal deformation creep but also by basal slip. The term ‘basal slip’ characterizes the process whereby an ice mass slips over a very thin lubricating layer or sheet of meltwater, or series of interconnected waterfilled cavities at the ice–bed interface (cf. Lliboutry, 1979; Weertman, 1979, 1986). With ice at pressure melting point basal ice deformation results in basal melting and the production of lubricating films or water-filled channels. The ice mass therefore changes its boundary conditions from non-slip to perfect

85

sliding. In considering basal slip, two distinct models can be examined: (1) sliding without separation (no cavities), and (2) sliding with separation (cavities). Basal slip is not a continuous movement but consists of a ‘stick-slip’ or jerky motion. This form of motion is attributable to several causes. First, obstacles at the ice–bed interface may cause localized reductions in the pressure melting point and subsequent freezing and immobility. Secondly, if the thin film of water (only a few microns thick) over which the ice mass slips is removed, either by freezing or drainage, localized ‘grounding’ of the ice mass results with subsequent increased longitudinal and transverse stresses and boundary interface friction. Finally, the formation or otherwise of basal cavities leads to marked changes in the value of basal friction, falling locally to zero where cavities form. The main input factors in basal sliding boundary conditions are: basal shear stress, bed roughness, normal pressure from overlying ice, interstitial porewater pressures and perhaps other unknown factors, the result of the lack of accessibility in observing the process. Basal sliding is seasonally controlled, being often higher in the spring, summer and autumn because of greater meltwater production. For discussions on basal sliding with and without cavity formation, the reader is referred to Menzies (1995, chapter 5, pp. 152–158). 4.4.3.1. Basal slip models and polythermal bed conditions Modelling the relationship of glacier bed hydraulics to aspects of basal ice velocity, basal shear stress and bed roughness is complicated further by the introduction of the influence of polythermal bed conditions (Blatter and Kappenberger, 1988; Blatter and Hutter, 1991). Monothermal Models 1, 2, 4 and 6 (Table 4. 1) are complex to model mathematically. They can be viewed as variants of either sliding without cavitation (Model 3) or with cavitation (Model 5) depending upon the competency or otherwise of the glacier bed, bed topography and basal hydraulic conditions. To these models must be added the possibility of polythermal bed conditions. In introducing more realistic bed conditions, the complexity of basal ice–bed rheological influences increase. The possibility that thermal, geotechnical (deformable,

86

ICE FLOW AND HYDROLOGY

TABLE 4.1. Classification of various ice mass types related to thermal bed conditions, the presence or absence of cavities and/or deformable bed conditions, and the likely bed motion of the ice mass Model

Ice mass type

Bed state

Bed motion

1 2 3 4 5 6

Cold Cold Warm Warm Warm Warm

Frozen to non-deformable Frozen to deformable P.M.P. no cavities non-deformable P.M.P. no cavities deformable P.M.P. cavities non-deformable P.M.P. cavities deformable

No basal slip No basal slip Basal slip Basal slip Basal slip Basal slip

P.M.P., pressure melting point.

In 1974, Boulton et al. reported a simple experiment, carried out beneath an Icelandic glacier, in which differential movement within subglacial sediment was observed. Building upon this work, Boulton and Jones (1979) suggested that some proportion of the forward motion of ice sheets might be accounted for by the existence of a deforming subglacial debris layer (a soft or deformable bed). Motion may be entirely by advective deformation within the subglacial debris layer or by ‘ploughing’, a process that is intermediate between basal slip on a lubricated interface and debris deformation. Beneath Brei∂/ amerkurj¨okull, Iceland, Boulton et al. noted that 88 per cent of the basal movement was due to deformation and only 12 per cent could be accounted for by slip at the ice–bed interface and internal ice deformation. Since then considerable evidence in the form of radar images from beneath Ice Stream B, West Antarctica where movement by the ice stream may be 100 per cent accountable to bed deformation (Alley et al., 1986, 1987a,b; Blankenship et al., 1986, 1987), and observations beneath many present-day ice masses, for example, at Storglaci¨aren, Sweden (Brand et al., 1987); Urumqui Glacier, China (Echelmeyer and Wang 1987); the Variegated Glacier, the Matanuska Glacier, the Columbia Glacier, Alaska (Kamb et al., 1985; Fahnestock and Humphrey, 1988; Meier, 1989) have noted or inferred the presence of deforming

t b = gs + hb es

hb

es

gs

4.4.4. Glacier Movement and Bed Deformation

subglacial debris layers. Discussion of the mechanics of deformation (Murray, 1994) indicates that the controlling variables influencing the thickness and mobility of a deforming layer are: porewater content and pressure; sediment yield strength, viscosity, porosity and permeability; ice mass thickness, basal sliding velocity; and basal ice interface thermal regimen. If it is assumed that this debris layer acts as a Bingham viscoplastic or non-linearly viscous material, then, once yield strength of the wet debris is surpassed, the critical variable controlling debris layer thickness and deformation rate is viscosity (Fig. 4.6). Once a subglacial debris layer deforms, the overlying ice sheet will flatten out owing to forward motion of the basal deforming layer rather than internal ice deformation or basal slip of the ice mass, such that the ice sheet

tb

non-deformable) and hydraulic conditions may vary markedly across the glacier bed introduces geological conditions hitherto unconsidered.

=

t b - gs hb

es FIG. 4.6. Relationship between shear strength and strain rate in a Bingham viscous slurry (after Menzies, 1989a; reproduced with permission from Elsevier Science Publishers).

ICE FLOW AND HYDROLOGY

thickness (hi ) will be related to subglacial debris deformation as follows: h 2i = 2

冕

xf

x

冢 ␳ G + tan ␾ 冤h – ␳ g 冥 冣 dx pw

c

i

i

(4.6)

i

where xf is the distance from a point beneath an ice mass to the terminus in a two-dimensional ice mass; ␾ is the angle of internal friction of the debris layer sediment, and C is the cohesion of the sediment. Where porewater pressures within the debris layer are less than the overlying ice pressures (pi > pw ), it is thought that regelation ice will invade the debris leading to immobilization. Depending upon the interrelationship between the internal porewater pressure of the underlying sediment and the overburden ice pressures, a steady state or an unsteady state may prevail at the ice–bed interface. In the former instance, two possibilities arise either resulting from the high confining ice pressures or the development of a thin skin of dilatant sediment at the upper interface between the bed and ice mass or deforming bed and bedrock in which zero deformation but positive effective pressures exist within the underlying sediment. Under an unsteady state, zero or negative effective pressures lead to sediment deformation; a state possibly the norm for ice masses overlying unlithified beds. If this latter scenario is typical then the vast percentage (70–80 per cent) of land surfaces covered by the Pleistocene ice sheets were underlain by potentially deformable sediments. As the ice mass thins out and effective stress and basal shear stress reduce, frictional drag at the upper ice–bed interface will diminish, as will meltwater production, thereby causing the mobile bed to reduce in thickness and 10 30 40

50

87

eventually cease to deform, returning the glacier system to a steady state once again. 4.5. THE FLOW OF VALLEY GLACIERS Valley glaciers, in contrast to ice sheets, are confined by valley walls. The walls act to influence and retard valley glacier flow. This contrast is not as great when one considers their similarities to ice streams where an element of sidewall drag also occurs. The physics of ice flow and basal slippage, as described for ice sheets, are essentially similar for valley glaciers (Raymond, 1980). The major difference is in the component of sidewall frictional drag. Raymond’s pioneering work on the cross-sectional transverse velocity fields of the Athabasca Glacier, Alberta, illustrate this effect (Fig. 4.7) Frictional drag may reduce edge velocities by ~80 per cent compared with central maximum surface velocity. In order to account for this frictional effect the basal sliding equation (Eq. (4.1)) alters to: us – ub =

冤

2A (n + 1)

冥

(␶b )nhi

(4.7)

where, in the centre line of the glacier, the basal shear stress can be equated as ␶b ≈ ␶xy 円 bed ≈ –␳i g(fsh hi␣)

(4.8)

where the shape factor (fsh ) is derived as follows: fsh =

S wb hi

(4.9)

where S is the cross-sectional area of the channel, and Wb is the width of the channel bed. Typical shape 40

30 20

100 m

FIG. 4.7. Isolines of equal downstream ice flow velocity observed in the Anthabasca Glacier, Alberta, Canada, are metres per year (after Raymond, 1978, 1981; reproduced with permission of the author and Academic Press).

ICE FLOW AND HYDROLOGY

Filchner Ice Shelf

90 ºW

Ronne Ice Shelf

ºS 80

P.I.G. T.G.

East Antarctic Ice Sheet

t Caof

B A C D

Cary Ice Rise

E F 0

500

Ross Ice Shelf

km Bedrock that is above sea level

0º

Until relatively recently, ice shelves were regarded as minor peripheral curiosities found only in high latitudes. However, there is increasing evidence that ice shelves may have occurred along the fringes of the Laurentide and Fennoscandian Ice Sheets during the Pleistocene and may have acted as major constraints upon ice sheet growth and ultimate decay. In North America, for example, ice shelves probably formed along the eastern seaboard from the Maritime Provinces north to beyond the Labrador coast of Canada (Andrews, 1997). Of prime importance in understanding ice shelf growth and diminution, as well as ice shelf impact upon sedimentological processes, is the position and spatial transience of the groundingline, the rate of bottom melting or accretion, the impact of mass balance changes, and the influence of grounded ice streams entering shelves. An ice shelf is a large body of glacial ice attached to the land but floating on the sea or a large lake. The ice deforms under its own weight and, in theory, bottom and surface shear stresses approach zero, if bottom and lateral drag is ignored, and it is assumed that the ice shelf is a perfect horizontal slab with top and bottom boundaries parallel. In reality, however, ice shelves exhibit variable thicknesses with gradients in the down-flow line direction. Since most ice shelves have been in existence for thousands of years it has been generally assumed that, if climatic and glaciodynamic controlling parameters have not significantly varied over time, those shelves may be in a state of dynamic equilibrium. The converse that many ice shelves are, instead, in a state of considerable instability and non-equilibrium is a view that many

18

4.6. THE FLOW OF ICE SHELVES

researchers now advocate. The latter view has profound implications concerning global warming and the stability of the Antarctic Ice Sheet (Vaughan and Doake, 1996). Any ice shelf, as it extends out to sea, can be regarded as the product of a complex balance of processes that control its extent, thickness and flow conditions, and thus its dynamic equilibrium. These processes involve the glaciodynamics of ice creep and tensile strength, edge drag, surface snow accumulation and ablation, bottom melting and freezing, the position of grounding-lines, ice input from upice and associated effects of glaciers ploughing into the shelf, the presence of ice rises and/or seaward islands, and the impact of ocean currents, storm tracks, waves and tidal fluctuations. Ice shelves are located in a variety of topographic positions: confined or unconfined.

c th hm e e Sh Ro nt el ss are f Ic a e

factor values for parabolic cross-sections vary from 0.5–0.6; V-shaped channels have smaller values and flat-bottomed U-shaped channels larger values. The effect of the shape factor alone on the sliding velocity equation is to reduce the overall velocity by as much as 15 per cent compared with ice sheet flow with similar exponent values. Finally, the impact of deformable beds upon valley glacier movement and stability engenders ongoing debate (Alean et al., 1986; Clarke, 1987a,b).

70 ºS

88

FIG. 4.8. West Antarctic showing areas of bedrock above sea level. Arrows indicate ice streams (after Thomas, 1979; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 24(90), 1979, p. 168, fig. 1).

ICE FLOW AND HYDROLOGY

4.6.1. Confined/Unconfined Ice Shelves

embayment but rather upon the vagaries of tidal movement, currents and wave action. Such an ice shelf extends out into a body of water at a sufficient velocity from feeder ice sources to effectively spread out and remain intact as a floating ice mass. For detailed discussion on unconfined ice shelves see the papers in Van der Veen and Oerlemans (1987).

In confined shelves, edge-support is provided by the land (Fig. 4.8). Today major ice shelves such as the Ross, Ronne and Filchner Ice Shelves in Antarctica are fed by major outlet glaciers from the Antarctic Ice Sheet. The outlet glaciers, flowing at higher velocities than the ice shelves, plough into the shelf, becoming floating ice streams flowing for some considerable distance within the shelf (e.g., Byrd Glacier) (Fig. 4.9). Ice rises, where the shelf locally grounds, occur where bedrock rises up to the base of the shelf or the shelf itself locally thickens and touches the sea bed. Ice rises and rumples cause local impounding and back-pressure up-ice within the ice shelf (Thomas, 1979; Doake, 1987). Unconfined ice shelves are relatively rare since their stability does not hinge upon a topographic

4.6.2. Ice Shelf Stability At the back of an ice shelf is the grounding-line where the ice shelf ceases to float. This line demarcates the point where hydrostatic equilibrium of the ice and sea begins. At this critical junction several major thermal and glaciodynamic phase changes occur at the base of and within the ice mass. In contrast to a grounded ice mass, vertical stresses in an ice shelf are almost at a constant from the upper surface to the base of the

150ºW

80ºS

180ºW 600

Scott Gl.

Ice stream

po

800

se

dr

oc

k

stre

Beardmore Gl.

ea

str m

80ºS

400

300

150ºE

Byrd Gl.

400 400

80ºS

400

0 20

200

Ice front

0

600

ne

400

Roosevelt Island

w li Flo

600

150ºW

0

40

400

800

Nimrod Gl.

Cary 600 Ice Rise

0

80

Ice

600

0

600

40

am

0

Ice

Ex

1000

Isopach (m) Boundary of ice shelf

40

Byrd land

89

ROSS SEA

McMurdo Sound

km W 80ºE

FIG. 4.9. Isopachs of ice shelf thickness for the Ross Ice Shelf. Dashed lines indicate the shelf boundaries (after Robin, 1975; reprinted with permission from Nature, 253, 168–172, 1975, Macmillan Magazines Ltd.).

90

ICE FLOW AND HYDROLOGY

shelf (Sanderson, 1979). The transformation from a grounded to a floating ice mass results in large-scale velocity field changes that are manifest in the thickness of the ice shelf for several kilometres downice of the hinge point. The position of the groundingline is especially susceptible to eustatic fluctuations (and isostatic rebound effects) resulting in potentially large increases or decreases in the amount of ice floating and thus the frontal position of the ice shelf cliff. The relationship of the sensitivity of groundingline positions to ice shelf collapse and climatic change has led to discussion on the potential stability/ instability of the marine sections of the West Antarctic and northern hemisphere Pleistocene ice sheets (Vaughan and Doake, 1996). As sea-level rises, the grounding-line should begin to extend up-ice resulting in an increase in the total area of the ice shelf and therefore increasing forward (down-ice) spreading. This spreading will cause the overall ice shelf to thin and be a less effective buttress for the grounded ice mass behind it. A critical point will be reached when

this buttressing effect will founder and a vast outpouring of ice from the ice sheet, itself, will occur causing possible rapid ice sheet decay. 4.7. VELOCITIES OF ICE SHEETS AND GLACIERS In considering velocity as a characteristic of a particular ice mass it must be clearly stated what time frame is being considered (e.g., a few days, one or two months, or years, several decades or centuries). All ice masses exhibit variations in velocity; some are extrinsic to the effect of local and short duration causes or the result of long-term continental or global impacts, while others are intrinsic to a particular ice mass and are a function of mass balance change and ‘internal adjustments’ within the glacial system to climate, bed topography, meltwater discharge and pressure fluctuations. The velocity of an ice mass is a function of mass balance, flow from internal deformation and, if the glacier bed is at pressure melting point,

TABLE 4.2. Sliding velocitiesa (modified after Paterson, 1981) Glacier

Ice thickness (m)

Velocity (mm day–1)

Reference

Argenti‘ere Blue Blue Blue Casement Grindelwald Mer de Glace Mont Collon Østerdalsisen Variegated Vesi-Skautbreen Susitna East Fork Storglaciaren ¨ Urumqui No. 1 Urumqui No. 1 Urumqui No.1 Ice Stream B (50 km)b Ice Stream B (300 km)c

100 26 65 65 20 40 100 65 40 356 55 520 410 125 12 23 33 1000 2000

600–1200 16 350 10 24 250–370 30–80 30 29–97 250 8 119 236 180–280 11 14 16 1986 137

Vivian and Bocquet, 1973 Kamb and LaChapelle, 1964

a

McKenzie and Peterson, 1975 Carol, 1947

Theakstone, 1967

Hooke et al., 1987 Echelmeyer and Wang, 1987 Echelmeyer and Wang, 1987 Echelmeyer and Wang, 1987 Lingle and Brown, 1987 Lingle and Brown, 1987

Velocities measured may vary by at least a magnitude of 100 depending upon location, season of the year and state of bed conditions (e.g., prior to a surge event or immediately following such an event). During surge events velocities at least ten times higher may occur. Velocity taken 50 km up-ice of grounding-line. c Velocity taken 300 km up-ice from grounding-line. b

ICE FLOW AND HYDROLOGY

additional flow owing to basal slip, and where underlain by deformable sediments a large proportion of ice movement is the result of a combination of both ice and sediment motion. 4.7.1. Surface and Balance Velocities Variations in mean surface velocities (us ) among valley glaciers, ice sheets, ice streams and ice shelves are considerable. Values range from 10 to 200 m a–1 for many valley glaciers, to 250 to 1400 m a–1 for parts of the Antarctic Ice Sheet (Paterson, 1994; Van der Veen, 1999). Jakobshavns Isbræ, West Greenland, is considered the world’s fastest glacier with velocities of ~8.4 km a–1, while several ice streams in Antarctica exhibit velocities of 400 m a–1 (Rutford Ice Stream) to over 800 m a–1 (Ice Stream B). Ice Stream C, however, has a present velocity of 5 m a–1, but in the past 2000 years possibly moved at velocities similar to those of Ice Stream B. Typical sliding velocities are shown in Table 4.2. Short-term velocities recorded during the fast phase of a surging glacier may be exceptionally fast, for example, >3.8 km a–1 for the Medvezhiy Glacier, Tadzhikistan in 1973 (Dolgushin and Osipova, 1975) and >23 km a–1 for the Variegated Glacier, Alaska in 1982–1983 (Kamb et al., 1985); both were peak velocities recorded over a few hours. The distinction as to whether glaciers are fast or slow moving, intrinsically unstable or not, is a major concern in glacial sedimentological processes. In the past, velocity variations have tended to be regarded as a function of basal ice conditions and have resulted in a search for increasingly complex and sophisticated mathematical abstractions of the basic sliding law (Hooke et al., 1983 and references therein). However, it is now apparent that fast velocities are a function of fast sliding and not fast creep; the former the effect of transient subglacial states and boundary interfaces in interaction with glaciological conditions. Depending upon glacier type and ice velocities observed over several years, it is possible to classify glaciers on the basis of their velocities in relation to basal shear stress. As a consequence the ‘state’ of an individual ice mass can be crudely predicted and is of significance from the geological, climatological and engineering viewpoints. Major variations in ice surface and balance

91

velocities may occur on a single ice mass owing to several factors depending on where and when measurements were taken. Other influencing factors causing velocity changes may be sudden increases in the thickness of subglacial water films, basal materials becoming unstable owing to saturation or fracturing, localized surface accumulations resulting from valley wall avalanches or rock-ice falls, or the increase in velocity observed down-ice of a confluent glacier where the side-drag effect is removed. In general, most glaciers react stably to external factors that influence net mass balance, basal movement and therefore balance velocity (see eq. 4.11). However, certain ice masses appear capable of switching from slow to fast mode (Fig. 4.10) both spatially along the length of the ice mass or temporally (Ice Stream C, West Antarctica, Fastook, 1987; Shabtaie and Bentley, 1987). Both ice streams and tidewater glaciers can exhibit marked velocity changes. In the latter case, drastic retreat and rapid upstream migration of the grounding-line can ensue where secular changes in tidal or eustatic levels cause grounding-lines to migrate, or lower sections of the glacier to flexure (Krimmel and Vaughn, 1987; Meier and Post, 1987). 4.7.2. Basal Velocities Basal sliding velocities are not well documented but appear to be approximately 10–20 per cent of the surface velocity (Table 4.1). It is difficult to accurately predict sliding velocities. Where no basal sliding occurs and a polar glacier state prevails, basal ice velocities can be related to surface velocities. Midsummer melt season velocities are usually highest. The influence of diurnal temperature variations, rainfall, surface meltwater discharge, sky cloudiness and geothermal fluctuations all appear to reduce frictional drag at the ice base and manifest as higher than annual mean velocities. These higher velocities are thought to be the result of changes to the subglacial hydraulic system, increased lubrication as a result of decreased basal friction, and higher shear stresses upon those parts of the basal ice that are still in contact with the bed. Velocities may vary between summer and winter by as much as 20–100 per cent. Velocity mode changes may also be rather sudden,

ICE FLOW AND HYDROLOGY

Velocity (m/yr)

92

10000

JB

1000

IB

CG

100

WG VG 10 3 1600

IC 1650

1700

1750

1800

1850

1900

1950

2000

Year FIG. 4.10. Flow speed for various glacier types. The graphs draw upon real data but are partly extrapolated. JB, Jakobhavns Isbrœ, Greenland – a fast outlet glacier that terminates as a tidewater glacier; CG, Columbia Glacier, Alaska, USA – a tidewater glacier that in 1984 began to retreat rapidly; IB, Ice Stream B, Antarctica – a fast ice stream entering the Ross Ice Shelf; IC, Ice Stream C, Antarctica – a slow ice stream that switched from fast to slow mode ~250 years ago; WG, White Glacier, Axel Heiberg Island, Canada – a normal valley glacier; and VG, Variegated Glacier, Yukon Territory, Canada – a surging glacier (after Clarke, 1987c, Journal of Geophysical Research, 92B, p. 8838, copyright by the American Geophysical Union).

especially from winter to summer. However, not all glaciers exhibit this winter-to-summer seasonal acceleration, perhaps because seasonal and related meltwater penetration is not achieved so catastrophically and effectively (e.g., Blue Glacier, Washington). Observations made on Jakobshavns Isbræ (Echelmeyer and Harrison, 1990) show that, unlike many glaciers that demonstrate delayed but direct responses to seasonal climatic changes and fluctuating surface meltwater discharge, this very fast glacier is unaffected by these transient input variables. In the case of valley glaciers and ice streams the influence of conduit geometry and frictional drag from walls and edge effects have to be included in velocity calculations (Echelmeyer and Harrison, 1990). Ice streams typically have very low basal shear stresses (~10 kPa) and may exhibit high basal velocities as a result of deformable bed conditions (M-beds). Fast outlet glaciers, in contrast, appear to have basal shear stresses of ~100 kPa and lie on H-bed types with high free meltwater pressures.

4.7.3. Basal Ice Velocities and Polar Bed Conditions Where ice is frozen to its bed (Polar), internal deformation is the dominant cause of movement. The relationship between basal shear stress (␶b ) and mean basal ice velocity (ub ) can be expressed as follows (Weertman, 1957): ␶b m ub = (4.10) B

冢 冣

where B and m are constants with m = 2. Equation (4.10) is difficult to test as yet because of scarcity of data and therefore the balance velocity (␨) is calculated instead from mass balance and ice thickness data. Balance velocity is defined as: ␨ =

1 ␳¯ i hi

冕

x

Mdx

(4.11)

0

where ␳¯ i is averaged ice density. Until relatively recently Polar ice masses were thought to move

ICE FLOW AND HYDROLOGY

4.7.4. Basal Ice Velocities and Temperate Bed Conditions A similar general relationship exists where ice is at pressure melting point (temperate) at its bed. Basal melting occurs and ice flow is due to both basal slip and internal deformation: ub ⬵

冢B冣 d ␶b

1000

=

1 ih

x

M dx

o

East Antarctica data envelope

-1

)

100

(m yr

exclusively by internal deformation (i.e., they were frozen to their beds). However, studies by Shreve (1984) and Echelmeyer and Wang (1987) among others illustrate that ice masses can ‘slide’ at their base at temperatures below pressure melting point. The predicted sliding velocities of cold-based ice were formulated by Shreve (1984) based upon a model by Gilpin (1979, 1980) inter alia. It was found that regelation could take place at sub-freezing temperatures owing to the presence of microscopic but discrete liquid or liquid-like layers forming around foreign materials.

93

Theoretical line incorporating basal lubrication factor

10

m

w

(4.12)

where dw is the depth of water film (Weertman, 1957). Improvements on the above equation have been developed by Kamb (1970) where: ub = ␶b /(cr • ␩)0.5␭2R

(4.13)

where cr is a constant related to regelation processes and ␭R is the bed roughness term. Both equations for polar and temperate conditions predict increasing basal shear stress with increasing ice velocities. Data supportive of these equations are shown from East Antarctica in Figure 4.11 (Drewry, 1983). However, the values of the balance velocity from West Antarctic ice streams do not fit the relationship, possibly because where fast basal sliding occurs a feedback may exist between dissipation of heat from sliding and basal water production. This fact may explain the past inconsistencies that were shown to exist between theoretical equations and actual observations.

1 5

10

 b (kPa)

100

FIG. 4.11. Relationship between ‘balance velocity’ and basal ice shear stress for West Antarctic ice streams (circles) and for the East Antarctic ice sheet (after Drewry, 1983, reprinted from Gardner, R. and Scoging, H. (eds) Mega-Geomorphology, 1983; by permission of Oxford University Press).

4.7.5. Basal Ice Velocities and Deformable Bed Conditions In recent years it has become apparent that many glaciers and parts of ice masses flow across mobile debris layers (1–6 m). It is thought likely that parts of the Quaternary Ice Sheets were also underlain by deformable beds (Benn, 1995; Marshall and Clarke, 1997a,b; Maher and Mickelson, 1997; Menzies et al., 1997; Hart, 1998). These soft beds act as a substantial lubricating layer causing ice masses to move at higher

94

ICE FLOW AND HYDROLOGY

velocities than would be typical of ice flow solely resulting from basal slip and internal deformation. It has been demonstrated that under soft-bed conditions basal ice velocity (ui ) is a consequence of the merger of the component of velocity attributable to the deforming sediment (us ) and the basal ice itself (ub ) where: ui = us + ub

(4.14)

in expanded form, eq. (4.14) can be rewritten as: ui =

冬

zs Bf

(␶o – ␶s )a ␴e

冭 + 具B␶ d 典 2 b w

(4.15)

where a, B, and Bf are constants and ␶s is the shear strength of the deformable debris. In general, where beds are smooth and/or soft and are spatially extensive, basal ice velocities are high. Under fast ice conditions disruption of subglacial water systems may occur resulting in ineffective meltwater evacuation (Kamb, 1987; Alley et al., 1989c). Smooth beds would appear to be typically undeformable, H-bed types; whereas beds under soft conditions (M-bed types) have limited free meltwater and are rough. Reality is, however, likely to be more complex with both bed types occasionally overlapping spatially and temporally (Q-bed conditions).

4.7.6. Basal Ice Velocities – A Classification Budd (1975) and Hutter (1982) recognized three distinct velocity categories or classes: 䊉

䊉

Class A: ‘ordinary glaciers’ where the ice flux (discharge) for a given bed profile is of a sufficiently low magnitude for the glacier to be in the ‘slow mode’. Flow can be described according to Glen’s Law. Class B: ‘fast glaciers’ where the ice flux is sufficiently high to maintain the ice mass in a steady state in the ‘fast mode’ (i.e., where ␶b decreases with increasing ub and some form of cavitation, as a result of high meltwater pressures, or bed decoupling, as in deformable bed development occurs) (e.g., West Antarctic Ice Streams).

䊉

Class C: ‘surging glaciers’ where a sharp ‘jump’ occurs cyclically between a long-term slow mode to a shorter-term fast mode. This type of ice mass would appear to exhibit characteristics of both Classes A and B with an intrinsic instability that permits such a sharp velocity shift. The key to understanding these ice masses lies in understanding the mechanism that permits and cyclically maintains this inherent instability.

It is possible for glaciers to ‘jump’ from Class A type to Class C with no transitional phase. Significant changes in mass balance (measured in centuries), and/ or bed characteristics (over tens of kilometres) and/or thermal conditions may, singly or in concert, provide the impetus for a glacier to change from one ‘mode’ to another. If such a change takes place, the impact upon sedimentological environments throughout the glacier system would be considerable. Historical evidence seems to point to such a major change in Ice Stream C (West Antarctica) and possibly parts of the Greenland Ice Sheet. The association of deeply incised through-valleys and fiord valleys with ice masses of outlet glacier or ice stream types (Class B) is well established. It has been suggested that surging glaciers (Class C) may lead to enhanced debris incorporation, increased compressive flow in the lower ablation area and rapid entrainment of debris and debris stacking near glacier snouts, thereby producing a distinctive depositional landscape. However, contradictory suggestions have been made that, under very rapid basal sliding with massive areal bed decoupling, bed debris entrainment would possibly be at a minimum and thus a surging glacier may have little or no effect upon its bed. 4.8. GLACIER SURGES Glacier surges have been described as ‘a brutal advance of the glacier front in a short period of time’ (Michel, 1978) or as ‘the most impressive type of instability’ (Paterson, 1980, p. 71) occurring in valley glaciers and ice sheets, ice streams and outlet glaciers (Clarke et al., 1986; Clarke, 1987a,c, 1991). There have been several conferences and review papers devoted to the ‘surging’ problem (see Menzies, 1995, chapter 5, p.179). Ice masses subject to surging

ICE FLOW AND HYDROLOGY

phases typically exhibit long periods of quiescent ‘normal’ flow behaviour that may last for several years, decades or centuries and are then interrupted periodically by short phases, lasting a few months to 2–3 years, of very rapid ice motion and advancement

95

(Table 4.3). During the surge phase velocities may occur that are >100 times in magnitude greater than the quiescent phase. This periodicity is what is characteristic of surging ice masses but rarely can a temporal pattern be established.

TABLE 4.3. Duration of the active phase of the surge cycle for glaciers in polar and mountain areas of the world (Dowdeswell et al., 1991) Glacier

Length (km)

Alaska, Yukon Territories, British Columbia Variegated, AK 20 Walsh, AK 89 Muldrow, AK 46 Peters, AK 27 Tikke, BC 19 Tyeen, AK 7 West Fork, AK 41 Rendu, AK 17 Hazard, Yuk. 8 Childs, AK 19 Unnamed, AK 6 Black Rapids, AK 45 Carroll, AK 42

Area (km2)

Duration (years)

49 830 393 120 75 11 311 50 – 20 – 341 200

2 4 2 2 3 2 1 2 2 1 2 1 1

Kamb et al., 1985 Post, 1960 Post, 1960 Echelmeyer et al., 1987 Meier and Post, 1969 Field, 1969 Echelmeyer and Harrison, 1989 Field, 1969 Clarke and Collins, 1984 Tarr and Martin, 1914 Krimmel, 1988 Moffit, 1942 Smith, 1990

Reference

Iceland Sidujokull ¨ Dyngiujokull ¨ Tungnarj ´ okull ¨ Bruarj ´ okull ¨ Eyjabakajokull ¨ Teigadaljokull ¨ Hagafellsjokull ¨ eystri Hagafellsjokull ¨ eystri

40 – 25 50 10 1 17 17

350 – 120 1500 50 1 110 100

1 2 1 2 1 1 1 1

Thorarinsson, 1964 Thorarinsson, 1964 Thorarinsson, 1964 Thorarinsson, 1969 Williams, 1976 Halgrimsson, 1972 Sigbjarnarsson, 1976 Theodorsson, ´ 1980

Pamirs and Caucasus Medvezhiy Byrs Shini-bini Sugran Muzgazy Abramova Ravak Tanyrnas Burokurmas Kolka

13 2 10 22 11 – 3 10 7 6

25 10 16 47 16 – 2 61 8 3

1 2 2 3 2 3 1 2 2 1

Dolgushin and Osipova, 1975 Rototayev, 1983 Uskov and Dil’muradov, 1983 Uskov and Dil’muradov, 1983 Desinov, 1984 Glazyrin et al., 1987 Dolgushin and Osipova, 1975 Dolgushin and Osipova, 1975 Dolgushin and Osipova, 1975 Krenke and Rotoatyev, 1973

Tien-shan Mushketov Bezymyanny

21 6

70 11

2 2

Dolgushin and Osipova, 1975 Dolgushin and Osipova, 1975

Karakoram Kutiah Baltbare

12 8

– –

1 2

Desio, 1954 Wang et al., 1984

Andes Grande del Junca Grande del Nevado

10 6

9 5

1 1

Espizua, 1986 Bruce et al., 1987

96

ICE FLOW AND HYDROLOGY

Temperate and polar ice masses of all sizes, in a wide range of topographic and tectonic settings, on a variety of geological structures and lithologies, exhibit surging behaviour. In a survey of Yukon glaciers (2356) it was found that there appeared to be a statistically higher probability of surging conditions occurring in long and wider glaciers with low surface slopes (Fig. 4.12); while the probability of surging increased from 0.61 per cent for short glaciers (0–1 km) to 65.1 per cent for long glaciers (10–75 km). These results might suggest that large ice

(a)

(b)

masses such as outlet lobes and ice streams of ice sheets and ice caps could have a greater tendency to surge than smaller valley glaciers. For example, it has been hypothesized that megasurges of Antarctica may help trigger Global glaciation (Wilson, 1969) but the lack of sedimentological evidence in the Southern Ocean appears to negate the idea. Two fundamental questions regarding surging behaviour must be understood: (1) what is the mechanism that enables fast ice motion to occur during a surge, and (2) what are the mechanisms whether external or internal that initiate and terminate surging? Related to these questions is the need to understand the pattern, if any, of the geographical distribution and topographic setting of surging ice masses, the duration and persistence of surge periodicity between normal and surging flow states, the continuum of glaciological states between surging, normal and continuous fast ice flow conditions, the surging potential of present day ice sheets, and whether past ice sheets exhibited surging conditions. A major problem that may partly account for the proliferation of hypotheses of surging mechanisms is the lack of sound data acquired from direct observations of the surging process in progress. Until relatively recently few observations had been accurately made immediately before, during and subsequent to a surge event. Data on the 1982–1983 surge of the Variegated Glacier, Alaska, and work from the Trapridge Glacier, Yukon Territory, Canada, and the Medvehiy Glacier, Pamirs, Tadzhikistan, have provided fresh insights into surge behaviour. However, it is possible that surging behaviour under different basal thermal regimens may be initiated by dissimilar mechanisms. 4.8.1. Surface Appearance and Morphology

FIG. 4.12. (a) The influence of glacier length of surging. The variation of glacier length is shown for three different but related data sets: primary data set (solid line and subsets T (dot-dash line) and NT (dashed line)) of the primary data set. Subset T is derived from the primary data set by selecting only glaciers that are tributaries of some larger glacier system; subset NT contains the remaining glaciers. The data set based upon Yukon glaciers has been divided into ten BIN lengths denoted by index 1. (b) The influence of slope on glacier surging. A subset of the Yukon glacier primary data was used for this analysis (after Clarke et al., 1986; Journal of Geophysical Research, 91B, p. 7165, copyright by the American Geophysical Union).

Glaciers that exhibit surging behaviour have a characteristic surface appearance and distinctive morphology. However, not all the characteristics are peculiar to surging ice masses nor do all such glaciers display all of these features. Surging glaciers typically display: contorted (loop-like) surface medial moraines; sheared and heavily crevassed edges; severely crevassed surfaces; a large kinematic bulge in the ‘reservoir area’ high in the accumulation zone,

ICE FLOW AND HYDROLOGY

associated with a depleted ice ‘receiving area’ usually in the ablation zone, that is manifested as surface ice gradients higher and lower than ‘normal’, respectively; the presence of tributary glaciers with sheared snouts and/or ice-dammed lakes and the presence of large ‘dead-ice’ areas immediately in front of the glacier snout. During the surge period other features indicative of the state are: high meltwater discharges into the proglacial area; crevasses opening and closing at much higher rates than normal; the rapid appearance of up-sheared debris and clasts in the ablation area; ice gradient changes consequent on the transference of vast volumes of ice from the reservoir to the receiving areas; and the sudden draining of icedammed lakes near the end of a surging phase. Based upon observations at the Variegated Glacier and other Northwestern North American glaciers, several other characteristics would appear typical of surging behaviour (e.g., Clarke and Blake, 1991): 1 it seems well established that fast ice flow during a surge is a function of rapid basal sliding; 2 subglacial meltwater transmission across the total extent of a surging ice mass appears to take place during the surge phase but becomes restricted and/ or constricted once normal flow conditions resume; 3 it has been noted that meltwater discharge through the subglacial conduit system is much slower during the surging phase than under normal flow conditions; 4 close to surge cessation or when major surge ‘slowdowns’ occur, vast discharges of meltwater take place; 5 it is apparent that high sliding velocities are associated with high basal meltwater pressures possibly indicative of the existence of extensive basal cavitation during surging, such that on cessation of surging glacier surfaces are lowered; 6 observation shows that immediately prior to surging an episode of longitudinal shortening followed by one of elongation precedes the peak velocity passage through the glacier. This transient fluctuation in longitudinal strain rate causes differential zonation to develop within the surging ice mass such that ice above the point of velocity peak initiation undergoes continuous and cumulative

97

elongation, while ice below the maximum velocity peak experiences continuous and cumulative shortening. Ice between these two points experiences shortening followed by elongation and thus relatively low cumulative longitudinal strain; and 7 at least beneath the Variegated Glacier, it has been observed that during surging basal meltwater pressures rise to within 20–50 kPa of the normal ice overburden pressure but fall, generally, to ⭐15 kPa under normal flow. This correlation between high meltwater pressures and surging would appear to indicate that both aspects are closely related to each other. In any formulation of a hypothesis of surging, these above features and attributes must be satisfied. 4.8.2. The Surge/Non-surge Cycle As noted above, any surging glacier or ice mass manifests a two-stage flow state in which a long quiescent phase of ice flow is periodically punctuated by a short but very rapid period of ice flow. Table 4.4 makes comparisons between the Variegated and the Medvezhiy Glaciers (Raymond, 1987) The developmental stages observed on these glaciers are manifest in changes in ice surface geometry and velocity (Fig. 4.13). As a potential surging glacier approaches the next surge, there is a build-up of ice within the reservoir area and depletion in the receiving area, resulting in increasing ice velocity within the reservoir. At the pre-surge phase, there develops a state in which ice within the reservoir area is increasingly active while the ice below in the receiving area may be virtually stagnant. The boundary between these two areas has been termed the dynamic balance line (DBL). On the Medvezhiy Glacier, in the pre-surge stage, the DBL moved progressively down the flowline prior to the surge event introducing, perhaps, a critical ice mass that may act as a ‘trigger’ to surge initiation (Dyurgerov, 1986). In contrast, motion of the DBL on the Variegated Glacier was not readily observable since, unlike the Medvezhiy where crevasse systems became active and opened up in the reservoir area, little or no difference was noted between this glacier and other mountain glaciers in the region. However,

98

ICE FLOW AND HYDROLOGY

TABLE 4.4. Comparison of surging glaciers – the variegated and Medvezhiy glaciers (after Raymond, 1987) Variegated

Medvezhiy

General Total length (km) Surging length (km) Surging period (years) Mean slope of surging length 10-m temperature (°C)

20 18 ~18 0.094 Temperate

16 7–8 9–14 0.11 –0.9

Quiescent phase Terminus retreat (km) Advance of dynamic balance line Elevation change (m) Maximum basal shear stress (105 Pa) Maximum annual velocity (m day–1) Summer velocity increase (%)

zero 1 +60(–40) 1.8 0.6 80

<1.5 >4 +120(–100) ? 1.5 100

Surge phase Number of events/duration Timing of initiation Timing of termination Advance of surge front (km) Advance of topographic peak (km) Maximum thickness change (m) Advance of velocity peak (km) Maximum speed (m day–1) Maximum ice displacement (km)

1982–1983 surge 2/6–8 months Early winter, late autumn Early summer 12 11 110 11 65 ~2

1963 and 1973 surges 1 Late winter ? Early summer ? ? 120 ? 100 >1.6

basal sliding velocities of the Variegated Glacier in the pre-surge stage were unaccountably higher than normal during the winter period (Bindschadler et al., 1977). As a glacier approaches the succeeding surge event, occasionally short pulses of rapid, surge-like events, have been observed. These phases occurred on the Medvezhiy, termed ‘wavy surges’, and on the Variegated, where the term ‘mini-surges’ was employed (Dyurgerov, 1986; Raymond and Harrison, 1988). In general, these mini-surges last for only a few hours. However, mini-surges are not necessarily related to the major surge event, nor are they indicative of surging behaviour since such events do occur on normal glaciers. On the Trapridge Glacier during the non-surge part of the cycle a sharp spatial differentiation was noted in basal thermal regimen with thick ice in the reservoir area underlain by a temperate bed separated by the DBL from thin ice in the receiving area underlain by a frozen bed (Clarke et al., 1984). Whether this form of basal thermal

development is characteristic of sub-polar surging glaciers remains unknown but elsewhere such glaciers have been observed to have warm temperate beds. As the ice mass enters the surge phase of the cycle there is a reversal of the geometric evolution of the quiescent phase. The cycle of the surge period is rarely continuous but often exhibits irregular patterns sometimes of a single major surge event separated by differing lengths of non-surge periods or, as in the case of the Bering Glacier, Alaska, two short surge events close together followed by a longer quiet interval. Surge periodicity (Table 4.3) would appear to be connected with periodicity in longitudinal strain rate change (Shoji and Langway, 1985). There does appear to be some evidence that suggests surges often begin in the winter and terminate in the early spring/ summer. However, some glaciers such as the Muldrow in 1956 (Harrison, 1964), and Bruarj¨okull in 1963–1964 (Thorarinsson, 1969) continued through a summer period, while the Black Rapids Glacier in 1936–1937 stopped in winter (Hance, 1937).

ICE FLOW AND HYDROLOGY

Medvezhiy Glacier

Elevation (m)

Elevation (m)

longitudinal profile 72

64

Variegated Glacier

2000

4000

3500

99

73

1500

longitudinal profile

500

9/81

9/73

9/84

bed

3000

100

Elevation change (m)

Elevation change (m)

3000

reference 1964

70

50

72

68

0 64

-50 -100

50

9/81

0

-50

71-72

0.5

69-70

70-71

67-68 66-67 64-65 68-69 65-66

0

2

78 79 78 77 76

9/73 74 75 76 77 78 79 80

75 74 9/73

9/81

winter speed Speed (m d-1)

Speed (m d-1)

annual speed 1.0

reference 6/1973

4

6

79-80

0.4

80-81

78-79 77-78

0.2

76-77 73-74

8

Distance (km)

20

15

10

5

76-77

0

Distance (km)

FIG. 4.13. Variation and development of longitudinal surface profiles, change in surface elevation and annual speed during quiescent periods of two surging glaciers, the Medvezhiy and Variegated Glaciers (after Raymond, 1987, Journal of Geophysical Research, 92B, p. 9123, copyright by the American Geophysical Union).

4.8.3. Hypotheses of Surge Behaviour Debate as to the specific mechanism(s) to explain surging behaviour has continued unabated for the past 30 years. In the process, a profusion of hypotheses have surfaced (Table 4.5). Whether one or several mechanisms can explain surging or whether surging is externally or internally triggered, remain sources of ongoing discussion. There is a general consensus that the surging mechanism is associated in some manner with changes in the subglacial hydraulic system. In summary, two dominant ideas both connected with subglacial hydraulic system disruption have

emerged: (1) Kamb’s (1987) suggestion that a linkedcavity subglacial meltwater system develops to replace a hydraulic system composed of a few large drainage channels and acts to trigger the surge and, in disruption (short-circuiting), terminates the surge; and (2) Clarke et al.’s (1984) concept of subglacial permeable bed conditions ceasing owing to massive basal shearing resulting in impermeable subglacial bed conditions and H-bed conditions prevailing, leading to high basal meltwater pressures triggering the surge that sequentially leads to basal friction reduction, drainage system deterioration and subsequent surge termination (Menzies, 1995, chapter 5, sections 5.10.3.1 and 5.10.3.2).

100

ICE FLOW AND HYDROLOGY

TABLE 4.5. Suggested surge release mechanisms Release mechanism

Instability manifestation

References

Tectonic activity

Causing increased avalanching and perhaps increased basal geothermal heating

Tarr and Martin, 1914; Post, 1965

Landsliding

Increased load added to reservoir area

Gardner and Hewitt, 1990

Creep instability

(a) increased ‘softening’ of basal ice thus increased basal ice deformation (b) change of basal thermal regime from frozen to warm bed (c) change in mass balance thus increased mass balance gradient

Robin, 1955; Clarke et al., 1977; Paterson et al., 1978 Yuen and Schubert, 1977 Clarke et al., 1977; Schubert and Yuen, 1982

Melted core

Instability beneath reservoir area

Schytt, 1969

Lubrication

Increased basal meltwater – bed decoupling

Budd, 1975; Budd and Jenssen, 1975; Weertman, 1962; Robin and Weertman, 1973

Friction melting

Increased basal velocity – increased meltwater

Clarke, 1976; Budd and McInnes, 1974

Basal shear stress fluctuation

Increased basal slip – increased meltwater

Meir and Post, 1969; Robin, 1969; Robin and Weertman, 1973

Kinematic wave transmission

Increased basal slip – increased meltwater

Weertman, 1969; Palmer, 1972

Super-cavitation

Increased area of bed with cavity development-decoupling

Lliboutry, 1978

Frontal instability, calving

Removal of frontal ice mass support

Hughes, 1981

Linked-cavity development

Transition between localized conduit expansion and short-circuit to linked cavities-decoupling

Kamb, 1987

Deformable bed/hydraulic instability

Decoupling of bed owing to decreased viscosity within deforming subglacial sediment

Clarke et al., 1984

4.8.4. Surging and Glacial Sedimentology The influence of surging ice masses upon processes in proglacial, supraglacial, englacial and subglacial environments remains obscure. However, chronostratigraphic evidence seems to indicate that several lobes of Quaternary Ice Sheets may have surged. For example, outlet lobes along the southern margin of the Laurentide Ice Sheet (e.g., Clayton et al., 1985) and the eastern lobe of the British Devensian Ice Sheet may have surged. The development of a model that would allow the recognition within the stratigraphic record of paleosurges, however, remains elusive. Work in Wisconsin, Michigan and Montana indicate lobes having low surface profiles, perhaps indicative

of low basal stresses, basal water lubrication, deformable beds and a surge-phase (e.g., MacAyeal et al., 1995; Jenson et al., 1996; Maher and Mickelson, 1997; Marshall and Clarke, 1997a,b). Similarly, a relationship has been advocated between surging and the formation of flutings owing to inherent flow instabilities of secondary flow patterns that may develop within the ice (Kr¨uger, 1979; Sharp, 1985). The relationship of composite moraine ridges, common along the margins of several surging glaciers in Iceland and Svalbard, has also been investigated (Croot, 1988b). The potential impact of surging involves (A) the effect of increased subglacial and proglacial meltwater activity allied with the disruption and possible destruction of all

ICE FLOW AND HYDROLOGY

supraglacial and englacial hydraulic systems during and for some time after the surge event; and (B) the impact of stress/strain fluctuations upon subglacial and proximal proglacial sediments and landforms, and the erosive effect of high basal sliding velocities upon the subglacial bed. (A) The consequences of meltwater activity can be summarized as follows: (i) prior to surging most ice masses have well developed hydraulic systems. Depending upon climatic and glacial thermal regimens an englacial hydraulic system may be in place. It is likely that a tunnel system of a few major conduits with smaller tributary channels will exist in equilibrium with subglacial stress conditions and bed topography. Meltwater evacuation would, in the temperate areas, possibly drain via N-channel systems and by pipe flow within subjacent unfrozen sediment within the frozen areas beneath the glacier. (ii) On surging, the supra- and englacial hydraulic systems are probably destroyed. In the former case if the surge begins in winter, as is typical, little or no supraglacial system will evolve and, in the latter case the englacial system may be shut down or operate at very low discharges. (iii) With surge termination, there is a time-lag before the pre-surge hydraulic system can re-establish. A drop in meltwater pressure will probably occur as the surge hydraulic system collapses and a j¨okulhlaup-style flush of turbulent and high-discharge meltwater occurs just prior to the cessation of the surge or immediately afterward. (iv) During and immediately following a surge, vast quantities of silts and clays are transported by meltwater. The turbulence and turbidity of surge meltwater systems appear to be related to the development of a hydraulic system spreading across the glacier bed transporting sediment from areas not previously flushed since the last surge event. (v) Sediment transport within supra- and englacial parts of a surging glacier system is two-

101

fold; first, sediment is transported into differing environments. In particular, englacial debris will be transported into the subglacial system a result of massive undermelting. Secondly, large quantities of sediment will be enclosed within sealed crevasse systems disrupting the ‘normal’ sediment transport flux. (vi) A final effect of surging upon meltwater activity is to introduce a vast quantity of meltwater transported sediment into proximal proglacial areas. (B) A second less precise effect of surging is the influence of transient but widely fluctuating stresses and strains that may impact upon subglacial and proglacial sediments and bedforms. Where sediments underlying surging ice masses are frozen or heavily consolidated some form of deformation may occur. If the surge mobilizes a sediment bed, sediments may suffer massive deformation with clast ploughing and intraclastic rafting occurring (Menzies et al., 1997). Deformation may be pervasive or nonpervasive and may be areally widespread or restricted to patches of the bed where localized sediment tectonization occurs (Chapter 14). Sharp (1985) and Solheim and Pfirman (1985) report dyke-like structures that are possibly casts of basal crevasse intrusions caused by differential loading and/or deformation. Fluted diamictons beneath surging glaciers may be a consequence of rapid deformation of a mobilized slurried subglacial sediment. The influence upon deposited sediment of overriding ice during a surge may create localized stress resulting in neofracturing and the re-opening of depositional or structural discontinuities. The effect of a surging glacier advancing rapidly upon either a stagnant snout region or a proglacial zone leads to considerable lateral stress, overriding and differential loading (e.g., van der Meer, 1987; Aber et al., 1989). Many major push moraines and nearhorizontal nappes, the latter exhibiting long-distance translation of stacked sediment sequences, may be tentatively associated with surge events (Fig. 4.14) (Chapter 12).

102

ICE FLOW AND HYDROLOGY

6

(a)

9

8 11

5 1

3

2

7

10 4

12

(b)

(c)

13 14

15

FIG. 4.14. A general model of sedimentation by surging glaciers. (a) Marginal environment immediately after a surge: 1, composite lobate gravitational slump of proglacial sediments; 2, glaciotectonically thrust ridges; 3, normal faulted sediments in ridge core; 4, water escape structures in ice-overridden sediment; 5, outwash sediments; 6, ridge composed of resedimented diamictons; 7, englacial debris bands elevated along thrust zone; 8, supraglacial debris dikes; 9, supraglacial sediment flows; 10, subglacial diamictons; 11, basal crevasse-filled dikes. (b) Marginal environment early in quiescent phase of surge cycle; 12, impact of ablation leading to sediment drapes, ridges and thermokarst relief features. (c) Marginal environment late in quiescent phase of surge cycle; 13, marginal rim of sediment isolated from the rest of the glacier characterized by ice-cored ridges and hummocky relief; 14, crevasse-fill ridges on fluted subglacial diamicton (see 11); 15, exposed subaerial surface resulting from glacier retreat (after Sharp, 1985; reproduced by permission of Academic Press).

4.9. HYDROLOGY OF GLACIERS Glacial hydrology is an integral element in understanding all glacial environments and processes. Water occurs as a fluid, as an interstitial liquid, as a solid and as interstitial ice within all glacial environments. As meltwater, it is active in debris erosion, entrainment and deposition. Meltwater causes channel bed erosion and scour, and the development of complex drainage networks within all glacial subenvironments. As interstitial water within debris or fractured bedrock, deformation and fracture may be facilitated and secondary sedimentological structures formed. Porewater migration and/or porewater saturation levels may lead to mineral and particle redistribution. Finally the influence of frost action and the

development of cryostatic stress within most glacial subenvironments is an area of new research. Many glaciodynamic processes are linked to various hydraulic effects whether it is the influence of porewater on the viscosity of subglacial debris or the subglacial hydraulic system fluctuations that appear to characterize the triggering of surges. Meltwater, from both subglacial, supraglacial and englacial areas of an ice mass, carries clues to other critical but unseen processes taking place within these oftenobscured environments. 4.9.1. Meltwater Water associated with ice masses is derived from two main sources: (1) atmospheric precipitation and

ICE FLOW AND HYDROLOGY

(2) melted ice and snow. Other minor sources within the subglacial environment include water of juvenile derivation, from underlying aquifers, and from the melting of overridden permafrost sediments. Glacial water originates from melting because of surface ablation, basal ice melt resulting from friction, melt from geothermal heat, basal melt owing to air temperature melting in cavities and near marginal locations, melting owing to the passage of meltwater over the surface of ice masses, and internal ice deformation and consequent energy dissipation (R¨othlisberger and Lang, 1987; Sharp et al., 1998) (Table 4.6).

103

4.9.2. Meltwater Systems Meltwater exists either within a series of discrete, segregated hydraulic systems or as a few large systems that interconnect with all or most subenvironments. The development of glacial hydraulic systems within any ice mass is dependent upon: the type, morphology and topographic setting of the ice mass; thermal conditions at the surface, within and beneath each ice mass; mass balance; velocity profile (surging/non-surging); basal conditions; seasonal climatic conditions present and past; and the amount and type of debris cover (Fig. 4.15). Meltwater systems may be

TABLE 4.6. Physical constants of ice and water at 0°C and related properties (after Drewry, 1986 and R¨othlisberger and Lang, 1987) Property

Symbol

Quantity

Units

Mechanical Density of water Density of ice Young’s modulus of ice Yield strength of ice Fracture toughness of ice Creep activation energy of ice Flow law constant of ice

␳wt ␳i E1 ␶1 K1 E* A

kg m–3 kg m–3 GN m–2 MN m–2 MN m–3/2 J mol–1 Pa–3s–1

Exponent in power of ice Vicosity of water Hydraulic roughness parameter for conduits

n ␩w kr

999.8 920 9.10 85.0 0.2 6.07 × 104 8.75 × 10–13a 5.20 × 10–13b 7.93 × 10–13c 7.54 × 10–13d 32.5 × 10–13e 3 1.787 × 10–3 Smooth 100 Medium 50 Very rough 10

°K kte ␬ – csw ct,o

273.1 2.51 1.33 × 10–6 37.7 4.217 × 10–3 7.4 × 10–8

K W m–1 deg–1 m2 s–1 J mol–1 deg–1 J kg–1deg–1 K Pa–1

ct,a

–9.8 × 10–8

K Pa–1

Lh

334 @ 0°C 285 @ –10°C 241 @ –20°C

kJ kg–1

Thermal Melting temperature of ice Thermal conductivity of ice Thermal diffusivity of ice Heat capacity of ice Specific heat of water Change in pressure melting point of ice with pressure (pure water) Change in pressure melting point of ice with pressure (air-saturated water) Latent heat (fusion) of ice

a

Drewry, 1986. Paterson, 1981. Hooke, 1984. d Lliboutry, 1983. e R¨othisberger, 1972. b c

– kg s–1 m–1 m–36 s–1

104

ICE FLOW AND HYDROLOGY

snow/firn: fine grained coarse grained water saturated

subsurface/surface channel transverse seepage glacier bed

vertical

cold

seepage zone

crevasse drainage

moulin

englacial conduit

subglacial conduit FIG. 4.15. Model of supraglacial, englacial and subglacial drainage routes (afer R¨othlisberger and Lang, 1987; reprinted from Gurnell, A.M. and Clark, M.J. (eds), Glacio-Fluvial Sediment Transfer, by permission of John Wiley and Sons).

in the form of (i) thin sheet flow; (ii) channel flow; (iii) linked cavity flow; or (iv) an integration of the above forms. A limited amount of meltwater may also seep into the ice from surface melt and runoff and through the walls of sealed or clogged crevasses (Paterson, 1994; Sharp et al., 1998).

Meltwater systems may be seasonal or of shorter duration, and may exhibit normal or catastrophic fluctuations. The frequency and magnitude of system discharge fluctuations may vary widely from season to season, on different parts of the same ice mass and between neighbouring ice masses. The evolution

(a) (b) PLATE 4.1. Supraglacial lakes. (A) On the surface of an unnamed glacier, Spitsbergen. (B) On the surface of the Matanuska Glacier, Alaska (photographs courtesy of Ed Evenson).

ICE FLOW AND HYDROLOGY

of different systems may also vary enormously in time and space. Under steady-state or quasi-steadystate conditions, a hydraulic system may be distinctive to a specific ice mass that, when interrupted, will return to that same type of system. For example, the hydraulic systems beneath surging glaciers mature during the non-surging quiescent phase but may be destroyed or altered during the surge, and

105

then returned to the former system after a few months or years of non-surging ice flow conditions. Supraglacial and englacial hydraulic systems exhibit considerable fragility, being easily susceptible to rapid change, whereas established subglacial systems are more resistant to alteration unless ice conditions are dramatically transformed as in a surge or rapid ice retreat.

N

1 2 3 4 5 6 W M

P E

0

200

400 metres

FIG. 4.16. The lower section of Mikkaglaci¨aren, Sweden, showing a generalized map of drainage collection crevasses and moulins (ca. 1958). 1, moraine covered ice; 2, crevasses; 3, direction of surface drainage; 4, moulins; 5, boundary line between crevasses striking east and west; 6, area of surface ice where drainage is characterized by predominant structure and slope-factors (after Stenborg, 1969; reprinted from ‘Studies on the internal drainage of glaciers’, by T. Stenborg, Geografiska Annaler, 1970, 52A, 1–30, by permission of the Scandinavian University Press).

106

ICE FLOW AND HYDROLOGY

4.9.2.1 Supraglacial hydraulic systems In the supraglacial environment, hydraulic systems develop an arborescent pattern largely in accordance with subaerial fluid dynamics. As overall discharge and/or gradient increases, the hydraulic system tends to evolve from disconnected flow to an arborescent drainage pattern of connected channels. Supraglacial lakes may develop, especially in summer months, and form part of the drainage system over a large area of the ice surface. These lakes may create localized radial but isolated drainage systems (Plate 4.1). Drainage on the surface may also be interrupted by meltwater disappearing into englacial and subglacial systems via moulins. Depending upon the ice mass

activity, vertical ice thickness, state of internal stress and other local factors, connections may or may not be made between the supraglacial and englacial and/ or subglacial hydraulic systems (Fig. 4.16). Drainage patterns on glacier surfaces are typically composed of several basins that may drain off laterally/frontally or into large surface lakes or major moulins (Fig. 4.16). The effect of the moulins is to frequently create, in the summer when they are best developed, a ‘karstic’-style surface drainage system. During massive ice downwasting and deglaciation, thermokarst drainage systems should develop to their greatest extent. Surface lakes may drain episodically during a summer season causing englacial and subglacial drainage systems to be flooded and altered

0.7 August 0.6 July

[ mm h-1 ]

0.5

0.4

September

0.3 June 0.2

0.1 May 0

6

12

18

24

6

Hours FIG. 4.17. Mean diurnal variation in runoff, 1974–1980 for Verngtbach (Austrian Alps), drainage area 11.4 km2, 84 per cent glacier covered, over the period May–September (after R¨othlisberger and Lang, 1987; reprinted from Gurnell, A.M. and Clark, M.J. (eds), Glacio-Fluvial Sediment Transfer, by permission of John Wiley and Sons).

PLATE 4.2. Moulin on the Athabasca Glacier, Alberta (photograph courtesy of Gerry Holdsworth).

ICE FLOW AND HYDROLOGY

accordingly (Clarke, 1982). As winter approaches the amount of surface meltwater decreases and concomitantly moulins and small crevasses ‘seal up’ with snow and ice; however, by the following summer many of these moulins re-open. Supraglacial meltwater discharge exhibits large seasonal variations that increase from early spring to a mid-summer maximum followed by a decrease until freeze-up in the autumn (Sharp et al., 1998). Considerable surface meltwater may be stored in snow and firn during the early spring. This water is later released into the glacier hydraulic systems. Superimposed upon these long-term fluctuations are diurnal fluctuations related to local weather conditions (Fig. 4.17). The amount and form of supraglacial meltwater activity would seem to be little influenced by the thermal state of the underlying ice mass. It might be expected that the volume of meltwater on a cold (polar) glacier might be less. However, little evidence exists to

107

support this impression. On ice masses where the surface is heavily crevassed, the surface drainage system is adversely affected and consequently poorly developed. In areas of ice falls (seracs), surging glaciers or along the edge of fast ice streams, crevasses intercept the natural drainage system (Plate 4.2). In valley glaciers confined by high surrounding valley sides where a distinct convex, transverse profile develops, the profile leads surface meltwater to flow towards valley sides. Lateral streams may then flow either along the edge, often disappearing within lateral moraines or submarginally beneath the ice (Plate 4.3). Where glaciers are extensively covered by morainal debris over large areas of the surface, supraglacial stream flow is interrupted. Moraines disrupt the supraglacial drainage system causing distinct separate drainage system patterns to develop (Plate 4.4). On debris-covered glacier surfaces much of the surface meltwater disappears into the debris (Plate 4.5).

PLATE 4.3. Lateral meltwater systems along the edges of Uversbreen, Spitzbergen (photograph courtesy of Ed Evenson).

108

ICE FLOW AND HYDROLOGY

PLATE 4.4. Separate supraglacial meltwater drainage systems on the surface of the G¨ornergletscher, Switzerland, August 1989. Note major drainage system on larger glaciers with smaller system developed on the tributary glacier at bottom left of photograph.

Supraglacial meltwater influences several significant sedimentological processes. 1 In the debris-covered terminal zones of many glaciers where debris acts to ‘soak-up’ surface meltwater, debris flows, slides and subsequent flow tills may occur. Such mass transport of debris is largely a function of angle of slope, gravity and the instability caused by debris becoming saturated, reaching high internal porewater pressures, low effective stress levels and low internal shear strength. 2 As a major source of meltwater to proglacial outwash areas, supraglacial meltwater acts to entrain large amounts of debris from the ice surface into this zone.

3 The total hydraulic system of a glacier has been likened to an enormous ‘plumbing system’ in which the supraglacial system acts as one of the sources whereby debris is flushed out or through the system. Evidence of this ‘flushing process’ has been observed in many glaciers, especially in the early spring when melting begins or following major thunderstorms (Collins, 1982). 4.9.2.2. Englacial hydraulic systems The type of hydraulic system that evolves within englacial and subglacial environments is much less understood. In both cases, meltwater may be under pressure because of the confining effect of the ice and

ICE FLOW AND HYDROLOGY

109

PLATE 4.5. Meltwater penetrating morainal debris in front of an unnamed glacier, Spitsbergen (photograph courtesy of Ed Evenson).

the closure forces of the internally deforming ice. The nature and style of development of englacial and subglacial hydraulic systems are controlled in part by the meltwater discharge, and the source(s) of that discharge. Both systems may be linked and may also be coupled periodically to the surface supraglacial system, thus deriving an extrinsic (exotic) supplementary source that is unaffected by internal processes of either system. Neither system can be easily investigated. The englacial drainage system can be viewed as a complex architecture of interconnected cavities and conduits within a three-dimensional gallery system. Englacial drainage systems are often compared to those found in karst terrains but, unlike such systems, the englacial system is highly dynamic (Smart, 1990).

Where ice is temperate and fast moving, englacial systems are likely to be temporary and largely autonomous. Under slow moving and temperate conditions, a well-developed system that may interconnect to the subglacial system may develop. Within polar glaciers, a limited summer near-surface (<2 m) englacial system can develop if connected with those supraglacial locations where sufficient summer ablation occurs. Englacial systems within surging glaciers probably suffer total destruction during the surge phase but may resume a well-evolved system during quiescent periods if those periods are sufficiently long (Sturm and Cosgrove, 1990). However, it is difficult to estimate the time necessary for an englacial system to become fully activated after a surge event. In

110

ICE FLOW AND HYDROLOGY

stagnant ice, englacial drainage systems are very well developed and are normally fully connected to both supra- and subglacial drainage paths (NB englacial watertables). From evidence gathered from descent into moulins and by entering frontal passageways, and the use of dyes and other soluble tracers, it has been possible to establish something of the nature of englacial hydraulic systems (Holmlund,1986, 1988). In only a very few cases have moulins been explored and some clue as to the englacial conduit system examined first hand (Reynaud, 1987; Holmlund, 1988) (Fig. 4.18). The degree of evolution of any englacial system would

appear to be a function of glaciodynamics and stress history; thus systems may alter or degenerate every few years. The internal geometry and evolution of englacial conduit systems is complex and still remains, somewhat, enigmatic. Englacial conduit development may begin as a result of crevasses being exploited by either direct supraglacial streams entering them or by up-ice englacial conduits intersecting crevasses further down the flowline, or by the evolution of seepage passages into small tube-like englacial tributary systems. In general, englacial hydraulic systems are best developed in ablation areas and/or areas of extending

68:5

10 m 0

68:6

-20 m

68:7

-25

m

Photo -30

m

-35

m

Ice flow

-40

m

FIG. 4.18. Perspective illustration of moulins on 2 August 1982 on Storglaci¨aren, Sweden (after Holmlund, 1988; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 34(117), 1988, p. 245, fig. 5b).

ICE FLOW AND HYDROLOGY

111

(a)

(b) PLATE 4.6. Submarginal englacial tunnels. (A) In the Matanuska Glacier, Alaska. (B) In the MacLaren Glacier, Alaska (photographs courtesy of Ed Evenson).

112

ICE FLOW AND HYDROLOGY

longitudinal strain rates. The style, geometry and degree of development of any englacial hydraulic system is a complex function of: the meltwater discharge entering the system; the depth, width and number of crevasses; the thermal conditions of the ice mass; the basal sliding velocity; the nature of the subglacial hydraulic system; and the overall stability of the ice mass in the long term. In order for an englacial conduit to remain open and stable under active ice conditions after the crevasse seals, it is necessary for those forces maintaining the opening of the conduit to equal those attempting to close it. Englacial conduits remain open as a result of thermal energy generated by: (a) meltwater turbulence; (b) frictional energy resulting from transported debris within meltwater causing sidewall abrasion; and (c) the diffusion of direct thermal energy from the passage of warm meltwater (viscous heat dissipation). The factors leading to conduit closing are those associated with the intrinsic internal deformation of the confining ice mass attempting to seal the conduit space and the rate of freezing-on of water to conduit walls. Also, when englacial meltwater discharge wanes, for whatever reason, debris transported in these channels will settle and possibly clog the system (Saunderson and Jopling, 1980). In general, englacial conduits usually approximate a circular cross-section (Plate 4.6), but variations do occur. Englacial and subglacial channels are often not fully filled by water and channel shape will thus deviate from the circular to lesser or greater degrees. In channels totally water filled, the hydraulic pressure of the meltwater may or may not approach the confining ice overburden pressure, whereas in partially filled channels water pressure may be at or close to atmospheric pressure or at triple-point pressure if no open outlet exists. Steady-State Channels or Not? Several theoretical studies of englacial drainage have been made in which the existence of a steady-state channel has been assumed and in which the primary unknown has been the meltwater pressure (R¨othlisberger and Lang, 1987). Following the analysis by R¨othlisberger (1972), it can be shown, from geometric considerations, that large channels will tend to grow

at the expense of small channels by piracy and that meltwater will be predisposed to flow in channels rather than sheets (Weertman and Birchfield, 1983). In general, steady-state models of these hydraulic systems are at best ‘crude approximations’. Increasing field evidence indicates that most englacial and subglacial channels, under thin marginal ice thicknesses, are not filled for much of the time and that open-channel flow is common. In other locations within ice masses where ice pressures owing to greater ice thicknesses cause more rapid channel closure, channels can be expected to be filled for longer periods of time and thus approach steady-state. These deeper channels will also tend to be supplied from longer distances by seepage and small englacial complex hydraulic systems; therefore, discharge fluctuates to a lesser extent and consequently the system approaches steady-state. 4.9.2.3. Subglacial hydraulic systems The imprint of subglacial hydraulic systems is unquestionably one of the most important features of any glaciated landscape. Lying as it does at the interface between the glacier and its bed, the subglacial system is capable of moving vast quantities of sediment, of cutting down into bedrock and already existing deposited sediment, and thus imparting a critical influence upon glaciodynamics that ultimately dictate the geomorphic ‘effect’ of glaciation. Understanding of subglacial systems remains limited; at best, secondhand and derivative. However, unlike englacial systems that survive intact in a limited capacity as sedimentary expressions in the form of eskers and kames; subglacial meltwater systems, typical of late stages in glaciation, are often left as relict features in the form of channel systems and esker ridges (Walder and Hallet, 1979). The presence of subglacial meltwater beneath ice masses is well documented, not only from observations of large, apparently subglacial, meltwater portals within valley glaciers (Plate 4.7) but from borings and measurements completed some distance up-ice in the upper ablation and accumulation areas (Sharp et al., 1998). Until 1963 (Gow, 1963) little, even indirect, evidence existed to indicate the presence of subglacial meltwater beneath the Antarctic Ice Sheet. Since then

ICE FLOW AND HYDROLOGY

PLATE 4.7. Subglacial meltwater channel portal. Note high summer volume of glacial meltwater, Nigardsbreen, Norway.

the presence of meltwater, either free or as a component within a saturated debris layer, or as lakes beneath Antarctica, has been confirmed. Beneath the Greenland Ice Sheet, there is little evidence from central areas of the ice sheet to support an active meltwater presence but marginal areas appear to have subglacial water present. Where temperate conditions occur meltwater is likely to be present but under polar conditions no meltwater, unless supercooled or saltsaturated, should be found other than in an interstitial state. Therefore subglacial hydraulic systems are only likely to be developed to any degree beneath ice masses with areas of temperate bed conditions. In those ice masses that have a subpolar thermal regimen, limited areas of the bed may have a subglacial hydraulic system; also where thermal regimens have altered from warm to cold, relict

113

features both on the bed and within the basal ice may testify to the past existence of a subglacial hydraulic system. Where polythermal regimens occur, it is likely that interrupted, diverted and pinched-out subglacial hydraulic systems will develop, flourish and be spatially extinguished over time. This form of basal thermal condition may typify major ice sheets over the long-term (MacAyeal et al., 1995; Jenson et al., 1996; Marshall et al., 1996), whereas temperate conditions may be more characteristic of valley glacier systems as commonly observed in middle latitudes. It is apparent that where sudden or cyclical changes in ice velocity occur, as in a mini-surge or major surge event, a subglacial hydraulic system is likely to be drastically altered or eradicated, at least temporarily. Principal basal stress conditions vary according to position beneath the ice mass and distance from the ice front, or over significant topographic highs and lows on the bed. In addition, the influence of overburden pressures will, in part, dictate meltwater hydraulic pressures. Where no direct connection exists to the outside atmosphere, it can be expected that meltwater will be under hydrostatic pressure equal to the ice overburden pressure. Whether bed topography is very smooth or rough, specific styles of meltwater evacuation systems may or may not be capable of development; thus, for example, Kamb’s linked-cavity system is unlikely to evolve successfully across a smooth impermeable surface (Menzies, 1996, chapter 5). In those areas of the glacier bed where either highly permeable bedrock or sediment are found, it is possible that little or no meltwater will flow at the interface between the bed and the glacier sole. It is also possible that meltwater may infiltrate into an advecting deformable bed, reducing again the presence of free meltwater at the ice–bed interface. Any glacial hydraulic system will develop and/or persist to a greater or lesser degree if it can obtain exotic source(s) of water that, in the subglacial instance, is typically from englacial or supraglacial systems, both of which are unaffected by those basal ice pressure effects likely to curtail or remove subglacial water-source discharges. Finally, in trying to understand how any subglacial hydraulic system develops and attains a specific form over time, it is essential to bear in mind that such a system exists

114

ICE FLOW AND HYDROLOGY

within a dynamically changing environment. Thus the nature of subglacial hydraulic patterns and form will vary across the bed of a glacier, in all directions, and will alter in any one location over both the short and long term, depending upon the nature and style of glaciodynamic changes. Subglacial Meltwater Discharge Subglacial discharges exhibit considerable diurnal, seasonal and annual variations. These fluctuations decrease as the time period examined increases and/or the size of the ice mass decreases. At the annual scale, a pattern of fluctuations in relation to solar energy, net mass balance and hydraulic system development can be observed. In general, total run-off variations from glaciers tend to smooth out on a year-to-year basis. By mid-summer, it would appear that the subglacial hydraulic system(s) (usually in association with fully developed supra- and englacial hydraulic systems) within a temperate ice mass are at the highest degree of development with high average discharge values owing to high insolation and supraglacial melting. As winter approaches there is a reduction in surface melting, moulins may begin to close or seal up and, because the relaxation time of ice movement maintains high velocity values, subglacial cavities and channels may increase their closure rates. With further loss of surface meltwater, heat transference through the system via meltwater passage is reduced and a subglacial hydraulic system of much-reduced capacity requirements begins to develop (Sharp et al., 1998). At the mid-winter stage, depending upon climatic conditions and the severity of winter temperatures, in many cases the subglacial system will seal up and cease to function. In other locales, sufficient basal friction and geothermal heat energies allow some subglacial meltwater to be produced but in a drastically altered and degenerated system compared with the summer. With spring, there are two significant events with associated time lags: first, crevasses and moulins have to open or re-open and an englacial hydraulic system begin to operate while the increase in snow accumulation will cause an increase in basal ice stress and marginally increase basal ice velocities leading to increasing basal meltwater production. The start of the spring will therefore tend

to be linked with high basal meltwater pressures that in turn lead to a new hydraulic system opening up. Second, if any remnants of a subglacial hydraulic system exist it is at a much reduced and restricted level of development. This effect ‘throttles’ the subglacial system and obstructs meltwater flow until a new system of subglacial hydraulic passageways evolves. Once formed, this new subglacial system tends to exhibit a characteristic spring flood event or peak discharge as the dammed meltwater is released via the newly developed system. It is during the ‘spring event’ that large volumes of subglacial sediment are flushed through the system where they had accumulated in cavities and on channel floors since the start of the winter period. This sequence of subglacial hydraulic cyclical change, however, is ‘ideal’. In reality, different ice masses exhibit considerable variations in terms of the timing of conduit opening and closing, and interconnection development during the ablation season. 4.10. THE NATURE OF MELTWATER FLOW AND ROUTING AT THE ICE–BED INTERFACE When discussing meltwater at the base of an ice mass, two related but, for the purposes of discussion here, separate aspects must be addressed: (a) the form of meltwater flow (whether sheet, channel, linked cavity or some combination), and (b) the routing of meltwater flow whether across a bedrock or sediment surface, totally or partially within a shattered bedrock or other material or by some combination of pathways. 4.10.1. Meltwater Routes The development of any subglacial hydraulic system is contingent upon the nature and form of the ice–bed interface. There are four possible states in which meltwater may exist at the base of an ice mass, viz: (1) meltwater (free, unconfined) flowing in some form at the upper ice–bed interface (H-bed state); (2) meltwater (confined) flowing within rigid debris or bedrock (H-bed state); (3) meltwater (confined) flowing within a deforming sediment layer (M-bed state); and (4) meltwater (quasi-free) as interstitial films (H-bed state) (Menzies, 1996, chapter 2).

ICE FLOW AND HYDROLOGY

1 For a free subglacial hydraulic system to fully develop, it is necessary for most of the meltwater to flow at the upper interface between the ice and bed (Fig. 4.19). This is likely to occur under H-bed states where ice is flowing upon bedrock of low to zero permeability (Ha), or over frozen debris (Hb) or unfrozen debris (Hc) of low (<10–6 m s–1 ) to near zero permeability (Shoemaker, 1986c). 2 Where debris or shattered bedrock exists, the passage of meltwater in the form of pipe flow may occur (Clarke et al., 1984). Such a condition may take place where an aquifer of sufficiently high conductivity (>10–6 m s–1 ) overlies an aquitard of much lower or near-zero conductivity allowing

DIR ICE

ECT

115

passage of meltwater (Hd), at least until fully saturated. This limiting condition may then cause the confined meltwater to flow freely at the surface. Where an aquifer overlies an aquitard that is sufficiently thick and of high permeability, meltwater may pass in a confined state down to a level defined by hydraulic pressure gradient itself related to overburden ice and sediment pressures (He). 3 Where a deforming debris exists beneath an ice mass only a limited movement of meltwater is expected to occur at the upper ice–bed interface while the vast proportion of water will pass into deforming sediment, thereby moving toward low pressure via advective means (M). The volume of

ION

ICE

MOBILE SEDIMENTS IMMOBILE SEDIMENTS MELTWATER UPPER INTERFACE LOWER INTERFACE

BEDROCK

GLACIOFLUVIAL SEDIMENT

FIG. 4.19. Model of possible active interfaces within an active net-melting subglacial environment (after Menzies, 1987; reprinted from Menzies, J. and Rose, J. (eds), Drumlin Symposium. Proceedings of the Drumlin symposium/1st Int. Conf. Geom., Manchester, 1985, p.17, fig. 6; courtesy of A.A. Balkema, Rotterdam).

116

ICE FLOW AND HYDROLOGY

meltwater likely to persist at the upper interface remains a matter of debate. It can be supposed that, under typically anisotropic sediment rheological conditions, patches of sediment of much lower conductivity will occur, thereby leading to similar patches of locally H-bed conditions. This latter state is what is envisioned under Q-bed states. 4.10.2. Models of Subglacial Hydraulic Systems There are several models of subglacial hydraulic flow conditions that describe various modes of meltwater flow at the base of an ice mass, all of which have some supporting field evidence in their favour. Some of these models may be part of a continuum of hydraulic system evolution either developing from one to another over time or in space across the upper ice–bed interface. Therefore, criticism should be tempered with the realization that certain hydraulic systems may develop only under specific glaciodynamic conditions. Each of the models have common and limiting factors that govern their development. In order that meltwater flows freely at the ice–bed interface, it is necessary that melting (r) (cf. discussion on englacial conduits), and the rate of meltwater supply, equals or exceeds the rate of meltwater discharge. The relationship, however, between meltwater discharge and hydraulic pressure varies in accordance with the type of drainage system that evolves (i.e., under a channel-dominant system an inverse relationship exists between these two variables, while in a linked-cavity system a positive relationship appears to occur) (NB discussion on surge mechanisms). Meltwater flow at the ice–bed interface may occur as: (i) sheet flow in the form of a thin water film (Weertman, 1972; Walder, 1982); (ii) channel flow cut down into the glacier bed (N channels) (Nye, 1976) or cut up into the overlying active ice (R channels) (R¨othlisberger, 1972), or a combination of both (as may occur where glacier beds are composed of unconsolidated material) (C channels) (Clarke et al., 1984); or (iii) a linked cavity system developed downstream of bedrock obstacles (Walder, 1986; Kamb, 1987). In all cases, no model is exclusive of the others; thus, for example, sheet flow may co-exist adjacent to channel flow.

(i) SHEET FLOW Weertman (1962) suggested that, under active ice motion, meltwater solely derived from basal ice melting would be evacuated across the ice–bed interface in the form of a thin, virtually continuous sheet of meltwater that in the process would have a positive feedback on ice motion by enhancing basal slip lubrication. It was theorized that surge behaviour might also be coupled with the subglacial water film ‘drowning’ obstacles at the bed thereby facilitating even greater basal ice motion. The increased basal motion would, in turn, cause increased melting owing to increased basal friction leading ultimately to unstable ice motion in the form of a surge. Following Weertman’s analysis, it is assumed that meltwater, originating solely from basal ice melting, the result of geothermal heat and frictional heat from sliding, flows as a sheet under a pressure gradient that is largely a function of the average ice surface gradient. Water film thickness is likely to vary both spatially and temporally across the bed and where high-pressure areas exist no water film will be found. Assuming that the rate of basal ice melting (mb ) per unit time is uniform across the ice–bed interface, and that all the meltwater flows at that interface, the average water film thickness (after Walder, 1982, p.283) is: d¯ ≈

冢

12mb x ⌫

冣

1 3

(4.16)

where x = 0 is a point farthest from the glacier terminus and ⌫ is the meltwater pressure gradient. Walder (1982) has shown that when typical melting rates and pressure gradients are taken into account, average film thicknesses rarely exceed a maximum of ~2 mm and are typically ~1 mm (Fig. 4.20). Figure 4.20 shows that where distances from the glacier snout increase to ~1000 km, meltwater film thicknesses increase to only ~5 mm with a basal melting rate of 15 mm a–1, while for a melt rate of 2 m a–1 film thickness would increase to ~20 mm. It seems likely that sheet continuity and integrity is unlikely to be retained except over relatively short distances of a few tens of square metres of bed area.

ICE FLOW AND HYDROLOGY

(a)

117

10 5 3

1 : Solid Line d : Dashed Line

-1

V=10ma

0.5 0.3

Pg

TIME (a)

1

-3

ba rm

0.01 0.1

0.3 0.5

1

-1

-1

-1

rm

rm

ba

ba

-2

ma

-3

10

00 V=1

x10

ma

-1

0.03

=5

= Pg

-1

V=50

0.05

0 =1

Pg

0.1

3

5

10

30

50

100

SHEET THICKNESS (mm)

(b)

100 50 30 -1

SHEET THICKNESS (mm)

M=2ma

10 -1

-3

Pg =

5

10

-2

Pg =

3

m bar

5x10

.6

a=0

.2

-1

a=3

-1

a=6

m bar

.5

-1

M=15ma

m -2 bar .6 10 a=0 -1 m -3 bar .2 a=3 10 -1 m Pg = -2 bar 0 5x1 .5 a=6 -1 Pg = m -2 bar 10 Pg =

Pg =

1

0.5 0.3

0.1

1

3

5

10

30

50

100

300

500

1000

DISTANCE ALONG GLACIER (km)

FIG. 4.20. (a) Typical development and disintegration of water-sheet perturbations at the ice–bed interface (after Walder, 1982; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 28(99), 1982, p. 280, fig. 2). (b) Thickness of a water sheet as a function of distance along the ice flow direction, assuming all subglacial meltwater flows as a sheet (modified from Walder, 1982; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 28(99), 1982, p. 282, fig. 3).

118

ICE FLOW AND HYDROLOGY

This lack of a widespread continuity in a meltwater film raises the query of non-steady-state versus quasisteady-state flow conditions. Under the parameters defined by Weertman and others where meltwater is solely derived from basal ice melting, increasing ice velocity should, in the short term, lead to increased basal ice friction and higher melting rates. Inversely, as the area of the bed in contact with the moving ice reduces because of greater film thickness, the rate of basal ice melting should, in the longer term, diminish along with basal ice velocity. This state therefore verges upon apparent quasi-steady-state conditions for a meltwater film at the bed on an active ice mass. However, if ice velocity increases beyond a certain critical level and the average obstacle size at the upper interface is effectively ‘drowned’ (leading to an effective zero basal friction value), a state may be reached where, if incipient surge conditions are already present, a surge cycle may be maintained or precipitated leading to non-steady-state conditions. In reality, since only a small fraction of the meltwater at the bed of a valley glacier is usually solely of subglacial derivation (<20 per cent), transient exotic discharges of meltwater into the subglacial system will tend to cause a non-steady-state to prevail. In ice sheets a much larger percentage of meltwater at the bed is of subglacial origin. Therefore quasisteady-state conditions across smooth, low permeability beds where cavities are rare or infrequent may be maintained for short time periods. Finally, it would appear that, even if quasi-steadystate water films can be maintained, they are likely to be of such limited depth to have little or no effect upon substantial alterations in basal ice velocity or bed roughness. The concept of a thin subglacial water film raises an interrelated aspect of subglacial hydrology viz: the relationship between sheet flow and channel flow stability. Walder (1982) suggested that, given the nature of sheet flow, inevitably localized sheet thickenings will occur causing slight uneven, anisotropic hydraulic pressure differences to evolve. As the water flows preferentially in these ‘thickenings’, a local reduction in hydraulic pressure will cause meltwater to migrate increasingly to these zones leading inevitably to degeneration of the sheet flow and the development of an embryonic subglacial

(a)

INTERFACE Z = Zo(Y)

Z ROCK

WATER SHEET

Y X

FLOW DIRECTION

(b)

(c)

FIG. 4.21. Idealized evolution of a water film toward discrete conduit system development. (a) Water film geometry; (b) incipient water film–conduit transition; (c) conduit establishment (modified from Walder, 1982; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 28(99), 1982, p. 276, fig. 1).

channel system (Fig. 4.21). The corollary of such sheet-to-channel evolution is that with increases in basal ice velocity, channels will increasingly encounter bed protuberances that, if greater in size than the channels, will lead to their disruption and a return perhaps to sheet flow. (ii) CHANNEL FLOW Meltwater may exist in sheet and channel flow side by side, or the former may evolve into the latter, or under non-steady-state conditions a hydraulic ‘switch’ from one to the other may occur. Where meltwater from surface-derived sources reaches the bed, meltwater

ICE FLOW AND HYDROLOGY

tends to preferentially flow in a channel system (R¨othlisberger and Lang, 1987; Hooke et al., 1990; Sharp et al., 1998).

119

all types at different places and times along their length (for a discussion on channel form and inchannel pressure systems see Menzies, 1995, chapter 6, pp. 221–222).

Channel Types Channel Maturation/Degeneration Channel types are of three basic forms, viz: (a) R-channels, (b) N-channels, and (c) C-channels; but channel cross-section configuration and how it alters over time in response to changing meltwater discharge, temperature and debris content; bed material composition and intrinsic variability; and glaciodynamic responses, remains largely unknown (Hooke et al., 1990). In modelling subglacial channels it has generally been assumed that channel cross-sections are of a symmetric semi-circular shape in response to competing processes of closure by ice pressures and opening by melting. However, this geometry may only apply to channels oriented parallel to maximum ice flow and the principal longitudinal axis of stress; and/or where ice is moving at very low basal velocities, and in channels totally water-filled considerable distances from the ice front. In other, more typical, conditions channels are likely to exhibit asymmetric cross-sections with considerable variation in channel width down the water-flow line, especially where channels cross at an angle to the principal ice flow direction and/or where channels are not fully filled at all times. A further consideration of channel architecture is the influence of bed material erodibility. Since channels flow across differing bed materials whether of different rock types or sediments, variations in channel cross-section and long-profile geometry must occur. Which type of subglacial channel forms under a particular set of conditions depends upon the nature of the bed materials, the magnitude of meltwater discharge, ice forces and the presence or otherwise of any preferential hydraulic routes at the bed prior to glaciation or the onset of a new channel hydraulic system. In those locations where bed material is particularly resistant, R-channels will tend to form. Where prior flow routes exist or bed materials are weak, N-channels may quickly form. Where extensive unconsolidated bed materials are found, C-channels may develop. It is likely, as examination of exhumed subglacial channels reveals, that channels may be of

As a subglacial channel system evolves, it may link with other channels into a drainage network. Such a network, by definition, has a hierarchical system of channel discharge that under ice pressures etc. will cause larger, trunk channels to develop at the expense of small channels. Shoemaker (1986c) found that approximately parallel channels may develop beneath large ice sheets, closest to their margins, separated by transverse distances of approximately 104 times the channel radius (NB major esker trains in N.W.T., Canada; Aylsworth and Shilts, 1989). Typically, valley glaciers, 2 km or less in width, are thought to have a single trunk subglacial channel. Channel Location Prediction of the location of subglacial channels remains problematic. Control of channel location appears to be regulated by ice stress conditions, the presence of pre-existing channels and bed topography. In general, two subglacial channel types may coexist within any one ice mass, viz: ‘gradient’ channels and bottom channels. The former flow in relation to the gradient of the equipotential piezometric line (R¨othlisberger and Lang, 1987, fig. 10 15, p. 242) and therefore may occur at high levels along valley sides or in the deeper areas of an ice mass close to the main axis of ice flow. ‘Gradient’ channels may migrate downward more easily than upwards because of energy considerations thus transforming from gradient to bottom channels. In mapping glaciated terrain this transition can often be noted where subglacial channels cut across the topographic slope of hillsides and then steeply descend to the lower levels of, or close to, the valley floor (Fig. 4.22). Two further considerations of subglacial channel location remain puzzling: (1) how do subglacial meltwaters traverse areas of overdeepening or bed rises; and (2) how are subglacial channels affected by floating ice margins?

120

ICE FLOW AND HYDROLOGY

pot-holes

compression extension

FIG. 4.22. A proposed relationship between subglacial bed pot-hole locations and the stress field within an ice mass (after R¨othlisberger and Lang, 1987; reprinted from Gurnell, A.M. and Clark, M.J. (eds), Glacio-Fluvial Sediment Transfer, by permission of John Wiley and Sons).

1 Lliboutry (1983) suggested that where glacial overdeepening is encountered, subglacial channels would tend to bypass such areas by a marginal route. However, R¨othlisberger and Lang (1987) argue that the opposite is true and that meltwater may well flow down the glacier axis and cross overdeepened areas. Where up-slope topographic obstructions occur, subglacial channels will often bifurcate and produce a ‘horseshoe’-shaped furrow (sichelwannen) on the stoss-side of the obstruction (Plate 4.8). Such furrows may range in scale from a few millimetres to tens of metres in depth. 2 In most discussions of subglacial and englacial hydraulic regimens within ice masses there is little or no discussion on the unique problems associated with the pressure fluctuations and basal stress conditions close to and at the margins of floating ice fronts. Whether the ice mass passes

PLATE 4.8. Sichelwannen forms (s) from Georgian Bay, Ontario, with differing degrees of development. Shaw has classified these forms as comma forms, being sichelwannen with only one arm developed or missing and are thus transitional forms of sichelwannen. Scale bar = 2 m (photograph courtesy of John Shaw).

ICE FLOW AND HYDROLOGY

into an ice shelf or floats as a tidewater glacier margin, there are distinctive hydraulic conditions that must impact upon the hydraulic systems developed up-ice from the grounding-line position (Powell, 1990; Anderson and Ashley, 1991). First, channels, whether subglacial or englacial, entering a body of water from a floating ice mass margin must be fully filled. Second, the hydraulic pressure developed by standing water, in a lake or sea, will be higher, for equivalent depth, than the ice just immediately up-ice of the grounding-line. Finally, this higher equipotential will cause quasiartesian conditions to exist within the channels since the water equivalent line, at and snoutward of the grounding-line, will be above the ice surface. This final effect explains the bubbling up of water often observed when glaciers enter lakes, and the jet- and plume-effects and related sed-

121

imentation processes observed subaquatically at many floating ice margins (Plate 4.9). Further effects of a floating margin are: (1) the tendency for climatically controlled temperature fluctuations, that influence meltwater flow and seasonal discharge, to be diminished close to the floating margin owing to lake or sea water thermal effects; (2) back-pressure, from standing water upice of the grounding-line, will help maintain melting rates and reduce closure rates of tunnel walls, but will also critically influence longitudinal channel velocity and sediment load competency. Back-pressure effects, acting on a channel’s hydraulic regime, will stretch back beneath an ice mass, where the glacier bed is essentially horizontal, for at least ~1.1 times the depth of water margin-ward of the grounding-line.

PLATE 4.9. J¨okulhlaup at Grimsv¨otn, Iceland in 1982 (photograph courtesy of Helgi Bj¨ornsson).

122

ICE FLOW AND HYDROLOGY

Channel Stability Where perturbations to the system occur through fluctuations in meltwater discharge, channel dimensions and/or transient meltwater pressures, channel form quickly adjusts. The size of change required varies in magnitude dependent upon channel size, location, discharge and other glaciodynamic aspects of the particular channel’s surrounding environment. It has been argued that over the long term, because of ice pressures on the bed, R-channels, solely supplied with subglacially derived meltwater, would become starved of meltwater and would cease to function in a progressive up-stream to down-stream degeneration of arborescent drainage systems. In contrast, Walder (1982) has shown that sheet flow is inherently unstable leading to incipient channel development where preferential flow in deeper parts of a sheet leads to channel evolution (Fig. 4.23). Two reasons for channel instability are (1) that pressure fields along the edges of R-channels will prevent meltwater migrating into the channels

log

-df ds

with heat advection without heat advection enlargement of conduit

nd

un s

ui

t

closure of conduit

tab

le

-df ds

co nd

uit

sta

bl

e

co

min

thereby ‘starving’ them; and (2) that developing channels encountering obstacles at the ice–bed interface will be disrupted by them at a rate equal to the sliding velocity of the ice mass and as bed roughness increases. Where sheet flow occurs with depths of <4 mm, quasi-stable sheet flow may occur, otherwise channel flow will more typically develop. The only set of conditions under which sheet flow, as opposed to channel flow, may occur is where few subglacial cavities exist, bed roughness is slight and large volumes of extraneous meltwater from englacial and supraglacial sources penetrate to the subglacial bed interface. A further form of channel instability occurs when, under increasing meltwater discharge typically derived from external, non-subglacial sources, channel wall melting increases at a very rapid rate in comparison with closure rate. This situation is typical of ‘flood events’ caused by seasonal discharge fluctuations, surges and ice-dammed lake outbursts (note the recent outburst at Skeldar´arj¨okull, in Iceland; Bj¨ornsson, 1998). Under these circumstances, a limited but unstable condition is reached, at least in the short-term (a few hours or days) leading to a rapid headward and lateral expansion of the drainage system. It is likely that when heat from flowing meltwater is advected along the conduit and is no longer available for continued ice-wall melting, conduit closure begins even at high discharges and thus unstable conditions resume until discharge drops to a lower volume and stable conditions once again resume (Kamb, 1987, fig. 12). Where most of the discharge is from external sources and thus discharge at unstable high volumes is not reduced by self-regulation, the instability inherent in this condition may ultimately lead to channel degeneration and either massive sheet flow or to a linked-cavity hydraulic system. (iii) LINKED-CAVITY FLOW

log R FIG. 4.23. Incipient instability within subglacial conduit draining major subglacial meltwater reservoirs (after Spring, unpublished, and R¨othlisberger and Lang, 1987; reprinted from Gurnell, A.M. and Clark, M.J. (eds), Glacio-Fluvial Sediment Transfer, by permission of John Wiley and Sons).

A linked-cavity flow hydraulic system may develop when meltwater discharges via a series of lee-side cavities linked by narrow, short R-channels (orifices) cutting transversely across a glacier bed (Kamb, 1987; Menzies, 1995, chapter 5, fig. 5.28; Bj¨ornsson, 1998). Evidence for a linked-cavity hydraulic system rests

ICE FLOW AND HYDROLOGY

upon: (a) mapping of previously glaciated areas; (b) modelling of time-related dispersion of dye tracers in present day ice masses; and (c) subglacial meltwater pressure observations under present day glaciers (Bj¨ornsson, 1998). Unlike channel systems, no intrinsic instability exists within a linked-cavity system at high meltwater discharges (Fig. 4.24). As discharges increase, the orifices act to self-regulate the through-flow of meltwater by effectively ‘throttling’ the flow from one cavity to another. However, a linked-cavity system would only appear capable of operation under relatively high basal meltwater discharges; once meltwater flux levels fall below the level necessary to maintain the orifice connections (owing to ice wall closure) either channel or sheet flow hydraulic systems will re-form. It would appear that the linkedcavity hydraulic system is peculiar to certain glaciers and glacier bed states and is only fully active under hyper-flow meltwater conditions, probably of the type found at or immediately prior to the commencement of a surge. 0

(Step-Orifice Model)

5

100

Linked-Cavity System (Wave-Orifice Model) 10

TUNNEL

15 200

0.01

0.1

1

10

= ( PI - PW) (bar)

150

EFFECTIVE PRESSURE

WATER DEPTH BELOW SURFACE (m) (for ice thickness 400m)

50

100

DISCHARGE QW(m3/s)

FIG. 4.24. Hypothetical graph of effective confining pressure against subglacial discharge via conduit system types (step and wave orifices, and tunnels) (after Kamb, 1987; Journal of Geophysical Research, 92B, p. 9094, copyright by the American Geophysical Union).

123

4.11. HYDRAULIC SYSTEMS AND DEFORMABLE BEDS All the discussion so far on hydraulic systems has been largely concerned with rigid beds (H-bed states). However, increasing evidence suggests that large portions of many glacier beds today and in the past were probably composed of soft unconsolidated beds (M- or Q-bed states), deformable beds (Menzies, 1995, chapter 5; Menzies, 1996, chapter 2). Where deformable sediment, saturated by meltwater, exists beneath ice masses, deformation occurs when the applied shear stresses from the overlying active ice overcome the sediment’s internal shear strength. A major controlling variable influencing this process is the intrinsic viscosity of the sediment that reflects the level of consolidation, grain size and porosity of the sediment. If it is assumed that the sediment at the ice– bed interface is deforming, then it is generally thought that little or no free meltwater, derived from solely subglacial sources, would exist at the upper interface between the mobile sediment and glacier sole (Murray and Dowdeswell, 1992; Murray, 1994). However, where meltwater from external sources reaches the glacier bed, a slightly different set of circumstances may occur. In the first instance, extraneous water cannot be totally absorbed by the already saturated sediment and thus must move along the upper interface. The absorbed meltwater will aid in reducing sediment viscosity, further allowing the sediment to deform at an even greater rate. The remaining unabsorbed water, however, must find a pathway along the upper interface or within the sediment itself in the form of pipe-flow. Inconclusive evidence for the latter form of flow does exist within many diamictons where pipes and small intercalated channels of stratified sediment are commonly observed (Eyles et al., 1982; Menzies et al., 1997). Meltwater flow at the upper interface of a deformable bed cannot be directly compared with water flow under rigid bed conditions since sediment deformation causing channel closure must also be considered (Alley, 1989a,b). Where temperate ice overlies unconsolidated sediment capable of deformation several models of upper interface/meltwater interrelationships have been developed that consider the microscale conditions at the upper interface, and the

124

ICE FLOW AND HYDROLOGY

complex macroscale relationships between differing immobile sediment–deforming sediment–ice interface boundaries. Channel closure, within a soft bed state, is a function of the rate of sediment movement into the conduit, which in turn is a consequence of the stress applied to the sediment, the sediment’s yield strength and porewater content (Menzies, 1995, chapter 6, pp. 227–229). Any sediment channel (Alley’s till channels) will remain open if infilling sediment can be evacuated at a rate comparable with the closure rate. Thus any meltwater channel system at the junction of unconsolidated deformable sediment and (a)

the base of an ice mass is likely to be short-lived and therefore of minor significance (Fig. 4.25(a,b)). R-channels, if they exist, at low effective stress levels may be stable if a few millimetres in radius or less. Where subglacial sediment has a high hydraulic conductivity and/or is overlying an aquifer of similar or higher conductivity, a high flux rate of meltwater evacuation via porous flow may be possible assuming the sediment does not deform and acts in a ‘rigid’ structural manner. Provided meltwater production at the upper interface does not exceed the through-flux of porewater within the sediment layer, meltwater may be removed successfully (Shoemaker, 1986c,

r (m) 10-3 11

10-1

101 11 10

Grow

Like

t* =

ly wa

9

ter

Shrin

0

109

t*m

0 t* =

w Gro r wate u im m x a M k Shrin

7

5

107

t* max

Grow

Shrink

105

3 -3

1

100 10

0.1 0.1 0.01 0

5

0.01

0

-0.1

-1

-0.1

-0.01

-10-5 -10-4 -10-3 3

Log10 (r m)

-10

100 0

Turbulent

-1

105

100

t*likely

Grow Laminar

1000

1

t*max

101 7 10

10-1

10

R channels

Till channels

r (m) 10-3 7

N (Pa)

k Shrin w o r G Shrink

(b)

ax

Log10 (N Pa)

Till channels

N (Pa)

Log10 (N Pa)

k

1

10

3 -3

-1

1

103

Log10 (r m)

FIG. 4.25. Effective confining pressure against subglacial channel radius for R and ‘Till’ channels. At high effective pressures ‘Till’ channels exhibit stability. A ‘Till’ channel plotted above a high effective confining pressure tends to grow and below the channel shrinks. Within the stippled zone both R and ‘Till’ channels shrink at equal rates (after Alley, 1989b; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 35(119), 1989, p. 113, fig. 2).

ICE FLOW AND HYDROLOGY

Type II). It is unlikely that meltwater flow through a porous sediment can ever account exclusively for all meltwater produced at the upper interface. A substantial proportion of the meltwater must flow across the upper interface in channels and/or as sheets. 4.12. SUBGLACIAL LAKES Subglacial lakes (basal or sub-ice lakes) occur as large pondings of meltwater that have accumulated beneath ice sheets. These lakes appear to form under conditions of temperate basal melting but are prevented from evacuating because of the constraining effects caused by a hydraulic seal as result of high normal ice overburden pressures, thermal fluctuations in basal ice or a topographic high (Dowdeswell and Siegert, 1999; Siegert, 2000). The areal extent of these subglacial lakes is difficult to establish, but a lake beneath Vostok Station, East Antarctica, is estimated to be 14 000 km2 (Goodwin, 1988; Siegert, 2000). Many other small lakes and pondings occur beneath ice masses of all sizes. In Antarctica at least 77 lakes have been detected, 70 per cent of which lie beneath the East Antarctic Ice Sheet (Siegert et al., 1996). However, only large subglacial lakes can appreciably affect the regional glaciodynamics of the ice sheet for some distance around the lake perimeter producing either flat areas or slight depressions within the general surface topography. Since a large body of subglacial meltwater will cause the basal shear stress level of the ice mass to drop locally to near zero, resulting in an increase in the longitudinal strain rate, this effect is transmitted threedimensionally in all directions away from the ponded area. The geomorphological effect of these large lakes is two-fold: (1) as localities of sediment accumulation, and (2) as sources of major subglacial flood events, when the barrier seal ‘breaks’ (Baker et al., 1988; Shaw et al., 1989, 2000). 4.13. HIGH-MAGNITUDE MELTWATER DISCHARGES There are perhaps few glacial phenomena quite so spectacular as the sudden and episodic outbursts of meltwater from subglacial and/or supraglacial lakes, ice-dammed lakes or overspilling proglacial lakes

125

(Tweed and Russell, 1999). These outbursts are collectively termed ‘j¨okulhlaups’. Such outbursts or ‘flash’ floods were thought to be characteristic of temperate ice masses but evidence of a non-cyclic j¨okulhlaup has been observed in East Antarctica under polar ice conditions (Goodwin, 1988). Such catastrophic events have a considerable impact upon the entire proglacial zone and beyond (Bj¨ornsson, 1998).Within the proximal proglacial zone large blocks of ice may be left stranded, later creating ‘reversed’ thermokarst terrain. Substantial thicknesses of sediment are deposited, new proglacial stream networks formed and previous ones abandoned, annual moraines breached and extensive areas of bedrock scoured. Perhaps the greatest impact j¨okulhlaups have is in their devastating floods that cause loss of human life, livestock and destruction of property (Clague and Evans, 1997). J¨okulhlaups destroy bridges, roads, buildings and breach pipelines. It is the association of j¨okulhlaups with human tragedy in one form or other that has left a mark on folk histories among the peoples of the European Alps, the Pamirs and Iceland. Table 4.7 chronicles the passage of repeated j¨okulhlaups for the Skiedarfårj¨okull area of Katnaj¨okull from 1332 to 1972. 4.13.1. Discharge As the Icelandic term ‘j¨okulhlaup’ indicates, a glacial meltwater stream experiences a very rapid and enormous increase in discharge. This increased water volume usually continues for three to four days and just as quickly subsides (Fig. 4.26). In this period of increased flow, conditions are suitable for significant changes in rates of debris erosion and entrainment. T´omasson (1974) reported >3000 tons of sediment moved in the outlet streams over a seven-day period following the 1972 Grimsv¨otn j¨okulhlaup. J¨okulhlaups cause destabilization of proglacial (sandur) sediments leading to extensive slumping and debris flows (Russell et al., 1990). Detailed analysis of this sediment transport and distance of particular grade size movement from the Grimsv¨otn events is illustrated in Fig. 4.27 (Tweed and Russell, 1999). Typical meltwater discharge volumes vary immensely according to the type and nature of their sources. Table 4.8 illustrates some examples of

126

ICE FLOW AND HYDROLOGY

TABLE 4.7. A record of J¨okulhlaups from 1332 until 1972 from Iceland (from Thorarinsson, 1974; translation by Helgi Bj¨ornsson) Month

Year

November May November July Uncertain February Uncertain (December) (September) October April July February April June June June Uncertain May August

1332 1341 1598 1619 1629 1638 1659 1684/85 1706 1716 1725 1766 1774 1784 1796 1816 1838 1851 1861 1867

June March

Jokulhlaup ¨ in Skeioar ¨ a´

Eruption in Gr´ımsvotn ¨

§ § § § # # § # § § # § # §?

10 6

§a § § § # # § # § § # § # # # # # # # #

1873 1883

6 10

# #

# #

March January May

1892 1897 1903

9 5 62

# # #

April September March May

1913 1922 1934 1938

10 92 112 4

# # # #

April September February July January September May

1941 1945 1948 1954 1960 1965 1972

3 4 22 62 52 52 62

# # # # # # #

a

Interval since previous Jokulhlaup ¨ (years)

812b 10 9

9 72 10 12

Remarks

§ § §

Main eruption south, southwest of Grimsvotn ¨ Greatest production of tephra Eruption of Thordarhyma

# # Probable eruption north of Grimsvotn ¨

The symbol § indicates that the event is uncertain. Only those intervals are included when it is most likely that no j¨okulhlaups are missing.

b

measured and estimated volumes of water from modern and Quaternary j¨okulhlaups. Characteristically, the discharge hydrograph of a j¨okulhlaup event exhibits an exponential increase in discharge with gradually ascending and very steeply descending limbs (Fig. 4.26). Clague and Mathews (1973)

derived an empirical relationship between peak discharge (Qmax ) and total volume of drained meltwater (Vmax ) during a J¨okulhlaup event where: Qmax ≈ 75

冢 10 冣 Vmax 6

0.67

(4.17)

flow summation curve

10 hydrograph for a jökulhlaup

5

0

127

5

10

4.0

2.0

15

20

Volume

Discharge 103m3/s

ICE FLOW AND HYDROLOGY

0 30

25

July, 1954 FIG. 4.26. Characteristic hydrograph of a j¨okulhlaup. J¨okulhlaup of 1954 from Grimsv¨otn via Skeidar´a (Rist, 1955). Rist suggests accuracy of discharge estimate to be ~20 per cent. Note rapidity of flow cut off within a few hours (afer Bj¨ornsson, 1974; reproduced by permission of the Iceland Glaciological Society).

8000 Grimsvötn

ICELAND

7000 ll

Key Map ku

ar Súla Gigj a

5000 4500

ll jöku rsár llsjökull e f ta

Mo

ll ku l s j ö kull inafe l jö a Sv f ¨Ore

Sk a f

5500

Skaftafellsfjöll

Skeidará

Sk

eid

ll

6000

ku

Falljökull

Skeidarársandur

4000

Atlantic

Oce

an

3500 3000 2500 2000

e Sk

á ar id gj a

Suspended Sediment Concentration of Fine Silt

ujö

á r jö

6500

Sid

(mg/l)

7500

G

1500

i

la Sú

1000 500 17

18

19

20

21

22

23

24

25

26

27

28

29

30

Days in March, 1972 FIG. 4.27. Variation in suspended sediment load for three outlet proglacial streams of Vatnaj¨okull resulting from a j¨okulhlaup related to Grimsv¨otn (Grimsv¨otnahlaup of 1972). Location of streams shown in inset map (after T´omasson, 1974; reproduced by permission of the Iceland Glaciological Society).

TABLE 4.8. J¨okulhlaup data for various examples of both modern and Pleistocence ice dammed lakes (after Clague and Matthews, 1973; Beget, 1986) L (km)

D (m)

V (106 m3)

29a

2.6b

150

Whalley, 1971

4.8

200

Church, 1972

11.6

1000

Strom, ¨ 1938

Lake

Year

H (m)

Strupvatnet, Norway

1969

186

1

Ekalugad Valley, Baffin Island, Canada

1967

120

2

120

Demmevatn, Norway

1937

406

–3

79 a

Q (m3 s–1)

Reference

Gjan ´ upsvatn, ´ Iceland

1951

167

5

20

20.0

370

Vatnsdalur, Iceland

1898

372

10

188

120.0

3000

Tulsequah Lake, B.C., Canada

1958

150

8

73

229.0

1556

Marcus, 1960

Summit Lake, B.C., Canada

1965, 1967

620

12

200

251.0

3260

Matthews, 1965

Graenalon, ´ Iceland

1939

535

19

230

1500.00

5000

Thorarinsson, 1939

40

9

40

1730.0

10100

a

Lake George, Alaska, USA

1958

Lake Missoula, Montana, USAc (Wallula Gap) Lake Missoula–Upper Spokane Valley/Rathdrum Prairie

640 Pleistocene

Lake Missoula–Wallula Gap

–

610 –

–

–

2 × 106

–

Arnborg, 1955 Thorarinsson, 1939

Stone, 1963

1.87 × 106

Bretz, 1925; Pardee, 1942

17±3 × 106

O’Connor and Baker, 1992

10±2.5 × 106

O’Connor and Baker, 1992

Mulakvisi, Iceland

1956

–

–

–

3.5 × 10

Kaldakvisi River, Iceland

1965

–

–

–

6.20 × 106

260

Freysteinsson, 1972

Hazard Lake, Yukon Terr., Canada

1978

–

–

–

19.4 × 106

640

Clarke, 1982

6

50

Kverka, Iceland

1980

–

–

–

20.6 × 10

Katla, Iceland

1955

–

–

–

28.0 × 106

2500

Snow River, Alaska, USA

1967

–

–

–

140 × 106

780

6

6

410

Rist, 1968

Rist, 1982 Thorarinsson, 1957 Chapman, 1981

Flood Lake, B.C., Canada

1979

–

–

–

150 × 10

1200

Clarke and Waldron, 1984

Sululaup, Iceland

1978

–

–

–

175 × 106

3000

Rist, 1983

Skeidararsandur, Iceland

–

–

–

–

200 × 106

2000

Bjornsson, ¨ 1977

6

1500

Bjornsson, ¨ 1977

6100

Sturm, 1986

Skaftardalur, Iceland

1970

–

–

–

237 × 10

Strandline Lake, Alaska, USA

1984

–

–

–

710 × 106

a

Depth to bedrock knob limits J¨okulhlaup magnitude = 13 m. Volume stored in lake = 4.6 × 106 m3; water released = 2.6 × 106 m3. Lowering of lake surface during J¨okulhlaup.

b c

ICE FLOW AND HYDROLOGY

129

This equation seems to give reasonably consistent results but is not founded in theory. Clarke (1982) working on ‘Hazard Lake’, dammed by the Steele Glacier, Yukon Territory, Canada, found the above empirical equation to work reasonably well on the basis of measured lake volume and discharge.

or supraglacially, or by some non-glacier route; or (b) by discharge on a periodic basis by overspilling; or (c) by j¨okulhlaup-style discharge. In most cases an almost continuous leakage takes place from icedammed lakes negating the idea of a watertight ice– bedrock seal.

4.13.1.1. Sources of meltwater

4.13.1.2. J¨okulhlaup initiation mechanisms

The major sources of j¨okulhlaup water are from either: (1) lakes impounded by active or stagnant ice and morainic debris, or (2) lakes in natural depressions in proglacial, subglacial or supraglacial positions (Bj¨ornsson, 1998). Except for supraglacial lakes, all other lakes form in depressions that are the result of glacial processes or glacial loading (isostatic depression) or in glacier-free tributary valleys (Fig. 4.28). The latter type are what are generally known as ice-dammed lakes, being impounded by a barrier of glacier ice interrupting and intercepting normal drainage patterns. Discharge from ice-dammed lakes tends to be (a) by subglacial leakage, along the ice margins

Several hypotheses have been advanced to explain j¨okulhlaup initiation triggers (Menzies, 1995, chapter 6, p.236; Bj¨ornsson, 1998). There are specific characteristics of j¨okulhlaups that need to be included in any triggering mechanism: (1) the rapidity of the meltwater outburst; (2) the sudden cessation of flow often before a lake is emptied; (3) the periodicity of the outburst events; and (4) the lack of a watertight ice– bed seal. At present three main hypotheses have been suggested to explain the triggering mechanism that drains ice-dammed lakes: (1) ice-barrier buoyancy, (2) plastic deformation of the ice barrier, and (3) the

0

1

2

3

Strandline lake High water 388 m (Aug. 21, 1982) Low water 264 m (Sept. 23, 1982) Small lake ~600 m

Plunge pool Upper lip ~320 m Lower lip ~200 m Exit tunnel (1980-1982) ~150 m Rock channels

4

5 kilometres

N

Approx. location of drainage tunnels

Supra-glacier pools 386 m (Aug. 21, 1982) Exit tunnel (1974) TRIUMVIRATE GLACIER

Upper Beluga Lake 75 m

FIG. 4.28. Effects of j¨okulhlaup flood as shown in the Strandline Lake and Triumvirate Glacier area of Alaska. J¨okulhlaup subglacial channel location shown by dot-dash line (after Sturm and Benson, 1985; reproduced by courtesy of the International Geological Society from Journal of Glaciology, 31(109), 1985, p. 273, fig. 2).

130

ICE FLOW AND HYDROLOGY

time-lag between subglacial water pressure reduction and exit channel closure. 1 It has been suggested that when the ice overburden pressure at the base of the ice barrier became less than the hydrostatic water pressure within the lake because of increasing lake water depth, the ice barrier would become buoyant. 2 Once the depth of an ice-dammed lake reached approximately 200 m, at which time a shear stress of 100 kPa at the ice barrier would develop, plastic deformation of the ice in the barrier would result. Since j¨okulhlaups take place from lakes of lesser depths this explanation seems untenable. 3 A further hypothesis relates to the time-lag that exists between variations of subglacial water discharge and pressure within subglacial conduits and the time necessary for the ice to plastically readjust to the reduced flow regime. 4.13.1.3. J¨okulhlaup stages of initiation and development Any j¨okulhlaup develops in a two-stage process, the first a triggering mechanism and the second the subglacial evacuation of lake water. This second phase, within temperate ice, occurs through the enlargement of already existing subglacial drainage networks. As water penetrates these networks from the impounded lake, ‘uncontrolled’ enlargement occurs by channel-wall melting caused by higher lake water temperatures, and side-wall friction and abrasion caused by suspended debris and water turbulence (Fig. 4.23). Flow continues from the lake provided channel enlargement is greater than the rate of

closure. These competing forces alter as the lake level drops and ice overburden pressures increase both at the entrance portal and within the adjacent subglacial drainage network. Typically, j¨okulhlaups end abruptly. The cessation of flow may be due to subglacial channels being squeezed shut, channel roof collapse and channel infill or bedrock lips cutting off continued meltwater supply either at the entrance portal or down-ice within the subglacial network system. 4.14. CONCLUDING REMARKS The rheological behaviour of any ice mass plays a fundamental role in glacial processes. To understand these processes it is crucial that ice mass glaciodynamics be understood within a context relevant to glacial sedimentology. The potentially important role that deformable bed states may have on ice motion and glacial bed conditions must be better understood. It is increasingly apparent that glacial hydrology is a critical element in our understanding of many glaciodynamic processes. The role of glacial meltwater can no longer be regarded as of secondary relevance in glacial environments solely as a means of sediment transport and redistribution within the glacial system but rather, for example, as the controlling variable in the mobility and persistence of deformable beds, in the development and evolution of certain wear phenomena across the glacier bed, in the initiation and consequent maturation of specific subglacial bedforms, and in the ‘construction’ and sedimentological architecture of sediment lithofacies types and associations.

5

PROCESSES OF GLACIAL EROSION N. R. Iverson 5.1. INTRODUCTION The basic processes of glacial erosion responsible for the spectacular landforms associated with alpine glaciation were identified over a century ago. Forbes (1846) recognized that rock clasts embedded in glacier ice abrade the underlying rock bed, and Tyndall (1864) noted that glaciers quarry rock blocks bounded by joints. Detailed observations of striation patterns and grooves made subsequently by Chamberlin (1888) established that ice behaves as a deforming fluid rather than a rigid body. These observations helped Gilbert (1903, 1906) relate ice flow at glacier beds to mechanisms of rock fracture and erosion. The most illuminating recent studies of glacial erosion have paralleled Gilbert’s pioneering efforts, explicitly linking glaciological theory with the processes that erode the bed. This approach, complemented by observations in the field and laboratory, has provided a fuller understanding of erosional processes than that possible solely from inferences based upon field observations (Plate 5.1). Three mechanisms of erosion are traditionally considered: abrasion, quarrying and the action of subglacial water. Turbulent flow of water in zones where ice has separated from the bed erodes sinuous channels, pot-holes, and elongate crescentic furrows around resistant obstacles on the bed (Sharpe and Shaw, 1989) and dissolution leaves a clear imprint on carbonate bedrock (Hallet, 1976). Despite the local

PLATE 5.1. Bedrock exhibiting striae and chattermarks. Glacier (Omsbreen, Norway) had flowed from left to right.

131

importance of erosion by subglacial water, it is usually considered to be volumetrically subordinate to abrasion and quarrying (Drewry, 1986, p. 90). Unequivocal data in support of this viewpoint are lacking, but the ubiquity of glaciated bedrock smoothed by abrasion and roughened by quarrying suggests that these two processes are dominant in many

132

PROCESSES OF GLACIAL EROSION

areas (Plate 5.1). Abrasion and quarrying involve the fracture and dislodgement of rock from the bed. In the case of abrasion, fracture results from large stress differences localized beneath ice-entrained rock fragments in frictional contact with the bed. In the case of quarrying, such stresses are induced directly by ice, usually over larger areas of the bed and for longer periods of time. Abrasion, therefore, produces fine debris, ranging from fine silt to coarse sand, whereas quarrying generally produces larger rock fragments. An exception would be large boulders entrained in the glacier sole that may fracture the bed on scales usually associated with quarrying. Dislodgement of quarried fragments requires that frictional forces be overcome by the combined action of water and sliding ice. The products of abrasion are usually accomplished by the abrading tool.

5.2. ABRASION Early this century, Gilbert discussed many of the factors governing abrasion, recognizing that abrasion is: (1) a brittle process, involving fracture of the bed beneath clasts (Gilbert, 1906); (2) that it depends on the velocity of the basal ice, the quantity of subglacial debris and the hardness contrast between the debris and the bed (Gilbert, 1903, pp. 203–205); (3) that it also depends on the stress that clasts exert on the bed, in excess of that exerted adjacent to points of clast–bed contact; (4) that such stress concentrations beneath clasts result from the flow of ice toward the bed; and (5) that the rate of ice flow toward the bed depends on the bed geometry and the sliding speed of the glacier. As discussed hereafter, these concepts have since been incorporated in most quantitative models of glacial abrasion (Hallet, 1979; Riley, 1982; Shoemaker, 1988). Virtually all studies consider the primary agents of abrasion to be clasts embedded in the basal ice of glaciers. Alley et al. (1987a, b), however, has suggested that a subglacial deforming sediment layer that is free of ice might effectively abrade underlying bedrock, consistent with earlier conjecture by Gjessing (1965). Given the possible role of such layers in glacier motion, this bed condition will also be considered.

5.2.1. Abrasive Wear Abrasion beneath glaciers occurs by fracture and subsequent displacement of minute particles of bedrock. Stresses sufficient to cause plastic deformation should not develop at contacts between clasts and the bed, except perhaps in a very limited zone directly adjacent to asperities on the contact. The dominance of brittle deformation is clearly indicated by close examination of glacial striae, which ‘present rough borders. . . whose edges are bruised, torn, or hackly’ (Chamberlin, 1888, p. 218), and also by the main product of abrasion, rock flour, which consists of angular silt particles. Clasts indent the bed and produce such particles by indentation fracture (Lawn and Wilshaw, 1975), a process well studied in materials science and rock mechanics. Consider a point on a clast, regardless of its exact shape, impinging on the bed under a normal contact force, Fc . A good approximation is to assume the bed and point behave as linear elastic solids. In a crosssection of the bed, the most compressive (+) principal stress σ1 is directed radially away from the contact, while the least compressive principal stress σ3 is concentric about the contact (Fig. 5.1). This latter stress becomes tensile immediately adjacent to and also at some depth below the contact. In the extreme case of a perfect point load, this stress is tensile everywhere near the contact. Like all rocks, the bed contains microscopic cracks with many orientations. The largest cracks that lie approximately normal to σ3 and in zones where σ3 is tensile will be most likely to grow, owing to tensile stresses that are concentrated at the tips of such cracks (Atkinson, 1984). Consequently, if Fc is sufficiently large, radially oriented macroscopic cracks progressively develop beneath the contact (Fig. 5.2). The cracks eventually begin to merge and become sufficiently pervasive for the point to displace the crushed debris, and thus indent the bed (Fig. 5.2). The cross-sectional area of the fractured zone is directly related to Fc . Crack growth is probably slow and intermittent; it is facilitated by the chemical action of water, which weakens strained bonds at the crack tips (stress corrosion; Atkinson, 1984). Thus, to some extent, the process is time-dependent, suggesting that the depth of abrasion may decrease

PROCESSES OF GLACIAL EROSION

133

FFcc Contact

s3

s1 -0.17 0.5 0.1 0.025 0 -0.005

FIG. 5.2. Cross-sectional view of progressive indentation fracture of limestone and granite by a truncated and sharp wedge, respectively, under a contact force Fc , as observed with a scanning electron microscope (SEM) (modified from Lundqvist et al., 1984; reprinted from the International Journal of Rock Mechanics and Mineral Sciences and Geomechanics, Abstract, 21 (4), ‘Indentation fracture development in rock continuously observed with a scanning electron microscope’, 165–182, 1984, with kind permission from Elsevier Science Ltd., The Boulvard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

-0.001 FIG. 5.1. Principal stresses beneath an elastic sphere pressed against a flat elastic bed (Hertzian contact). Cross-sectional view of trajectories of ␴1 and ␴3 (top), and contours of ␴3 normalized with respect to mean normal stress at the contact (bottom). Shaded area in bottom figure is zone where ␴3 is compressive. Crack growth should be initiated at edge of contact where stress is most tensile (–0.17) and should occur normal to ␴3 (modified from Lawn and Wilshaw, 1975); reprinted from Journal of Materials Science, 10, ‘Review of indentation fracture: principles and applications’, 1049–1081, 1975, with kind permission from Elsevier Science Ltd., the Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

with increasing clast speed (Scholz and Engelder, 1976). If a tangential load is applied in addition to Fc , tensile stresses adjacent to the trailing edge of the contact increase and extend deeper into the bed, and conversely beneath the leading edge (Lawn, 1967).

This produces the crescentic fractures that often occupy the bottoms of glacial striae (Plate 5.2). They are concave down-glacier in plan view, and roughly demarcate the trailing edge of the former contact. Stress fields may also be asymmetric owing to the torque that is expected on clasts, which focuses Fc on the leading edge of the contact. This may produce tensile stresses that reach maximum values down-ice of the contact and is responsible for the primary cracks of crescentic gouges (Johnson, 1975). Tensile stresses may also develop at the leading edge of a contact if the point has indented the bed, resulting in crack growth and coalescence that allows the point to plough forward (Iverson, 1991a). A wear law relates point motion and contact force to the rate at which material is eroded. Most theoretical treatments of glacial abrasion (Boulton,

134

PROCESSES OF GLACIAL EROSION

depths to which different clasts abrade vary significantly, even when Fc is steady (Iverson, 1991a). 5.2.2. Effective Contact Force and Clast Velocity

PLATE 5.2. Crescentic fractures (chattermarks) at the bottom of a striation. SEM photomicrograph, magnification 700×).

1974; Hallet, 1979, 1981; Drewry, 1986; Shoemaker, 1988) use essentially the same relation. Hallet (1979) ˙ as: gives the rate of abrasion, A, A˙ ≈ ␣ch Cr uc Fc

(5.1)

where ␣ch is a constant dependent on the hardness of clasts and the bed and the geometry of the striator point, Cr is the areal concentration of clasts in contact with the bed, uc is the clast velocity and Fc is the effective contact force between the indenting clast and the bed (Drewry, 1986, p. 55). The abrasion rate, therefore, depends on the flux of abrading clasts, Cr·uc , and the depth to which clasts indent the bed, proportional to ␣ch·Fc . This relation follows from Archard’s (1953) original wear model and is supported by geophysical data on frictional wear between sliding rock surfaces (Scholz, 1987). It is expected to be a reasonable approximation for average rates of abrasion after many clasts have passed over the bed, but may be a poor approximation when applied to an individual event. This is because the depth to which a clast indents the bed also depends on the shear loading at the contact. This loading depends not only on Fc but also on the geometry of the striating element and the potential for rotation of the clast in the ice. Thus, the

The glaciological parameters that control the effective contact force and the clast velocity are discussed below, initially without reference to how these variables may vary spatially on a rough glacier bed. The ice is considered to be at its pressure melting temperature and sliding over the bed. Sliding at subfreezing temperatures occurs (Shreve, 1984; Echelmeyer and Wang, 1987), but at a speed that is probably several orders of magnitude slower than that of temperate ice. The glaciological parameters that control Fc for isolated clasts are most important because Fc influences both the depth to which clasts indent the bed and the clast velocity. Clasts will typically be completely embedded in basal ice, particularly along stoss surfaces where abrasion is most effective (Hallet, 1979, 1981). The deviatoric stresses in the bed that result in indentation fracture only develop if contact stresses exceed the pressure on the bed adjacent to the contact. Thus, hydrostatic ice pressure, which contributes equally to the contact stress and the pressure adjacent to the contact, does not contribute to Fc . Stresses, instead, are concentrated beneath clasts by motion of ice toward the bed. This induces a bednormal gradient in pressure across the clast that forces ice to regelate and deform past the clast. The resultant drag is a major component of Fc . It is represented by the term, ⌿uN in the relation: Fc ≈ By + ⌿uN

(5.2)

where By is the buoyant weight of the clast, uN is the component of the ice velocity normal to the bed, and ⌿ is a drag coefficient that increases with the effective ice viscosity and clast size. The non-linearity of the flow law for ice can be included explicitly in the formulation of ⌿ following Lliboutry and Ritz (1978). By is negligible, except in the case of bouldersized clasts (Hallet, 1979). Laboratory simulations of abrasion by sparse debris have demonstrated the direct relationship

PROCESSES OF GLACIAL EROSION 4000

3000

3500

2500 Shear force

2500

1500

2000

1000 Velocity

1500 Shear force (N)

2000

1000

43

45

47

49

500 51

53

55

57

59

61

6000

0 3000

5500

Shear force

2500

5000

2000

4500

1500

4000

Velocity

1000 500

3500 3000

Ice velocity toward the bed (mm a-1)

3000

67 69 71 73 75 77 79 81 83 85 87

0

Time (hours)

FIG. 5.3. Bed-normal ice velocity and shear force during two experiments in which a flat rock bed was slid beneath temperate ice containing sparse clasts. Shear force results from friction between clasts and bed and is approximately equal to ␮Fc , where ␮ is the coefficient of rock friction.

between uN and Fc (Fig. 5.3), although theoretical and experimental results differ by as much as a factor of six (Iverson, 1990). Complications arise because the bed is a rigid boundary and a source of heat. This prohibits an exact determination of ⌿ using classical regelation theory (Morris, 1979). Coupled with the possible effect of solutes on regelation (Drake and Shreve, 1973) and the effect of the bed on viscous flow past clasts (Hallet, 1981), this leads to order-of-magnitude uncertainty in theoretical estimates of ⌿. Observations of clasts in basal ice (Boulton, 1974, 1979) and experiments (Iverson, 1990) suggest that cavities may sometimes form between clasts and the bed. Boulton (1974) argued that Fc should depend on the local effective pressure, equal to the difference between the pressure exerted by the overlying ice column and the local water pressure in such cavities. This formulation generally is not applicable, however, because the drag on fragments as ice

135

flows toward the bed is neglected. Furthermore, if cavities are common beneath clasts, it is likely that the water pressure within them is nearly as large as the local mean ice pressure, as such cavities are likely to be isolated from the through-flowing subglacial hydraulic system. An exception would be where ice regains contact with the bed down-ice from zones of ice–bed separation. The bed-parallel velocity of clasts will depend on the extent to which they are frictionally impeded by the bed relative to the adjacent sliding ice. If ␮ is the coefficient of rock friction, then the tangential force on clasts is ␮Fc , and the velocity of clasts relative to the ice is (␮Fc )/⌿, where ⌿ in this case is applied to bed-parallel ice flow past clasts (Boulton, 1974; Hallet, 1979). Thus, using Eq. (5.2) and ignoring By the clast velocity is given by: u c = ut –

␮Fc ⌿

= ut – ␮uN

(5.3)

where ut is the ice velocity parallel to the bed (Hallet, 1979). Thus, uc is independent of the clast size. Strictly, this relation provides a minimum estimate of the clast velocity because clasts may rotate in the ice, resulting in a shear force at the clast–bed contact that is smaller than ␮Fc (Iverson, 1991a). In the extreme case of a smooth, spherical clast, uc = ut because there would be no resistance to clast rotation because of the water film that lubricates the clast–ice interface. Debris in the basal ice of many Alpine glaciers is generally sparse (Lawson, 1979a; Wold and Østrem, 1979; Anderson et al., 1982). Thus, ice flow past clasts can be treated independently of flow past neighbouring clasts, as is done in theoretical models (Hallet, 1979; Shoemaker, 1988). Debris may sometimes, however, occupy more than 30 per cent of the volume of the basal ice layer (Echelmeyer and Wang, 1987; Ronnert and Mickelson, 1992). An important question is whether the physics embodied in Eq. (5.2) are fundamentally different in this case. Philip (1980) has analysed regelation of temperate ice through a dense three-dimensional array of particles. If the particles of a debris-rich ice layer are idealized as spheres that are in mutual contact, Philip’s theory can be adapted to obtain the following

136

PROCESSES OF GLACIAL EROSION

approximation for Fc : Fc = Bh +

4hd r 2p uN Kp

(5.4)

where hd is the thickness of the debris-rich ice layer, Bh is the buoyant weight in ice of a ‘column’ of particles within the debris layer, rp is the particle radius and Kp is the apparent conductivity of the particle array, proportional to the permeability of the array to ice. The permeability depends on the array porosity and on the thermal properties of the ice and particles. The expression should apply only to small particles (rp << 0.10 m), for which regelation is the dominant mechanism of ice motion. The presence of the bed and the relative motion of particles add uncertainty to this expression, but experiments (Iverson, 1993) indicate that its general form is correct. It is similar to that for isolated clasts (Eq. (5.2)). 5.2.3. Abrasion of a Rough Bed Abrasion of a rough bed depends on the magnitude and distribution of the basal ice velocity, as this controls both the flux of particles across the bed and particle–bed contact forces. Nye’s (1969) sliding theory provides an exact solution for the distribution of pressure and velocity in linearly viscous ice sliding over a low-roughness sinusoidal bed by regelation and viscous deformation. It has provided a useful foundation for abrasion models, although it neglects subglacial cavity formation and water pressure, both of which have a strong and well-documented effect on glacier sliding (Iken and Bindschadler, 1986; Hooke et al., 1987; Schweizer and Iken, 1992). The spatial variation of each of the parameters in the wear law (Eq. (5.1)) is evaluated within the context of Nye’s model. The discussion relies heavily on the abrasion models of Hallet (1979) and Shoemaker (1988) and the field observations by Rastas and Sepp¨al¨a (1981). It applies exclusively to sparse debris in ice. As discussed, contact forces depend primarily on uN . The principal component of uN results from regelation and deformation of ice around bumps on the bed during sliding. The combined effect of the resultant melting and bed-parallel extension of ice along stoss surfaces produces rates of ice convergence

with the bed that are a significant fraction of the sliding velocity. An additional minor component of uN results from the rate of basal melting owing to geothermal heat and heat from sliding friction, which should be on the order of 10–100 mm a–1 (R¨othlisberger, 1968). Although basal melting owing to heat from sliding friction should vary along the bed, most abrasion models (Hallet, 1979; Shoemaker, 1988) approximate it as uniform because it is typically much smaller than rates of ice convergence from sliding. Assuming u0 is uniform and neglecting any largescale straining of ice that might perturb the velocity field, uN at the surface of a sinusoidal bed of wavelength ␭ is given by uN = ub ap kw⍀w cos kw x + u0

(5.5)

where ub is the sliding velocity, ap is the bump amplitude, kw is the wave number equal to 2␲/␭ and ⍀w is a constant between 0 and 1 that depends on ␭ and the transition wavelength ␭* , the wavelength for which regelation and ice deformation are equally efficient (Hallet, 1979, eq. (17)). ␭* depends on the effective ice viscosity; thus, owing to the nonlinearity of the flow law of ice, it may range from 0.05 to 1.0 in for a reasonable range of sliding velocity and bed roughness (Hallet, 1996). Equation (5.5) suggests that contact forces should be largest halfway up stoss surfaces, where cos kw x = 1.0. It also indicates that clasts should lose contact with the bed a short distance down-ice from the crests of bumps where cos kw x becomes sufficiently negative so that ice flow diverges from the bed. This distance should scale with u0/ub . The velocity of clasts along a rough bed depends on the normal component of the ice velocity that presses clasts against the bed and the tangential component that pushes them forward (Eq. (5.3)) (Hallet, 1979). Observations of glacier beds suggest that the product ap kw seldom exceeds 0.5. For a bump of short wavelength in which regelation is the dominant mechanism of ice motion (␭<<␭* ), ⍀w = 1, and uN is maximized. Thus, on the steepest portions of some stoss surfaces uN may approach one-half of the sliding velocity. Usually, however, ap kw⍀w < 0.1. Therefore, a good approximation is that clasts move at the speed of the basal ice (Hallet, 1981). This was true in

PROCESSES OF GLACIAL EROSION

laboratory experiments with sparse debris (Iverson, 1990) in which uN/ut was quite large, 0.11, and clast velocity was 92 per cent of the sliding velocity. The concentration of clasts in contact with the bed, Cr , primarily controls the spatial variability of the flux of clasts. As a first approximation, Hallet (1979, 1981) treated Cr as an independent variable, and assumed it was spatially uniform on stoss surfaces. Shoemaker (1988), however, noted that as ice slides over a bed with periodic roughness elements, debris will contact only a limited portion of each stoss surface. This arises because as a clast loses contact with the bed just down-ice from the crest of a bump, its trajectory in the ice allows it to graze only the crests of stoss surfaces further down-ice (Fig. 5.4(a)). The length of the abraded zone on each stoss surface should scale with u0␭/ap ub . Thus, Shoemaker treated the areally averaged concentration of clasts in contact with the bed as a variable and found that it may be up to several orders of magnitude less than the concentration of clasts in basal ice. Shoemaker’s model, therefore, predicts spatially averaged rates of abrasion that are orders of magnitude smaller than those of Hallet’s (1979) model. Evaluating the realism of the treatments of Cr in the two models requires conjecture about how debris is supplied to the bed. The variable u0 is extremely important in Shoemaker’s model because melting owing to geothermal heat and heat from sliding friction is the only means by which clasts make contact with the bed. In the extreme case, when u0 = 0, there is no contact between clasts and the bed, and Å = 0. This counterintuitive result and the ubiquity of quarried surfaces on deglaciated bedrock indicate that debris entrainment from the bed must also be an important control on Cr and on the rate of abrasion. Hallet (1996) has recently modified his original model by coupling the ‘shadowing effect’ of Shoemaker (1988) with debris entrainment from the bed. He also includes the effect of debris comminution on Cr. He argues that in the common situation in which u0 << ub , debris entrained from the bed should be the principal source of abrasive debris (Fig. 5.4(b), (c)). In this case, most clasts abrade a significantly larger fraction of stoss surfaces than predicted by Shoemaker’s model. Hallet’s results indicate that his original model overestimated average abrasion rates

137

(a)

ub Ice

Regelation ice Bedrock

(b)

Abraded zone

ub Ice

Regelation ice Abraded zone

Bedrock

(c) ub Ice Regelation ice

Bedrock

Abraded zone

FIG. 5.4. (a) ‘Shadowing effect’ of Shoemaker (1988) for sliding dominated by regelation (␭ << ␭* ). Trajectories of clasts are inclined downglacier at a slope equal to u0/ub and allow abrasion of only a small portion of each stoss surface. No debris is entrained from bed. (b) Model of Hallet (1996) for sliding dominated by regelation. Clasts are quarried from lee surfaces enabling abrasion of all of stoss surfaces immediately downglacier. (c) Model of Hallet (1996) for sliding by viscous deformation with some regelation (␭ > ␭* ). Large clasts follow streamlines higher above bed and thus lose contact with bed further upglacier than smaller clasts (modified from Hallet,1996).

by an order-of-magnitude while Shoemaker’s model underestimated them by a similar amount. The spatial variability of Fc and Cr indicated by theoretical models provides a framework for evaluating local patterns of abrasion. Taking into consideration the shadowing effect and debris entrainment from the bed, the concentration of abrading clasts increases monotonically with height along stoss surfaces with a sharp increase in clast concentration where clasts derived from the crest of the preceding bump regain contact with the bed (Hallet, 1996). This, and the knowledge that contact forces should be largest midway up sinusoidal stoss surfaces, indicate that

Mean ice pressure (MPa)

138

PROCESSES OF GLACIAL EROSION

4.0 3.5 3.0 Ice

2.5 Rock

rw

Water

FIG. 5.5. Geometry of steady, water-filled cavity in lee of step and mean ice pressure against bed upglacier from cavity. Sliding velocity was 19 m a–1; water pressure in cavity (␳w ) was 2.1 MPa and ice overburden pressure was 2.7 MPa.

abrasion should be focused near the tops of bumps (Fig 5.5). Thus, the result of sustained abrasion should be to smooth glacier beds in the direction of sliding by reducing the height of bedforms. In addition, the tops of bumps that extend higher than surrounding bumps on the bed should be traversed by larger concentrations of clasts and abraded preferentially. This may explain the field observation that roche mouton´ees tend to be of roughly equal height in a particular region (Burke, 1969; Rastas and Sepp¨al¨a, 1981). Rastas and Sepp¨al¨a (1981) reported glacial polish on gently sloping lee surfaces near the crests of roches mouton´ees. Such polish is presumably the result of abrasion in the absence of particles larger than silt. Abrasion models suggest that clasts should abrade the bed a short distance down-ice from the crest of bumps. The distance will depend on the particle size. Small particles will follow streamlines near the bed that should diverge only slightly from gently sloping lee surfaces, particularly if regelation is minor. Larger clasts will follow streamlines higher in the ice that should diverge more sharply from the bed (Fig. 5.4(c)). Thus, large clasts will lose contact with the bed further up-ice than smaller particles. Glacial polish on lee surfaces may reflect this sorting. Hallet (personal communication) notes that this ‘shadowing effect’ may explain the distribution of crescentic fractures that tend to be clustered near the crests of roches mouton´ees. Growth of such fractures requires very large contact forces, but rates of ice convergence with the bed are expected to be small near the crest of bumps. This suggests that the fractures were produced by large rock fragments,

because both the fragment buoyant weight and viscous drag on fragments increase with size. Consistent with this expectation, the shadowing effect suggests that large rock fragments, which presumably travel long distances before being crushed, should tend to be preferentially concentrated near bump crests. Abrasion models do not explicitly yield information about transverse variations in abrasion. However, along the sides of bumps contact forces should be small because ice convergence with the bed as a result of sliding is minimal there. The sides of bumps, therefore, should be more lightly striated than stoss surfaces (Rastas and Sepp¨al¨a, 1981). Striae show that clasts tend to be deflected laterally around stoss surfaces. This should increase the concentration of clasts along the sides of bumps. For large bumps this deflection results from the lateral flow of ice. In the lee of such bumps, ice flow should converge and result in a lee side distribution of sediment equivalent to that up-ice from the bump. For small bumps, where regelation is the dominant mechanism of sliding, the deflection is a function of the angle between ub and the local slope of the bed and does not result from lateral ice flow. This is an important distinction when considering the ‘streaming’ process advocated by Boulton (1974, fig. 12). It implies that only in the case of small bumps should debris concentrations peak in zones down-ice from bump sides. Thus, although the convergence in the lee of large bumps may occur some distance down-ice if ice separates from the bed, it is difficult to appeal to ‘streaming’ to explain grooves that sometimes extend well down-ice from the sides of large hummocks and lie parallel to the overall direction of flow (Boulton, 1974; Goldthwait, 1979; Sharpe and Shaw, 1989). Similar grooves down-ice from centimetre-scale bumps, however, may be the result of streaming. Several arguments suggest that local separation between ice and the bed in the lee of bumps should increase abrasion rates by allowing clasts to abrade larger fractions of stoss surfaces than when ice contact with the bed is complete. If there is no local separation then, after traversing one stoss surface, clasts will follow streamlines that dip down-ice with a slope of u0/ub (relative to the mean bed slope) and thus will intersect only the crests of succeeding stoss

PROCESSES OF GLACIAL EROSION

surfaces (Fig. 5.4(a)) (Shoemaker, 1988). Clasts that are entrained where ice regains contact with the bed, as was observed by Vivian and Bocquet (1973), should abrade stoss surfaces further up-ice than when no cavity is present. Such clasts should be melted from the ceilings of cavities by viscous dissipation of heat associated with subglacial water flow. This process may be complemented by the ejection of particles owing to the decrease in ice pressure toward the cavity ceiling. As discussed hereafter, ice–bed separation should accelerate quarrying of lee surfaces, which also should increase the supply of abrasive debris on stoss surfaces. The most important contribution of abrasion models is the clear delineation of the glaciological parameters that control the process. Hallet’s conclusion that contact forces should be independent of the effective ice pressure and should depend mainly on the component of the ice velocity toward the bed is most significant. The rate of abrasion, therefore, appears to be directly proportional to the square of the local sliding velocity and independent of the ice thickness. This provides a tool for quantitative analyses of the evolution of glacially eroded topography (Harbor et al., 1988; Harbor, 1992).

5.2.4. Abrasion by a Subglacial Sediment Layer Some ice masses are underlain by a layer of deforming, water-saturated sediment. Field evidence suggests that such sediment may abrade the bed (MacClintock and Dreimanis, 1964). In this case, contact forces, in part, will depend on the effective pressure on the sediment. As the difference between the ice overburden pressure and the porewater pressure increases, contact forces between sediment particles and the bed should, on average, also increase (Chapter 4). Mobilization of particles at the base of a subglacial sediment layer is a requirement for abrasion of the underlying bedrock. Field observations indicate that the lower parts of some layers do not deform pervasively, although some movement may occur there along discrete surfaces (Menzies and Maltman, 1992; van der Meer, 1997). If water escapes into permeable rock at the base of the subglacial sediment

139

layer (Boulton and Dobbie, 1998), deformation at depth may be suppressed by the reduction in porewater pressure and resultant increase in strength with depth. This effect may be amplified by a feedback in which the sediment loses its dilatancy below a critical strain rate; the resultant compaction strengthens the sediment, reduces the strain rate further, and ultimately suppresses pervasive deformation at some depth (Alley et al., 1989a,b; Boulton and Dobbie, 1993, 1998; Hart, 1995). Perhaps the safest generalization that can be made is that as the subglacial deforming sediment thickness increases, the potential for abrasion of the underlying bedrock is reduced. Given that mobilization of the sediment base may not commonly occur, abrasion by subglacial sediment layers should generally be less effective than abrasion by clasts in ice. This conclusion is in agreement with general observations of abrasive wear that show that when two sliding surfaces are separated by a layer of loose debris, abrasion rates are usually an order of magnitude smaller than when no loose debris is present (Rabinowicz, 1965). 5.3. QUARRYING The lee sides of bumps on glacier beds are typically comprised of irregular, fractured surfaces that show little or no sign of abrasion. Such surfaces are presumably the result of quarrying, the process by which basal ice fractures the bed and dislodges rock fragments. Growth of cracks, which probably intersect joints and bedding surfaces, should detach rock fragments from the bed. Such fragments are dislodged when bed-parallel forces exerted on them by sliding ice exceed the frictional forces that hold them in place. 5.3.1. Rock Fracture Weathering processes, particularly frost cracking, may fracture subglacial bedrock to some extent. Modern studies of frost cracking show that the expansion of water upon freezing is not principally responsible for crack growth (Walder and Hallet, 1985, 1986). Rather, slow crack growth occurs because absorbtive forces at freezing fronts draw

140

PROCESSES OF GLACIAL EROSION

water toward ice bodies in cracks. The process requires rock on the surface or at depth to be steadily at subfreezing temperatures. This is potentially true anywhere subfreezing air can penetrate to the bed, such as at bergschrunds (Hooke, 1991) and near glacier margins (Anderson et al., 1982). Elsewhere at the beds of temperate glaciers, however, rock should typically not be at subfreezing temperatures, and frost cracking is not expected. However, under polythermal basal conditions there is the real possibility of freeze– thaw cycles occurring. Subaerial weathering processes undoubtedly facilitate quarrying during a glacier advance, but this cannot be invoked to explain deep glacial valleys with quarried surfaces. The role of joints in assisting quarrying has been cited extensively (Matthes, 1930; Crosby, 1945; Zumberge, 1955; Addison, 1981; Rastas and Sepp¨al¨a,1981). It is unlikely, however, that pre-existing joints will always be so dense and continuous that quarrying only requires that frictional forces be overcome. This is supported by highly fractured lee surfaces on some rocks that are otherwise unjointed or contain unrelated joint patterns. An example are groups of so-called ‘bullet boulders’ embedded in lodgement tills (Boulton, 1978). These boulders have fractured lee surfaces, but dissimilar structural characteristics as a result of their exaration and transport. Their asymmetry is strong evidence that crack growth is induced by glacier ice. Although it is unlikely that the bed is sufficiently fragmented that no crack growth is necessary for quarrying, it is equally unlikely that the bed contains no macroscopic cracks. This points out one difficulty in applying the Coulomb failure criterion to subglacial rock fracture, as was done in early studies (Boulton, 1974; Morland and Boulton, 1975; Morland and Morris, 1977). The criterion is empirical, based on uniaxial and triaxial compression tests on small rock specimens that generally do not contain macroscopic cracks. Such cracks are more likely to grow than microscopic voids and cracks. Therefore, the criterion overestimates the strength of larger rock bodies that have a higher probability of containing macroscopic cracks. In addition, the criterion is based on deviatoric stresses necessary to cause rapid, unstable crack growth, and, therefore, is not suitable for addressing slow,

stable crack growth at lower stresses, which is more likely under long-term loading by glacier ice and pressurized water. The approach taken here is to evaluate the subglacial geometry most likely to induce the optimum state of stress for growth of pre-existing cracks in the bed. A particularly favourable site for crack growth should be just up-ice from the point where ice separates from the bed near the crest of a bump. This, indeed, is precisely where field observations suggest quarrying is most prevalent. Ice–bed separation focuses some of the weight of the glacier on zones where the ice is in contact with the bed. Thus, normal stresses should be significantly larger on rock adjacent to cavities than elsewhere. The normal stress depends upon the extent of ice–bed separation and the effective pressure, defined in this context as the difference between the ice overburden pressure and internal water pressure within basal cavities. Numerical models based on the finite-element method permit the study of stresses in rock steps adjacent to such cavities (Iverson, 1991b). First, the geometry of a cavity of steady size in the lee of the step and the distribution of ice pressure against the bed upstream from the cavity are calculated. For an effective pressure of 0.6 MPa and a sliding velocity of 19 m a–1, pressure near the step edge is about 50 per cent greater than the ice overburden pressure (2.7 MPa) (Fig.5.5). Although larger stress concentrations are possible when ice–bed separation is more extensive, the applied stress can never exceed the strength of the ice, which is in the order of 10 MPa in uniaxial compression (Palmer et al., 1983). All principal stresses in the bedrock near such a lee surface are compressive, with the maximum principal stress oriented approximately vertically, reflecting the bed-normal ice pressure, and the least principal stress oriented approximately horizontally, reflecting the water pressure on the lee of the step or bump. In a vertically oriented crack near the step edge, tensile stresses should develop near the crack tips because compression parallel to the crack exceeds that normal to the crack faces (Griffith, 1924). Such tensile stresses are directly related to the crack length. In addition, the crack will normally be filled with water under pressure and this pressure also contributes to

PROCESSES OF GLACIAL EROSION

the tensile stress at the crack tips. Whether a crack actually grows depends on the rock strength, as well as stress corrosion. During this process, water reacting with silicate rock, for example, at crack tips results in hydrolysis of strong Si–O bonds to weaker hydrogen bonded hydroxyl groups. This may reduce the tensile stress necessary to cause slow crack growth by a factor of five (Atkinson, 1984). Assuming that the water pressure in cracks equals the cavity water pressure, numerical calculations suggest that for the aforementioned steady cavity, stress differences in the bed would be sufficiently large to cause slow growth of centimetre-scale cracks in most sedimentary and some crystalline rocks (Iverson, 1991b). Pre-existing cracks that are closest to being aligned parallel to the greatest principal stress are most likely to grow. As cracks extend, the orientations of principal stresses in the rock change, but final fracture surfaces should lie at a small angle to the greatest principal stress. This is indicated by so-called shear fractures produced in confined compression tests on rock, which initiate as tensile fractures and lie at less than 45° to the greatest principal stress (Jaeger and Cook, 1979, p. 92). Thus, fractures should tend to dip steeply into the bed beneath zones of ice–bed contact, because the stress that ice exerts normal to the bed, locally orients the greatest principal stress so that it is nearly normal to the bed surface. As water discharge in glaciers varies diurnally, subglacial water pressures fluctuate. The observed range is commonly greater than 0.5 MPa (Kamb et al., 1985; Iken and Bindschadler, 1986; Hooke et al., 1987, 1989; Sharp et al., 1998). Reductions in water pressure in subglacial cavities during such diurnal variations should transiently accelerate crack growth in the bed (Lliboutry, 1962; Iverson, 1991b). During water-pressure reductions, some of the weight of the glacier that was formerly supported by pressurized water is shifted to the bed up-ice from cavities. Together with the decrease in water pressure against the lee surface, this increases the difference between principal stresses in the rock, which should enhance crack growth (Fig 5.6). The process is somewhat analogous to a confined compression test on rock in which the axial load is increased while the lateral confining pressure is decreased.

141 Ice

Bedrock

FIG. 5.6. Ice–bed configuration that may be particularly favourable for subglacial rock fracture. Least principal stress in the bed is tensile just upglacier from point where ice regains contact with bed near lee surfaces.

The effect will be largest when ice–bed separation is extensive. The mean increase in ice pressure against the bed is simply the product of the water pressure reduction and the ratio between the areas of ice–bed separation and contact. Thus, for a glacier that has separated from the bed over 75 per cent of its area, which is not unrealistic for some glaciers (Anderson et al., 1982), a 0.5 MPa decrease in water pressure should produce a 1.5 MPa increase in the mean normal stress between the ice and the bed. Subglacial water-pressure fluctuations during the melt season may extend cracks in fracture-resistant lithologies by this mechanism. It has been suggested where a bed consisting of periodic steps and has a sufficiently low effective pressure that the ice may contact only a small portion of each ledge (Fig. 5.7). Normal stresses between the ice and rock will be very high in these areas. Immediately up-ice, the least principal stress in the bed is tensile and parallel to the bed. There the potential for bed-normal crack growth exceeds that for an isolated step or bump in which all principal stresses are compressive.

Ice

A

B Bedrock

C

FIG. 5.7. A, loosened rock fragment spalling off into a water-filled cavity; B, rock fragment frozen to the ice as a result of pressurerelease freezing (modified from R¨othlisberger and Iken, 1981; reproduced by courtesy of the International Glaciological Society from Annals of Glaciology, 2, 1981, p. 59, fig. 3); C, loosened rock fragment that must undergo frictional sliding to be dislodged.

142

PROCESSES OF GLACIAL EROSION

The above discussion suggests that ice–bed separation is important, if not essential, for subglacial rock fracture. Separation requires low effective pressures, high sliding velocities and high bed roughness. Because cracks are expected to propagate approximately normal to the crests of bumps, quarrying should steepen the down-ice sides of bumps. This will tend to favour continued ice–bed separation, continued crack growth and maintenance of steep lee surfaces. Growth of such cracks and their coalescence with joints and bedding planes may ultimately be responsible for maintaining bed roughness and limiting the rate of quarrying for a particular lithologic and structural setting. 5.3.2. Rock Dislodgement Ice–bed separation and transient effects associated with subglacial water pressure fluctuations are expected to be important elements in dislodging rock fragments that have become separated from the bed by cracks. R¨othlisberger and Iken (1981) noted that once ice–bed separation occurs, the resultant lack of confining pressure on lee surfaces should allow some loosened rock fragments to spall off into the cavity. This would be particularly true for thin slabs of rock on lee surfaces (Fig. 5.7, A). They also suggest a mechanism for increasing the bed-parallel force on such fragments. It can be argued that increases in subglacial water pressure should transiently reduce the pressure on lee surfaces that are in contact with ice, but not in communication with the subglacial hydraulic system. Because some water is extruded from the basal ice under high pressure, there is not sufficient interstitial water after such reductions in ice pressure to provide the latent heat necessary to warm the ice to the new pressure-melting temperature. Thus, the ice warms by freezing the meltwater film at the bed. The resultant tensile strength of the ice–rock bond, although it is probably small at temperatures near the melting point (<0.2 MPa) (Jellinek, 1959), may be sufficient in some cases to pull loosened rock fragments from the bed if frictional resistance is minimal (Fig 5.7, B). Other detached blocks should be pressed firmly against the adjacent rock because of the concentration in ice pressure expected near the crests of bumps (Fig

5.7, C). The viscous drag exerted on such blocks by sliding ice is likely to be small owing to smoothing and polishing of exposed surfaces by abrasion, and it is unlikely that shear stresses on such blocks ever exceed more than several tenths of a megapascal. Dislodgement of such blocks, therefore, should only occur when the water pressure in bounding fractures is a large fraction of the normal stress on blocks, resulting in low effective pressure and, hence, low friction on fracture surfaces. Effective pressure and friction on fractures may be minimized during periods of rising water pressure in adjacent cavities (Iverson, 1991b). This is, in part, because sufficiently continuous fractures should be in hydraulic communication with cavities. It is mainly a result, however, of the large reduction in the normal pressure against bump crests that occurs as water pressure in adjacent cavities increases. In a numerical calculation presented by R¨othlisberger and Iken (1981), an increase of 0.07 MPa in cavity water pressure (7 m of water column) produces a 1.26 MPa reduction in normal pressure on the crest of a sinusoidal bump. Thus, an increase in cavity water pressure of 5 per cent would result in a 95 per cent reduction in effective pressure across hypothetical fractures aligned parallel to the bump crest. It should be emphasized that the effect of subglacial water pressure fluctuations on normal stresses between ice and the bed are greatest when ice–bed separation is extensive. Thus, as was the case with rock fracture, rock fragments may be most susceptible to dislodgement when effective pressures at the bed are low and sliding velocities are large. 5.4. RATES OF EROSION The most reliable estimates of erosion rates come from measurements of sediment discharge in meltwater streams. Such measurements, compiled by Drewry (1986, p. 87), suggest that erosion rates range from 0.07 to 30 mm a–1. A major assumption associated with these measurements is that over the period of measurement, water has access to most of the glacier bed so that there is negligible storage of debris. Furthermore, it is assumed that most quarried debris is sufficiently comminuted so that it can be fluvially transported. Direct measurements of abrasion were

PROCESSES OF GLACIAL EROSION

made by Boulton (1974, 1979) beneath Brei∂/ amerkurj¨okull and range from 0.2 to 4 mm a–1 for platens of several different lithologies fixed to the bed. The experiments, however, were necessarily limited to small areas of the bed and, therefore, may not be representative of areally averaged abrasion rates, given the spatial variability of both contact forces and the concentration of abrading clasts. Very few data are available with which to assess the relative importance of abrasion and quarrying, but quarrying is probably dominant. Gilbert (1903, p. 206), for example, observed that quarried surfaces were more abundant on eroded bedrock than abraded surfaces and stressed the importance of quarrying as an erosional mechanism. In a semi-quantitative study using the pre-existing sheet structure of roches mouton´ees in Massachusetts, Jahns (1943) was able to calculate that several times more rock had been quarried from lee surfaces than had been abraded from stoss surfaces. Boulton (1979) argued that if the only source of abrasive tools is quarrying, as is true for ice sheets, and if tools are comminuted at a rate comparable with the rate at which the bed is abraded, then abrasion must be subordinate to quarrying because resultant tills contain large blocks that have survived commutation. In Hallet’s abrasion model, quarrying is the principal source of debris because clasts derived from ice above the bed can only contact a small portion of each stoss surface owing to the shadowing effect described by Shoemaker (1988). The model, therefore, provides a quantitative coupling between quarrying and abrasion, and suggests that even for very high sliding velocities, abrasion never exceeds about 40 per cent of the total erosion rate. 5.5. LARGE-SCALE PROCESSES In relatively few studies has current knowledge of small-scale erosional processes been used to interpret the morphology of large-scale landforms. To do so requires that large-scale patterns of glacier flow be measured or calculated and then linked with smallscale processes in order to estimate spatial variations in erosion rates. More commonly, glacier flow models, sometimes with erosional feedbacks, have been presented without specific reference to the

143

erosional processes involved (e.g., Nye and Martin, 1968; Hirano and Aniya, 1988; Mazo, 1989). Hooke (1991) invoked the potential effect of subglacial water pressure fluctuations on quarrying (R¨othlisberger and Iken, 1981; Iverson, 1991b) to suggest a positive feedback that may lead to the erosion of cirques and overdeepenings in the longitudinal profiles of glacier beds. He argues that water access to the bed is generally at bergschrunds in the case of cirques and through crevasses that develop above the high points on the bed between overdeepenings (reigels). Most water, therefore, encounters the bed immediately down-ice at cirque and overdeepening headwalls. As a result, water pressure fluctuations caused by variations in surface water input tend to be focused there. This accelerates quarrying of the headwall, which deepens the up-ice side of the overdeepening thereby increasing the potential for crevassing and further input of surface water to the headwall. The feedback is complemented by the tendency for water flow to be englacial, rather than subglacial, along the down-ice sides of overdeepenings with sufficiently steep adverse slopes. This is because as water flows up such slopes, some of it freezes to conduit walls to provide the heat necessary to keep the water at the pressuremelting temperature. The resultant constriction of the flow increases basal water pressure and forces the water into englacial conduits. This may allow a subglacial sediment layer to accumulate, protecting the bed from erosion, and potentially producing the characteristic asymmetry of overdeepened basins, which tend to be deeper on their up-ice ends. Measurements of water pressure fluctuations and the distribution of basal till beneath Storglaci¨aren, a valley glacier in northern Sweden, are consistent with the hypothesis. Harbor (1992) completed a major numerical study of the progressive erosion of U-shaped valleys. He first modelled glacier flow through a V-shaped transverse cross-section in order to obtain the lateral variation in sliding velocity. In agreement with the form of Hallet’s (1979, 1981) model of abrasion and Shoemaker’s (1986b) suggestion for a quarrying law, the rate of erosion was assumed to be proportional to the sliding velocity raised to an exponent ranging from 1 to 4. Valley evolution was iteratively simulated by calculating relative erosion rates across the profile,

144

PROCESSES OF GLACIAL EROSION 50 2

T=0

0

50

2 1 T=40

2 1 T=80

0

Valley Center 2

Dimensionless Basal Velocity

0

50

50

and Erosion Rate

1

(a)

1 T=120

0

50

T=300

FIG. 5.8. Numerical simulation of erosion of a glacial valley. Erosion rate was scaled to local sliding velocity squared. Figure shows cross-section after different numbers of iterations (T) with model. Concentric lines are velocity contours in units of 10 per cent of maximum velocity for section, with the most central contour being 90 per cent. Right column shows distribution of basal velocity and erosion rate, both scaled to an average cross-sectional value of 1.0. Basal-velocity and erosion-rate distributions for T = 300 (not shown) are essentially identical to that for T = 120 (from Harbor, 1992; reproduced by permission of the author).

modifying the profile accordingly and recalculating the distribution of sliding velocity. A sliding law was adopted in which the sliding velocity is directly proportional to the basal shear stress and inversely proportional to the effective pressure. Assuming a level piezometric surface across the section, the sliding

(b) PLATE 5.3. (a) The U-shaped valley of Lierdalen, Norway. Note kame terraces and trimline. (b) Profile of a U-shaped valley near Sogndal, Norway.

velocity increases away from the glacier margins. However, near the bottom of the V-shaped valley, the constriction impedes the flow so that there is a central minimum in the basal velocity distribution. This results in erosion rates that increase away from the margins, and peak on the valley sides. This effectively broadens the valley bottom and steepens the valley walls into the U-shaped form (Fig. 5.8, Plate 5.3).

PROCESSES OF GLACIAL EROSION

The shape of the active channel eventually becomes steady with continued erosion, although the valley continues to incise. Steady valley profiles were most similar to natural examples when the exponent in the erosion law was set equal to one or two. This is in general agreement with Hallet’s abrasion model and Shoemaker’s (1986b) suggestion that quarrying rates vary approximately linearly with the sliding velocity. 5.6. SUMMARY The rate of abrasion is controlled principally by the sliding velocity, which controls both contact forces and the flux of debris across the bed. Contact forces are largest midway up stoss surfaces of sinusoidal bumps, where the local bed slope dips most steeply up-ice. The flux of debris is largest near the crests of bumps, in part because once debris has traversed one stoss surface, its trajectory in the ice prohibits its further contact with all but the crests of similarly sized bumps further down-ice. This distribution of contact force and debris flux results in abrasion rates that are largest high on stoss surfaces. Thus, the longterm effect of abrasion should be to reduce the bed roughness. The tendency for debris to lose contact with much of the bed after abrading one stoss surface indicates that quarrying plays an important role in the process by providing tools to abrade the lower portions of bumps that would not otherwise be abraded. Ice–bed separation should increase abrasion rates by allowing debris greater access to stoss surfaces.

145

Quarrying must normally require that cracks in the bed grow and coalesce with each other and with preexisting joints and bedding planes. The potential for crack growth is largest in bedrock immediately up-ice from the point where ice separates from the bed. In these zones, normal stresses exerted by ice on the bed are maximized, while orthogonal compressive stresses are minimized, owing to the smaller pressure exerted by water on lee surfaces. As a result, the stress differences in the bed are maximized and conditions favour the growth of cracks that lie approximately normal to the crests of bumps. Cracks grow as a result of tensile stresses that develop locally at crack tips. The pressure of water in cracks should contribute to such tensile stresses, while the chemical action of water at crack tips should significantly reduce the stress necessary for slow crack growth. Rapid reductions in water pressure in zones of ice–bed separation intensify normal stresses that ice exerts on the bed by shifting some of the weight of the glacier to zones of ice–bed contact. This may result in periodic crack growth as subglacial water pressure fluctuates recurrently during the melt season. The effect is largest when separation between the ice and bed is extensive and is limited by the strength of the ice. Ice–bed separation and water-pressure fluctuations should also help dislodge rock fragments that have become loosened from the bed. As a generalization, quarrying rates should be largest beneath glaciers in which low effective pressures, large variations in water pressure and high rates of sliding combine to produce extensive ice–bed separation and fluctuations in stress on the bed.

This Page Intentionally Left Blank

6

PROCESSES OF GLACIAL TRANSPORTATION M. P. Kirkbride 6.1. INTRODUCTION The entrainment and transport of sediment by glaciers is an important link in the sediment cascade in high latitudes and at high altitudes today and, in the past, over vast areas of the mid-latitude continents. Glacial action is not only a potent erosional agent, but also provides the means by which eroded debris is removed from its source and transported to areas of deposition. Potential transport distances vary from a few hundred metres in cirque glaciers to many hundreds of kilometres in ice sheets. Without the evacuation of debris by glaciers, wear processes would not continue to attack fresh bedrock and sediments to create many of the characteristics of glaciated landscapes. Most wear and subsequent erosion is performed by rock particles (clasts) in transport at the glacier sole. Depositional landforms equally owe much to the processes of sediment delivery to ice mass margins and interfaces (Fig. 6.1). Where, for example, deposition is areally restricted around stable margins, continuous sediment supply allows large moraines to be constructed; or where glaciers overrun soft sediment or easily eroded material, large basal ‘layers’ of glacial sediment may be derived, transported and subsequently deposited. The amount, nature and distribution of sediment within, upon and beneath 147

moving ice fundamentally influences the assemblages of landforms deposited at different stages of the massbalance cycle. Two examples serve to illustrate this association. First, the large sediment loads and higher than average velocities of many warm-based valley glaciers account for their huge lateral moraines formed during the Neoglacial period. These moraines provide detailed sedimentary records of global Holocene climatic variation (Chapter 2). Secondly, the streaming of basal debris around large obstacles on the glacier bed may lead to the downstream persistence of longitudinal ‘ridges’ in the basal transport zone that results in the deposition of fluted moraine during ice retreat. Glacier transport is remarkable in the diverse manner in which the sediments themselves are modified. Much basal debris can be drastically shaped and comminuted over short distances, while part of the debris load may travel large distances with virtually no sedimentological change. In the former case, sustained mechanical crushing and fracture under the relatively low regional stresses (but large tractive forces and high local stresses) beneath ice masses produce sediment quite distinct from the products of other comminuting processes. Indeed, glacial abrasion may be responsible for most of the silt-size rock particles created over geological time. Conversely, glacial transport of large boulders over

148 (a)

PROCESSES OF GLACIAL TRANSPORTATION ICE SHEET

Equilibrium line

Equilibrium line

Iceberg

Land periphery

(b)

Maritime periphery

ICE SHELF

Equilibrium line Iceberg Sea

(c)

Land ice Melting

VALLEY GLACIER

Strand crack

Grounding line Equilibrium line

Accumulation

Basal Ice Velocity

Ablation

Particle Trajectory

FIG. 6.1. Models of (a) ice sheet, (b) ice shelf and (c) valley glacier, showing the distribution of accumulation and ablation and related flow characteristics. Basal sliding is assumed to occur in models (a) and (c) and is at a maximum in the vicinity of the equilibrium line (from Sugden and John, 1976, Glaciers and Landscape, Edward Arnold).

hundreds of kilometres without modification is unique in the natural sediment cascade. It is misleading to suggest that glaciers in all parts of the world are highly erosive. Glaciers frozen to their substrates are incapable of significant geomorphic work and may even have reduced erosion rates compared with neighbouring ice-free land. Indeed, doubt exists as to whether warm-based glaciers necessarily erode and transport more material than non-glacial processes in landscapes of similar relief and precipitation (Hicks et al., 1990). Nevertheless, processes of glacial transportation have played a major role in shaping landscapes, in moving vast supplies of sediment both on land and in the oceans and in the genesis and maturation of many

glacial land- and bed-forms. The understanding of the principles of drift exploration and in mining placer deposits (Menzies, 1996, chapters 15 and 16) in the establishment of the geotechnical character of glacial sediments and in the decisions concerning the suitability or otherwise of landfill sites for the disposal of toxic and other waste products hinges upon the impact and influence of glacial transport processes (De Mulder and Hageman, 1989). 6.2. ICE PROPERTIES AFFECTING SEDIMENT TRANSPORT Glaciers are commonly likened to ‘rivers of ice’. Such a comparison could not be further from the truth when the mechanics of flow and the nature of sediment transport are considered. An understanding of glacial transport depends on examining those physical properties of ice that have a direct bearing on how rock particles are entrained and transported. Three properties are important in this context: ice temperature, density and viscosity. 6.2.1. Temperature Within any ice mass, ice temperature varies over both time and space (Chapter 3). Ice at the base of many ice masses is at or close to the pressure melting point. Pressure melting occurs over large areas of glacier beds where overburden pressures are high (Fig. 6.1) and in localized zones in response to pressure variations caused by substrate irregularities. In terms of sediment transport, the distribution of basal melting and the presence of water affects the processes and effectiveness of erosion and entrainment, and has an overall influence on the discharge of sediment through ice masses. Where the ice mass bed is frozen, later transport of previously frozen debris may occur resulting in ‘patches’ of debris being moved long distances, in some cases, with limited comminution or evidence of deformation; or in other cases frozen debris, after transport, may melt out and produce a distinctive lithofacies. At glacier surfaces, the rate of melting is strongly influenced by the distribution and thickness of supraglacial debris (Paul and Eyles, 1990).

PROCESSES OF GLACIAL TRANSPORTATION

149

6.2.2. Density

6.3.1.1. Weertman regelation

Ice is less dense in the solid phase than in the liquid phase. Implications for glacial sediment transport are important where ice is in contact with water sufficiently deep for flotation, such as glaciers terminating in the sea or in deep proglacial lakes. The stress field that drives the flow of floating ice is very different from that of grounded ice, resulting in different flow paths and particle trajectories. Some ice shelves have upward particle trajectories, the reverse of those in grounded glaciers.

The Weertman regelation mechanism operates under warm-based glaciers at the scale of bedrock obstacles of <1 m in length (Weertman, 1957, 1964). Ice coming into contact with the upstream (stoss) side of such an obstacle experiences increased pressure that, if the ice is at the pressure-melting point, causes melting. Meltwater migrates to the downstream (lee) side of the obstacle where reduced pressure allows refreezing and the incorporation of usually fine-grained debris into the regelation layer accreted to the glacier sole. The repeated formation and destruction of a regelation ice layer as it encounters successive obstacles limits the layer thickness to only a few centimetres. Weertman regelation has been confirmed by observation, and may be an important contributor to the sliding of warm-based glaciers.

6.2.3. Viscosity Newton’s law of fluid friction defines viscosity as the ratio of the shear stress acting on a fluid to the rate of shear that results. Thus, highly viscous materials deform only slowly under a given applied shear stress, whereas low-viscosity materials deform very rapidly under the same applied stress. Ice, with a viscosity of 1012 –1014 poises, may be likened to a viscous fluid that deforms only on attaining a critical level of stress (the yield stress). The pressure exerted by even the largest boulders resting on glacier surfaces rarely exceeds the yield stress of ice, so that glaciers transport material of about three times the density of ice without the debris sinking. 6.3. DEBRIS SOURCES 6.3.1. Entrainment at the Glacier Sole The mechanisms by which glaciers pick up debris from their beds, and the physical controls governing these mechanisms, remain one of the most problematic, yet intriguing, areas of glaciological research. Basal debris entrainment is notoriously difficult to observe, but borehole camera observations, direct observations, proxy evidence such as meltwater chemistry and landform studies, give some substance to theories of debris entrainment. Much of this research tackles the problem of how ice is added to the glacier sole, and it is through study of basal ice formation that debris entrainment can be understood (Hubbard, 1991). Most theories of basal entrainment involve some mechanism of re-freezing (regelation) of subglacial water.

6.3.1.2. Robin ‘heat pump’ effect Under some circumstances, meltwater produced by localized pressure melting upstream of an obstacle is lost from the system, resulting in a net loss of latent heat in the lee of the obstacle and the formation of a cold patch of several square metres extent (Robin, 1976; Goodman et al., 1979). Re-freezing in the vicinity of the cold patch incorporates subglacial water not produced by pressure melting of existing ice, thus adding ‘new’ ice to the glacier in which thin layers of fine-grained sediment may be incorporated.

6.3.1.3. Large-scale zones of melting and re-freezing The overburden pressure exerted on the glacier bed by glacier ice reduces where ice becomes thinner, typically around the margins or differentially across bed obstacles. Lowering of the pressure melting point commonly results in re-freezing of meltwater derived from basal melting upstream and in marginal areas from surface ablation. Ice masses experiencing such a large-scale contrast between their interiors and margins are termed polythermal glaciers. Such glaciers

150

PROCESSES OF GLACIAL TRANSPORTATION

are capable of effective basal ice formation producing extensive sequences of heavily debris-charged ice of several metres thickness. It is thought that polythermal glaciers carry larger debris loads than fully warm-based glaciers, and may also be efficient erosional agents on both local and continental scales. Variations in subglacial water pressure may also cause widespread basal adfreezing. Increased pressure in the subglacial water film reduces relative pressures on adjacent bed protrusions. A resulting rise in pressure melt point at these locations causes conduction of heat from the bed, lowering the temperature at the ice–bed interface sufficiently to cause freezing. Variations in subglacial water pressures enhance basal entrainment. Where water pressures increase in lee-side cavities, the effective overburden pressure on the bedrock substrate is reduced and freezing encouraged. These effects in association with higher ice velocities linked with high basal water pressures increase the tractive forces that physically ‘pluck’ boulders away from cavity walls. Zones of effective bedrock erosion may, therefore, correspond to those parts of glacier hydrological systems where water pressure variations are most marked, such as below crevassed areas (Hooke, 1991) and/or where bedrock is impermeable (Iverson, 1991b). In addition, ice may also be incorporated onto the glacier sole by several other mechanisms. Overriding and inclusion of existing debris-rich ice by glacier surges has been recorded (Sharp, 1985), and at a smaller scale blowing snow, cold air intrusions into glaciers along crevasses and tunnels, and incorporation of frozen ground may all result in the addition of ice and debris to the glacier. All these mechanisms, and possibly others yet to be identified, produce ice with textures and structures quite distinct from ice (basal ice zone) produced by the firnification of snow in glacier accumulation zones. Floating ice shelves, which normally release sediment by basal melting at their grounding-lines, may, under fairly rare conditions, accrete new basal ice by the adfreezing of seawater. Very small amounts of suspended marine sediment and solutes may be entrained by this process, though occasionally larger quantities of clastic and biogenic sediment are entrained at grounding-lines.

6.3.2. Supraglacial Entrainment The extensive debris-free surfaces of ice sheet ablation zones contrast with the dirty, moraine-strewn appearance of many cirque and valley glaciers. The contrast owes much to the spatial variability of supraglacial sediment supply, and the influence of the distribution of extraglacial sediment sources and to rates of sediment supply. The addition of debris to valley glacier surfaces is dominated by rockfall and avalanching from mountain faces, especially in temperate alpine regions. Cliffs in alpine regions typically retreat at an average rate of between 0.5 and 3 mm a–1 (French, 1976) providing large amounts of coarse debris. Nevertheless, many other processes acting around glacier margins can supply debris to supraglacial transport zones. Close to the margin, mass movements onto glacier surfaces are common in alpine regions. Other more distant sources of debris include transport by rivers from ice-free valleys tributary to glacierized main valleys (Evenson and Clinch, 1987). Aeoliantransported fine-grained material is delivered to glaciers and ice sheets by snowfall, by direct deposition or by redistribution of unconsolidated sediment (Menzies, 1996, chapter 6). Finally, ice sheets contain debris from extraterrestrial sources. This debris is concentrated over millennia from catchment areas of several million square kilometres into zones of flow convergence and ablation, often in the lee of nunataks. This debris includes meteorites of several varieties (chondrites, irons), smaller spherules and cosmic dust trapped by the Earth’s gravitational field (Nishiizumi et al., 1989). Supraglacial debris from mountainsides tends to emanate from discrete points of entrainment. Commonly, rock gullies supply rockfall, avalanche and debris-flow material repeatedly to the same point on the glacier surface. Intervening parts of glaciers receive relatively little debris. By far the majority of supraglacial debris enters glacier accumulation zones, where high local relief, concave glacier slopes, an absence of marginal moraines and centripetal flow lines combine to produce a strong coupling of the mountain slope and glacier transport systems (Fig. 6.2). In most cases, small quantities of supraglacial debris are supplied fairly constantly to the ice mass

PROCESSES OF GLACIAL TRANSPORTATION

151

ACCUMULATION ZONE Avalanche and rockfall

Fall zone

ABLATION ZONE

Equil

ibriu

m Lin

e

5

6

4 3 2 1

Medial debris septum

1 2 3 4 5 6

4 3

Bedrock Lodgement till Basal transport zone Bed-parallel debris septum Superglacial lateral moraine Superglacial medial moraine

FIG. 6.2. The geometry and terminology of the glacier transport system (afer Boulton and Eyles, 1979; reprinted from: Moraines and Varves (Ch. Schl¨uchter, ed.), 1979; courtesy of A.A. Balkema, Rotterdam).

surface. However, rock avalanches may instantaneously increase the supraglacial debris load. These sudden and vast inputs of debris lead to changes in the pattern of ablation. Significant modification in net mass balance can result, over a time lag of some years, in ice margin movement. Large rock avalanches are fairly common phenomena in alpine regions, several occurring every decade in tectonically active and high-precipitation alpine ranges such as in Alaska, western Canada, the Andes and New Zealand. The 1964 Alaska earthquake, for example, caused a rock avalanche of 106 m3 of rock to blanket 8.25 km2 of the Sherman Glacier (McSaveney, 1978). The largest recent example of a rock avalanche in New Zealand took place in December 1991 (Kirkbride and Sugden, 1992) and appears to have had no seismic trigger (Plate 6.1). The Tasman Glacier, onto which it fell, receives the majority of its huge supraglacial debris load from relatively large rockfalls

in its accumulation zones. The 1991 avalanche almost doubled the supraglacial load in only a few minutes. 6.3.3. A Sediment Supply Model The amount and variety of sediment entering highlevel transport depends on the nature and extent of extraglacial terrain, particularly the effectiveness of weathering and erosion. In contrast, subglacial entrainment depends on thermal regime and substrate erodibility. The effectiveness of high-level and basal entrainment processes, therefore, experience quite different controls. Figure 6.3 summarizes how two of the main determinants of debris supply (thermal regime and degree of topographic inundation) in combination influence the amount of debris available for entrainment and transport. Two possible interpretations are shown that illustrate that overall sediment supply is less where ice cover is great and

152

PROCESSES OF GLACIAL TRANSPORTATION

PLATE 6.1. A large rock avalanche of 14 December 1991 which fell on Tasman Glacier, New Zealand. Such events are important debris sources for valley glaciers in tectonically active mountain ranges (photo courtesy of Lloyd Homer, Institute of Geological and Nuclear Sciences Ltd).

that most of this reduction is accounted for by a relative lack of debris in high-level transport. The supply of sediment to the basal transport zone of warm-based glaciers is assumed to be a function of sliding velocity and declines initially with increasingly cold thermal regime. An increase in basal sediment supply under polythermal ice reflects potent basal adfreezing in this regime: the magnitude of this increase is the fundamental difference between Figures 6.3(a) and (b). In mountainous areas, the extent to which terrain is inundated by ice is in inverse proportion to the area of ice-free cliff faces from which rockfalls supply debris to the high-level transport zone. Glaciers with high supraglacial and englacial sediment loads are invariably surrounded by extensive steep cliffs and usually have ‘dirty’ surfaces. The degree of inundation probably plays a more important role for warm-based

than for cold-based glaciers because extraglacial climates around the former favour abundant sediment transport by water, avalanching, debris flow, rapid weathering and slope erosion. In mid-latitude glacial environments, greater burial of the landscape by ice probably causes a proportionately greater reduction in sediment availability for the high-level transport zone than in high latitudes. Aeolian and subglacial processes supply debris to subpolar glaciers, but entirely cold-based glaciers depend on a very small supply of aeolian and extraterrestrial sediment. Ice sheets, in comparison, have zero supraglacial debris in accumulation areas (Chapter 10). In ablation areas surface debris accumulates where outlet lobes cut through marginal mountains or isolated nunataks. In ice sheet marginal areas where no such topographic relief exists, the only supraglacial debris is likely to be found very close to the margins where basally entrained debris is elevated along upward flowlines. Therefore, along the southern margins of the Laurentide and Fenno-Scandian Ice Sheets, supraglacial debris would have been restricted to a narrow band close to the ice sheet margins. In contrast, along the seaboards of Labrador-Ungava and western Norway, both ice sheets would have had extensive supraglacial debris cover derived from exposed ice-free mountains. A similar contrast exists today between the western and eastern edges of the Antarctic Ice Sheet. 6.4. TRANSPORT PATHWAYS THROUGH ICE MASSES 6.4.1. Controls on Particle Trajectories The distribution of sediment in glaciers reflects the trajectories taken by particles away from the locations of debris sources. Ice sheets, valley glaciers and ice shelves have contrasting transport pathways reflecting their different flow patterns (Fig. 6.1). The geometry of transport paths within any glacier depends on three main factors: 1 Locations of points of entrainment: debris eroded from the ice mass bed generally remains within basal ice where pressure melting prevents significant upward translocation. Debris supplied to

PROCESSES OF GLACIAL TRANSPORTATION

153

(b) after Boulton

RATIO OF DEBRIS:ICE DICHARGE

(a) after Andrews, Anderson

HIGH

LOW 0

DE

G

RE

E IN OF UN T DA OP TI OG O R N AP

SED

-BA

HI

C

AL ERM

TH

10

0%

WE

T-

ED BAS

RM

WA

Y POL

LD

CO

High-level transport zone

IME

Basal transport zone

EG LR

A

RM

THE

DRY

100% 100% 25%

100% 50%

Topographic inundation

75%

FIG. 6.3. Conjectural relationships between the relative discharge of debris to ice in glaciers, thermal regime and the degree of burial of topography by ice. Note the relative increase in basal transport zone debris as fewer extraglacial debris sources are exposed, and the extremely low debris load of cold glaciers. (a) Illustrates the case where transport by temperate glaciers exceeds that by polythermal glaciers (e.g., Andrews, 1971; Anderson, 1978). (b) Illustrates the reverse case (Boulton, 1970). The figure shows relative discharges; total debris discharge of large glaciers and ice sheets will be greater than that of smaller glaciers even though their thermal regime and extraglacial sources are often less favourable of entrainment.

the accumulation zone is buried by snow and enters englacial transport. Debris supplied to the ablation zone surface remains in supraglacial transport. Englacial and supraglacial transport together form the high-level transport zone. 2 Glacier flow lines: a cirque or valley glacier of simple structure with no tributaries transports debris from source along flow lines that converge and flow downward in the accumulation zone but which diverge and flow upwards in the ablation zone. Debris in transport above the equilibrium line therefore tends to move towards the centre of a glacier cross-section and, below the equilibrium line, tends to move outwards to the glacier margins. Particles in dome-shaped ice sheets with purely radial flow follow trajectories (Fig. 6.1(a)). The simple pattern may be disrupted by multiple centres

of ice accumulation, by subglacial topography and by concentration of ice discharge into fast-flowing ice streams. Ice shelves (Fig. 6.1(b)) usually have downward trajectories, but this pattern is reduced or reversed where basal adfreezing predominates. 3 Glacier flow structure: many valley glaciers and topographically constrained ice fields are composites of multiple tributaries, each with its own set of debris sources and transport paths. Ice flow is laminar, so the transport zones of each ice stream maintain their individuality after confluence with other ice streams. The general model of transport path geometry (Fig. 6.2) distinguishes between a basal transport zone and a high-level transport zone. The former includes the zone of traction, where interparticle and particle–bed collisions cause abrasion and fracture. In contrast, sediment

PROCESSES OF GLACIAL TRANSPORTATION

-1

10 m

SUPERGLACIAL ZONE

Diffused

Banded

ENGLACIAL ZONE

Diffused

2

The bulk of the debris load of most ice masses is concentrated in the basal few metres, reflecting (1) the entrainment of debris at the glacier sole, and (2) concentration of debris brought from higher englacial levels to the ice–bedrock interface by basal melting. The greatest shear strain within glaciers occurs in the basal few metres, causing ice deformation and recrystallization. High strain rates yield textures and structures that reveal a history of both ice formation, debris entrainment and subsequent deformation. The basal transport zone is defined by this combination of sedimentological and crystallographic characteristics. A primary constraint on the extent and nature of the basal transport zone is the thermal regime of the glacier. Cold-based glaciers, frozen to their substrates, entrain a very small volume of debris thus englacial debris is uncommon. What basal debris they do transport is usually entrained by overriding of marginal aprons and upward shearing to form ‘inner moraines’ (Chinn, 1989; Evans, 1989). Warm-based glaciers, on the other hand, rarely develop basal debris sequences thicker than a few tens of centimetres because basal melting is sufficiently effective to keep basal debris close to the glacier sole. Furthermore, comminution of grains during traction

10 m

6.4.2. Basal Transport Zone

Banded Diffused

BASAL ZONE

Dispersed

1

The distribution of debris in glaciers is highly inhomogeneous. Debris point sources and glacier flow structure tend to concentrate debris into planar geometric shapes termed debris septa, or into discrete bodies of variable geometry. The nature of ice deformation creates lensate and ellipsoid englacial debris bodies, though compressive flow has the potential to create more complex forms through folding. Only where debris is supplied almost continuously from a point source is a medial moraine truly continuous in form. Where inputs of debris are pulsed, such as infrequent rockfalls from the same source, medial moraines have a ‘beaded’ form.

reduces weaker rocks to silt, much of which is flushed from the basal zone by meltwater. Polythermal glaciers, characterized by mixed thermal regimes, are able to accumulate thick basal sequences in their marginal zones, where re-freezing adds much ice to the glacier sole. Perhaps the most complete picture so far of the character of the basal transport zone is provided by the Matanuska Glacier, Alaska (Lawson, 1979a). Here, two main facies, the englacial and basal facies, were identified on the bases of ice crystallography and debris content. Englacial facies comprise ice formed by firnification, forming by far the greater part of the glacier. Basal facies comprise ice formed at the glacier sole (Fig. 6.4). This facies is further subdivided into ‘dispersed’ and ‘stratified’ subfacies. Weertman regelation during basal sliding forms the dispersed facies, which are subsequently elevated within the marginal zone by net basal adfreezing accreting the debris-rich stratified facies to the glacier sole. Thus, debris-rich bands are interstratified with relatively clean ice over several metres of vertical

0

in high-level transport is little modified during transport. Distinction between the two exists on the basis of sedimentological process–form relationships (Mills, 1977) as well as ice crystallographic characteristics.

10 -10

154

Stratified Substrate

FIG. 6.4. Schematic vertical section through Matanuska Glacier, Alaska, showing the distribution of ice facies. High debris concentrations occur in the banded facies (from supraglacial sources) and in the stratified facies (from subglacial sources) (after Lawson, 1979a; reprinted by permission of the author).

PROCESSES OF GLACIAL TRANSPORTATION

thickness close to glacier margins. Such sequences are not, however, representative of the whole basal transport zone in polythermal glaciers, because thin basal transport zones associated with pressure melting and basal sliding often occur extensively beneath thicker ice upstream. Stratified facies occupy the lower 3–15 m of Matanuska Glacier and comprise irregular alternating layers and bands of almost pure ice and highly debrischarged ice. A sharp contact separates stratified from dispersed facies. Intense folding and shearing characterizes the stratified facies, and abrasion, fracture and crushing of debris causes increased rounding and comminution of particles during transport. Debris may be derived from high-level transport by downward ice flow, or from glacier-bed erosion. Debris concentrations vary widely; ranging from 5 to 55 per cent as reported from Matanuska Glacier. The thickness and debris concentration of basal ice changes as ice moves across its bed by both vertical and lateral migration of debris. Vertical migration involves the slow upward dispersion of particles in the dispersed facies under vertical strain to form socalled ‘amber-ice’ facies. Horizontal migration

155

involves the lateral diversion of sediment-rich basal ice around bedrock obstacles. Enhanced plastic deformation of ice owing to high stresses in the vicinity of bed hummocks causes ice to accelerate around the flanks. Much less ice accelerates over the top of the obstacle, resulting in ‘streaming’ of sediment-charged basal ice through gaps while intervening spaces are occupied by debris-poor ice from above. Ice moving through a field of bed hummocks thereby develops a laterally dispersed basal sediment layer. The presence of a soft bed of deforming mobile sediment moving as a debris traction layer between the overlying ice and immobile bed permits the transport of large volumes of sediment. Within this layer considerable comminution and deformation occurs with associated kinetic sieving, grain-size differentiation with transport and sediment block rafting. 6.4.3. High-level Transport Zone High-level transport includes supraglacial and englacial debris that is not subject to basal traction. Most debris at high level comes from supraglacial sources, but debris is elevated from basal ice at glacier

ice cliff

4 rognon

3

1

2

5 6

5

avalanche scar

tributary stream

FIG. 6.5. Idealized plan view of the variety of shapes of moraines in high-level transport in valley glaciers. Dashed line is the equilibrium line. 1, Ice–stream interaction medial; 2, ablation-dominant medial; 3, beaded (avalanche-fed) moraine; 4, ice-cliff fed moraine (basal debris elevated to surface); 5, rock-avalanche debris; 6, fluvial inwash debris.

156

PROCESSES OF GLACIAL TRANSPORTATION

confluences, avalanching ice cliffs, to the lee of large bed obstacles and near the termini of cold-based and polythermal glaciers. Medial moraines are the most commonly recognized forms within the high-level transport zone, yet they are part of a continuum of forms comprising isolated rockfall deposits, beaded and linear medial moraines and extensive supraglacial debris mantles (Fig. 6.5). Bodies of debris manifest shapes reflecting the total strain experienced by the ice mass since entrainment. Thus, extending flow produces characteristically linear forms, elongated down-glacier but compressed laterally, whereas compressing flow tends to cause crescentic and broadening forms. The shapes of debris bodies also depend on their volume, the periodicity of debris supply, the duration of englacial and/or supraglacial transport and on the depth and distribution of crevasses along the transport path. A medial moraine forms in one of two ways: (1) at the confluence of two tributary glaciers, where lateral

moraines-in-transit merge; and (2) where rockfall and avalanche debris from above cirque headwalls supply medial debris septa above the firn line (Fig. 6.6). Eyles and Rogerson (1977) classified medial moraines according to the position of emergence of the moraine relative to the firn line. ‘Ablation dominant’ (AD) moraines are formed by debris septa from above the firn line emerging downstream in the ablation zone. ‘Ice-stream interaction’ (ISI) moraines form where two tributaries are confluent below the firn line so that two lateral moraines-in-transit join. ‘Avalanche’ moraines form by infrequent rockfalls leaving discrete deposits on the glacier surface. A second classification (Small et al.; 1979) is based on the position of rockfalls relative to equilibrium line position, as well as to whether or not debris enters englacial transport via crevasses or remains in supraglacial transport throughout (Fig. 6.7) and relates debris supply and differential ablation to moraine morphology. ‘Waxing’ and ‘waning’ stages

bedrock source

ice

ACCUMULATION ZONE

flow

point of emergence A D E B R I S

S E P T U M

ABLATION ZONE

B ra

l

e em

l at

mo

v

en

t

MEDIAL MORAINE E D C

Long profile

E D

A

B

C

FIG. 6.6. Anatomy of a medial moraine fed by rockfalls above the equilibrium line. Vertical lines represent increments of ice movement (extension in accumulation zone, compression in ablation zone). B, ice core; AE, moraine crest; AD, ice–debris interface; AC, level of nearby glacier not affected by differential ablation. Lateral movement of debris away from the moraine crest increases the area affected by differential ablation.

PROCESSES OF GLACIAL TRANSPORTATION TYPE A

ABLATION ZONE

FIRN LINE

DEBRIS SUPPLY TO GLACIER SURFACE LONG-PROFILE OF IC E-C

BARE-I

little debris

CE SU RFACE

OR

OF GLA C

continuous cover = waxing stage of moraine

ED

MO

IER

RA

INE

lateral sliding of debris = waning stage of moraine (not always present)

ABLATION ZONE

TYPE B FIRN LINE

ENGLACIAL DEBRIS

little debris TYPE C

ablation = debris debris release = continuous enters cover and waxing crevasses moraine

ABLATION ZONE

ACCUMULATION ZONE

IC E

E

N

AC IA

D

L

TYPE D

FA L EB

L

RIS

6.4.4. Transport in Subglacial and Englacial Conduits

debris buried by annual englacial transport firn layers

as above

as above

ABLATION ZONE

ACCUMULATION ZONE

IC E

EN

of medial moraines are identified by the exhaustion of the englacial debris septum on moraine morphology. In general, medial moraines broaden down-glacier owing to the reduction in lateral compression away from a confluence and to the aggregated effects of differential ablation and lateral movement of debris down the flanks of the moraine. Moraines of superficially similar appearance may contain either basally derived debris or rockfall and avalanche debris indicating fundamentally different origins. Debris from the traction zone may be elevated to the surface at confluences by the convergence of flow lines (Fig. 6.8). While basal debris is commonly elevated to high-level transport close to the termini of polythermal glaciers. Often this involves upwarping of warm-based (sliding) ice upstream of an obstruction of thinner, and more rigid, cold-based marginal ice. The mechanisms of upwarping involve deceleration and compression leading to the formation of a shear zone between the contrasting parts of the glacier, although the extent to which debris is passively elevated, or actively affects the distribution of shear strain, remains uncertain.

FIRN LINE

GL

little debris

as above

157

FIRN LINE

G

LA CI AL

DE

FA L BR

L

IS

debris injested via crevasses = vertical little debris bands buried englacial transport by firn

as above

as above

FIG. 6.7. Explanation of varieties of medial moraine found on Glacier de Tsidjiore Neuve, Switzerland (after Small et al., 1979; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 22(86), 1979, p. 50, fig. 4).

Water may be supplied either by melting of glacier ice or from ice marginal rivers (Chapter 4). Valley glaciers play an important passive role by diverting and channelling meltwater. This is the case where wholly or partly ice-free tributary valleys feed streams into a glacier occupying a main valley. The importance of the fluvial sediment input will increase with the area of marginal ice-free terrain and sediment and water availability. From provenance studies, Evenson and Clinch (1987) suggest that ~90 per cent of sediment delivered to some Alaskan glacier termini has been routed through englacial and subglacial conduits. Such a high proportion emphasizes the rapid throughput of intraglacial fluvial sediments compared with the glacier sediment load. A direct estimate of subglacial fluvial sediment discharge has been made possible by hydro tunnels beneath Bondhusbreen, Norway (Hooke et al., 1985; Bezinge, 1987). Bondhusbreen, an ice-cap outlet glacier fed by few

158

PROCESSES OF GLACIAL TRANSPORTATION

Superglacial debris supply

Subglacial debris supply +

+

+

Englacial debris supply + +

Thickness of supraglacial till layer

_

Lateral ice compression

+

+

+ + _

Differential ablation Da =

Da above one

Thermal erosion of bare ice by meltwater streams

+

Extending flow

clean ice ablation covered ice ablation

+

Da below one

+ Moraine height

+

Compressive flow

_ +

+ Moraine lateral slopes +

_

Moraine width

+

Lateral sliding of supraglacial till layer

FIG. 6.8. Influences on the down-glacier development of medial moraines (after Small and Clark, 1976; reproduced courtesy of the International Glaciological Society from Journal of Glaciology, 17(75), 1976, p. 163, fig. 1).

supraglacial debris sources and flowing across resistant crystalline rocks, has low glacial sediment load. However, subglacial conduits develop each spring and transport as much as 95 per cent of the total sediment flux of the glacier. The transfer of sediment between englacial and intraglacial fluvial transport has also been observed in Mueller Glacier, New Zealand (Plate 6.2). Large rockfall deposits that have travelled along englacial pathways have been exposed in ice caves within the ablation zone. Englacial rivers flowing within these caves repeatedly switch position, maintaining their gradient towards the terminal outlet while ice motion has an upward vector: thus, lenses and ribbons of fluvial sediment have been successively abandoned by the river and elevated to the glacier surface. Fluvial transport of sediment through glaciers has important implications for interpreting

ice-marginal sedimentary facies. There has been an over-emphasis in the past on glacial reworking of outwash to produce diamictons comprised of rounded clasts and a neglect of the effect of water-reworking of sediment before it leaves the glacier. 6.4.5. Transport in Floating Ice Shelves Ice shelves are formed by the coalescence of floating glacier tongues in the sea or large lakes, or by the accumulation of snow and ice on perennial sea ice. Ice shelves are attached to land-based ice at groundinglines. At the grounding-line there are major changes in debris transport paths. When the base of an ice shelf melts in contact with warmer water, sediment is released at the grounding-line from basal transport zones into the water (Fig. 6.9). Conversely, in some

PROCESSES OF GLACIAL TRANSPORTATION

159

PLATE 6.2. Vertical aerial photograph of Mueller Glacier, New Zealand, showing emergence of debris at the glacier surface in zones of intense compression (A) and sinuous band of debris (B) reworked by an englacial river, now visible in collapse dolines at C. The river has never flowed supraglacially (photo courtesy of Lloyd Homer, Institute of Geological and Nuclear Sciences Ltd).

glacier s

urface

interfluve

tributary B tributary A

flow

s

line

FIG. 6.9. Transfer of debris from the basal to the high-level transport zones along flowlines at the confluence of two glacier tributaries (after Boulton, 1978; and Gomez and Small, 1986).

ice shelves the majority of the debris load is entrained by basal adfreezing at the grounding-line. The thermal conditions that govern whether sediment is entrained or released are complex and depend on water and ice temperatures, and on salinity. Particle trajectories follow flow lines that reflect the spatial distribution of accumulation and ablation (Fig. 6.10). These are often markedly different from terrestrial glaciers. Ice shelves fringing Antarctica receive most accumulation close to their seaward margins. The accumulation zone extends over the entire upper surface of most ice shelves, and ablation occurs by a combination of basal melting and/or iceberg calving from the seaward margin. The pattern

160

PROCESSES OF GLACIAL TRANSPORTATION

HIGH

PROCESSES OF TRANSPORTATION

Velocity

Terminal moraine deposition

Over-riding +/incorporation

Terminal moraine deposition

Terminal moraine deposition

Recessional moraine deposition

Mass Balance

ADVANCE

Stable

A

Re

tre

ing

c an

dv

at

ing

Stable

RETREAT

Terminus Position

LOW

Sediment Release Per Unit Area of Terminus

HIGH

LOW

Ablation Rate

Stable

Response Time

EQUILIBRIUM

POSITIVE

EQUILIBRIUM

NEGATIVE

EQUILIBRIUM

Time

FIG. 6.10. Schematic relationships between mass balance and sediment release from a glacier terminus. As positive mass balance commences, ablation initially decreases as climate cools, then increases as the glacier advances to lower altitude. Similarly, as negative mass balance commences, ablation initially increases as climate warms, but decreases as glacier terminus retreats to higher altitude. Sediment release closely mirrors ablation, but moraine construction and preservation also depend on changes in terminus position.

PROCESSES OF GLACIAL TRANSPORTATION

of ice flow is determined by mass redistribution from the accumulation zone to the ablation zone (as in land-based glaciers); therefore, particle trajectories are downward and seaward in most ice shelves. Debris discharges in ice shelves are generally very low compared with land-based glaciers because of the lack of sediment sources in their accumulation zones. However, ice shelves nourished by land-based glaciers may receive glacially transported debris from adjacent land masses, although basal debris commonly melts out close to the grounding-line. Morainic mounds and stripes occur supraglacially in zones of surface ablation, exceptionally in large quantities such as on McMurdo Ice Shelf, Antarctica, where basal adfreezing also entrains marine sediment (Swithinbank, 1988). Ice shelves formed from perennial sea ice have very low debris concentrations, restricted to aerosols and aeolian particles. 6.4.6. Quantitative Measurements of Debris Discharges Direct measurement of basal sediment volumes

161

transported by glaciers is hindered by difficulties of access. Sediment yields and denudation rates for glacierized catchments, therefore, are usually estimated from outwash sediment volumes rather than from glacier loads. Measurements of sediment transport are based on samples exposed around glacier margins, especially at termini where sediment is delivered to the proglacial environment. In a few rare cases, natural cavities and artificial tunnels have been used to observe basal debris moving over glacier beds. Sediment quantity can be expressed in several different ways (Table 6.1) so that direct comparison between different studies can only be made where similar sampling strategies have been performed: unfortunately this is almost never the case. Quantification at a hierarchy of spatial and temporal scales gives a range of information from point samples of sediment concentration to continuity equations of sediment flux through an entire glacier. There is, therefore, no simple statistic to answer the question ‘how much sediment does a glacier transport?’. Nevertheless, broad generalizations can be drawn from available data (Table 6.2). By far the highest

TABLE 6.1. Various measures of the quantity of sediment in transport in glaciers Measure

Scale

Typical units

Concentration

Point sample to transport zone

% by volume; % by weight; g cm–3; kg m–3

Load

Transport zone to whole glacier

tonnes or m3

Discharge

Glacier cross-section or debris melt-out at terminus

Gross measures; kg year–1; t year–1; m year–1; specific measures: kg m–1 year–1; t m–1 year–1

Sediment yield

Whole catchment

tonnes or m3

Denudation rate

Whole catchment

mm year–1; kg m–2 year–1

TABLE 6.2. (A) Measured englacial debris concentrations Glacier

Debris concentration (% by weight)

Source

Ayutor, Tien Shan

0.02–0.33

Glazyrin and Sokolov (1975)

Watts, Baffin Island

0.0001–0.0097 (x = 0.0018)

Dowdeswell (1986)

Djankuat, Caucasus

0.12

Bozhinskiy et al. (1986)

Tasman, New Zealand

0.028

Kirkbride (1989)

Ivory, New Zealand

0.009

Hicks et al. (1990)

162

PROCESSES OF GLACIAL TRANSPORTATION

TABLE 6.2. (B) Measured basal debris concentrations Glacier

Debris concentration (% by weight)

Debris concentration (% by volume)

Basal zone thickness (m)

Source

E. Antarctic Ice Sheet (margin)

–

0–12

15

Yevteyev (1959)

E. Antarctic Ice Sheet (Byrd Core)

–

7

5

Drewry (1986)

Antarctic Ice sheet

–

7

4.8

Gow et al. (1979)

Greenland Ice Sheet

–

0.1

15.7

Herron and Langway (1979)

Nordenskjoldbreen, ¨ Spitsbergen

–

40

0.4

Boulton (1970)

Barnes Ice Cap, Baffin Island

–

8±2

8

Barnett and Holdsworth (1974)

¨ Iceland Brei∂amerkurjokull,

–

50

0.15–0.3

Boulton et al. (1974)

Brei∂/ amerkurjokull, ¨ Iceland

–

8–10

0.05–0.2

Boulton (1979)

Matanuska, Alaska: dispersed facies

–

0.04–8.4

0.2–8

Lawson (1979a)

Matanuska, Alaska: stratified facies

–

0.02–74

3–15

Lawson (1979a)

Glacier d’Argentiere, ´ France

–

43

0.02–0.04

Boulton et al. (1979)

/

Myrdaisjokull, ¨ Iceland

–

15–31

2–5

Humlum (1981)

Bondhusbreen, Norway

1

0.39

5

Hagen et al. (1983)

Watts, Baffin Island

32–80

14–57

0.8–2.9

Dowdeswell (1986)

TABLE 6.2. (C) Measured debris discharges Glacier

Discharge

Comments

Source

Barnes Ice Cap, Baffin Island

11.5±5 m3 m–1 year–1

Specific discharge measured at ice cliff terminus sliding at 17 m year–1

Barnett and Holdsworth (1974)

Bondhusbreen, Norway

1500 kg m–1 year–1

Specific discharge measured in subglacial tunnel

Hagen et al. (1983)

Ivory Glacier, New Zealand

300 kg m–1 year–1

Specific discharge measured at ice cliff terminus with ice velocity of 30 m year–1

Hicks et al. (1990)

Hatunraju, Peru

5080 m3 year–1

Gross discharge based on ice core data, measured flow rates

Lliboutry (1986)

Ayutor-3

9.6 × 106 kg year–1

Gross discharges measured over periods of 5 to 82 years. Correspond to erosion rates of 0.7–2.9 mm year–1

Chernova (1981)

Fedchenko

8300 × 106kg year–1

Zarovshanskiy

2200 × 106kg year–1

RGO

1100 × 106 kg year–1

IMAT

15.4 × 106 kg year–1

Karabatkak (all central Asia)

13.6 × 106 kg year–1

PROCESSES OF GLACIAL TRANSPORTATION

debris concentrations occur in the basal transport zone. Englacial facies are generally very low in debris content (values as low as 0.0001 per cent by weight have been recorded), even in glaciers where supraglacial moraines indicate apparently high debris loads. Sediment concentrations vary through several orders of magnitude. Ice sheets with negligible supraglacial debris supply may have even lower concentrations. Much higher concentrations are localized in englacial debris septa, especially those nourished by frequent rockfalls onto slow-moving ice in accumulation zones. The majority of the sediment loads of warmbased and polythermal glaciers are carried as high debris concentrations at the ice sole. Concentrations vary greatly even within the basal ice zone; ~80 per cent by volume of the stratified ice facies in basal zones may consist of suspended debris, though substrate erodibility and the erosivity of ice combine to produce great variability. Estimates of supraglacial and basal discharges for six Icelandic glaciers varied from 200 to 26 000 m3 a–1 (Eyles, 1979). Alpine outlet glaciers have similar debris loads to piedmont glaciers, but the latter have more basal debris (~93 per cent of total debris load) at their termini. Sediment discharge is a measure of sediment volume passing through a unit of glacier cross-section in a given time. For valley glaciers in zero net balance (equilibrium), sediment discharges are probably at a maximum at equilibrium lines, where ice velocities are usually greatest. Most sediment discharges have been measured at glacier termini, where debris released from ice by ablation is deposited as diamicton or transported into outwash streams. The volume of sediment released at a terminus is related to the ablation rate, ice velocity and sediment concentration. Where a glacier terminus becomes stagnant (during negative mass balance or the quiescent phases of surging glaciers) down-valley sediment discharge may approach zero some distance upstream, though debris is released at the terminus as long as ice ablates. Chernova (1981) summarizes the results of extensive surveys of several Asian glaciers, where sediment delivery to termini ranges over three orders of magnitude (Table 6.2). Sediment discharges fluctuate over time in accord with mass balance changes. Debris tends to accumulate in glacier ablation zones under negative mass

163

balance owing to increased compression and ablation. Such debris ‘reservoirs’ can be rapidly evacuated and deposited at glacier termini when mass balance becomes positive. Thus, the amount of sediment in high-level transport varies inversely with glacier mass balance. In fact, short-term sediment discharge is only likely to equate to the catchment denudation rate if (a) the glacier is in equilibrium, eliminating temporal variability; and (b) discharge is measured at the equilibrium line cross-section, eliminating spatial variability. This ideal is never achieved and allowances must be made for disequilibrium mass balance and sediment reworking. Information on glacier sediment discharges is rather sketchy, so that comparisons between glaciers in different climatic and tectonic settings are difficult to make. Cirque glaciers, having logistical advantages of size and form, provide useful information on processes and rates of sediment transport in arid and maritime environments. For example, Reheis (1975), using sedimentological observations, estimated that 70 per cent of debris supplied to the terminus of Arapaho Glacier (Colorado, USA) had at some time been in traction at the glacier sole. From studies of proglacial lake sedimentation, Hicks et al. (1990) found that of a total sediment delivery to the terminus of Ivory Glacier (New Zealand) of 840 t a–1, 40 per cent was transported in the basal transport zone, 3 per cent in high-level transport and 57 per cent washed through, over or underneath the glacier by avalanches, water and mass movements. This latter figure reflects the potency of extraglacial processes in small, steep catchments in maritime environments. Possibly the greatest sediment discharges in modern ice masses are achieved by glaciers flowing in a so-called ‘fast mode’, either as surging glaciers or fast ice streams draining ice sheets. Surging glaciers entrain large amounts of basal debris, particularly where they override unconsolidated sediment. Entrainment is favoured by conditions of increased cavitation and meltwater production associated with the very high velocities attained during surges (Clapperton, 1975). Ongoing research into fast ice streams reveals very high sediment discharges from ice sheets associated with deformable subglacial sediment. Quantitative studies of sediment transport have posed fundamental questions about the geomorphic

164

PROCESSES OF GLACIAL TRANSPORTATION

role of glacier entrainment under different thermal regimes. It has been suggested that polythermal glaciers in general have higher sediment loads than warm-based glaciers because of the erosion and entrainment made possible by their mixed thermal regime. Conversely, it has been suggested that warmbased glaciers possibly produce between two and ten times more debris than cold-based and polythermal glaciers, based on a comparison of field studies in Colorado and Baffin Island. Aside from the questions of comparability between studies, this debate also raises questions about the distinction between erosion and entrainment capability. Warm-based glaciers appear to lose much of their potential debris load to flushing from the glacier sole by meltwater and derived rainwater, giving the appearance of being relatively free of debris even though they are highly erosive. 6.5. MODIFICATION OF SEDIMENT DURING TRANSPORT Sedimentological modification during glacial transport gives clues to the nature and intensity of the stresses acting on the debris load during glacier movement (Menzies, 1996, chapters 2, 13 and 15). 6.5.1. Particle-size Distributions Debris carried in the high-level transport zone is not usually modified significantly during transport since englacial debris generally occurs at low concentrations, reducing the likelihood of interparticle collisions. Local shear stresses at high levels within glaciers are usually very low, so that even if two particles come into contact with each other, their resistances to wear far exceed the forces potentially causing abrasion or crushing. Debris brought close to the glacier sole by ice flow and basal melting is concentrated in suspension and may even impinge upon the substrate. Here, traction occurs. Ice deformation around bed obstacles and repeated episodes of pressure melting and regelation cause numerous collisions between particles carried in basal ice. Particle shape is rapidly modified, and similar grain size distributions are created regardless of how the debris arrived at the glacier sole. Traction-zone debris

shows a high degree of granulometric variation. This reflects both the localized intensity of abrasion and crushing, and the differing resistances of mineral grains compared with clastic fragments, and of weak lithologies compared with strong. Abrasion of particles produces distributions showing a characteristic negative skew reflecting the production of fine sand to clay sizes from larger clasts. Most grain-size distributions are polymodal, though bimodal grainsize distributions have been reported (Fig. 6.11) (Brodzikowski and van Loon, 1991). Boulton (1978) sampled debris directly at the soles of three glaciers and found that crushing produced rock fragments generally coarser than 1␸ (0.5 mm), while abrasion produced material finer than 1␸. When the percentages of samples above and below the 1␸ cut off are compared, the debris modified in the traction zone is clearly discriminated. Dreimanis and Vagners (1971) concluded that the coarse mode of a bimodal distribution dominates close to the sediment source, but is progressively replaced by a fine mode with distance along a transport path as rock fragments are progressively comminuted. Many workers recognize the concept of ‘terminal grade’ of glacially transported debris, referring to a critical particle size at which tractive forces are equalled by resistance to further wear. The concept is important because if comminution causes basal debris to attain terminal grade, particle-size distributions may eventually become independent of transport distance. However, Haldorsen (1981) and Drewry (1986) doubt the reality of terminal grades. Haldorsen indicates that for a particular mineral variable mineralogy and fracture geometry (among other factors) negate such a concept as a single terminal grade. Drewry cites experiments that suggest that further comminution would occur given sufficient time and transport distance. He concludes that bimodal particle-size distributions represent only a transient stage of the comminution process. Whether glaciers achieve such ‘tertiary’ comminution is uncertain. Both theoretical and experimental grinding studies suggest that coarse clastic fragments are crushed much more easily than fine ones. Below about 1 mm grain size, the energy required to comminute debris increases exponentially with decreasing particle size and grinding of quartz grains ceases at ~1 ␮m (Drewry, 1986). Terminal

PROCESSES OF GLACIAL TRANSPORTATION

(b)

Sphericity

IMMATURE SHAPES

1.0

MATURE SHAPES

0.0

PRIMARY SHAPES granular disintegration

40 points

Minimum sphericity

32 points

39 points 0.0

0.2

1.0

Range of unabraded clast shapes in glacial transport

0.4 0.6 Roundness

0.8

1.0

SORE BUCHANANISEN

0.8

after fracture Sphericity

roc k

fall

a abr ice by

Increasing “sphericity”

0.4 0.2

sion

Decreasing “flatness”

0.6

ter

wa n by

sio

abra

WEAK FORMS (elongated/platy)

BREI AMERKURJÖKULL

0.8 without abrasion

STRONG FORMS (equidimensional)

Maximum roundness

(a)

165

0.6 0.4 36 points 0.2

snow avalanching

ANGULAR CLASTS

30 points 34 points

0.0

Increasing roundness

0.0

0.2

0.4 0.6 Roundness

0.8

1.0

ROUNDED CLASTS

1

2

3

FIG. 6.11. (a) Evolution of clast shape during glacial transport, based on measurements of 88 samples of 50 clasts each on Tasman and Mueller Glacier, New Zealand. There is no time axis, because some clasts pass rapidly through the system and others never attain mature shapes depending on their exposure to high stresses in the basal transport zone or in englacial or subglacial rivers. (b) Krumbein formroundness plots of clasts from high-level and basal transport zones of glaciers in Iceland (top) and Svalbard (bottom), showing the effect of basal traction and abrasion on particle shape. 1, Englacial debris; 2, basal debris; 3, lodgement till (after Boulton, 1978; reproduced by permission of the International Association of Sedimentologists).

grades are determined as much by the lithology of source rocks, particularly the density of cracks and microcracks in mineral grains, as by the distance over which debris has been transported. Diamictons derived from crystalline rocks tend to be dominated by sand and coarse silt owing to mechanically resistant quartz and (to a lesser extent) feldspar in these sizes, while other diamictons derived from finegrained rocks tend to contain a higher proportion of silt and clay. Terminal grades appear to be attained after fairly short transport distances, for example, <1.6 km in New Hampshire (Drake, 1972) and only 0.5 km in Iceland (Humlum, 1985). The terms ‘crushing’ and ‘grinding’ do not correspond directly to the

nomenclature of recognized processes of wear, and redefinition for the purpose of clarification has been necessary (Sharp and Gomez, 1986). Crushing is directly related to the formation of cracks in grain surfaces as grains slide over a hard substrate and to fatigue fracture involving cracking beneath grain surfaces. Grinding, on the other hand, corresponds to abrasive wear, whereby a harder protrusion ploughs a groove in a softer surface by plastic deformation. 6.5.2. Particle Shapes A distinction comparable with that between particle sizes in the traction zone and in high-level transport

166

PROCESSES OF GLACIAL TRANSPORTATION

zones is revealed by the shapes of coarse clasts. Shape comprises elements of roundness (particle faces and edges) and form (the relative dimensions of the three principal orthogonal axes). Debris entrained by rockfall and avalanching onto glacier surfaces, and by plucking from a bedrock substrate, is typically angular. Diamictons deposited at great distances from points of entrainment often have a large proportion of angular clasts. Frequent clast collisions occur in the zone of basal traction and during englacial water transport which, though very different in terms of process, produce rather similar rounded clast forms. The trajectories of clasts in the traction zone are complex, being influenced by regional ice flow vectors, substrate topographic irregularities, localized regelation, deflection by neighbouring particles and even by the form of individual clasts. Modifications to form and roundness are correspondingly complex because clasts interact with the glacier bed and with each other. The velocity of clasts held in suspension in sliding ice close to the glacier sole varies according to clast size, form and frictional drag. Thus, some clasts change their positions relative to surrounding ice by retardation and rotation. This is especially so when clasts at the glacier sole come into contact with the substrate. Tractional movement across the substrate is most effective on clasts in the size range –3 to –7␸, which move at 80–100 per cent of the basal ice velocity. Clasts of initially equant form are more prone to being rotated, or ‘rolled along’ at the glacier sole with the result that edges are chipped and abraded and both roundness and, to a lesser extent, sphericity are increased. Debris from both supraglacial and subglacial sources assume similar shape properties in traction, indicating that the nature of clast contacts is the primary determinant of shape, rather than debris sources. Blade- or plate-shaped clasts (‘flatirons’) are common forms produced in traction. These shapes probably develop from initially platy joint-bounded fragments in much the same way that spheroidal clasts are produced where primary joint spacings are equant (Holmes, 1941, 1960). Smooth surfaces are typical of all clasts from the traction zone where sustained stresses and an abundance of wet silt and fine sand act as efficient abrasives. In deformable bed traction zones, complex interaction between advecting particles, retarded immobile

substrate and superjacent mobile ice combine to produce a highly variable wear environment where many sub-wear processes act over a range of distances and areas. Comminuted sediment and individual particle shapes have characteristics similar to those found in other glacial sub-environments. A second sub-environment where boulder shapes are efficiently altered is the englacial fluvial environment. Supraglacial streams and englacial conduits commonly entrain debris from the ice walls of channels. This debris rapidly becomes rounded and platy clasts readily fracture into stronger equant forms. Fluvial sediment, both on and within glaciers, therefore tend to contain clasts superficially resembling traction-zone debris, but with a conspicuous absence of striations and of silt and clay. Roundness and sphericity are slightly more highly developed than in traction-zone debris. 6.5.3. A Cascading System of Particle Shape Evolution The variety of boulder shapes within a glacier reflects the persistence of some shapes over long transport distances and time periods while, at the same time, rapid and dramatic changes to boulder shapes occur almost instantaneously in some parts of the transport system. Boulders travelling between their points of entrainment and point of deposition follow an almost infinite variety of transport routeways, along which shape and size may be drastically altered or remain virtually unchanged. Some order may be brought to this complexity by considering the ‘maturity’ of boulder shapes, rather than transport time or distance, as a frame of reference. The term maturity implies the degree to which angular, weak shapes are modified to rounded, strong shapes. Weak shapes are elongate and platy forms which are easily fractured on planes parallel to their short and intermediate axes. Clasts tend to be stronger when they are more equant in form. Breakage of weak clasts tends to create smaller but mechanically stronger clasts. Measurements of form and roundness of 4400 clasts on and around Tasman Glacier (a warm-based valley glacier in New Zealand), when plotted in a form/roundness field, reveal several groups of boulder shapes characteristic of different stages in the transport system (Fig. 6.11).

PROCESSES OF GLACIAL TRANSPORTATION

Any clast may pass from one group to another either directly or via other groups depending on the combination of transport routeways followed. However, there is an overall trend from immature shapes towards the lower left corner of the graph to mature shapes towards the upper right corner. Of course, much debris is deposited before it reaches maturity, and the time period between entrainment and deposition varies greatly both within and between glaciers. It is noticeable that most subaerial processes do not greatly modify clast roundness but effectively alter weak forms into strong forms by fracture. Thus, immature debris appears more susceptible to changes in form than in roundness. Having achieved relatively equidimensional forms, subsequent shape changes in mature debris involve increasing roundness and water-transport appears more effective than subglacial traction in increasing clast roundness. The model illustrates the sedimentological distinction between the high-level and basal traction zones as well as demonstrating the complexity of shape evolution throughout the whole glacier transport system. Variations in both intensity and spatial distribution of shape-modifying processes operating during glacial transport are responsible for this complexity. Glacial particle shape evolution demonstrates the enormous range of wear processes that impinge on any single particle over both temporal and spatial ranges and the common tendency for particle shapes from many sources and wear process combinations to exhibit equifinal shape geometries. Great caution must therefore be used in discriminating between different glacial environments solely upon particle shape and roundness. 6.6. GLACIOLOGICAL EFFECTS OF DEBRIS IN TRANSPORT

167

deforms along glide planes parallel to the basal crystallographic plane. The presence of undeformable sediment within an aggregate of ice crystals has been shown experimentally to reduce the creep rate of ice even at low sediment concentrations (Nickling and Bennett, 1984). In heavily sediment-charged ice, friction caused by interparticle interactions stiffens ice. In marked contrast, borehole closure measurements from ice sheets show that sediment-charged ice deforms more easily than ‘clean ice’. Experiments to test the influence of englacial sediment on ice rheology have given quite different results from borehole observations. Nickling and Bennett (1984) found that ice attained a peak strength at a sediment concentration of 25 per cent by volume, with ice becoming more deformable at higher concentrations. No simple inverse relationship was found between debris content and creep rate. Englacial sediment indirectly affects ice rheology through its influence on ice crystal growth. Basal ice sequences usually have small crystals where sediment concentration is high. Understanding the rheology of debris-laden ice is important in deciphering the long-term flow of large ice sheets using deep ice cores, and in establishing core chronologies for environmental reconstruction. Boreholes from Greenland and Antarctica have revealed contrasting rheologies between Holocene and deeper Pleistocene ice (Van der Veen, 1999). Late Pleistocene ice has a higher concentration of fine wind-borne sediment, deforming three times more readily under stress than ‘clean’ Holocene ice (Fig. 6.12). The reason for these differences is probably low Pleistocene sea levels exposing continental shelves to subaerial wind erosion, supplying aeolian dust to the ice sheets. 6.6.2. The Effects of Supraglacial Debris on Ablation

6.6.1. Rheological Effects of Englacial Debris In recent years it has become increasingly apparent that debris is not transported entirely passively by glacier ice. Rather, particles in transport significantly alter the deformational behaviour (rheology) of ice both directly (through the inclusion of undeformable rock and mineral grains within ice) and indirectly (through affecting ice crystal growth). Glacier ice

The presence of supraglacial debris affects the transfer of heat between the atmosphere and ice surface, consequently affecting the rate of ice melt. Ablation may be increased or decreased depending on the texture, thickness and composition of supraglacial debris. Small-scale morphological features commonly develop where differential ablation is caused by debris, including dirt cones and boulder tables where

PROCESSES OF GLACIAL TRANSPORTATION

Distance from base (metres)

(a)

(b)

(c)

1500

1500

1500

1000

1000

1000

500

500

500

Holocene ice

168

Pleistocene ice 0

0

0.05

du/dz

0.10

0

-1

(yr )

0

5.0

B

(Pa-3yr-1)

*

10-17

0

0

2

E

4

6

FIG. 6.12. Rheological changes with depth in a large ice sheet, measured from tilt experiments in the Dye 3 borehole, Greenland. (a) Variation of ice velocity with depth; (b) variation of flow-law parameter B with depth; (c) variation of flow enhancement factor E with depth. Values of all parameters increase abruptly at the Holocene–Pleistocene transition, indicating more deformable ice at depth owing mostly to higher microparticle concentrations (from Dahl-Jensen, 1985; reproduced by courtesy of the International Glaciological Society from Journal of Glaciology, 31(108), 1985, p. 94, fig. 3).

ablation is hindered, cryoconites and penitentes where it is enhanced. Debris affects melting by altering the amount and rate of heat transfer by conduction and advection. Three main controls on conductive heat transfer are: (1) the thickness of the debris layer; (2) the thermal conductivity of the debris, which varies with mineralogy, porosity and moisture content; and (3) the thermal gradient through the debris layer, which varies both daily and seasonally according to air temperature cycles. Heat is also transferred by percolating rain- and melt-water, and by air circulating within the debris, all parameters that need to be measured or estimated in order to calculate sub-debris ablation. As a result of such difficulties few studies of this type have been made, yet results indicate that most melting results from downward conductive heat transfer (especially in arid environments) and that debris is a very effective insulator of ice (Bozhinskiiy et al., 1986). In cold environments with short ablation seasons, thin debris layers a few centimetres thick may prevent ablation from occurring, allowing long-term preserva-

tion of ice-cored landforms. In warmer environments, much thicker layers are required to provide an equivalent insulating effect (Table 6.3). It is now possible to estimate sub-debris ablation indirectly from meteorological data (Nakawo and Takahashi, 1983). 6.6.3. Supraglacial Debris and Glacier Dynamics Many glaciers in mountainous regions such as in Alaska, the Himalayas, Karakoram, Andes and the Southern Alps of New Zealand receive so much debris into high-level transport that their ablation zones are characteristically debris-mantled. The dynamics of glaciers and their sensitivity to climatic change can be greatly altered by an extensive covering of supraglacial debris. A fundamental effect is that debris-mantled glaciers tend to be longer than their uncovered counterparts. Such ‘over-lengthening’ is a consequence of reduced ablation rates being compensated by increased ablation areas to maintain equilibrium mass balance. As

PROCESSES OF GLACIAL TRANSPORTATION

169

TABLE 6.3. Estimates of the insulating effect of supraglacial debris Location

Percentage reduction in melting of bare ice

Source

5 15 30 40

40 70 80 85

Østrem (1959)

Kaskawalsh Glacier, Yukon

5 10 20 30

42 60 82 94

Loomis (1970)

Ayutor-2, Tien Shan

4 8 12

8 15 50

Glazyrin (1975)

100

80

McSaveney (1975)

8 12

0 10

Bozhinskiy et al. (1986)

20 55

Small (1987)

60 78 89 96

Kirkbride (1989)

Isfallsglaciaren, Sweden

Sherman Glacier, Alaska Djankuat Glacier, Caucasus Glacier de Tsidjiore Neuve, Switzerland Tasman Glacier, New Zealand

Debris thickness (cm)

2–3 4 25 50 100 200

a result, the accumulation area ratios (AAR) of such glaciers are typically only 0.3 to 0.5, compared with an AAR of 0.5 to 0.8 for unmantled glaciers (Kirkbride, 1989). Debris-mantled glaciers also respond differently to climate change, both during advances and retreats. Where ablation is reduced beneath debris, glacier advances tend to be more rapid and sustained. Furthermore, large rockfalls onto glaciers can initiate advances by reducing ablation independently of climate change. Thus, the Brenva Glacier (Italy) advanced for 20 years after a 1920 rockfall while neighbouring glaciers retreated (Porter and Orombelli, 1981). During negative mass balance, the amount of supraglacial debris tends to increase as flow velocities reduce and ablation zones become increasingly protected from melting. The terminus of the heavily mantled (but non-surging) Tasman

Østrem (1975)

Glacier (New Zealand) has not retreated from its late-nineteenth-century maximum position even though nearby glaciers have retreated by as much as several kilometres over the same period. Similarly, surging glaciers become heavily mantled during their quiescent phases as their abundant debris loads melt out at the surface to produce characteristic landform assemblages. Such protected, stagnant tongues have lead to the misleading impression that all debris-covered glaciers are necessarily stagnant: in fact, many simply have very high rates of debris supply. The effect of supraglacial debris loads on multiple glacier fluctuations deserves further research. Revealingly, Kick (1989) suggests that the difference in timing of the last Holocene advances between Europe and the Himalaya-Karakoram may reflect the complicating effect of debris mantles (Chapter 2).

This Page Intentionally Left Blank

7

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION C. A. Whiteman tion processes, (b) mechanisms of glaciogenic deposition, and (c) examples of problems in determining depositional processes based upon sedimentary characteristics.

7.1. INTRODUCTION The study of glacial sediments and processes of glacial deposition has generated an extensive literature (Goldthwait and Matsch, 1988). One of the results of this keen interest is an increasingly complex genetic classification of glacial and glaciofluvial sediments. There have been intense discussions, not only concerning the precise nature of glaciogenic depositional processes but even about what constitutes a glaciogenic sediment. One very narrow interpretation of this debate classifies only so-called primary sedimentation by lodgement and melt-out (Lawson, 1979a). In this classification secondary processes such as flowage and deposition through water (glaciolacustrine and glaciomarine) are excluded on the grounds that particle disaggregation during flowage and under the influence of gravity removes direct evidence of glaciogenic stress (Boulton, 1980). In contrast, a more catholic view might include glacioaeolian and glaciofluvial sediments in addition to those mentioned above. In this chapter a broader interpretation will operate: processes traditionally referred to as lodgement, meltout and flowage will be addressed, together with a brief discussion of glacial deformation and subglacial fluvial processes. The chapter is divided into sections dealing with (a) factors that influence glacial deposi-

7.2. FACTORS CONTRIBUTING TO TERRESTRIAL GLACIOGENIC DEPOSITION In essence, glaciogenic diamictons are deposited (Fig. 7.1) when either or both of the following conditions are fulfilled: (a) englacial debris is released from active or stagnant glacier ice in a subglacial or a supraglacial position; and/or (b) debris in motion on, or beneath, glacier ice stops moving at a location beyond the direct influence of the glacier either in a subglacial or a proglacial position. A third category of deposition involves sorted sediments (glaciofluvial) deposited in tunnels, but under strong glacial influence. Englacial debris is released when ice, the encasing medium, melts. This process is conditioned by the thermal regime of the glacier and its ambient environment. This thermal regime is the sum of many heat sources, including atmospheric components (air, precipitation), meltwater, the geoid, friction generated by deforming ice and glacier sliding, normal pressure generated by the overburden of ice and debris, and the ice itself. A reduction in any of these heat sources may

171

172

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

Ambient environment Mass balance

Gravitational stress Fall Slump

active

Pressure melting point Freezing front

Stagnant/retreating ice-front

Ice dynamics passive

Shear stress Normal stress

Thermal regime

Slide Flow

c en t r ati o n Deb r i s c o n hard Lodgement

cavity

Ice : water ratio

soft Substrate Frictional resistance Material strength Permeability Transmissivity Pore water pressure Sediment flux

Deformation Position Lodgement Supraglacial Englacial Subglacial

Melt-out Flow Glaciofluvial deposition

Advancing ice-front

FIG. 7.1. Major factors, associated variables and processes responsible for terrestrial glaciogenic deposition.

enable a freezing front to advance from the glacier into previously released debris (Menzies, 1981a; French and Gozdzik, 1988) and return it to an englacial position, unless the porewater pressure within the deposited diamicton is sufficient to inhibit the advance of the freezing front. Debris that is released by stagnant ice subglacially or supraglacially onto a horizontal ice surface that is gradually lowered to the ground, will remain in a stationary position and can be deemed to have been deposited. However, conversely, on release, debris may remain in motion, either beneath the sole of the glacier or flowing across its surface, if the shear stress to which it is subjected is greater than its material strength. This condition applies whether the debris is responding to either glacier shear stress at the base of the moving glacier, or to gravitational stress on the glacier surface or within subglacial cavities. Frictional resistance between a clast and the glacier bed is dependent upon the area of contact between them, and is proportional to the normal load (Drewry, 1986). Movement of a particle will cease when the load exerted on the bed reaches a critical threshold. In a debate about the lodgement of till Boulton (1975) equated normal load to the product of clast weight and effective normal pressure (overburden pressure less porewater pressure): ice velocity was a secondary

component contributing to total glacier shear stress. In contrast, Hallet (1979, 1981) considered normal load to comprise the weight of the clast together with the velocity with which it approaches the bed as basal ice melts. In Hallet’s view, normal load is independent of cryostatic pressure and the downward and lateral velocities are the key components of glacial stress. Material strength, expressed in terms of Coulomb’s equation (Menzies, 1995, chapter 5), is related to the cohesion of the sediment, its angle of internal friction and its effective normal pressure. In the glacial environment the most important components are porewater pressure and material characteristics of size and sorting. These are dependent on factors such as the rate of supply of water, either from melting ice or from external hydraulic sources, the thickness and texture of the sediment, the ease with which it drains, and the lithological characteristics of the available material. These in turn are related to and determine the permeability of the substrate, either ice in the case of subglacial debris, or subjacent debris or bedrock in the subglacial context. 7.3. MECHANISMS OF TILL DEPOSITION The preceding outline of factors contributing to processes of glacial deposition indicates the complexity of

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

FLOW TILL

MELT - OUT TILL

DEFORMATION TILL

LODGEMENT TILL

FIG. 7.2. The conceptualization of till deposition as four principal endmembers of a tetrahedron (after Dreimanis, 1988; reprinted from Goldthwait, R. P. and Matsch, C. L. (eds), Genetic Classification of Glacigenic Deposits, with permission of A.A. Balkema, Rotterdam).

the glacial environment. The direct and indirect influence of ice in the debris-release mechanism has been discussed at great length. Similarly, the roles of active and passive ice and the position of debris release have generated much debate. The outcome of these protracted discussions (Goldthwait and Matsch, 1988) is a series of genetic classifications of till. These show

SUPRAGLACIAL MELT-OUT and resedimentation FLOW

SEQUENCE VERTICAL

considerable variety in detail but most involve some or all of the following types of till: deformation, lodgement, subglacial (confined) melt-out/sublimation, supraglacial (unconfined) melt-out and flow till. The genetic interrelationships of these four basic processes have been conceptualized by Dreimanis (1988) and later Hicock (1990) as end members of a complex ‘till prism’ (Figs 7.2 and 7.3). The processes of supraglacial melt-out and flow have been described in greatest detail (e.g., Lawson, 1981a,b). In contrast, deformation, lodgement and subglacial melt-out processes are more difficult to investigate directly because of inaccessibility and neither are they easy to simulate because of problems of scale and time. Debris flows under the influence of gravity when its constituent particles lose cohesion, usually owing to high porewater pressure. Debris melts out when ice undergoes a phase change from solid to liquid and/or gas. Debris lodges when resistance to movement exceeds the shear stress imparted by the glacier or gravitational force. Movement may be inhibited by adhesion between particles and a hard substrate, by the increasing resistance met by a particle ploughing into a soft bed, by localized freezing and by the gradual attenuation of stress with increasing depth in a soft, mobile bed material. Therefore, the

BASAL MELT-OUT of stagnant ice LODGEMENT (REPETITION) BASAL MELT-OUT of lodged sheets Subglacial resedimentation in cavities LODGEMENT SUBSOLE

DEFORMATION

by subsole drag

RESEDIMENTATION due to gravity and high pore-water pressure

GRAVITY

173

DRAG

TIME FIG. 7.3. A view of the spatial and temporal relationships between till types (after Dreimanis, 1988; reprinted from Goldthwait, R. P. and Matsch, C. L. (eds), Genetic Classification of Glacigenic Deposits, with permission of A.A. Balkema, Rotterdam).

174

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

fundamental mechanisms involved in glacial deposition are flow, melt (and sublimation) and frictional resistance, acting separately or in concert.

in appearance, a characteristic that is often said to typify lodgement till (Menzies, 1989a). This concept of pervasive shear or homogenization has not been accepted uncritically in its application to tills of the Laurentide Ice Sheet (Clayton et al., 1989) in North America and the Lowestoft Till Ice Sheet in eastern England. If, however, the process of homogenization, as described by Hart and Boulton (1991), does occur, then tills that have traditionally been described as lodgement tills on account of their massive appearance, overconsolidation and strong clast preferred orientations, could equally be classified as deformation tills (Plate 7.1a,b). For this reason, Hart and Boulton (1991, p. 34) advocate the abandonment of the term non-deformed lodgement till as they envisage, ‘most (if not all) tills have at some time been subjected to some degree of soft bed depositional processes and thus implicitly some deformation’. Less controversial, however, is the idea of deposition at the base of the deforming layer. As the body of subglacial debris thickens, towards the glacier margin the strength of this substrate increases and strain in the lower part of the debris is attenuated until frictional resistance exceeds shear stress and the material becomes immobile (Boulton and Hindmarsh, 1987; Menzies, 1989a; Hart et al., 1990) and can be considered to have become lodged.

7.3.1. Deformation

7.3.2. Lodgement

Deformation structures are widespread in soft, unlithified sediments. Boulton and Jones (1979) demonstrated experimentally that in subglacial debris beneath Brei∂/ amerkurj¨okull, Iceland, strain increased upwards towards the base of the glacier. Earlier, Banham (1977) (Fig. 7.4) had proposed a model of glaciotectonic deformation, where strain also increases upwards. Banham distinguished a principal structural boundary at the base of ‘penetratively cleaved’ sediments marking the base of the zone that clearly shows the imprint of glacier stress (Chapter 14). Boulton (1987a, fig. 7) proposed a similar model of subglacial deformation showing zones of stability (B) and increasing strain (A), respectively below and above a plane of d´ecollement. In zone A deformation may be so intense that the sediment becomes homogenized (Hart and Boulton, 1991) and massive

In contrast to the process described above, the traditional view of the lodgement process involves the ‘plastering on’ of glacial debris from the sliding base of a moving glacier by pressure melting and/or other mechanical processes (Dreimanis, 1988). Hart and Boulton (1991) differentiate subglacial deposition according to the nature of the substrate, recognizing rigid and soft (i.e., deformable) beds (Fig. 7.5). For rigid beds, lodgement is a frictional process accompanied by pressure melting during which the debris is ‘plastered’ against the substrate, while under deformable bed conditions, Hart and Boulton (1991, p. 345) suggest that ‘purely frictional lodgement cannot occur as there is nothing rigid for the clasts to lodge against’. Under soft bed conditions lodgement may occur when large clasts plough into the sediments, which retards and eventually halts it (Plate 7.1(c),(d)).

D 2 C 1 B

A FIG. 7.4. Model of glaciotectonic deformation. A, undeformed bedrock; B, weakly deformed bedrock; C, penetratively deformed bedrock; D, lodgement till. 1, principal structural boundary; 2, principal lithologic boundary (after Banham, 1977; reproduced with permission of Scandinavian University Press).

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

(a)

(b)

(c)

(d)

175

PLATE 7.1. (a) Overturned deposits (mottled Valley Farm Paleosol and sand wedges), the sheared zone and ‘homogenized’ zone of Hart and Boulton (1991) at Great Blakenham, Gipping Valley, Suffolk, UK. (b) Middle Pleistocene sediments at Newney Green near Chelmsford, Essex, UK, illustrating Banham’s model of glaciotectonic deformation. (c) Vertical exposure of lodgement till, Geilston, Scotland. (d) Horizontal exposure of lodgement till, Fjallsj¨okull foreland, Iceland. Note ‘bow wave’ in sediment in front of the nearest block. Scale bar is 1 m in length.

176

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

Rigid bed

Ice

Rigid bed

subglacial lodgement forming due to frictional processes

Soft bed

Ice melt release

Deformable bed

sediment moved laterally due to high shear strains movable base of deformation below this sediment is “frozen”

FIG. 7.5. A schematic diagram illustrating the formation of till on rigid and deformable beds (after Hart and Boulton, 1991; reprinted from Quaternary Science Reviews, 10, Hart, R. J. and Boulton, G. S., ‘The interrelation of glaciotectonic and glaciodepositional processes within the glacial environment’, p. 341, 1991, with kind permission of Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, UK).

However, this latter statement implies that the relative velocity of ice and the deforming substrate is equal at their interface. If the velocity of the ice exceeds that of the debris beneath, some frictional resistance will be exerted by the debris, which may induce lodgement in the traditional sense and allied streamlining. 7.3.3. Melt-out Till is also deposited by melt-out processes (Boulton 1970; Lawson, 1979a; Shaw, 1985). Boulton (1970) proposed the term ‘melt-out till’, although the concept is not new (e.g., Goodchild, 1875; Carruthers, 1953). Melt-out is the slow release of debris from the lower or upper surface of stagnant ice, referred to respectively as subglacial and supraglacial melt-out. Sublimation is an associated mechanism involving the solid to gaseous phase change rather than the solid to liquid phase change as in melt-out. During periods of marked glacier retreat melt-out may be the most productive till-forming mechanism, especially in marginal situations where warm ambient

air can penetrate the glacier system, although thinning of the glacier in its marginal zone may counteract this by reducing overburden pressure. Melting may also operate preferentially over crustal hot spots where geothermal heat flux is higher than normal. Melting will also be influenced by the thermal conductivity of the debris, which is controlled by factors such as its texture, composition and porosity. In the supraglacial position, slope angle and aspect influence the effect of meteorological variables such as temperature, precipitation and wind. The passive nature of the ice is important to melt-out processes and contrasts with pressure melting resulting from sliding or internal deformation associated with the traditional active lodgement process. During melt-out subglacial debris is confined by the ice above, while supraglacial debris is confined by the accumulating debris. Such confinement should inhibit deformation and mixing. Melt-out till should preserve the characteristics of the englacial debris, depending on the ratio of ice to debris. Solid layers or lenses of debris are likely to show little change of configuration on release while ice with only scattered clasts or pockets of sediment may lose all pre-melt structural characteristics and produce a massive deposit. A high debris-to-ice ratio should allow greater preservation of englacial characteristics as fewer voids are created during ice melt. Only in the unusual event of there being no interstitial ice within the debris, will no modification of the englacial properties occur during melt-out. Lawson (1979b), from observations on the Matanuska Glacier, Alaska, suggested that clast fabric, debris stratification, bulk texture and deposit geometry should be well preserved though clast scatter may increase and clast dip decrease. Water released during melting may be available for sorting sediment. The discharge of meltwater determines the degree of sorting during the melt and is dependent on the volume and rate of icemelt, ice surface gradient, hydrostatic head and the permeability of subjacent debris. The higher the discharge the larger the calibre of material that is removed and the coarser the sorted layer that is left behind or the larger the cavity that may be developed. Shaw (1979) considers that sorted layers, and drapes of these and interbedded diamict layers over large clasts, are indicative of the basal melt-out process exemplified by what he called the Sveg Tills of

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION Stage 1

Stage 2

Stage 3

Ice layer Debris-rich ice

Till

Sorted sediment Large clast

FIG. 7.6. Origin of relationships between large clasts and stratification of sorted and diamicton layers in stratified till at Overberg, Sweden, indicative of the basal melt-out process (after Shaw, 1979; reproduced with permission of Scandinavian University Press).

Sweden (Fig. 7.6). Paul and Eyles (1990) argue that the preservation potential of melt-out till is low, especially melt-out till formed in a supraglacial position where potentially high porewater pressures, above relatively impermeable and steep ice-marginal gradients, encourage debris flow. 7.3.4. Flow The concept of flow till has generated a great deal of controversy but much of the argument is terminological and revolves around the question of whether glacial debris that has flowed should be called till

177

(Lawson, 1979b; Dreimanis, 1988). Rather less effort has been applied to the elucidation of flow processes. Two groups of flow process are recognized and attributed to squeeze and gravity. ‘Squeeze’ implies active ice or at least the normal stress of stagnant ice acting on a sediment layer of low compressive strength. An upward component of movement may be present during this process. It is arguable whether such a process is more accurately described as deformation or flow. In contrast, flow is traditionally viewed as a purely gravitational process operating down the maximum ice surface gradient. If water released during ice-melt is not easily evacuated glacial debris may become liquefied and flow. Whether or not flow occurs depends on the ratio of sediment strength to shear stress. The key parameters are porewater pressure, thickness of the sediment layer and slope angle of the buried ice. Material strength is reduced when porewater pressure is increased and this may be achieved in several ways. Meltwater may be added to the debris if not evacuated from the area; the debris may consolidate as the sediment adjusts to melting or the debris may be loaded by the superimposition of debris flows from higher up the slope. The stress imposed on the debris is increased by the thickening of the debris pile, either by the addition of further melt-out at the base or by the superimposition of flows from upslope. An increase in the slope of the underlying ice caused by thickening or differential melt will also add to the stress on the debris. Glacier ice typically steepens towards its margin and ice surface relief is increased by irregular melting and crevassing. This provides locations for a wide variety of mass movement processes to occur, including fall, topple, slump/collapse, rotational and translational sliding, rolling and creep in addition to flow. However, the special conditions of a melting ice margin with excessive quantities of water overlying a generally impermeable glacial substrate usually ensure that flow is the dominant process, or rather the continuum of processes, of re-sedimentation. Flow mechanics are complex. Several mechanisms of grain support and transport (grain–grain interaction, buoyancy, turbulence, fluidization, liquefaction, dispersive pressure) operate simultaneously within flowing debris. Relationships between

178

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

texture, surface shear strength, porosity and bulk wet density all suggest that flows vary systematically with changes in water content and produce a continuum of flow morphology as flow varies from plastic to viscous in type. For discussion purposes only, Lawson (1979a) illustrated the transverse and flow-parallel cross-sections of four discrete types of sediment flow, emphasizing their transitional relationships. Flows with low water content (8–14 per cent by weight) and high bulk density retain a measurable shear strength enabling the body of material to operate as a discrete, lobate self-supported unit with shear confined to a thin zone at the base of the flow (Lawson Type I Flow). As water content increases (14–19 per cent) and strength weakens the basal shear zone increases in thickness, supporting a plug of unsheared, rafted material (Lawson Type II Flow). Meltwater may move over the sediment flow surface. The movement of the undeformed plug may induce sliding and rolling of clasts lying in front of the flow. Further increases in water content (18–25 per cent) reduce plug thickness and continuity until the whole flow is in shear and moving within an incised channel (Lawson Type III Flow). When water content exceeds saturation throughout the flow, the debris becomes liquefied, fluid flow operates, grain–grain contact is lost and other grain support mechanisms largely cease (Lawson Type IV Flow). Deposition may result from one or several of the following ambient or internal changes; a reduction in bed gradient, a decrease in flow mass thickness, the loss of interstitial fluids by drainage through permeable substrates, the blockage of the flow path, or the addition of dry sediment that increases the bulk density and reduces the relative water content of the debris. This description of the re-sedimentation processes that can dominate some glacial margins (Lawson estimates that 95 per cent of the deposits of the Matanuska Glacier are ‘re-sedimented’) shows clearly that they conform to mass wasting processes that are not exclusive to the glacial environment. Nevertheless, the debris has been accumulated by glacier activity and transported and deposited from a glacier. It is unlikely that debris flow mechanisms alone could have transported the same debris to the same place as

than by glacial means. Therefore it seems reasonable to retain the term ‘till’ for deposits deriving from flow mechanisms as they are integral members of the assemblage of ice marginal sediments and retain at least one property (lithology) imparted exclusively by glacial processes (Chapters 10 and 14). 7.4. MECHANISMS OF GLACIOFLUVIAL DEPOSITION It is clear from the preceding discussion, that water is a contributory factor in most depositional mechanisms. In glaciofluvial contexts it becomes the predominant element, especially in subaerial supraglacial and proglacial locations. Proglacial fluvial processes, especially those associated with river braiding, are extensively discussed in Chapter 9, and will not be discussed further here. In the subglacial context, ‘normal’ subaerial conditions may be significantly altered. In englacial and subglacial spaces confined by a glacier, meltwater may flow under hydrostatic pressure generated by a steep hydraulic gradient. When full of water englacial and subglacial conduits function as pipes and the special conditions of ‘fullpipe flow’ apply. Glaciofluvial sediments are characterized by: inversely graded beds, bimodal particle size distributions, a-axis imbrication of clasts and matrix-supported gravels, attributed to transport under conditions of dispersive stress (Bagnold, 1954) or as a consequence of kinetic sieving (Middleton, 1970). Textural bimodality and matrix-supported gravels are also attributable to dispersive flows and flow-parallel, a-axis clast fabric results from repeated intergranular contact between clasts in suspension. It has been argued that the movement of dispersive flows will only be sustained on a slope that greatly exceeds the longitudinal slope of the conduit in which sediment is being transported. In their investigation of flow in full glacial pipes, McDonald and Vincent (1972) found that upper flow regime dunes were succeeded not by antidunes but by full particle suspension, standing waves being inhibited by the presence of the tunnel roof. This full suspension of particles represents the final stage of four pipe-transport regimes (a–d) (Newitt et al., 1955; Saunderson, 1982):

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

l

l

Dir

l

ec

Cromer

l

l

l

Ice

Happisburgh l l

l

l

l

l

l

l

Lowestoft

l

l

l

l

Margin of Lowestoft Till

l

Area of North Sea Drift Administrative boundary

Suffolk

R

p pi Gi

Cambridge

l

l

l

l

l

l

l

l

l

ng

Great Blackenham

l

l

l

l

l

l

l

l

Essex

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

Chelmsford l

l

l

l

l

l

l

l

l

l

l

N

l

l l

l

l

l

This discussion of the processes of glacial deposition reveals that many details associated with lodgement and deformation remain to be clarified. Eastern England provides an excellent case study of the current debate concerning till-forming processes. One debate pertains to the subglacial or subaqueous origin of deformation structures in tills along the north Norfolk coast (Fig. 7.7). The other involves a model of till formation applicable to most of the remainder of East Anglia. Both of these debates are crucial to the accurate interpretation of glacial environments in the region as well as contributing to general questions of ice sheet dynamics and till genesis. Most authorities now recognize tills from two Pleistocene stages in the East Anglian region of Britain; the Anglian Stage (OIS 12) and the Devensian Stage (OIS 2). Anglian-age till of the Lowestoft Formation is generally considered to be widespread but Devensian Till is restricted to the northwest of the region and will not be considered further. In northeast Norfolk, Scandinavian ice deposited the North Sea Drift Formation. Penecontemporaneous ice of British origin deposited the Lowestoft Till over most of East Anglia including some of the area previously occupied earlier by the Scandinavian ice. A subglacial origin was usually assumed as the mechanism of deposition of these tills. With rare exceptions (Rose and Allen, 1977; Allen et al., 1991)

l

l

l

Norfolk l

l

7.5. TILLS OF EASTERN ENGLAND – DEBATES ABOUT MECHANISMS OF TILL DEPOSITION – A CASE EXAMPLE

Ice Direction

l

tio

n

North Sea

l

(a) at low velocities a stationary bed with bedload and saltation, and ripples as the primary bedform; (b) at higher velocities a moving or sliding bed, particularly for coarser particle sizes; (c) heterogeneous suspension, in which all of the sediment available is in suspension, the finer sizes distributed uniformly throughout the pipe crosssection and the coarser confined to the lower half of the cross-section; (d) homogeneous suspension in which the sediment available is all in suspension but all sizes distributed uniformly throughout the crosssection.

179

0

km

30

London

FIG. 7.7. Map of East Anglia, UK, showing features and locations mentioned in the text.

the Lowestoft Till was referred to as lodgement till, while the North Sea Drift tills were generally thought to have melted-out from subglacially deformed ice (Banham, 1977). However, following detailed analysis during the last couple of decades a strong divergence of opinion has emerged concerning mechanisms of deposition of North Sea Drift tills (Banham, 1988; Eyles et al., 1989; Hart and Boulton, 1991; Lunkka, 1994) and, to a lesser extent, the Lowestoft Till (Whiteman, 1987; Hart et al., 1990; Allen et al., 1991; Hart and Roberts, 1994; Roberts and Hart, 2000a,b), and there is clearly considerable scope for further work. The first of these debates concerns the North Sea Drift, a complex sequence, up to 20 m thick, of diamictons (the Cromer Tills) and sorted sediments including clays, silts, sands and gravels. In the south of the region near Happisburgh (Fig. 7.7), the bedding of these sediments is subhorizontal but is substantially deformed northwest of a set of listric faults where the deposits have been referred to as the ‘Contorted Drift’, or laminated diamicton (Hart and Boulton, 1991). They are composed mostly of local lithologies but with some Scandinavian and possibly Scottish erratics.

180

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

Hart and Boulton (1991, p. 338) suggest that ‘the most striking features of the diamicton are the laminations and the ‘pods’, representing respectively, highly attenuated, conical sheath folds and more competent masses (often chalk and sand) with strain shadows, generated by simple shearing within a subglacially deforming zone. That the ‘laminated diamicton/Contorted Drift’ may be glaciolacustrine (Gibbard, 1980) or glaciomarine (Eyles et al., 1989) was rejected by Hart and Boulton (1991) owing to an apparent lack of sedimentary structures and grading, and paucity fold noses in the laminations. Deformation extending throughout the sequence and rather than confined to individual beds was considered to disprove a glacioaquatic origin. They also argued against these sediments resulting from meltout of deformed englacial structures generated at the margin of a cold-based ice sheet asserting that observed basal debris layers are an order of magnitude less than the ‘laminated diamicton’ (although approximating the thickness of individual Cromer Till diamict units). An alternative view of the origin of the North Sea Drift tills (Eyles et al., 1989) stresses that in the southeastern part of the region beyond the marginal listric faults of the ‘Contorted Drift’, a sedimentary sequence of multiple laminated diamict facies (Cromer Tills) and sorted sediments is subhorizontally bedded and is traceable for several tens of kilometres in cliff sections. Eyles et al. (1989) differentiate the sequence into ‘diamict facies’ and ‘associated facies’. The latter largely comprise sorted sediments where an overall coarsening upwards sequence is recognized. Grading is not discernable in the laminated and bedded diamict facies of the ‘Contorted Drift’. In the flat, undisturbed sediments of the southeast, however, the diamicton matrix coarsens upwards into a sandmud with large pillow structures filled with deformed sands loaded into the underlying finer diamict. Load structures are truncated by an erosion surface on which there is an accumulation of shell fragments (<2 mm) and a pebble lag. A higher diamicton unit with similar characteristics is separated from the lower unit by a waterlain facies that shows graded laminations, wave-formed ripples, soft-sediment deformation structures and evidence of ice-rafting. In their view the intimate association of diamict with

Homogeneous Zone (c) Diamicton

Sheared Zone (b) Diamicton and Pre-glacial sediments

Overturned Zone (a) sand wedge

paleosol

Undormed Zone sand

FIG. 7.8. Schematic diagram of subglacial deformation caused by the Lowestoft Till Ice Sheet, at Great Blakenham, Suffolk (after Hart and Boulton, 1991; reprinted from Quaternary Science Reviews, 10, Hart, J. K. and Boulton, G. S., ‘The interrelation of glaciotectonic and glaciodepositional processes within the glacial environment’, p. 346, 1991, with kind permission of Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, UK).

waterlain facies points to a subaqueous origin for the North Sea Drift. Plumes of fine suspended sediment are thought to have emanated from an approaching ice front with ice rafted debris deposited as a series of laminated sediments with dropstones in a proglacial marine environment. Intraformational folds are interpreted as creep and slumped forms developed down the regional chalk bedrock palaeoslope based on the direction of fold hinges. Sand and gravels towards the top of the sequence are viewed as shoreface and beach sediments deposited after the ice retreat. These sediments are well preserved by loading into the ‘overpressured’ diamictons. This debate has continued with Lunkka (1994) largely supporting the subglacial deformation hypothesis, though acknowledging a significant subaqueous contribution to at least one of the diamictons, and Hart maintaining a stance in favour of subglacial deformation but conceding, along with Roberts (Hart and Roberts, 1994; Roberts and Hart, 2000a,b) that some laminated diamictons possess evidence of subaqueous

PROCESSES OF TERRESTRIAL GLACIAL DEPOSITION

deposition not destroyed by secondary subglacial deformation. The second debate concerns the Lowestoft Till in southern East Anglia, part of a till sheet extending over most of that region. The Lowestoft Till, traditionally, has been interpreted as a lodgement till, though there are references to deformation, melt-out (Whiteman, 1983), flow (Allen, 1984) and slump facies (Rose, 1974) in relation to a few restricted areas. Whiteman (1987) and Allen et al. (1991) applied Banham’s (1975) three-zone model (Fig. 7.4) to a sequence of sediments deposited by the Lowestoft Till Ice Sheet. Hart and Boulton (1991) extended the model suggesting that the apparently homogenous zone (their zone c (Fig. 7.8), equivalent to zone A of Boulton (1987a) above the zone of shearing (zone b) represents a zone of extreme shear and fold attenuation culminating in a deposit that appears massive in the field, although this degree of deformation needs study at the microscopic scale. Hart and Boulton

181

(1991) interpret the whole sequence as ‘representing a continuous deposition as. . . ice. . . advances over the area’, on the basis that ‘the boundaries between these different zones are not abrupt but each zone grades into another’. In fact, the boundaries between the zones are usually abrupt and reveal very marked differences in texture, lithology and clast fabric (Allen et al., 1991), in part, reflecting differences between local glaciotectonized sediment (Benn and Evans, 1996) and far-travelled till. Even within the supposedly ‘homogeneous’ zone significant differences of colour, lithology and clast fabric exist between a lower, dark subunit and an upper lighter subunit, reflecting either change in ice sheet provenance, differences between facies associated with advancing and retreating ice, or differences in depositional processes. The interpretation of till sequences will remain problematic until unequivocal discriminant properties are established for the full range of till facies.

This Page Intentionally Left Blank

8

SUBGLACIAL ENVIRONMENTS J. Menzies and W. W. Shilts 8.1. INTRODUCTION No other glacial environment has left its imprint so indelibly on glaciated areas as the subglacial environment. Other than processes active in the glaciomarine environment, subglacial processes and sediments dominate previously glaciated continental land surfaces and epicontinental margins. Subglacial erosional, depositional and bedforming processes are among the most complex, yet least understood set of glacial processes. Our poor knowledge stems from the inaccessible nature of what occurs beneath an ice mass and the limited extent of modern analogues accessible when attempting to comprehend Pleistocene and pre-Pleistocene subglacial environments. The products of the subglacial environment and its impact upon the land surface are of paramount importance to society. The spatial and volumetric dominance of subglacial sediments in those countries once glaciated cannot be understated. Most building foundations, roads, railroads, aircraft runways, sewage disposal sites, toxic waste sites, agricultural land, groundwater aquifers and construction materials in those areas last glaciated in the Pleistocene, utilize and are in direct contact with subglacial sediments. As our increasing awareness of environmental hazards and rationally sound land use develops into the twenty-first century, the demand for expert and applied knowledge of glacial sediments and environments will dramatically rise with a prime place given 183

to knowledge of the subglacial environment (De Mulder and Hageman, 1989; Coates, 1991). Perhaps no other glacial environment impacts on the daily lives of millions of people to such a high degree as does the subglacial. 8.2. DEFINITION Any definition is fraught with problems in terms of defining and delimiting where the subglacial environment ceases and the proglacial, submarginal and/or subaqueous glacial environments commence. The subglacial environment is that glacial sub-system directly underlying an ice mass in close contact with the overlying ice, including those cavities and channels beneath the ice that are not influenced by subaerial processes. On land, the subglacial environment may continue for some distance beyond the apparent surface ice margin since buried active ice may underlie debris that has been deposited either from supraglacial sources or from various meltwater streams exiting the ice (Fig. 8.1(a)). Likewise, on entering a body of water an ice mass may be sufficiently thick or the water sufficiently shallow to permit ice to remain in contact with the bed, the basal environment thereby remaining subglacial. However, where ice begins to rise off its bed, the subglacial environment can be said to cease down-ice of this dynamic hinge point, known as a groundingline (Fig. 8.1(b)). The nature of the frontal zone of an

184

SUBGLACIAL ENVIRONMENTS

(a)

Debris bands in active ice

Cavities and tunnel

Sandur plain

Deformable bed Subglacial diamicton Debris band in stagnant glacial ice Flow till Buried stagnant ice

(b) Moulin

Sediment-laden fresh meltwater

Englacial tunnel Englacial stratified sediment

Slumps

Bedrock

Unstratified diamicton Basal diamicton

Stratified diamicton

Seawater

Laminated mud with dropstones

FIG. 8.1. (a) Model of an ice mass with a land-based margin (adapted from Boulton, 1972; Drewry, 1986). (b) Model of an ice mass with an aquatic margin (adapted from Edwards, 1986).

SUBGLACIAL ENVIRONMENTS

ice mass, whether tidal or floating, whether in the form of an ice cliff or a thinning tapered snout, impacts upon subglacial environments. Down-ice from a grounding-line an extensive proglacial subaqueous proximal environment may exist. Such a proximal subaquatic zone may become transiently subglacial for short periods of time and at different parts of the bed (Chapter 11). Finally, the subglacial environment is a boundary interface where a complex set of processes interact altering the morphological, thermal and rheological state of the interface, evolving into several and, at times, a single interface junction. The nature of this interface(s) controls glacial erosional, transportational and depositional processes. The key to understanding subglacial bed types and associated forms lies in comprehending the mechanics of this interface. These complex boundary zones migrate across the landscape with every advance and retreat of the ice masses. Therefore all terrains covered by glaciers have been affected and altered by the passage of this boundary zone. Subglacial depositional and erosional phenomena are often destroyed, radically altered or obscured by later processes associated with proglacial environments following gradual ice retreat or by more complex subglacial overriding and subsequent ice retreat, or by subaerial diagenesis. 8.3. THE SUBGLACIAL INTERFACE – BED TYPES AND CONDITIONS The subglacial environment at any one point in time and space is the product of a large group of interrelated factors (Fig. 8.2). Many factors that directly influence the subglacial environment are indirectly affected by external factors, for example, climate, rates of snow transformation and tectonic setting. The end product of these diverse influences reduces to a series of subglacial interface bed types that can be distinguished by thermal and rheological variations that are then subject to an enormous dynamic range of temporal and spatial variations. Any subglacial interface and related bed type is a function of the prevailing basal ice and bed conditions. These conditions are the result of complex relationships between basal ice dynamics, subglacial sediments and bedrock, subglacial hydraulics and the ambient ther-

185

mal state for any given area of glacier bed. Changes in basal ice and/or bed conditions may be widespread or local and may develop rapidly or slowly. These changing conditions may be of enormous magnitude or may simply be minor variations within the interface boundary zone, the former being sometimes detectable whilst the latter may leave little or no imprint in the sedimentary record. Subglacial thermal conditions vary areally across a bed interface and over time, creating a transient subglacial environment in which a large and complex array of differing sedimentological conditions associated with distinct thermal regimens can pass repeatedly over a given point on the glacier bed. Directly connected with these thermal phase changes are rheological changes associated with unconsolidated bed sediments (Alley, 1989a,b; MacAyeal, 1989; Kamb, 1991). Research from West Antarctica suggests that parts of that ice sheet, in particular Ice Stream B, is underlain by a layer of deforming sediment (Alley et al., 1987a; Blankenship et al., 1987; MacAyeal, 1989). The deforming debris layer extends downstream for about 200 km and averages 6 m thick with a variation of between 1 and 12 m. The upper surface of the layer appears comparatively smooth but flutes (~10 m high by 300–1000 m wide) were detected at the lower interface. Typically, the debris has a high porosity (n~0.4), high porewater pressure (ρw = 50 kPa), debris cohesion <15 kPa, low effective pressure and is probably a water-saturated unconsolidated diamicton. Additional proof of deformable bed conditions under present-day ice masses has been reported from Iceland (Boulton et al., 1974), Alaska (Engelhardt et al., 1978; Kamb et al., 1985), China (Echelmeyer and Wang, 1987), Sweden (Brand et al., 1987); Ellesmere Island, Canada (Koerner and Fisher, 1979) and Washington State, USA (Brugman, unpublished 1985). The data gathered from these diverse locations show actual or calculated values for strain rate, deforming layer depth and basal ice velocities (Table 8.1). Evidence of deformable debris layers beneath present day ice masses has been used to suggest that deforming bed conditions possibly existed beneath the Pleistocene ice sheets (Jenson et al., 1995, 1996; Greve and MacAyeal, 1996; Clark, 1997; Maher and

186

SUBGLACIAL ENVIRONMENTS

ice–bedrock

ice–bed

ice–bed

ice–bed

FIG. 8.2. (a) Model of interrelationships between main parameters influencing subglacial processes and erosion (based upon Sugden and John, 1976; Embleton and King, 1975 and Embleton and Thornes, 1979, modified from Brodzikowski and van Loon, 1991). (b) Model of interrelationships between main parameters influencing subglacial processes and deposition.

SUBGLACIAL ENVIRONMENTS

187

TABLE 8.1. Deformable bed data: basal sediment velocity, thickness and calculated strain rates Location Urumqui Glacier No. 1 Chinaa) Blue Glacier, Washington, USAb Brei∂/ amerkujokull, ¨ Icelandc Upstream B. Camp, Ice Stream B, West Antarcticad Pleistocene Puget Lobe, Washington, USAe

us (m year–1) ~3.00 ~4.00 ~16.00 ~450.00 ~500.00

hs (m) ~0.35 ~0.10 ~0.50 ~6.00 ~15 undilated ~22.5 dilated

es (year–1) ~8.57 ~8.39 ~40.00 ~32.00 ~75.00 ~33.33 ~22.22

a

Echelmeyer and Wang, 1987. Engelhardt et al., 1978. Boulton, 1979. d Alley et al., 1987a. e Brown et al., 1987. b c

Mickelson, 1997; Marshall and Clarke, 1997a,b). Under marginal and lowland areas of Pleistocene ice sheets it is thought probable that soft sediment or deformable bed conditions were widespread (Hart et al., 1990; Ridky and Blankenship, 1990; Hicock, 1992; Menzies et al., 1997). Several subglacial bed types would seem to best typify environmental conditions at the subglacial interface (Chapter 4). The bed types discussed below should be viewed in terms of our present state of knowledge and are likely to be modified and constrained as new data and understanding emerge. Four basic bed types appear to exist: (1) Polar Active, (2) Polar Passive, (3) Temperate Active and (4) Temperate Passive. Most research efforts to date have centred around Temperate Active bed types since these are probably the most widespread. However, some consideration of the other forms is necessary. It has never been defined specifically when an ice mass is characterized by widespread polar, as opposed to temperate, bed conditions. The standard definition of either condition is based on the thermal state of the specific ice mass but no definitive spatial, volumetric and/or temporal parameters have been employed. As a general guide where an ice mass exhibits a specific thermal condition areally over several tens of square kilometres and from close to the surface to its bed, over at least a decade of observations, perhaps a specific thermal state can be said to prevail. However, polythermal bed conditions, as a general encompassing state, best typify most ice masses.

8.3.1. Polar Bed Types Under active polar glacier ice conditions it has been assumed in the past that no movement occurs at the ice–bed interface, that frozen conditions prevail and thus no free meltwater is present, and that the glacier ice is below the pressure melting point (Fig. 8.3). However, Shreve (1984) demonstrated, theoretically, that movement at the frozen interface can occur, and Echelmeyer and Wang (1987) observed movement along the ice–bed interface in a subpolar glacier (Urumqui Glacier No.1, China). The rate of ice movement under polar conditions in comparison with that under temperate ice masses is magnitudes less. In terms of the potential volume of tectonized sediment, however, considerable amounts of sediment and/or bedrock may be tectonically stacked and ‘re-transported’ under polar thermal conditions (Van der Wateren, 1987; Croot, 1988a,b; Aber et al., 1989). 8.3.2. Temperate Bed Types Temperate or warm glacier ice conditions are the result of glacier ice being at or above the pressure melting point. Under these conditions a spectrum of environments reflecting both thermal and rheological fluctuations can be expected to exist. Many thermal fluctuations are of either short-term spatial and/or temporal variation. Localized patches of a frozen bed state may occur temporarily, as may passive stagnant conditions. However, the major distinction between

188

SUBGLACIAL ENVIRONMENTS

ICE MOVEMENT A

B WARM MELTING ZONE

FREEZING ZONE Glacier Flow Lines Glacier Sole

C WARM FREEZING ZONE

Deposition Local Deposition/Erosion

Near-Zero Erosion/Deposition

D FREEZING ZONE

Erosion

Active Entrainment Active Deposition

GLACIER BED

A

Glacier totally frozen to its bed

B

Glacier melting at bed with isolated patches remaining frozen, larger toward zone A, smaller toward zone C

C

Glacier beginning to freeze to bed, patches melting larger toward zone B, smaller toward down-ice edge zone D

FIG. 8.3. Models of subglacial thermal regimes, their spatial relationships and processes of subglacial erosion, transportation and deposition.

polar and temperate subglacial bed states is that meltwater in some form is present in the latter. Three dominant bed states best typify our present understanding of temperate bed conditions (1) Hard or Rigid, ‘H’ beds, (2) Soft or Mobile, ‘M’ beds, and (3) Quasi-Rigid/Quasi-Soft, ‘Q’ beds. These three bed states include a series of variants that are spatially correlative to transient changes in subglacial interface thermal, rheological, geological and morphological constraints. A complex series of diverging and interconnecting interface boundaries are thought to evolve beneath these ice masses (Menzies, 1987; Menzies, 1995, chapter 4). 8.3.2.1. ‘H’ beds ‘H-beds’ are characterized by a hard or rigid bed state in which almost all basal motion is attributable to overlying ice stresses. Movement is not facilitated by the development of lowered effective stress levels in sediment or bedrock below the upper ice–bed interface owing to meltwater infiltration. These bed types may develop where an ice mass directly overlies bedrock of low permeability (‘Ha’), frozen debris

(‘Hb’) of low or zero permeability, or unfrozen debris (‘Hc’) of low hydraulic conductivity (e.g., <10–6 m s–1 ). In all instances, most free meltwater moves at the ice–bed interface and is prevented from penetrating below this boundary. In some cases, permeable debris (‘Hd’) subjacent to the ice–bed interface overlying an aquitard is quickly saturated preventing any further meltwater movement except at the ice–bed interface. Finally, debris (‘He’) with a high hydraulic conductivity may overlie an aquifer of similar or higher conductivity permitting high rates of porewater movement through the debris layer. In this instance, provided meltwater production at the upper ice–bed interface does not produce discharges greater than the through-flux of porewater within the debris layer, a process of ice–bed interface meltwater evacuation can be established. The persistence and integrity of an ‘He’ bed state is largely a function of the basal debris layer’s permeability and the maintenance of a hydraulic gradient below the critical hydraulic gradient. In all of the above conditions (except ‘He’) the ice–bed interface is effectively sealed off from the underlying bedrock or debris thus meltwater flows along the ice–bed interface leading

SUBGLACIAL ENVIRONMENTS

to the possible development of decoupling instability (Chapter 5). A rigid sediment structure, in which there is zero intergranular movement, is assumed to exist under ‘Hd’ and ‘He’ bed conditions. Under these conditions porewater, derived from the ice–bed interface, passes through the sediment in a manner described by Darcy’s equation for flow through a porous medium. Q = iK

(8.1)

where Q is the specific discharge, assuming that the inflow rate and outflow rate of porewater is equal through a unit area of sediment, i is the hydraulic gradient defined as the excess hydrostatic pressure in the meltwater as it flows through a specific length of sediment and K the hydraulic conductivity. Flow rates have been estimated in thousands of years except under shallow ‘He’ bed conditions. It seems unlikely that subglacial debris can, exclusively, evacuate subglacially produced meltwaters. Therefore, under all ‘H’ bed states it is probable that substantial meltwater discharge must take place at the ice–bed interface. The style and nature of subglacial channel flow at the upper ice–bed interface (Chapter 5) remains the subject of considerable controversy. Both ‘N’ and ‘R’ type channels are likely to develop or a combination of both (‘C’-channels). Under ‘H’ bed conditions, tunnel valleys (‘N’ channels) may develop across the ice–bed interface (Booth and Hallet, 1993; Piotrowski, 1994). 8.3.2.2. ‘M’ beds In understanding subglacial environments, the presence and possible widespread occurrence of mobile, soft deformable beds requires a major reappraisal of temperate subglacial conditions and the likely effects such deformable beds may have upon sedimentation processes. Typically, these beds are composed of saturated debris of a slurry-like consistency in which free meltwater discharge is less dominant and porewater movement critical. With mobilization of the upper parts of the subglacial debris layer, via intergranular deformation in the form of a saturated debris flow, en masse, then meltwater discharge may be possible. Debris slurry flow direction would,

189

generally, be parallel to the principal subglacial pressure head direction (down the ice flow line toward the ice front). Where major obstructions at the ice–bed interface are encountered local deviations from the main flow direction occur and are evident in the deformation structures and fabric extant within the deformed sediment (van der Meer, 1992; Van der Wateren, 1992; Menzies and Woodward, 1993; Hart, 1994). Subglacial debris flows in a manner similar to a Bingham visco-plastic material; thus, mobilization, once begun, becomes a function of the viscosity of the slurry and its strain rate (Johnson, 1970; Kamb, 1991). In clay-rich debris, it is thought that the rate of deformation is more a function of viscosity (controlled largely by porewater content) than strain rate (Maltman, 1988). Sediment deformation may occur as a pervasive or non-pervasive process. In the former, the sediment as a whole (each individual grain) will deform en masse, therefore existing structures and fabrics within the sediment will be grossly affected, if not erased and homogenized. In the latter process, in contrast, discrete shear planes or zones develop within the subglacial debris layer that ‘carry’ the principal components of applied stress resulting in shear planes or zones separated by areas of sediment that remain largely unaffected. This latter process thus permits pre-deformation structures and fabric between the sheared areas, if any, to remain intact. Non-pervasive deformation has been observed both experimentally and in Quaternary sediments (Talbot and Von Brunn, 1987; Maltman, 1988; Menzies, 1990a; Hicock, 1992; Menzies and Maltman, 1992; van der Meer, 1993, 1997). 8.3.2.3. ‘Q’ beds It is probable that ‘Q’ bed subglacial states, in which hard and soft bed conditions alternate both temporally and spatially at the ice–bed interface, are the most common beneath polythermal ice masses. The importance of ‘Q’ bed states is apparent in the variability of subglacial environmental conditions where changes in subglacial debris rheology, and in basal ice thermomechanical behaviour can occur with varying degrees of frequency, duration and spatial distribution. These

190

SUBGLACIAL ENVIRONMENTS

fluctuating subglacial conditions have major sedimentological implications for Quaternary research where terrain is examined that has been overprinted by several glaciations. 8.4. SPATIAL VARIATIONS IN SUBGLACIAL POLYTHERMAL-RHEOLOGICAL BED CONDITIONS It is apparent that most ice masses are underlain by polythermal bed states. These polyphase beds might be compared to a chequered quilt where each square represents a thermal and/or rheological state that is in continual state of spatial and temporal transition. Any one patch may, over time, alter from hard to soft bed conditions and vice versa. From a sedimentological viewpoint, these subglacial variations can provide glaciological explanations of the spatially complex geomorphological processes that take place. Polyphase subglacial bed conditions account for distinctive bedforming processes, spatial bedform associations, subglacial sediments and associated structures (Eyles and Miall, 1984; Hicock, 1990). A ‘problem’ in the past has been the inability to explain localized glacial deposition. Previously, it had been thought that erosion took place in some distant up-ice location while deposition occurred in marginal areas of ice sheets. Even as late as 1971, debate still persisted concerning short- and long-distance subglacial transport (Goldthwait, 1971). One reason for this perception is that metamorphic and igneous rocks tend to produce limited debris and mostly boulders, whereas sedimentary rock produces much more debris (but few boulders). It is now generally accepted that most subglacial deposition (the bulk of the till matrix) is the result of relatively short-distance transport in the order of 10–15 km or less. Polythermal beds essentially provide conditions in which local erosion, transport and deposition can occur simultaneously across a glacier bed varying both in time and space (Chapter 6). It can be shown that in traversing the bed of an ice sheet from its centre to its margins transient polyphase bed conditions result in distinctive sedimentological and bedform changes and transformations. Figure 8.4 is an example across a section of the Laurentide Ice Sheet in Canada relating landform zones to the bed

geology and changing ice dynamics. At the centre of an ice sheet, where basal ice velocities are at a minimum and polar bed conditions may prevail, limited sediment erosion and, thus, transportation is expected to occur. The geological results of zero velocity and/or polar bed conditions are, however, rarely observed since ice flow centres migrate thus terrain where flow centres existed for some time are affected by previous or later ice movements (Boulton and Clark, 1990a,b). Under ideal conditions a series of polyphase bed states should be found sequentially outward from an ice flow centre. Each individual bed state will exhibit, as the distance from the ice centre extends, increasing basal ice velocity and frictional heat. From the central portions of an ice sheet toward the margins, changes adjacent to and interfingering with one another can be expected to pass into a succession of bed states: polar → temperate ‘Q’ beds with a dominance of temperate ‘H’ beds → temperate ‘Q’ beds with a dominance of ‘M’ beds → temperate ‘M’ beds → temperate ‘Q’ beds with a dominance of ‘M’ beds → temperate ‘Q’ beds with a dominance of temperate ‘H’ beds → polar beds (Fig. 8.5). Such a succession may not always occur nor does this sequence necessarily always follow but, in many instances, such an outward developing series of subglacial bed conditions can be anticipated. The series may alter according to local variations in the geology of the substrate including lithology and structure of outcrops, local topography and the physical nature of pre-existing unconsolidated sediments. It may also vary according to basal sediment grain size and therefore rheological parameters, variations in ice basal stress fields, basal ice thermal patterns, interface morphology and topology, basal ice velocity and subglacial meltwater flux (Fig. 8.6). From the central areas of ice sheets outwards, as basal ice velocities increase, production of subglacial meltwater should also increase, perhaps to the point where sufficient portions of the ice–bed interface begin to slip over the bed (Lliboutry, 1968; Weertman, 1972). Close to ice flow centres where limited debris has been produced and fast zones of ice flow are unlikely to be found, deformable bed conditions are

SUBGLACIAL ENVIRONMENTS

(a)

191 86º 68º

1 68º 20º

Great Bear Lake

Great Slave Lake

HUDSON BAY

Esker system Predominantly bedrock 0

(b)

200 km

58º 94º

58º 104º

1 68º 20º

86º 8º 6

3

Great Bear Lake

4

3 V DI

IC

E

3

E

3 4

K E E W AT

IN

1

ID

Sla ve

Great Slave Lake

R iver

0

3

2

200 km

Edge of Canadian Shield

HUDSON BAY

58º 104º

58º 94º

Hummocky moraine

FIG. 8.4. (a) Eskers radiate outwards from area of the Keewatin Ice Divide, Canada and eventually disappear in the sediment-poor region along the western edge of the Canadian Shield. (b) Landform/sediment zones around the Keewatin Ice Divide. Dark areas represent hummocky moraine, the characteristic landform of the Ice Divide (after Aylsworth and Shilts, 1989; reproduced with permission from Elsevier Science).

192

SUBGLACIAL ENVIRONMENTS

1

2

MELTED

MELTING

FROZEN

FROZEN

MELTING

FREEZING

FROZEN

MELTING

MELTED

FREEZING

BASAL WATER THICKNESS

3

CONTROLLING OBSTACLE HEIGHT

4 ICE LOBE

ICE STREAM TERRESTRIAL ICE SHEET

ICE STREAM MARINE ICE SHEET

FIG. 8.5. Basal ice thermal zones in relation to processes of subglacial erosion and deposition beneath a steady-state ice sheet having terrestrial and marine portions. 1, internal ice flow trajectories; 2, basal topography created by glacial erosion; 3, areal distribution of subglacial meltwater. Subglacial meltwater shown in black in 3 and 4 (after Hughes, 1981; reprinted from Denton and Hughes, 1981 by permission of John Wiley and Sons, Ltd).

unlikely to develop (Fig. 8.6). Instead, bed conditions similar to those defined as ‘H’ bed states may exist. In these inner reaches of an ice sheet some sediment transport by meltwater sheets and channels can be expected but significant fluvioglacial deposition is unlikely (Fig. 8.4). If subglacial meltwater production increases to the point where the ice is decoupled from its bed, frictional resistance will be reduced locally to near zero. Once this state is reached, frictional heat that produces meltwater will drop and meltwater produc-

tion will decrease, allowing the ice to contact its bed once again. With renewed contact, debris production increases, frictional heat increases and the cyclic production of meltwater begins again. This cycle will vary with changing ice flow velocities, in a general way outward from the ice flow centre and locally resulting from topographic irregularities such as depressions where constricted ice flow leads to increased velocities. Likewise, in areas or zones of faster-moving ice such as ice streams, sufficient debris may possibly

SUBGLACIAL ENVIRONMENTS KEEWATIN ICE DIVIDE

1

N

NOUVEAU-QUEBEC ICE DIVIDE

0

Unpatterned Thick Drift Zone

100

2

200

300

3

400

500

4

600

Inner Streamlined Zone

Rogen Moraine Zone Outer Streamlined Zone Sakam

i Mo

rain

e

km

WEST OF HUDSON BAY

S

EAST OF HUDSON BAY

FIG. 8.6. Landform/bedform zonation around the Keewatin Ice Divide, west of Hudson Bay (from Shilts et al., 1987), and around Nouveau Qu´ebec Ice Divide, east of Hudson Bay Canada (after Bouchard, 1989; reproduced with permission from Elsevier Science Publishers).

exist at the bed to enable deformable bed conditions to be generated. In this situation ‘Q’ beds with a dominance of ‘M’ bed conditions may begin to prevail locally. In basal environments of rapid ice flow, subglacial bedform development under deforming sediment conditions may transmutate through a sequence or continuum of forms (Aario, 1977a,b; Menzies, 1987; Rose, 1989b). Where subglacial environments are unsuitable for ‘M’ bed conditions, ‘H’ bed states may exist, with large volumes of subglacial sediment being transported within meltwater channels. Conversely, meltwater streams may carve erosion marks or remnants of previously deposited sediment into fluvial bedforms (Shaw et al., 1989). It is under ‘Q’ beds that subglacial sedimentation of both unsorted and sorted sediments occurs. The timing of deposition remains enigmatic since sedimentation is likely to occur both during advance and retreat of the ice margin. Near ice sheet margins, where similar polyphase sequences of bed conditions occur, increasing debris and meltwater volumes develop progressively, creat-

193

ing widespread examples of bedforms and thicker sediment piles. Also near the ice front, where ice is thinner, topographic controls begin to influence ice flow and meltwater directions, often allowing radial deposition of various subglacial bedforms (e.g., drumlins in Wisconsin, USA). Not all radial flow patterns are indicative of ice marginal positions, for example, eskers and drumlins in parts of northern Manitoba, Ungava, Quebec, and the Northwest Territories, Canada, may represent basal ice motion away from ice flow centres and divides. Finally, where ice margins float, basal ice dynamic conditions alter down-ice of the groundingline. Changes in subglacial stresses are transmitted back up-ice from the grounding-line for some distance, thereby affecting bed conditions and influencing subglacial processes. It is apparent from the above discussion that similar spatial down-ice distributions cannot be expected within valley glaciers. In general, because of the short distances of transport, as compared with ice sheets, basal ice thermal zonation patterns are unlikely to develop. Within the middle and lower reaches of valley glaciers, polyphase conditions may prevail with ‘H’ bed states dominant. Occasionally, where large volumes of debris have been derived or overrun, deformable ‘M’ beds may occur. Also, unlike large ice sheets, the impact of bedrock topographic and geological control is probably far greater.

8.5. SUBGLACIAL SEDIMENTS AND LANDFORMS Sediments and landforms of subglacial environments can be subdivided into those developed under (1) active ice flow (advance and active retreat phases); and (2) passive or ‘dying’ ice flow (retreat phase). In each group the origin of the forms will be discussed as will the sediments and their dominant characteristics. In seeking diagnostic criteria, rarely can a single or a few characteristics be used when studying sediments in the field. Rather, it is a combination of properties, stratigraphy and lithofacies associations that aid in the identification of a specific sediment type and formative process(es).

194

SUBGLACIAL ENVIRONMENTS

8.5.1. Sediments and Landforms of Active Ice Flow The processes and mechanisms that lead to sediment deposition beneath an active land-based ice sheet can be subdivided into those processes producing: (a) in situ tills (diamictons), (b) glaciofluvial and glaciolacustrine sediments, and (c) glacial diamicton m´elange units indicative of syndepositional deformation. 8.5.1.1. Tills – diamictons Till (often termed diamict or diamicton today, to avoid the genetic connotation that the term ‘till’ carries) has been defined by Dreimanis and Lundqvist (1984, p. 9) as ‘. . . a sediment that has been transported and is subsequently deposited by or from glacier ice, with little or no sorting by water’. Tills exhibit a wide

range of grain size and may be very poorly to wellsorted (Plate 8.1). Typically, tills contain a variable percentage of exotic clasts and mineral grains that have been transported considerable distances from up-ice sources. In general, however, the majority of the clasts and finer particles are of local origin (<15 km up-ice). Tills may form within several distinct glacial subenvironments such as subglacial, englacial, proglacial and supraglacial; and may be deposited from active temperate wet-based and polar cold-based ice masses, as well as under stagnant ice conditions. Hicock (1990) has developed a graphic means (‘Till Prism’) of illustrating the complexity and interrelationships that exist within till-forming subglacial environments (Fig. 7.2). The till prism portrays the multiple pathways and associated subglacial environments through which sediment may pass before final

PLATE 8.1. Large exposure (>30 m in height) of diamicton in stoss end of a drumlin, Chimney Bluffs State Park, New York State. Note crude banding in centre of photograph and marked diamicton colour change near upper part of exposure.

SUBGLACIAL ENVIRONMENTS

deposition. The ‘prism’ demonstrates the intrinsic complexity of till formation that has, in the past, been oversimplified. Although disregarded because of their comparative rarity, tills from polar ice bed conditions can develop under two states, sublimation and meltout. In the former instance it has been shown that under dry polar climatic conditions (e.g., the Dry Valleys of West Antarctica) direct sublimation of debris-charged ice can result in the formation of sublimation till that is a ‘sub-species’ of meltout till. Where polar ice stagnates, direct meltout tills may also form. In either instance a limited input from meltwater can be expected and thus stratification and/or sorting will be restricted. Under wet-based, active temperate conditions considerable input from meltwater can be expected, as well as thermal fluctuations, and transient fluctuations in ice velocity and stress conditions. Subglacial temperatures are likely to be in the –1° to –3°C range, with average basal ice stress levels of ~100 kPa. Tills formed under active temperate conditions are dominantly recognized as Lodgement and Basal Meltout Tills with a variety of local till sub-types and subglacial M´elange Diamictons. Sub-types of tills are often identified owing to some peculiar characteristic of grain size, provenance and/or structural attribute. Both lodgement and meltout tills have been extensively described in the literature (Shaw, 1987, Dreimanis, 1988; Brodzikowski and Van Loon, 1991), whilst subglacial m´elanges have only recently been regarded as separate and distinct subglacial sediments (Menzies, 1989a; Hicock, 1992; Curry et al., 1994; Hoffman and Piotrowski, 2001). Lodgement tills. Lodgement tills are formed subglacially under pressure melting conditions. As ice comes in contact with its bed and frictional heating occurs ice-encased debris is released into the contact zone between the ice and the bed by a process of regelation. This debris is released at a rate of only a few centimetres per year and is smeared along the ice–bed contact where it is lodged at the bed. Boulton (1974) suggested that under increasing effective pressures, depending upon basal ice velocity and ice thickness, a ‘critical lodgement index (Lc )’ can be found that, once reached, leads to increasing lodge-

195

ment of debris at the ice–bed interface, i.e., as friction increases so lodgement increases. The index is an empirical relationship where: V 1r ≈

冬L 冭 N

i/m

(8.3)

c

where V 1r is the relative forward velocity of a particle in traction at the ice–bed interface; N is the effective stress and m is an empirical constant of ~0.3. Boulton found values of Lc in the range of 5–25. However, some debate as to the validity of this concept has since arisen (Hallet, 1981; Drewry, 1986). Lodgement tills form widespread continuous till plains often associated with bedforms (drumlins, Rogen moraine, fluted moraine) These plains may be of considerable thickness (≥30 m) and cover large areas of terrain. Typical characteristics of lodgement tills are noted in Table 8.2. Structural discontinuities are common within lodgement tills. (It should be pointed out that, although many of these structures have been described as occurring in lodgement tills, the likelihood is that many of these tills have formed under different sedimentological processes and are not lodgement tills.) Many lodgement tills display distinctive fissile structures (Dreimanis, 1993; Kr¨uger, 1994). Fissility, typically, appears as sub-horizontal thin partings within fine-grained till, that are interpreted as being indicative of stress application during deposition (autokinetic discontinuity). The fissile structures may be weak plate-like forms or may appear as bedding units in the form of laminations. These laminations are not regular and often pass around boulders. Other structures are sub-horizontal foliations of similar origin to fissures; and macro- and micro-shear joints (Plate 8.2). Occasionally, bedding planes and other structures have been thought to be caused by dewatering and unloading. A characteristic of all tills is the presence of a discernable clast fabric (Fig. 8.7). Clast orientation applies to both macro- and micro-clasts. In lodged tills a strong fabric orientation is present indicative of autokinetic stress application during deposition (Fig. 8.8). It is thought that as the fine-grained matrix of the till is being lodged, large particles within the mobile sediment or at the sediment/ice bed interface are

196

SUBGLACIAL ENVIRONMENTS

TABLE 8.2. Criteria for identification of lodgement till, melt-out till and gravity flowtill (after Dreimanis, 1988) Criterion

Lodgement till

Melt-out till

Gravity flowtill

Position and sequence in relation to other glacigenic sediments

Under advancing glaciers: lodged over older pre-advance sediments and over glacitectonites, unless they have been eroded. Under retreating glaciers: the lower-most depositional unit, if the deposits related to glacial advance have been eroded. Locally underlain by meltwater channel deposits. May be overlain by any glacigenic sediments.

Usually deposited during glacial retreat over any glacially eroded substratum or over lodgement till. Also as lenses in lodgement till. May be interbedded with lenses of englacial meltwater deposits, and locally is underlain by syndepositional subglacial meltwater sediments and subglacial flowtill.

Most commonly it is the uppermost glacial sediment in a non-aquatic facies association. Associated also locally with subglacial tills, where cavities were present under glacier ice, or where the glacier had over-ridden the icemarginal flowtill. May be interbedded or interdigitated with glaciofluvial, glaciolacustrine or glaciomarine sediments, particularly away from its original source at the glacier ice.

Basal contact

Since both lodgement and melt-out tills begin their formation and deposition at the glacier sole, their basal contact with the substratum (bedrock or unconsolidated sediments) is similar in large scale, being usually erosional and sharp. The glacial erosion-marks underneath the contact and the alignment of clasts immediately above the contact have the same orientation. Glacitectonic deformation structures formed by the fill-depositing glacier may occur under both tills, and they strike transverse to the direction of local glacial stress.

Variable; seldom planar over longer distances. The flows may fill shallow channels or depressions. The contact may be either concordant, or erosional, with sole marks parallel to the local direction of sediment flow. Loading structures may be present at the basal contact of waterlain flowtill and the underlying soft sediment.

Basal contact, representing the sliding base of the glacier, is generally planar if over unconsolidated substratum, but it may be grooved. The bedrock contact is usually abraded, particularly on stoss sides of bedrock protrusions. Since the sliding base of the glacier represents a large shear plane, sheared and strongly attenuated substratum material may be deposited as a thin layer along this plane, and from place to place it is sheared up into the lodgement till. Clast pavements, both erosional and depositional, may be present along the basal contact, but they occur also higher up in lodgement till. If lodgement till becomes deformed by glacial drag shortly after its deposition, the basal contact may become involved in the deformation with tight recumbent folding, overthrusting and shearing.

If the basal contact of glacier ice was tight with the substratum during the melting, the predepositional erosional marks characteristic for moving glaciers are as wellpreserved as under lodgement till. However, subsole meltwater may modify the basal contact locally, and produce convex-up channel fills and various other meltwater scour features.

SUBGLACIAL ENVIRONMENTS

197

Table 8.2 Continued Criterion

Lodgement till

Melt-out till

Gravity flowtill

Surface expression, landforms

Mainly in ground moraines and other subglacial landforms. Also along the proximal side of some end moraines are always associated with lodgement till.

In those ice-marginal landforms where glacier ice had stagnated.

Associated with most ice marginal landforms. Also, as a thin surface layer on many other direct glacial landforms.

Thickness

Typically one to a few metres; relatively constant laterally over long distances.

Single units are usually a few centimetres to a few metres thick, but they may be stacked to much greater thicknesses.

Very variable. Individual flows are usually a few decimetres to metres thick, but they may locally stack up to many metres, particularly in proglacial ice marginal moraines and some lateral moraines.

Structure, folding, faulting

Typically, described as massive but, on closer examination, a variety of consistently oriented macro- and microstructures indicative of shear or thrust may be found. Folds are overturned, with anticlines attenuated downglacier. Deformation structures are particularly noticeable, if underlying sediments are involved, or incorporated in the till, developing smudges. Subhorizontal jointing or fissility is common. Vertical joint systems, bisected by the stress direction, and transverse joints steeply dipping down-glacier, may be formed by the glacier deforming its own lodgement till. The orientation of all the deformation structures is related to the stress applied by the moving glacier, and therefore it is laterally consistent for some distance.

Either massive, or with palimpsest structures partially preserved from debris stratification in basal debris-rich ice. Lenses, clasts and pods of texturally different material preserve best, for instance softsediment inclusions of various sizes and englacial channel-fills. Loss of volume with melting leads to the draping of sorted sediments over large clasts. Most large rafts or floes of substratum are associated with melt-out tills, and they may be deformed by glacial transport and by differential settlement during the melting.

The structure depends upon the type of flow and associated other mass movements, the water content and the position in the flow. Either massive, or displaying a variety of flow structures, such as: (a) overturned folds with flat-lying isoclinal anticlines, (b) slump folds or flow lobes with their base usually sloping downflow, (c) rollup structures, (d) stretched-out silt and sand clasts, (e) intraformationally sheared lenses of sediments incorporated from substratum, with their upper downflow end attenuated, if consisting of finegrained material, or banana shaped.

Grain size composition

Usually a diamicton, containing clasts of various sizes. Grain-size composition depends greatly upon the lithology and grain-size composition of the substrata up-glacier and the distance and mode of transport (basal, englacial) from there. Comminution during glacial transport and lodgement has produced a multimodal particle size distribution. Most resulting subglacial tills are poorly to very poorly sorted (σ = 2–5), described also as well graded and their skewness has a nearly symmetrical distribution (Sk = –0.2 to 0.2), except for those tills that are rich in incorporated presorted materials.

Usually a diamicton with polymodal particles size distribution. It is texturally similar to that primary till to which it is related, but with a greater variability in grain size composition owing to washing out of, or enrichment in, fines or incorporation of soft substratum sediments during the flow. Some particle size redistribution takes place during the flow. The grain size composition depends greatly upon the type of flow, and the position or zone in it. Sorting, inverse or normal grading may develop in some zones of flows, and parts of clasts may sink to the base of flow.

198

SUBGLACIAL ENVIRONMENTS

Table 8.2 Continued Criterion

Lithology of clasts and matrix

Lodgement till

Melt-out till

The abrasion in the zone of traction during lodgement produces particularly silt-sized particles typical for lodgement tills. Most lodgement tills have a relatively consistent grain-size composition, traceable laterally for kilometres, except for the lower 0.5–1 m that strongly reflects the local material. Clusters or pavement of clasts are common.

The winnowing of silt- and clay-size particles in the voids during the melt-out may reduce the abundance of these particle sizes in comparison with their lodged equivalents. Some particle size variability is inherited from texturally different debris bands in ice. Extreme variations in grain size may occur over short distances in the vicinity of large rafts and other inclusions of soft sediment.

Lithologic composition tends to be less variable than in other genetic varieties of tills; most constant is the mineralogic and geochemical composition of the till matrix. Materials of local derivation increase in abundance towards the basal contact of the tills with substratum.

Since glacial debris of distant derivation is more common in the englacial zone than in the basal zone of a glacier and since the englacial zone has a greater possibility to be deposited as melt-out till, rather than by lodgement, materials of distant derivation may be more abundant in the melt-out till than in the lodgement till of the same till unit, particularly in supraglacial melt-out till. Great compositional variability occurs in the vicinity of incorporated ‘megaclasts’, ‘rafts’ or ‘floes’ of sub-till material. Soft sediment clasts, for instance consisting of sand, may be found in melt-out till, but not in typical undeformed lodgement till.

Gravity flowtill

The lithologic composition is generally the same as that of the source material of flowtill – a primary till or glacial debris, plus some substratum material incorporated during the flowage. Material of distant derivation dominates in the flowtills derived from supraglacial and englacial debris, but dominance of local material indicates derivation from basal debris. Soft sediment clasts derived from the substratum, or from sediment interbeds in multiple flows, are common.

SUBGLACIAL ENVIRONMENTS

199

Table 8.2 Continued Criterion

Lodgement till

Clast shapes and their surface marks

Following criteria apply to lodgement till and basal melt-out till where most clasts are derived from a single-cycle transport: subangular to subrounded shapes dominate, depending mainly upon the distance of transport in the basal zone of traction. Bullet-shaped (‘flat-iron’, ‘elongate pentagonal’) clasts are more common than in other tills and non-glacial deposits, and their tapered ends usually point upglacier. Some of the elongate clasts have a keel at their base. Glacial striae are visible mainly on medium-hard fine-grained rock surfaces. Elongate clasts are striated mainly parallel to their long axes, unless they have been lodged or transported by rolling.

The bull-shaped and faceted clasts, also crushed, sheared and stressedout clasts are more common in lodgement till than in other tills. Lodged clasts are striated parallel to the direction of the lodging glacial movement, and they have impact marks on both the upper and lower surfaces, but in opposite orientation; on the surface the stoss end is upglacier, but on the underside the stoss end is downglacier. Clast pavements with sets of striae parallel to the direction of the latest glacial movement over them may occur at several lodgement levels. Their top facets are either parallel with the general plane of lodgement, or they dip upglacier. Fabrics: macrofabric (orientation of clasts) or micro-fabric (orientation of particles in the matrix)

Melt-out till

Gravity flowtill If present, soft sediment clasts are either rounded or deformed by shear or dewatering. The more resistant rock clasts are in the same shape as they were in the source material when resedimented by the flowage. Therefore, the relative abundance of glacially abraded subangular to subrounded clasts versus completely angular clasts in flowtills of mountain glaciers will indicate the approximate participation of basal debris versus supraglacial debris in the formation of the flowtill. Some rounded waterreworked clasts, without striations, may derive in flowtills from melt-water stream deposits.

If, in an area of mountain glaciation, the source of supraglacial melt-out till is englacially or even supraglacially transported, supraglacially derived debris, then the clasts are angular. Most commonly, supraglacial melt-out till in such areas also contains an admixture of glacially abraded basal debris, also englacially transported.

Strong macro-fabrics with the long axes parallel to the local direction of glacial movement are reported from diamictons identified either as lodgement or melt-out tills. Occasionally transverse maxims have developed, associated with folding and shearing. The fabric strength may vary also, depending upon the grain-size of till, the abundance of clasts and postdepositional modification.

Variable, and depending greatly upon the type of flow and the position in the flow. It may range from randomly oriented to strong fabric, in thin flow tills. Fabric maxima are either parallel or transverse to the local flow direction, unrelated to glacial movement; the a–b planes are either subparallel to the base of flow, or they dip upflow. Fabric maxima may also differ laterally on short distances.

200

SUBGLACIAL ENVIRONMENTS

Table 8.2 Continued Criterion

Consolidation, permeability, density

Lodgement till

Melt-out till

The lodgement till fabric may be of complex origin: produced by lodgement or by deformation of the already deposited dilated till, under the same glacier. If both stress directions coincide, a strong fabric will develop; if not, the lodgement fabric becomes weakened. Typically, the a–b planes dip slightly upglacier, if lodgement alone is involved. The micro-fabric is usually as strong as the macro-fabric.

In melt-out tills, fabric is inherited from glacier transport, where fabric dominates parallel to the direction of glacial movement, unless deformation changes it to transverse fabric locally. However the melting-out process may weaken the fabric, particularly the micro-fabric. Also, the dip of the inclination of clasts becomes reduced by the reduction of the volume of ice during melting.

Most lodgement tills, particularly the poorly melted-out tills are usually sorted matrix-supported varieties, are over-consolidated, provided there was adequate subglacial drainage. Their bulk density, penetration resistance and seismic velocity are usually high, permeability is low, relative to other varieties of till of the region.

Supraglacially formed melt-out tills are usually less (normally to weakly) consolidated than the subglacially formed, commonly overconsolidated melt-out tills, provided there was adequate drainage of meltwater. Bulk density and penetration resistance may be lower and more variable than in related lodgement till. Also, permeability is more variable.

oriented with their long axes in the principal stress (typically, in the down-ice) direction. Since this involves clast ploughing during the lodging process (Alley, 1989a,b), most clasts also exhibit a slight upice dip (Hicock et al., 1996; Kjaer and Kr¨uger, 1998). Lodgement tills, typically, contain exotic sediments apparently rafted into or deposited within the lodged till (Ruszczynska-Szenajch, 1987; Eyles et al., 1982). These inclusions often have surrounding joint sets, fractures and clay injection features. The size of these inclusions vary enormously and may be composed of fluvioglacial sands and gravels, laminated lacustrine clays and silts, re-sedimented tills, boulder beds and pavements, and weathered bedrock (Plate 8.3).

Gravity flowtill

Primarily normally consolidated and relatively permeable. If clayey, may become over-consolidated owing to postdepositional desiccation. Density lower than in primary tills.

Lodgement tills, if fine-grained, are usually overconsolidated, massive in appearance, have low permeability and rarely display large-scale fracture geometries. In terms of micromorphology, however, lodgement tills tend to exhibit a marked degree of foliation in thin section, a distinct unistrial fabric and often ‘card-house’ structures within massive clay units (van der Meer, 1996, 1997). Meltout tills. Meltout tills are formed as a result of massive in situ melting and subsequent deposition from the upper (supraglacial) and lower (subglacial) englacial zones of an ice mass (Haldorsen and Shaw, 1982; Sharpe and Barnett, 1985). The likely importance of meltout tills in the stratigraphic record is now

SUBGLACIAL ENVIRONMENTS

(a)

(b)

(c)

(d)

201

PLATE 8.2. (a) Brecciation in diamicton immediately above a sand intraclast. Note scale card is 8.5 cm long. (b) Shear planes within a diamicton, note two major planes in the centre and top left of the photograph. Several smaller planes intersect these large planes. Scale width of photograph is 45 cm. (c) ‘Bed limits’ within a diamicton. Scale bar of 13.0 cm. (d) Distinct lamination within a diamicton with several small faults and a shear plane in bottom right of photograph. Note knife for scale, 9.0 cm. All photos from Mohawk Bay, southern Ontario.

202

SUBGLACIAL ENVIRONMENTS

Site A

5

S 1 = 0.8668

1

S 1 = 0.4660 Diamicton

Sand Intraclast

Slump

7

S 1 = 0.7011

0

metre

3

1

S 1 = 0.9406

Site B

6

S 1 = 0.8417

4

Diamicton

S 1 = 0.7914

Sand Intraclast

Slump

8

S 1 = 0.7729

2

S 1 = 0.5882

0

metre

1

FIG. 8.7. Schmidt equal-area projections of clast fabric data at four sites around each of two sand intraclasts within a glacial m´elange diamicton, Mohawk Bay, Ontario. (After Menzies, 1990c).

SUBGLACIAL ENVIRONMENTS

203

0.30 Legend

WATERLAIN GLACIGENIC SEDIMENT

Melt-out till

0.25

Undeformed lodgement till Fossiliferous diamicton

0.20 EIGENVALUE S3

Glacigenic sediment flow

*

Ice slope colluvium

*

Deformed lodgement till Lodgement till, Hertford

*

0.15

1

0.10

*

* 2

"Slumped till", Hatfield Svalbard tillite Mohawk diamicton

GLACIGENIC SEDIMENT FLOWS

* *

*

*

8

0.05

6

5

DEBRIS-RICH BASAL ICE

LODGEMENT TILL MELT-OUT TILL

7

0.00 0.4

0.5

0.6

0.7

3

4 0.8

0.9

1.0

EIGENVALUE S1 FIG. 8.8. Plot of S1/S3 eigenvalues for clast fabric data (data from Dowdeswell et al., 1985; Dowdeswell and Sharp, 1986 and Menzies, 1990c).

under some doubt and their ‘once thought’ ubiquity questionable (Rappol, 1987; Paul and Eyles, 1990). Subglacial meltout till formation is thought to occur as a result of basal ice stagnation and subsequent thawing of the encased englacial debris down onto the glacier bed (Shaw, 1982). The process of meltout till formation would seem to occur under both active and passive ice conditions. Meltout tills are saturated and, therefore, subject to syn- and postdepositional processes of disaggregation and resedimentation. The process of meltout is normally visualized as a grain by grain sedimentation process that, owing to large volumes of meltwater, results in considerable winnowing and loss of fines. Local pondings, channel and pipe flow conditions occur and thus stratified sands, gravels and laminated clays and silts are often part of an entire meltout till package. The process of meltout till formation is sufficiently

long to allow porewaters to dissipate and stress adjustments be compensated for within the structural framework of the sediment. In some instances, the thaw process may last for several hundreds of years and be accompanied continuously by thaw consolidation and other stress adjustment effects. Meltout resulting from sublimation can occur under arid polar conditions. Meltout tills are not found extensively over wide areas of glaciated terrains principally because of their poor preservation potential after initial deposition, and because of the restricted spatial distribution of thick englacial sediment packages from which meltout tills might form (Rappol, 1987). Meltout tills are usually relatively thin, coarse-grained, occurring as spatially disjunct patches within complex marginal stratigraphic packages. The typical characteristics of meltout tills are noted in Table 8.2.

204

SUBGLACIAL ENVIRONMENTS

(a)

(b) PLATE 8.3. (a) Bottom sharp contact between a sand intraclast and surrounding diamicton. Scale bar of 15 cm. (b) Sand intraclast within Cromer Till, West Runton, East Anglia, UK. Note distinct lamination within diamicton, and flow structures around the intraclast and behind shovel handle. Shovel handle approximately 12.5 cm wide (photo courtesy of Phil Gibbard).

SUBGLACIAL ENVIRONMENTS

205

PLATE 8.3. (c) Contorted sand intraclasts within transition between laminated clays and Port Stanley Till, Bradtville, southern Ontario. Knife is approximately 25 cm long (photograph courtesy of Phil Gibbard).

The most pertinent aspects of meltout tills recognizable in the field are: (a) the presence of stratified units, (b) large numbers of faults, fractures and slump structures; and (c) englacially inherited structures (Fig. 8.8). Discontinuities are common within meltout tills as a result of thaw consolidation, mass movement, loading and dewatering. Discrimination between meltout till deposition and glaciotectonic allokinetic disturbance is difficult. In comparison with lodgement tills, meltout tills probably have a higher percentage of near-vertical discontinuities that may extend across different sub-units within the sediment package (>1–2 m). The presence of large, often subhorizontal, stratified sand lenses and other laterally extensive units allows for a greater degree of susceptibility to post-depositional dewatering, cryogenic effects and other deformation processes. Diapirism, brecciation and other interstratal disruptive processes are likely to be more pervasive in meltout tills than in lodgement tills. A characteristic, used in the past, to differentiate meltout tills from other till types is their strongly

preferred clast fabric, often retained from prior englacial encasement (Lawson, 1979b; Dowdeswell and Sharp, 1986). Recent discussions, however, question the problem of fabric replication in other subglacial environments (Fig. 8.8); and the validity of this diagnostic criterion (Menzies, 1990c). Likewise, the presence of undeformed inclusions within meltout tills when used as a diagnostic characteristic is suspect. 8.5.1.2. Subglacial glaciofluvial and glaciolacustrine sediments Many tills contain intraclasts indicative, in numerous instances, of subglacial stratified sediment deposition. Today, it is commonly recognized that within subglacial cavities, glacier bed depressions, channels and tunnels a considerable volume of stratified sediments may be deposited, ranging from delicately laminated lacustrine clays to coarse bouldery gravels (Plate 8.4). These sediments, deposited within fluvial environments, exhibit evidence of enormous short-term

206

SUBGLACIAL ENVIRONMENTS

(a)

PLATE 8.4. (a) Inclusion of stratified sands within a diamicton, Port Burwell, southern Ontario. Note coin for scale is 2.5 cm in diameter. (b) Stratified sands containing angular clay fragments and some cross-bedding, near Hanover, southern Ontario. Note change in sand toward top and contact with a diamicton. Scale bar of 15 cm. (b)

fluctuations in meltwater discharge, velocity and depth (Sharp et al., 1998). This variability in meltwater stream competency is manifest in a wide range of sedimentary facies and internal structures, and grain size distributions. Sedimentation in the subglacial environment is usually rapid (within hours), causing these sediments to exhibit rapid porewater expulsion and migration features, as well as load and shear structures owing to overriding ice. Subglacial glaciofluvial sediments tend to be areally restricted; occurring as thin patches or linear ribbons (eskers) parallel or subparallel to ice direction. Subglacial and proglacial stratified sediment can be segregated on the basis that the latter are tabular, widespread, thick (>15 m) sediment packages that may have pitted surfaces pock-marked by kettles. Subglacial stratified sediments commonly reach

thicknesses <15 m, and are areally limited in distribution. Discontinuities are often found in stratified sediments within lodgement tills (Table 8.3). In particular, owing to overriding by ice, stratified sediments may exhibit evidence of liquefaction, fluidization, faulting and shearing. The lack of sediment cohesion usually precludes folding but post-depositional cryostatic conditions may support folding and buckling. In some cases these sediments may be ‘rafted’ into glacial m´elanges (Plate 8.5). Within subglacial and proglacial stratified sediments, inclusions are common in the form of till balls, injections from overlying sediments and underlying diapirs, clay shards, and other exotic sediments that may have fallen into subglacial meltwater streams and cavities. Where dense clay sediment overlying saturated stratified sediment have

TABLE 8.3. The types and origins of discontinuities within glacigenic sediments KINETIC ENERGY SOURCES

ENVIRONMENT

DISCONTINUITIES UNSHEARED

DISCONTINUITIES SHEARED

SUBAQUATIC

SUBGLACIAL

PROCESS AGENCY

PROCESS

ALLOKINETIC (EXTRINSIC)

AUTOKINETIC (INTRINSIC)

LOADING/ UNLOADING

ENDOGENIC

EXOGENIC

GLACIAL DEPOSITION

GLACIAL EROSION

GLACIOTECTONISM

SUBAERIAL

TECTONISM

WEATHERING

DEWATERING

SYNERESIS

FRACTURING

CRYOGENIC

SEISMIC

CHEMICAL

DESICCATION

BEDDING DISTURBANCE

BRECCIA

WEDGING

FISSURES

BEDDING PLANES

BEDDING PLANE SHEARS

FAULTS

SLICKENSIDES

208

(a)

SUBGLACIAL ENVIRONMENTS

(b)

(c) PLATE 8.5. (a) Ductile banding in diamicton m´elange, Mohawk Bay, southern Ontario. Photo is approximately 1.5 m across. (b) Similar to (a) but containing a large number of angular sand units. Scale card is 8.5 cm long. (c) Faulted and laminated ductile m´elange sediments. Each bar scale segment is 5 cm.

SUBGLACIAL ENVIRONMENTS

suffered rapid loading, dish and pillar structures and, in some cases, complete ball and pillow structures may develop. Subglacial glaciolacustrine sediments are often found in the lee-side of large topographic obstructions, boulders or on the floor of cavities. The presence of small units of glaciolacustrine sediment, within other subglacial sediments, is common. 8.5.1.3. Glacial diamicton m´elange sediments In the past two decades there has been increasing evidence that major ice sheets may have been underlain by areas of deformable debris layers or beds (Chapter 4). These deformable beds or debris traction layers, beneath ice masses, essentially act as a lubricant, resulting in high basal ice velocities and associated thin marginal ice surface profiles. The emplacement (deposition) of deforming bed sediment is similar to lodgement in that deposition occurs when the level of applied strain to the sediment falls below its yield strength. Therefore, deposition is likely incremental, occurring, at varying times, randomly across different parts of the glacier bed. The deposit resembles a m´elange (a sediment composed of a wide variety of sub-units of differing sediment types, often from various distinctive facies environments although usually from the same sedimentary basin; Cowan, 1982, 1985). The term ‘m´elange’ implies that deformation or massive ductility has occurred in the formative process. Cowan (1985) distinguished four m´elange types of which Type III described as ‘Block-in-Matrix’ perhaps best typifies glacial diamicton m´elanges (Table 8.4).

Diamicton m´elanges are, perhaps, much more widespread than hitherto recognized. Extensive areas of thick diamicton, previously considered lodgement till, may in fact be m´elange diamictons (Hicock, 1992). Specific subglacial bedforms and sediment suites may be part of m´elange units. As in the case of lodgement tills, the intrinsic variability of m´elanges almost defies any general description (Table 8.2) of sedimentological and geotechnical characteristics. It must be pointed out that many characteristics of matrix, lithology, provenance, inclusions and clast fabric are similar to those in lodgement and meltout tills (Fig. 8.2) (Hicock, 1990). Discontinuities are common and a distinctive characteristic of diamicton m´elanges (Menzies and Woodward, 1993). The type of discontinuity may vary depending upon the nature of the particular sediment within the m´elange (Maltman, 1988; van der Meer, 1997; Menzies et al., 1997). There is increasing evidence that no distinctive clast fabric pattern should be expected for these m´elange diamictons (Rappol, 1985; Menzies, 1990c; Hart, 1994). Clast fabrics often reveal a wide array of orientations consistent with massive deformation. Secondary fabric replication owing to deformation may occur in many cases while in others primary fabrics are preserved (Fig. 8.8). Thus discrimination of these fabric artefacts would be almost impossible. Inclusions within subglacial m´elanges may consist of any sediment entrained by the mobile debris layer beneath the ice mass and, therefore, could be both glacial or non-glacial in primary origin. Some inclusions may be entrained and immediately destroyed or

TABLE 8.4. Concept of m´elange (after Cowan, 1985) Definition TYPE I TYPE II TYPE III TYPE IV

209

Fragments enveloped within a fine-grained matrix; typically of obscure stratigraphy, stratal disruption and/or chaotic ‘block-in-matrix’ fabric Stratified sequence of sands/gravels and fine-grained clays/silts. Origin probably a result of disruption/fragmentation caused by layer parallel extension Analagous to TYPE I with progressively disrupted sequences but of a more extreme form of extension deformation ‘Block-in-matrix’ fine-grained ‘polymict’. Exotic inclusions are typical. Formation is multi-stage by progressive highly ductile (en masse) deformation Analagous to TYPE III but typically inclusions bounded by systems of subparallel faults. Dominant deformation process is cataclasis (structural slicing) within a brittle fracture zone

210

SUBGLACIAL ENVIRONMENTS

drastically altered while others remain, in various states of preservation, while others may be entrained in a frozen state thawing later. 8.6. SEDIMENT STRUCTURES AND RELATED SEDIMENTOLOGICAL AND GEOTECHNICAL CHARACTERISTICS WITHIN SUBGLACIAL SEDIMENTS Sediment structures are geometric or internal architectural arrangements indicative of sediment depositional and deformational processes (Allen, 1982; Tucker, 1986; Collinson and Thompson, 1982). These structures are delimited by distinct sets of discontinuities, the lineation, orientation and precise geometry of which can also be used in recognizing sedimentation processes in the subglacial environment (Dreimanis, 1993). 8.6.1. Structural Discontinuities Structures within subglacial sediments are the consequence of in situ contemporaneous development (Autokinetic or Intrinsic); the result of post-depositional effects (Allokinetic or Extrinsic); or a compound of both intrinsic and extrinsic forms (Compound). Structures may be bedding planes, faults, joints, cracks, cleavages, fractures, fissures, ice wedges and breccia (Plate 8.6). Discontinuities may also be zones or areas of less competence. These zones may develop during or immediately following deposition (e.g., slickensides, clay eluviation, liquefaction, fluidization, boudinage effects and other ‘tectonic’ events, and dewatering processes leading to pipe channels, flame structures and ball and pillow structures (Table 8.3). Discontinuities are structural attributes of subglacial sediments whose presence may grossly affect the bulk geotechnical nature and response of the sediment under stress. The degree of their development appears to be a function of the intensity and magnitude of the processes that generated them. Structures may be geometrically interconnected in various dendritic branching systems, or discrete and autonomous, influencing large or only localized areas of sediment (for a discussion on types and forms of discontinuities in subglacial sediments see Menzies and Shilts, 1996, pp. 46–54).

8.7. SUBGLACIAL–PROGLACIAL TRANSITION ENVIRONMENTS The boundary between subglacial and proglacial environments marks the site of greatest localized deposition within the glacial system (Fig. 8.1(a),(b)). At this transition zone, large volumes of sediment derived and transported from within the subglacial system enter new sedimentological and glaciodynamic environments. 8.7.1. The Subglacial and Terrestrial Proglacial Environment Subglacial debris is transported toward the margins of ice masses either along the ice–bed interface or by being sheared upward to become englacial debris close to the margin and eventually released as supraglacial meltout debris (Fig. 8.9). The mechanism by which subglacial debris is sheared up into basal ice close to ice margins remains controversial. Debris transported to the ice front via the subglacial interface enters the proglacial system and may, at the ice margin, be squeezed and/or pushed by the ice mass (Chapter 14). In both instances, moraines may develop to a size concordant with the volumes of debris and ice, stress levels, the rheology of the debris and the length of time the ice front is stationary at a particular location (Bennett and Boulton, 1993) (Fig. 8.10). Sediments within these morainic forms are composed of a wide range of individual lithofacies types, grain sizes and provenances. Where subglacial debris has been transported englacially and ablated as supraglacial debris, flow tills develop as the debris moves off the ice under low effective stress levels. Lawson (1981a,b), working at the edge of the Matanuska Glacier, Alaska, demonstrated that ~95 per cent of the sediments in the snout area of the glacier have been influenced by, or are directly the result of debris flow and resedimentation (Table 8.5) (Menzies, 1995, chapter 11). 8.7.1.1. Flow tills Flow tills are found in several differing glacial environments, for example, within subglacial cavities, on the surface of ice masses and, dominantly, at the

SUBGLACIAL ENVIRONMENTS

211

(a)

(b)

(c) PLATE 8.6. (a) Photomicrograph of thin section of m´elange sediment sand intraclast within a diamicton. Note banding, diffusion layering, micro-faulting, clay clots and balls. (b) Stratified and faulted sand intraclast with clay intrusion from below. Note banding within clay, and presence of clay balls. (c) Heavily deformed sand intraclast containing large clay content. Effect of clay is to cause the sediment to have a ‘marbled’ appearance. Note numerous small sub-horizontal shear planes. Large oblique shear plane at bottom right of micrograph.

SUBGLACIAL ENVIRONMENTS

STAGES OF DEVELOPMENT

a

ice front

A

ice

time

212

2 1

B

1+ 2

a

distance

b

C

3

1 push moraine

D

2

3

time

b 2 1

distance

FIG. 8.10. Diagram illustrating the formation of end moraines. Diagram to the left illustrates a plan view of moraines and ice front. To the right, the two graphs chart the position of the ice front through time following advances, still-stands and retreat (modified from Bennett and Boulton, 1993).

E

F

G

H

I ? 0

100 metres

Undeformed and deformed superimposed ice Debris-bearing ice Vertical exaggeration = 2X

FIG. 8.9. Model of evolutionary stages of an ice mass margin with progressive development of a debris-covered, ice-cored moraine (after Hooke, 1973, reproduced by permission of the author).

ice front (Dreimanis, 1988). Flow tills moving under very low gradients (~2°) can spread over extensive proglacial areas. Within the proglacial environment complex stratigraphies can develop with large volumes of buried glacier ice being overlain by these tills. Flow tills are generally formed as wedge style deposits and are often stacked one flow on another. Thicknesses of stacked flow tills have been reported in excess of 10 m. Typically, stacked flow tills occur with each layer laterally offset by several metres in an on-lap–off-lap sequence (Ing´olfsson, 1988; Owen and Derbyshire, 1988). Flow till units are usually composed of a central unsorted body or plug surrounded by heavily sheared and sorted outer units (Plate 8.7). Structures associated with flow tills are complex, exhibiting a wide range of folds, faults, riedel shears and kink band arrays. Flow till clast fabrics tend to exhibit strong, unidirectional orientations (Fig. 8.8). However, similar fabrics have also been measured in sediments from other environments negating the value of clast fabric, alone, as a potential identifying characteristic. Flow tills may contain clasts from many sources. Some are angular frost-riven whose source is supraglacial (in valley glaciers and ice sheets where

SUBGLACIAL ENVIRONMENTS

nunataks exist), others are heavily comminuted, striated, sub-angular to sub-rounded clasts of englacial and subglacial derivation. Frequently, small inclusions and rafted units of sediment occur within flow tills especially where flows have crossed weaker subjacent sediments. These units may be strongly contorted but often are intact.

213

8.7.2. The Subglacial and Subaquatic Proglacial Environments As ice sheets enter large bodies of water, the ice margin may float (Fig. 8.1(b)). Ice masses with a floating tongue whether as a free-floating ice front (tidewater) or contiguous with an ice shelf become

TABLE 8.5. (a) Principal characteristics of Lawson-type sediment flows (after Lawson, 1979a) Lawson flow type I

II

III

IV

Morphology

Lobate with marginal ridges, nonchannelized

Lobate to channelized

Channelized

Channelized

Channel-wise profile

Body constant in thickness with planar surface, head stands above body, tail thins abruptly upslope

Body constant in thickness with ridged to planar surface, head stands above body (less than Type I), tail thins upslope

Mass thins from head to tail, irregular surface

Thin continuous ‘stream’, planar surface

Thickness (m)

0.01–2.0 (0.5–0.7 typical)

0.01–1.4 (0.1–0.7 typical)

0.03–0.6

0.02–0.1

Bulk water content (wt %)

~ 8–14

~ 14–19

~ 18–25

725

Bulk wet density (kg m–3)

2000–2600

1900–2150

1800–1950

<1800

Surface flow rates (mm s–1)

1–5

2–50

150–1250

10–2000

Typical length of flow (m)

10–300+

10–300+

100–400+

50–400+

Surface shear strength (MPa)

0.04–0.15

0.06 or less

Not measurable

Not measurable

Approx. bulk grain size (mm)

2–0.3

0.4–0.1

0.15–0.06

<0.06

Flow character (laminar)

Shear in thin basal zone with override at head

Rafted plug with shear in lower and marginal zones

Discontinuous plug to shear throughout

Differential shear throughout

Grain support and transport

Gross strength

Gross strength in plug; traction, local liquefaction and fluidization, grain dispersive pressures and reduced matrix strength in shear zone

Reduced strength, traction, grain dispersive pressures; possibly liquefactionfluidization, transient turbidity

Liquefaction; some traction; buoyancy (?)

214

SUBGLACIAL ENVIRONMENTS

TABLE 8.5. (b) Characteristics of sediment flow deposits in the terminus region of Matanuska Glacier, Alaska (after Lawson, 1979) Lawson flow type

Bulk texture (1) mean (␶) (2) S.D. (␶)

Internal organization General

Structure

Pebble fabric

I

Gravel-sand-silt, sandy, silt (1) –1 to 2 (2) 3 to 4.5

Clasts dispersed in fine-grained matrix

Massive

Absent to very weak; vertical clasts S1 + –0.49–0.55

II

Gravel-sand-silt, sandy silt, silty sand (1) 2 to 3 (2) 3 to 4

Plug zone; clasts dispersed in finegrained matrix. Shear zone; gravel zone at base, upper part may show decreased siltclay and gravel content; overall, clasts in fine-grained matrix

Massive; intraformational blocks. Massive; deposit may appear layered where shear and plug zones distinct in texture

Absent to very weak; vertical clasts. Absent to weak; bimodal or multimodal; vertical clasts. S1 = 0.50–0.65

III

Gravelly sand to sandy silt (1) –2.5 to 2.5 (2) 3.5 to 2.0

Matrix to clast dominated; lack of fine-grained matrix possible; basal gravels

Massive; intraformational blocks occasionally

Moderate, multimodal to bimodal parallel and transverse to flow. S1 = 0.60–0.70

IV

Sand, silty sand, sandy silt (1) >3.5 (2) <2.5

Matrix except at base where granules possible

Massive to graded (distribution, coarsetail)

Absent

Lawson flow type

Surface forms

Contacts and basal surface features

Penecontemporaneous deformation

Geometry* and maximum observed dimensions (length × width, thickness, m)

I

Generally planar; also arcuate ridges, secondary rills and desiccation cracks

Non-erosional, conformable contacts; contacts sharp; load structures

Possible subflow and marginal deformation during and after deposition

Lobe: 50 × 20, 2.5

II

Arcuate ridges; flow lineations, marginal folds, mud volcanoes, braided and distributary rills on surface

Non-erosional, conformable contacts; contacts indistinct to sharp; load structures

Possible subflow and marginal deformation during and after deposition

Lobe: 30 × 20, 2.5; sheet of coalesced deposits

III

Irregular to planar; singular rill development; mud volcanoes

Non-erosional, conformable contacts; contacts indistinct to sharp

Generally absent; possible subflow deformation on liquefied sediments

Thin lobe; 20 × 10, 0.5; fan wedge; 30 × 65, 3.5; rarely, sheet of coalesced deposits

IV

Smooth, planar; mud volcanoes possible

Contacts conformable indistinct

Absent

Thin sheet; 20 × 30, 0.3; Fills surface lows of irregular size and shape

* Length and width refers to dimensions parallel and transverse to direction of movement prior to deposition.

SUBGLACIAL ENVIRONMENTS

215

PLATE 8.7. Flow nose with a diamicton, Vancouver Island, British Columbia. (Photo courtesy of Steve Hicock.)

buoyant at the grounding-line. At this point of detachment from the bed a considerable volume of subglacial debris exits into the proglacial subaquatic environment. Various sediments are deposited at or close to the grounding-line with characteristics derived from both subglacial and subaqueous sedimentary environments (Fig. 8.11) (Talbot and Von Brunn, 1987; Albino and Dreimanis, 1988; Powell, 1990; Dowdeswell and Scourse, 1990). At subglacial meltwater portals extensive linear or deltaic assemblages (subaqueous fans) of stratified sediments are deposited into the proximal subaquatic proglacial zone as meltwater flow velocities rapidly drop. In

more distal subaquatic areas a different imprint characterizes the sediments and a truly subaquatic sedimentation environment develops (Fig. 8.11). Tills deposited in the vicinity of a grounded floating ice front can be subdivided into: (1) Waterlain Tills and (2) Proximal Subaquatic Diamicton M´elanges. The distinction between these tills is made on the basis that waterlain tills are formed as a result of rain-out through a water column, while proximal subaquatic diamictons are formed as a result of ductile extrusion or gravitational deformation (both soft sediment deformation) extending from the grounding-line into the subaquatic environment.

216

SUBGLACIAL ENVIRONMENTS

(a)

grounding line

fl o w li n es

GMS = Glacimarine sediment

e ng lac i a l debris

Dm

Dm

[D]b

basal till

SMc

[D]b

Dm proximal GMS waterlain till waterlain till

proximal GMS

transitional GMS

SM distal GMS

(b)

flo w lin es

Dm basal till

[D]b

waterlain till

proximal GMS

SMc-SM transitional-distal GMS

FIG. 8.11. Schematic cross-section of (a) a floating ice shelf and (b) a grounded ice front showing the relationship between basal till; waterlain till and glaciomarine sediments (reproduced with permission from the Geological Society of London).

8.7.2.1. Waterlain tills Waterlain tills are deposited within a subaquatic environment by continuous rain-out of basal glacial debris from floating icebergs, melting under-surfaces of the main ice body, debris plumes from exiting subglacial meltwater portals, debris flowing off the main ice body and debris falling into the water from these various sources. Little or no reworking is accomplished by bottom currents and, therefore, the influence of subaquatic (lacustrine/marine) processes are minimal except for post-depositional bioturbation (Gravenor et al., 1984). These tills have been classified under a series of different terms in recent years such as waterlaid, subaqueous, glacioaquatic, para- and aqua- tills; also submarine, shelf and basin moraine, and lacustrotill among many (Dreimanis, 1988). Waterlain tills can be deposited over wide

areas depending upon the rate of debris rain-out and the rate of grounding-line retreat across the terrain. Waterlain tills tend to be relatively thin (<1 m) with greater thicknesses occurring in bands transverse to ice flow direction. The characteristics of waterlain tills are determined by the rate of debris rain-out; debris grain size distribution; the style of landward debris transport; water column depth, temperature gradient and turbidity; the angle of lake/sea bed slope; the nature of the underlying sediment; the presence or absence of subglacial meltwater portals and meltwater discharge fluctuations; the rate of sediment deposition; the rate of grounding-line retreat/advance; and the nature and style of iceberg debris addition. Waterlain tills may be rhythmically bedded, almost laminated in appearance, but also may be massive and structureless. Apart from a crude flow-like stratifica-

SUBGLACIAL ENVIRONMENTS

tion, these tills usually contain abraded and striated clasts and have a preferred clast fabric (Fig. 8.8) (Domack and Lawson, 1985). Fine-grained waterlain tills may contain partially rounded clay and silt balls. Occasional deformed and/or flow lenses and sediment intraclasts can be found, the former often of considerable lateral continuity. In comparison with terrestrial tills, waterlain tills often have higher clay and silt concentrations (Stevens, 1990; Cowan and Powell, 1990). The distinction between terrestrial tills and waterlain tills remains problematic. One, often quoted, method of discrimination is the presence of ‘dropstones’ and other artefacts indicative of deposition through a water column, but even this attribute is far from conclusive. Primary structures formed within waterlain tills that have not been subject to the loading/unloading stress cycles of a nearby grounding-line or lateral gravitational slumping are dominantly autokinetic related to bedding planes and porewater transmission processes (Dreimanis, 1993). Waterlain tills, because of their high initial water contents, are highly susceptible to allokinetic stress events. These tills tend to be relatively sensitive immediately following deposition thus dewatering, synergic collapse, lateral shear processes and seismically induced structures are all potential post-depositional features. Where flocculation has been a major physico-chemical process in the rain-out of debris into a water body (especially into brackish water), distinct microscopic ‘cardhouse’ packing of floccules may occur. Contemporaneous with particle structure collapse, porewater migration usually occurs resulting in clay particle migration, clay coating formation and piping. In areas where sea or lake floor slope is sufficiently steep, sediment build-up may cause gravitational slumping and a series of structures may form related to folding and shear deformation akin to those formed as diamicton m´elanges. The main distinction between these deformed sediments is that in gravitationally slumped waterlain tills, the degree of stress is much lower, extrusion flow is not involved and the sediment units are much thinner (<1 m). Where dropstones or inclusions caused by rain-out of sediment or large clasts/boulders falling through the water column and embedding into the till occurs, a distinct suite of stratal warping/buckling structures

217

develop with associated diapiric water escape features developing around the dropstones (Thomas and Connel, 1985; Parkin and Hicock, 1988). Clast fabrics within waterlain tills reveal weak principal preferred orientations, tending toward random, as can be expected with sedimentation through a water column rather than under a unidirectional shear stress (Fig. 8.8). 8.7.2.2. Proximal subaquatic diamicton m´elanges These tills form as a result of two dominant processes: (a) extrusion of subglacial diamicton m´elange into a water body as a ‘Till Delta’ (King and Fader, 1986; Vorren et al., 1989) and (b) massive internal pervasive deformation by weight of overlying waterlain sediments to such a degree that virtually all primary structures are destroyed (Talbot and Von Brunn, 1987). The formation of type (a) is dependent upon a subglacial diamicton m´elange continuing to deform as it is extruded from beneath the ice front at the grounding-line (Fig. 8.12). In the latter case (b), the diamicton m´elange forms under a different set of processes where hydraulic pumping occurs laterally into proximal subaquaeous sequences leading to strata disruption, faulting, folding and some lateral deformation. The pumping effect is thought to be a manifestation of the grounding-line lifting and falling owing to tides thereby causing lateral surges of porewater migrating into the sediment package. The areal extent of these tills is likely to be very limited. The tills will tend to be relatively thin layers (<2 m) covering zones parallel to the grounding-line in an ice front position (Benn, 1989). Only when lengthy stillstands have occurred is there any likelihood the tills will reach substantial thicknesses. These m´elange sub-types exhibit a wide array of autokinetic structures similar to those already discussed under subglacial diamicton m´elanges. The m´elanges developed under lateral hydraulic pumping effects have micro-scale structures. As discussed elsewhere (Fig. 8.8), the intrinsic value of clast fabrics in these heavily deformed tills is subject to great debate. Measured clast fabrics tend to range from well-oriented strong fabrics to scattered, almost, random fabrics (Hart, 1994; Kjaer and Kr¨uger, 1998).

218

SUBGLACIAL ENVIRONMENTS

COUPLING LINE ICE STREAM

GROUNDING LINE DELTA

ICE SHELF

ICE FRONT SEA WATER

ICE ACTIVE TILL

“TILL TONGUE”

WATER FILM

LODGED TILL

MAX. THICKNESS OF ACTIVE TILL

BEDROCK FIG. 8.12. Schematic diagram of ice stream/ice shelf model with a till ‘delta’ or ‘tongue’ (after Alley et al., 1987c, copyright by the American Geophysical Union).

8.8. SUBGLACIAL LANDFORMS AND BEDFORMS A fundamental characteristic of subglacial sedimentation is the formation of depositional or constructive landforms or bedforms (Fig. 8.13). These forms are indicative of subglacial environmental processes. Over the past century there has been a proliferation of ideas and hypotheses on the formation of these landforms. These ideas have, almost exclusively, developed ‘unique’ hypotheses for individual landforms. Most of these hypotheses reflect a view based upon landform morphology or certain apparent unique characteristics. This ‘morpho-sedimentological’ approach, which in the past has been one of the fundamental paradigms of research investigations by geologists and geomorphologists may, in isolation, be tautological and potentially sterile. In recent years a new paradigm has been advocated that considers glaciodynamic conditions prevailing under active

temperate ice masses and attempts to interrelate these subglacial conditions to subglacial sediments, facies and landforms/bedforms. This ‘glacio-sedimentological’ paradigm may substantiate and refine some of the above hypotheses and advance our understanding subglacial environments. Subglacial landforms can be subdivided into two distinctive groups, bedforms and non-bedforms. The former group are integral elements of a continuum of bedforms evolving into each other as conditions alter at the subglacial interface. The latter group are landforms that are interrelated and have certain common characteristics but are not part of a spectrum of forms, but are individual reflections of and responses to varying conditions at the ice–bed interface. Within the first group are drumlins and Rogen moraines and in the second, Kalix till ridges and De Geer moraines. In general, both bedform and non-bedform landforms are lineated either parallel or transverse to major ice flow directions (Table 8.6).

SUBGLACIAL ENVIRONMENTS central area

bottom unfrozen

inner marginal zone outer marginal zone bottom unfrozen sporadically frozen

frozen margin or front unfrozen

U-valleys, roches moutonnees etc. crescentric troughs smallscale flutings cover moraine drumlins and megafluting

219

signify uniformity of till type. Till plain sediments are a complex reflection of local and regional bed geology, topography, ice dynamics and subglacial bed conditions. Lodgement, meltout, flow tills and m´elanges can be found over large or restricted areas of till plains. Typically, they contain rafted units of subjacent sediments and bedrock. 8.8.2. Subglacial Bedforms

lee-ridges Rogen moraine Sevetti moraine radial moraine Veiki plateau moraine Pulju moraine ablation moraine thermokarst included end moraine hills De Geer-, Vikaand other small transverse ridges interlobate formations in general Kianta moraine

FIG. 8.13. Glacial landform/bedform types with reference to their respective basal thermal environment and position at the bed of an ice sheet (after Aario, 1990).

8.8.1. Till Plains Although many characteristic landforms of glaciation are products of active ice formed within the subglacial environment, perhaps the most widespread evidence of subglacial action is the formation of low relief, rolling till plains. These plains cover vast areas and are perhaps indicative of the more typical nonlandforming or bedforming aspect of subglacial processes. Across till plains distinctive features (thrust block ridges, depressions and push ridges) can be found that reflect localized subglacial differential stresses, local variations in sediment rheology or the influence of topographic control (Aber et al., 1989; Aylsworth and Shilts, 1989). These plains have formed as a result of a widespread uniformity of subglacial depositional conditions. These conditions, however, do not necessarily

Subglacial bedforms have been and remain the subject of intense study and debate (Menzies, 1984; Menzies and Rose, 1987, 1989; Clark, 1993). It has become apparent that these subglacial forms can be considered as bedform suites produced at the ice–bed interface (Lundqvist, 1969; Aario, 1977a; Menzies, 1989a; Rose, 1989b; Boulton and Clark, 1990a,b). This view is not universally held, as alternate hypotheses have also been advocated (Dardis et al., 1984; Shaw et al., 1989). The study of subglacial bedforms is the subject of strongly held and often divergent views concerning the processes of formation. However, certain central issues emerge that, in most instances, are common to all approaches to the problem: 1 specific subglacial environmental conditions favour the development of bedforms and nonbedforms within specific regional topographic, sedimentological and glaciodynamic settings; 2 macroscale spatial patterns of subglacial bedforms within bedform belts, zones or fields may provide insight into the subglacial formative dynamics of individual ice masses; 3 the relationship between subglacial interface conditions of sediment rheology, thermal conditions and subglacial hydraulics is critical in bedform/nonbedform generation. Ultimately, any subglacial bedform/non-bedform must be a reflection of the interaction of basal ice/bed stress conditions of a particular ice mass at a specific location and time. When subglacial bedforms are observed at the very large scale it is apparent that a mega-scale pattern of lineation of >50 km in length can be seen (Fig. 8.14). At this large scale, two previously undocumented ice-

220

SUBGLACIAL ENVIRONMENTS

TABLE 8.6. Types of subglacial landforms/bedforms

A

Direct glacial land-/bedforms

Large Plucked Bedrock Forms

Subglacial Forms

Subglacial

Subglacial

Ice-marginal

Parrallel to Ice Motion

Forms Transverse to

Forms Non-linear –

Forms (ice proximal)

Ice Motion

unoriented to Ice Motion

Streamlined Sediments

Moraine Ridges

Ground Moraine

End Moraines

Streamlined Bedrock

Lateral Moraines

Other Moraines

cirques

rock drumlins

drumlins

washboard

corrugated

end moraines

perched

subaqueous

headwalls aretes ˆ rock basins

roche moutonees ´ crag and tails grooves

drumlinoids fluted moraines radial moraines

moraines De Geer moraines

moraines cover moraines till veneers

terminal moraines recessional

moraines valley-sided moraines

moraines grounding-line ridges

tarms nivation basins

p-forms s-forms

lee-side cones

Rogen moraines ribbed moraines

hummocky moraines

moraines push moraines –

looped moraines interlobate

De Geer moraines

glacial troughs (u-valley)

thrust moraines Kalixpinmo Hills

Pulju moraines Veiki moraines

(Stauchmorane ¨ boulder belts/

moraines radial moraines

cross-valley moraines

lake chains (paternoster)

shear moraines squeeze

Blattnick moraines

aprons kame moraines

trimline moraines

till plateau disintegration

finger lakes hanging

moraines

till plains squeeze

tributaries fjords

moraines ice-pressed

glacial lake basins

moraines (ridges)

B

ridges De Kalb mounds

Indirect glacially induced landforms

Subglacial Meltwater Forms

Basal Meltwater Forms

Esker Systems

Submarginal/Ice-marginal Forms

Meltwater Channels

Kame Fields

tunnel valleys

eskers

laterals

kames

potholes glacial chutes

beaded eskers engorged eskers

submarginals chutes

moulin kames kame plateau

erosdion marks p-forms

squeeze-up eskers

subglacials in-out channels

kame terraces kame deltas

s-forms

kame ridges esker chains crevasse fillings

urstromtaler ¨ (pradoliny)

kettles esker fans subaqueous fans subaqueous eskers

Note: based upon Prest, 1968; Sugden and John, 1976; Goldthwait, 1988.

SUBGLACIAL ENVIRONMENTS

221

(a)

(b) 0

500 km

FIG. 8.14. (a) Summarized ice flowlines interpreted from the work of Prest et al. (1969) based upon aerial photographic interpretation. Note the two main radial flow patterns, and that there are no areas of cross-cutting or overlain flowlines. (b) Summarized ice flowlines interpreted by Clark (1993) based upon glacial lineation mapping using Landsat images. Note the extensive area of cross-cutting patterns and, in comparison with (a), the numerous flowline patterns identified (from Clark, 1993; reproduced by permission from Earth Surface Processes and Landforms).

222

SUBGLACIAL ENVIRONMENTS

moulded forms are apparent, mega-scale lineations and cross-cutting lineations. 8.8.2.1. Bedform continua Before discussing individual subglacial bedforms, it is worth considering the spatial and morphological continua that appear to exist in some glaciated terrains between Rogen moraines, drumlins and fluted moraines (Fig.8.15). While working in the Rogen area of Sweden, Lundqvist noted a perceived spatial relationship between these bedforms over terrain in northwestern Sweden (Lundqvist, 1989). In general, in areas where Rogen moraines and drumlins exist, often the Rogen moraines occur only in topographic basins and valley bottoms while drumlins occupy interfluves; there are exceptions, however, to this rule. A relationship seems to exist between concave trending terrain and compressive basal ice flow where the Rogen moraines occur and convex terrain and extending basal ice flow where drumlins are found. The transition from one bedform to another is almost imperceptible. Three styles of bedform transition appear to occur: (1) drumlins become incomplete with their down-ice, lee ends truncated or replaced by a

concave end. These incomplete drumlins begin to become aligned side by side forming ridges transverse to the main ice direction. This transition zone has been separately termed Blattnick moraine (Markgren and Lassila, 1980). The ‘horns’ of the incomplete drumlins always point in the down-ice direction. The ‘change-over’ from drumlins to Rogen moraine and vice versa may repeat itself several times where terrain conditions permit; (2) where drumlins begin to coalesce forming side-by-side ridges, Rogen moraines essentially ‘grow’ from this adjacency; (3) finally flutings, in some places, begin to appear on the surface of the Rogen moraine and assume increasing topographic expression in the down-ice direction to the point where the flutes become the predominant bedform. This third transition may also occur where flutes first ‘convert’ to drumlins and then into very long elongated drumlins and finally into fluted moraine (Fig.8.16). Within 200 km of the ice divide in Finland, Rogen moraine often appear to alter, in the up-ice direction, to Pulju moraines (Aario, 1990) and in the down-ice direction into Blattnick moraines. Similar continuum transitions to those described in Sweden have been noted in bedforms in many parts of the world (e.g.,

ED

IN

GA

ES

ME

UT

FL

GA

L AM RE ST LLS HI

ME

Do

DR S

LIN

S

LIN

UM

UM

DR

ES

UT

y ntl ng na ergi i m v w Do con flo

Ice Thickness

ex

y

ntl

na mi

w

flo

ing

d ten

FL

Bedform Elongation Ratio (a/b)

Rate of Ice Movement

Bedform Length (a axis) FIG. 8.15. Continuum of subglacial bedforms in relation to ice thickness and rate of ice movement (after Rose, 1987a, reprinted from Menzies and Rose (eds), courtesy of A.A. Balkema, Rotterdam).

SUBGLACIAL ENVIRONMENTS

(a)

223

Ice Flow Direction

Hummocky active-ice assemblage

Drumlin assemblage

Fluting assemblage (after Aario, 1977)

(b)

Rogen Created Incomplete moraine drumlins drumlins Drumlins

(after Lundqvist, 1970)

FIG. 8.16. Bedform transition patterns beneath active ice. After (a) Aario, 1977a; (b) Lundqvist, 1970 (after Menzies, 1987. Reprinted from Menzies and Rose (eds), courtesy of A.A. Balkema, Rotterdam).

central and northern Finland (Punkari, 1984; Heikkinen and Tikkanen, 1989); central Norway (Sollid and Sørbel, 1984); Quebec (Bouchard, 1989), Keewatin and Labrador (Aylsworth and Shilts, 1989), Saskatchewan (Moran et al., 1980) and Ontario, Canada (Menzies, 1987); North Dakota, USA (Moran et al., 1980); and Patagonia, Chile (Clapperton, 1989)). In general, the transition from Rogen moraine to drumlins and/or fluted moraine ceases closer to ice margins where perhaps basal ice velocities are too great, basal extending flow too dominant or basal debris deformation too rapid (Fig.8.17). The significance of this continuum of forms demonstrates the close relationship between the ice– bed interface processes of sediment rheology, basal ice glaciodynamics and subglacial hydraulics (Fig. 8.17). Table 8.7 illustrates some possible relationships between ice–bed interface conditions and bedform/ non-bedform type development in the subglacial environment.

FIG. 8.17. Distribution of Rogen morainic forms in Sweden (courtesy of C. H¨allestrand).

8.8.2.2. Rogen moraine The term Rogen moraine or ribbed moraine is applied to a series of conspicuous morainic ridges found transverse to the main ice direction occurring in such large numbers to constitute a field. These ridges may reach heights of 10–20 m, are 50–100 m in width, with lateral extents of sometimes several kilometres and often with an interval spacing of 100–300 m (Plate 8.8). Where the moraines transversely cross valleys and other topographic depressions gaps are often found in some of the central parts of the ridge sets. Occasionally, the ridges exhibit a slight down-ice arcuate pattern. Lateral ridges usually are composite

224

SUBGLACIAL ENVIRONMENTS

TABLE 8.7. Interface locations

Interface conditions

Bedform type

Internal sediments and structures

Orientation of bedforms to flow of ice

Lower

1. Ice-bedrock/ immobile bed

No depositional bedforms but erosion forms

Diamicton and occasional stratified sediments, deformation structures, intraclasts, distinct herringbone

Linear and transverse

2. Ice-thin traction zone

Isolated forms, possibly bare bedrock between

3. Ice-thick mobile traction zone

Well-developed continua of forms of unstratified sediment

4. Ice-meltwaterbedrock

Forms but unlikely to be part of a continuum

Stratified, melt-out deformation structures, intracists, faulting, drag-folds

Linear and transverse but not part of continuum of forms

5. Thin mobileimmobile traction zones

Small isolated forms of unstratified sediment with bare bedrock between

Diamicton dominant, deformation structures, strong clast fabric, lodgement, melt-out, flow tills present

Linear and transverse

6. Thick mobile– immobile traction zones

Well-developed continua of forms of unstratified sediment

Occasional stratified intraclasts

7. Mobile traction zone-bedrock (thin?)

Isolated well-developed forms of unstratified sediment with bare bedrock between

Upper

interfingered ridge mosaics rather than single forms. Individual ridges are often asymmetric with a steeper lee-side giving a fish-scale appearance to the terrain. Over large areas of terrain ridge crest heights are often remarkably accordant. Not all Rogen moraines are well developed and occasionally moraines are found superimposed upon drumlinoid ridges. At the regional scale, Rogen moraine occur at distances of ~200 km from ice sheet centres and rarely within 200–300 km of the ice margin (Fig. 8.6). In other words, Rogen moraine occur typically just beyond the centres of glaciation. At the local scale, as discussed above, Rogen moraine often occur within topographic depressions and are intrinsically associated with drumlinized terrain. The composition of Rogen moraines exhibit, in general, a wide range of sediment types and structures. However, stratified sediments are often the

dominant type. Aylsworth and Shilts (1985) noted that in some places Rogen moraines seemed to have a coarser texture than neighbouring drumlins. Shaw (1979), from three areas studied in Sweden, concluded that the sediments were of definite subglacial origin but of a complex lithofacies association. The till within the ridges exhibited folding and the presence of dislocated till units, with clast fabrics exhibiting random orientations. The till, termed ‘Sveg Till’, was weakly stratified with thin lenses of sorted material intercalated within the units. Bouchard (1989), in Quebec, discovered similar sediment structures with distinctive thrust shear planes manifest as ‘slabs’ of till stacked one on another. Several hypotheses explaining the formation and spatial pattern of Rogen moraines have been advanced (Fig.8.18) (see reviews by Bouchard, 1980 and Lundqvist, 1989). Three main hypotheses (1)

SUBGLACIAL ENVIRONMENTS

225

PLATE 8.8. Roger moraine, Uthussl¨on, Krattels¨on Sweden. (Photo courtesy of Jan Lunqvist.)

subglacial, (2) stagnant ice and (3) subglacial tectonics, have gained some support. Other origins connected with subglacial meltwater floods, crevasses fillings and marginal moraine formation have also been advanced (Fisher and Shaw, 1992). 8.8.2.3. Drumlins Drumlins are roughly ovoid-shaped hills of dominantly glacial debris that typically occur within groups or fields of several thousands. Drumlins exhibit very strong en echelon long-axis preferred

orientation paralleling the main direction of ice flow (Menzies, 1984). Isolated drumlins are known to exist and occasional fields may contain only a few dozen drumlins. The classical shaped drumlin usually has a steeper stoss end and a tapered lee-side, however variants on this shape are perhaps more common than the classical shape itself (Plate 8.9). Drumlins may range in height from 5 to 200 m, in width from 10 to 100 m and in overall length from 100 m to several kilometres. An extensive statistical analysis of drumlin dimensions has been pursued from which several explanations of drumlin origin have been derived

226

SUBGLACIAL ENVIRONMENTS

Frontal formation End moraines

Supraglacial till and/or glaciofluvial sediments

Formation by active ice

vs.

Basal till

vs.

Formation behind the ice margin

Supraglacial formation

Fluted surface Transition to drumlins

Formation by stagnant ice

Subglacial formation

vs.

Situation in concave parts of terrain

Formation in open crevasses

vs.

Distribution in the central area of glaciation

Formation in basal crevasses

FIG. 8.18. Schematic illustration of proposed interpretations of Rogen moraine (in ovals) compared with facts observed in the field (in rectangles). The interpretations are shown in pairs, representing opposing hypotheses. The lines indicate the implications of each parameter; broken lines are less apparent relationships (after Lundqvist, 1989, reproduced with permission from Elsevier Science Publishers).

(e.g., Smalley and Unwin, 1968; Trenhaile, 1975; Evans, 1987; Mills, 1987). Few drumlins have been observed appearing from beneath modern-day ice masses. Drumlins on James Ross Island, Antarctica (Rabassa, 1987), in the proglacial zones of Myrdalsj¨okull, Iceland (Kr¨uger, 1987) and the Bifertensgletscher, Switzerland (Van der Meer, 1983) have been observed but none, as yet, emanating from beneath the Greenland or Antarctic Ice Sheets (Plate 8.10). Vast drumlin fields, numbering in the thousands exist, for example, in Canada, Estonia, Finland, Ireland, Germany, Poland, Russia and the USA. In almost all glaciated terrains smaller drumlin fields are to be found. The topographic locations within which drumlins are found are many and varied. Drumlins occur in both lowland and highland terrains beneath ice sheets and valley glaciers. They occur close to terminal moraines and may, in places, appear contiguous with these moraines, while elsewhere drumlins occur on the edge of ice sheet centres (Fig. 8.6). It has been suggested that beneath an ice sheet certain areas

might preferentially be conducive to drumlin formation (Fig. 8.19) (Boulton et al., 1977, p. 243). Occasionally, a radiating pattern can be observed within a drumlin field (Fig. 8.20) (Goldstein, 1989) that has been interpreted, in the past, as evidence of basal crevasse infilling owing to divergent ice flow close to an ice margin. It has also been suggested that drumlins, in association with Rogen moraine and fluted moraines, may be related to deformable beds beneath ice sheets and are, therefore, linked to fast basal ice (>500 m year–1 ) and a preferential location within ice streams in ice sheets (Dyke and Morris, 1988; Menzies, 1989a). Limited relationships appear to occur between drumlins and topography although in certain fields some, perhaps causal, relationships do appear to exist. Drumlins are such striking landforms that for many years little cognisance was taken of their internal composition. Drumlins are composed of a vast range of sediment types of varied provenance, containing an array of sediment structures and forms (Fig. 8.21). In the past, drumlins were mistakenly perceived as being

SUBGLACIAL ENVIRONMENTS

227

(a)

(b) PLATE 8.9. (a) Drumlin of the Green Bay Lobe, Wisconsin. (Photo courtesy of Donna Stetz.) (b) Drumlins from Kuusamo drumlin field, eastern Finland. (Photo courtesy of Risto Aario.)

228

SUBGLACIAL ENVIRONMENTS

PLATE 8.10. Drumlin in the proglacial zone of the Biferten Glacier, Switzerland. Photo shows drumlin width to be approximately 150 m.

composed almost exclusively of subglacial tills. Although the dominant sediment within drumlins is often till, many drumlins may contain stratified sediment. Stratified sediment may compose the internal materials of whole drumlin fields, as in Velva, North Dakota or Livingstone Lake, Saskatchewan, with only a few isolated till intraclasts; while individual stratified drumlins may ‘sit’ adjacent to till drumlins as in Peterborough, Ontario. The range of sediment types and structures found within drumlins is illustrated in Figure 8.21. Many drumlins have observable cores of bedrock, boulder dykes and other non-glacial nuclei around which subglacial debris has accreted or been emplaced by some mechanism. In some cases, drumlin or drumlinoidal forms can be observed ‘carved’ from bedrock in the form of rocdrumlins. However, most drumlins do not appear to

have obvious cores around which they have been ‘built’ and these forms remain puzzling to explain. Clast fabrics within drumlins appear, in some cases, to follow the outer morphology of the drumlin (Walker, 1973), while others exhibit transverse orientations (Andrews and King, 1968); or a ‘herringbone’ style pattern (Shaw and Freschauf, 1973; Aario, 1977a). In many cases the complexity of internal sedimentological structures provides a random orientation. It can be questioned as to whether clast fabrics within drumlins have anything other than local significance and may indicate little concerning the origin of the drumlin form or its initiation. Drumlins exhibit such a wide complexity of form and internal composition that it is almost impossible to characterize what is an ‘ideal’ drumlin. Many drumlins, for example, are found lying on top or

SUBGLACIAL ENVIRONMENTS

229

FIG. 8.19. Schematic diagram of general conditions for drumlin formation following the dilatancy theory (modified from Smalley and Unwin, 1968; and Piotrowski and Smalley, 1987).

230

SUBGLACIAL ENVIRONMENTS

20

0 km

FIG. 8.20. Radial pattern of drumlin distribution in the Wadena drumlin field, Minnesota (after Goldstein, 1989, reproduced with permission from Elsevier Science Publishers).

obliquely across other larger drumlin forms (megadrumlins). Drumlin shapes may vary enormously and may reflect formative penecontemporaneous processes or simply post-depositional subaerial mass movement (Fig. 8.22). Many drumlin fields progressively change as part of a continuum of bedforms, thus drumlin genesis would appear tied, in those instances, to subglacial environments conducive to Rogen and fluted moraine formation. The question of drumlin formation has attracted a vast array of research work (Menzies, 1984; Menzies and Rose, 1987, 1989) (Table 8.8). Before discussing this most definitive of all ‘glacial’ questions, it is pertinent to state the ‘conditions’ that must be met by any hypotheses attempting to explain drumlin formation, assuming that a single explanation does exist for such a diverse bedform type. Any explanation of drumlin formation needs to account for: (1) the varied

location of drumlins and their close association in fields; (2) the diverse shape and morphology of drumlin form; (3) the enormous range in sediment type and structures within drumlins; (4) the existence of rock-cored and non-rock-cored drumlins, often in proximity to each other; (5) the presence of drumlins in bedform continua in some but not all cases; (6) the relationship of drumlins to subglacial glaciodynamics and hydraulics; (7) the chronology of drumlin formation whether drumlins form simultaneously as a single field or develop into a field by repeated ‘overprinting’ in a single glacial phase or repetition over several glacial phases; (8) stages of drumlin development whether formed en masse or by gradual accretion in a single continuous event or interrupted accretionary events; and, finally, (9) a ‘trigger’ mechanism(s) needs to be found that is operative in certain specific conditions yet not under others.

SUBGLACIAL ENVIRONMENTS

231 LEGEND SEDIMENTS Subglacial diamicton - type A

ICE FLOW

Subglacial diamicton - type B Sands, gravels, silts (of almost any fluvial depositional environment, glacial or non-glacial) Rock (intact) or boulders Clasts STRUCTURES Folding Faulting Fissuring and jointing Lenses and laminae Cross-bedding and other fluvial bedform structures Injections or intrusions Clast pavements

FIG. 8.21. General model of internal sediments and structures found within drumlins.

At present, three main groups of drumlin forming hypotheses can be identified: 1 Formation by moulding of previously deposited material within a subglacial environment in which a limited amount of subglacial meltwater activity occurs (Q-bed state possibly where a frozen bed transforms to a melted bed). Meltwater may influence moulding and deformational processes that are subsequently produced by acting either as a lubricating basal film at the upper ice–bed interface or as debris-held porewater thereby reducing subglacial effective stresses (Whittecar and Mickelson, 1979; Kr¨uger and Thomsen, 1984; Boulton, 1987a; Kr¨uger, 1987). Debris is moulded by direct deformation of previously deposited sediment (both glacial and non-glacial) into drumlinoidal shapes by direct basal ice contact following some type of smearing-on or sculpting process(es). 2 Formation resulting from anisotropic differences in the subglacial debris (under dominant M-bed

conditions) owing to: (a) dilatancy (Smalley, 1966; Smalley and Unwin, 1968; Piotrowski and Smalley, 1987); (b) porewater dissipation (Menzies, 1989a,b); (c) localized freezing (Baranowski, 1979; Menzies, 1989a,b); (d) localized helicoidal basal ice flow patterns (Shaw and Freschauf, 1973; Aario, 1977a,c); or (e) localized subglacial debris deformation (Menzies, 1987, 1989b; Aario, 1987; Boulton, 1987a; Rose, 1987a). Within this specific group meltwater activity is of limited impact, whereas porewater is considered critical in mobilizing or immobilizing local bed debris. The debris is not usually considered as being in a state of mobilized deformation except in case (e). It is suggested that changing stress field and/or stress/ strain histories owing to transient basal glaciodynamics locally affecting subglacial debris rheology are the important parameters in determining whether drumlins begin to form or not. 3 Formation resulting from the influence of active basal meltwater (under H-bed conditions) carving

232

SUBGLACIAL ENVIRONMENTS TABLE 8.8. Hypotheses of drumlin formation* 9

11

1. Erosion of pre-existing glacial materials 2. Deposition en masse

10

6

3

4

17

18

2

12

8

5

7

1

13 14 15

16

33 34

29 30 31

(a)

32

20

19

21

22

23

35 38 36 37 39 40

41 42 43

44

45

24

46

47

25

26

27 28

52 48 49 50 51

53

3. Accretion/deposition (a) with a pre-existing rock/boulder/sediment core (b) with a ‘generated’ core of sediment/boulders kinematic fluting porewater dissipation freezing front effects dilatancy thrust-block – sediment emplacement 4. Subglacial bed deformation 5. Subglacial meltwater processes erosion mark – cavity infilling erosion of pre-existing sediment combined effects of cavity infilling and erosion of pre-existing sediment * modified from Menzies, 1984

9 10 3

7 5

8

4

(b) 0

11

2

1 1 KM

2

14

15 Outcrop of bedrock

6

16

17

18 19 20 12 13

Highest point(s) of drumlin

FIG. 8.22. Examples of drumlin morphology from the Pieks¨am¨aki (a) and Keitele (b) drumlin fields, Finland (reprinted from Gl¨uckert, 1973, reproduced by permission of the Geographical Society of Finland).

cavities beneath an ice mass and later infilling with assorted but predominantly stratified sediment or by the subglacial meltwater erosion of already deposited sediment at the upper ice–bed interface. This hypothesis stems from the presence of stratified sediments in drumlins either in part, in leeside positions (Dardis and McCabe, 1987; Sharpe, 1988a), or through the entire drumlin, or the sculpting by fluvial processes of previously deposited sediment (Shaw et al., 1989). This hypothesis demands meltwater flow at catastrophic discharges from beneath certain areas of an ice mass across the upper ice–bed interface yet permitting the overall ice mass to remain glaciodynamically stable (for

detailed calculations see Shaw et al., 1989). This form of drumlin development, as with the hypothesis in (1), requires a two-stage process of initiation, beginning first with either a pre-formed cavity or pre-existing sediment at the upper ice–bed interface. The latter stage need not be linked directly to the former stage, therefore in some cases although conditions may be suitable for initiation for the first stage, the second stage may not continue toward the critical point (trigger) of drumlin development. In all of these hypotheses the conditions at the subglacial interface(s) are the key to subsequent drumlin formation and, in the long term, to drumlin ‘survival’. A complex relationship must exist between basal glaciodynamics, subglacial sediment rheology and hydraulics for any particular area of ice bed. Fluctuations in state or stress levels or meltwater production and pathways will affect all other parameters to some degree. Certain fluctuations may cross critical thresholds that cannot be reversed, while others may exhibit varying degrees of hysteresis. The likelihood or otherwise of subglacial conditions occurring in any or all of these hypotheses remains a fundamental research problem.

SUBGLACIAL ENVIRONMENTS

233

In certain locations it is difficult to distinguish fluted moraine from extremely elongated drumlins (Plate 8.11). A close relationship does appear to exist between these bedforms but whether such a relationship exists in all instances remains debateable. Fluted moraines, on average, are long ridges of dominantly glacial sediment 100–500 m in length, 1–3 m in width and less than 1–2 m in height with a transverse spacing of 0.5–1.5 m (Table 8.9). Exceptions do occur where flutes may be much larger in all dimensions and stretch for several kilometres. Flute ridges usually exist in groups paralleling the main direction of ice flow and may, as discussed above, be part of a continuum of bedforms. The ridges are usually symmetric in cross-section but often mass wasting alters the flanks where meltwater from the glacier snout has undercut ridge flanks. Flutes may occur in front of a drumlin belt, in places on top of drumlins, occasionally on the surface of Rogen moraine or as isolated bedforms in the proglacial zones of present day glaciers. Typically, flutes exhibit a strong preferred unidirectional ori-

entation but occasionally a radiating pattern may occur. Fluted moraine may be composed of a wide variety of glacial sediments and often a structureless or massive inner core can be observed. In other places, deformed sediments, with folding and faulting, have been described (Fig. 8.23) (Paul and Evans, 1974). Sediment grain size, typically, coarsens both toward the surface of individual ridges and, in present proglacial areas, in a down-valley direction. In many cases ridge surfaces are covered with a gravel lag deposit. Interflute troughs may contain glaciofluvial sediments, and meltout and supraglacial flow tills. Clast fabrics within flutes exhibit strong orientations usually paralleling the ridge long axis (Fig. 8.24). Hypotheses on fluted moraine origins include those related to drumlins, basal crevasse infilling, surface crevasse infilling and collapse, localized subglacial frozen patches, subglacial tectonization and simple lee-side ridges in the down-ice side of large boulders or similar obstructions (Fig. 8.25). Since fluted moraine is only a morphological term, multiple origins are likely. Flutes can be divided into three groups: (1) large flutings related in origin to drumlins

(a)

(b)

8.8.2.4. Fluted moraine

PLATE 8.11. (a) Fluted moraine in the proglacial zone. (b) Large flute extending from a boulder (approximately 1.5 m in height). Note other flutes adjacent to central flute. Both photographs from proglacial zone of Storbreen, Norway.

234

SUBGLACIAL ENVIRONMENTS

TABLE 8.9. Maximum height (m) of flutes observed by various authors Author

Location

Height (m)

Paul and Evans (1974)

Blomstrandbreen, Spitsbergen

0.2 1.0

Silt and fine sand with pockets of coarse sand and gravel Till

Boulton and Dent (1974)

Brei∂/ amerkurjokull, ¨ Iceland

0.5

Till

Morris and Morland (1976)

¨ Iceland Brei∂amerkurjokull,

0.67

Undrained till

/

Comment

Ray (1935)

Mendenhall Glacier, Alaska

0.075

Gravel

Grant and Higgins (1913)

Petrof Glacier, Alaska

0.45

Gravel

Todtmann (1952)

Bruarjokull, ¨ Iceland

2.0

Hoppe and Schytt (1953)

Bruarjokull, ¨ Iceland

1.0

Isfallsglaciaren, ¨ Kebnekajse

0.45

Bogacki (1973)

East of river Sandgigjukvisl, Iceland

1.0

Piles of stone and rubble

Kozarski and Szupryczynski (1973)

Sidujokull, ¨ Iceland

1.3

Stone layer at top, mainly pebbles 0.2 m diameter strongly compacted moraine deposit, tabular structure thin layer very strongly pressed slate-like material, moraine with numerous cobbles and pebbles

Baranowski (1970)

Werenskioldbreen, Spitsbergen

0.3

Abundance of silty material

Dyson (1952)

Grinnell Glacier/Sperry Glacier, Montana

0.9

Largest and most pronounced ridges on moraine which contains a relatively high proportion of rock flour

Shaw and Freschauf (1973)

Athabasca, Alberta

20.0

Coarse till (these features may not be flutes in our sense, however)

and Rogen moraines; (2) flutes related to active subglacial processes close to ice margins; and (3) small flutes formed at the ice margin where saturated debris is squeezed into basal ice cavities formed by boulders or empty meltwater conduits. 8.8.3. Non-bedform Subglacial Landforms It is generally accepted that many subglacial landform features are the products of non-bedforming processes that are related to specific topographic, geological and other non-glacial controlling variables. Landforms such as De Geer moraines and Crag and Tails are major examples of these types of subglacial forms.

Finest material, sand and finer, on crests of ridges Unsorted material, all grades from fine clay to large boulders

8.8.3.1. De Geer moraines These moraines form in the subglacial/subaquatic proximal environment close to either local groundinglines or basal crevasse fractures. These moraines are also described as Cross-Valley and Washboard moraines. De Geer moraine ridges are, typically, transverse to the main ice direction often with a slightly arcuate down-ice plan form with an asymmetric cross-section, the distal slope tending to be the steepest. Ridges are of highly variable height, length and width but, typically, are <5 m in height, 50 m wide and laterally may extend as a continuous ridge for <1 km (Zilliacus, 1989). Ridges usually occupy a low

SUBGLACIAL ENVIRONMENTS

W

point in the terrain. De Geer moraines may, in many cases, be seasonal in formation. In glaciated fjord-type terrain they are found as a series of ridges across the fjord valley floor (Fig. 8.26(a)) and in more open terrain may exist as contiguous ridges stretching over many kilometres paralleling successive ice marginal positions (Fig. 8.26(b)). De Geer (1889) was the first to recognize these ridge moraines and thought they must have formed in front of the ice. Later, others described De Geer moraines as forming in ‘calving bay’ locations. Since De Geer moraines form at the boundary between the subglacial and subaquatic environments, they form in an area of sediment concentration both in transit and deposition. These moraines are, therefore, composed of a vast array of sediment types but with a dominance of stratified sediment units and often close association to meltwater deposits such as eskers (Fig. 8.27). The sediments usually contain rafted intraclast units but often do not exhibit evidence of glacial tectonization. Clast fabrics within the moraines usually exhibit a strong preferred orientation normal to the ridge crest with proximal fabrics

E

Brown Silty Sands

Medium-Coarse Sands & Pebbles

Sandy Silts

(a)

(b)

25 cm

FIG. 8.23. (a) Sediments within fluted moraine from Blomstrandbreen, northwest Spitsbergen. (b) Schematic representation of structures observed in (a) (after Paul and Evans, 1974; reproduced by courtesy of the International Glaciological Society).

C A B

E

G H I

D

L

M N

K

O P

J F

235

1m

/

FIG. 8.24. Clast fabrics, plotted on Schmidt equal-area projection, from flute-forming till from beyond the margin of Brei∂amerkurj¨okull, Iceland. Arrow gives the direction of ice movement and of flute crest. 2␴ contour interval (after Boulton, 1976; reproduced by courtesy of the International Glaciological Society).

236

SUBGLACIAL ENVIRONMENTS D

I

S

T A

N

C

E

(a) Glacier Flow Lines

Boulder Vp = Vi

Vi

Till

(e)

PLAN VIEW Vp < V i Zone of enhanced pressure

I

M

E

(b)

Boulder ploughs up till

T

(c)

Enhanced pressure in ice around boulder Hydrostatic pressures at the ice-till contact

Hydrostatic conditions

(d)

Pi > Pt

Pi = Pt

Pi > Pt Ice pressure into till

Vp = 0

Till squeezed into cavity

Pi = P t Flute retains constant height

Advancing nose of flute

FIG. 8.25. A graph of time versus distance superimposed on diagrams of (a) an englacially transported boulder, which is (b) retarded by ploughing into the till, and which (c) subsequently develops a wedge of till on its lee side. This till can (d) be traced into a flute. Diagram (e) shows a plan view of diagram (d) (after Boulton, 1976; reproduced by courtesy of the International Glaciological Society).

showing up-ice plunges and the distal fabrics indicating down-ice orientations. De Geer moraine formation would appear to occur within the subglacial/subaquatic grounding-line zone. All hypotheses of moraine formation tend to concede that deposition has occurred at the grounding-line, only the mechanism of transport to that position is in contention. It has been suggested that debris concentration is the result of englacial thrust zones, or shear planes, or near-marginal basal crevasses. Zilliacus (1989) presented a model of De Geer moraine formation in which it is suggested that debris was squeezed up into transverse basal crevasses (Fig. 8.28). A significant conclusion from this work suggests that De Geer moraines are not annual

moraines but instead merely reflect the temporary positions of a nearby dynamic grounding-line and thus have limited geochronological significance. 8.8.3.2. Crag and tail forms The distinction between crag and tail forms and fluted moraine forms, developed in the lee of boulders, is one possibly only of scale. The formative mechanisms of both forms would appear to be similar. The characteristic crag and tail may be tens of metres high at the stoss end and may stretch for several kilometres in the down-ice direction. The classic form is well represented by the crag and tail that Edinburgh Castle and the Royal Mile in Edinburgh, Scotland, sits upon

SUBGLACIAL ENVIRONMENTS 0

½ Miles

1

'

58

16

16

58

'

Striation ................................................ Glacial Lake Lewis (1658 ft. a.s.l.) .. Upper "terraces" ....................... Cross-valley moraines ............. Central kames ........................... Entrance to 1658 ft. over flow of Glacial Lake Lewis ..........................

(a) Striation ................................................ Glacial Lake Lewis (1658 ft. a.s.l.) ... Upper "terraces" ........................... Cross-valley moraines .................. Central kames ..............................

237

(Fig. 8.29). In many cases the crag is composed of bedrock while the tail may contain a variety of dominantly glacial sediments, often of a glaciofluvial nature. However, as in the case of the form in Edinburgh, the bedrock protrusion that represents the crag has been moulded in the down-ice direction reflecting a large bedrock-cored tail. Crag and tails are typical forms in terrain where resistant but isolated up-standing bedrock obstructions have been overridden by active ice. Therefore in areas where extinct volcanic cores, reefs or igneous intrusions such as dykes have occurred crag and tails of a range of sizes generally are to be found. The moulding and erosive effect of active ice was thought, in the past, to be the formative mechanism involved in crag and tail development. However, usually associated with the crag is an up-ice crescentic hollow and hollows along the sides of the tail that may represent the additional impact of subglacial meltwater under high hydrostatic pressures. The crag and tail form is a morphological expression of active ice motion past a large subglacial bed obstacle and, therefore, is both erosional and depositional in development. Where the tail contains substantive debris a combination of ice moulding at the stoss end and cavity infilling in the lee-side owing to pressure reduction can be expected. 8.9. SUBGLACIAL EROSIONAL FORMS UNDER ACTIVE ICE

0

½ Miles

1

(b) FIG. 8.26. (a) Baffin Island, 19, 1963) (b) Baffin Island,

Map of cross-valley moraines in the Rimrock area, Canada (reprinted from Geographical Bulletin, Vol Map of cross-valley moraines in the Isortoq area, Canada (reprinted from Geographical Bulletin, Vol 19, 1963).

Subglacial erosion processes are pervasive at the active ice–bed interface. In some terrains under certain subglacial conditions, erosive processes become dominant. Almost any description of the impact on the land surface of past glaciers invariably notes, for example, the grandeur and size of fjords and the rugged sculpted bedrock features of once glaciated terrains. However, our understanding of how these distinctive features were fashioned remains limited. Erosional forms exist at an immense range of scales from the surface microscopic features of particles composing glacial debris to the mega-scale landforms of the District of Keewatin in Canada’s North West Territories. Regional erosional features can be subdivided, for the purposes of discussion, into regional and local forms. However, some forms, such as roche

238

SUBGLACIAL ENVIRONMENTS

(a)

E

W Distal Slope

Proximal slope Primary structure destroyed

3m

2

1

(b)

E 1 2 3

1

2

10 11 9 5 4

8

3

W 6

4

1

5

6

7

8

2

9

10

3

11

12

13

4

1

2

14

15

3 16

4 17m

5

7

1-10 Sample sites Orientation of resultant vector Orientation of ridge crest 10% Percent of sample

6

Ø = N124º - N304º L = 26.3% n = 86

7

Ø = N120º - N300º L = 36.7% n = 87

Ø = N144º - N324º L = 30.2% n = 87

Ø = N125º - N305º L = 33.7% n = 87

8

Ø = N135º - N315º L = 36.2% n = 87

Ø = N136º - N316º L = 30.1% n = 86

Ø = N127º - N307º L = 28.7% n = 87

Ø = N130º - N310º L = 20.5% n = 87

9

10

11

Ø = N106º - N286º L = 18.7% n = 87

Ø = N111º - N291º L = 25.8% n = 87

Ø = N116º - N296º L = 29.0% n = 58

FIG. 8.27. (a) Cross-section through a De Geer moraine showing internal structures and macrofabrics. 1, Gravelly till; 2, lenses of gravel; 3, sand/silt; 4, lenses of sand/silt. The section consists mainly of sandy till. (b) Fabrics of pebbles (2.0–6.4 cm) from the cross-section. 20° class intervals, n = number of particles. Resultant vector (␪) and vector magnitude (L) (reproduced by courtesy of the Societas Upsaliensis pro Geologic Quaternaria, from Striae 20).

mouton´ees, can be found at all scales, illustrative of the pervasiveness of most erosional wear processes. Regional erosional features occur as either areal (spatially pervasive) or linear (spatially discrete) landscape types. In the former case, an areally scoured terrain develops where limited debris existed at the ice–bed interface and H-bed conditions prevailed. In contrast linear erosion processes, by definition, are differential in their impact upon terrain being confined within specific areas. Linear erosional forms are indicative of subglacial bed states in which rapid but spatially restricted basal ice movement and/ or meltwater channelling has occurred (Fig. 8.30). 8.9.1. Regional Areal Erosion Areas of terrain where regional areal erosion has dominated the landscape, typically exhibit low relief

amplitude, limited sediment deposition and are dominated by a moulded and scoured appearance (Plate 8.12). Geological structure has often been partially exhumed in these terrains and irregular depressions and small roche moutonn´ees are common. These terrains, known as ‘knock and lochan’ in northwest Scotland, are typical of this form of glacial erosion. Similar landscapes exist in shield terrains in Canada and Fennoscandia, along the edges of the Greenland and Antarctic Ice Sheets, in Patagonia, and the South Island, New Zealand. The formation of this regional erosional landscape would appear to be related to relatively slow-moving ice masses under H-bed conditions with limited debris present. Zones with these characteristics within ice sheets may be present either close to ice sheet centres or in thinner marginal areas. A range of wear processes appear to have operated across a dominantly bedrock surface where

SUBGLACIAL ENVIRONMENTS 1

2

INITIATION OF ACTIVE FLOW PHASE

ACTIVE FLOW PHASE; CREVASSE FORMATION

ICE CREVASSES

SUBGLACIAL TILL

239

SUPRAGLACIAL BASAL

ROCK 3

4

QUIESCENT PHASE; DE GEER MORAINE DEPOSITION FOLLOWING ICE SUBSIDENCE

END OF QUIESCENT PHASE STARTING NEW CYCLE; ICE UP-LIFT FOLLOWING WATER PENETRATION; DE GEER MORAINES FREED OF ICE

REMOULDED DEPOSIT

MELT-OUT BOULDER

ICEBERG

DE GEER MORAINES

FIG. 8.28. Model of the genesis of a series of De Geer moraines (after Zilliacus, 1989, reproduced with permission from Elsevier Science Publishers).

60

Ge

g or

e

St

re

et

30

60

40

50

50

40

60

70 Mil Roy

120

s Gra

40

50

al

Cowga

0

10

CASTLE

e

sm

ark

60

te

et

60

70 contour interval 5m

80 Lauriston

0

m

200

FIG. 8.29. Crag and tail shown by rockhead contours, Edinburgh, Scotland (ice flow direction from left to right) (after Sissons, 1976).

240

SUBGLACIAL ENVIRONMENTS

Legend Main erosion zone (area) Queen Elizabeth Islands

Selective linear erosion Main depositional zone

?

Zone of little/no erosion

Viscount Melville Sound Victoria Island

Somerset Island

Alpine

Baffin Island

?

?

Ungava Hudson Bay Labrador

?

?

0

500 km

FIG. 8.30. Main glacial landscape zones, Laurentide Ice Sheet. Compiled from Landsat–1 images and topographic maps (after Sugden, 1978; reproduced by courtesy of the International Glaciological Society).

occasional protuberances have led to pressure melting of the ice and discharges of meltwater often under high hydraulic pressure heads. Within this landscape, at a lower scale, crag and tails, roche moutonn´ees and grooves are prevalent. Typically, P-forms and associated forms related to rapid subglacial meltwater flow are also common.

8.9.2. Regional Linear Erosion Beneath specific zones within ice sheets, regional linear erosion appears to occur probably as a consequence of ice streaming and related fast-moving, but spatially restricted, basal ice. Under these conditions, major linear forms of glacial erosion appear to

SUBGLACIAL ENVIRONMENTS

241

PLATE 8.12. Areally scoured glaciated terrain in the Canadian Shield. Note the tracery of bedrock faults and fractures. Lac Troie area, Nouveau-Qu´ebec, Qu´ebec, Canada. (Photograph from the Government of Canada.)

develop, for example, incised bedrock troughs, fjord valleys and tunnel valleys. 8.9.2.1. Bedrock troughs and fjord valleys Deeply incised bedrock troughs and coastal fjords are both spectacular landscape forms indicative of intense highland glaciation (Sugden, 1978). These, almost archetypal, forms of linear erosion are typically deep,

parallel-sided valleys cut into bedrock that have short straight sections and often several deep basins (Fig. 8.31). Troughs and fjords, although similar in many respects, differ in a few critical aspects: bedrock troughs usually have a stepped longitudinal profile (riegels) and relatively few transverse sections, while fjords may have one or several bedrock sills or thresholds (Fig. 8.32) (Nesje and Whillans, 1994) and typically many transverse valleys (Plate 8.13).

242

SUBGLACIAL ENVIRONMENTS

132º

128º Portland Canal Alice Arm

124º

Observatory Inlet

Dixon

Entr

Douglas Channel Kitimat Arm Pitt I.

Hec

54º

ance

ate

Stra

it

Queen Charlotte Islands

54º

Princess Royal Island

Bella Coola

52º QUEEN CHARLOTTE SOUND Fitz Hugh

52º

Knight Inlet

Sound

Bute Inlet Quatsino Sound Rupert Inlet

50º

PA C I F I C

er

Isl

Alberni Inlet Barkley Sound

d

St

ra

it

of

St ra

Saanich Inlet it of J ua n de Fuca 124º

ia

128º

an

rg

132º

uv

50º Jervis Inlet Howe Sound

eo

48º

co

G

OCEAN

Powell Lake

Va n

48º

FIG. 8.31. Fjord Coast of British Columbia, Canada (reprinted from Syvitski et al., 1987).

Both forms are indicative of intense glacial downcutting and must be related to specific zones or lineations of fast ice movement either as outlet glaciers or ice streams within ice sheets. It has been suggested in the past that the location of both forms must in some way be related to bedrock influences

and control, specifically in the form of faults, shatter and mylonized zones (Augustinus, 1992). However, what becomes apparent on comparing geologic structure and tectonic zonation is that although geological control must have a strong influence upon trough orientation there is no evidence that this influence is

SUBGLACIAL ENVIRONMENTS

(a)

243

(b)

1700m CLEDDAU VALLEY

EIDFJORD FJORD GLACIOFLUVIAL ICE CONTACT DELTA

MITRE PEAK STIRLING BASIN

LAKE

EN

TR

AN

CE

SI

LL

EIDFJORD FJORD RAISED GLACIOFLUVIAL OUTWASH TERRACE

ENTRANCE BASIN

DEPTH (m)

0

0

10

5

15

20

SINGLE BASIN FJORD

100 200

km

SILL

ENTRANCE

25

0 400

0

20 EIDFJORD

40

60

80

km 100

120

140

160

180

MULTI BASIN FJORD

800 HARDANGERFJORD, NORWAY

STIRLING BASIN

MILFORD SOUND, NEW ZEALAND

FIG. 8.32. Schematic presentation of two of the world’s ‘classic’ fjords. (a) A simple basin fjord, Milford Sound, New Zealand; (b) a complex multi-basin fjord, Hardangerfjord, Norway (from Holtedahl, 1975; reproduced by permission of the Geological Survey of Norway).

anything other than secondary. Rather, pre-glacial geomorphic control in the form of pre-existing landscape dissection of pre-glacial topography determines to a greater degree the topographic lows that localized fast-moving ice masses may begin to follow and exploit, eventually developing into troughs. These trough are often described as U-shaped (Fig. 8.33). This characteristic cross-profile shape is not found in all troughs and in others sediment infill has disrupted the original cross-profile. However, it has long been held that an explanation of this typical cross-profile shape may hold a key to the understanding of the formation of glacial troughs. A comparison with fluvially eroded valley cross-profiles has been made to some degree in that it has been

suggested that the volume of ice moving through a glacial trough should be directly related to the drainage or ice evacuation area up-ice of the trough (Plate 8.14). Roberts and Rood (1984), using data from British Columbia, compared fjord length, width and depth with ice contributing area and found a significant relationship between fjord length and ice contributing area (r 2 = 0.81) (Fig. 8.34). While in Baffin Island, Sugden (1978) found a strong relationship between ice discharge and fjord cross-profile (r2 = 0.96). The catenary, parabolic or U-shape cross-section can be defined as Y = aXb, where Y is the vertical and X the horizontal distance, respectively, from the midpoint of the valley bottom and a and b are

244

SUBGLACIAL ENVIRONMENTS

(a)

(b)

PLATE 8.13. (a) Grenville Channel (fjord), British Columbia. (Photo courtesy B. C. Ferries.) (b) Princess Louise Inlet (fjord), British Columbia. (Photo courtesy of W. H. Wolferstan.) (c) Princess Louise Inlet (fjord) and Jervis Inlet (top left). (Photo courtesy of W. H. Wolferstan.) (c)

constants (Harbor, 1992; Harbor and Wheeler, 1992). It is assumed that a valley progresses from a V- shape (b~1.0) to a U-shape (b~2.20), then the value of component b is indicative of the stage of valley form development (Fig. 8.35). Harbor (1992) has shown, using an iterative finite-element model of glacial trough development, that, irrespective of the initial

valley form, over a period on the order of 1000 years a steady state, quasi-parabolic cross-profile can develop. The key element in the generation of the U-shaped profile was recognized by early workers that greater erosion must have occurred not at the base of the valley form but some distance laterally from the central lowest point thereby reducing over time the

SUBGLACIAL ENVIRONMENTS

A

245

A Unglaciated Valley A-A Unglaciated Valley

Bedrock A

A M

Maximum Vertical Ice Extent A-A Glacial Valley M-M Active Glacial Channel M-M Zone of Glacial Influence

M

Glacier Bedrock A

A Ice Extent Less Than Maximum A-A Glacial Valley M-M Zone of Glacial Influence B-B Active Glacial Channel

M M B B Glacier Bedrock A

A M

Bedrock

M

After Deglaciation A-A Glacial Valley M-M Zone of Glacial Influence

PLATE 8.14. Glaciated trough, head of Glen Muick, Cairngorm Mountains, Scotland. Snow in centre foreground beyond the loch highlights a zone of hummocky moraine possibly of Loch Lomond Stadial age.

Talus and Alluvium

FIG. 8.33. Schematic cross-sectional evolution of a valley resulting from glacial erosion (from Harbor, 1992; reproduced by permission of the author).

Length, Width and Average Depth (km)

elevation difference between the centre and adjacent areas. The location of glacial trough erosion seems now to be accepted but the actual mechanisms of subglacial erosion within a confined valley need to be further explored. As Harbor (1992) has pointed out, as ice thickness increases in association with greater trough depth there is a limiting point if a constant ice discharge and erosion rates are to be maintained. Since, for a given ice discharge, a specific surface gradient must be maintained, conditions at the subglacial interface are sensitive to changes in this gradient. Where changes in discharge and therefore surface gradient occur, it can be assumed that trough erosion rates will also vary (Harbor, 1992, fig. 11). The erosion of a glacial trough must occur owing to a combination of subglacial processes of abrasion, plucking and hydraulic erosion by debris-charged high-pressure meltwater. Unlike areal erosion beneath

100

10

0.50

L=

1.0

9A f

W = .2

0.1

0.01 10

.81A f

= AVD

0.27

A .013 f

0.39

Length Width Depth

100

1000

10 000

Ice Contributing Area (km2)

FIG. 8.34. The relationship between ice contributing area and fjord properties of length, width and average depth. Data from British Columbia (after Roberts and Rood, 1984; reprinted from Roberts and Rood, 1984, by permission of the Scandinavian University Press).

246

SUBGLACIAL ENVIRONMENTS 50

2

T=0

0

50

2 1 T=40

2 1 T=80

0 Valley Centre 2

Dimensionless Basal Velocity

0

50

50

and Erosion Rate

1

1 T=120

0

50

T=300

FIG. 8.35. Model results of a simulation of form development with erosion scaled to local basin velocity squared. The figures show the glacial valley cross-section at different time steps (T), with velocity contours on the glacier in units of 10 per cent of the maximum velocity for the section, and with the central contour being 90 per cent. The plots are dimensionless and are shown with no vertical exaggeration. The graphs show corresponding cross-glacier variations in basal velocities and erosion rates, in each case scaled to an average cross-section value of one (from Harbor, 1992; reproduced by permission of the author).

unconstrained ice sheets where sensitive and transient feedback fluctuations in the basal thermal regime in close association with basal meltwater occurs, this delimiting control on effective stress levels may not exist in confined troughs. In the basal and side-wall environments of glacial troughs where areal changes in basal temperature, for example, cannot be dis-

sipated or conserved and where fast basal ice motion occurs, it is likely that erosion processes are unconstrained. Under this scenario a ‘run-away’ style of basal erosion occurs that can only be limited where ice discharge or other external constraints upon ice velocity can reduce erosion (the back-pressure exerted by ice shelves or the buoyancy effect of a grounding-line). The effect of greater debris concentrations at the base of an ice mass may retard basal sliding especially in the presence of a deformable bed or ice mass divergence on reaching the coastline. A characteristic of many glacial trough systems is the presence of transverse trough sections (Plate 8.15). These transverse troughs appear to be related to erosion by diffluent lobes of ice crossing interfluves possibly representative of pre-glacial tributary valleys. The sill at the sea-entrance of many fjords would seem attributable in many cases to ice divergence and buoyancy effects on reaching the edge of coastal mountains and deep seawater conditions (Shoemaker, 1986a,b). In some cases subglacial diamictons are found on sill bedrock surfaces indicative of the limited erosion that may have occurred close to the final stages of glaciation in these troughs. The presence of riegel bedrock steps and associated basins in many glacial troughs remains a difficult problem. There have been many hypotheses put forward, for example, differential bedrock resistance, faulting, valley constrictions, localized freeze/thaw at the bed beneath heavily crevassed ice, and proximity to the entrance of tributary glaciers. 8.9.2.2. Tunnel valleys Tunnel valleys can be found in many marginal areas of once continental glaciated terrains, for example, Denmark, Germany, Poland, Britain, the North Sea Basin, southern Ontario, the Canadian prairie provinces and northern states of the USA. These valleys, sometimes known as rinnen, rinnentaler, tunneltaler, or tunneldalen in Europe, are often referred to as buried valleys. The term ‘valley’ is perhaps a misnomer since these forms usually do not have a dendritic drainage system, although tributaries and braided reaches do occur. These forms are deeply incised usually into bedrock and may better be termed as deep slits cut into the rock since they are usually

SUBGLACIAL ENVIRONMENTS

247

PLATE 8.15. Landsat image of interior Rocky Mountains showing typical glaciated landscape of linear glacial troughs often linked by cross-valleys (diffluent routes). This image includes part of the continental divide between Alberta and British Columbia. The main feature in the centre of the image from right to left is part of the Rocky Mountain Trench now containing the Fraser River. This trench is a major fault zone extending from Alaska to Montana. Snow covers only the highest mountains (> 3000 m) – many carrying glaciers (image taken July, 1975).

only a few tens of metres in width but sometimes over a hundred metres in depth. Exceptions do occur to these dimensions but they are rare. These valleys are usually infilled with glaciofluvial sediments with occasional diamictons present either as flow sediments or rafted sediment intraclasts (Fig. 8.36) (Brennand and Sharpe, 1993; Piotrowski, 1994). The effect of infilling is to bury the valley below ground surface. Additional glacial sediments may often overly the valley therefore the term ‘buried valley’. The long profile of these valleys is usually undulating with deep depressions separated by short or long shallow reaches. Tunnel valleys may extend

for tens of kilometres and exhibit reverse gradients to that of the ground or bedrock surface. These gradients have, in the past, been employed to suggest that these valleys were incised under hydrostatic pressures beneath an ice mass. Valley walls are smooth with evidence of meltwater erosional scour marks and P-forms often present. In some cases the remnants of older diamictons may occur in the deepest depressions. An association has been noted between tunnel valleys and eskers, and between drumlin formation, tunnel valleys, and subglacial hydraulics and thermal conditions (Mooers, 1989a,b; Brennand and Sharpe, 1993).

248

SUBGLACIAL ENVIRONMENTS

(a)

(b) FIG. 8.36. Cross-sections through the Bornh¨oved tunnel valley, northwest Germany. Stratigraphy: qe, Elsterian Glaciation; qhol, Holsteinian Interglacial; qs1, qs2, first and second ice advances of Saalian Ice (local terminology); qw1, qw2, qw3, first, second and third Weichselian Ice advances; suffix v, advance outwash; suffix r, retreat outwash; Btv ss., Bornh¨oved tunnel valley sensu stricto; Btv sl., Bornh¨oved tunnel valley sensu lato. Lithology: 1, clay; 2, silt; 3, fine sand; 4, medium sand; 5, coarse sand; 6, gravel; 7, diamicton. Vertical exaggeration 12.5× (after Piotrowski, 1994, reproduced with permission from Elsevier Science Publishers).

The location of tunnel valleys is far from obvious on the present day ground surface leading to problems in hydrogeology, in particular difficulties in waste disposal, and the undetected presence of local and/or linear and perched aquifers. Tunnel valleys appear to be related to subglacial ice dynamics and the evacuation of subglacial meltwater in large volumes (the Sable Island tunnel valleys cut across the Scotian Shelf off the east coast of Canada may have had volumes of ~0.45 × 107 m3 s–1 when all tunnels were operational (Fig. 8.37) (Boyd et al., 1988; Barnett, 1990; Rains et al., 1993). In some instances, as in the mid-west of the USA, tunnel valleys extend to the

margins of the Late Wisconsinan Ice Sheet margins and form outwash aprons within the terminal moraine complex (Mickelson et al., 1983), however in others areas such as Indiana and Illinois such an association has not been observed. Tunnel valleys form as a result of intense localized subglacial meltwater erosion within subglacial meltwater channels. The water in these channels is under hydrostatic pressure and is heavily charged with angular debris resulting in an efficient and high-speed erosive tool. The undulating profile and the watersmoothed walls are consistent with high-pressure but fluctuating meltwater discharges. The location of

SUBGLACIAL ENVIRONMENTS 62º

La

ur

P.E.I. 46º

of

y

nd

Fu

A OV

N

N TIA

O

(a)

Unnamed Basin

IA

OT

SC

SC

43º

tia

n

Ch

an

NEW BRUNSWICK

y Ba

en

ELF

SH

Emerald Basin

Middle Bank Brandal Basin

ne

l

46º

Misaine Bank

Georges Bank

8.9.3 Local Linear Erosional Forms

43º

59º 30'W

60ºW

44º 10'N

412 424 X

(b)

328 432 Y 384

SABLE

SAB 85

44ºN

D

ISLAN C67

margin, it is thought that subglacial meltwater can be evacuated toward the margin by crossing the He-bed and incising into the bed and producing tunnel valleys.

Banquereau

Sable Island Sable Island Emerald Bank STUDY LaHave Bank SITE Bank pe 200 m Slo l a m nt e 1,000 ntin m Co m 2,000 00 0m ATLANTIC OCEAN 3,0 4,00 62º

Browns Bank

249

Tunnel Valley

FIG. 8.37. Tunnel valleys near Sable Island, Nova Scotia (from Boyd et al., 1988; reprinted with permission from Macmillan Magazines Ltd).

these channels may in part be influenced by underlying topography but is generally thought to be controlled by ice pressures. Many tunnel valleys obliquely cross slopes or are cut into the side-wall of a larger glaciated valley, in both instances without regard for local gradients or topographic control (Booth and Hallet, 1993). The difference between these valleys and subglacial channels (discussed below) is probably only one of scale and the length of time taken to cut tunnel valleys. Conditions most suitable for valley incision appear to be related, in many cases, to frozen bed environments down-ice of a temperate bed thus probably close to an ice sheet margin. Owing to the volume of water up-ice under M-bed conditions, the thinness of the marginal ice and probably He-bed conditions prevailing close to the

Linear erosional forms are the products of direct icecontact erosion or channelled meltwater erosion, or a combination of both. The scale of these forms varies enormously from the micro-lineations of striae to large-scale forms such as roche mouton´ees. 8.9.3.1. Striae and associated percussion erosional forms Bedrock striations are perhaps one of the best indicators of glacial erosion (Plate 8.16). These scratches indicate the local vagaries of basal ice movement that occur as a result of variations in principal stress directions and micro-topographic bed control. For too long striae have been ignored as being of rudimentary significance and limited value. However, when time is taken to measure the typical crosscutting relationships of superimposed striae, long-axis depth variations and the presence of micro-faulting, striations can be a useful tool in deciphering ice movements, chronology and ice–bed conditions (Rappol, 1993; Kjaer, personal communication, 1999). Striae exhibit a wide variation in form from straight micro-grooves of a few millimetres in depth extending for only a few centimetres, to some that are almost a centimetre deep and extending for over a metre (Laverdi`ere et al., 1985). In some cases smaller striations occur at the base of large striations. Delicate tracery of cross-cutting striae commonly occur on boulder and clast surfaces and, where resistant knobs occur, striae may circumvent these obstructions. Many striae have a variable depth becoming shallower in the direction of ice movement, possibly indicative that the striating tool became blunt with travel down-ice. Striae on individual clasts and bedrock surfaces are products of either direct scratching of clasts and boulders held in the basal ice or clasts within a moving basal debris layer. The depth, width and

250

SUBGLACIAL ENVIRONMENTS

PLATE 8.16. (a) Striations visible on the stoss-side of large streamlined bedrock protrusion, near Espanola, Ontario. Scale card is 8.5 cm long. (Photo courtesy of Greg Hamelin.) (b) Striations of Pleistocene ice across a Precambrian diamictite of Huronian age, near Whitefish Falls, Ontario. Note coin 3.0 cm in diameter. (Photograph courtesy of Greg Hamelin.)

(a)

(b)

SUBGLACIAL ENVIRONMENTS

251

length of individual striations, therefore, reflect the combined influence of basal ice stress levels, ice velocity, meltwater presence, debris concentrations, effective stress levels, the sharpness of the individual clast indenter and the properties of both the indenter and the surface being scratched. In general, a striation is thought to be produced by plastic deformation of the rock surface against which the tool is applied. Often small leve´es are produced on either side of the main trough (Plate 8.17). In those cases where the difference in hardness between the indenter and the surface is small, a shallow striation without distinct leve´es tends to be produced. In some cases the track of an indenting tool producing a striation reveals that the tool did not scratch continuously across a surface but rather moved in a percussive manner, producing a series of subparallel transverse percussion gouges (Plate 8.18). A final aspect of striation formation is that although the indenter must be sufficiently sharp and be able to hold a cutting edge for some distance, a striation can only occur if the surface of the clast or bedrock can retain the scratch mark, thus soft materials and those easily weathered do not carry striations.

Although not linear forms of glacial erosion, there exists a suite of local surface erosional forms associated with striations that can be considered along with striae; these include chattermarks, crescentic gouges, lunate and conchoidal fractures and friction cracks (Laverdi`ere et al., 1985). These forms are sharp-edged with surface evidence of conchoidal fracturing usually related to percussive action of an indenting boulder or clast. The effect of percussion tends to be localized, therefore these erosional forms occur across small areas of a bedrock surface. Gouges and cracks typically are only a few centimetres wide and millimetres in depth. These forms occur in groups or sets and can be used to estimate approximate ice direction. It is likely that cracks and gouges of this type may be the site of later subglacial meltwater flow disruption thereby acting as instigators of larger Sand P-forms.

(a)

(b)

8.9.3.2. Roche moutonn´ees Roche moutonn´ees may be regarded as streamlined bedrock hills. Typically, these forms have a smooth stoss-end and a precipitous lee-end slope (Plate 8.19).

PLATE 8.17. (a) SEM photomicrograph of striations (Scale 2 cm = 50 ␮m, magnification 200×). (b) Higher magnification SEM photomicrograph of striation showing upturned edges (Scale 2 cm = 3.33 ␮m, magnification 3000×).

252

SUBGLACIAL ENVIRONMENTS

(a)

(b)

PLATE 8.18. (a) A series of chattermarks across a bedrock surface, view in the down-ice direction. Note coin, centre-left, for scale (3.0 cm diameter). (Photo courtesy of Greg Hamelin.) (b) Line of chattermarks across large bedrock surface, view in the up ice direction. Note ruler for scale (25 cm in length) (photo courtesy of Greg Hamelin); both photographs from near Whitefish Falls, Ontario.

These bedrock forms may range in scale from a few centimetres high to hills several hundreds of metres in elevation. The formation of a roche moutonn´ee would appear to be partially the result of ice erosive moulding on the up-stream end of the form and glacial plucking on the lee side. In order that a lee side be affected by plucking action it is probable that no cavity of any size formed in the lee of the obstruction. As a consequence, tensile stresses and freeze–thaw processes in combination create a plucked lee face. Rastas and Sepp¨al¨a (1981), in an extensive survey of roche moutonn´ees in southern Finland, noted the close relationship between bedrock joint systems, fissures and the size and morphology of individual forms (Fig. 8.38).

8.9.3.3. Meltwater erosional forms As ice pressure melts on the up-ice side of bedrock and other topographic obstacles, meltwater is released that may flow as turbulent, debris-charged, meltwater streams under hydrostatic pressure. Where large subglacial lakes form beneath ice sheets, it has been suggested that vast, regional, catastrophic floods may periodically occur resulting in enormous discharges of meltwater across the subglacial bed (Shaw et al., 1989; Shoemaker, 1992a,b). Likewise j¨okulhlaups may provide sudden catastrophic meltwater drainage. Several meltwater-associated erosional forms, under H-bed conditions, are typically found, for instance potholes, P-forms and subglacial meltwater channels.

SUBGLACIAL ENVIRONMENTS

253

(a)

(b) PLATE 8.19. (a) Small roche mouton´ees in northern Lake Huron, Ontario. Direction of ice flow from left to right. Buoy in centre approximately 1 m in height. (b) Roche mouton´ees in the proglacial zone of Nigardsbreen, Norway. Direction of ice flow from right to left. The roche mouton´ees are approximately 25 m in length.

254

SUBGLACIAL ENVIRONMENTS N

(a)

N

E W

E W º

E

A

S

S

º 60

B

º 90

90

º

90

º

60

60

º

º

30

30

30

º

º

W

N

S

C

N

(b)

º 90 º 60

º 30

E

W

S No striation

Plucked surface

Zone of striae

Edge of plucked area

Facet zone

Zone of friction cracks

FIG. 8.38. Roches moutonn´ee surfaces showing: (a) stereographic diagrams of: A, striated bedrock surfaces with striae directions; B, surfaces with friction cracks, lines indicate the direction normal to cracks; and C, surfaces truncated by plucking. (b) Contoured stereographic diagram of the orientation of 232 polished facets. On all diagrams arrows indicate general direction of ice movement (after Rastas and Sepp¨al¨a, 1981; reproduced by courtesy of the International Geological Society).

These forms develop in association with other erosional forms such as rock drumlins, roche moutonn´ees and crag and tails. There are certain characteristics that these forms have in common; smooth, rounded bedrock surfaces, channels with fluctuating longitudinal thalwegs, limited development of channel networks and the siting of many of these forms without apparent regard for present topographic position, slope or gradient (Kor et al., 1991). For example, in terms of topographic siting channels may cross hillsides at angles inappropriate for ‘normal’ drainage pathways and may traverse topographic

divides or interfluves without regard to channel slope. (1) Subglacial meltwater erosional forms In many glaciated areas, exposed or recently uncovered bedrock surfaces carry distinctive erosion marks that are finely detailed, often tortuous, smoothed depressions and hollows of a wide range of morphologies (Plate 8.20). These forms would appear to result from intense subglacial meltwater erosion either the result of high-pressure turbulent debris-charged melt-

SUBGLACIAL ENVIRONMENTS

255

(a)

(c)

(b)

(d)

PLATE 8.20. (a) Straight small-scale flutes near Wilton Creek, Ontario. (Photo courtesy of John Shaw.) (b) P-form in bedrock near Espanola, Ontario. Scale card is 8.5 cm long. (c) Scallops, subglacial meltwater erosional marks near Wilton Creek, Ontario. (Photo courtesy of John Shaw.) (d) P-form cutting across a bedrock surface near Espanola, Ontario. Note conchoidal fracture in lee side (far centreleft) and iron staining (top left). Scale card is 8.5 cm long. (Photo courtesy of Greg Hamelin.)

256

SUBGLACIAL ENVIRONMENTS

water or, in some cases, high-pressure fluidized sediment slurries. These forms were first studied in detail in Scandinavia where the term P-form was used to denote that these forms were ‘plastically formed’. The continued validity of this term is in some doubt and perhaps ‘S-forms’ as a more exact term representing ‘sculpted forms’ may be now appropriate (Kor et al., 1991) (Fig. 8.39). In terms of scale these forms most typically are small, only a few metres or less in length and width but occasionally very large-scale multiple associations of these forms exist stretching over tens of metres of bedrock surfaces (Fig. 8.40). The debate as to whether S-forms develop because of a meltwater dominated flow or a fluidized debris

slurry still remains but convincing evidence of the morphology and continuity of one form with another produced under high-pressure, high-velocity flow conditions seems to negate most forms being produced by slurried sediments. These erosional forms appear to develop owing to meltwater flow separations over distinct time periods related, in the first instance, to defects or bed irregularities or up-flow vortices impinging on the bed (Shaw, 1988b,c) (Fig. 8.41). As the fluid passes across the bedrock surface, cavitation, fluid-stressing and corrasion processes act upon the rock surface as erosional wear processes, the former being likely the most important. A typical association of S-forms is shown in Figure 8.42 across

Transverse Forms

Nondirectional Forms

(d) TRANSVERSE TROUGH (a) MUSCHELBRUCH

r

(h) UNDULATING SURFACE

rs ss lf

STOSS-SIDE FURROW

(i) POTHOLE

Longitudinal Forms

(b) SICHELWANNE

r

(e) SPINDLE FLUTES

mr lf

ndle

Open Spi

(g) FURROW (c) COMMA FORM Closed

Spindle

(f) CAVETTO

FIG. 8.39. Examples of S-forms as identified in the French River complex, Ontario. Flow direction is from left to right in all but nondirectional forms (reproduced by permission of the authors (Kor et al. 1991), from Canadian Journal of Earth Sciences, 28).

SUBGLACIAL ENVIRONMENTS

257

LEGEND Occurrence of S-forms Paleoflow direction (measured) Paleoflow direction (from airphotos)

0

10

20 kilometres

FIG. 8.40. Northeastern section of Georgian Bay, Ontario, showing observed distribution and orientation of S-forms (reproduced by permission of the authors (Kor et al. 1991), from Canadian Journal of Earth Sciences, 28).

the stoss-side of a rock drumlin. Kor et al. (1991) have classified these forms into: (a) transverse Sforms (muschelbr¨uche, sichelwannen, comma forms and transverse troughs), (b) longitudinal S-forms (spindle forms, cavettos, stoss-side furrows and furrows), and (c) non-directional S-forms (undulating surfaces and potholes). Potholes of variable depth and diameter appear to develop from vertical meltwater vortices impinging upon a rock surface and drilling into the rock surface by a combination of corrasion and, once a hollow has formed, by the grinding action

of boulders caught in the vortex of swirling meltwater. The location of S-forms appears to be related to a complex interaction between topography, bedrock structures and fissure geometry, subglacial stress fields and subglacial meltwater hydraulics and evacuation routes. Typically, sichelwannen and most S-forms exhibit an independence from topographic control. Sichelwannen often occur preferentially on the stoss-side of bedrock forms such as roche moutone´es and rock drumlins but are found in distal

258

SUBGLACIAL ENVIRONMENTS horseshoe vortex

neck

obstacle ge

ne

ral

obstacle

remnant ridge (rat-tail)

flow

vortex

furrow

furrow

(a)

(c)

flow line accelerated erosion surface lowering

emerging obstacle

(b)

FIG. 8.41. (a) Conceptual model of the formation of crescentic furrows and furrows to the lee-side of an obstacle related to a horseshoe vortex. A remnant ridge (rat tail) is produced. (b) and (c) Models of the formation of furrows related to meltwater flow around an emerging obstacle. In (b) the incipient furrows occur at zones of flowline convergence. (c) Extended furrows, related to two independent vortices, converge on a neck upstream of the obstacle (from Sharpe and Shaw, 1989, reproduced by permission of the authors).

us FLOW

tt sf s

f

Spindle

cf

c

m

(a)

ssf

(b)

Muschelbruche

(c)

Sichelwanne

(d)

FIG. 8.42. (a) Schematic diagram of S-forms distributed on rock drumlins. Flow away from the reader. Abbreviations: tt, transverse troughs; s, sichelwannen; m, muschelbr¨uche; ssf, stoss-side furrows, cf, comma form; us, undulating surface; c, cavetto; f, furrow; sf, spindle furrow. Inferred relationship between form and flow structure: (b) spindle flutes form as a result of low-angle vortex impingement; (c) muschelbr¨uche are formed by higher-angle vortex impingement; (d) sichelwannen are formed when flow separation beyond the rim produces a roller eddy and vortices (reproduced by permission of the authors (Kor et al. 1991), from Canadian Journal of Earth Sciences, 28).

SUBGLACIAL ENVIRONMENTS

(a)

259

(b)

PLATE 8.21. (a) Large pothole, Finland. Note ‘roller’ stones on edge. For scale note measuring tape top centre. (Photo courtesy Geological Survey of Finland.) (b) Small assymetric pothole cut into bedrock wall, Norway.

flanks of some larger bedrock forms and may on occasion be related to joint-specific sites (Gray, 1981; Shaw and Sharpe, 1987). Potholes often occur where major breaks of slope exist across a longer bedrock surface (Fig. 4.42). In the past it was thought that potholes might be associated with surface crevasses impinging on the bedrock below the ice but this seems unlikely since crevasses rarely reach bedrock and if so rarely remain at one site for any length of time (Plate 8.21). (2) Subglacial meltwater channels Beneath ice sheets and valley glaciers where meltwater flows in channels, numerous examples of channels cut in bedrock and sediment can be observed. These channels, whether under active or passive ice, can be distinguished on the basis of several unusual charac-

teristics that are inconsistent with subaerial channels: (a) independent of topography, (b) undulating long profile, (c) lack of a significant drainage basin, (d) abrupt inception and termination, (e) proximity to other channels with no fluvial connection, (f) crosscutting relationship with other subglacial and subaerial channels, (g) occasional association with eskers and glaciofluvial deltas, (h) abrupt changes in channel direction, and (i) general lack of tributaries (Fig. 8.43). In general, subglacial meltwater channels can be classified according to their position beneath an ice mass and in relation to specific topographic locations and association with certain glaciofluvial sediments and deposits (Table 8.10). Beneath ice sheets only icedirected subglacial channels are likely to occur showing little or no dependence on topography and in marginal areas subglacial/proglacial channels can be

260

SUBGLACIAL ENVIRONMENTS

A B mause F C D

G E

(a)

N

Channel Wall in Drift Channel Wall in Rock

Small Meltwater Channels Cliffs

Contours in feet 0

Scale in miles

1

(b) FIG. 8.43. (a) Aerial photograph of area of gradient subglacial/submarginal meltwater channels near Blairgowrie, Scotland. (b) Map of the same meltwater channels as shown in (a).

found. In valley glaciers or where ice lobes of ice sheets have been confined, both submarginal and marginal channels occur. The distinction between channel types is at times subtle, especially since channels may have been exploited repeatedly during

active and passive phases and modified or had sediments deposited within the channel itself. Meltwater channels can range in size from barely a metre deep and wide to huge valleys tens of metres deep and as much as a kilometre wide. Channels may

SUBGLACIAL ENVIRONMENTS

261

FIG. 8.43. (c) Distribution of meltwater channels and glaciofluvial deposits in Nithsdale, Scotland. (After Stone, 1959; reproduced with permission of the Royal Scottish Geographical Society.)

extend for only a few metres before terminating while others extend over tens of kilometres. They exhibit incised meanders, bifurcations and complex stream patterns (Fig. 8.43), have multiple intakes at the same or differing elevations, and may have long straight

reaches, discordant junctions and tributary entries. Some channels are exceptionally short in length yet very deep, often seen occurring on valley spurs, drumlin crests, cols and other topographic highs. These channels have been interpreted as evidence of

262

SUBGLACIAL ENVIRONMENTS

TABLE 8.10. Meltwater channel types Channel type Closed channel Subglacial

Submarginal

Open channel Marginal

Some typical characteristics

Deeply incised aligned roughly parallel to the main ice direction, limited topographic control, oblique to topographic gradient, in valleys close to valley bottom, often reverse gradient, may be associated with submarginal channels as chutes in lower reaches with engorged eskers, may cut valley-side spurs or cross-interfluves at high levels from one valley to the next Normal general gradient less than terrain gradient, some topographic control, may intersect or grade into marginal channel system, or alter to subglacial chutes flowing directly down-slope, typically run parallel or near-parallel to ice mass margins (or coincident with trim lines) Normal or steep gradient, some topographic control, may intersect with proglacial channel systems, associated with kame terrace deposits

Subglacial/proglacial (land-based)

Topographic control in lower reaches, associated with glaciofluvial outwash fans

Subglacial/proglacial (subaquatic)

Topographic control in lower reaches, associated with glaciofluvial deposits subaquatic outwash fans and deltas

englacial meltwater channels touching the subglacial bed or being superimposed upon parts of the glacier bed for only short distances. Subglacial channels and submarginal channels classified as ice-directed forms are closed systems that may be under hydrostatic pressure, if far enough back from the ice margin or, if connected to an open portal, may be only under atmospheric pressure (Chapter 4). The dimensions of these channels, however, are regulated by the meltwater volume, temperature and debris content, and both ice wall closure rates and squeezing of fluidized sediment into the channel space (for a detailed discussion on channel persistence see Menzies, 1995, chapter 6). In places where subglacial channel meltwater is under hydrostatic pressure, the characteristic undulating long profile, incised meanders and strong independence from topographic control can occur leading to channels with reverse gradients, cross-cutting topographic slopes and passing from one valley system to another at high elevations. Under these circumstances full pipe-flow conditions prevail and active scouring of the bed is likely to occur rapidly (S-forms, tunnel valleys, potholes etc.). Pressure or discharge fluctuations in meltwater conditions are likely to lead to rapid sedimentation within the

channel itself leading to glaciofluvial sedimentation and esker formation (Clark and Walder, 1994). The complex pattern of many of these channels may in part be due to fluctuations in basal ice pressures and the effects of localized channel wall and ceiling collapse or slumping of debris from meltout along channels, walls and ceilings temporarily blocking channel routes. It is also possible that many channel patterns are products of multiple phases of subglacial meltwater activity utilizing similar routeways across the subglacial bed or are the result of englacial superimposition creating overprinting of one channel system on another leading to channel configuration and morphological complexity. Submarginal subglacial channels appear to occur in most valley glaciers (Lliboutry, 1983; R¨othlisberger and Lang, 1987) where overdeepened sections or bedrock constrictions can be bypassed at higher levels leading to channels flowing close to valley sides. These channels are similar to marginal channels in both their close proximity to these open channels and in relation to topography. Often these channels are intersected by later marginal channel suites. Submarginal channels may grade into marginal channel systems or abruptly turn down-slope as subglacial chutes. As subglacial chutes, eskers and fan deltas

SUBGLACIAL ENVIRONMENTS

occasionally lie close to their abrupt termini. Rarely do these chutes grade into subglacial channel systems but appear to be connected to englacial channels. The extent down-slope of chutes may, as has been speculated, be related to the presence of the englacial piezometric water surface within the ice mass. This same surface level has been termed an englacial water table (Sissons, 1976; Gray, 1991) and may be related to the upper elevation of glaciofluvial sedimentation in valleys. Marginal and subglacial/proglacial channel types are open channel systems often with only enclosing ice walls for short distances or, in the former type, only along the down-slope side of the channel. Where subglacial meltwater channels enter bodies of water, jets of turbid meltwater have been reported with occasional fountains and upwelling at the ice margin illustrative of the artesian effects resulting from hydrostatic pressure release at the portal mouth. The terrestrial channel types develop under open channel flow conditions with only glacial ice controlling topographic position and channel gradient. Typically, marginal channels in confined valley sites reveal the three-dimensional retreat of ice masses both downslope and up-valley resulting in a suite of subparallel channels being formed as each successive higher channel is abandoned in favour of a lower, new channel at the ice’s edge (Fig. 8.43). 8.10. SEDIMENTS AND LANDFORMS OF PASSIVE ICE FLOW 8.10.1 Subglacial Sedimentation Processes The processes and mechanisms that lead toward sediment deposition beneath a passive land-based ice sheet can be subdivided into those processes producing (a) meltout tills (diamictons) and (b) glaciofluvial and glaciolacustrine sediments. Under passive or stagnant ice conditions, an ice mass begins to decay in situ resulting in the production of vast quantities of meltwater. Stagnation would appear to occur under two principal conditions. First, total in situ decay of an ice mass, wholly or in part, may occur within confined valley systems where the ice supply to parts of the glacial system has been cut-off once the regional glacier surface descended below a certain

263

elevation; or secondly where, in active retreat, portions of the frontal sections of an ice mass become detached from the main ice mass as a result of burial under massive volumes of debris thus leading to buried ice stagnation. 8.10.1.1. Tills – diamictons 1 Meltout tills: the fundamental characteristics of meltout tills have already been described in Section 8.5.1.1. Meltout tills can occur under both active and passive ice conditions. At present, discrimination as to under which ice conditions particular meltout tills were formed are difficult, if not impossible, without confirmatory evidence in the form of stratigraphic or facies associations. 2 Flow and waterlain tills: under stagnating conditions, where high porewater contents prevail, flow tills are released within subglacial cavities and from channel walls. These tills interfinger with meltout tills and may become indistinguishable from the meltout till package. Where massive stagnation occurs in subglacial/subaquatic environments at the grounding-line an increase in sediment rain-out causes large volumes of waterlain tills to be deposited. 8.10.1.2. Subglacial glaciofluvial and glaciolacustrine sediments Under conditions of passive ice decay large quantities of meltwater produce large outpourings of stratified sediment into the subglacial and eventually the proglacial systems. Within the subglacial system, cavities and channels are likely to become infilled and choked with stratified sediments exhibiting a wide range of grain size and depositional settings. 1 Glaciofluvial sediments: stratified sediments are found in the subglacial environment in a variety of locations and forms. Sheets, lenses and small inclusions of stratified sands and gravels are found intercalated within diamicton sequences (Fernlund, 1994; Van der Meer et al., 1994) (Plate 8.22). It is likely that these glaciofluvial sediments are formed intermittently whenever the ice separates from its bed, where tunnels develop then eskers may form,

264

SUBGLACIAL ENVIRONMENTS

PLATE 8.22. (a) Glaciofluvial sediments overlying a tectonized glaciolacustrine unit, near Rosenheim, Bavaria. Note rapid shifts in sediment type from coarse gravels to fine sands, visible as crude bedding. Section is approximately 20 m in height. (b) Delicate bedding structures in glaciofluvial deltaic sediments, Allen Park, Hanover, Ontario. Note load structures, dropstones and clast casts in centre, sharp contact between coarse and fine sands at base and rapid transition from fine sands to gravel at top of photo. (Photo courtesy of Greg Hamelin.) (c) Fine-grained planar bedded sands with smaller assymetric cross-bedding units in ice-contact delta, Allen Park, Hanover, Ontario. Note near upper part of the section a large channel infill gravel unit. (Photo courtesy of Greg Hamelin.)

(b)

(a)

(c)

SUBGLACIAL ENVIRONMENTS

but elsewhere subglacial openings, many of a temporary nature, may develop that permit meltwater flow, sediment transportation and thus sediment sorting. In most cases these glaciofluvial sediments exist as sand stringers, lenses and occasional small spreads within the more general depositional environment. In the field it is often difficult to tell if stratified sediments within diamictons are necessarily of primary subglacial origin. Where large sheets of glaciofluvial sediments occur that are not related to esker deposition it is more likely that such sediments are not subglacial in origin (Menzies, 1990c). Stratified sediments appear in many instances to grade into finer-grained diamictons thus indicating a gradational switch from one form of subglacial process to another (Plate 8.23). 2 Glaciolacustrine sediments: Subglacial glaciolacustrine sediments are common but usually of small dimensions and areal distribution. Typically, these sediments are laminated and often grade into surrounding fine-grained diamictons. As in the case of glaciofluvial sediments, laminated sediments can be deposited where a small cavity forms through ice–bed separation and a pool or small lake develops (Huddart, 1983; Morawski, 1988). In many cases it is almost impossible to resolve whether the glaciolacustrine sediments are of primary origin or if they have been rafted or otherwise incorporated into a subglacial sediment.

265

(a)

8.10.2. Subglacial Landforms As a consequence of the dominance of meltwater action under passive ice conditions, subglacial landforms are characteristically glaciofluvial in composition with intercalated tills present. These forms can be classified into two groups: (a) eskers, and (b) random ice disintegration forms such as kames, De Kalb mounds and Veiki and Sevetti moraine. 8.10.2.1. Eskers Eskers are straight to sinuous ridges of glaciofluvial gravel and/or sand that vary from a few tens of metres to over 100 km in unbroken length (Plate 8.24). The ridges locally range from a few metres to over 50 m in height and from less than 50 m to hundreds of metres

(b) PLATE 8.23. (a) Large sandy diamicton inclusion within a sandy, banded stratified sediment unit near Clyde, New York State. Note sharp contacts but also gradational sub-facies unit surrounding inclusion. Scale card is 8.5 cm long. (b) Gradational contact zone between a lower diamicton and overlying stratified sand. Note banding in diamicton and bedding in sand, the latter prominent owing to heavy mineral preferential deposition. Several clay ball inclusions possibly intruded in a line above major contact. Pencil for scale is 15 cm long. Photo from near Niagara-on-the-Lake, Ontario.

266

SUBGLACIAL ENVIRONMENTS

(a)

(b) PLATE 8.24. (a) Typical winding esker ridge, Northwest Territories, Canada. Note arrow indicating person for scale. (b) Steep-sided, meandering esker in north-central District of Keewatin, Northwest Territories, Canada.

SUBGLACIAL ENVIRONMENTS

in width at their base. Eskers may be transitional to large ice-contact stratified drift complexes deposited by interlobate meltwater streams, and this is reflected by ambiguous references to the origin of large, linear glaciofluvial complexes in the literature (Brennand and Sharpe, 1993; Clark and Walder, 1994; Warren and Ashley, 1994, 1997). Esker ridges generally trend parallel to the final direction of ice flow in the area where they are preserved, suggesting that although they may be formed by meltwater drainage throughout a glaciation, only those formed during the latest stages of glaciation are preserved. Notwithstanding the above observation, eskers can be seen to cross ice-flow features such as drumlins at sharp angles, suggesting that regional ice flow was, at best, sluggish during their formation (Plate 8.25). Eskers tend to occupy the lower portions of a landscape. Where esker ridges pass from one depression or valley to another, they are draped, with no apparent deformation, over divides or cols. This observation has led to the concept that eskers are formed in closed subglacial tunnels, in streams flowing full and under hydrostatic head. This model dates from the nineteenth century and is widely accepted today as the principal environment of esker formation. Observations of modern glaciers show them to be honeycombed by tunnels, passages, crevasses, etc., conduits that pass through the ice mass in various orientations from vertical to subhorizontal. During the melt season, in the ablation zone, these conduits are commonly full of water, and those at or near the base of the glacier are under a hydrostatic head. Meltwater in conduits that reach or pass near the base of the glacier become charged with mud melted out of the glacier’s debris-rich basal layers. The finer portions of the mud are carried through and out the mouth of the conduit into proglacial streams or directly into lake or marine water, depending on whether the glacier is retreating in the sea, up gradient (allowing unimpeded proglacial drainage), or down gradient (blocking drainage and forming proglacial lakes). The coarser components are largely trapped in the conduit, partially or completely filling it, and are left as sandy or gravelly esker ridges when the glacier melts away (Plate 8.26a,b).

267

Where glacier ice is retreating in a proglacial lake or marine waters, any coarse material (silt to gravel) that exits the conduit mouth is quickly deposited and may form a fan or blanket of varying thickness over the part of the esker ridge exposed by ice recession. This blanket may not alter the morphology of the esker greatly if the recession is rapid and steady. However, where eskers were deposited in marine environments, isostatic uplift may have reworked and removed this blanket and altered the original form of the esker by wave erosion (the same process affects eskers exposed in the shallower parts of proglacial lakes). Where ice retreat is interrupted by stillstands of the ice front, however, significant piles of sand and gravel are commonly dumped at the conduit mouth, forming subaqueous fans that all but obscure the underlying esker (Henderson, 1988). If the water is sufficiently shallow, the sedimentation rate sufficiently high, or the stillstand sufficiently long, the subaqueous conduit-mouth deposits could build up to the water surface, forming an ice-contact delta, which likewise may obscure the esker locally. Where such interruptions of the regular retreat patterns occurred around the shrinking Keewatin (Canada) Ice Sheet, short eskers formed in profusion between the larger systems for which the stillstands were represented by fans or deltas that interrupt and bury the main esker ridges. Since the diamicton adjacent to eskers represents the total basal load of the glacier that deposited the esker, and since individual till sheet thicknesses rarely average more than a few metres, it is evident that some process(es) must enhance sediment accumulation in esker conduits. Since meltwater flowing through glacial conduits is generally at a temperature above the pressure melting point of the ice in which the conduits formed, and since frictional heat is generated by the fast-flowing water-sediment suspension, tunnel walls are subjected to melting, thus enlarging the meltwater cavity. If a tunnel becomes too large, it is liable to collapse, but because the basal ice of a glacier can deform plastically, as the tunnel walls melt, ice flows from the sides maintaining a conduit-cross-section that allows the tunnel to stay open. As ice flows toward the conduit, it brings basal debris with it so that new material from lateral sources is constantly being dumped into the conduit,

268

SUBGLACIAL ENVIRONMENTS

(a)

(b) PLATE 8.25. (a) Esker superimposed at almost a right angle on drumlins, west of Rankin Inlet, southern District of Keewatin, Northwest Territories, Canada. (b) Esker draped and the nose of a drumlin in northern Manitoba, Canada.

SUBGLACIAL ENVIRONMENTS

269

(a)

(b) PLATE 8.26. (a) Ice-cored esker ridge melted out of glacier on Bylot Island, Canada. (Photo courtesy of C. Zdanowicz, summer 1992.) (b) Close-up photograph of same esker ridge. (Photo courtesy of C. Zdanowicz, summer 1992.)

270

SUBGLACIAL ENVIRONMENTS

sorted, and deposited in or carried out of the conduit to be deposited ultimately in proglacial outwash streams, lakes or marine waters. By this process, the esker grows to be orders of magnitude thicker than it would be if it were just sweeping out the debris contained in basal ice into which the tunnel was excavated. As a consequence of lateral movement of ice toward conduits, eskers often contain components from sources that are located as far as several kilometres off the final location of the esker ridges. Regional striation patterns also can be deflected toward the conduit (Repo, 1954), and the area adjacent to the esker ridge may be impoverished in drift cover. The conduits in which eskers form are regarded by many authors to exist only in a zone a few kilometres back from the retreating ice front. The water that fills the conduits is thought to be derived largely from the ice surface and is thought to reach the glacier’s base through crevasses kept open by thermal melting, even in zones of compressive flow. At the height of the meltwater season, conduits and their surface connections are so full of water that rain or snow storms ‘overload’ the system, causing water to back up and erupt from the glacier’s surface and sides as fountains of water under hydrostatic pressure. The fountains generally carry sediment-charged basal water, which is conspicuously muddy compared with normal surface drainage. Very few eskers have been observed melting out from modern glaciers, perhaps because those in the Arctic are not particularly sediment-rich, and modern retreating glaciers are so active that conduits are not easily preserved in one place for long enough for lateral flow to build up significant deposits. Conduits in the snouts of glaciers on Bylot Island, Arctic Canada, were observed to be abandoned after functioning for a melt season or two, leaving behind an ice cored, 200 m long, meandering sediment ridge in front of rapidly retreating glaciers (Plate 8.26a,b). The total sediment content of these ridges is so slight that, were the ice core to melt, a short sinuous ridge less than 1 m high would remain. Esker patterns: the spatial relationships of the substantial eskers left behind by the great Fennoscandian

and North American continental ice sheets fall broadly into four groups. 1 The outer parts of the ice sheets, particularly where located on unmetamorphosed sedimentary rocks, have rare, isolated esker segments, often located in tunnel valleys. Because eskers do occur profusely on some sedimentary terrain, however, their paucity in the outer reaches of continental ice sheets may have more to do with ice dynamics in those regions than with the nature of the glacial substrate (see discussion in Clark and Walder, 1994). 2 In areas of high relief, such as the Appalachian Mountains of eastern North America or the Cordillera of western Canada, eskers tend to be confined to major valleys. Although an individual esker may cross a divide to pass from one valley or depression to another, these eskers generally present an areal pattern reminiscent of the modern drainage but, generally, lack lower order (smaller) tributaries. 3 In areas of relatively low relief around ice divides of continental ice sheets, eskers form a radiating pattern of ridges trending generally in the main direction of late-glacial ice flows (Fig. 8.44). Major trunk eskers are joined by tributaries forming integrated sets representing Horton drainage systems as complex as fourth-order (Fig. 8.45). Though individual esker ridges tend to follow the lower part of the landscape, as in mountainous regions, they commonly cross divides and have numerous up-gradient reaches. In the case of the esker systems radiating from the Keewatin Ice Divide, the pattern of esker drainage of the ice sheets bears virtually no resemblance to that of the major elements of modern drainage (Fig. 8.44). 4 A fourth esker pattern that is common in the central sector of the Laurentide Ice Sheet is that of a dense cluster of individual and simple dendritic esker systems that terminate at major end moraines that define readvance lobes or ice streams that were active during the final stages of retreat of continental ice sheets (Fig. 8.46). These eskers are clearly confined to lobate areas that also are characterized by distinctive drift compositions and landforms.

SUBGLACIAL ENVIRONMENTS

271

1 68º 20º

86º68º

Great Bear Lake

Southhampton Island

Goats Island

Great Slave Lake

HUDSON BAY

Esker system Study Area 0

200 km

58º 94º

58º 104º

FIG. 8.44. Eskers radiating away from the Keewatin Ice Divide of the Laurentide Ice Sheet, Canada. Northeast-trending area devoid of eskers and passing through Baker Lake and west end of Wager Bay is the ice divide (modified from Shilts et al., 1987).

63º 104º

Rennie Lake Esker System

0

esker meltwater channel (includes lakes)

100 km 104º 60º

probable connection

FIG. 8.45. Detail of the Rennie Lake esker system showing high order tributaries (from Shilts et al., 1987).

272 60º

SUBGLACIAL ENVIRONMENTS 112 º

Clay plain (?: extent unknown)

60º

edg

Cochrane till

e

Esker

HUDSON BAY

of

Moraine

C

na di an

a

P Sh

r ne m ha eca r

P

oz

ield

br

oic ian rock roc k

JAMES BAY

72º

52º

Lake Winnipeg

0

200

400

iel

72º

Sh

r

Ca

na

di

an

Lake Superio

d

kilometres

ke

La

e

of

r e Onta La k

ron

Lake Michigan

Hu

edg

io

rie e E L ak

FIG. 8.46. Relationship of eskers and end moraines associated with lobate readvances, surges or ice streams active during deglaciation in south-central Canada. Some eskers may be buried under lacustrine or marine silty clay or Cochrane Till (from Shilts et al., 1987).

Though numerous studies of the internal structures and textures of eskers and associated glaciofluvial sediments have been carried out, few sedimentological criteria have been advanced as being particularly diagnostic of eskers. This partially relates to the extreme variability of the debris that glaciers carry and which serves as the source of esker sediments. Also, since debris is delivered directly into subglacial streams by ice, some components of esker deposits, for example large boulders and diamicton rafts, are out of equilibrium with the hydraulic characteristics

of the former esker streams. These, almost random, sedimentological characteristics make it almost as difficult to generalize about the dynamic sedimentology of eskers as it is to generalize about the sedimentological characteristics of the subglacial diamicton from which eskers derive the bulk of their components. In lieu of definitive sedimentological clues as to esker genesis and hydrology, the form of eskers and associated sediments has been the most effective characteristic in indicating their origin. The ridge-like

SUBGLACIAL ENVIRONMENTS

form of eskers indicates the presence of ‘temporary’ ice walls or tunnels and their tendency to be draped over topography with little change in form confirms the former existence of a closed hydrological system at the base of the ice (Plate 8.26b). The areal pattern of esker distribution also can be used to deduce the hydrology of esker systems. If conduits in which eskers are actively deposited extend only a few kilometres or a few tens of kilometres back from the ice margin at any one time, then it is logical to assume that the dendritic patterns relate to either large-scale basal drainage or to traces of surface drainage of a dying ice sheet. The definitive study of the relative importance of esker meltwater sources has yet to be made. 8.10.2.2. Non-directional subglacial forms and ice pressed forms Within this group of subglacial forms generated in the basal zone of an ice mass are forms such as Pulju, Veiki and hummocky moraines. Considerable debate continues as to the origin of these forms and the location in which the forms are created (Lagerb¨ack, 1988). The composition of these forms is varied. Some forms are composed of lodgement and meltout tills, and stratified sediments in distinct layers, while in others evidence of thrusting, diapiric action and deformation by both active and passive ice gravitational consolidation can be found (Aario, 1990). The location of these subglacial moraines appears to correspond to two sites. First, in the central zones of ice sheets close to ice divides, and secondly close to the margins of ice sheets where large terminal moraines occur. Pulju moraines appear to coincide with the former, while Veiki and hummocky moraines often relate to the latter. Since these moraine types exhibit non-directional surface morphologies it is unlikely that their origin can be totally attributed to active ice movement but rather, although the process of instigation may have occurred under active ice, the final morphology and often chaotic distribution seems indicative of passive ice conditions following ice disintegration and collapse. Since the dominant formative processes involved in the generation of these various moraines involves passive ice wastage, all of these forms are dealt with in Section 8.10.3.2.

273

8.10.3. Subglacial Erosional Forms under Passive Ice As ice stagnates under passive conditions the volume of meltwater produced tends to lead to a dominance of meltwater erosion. However, even during passive ice, small erosional forms owing to ice wear processes are generated. Most of these forms are small-scale surface etchings and abrasive scratching manifesting as deflected striae. Unlike striae formed under active ice conditions, under passive ice striae, to a greater extent, indicate the deflecting influence and control of micro-topography and meltwater hydraulic gradients (Sveian et al., 1979). The impact of meltwater erosion in the subglacial environment is possibly more intense than under active ice conditions. Although hydrostatic pressures are likely to be much less and channels and conduits are not so rapidly sealed off as a result of ice motion, ice overburden pressures at key locations on the bed of an ice mass will still lead to fluctuating hydraulic pressures. The effect of net ablation also leads to greater sediment input into the meltwater systems than under normal active (non-surging) ice conditions. Finally, unless the ice mass is reactivated, most passive erosional forms survive into the proglacial environment.

8.10.3.1. Meltwater erosional forms Meltwater erosional forms such as S-forms fashioned under passive ice conditions are similar to those formed under active ice motion. Perhaps the greatest difference between forms created under these active and passive ice conditions is in survivability, ‘freshness’ and degree of integrity of forms under the latter ice conditions. Although subglacial meltwater channel development has been discussed under active ice conditions, within passive ice these channels reach their greatest level of development individually and as networks. The value of meltwater channel system morphology pattern in relation to ice disintegration landforms and sediments can be of immense benefit in deciphering the final stages of ice retreat (Young, 1975; Sissons, 1982; Rodhe, 1988; Gray, 1991).

274

SUBGLACIAL ENVIRONMENTS

8.10.3.2. Ice disintegration forms

consist of an area of chaotic ridges that have a winding, hummocky appearance (Fig. 8.47). They seem to occur preferentially within 200 km of an ice divide and, although usually found in valley lowlands, do occur on rising ground. The origin of these moraines remains problematic. It has been suggested that Pulju moraines are formed through sediment being squeezed up into basal crevasses or that the moraines are a result of general ice disintegration during deglaciation or lateral pressure-induced ridging, both under passive ice conditions. However, strong clast fabrics in these moraines appear to preclude passive ice conditions.

There are several morainic forms that, although initiated under active ice conditions, appear to attain their final form and morphological character under passive ice disintegration. These moraines are not commonplace but are important in several glaciated areas, for example, Finland, Canada, Alaska and Sweden. 1 Pulju moraines. These moraines have been best described from locations in northern Finland (Aario, 1990) (Plate 8.27). In plan form these moraines

(a)

(b)

(c)

PLATE 8.27. (a) Pulju moraines, northern Finland. (Photo courtesy of Risto Aario.) (b) Closer view of a Pulju moraine, northern Finland. (Photo courtesy of Risto Aario.) (c) Pulju moraine at the eastern end of Lake Råstojaure, north of Kiruna, Sweden. The formation is about 100 m across. (Photo courtesy of Robert Lagerb¨ack.)

SUBGLACIAL ENVIRONMENTS

275

FIG. 8.47. Sketch of Pulju moraine ridges. Ridges are typically 2–5 m high, 10–15 m wide and 50–100 m in length (drawing by Kari T¨orm¨anen, after Aario, 1990).

2 Veiki moraines. The origin of these moraines has been the subject of a long and intense debate principally in Sweden. Veiki moraines can be described as a distinctive near-circular plateau surrounded by a single or, on occasion, double rim or ridge (Plate 8.28). Mechanisms of Veiki moraine formation can be subdivided into: (a) subglacial origins under active or passive ice, or (2) supraglacial origins. The relationship with drumlins and other active subglacial bedforms needs investigation before an accepted formative hypothesis can be obtained. 3 Hummocky moraines. The term hummocky moraine tends to be applied to a rather wide array of glacial forms and is perhaps more a morphological than sedimentological term. Hummocky moraines are chaotic steep-sided piles of dom-

inantly subglacial debris that lack a coherent directional pattern and are often associated with marginal areas of ice masses. In surface appearance hummocky moraines, seen from above, can be easily confused with the chaotic pattern of stratified outwash aprons and dead-ice topography associated with ‘kame and kettle topography’. Considerable debate as to the origin of these moraines has occurred. It has been argued that the moraines are formed under active ice conditions (Sutinen, 1985), or are glaciofluvial ice decay forms, while others have suggested that the moraines were basal ice stagnation forms (Gray and Lowe, 1977; Bennett and Boulton, 1993) and finally, it has been suggested that the moraines mark the position of an active but retreating ice terminus, the debris being largely derived from

276

SUBGLACIAL ENVIRONMENTS

(a)

PLATE 8.29. Hummocky moraines marking the outer limit of the Loch Lomond Stadial in Glen Turret, Tayside, Scotland. Hummocks, individually, are approximately 4 m in height.

(b) PLATE 8.28. (a) Veiki moraine, Finland. (Photo courtesy of Risto Aario.) (b) Veiki moraine plateau with rim-ridge, northwest of Nattavaara, northern Sweden. (Photo courtesy of Robert Lagerb¨ack.)

side-wall periglacial activity (Eyles, 1983a,b) (Plate 8.29). In the past, the term ‘ablation moraine’ has probably been misapplied to these moraines (Lundqvist, 1981). Detailed sedimentological work by Hodgson (1982) indicated that the debris composing these

moraines is often subglacially derived and was later overridden by active ice; thus an active origin may be correct. The common association of hummocky moraine with down-ice fluted moraines adds credence to this active ice hypothesis. However, Thorp (1991) in the western Grampian Mountains of Scotland, suggested that there is often an association of hummocky moraine with the availability of rock debris from soliflucted valley walls suggesting that supraglacial debris down-wasted and formed supraglacial diamicton dumps. In contrast, on Rannoch Moor in Scotland where no higher elevation sources of soliflucted debris exist, this model (similar to that proposed by Eyles, 1983b) does not apply and instead a subglacial active ice origin seems appropriate (Fig. 8.48). 4 Stratified ‘kame and kettle’ terrain. It is difficult to be certain whether this type of terrain can be attributed to subglacial passive ice conditions or supraglacial ablation and not, as is often the case, to proglacial surface collapse owing to subsurface melting of stagnant ice (Menzies, 1995, chapter

SUBGLACIAL ENVIRONMENTS

277

FIG. 8.48. Development of ‘hummocky moraine’ terrain. Ice cored sediment dumps melt under thinner covers of sediment creating topographic lows thus inverting the topography (Stages 1→2→3) (after Eyles, 1983b).

12). Where hummocky moraines are composed dominantly of stratified sediment, a subglacial glaciofluvial origin might be plausible (M¨akinen, 1985; Sutinen, 1985; Lagerb¨ack, 1988). 8.11. REPETITIVE EVENT HISTORIES IN SUBGLACIAL ENVIRONMENTS In any sedimentary environment there is often witnessed the repetition of a sequence of sedimentological events associated with a set of processes and resultant sedimentary forms and structures at the same site over and over again (Menzies, 1996, fig. 2.10). Within glacial environments this repeated succession of events is commonplace. In subglacial environments there are two levels of repetitive event histories that can be distinguished: (1) those occurring during a single phase of glacial activity and (2) those during each successive phase of glaciation. Too often subglacial environments are discussed with a somewhat static view of sedimentological activity when in fact the subglacial environment is one of repeated dynamic activity in which multiple events occur with such enormous variations in time and space that differentiation of each event history is an intricate and, at times, almost impossible task.

During single glacial events sediments can pass through many sequences of erosion, transport and deposition. The ambiguous term ‘resedimentation’ has been extensively used in attempts to convey a sense of the non-static sedimentary systems active within subglacial environments. Landforms and bedforms may be formed or, if from a previous glacial phase or earlier within the same phase, partially or wholly destroyed or unaffected, depending upon time of formation, sediment shear strength, effective stress levels etc. In some cases the change of one bedform to another may be only partially complete, while in other locations a new form emerges replacing the previous form with no vestige of the earlier form visible. However, sediments, their characteristics, structures and properties may still retain the imprint of past sedimentological histories that indicate frozen conditions, rafting, thrusting etc. Overprinting, reorientation, reworking, realignment and redirection are all terms indicative of episodes of repetitive event histories. Where deglaciation has occurred and subsequent ice readvance has occupied an area once again, a quite different set of circumstances must be considered. First, the length of time involved between separate advances. Second, the impact of climatic changes on

278

SUBGLACIAL ENVIRONMENTS

soil development, vegetation colonization and the effect of subaerial processes in altering topographic features and slopes. Third, the presence of postglacial lakes or sea-level rise in flooding areas of terrain. Finally, whatever length of time and changed surface conditions that have taken place, previous subglacial sediments and landforms may be drastically or only slightly altered. In many instances all record of previous subglacial sediments may be totally removed and reincorporated into a new subglacial environment. It is possible that no sediment units may survive even as rafted inclusion, however, geochemical signatures may be detectable in the form of altered minerals and neoformed clay minerals. In some cases, the presence of organics in the form of woody materials, pollen or bone fragments may be indicative of past warmer climate conditions, however, all too often such evidence is lost. Some artefacts cannot be so easily destroyed (striae, chattermarks, conchoidal fractures) but instead cross-cutting and overprinting relationships may be detected. It is critical that field researchers be aware of the influence and extent of overprinting and multiple event histories. The overlaying effect of multiple sedimentary events within one or multiple glacial phases must be considered whenever the origin of subglacial sediments and forms are debated and when glacial stratigraphy is examined. 8.12. SUMMARY The importance of subglacial sediments in terrigenous glaciated areas of the world cannot be undervalued. Volumetrically, other than glaciomarine sediments, these sediments cover the largest areas of the Earth’s glaciated surface. Their influence upon all forms of human activity cannot be underestimated from agriculture, to construction as foundations and aggregate resources, to sites for waste disposal, and as aquifers of major groundwater sources for large urban areas. There remains an enormous wealth of knowledge to be gleaned from these sediments and forms. As our understanding of subglacial processes, from both modern and past glacial environments, increases, so new avenues of research emerge. Likewise, paradigms shift and mutate such that what has been

accepted understanding is now in some cases in a state of flux. For example, the existence and detailed analyses of microstructures within diamictons has opened up a huge area of new research that demands new perspectives on how diamictons are deposited. Likewise, there is increasing need to grasp the intricate interplay between thermal, geological, glaciological and rheological conditions that coexist at subglacial interfaces of which subglacial sediments are the end product. Interpretation of subglacial facies associations, sub-facies units and the diagnostic criteria that might be used to discriminate environment, process and form remains a fertile research area. Although a great deal is known about subglacial processes, little is understood concerning rates of processes. It is likely that in many cases sediments and attendant structures may be deposited and generated over very short periods of a few hours or days while others such as meltout may take centuries. Since subglacial sediments of past glacial environments are heavily utilized today for innumerable purposes it is critical that the effects of postdepositional diagenesis be much better understood. Diagenesis appears to begin almost immediately after sediments are deposited even before ice retreat and accelerates especially in the first years of subaerial exposure. However, the extent, depth and degree of alteration caused by diagenetic processes still remains largely unknown. It would be fair to state that our still limited understanding of past ice sheets, the dynamics of their growth and decay and the interrelationship of ice sheets with oceans is revealed by our as yet restricted grasp of the processes and dynamics of subglacial environments. The ‘drumlin enigma’ stands as a conspicuous instance of how much there is still to understand about ice sheet dynamics and the interplay of glacier motion, sediments, topography and environmental ‘conditions’ at the base of ice sheets and valley glaciers. Until we fully recognize the complexities of environments beneath ice sheets and other ice masses it is unlikely we can correctly postulate the explanantion(s) necessary for subglacial landform development, sediment deposition and other geomorphic processes at the ice–bed interface.

9

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS J. Maizels 9.1. DISTINCTIVENESS OF PROGLACIAL ENVIRONMENTS Proglacial environments lie at and beyond the ice margin. Sediments and landforms in proglacial environments are dominated by meltwater and sediment derived from the ice mass itself The physical characteristics of proglacial environments are therefore largely dependent on extraneous controls, affecting input of water and sediment to the system from beyond its boundaries, and reflect the nature of those water and sediment inputs, the volume, sources and character of the inputs and the pathways and rates of transfer through the system (Fig. 9.1) (Brodzikowski and van Loon, 1991). 9.1.1. Nature of Meltwater Inputs Meltwater inputs to proglacial environments are characterized by large-scale regular and irregular variability in runoff magnitude. Regular variations reflect seasonal and diurnal ablation cycles and responses to periods of glacier advance (positive mass balance) or retreat (negative mass balance). In most climatic zones, discharge fluctuations closely follow the annual temperature curve, with little or no flow 279

occurring during winter. Four meltwater runoff periods have been identified following the winter lowflow period (Sharp et al., 1998): (1) spring meltwater flows occur in relation to break-up of river ice, while up to 25 per cent of meltwater produced is retained as superimposed ice, in snow and slush areas, and in ice capillaries and channels. Runoff amounts are much lower than expected because of this storage effect; (2) by early to mid-summer, rising temperatures lead to a rapid rise in snow melt followed by ice melt. Meltwater flows exceed those expected from climatic conditions because of the supplementary inputs of water from storage at a time of full development of englacial watercourses; (3) by late summer, runoff of stored water is expended, while the well-developed englacial channel system allows minimum delay in runoff. Runoff is directly related to ablation and rainfall, but since temperatures are decreasing towards the end of the meltseason, discharge declines, and (4) as autumn comes to an end, the winter freezeback begins and flows approach their seasonal minimum. Hence, the glacier acts as a natural reservoir for meltwater, storing it in winter and releasing it in summer (Chapter 4). Most glaciers also exhibit a strong diurnal cycle in runoff, reflecting diurnal temperature fluctuations.

280

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS Controls on 1) QW and QS in glacier system 2) Channel form pattern and sedimentation in proglacial system, by QW and QS SW

SG SNG

CD D

ABD

DQW

QNG

DQW

QS

GLACIER SYSTEM

EG EGN WNG

QSNG g

D.p.s

r T

F

n d

Slope a

Fr

CSS

Flow Regime

tc V*

Gs

Sed Structures

Ds

BI

PROGLACIAL SYSTEM

n

Re

BL

SIN

w/d

V

B

w

Indeterminate variables in paleohydrologic analysis Determinate variables (can be determined by direct measurement in the field) Variables that can be estimated quantitatively in paleohydrologic analysis

FIG. 9.1. A model of the main interrelationships by which meltwater and sediment discharge from a glacier system can affect the characteristics of the proglacial drainage system. (Note: symbols used here are defined specifically for this figure alone). Key: Glacier system. EG , glacier erosion; EGN , non-glacial erosion; WNG , non-glacial weathering; CD , degree of ‘channel development’; ⌬, storage of water; BL, ablation; QNG , water from non-glacial source, SG , sediment supply from glacial source; SNG , sediment supply from non-glacial source; SW, sediment supply to glacial meltwater system; QW, discharge of meltwater; QS , discharge of sediment transported in meltwaters; ⌬QW, variations in QW. Proglacial system. B, boundary conditions; BI, ‘braiding index’; CSS , suspended sediment concentration; d, flow depth; D, sediment size (characteristics); DS , amount of sediment aggradation; F, channel form parameter; Fr , Froude no. = (v/dg)1/2; g, acceleration due to force of gravity; GS , rate of sediment transport; n, roughness coefficient, QsNG , non-glacial sediment supply; Re , Reynolds no. defined as: Re =

Vd␳ ␮

=

Vd ␯

SIN, channel sinuosity; T, water temperature, ␶c , critical tractive force = ␥dS; V, mean flow velocity; V*, critical shear velocity (␶c/␳)1/2; w, flow width; ␥, specific weight of water; ␳, density of water; ␳s , density of sediment; v, kinematic viscosity of water = ␮/␳; ␮, dynamic viscosity of water.

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS 5

1978

3

3

m s

-1

4

2

1

0

May

June

July

Aug.

Sept.

Oct.

5

1983

-1

4

3

3

m s

The timing of peak daily flow largely depends on the transit time for water to flow through the glacier system from different parts of the ablation zone. Transit time, in turn, varies according to the size of the ablation zone and the extent to which the englacial drainage network has developed (Sharp et al., 1998). Runoff earlier in the season or draining very large glaciers and ice caps may exhibit little or no detectable diurnal cycle. However, where diurnal runoff cycles are present, the meltwater system is subject to the equivalent of a complete flood cycle every 24 hours. These daily flood flows are highly significant in terms of instigating bedload sediment transport and morphological changes in the proglacial channel system. Superimposed upon these regular seasonal and diurnal cycles of runoff are a number of irregular discharge events. The most common periods of irregular high flow are associated with summer or autumn storm events, particularly those that occur after periods of relatively warm conditions. These rain-induced runoff events often generate the main floods of the meltseason (Fig. 9.2). Rainstorm floods can be instrumental in accelerating rates of sediment transport and proglacial channel changes. Ashworth and Ferguson (1986), for example, found that a rain flood of 10 m3 s–1, compared with peak ablation flows of 5–8 m3 s–1, generated an order of magnitude increase in bedload sediment transport rates in a meltwater channel in Lyngsdalselva, northern Norway, from 0.3 to 3.0 kg m–1 s–1. High magnitude, periodic runoff events may also be superimposed on the regular seasonal and diurnal cycles. These are ‘glacier-burst’ or j¨okulhlaup floods (Chapter 4) (Maizels and Russell, 1992; Bj¨ornsson, 1998). J¨okulhlaup floods have characteristic hydrographs, reflecting the rate of development of the englacial. tunnelways that accommodate the flood waters (Fig. 9.3) (Chapter 4). Hence, a j¨okulhlaup hydrograph is typified by a gradual rise to a peak flow associated with the progressive enlargement of tunnels, followed by a rapid recession curve as the water supply from the source reservoir becomes depleted. However, complex and interrupted hydrograph forms occur where tunnels become temporarily blocked by ice and debris, leading to short-lived shut-offs of flow. J¨okulhlaups may last for less than 24 hours or, when draining larger bodies of water, may extend over several weeks (Fig. 9.3)

281

2

1

0

May

June

July

Aug.

Sept.

Oct.

FIG. 9.2. Contrasts in annual meltwater runoff from the Vernagtgletscher, reflecting years of positive mass balance (1978) and negative mass balance (1983). Periods of summer melting are interrupted in both years by heavy snowfalls (after R¨othlisberger and Lang 1987; reprinted from Gurnell, A. M. and Clark, M. J. (eds) Glacio-Fluvial Sediment Transfer, by permission John Wiley and Sons).

(Bj¨ornson, 1992, 1998). Lower magnitude, shorterterm runoff irregularities may reflect drainage of small reservoirs, collapse, blockage and breaching of ice walls and tunnel roofs or changes in tunnel capacity. These variations may extend only over minutes or hours, and associated flow fluctuations may become rapidly dissipated downstream. Large-scale, long-term variations in runoff also occur during periods of glacier or ice-sheet growth

282

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

Meltwater Discharge (m3 s-1)

(a) 1,000

500

0 1200

1800 17/7/87

0000

0600

1200 1800 18/7/87 Date

(b)

Katla 1918

Meltwater Discharge (m3 s-1)

106

0000

0600 1200 19/7/87

Drainage from subglacial geothermal area Drainage from ice-dammed lake Drainage from subglacial volcanic eruption

105 Skeiðará (Grimsvötn) 1922, 1934

104

103

102

Skeiðará (Grimsvötn) 1954

Katla 1955

Skeiðará (Grimsvötn) 1972

Graenalon 1939 Skaftavkvos 1955

Vatnsdalslon, Kalgrimo 1975

Time (in Days)

FIG. 9.3. Hydrographs typical of j¨okulhlaup drainage. (a) Hydrograph produced during j¨okulhlaup from sudden drainage of icedammed lake in west Greenland, July 1987 (based on Russell, 1989). (b) Hydrographs produced during Icelandic j¨okulhlaup generated from sudden drainage of subglacial reservoir in area of geothermal activity (Grimsv¨otn); of ice-dammed lakes (Graenalon, Skaftavkvos and Vatnsdalslon); and from subglacial volcanic eruptions (of Katla). (based on Thorarinsson, 1957; Bj¨ornsson, 1988; from Maizels and Russell, 1992).

and decay. Reconstruction of bankfull flows at the edge of the Late Weichselian Ice Sheet in central Poland, for example, indicates a decline from about 145 m3 s–1 in the Aller¨od to 20 m3 s–1 at the present time (Rotnicki, 1991). Similarly, Teller (1990) reconstructed meltwater flows from the Laurentide Ice Sheet during the Late Wisconsinan, demonstrating the large changes in runoff associated with repeated catastrophic drainings of Lake Agassiz through the St

Lawrence route to the North Atlantic, reaching peaks of 4000 km3a–1 between 11 000 and 10 000 years ago. Even in the last century or so, overall glacier retreat in the European Alps has led to declines in mean summer runoff and annual specific yields. The sediments and landforms of proglacial drainage systems are therefore clearly subject to a wide range of runoff fluctuations, operating over a variety of timescales and each exhibiting a varying capacity to entrain, transport and deposit sediment, thereby modifying the proglacial landscape. Most proglacial channel systems can be expected to reflect some degree of equilibrium with flows generated during the diurnal flood cycle; even those environments subject to less frequent, major flood events may be adjusted to these rare events, depending on the ability of the system to return to its pre-flood state. Depending on the resistance and sensitivity of the system to change, and on its relaxation time, landforms may also be modified by later flows, either through lateral scour and re-working or by incision during large floods. This concept of landform response to runoff events is important in the analysis of proglacial environments, since it provides the framework for identifying: (1) how glacier/ice sheet behaviour can affect the pattern of geomorphic response in proglacial environments, and (2) the proportion and character of runoff events that affect the sediments and landforms of proglacial environments 9.1.2. The Nature of Sediment Input Sediment inputs to proglacial environments are largely derived from the glacier source generated both by glacial and fluvioglacial transport to the ice marginal zone. Most suspended sediment is derived from the ice mass basal zone. Sediment inputs reflect both the efficacy with which sediment is eroded, entrained and transported by these agents, and also the temporal and spatial variability of sediment transport within the glacier system. Sediment can also be delivered to the proglacial environment from nonglacial sources, particularly in mountain areas where slope mass movements, colluvial or ‘paraglacial’ fan development, rain- and snow-fed valley side streams contribute sediments; from volcanic eruptions; and in coastal areas, aeolian and shoreline sediments. Other

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

windblown sediments can be derived from both glacial and exposed fluvioglacial materials. Sediment inputs from this wide variety of sources are highly variable over space and time. Spatial variability reflects the geographic distribution of sediment pathways to the proglacial environment, while temporal variability indicates variations in runoff and in availability of sediments (Fig. 9.4). One major control on sediment inputs to proglacial environments is the resistance of bedrock to glacial and fluvioglacial erosion. In ancient crystalline shield areas, for example, suspended sediment loads of meltwater rivers have been found to be significantly lower than those of young, mountainous volcanic regions such as found in Iceland, New Zealand and parts of the Rockies–Andean chain (for otherwise

283

similar flow and catchment conditions). Nevertheless, the volumes of sediment input to almost all proglacial systems are significantly higher than for most nonglacial fluvial systems, other than for rivers flowing through thick loess sequences as in China (Lawler, 1991). Suspended sediment loads, which range in size from about 0.001 to 1 mm, generally exceed 1000 ppm (Church and Gilbert, 1975), and commonly exceed 400 t km–2 of catchment area (Østrem, 1975; Lawler, 1991). These high values have been estimated to represent mean long-term rates of total catchment sediment yields (Table 9.1). The values of glacial sediment yields in Table 9.1 can be seen to exceed the estimates of mean annual suspended sediment yields from a wide range of rivers. Hence, it can be noted that proglacial meltwaters receive very high inputs of

SOURCE STREAMS 9,000

8,000

SOURCE STREAMS

9 8 7

Major Tributaries

10

Minor Tributary

MIXING ZONE

MAIN CHANNEL

5 4

3 2 1

For Source Streams heavy lines show dominant streams

6,000

5,000

4,000

3,000

MIXING ZONE

MAIN CHANNEL 1,200 1,000 800 600

2,000 400 1,000

0

Total Discharge (l s-1)

Suspended Sediment Concentrations (mg l-1)

7,000

14th July 1981

200 0

FIG. 9.4. Spatial variability in suspended sediment concentration during diurnal discharge cycles, Glacier de Tsidjiore Nouve, Swiss Alps. Time intervals shown on horizontal axis are hourly, starting at 08:00 h (after Gurnell, 1987; reprinted from Gurnell, A. M. and Clark, M. J. (eds), Glacio-Fluvial Sediment Transfer, by permission of John Wiley and Sons).

284

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

TABLE 9.1. Sediment of yields of selected glacial and non-glacial rivers Location

Sediment yield

Reference

<7.0 × 102 t year–1

Hammer and Smith, 1983

Annual sediment yields of glacial rivers Hilda Glacier, Alberta, Canada Engabreen, Norway

1.7 × 104 t year–1

Østrem, 1975

Gornergletscher, Switzerland

4.8 × 104 t year–1

Collins, 1978

Solheimajokull, ¨ Iceland

1.1 × 10 t year

Lawler, 1991

Hunza River, Pakistan

>6.3 × 107 t year–1

6

–1

Ferguson, 1984

Specific annual sediment yields of glacial rivers 16–1545 t km–2 year–1

Norwegian rivers

4773 t km–2 year–1

Hunza River, Pakistan Solheimajokull, ¨ Iceland

14 482 t km

–2

year

–1

Østrem, 1975 Ferguson, 1984 Lawler, 1991

Specific annual sediment yields of non-glaciated catchment basins St Lawrence River, Canada Congo River, Zaire Amazon River, Brazil Ganges River, India Brahmaputra River, India

suspended sediments in response to glacial and fluvioglacial erosion within the source catchment. Bedload sediment yields also appear to be significantly higher than in many non-glacial streams, although there are few reliable measurements of bedload movements from proglacial environments (Gomez, 1987). Church (1972) recorded values of 29–1114 t km–2 a–1 for the Lewis and Ekalugad meltwater rivers in Baffin Island, representing 13.5–93.6 per cent of the total sediment yields from the catchments. For many coarse-bed rivers, however, bedload sediment transport rates are estimated, rather than measured, while measurements of bedload transport rates in proglacial rivers have tended to be restricted to short sample periods within small, minor (wadeable) channels (Ashworth and Ferguson, 1986). The high sediment yields recorded from glacial meltwaters to some extent mask the high temporal variability that characterizes these inputs of sediment to the proglacial channel system. For example, not only do these amounts vary considerably from year to year, but the suspended loads also vary in response to

5 t km–2 year–1

Walling, 1984

–2

–1

Walling, 1984

79 t km–2 year–1

Walling, 1984

–2

–1

Walling, 1984

1370 t km–2 year–1

Walling, 1984

13.2 t km year 537 t km

year

progressive ‘washing out’ of the supply of fines from the glacier and proglacial channel boundaries. This depletion process can occur over each meltseason, with maximum fines being made available at the start of the season, and over each periodic and diurnal flood cycle. Progressive flushing of fines leads to development of marked, clockwise hysteresis loops during cyclic discharge fluctuations. For example, suspended sediment concentrations are higher at the start of the meltseason, and on the rising limb of cyclic events, than during similar magnitude flows at the end of the meltseason, or on the falling limb of cyclic events. Bedload only moves sporadically, once the local thresholds for entrainment have been exceeded (Fig. 9.5). The threshold conditions are not constant in a given channel, however, since they depend not only on the force of flow, but also on the resistance of the bed. Bed conditions can vary markedly through time and space, from being loose and highly mobile, to forming a highly resistant, imbricated, armoured bed, which may only be disturbed by infrequent flood events. These events may then disrupt the armour layer and

Bedload Sediment Transport Rate Gb (Kg m-1 s-1)

100

200

80

150

80 60

100 40 20

0

10

20

30

40

50

60

70

x

0

Armour breakage 0

20

40

60

70

40 20

80

100

0

0

20

40

0

20

40

60

80

100

120

60 80 Time (mins)

100

120

x

60

40 x

20 0

Armour development

50

60 Stream Power

Armour breakage

60

0

D84 (cm)

100 Armour development

x 0

x

x x

x 10

x

50

x

x

x x x x x xx x x xx x x x x xx 20 30 40 50 60

x

40

xx

30 x 70

20

0

x 20

x x 40

x

x 60

x 80

x 100

20 15

5 4

40

10

3

20

5

2

0

0

10

20

30 40 Time (mins)

50

60

70

0

20

40 60 Time (mins)

80

100

0

FIG. 9.5. Fluctuations in bedload sediment transport rates in a meltwater stream on a proglacial fan in west Greenland. The three examples (left to right) illustrate periods of declining, fluctuating and rising rates of sediment transport through time, respectively, independently of stream power but reflecting patterns of bed armour development.

286

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

expose the wide range of sediment sizes contained in the underlying substrate. In this circumstance, almost all bed sizes become mobile and sediment transport rates increase very significantly. Østrem et al. (1967), for example, estimated that 60 per cent of the total summer sediment load in a Baffin Island meltwater channel was removed during a 24-hour outburst flood. Similarly, Church (1972) found that between 25 and 75 per cent of the total annual sediment transport on Baffin Island sandurs occurred during the 4–5 days of peak flows, while Østrem (1975) found that 50 per cent of the load transported from Erdalsbreen, Norway, was removed in 4–5 days of high flows. Church (1972) estimated that j¨okulhlaup flows removed up to 90 per cent of total annual sediment yield of a Baffin Island catchment, while two outburst events in an Alpine catchment were estimated to have removed about 50 per cent of annual sediment yields (Beecroft, 1983; Gurnell, 1987). Nummedal et al. (1987) suggest that a two-week j¨okulhlaup on Skei∂/ ararsandur is able to move as much sand as would otherwise be transported in 70–80 years of normal meltwater flows (Russell and Knudsen, 1999). The volume of material available for transport into the proglacial environment plays a crucial role in influencing the sedimentary characteristics of proglacial deposits. For example, where only small volumes of sediment are introduced, whole phases of glaciation and deglaciation may be represented in the proglacial environment (and in the sedimentary record) by only a metre or two of sediment, much of which may have been repeatedly re-worked. It becomes increasingly difficult to interpret such deposits and associated landforms in terms of the controlling processes that generated them because of their composite and interrupted history (Menzies, 1996, chapter 8). Hence, the smaller the volume of sediment input to the proglacial environment over a given timescale, the greater the likelihood of complex, reworked, composite landforms evolving. Where large volumes of sediment are repeatedly introduced to the proglacial environment, thick deposits, often forming steep outwash fans, can accumulate rapidly. These may contain sediment sequences representing individual flow or flood events, or flows from successive meltseasons. With rapid steepening of fans, stream powers (Ws ) in turn increase markedly

((Ws = ␥w QSg ), where ␥w is the specific weight of water, Q is stream discharge and Sg is stream gradient) and fan entrenchment and terrace formation can ensue. Terrace sequences commonly occur in proximal zones of proglacial fans, in response to fan steepening during periods of rapid accumulation (i.e., before glacier fluctuations directly affect the fan morphology), accentuated by, and possibly encouraging, entrenchment of the ice margin itself. An important effect of these high sediment inputs lies in the response of meltwater streams to entrain, transport and deposit their loads. The balance between available stream power and that required to entrain and transport sediment through a given reach, largely controls the nature of the developing channel system. All river flows are characterized by longitudinal alterations in sediment transport capacity reflecting development of flow cells. Once deposition occurs, these alternating flow zones are accentuated by lateral flow deflections towards one or other bank. Hence, when large volumes of sediment are introduced, much of this is deposited as accretionary bar deposits in zones of flow divergence within the channel. Where lateral bank scour is possible, braided channel systems develop. The majority of outwash plains are characterized by braided river systems, replaced by single-thread channels only where entrenchment or confinement has occurred. When little new sediment is introduced, meltwaters gain their load from the channel boundaries. Large-scale re-working of older deposits can provide the main source of sediment transported through the fluvial system and re-deposited across the outwash surface. However, where channels are composed of mixed grain sizes, the bed can become armoured as fines become sheltered behind larger grains, filter into the bed or are winnowed away. Once the bed sediment is formed of an impacted, imbricated layer of coarse pebbles or cobbles, it becomes highly stable and resistant to entrainment. In some meltwater streams, therefore, material derived from the bed is only acquired during high magnitude flows. If the bed cannot supply the required material, additional sediment is gained from the channel banks. This process leads to channel widening and shallowing and, in turn, reduction of bed shear stresses that could otherwise promote bed scour. Deposition within the channel may ensue.

N Ice-margin at time of survey Limit of valley-train Generalized contours at intervals of 1 m on the valley-train Bluff or edge of terracette Shoal or bar on the valley-train Terminal moraine (active and abandoned)

0

10 metres

20

Course of supraglacial and mudflow streams Boulder or bedrock outcrop Stand of vegetation

FIG. 9.6. A plane-table map of the meltwater channel system on the Bossons valley-train, near Chamonix, France, April 1974.

288

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

High volumes of fine sediments transported in suspension may be deposited as a temporary veneer on channel bar deposits during waning flows, or may concentrate in proglacial pools and lakes, or be transported into the marine off-shore zone. Fines exposed on outwash surfaces are frequently removed by winds and accumulate as loessic or dune distal deposits (Menzies, 1996, chapter 6). With the largescale transport of fines from the fluvial system, few outwash deposits contain fines and are typically coarse-grained. Exceptions to this pattern occur during meltwater flows containing hyperconcentrations of fines. Hyperconcentrated flows occur often during glacier-outburst floods. The effect of these flows leads to the formation of lobate, hummocky or sheet-like topography, characterized by matrix-supported, unsorted or inversely graded deposits. 9.2. MORPHOLOGY AND LANDFORMS OF PROGLACIAL ENVIRONMENTS 9.2.1. Non-j¨okulhlaup Glacial Outwash Many proglacial environments appear to be composed of extensive, featureless outwash plains. However, proglacial environments contain a range of distinctive features, representing both active and former fluvioglacial activity. In addition, some landforms may be relict features of glacial or colluvial origin, modified by aeolian or permafrost activity, or affected by large scale, long-term fluctuations in ice-limits, or in local or global base-level or isostatic changes. The most distinctive characteristic of proglacial environments is the zone of outwash accumulation. The morphology of the outwash deposit depends on: (1) the volume of sediment inputs to the outwash system, (2) the topography and accommodation of the receiving area, and (3) the nature of the processes distributing the sediment through the proglacial zone. Where the accumulation zone is confined between valley sides, outwash sediments are concentrated on the valley floor forming a ‘valley-train’ deposit (Fig. 9.6; Plate 9.1). Infilling of glacial ‘troughs’ by glaciofluvial outwash leads to the development of a relatively flat valley floor, which may mask a highly irregular bedrock topography. In addition, progressive infilling may gradually bury pre-existing landforms,

such as former terminal moraines, eskers, drumlins and kames. Valley-trains exhibit a wide range of down-valley slopes. Where proglacial outwash is unconfined and can extend laterally for a considerable distance, it becomes known as an outwash plain or sandur, the latter an Icelandic term for a sand plain. Modern sandur plains are located typically in piedmont zones, commonly those that extend seawards from young glacierized mountain ranges to form coastal plain infill deposits. The coastal plains of southern Iceland, southwestern Alaska and western Greenland, for example, contain some of the most extensive modern outwash plains. The largest is Skei∂/ ararsandur in southern Iceland, which is 40 km wide and 30 km from the ice margin to the sea (Plate 9.2). Pleistocene sandur plains were more extensive, including, for example, the Canterbury Plains in South Island, New Zealand (Plate 9.3), the Patagonian gravels (Argentina) and those bounding the margins of the Laurentide and Eurasian Ice Sheets. Where sandurs extend to a coastal margin, their distal limits are marked by steep deltaic slopes, reworked by shoreline processes to form a fan-delta or plain-delta deposits. Outwash fans reflect relatively rapid deposition in the proximal zone, fed by a major glacial meltwater route reaching the ice-margin at a single point (Fig. 9.7; Plate 9.4). Composite fans fed by subparallel, subglacial drainage routes can merge and form an ‘apron’ along the ice-margin. Such fans require the ice-margin to remain relatively stable, since a glacier advance leads to overriding and burial, while retreat is likely to lead to entrenchment with meltwater flows emerging at a lower level than those reaching the original fan head. Meltwater is likely to become ponded behind the ice-contact slope of the fan and, on reaching the fan surface, instigating rapid entrenchment of the fan itself. If the fan ‘apron’ deposits are banked up against a stagnating ice margin, progressive burial and differential decay of the ice leads to the formation of a hummocky, pitted outwash surface (Plate 9.5). Where gradients exceed the threshold for incision during episodes of competent flows, bed scour may lead to development of a new channel bed at a lower level. This may happen at a local scale, around a braid bar or within a single channel reach, for example, or

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

289

PLATE 9.1. The ice-terminus and valley train of the Bossons Glacier, French Alps, 1981. The valley-train is ca. 50 m wide and is bounded by steep Little Ice Age lateral moraines (photo by W. Mitchell).

PLATE 9.2. Complex anastomosing channel systems characteristic of the Icelandic sandur plains, illustrating Sulasandur, the western edge of Skei∂/ ararsandur in southern Iceland, looking north towards Vatnaj¨okull.

290

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

PLATE 9.3. The braided Waimakariri River extending eastwards across the Canterbury Plains, South Island, New Zealand. 100

Ekalugad Sandur Length 8.8 km v.e. = 20

Elevation (m)

Exponential portion only of compound equations plotted Ekalugad Sandur y = 30 - 4x + 35e-0.3x

10

Lewis Sandur y = 17e-0.27x (arbitrary datum)

100 1

10

Lewis Sandur

Length 1.0 km v.e. = 20

10

1

Tingin Sandur

y = 12 - 2.5x + 45e-0.9x

Tingin Sandur Length 4.8 km v.e. = 20

Hoffellsandur

1

Length 15.0 km v.e. = 20

Hoffellsandur 100

Slims River Sandur

100

Length 26.0 km v.e. = 21

y = 60e-0.34x White River y = 160e-1.5x (arbitrary datum)

10

Slims River 10

White River Sandur

y = 34e-0.14x (Lake Kluane datum)

10 1

Length 2.1 km v.e. = 2 1

Distance ( km: 100m for Lewis only: 2km for Slims only)

1

FIG. 9.7. Longitudinal profiles of sandur surfaces, showing both plotted survey data (left) and exponential fitted equations (right) (from Church, 1972; reproduced from the Geological Survey of Canada, Bulletin 216, p. 108, fig. 70).

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

PLATE 9.4. Steeply graded outwash fan at the edge of the West Greenland ice-sheet. Note figures for scale.

over the stream profile as a whole. If incision is repeated, a terrace sequence may develop. At present, however, little is known about the flow conditions that trigger incision and create terrace sequences in proglacial environments. The longitudinal profiles of proglacial rivers, terrace sequences and outwash surfaces exhibit shapes ranging from single, straightline profiles to highly complex stepped profiles. Distal zones exhibit significantly lower gradients than those of proximal zones and an abrupt discontinuity between the two zones has often been identified. The discontinuity may reflect statistical methods of analysis rather than physical processes. Several explanations infer an abrupt change in processes at this locality. Interestingly, this discontinuity has also been

291

identified in many non-glacial river profiles. It may be that the boundary between proximal and distal slopes reflects a change in the size distribution of bed material. Steeper proximal reaches may be characterized by multi-mineral or multi-grain particles, while distally the grains may have been broken down into their separate mineral or granular components to form a significantly finer load. In contrast, it has been argued that proximal channels are controlled more by bedload transport, while suspended sediment transport dominates in distal channels, the boundary being marked by grain sizes of = 2ϕ. A second explanation is that the proximal zone is characterized by the addition of sediments directly from the ice margin and moraines or from valley walls, such that processes are not confined to fluvial activity and sediments are not limited to those provided solely by fluvioglacial processes. A third, and widely accepted, explanation is that the slope is adjusted to the amount of stream power required to transport inputs of sediment to the system in order for the system to achieve equilibrium. While in non-glacial, humid catchments, discharge increases downstream in response to an increased catchment area, in proglacial systems discharge remains at a similar magnitude throughout the system unless supplemented by valley-side tributaries, snowmelt inputs or emergence of groundwater. Of the three variables of discharge, slope and sediment calibre, only the latter two variables are adjustable in the downstream direction. Therefore, large inputs of coarse sediment require steep gradients to generate sufficient stream power for transport, while progressive selective deposition of coarse particles creates an increasingly fine-grained load, which in turn requires lower gradients and stream powers for transport. An additional explanation for variations in long profile form reflects the alternating pattern of fluvial sedimentation and scour downstream. Variations in the long profiles of Baffin Island sandurs have been linked to a prominent change in channel habit (Church and Jones, 1982). Outwash surfaces, also, exhibit numerous smallerscale relief features. Variations in relief of up to several metres occur in response to the distribution of meltwater channel networks. The relief range depends on the depth of maximum channel scour, the maximum height of flood flows, and the calibre of the bed

292

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

PLATE 9.5. Terminal zone of part of Skei∂/ ararj¨okull, southern Iceland, illustrating complex hummocky topography related to ice stagnation (left); abandoned lateral drainage channels; a steep ice-contact slope; level outwash surface, underlain by horizontally bedded sands and gravels; and zones of eskers and ponded water (right).

material. Maximum bed relief occurs either in coarsegrained proximal zones where single-thread channels have formed, or in deeper channel systems of finegrained braided distal networks. However, exceptions are common, especially where outwash is bounded downstream by rock bars or local baselevels, behind which flood waters may be temporarily ponded to form deep backwater zones. Kettle holes are commonly found on outwash surfaces. They follow from the in situ melting of transported ice blocks or from collapse of marginal ice masses buried beneath outwash deposits. The extensive development of kettle holes can lead to a pocked or pitted surface. Where the ice blocks are rich in glacial debris, kettle holes may be rimmed or filled with diamicton (Maizels, 1992). If flood flows transported the ice blocks, meltwater flows around

grounded blocks can lead to longitudinal scouring. The scours may in turn be exploited by high flows, which use the scours to trigger incision of the bar surface to form chute channels, thereby promoting further braid bar division. Typically, the fossil patterns of braided or anastomosing streams are preserved on abandoned outwash surfaces. The differential extent of vegetation colonization has been widely used to evaluate the relative age and stability of various parts of proglacial braided river systems, particularly since channels are unstable and frequent migrations of the active channel zone promote repeated periods of renewed colonization by vegetation (Plate 9.6). In sub-polar and polar regions, outwash may be underlain by permafrost, and subject to pingo formation, growth of ice-wedge networks and thermokarst development.

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

293

PLATE 9.6. Aerial view of braided river system subject to a variety of levels of flooding. Most stable bars are partially vegetated. Flow is towards the base of the photo. Tekapo River, South Island, New Zealand.

9.2.2. J¨okulhlaup-related Outwash A substantial amount of literature exists on the proglacial features generated by j¨okulhlaups during the last glaciation indicating that j¨okulhlaups can generate a distinctive suite of erosional and depositional landforms in proglacial areas (Maizels and Russell, 1992). J¨okulhlaups possess an enormous erosive capacity and can create large-scale erosional features. The channelled scablands of the Columbia Plateau, Washington State, USA, for example, were drained by up to 70 floods from glacial Lake Missoula to create hundreds of streamlined, residual erosional hills, longitudinal grooves and giant potholes, scours and plunge pools. All the major glacial lake basins bounding the southern edge of the Laurentide Ice Sheet are distinguished by complex systems of huge spillway

channels up to 3 m wide and over 100 m deep (Kehew and Lord, 1987; Teller and Thorleifson, 1987). The type of landform produced during a j¨okulhlaup depends on the nature of the flood flow itself, its sediment load and the topography of the flood routeway. The largest depositional landform is a j¨okulhlaup-sandur (e.g., Myrdalssandur) (Maizels, 1992). Locally, aggradation during j¨okulhlaups can lead to the formation of raised boulder deltas in flood routeway lakes, boulder lags, spillway expansion bars, boulder bars and pendant bars (formed downstream of bedrock obstacles) (Shakesby, 1985). Unconfined j¨okulhlaup depositional landforms may include streamlined boulder bars, ‘washed’ sandur spreads and boulder lags, pitted or kettled outwash, and fields of ice-block obstacle marks, terrace sequences, lobate and hummocky fans and a variety of

294

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

distinctive sediment assemblages (Elfstr¨om and Rossbacher, 1985; Maizels, 1992). In confined proglacial routeways, the depositional impact of a j¨okulhlaup reflects variations in local flow dynamics and sediment supplies in different topographic settings. Where flows are constricted, depositional features may include local expansion bars, boulder berms and boulder lags (Teller and Thorleifson, 1987). Tributary bedrock channels or valleys lying at a high angle to the main flood channel commonly act as sediment traps to backwater flows to form sequences of slackwater infill deposits (Kochel and Baker, 1988). 9.3. PROGLACIAL MELTWATER CHANNEL SYSTEMS 9.3.1. Channel Pattern and Morphology Proglacial outwash is commonly characterized by complex braided channel networks, particularly in distal zones (Plates 9.2 and 9.6). The proximal zones may have been subject to fan entrenchment, with channels becoming confined between walls of older outwash or moraine. Proximal channels may therefore exhibit deep, single-thread courses, high in energy and transport capacity. In other proximal zones, where no topographic confinement exists, and where little glacial debris has contributed to form a ‘blocking’ terminal moraine, the proximal outwash zone may develop a braided network at the ice-margin itself. At Solheimaj¨okull in southern Iceland, for example, the glacier has recently advanced, overriding older outwash; it has generated only a small terminal moraine (2–3 m high), and is drained by over 100 small streams draining off the near-vertical ice-front. A third form of proximal channel pattern occurs where the proglacial topography is oriented parallel to the ice margin, causing lateral drainage diversion. At Skaftafellsj¨okull in southeastern Iceland, for example, successive retreat stages of the glacier over the past century (since 1870 AD) are marked by subparallel terminal moraine ridges up to 3 m high, cut by a single central channel (Thompson, 1988) (Plate 9.7; Fig. 9.8). During each recessional stage, drainage has been confined to a narrow lateral routeway, parallel to the ice margin and the confining moraine. Distal drainage patterns are commonly characterized by complex braided channel

networks, exhibiting highly unstable, broad, shallow channels (Plate 9.7). Width–depth ratios are widely used as a measure of channel cross-section shape. Braided proglacial rivers have width–depth values ranging from 5–15 in proximal Alpine streams to 17–50 on Icelandic sandar, except during a j¨okulhlaup when values may reach 207 (Churski, 1973). Several explanations have been proposed for the high degree of instability of proglacial braided channel networks. Primarily, instability is related to channel boundaries (especially the banks) being readily erodible. Bank erodibility is enhanced where bank sediments are non-cohesive containing a minimum content of silts and clays. In addition, if bank collapse occurs, flows of sufficient competence are required to entrain the sediment so that further bank faces can be exposed to erosion. Proglacial drainage systems clearly promote the development of channel instability, since fines are normally removed in suspension, leaving only medium- and coarse-grained sediments to form bank materials. Most proglacial zones lack full vegetation cover. Frequent fluctuations in meltwater flow levels promote repeated saturation of bank sediments. Finally, high magnitude diurnal flood flows allow regular removal of bank-foot material and continued attack on bank sediments. When channel widening occurs, a complex chain reaction of feedback responses ensues. Channel widening initially leads to a shallowing of flow, which in turn reduces the bed shear stress and hence the transport capacity of the flow. Deposition may occur within the channel crosssection or against one bank, leading to the development of a mid-channel or lateral bar, respectively. This process of braiding (Plate 9.8) is known as ‘primary anastomosis’. ‘Switching’ of dominant flows between each braid channel contributes markedly to the instability of braided networks, particularly in rapidly aggrading, relatively fine-grained systems, where bed scour is readily achieved and not obstructed by excessive armouring. Many bars in proglacial rivers are highly complex in morphology and sedimentology, reflecting a complicated history of aggradation (vertical), incision and accretion (especially lateral), as well as avulsion, splaying, channel abandonment, and chute and lobe development (Fig. 9.9). During high flows, bar-top

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

295

PLATE 9.7. Terminal zone of the Skaftafellsj¨okull Glacier, southeast Iceland, illustrating braided meltwater drainage routes oriented parallel with the ice-margin (left), confined behind a recent terminal moraine ridge (right).

sediments can become mobilized, leading to downstream migration of the bar-front to form a crudely bedded avalanche face. In gravelly environments, shallow bar-top flows lead to aggradation of a crudely horizontally bedded or sheet deposit, or an irregular spread of coarse gravels; in sandy environments, dunes associated with relatively deep, fast flows may be capped by waning stage ripples. Commonly, as channels shift and switch during high flows, large areas of bar sediments may be removed, while new bar units are accreted laterally (Bluck, 1982). A braid bar may therefore contain the record of a complex history, particularly in areas with relatively low rates of vertical accretion. In zones of highly active sediment transport and widespread inundation, each bar may represent the outcome of only the last single high-flow event, rather than of an unknown series of multiple past events.

The individual channels within a braided network commonly exhibit relatively low sinuosities, reflecting bank instability. The highest sinuosity channels are found in silt- and clay-rich alluvium. In nonvegetated, non-cohesive sediments, bank collapse occurs so readily that incipient meanders cannot be sustained, and the channel expands rather than resists, resulting in formation of a straighter path. 9.4. CHARACTERISTICS OF PROGLACIAL OUTWASH SEDIMENTS 9.4.1. Characteristics of Individual Particles 9.4.1.1. Particle size and size distribution It is widely assumed that ‘all’ sizes of sediment, ranging from large boulders to glacially abraded

296

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

1904

Not surveyed

1960

1945

1954

1968

1980

N

Not surveyed 0

1000 metres

Rivers

Lakes

Moraines

Supraglacial moraines

FIG. 9.8. Stages of proglacial moraine and drainage pattern development during historical glacier recession, Skaftafellsj¨okull, southeast Iceland (from Thompson, 1988; reproduced by permission of the Iceland Glaciological Society). (Compare with Plate 9.7.)

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

297

PLATE 9.8. Broad, shallow, gravel-bed meltwater stream, west Greenland valley-sandur. Flow is towards the camera.

‘powder’, are initially made available from the source glacier to the proglacial drainage system. Although this assumption may prove generally acceptable, there are a number of environments in which the size range of input sediments to proximal meltwater streams is curtailed or limited. First, where bedrock geology is composed of highly friable or soft materials much of the material released at the ice margin is likely to be composed of relatively fine sediments. Secondly, where significant debris inputs to the glacier are derived from air-fall deposits such as volcanic tephra (as on many glaciers in south Iceland and in Washington State, USA) or loessic sediments, high proportions of output materials will be fine-grained. There is also some tentative evidence that suggests that the sizes of material issuing at the ice margin also partly reflect the distance of englacial transport (Maizels, 1976). In addition, coarser clasts may be

found in the proximal zones of channels of larger glacierized catchments, reflecting the higher discharge and competence of these meltwater streams. Finally, there is evidence suggesting that natural size distributions of fluvial sediments are depleted in the 2–4 mm (–1 to –2N) size range, possibly representing the boundary between coarser multi-mineral clasts and finer single-mineral grains or reflecting the high entrainment mobility of these grain sizes. The size distributions of clasts on proglacial valley-trains and sandurs have been examined in a number of studies. Mean proximal clast sizes are reported to vary between about 50 mm (Bossons, French Alps; Maizels, 1979) and 1800 mm (Donjek River, Yukon; Rust, 1972). Many studies have demonstrated that clast sizes decrease significantly downstream (Figs. 9.10 and 9.11). The rate of downstream decrease in size is significantly related

298

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

to the rate of decline in gradient (Fig. 9.12), as well as to the lithology of the sediment and the initial clast size, with maximum rates of decline in mean clast diameter (b-axis) in areas with the coarsest initial sizes. Bluck (1987) has also shown that the rate of size decrease is proportional to the length of the river system itself. Bluck found that rates of size decrease are greatest on small alluvial fans and

lowest on vast alluvial plains such as that of the Indus river. The mean rates of downstream size decrease recorded in modern proglacial outwash deposits appear to conform to Bluck’s model. Downstream decline in sediment size has been widely attributed to the combined effects of abrasion and selective sorting (Bradley et al., 1972). Since rates of size decrease occur more rapidly than accounted

CRESCENTRIC BAR

LONGITUDINAL BAR

MEDIAL BAR

TRANSVERSE BAR

chute POINT BAR or LATERAL BAR

POINT BAR or RIVER SPUR BAR

chute

DIAGONAL BAR

DIAGONAL BAR

FIG. 9.9. Variety of bar types found in gravel-bed rivers, illustrating simple accumulation bars on the left (‘unit bars’) and erosively modified bars on the right (from Church and Jones, 1982; reprinted from Hey, R. D. et al. (eds), Gravel-Bed Rivers; Fluvial Processes, Engineering Mand Management, by permission of John Wiley and Sons).

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

299

2000

Mean Particle Diameter (mm)

1000

D

500 P

Tp SCp

Kp

BW L 100

WR

M

SKp

SCd Gp

50

B71 B73

SKd Gd

10 0.01

0.1

1

10

100

Distance from Ice-Margin (km)

FIG. 9.10. Downstream decline in mean particle diameter in a variety of modern glacial meltwater streams. B71, Bossons valley-train 1971; B73, Bossons valley-train 1973; BW, Bow River, Alberta (after McDonald and Banerjee, 1971); D, Donjek River, Yukon (after Rust, 1972); G, Gigjukvisl, Skei∂/ ararsandur, Iceland (after Bothroyd and Nummedal, 1978); K, Knik River, Alaska (after Bradley et al., 1972); L, Lewis River, Baffin Island (after Church, 1972); M, Mendenhall outwash, Alaska (after Ehrlich and Davies, 1968); P, Peyto outwash, Alberta (after McDonald and Banerjee, 1971); SC, Scott outwash, Alaska (after Boothroyd and Nummedal, 1978); SK, Skei∂/ arar River, Iceland (after Boothroyd and Nummedal, 1978), T, Tingin outwash (after Church, 1972); WR, White River, Washington State (after Fahnestock, 1963); d, distal zone; p, proximal zone.

180 90 45 pebble size mm Y 0

50 metres

X

W

FIG. 9.11. Downstream decline in maximum surface grain size on bars over a 300 m reach in the Lyngsdalselva River, northern Norway (from Ashworth and Ferguson, 1986; reprinted from ‘Interrelationships of channel processes, changes and sediments in a proglacial braided river’, by P. J. Ashworth and R. I. Ferguson, Geografiska Annaler, 1986, 68A, 361–371, by permission of the Scandinavian University Press).

300

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

for solely by abrasion, particularly for resistant lithologies, most downstream declines in sediment size are considered to reflect processes of selective deposition of coarser clasts. The processes that generate selective downstream sorting of grain sizes are, however, poorly understood and are currently the focus of research. The traditional explanation suggests that a decline in stream power downstream leads to the selective deposition of coarser clasts; thus bedload becomes progressively fine-grained. However, the fact that distal bar-heads commonly contain coarser grains than those found in proximal bar-tails indicates that coarse sediment can ‘by-pass’ zones of fine-grained sedimentation located upstream. The complexities of sediment transport pathways during sediment transporting events (i.e., normally during floods) have not yet been resolved

sufficiently to allow identification of the precise mechanisms of downstream particle sorting. Sediment size distributions of proglacial outwash deposits are highly variable, reflecting the range of source materials and depositional environments (Fig. 9.13). Hence, size distributions range from wellsorted, unimodal sands to poorly sorted, bimodal or polymodal, coarse gravelly deposits with a variety of sand-sized matrix materials. The former may occur more commonly in distal channel zones, the latter in proximal bar and channel environments, where rapid deposition of torrential mobile-bed sediments can lead to matrix sediments becoming trapped amongst coarser clasts during deposition. The exception is where fines have selectively accumulated in sheltered pockets or pools on the outwash surface, remain as waning-stage siltskins on bar-tops prior to removal by

2000

1000

Mean Particle Diameter (mm)

500 SC

P K

Bl BW WR 100 L 50

G

T

B

SK 10

10-2

10-1

1

10

102

Mean Gradient (m km-1)

FIG. 9.12. Changes in mean particle diameter of outwash sediments in relation to sandur gradient, for a variety of modern meltwater streams (for key see Fig. 9.10).

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

301

100

B

BAC KWA TER

D

FLO OD P

ET

LAI

N

C

TOP S

Percentage Fines

A

50 BEDLOAD ENVELOPES



< 15 Nm-2

7 to 202 53 to 319 121 to 402

A B C D

0 0.25 0.50

1.0

2

4

8

16

32

64

128

Particle Size (mm)

FIG. 9.13. Sediment size distribution curves (by weight) of surface (floodplain) outwash deposits, compared with those of bedload sampled in different ranges (A to D) of shear stress. Measurements of bedload were truncated at 0.25 and 76 mm (from Ashworth and Ferguson, 1986; reprinted from ‘Interrelationships of channel processes, changes and sediments in a proglacial braided river’, by P. J. Ashworth and R. I. Ferguson, Geografiska Annaler, 1986, 68A, 361–371, by permission of the Scandinavian University Press).

deflation, or where influxes of air-borne silt or dust are derived from erosion of older coastal, subglacial or valley side deposits. 9.4.1.2. Particle shape Many studies have demonstrated that clast form (especially roundness) is closely related to the agent and distance of transport and the environment of deposition, in response to the degree of abrasion (Menzies, 1995, chapter 15; 1996, chapter 13). Clasts deposited by meltwaters in glaciofluvial outwash environments may therefore exhibit a wide range of roundness values depending on the distance of transport (both within the glacier and the proglacial zone), the lithology of the clast and its resistance to abrasion and its transport history (King and Buckley, 1968). Two indices of shape have been widely adopted. First, the Cailleux–Tricart roundness index (RI ) measures the radius of curvature, rc , of the sharpest corner on the a–b plane, in relation to the

maximum particle diameter, a, such that RI = (2rc/a) × 1000. Roundness values lie between 0 and 1000, and are applied to limited size ranges only, as rc is likely to be a function of particle size itself. Cailleux roundness values recorded for clasts deposited in modern proglacial outwash environments range from as low as 0.03 to as high as 0.40, with the majority of values lying between 0.15 and 0.25. Variability within a single deposit can also be relatively high since roundness has been widely demonstrated to increase exponentially downstream in response to progressive attrition. A second and more widely adopted quantitative approach to clast shape was proposed by Sneed and Folk (1958). Three form ratios were derived, based on the three axial lengths of clasts, a, b and c, giving measures of symmetry (a/c), elongation ((a–b)/(a–c)), and sphericity (c2/ab). When plotted on a triangular graph or ‘form diagram’, ten shape classes were distinguished. Results from proglacial outwash sediments indicate clast shape partly determines its relative

302

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

ease of entrainment, with disc-shaped clasts being most mobile (Ehrlich and Davies, 1968; Bradley et al., 1972) (Fig. 9.14). However, it appears shape is largely determined by the original structure and composition of the source rock (Haldorsen, 1981). Where there are influxes of debris from valley sides, tributaries, reworking of older weathered materials or inputs directly from the ice, the spatial pattern of roundness variation can be highly complex.

Thus a stable imbricated or ‘shingled’ bed containing a strong, preferred, upstream fabric develops. A weak or random fabric is characteristic of rapidly fluctuating flows or deposition rates that are too rapid or sediment-laden to allow free clast rotation. Where clasts are pushed along the bed their long axis becomes oriented parallel with the flow, offering minimum resistance to fluid flow. 9.4.2. Facies of Proglacial Outwash Sediments

9.4.1.3. Clast fabric Where particles are subject to rotational stresses, such as imposed during bedload traction, clasts commonly exhibit a preferred orientation or fabric. In many fluvial environments, platy or discoidal clasts are typically orientated with the long axes of clasts perpendicular to the flow direction, while their a–b planes assume an upstream dip. Each clast becomes banked up against other clasts or obstacles on the bed.

(a)

(b)

STATION 55 N = 29 q = 141.6º L = 83.88%

COMPACT TYPICAL CLAST SHAPE

STA 11 N = 77

S/L

G SC LA O C TT IE R

STA 65 N = 35 q = 130.6º L = 78.01%

Three facies types have been widely recognized from modern proglacial outwash deposits on the basis of their dominant grain size: silt, sand and gravel (Rust, 1978; Menzies, 1996, chapters 8 and 9). The different facies commonly exhibit a distinct downstream distribution from gravels to sands to silts. However, lateral variations in facies type can also occur according to the nature of the local bar-and-channel network, including its scale, stability, relative relief

C

CP

CB

P

CE

E

B

FOLK FORM DIAGRAM

N 4

STA 8 N = 30 q = 078.4º L = 84.35%

5

VP

VB

PLATY

6

VE

L - I/L - S BLADED COMPACT

7

ELONGATE

8 9 10

VEC T ME OR AN

STA 65 N = 132

EXPLANATION

C

CLAST SIZE (L - AXIS) 30-39 MM 70-79 MM 40-49 80-89 50-59 90-99 60-69 >100

11 12

S/L

STA 11 N = 42 q = 097.7º L = 76.71%

TR OF EN FA D N

13 14 15

TOTAL N = 136 q = 111.2º L = 72.93%

P

VP

PLATY

CP

CB

CE

B

VB

E

VE

L - I/L - S BLADED

ELONGATE

FIG. 9.14. (a) Clast long-axis orientation, Scott outwash fan, Alaska, measured within 50 × 80 cm2 area on bartop surfaces. (b) Clast form distributions of particles from two sites on the Scott outwash fan, following classification of Sneed and Folk (1958) (both parts of figure from Boothroyd and Ashley, 1975, reproduced by permission of the Society of Economic Palaeontologists and Mineralogists (SEPM)).

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

and degree of inundation at different flow stages. Local variability in facies type can also occur where influxes of non-fluvioglacial sediments are deposited on the outwash. Aeolian sediments may be deposited as sheets or dunes, especially in coastal environments, while mud-flows and debris flows may be input from ice-contact slopes or valley sides. In addition, temporal variations in facies type can occur during individual

303

flood events, when falling-limb deposition may change from coarse- to fine-grained as flow competence declines. Hence, although large-scale downstream trends in facies types occur in most outwash deposits, local variations prove important. In 1977, a coding system for annotating facies types was introduced (Miall, 1977) (Table 9.2). Although problems arise in using and applying

TABLE 9.2. Miall’s (1977, 1978) classification and coding of facies, lithofacies and sedimentary structures of modern and ancient braided stream deposits (From Miall, 1978) Facies code

Lithofacies

Sedimentary structures

Interpretation

Gms

Massive, matrix supported gravel

None

Debris flow deposits

Gm

Massive or crudely bedded gravel

Horizontal bedding, imbrication

Longitudinal bars, lag deposits, sieve deposits

Gt

Gravel, stratified

Trough crossbeds

Minor channel fills

Gp

Gravel, stratified

Planar crossbeds

Linguoid bars or deltaic growths from older bar remnants

St

Sand, medium to v. coarse, may be pebbly

Solitary (theta) or grouped (pi) trough crossbeds

Dunes (lower flow regime)

Sp

Sand, medium to v. coarse, may be pebbly

Solitary (alpha) or grouped (omikron) planar crossbeds

Linguoid, transverse bars, sand waves (lower flow regime)

Sr

Sand, very fine to coarse

Ripple marks of all types

Ripples (lower flow regime)

Sh

Sand, very fine to very coarse, may be pebbly

Horizontal lamination, parting or streaming lineation

Planar bed flow (l. and u. flow regime)

Sl

Sand, fine

Low angle (<10°) crossbeds

Scour fills, crevasse splays, antidunes

Se

Erosional scours with intraclasts

Crude crossbedding

Scour fills

Ss

Sand, fine to coarse, may be pebbly

Broad, shallow scours including eta crossstratification

Scour fills

Sse, She, Spe

Sand

Analogous to Ss, Sh, Sp

Eolian deposits

Fl

Sand, silt, mud

Fine lamination, very small ripples

Overbank or waning flood deposits

Fsc

Silt, mud

Laminated to massive

Backswamp deposits

Fcf

Mud

Massive, with freshwater molluscs

Backswamp pond deposits

Fm

Mud, silt

Massive, desiccation cracks

Overbank or drape deposits

Fr

Silt, mud

Rootlets

Seatearth

C

Coal, carbonaceous mud

Plants, mud films

Swamp deposits

P

Carbonate

Pedogenic features

Soil

304

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

PLATE 9.9. Large-scale trough cross-stratification formed by migrating dunes, capped by horizontally bedded sands and silts associated with shallow, waning-stage sheet flows, Myrdalssandur, south Iceland.

PLATE 9.10. Type ‘A’ ripple-drift cross-lamination overlain by Type ‘S’, sinusoidal ripple laminations, Late Devensian outwash, NE Scotland. Flow was from left to right.

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

Miall’s classification scheme to a wide variety of complex sediment types, the approach has been widely adopted to describe in simple terms the sedimentary characteristics of fluvial and fluvioglacial deposits (Miall, 1983; Menzies, 1996, chapter 9). Many workers have extended or modified the classification scheme (Eyles and Miall, 1984; Brodzikowski and Van Loon, 1991). The use of facies codes has been widely adopted as the basis for identifying facies assemblages that characterize particular kinds of fluvial environments. These assemblages in turn form the basis of fluvial facies models. However, it must be remembered that the codings act only as descriptive summaries of the sediment characteristics that must be supplemented by field and laboratory data. They do not necessarily indicate the regional and local hydraulic or sedimentary controls that may have generated those deposits.

305

moved only during episodic high-flow events, while short-term storage relates to finer-grained sediments moved at all but the lowest flows. These storage effects may be altered where the bed is strongly armoured. A wide range of bar types has been recognized in fluvial and fluvioglacial environments and a myriad of terms has been used to describe them according to their plan form shape, location, relief and texture. Sandy bars

9.4.2.1. Fine-grained facies The fine-grained facies include laminated (Fl) and massive (Fm) silts and clays. In proglacial environments, massive fines may be found as drapes or ‘siltskins’ across bar-tops, deposited from suspension during waning flows. Laminated fines accumulate in pools, abandoned channels, kettle holes and lakes. The laminae, of coarse and fine layers, are commonly interpreted as annual varves, in which the coarse grains represent summer influxes, the fine grains settling out slowly during the winter. However, in many proglacial environments such sediment couplets reflect sedimentation from short-term ‘pulsed’ sediment influx events associated with turbidity currents produced during diurnal or periodic floods (Russell and Knudsen, 1999). 9.4.2.2. Sand facies A wide range of sand facies has been recognized in proglacial meltwater environments, the most common are associated with the downstream migration of bars, and the formation of fluvial dunes and ripples (Plates 9.9, 9.10 and 9.11). Bars are the fundamental depositional unit within proglacial meltwater environments acting as the primary sediment storage zones. Long-term storage relates to coarse sediment that is

PLATE 9.11. Type ‘A’ ripple-drift cross-laminations in coarse sand and gravel of Late Devensian outwash, NE Scotland. Note local imbrication of latter clasts.

306

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

typically form as point bars in meandering and anastomosing channels, and as transverse, ‘linguoid’ or ‘spool’ bars in braided environments (Fig. 9.15). While dunes develop within the major channels, rippled sands may accumulate locally during shallow flows across bar tops and within shallow channels cut into the bar surface. However, the preservation potential of these deposits is low, typically being swept away during subsequent high-flood events. Horizontal bedding most commonly develops as upper (coarse sand) and lower (fine sand) flow regime plane beds that are confined to shallow flows across bar-tops or sandflats. Massive sands, by contrast, are more characteristic of suspension flows, in which rapid deposition occurs within deep channels. 9.4.2.3. Gravel facies The most common gravel facies within glacial outwash deposits are clast-supported, imbricated, heterogeneous, sub-rounded, bimodal or polymodal gravels exhibiting poorly defined bedding. This facies was described by Miall (1977) as Gm, representing massive or horizontally bedded gravel facies (Plate 9.12). However, the wide variety of textures, grading, fabric and internal structures typical of glacial outwash are poorly represented by this single descriptive code. Some outwash deposits consist of wellimbricated, horizontally bedded gravels composed of a sand matrix and a fine gravel clast fraction deposited by gradual accretion during moderate flows; while others are composed of massive, struc-

tureless, polymodal, weakly imbricated cobble gravels, where all size fractions have been deposited rapidly during torrential flood flows. Horizontally bedded gravels occur widely in proglacial braided river environments, although bedding planes are normally indistinct or poorly developed. Crudely bedded imbricate gravels form the nucleus of many gravel bars and gravel sheets, where flows are shallow relative to the mean particle size of the bed material. Since many proglacial bar-channel networks exhibit a low relative relief and are subject to sheet floods across the outwash surface, this gravel facies tends to dominate many bar and channel deposits (Rust and Koster, 1984). The most common gravel facies bar type, the longitudinal bar, extends parallel to flow, and is bounded by low sinuosity channels (Plate 9.7; Fig. 9.9). Longitudinal bars are composed of a core or platform of gravels on to which are accreted horizontally bedded, imbricated gravels formed through shallow sheet flows. Bar front sediments may exhibit crude planar cross-bedding deposited during downstream and lateral bar progradation (Plate 9.13). Proglacial outwash, especially in proximal zones, may contain massive, matrix-supported gravels (Miall, 1977, Gms). These deposits have a high (<50 per cent) matrix context, are poorly sorted, without bedding or imbrication and often contain large isolated, angular clasts. These deposits are widely interpreted as debrisor mud-flow facies, associated with viscous, sediment rich (>70 per cent) meltout flows from the ice-margin, solifluction deposits from adjacent slopes or facies

Flow

3m

FIG. 9.15. Large-scale trough cross-stratification formed by migrating dunes.

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

307

PLATE 9.12. Contact imbrication of clasts in massive gravel outwash, Late Devensian, NE Scotland.

PLATE 9.13. Coarse, crudely bedded gravels forming longitudinal bars, Solheimaj¨okull outwash, south Iceland. Flow is from right to left. Note glacier snout in background.

308

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

produced by saturation and mobilization of preexisting sediments. Where sediment-laden flows still contain sufficient fluid content to sustain low viscosity flows (i.e., ca. 40–70 per cent sediment concentrations), hyperconcentrated grain flows occur (Plate 9.14) and an intermediate gravel facies may be deposited. Grain flow deposits are commonly composed of massive, matrix-supported gravels, exhibiting distinct inverse grading overlying a basal, imbricated lag gravel. These deposits reflect the importance of internal dispersive stresses acting on grains in suspension, as well as basal shear stresses operating at the flow–bed contact. Debris flows commonly occur as steep, lobate deposits or ‘plugs’ where they are channelized, while grain flows may also be deposited as broad gravel sheets across the outwash plain.

be composed of a wide variety of lithologies, reflecting the geological structure of the glacierized basin or of a single lithology, where meltwaters have been directed along selected bedrock routeways. Boulders may occur singly or as localized clusters, reflecting transport on ice-rafts or bergs, or by buoyant high viscosity debris or grain flows, particularly during j¨okulhlaups. Boulders accumulate particularly on the bar heads where they can form large, imbricated spreads, or clusters. The clusters act as obstacles to the transport of clasts from upstream, which then accumulate on the upstream side of the boulder cluster and accentuate the development of a pendant bar and flow separation into lateral channels. 9.4.3. Facies Models in Proglacial Environments

9.4.2.4. Boulder facies Boulder deposits are often found within proglacial meltwater environments. They may form distinctive spreads, both in proximal and distal zones, and may

Facies models of proglacial meltwater environments have largely focused on the braided river sedimentary environment (Fig. 9.16) (Menzies, 1996, chapter 9). Studies of braid bar and channel development, bar

PLATE 9.14. Massive, hyperconcentrated grain-flow deposit, representing main flood surge during 1918 j¨okulhlaup from Katla, on to Myrdalssandur, south Iceland.

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

309

bars

flat stratified sand A

0.3 - 2.5 m

flat stratified sand A flat stratified sand B

cross stratified sand A

B

flat bedded sand and gravel INCREASING FLOW DEPTH

FIG. 9.16. Bluck’s (1979) model of fine-grained braided stream alluvial deposits in wide, shallow channels. Flat stratified sand A deposited by bars of low relief shown on the surface of the diagram; flat stratified sand B refers to the gently inclined and flat strata. Cross-stratified sands A are deposited by dunes or megaripples; B, cross-stratified sand sheets. The left-hand sequence reflects shallower flows than that on the right (from Bluck, 1979; reprinted by permission of the Royal Society of Edinburgh).

sedimentology, braiding intensity and spatial trends in network characteristics have been carried out in the proglacial outwash environments of Iceland (Rust, 1978; Maizels, 1993), Alaska (Gustavson and Boothroyd, 1987), Baffin Island (Church and Gilbert,

1975), the Cordillera (Hammer and Smith, 1983), the European Alps (Fenn and Gurnell, 1987), Scandinavia (Ashworth and Ferguson, 1986), Greenland (De Jong, 1992), the Himalayas (Hewitt, 1982) and New Zealand (Rundle, 1985). b

Bank

BAR

BAR HEAD mostly recording high flow stage activity

a Main

chan

nel

TAIL

record high and mostly low flow stage activity

flow

mud sand gravel

a

some (low in section) retained at BF stages

most replaced on BF stages

b

FIG. 9.17. Bluck’s (1982) simplified model of lateral bar sedimentology in gravel-bed rivers. BF refers to bankfull discharge (from Bluck, 1982; reprinted from Hey, R. D. et al. (eds), Gravel-Bed Rivers, by permission John Wiley and Sons).

310

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

9.4.3.1. Bar and channel facies models for braided rivers and glacial outwash

inclined sheets of sand and gravel, horizontally bedded lag gravels and crude foreset sands and gravels. These distal bar sediments are subject to reworking during all but the lowest flows (Figs. 9.17 and 9.18). In addition, channel facies are characterized by extensive coarse armouring and localized formation of transverse ribs in steep, shallow reaches. A large number of facies models have been proposed for sandy, distal braid bar (Fig. 9.19) (Cant and Walker, 1978) and anastomosing channel (Smith, 1983) environments.

Analysis of small-scale variations in the sedimentology of braid bars and their sedimentary structures has been carried out on proglacial outwash (Bluck, 1982). Bar-head sediments are composed of coarse imbricated gravels forming horizontally bedded spreads, transverse ribs and local sand shadows subject to entrainment and transport only during high flows. The bar-tail sediments, by contrast, consist of gently

FLOW STAGE

BAR

SUPRA

head

P L AT F O R M

tail

inner channel

coarse

high flat or greatly inclined sheets of sand and gravel

imbrication

megaripples

transverse ribs on earlier imbricate frabric

A

bar lee

chute bars

fine gravel infilling coarse

fine

B cross stratified sand sheet A

sand shadows and sand sheets

low

modified sand shadows

B

A

B A B

deltas lag gravels

ripples and drapes

splits

FIG. 9.18. Bluck’s (1979) model of sedimentary structures of braid bars, indicating structures representative of high- and low-flow stages on the bar-head, bar-tail and inner (chute) channel (from Bluck, 1979, reprinted from Hey, R. D. et al. (eds), Gravel-Bed Rivers, by permission John Wiley and Sons).

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

1

EXPOSED COMPLEX SAND FLATS

SAND WAVES

6

5

SINUOUS-CRESTED DUNES

1

CROSS-CHANNEL BAR

2

VEGETATED ISLAND

311

4 3 4

5

5

6 5

A

3

LAG

BANK

B C 3

EMERGENT NUCLEUS

BANK

FIG. 9.19. Cant’s (1978) facies model of a sandy, braided river, illustrating the sedimentary structures characteristic of both the active channel environment and of preserved deposits. The inset diagram shows a vertical view of the same area. The circled letters A, B and C show the locations of formation of the sand flat, mixed influence and channel sequences, respectively (see text for description) (from Cant, 1978, permission of the Canadian Society of Petroleum Geologists).

9.4.3.2. Proximal-distal facies models for braided rivers and glacial outwash Facies models of glacial outwash stress the importance of proximal–distal changes in sedimentology, reflecting downstream declines in gradient, stream power and maximum clast size, and a downstream succession from massive or horizontally bedded, longitudinal gravel bars, deposited by vertical accretion in the proximal zones, to mainly trough crossbedded, cyclic fining-upward sequences of gravels, sands and silts in the distal zones. The proximal zones of the Scott and Yana outwash fans in Alaska, for example, are dominated by a coarse-gravel facies consisting of well imbricated pebbles, cobbles and boulders associated with high flows across low bed relief bars (<30 cm), and longitudinal bars up to 50 m long. Towards the mid-fan zone the gravel facies exhibit a marked downstream decline in maximum clast size. The gravel is interbedded with

plane- and large-scale trough cross-bedded sands resulting from ‘megaripple’ migration in low-stage channels. The lower fan is dominated by slip-face migration of longitudinal and linguoid braid bars, giving rise to planar cross-beds, while bar-top formation of ripples produces abundant ripple-drift cross-lamination. The meltwater channels begin to meander as the outwash sediments become increasingly fine-grained, and the most distal facies exhibit ‘megaripple’ migration on point-bar surfaces producing large-scale trough cross-beds and migration of the bar slip face producing planar to tangential cross-beds (Miall, 1977, 1978; Rust, 1978). Miall recognized six principal facies assemblages in gravel- and sand-dominated braided river deposits (Fig. 9.20; Table 9.2). Two kinds of proximal facies were identified: the Trollheim Type (equivalent to Rust’s G1 lithotype (Table 9.3)), which is dominated by massive, clast-supported and matrix-supported gravels representing deposition in longitudinal bars

TROLLHEIM TYPE Sr

Gms

SCOTT TYPE

DONJEK TYPE

S. SASKATCHEWAN TYPE

PLATTE TYPE

BIJOU CREEK TYPE

Gm

superimposed debris flows

St

minor channel minor channels

Gm Gm Gms Fm St

Sp Gm Sh

bar-edge sand wedge

Fl Sr Ss St Sp

Gm

St Sh St

debris flow deposit Gm

stream flow channel

Sr

Gm

St

Sp Sh Sr Sh

minor channel or channel system

major channel Se Sr

Sr

superimposed flood cycles

Sp Fl St Sh Sp

Sp Gm Fl

Gm

Sh

Gm

compound bar

Gt

Gm

longitudinal bar

Sp

Gm

superimposed linquoid bars

Sl

superimposed bars

5 m

St Gm

0

FIG. 9.20. Miall’s (1978) vertical facies profile models for braided stream deposits. Facies coded to the left of each column are given in Table 9.2. Arrows show small-scale cyclic sequences (from Miall, 1978), permission of the Canadian Society of Petroleum Geologists).

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

313

TABLE 9.3. Miall’s (1977, 1978) classification of the six main facies assemblages in gravel- and sand-dominated braided river deposits (from Miall, 1978) Name

Environmental setting

Main facies

Minor facies

Trollheim type (Gl)

Proximal rivers (predominantly alluvial fans) subject to debris flows

Gms, Gm

St, Sp, Fl, Fm

Scott type (Gll)

Proximal rivers (including alluvial fans) with stream flows

Gm

Gp, Gt, Sp, St, Sr, Fl, Fm

Donjek type (Glll)

Distal gravelly rivers (cyclic deposits)

Gm, Gt, St

Gp, Sh, Sr, Sp, Fl, Fm

South Saskatchewan type (Sll)

Sandy braided rivers(cyclic deposits)

St

Sp, Se, Sr, Sh, Ss, Sl, Gm, Fl, Fm

Platte type (Sll)

Sandy braided rivers (virtually noncyclic)

St, Sp

Sh, Sr, Ss, Gm, Fl, Fm

Bijou Creek type (Sl)

Ephemeral or perennial rivers subject to flash floods

Sh, Sl

Sp, Sr

and debris flows, respectively; and the Scott Type (Rust’s G11 Lithotype), which is largely composed of massive or crudely bedded, clast-supported gravels typical of longitudinal braid bars. Five distal facies have been recognized by these authors. A distal gravelly (gravel content of 10–90 per cent) river facies (Miall’s Donjek Type; Rust’s G111; Rust, 1972), dominated by fining-upward cycles of crudely bedded coarse gravels and trough crossbedded sands and gravels, reflects deposition of dunes on to gravel lag during infilling of sandur channels; minor effects include the migration of linguoid bars and sand waves. Distal braided rivers dominated by linguoid bars, dunes and ripples are Miall’s South Saskatchewan Type (Rust’s S11 lithotype) in which vertical sequences are mainly composed of cyclic channel fills and trough cross-bedded sands, while planar cross-bedded, horizontally bedded, and scour fill sands, and silty laminae are of minor importance. Non-cyclic vertical sequences (Miall’s Platte Type; Rust’s S11 lithotype) contain higher proportions of planar crossbeds. Miall also recognized a distal facies assemblage associated with flash floods dominated by sheet flow deposits and horizontally bedded sands (Miall’s Bijou Creek Type; Rust’s S1 lithotype). The finest-grained distal facies assemblage as described by Rust (1978) is dominated by laminated and massive silts and clays. These are found particularly where aeolian

reworking and thixotropic deformation are common, as in tidal flat areas. 9.4.3.3. J¨okulhlaup facies model for glacial outwash Most sedimentological studies of modern outwash deposits have focused on braided river sedimentation, only briefly referring to the incidental effects of large flood events. Examination of many such studies suggests, however, that sedimentation in some outwash environments may be dominated by j¨okulhlaup activity. In Iceland, for example, studies of sandurs (Russell and Knudsen, 1999) emphasize the role of j¨okulhlaups in the development of terrace sequences, kettle holes and pitted sandur, boulder lobes and spreads, and large bar deposits. In addition, largescale ‘megaripple’ cross-strata, some with coarsening-upward foresets, have been interpreted as flood deposits, as have planar gravel cross-strata on proximal longitudinal bars. Miall’s Bijou Creek facies type also recognizes the impact of repeated flood events on the cyclical nature of the vertical sediment sequence, albeit in a distal, sandy environment. In addition to large-scale gravel cross-beds, j¨okulhlaups generate sediment sequences associated with deposition from hyperconcentrated grain and debris flows, particularly in areas of abundant sediment supply. Some of the Icelandic sandurs have been found to be largely composed of thick units of black

314

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

pumice, granular gravels deposited during volcanically generated j¨okulhlaup events. Analysis of outwash sediments on Solheimasandur (Maizels, 1992) and Myrdalssandur (Jonsson, 1982; Maizels, 1993) forms the basis of a facies model of volcano-glacial j¨okulhlaup sandur deposition. The sandurs are composed of a number of individual flood deposits, up to 12 m thick in proximal zones and only 1–2 m thick is distal zones (Fig. 9.21). Each flood deposit is

composed of four sedimentary units, reflecting successive stages of the j¨okulhlaup event. The rising flood wave is marked by basal imbricated gravels with a fine pumice matrix; the main flood surge is represented by thick (<5 m), homogeneous, massive, granular pumice gravels representing grain flows; the onset of the waning stage characterized by migration of large pumice dunes along the main flow paths leading to formation of large-scale trough cross-beds;

Depth Below Surface (m)

0

1

GRh, Sh

2

GRt

3

4 GRm 5

6

7

8

Gm, Gic +2

0

-2

Median Particle Size (f mm)

1

3

Sorting Index (f mm)

0

10

% Silt & Clay

FIG. 9.21. Maizels’ (1989a, b, 1991, 1992) vertical facies profile model for j¨okulhlaup sandur deposits where main flood surge is characterized by hyperconcentrated grain flows, to form central massive pumice, granule unit (GRm), overlying basal gravels, and overlain by waning-stage trough cross-bedded and horizontally bedded granular gravels, sands and silts (from Maizels, 1992; reprinted from ‘Boulder ring structures produced during j¨okulhlaup flows – origin and hydraulic significance’, by J. K. Maizels, Geografiska Annaler, 1992, 74A, 21–33, by permission of the Scandinavian University Press) (compare with Plates 9.9 and 9.14).

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

315

3 2 10

6

5 9

4 8 7 1 1 2 3 4

Massive pumice gravels Palagonite boulder gravels Top sandur terrace Distal sandur

5 6 7

Lobate pumice fan 8 Incised jökulhlaup channel + streamlined residual hummocks, boulders and megaripples 9 10 Distal jökulhlaup deposit ('Lahar')

Washed sandur Incised jökulhlaup channel Incised active meltwater river

FIG. 9.22. Schematic model of j¨okulhlaup sandurs, illustrating range of depositional sequences and landforms and erosional features (from Maizels, 1991; reprinted with permission of Kluwer Academic Publishers).

while shallow sheet flows marking the final stage of the j¨okulhlaup are represented by horizontal laminations of pumice granules, sands and silts. The main flood surge unit is thickest, and the dune and sheetflow units thinnest, in proximal zones; clast sizes exhibit little downstream decrease but decline rapidly in the dune and sheet-flow units, reflecting the mass transport of sediment in the former and bedload traction in the latter. Debris flow deposits also occur locally, wherever older outwash, slope or soil deposits are in contact with and entrained or mobilized by j¨okulhlaup flows. J¨okulhlaup outwash facies may also include a distinctive assemblage of bedform structures and morphological features. In addition to large-scale bar deposits, boulder accumulations and fields of ‘megaripples’, j¨okulhlaup outwash deposits may exhibit a range of kettle-hole forms, including: ‘till-fill kettles’ formed by the in situ melting of debris-rich ice blocks

transported by the flood; streamlined obstacle marks, residual bars and hillocks; washlimits on bedrock obstacles and valley sides; and thick fine-grained or laminated sequences associated with temporary ponding of flood waters behind valley constrictions (Maizels and Russell, 1992; Russell and Knudsen, 1999) (Fig. 9.22; Plate 9.15). 9.5. ISSUES AND FUTURE PROSPECTS General models of the landform and sediment assemblages that characterize proglacial terrestrial environments are now relatively well-established, particularly with respect to changes in outwash morphology, channel pattern characteristics and facies variations in proximal and distal zones. There are substantial data on seasonal and diurnal fluctuations in meltwater runoff in proglacial environments in a variety of global settings. However, there are two

316

SEDIMENTS AND LANDFORMS OF MODERN PROGLACIAL TERRESTRIAL ENVIRONMENTS

PLATE 9.15. Scattered boulders, scour marks and residual hillocks formed during j¨okulhlaup event on Skogasandur, south Iceland (from Maizels, 1989a). Flow was from right to left.

major areas of study that remain poorly represented. First, there are no substantive studies that have successfully modelled channel responses to fluctuating runoff and sediment inputs to the proglacial drainage system. The complex interactions between variations in runoff, sediment transport and the processes of channel and drainage pattern development remain at a semi-quantitative stage. Secondly, the temporal and spatial patterns of channel change and facies distributions need to be viewed within an appropriate conceptual framework. One approach may be through determination of the magnitude and frequency of glacial runoff and sediment yield events that generate: (a) changes in channel morphology and pattern; (b) deposition of typical facies types; and (c)

patterns of scour, sedimentation or storage that may reflect longer-term or more permanent changes within the proglacial system. This approach should allow facies distributions and sequences to be modelled in relation not only to channel system dynamics and the controlling patterns of the runoff and sediment inputs, but also to glacier system dynamics. Over the longer term, the response of proglacial meltwater systems to large-scale environmental changes needs to be modelled. While qualitative and semi-quantitative models are available at present, much still needs to be done in predicting fluvial response to deglaciation, eustatic and isostatic changes, climate change and associated changes in meltwater runoff and sediment supply regimes.

10

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS W. H. Johnson and J. Menzies

10.1. INTRODUCTION Supraglacial and ice-marginal deposits and landforms are significant and important components of the glacial record. These deposits commonly are the uppermost portion of the sediment sequence locally and dominate the landscape; their properties and characteristics have a profound influence on land use. In areas of multiple glacial events, the occurrence of proglacial and/or supraglacial sediment between deposits of subglacial sediments permit the delineation of ice-marginal fluctuations in the stratigraphic record (Chapter 15). Supraglacial and ice-marginal landforms leave a distinctive imprint on the landscape and record the nature of glacial retreat, whether by marginal backwasting or widespread downwasting, and a history of ice-marginal fluctuations. Supraglacial and ice-marginal environments are transitional to and have close ties with both the subglacial and proglacial environments (Fig. 10.1). The supraglacial environment, by definition, occurs on the surface of glacier ice, which separates it from the subglacial environment. These two glacial subsystems come closer together as the intervening ice ablates and commonly are interconnected by englacial water passages within the ablation zone. Toward the icemargin, continuity with the proglacial environment

develops as the ice thins, and the supraglacial environment, as well as the subglacial environment, give way to proglacial conditions. In many situations it is impossible to distinguish sediments deposited in one environment from those of another in these transitional zones. From a sedimentary and geomorphic standpoint, the supraglacial environment is most important along the outer zones of glaciers and ice sheets where sediment becomes concentrated on the surface of the ice, either derived from valley and nunatak slopes that rise above the ice mass or through ablation and concentration of basal and englacial debris in the icemarginal zone. The former situation was a common occurrence in most major mountains of the world during the Pleistocene, as cirque and valley glaciers and mountain ice caps formed, and large valley glaciers extended for many kilometres to lower elevations. These glaciers left a sediment and morphologic record that shows the extent of the various glacial advances and their subsequent fluctuations, and from which a chronology has been developed for the most recent portion of the Pleistocene in many glaciated valleys and mountain areas. For the large ice sheets and smaller ice caps that formed in the past, it was basal and englacial debris primarily that became concentrated on the ice surface and was

317

318

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

ion

tat

gla

pra

su

f en a o dim Arerese l cia

Active glacier Fluvial deposits Debris band in stagnant ice Debris flow deposit Lacustrine deposits Supraglacial sediment (undifferentiated) Buried stagnant ice

FIG. 10.1. Schematic diagram of supraglacial and ice-marginal environments; subglacial sedimentation beneath active-ice to the right and proglacial deposition to the left. Basal debris is concentrated on the glacier surface by ablation and undergoes resedimentation by mass wasting, fluvial and lacustrine processes as the glacier downwastes (reprinted with modifications from Edwards, 1986, in Reading, H. G. (ed.), 1986, by permission of Blackwell Scientific Publications).

subsequently reworked in the supraglacial environment during backwasting and downwasting of the ice sheet (Fig. 10.1). Areas where Pleistocene supraglacial and icemarginal deposits are most common and important are the lower and lateral portions of glaciated valleys and troughs, the marginal areas of former ice sheets where the ice-margin position was maintained for some length of time, areas where the ice mass flowed up an escarpment or regional slope, and in former interlobate positions where adjacent glacier bodies flowed toward each other or joined in the case of valley glaciers. In these situations compressive flow moved sediment toward the glacier surface where it became concentrated through ablation. These deposits are not

restricted to the outer and lower zones of glaciers and ice sheets, however, and locally they can be a significant component of the glacial sequence in any former glaciated area. 10.2. SEDIMENT AND SEDIMENT ASSOCIATIONS 10.2.1. Types of Sediment and Sediment Variability Pleistocene supraglacial sediment sequences contain the entire range of clastic sediment types, from coarse gravels to fine clays to cobbly and pebbly diamictons with a wide range of matrix textures (Plate 10.1). The

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

319

Although some supraglacial sequences are dominated by a particular type and texture of sediment, most are characterized by lateral and vertical variability. Beds are relatively thin and discontinuous, and consist of well- to poorly sorted lacustrine and/or fluvial deposits interbedded with poorly sorted mass wasting deposits of variable texture. Among the latter sediments, debris flow deposits are particularly common but they, in turn, vary considerably depending on sediment texture and water content of the flow when active. Many are massive diamictons, <2 m thick, that have a primary ‘till-like’ appearance and are called ‘flow till’ in many glaciogenic sediment classifications (Dreimanis, 1988) or supraglacial subaerial mass-transport deposits by Brodzikowski and Van Loon (1991). Crude stratification and internal sorting (grading) are evident in some of these deposits, with diamictons commonly containing irregular unlithified sediment clasts. Rock fall, slump and colluvial accumulations locally are associated with the debris flow and sorted deposits. Collapse structures from melting of underlying or adjacent ice, deformational shear structures from the moving debris flows or overriding active ice and soft sediment deformational structures from loading of saturated sediment are common. PLATE 10.1. Sediment sequence near Two Creeks, Wisconsin, USA, with subglacial till in lower part and supraglacial sediment composed of variable silts, sands and diamicton in the upper part. Handle of shovel 0.5 m (photography by D. W. Moore).

broad range of sediment type reflects the different processes that were active in the environment, namely: fluvial, lacustrine and mass wasting, and the dynamic and continually changing environmental conditions in response to controls by glacier dynamics, bed topography and materials, regional and local climate, and sediment availability. The type of sediment record that developed in any one particular area and/or site depended on the interaction among these factors and the processes that were most active during resedimentation. Because both conditions and processes varied spatially and temporally, almost every site has its unique lithologic and morphologic characteristics. In spite of this variability, some general characteristics are common to most supraglacial sequences.

10.2.2. Areas of Low Relief Extensive lowland areas of northern North America, Europe and Asia were glaciated during the Pleistocene. As a case example, studies along both the eastern marginal area and the north-central portion of the Pleistocene Des Moines lobe in Iowa illustrate the variability of the supraglacial sediment sequence (Kemmis et al., 1981; Kemmis, 1991). The Des Moines lobe of the Laurentide Ice Sheet reached its maximum extent during the last glaciation about 14 000 BP. This advance and several subsequent readvances have been interpreted as surges followed by regional stagnation. Along the eastern margin of the lobe, the glacial ice flowed up a regional slope. Borings and exposures in this area contain an upper zone of silts and sands interbedded with thin and variable diamictons that are interpreted to be of supraglacial origin. These deposits overlie a zone of texturally uniform diamicton interpreted to be

320

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

UNIT Surficial Sediments Sandy Loam Diamicton Silty Diamicton

HORIZON or ZONE

DEPTH M Ft 0

1 2 3 4

Solum OL

1

Supraglacial sediment

OU

25

PARTICLE SIZE %

50

SAND CLAY

* 5 B.d.

FINE SILT

6

2

Stratified Silty Diamicton

COARSE SILT

8 1.69

Stratified Loam Diamicton

3 10

Sand and Gravel MOU

1.57

Silt Loam Silty Diamicton

12

MRU

Silts UU

Till

4 14 1.82

UU

5

Subglacial till

100

1.51

Sand Silt Loam

Silty Diamicton

75

(7/13/79)

16 1.88 1.85

18 1.90 6

20 1.96

*Bulk density: g/cc

22

64LH3

FIG. 10.2. Sediment log, weathering zones, matrix grain-size data and bulk density for a core taken through a supraglacial and subglacial sediment sequence in Story County, Iowa, USA. Weathering symbols: OL, oxidized leached; OU, oxidized unleached; MOU, mottled oxidized unleached; MRU, mottled reduced unleached; UU, unoxidized unleached. The supraglacial sequence above 4 m is texturally variable and less dense than the lower subglacial till (after Kemmis et al., 1981).

subglacial till (Fig. 10.2). The thickness of the two zones varies, for example in some areas, particularly toward the margin of the glacial advance, the supraglacial sediment facies dominates and little or no subglacial till is present; in other areas both facies are well developed (Fig. 10.2). Elsewhere, and particularly in the up-ice direction, the supraglacial facies is thin and completely absent in some areas. Most of the diamictons in the supraglacial facies are likely to be of debris flow origin, but some may be meltout till or ice-slope colluvium. These diamictons differ from those of subglacial till in being thin, a few centimetres

to tens of centimetres thick, discontinuous and often stacked with diffuse to sharp contacts. Locally, vertical variations in both matrix texture and pebble size occur, as well as crude layering and deformation structures. These diamictons are less dense than the subglacial diamicton and, although regionally the mean textures of the supraglacial diamictons and the subglacial diamictons (till) are about the same, the former are more variable (Fig. 10.3). Figure 10.4 shows the complex sediment relationships in the upper portion of a hummock in an upland area of moderate relief. The upper sediment sequence

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS Summary, Supraglacial till and diamictons only:

SUPRAGLACIAL SEDIMENTS D = Till and diamictons

100 0

S = Sorted sediments

90

n = 65 10

80

20

70

x s.d.

CLAY 14.1 ±5.2

SILT 42.3 ±14.4

SAND 43.6 ±15.0

= Till B = Till with basal inclusions S = Sorted sediments with till

90

80

20

30

CL AY

40

50

50

%

%

SAND 47.7 ±3.4

T SIL

T SIL

50

SILT 36.7 ±3.2

%

%

CL AY

CLAY 15.5 ±2.4

x s.d.

10

60

40

60

30

70

10

60

50

40

30

20

10

80

20 90

70

70

30 80

80

60

40

20

90

n = 185

100 0

70 40

50

0 100

Summary, Subglacial till:

SUBGLACIAL TILL

30

60

321

90

10 100 0

0 100

90

80

70

60

% SAND

50

40

30

20

100 0

10

% SAND

FIG. 10.3. Regional matrix grain-size data for 250 supraglacial and subglacial (basal) till samples from the southeastern margin of the Des Moines Lobe in central Iowa, USA. Although mean grain sizes are relatively similar, the large standard deviations for the supraglacial deposits indicate their textural variability (after Kemmis et al., 1981).

contains pebbly gravel and coarse sand associated with mostly massive silty sand and sandy silt; both are interpreted to be glaciofluvial sediment deposited in a supraglacial position. The underlying sediment, also supraglacial in origin, is composed primarily of interbedded sands and diamicton. Within the dia-

micton sequence, individual diamicton beds are distinguished by either subtle colour differences, reflecting variations in texture or porosity, or sand laminae at contacts. Contacts often become diffuse laterally and individual beds become indistinguishable in massive diamicton. Where distinguishable,

A Horizon of Modern Soil Profile

diamicton, loam and sandy loam textured matrix

dominantly fine sand

fine gravel (with some coarse sand)

loam sediments (silty sand and sandy silt)

dominantly coarse sand

laminations of sand within the diamicton

dominantly medium sand

animal burrows

0

.5

1

metres

FIG. 10.4. Sketch of a road-cut exposure of supraglacial sediment from a hummock in Winnebago County, Iowa, USA. Diamiactons, including the loam sediment, vary in pebble content; animal burrows indicate the friable nature of the upper sediment. Notice stacked channel deposits, faulting and evidence for subsidence and collapse (reprinted from Kemmis, 1991, by permission of the author).

322

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

beds commonly are 5–20 cm thick. Associated sorted deposits, some as distinct channel fills, have abrupt but often contorted or folded contacts with the diamicton. These deposits dip in different directions at varying angles, and are cut by high-angle normal and reverse faults. The diamicton is interpreted as consisting of debris flow deposits with thin meltwater deposits; the sand laminae, interbedded with glaciofluvial and glaciolacustrine deposits. The contorted beds and faults suggest deposition on ice with accompanying soft-sediment deformation, followed by collapse as the supporting ice melts. Although no data are available at this site, Kemmis (1991) reports generally weak pebble fabrics in diamicton of similar origin at other sites; S1 significance values for the eigenvector of maximum clustering generally were <0.59. Where pebble orientation was stronger with S1 values about 0.7, the mean azimuth and plunge of the fabric data were in the direction of dip of the diamicton beds. Boulton (1971a) has reported relatively strong pebble fabrics from some modern glaciogenic debris flows. The stronger fabrics in Iowa likely represent more fluid water-sediment mixtures and greater shear within the flow, possibly a Type III Flow of Lawson (1982). Similar sediment sequences have been described in Denmark (Marcussen, 1975; Houmark-Nielsen, 1983). There thin diamicton beds are interbedded with stratified sorted sediment. The diamicton layers are discontinuous, generally <0.5 m thick, friable, of variable texture and some have stone-rich layers in their lower part. The latter characteristic indicates that the viscosity of the debris flows, as controlled by water content and texture, was sufficiently low that larger clasts settled within the flows and moved through traction in the basal part of the flow, as has been observed in flows on modern glaciers. These diamictons commonly have irregular contacts with the stratified deposits with both concordant and discordant relationships (Fig. 10.5). Slump structures and flow deformation structures are common. Pebbles in the diamicton generally lack a strong orientation and most fabrics are unrelated to local topography or ice-flow direction. Stratified deposits are more common in these sections and, with diamictons, form the uppermost facies association occurring as kames or hummocks with steep ice-contact slopes.

DGU 1972 MARCUSSEN

0

10

20 cm

Ganlöse

Stratified meltwater silt

Diamicton

Stratified meltwater sand

Stone

Stratified meltwater gravel

FIG. 10.5. Sketch of supraglacial sediment exposed in a ditch in Ganl¨ose, Denmark. Stratified silts and sands and diamicton with concordant and discordant contacts. Diamicton is friable and is texturally variable; notice soft-sediment and other deformation (reprinted with modifications from Marcussen, 1973, by permission of Scandinavian University Press).

Although the above types of sequences are typical of supraglacial deposits in low relief areas, sorted sediment also dominates some sequences and flow deposits are relatively minor. These sequences vary depending on whether fluvial, lacustrine or marine processes dominate. Ponded drainage was common near many Pleistocene ice margins as a result of crustal isostatic depression or end moraine damming (Teller, 1987). Such lakes locally extended onto the ice and, in some cases, many lakes existed supraglacially. Sediment deposited in these lakes is typical of ice-contact proglacial lakes with sorted gravels and sands near conduits to the lake and along the lake margin, and finer silt and clay in more distal positions. Interbedded with the sorted deposits are diamicton beds of varying character resulting from various subaqueous mass-wasting processes. Supraglacial lacustrine sediment sequences differ in that faulting and flowage of material was common as a result of collapse during subsequent melting of the adjacent

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

and subjacent ice. The resulting structures and deformed sediment, as well as irregular hummocky topography, serve to distinguish supraglacial lake sediment from proglacial lake deposits. 10.2.3. Moderate-relief Areas In areas of more relief, such as portions of the shield terrane in northern Europe and North America, as well as glaciated dissected plateaux, topography played a stronger role with respect to glacial dynamics and sedimentation. Near ice margins and particularly during deglaciation, stagnant ice became concentrated in valleys and basins and eventually separated into isolated ice blocks. In such situations, sedimentation was concentrated between stagnant ice and valley sides and around buried ice blocks, as well as subglacially. Kaszycki (1987, 1989) has developed models for sedimentation in such regions based on studies of Pleistocene deposits in the shield terrane of southern Ontario. Knobs and perched terraces along valley sides are characterized by sediment flow deposits of varying character interbedded with either glaciofluvial or glaciolacustrine deposits. These deposits also fill depressions on some surfaces or overlie till or bedrock. Kaszycki (1989) delineated three facies within these deposits (Fig. 10.6): 1 Facies I, rhythmically interbedded diamicton and silt (A in Fig. 10.6), consists of sandy diamicton beds about 10 cm thick that contain small pebbles (1–2 cm) and pinch out laterally over distances of 5 m or more. They are interbedded with thin beds of sand and silt that are inversely graded and have gradation contacts with the diamicton. The rhythmic sequence is interpreted to be the result of a series of subaqueous debris flows in shallow icemarginal or supraglacial ponds and lakes. Inverse grading likely resulted from dispersive pressures in a zone of basal shear within the flow and the diamicton beds represent either the plug zone of the flows or possibly cohesive flow with laminar shear throughout. 2 Facies II (B in Fig. 10.6), with extreme sediment variability, consists of chaotically interbedded sand, gravel and diamicton. Diamicton beds range in thickness from 5 to 20 cm, are matrix dominated,

323

and contain large clasts in the cobble to boulder size. Although diamicton is not graded, cobbles locally mark the lower contact and inclusions (clasts) of sorted sediment commonly occur near the base of diamicton where it overlies similar deposits. Pebbles have a weak to random orientation within the diamicton. Sorted sediment beds, up to 8 cm thick, commonly grade upward from granules and coarse sand to fine silty sand. Their lower contact with diamicton is abrupt and erosional in character; internal bedding generally is deformed and chaotic with small-scale normal and reverse faulting and folding. Sand dykes, about 6 mm thick, extend upward from the sand beds into diamicton for several metres. These sediments are interpreted to be the result of deposition in icecontact debris-fan environments with an abundant supply of meltwater. The diamicton beds were deposited from debris flows that likely were erosional, somewhat channelized, relatively thin and turbulent in their lower portions, possibly similar to Lawson’s (1982) Type II Flows. The interbedded granules and sand with normal grading are interpreted to represent either sandy high density turbulent flows or channelized sheet flow deposits. Folds and faults within the sand beds suggest shear deformation from suprajacent and adjacent sediment flows, and the dykes and flame structure indicate rapid loading of water-saturated sediment and the development of high pore water pressures and fluidization. 3 Facies III consists mainly of graded diamicton beds with occasional thin beds of sorted sediment (C in Fig. 10.6). Sediment variability in this facies primarily is the result of the varying concentration of large clasts in diamicton. Kaszycki (1989) recognized three textural classes of diamicton: cobbly, pebbly and sandy, with decreasing size and concentration of large clasts in the classes. Normal and inverse grading occurs in the cobble and boulder fractions of the coarsest diamicton. Diamicton beds range from 0.2 to 1.0 m thick. Inclusions of sorted sediment are uncommon, and the thin sorted sediment beds exhibit soft sediment deformation structures. Clast fabrics are variable. This facies is regarded as the result of relatively low-watercontent debris flows, similar to Lawson’s Type I and

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

5

10

50

10

20

100

15

30

150

40

s fsd msd csd gran gr

20

gran

msd

csd

s

cm 0

fsd

msd

csd gran gr

s

cm 0

fsd

msd

cm 0

csd gran gr

C Facies C3 Graded Diamicton

s

B Facies C2 Chaotically Interbedded Sand, Gravel & Diamicton

fsd

A Facies C1 Rhythmically Interbedded Diamicton & Silt

gr

324

silt fine sand medium sand coarse sand granule gravel sandy diamicton

25

50

pebbly diamicton cobbly diamicton

FIG. 10.6. Typical sediment logs of three facies within supraglacial and ice-contact deposits in the Haliburton region, south-central Ontario, Canada (reprinted from Kaszycki, 1989, by permission of the author).

II Flows. Ungraded diamicton possibly reflects deposition of a non-deforming, rigid plug. This facies represents deposition in a variety of icecontact supraglacial and proglacial situations. Similar supraglacial diamictons and sorted deposits have been described from exposures the Petteril Valley, Cumbria, UK (Huddart, 1983). 10.2.4. High-relief Mountain Regions In mountainous regions, topographic relief played a strong role in the character, nature and preservation

of the Pleistocene depositional record. Drainage was concentrated within valleys, often between valley sides and the lateral ice margin, as well as downstream from ice-margins. Proglacial lakes were common as a result of glacial dams across tributary valleys and in major valleys when drainage was toward the ice margin. Many such lakes extended onto the glacier and thus were in part supraglacial. Lateral and medial zones of concentrated debris within valley glaciers, basal debris, recently deposited drift and steep valley sides provided abundant sediment for reworking in the supraglacial,

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

ice-marginal and proglacial environments. Sedimentological studies of Pleistocene supraglacial and icemarginal deposits indicate that, except for commonly being coarser, the deposits generally are similar to those described previously in areas of continental glaciation. Mass movement deposits of varying origin are particularly common (Levson and Rutter, 1988). 10.3. LANDFORMS Supraglacial and ice-marginal landscapes include a varying and complex assemblage of landforms that reflect the interplay of local and regional topographic settings, geology, glacier dynamics, and local and regional climatic controls. These complexities are well illustrated by the differing and strongly contrasting landform assemblages that occur along the southern marginal areas of the Laurentide Ice Sheet in the USA (Mickelson et al., 1983; Mickelson and Attig, 1999). 10.3.1. Ice-marginal Glacial Forms 10.3.1.1. End moraines End moraines are ridges of glacial sediment that accumulate along and at the margins of glaciers. The term ‘moraine’ was first used by peasants and farmers who observed mounds and accumulations of rocky debris around glaciers in the French Alps. It has since come to be widely used for referring to and describing many types of constructional glacial morphology, and commonly is used with a preceding adjective. When used alone, end moraine is implied and an icemarginal origin inferred. The term originally was used for both the morphology and the sediment making up the landform. The practice of using the same term for both morphology and material is confusing, and ‘moraine’ should be restricted to morphology. Pleistocene end moraines are commonly given a geographic name for some locality on the moraine, for example, the Johnstown moraine of the Green Bay lobe in Wisconsin, or the Guelph-Paris moraine of the Lake Ontario Lobe in southern Ontario, or the Kirkham moraine of Scottish Readvance Ice in Cumbria, UK. They commonly are complex land-

325

forms and may partially reflect bedrock topography, earlier glacial events or multiple episodes of deposition. Those composed of deposits from two or more glacial advances are called superposed end moraines; those that occur over bedrock highs are referred to as rock-cored end moraines; and those that are buried by a relatively thin drift cover of a younger glacial event are known as palimpsest end moraines. Wide tracks of end morainic topography, usually with several distinct ridges, are referred to as a morainic system, and morainic topography that develops between adjacent glacial lobes is referred to as interlobate moraine (Russell et al., 1998). In mountainous regions, the term lateral moraine is used for the ridge or hummocky topography that result from the accumulation of debris along the lateral glacier margin, and medial moraine for ridges and topographic forms resulting from the deposition of linear debris-rich concentrations in ice, which result when two valley glaciers join and their debris-rich lateral margins merge together. Pleistocene end moraines are widely used to interpret glacial history, particularly the fluctuations ˘ of the ice margin during recent glaciations (Sibrava et al., 1986; Ehlers et al., 1991; Ehlers, 1996). Correct interpretation is essential if a viable glacial history is to be reconstructed. Their formation generally is considered to mark a time when an ice mass was active, the ice margin was relatively stable, fluctuating within a narrow range, and the former ice mass balance was more or less in steady state (i.e., they represent stillstands of the ice margin). As discussed below, this is not true for all end moraines and some likely formed in short time intervals. Those end moraines marking the farthest advance, as well as major readvances during a glacial event, are sometimes classified as terminal moraines, and those formed during deglaciation with limited or no readvance as recessional moraines. End moraines vary in both morphology and sediment content. Those related to ice sheets and ice caps commonly are arcuate in outline, and rise anywhere from 5 to 50 m or more above the surrounding topography. Some are short or composed of discontinuous ridges, whereas others can be traced for hundreds of kilometres with only breaks at former meltwater valleys. Some are narrow, only a kilometre

326

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

LAKE

Green Bay Lobe

MICH

Driftless Area

IGAN

WIS ILL ILLINOIAN

ke La

PLATE 10.2. Hummocky topography and kettle in the St Croix moraine, western Wisconsin, USA.

n iga ich

M be Lo

or two wide, others form belts >10 km wide. Local relief within end moraines is just as variable; some have rounded, relatively smooth slopes with relief related primarily to a developed drainage net. Others are highly irregular with many hills or hummocks and depressions or kettles (Plate 10.2).

ILLINOIAN

10.3.1.2. End and ground moraine composed primarily of diamicton and sorted sediment

N

50 km

IL L IN D

0

Materials that compose end moraines are highly variable. Many end moraines are composed of diamicton and sorted sediment that originated in the supraglacial environment. Although some subglacial till may occur, debris flow deposits and sorted fluvial and lacustrine sediments are more common. Beds usually are relatively thin, discontinuous and show deformational structures as a result of collapse, ice push and glacier overriding. These deposits, as is the topography, are the result of ice-contact resedimentation processes concurrent with downwasting and backwasting of the ice. Other end moraines are dominated by diamicton interpreted as mostly subglacial till, although some sediments are likely of supraglacial melt-out origin. For example, Lundqvist et al. (1993) report that the Johnstown moraine on the western side of the Green Bay Lobe in Wisconsin (Fig. 10.7) is composed of several facies, but that the dominant ones are diamicton of varying character interpreted to be till of melt-out and lodgement origin. The only supraglacial deposits are a thin veneer (
n-Erie Huroobe L

FIG. 10.7. Late Wisconsin end moraines (dark) and drift plains (ruled pattern) of the Green Bay lobe (in Wisconsin), Lake Michigan lobe and Huron–Erie lobe (in Indiana). Rectangle area shown in Plate 10.3 (after Frye and Willman, 1973).

glaciation to the south in Illinois (Fig. 10.7; Plate 10.3) appear to be composed primarily of uniform till with varying and generally minor amounts of sorted sediment and debris flow deposits. Wickham et al. (1988), for example, reported over 90 m of diamicton in Marengo moraine in northern Illinois, and thicknesses of 20–60 m are common in the Bloomington and Shelbyville morainic systems of Illinois. These moraines generally lack well-developed hummocky topography, and typical supraglacial and ice-marginal deposits are relatively thin. These Pleistocene moraines contrast strongly with many modern end moraines where till is rare in the marginal zone, for example at the Matanuska Glacier, Alaska where only about 5 per cent of the marginal deposits are primary tills.

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

327

In many lowland areas, such as the southern Great Lakes basin of North America, northern Germany and southern Sweden, Pleistocene end moraines are separated by lowlands with less local relief, known as ground moraine or drift (till) plains. These undulating areas usually have relief less than 10 m, and often less than 5 m (Chapter 8). Although exceptions occur, the supraglacial sediment association is thin and often discontinuous. Subglacial deposits commonly occur at or near the surface and subglacial bedforms often are conspicuous and may dominate the landscape. Locally, minor transverse moraines may occur and, if of ice-marginal origin, represent short-lived icemarginal positions during deglaciation. Some intermoraine areas were the sites of proglacial lakes, as drainage was dammed behind end moraines and the retreating ice margin. Often the supraglacial sediments are masked by lacustrine deposits, and subsequently mapped as glacial lake plains, rather than areas of ground moraine. 10.3.1.3. End moraines primarily of glaciotectonic origin

PLATE 10.3. Satellite view of end and ground moraines in eastcentral Illinois (see rectangle area in FIG. 10.7; ChampaignUrbana, Illinois located in lower centre part of scene). Note arcuate end moraines (light tone), re-entrant between two glacial sublobes in central and west-central part of area; relict braided pattern on fluvial surface, west-central part of area; parabolic sand dunes on glacial lake plain in northeast part of area, and local small-scale fluting and disintegration features.

Other moraines, particularly around the margin of the Lake Michigan basin in northeastern Illinois (Fig. 10.7), contain more variable sediment, and debris flow deposits are common (Hansel and Johnson, 1987). In these areas, the regional slope was toward the ice margin, and more basal debris became concentrated in a supraglacial position where resedimentation processes were active. Successive ice-marginal positions often more or less coincided, as previous moraines served as obstacles to glacial movement, resulting in a morainic complex of the superposed type.

Glaciotectonic deformation near ice margins commonly results in ridges known as push moraines (Chapter 14). They are composed primarily of large thrust masses of sediment and bedrock, often imbricated, that have been derived from beneath the glacier. 10.3.1.4. End moraines composed primarily of stratified drift Although the term moraine commonly implies diamicton, many end moraines are composed primarily of stratified and sorted deposits. Such ridge moraines exist in a terrestrial setting, where they are the proximal parts of outwash or sandur fans and plains, or in a lacustrine or marine setting, where they commonly are the result of ice-contact deltaic or subaqueous fan sedimentation. Diamicton occurs within these moraines, most commonly in the proximal zone, and most is of subaerial or subaqueous sediment gravity flow origin. Some of the best known end moraines of this type occur in Canada and Scandinavia where the retreating

328

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

(a)

(b)

(c)

FIG. 10.8. Schematic diagrams suggesting the origin of different portions of the Salpausselk¨a I moraine in Finland (reprinted from Fyfe, 1990, by permission of Scandinavian University Press).

Late Pleistocene ice margin was in contact with water, either lake or sea. The Late Pleistocene Ra and younger moraines in Norway, the Levene, Skovde and Billingen moraines of Sweden and the Salpausselk¨a moraines of Finland are particularly well known (Fyfe, 1990; Lundqvist, 1990). The moraines vary in character laterally but most contain stratified gravels, sands and silts with some interbedded diamicton. Deltaic foreset beds are common, and the moraines often are wider and more deltaic in character near zones where meltwater discharge was discrete and eskers occur in the up-ice direction. The morphology and sediment character of the moraines is controlled in part by water depth and the nature of the subglacial drainage system, as illustrated by the Salpausselk¨a I moraine in Finland (Fyfe, 1990). Where the glacier

margin was slightly above water level, Gilbert-type deltas formed with topset and foreset beds in the distal part and diamicton and coarse gravels in the proximal portion of the moraine (Fig. 10.8(a)). The moraine complex is commonly several kilometres wide with about 35 m of relief. The proximal slope is steep and of ice-contact origin with kettle holes common; the distal slope is less steep and has distinct upper fan and delta front surfaces. These fan delta plateaux commonly connect with eskers and were the result of sedimentation from a distinct subglacial conduit drainage system. Closely spaced conduits resulted in overlapping fan deltas and a continuous end morainic ridge. Where the ice margin was grounded at a relatively shallow water depth, sedimentation resulted in a moraine that is narrower, has more relief, and both proximal and distal slopes are steep (Fig. 10.8(b)). Under these conditions the moraine consists primarily of foreset deposits. In areas where the water depth was greater, in the order of 60 m or so, sedimentation continued to be restricted and a narrow, but lower, ridge resulted. In even deeper water, the ice-marginal ridge loses much of its morphology and is several kilometres wide with about 10 m of relief (Fig. 10.8(c)). These latter subaqueous moraines generally are not connected with eskers in the up-ice direction. Fyfe (1990) interprets the changing morphology to be a reflection of the nature of sediment delivery to the ice-margin. As water depth increased, the basal shear stress of the ice sheet near the margin was reduced, which led to a lowering of the ice sheet profile and a destabilization of large conduit subglacial drainage. Similar moraines have been described from Canada and the USA. Sharpe and Cowan (1990) report long, arcuate moraines about 10–40 m high and 0.5–3 km wide in Ontario. The moraines have a steep proximal slope and a more gentle distal slope, and are composed primarily of stratified sand and gravel, with large boulders on the surface. Similar ridges and forms in Maine, called ‘stratified end moraines’ by Borns (1973), primarily are the result of marine fan sedimentation (Ashley et al., 1991). Within bodies of water where an ice mass has temporally grounded, moraine ridges variously termed De Geer, crossvalley, ribbed or washboard moraines can form (Chapter 8).

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

329

PLATE 10.4. Lateral and end moraines of Bloody and Sawmill Canyons, Mono Basin area, California, USA. Moraines represent several major glaciations; Walker Lake dammed behind youngest moraine (photograph by Peter Birkeland).

Other moraines are similar to these, except instead of forming in a water body, they are land-based and composed of fluvial sands and gravels. In essence, these moraines represent ‘heads of outwash’ where closely spaced drainageways from the glacier coalesced to form more or less continuous ridges with a steep ice-contact proximal side and a more gentle distal side. Kettles, near the former ice margin, give the ridge an irregular and distinctly morainic surface morphology (Blewett and Rieck, 1987).

tributary glaciers. Because drainage and fluvial activity was concentrated in valley bottoms, end and medial moraines commonly were not preserved as part of the geological record. Nonetheless, there are many mountain valleys where late Pleistocene forms are extremely well preserved, and the former position of the ice margin easily can be reconstructed (Plate 10.4).

10.3.1.5. Lateral, end and medial moraines of mountain glaciation

10.3.2.1. Ice-marginal meltwater channels and systems

Most moraines are ice-marginal in origin and consist of end, lateral moraines and medial moraines. The latter are the longitudinal ridges along valley bottoms that formed down-ice from the junction of two

Along the margins of Pleistocene ice masses, considerable volumes of meltwater, especially during the ablation season, carved channels of various sizes and/ or deposited an assortment of stratified sediment

10.3.2. Ice-marginal Fluvial Forms

330

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

(Chapter 8). Such channels were most common in areas of moderate to high relief where the drainage was confined between the glacier and valleys sides or in situations where the glacier flowed up a regional slope and drainage again was confined between the ice and land areas. Many of the marginal channels reflect a compromise between the slope of the topography and the ice margin itself, thus in many cases channels appear to disregard the normal slope gradient cutting instead across hillslopes at low angles. Such channels or erosional benches vary in character, both in cross-section and longitudinal profile, depending on whether drainage was completely on land, on or in ice, or some combination of both.

PLATE 10.5. Kame terrace surface with kettle to right, White River valley, Vermont, USA. Sand and gravel locally is mined from this landform.

10.3.2.2. Kame terraces and kame deposits Where glaciofluvial deposition was the dominant process along former ice margins, various landforms of stratified drift developed. Along valley sides, and in some cases between separate but adjacent ice lobes, large volumes of glaciofluvial sediment were deposited. The former situation was particularly common and kame terraces, are found in many glacial valleys (Plate 10.5). They commonly have an irregular upper surface as the result of kettle development, and a relatively steep ice-contact side facing the central valley. Their gradient usually reflects the ice-margin slope in the down-ice direction (Gray, 1991). Kame terraces formed at the same time but on opposite sides of a valley commonly are not at the same elevation because of variations in microclimate related to the aspect of the valley. This relationship, as well as their morphology and spatial relationship to other glacial features, serves to help distinguish them from alluvial terraces composed of proglacial fluvial sediment. In many valleys, as the ice downwasted and the margin retreated, a sequence of kame terraces formed along the valley sides (Sissons, 1958). These terraces are composed of a variety of stratified and sorted sediment with a wide range of particle sizes, as well as masses of diamicton of sediment flow origin. The surface of the terraces are often pitted and contain kettle holes where abandoned ice blocks were left to slowly ablate while sedimentation occurred around them. Also the terrace surface may have small

meltwater marginal channels that may dissect the terraces into a large number of accordant segments. In many instances kame terraces merge down-stream into proglacial deltaic terraces, valley trains, or sandur or lacustrine plains (Fig. 10.9). Kames are roughly circular mounds that are composed of glaciofluvial sediments. They form in contact with ice and usually occur in association with other forms of glaciofluvial deposition. Kames originate in a variety of ways, sometimes from sediment being deposited in cavities within or below ice, or in supraglacial depressions, with the resultant mound forming as the sediment is let down onto the ground surface. Conical and circular kames often form below moulins, as surface drainage, upon reaching the ground surface, spreads laterally and deposits its load in a cone-shaped deposit. As in the case of kame terraces, sediments are of a wide range in particle size and sorting and, owing to the removal of ice walls, sediment in kames commonly exhibits faulting and slumping. 10.3.3. Supraglacial Forms and Landscapes 10.3.3.1. Disintegration features – Hummocky moraine Large areas glaciated during the Pleistocene are characterized by distinctive topography, which

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

(a)

331

VA L L E Y S I D E

GLACIER B R A I D E D O U T WA S H T R A I N

Subglacial river

D E LTA

SEA

VA L L E Y S I D E

(b) KAME TERRACE

ICE-

CO

NT A

OUTWASH TERRACE

GLACIOMARINE DELTA

CT SL O

SEA

PE

OUTWASH TERRACE

DELTA

ASH UTW FAN

KAME TERRACE

O

(c) KAME

TERR

ACES

KAME, KETTLES, ESKERS

KAME

ESKER

KETTLE HOLE

PALAEOCHANNEL

OUTWASH & KAME TERRACES

KETTLED OUTWASH FAN ICECONTACT SLOPE

ICE-CONTACT

OUT

WAS H

TRAIN

GLACIOMARINE DELTA

PROGLACIAL

FIG. 10.9. Schematic maps (a) during glaciation; (b) after glaciation, and (c) longitudinal profile of (b); showing inferred origin of glaciofluvial morphosequence composed of ice-contact and proglacial deposits (reprinted from Gray, 1991, in Ehlers, J., Gibbard, P. L. and Rose, J. (eds) 1991, courtesy of A.A. Balkema, Rotterdam).

resulted from stagnation of the marginal zone and downwasting of the glacier (Hodgson, 1982; Eyles, 1983b; Attig and Clayton, 1993). Areas of stagnant ice developed either after an extensive advance, possibly a surge (Sharpe 1988b), or incrementally as the marginal zone became insulated by supraglacial debris, eventually stagnated, and the region of stagnation grew progressively larger in the up-ice direction. The latter situation locally was common

and eventually large regions were covered by debriscovered dead ice. Stagnant ice landforms are particularly well developed where the glacier flowed up a regional slope and large quantities of basal debris moved toward the surface as a result of intense compressive flow. The landscape is irregular and chaotic in appearance, with many small hills and depressions with steep to gentle slopes. Most hills reach about the same elevation and lack any

332

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

PLATE 10.6. Hummocky ground moraine resulting from downwasting of stagnant ice concurrent with reworking and collapse of supraglacial sediment. Mountrail County, North Dakota, USA. (Photograph by Lee Clayton.)

consistent trend (Plate 10.6). These areas generally lack end moraines, and the areas could be considered as ‘high relief’ ground moraine. Disintegration features are considered controlled if the landforms maintain some regular and repetitive pattern that resulted from ice control, such as a crevasse pattern or differential debris distribution in the ice as a result of shearing or folding. If the forms A

ice

E

ice B

F

C

G

PLATE 10.7. Aerial photograph of hummocky moraine with small circular disintegration ridges and larger perched lake plains with raised rims. The larger ice-walled lake extended onto buried ice to the lower left and lake sediment subsequently collapsed resulting in hummocky topography. Mountrail County, North Dakota, USA. (US Department of Agriculture Photograph.)

ice

ice

ice D

H

ice

FIG. 10.10. Schematic diagrams showing the formation of hummocks and circular disintegration ridges during the redistribution of supraglacial sediment as stagnant ice downwastes. A, B, C, D, inversions of topography as a result of non-uniformity of debris and variable melt rates; E, F, formation of hummocks; G, H, formation of circulation disintegration ridges (from Clayton and Moran in Coates (ed), 1974, by permission of the State University of New York, Binghamton, NY).

are randomly distributed on the landscape, they are described as uncontrolled. The overall landscape is best described as hummocky moraine. Genetic terminology commonly refers to such areas as stagnant ice or dead-ice moraine, or collapse moraine in order to emphasize the process. All hummocky topography is not of supraglacial origin and can form as a result of subglacial processes (Attig and Clayton,1993; Mickelson and Attig, 1999). A supraglacial landscape usually reflects the final stages of ice wastage, which commonly is the last of a sequence of topographic inversions as sediment moves from high areas on the glacier to low areas,

SUPRAGLACIAL AND ICE-MARGINAL DEPOSITS AND LANDFORMS

333

Ice-Contact Rim Hummocky Ground Moraine Stagnant Ice Origin

150 ft

Perched Lake Plain

ld Bou

er

g La

Supraglacial Diamicton

Dirty Boulder Gravel

Clean Sand and Gravel

Silt and Clay

Glacial Lake Clay

FIG. 10.11. Schematic cross-section showing sediment relationships among rim, perched lake plain and hummocky ground moraine (reprinted from Parizek, 1959, permission of the author).

only to be reworked later as a result of differential melt rates owing to differential insulation of ice by supraglacial debris. Thus, high areas of the landscape today reflect low areas immediately following glacier wastage, and vice versa. Two common forms are circular hummocks and disintegration ridges or ‘doughnuts’, the latter being hummocks with a central depression (Fig. 10.10). Sediment that accumulates in the depression is let down to form the hummock as the surrounding ice melts. If the hummock, late in its formation, has an ice core, a depression forms in the central area resulting in a circular disintegration ridge. Many of the hills, ridges and depressions in stagnant-ice moraine are irregular in shape, and reflect the complex and changing supraglacial topography during ice surface downwasting. 10.3.3.2. Raised plateaus and perched lake plains Many hummocky moraine areas contain small to large irregularly shaped plains or plateaus with rims usually sited higher than the tops of the hummocks (Plate 10.7). They are composed of a variety of sediment

types; often consisting of interbedded colluvial, fluvial and lacustrine deposits. Deposition possibly took place on firm ground, not on ice, but the depositional basin may have had walls of stagnant ice (Markgren and Lassila, 1980). Such raised features in North America have been called moraine plateaus, ice-walled lake plains, and moraine-lake plateaus (Fig. 10.11). These forms are similar to Blattnick and Veikki moraines (Chapter 8). 10.4. SUMMARY Supraglacial and ice-marginal deposits and landforms are among the more complex, and distinctive of the glacial system. In ideal situations they have been used to reconstruct the profiles and thickness of the past glaciers. Many of the sorted deposits are an important resource as an aggregate material and, where buried, locally serve as major sources of groundwater (Melhorn and Kempton, 1991). The varied lithologic and textural character of supraglacial deposits and their discontinuous nature and varied morphology make them an important consideration in land-use and construction site evaluations.

This Page Intentionally Left Blank

11

GLACIOLACUSTRINE ENVIRONMENTS G. M. Ashley

(Tables 11.2 and 11.3). Time series measurements of spatial variation in water temperature, suspended sediment concentration and current velocities reveal insights into limnological processes. Sediment trap arrays and isotopic dating of bottom sediments (cores) yield data on sediment dispersal processes and sedimentation rates (Weirich, 1985; Leonard, 1986a,b). These technological approaches and modern process studies have led to descriptive models that can now be used as a guide to interpret the environments of deposition of ancient sedimentary sequences.

11.1. INTRODUCTION Meltwater impounded by ice or glacial deposits, or in depressions created by glacial erosion, or by isostatic loading creates a wide variety of glacier-fed lake types in terms of size, longevity and limnological characteristics. All are fed by glacial drainage and are sediment ‘sinks’. Thus, typically, their deposits record important aspects of the deglacial history of an area. The variations of sedimentation rate and organic content through time provide proxy information on climate. Interpretation of oxygen isotope records in lakes have been attempted recently in order to correlate land records with the well-established marine isotope record of climate change (Lister, 1988)(Chapter 2). Glacier-fed water bodies are classified according to their position relative to the ice, the source of water and sediment, as either ice-contact lakes or distal lakes (physically separate from the ice) (Table 11.1, Fig. 11.1). Magnetic measurements on mineralogy have been used for provenance and sediment flux studies (Bradshaw and Thompson, 1985). Also highresolution seismic surveys have revealed the threedimensional geometry of an entire lake deposit (Mullins et al., 1990). Although logistically difficult, studies of contemporary glaciolacustrine environments provide an excellent opportunity to link physical processes directly to the resulting deposits

TABLE 11.1. Classification of glacier-fed lakes (modified from Ashley, 1988) Proximal, ice-contact lakes Subglacial and englacial lakes Supraglacial lakes (a) Active-ice lakes (b) Ice-stagnation kettle lakes (c) Ice-stagnation lake network Ice-marginal lakes (a) Dammed by ice (1) River-lakes (2) Semi-permanent lakes (b) Dammed by topography (1) River-lakes (2) Semi-permanent lakes Distal lakes

335

336

GLACIOLACUSTRINE ENVIRONMENTS

ICE-CONTACT LAKES (a) DAMMED BY ICE

(d) ICE-STAGNATION KETTLE NETWORK

IC

E

ICE

ICE

LAKE

(e)

ICE-STAGNATION LAKE NETWORK

(b) DAMMED BY TOPOGRAPHY ICE ICE LAKE

ICE

ICE

DISTAL LAKE

LAKE ICE

(c)

RIVER LAKE

(f)

DISTAL LAKES

ICE

LAKE ICE LAKE

LAKE

FIG. 11.1. Diagrammatic sketches of the various types of glacier-fed lakes described in Table 11.1.

11.1.1. Ice-contact Lakes An ice-contact lake is a ponded body of meltwater with at least a portion of the lake in direct contact with either active or stagnating glacial ice (Plate 11.1(a),(b)). The lakes may be dammed by the ice such that a change in the glacier causes complete drainage, or the lakes may be simply situated adjacent

to ice. These water bodies are either subglacial, englacial, supraglacial or ice marginal (Fig. 11.1). Subglacial lakes exist below the East Antarctic Ice Sheet and in the past below Pleistocene and earlier ice sheets probably produce episodic catastrophic discharges (j¨okulhlaups) (Chapters 8 and 9). There have been few documented examples of englacial lakes and although they may produce j¨okulhlaups they are

GLACIOLACUSTRINE ENVIRONMENTS

unlikely to leave a recognizable sediment record. Supraglacial lakes develop as ablation occurs and are generally found only in the portions of active glaciers that are under compression; meltwater normally moves downward under the influence of gravity through fractures and solution conduits. Supraglacial meltwater creates isolated depressions on inactive portions of glaciers (ice-stagnation kettle lakes) and complex interconnected lakes (ice-stagnation lake network) (Plate 11.1(c),(d)). Supraglacial lakes are relatively short lived and change configuration continuously as the ice melts. Little is known about the physical limnology of these lakes, but they are likely to be shallow and thus well mixed by wind during the summer when the lake surface is generally ice free. TABLE 11.2. Factors affecting sedimentation Lake type (proximity to ice) Lake stratification (thermal or sediment stratification; no stratification) Relative densities of lake and meltwater Position of meltwater inflow into lake water column Sediment sources (number and direction) Lake basin geometry Slope stability Ice rafting Extent of ice cover

337

Ice-marginal lakes may be dammed by ice if the local slope is toward the ice or topographically if the regional slope is away from the ice, but drainage is impeded locally by glacial deposits. Lakes dammed only by ice are likely to be ephemeral (a year to a few tens of years) and some may drain and refill more than once during a melt season (Gilbert and Desloges, 1987). Ice-marginal lakes that occupy pre-existing natural basins, dammed river valleys or isostatic depressions tend to be semi-permanent (hundreds to a few thousand years) and enlarge as the glacier recedes. A river-lake, a type of ice-marginal lake, commonly develops in poorly drained regions peripheral to the ice front where meltwater discharge moves slowly but continuously away from the ice. This body of impounded meltwater may be thought of as a lowvelocity river or a high-velocity lake. These shallow, continuously moving (albeit slowly) water bodies commonly cover large areas fringing the glacier terminus. 11.1.2. Distal Lakes Distal lakes are physically separate from the ice but fed primarily by glacial meltwater via outwash streams (Fig. 11.1). Distal lakes may occupy preexisting natural basins and thus have the potential to

TABLE 11.3. Sedimentary process and lithofacies produced in glacier-fed lakes (modified from Ashley, 1988) Sedimentary processes Mass movement Creep, slump, debris flow Surge current Currents Quasi-continuous Wind-driven Nearshore waves, lake circulation Combined

Lithofacies units

Matrix supported and clast supported massive to stratified diamicts Crossbedding, load structures, fluidization structures, massive beds, parallel and graded bedding, climbing ripples, draped lamination Parallel bedding, graded bedding, large-scale tabular bedding, climbing ripples, multiple graded bed, load structures, fluidization structures Small-scale crossbedding, parallel lamination parallel bedding, multiple graded beds Rhythmites

Ice rafting

Debris blebs, dropstones

Biogenic processes

Body or trace fossils, bioturbated sediments

Sedimentation From suspension

Massive to parallel laminated silts and clays

338

GLACIOLACUSTRINE ENVIRONMENTS

(a)

(b) PLATE 11.1. Ice-contact lake environment. (a) Ice-contact lake with ice bergs (Bering Glacier, Alaska). (b) Debris flows into an ice-contact lake (Matanuska Glacier, Alaska).

GLACIOLACUSTRINE ENVIRONMENTS

339

(c)

(d) PLATE 11.1. (c) Ice-stagnation kettle lakes (Malaspina Glacier, Alaska) (photo by T. C. Gustavson). (d) Ice-stagnation lake network (Malaspina Glacier, Alaska).

340

GLACIOLACUSTRINE ENVIRONMENTS

last for hundreds to thousands of years. In contrast, distal lakes impounded by sediment dams are likely to be shorter-lived (tens to hundreds of years). As the glacier recedes from the region, the proportion of meteoric water and groundwater seepage with respect to the volume of meltwater inflow increases in distal lakes. Meltwater effects typically include strong seasonal and weather-dependent discharge variations, low salinity and temperature and high sediment loads.

(a)

0

5

10

15

20

11.2.1. Physical Limnology Water has a high heat capacity and thus lakes are able to absorb or give off heat in response to changes in solar radiation input with little net change in temperature. Meltwater is near 0°C as flow exits the ice and enters proximal, ice-contact lakes. Streams that flow across outwash plains before entering the lake

(b)

Temperature °C 1.000

11.2. LIMNOLOGY OF GLACIER-FED LAKES

25

epilimnion

Density

0.999

metalimnion 0.998

hypolimnion

0.997

Temperature

0.996

(c)

Surface

Temperature °C¢ 0

e

5

d

10

c

15

b

a

Depth

ice

Bottom

FIG. 11.2. Physical limnology. (a) Density of fresh water as a function of temperature. (b) Thermally stratified water column. (c) Seasonal variation in hypothetical thermal profiles ranging from mid-summer, a, to winter, e.

GLACIOLACUSTRINE ENVIRONMENTS

warm slightly and may reach 10–12°C (Irwin and Pickrill, 1982). Meltwater contains varying concentrations of suspended sediment and dissolved load entrained during its passage through and under the ice. Lakes are influenced by these short-term thermal and sediment load fluctuations and generally develop some type of density stratification related to variation in temperature, chemistry or sediment concentration with depth. 11.2.1.1. Thermal stratification Thermal stratification is produced by solar warming. Water density is a function of temperature; however, the relationship is non-linear with the maximum density (~1000 kg m–3 ) at 4°C and lower densities at temperatures both colder and warmer than 4°C (Fig. 11.2(a)). On a seasonal time-scale, solar radiation absorbed near the lake surface warms the water from top down, creating a lower-density layer on the top

(epilimnion) overlying a cold, higher density layer at the bottom (hypolimnion). Daily fluctuations in solar radiation coupled with turbulent mixing by wind enhance the density contrast between the surface water and bottom so that the layers are separated by a region of rapid temperature change called the metalimnion. The. thermocline is contained within the metalimnion and is located where the change of temperature with depth is the greatest (Figs 11.2(b) and 11.3(a)). Increased surface warming creates evergreater density differences between the epilimnion and hypolimnion. In contrast, any wind mixing of warm epilimnial water with the water of the metalimnion tends to reduce the differences in density. On a yearly basis, the thermal profiles vary from midsummer to winter (Fig. 11.2(c)). Complete mixing of the lake waters occurs when the lake is isothermal (d in Fig. 11.2(c)), which is likely in spring and autumn. Water bodies that turnover (completely mix) twice a year are termed dimictic and are common among

THERMAL STRATIFICATION

(a)

341

wind

relative density

OVERFLOW INTERFLOW

depth

thermocline

UNDERFLOW

SEDIMENT STRATIFICATION ICE

wind

relative density

depth

(b)

ICE UNDERFLOW

ICE

(c)

NO STRATIFICATION

relative density

depth

wind

FIG. 11.3. Density stratification in glacier-fed lakes. (a) Thermal stratification, most common in distal lakes. (b) Sediment stratification, most common in ice-contact lakes. (c) No stratification, typical of river-lakes.

342

GLACIOLACUSTRINE ENVIRONMENTS

glacier-fed lakes, particularly distal lakes. Those that circulate once per year are monomictic. Hutchinson (1957) divided monomictic lakes into ‘warm’ and ‘cold’, the former implying winter circulation above 4°C, and latter with summer circulation below 4°C. Little limnological data exist for ice-contact lakes. They may be monomictic but most are probably polymictic, at least near the ice margin (Churski, 1973; Harris, 1976).

drained and refilled frequently with concentrations ranging from a maximum of 28.7 mg 1–1 prior to draining to 1248 mg 1–1 after draining (Gilbert and Desloges, 1987).

11.2.1.2. Sediment stratification

Dissolved materials are transported to lakes with mineral detritus but total volume is low. Chemical stratification is generally rare except in coastal icecontact lakes that have or have had a connection to the sea where dense saline water may be trapped in the bottom water (Harris, 1976). Authigenic carbonate and sodic minerals that typify other lakes (Allen and Collinson, 1986) are rare in glacial-fed lake deposits. Manganese nodules have been observed, as have concentrations of trace metals (Renberg and Segerstrom, 1981). Iron-rich laminations are associated with incomplete circulation (meromixis) and low sedimentation rates (Anthony, 1977). In general, thermal and sediment stratification dominate glacierfed systems and the importance of chemically precipitated sediments is often masked by detrital sediment. Little is known of the chemical processes involved in glacial-fed lakes.

Water density is also a function of suspended sediment concentration. Lakes being fed by meltwater streams with high sediment loads (>1 g l–1 ) may develop a density stratification that is a function of a gradual increase of suspended sediment concentration with depth. As the sediment settles it is being continuously replaced by new inflow during the melt season (Fig. 11.3(b)). The density profile owing to sediment is more gradual than the temperature profile (Fig. 11.3(a)). However, the magnitude of density owing to sediment may be significantly higher than temperature effects (Table 11.4). Gustavson (1975a,b) measured nearly 200 mg l–1 at the surface of Malaspina Lake (an ice-marginal lake) and 800 mg l–1 at mid-depth of the 80 m-deep lake. Ape Lake, British Columbia, an ice- marginal lake (dammed by ice)

11.2.2. Chemical and Biological Limnology 11.2.2.1. Chemical processes

TABLE 11.4. Maximum recorded concentration of suspended particulate matter in the waters of glacial lakes (modified from Gilbert and Desloges, 1987) Lake Tasikutaaq Lake, Baffin Island Hector Lake, Alberta Bow Lake, Alberta Lower Waterfowl Lake, Alberta Ape Lake, British Columbia Garibaldi Lake, British Columbia Lillooet Lake, British Columbia Lake Wakatipu, New Zealand Stewart Lakes, Baffin Islanda Cascade Lake, Washingtona Hazard Lake, Yukona Unnamed, Purcell Mountains, B.C. Sunwapta Lake, Alberta Malaspina Lake, Alaskaa a

Ice-contact lakes.

Concentration (Mg 1–1) 5.7 >8 15 17.2 28.7 30 49 57 109 110 122 140 380 723

Source Lemmen (1984) Smith et al. (1982) Smith et al. (1982) Smith et al. (1982) Gilbert and Desloges (1987a) Mathews (1956) Gilbert (1975) Pickrill and Irwin (1983) Gilbert et al. (1985) Campbell (1973) Liverman (1987) Weirich (1985) Gilbert and Shaw (1981) Gustavson (1975b)

GLACIOLACUSTRINE ENVIRONMENTS

11.2.2.2. Biological processes By their nature, glacier-fed lakes are inhospitable and thus biological productivity is relatively low. However, both carbonate and silicious micro-organisms occur in the water column. Fecal pellets produced by organisms aid in binding individual particles together into large flocs and thus enhance sedimentation (Smith and Syvitski, 1982). Trace fossils in the bottom sediment reveal that other forms of life such as worms, gastropods and some flying insects spend at least a portion of their life cycle in the lake (Gibbard and Stuart, 1974; Gibbard and Dreimanis, 1978). However, in general, biological processes, similar to chemical processes, are often masked by physical processes. 11.2.3. Sediment Transport and Deposition Several detailed studies of modern glacier-fed lakes have provided insight into the physical processes related to sediment dispersal that occur when a meltwater stream enters a lake (Pickrill and Irwin, 1982; Smith et al., 1982; Dominik et al., 1983; Weirich, 1984, 1985; Østrem and Olsen, 1987). Although the specifics of each scenario vary from lake to lake, the overall theme is the same. Short-term fluctuations (hours and days) in meltwater inflow are characteristic and the nature of the lake stratification, which may also vary with time, affects the dispersal of sediment in the lake. These temporal fluctuations in stream and lake properties control the various rhythms of sedimentation that are common to glaciolacustrine deposits (Table 11.2). As fluvially transported bedload and suspended sediment load enter a lake, the relative densities of lake and meltwater, as well as the nature of lake stratification, dictate the transport path (Fig. 11.3). If the density of inflowing water is less than the bottom water (hypolimnion) the stream moves as a plume at a level determined by its density, yielding overflows (hypopycnal) at the surface, interflows at any level below the surface, and underflows (hyperpycnal) along the bottom (Fig. 11.3(a)). In thermally stratified lakes, some of the fine suspended sediment is dispersed in the epilimnion by overflows or interflows (Smith et al., 1982). The rest

343

of the fine sediment is transported by high-density turbid flows that ‘dive’ and move as quasi-continuous underflows onto the lake bottom (Weirich, 1984). These density underflows are initially driven by the momentum of the inflowing stream, which may reach 0.15 m s–1 and later by gravity as flow continues down the slope of the lake basin. Gravel and coarse sand drop out of the flows early and accumulate in shallow water. Medium to fine sand and silt are deposited by the underflows as they continue into the lake. Fine silt and clay are kept in suspension until flow drops to <0.1 m s–1. In sediment-stratified lakes, glacial meltwater flows entering at the top, middle or bottom of the lake are affected by gravity and move downward and away from the ice until momentum is lost (Fig. 11.3(b)). Sedimentation is focused in topographic lows because sediment is transported at the base of the hypolimnion. Equal density mixing (homopycnal flow) occurs when the inflowing river water is of equal density to that of the lake and the lake is either unstratified or weakly stratified (Fig. 11.3(c)). Rather than following a specific level within the lake, the flow expands in three dimensions and the river and lake water gradually mix as the sediment moves by advection throughout the water column. Winds are generated almost continuously in the vicinity of glaciers. Cold air associated with the ice moves toward warmer areas producing ‘katabatic’ winds that blow across the surface of lakes setting up epilimnial currents that move slowly away from the ice. Water motion in lakes (including wind-driven flows) are affected by Coriolis forces of the Earth’s rotation such that flow is directed to the right in the northern hemisphere and to the left in the southern hemisphere. These currents play an important role in distributing sediment on lake bottoms (Smith and Ashley, 1985). The extent and longevity of ice cover that normally develops on the lake during the winter months affects the development of thermal stratification within the water column and inhibits wind stress on the lake surface. These factors affect the physical processes of sediment dispersal. Consequently, the amount of time that the lake is ice free each year is important (Table 11.2). Icebergs, pieces of a glacier that have broken off (calved), are free to float and drift about the lake under the influence of the wind and/or lake currents

344

GLACIOLACUSTRINE ENVIRONMENTS

(Plate 11.1(a)). As the density of ice is 0.9 g cm–1 then nine-tenths of a berg is below the surface. Icebergs commonly disturb (scour and erode) sediment in the shallow parts of the lake. Glacial debris is released from the bergs as they melt creating lens-shaped coarse deposits called ‘berg dumps’ or singular dropstones which are highly visible if found in otherwise fine grained glaciolacustrine deposits (Thomas and Connell, 1985). Icebergs commonly collect in shallow water, particularly at the lake outlet, and sedimentation is concentrated as debris is released from the ice during melting. Relatively large landforms may accumulate by this process (McManus and Duck, 1988). Mass movement is another key mechanism for moving sediment within the lake (Fulton and Pullen, 1969). Glacial debris deposited at the lake margin by fluvial or colluvial processes may become unstable (particularly on steep slopes) and slump, slide or flow into deeper water (Siegenthaler and Sturm, 1991). Slump terraces are common on steeper slopes reflecting rotational failures. Sliding may shift deposits to a less steep slope building compressional ridges at the base (Nemec, 1990). Mass movement processes commonly trigger short-lived (minutes) flows of turbid water, called ‘surge currents’ (Smith and Ashley, 1985), they are identical to what has been identified as ‘turbidity currents’ in the marine environment. However, these surges are short lived compared with the quasi-continuous density currents that are produced by meltwater inflow that may last for days, weeks or even months. Although both types of density currents are a function of high turbidity, they operate on different time scales and produce different types of deposits. 11.3. ICE-CONTACT GLACIER-FED LAKES The margins of ice-contact lakes continually change as ice masses advance or retreat and as ice melts and sediment is released. Most ice-contact lakes are polymictic and are sediment stratified rather than thermally stratified. Few generalizations can be made about the deposits of these lakes except that sedimentation patterns are highly variable in proximal areas. The high variability of sediment influx to the proximal zone is damped with distance from the ice

front and as a consequence the lake bottom subenvironment is typified by steady and relatively low sedimentation rates. The discussion, here, focuses on river-lakes and relatively large semi-permanent icemarginal lakes as these are the types most likely to leave a sedimentary record. 11.3.1. River-lakes River-lakes are short-lived bodies of water that result from partially impounded drainage near the ice front (Fig. 11.1). They are shallow (few metres) slowmoving lakes with a low residence time and thus lack stratification. Most are likely to be characterized by equal-density mixing (Fig. 11.3(c)). Sediments tend to be fine sands and silts because coarse clasts are left as lag and finer material is carried through the lake (e.g., Matanuska Glacier, Ashley, unpublished data; Smith, 1990). Deposits are thin tabular bodies composed of poorly stratified to massive sandy silts that are interbedded with a variety of ice-contact deposits. 11.3.2. Semi-permanent Lakes; Proximal Sub-environments The ice proximal environment can be divided into four distinct sub-environments: (a) ice-contact delta, (b) subaqueous fan, (c) ice-cliff margin, and (d) submerged ice ramp (Fig. 11.4). Each is characterized by high sedimentation rates and its own suite of depositional processes and deposits (Table 11.5). The ice-proximal deposits are commonly deformed by seasonal ice marginal fluctuations. The proximal depositional settings of lakes where water lies in direct contact with the ice can be further subdivided into (1) those dominated by meltwater discharge (deltas and subaqueous fans) and (2) those dominated by gravity-driven mass-movement processes (ice-cliff margins and submerged ice-ramps). 11.3.2.1. Ice-contact delta Deltas are landforms built into the lake from fluvially transported detritus disgorged at or near the lake surface. Deltas may build at a single point reflecting a single, stable meltwater stream as the sediment source or may form at several locations along the lake margin

GLACIOLACUSTRINE ENVIRONMENTS

ICE-CLIFF MARGIN

345

DELTA

ICE SUBMERGED

SUBAQUEOUS FAN

ICE-RAMP ABANDONED FAN LAKE BOTTOM

FIG. 11.4. Ice-contact lake depositional environments.

TABLE 11.5. Ice-contact glacier-fed lacustrine subenvironments and associated lithofacies (from Ashley, 1988) Subenvironments Delta Coarse-grained

Processes

Deposits

Mass-movement (creep, slump, debris flows, grain flows), surge currents

Well-sorted coarse sediments. Parallel bedded, steeply dipping foresets (20–33°), normal and inverse graded beds of sand and gravel. Clasts parallel to bedding planes. Well-sorted medium to fine sediments. Sets of climbing ripple sequences, draped lamination occurring as low angle <20° foresets of sand and silt. Deformational structures and dewatering structures common.

Fine-grained

Rapid deposition from density underflows

Subaqueous fan

Deposition from a high velocity expanding jet

Proximal to distal open work gravel, chaotic bedding, convolute lamination, dewatering structures, climbing ripple sequences and drape lamination.

Ice-cliff margin

Mass-movement, rock fall, debris slides, calving, squeezed debris at base

Poorly sorted, stacked diamicts interbedded with sand beds, mud, and dropstones.

Submerged ice-ramp

Mass-movement, debris flow, creep, slumps, surge currents

Poorly sorted, stacked diamicts interbedded with sand beds, mud drapes.

Lake bottom

Deposition from overflows interflows and underflows and occasional icebergs

Fine-grained rhythmic, parallel bedded; occasional dropstones and biogenic structures.

346

GLACIOLACUSTRINE ENVIRONMENTS

suggesting multiple discharge points or one stream that moved frequently. Some deltas build at active ice margins and others on stagnating ice margins. Deltas formed at active ice margins are fed from subglacial or englacial tunnels where sedimentation eventually builds the deposit to the lake surface. Deltas formed at stagnating ice margins are fed by meltwater streams that originate in active ice but end up flowing over the surface of dead ice and into the lake. Sediment influx occurs mainly during the melt season and deposition generally results in a fan-shape landform because distributaries commonly splay at the mouth of the meltwater stream supplying the sediment. Sediment is transported and deposited along a portion of the delta

until the distributaries become inefficient and the flow seeks a steeper gradient and switches to another portion of the delta. Generally, only one portion of the delta is active at any time. Deltas consist of both subaerial and subaqueous portions. The uppermost portions (topset beds) of the delta are deposited by fluvial processes and tend to be thin because vertical aggradation is limited by the local base (lake) level (Fig. 11.5). The bottom contact of the topset sediments is controlled by the depth of fluvial scour in the river channels feeding the delta. The top of the fluvial beds is determined by elevation of overbank flooding. Thus, given a stable lake level, the fluvial topset beds will be about the same

(a)

ICE-CONTACT FACE TOPSET

SE

RE

FO

DENSITY CURRENTS

T

TOPSETFORESET CONTACT

BOTTOMSET

(b) LAKE

TOPSET

FORE

SET

BOTTOMSET

FIG. 11.5. Ice-contact delta depositional environments. (a) Coarse-grained delta, with high-angle foresets. Density underflows can be generated by inflowing meltwater (quasi-continuous currents) or delta foreset slumps (surge currents). (b) Fine-grained delta, with lowangle foresets that can form in the distal portion of an ice-contact delta or more typically where the delta is physically separated from the ice by an outwash stream.

GLACIOLACUSTRINE ENVIRONMENTS

thickness as the yearly vertical range of the meltwater streams building the delta. The classic ‘Gilbert’ delta forms in higher energy, coarse-grained systems; the foreset beds have steep dips (up to 33°)(Fig. 11.5(a)). Sand and larger clasts that were deposited by the meltwater stream at the margin of the lake move downslope by a variety of mass-movement processes, such as avalanching, creep, slump, debris flows and grain flows (Nemec, 1990). Larger clasts are not imbricated and tend to lie parallel to bedding, indicating that they slid downslope individually rather than being transported as bedload in a density current. Massive to parallelbedded gravity layers commonly alternate with openwork gravel and coarse sand lenses (Martini, 1990). Fining upward and coarsening upward gravel sequences are also common (Clemmensen and HoumarkNielsen, 1981; Thomas, 1984). Deltas with shallowly dipping foresets tend to form in lower energy, fine-grained systems (Fig. 11.5(b)). Sedimentation occurs from bottom-hugging quasicontinuous underflows creating small fan-shaped lobes on the delta front (Gustavson et al., 1975; Leckie and McCann, 1982). The upper portions of the delta front commonly reveal evidence of rapid sedimentation and dewatering: slumps, ball-and-pillow structures, roll-up structures, pipes and dish structures. In addition, ice-contact deltas typically have interbeds of diamictons and ice-rafted debris and usually have evidence of postdepositional collapse from the melting of underlying or adjacent ice. The glacier functions as a temporary support for the glaciofluvial sediment. When the ice melts, either during or following sedimentation, the bedding is deformed by gravity-driven processes. The removal of support produces a distinctive landform that has a coarse-grained steep, ‘ice-contact face’ on the proximal side and gentle, depositional slope on the lake side. Ice-contact deltas are often called ‘heads of outwash’, because they represent the source of the sediment supply for the fluvial outwash environment (Chapter 10). On the middle to distal delta front, lobe-shaped sediment bodies composed of climbing-ripple sequences and drape lamination are constructed. The distal delta front deposits merge imperceptibly with low angle to horizontal bottomset beds that consist mainly

347

of drape lamination built up from waning density underflows (Kelly and Martini, 1986). Clays and fine silt accumulate on the delta foreset beds at locations and at times when current activity at the bed is low. Thus, clay dispersed in the water column may settle on areas of the delta removed from points of inflow (distributary channels) during the non-melt season (autumn, winter and early spring). These clay-rich horizons punctuate the deltaic sediments with what may (but not necessarily) be an annual rhythm. 11.3.2.2. Subaqueous fan Glaciofluvial detritus carried to the lake via tunnels near or at the base of the ice form subaqueous fanshaped deposits well below the surface of the lake (Figs 11.4 and 11.6(a)) (Rust, 1977; Gustavson and Boothroyd, 1987; Donnelly and Harris, 1989). The fact that vertical position (elevation) of fans is unrelated to lake level clearly distinguishes them from deltas where the landform surface (topsets) is controlled directly by lake level (i.e., the topset–foreset contact forms at the lake surface). However, both icecontact deltas and subaqueous fans mark ice-front positions. The individual landforms commonly coalesce forming a ridge parallel to the ice margin. These features have only recently been recognized as moraines (Fyfe, 1990; Sharpe and Cowan, 1990). Flows, under a hydraulic head, enter the lake at a point source (Fig. 11.6(b)) creating a deposit cored with coarse open-work gravel. Rust (1977) found that the long axes of the larger clasts were parallel to flow in subaqueous fan deposits as opposed to the perpendicular orientation to flow typical of fluvial deposits. Chaotic bedding with large clasts dispersed in sand and massive ‘quick’ sands and dewatering structures occur in the proximal fan area representing rapid deposition from the high-discharge flows entering the lake (Cheel and Rust, 1986). Mid-fan sediment contains mass flow deposits, ball and pillow structures, dune and ripple crossbedding, flame structures, etc. (Postma et al., 1983). Distal fan deposits likely contain climbing ripple sequences and drape laminations that eventually become indistinguishable from lake bottom sediments. Fans are abandoned during glacial recession and are later blanketed with fine-grained suspension deposits.

348

GLACIOLACUSTRINE ENVIRONMENTS

(a) MOULIN

POTEN TIOM ET

RIC

SUR FAC E

ICE LAKE ICE SUBGLACIAL TUNNEL

(b) ICE

LAKE

FIG. 11.6. Subaqueous fan depositional environment. (a) Schematic cross-section showing relationship of fan to glacier hydrologic system (after Gustavson and Boothroyd, 1987; reprinted with permission of the authors). (b) Depositional model for fan (after Shaw, 1985; reprinted with permission from SEPM).

11.3.2.3. Ice-cliff margin The ice-cliff sub-environment develops when lake water is relatively deep adjacent to the ice (Figs 11.4 and 11.7(a)). Frequently, calving icebergs maintain the cliff face and thus ice-cliffed margins are likely to

occur only on active glacier termini where ice is replenished. Sediment accumulates at the base as an apron, with a source from some glacial debris falling from above or, more likely, from sediments squeezed out or shoved forward during ice-marginal fluctuations. The ice cliff debris aprons are composed of

GLACIOLACUSTRINE ENVIRONMENTS

349

(a) ICE

LAKE

SQUEEZ

E-UP

(b) MASS

MOVE

MENT

LAKE

MUD FLOWS

ICE

FIG. 11.7. (a) Ice-cliff depositional environment. (b) Submerged ice-ramp depositional environment. Sediment is derived from diamicts concentrated at the ice terminus and moves by mass movement processes as the ice margin melts and collapses (after Evenson et al. 1977; reprinted from ‘Subaquatic flow tills: a new interpretation for genesis of some laminated till deposits’, by E. B. Evenson, A. Dreimanis and J. W. Newsome, Boreas, 1977, 6, 115–133, by permission of the Scandinavian University Press).

coarse clasts and diamictons in the form of ridges oriented parallel to the local ice margin. They are likely to show evidence of glaciotectonization, particularly the squeezing of diamicton from beneath the glacier. These ridges would often be considered push or Rogen moraines (Chapter 14). The ridges consist of interbedded mass flow deposits, re-sedimented diamicton, lacustrine mud and ice rafted debris, such as isolated dropstones and berg dump. 11.3.2.4. Submerged ice-ramp The ice-ramp sub-environment occurs only at lake margins of stagnating glaciers where water partially covers melting ice (Figs 11.4 and 11.7(b)). Basal and supraglacial debris elevated and concentrated by repeated overriding of active ice over dead ice at the terminus is released by melting. The ice margin, as well as ice-cored debris islands and peninsulas in the

lake, melts and episodically sheds debris (Lawson, 1982). The sediment distribution processes are gravity-driven for the most part, except for wind-driven surface currents that circulate fine suspended sediment. Mass movement processes of creep, slumps, debris flows and surge currents produce a complex lithofacies association of massive to bedded diamicts, massive to cross-bedded sands and mud drapes (Eyles et al., 1987). The characteristically poorly sorted, weakly bedded sediments, flame structures and rip-up clasts indicate deposition by a variety of subaqueous flow processes (Table 11.5). 11.3.3. Lake Bottom Sub-environments The lake basin depositional setting is the largest glacier-fed lacustrine sub-environment in terms of area (Fig. 11.4). It represents sites in the lake located at a distance sufficiently removed from the points of

350

GLACIOLACUSTRINE ENVIRONMENTS

(a) (b) PLATE 11.2. (a) Rhythmites, glacial Lake Hitchcock, Massachusetts, USA. (Photo by J. Hartshorn.) (b) Couplets. Winter clay layers indicated with arrows. Summer silty-fine sand deposits (between the clays) consist of multiple graded laminae.

sediment influx such that highly variable flows entering the lake are damped and the effects of gravity-driven processes are minimized. Sedimentation is more regular and the deposits are finergrained than in proximal sites (Smith and Ashley, 1985). Rhythmic parallel laminated sediments are characteristic of lake bottom deposits (Plate 11.2(a)). Multiple graded laminae (3–5 mm thick) are common and likely represent temporal pulses in the inflow (Plate 11.2(b)). Limnological currents determine the transport path of the suspended fraction (silts and clay). Although these fine particles continuously settle out, a large portion remains trapped in the epilimnion until

thermal stratification is destroyed during autumn cooling (Fig. 11.2(c)). In the autumn, the lake water is overturned bringing surface water closer to the lake bottom. Because it is at the end of the melt season the inflowing water is clearer (i.e., less dense) than the deeper lake water and thus overflows on the surface. The net result of the overturn is to clear the lake of most of the suspended sediment. These fines end up as a drape on the lake bottom and become part of the annual deposit. Whatever fine material remains in the water column continues to settle out until the next melt season. There is a relatively sharp contact between the coarser sediments deposited prior to and during the overturn compared with the finer

GLACIOLACUSTRINE ENVIRONMENTS

sediments deposited after the overturn. Under ideal conditions one coarse–fine couplet represents an annual deposit and is termed a varve (Plate 11.2(a)). However, if sediment enters the lake periodically throughout the year (including winter), the annual signature may become confused and obliterated (Shaw et al., 1978). Individual, temporally isolated surge-current events, because of an increase in meltwater from the glacier or a mass movement event within the lake itself, may produce an additional coarse–fine couplet. This surge-current deposit can be distinguished from the annual coarse–fine couplet because it is likely to have a gradational contact between the coarse and fine components because of the rapid and continuous settling of sediment out of the flow. It is wiser to use the more general descriptive term rhythmite to describe repetitious bedding rather than the more specific temporal term, varve, unless the deposits are demonstrably annual. The identification of rhythmically bedded deposits as annual (varves) provides a means of unravelling the history of the lake in terms of short-term glacierrelated processes (ablation and runoff) and long-term climatic change record. A variety of techniques have been used to demonstrate the annual nature of the rhythmites: (1) sediment budget considerations in modern lakes comparing total annual influx with calculated net accumulation represented by the rhythmite (Østrem and Olsen, 1987); (2) isotopic dating using Cs137, Pb210, or C14 (Coard et al., 1983; Leonard, 1986a); (3) occurrence of seasonally generated microfossils (Lemmen et al., 1988; Lotter, 1989); or (4) sedimentary structures (Smith and Ashley, 1985). The list of characteristics, below, can be used to interpret which rhythmites are annual and not a deposit from a single sedimentation event (river flood, slump, etc.): 1 although the couplet may, in general, fine from bottom to top, little to no fining occurs in the coarse layer because the layer represents an accumulation over a long time period (weeks to months); 2 there is a sharp contact between coarse and fine portions of the couplet suggesting a break in sedimentation, rather than continuous settling. The break may be the autumn overturn of the water column;

351

3 the fine layer fines upwards reflecting seasonal control, diminishing sediment supply in the lake, and clearing of the water column during winter months; 4 trace fossils (lebenspuren) on bedding planes both within the coarse layer and at the top of the fine layer suggest several episodes of low to zero sedimentation; 5 diatom microfossils occurring at the same position within each rhythmite reflect a seasonal bloom (spring?/summer?); 6 at any one site within the lake the fine-layer thickness, which represents slow sedimentation for the same amount of time each year, remains consistent from year to year, whereas the coarse layer thickness varies reflecting the seasonal-toseason variability of rapid sedimentation within the overall framework of the regular annual cycle.

11.3.4. J¨okulhlaups High-discharge events where meltwater dammed beneath or at the margin of the glacier is released suddenly are termed j¨okulhlaups (Chapters 8 and 9). These glacial outburst floods may occur several times within a season, regularly once a year (Clement, 1984; Liverman, 1987) or at several year intervals (Bj¨ornsson, 1998). They are accompanied by dramatic erosional and depositional effects. The draining of the lake during a single j¨okulhlaup may be irregular with several episodes of fluctuating discharge superimposed on the main discharge curve, presumably reflecting the blocking and clearing of flow paths within the glacier (Sturm et al., 1987). Erosion takes place near to the source of the j¨okulhlaup and deposition may occur in the ice-dammed lake, if any water remains, on proglacial fluvial plains and in distal lakes downstream (Waitt, 1985; Schmok and Clarke, 1989). Large coarse-grained bedforms occur in proximal areas and large erratics are often strewn along the river by icebergs transported by the flood. Distal deposits are characterized by a blanket of rapidly deposited sediment. Thick coarse–fine couplets (up to a few metres in thickness) may be found tens of kilometres downstream. The couplets are composed

352

GLACIOLACUSTRINE ENVIRONMENTS

of bedded to massive sand overlain by massive silt and clay representing an entire j¨okulhlaup or perhaps only a single pulse within an outburst flood.

by drift or bedrock and thus tend to be semipermanent. The distance from the ice results in density reduction of the meltwater streams by warming and by decreasing the suspended sediment concentration either by deposition or by dilution by non-glacial tributaries carrying low sediment load. Depending upon longevity, distal lakes may evolve gradually from a system supplied directly with glacial detritus into one dominated by meteoric runoff and reworked sediment from ice-free drainage basins. The most important change in parameters through time is the reduction in total annual sediment load to the lake.

11.4. DISTAL GLACIER-FED LAKES Many lakes fed with meltwater are separated from the glacier by a fluvial outwash plain (sandur). Because of the fluvial ‘buffer’ separating distal lakes and the source of sediment and water (the glacier), distal lake deposits tend to be fine grained and well-sorted in comparison with deposits of ice-contact lakes. Distal lakes occur in structural basins or in valleys dammed

1

2

3

Type A Clay

Type A Type B Draped Lamination Type B Type A

I

II

III

IV

V

WINTER

SUMMER

Type B

Type A Clay Draped Lamination with Crossbeds

SUSPENSION

GRAIN FLOW

TURBIDITY CURRENT

I

II

I

III IV

2 3

V

FIG. 11.8. Distal glacier-fed lake depositional environment. Sediments become thinner and finer grained from proximal (1) to distal (3) deposits. Dashed line brackets one year’s accumulation.

GLACIOLACUSTRINE ENVIRONMENTS

11.4.1. Proximal Sub-environments Deposition in proximal areas of lakes separated from the glacier occurs primarily in deltas with a minor amount of deposition on the lake margin in between the deltas. 11.4.1.1. Deltas In general, the processes of formation of ice-contact deltas and distal deltas are similar. The differences are related to the physical separation of the lake water body from the vagaries of a melting glacier. In distal lake deltas, large clasts and icebergs are trapped in the intermediate fluvial system (Fig. 11.8). The suspended load is decreased and the meltwater warms up as it moves over the sandur. Because distal lakes are not in contact with the glacier they are likely to be thermally stratified with a well-developed thermocline (Fig. 11.3(a)) as opposed to ice-contact lakes that are sediment stratified (Fig. 11.3(b)). The clast sizes composing the deltas vary with the setting and range from pebble-cobble gravels in ‘Gilbert’ deltas with steep foresets, to sands and

353

silts in low-angle foresets (5–15°)(Fig. 11.5(a),(b)). However, because the deltas in distal lakes are at a distance from the source, they tend to be finer-grained (sands to clay) than ice-contact deltas and dominated by climbing-ripple sequences (Plate 11.3(a),(b)). Many of the fines discharged into the lake are trapped above the thermocline and transported across the lake aided partially by the momentum of the infilling meltwater. Wind stress helps the movement within the epilimnion. Flow gradually slows down and sediment settles through the metalimnion decreasing flow density. Thus, spatial distribution of a ‘summer’ sediment layer shows a proximal–distal thinning and fining trend, as well as a ‘piling-up’ of sediment on one side of the lake owing to Coriolis effects (Sturm and Matter, 1978; Smith et al., 1982). Deposition from the epilimnion occurs as a drape distributed at all elevations in the lake. The upper delta foreset deposits are commonly composed of rhythmically bedded (up to a metre thick) fining-upward sand and gravel beds. These deposits are probably not seasonal but represent episodes of sedimentation during the melt season. Two common causes are: (1) high discharge events, or (2) changes in points of sediment

(b) PLATE 11.3. (a) Ripple–drift sequences typical of fine-grained low-angle deltas. A sequence (representing waning flow) consists of Type ‘A’ ripples at base that grade to Type ‘B’ and culminates with drape lamination. Scale is 30 cm. (b) Shallowly dipping sandy foreset beds composed of climbing ripple sequences (Glacial Lake Hitchcock, Massachusetts, USA). Scale is 30 cm. (a)

354

GLACIOLACUSTRINE ENVIRONMENTS

influx resulting from relocation of distributary channels in the topsets. The deposits on the mid-delta foresets are finer grained (mainly sand), and foreset dips are less steep (<10°) than upper-delta foresets (Fig. 11.4 and Plate 11.3(b)). Density underflows carrying turbid mixtures of sand, silt and clay frequently deposit climbing ripple-drift sequences capped by silt drapes under waning flow. Climbing ripples have been classified into two end-member types (Jopling and Walker,1968): Type A, where stoss-side laminae of ripples are eroded, and Type B where stoss-side laminae are preserved (Fig. 11.9). Type A and Type B are end-members in a continuum of ripple types representing a shift during a waning flow from rapidly migrating ripples with low vertical aggradation rates to slowly migrating ripples with high vertical aggradation rates. The draped lamination capping the sequence is formed when ripple migration ceases and the form is blanketed by a drape of sediment from suspension. A series of flume experiments with varying time–

velocity curves simulating waning flow produced climbing ripple-drift sequences similar to natural ones (Ashley et al., 1982). The flume runs with the most ‘natural’ appearing sequences had vertical accretion rates of 0.3 m h–1. Thus, most individual sequences deposited in natural accumulated within 1–2 hours. The climbing-ripple sequences are composed mainly of fine sand (3.0–3.5 ø) and are commonly repeated. The mid-delta foresets are characterized by lobes marking point sources of sediment (delta distributary channels) that are likely to shift frequently. Thus, the sandy mid-delta foresets are a three-dimensional mosaic of overlapping lobes (Stone and Force, 1983). Thin, continuous clay drapes presumed to be time-stratigraphic horizons (winter clay) occasionally interrupt the stacked climbing-ripple sequences. An individual sequence is not an annual deposit. Thus, the mid-delta foresets are sites of rapid sedimentation receiving deposition from both episodic surge-currents possibly generated by slumping of upper delta foresets or from riverine

Increasing Suspended Load Increasing Bed Load Type A

VY

VX

Draped Lamination

Type B

q

VY

q VX

VY

q VX

VY

FIG. 11.9. The ripple types reflect a changing proportion of suspended and bed load. Vx , ripple migration velocity; Vy, aggradation rate; ␪, angle of climb of ripple crossbedding (from Ashley et al., 1982; reprinted with permission from the International Association of Sedimentologists).

GLACIOLACUSTRINE ENVIRONMENTS

flood events and quasi-continuous underflows (the subaqueous continuation of meltwater streams) (Fig. 11.8). Because of high sedimentation rates in this proximal site, the two types of underflow deposits are likely to be interbedded and may be difficult to differentiate. However, some studies in modern lakes comparing the long-term sediment record with historical river inflow data have been able to distinguish the isolated surge current deposit generated from either a mass movement event or river flood from the background of seasonally controlled sedimentation (Østrem and Olsen, 1987; Siegenthaler and Sturm, 1991).

355

11.4.1.2. Lake margin Shoreline sedimentation in between deltas is restricted to colluvium washed from adjacent terrain and sediment dispersed within the epilimnion. The uppermost deposits, beaches, form a narrow bench of coarse, sorted sediment, whereas the shallow shoreface deposits tend to be homogeneous mud and silt (Fig. 11.10) (Sturm and Matter, 1978; N. D. Smith, unpublished data). This lack of stratification in deposits accumulating in shallow water may reflect: (1) bioturbation, (2) wave disturbance or (3) mixing of sediment by wind-generated currents.

(a) thermocline

Homogeneous Mud

Rhythmites

(b) thermocline

Homogeneous Mud

Rhythmites

(c) thermocline

Homogeneous Mud

Rhythmites

FIG. 11.10. Spatial variation in lake-bottom environment deposits with different dispersal mechanisms. (a) Lakes dominated by underflow. (b) Lakes dominated by overflow/interflow. (c) Lakes that have a combination of all types of flow.

356

GLACIOLACUSTRINE ENVIRONMENTS

11.4.2. Lake Bottom Sub-environments Rhythmites deposited in the lake basin are the distal equivalents of the lower delta foreset rhythmites and the sedimentary sequences sandwiched between persistent clay layers found in the mid-foreset deposits. These equivalent sedimentary units are shown as 3, 2 and 1 on Fig. 11.8. The sand-silt layer changes gradationally from a complex unit on the mid-delta foreset consisting of climbing ripple-drift sequences to lower delta foresets (rippled and multi-laminated beds) to lake bottom (parallel multi-laminated beds). Slump-generated underflows may complicate the rhythmicity of lake bottom sediments. Lake bottom deposits at sites removed (up to a few kilometres) from the point of sediment influx tend to be thin (<1 cm) rhythmites consisting of few multiple laminations. The coarser laminae are deposited from summer underflows whereas underflow-interflows deposit laterally equivalent silts and clays along the lake margin and over topographic highs. A thin fining-upward clay layer representing sedimentation during the rest of the year evenly blankets the more variable summer layer. The contact between the two layers of the couplet is relatively sharp, which results from alternation of dispersal mechanisms transporting sediment to the site. The silt is transported by bottomhugging underflows whose dispersal patterns are controlled by topography; thus, total silt thickness varies with topography throughout the basin. The finer layer blankets the coarser sediment, accumulating rapidly following the autumn overturn and then gradually during the rest of the year. 11.5. STRATIGRAPHY AND LANDFORMS 11.5.1. Spatial and Temporal Variations Glacial-fed lakes exist both during the advancing and retreating phases of an ice mass but, during ice advance, proglacial deposits are commonly overridden and incorporated into the base of the ice. Consequently, most glacial lake records are of deglaciation. One landform associated with glacierfed lakes (deltas) consists of coarse-grained proximal deposits that are perched above surrounding topography. Gilbert’s (1890) classic study of the deltas of

glacial Lake Bonneville revealed that the topset– foreset contact in the deltas marks the elevation of the lake surface at the time of deposition. Lacustrine beach deposits (coarse, sorted sediment) are commonly found at the same elevation as the topset– foreset contact thus supporting the interpretation of the former lake level. Other landforms (deltas or fans) lie as amorphous mounds in valley bottoms (Edwards, 1986). Fans are often blanketed with younger lake bottom sediments and thus not visible as landforms but only show up in cores or borings in the deposits (Sharpe and Cowan, 1990). Extensive flat plains that are usually dissected by younger drainage systems are the landforms produced in the lake bottom environment. The spatial distribution of sediments in glacier-fed lakes is a function of the mechanisms of sediment distribution. Underflows are controlled by gravity and the deposits are thus constrained to topographic lows (Fig. 11.10(a)). Sediments dispersed by overflows/ interflows blanket topography (Fig. 11.10(b)). A combination of processes accentuates the thickening in lows and thinning over highs (Fig. 11.10(c)). Rhythmites typically occur below the thermocline and homogeneous muds above. Depending upon the regional topography (mountains, rolling hills or plains) the details of the deglacial deposits will vary; however, in general, the deposits consist of a series of horizontally stacked morphosequences that stratigraphically onlap and become younger in the direction of ice retreat (Fig. 11.11). A morphosequence is a suite of deposits that are related to a single ice-front position. A morphosequence is generally deposited during a short time-period (few hundred years) and is graded to a common local base level (Koteff and Pessl, 1981). A typical glaciolacustrine morphosequence consists of a proximal/distal, coarse-to-fine suite of deposits. Sedimentation in glacier-fed lakes is dependent on the proximity to ice. As the glacier melts, the sediment source moves and, at any given site, the sedimentation rate decreases or essentially drops to zero with time. The glacial water and sediment source is replaced by meteoric water and reworked upland deposits. The majority of lakes eventually drain leaving the deposits exposed to subaerial processes.

GLACIOLACUSTRINE ENVIRONMENTS DELTA

DELTA

FAN

LAKE

MORPHOSEQUENCE

357

DELTA LAKE

MORPHOSEQUENCE

LAKE

MORPHOSEQUENCE

RETREAT OF ICE

FIG. 11.11. Deglacial glaciolacustrine morphosequences.

11.5.2. Sediment Record Most of the glacial sediment carried to lakes remains in the basin, except for a minor fine-grained portion that may be transported through the lake by epilimnial currents. Lake basins are sediment ‘sinks’. These sediments and the organic remains deposited simultaneously with them provide an excellent opportunity for recording a long-term (hundreds to thousands of years) uninterrupted environmental record (Lowe and Walker, 1984) (Chapter 2). Sediments of the lakes that once fringed the great ice sheets provide a tangible record from the now extinct water bodies (Teller, 1987) (Fig. 11.12). Modern alpine glacial lake deposits provide a direct link from past to present by revealing changes in sedimentation, rates, lithic composition and organic content through time (Leonard, 1986a,b; Donnelly and Harris, 1989). The stratigraphic record of glacierfed lakes is the best continental source of continuous record of climate change spanning from the Late Glacial through the Holocene and up to the presentday influences of humans (Renberg and Segerstrom, 1981; Leonard, 1986b). The parameters commonly used in reconstructing geochronology are varve thickness (Pickrill and Irwin, 1983), diatom (Lemmen et al., 1988) or pollen records, paleomagnetic record (Thompson and Kelts, 1974; Sandgren et al., 1988; Ridge et al., 1990), isotopic dating (Cs117, Pb210

and C14 ) (Ashley, 1979) and tephrochronology (Leonard, 1986b), as well as stable isotope records (Lister, 1988; Menzies, 1996, chapter 14). A several-thousand-year varve chronology was developed in Sweden by measuring the absolute thickness of individual varves in sequence at hundreds of localities and correlating the sequences between localities (De Geer, 1912). The underlying assumption in developing the chronology was that the rhythmites were varves and that couplet thickness was controlled by regional climatic factors, temperature and precipitation, and not local factors. The chronology was later corroborated by C14 dating. Antevs (1922) developed and successfully used a similar varve chronology for eastern North America (Ridge and Larsen, 1990). The apparent faithful recording of environment change in the continuous sedimentation records of glacial-fed lakes makes glaciolacustrine deposits invaluable sources of information for reconstructing the past.

11.6. RECOGNITION OF PAST GLACIOLACUSTRINE ENVIRONMENTS There are few descriptions of pre-Pleistocene glaciogenic sediments that are unequivocally of glaciolacustrine origin. This is largely because of the absence of distinctive fossil fauna or flora in ancient

358

GLACIOLACUSTRINE ENVIRONMENTS

º

60

B A Y

H U D S O N

52º

Calgary

Marine Limit Regina Winnipeg

Maximum Extent of Wisconsinan Glaciation

PERI

Marine Limit O

R

IG

LAKE

RON

MIC H

HU

s ip

pi

Detroit LA

Chicago

LA

KE

ER

IE

KE

O N TA

RIO

New York

R. er

e

iv

r

R

iv

R

o

Cincinnati

i

r

R.

Oh

i

ve

is

Wa a s h b

o

R

Illin

ri

kilometres

ou

500

E

ssis

Miss

0

L

N

AK

Minneapolis M i

N

44º

Montreal

A

L

SU

E AK

FIG. 11.12. Area covered by proglacial lacustrine sediment (stippled) deposited during the last retreat of the Laurentide Ice Sheet (from Teller, 1987; reprinted with permission of the author). Major avenues of overflow into the Mississippi River and Atlantic Ocean basins shown by arrows.

glaciolacustrine and glaciomarine sediments, and because of some general similarities in facies. Hambrey and Harland (1981) present three key criteria to distinguish glaciolacustrine from glaciomarine deposits: (1) the presence of dropstones, (2) finely graded stratification, such as laminated clay or siltstones, and (3) association of diamictites with resedimented deposits. Most interpretations are founded on the existence of varvites/laminites with dropstones associated with

facies consistent with continental glaciation. But laminated sediments with dropstones can be associated with either glaciolacustrine or glaciomarine environments. Unfortunately, this ambiguity in possible interpretations does little to assist in the recognition of ancient glaciolacustrine rocks. Consequently, there is a need to develop specific criteria that distinguish ancient glaciolacustrine from glaciomarine sediments (for details see Menzies, 1996, chapter 4).

GLACIOLACUSTRINE ENVIRONMENTS

11.7. CONCLUSION Concerns about future global warming raise questions pertaining to the interaction of earth systems (hydrosphere, atmosphere, lithosphere, biosphere and cryosphere). Large complex computer models (Global Climatic Models) have been developed to examine the nature of these interactions and analyse how climate is affected. The only test of the validity and reliability of these model results, aside from waiting patiently, is the history of past climatic change as recorded on the geologic record. Deep sea sediments and continental lacustrine and loess sediments are likely to be the best long-term records (Chapter 2). It is essential that new research concentrates upon developing techniques in order to interpret continental

359

climates (temperature and moisture) from the sedimentary record. Similarly, improved techniques are required in dating (Menzies, 1996, chapter 14) and correlating deposits from widely separated basins (Menzies, 1996, chapter 8). Paleoclimatic studies are likely to employ geochemistry and biochemistry to interpret the record left by climate-sensitive organisms. New geophysical methods need to be explored to help unravel the finer-scale details of the paleomagnetic record (1–10 years) imprinted upon longterm records (>100 years). Central to much of the above is an increased need to understand the complexity and significance of glaciolacustrine sediments as significant sources of, as yet, untapped data. With this data, the past can hopefully be used to help predict the future.

This Page Intentionally Left Blank

12

MODERN GLACIOMARINE ENVIRONMENTS R. Powell and E. Domack

12.1. INTRODUCTION Glaciomarine sediment contains the most detailed and continuous direct record of Late Cenozoic glacial fluctuations and often the first direct indications of a glacial period. In addition, many of the Earth’s preCenozoic glaciogenic sedimentary rocks are interpreted as glaciomarine. Knowledge of glaciomarine environments and processes that occur within them is, therefore, invaluable.

Glaciomarine (glacimarine, glacial marine) environments may be defined as marine environments in sufficient proximity to glacial ice that a glacial signature can be detected within the sediments. Some or all of the sediment is released from grounded or floating glacial or sea ice (Table 12.1). The environment has a plethora of processes and sediment sources owing to the combined action of ice (glacial and sea), water (fresh and sea), wind and biological activity. Knowledge of glaciomarine processes has advanced

TABLE 12.1. The relationship between climatic regions, summer temperatures, glacier temperatures and equilibrium line elevations Climate region

Summer temperatures

Glacial temperature

Equilibrium line elevations

Polar Ice Cap

<0°C

<0°C, except at base under special conditions

At calving line or seawarda

Polar Tundra (Subpolar)

<10°C

<0°C

At calving line or 100s of m above sea level

Boreal

>10°C (1–3 months)

At and below pressure melting

100s of m above sea level

Temperate Oceanic

>10°C

At pressure melting throughout

100s–1000s of m above sea level

a

The potential equilibrium line elevation below present sea level can be calculated, for instance see Robin (1988). Climate regions and summer temperature characteristics are taken from Trewartha, 1968. After Anderson and Domack (1991).

361

362

MODERN GLACIOMARINE ENVIRONMENTS

TABLE 12.2. Sedimentation rates of recent Antarctic glacio-marine deposits of the continental shelf Location

Sediment type

Rate (mm year–1)

East Antarctica Mertz-Ninnis Trough (George V Coast)

Diatom mud, ooze Diatom ooze Silty clay (pebbly mud)

3.1–3.4 3.0 1.5–0.67 2.1

Ross Sea Terra Nova Bay

Diatom mud

0.05–0.20

Antarctic Peninsula Andvard Bay Hughes Bay Brialmont Cove Marguerite Bay

Diatom, pebbly mud Diatom, pebbly mud Sandy mud Diatom mud

Gerlache Strait

Diatom mud

Bransfield Strait

Diatom mud

2.0–5.0 2.8 1.7–49.8 0.3 2.0 2.7 2.7 1.8

South Shetland Islands Admiralty Bay

Sandy, pebbly mud

4.7–33.2

Prydz Bay

in the past decade. However, large gaps still remain in understanding these environments and many interesting aspects remain unsolved. Sedimentation processes in cool-temperate, boreal and, to a lesser degree, polar-tundra climatic regions are perhaps best documented whilst processes in polar climatic regions are probably least documented. 12.1.1. Glaciation and Climate Glaciation at sea level depends on latitudinal and regional climatic factors such as winter and summer temperatures, insolation, snowfall and cloudiness, and a wide range of climatic conditions that, when combined with relief, are capable of maintaining glaciers at sea level (Table 12.2). Most glaciers of coastal Antarctica, for example, are in the polar ice cap category where mean summer temperatures are less than 0°C and ice temperature is subzero. Surface ablation is limited mainly to sublimation; therefore, even though accumulation averages <2 cm a–1 on the surface interior of the ice sheet, glaciers can be maintained at sea level. Meltwater and terrestrial vegetation are limited or lacking under these conditions. Unglaciated areas are cold, polar deserts with

Method

14

C Pb C 14 C

210

14

14

C

14

C Pb 14 C 14 C 210 Pb 210 Pb 14 C 210 Pb 210

14

C

little atmospheric moisture and rapid evaporation. Thicker portions of the ice sheet and its major drainage systems are melting or melting/freezing at their base. Sea ice is extensive and can be strongly seasonal, such as surrounding Antarctica, or persistent as in the Arctic Ocean. The transition from polar ice cap to polar-tundra or subpolar climatic conditions occurs where mean summer temperatures are >0°C. These conditions occur along the western side of the Antarctic Peninsula, sub-Antarctic islands, coastal Greenland, arctic Canada, Iceland, Svalbard, Nova Semlya and Severnaya Zemla. Some summer melting occurs and glacier equilibrium lines are at or just above sea level, but terrestrial vegetation is limited. Glaciers, under these climatic conditions, experience surface melting during the summer but freezing temperatures remain throughout most of the glacier. Boreal climates have one to three months with mean temperatures >10°C and, except for summers, are rather dry. These climatic conditions support large areas of mountain glaciers but only a few reach sea level, for example, along the Kenai Peninsula, north-central Gulf of Alaska. Glaciers experience surface melting, but vegetation is shrubs and lowland forest. The mildest

MODERN GLACIOMARINE ENVIRONMENTS

climate supporting glaciation at sea level is temperate oceanic, occurring along the eastern Pacific mountainous coasts of North and South America as far as latitude 57°N in southeastern Alaska and northern British Columbia, and 47°S in western Patagonia. Mean temperatures for the four- to seven-month summers are >10°C, temperature minima are 0–2°C and, locally, rainfall occurs year round. Extremes in ablation are balanced by very high snow accumulation, internal ice is temperate, glaciers are melting or melting/freezing at their base and equilibrium lines are well above sea level. It is the presence of coastal mountains that allows glaciers to thrive under temperate oceanic conditions. These coastlines are biotically productive with dense ground vegetation and even forested glacier termini (e.g., Malaspina Glacier). 12.2. SUBGLACIAL PROCESSES Generally, transportation and deposition of englacial and subglacial debris by marine-ending glaciers is similar to terrestrial glaciers (Chapter 8). However, some differences may occur owing to the fast and extensional flow behind grounding-lines of major marine-ending glaciers. Knowledge on this question is poor and conjectural. Lodgement of basal debris may occur in the form of till sheets that may be discontinuous. However, because of fast basal ice flow and high basal water pressures, deforming beds may be common. Where subglacial water is in excess, most subglacial sediment appears to be flushed out by meltwater streams. The presence of deforming beds and large volumes of subglacial water may produce different types of depositional processes at groundinglines. 12.3. MARINE-ENDING TERMINI Glaciers that terminate in seas or oceans may, in the final few hundred metres or kilometres, enter the body of water as: (a) a floating terminus, or (b) a grounding tidewater cliff (Plate 12.1). These different modes of terminating in water result in: (1) variations in how sediment arrives at the point of delivery from the ice to the water body; (2) the effect of marine conditions such as temperature, currents and salinity on sedimentological and glaciological processes; (3) the

363

transient interaction between marine and glacial environments; and (4) specific sedimentological processes ongoing within the proximal zone of the ice, the immediate proglacial subaquatic zone and in more distal aquatic areas. 12.3.1. Floating Termini 12.3.1.1. Formation and maintenance Floating termini form when glacial flow lines converge and flow is unconfined owing to very low frictional forces between glacier and sea. Commonly the annual 0°C isotherm is at sea level, but even if not, the ice mass is cold (polar). The termini form a continuum from confined floating glacier-tongues, to confined ice shelves, to unconfined floating glacier tongues, to fringing ice shelves, to large embayment ice shelves (Anderson and Domack, 1991) (Fig. 12.1). At present 57 per cent of the Antarctic coastline can be delineated as floating termini, 38 per cent is tidewater termini and 5 per cent is sediment/rock coast (Drewry et al., 1982). Most marine-ending glaciers elsewhere have tidewater termini, although some in the Arctic terminate as fringing ice shelves or floating glacier tongues. Having a large horizontal surface area at low elevation (<100 m) means that floating termini are susceptible to rapid surface ablation should equilibrium lines rise. This condition probably places a climatic limit for ice shelves at about a mean annual temperature of –8°C and mean summer temperature of –2°C (Swithinbank et al., 1988). Because the temperature of ocean water is often at or above 0°C under floating ice, basal ablation of these ice masses may be just as important as surface ablation (Jacobs, 1989). Nearly all of the mass loss from the George VI Ice Shelf (ca. 4.5 to 1.8 × 10 kg3 m–2 a–1 ) is by basal melting aided by warm ocean waters (Bishop and Walton, 1981; Potter et al., 1984). 12.3.1.2. Debris transport and release Sediment deposited under floating termini comes from glacial sources and marine currents. Basal debris is melted out from floating termini within a few kilometres of a grounding-line. However, englacial

364

MODERN GLACIOMARINE ENVIRONMENTS

PLATE 12.1. Examples of two types of glacier termini ending in the sea. (a) A floating glacier-tongue of an outlet glacier from the East Antarctic Ice Sheet producing icebergs in the Ross Sea, in which there are also remnants of seasonal sea ice. (b) A grounded vertical cliff of the Margerie Glacier, Glacier Bay, Alaska, with its bergs and a tourist craft in Tarr Inlet.

MODERN GLACIOMARINE ENVIRONMENTS

365

FIG. 12.1. Map of Antarctic and surrounding oceans showing coastal regions mentioned in text and illustrations. Dotted lines on continent are ice flow directions (after Drewry, 1983) and arrows in ocean are current directions. Dashed line in ocean is maximum limit of seasonal pack ice while dotted line is northern extent of drifting icebergs. Solid line in ocean is the approximate location of the southern polar front or Antarctic Convergence. Major ice shelves are indicated by the shaded areas.

debris entrained by converging outlet glaciers along the coast and then buried by snowfall, can be released from the base by bottom melting in the distal zones of a floating terminus (Powell, 1984). These influences of debris release can be seen in the confined George VI and fringing Larsen Ice Shelves in Antarctica (Domack, 1990). Debris is probably contributed in

similar styles to the Ross Ice Shelf, Antarctica, where several large outlet glaciers have grounding-lines located up-valley from where they join the ice shelf. Release of debris from beneath floating termini depends on oceanic temperature and pressure gradients that are controlled by water depth and circulation. For example, water flowing toward a

366

MODERN GLACIOMARINE ENVIRONMENTS

grounding-line under a floating terminus flows down a pressure gradient and can potentially melt the basal ice, whereas outflowing water moving up the pressure gradient will probably freeze to the base. The ensuing pattern of debris release may vary amongst floating termini, especially embayment-type ice shelves, and may unpredictably change through the history of a floating terminus depending on oceanic circulation conditions. There is growing indirect evidence from Antarctica indicating that the grounding-lines of large ice shelf systems are dynamic settings with regard to sediment transport and delivery to the marine environment. An early intuitive model by Drewry and Cooper (1981) recognized the role that grounding-line cavities should play in the circulation of basal melt-out and suspended debris. Geophysical investigations along the Siple Coast of West Antarctica have demonstrated that grounding-lines of ice streams (fast-flowing currents in an ice shelf) may occupy distinct zones of subglacial debris transport in the form of deformable, water-saturated diamicton (Alley et al., 1986). ‘Till deltas’ are one possible product of ice stream/ice shelf interaction (Alley et al., 1989b, fig. 5). The presence of thin subglacial water films is also recognized within the system but the transport of suspended particulates out into the sub-ice shelf or floating glacier-tongue environment remains a major point of future investigation. Shallow water environments, especially those bordering ice cliff or piedmont settings, are greatly affected by currents of wind or tidal origin. Low siliciclastic sediment supply at the calving line allows such currents to effectively sort and transport sediment into broad aprons of sandy sediments rich in carbonate bioclastic detritus. 12.3.1.3. Ice shelf pumping Studies of ice shelf grounding-line environments, including the Ronne Ice Shelf, have demonstrated significant vertical movement of the shelf indicating that the grounding-line may in fact undergo periodic uplift and set-down under the influence of tides (Stephenson and Doake, 1982; Kobarg, 1988). The effect within any grounding-line cavities would be similar to a bellows, with the expulsion of water and

suspended sediment out into the marine setting during falling tide or ice shelf let down (Talbot and Von Brunn, 1987). Unfortunately this pulsing process is difficult to observe directly but, when combined with debris rain-out from undermelt, it should produce laminae and deformation structures within sediment such as diamicton. In many cases a disrupted diamicton m´elange is produced. Indications of how tidal pumping may influence sediment transport are provided by studies of partially floating termini along the Antarctic Peninsula. Oceanographic data demonstrate that, within the water column, mid-water fine-structure consists of distinct interflows of turbid water adjacent to the glacial termini in deep water (Fig. 12.2). A mechanism similar to tidal pumping is implicated in the formation of these features (Domack and Williams, 1990) (Fig. 12.3). Suspended particulates sampled from these features include fresh quartz silt grains, even very fine-grained sand (Plate 12.2). 12.3.1.4. Calving styles and processes Calving (sometimes termed berging) is a rather infrequent occurrence for ice shelves; the mean annual iceberg production in Antarctica is about 2800–3040 km3 (Ørheim, 1985), 60–80 per cent of which comes from embayment ice shelves and 15–25 per cent from ice streams and outlet glaciers (Drewry and Cooper, 1981). Calving mechanisms are not fully understood but possible causes can be grouped into those related to: (1) glacial flow (creepfailure with extensional flow, rift zone development, surface and bottom crevasses); (2) a combination of glacial and oceanic processes (Reeh-type calving from force imbalances on the ice face (Reeh, 1968)); (3) bending and flexure by tides and cyclic vibration (Holdsworth and Glynn, 1978); and (4) impacts of large icebergs; and oceanic processes (storm and tsunami wave energies and ocean currents) (Kristensen, 1983). Calving processes from floating termini produce the largest icebergs that can survive long transport distances. For instance, one calving event from the Ross Ice Shelf, Antarctica, produced a tabular iceberg that was 154 × 35 km2 and it drifted across 2000 km of the Ross Sea during a 22-month period (Keys et al.,

MODERN GLACIOMARINE ENVIRONMENTS

367

Salinity (ppt) 33.750 0.0

34.000

34.250

34.500

(S)

34.750 Surface Layer (T)

(mg/l)

100.0

Depth (m)

Cold Tongues

A

200.0

Cold Tongue

B

300.0

Near-Bottom Turbidity

C Bottom 400.0 -1.500

0.0

-1.000

1.0

2.0

Temperature 3.0

-0.500

4.0

Suspended Sediment Concentration (mg/l)

0.0

5.0

6.0

FIG. 12.2. Vertical profile of temperature (T), salinity (S) and suspended particulate matter (mg l–1 ) within an Antarctic (subpolar) fjord. Location of SEM images (Plate 12.2) of suspended particulate matter are also shown (A, B, C). Note significant suspended particulate matter within mid-water layers (cold tongues) and comparison with surface layer and near bottom turbidity (modified after Domack et al., 1994).

368

MODERN GLACIOMARINE ENVIRONMENTS

Kilometres 1.5

1.0

Sea Ice Biogenic SPM

0

0.5

0

0.5

Icebergs Glacier Terminus

Metres

100

Cold Tongues

Coo

200

ling

Terrigenous SPM 300

Warm Clear Outer Bay Water Sub-ice Cavity

Sill

Grounding Line

400 Turbid Bottom Water

FIG. 12.3. Model for cold tongue (interflow) formation in polar fjord settings as based upon observations in Antarctic fjords (Domack and Williams, 1990; reprinted with permission of the authors and the American Geophysical Union). Circulation is driven by storm or tidal surges within the basal cavities of tidewater termini or ice shelves.

1990). It contained 1200 km3 of ice, which is nearly 50 per cent of the net annual accumulation of the Antarctic Ice Sheet (Giovinetti and Bentley, 1985). The actual calving event is relatively passive and has no influence on sea floor sediment several hundred metres below (Plate 12.3). 12.3.2. Tidewater Termini

terminus (Plate 12.4). Glaciers form a vertical cliff once they reach sea level because the glacier calves faster than it melts. As the ratio of melting to calving increases, a more parabolic surface profile is produced. Under these latter circumstances a cliff will remain grounded unless the conditions required for floating are met. Maintaining a tidewater terminus appears to be primarily a function of the water depth in which the grounded cliff ends (Brown et al., 1982).

12.3.2.1. Formation and maintenance 12.3.2.2. Debris transport and release Glaciers that end as tidewater termini occur in every climatic setting (a cool-temperate example is shown in Fig. 12.4). Glaciers of temperate ice do not have the tensile strength to float as intact ‘slabs’ (Powell, 1980), while those of cold ice apparently do not have sufficient flow velocities to maintain a floating

Downwasting of a glacier surface produces supraglacial debris by melt-out, especially on valley and outlet glaciers. Most of that debris is from supraglacial sources and, for marine termini, it often falls into the many crevasses caused by related extensional flow.

MODERN GLACIOMARINE ENVIRONMENTS

369

Little englacial and no basal debris is moved to the surface by ice flow at the terminus because flow lines, under these terminus conditions, remain near-horizontal. Englacial debris derived from upglacier tributaries is often exposed in cross-section on the cliff face in near-vertical layers that may be deformed by local ice flow. All basal debris being horizontal remains near the sea floor and is only exposed above sea-level in shallow water depths. Melt-out/fallout A tidewater terminus is melted by solar heat, stream discharges and seawater. In general, melting by air is much slower than calving and other melting rates. Meltwater streams rising up the cliff can melt small vertical tunnels or larger channels that produce ‘caves’ at sea level. This process can enhance meltout of debris and assists in creating calving embayments around the efflux (Sikonia and Post, 1980). Melt-out from the face produces rockfall and debris grainfall. Dumps of supraglacial debris are produced from the surface and crevasse-fills during calving. Debris flows and slurries slide down the face but disaggregate owing to the steep angle of the face before entering the sea. All supraglacial contributions to grounding-line systems are minor except locally where glaciers have abundant supraglacial debris. Melt-out till similar to that from terrestrial glaciers is difficult to produce in these situations because of fast flow and non-stagnant conditions. Some melt-out deposits may form on the upglacier side of morainal banks or where glacial ‘toes’ become buried by glaciomarine sediment. Squeeze/push

PLATE 12.2. Scanning electron microscope photographs of suspended particulate matter within oceanographic features shown in Fig. 12.2. Note dominance of siliciclastic particles in all samples especially sample B from mid-water cold tongue. Greater biogenic (diatom) constituents are observed in sample from less turbid water at 196 m (A).

Sediment that is pushed by marine-ending glaciers is generally preserved by seasonal fluctuations of termini that are undergoing net retreat. Because marineending glaciers advance across sediment with a high water content, the probability of ejection out from a grounding-line by squeeze and liquefaction processes is likely (Powell, 1984; Smith, 1990). Squeeze-up into basal crevasses is also strongly implied by the distinctive sea floor pattern observed after surges (Solheim and Pfirman, 1985).

370

MODERN GLACIOMARINE ENVIRONMENTS

PLATE 12.3. Fluvial discharges from temperate tidewater termini in Glacier Bay, Alaska, that, of all the glacial sources, contribute the largest volume of sediment to the sea. (a) An upwelling at the sea surface is an expression of the meltwater jet discharging from under Grand Pacific Glacier at about 75 m depth. Gulls are feeding on shrimp that live at the ice face and which are being forced up to the surface by the upwelling. (b) An example of a smaller, englacial stream discharging at sea level from Muir Glacier. The diameter of the conduit is about 4 m.

MODERN GLACIOMARINE ENVIRONMENTS

371

4 1

5

2 3

14

24

20 21 25

6

23

21

13

7

15

18 19

16

8

10

9

12 11 22

17

12

13

FIG. 12.4. Three-dimensional sketch showing the major environments for glaciomarine sedimentation in a cool-temperate climate, based on the present north-northwest Gulf of Alaska (not true orthographic scale). Numbered features are as follows: 1, tidewater terminus (see Figs 12.7 (a), (b) for details); 2, side drainage with gravel pocket beach and talus fan on fjord floor; 3, bergstone mud on glacial fjord floor; 4, sandur plain in marine outwash fjord; 5, sediment gravity flow channels in marine outwash mud; 6, entrance sill capped with morainal banks and grounding-line fan; 7, continuation of outwash mud; 8, rocky shore with gravel pocket beaches; 9, tectonically uplifted banks of older rocks exposed by winnowing on the shelf; 10, extensive tidal mudflat in estuary; 11, recurved and cuspate spits from alongshore transport; 12, marine outwash mud on the continental shelf; 13, relict sand deposits; 14, thick temperate rain forest of spruce/hemlock with muskeg swamps; 15, raised marine terrace; 16, eolian dunes fed from spit at wave-dominated delta; 17, large slide-slump areas on the continental shelf; 18, ice-dammed lake with sublacustrine grounding-line fan and lake laminites with iceberg-rafted debris; 19, terminal moraine and gravelly beach; 20, sandur/delta system with small moraines and gravelly beach; 21, modern littoral sand; 22, bank on continental shelf of relict glacier-marginal sediment, currently being winnowed; 23, tidewater terminus just at sea level with short-headed stream, fan deltas; 24, river-dominated estuary with stunted barrier island system offshore; 25, sea valley on continental shelf being infilled with marine outwash mud (after Powell and Molnia, 1989; reprinted by permission of Elsevier Science Publishers, Amsterdam).

Mass movements Sliding, slumping and sediment gravity flowage are common events at tidewater termini of temperate and subpolar glaciers. These processes are common because of high sedimentation rates, iceberg calving impacts and glacier-push of grounding-line sediment accumulations such as morainal banks, groundingline fans and ice-contact deltas. Small events with short (within 1–2 km) run-out distances are common and may occur at intervals of days, hours or even continuously (Phillips et al., 1991). Larger events occur more sporadically but may remove large

volumes of sediment from a grounding-line and transport it tens of kilometres away from the glacier. Sediment accumulated against a grounding-line can collapse back toward the glacier in small events as basal icebergs calve, as well as in quite large volumes when the grounding-line retreats from the site. Hydrographic processes Waves, tides and ocean currents may affect melting of tidewater termini producing distinctive sediments. Commonly, waves, tides and ocean currents are second-order controls on grounding-line sedimenta-

372

MODERN GLACIOMARINE ENVIRONMENTS

PLATE 12.4. A calving event from the grounded, 55-m-high tidewater cliff of Lamplugh Glacier, Glacier Bay, Alaska. The slab is photographed about half way through its rotating fall (a) and then a large splash and wave is produced on impact (b).

tion except where large storm events occur on the continental shelf, or where the volume of sediment supplied from the glacier is low, as in polar settings, or in shallow water where the processes can rework much of the sediment. Fluvial discharges Fluvial discharges from tidewater termini are greatest from temperate glaciers. Such discharges dominate sediment production in temperate and, to a lesser extent, subpolar glaciers. Streams enter the sea from englacial and subglacial conduits; the latter position providing most of the sediment. Discharge is so high in the melt season that it forms turbulent jets where momentum forces in the flow are stronger than the buoyancy forces of the freshwater entering denser saltwater (Plate 12.5). For slower flows, such as

during early rising and late falling discharges, or from colder glaciers, the effluent is a forced plume where buoyancy dominates momentum forces. Underflows have been inferred and may be enhanced by hyperconcentrating sediment as particles rapidly fall out of transport when the flow detaches from the sea floor, but these remain unconfirmed. Eventually the jet transforms to a plume as it rises to sea level. Occasionally the flow may cause ‘boiling’ or, at slightly higher discharges, ‘fountaining’ at sea level above the upwelling (Plate 12.3). The plume spreads as a barotropic flow away from the discharge at a level of neutral buoyancy, which in sea water is commonly at sea level, that is, it forms an overflow or hypopycnal flow. It is from this jet/plume system that laminated sediments (cyclopsams and cyclopels) can form (Mackiewicz et al., 1984; Cowan and Powell, 1990).

MODERN GLACIOMARINE ENVIRONMENTS

373

PLATE 12.5. A berg produced by a submarine calving from the McBride Glacier tidewater cliff in the background, Glacier Bay, Alaska. These types of bergs provide excellent exposures of sections of glaciers otherwise unobtainable. Here, the sole of the glacier with subglacial till is exposed on the left side of the berg and all the basal debris layers to the right are in a stratigraphic-up sense. All of this debris will produce iceberg-rafted debris.

12.3.2.3. Calving styles and processes Tidewater termini differ from floating termini in that the calving line coincides with the grounding-line. Calving from ice cliffs occurs by three processes: (1) spalling of seracs by fracturing above sea level and the iceberg falls into the sea; (2) large sheets shear off the glacier face and the iceberg sinks vertically or topples forward into the sea (Plate 12.4); and (3) detachment below sea level and then the iceberg rises to the surface. While Muir Glacier, Glacier Bay, Alaska was retreating at about 2 km a–1, about 106 m3 day–1 of ice was calved (Powell, 1983). Subaerial calving occurs by fracture propagation (Hughes, 1992) and if icebergs do not shatter on

impact they can travel down through the seawater column. Some blocks either impact upon the sea floor, or come sufficiently close, such that their preceding pressure wave redistributes sediment. Surface waves caused by the impact generally cannot influence the sea floor initially because it is too deep; however, at shorelines the waves cause erosion and transport icebergs above the high-tide line. Icebergs from submarine calving are often termed ‘bergy bits’ and commonly contain englacial and basal debris. These icebergs can disturb bottom sediment as they detach, and some are seen with sea floor sediment on top of them. They are major sources of iceberg-rafted debris (IBRD) from tidewater termini (Plate 12.5).

374

MODERN GLACIOMARINE ENVIRONMENTS

FLUVIAL grounding-line fans (subaqueous outwash fan) moraine banks

LODGEMENT ramp-type moraines (?)

MORAINE BANKS (Ra moraines) (DeGeer moraines) (submarine moraines) (moraines)

SQUEEZE some cross-valley moraines

MELTOUT & CALVE-DUMPING frontal dump moraines

PUSH push morainal bank

FIG. 12.5. End-member processes that contribute to forming grounding-line systems in the form of morainal banks (not to scale). Beneath each process are terms that have been used to describe the deposit in the literature; here it is recognized that morainal banks can form by the end-member process or any combination of those processes. Alternative terms for morainal banks in the literature are also listed.

12.4. TYPES OF GROUNDING-LINES Grounding-line systems are defined as sedimentary depositional systems formed at grounding-lines of glaciers ending in large bodies of water, such as large lakes or seas. Such a system can be distinguished by its processes of formation, geometry, facies associations and internal architecture. Studies of glaciomarine sediments have shown there are several different varieties of grounding-line systems that include: (1) grounding-line fans and glacier-contact deltas, (2) morainal banks, and (3) grounding-line wedges (‘till deltas’) (Figs 12.5, 12.6, 12.7 and Plate 12.6). Understanding how to recognize each type by lithofacies analysis and being able to infer processes of formation are important for inferring the potential stability of grounding-lines and for interpreting paleoglaciological and paleoclimatological conditions from a sedimentary record.

Grounding-line fans are point-source depocentres with a fan geometry, formed at conduit mouths where meltwater streams discharge subaquatically (Fig. 12.6). They are composed of subaquatic outwash (deposited from the discharge while in contact with the floor of the water body) and a variety of sediment gravity flow and suspension settling deposits. They have also been named transverse eskers and deltas, beaded eskers, subaqueous outwash fans, subaqueous esker deltas and glacier-contact fans. Glacier-contact deltas are commonly formed by short-headed streams, and are typified by processes, facies associations and the architectural arrangements of fan deltas with glacial influences such as icerafting and pushing (Nemec and Steel, 1988). These fans often develop from grounding-line fans building up to sea level (Powell, 1990). Delta processes are grouped into bedload dumping, hemipelagic sedimentation, by-pass and diffusion of

MODERN GLACIOMARINE ENVIRONMENTS

375

MELTOUT

(a) LVE CA PING M DU

MELTOUT

WS

FLO

SUSPENSION SETTLING

L FAL CK RO LL A F AIN GR

ROLL DUMPING

ROCKFALL GRAINFALL TRACTION FLOW

T

U LTO

ME

MELTOUT

SEDIME NT GRAVITY FLOWS

LODGEMENT MELTOUT

PUSH SQUEEZE

(b) UPWELLING

RISING TURBULENT FLUME

TURBID OVERFLOW FLUME

SETTLING LAYER

SUSPENSION SETTLING

INTERSTRATIFIED SEDIMENT CYCLOPSAMS CYCLOPELS

BEDLOAD DUMPING

SEDIMENT GRAVITY FLOWS

UNDERFLOWS

FIG. 12.6. Processes and lithofacies associations contributing to grounding-line systems at tidewater termini of temperate glaciers (not to scale). (a) A morainal bank, (b) submarine outwash of a grounding-line fan and generation of cyclopsams and cyclopels (after Powell and Molnia, 1989; reprinted by permission of Elsevier Science Publishers, Amsterdam).

376

MODERN GLACIOMARINE ENVIRONMENTS

(a) Active deforming bed Basal Squeeze debris

Flows

G.L. Rainout (grain fall, rockfall)

(b)

C.L. - Coupling line G.L. - Grounding line

Trace of wedge brink-point as a function of sediment accumulation rate

C.L.

G.L.

(c)

Old deforming beds with amalgamated contacts accreting by continuous erosion up-glacier

Density flow and rainout deposits G.L. C.L.

FIG. 12.7. A possible mechanism for formation of grounding-line wedges and how they can expand with glacial advance to form sediment sheets that terminate as wedges. An ice stream feeding an ice shelf has basal debris that is feeding a subglacial deforming bed (a). At the grounding-line the deforming bed is squeezed out to form sediment gravity flows on an inclined surface. The wedge of sediment produced is also contributed to by grain fall and rock fall from under-melt of the ice shelf. The volume of that rain-out sediment decreases with distance from the grounding-line and also contributes to the thinning of sediment distally. As the sediment builds up to meet the base of the ice shelf, the grounding-line advances as does the coupling line (b). Upglacier from the coupling line, erosion can occur. Between the coupling line and the grounding-line, the deforming bed aggrades by being added to from basal debris but, because transmitted sheer decreases with depth, lower sediment in the deforming bed is deposited. With continued advance, the distance between the grounding-line and coupling line increases producing a deposit with a sheet-wedge geometry (c). Predictably, the deposit is composed of sediment gravity flow and rain-out facies topped by deformed-bed deposit with amalgamated contacts (not to scale).

MODERN GLACIOMARINE ENVIRONMENTS

377

PLATE 12.6. View of a grounding-line at a tidewater terminus of a temperate Alaskan glacier. The bottom left-hand segment is sorted glaciomarine sediment of gravels up to boulder size draped with a coating of mud from meltwater streams and pebbles from meltout of icebergs and the glacier cliff. The rest of the photograph is a wall of clear glacier ice with bubbles and pebbles within. This was the first observation made of a grounding-line, which in the photograph is about 15 cm long.

sediment. As a river enters the sea its competence is dramatically reduced and bedload is deposited and dumped rapidly near the river mouth. The suspended load can continue into the sea in a hypopycnal flow of fresher, less dense, water flowing over saline water. Particles suspended in this plume eventually settle out and deposit hemipelagic sediment. Meanwhile, some processes on the sea floor act to remove sediment totally from the nearshore delta area. Such processes as low density turbidity currents and large-scale slides behave in this manner. Other bottom processes act to ‘smear’ bottom sediment away into deeper water, such as tidal or wave action, creeps or slumps and perhaps hyperconcentrated sediment underflows from the river. Morainal banks are made of various glacier contact lithofacies (Figs 12.5, 12.6 and 12.7) that include chaotic mixtures of diamicton, gravel, rubble, sand

and mud facies formed by rock and grain fall processes resulting from calve-dumping and melt-out, lodgement, and squeeze/push processes from under the glacier. The deposits have a bank geometry (for terminology see Mitchum et al., 1977) and are analogous to terrestrial end moraines, but are formed subaquatically. Banks created mainly by pushing are pushmorainal banks. Morainal banks also include the ramp-type sublacustrine moraines that form more by subglacial lodgement processes (Barnett and Holdsworth, 1974), frontal-dump moraines from melt-out at the terminus and dumping of supraglacial debris during the calving process (Syvitski and Praeg, 1989). Other banks are thought to form by squeezing out of subglacial sediment beyond the grounding-line to produce landforms like the cross-valley (Rogen) moraines. Other forms of morainal banks may have

378

MODERN GLACIOMARINE ENVIRONMENTS

their entire lengths made of facies similar to grounding-line fans (moraine banks), indicating subglacial discharge of meltwater in either sheet flows or rapidly migrating conduits across a grounding-line. These types of banks confirm a continuum from lone grounding-line fans (that can form beaded eskers), to interstratified grounding-line fan and glacier-contact facies to solely glacier-contact facies. ‘Till deltas’ were defined as wedges of sediment made up of subglacial diamicton overlying dipping strata inferred to be mainly sediment gravity flow deposits (Kamb and Engelhardt, 1989; Alley et al., 1989b; Blankenship et al., 1989; Scherer, 1991). They are thought to be forming today beneath the Ross Ice Shelf. The term ‘grounding-line wedge’ is used rather than ‘till delta’ to avoid the problem of defining all of the sediment as till, and to avoid an association with sea level that could be inferred from the term ‘delta’. Grounding-line wedges are thought to form mainly from deforming subglacial diamicton that reaches the grounding-line (Fig. 12.7). There, sediment redistribution by gravity-flow processes produces deposits with dips as low as 1° or less, because the deforming till should contain abundant fine-grained matrix. Grounding-line advance across these deposits will cause subglacial deformation. The deforming layer may thicken downglacier. It has been suggested that larger-scale deposits independently termed ‘till tongues’ (King and Fader, 1986) defined as wedgeshaped deposits of till interbedded with stratified glacimarine sediment (King et al., 1987). The apparent major control over the two end member types of grounding-line systems, morainal banks and grounding-line wedges, is subglacial meltwater. With abundant, free-flowing water morainal banks appear to form; in contrast, grounding-line wedges appear to form with smaller volumes of more confined water and deforming beds. Using these simplistic limitations, it has been suggested that deposits such as ‘till tongues’ may be formed mainly by polar or subpolar glaciers flowing on deforming beds; the glaciers may or may not end as floating termini. Alternatively, morainal banks are most likely formed by temperate or subpolar glaciers where supraglacial water contributes to subglacial water in channelized or sheet flows, and the glacial termini are most commonly tidewater cliffs.

12.5. PROGLACIAL AND PARAGLACIAL ENVIRONMENTS Proglacial and paraglacial environments lie beyond the glacier proximal settings of grounding-line systems and subglacial ice shelf systems. The term paraglacial was used (Church and Ryder, 1972) to define non-glacial processes that are directly conditioned by glaciation and it is used here to include marine systems that receive glaciogenic sediment but are not in direct contact with glaciers nor icebergs (Powell and Molnia, 1989). In many instances processes in proglacial and paraglacial settings are similar. However, the presence of glacial ice sometimes creates distinctive processes in proglacial settings. 12.5.1. Hemipelagic Suspension Settling Sediment plumes from subglacial and marginal streams are the transportation agents for most glaciomarine sediment in front of temperate and subpolar marine-ending glaciers. The overflow originating from freshwater sources at tidewater termini has been modelled as a barotropic flow originating as a buoyant jet (Calabrese and Syvitski, 1987, figs 1 and 2). Proximal marine outwash occurs as a traction facies deposited along the run-out distance of the jet from the submarine efflux of a subglacial stream (Fig. 12.8). Sometimes the jet detachment will be at the glacier face but with larger discharges the zone may be away from the face. At the detachment zone small pebbles to sand sizes appear to fall out and are virtually ‘dumped’ (Powell, 1990). Particles as coarse as medium sand can remain in the plume as it rises to the surface and sand can be transported by the plume as much as 1 km from a face (Cowan et al., 1988). Sediment is released from overflow plumes episodically and the areal position of a plume changes as a result of tides and surface wind sheer (Fig. 12.6). Another control on the paths of plumes is the Coriolis effect that can lead to an increase in sediment accumulation rates at particular sites by two to three times in high latitude fjords (Syvitski, 1989). Oceanic currents such as those moving along continental shelves are important for mud transport, as are eddies both in distal areas and near a glacier face.

MODERN GLACIOMARINE ENVIRONMENTS

(a) Low discharge

(b) Moderate water discharge, high sediment discharge

(c) High discharge

FIG. 12.8. Alternative fan types with different discharges (not to scale). (a) At low discharges the fan grows progressively with coarse traction deposits dumped at the efflux and most deposits from settling or slumping/sliding and sediment gravity flows. (b) At higher sediment discharges the horizontal jet remains in traction to produce rapid sedimentation deposits in sheets or crude scour-andfill structures. Settling and sediment gravity flow deposits make up more distal deposits. (c) At high discharge the horizontal jet can include a migrating barchanoid bar at the detachment zone (after Powell, 1990; reprinted with permission of the Geological Society of London).

Settling of finer particles from the plume can be enhanced by three processes: flocculation, agglomeration and pelletization (Syvitski et al., 1987). The rising plumes from submarine discharges can have sufficient sediment concentration that they reach a level of neutral buoyancy within the brackish water column and form an interflow. Increased sediment concentration may originate from rainstorm events or sudden release of sediment stored subglacially by channel migration if there is a change of head or gradient such as during submarine iceberg calving. During winter, these discharges can shut down completely, especially from subpolar glaciers. 12.5.2. Biological Production Biological activity adds primary sedimentary particles to glaciomarine environments and they can be a major component of glaciomarine sediments where siliciclastic production is low. Organisms also speed settling rates of siliciclastic particles by pelletization

379

and mix bottom sediment by bioturbation. An important aspect of polar glaciomarine environments is the abundance of biogenic facies. The primary reason for this facies, particularly in the ice-proximal environment, is the limited amount of siliciclastic sediment supply to the sea. High supply rates of siliciclastic sediment dilute any biogenic sediment produced in warm glaciomarine settings; but, by contrast, bioclastic carbonate and siliceous ooze are the dominant facies found in Antarctica today. Bioclastic carbonates occur as shell coquina that fringes tidewater cliff margins of the East Antarctic Ice Sheet and as localized deposits of coquina on shelf banks, such as those found in the Ross Sea (Taviani et al., 1993) and near Prydz Bay (Quilty, 1985). Biogenic siliceous muds and oozes are the dominant deposit within deep water basins along the Antarctic continental shelf, including deep fjords along the Peninsula (Fig. 12.9). At depths >500 m, icebergs are prevented from scouring biogenic accumulations and mixing the sea floor sediment and, as a result, diatomaceous particulates accumulate from vertical settling and dilute, sediment gravity flow deposition. In polar-tundra, boreal and cool-temperate climates, aquatic biological activity is high, but successions are dominated by siliciclastic sediment owing to higher sediment yields from glaciers. Macrofaunal communities in cool-temperate to polartundra climates are dominated by infaunal populations in soup- and soft-grounds and epifaunal communities on firm- and hard-grounds (Dale et al., 1989; Aitken, 1990). 12.5.3. Ice Rafting Ice-rafted debris (IRD) originates from floating termini such as ice shelves as well as icebergs and sea ice; ice-shelf rafted (ISRD), iceberg rafted (IBRD) and sea-ice rafted (SIRD) debris will be used to specifically identify the debris source. The glacial character of IRD is important to recognize because sea ice can raft debris in areas where glaciers are absent on land. Generally, coarser particles such as gravel and sand sizes are most often recognized as being ice rafted, although contributions of finer particles can be significant. Resulting sediments include discrete lonestones that may be recognized as

380

MODERN GLACIOMARINE ENVIRONMENTS

Bays/Fjords Danco Coast and Palmer Archipelago Outer Bay

Inner Basin

Ice Terminus

Water Depth (m)

Iceberg and phytoplankton zone

Late summer snowline ~ 50m Surface melt

Concentrated ice

0 100 200 300

Phytoplankton

Occasional surface plume

Warm

Cold ice, no englacial melt tunnels

Cold

Crevasse flushing

Surface gyre

Maximum biogenic flux

Cold tongue

Superimposed ice zone

Grounding line cavity

Infrequent resuspension

Diamicton Sediment gravity flow

400 500 Biosiliceous pebbly muds

Terrigenous facies Coast parrallel sand belt as lateral equivalent (in shallow areas) ~ 5 km

FIG. 12.9. Facies relationships representative of fjords found along the northwestern side of the Antarctic Peninsula, Gerlache Strait. The dominant source of siliciclastic sediment includes ice-rafted debris, cold tongues (<–0.5°C), and sediment gravity flows; all of these processes are most active in ice-proximal settings near the terminus and within the inner basins. The absence of abundant meltwater prevents the formation of surface overflows and, hence, restricts siliciclastic sedimentation to within the inner basins. Phytoplankton (diatoms) are most productive within the warm (>0°C) surface waters of the outer bays where iceberg and sea-ice rafting also takes place. These two processes lead to biosiliceous pebbly muds within the outer basins.

dropstones if they deform strata, or as outsized-clasts when their diameter is larger than the thickness of strata in which they are embedded, or in the form of clusters of particles. Other features of IRD are till pellets and mud pellets that were frozen debris encased in glacial ice.

basal debris. Consequently, on the sea floor, ISRD may show a change in provenance as well as shape (from more rounded to more angular) with distance from the grounding-line. Deposits from these processes have been termed shelfstone muds or shelfstone diamictons, depending on their texture.

12.5.4. Ice-shelf Rafting

12.5.5. Iceberg Rafting

Ice-shelf rafting occurs when basal and englacial sediment is transported beyond a grounding-line by ice deforming into a floating glacier-tongue or an ice shelf. Most particles are released by meltout and descend to the sea floor in rock falls, grain falls or suspension settling of individual particles. There is a significant lack of particle clusters as can occur when icebergs roll. Because layers at higher elevations within the ice move farther from the grounding-line than lower layers, as happens when you slide the top of a deck of cards, englacial debris encased high above the ice sole can be transported farther than

Rafting of glacial debris by icebergs can provide a significant volume of glaciomarine sediment and IBRD and can be used to indicate glaciation when mixed with marine sediment of a non-glacial origin. The distribution of icebergs in the circum-Antarctic and North Atlantic (Figs 12.1 and 12.10) clearly demonstrate the far-travelled character of IBRD. Hence, while local climatic conditions may not be suitable for glaciation, IBRD can be deposited with ‘warm water’ sediments when icebergs are transported quickly by ocean currents. Fourier shapes of silt particles can distinguish IBRD from glaciofluvial

MODERN GLACIOMARINE ENVIRONMENTS

120º

60º

0º

ffin

30º

60º

Greenland Sea

Greenland

Ba

381

75º

PF

Ba

y

W

Norwegian Sea EG C

C G C NA

LC

60º

Europe North America GS

*

most southerly iceberg sighting

30º

Africa

FIG. 12.10. Map of the North Atlantic region and adjacent glaciated (shaded area) and non-glaciated masses. Dashed line in ocean is maximum limit of seasonal pack ice while dotted line is southern extent of drifting icebergs. Solid line in northern Baffin Bay is the location of the polar front. EGC, East Greenland Current; GS, Gulf Stream; WGC, West Greenland Current; LC, Labrador Current; NAC, North Atlantic Current.

quartz grains (Cai et al., 1992) and this technique could be useful for distinguishing iceberg-rafted silt from hemipelagic muds accumulated from other sources on continental shelves. Melt-out, sediment gravity flows and dumping are the main mechanisms of debris release from icebergs, and they contribute particles of clay to boulder size to the sea floor by rock fall, grain fall and suspension settling. The ultimate volume of IBRD depends on the size and numbers of icebergs and also their residence times (horizontal velocity), debris concentrations, melting rates and rolling rates. Because melting rates in air are slower than those in water, icebergs eventually get top-heavy and roll. Submarine melting rates vary around and under the iceberg depending on the surface attitudes and currents. Potentially, IBRD contributions vary with climatic and glacial regime. Large tabular icebergs produced in Greenland and Antarctica can travel to quite low

latitudes (30°N, 40°S) (Figs 12.1 and 12.10) but those produced by large ice shelves contain little debris since the debris is preferentially released from the ice shelf near the grounding-line. Most debris from Antarctic icebergs is thought to be released on the continental shelf owing to long residence times and recirculating drift paths of icebergs offshore. Occasionally, icebergs collect in restricted areas because sea ice prevents them drifting after they calve. In the North Atlantic large icebergs from Greenland are transported far south by the Labrador Current and can cause problems in shipping lanes (e.g., the ‘Titanic’ disaster). Tidewater termini produce smaller icebergs than floating termini and with iceberg transport lengths of only a few tens of kilometres in cool-temperate climates. In boreal and polar-tundra climates these distances can be much longer (Robe, 1980). Because of the rapid melting rate and the coincidence of

382

MODERN GLACIOMARINE ENVIRONMENTS

PLATE 12.7. Calving bay of the Columbia Glacier, Alaska (July, 1989). The bay is choked with recently calved icebergs that are prevented from drifting out of the bay by a shallow sill. The concentration of floating icebergs covers 100 per cent of the bay’s surface and results in conditions that would be the closest thing to a temperate ice shelf. Water depths within this iceberg-covered bay are up to 380 m deep (Krimmel and Meier, 1989). The dotted line indicates the water line against the grounded tidewater cliff.

calving and grounding-lines for tidewater termini, roll-dump structures are common. These structures can be used to infer this environment rather than a floating terminus at a grounding-line since dumping of large amounts of debris en masse is not likely to occur from the base of an ice shelf. 12.5.6. ‘Two Component’ Mixing ‘Two component mixing’ refers to the admixture of sediment particles derived from processes that are sufficiently distinct that they can be recognized by the character of the particles they contribute to a deposit. The two most prevalent in the glaciomarine environment are ice rafting and various types of current-derived sediment, such as that produced by

vertical settling from suspension in low density turbidity flows and bottom currents. The coarser ice-rafted component can usually be recognized by poor sorting of the glacially transported debris. However, ice-rafted silt and clay is more difficult to discern from hemipelagic particles in bottom sediment. In most temperate and subpolar regimes fluvial sediment production will dominate iceberg rafting and bergstone muds accumulate. If icebergs are concentrated for any reason, then the proportion of IBRD can increase to produce a bergstone diamicton interfingering with bergstone mud. Circumstances in which a bergstone diamicton may form are: (1) meltwater discharge is low because of a cold glacial regime; (2) seasonal discharge decreases but iceberg

MODERN GLACIOMARINE ENVIRONMENTS

rafting continues; (3) there is a very high, continuous flux of icebergs in proximal areas; and (4) icebergs occur in a continuous flux in more distal areas where rates of accumulation of marine rock flour have decreased exponentially. Icebergs may be concentrated or have a high continuous flux in situations such as: (1) in constricted basins (e.g., trapped by sills) (Plate 12.7); (2) in oceanic eddies; (3) during catastrophic retreats; (4) during storms blowing icebergs onshore; and (5) during periods when sea ice freezes-in high concentrations of icebergs. In polar environments a third, biogenic, component is involved. Because of the limited role of meltwater sedimentation, dropstone diamictons are found in a variety of situations such as near ice shelf groundinglines and bays with restricted circulation. In general, sediment accumulation rates again decrease from the termini of valley glaciers. This allows for the accumulation of biosiliceous pebbly muds, mixtures of ice rafted and biogenic sediment derived from phytoplankton (Fig. 12.9). In some areas, coarse IBRD and bioclastic debris are concentrated without a fine-grain component. These deposits have been termed residual glaciomarine sediments (Anderson et al., 1980) and occur in nearshore areas or shallow banks where marine currents are sufficiently strong to stop fine-grained sediment from accumulating and bioturbation brings finer sediment to the sediment surface where it is removed by bottom currents. 12.5.7. Sea-ice Rafting Formation of sea ice most commonly occurs when air temperatures drop below the freezing point of marine or brackish water. Its formation expels salt and produces brines that form gravity flows that sink down through the water column owing to their density and cold temperatures (Lewis and Perkin, 1985). Sea ice can also form under cool-temperate climates where during winter in protected areas heavy snow falls produce a slush of ice crystals in surface water that then freezes as temperatures drop. One problem still remaining is to distinguish SIRD deposited under glacial conditions from that deposited where sea ice forms without glaciers on land (Reimnitz and Kempema, 1988; Gilbert, 1990).

383

Sea ice is important in glaciomarine sedimentation because it (1) damps waves (especially during winter storms), (2) changes water column structure, (3) influences biotic productivity, (4) traps icebergs, (5) transports sediment, (6) scours and turbates bottom sediment, (7) deforms shoreline sediment, and (8) reduces calving rates. Sediment is incorporated into sea ice by: (1) freezing of an ice foot, (2) anchorice lifting, (3) trapping of sediment during frazil ice/slush ice formation, (4) bottom erosion, (5) stream wash overs, (6) wave wash overs, (7) offshore eolian transport, and (8) rock falls and slides (Table 12.1). Finally, SIRD is being increasingly considered as an important component of glaciomarine sediments in distal areas or under colder conditions where other siliciclastic input is low. 12.5.8. Transitional Environments Transitional environments between terrestrial and fully marine environments mainly include beach, delta, estuarine and tidal flat systems. In some areas these systems are quite important in producing a sedimentary record as they are large depocentres (Menzies, 1995, chapter 14, pp. 476–477). They can also provide important data for inferring climatic conditions, for example, large deltas are not produced in polar climates but can be significant in cooltemperate and boreal climates. The presence of ice in coastal settings can produce a distinctive record. For example, when grounding-lines of tidewater cliffs are directly at sea level, very gravelly, poorly sorted beaches are formed. Likewise, icebergs and sea ice leave deposits such as boulder lines, and deform intertidal sediment into gouges, wallows and push ridges (Dionne and Brodeur, 1988). 12.6. OTHER NON-GLACIAL AND MODIFYING PROCESSES 12.6.1. Hydrographic Processes In addition to controlling ice melting rates, iceberg paths (Figs 12.1 and 12.10) and paths of turbid freshwater plumes, oceanic currents are important in modifying and redepositing sediments on the sea floor. Hydrography of continental shelves and shallow

384

MODERN GLACIOMARINE ENVIRONMENTS

seas is very site-specific with regard to details of circulation patterns, current strengths and whether the environment is dominated by tides or storm waves. 12.6.2. Sediment Gravity Flows Sites for generating sediment mass movements are depocentres such as grounding-lines or deltas where high sedimentation rates form unstable slopes that collapse. Sediment gravity flows from these depocentres may have short run-out distances because of either significant bottom relief produced by glacial erosion especially in fjords that trap flows, or the continental shelf sloping toward the continent as a result of either glacial loading or preferential glacial erosion which also traps flows, or a low slope of the continental shelf on which flows rapidly slow. However, events other than high sedimentation rates, such as storm waves, biogenic gas and earthquakes may result in sediment instability and consequent mass movements in subaquatic proglacial areas. Sideentry inputs are important additional sources in fjords. Processes of sediment redistribution in proglacial areas are the same as in non-glacial marine settings and large volumes of sediment can be involved. Significant features specific to glaciomarine settings are: (1) the production of mass movement deposits on continental shelves owing to the glacier being on the shelf. In contrast, non-glacial shelves have their sediment sources usually some distance from sites of deposition except during sea level changes; (2) during glacial maxima, glaciers may often extend to the edge of the continental shelf so that all sediment is released directly down the continental slope to form thick wedges of mass flow deposits with glacial signatures. 12.6.3. Iceberg and Ice Keel Scouring and Turbation Plough and furrow marks are common on high latitude shelves in northern and southern hemispheres. In fact, turbation of bottom sediments can be so intense that primary stratigraphy and biotic communities may be destroyed and/or remoulded into diamictons (Menzies, 1996, chapter 4).

12.6.4. Production and Accumulation of Organic Carbon The content and character of organic carbon in sediments are perhaps two of the more important differences between polar and temperate glaciomarine facies, particularly within mud-dominated facies. Studies within shelf and fjord basins in Antarctica have demonstrated that total organic carbon contents within biosiliceous muds can be as high as 3–4 per cent of the total sediment weight (Dunbar, 1988) and up to 6 per cent of the mud fraction (Domack, 1988). Typically, the organic carbon values range between 1 and 1.5 per cent. More important, however, is the fact that almost all of the carbon is of autochthonous marine (algal) origin. In contrast, temperate and subpolar glaciomarine muds are typically poor in organic carbon, with typical values of <1 per cent (Syvitski et al., 1990). When Total Organic Carbon (TOC) content exceeds 1 per cent, the organic matter is invariably detrital or reworked, being derived from stable refractory sources such as bedrock and terrestrial vegetation. The reasons for this are related to climate and high rates of siliciclastic sedimentation associated with meltwater input. Restricted productivity owing to limited sea ice may also be a contributory factor to low indigenous carbon contents. Warmer water temperatures would also allow for higher rates of bacterial metabolism. 12.6.5. Bioturbation Infaunal organisms disturb sediment and progressively destroy sedimentary structures with increasing intensity away from a glacier (Dale et al., 1989). Bioturbation may also allow winnowing of finer sediment and homogenize sediment into a diamicton. The best potential for preserving fine sedimentary structures is where sedimentation rates are high and infaunal populations are consequently absent or low. 12.6.6. Eolian Sources Offshore transport of sediment by wind can be locally important in high-latitude polar ice cap and polartundra climates. Fine sediment is transported to sea directly, or via sea ice. In more temperate climates,

MODERN GLACIOMARINE ENVIRONMENTS

Sediment Yield (m3 a-1 x 105)

20

10

M 0

0

200

400

600

800

Drainage Basin Area (km2)

FIG. 12.11. Relation between sediment yields and drainage basin area for a temperate marine-ending glacier. Sediment yields in icecontact basins (fjord floor basins in which a glacier terminates) are estimated from seismic profiles from Molnia et al. (1984) averaged over the period of time Muir Glacier, Alaska, terminated in or at the head of each basin. Drainage basin areas of Muir Glacier, as determined from Brown et al. (1982), become smaller during glacial retreat. Curve is a best fit logarithmic function. M is the yield from present McBridge Glacier determined from modern sedimentation rates and is consistent with yields estimated for Muir Glacier from seismic profiles (after Powell, 1991; reprinted with permission of the author).

wind also transports sediment offshore but it is heavily diluted by sediment from other sources. 12.7. SEDIMENTATION RATES AND FLUXES Potential siliciclastic sediment yields from glaciated basins decrease with decreasing size of the glaciated basin, assuming the following conditions: (1) uniform subglacial sediment/rock types, (2) constant subglacial debris conditions (erosion-transportationrelease) during glacial retreat, and (3) minimal sediment storage (probably true for marine-ending glaciers except during glacial minimum conditions where extensive land areas are exposed in a drainage basin). A logarithmic decrease in sediment yield with decreasing drainage basin area has been documented (Fig. 12.11) for the retreat of the temperate Muir Glacier, Alaska (Powell, 1991). The sediment yield curve is considered to be a continuous function reflecting rates of glacial erosion and flow. Rates of glacial sediment accumulation on the sea floor depend on the processes of debris release from the glacier (Fig. 12.12). Accumulation rates are a function of: (1) glacier debris content and rate of

385

meltout by sea water; (2) subglacial squeeze/push; (3) sediment discharge from subglacial streams; (4) hemipelagic/pelagic suspension settling from meltwater discharges; (5) debris fall from floating ice; (6) marine hydrographic processes; and (7) sea floor processes involved with mass movement. Of these, sedimentation rates from suspension settling are easiest to measure; few estimates of other processes have been made. An estimate of relative significance of these processes in a temperate setting is that melting of basal debris produces 0.4 per cent of total sediment yield, stream bedload plus mass flow 30.6 per cent, and suspension settling 69 per cent (Powell and Molnia, 1989). Estimates of annual increases in volume of grounding-line systems under these conditions are of the order of 106 –107 m3 a–1 (Powell, 1990, 1991). As expected, rates of sedimentation of hemipelagic siliciclastic glacial sediment logarithmically decrease away from grounding lines (Cowan and Powell, 1991; Domack et al., 1991). The low production of siliciclastic sediment in polar ice cap climates means that sedimentation rates become more a function of biosiliceous productivity and dissolution rates. Surface productivity, in turn, is controlled by sea ice fluctuations and subsequent surface layer mixing. Table 12.3 lists estimates of sedimentation rates for some modern Antarctic glaciomarine sediments. Though far from representative of all depositional environments, especially ice proximal settings, the data do give some basis for comparison to other glaciomarine sequences from cool temperate zones. Sediment accumulation rates vary from 3.4 to 0.5 mm a–1 over two orders of magnitude. Most rates range between 0.3 and 3.0 mm a–1. Because most of these sediments are biosiliceous muds or oozes the rates of siliciclastic sediment accumulation are approximately 70–50 per cent of those listed in Table 12.2. The overall low sedimentation rates emphasize the low rate of siliciclastic sediment supply and the absence of extensive fluvial drainage systems. Total particle flux in the temperate-oceanic fjords of Alaska are several orders of magnitude greater than those of subpolar and polar climates mentioned above (Table 12.4). The results of Powell and others are taken from sediment traps and, therefore, would be expected to represent upper limits to sedimentation

MASS BALANCE WITH TIDEWATER TERMINI SNOW ACCUMULATION + CLOUD COVER = FLOW VELOCITY + MELTING + EVAPORATION/SUBLIMATION + CALVING

SNOW ACCUMULATION COULD COVER

MELTING + EVAPORATION/SUBLIMATION

ELA

Seawater

Atmosphere Temperature (ELA Slope/Surface Slope)

Rainfall

Cliff Orientation

Marine Currents

FLOW VELOCITY

Ice Regimen

Bed Condition

Bedrock

Water

Drawdown

Deforming Sediment

CALVING

Ice Regimen Crevassing Water Depth Flow Velocity Tectonism Sediment Yield Drainage Basin Area Subglacial

Erosion Rates

Transportation Rates

Isostasy

Sedimentation

Eustasy

Grounding-line System Volume Release Rates

Cliff Meltout Calve-dumping Conveyor Belt Squeeze/Push

Length Critical Depth Sediment Repose Angle Fluvial

Sediment Dispersal Patterns Diffusion

High Density Slides/Slumps Icebergs Gravity Flows

Direct Deposition

Plumes

By-Pass

Low Density Landslides Gravity Flows

Meltout Lodgement

FIG. 12.12. Factors controlling mass balance of glaciers, with tidewater termini emphasizing sedimentological controls at grounding lines (not to scale). Many factors are dependent variables, but feedback loops are not shown for simplicity (after Powell, 1991; reprinted with permission of the author).

MODERN GLACIOMARINE ENVIRONMENTS

387

TABLE 12.3. Sediment flux in glaciomarine environments Region and setting

Distance from glacier (km)

Total flux (kg m–2 a–1)

Terrigenous flux (kg m2 a–1)

Biogenic flux (kg m2 a–1)

5 16

2.285 1.219

n.a. n.a.

n.a. n.a.

Ross Seab SW Basin Sulzburger Bay

n.a. n.a.

1.010 0.633

0.741 0.619

0.268 0.013

East Antarctica Amery Troughc Mertz-Ninnis Troughd

70 50

0.471 1.089

0.118 0.033

0.353 1.056

Southeast Greenland e Shelf Trough

n.a.

0.060–0.080

n.a.

n.a.

n.a. n.a. n.a.

0.600–1.300 0.400 <0.050

n.a. n.a. n.a.

n.a. n.a. n.a.

West Antarctica Peninsula Fjordsa

Baffin Island, Canada Fjord Shelf Trough Shelf

e

Alaska, Glacier Bay f Fjord

0 0.06 0.1 0.8 1.0 2.75 4.5 8.2 11.0 14.5

490 430 180 80 28 14 7.4 3.7 2.2 0.17

× × × × × × × × × ×

106 106 106 106 106 106 106 106 106 106

n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

a

Domack et al. (in preparation). Ledford-Hoffman et al. (1986). Domack et al. (1991). d Domack et al. (1989). e Andrews and Syvitski (1991). f Cai and Powell (unpublished data). b c

when compared with preserved flux as determined by radiogenic isotopic studies of sediment cores. Regardless, the contrasts are extreme and demonstrate the proficiency by which temperate glacial regimes supply sediment into the marine realm. In order to obtain a reliable comparison of sedimentation rates in different climatic regimes, Cowan and Powell (1991) restricted the data set to those studies that obtained measurements from settling tubes in modern fjord settings and close to grounding lines (Table 12.4). The results show a dramatic three to four orders of magnitude difference between temperate and polar settings, one to two

orders of magnitude difference between temperate and subpolar settings, and between subpolar and polar settings. 12.8. ADVANCE AND RETREAT OF MARINE-ENDING GLACIERS The stability of marine-ending glaciers has been undergoing investigation since Mercer (1978) suggested that potential climatic warming (from greenhouse gases) may cause the West Antarctic Ice Sheet to disintegrate rapidly and cause a rise in global sea level. While stability of marine tidewater termini, ice

388

MODERN GLACIOMARINE ENVIRONMENTS

TABLE 12.4. Comparison of Holocene ice-proximal sedimentation rates in fjords or restricted basins (after Cowan and Powell, 1991) Location

Rate

Distance from grounding-line

Climatic classification

Reference

McBride Inlet: Southeast Alaska (tidewater terminus)

200–2000 cm a–1

Within 1 km

Temperate

Cowan and Powell, 1991

10–25 cm a–1

Within 10 km

Subpolar

Gorlich, ¨ 1986

Kongsfjorden: West Spitsbergen (tidewater terminus)

5–10 cm a–1

Within 10 km

Subpolar

Elverhøi et al., 1983

Mackay Glacier: East Antarctica (floating glacier-tongue)

1.8 cm a–1

Within 2 km

Polar

Powell et al., 1992

Prydz Bay: East Antarctica (inferred sub-ice-shelf or floating glacier-tongue

0.21 cm a–1

Within 10 km

Polar

Domack et al., 1991

Hornsund: West Spitsbergen (tidewater terminus)

shelves and terrestrial glacial termini is ultimately a product of mass balance, other variables influence stability and hence the advance and retreat of marine termini. Furthermore, some of these variables involve feedback mechanisms that are more complex in marine systems than in terrestrial systems. As with terrestrial termini, glacial mass balance for marineending termini is a function of snow accumulation rate, ablation rate, and glacial flow velocity. Considerations of accumulation rates are the same as for terrestrial systems. Non-climatic responses of marine termini have been noted for Pleistocene ice sheets. Retreat of some temperate tidewater termini has been documented historically, and some, after reaching a position of maximum retreat, have stabilized or readvanced. Readvances have been attributed to: (1) decrease in size of the ablation area, (2) increased precipitation at high altitude following a rise in the equilibrium line altitude (ELA), and (3) accumulation of sediment at the grounding line that decreases water depth and increases terminus stability (Warren and Hulton, 1990; Warren and Glasser, 1992). Marine-ending glaciers differ from terrestrial glaciers in that they have longitudinal profiles near their termini that are concave-up rather than convexup. Once retreat is initiated by a rise in the equilibrium line, the effect can be cumulative because of the near-horizontal glacial profile which

continues to lower by the ‘draw-down’ effect associated with fast flow. Water depth at the grounding line is also considered important because of the potential calving instability at tidewater termini and rapid migration of the grounding-line of floating termini. Changes in water depth can be affected by isostasy, eustasy and sediment accumulation/glacial erosion (Menzies, 1996, chapter 11). In theory, a glacier could initiate its own retreat from a maximum advance by crustal loading as the glacier increased in thickness. If the advance is local and independent of global cooling, then a concomitant eustatic lowering of sea level may not occur. Once initiated, retreat may proceed because rebound is not sufficiently rapid. When rebound had decreased water depth glacial advance could then ensue. Global cooling and glaciation could allow glaciers to advance onto continental shelves simply by eustatic lowering of sea level. If deglaciation of a sufficiently large ice sheet occurs and global sea level rises, then that can force marine-ending glaciers elsewhere to retreat. The timing of eustatic and isostatic movements is important in the generation of packages of glaciomarine sediments because such movements are forcing functions that initiate or terminate packages. The detailed sedimentary history within packages, however, can be more a function of local and regional sedimentary processes.

MODERN GLACIOMARINE ENVIRONMENTS

High-latitude continental shelf and slope systems are relatively poorly explored but they potentially hold the key to documenting linkages between major ice sheet fluctuations and global climate changes inferred from other records such as deep-sea cores and ice cores. Some models have been suggested to formulate ideas for interpretations of sediment packages with multiple glaciations over high latitude continental shelves (Andrews, 1990, 1997; Boulton, 1990; Anderson et al., 1991; Henrich, 1991; Powell, 1991; Syvitski, 1991; Bartek et al., 1991). The effect of rising eustatic sea level has been suggested to initiate instability in a grounding-line/ice shelf system that results in rapid disintegration of a marine ice sheet, specifically, the West Antarctic Ice Sheet. However, models that place more emphasis on longitudinal stresses and basal sliding indicate that marine ice sheets behave less catastrophically and are more stable to climatic and sea level forcings than the earlier models indicate (Van der Veen, 1999). On a smaller scale, water depth at a terminus is a function of sea-floor topography that can be modified by either glacial erosion (e.g., over-deepening) that creates deep water to force retreat, or glacial deposition (e.g., grounding-line systems of morainal banks, grounding-line fans, ice-contact deltas) at tidewater termini that causes shoaling and decreases calving rate to enhance glacial stability. Evidence from various sources points to instability of ice sheets on different time scales. The Laurentide Ice Sheet may have collapsed several times over 10 ka

389

intervals to produce ice-rafted debris layers in the North Atlantic (Heinrich, 1988; Andrews and Tedesco, 1992; Bond et al., 1992; Broecker et al., 1992; Andrews, 1997). The surface of the Greenland Ice Sheet appears to have experienced irregular temperature changes (interstadial episodes) over very short time intervals (500 to 2000 years) (Johnsen et al., 1992). A recent model of the West Antarctic Ice Sheet indicates that it could be self- regulated by the rates of erosion and transport of subglacial sediment over periods of the order of 100 ka (MacAyeal, 1992). On a longer time-scale, the West Antarctic Ice Sheet may have been absent about 2 Ma ago with open ocean in the area where it is today (Scherer, 1991). Likewise, even the East Antarctic Ice Sheet, which has been considered to have been stable for tens of millions of years, may have been absent as recently as 3 Ma ago (Barrett et al., 1992), although that is in debate (Marchant et al., 1993). The questions arise as to the linkages between ice sheets and climate: Which drives what? Our best long-term record of climatic change comes from deep-sea sediment cores and the Milankovitch theory has become the established model, using these data, to explain climatic change during the periods of time when Earth has been in ‘icehouse’ periods through its history. However, re-evaluation, especially in light of the above information in combination with data from caves (Winograd et al., 1992) tends to indicate that Milankovitch forcing is not the only process operating (Chapter 2).

This Page Intentionally Left Blank

13

PAST GLACIOMARINE ENVIRONMENTS A. Elverhøi and R. Henrich 13.1. INTRODUCTION The origin and genesis of ancient (Pre-Holocene) glaciomarine sediments is still a subject of controversy. However, as information on modern environments has progressed, reinterpretations of ancient sequences have demonstrated that glaciomarine sediments are widespread (Anderson, 1983). Glaciomarine sediments are now well documented as an important constituent of all the major glacial episodes in the Early and Late Proterozoic, Paleozoic and Cenozoic (Menzies,1996, chapter 7). A basic problem in interpreting ancient glaciomarine sediments has been, and still is, the lack of reliable criteria for differentiating between marine and terrestrial glacial sediments. For open marine conditions, and especially for glaciomarine sedimentation in the deep ocean, knowledge is still limited. However, new marine sampling techniques have provided a large number of undisturbed deep sea sediment cores. Hence, it has been possible to combine conventional methods of glaciomarine facies analysis with micropalaeontological and geochemical data analysis to interpret palaeo-oceanographic, ecologic and glaciomarine processes. Furthermore, a high degree of stratigraphic resolution has become possible through the use of oxygen isotope stratigraphy and accelerator mass spectrometry (AMS) 14C dating (Menzies, 1996, 391

chapter 14). It is now possible to identify the basic elements of an interglacial/glacial cycle within shelf, slope and deep-sea sediments (Henrich, 1990). 13.2. PAST GLACIOMARINE SEDIMENTS: CLASSIFICATION AND IDENTIFICATION Progress in geological marine research has demonstrated that deep sea sediments represent an important source of information concerning the long-term glacial record. In high latitude areas, the deposition of the deep sea sediments is largely controlled by the glacial conditions onshore and on the continental shelf (e.g., ice-shelf regimes, grounded ice-margins, tidewater glaciers). It is crucial to understand the origin of these sediments in the Cenozoic record in order to relate them to the correct glacial regime. In the analysis of deep sea glaciomarine sediments, it is important to realize that many glaciomarine processes have daily and/or seasonal fluctuations and the progradation or regression of glacial environments can occur on a scale of years to tens of years. In addition, shifts in pelagic sedimentation may occur, but on a much longer time-scale. Although seasonal pulses in pelagic particle flux within polar oceans are the rule rather than the exception, the average pelagic accumulation rates are more continuous over longer periods (e.g., centuries or thousands of years); while

392

PAST GLACIOMARINE ENVIRONMENTS

glaciomarine sedimentation is generally characterized by fluctuating conditions with episodic peaks of deposition. As a consequence, flux calculations using a linear sedimentation rate for a mixed pelagic/glaciomarine lithology tend to overestimate the pelagic fluxes and underestimate the glaciomarine fluxes. Another problem is that stratigraphic data points may be displaced by bioturbation, especially in sediment units containing alternating glaciomarine- and pelagic-dominated lithologies (Henrich et al., 1989a). Problems in dating are highly relevant when using the deep sea record for deciphering glacial histories. Varying lithologies and facies must be well-dated if glacial events are to be correctly identified. Within the Cenozoic, a set of intercalibrated chronostratigraphic methods is available, in addition to the combination of bio- and magnetostratigraphy (Menzies, 1996, chapter 14). The following parameters have frequently been used for interpreting ancient glaciomarine sediments: dropstones, grain-size distribution, particle shape, surface texture, clast fabric, mineralogy/geochemistry, stratification/lamination, thickness and lateral extent, stratigraphic and facies relationships, fossil content, and geomorphological features. In reconstructing former environments, interglacial/glacial cycles need to be identified. In recording the composition and distribution of deep sea glaciomarine sediments, which are strongly related to the oceanic circulation, parameters reflecting changes in the water masses should be included, namely: carbonate production and dissolution, organic carbon and amorphous silica content, the oxygen/carbon isotope signal and sedimentation rate or flux. 13.2.1. Ice-rafted Detritus Deposits of pebbles that disrupt basal laminations, while being draped by overlying undisturbed laminae, are prime indicators of a glaciomarine environment, reflecting ice-rafting of coarser materials in combination with fall-out from suspended particles in the water column. In the study of deep sea glaciomarine sediments, the proportion of coarser materials (icerafted debris, IRD) present is a meaningful parameter. Variations in IRD concentration are commonly related

to changes in the glacial regime of the adjacent shelves. Well-dated cores from the last glacial period document that the IRD peaks at the early phase of deglaciation are associated with extensive iceberg influx (Fig. 13.1). Compared with the oxygen isotope record, which reflects global ice volume, the IRD parameter reflects more localized glacial events. A shift from high to low IRD content may reflect a change from an outlet glacier to an ice-shelf regime, or a change from calving outlet glaciers to a dominant source from ice margins entirely on land. A fundamental problem in interpreting icerafted material is in distinguishing between clasts rafted by sea ice and those rafted by icebergs (Chapter 12). Difficulties arise from the lack of reliable criteria for differentiating between the two modes of rafting. It is difficult to discriminate iceberg versus sea ice-rafted sand-sized material. Ice-rafted material must therefore be interpreted with caution. 13.2.2. Grain-size Distribution and Sedimentation Rate Glaciomarine sediments may range from coarsegrained diamictons to fine-grained deep sea muds with only a small percentage of sand. In the deep ocean the grain-size distribution varies in a cyclic pattern, apparently corresponding to interglacial/ glacial changes (Morris et al., 1985; Henrich, 1989). Coarse-grained units (diamictons) seem to correspond to periods of deglaciation and maximum extension of the ice sheets across the continental shelves, while fine-grained muds are characteristic of interglacial periods. The sedimentation rate, also, reflects interglacial/glacial changes. In deep sea sediments, beyond formerly glaciated margins, interglacial sedimentation rates are in the range 1–2 cm ka–1, while during glacials, rates increase to several tens of cm ka–1 (Jones and Keigwin, 1988; Henrich et al., 1989a). In open shelf regions (distal), typical interglacial rates range from 1 to 5 cm ka–1 (Elverhøi, 1984), while during glacials, rates are variable, with >10 cm ka–1 reported (Vorren et al., 1984). 13.2.3. Deep Water Formation During interglacial periods, deep water formation in high latitude areas plays a significant role in deep

PAST GLACIOMARINE ENVIRONMENTS

393

FIG. 13.1. Oxygen isotope, rock fragments and quartz content variations at ODP site 646 (Leg 105, Labrador Sea). Note that peaks in rock fragments and quartz grains correspond to high ␦18O values, e.g., glacial periods and cool phases during interglacials (modified from Wolf and Thiede, 1991).

water circulation. Deep water is formed in polar regions in two ways: (1) dense brines are derived from adjacent shelves as a result of salt rejection during sea ice formation and super cooling beneath the ice shelves; and (2) as a result of the convection of open ocean deep currents in regions with low density stratification. Dense shelf-water formation is also important for sediment erosion and transport on polar continental shelves and slopes. In Antarctica, the cold water forms coast-parallel currents that flow underneath large ice shelves, such as the Filchner and Ross Ice Shelves. During its sub-ice-shelf flow, the water temperature decreases (<–1.9°C). In the Weddell Sea, the outflow follows the western slope of Crary Trough (Fig. 13.2), with a water discharge rate calculated to be ten times that of the Amazon River (Foldvik, personal communication, 1991). Current velocities >100 cm s–1 have been measured at a depth of 2000 m on the slope in the Weddell Sea (Foldvik and Gammelsrød, 1988). This flow strongly influences sedimentation, resulting in erosional features tens of metres in incision depth on the

continental slope and rise (Fig. 13.3). During glacials, Antarctic glacier ice most likely extended to the shelf break, and the flow of such cold water stopped. Thus, a major change in flow regime occurred on the shelf and slope from interglacial to glacial periods. In the Barents Sea, dense shelf water forms during the winter season in response to sea ice formation (Midttun, 1985). This cold water episodically cascades into the adjacent deep sea regions. Sediment trap studies in the Norwegian-Greenland Sea show short, intensive episodes of sediment discharge into the deep ocean during the winter months, thought to be associated with such influxes of cold water (Honjo et al., 1988). 13.2.4 Circulation System/Current Regimes The interglacial transport of warm, saline water (and air) masses, for example, into the Norwegian-Greenland Sea and Arctic Ocean (Fig. 13.4), typical of today, was opposed during glacials by a general southerly movement of cold air and water masses.

394

PAST GLACIOMARINE ENVIRONMENTS

FIG. 13.2. Bathymetric map of the Weddell Sea showing the possible flow of the cold shelf water (ISW, ice shelf water). ISW flows along the bottom. The discharge is on average approximately ten times greater than that of the Amazon River (from Foldvik and Gammelsrød, 1988 reproduced with permission from Elsevier Science Publishers). (The location of the seismic section shown in Fig. 13.3 is shown by a bar). W 00.00

2-WAY REFL. TIME

1.2

LINE AN5-85

01.00

02.00

03.00

04.00

5 km

Time (Hours)

E 05.00 1.2

1.4

1.4

1.6

1.6

1.8

1.8

2.0

2.0

2.2

2.2

2.4

2.4

1.2

1.2

1.4

1.4

1.6

1.6

1.8

1.8

2.0

2.0

2.2

2.2

2.4 00.00

2.4 01.00

02.00

03.00

04.00

05.00

FIG. 13.3. Seismic section (sparker) along the upper continental slope of the Weddell Sea (see Fig. 13.2 for location), running perpendicular to the ice shelf water. The area is characterized by a ‘ridge and trough terrain’ where individual sediment layers are cut and seen in the walls. The erosion is thought to have been caused by the downflowing ice shelf waters (Solheim and Elverhøi, unpublished data).

PAST GLACIOMARINE ENVIRONMENTS

40º 30º 20º 10º

0º

10º

20º

30º

40º RE

NT

50º

395

R

PO

T

E

RR

CU ND ISLA

T

T REN

CUR

TIC

IAN AT L AN T

ON

NORWE G

A

P

R

O

C T I C

EA

L A R

R

500 0

10º

50

0º

tu re

Greenland Basin

h Ridge

75º

00

500 1000

y R id

ay en F

rac

Jan Mayen

lbe ins e

JMR

tur e

ge Rid s 1 n 5 h 00 Mo Lofoten Basin

Zon e

B

643 elf

644

tla nd

0

100

an

Sc

ia

av

n di

nt

e me ero rp Fa sca E

Re

yk ja

ne s

Ri

10

20º

A

NB

Sh e

dg e

1000

R

I-F

500 00

1000

sh

is Ax

ND

LA

60ºN 30ºW

642

Ex tin ct

.

E IC

0

ed ge

Ko es F.Z

Barents Sea

50

VP

Tjorn

65º

BJØRNOYA

re Zone

ge

shelf

edg e

20

tu Frac nja Se 1000

70º

Ja nM

HOPEN

Knipovic

ac Fr nd e nla Zon ee

0

20º

100

00

Gr

58º

62º

RD

BA

AL SV

ne

62º

10

F

15 2000 00

500

Yermak Plateau

Zo re tu ac 500 1

IR

g

D GAP

Fr 00 15 1500 en

LAN

1

NT

ICE

1000

00 er b it s Sp000

I

66º

20

ST EA

CE LA TYPICAL LIMIT OF ND THE ICE EDGE R CURRENT IC E G IN A SEVERE N CU MI WINTER RR E

66º

1500

LA

R

F N

ST

J A N M AY E

70º

ON

BEAR

TS SEA

CU

T

70º

PE CA T TH REN

PO

EN

NO R CU R

R

LAN

CU

EN

TYPICAL LIMIT OF THE ICE EDGE IN A SEVERE WINTER

LAR

ER

N

FR

ST

SB

R

GR EEN

IT

BA

74º

SP

EN

D CUR R O N

T

T

G

78º

ES

F RRRENT O N T

R EN T

W

EA

82º

UR S P I TSB E R G E N C

TYPICAL LIMIT OF THE ICE EDGE IN A MILD SUMMER

10º

0º

10º

20º

30ºE

FIG. 13.4. Bathymetric map (left) and map showing the surface currents (right) of the Norwegian-Greenland Sea and the Barents Sea (NB, Norwegian Basin; VP, Vøring Plateau; I-FR, Iceland Faroe Ridge; JMR, Jan Mayen Ridge).

396

PAST GLACIOMARINE ENVIRONMENTS

However, during glacial periods, a northerly flow of surface water (Atlantic water or northeast currents parallel to the ice margin) along the eastern margins of the Norwegian-Greenland Sea was maintained for long periods. The current velocity was strongly reduced, causing limited erosion of the shelf sediments and upper slope, as is typical today. 13.2.5. Oxygen and Carbon Isotopes in Glaciomarine Settings The oxygen isotope values measured on pelagic and benthic calcareous tests record a mixed signal of: (1) the overall characteristics of the water masses (temperature and salinity), and (2) global ice volume. An increase in the ␦18O record corresponds with either a decrease in temperature or an increase in global ice volume, or a combination of both. However, if we combine the ‘global ice volume signal’ of ␦18O with the IRD content in a deep sea core, a more local ice record can be deduced (Wolf and Thiede, 1991). As seen in Figure 13.1, an increased content of IRD corresponds with intervals of high ␦18O isotope content, that is, with glacials. The ␦13C signal of benthic and planktonic organisms is a reflection of much more complex processes. A high ␦13C signal in planktonic organisms records a rapid exchange between CO2 of the surface waters and the atmosphere, while low ␦13C values may indicate a more stratified water column, for example, caused by high meltwater introduction. In addition, a high productivity of marine organic matter preferentially extracts 12C from sea water, which results in even more positive ␦13C values of the calcite precipitated in equilibrium with the 12C-depleted sea water. Production of young deep water, for example, in the Norwegian and Greenland Seas is documented by positive ␦13C values of benthic foraminifers (Jansen and Erlenkeuser, 1985). ‘Aging’ of deep waters is recorded by negative 13C values of benthic foraminifers, which results from the incorporation of 13 C-depleted CO2 derived from oxidation of organic matter at the sea floor. In glaciomarine environments, interpretations should always consider the problems concerned with the distinction between effects of salinity and temperature on the planktonic oxygen isotope signal.

13.2.6. Carbonate Sedimentation in Glaciomarine Environmental Settings A classical controversy in the interpretation of ancient glacial sediments is the association of diamictites and carbonates (Menzies, 1996, chapter 7). The association of carbonates and diamictites is common in LatePrecambrian sediments. Carbonates are found as interbedded layers as well as forming the base or top of a glacial sequence. However, the facies association of carbonate-diamictite has also been interpreted as evidence against a glacial origin of the diamictites (Schermerhorn, 1974) because carbonates are traditionally thought to reflect warmer climates. Studies have shown that skeletal carbonate (from molluscs, red algae and barnacles) may well accumulate in cold climates (Bjørlykke et al., 1978; Freiwald et al., 1991). The frequent and extensive accumulations of carbonate on Arctic shelves were formed in response to the Holocene transgression. Such carbonates cap the underlying Weichselian/Wisconsinan glacial sediments as far north as 80°N. In these very high latitudes, 75–80°N, carbonate-bearing glaciomarine sediments are being deposited at the present day in an ice distal setting, and carbonates can therefore be regarded as an integral component of glaciomarine sedimentation (Elverhøi et al., 1989). In Antarctica, carbonate and, in particular, siliceous ooze, form the main constituents of shelf sediments in a number of locations (Dunbar et al., 1989). Similarly, as in the high Arctic, the biogenic sediments of Antarctica form a component part of the today’s glaciomarine sedimentation and are found as ice-proximal deposits. There is therefore no environmental conflict between carbonate accumulation and clastic glaciomarine sedimentation. In principle, the contrast between glacial and interglacial carbonate production in high latitude open sea environments can be illustrated by the modern carbonate production in the Norwegian-Greenland Sea (Figs 13.5 and 13.6(a),(b)). Atlantic water with no ice cover characterizes the eastern areas (Norwegian Current), while cold Arctic waters and permanent or semi-permanent sea ice (i.e., glacial conditions) describe the environments to the north and west (Fig. 13.4). Data from sediment traps in the Norwegian Sea (Fig. 13.4) (Honjo et al., 1988; Samtleben and

PAST GLACIOMARINE ENVIRONMENTS

397

FIG. 13.5. Vertical variations in the carbonate content of Quaternary sediments from the Norwegian-Greenland Sea. From the vertical section, carbonate peaks and high contents of subpolar planktonic foraminifera are found in interglacial periods (1, 5, 7 and 11).

Bickert, 1990) show decreasing pelagic carbonate production within the Norwegian Current along its path northwards. This gradient records the cooling of the Norwegian Current. Further to the north and west, in the Fram Strait and Greenland Sea areas (Fig. 13.4) (‘glacial’ conditions), the carbonate flux decreases to one-half to one-third of the total flux determined for the Norwegian Current. The main biogenic carbonate content of surface sediments under open ocean cold water conditions is formed by a monospecific assemblage of the leftcoiling foraminifera, Neogloboquadrina pachyderma. High reproduction rates of this species have been reported from Arctic and Antarctic sea ice environments (Spindler, 1990). Sediment trap measurements have shown coccoliths as an important carbonate source. Areas with almost permanent sea ice cover are barren of coccoliths, while areas with some open water have regular summer coccolith blooms, dominated by Coccolithus pelagicus. However, most of the coccolith carbonate is generally rapidly dissolved in the water column and at the sediment surface. Introduction of meltwater plumes into the pelagic

system has the effect of strongly reducing both foraminifer and coccolith abundance in surface waters, primarily through the resultant loss of light caused by particle-rich glacial runoff. Thus, carbonate flux calculations, compositional analysis and species abundances provide useful tools for the recognition of water mass characteristics in glaciomarine environments, and permit reconstruction of carbonate production during glacial and interglacial climatic shifts (Gard, 1988; Ramm, 1989; Henrich et al., 1989a; Gard and Backman, 1990). 13.2.7. Biogenic Opal Production and Preservation in Glaciomarine Environmental Settings Maps of the modern distribution of biogenic opal in surface sediments (Leinen et al., 1986) indicate a clear correlation between opal-rich sediments and zones of higher primary productivity in the equatorial divergences, coastal upwelling areas and circumAntarctic and circum-Arctic diatom ooze belt. In addition, phytoplankton production in the Arctic and

398

PAST GLACIOMARINE ENVIRONMENTS 50º

40º

30º

20º

10º

0º

10º

20º

30º

40º

50º

N 75º

75º

14-18

43

0

20

200

70º

24

35

15

40-42 50

30

20

20

16

33

65º

42 48

30-33

38

40

20 28

30 40

4 3 60 50 0 0 60

26

28 28

50 50

56

22

38

32 30

(a)

30

40

50

10º

40º

65º

45

W 0º E

Stage 5.5.1 (122 Ka) 50º

22

25 23

45

61

20º

35

50

45

53

70º 20 200

16 18

40 30

15 10

10º

20º

CaCO3 [ % ]

30º

20º

10º

0º

10º

20º

30º

40º

50º

N 75º

75º

0.7

0-0.1

0

20

200

70º

1

1

5

8-10 8-12

5

8-10

3-5

65º

1

3

2

0 0 2 0

3-5

2

1 1

5-10 5-10 10

8-10

70º

1

1

10

5

3-4

200

1

?

2

0-5 0

0

0 1 2

6-8

0

1

65º

0

5 0

20º

(b)

10º

W 0º E

10º

20º

Stage 6.2/6.1 (131 - 135 Ka)

FIG. 13.6. The horizontal distribution maps reveal that the highest carbonate contents can be seen on the eastern side of the basin, deposited under the warm Atlantic surface waters ((a), Eemian interglacial, OIS 5). During glacial periods ((b), OIS 6.2), generally much lower carbonate contents are recorded with a much narrower extension along the western and central regions, i.e., areas that have been less affected by ice-rafting and meltwater dispersion.

PAST GLACIOMARINE ENVIRONMENTS

Subarctic (Sakshaug and Holm-Hansen, 1984) is related to spring blooms in the vicinity of the ice edge. During spring melting of sea ice, nutrient-rich water develops at the surface, inducing a vigorous phytoplankton bloom in the wake of the retreating ice edge. In general, biogenic opal is rapidly dissolved within the water column and, in particular, at and/or below the sediment water interface, as a result of the undersaturation of sea water with respect to opaline silica (Broecker and Peng, 1982; Calvert, 1983). The fraction of opal preserved in the sediments varies sharply. In general, the amount of biogenic opal and its degree of preservation increases in areas with high silica fluxes (Broecker and Peng, 1982). Despite the fragmentary record of opal production in glaciomarine environments, specific opal occurrences have been used successfully in glaciomarine palaeo-oceanographic reconstructions. 13.2.8. Carbonate Dissolution Records and Early Diagenetic Reactions as a Response to Changing Bottom Water Properties in Glaciomarine Settings Bottom water properties and circulation patterns in deep sea basins with glaciomarine ice margins on their surrounding shelves are strongly influenced by glaciomarine processes. Dense brine formation during seasonal sea ice growth and open ocean deep convection episodically injects a young oxygen-rich and dense deep water into the basins, whereas the introduction of high quantities of meltwater into the open ocean can stabilize the water column and thus inhibit deep water circulation and exchange. Hence, it is important to record changes in bottom water properties and to evaluate the effect that glaciomarine processes contribute to these changes. Changes in the O2 and CO2 content of bottom waters can be deduced from studies of carbonate preservation. In addition, carbonate dissolution may strongly modify carbonate fluxes, especially in glaciomarine settings where carbonate production rates may be low. In addition to changes in bottom water chemistry, carbonate dissolution may indicate early diagenetic reactions. Hence, dissolution studies encounter two principal problems: first, the development of reliable methods that allow the quantification of dissolution; and second, distinguishing between dissolution in bottom water and in

399

porewater. Most parameters commonly used for dissolution studies in low- and mid-latitude oceans (e.g., the planktonic/benthic ratio, quantitative determination of the insoluble residue of pelagic carbonates and the percentage of dissolution-sensitive taxa) cannot be applied to polar and subpolar sediments because of large-scale variations between ecological conditions and sedimentary inputs of sources from glacial to interglacial stages (Henrich et al., 1989a). Conventional fragmentation indices of left-coiling Neogloboquadrina pachyderma, the only planktonic foraminifer species occurring in abundance in both glacial and interglacial sediments, and the more sensitive SEM-based dissolution indices (Henrich, 1986) of this species, allow estimation of dissolution on a semi-quantitative scale. 13.2.9. Ecologic Conditions and Environmental Reconstructions When reconstructing glaciomarine environments, the use of fossils and knowledge of their ecological requirements represent important parameters. Distinctions between paleoenvironments must occur on smaller as well as larger scales, including surface versus deeper waters and glacial/interglacial climatic shifts. Examples of changing conditions include: (1) advection and mixture of warm and cold, seasonally pack-ice-covered surface waters, (2) introduction of meltwater at tidewater ice fronts and melting sea ice/ icebergs, (3) the production of cold, dense brines on the shelves, and (4) deep convection in subpolar and polar seas. Micropaleontological methods, including statistical analyses of specific planktonic and benthic foraminifer assemblages, have now been developed in order to estimate physical properties such as the temperature and salinity of surface and deep waters; these are based on planktonic and benthic assemblages, respectively (Haake and Pflaumann, 1989). Diatom assemblages have been used to trace changes in surface water mass conditions during the Holocene (Koc Karpuz and Jansen, 1992). However, because of low sedimentation rates and strong diagenetic solution of biogenic silica, this approach cannot be applied in most older sequences. Benthic foraminiferal assemblages reveal pronounced glacial/ interglacial shifts related to changes in bottom water

400

PAST GLACIOMARINE ENVIRONMENTS

conditions (Streeter et al., 1982). Information on glacial/interglacial variations in deep water circulation can also be acquired from the Cd/Ca ratios of benthic foraminifers (Fig. 13.7) (Boyle, 1988). On the shelves, typical stratigraphic successions of macroand microfaunal assemblages outline changes in benthic habitat conditions, which are related to shifts in the sedimentary environment during the waxing and waning of continental ice sheets (Thomsen and Vorren, 1986; Vorren et al., 1989). 13.3. EXAMPLES OF PRE-CENOZOIC GLACIOMARINE SEQUENCES Commonly, interpretations of ancient sequences can be characterized by conflicting points of view. In the following, a well-known Pre-Cenozoic sequence is described and various interpretations discussed. The main objective is to show how the interpretation of diamictites has changed radically, from being ascribed

Cd/Ca (mmol/mol) 0.0

Approx. Time Scale Thousands of Years Ago

0

0.1

0.2

5

10 Cd/Ca 15 13C/12C

20

1

0 -1 d13C (o/oo)

FIG. 13.7. Variations in the Cd/Ca ratio of benthic foraminiferas in the North Atlantic. Relatively high Cd/Ca values correspond to glacial periods, e.g., lack of cold deep water formed from surface cooling (and nutrient depletion from the surface biomass) during these intervals (from Boyle, 1988; Broecker and Denton, 1989).

largely as terrestrial, to the present view of essentially a glaciomarine origin (a further example is discussed in Menzies, 1996, chapter 5, section 5.3.2; and chapter 7). 13.3.1. Gowganda Formation The Gowganda Formation of Ontario, Canada, is an Early Proterozoic deposit, with a probable age of 2.7–2.3 × 109 years BP (Young, 1981). The formation is primarily exposed in three locations close to eastern Lake Superior in Canada. The sediments are characterized by various types of diamictites (tillite deposits) and varved argillites with abundant dropstones. The coarser diamictites are normally structureless, while laminations are often seen in the mudstones. Sedimentary structures are generally well preserved, but folded and metamorphosed rocks are found locally (up to amphibolite facies). Typically, the individual members of the Gowganda Formation are discontinuous on a scale of kilometres. However, some members can be followed for tens of kilometres. The thickness of the formation normally varies between 300 and 1000 m, but thicknesses up to 3000 m have been reported. In general, the diamictite members are massive, while the interbedded argillites are laminated and contain IRD (Fig. 13.8). These laminated units also contain mudstone pellets. Previously, the formation was interpreted as a lodgement tillite with interbedded lacustrine varved argillites (Lindsey, 1969; Young, 1981). Glaciomarine sediments were thought to be present only in a restricted area. As depicted in the schematic stratigraphic section (Fig. 13.8), the diamictites represent ice advances, with the interbedded argillites as lacustrine ice-recessional deposits. However, in more recent studies, a subaqueous origin with predominantly glaciomarine deposition has been suggested, but there are still clearly different opinions on the depositional environment (Young and Nesbitt, 1985; Eyles et al., 1985; Mustard and Donaldson, 1987; Menzies, 2000). 13.3.1.1. Continental slope/channel fill Eyles et al. (1985) suggested the following depositional regime: (1) sediment gravity flow deposits are indicated by debris flow, sandy fluidized flow and

PAST GLACIOMARINE ENVIRONMENTS

thought to represent paleovalleys or deep water lacustrine environments. However, the lack of coarse debris along the margins excludes the valley hypothesis, and, owing to the large scale (400 × 500 km) and thickness, a lacustrine formation seems unlikely. According to Eyles et al. (1985), the Gowganda Formation was formed as an open marine deposit, and the lower section in the Elliot Lake region (southwestern exposure) is interpreted as the continental slope and submarine channel-fill depositional system of a glaciomarine environment.

LORRAIN FM.

no glacial ice

14

100 m

13

401

12 11

13.3.1.2. Resedimentation of glacially transported debris

10

4 3

2

1

SERPENT FM.

b a

no glacial ice ?

1

7 6 5

c

ice advance 2

8

ice advance

GOWGANDA FM.

9

FIG. 13.8. Schematic stratigraphic section through the Gowganda Formation. Two major ice advances are postulated (member 1 and 3 to 9), while members 10 to 14 are considered to be non-glacial and consist of two deltaic cycles. Members 2, 6, 9 and 11 represent laminated argillite. Black dots at the right side of the column represent dropstones in argillite sediments (from Young, 1981 and Young and Nesbitt, 1985).

Bouma-turbidites, indicating a marine environment; (2) the massive- to faintly bedded diamictites tens of metres in thickness are interpreted as IRD and distal mud, while some thinner units may represent debris flows; (3) no convincing evidence of any lodgement till formation has been found. In addition, long ‘fingers’ of the Gowganda Formation extending northward on the Archean basement were earlier

According to Young and Nesbitt (1985), the lower diamictite (Fig. 13.8) results from resedimentation of glacially transported debris in an actively subsiding region. The sediments were supplied from outlet glaciers entering the sea, probably as ice tongues. Subsequent glacial recession was followed by deep water sedimentation of fine-grained materials; however, clasts in its lower and upper units reflect nearby glaciers. The second ice-advance (Fig. 13.8) was responsible for deposition of the upper diamictite complex. Here, conglomerates and turbidites are widespread, and are thought to be related to ice advances and active tectonic subsidence. 13.3.1.3. Ice proximal regime Mustard and Donaldson (1987) suggested a submarine ice proximal depositional regime, at least for the lower part of the Gowganda Formation. They separate the lower member into the following four units: (1) a ‘basal diamictite’ unit overlain by (2) a coarsening-upward rhythmite/sandstone/clast-supported diamictite, which may grade laterally into either (3a) a stratified, or (3b) massive matrix-supported diamictite (‘fan’ or ‘interfan’ association, respectively). The top unit (4) consists of a complex series of diamictites (upper diamictite), including matrix-supported as well as clast-supported diamictites, with additional fine-grained non-cyclic, poorly laminated mudstones, containing IRD. Subglacial erosional features such as plucking and glacial lee-side quarrying at the top of the Archean–Proterozoic unconformity surface demonstrate the presence

402

PAST GLACIOMARINE ENVIRONMENTS

of grounded ice. Consequently, the ‘basal diamictite’, 1–12 m thick, is interpreted as a basal tillite, while stratified interbeds within this diamictite zone are explained as deposits in subglacial meltwater channels. The two overlying ‘fan’ and ‘interfan’ associations are interpreted as: subaqueous outwash deposits (‘fan’ association) and a more complex regime of icemarginal debris/morainal bank/debris flow/high-density sediment gravity flows (‘interfan’). Scattered dropstones reflect rafting from an adjacent calving glacier. The ‘upper diamictite’ is thought to represent deposition under the readvance of a floating or partly floating ice sheet. The lack of glaciotectonic structures or features reflecting erosion or incorporation of the underlying sediments suggests an ice-shelf depositional regime rather than an advancing grounded ice sheet. Recently, Menzies (2000) has used micromorphology techniques on the ‘basal diamictite’ and concluded that evidence would support a ‘till delta’ interpretation with the diamictite being regarded as evidence of subglacial deformation out into an ice proximal marine (grounding-line) setting (similar to the interpretation of ‘till deltas and tongues’). 13.3.1.4. Gowganda Formation – concluding remarks The study of the Gowganda Formation illustrates the recent paradigm shift in means of interpretation, from a solely terrestrial origin to an inferred glaciomarine depositional environment; the massive diamictites being interpreted as glaciomarine deposits, possibly resedimented. As corroborative evidence along the margins of the Norwegian-Greenland Sea, thick diamictites combined with large-scale slump and slide scars are found as a typical facies association for a glaciated margin. 13.4. GLACIOMARINE SEDIMENTS AND THE SEDIMENTARY ENVIRONMENT OF A PASSIVE MARGIN AND ITS ADJACENT OCEAN. THE NORWEGIAN-GREENLAND SEA AND THE NORWEGIAN/BARENTS SEA MARGIN – A CASE EXAMPLE The sedimentary record of the Norwegian-Greenland Sea and its adjacent margins offers a good opportunity for gaining insight into the development of glacially

influenced passive margins. Four major ice sheets terminate or have terminated on these margins: the British, Scandinavian, Greenland and SvalbardBarents Sea Ice Sheets. Deep sea drilling (ODP Leg 104), combined with extensive sediment coring and shallow seismic profiling, enable reconstruction of the sedimentary record and environment of the past 2.6 Ma and postulation of different explanations for the glaciated continental margins. (For a detailed discussion on the background of the Cenozoic Glaciation in this region the reader is referred to Menzies, 1996, chapter 5, pp. 195–196.) 13.4.1. General Aspects of the Late Cenozoic Glaciomarine Sedimentation of the Norwegian-Greenland Sea One of the most interesting results from ODP Leg 104 was the observation of a rhythmic occurrence of dark diamictons frequently intercalated within the glacial/ interglacial section of the past 2.6 Ma. This was the first time such layers had been discovered in deep sea sections. Glacial/interglacial sediments in the Norwegian-Greenland Sea reveal cyclic and rhythmic variations in biogenic and terrigenous composition, indicating corresponding changes in surface and bottom water properties. Interglacial sediments are characterized by high pelagic carbonate shell production (e.g., planktonic foraminifers and coccoliths), with high percentages of warm subpolar species reflecting inflow of warm Atlantic water. Glacial sediments indicate much lower carbonate shell production, dominated by colder water species. In addition, variable inputs of IRD and bulk terrigenous sediment supply are observed during glacial stages, with rhythmic deposition of dark diamictons. In general, the peak deposition of IRD material occurs during early periods of deglaciation. A number of environmental shifts (glacial/interglacial) are observed that may reflect changing conditions on land and in the sea. Interglacial deposits do not appear prior to 1.2 Ma at sites located offshore on the edge and slope of the outer Vøring Plateau (Fig. 13.4). They are restricted to a few occurrences at a high-sedimentation-rate site on the inner plateau close to the shelf. Low amplitude oscillations in IRD input with peak supplies in the dark diamictons are

PAST GLACIOMARINE ENVIRONMENTS

indicated throughout the period since 2.57 Ma. Considering all environmental parameters, the section from 2.57 to 1.2 Ma is characterized by high frequency and low amplitude oscillations of IRD input, a lowered salinity and, probably, a decreased carbonate shell production in surface waters as well as a very

20º

shallow lysocline. During this period, the northern continents were covered by ice caps of smaller dimensions than those of the last million years and the oceanic polar front was maintained within the northeastern North Atlantic and the Norwegian Sea. However, the overall cool climate in the Norwegian

30º

1

S V A L B A R D

50º FRANS JOSEF LAND

100 200 30 0

100

30

NORDAUSTLANDET

40º KVITØYA

00

0

80º

80º 300 20 0

300 200

400

0

KONG KARLS LAND

78º

30

200

100

BARENTSØYA

SPITSBERGEN

403

100

100 200

300

300

78º

EDGEØYA

STORBANKEN

100

0

10

BA

EN

300

300 200

200

76º

S

PI

200

TS

BE

76º

HOPEN

RG

A 100 NN RE 300 RD FJO R STO

NK

EN

200

SENTRALBANKEN

20 - 78 100 200 300 400

00

10 00 500

15

BE

R IS

LAN

74º

DT

RO

0

UG

H

74º

400

BEA 72º

A

RI

N SLA

20

D FA

N

BEAR ISLAND

7200 - 73/77

40

20º

GR

EEN

LBARD

0 30 00 2 0 10

0

300

S VA

NOVAJA 72º 100 SEMLJA

300

N O R W AY

30º

40º

100

50º

S NT RE A A B SE

LAND

NO RW AY IC E L A

ND

FIG. 13.9. Bathymetric map of the Barents Sea (100 m contour interval), including location of seismic profiles.

404

PAST GLACIOMARINE ENVIRONMENTS

Sea was episodically interrupted by weak intrusions of warmer Atlantic waters along the eastern margin. The last 1 Ma is characterized by a higher supply of IRD and bulk sediment. Large amplitude variations characterize the IRD and carbonate records, indicating growth and decay of huge ice masses in the northern hemisphere.

glacial advances and iceberg ploughing. In the Barents Sea, only a thin cover (<15 m) of sediment is found in intermediate and shallower areas (50–300 m water depth). On the deeper parts of the shelf and towards the slope, thick prograding glaciomarine sediment wedges, which were formed during successive glacial advances, terminate at or close the shelf edge (Fig. 13.9). The entire shelf was probably covered by grounded ice during the last glaciation (Vorren et al., 1988) and today (and during the Holocene) the northern and central parts of the area are influenced by sea ice and iceberg deposition.

13.4.2. Barents Sea/Northern Norwegian Shelf The glaciomarine shelf record may be incomplete or represent disturbed sequences because of repetitive

12º 0

1.0

13º

14º

6.0

Two-way Travel Time (s)

E G G A

20º 0.0 1.0 2.0 3.0 RLFC

4.0 5.0

LINE 7200-73/77

SENJA FRACTURE ZONE

7.0

B A S I N

W 90 km

TROMSØ BASIN

6.0 7.0

E 50 20 - 78 (Storfjordrenna - original data) Bedrock

0.5

E

19º

TeA

TeC TeB

5.0

0 Acoustically transparent deposits

Bedrock

0.9

W Two-way Travel Time (s)

18º

TeD

4.0

(c)

17º

TeE

3.0

(b)

16º

20 km

2.0

(a)

15º

SEC. TWT

0.0

URU

W 11º

LINE 20 - 78

E

50

0

90 km 20 - 78 (Storfjordrenna) - interpreted section 0.5

0.9

IA IB

II III

a d

d

IV

Bedrock (URU = upper regional unconformity)

URU

FIG. 13.10. Interpreted seismic sections from the Bear Island Trough (a) and Storfjordrenna (b) and original data from the latter area (c). (for location, see Fig. 13.9). (a) Schematic profile showing the main structure of the fan complex outside the Bear Island Trough (from Vorren et al., 1991). Unit TeE and TeD are characterized by chaotic seismic reflection pattern. According to Vorren et al. (1991) unit TeE and TeD represent Late Cenozoic glacigenic sediments, while Eidvin and Riis (1989) also include unit TeC into the Late Cenozoic deposits. Units TeB and TeA represent probably Middle and Early Tertiary deposits. (b, c) The glacial nature of the sediments is clearly seen from the large-scale ridge features forming up to 50 m local highs, and a number of erosional uncomformities, which reflect periods of glacial advance. Four regionally correlative reflectors have been identified: a, b, y and d, while I, II, III and IV define four seismic sequences. Acoustically transparent deposits in front of a bedrock threshold are interpreted to represent glaciomarine sediments deposited in front of a glacier that had its front pinned at the threshold (from Solheim and Kristoffersen, 1984, reproduced by permission of the Norsk Polarinstitutt).

PAST GLACIOMARINE ENVIRONMENTS

The southwestern Barents shelf and slope reveal a variety of glaciogenic units typical for passive polar margins. On the outer shelf up to 300 m of stratiform glaciogenic sediments overlie an upper regional unconformity (URU) with a glacially eroded morphology (Fig. 13.10). In the central and shallower regions (300–100 m water depth), the sediment thickness is generally <15 m. However, local sediment accumulations of 30–50 m in thickness are found as moraine ridges and ice proximal sediments or acoustically transparent deposits (Profile 20–78 in Fig. 13.10). The ice proximal sediments, which may cover up to 2000 km2, are found in the inner part of major embayments commonly at a water depths of 300 m (present day). These accumulations may have formed from settling of fine-grained materials from turbid surface plumes along the ice margin. The ice proximal deposits, at 300 m water depth, appear to represent a major halt in the recession of the Barents Sea Ice Sheet ~12–15 ka. At the outer shelf, the sediments are comprised of 艋150 m-thick seismic units that are separated by erosional surfaces (Fig. 13.10). Various depositional environments have been inferred, based on the

78 - 82

TWT (sec.)

0.4 0.5 0.6 0.7

0

SSE 78 - 82

TWT (sec.)

0.4

1

2

3

4

5 km

NNW

DEPTH (m) 300 350

0.5

5E

0.6

4E?

4E? 0.7

1E

FIG. 13.11. Examples of interpretations of acoustic character and sediment type from glacigenic sediments in the outer part of the Bear Island Trough, 5E, acoustically transparent/semi-transparent, e.g., glaciomarine sediments. 4E, chaotic reflection pattern, e.g., till deposits (from Vorren et al., 1989).

405

INTERGLACIAL NORW. CURRENT WINNOWING

NORW. COASTAL CURRENT

GULLYING

w

E

GLACIAL

DEBRIS LOBES

Y IT AV GR

FL

OW

PROGRADATION

FIG. 13.12. Generalized model for glacial/interglacial sedimentary regimes on a passive margin. Glacials are characterized by wedge progradation while interglacial periods are associated with currentwinnowing and gullying (from Vorren et al., 1989, reproduced with permission from Elsevier Science Publishers).

seismic signatures (Fig. 13.11) including glaciomarine sediments (semi-transparent signature), submarine outwash fans (diverging stratified signature), glaciomarine/marine trough fills (stratified onlap pattern) and till (chaotic reflection). On the upper slope and at the shelf margin, thicknesses of 1000 m are reported, with a thickening towards deeper water. Slope sediments are composed of prograding sequences with a complex sigmoidal-oblique character. Glacial periods are characterized by progradation and build-up, while during interglacials, sediments were removed from the continental shelf and transported into the deep ocean, bypassing and eroding the upper slope sediments (Fig. 13.12) (Vorren et al., 1989). Owing to glacio-isostatic rebound, a shallowing of 60–100 m has taken place in the central and northern Barents Sea in the last 10 000 years. The sedimentary section on the shallow Spitsbergenbanken (Fig. 13.9) (30–80 m water depth) reflects these conditions in a typical regressive facies succession (Fig. 13.13(a)). This sequence is comprised of Late Weichselian till, overlain by glaciomarine sediments, capped by a residual diamicton, with a final layer of approximately 0.5 m polygenetic carbonate-rich sediments. In the Barents Sea, the carbonates are interpreted as postglacial. Icebergs and also sea ice pass over this

406

PAST GLACIOMARINE ENVIRONMENTS

(a)

Shallow banks (Spitsbergenbanken, 30-80m water depth) Lithofacies

Thickness

Reworked coquina/sand Residual diamicton/ gravel lag Diamicton with rare molluscan infauna

0.5m 0.5-1m

Overconsolidated diamicton/lodgement till

2-8m

4-8ka

2-5m

~10ka

Grounding line retreat

Deep banks (200-300m water depth) Lithofacies

Thickness 0.1-0.3m

Mud Homogeneous/ interstratified diamicton, dropstone mud

3-15m

Overconsolidated diamicton/lodgement till

3-10m

Events Increase in bottom current vigor Ice-isostatic rebound Glacial flutes Grounding line retreat Barents Sea Ice Sheet lodgement

Bed rock

(c)

~9ka

Barents Sea Ice Sheet lodgement

Bed rock

(b)

since 3ka

Events Hard-substrate epifauna on coquina/gravel lag Ice-isostatic rebound, increase of bottom currents Ingression of marine molluscan infauna

Deep banks (>300m water depth) Lithofacies Resuspended muds Dropstone-muds

Thickness

Events

0.3-1m 1-2m

Re-entry of Atlantic waters Ice-isostatic rebound

Overconsolidated diamicton (till/dropstone muds)

Lodgement till Bed rock

10-50m

?

Grounding line retreat Barents Sea Ice Sheet lodgement

FIG. 13.13. Regressive glaciomarine sediments formed in response to glacioisostatic rebound in the central Barents Sea (from Bjørlykke et al., 1978; Henrich, 1990; Elverhøi et al., 1990; Hald et al., 1990).

PAST GLACIOMARINE ENVIRONMENTS

region, and coarser IRD has been dropped and mixed into the postglacial sediments. In intermediate and deeper parts of the central and northern Barents Sea (200–300 m water depth), the sediments have not been affected by the postglacial erosion, and the Late Weichselian glaciomarine sediments also grade gradually into fine-grained Holocene mud (Fig. 13.13(b)). In deeper troughs, such as the Bear Island Trough (300–500 m water depth), the 50–100 m thick sediment sequence consists of relatively homogeneous diamicton of more than one depositional event (Fig. 13.13(c)) (Hald et al., 1990). The sediments are overconsolidated and a till origin has been proposed; however, the presence of foraminifers, typical of a glaciomarine environment, show that parts of the till deposits are derived from glaciomarine sediments. The glacially reworked nature of these sediments is seen also from shearing, folding and fissile textures. The glaciomarine sequence on the shelf off northern Norway demonstrates the various facies formed during deglaciation of a shelf region and their distribution. Here, it is seen how the shelf topography has controlled the facies distribution. Compared with the sequence from the central Barents Sea, the sequence off northern Norway shows a

Trough >300m

Laminated clay (B) Basal till/ diamicton (A)

(ri

ine mar ted) lacio a

Basal till/ diamicton

a 10 k istal g gen oxy D na, u a tic f (arc

ka gen13 stricted and oxy na) u (re nthic fa e b d te deple (arctic benthic fauna)

Seasonal sea ice cover

(restricted arctic benthic fauna)

Grounded ice sheet (reworked fauna)

Shallow bank <150m

Ice free Atlantic waters

Foraminiferal sand

Proximal glaciomarine

Sandy diamicton

10ka

Grounded ice sheet/ proximal glaciomarine (iceberg turbate)

Legend

Proximal glaciomarine 2m

Dropstone mud (*C)

fauna) ch boreal

1

Sandy dropstone mud (D)

Foraminiferal sand Lag gravel Sandy diamicton

Ice free Atlantic waters 7.8 ka

1 basal till/overconsolidated diamicton, including scattered reworked shell fragments, indicates lodgement and/or deposition close to a tidewater glacier front (Stage A); 2 laminated clay with a restricted Arctic benthic fauna and scattered dropstones indicates seasonal sea ice cover during a stillstand in the ice recession (Stage B); 3 dropstone-rich mud with a rich bottom fauna indicates readvance of the ice front, with frequent iceberg rafting. During this period the sediment on bank areas was intensively reworked by iceberg ploughing (Gipp, 1993; Menzies, 1996, chapter 4), forming typical iceberg turbates (accumulations of sediments reworked by icebergs) (Stage C). A more restricted fauna towards the upper part of the dropstone-rich mud may indicate less saline and less oxygenated water masses; 4 sandy dropstone mud, including sand lenses both on banks and in the troughs, demonstrates an open marine environment with icebergs and intensified

diamicton

lag gravel

dropstone mud

laminated clay

sandy dropstone mud

foraminiferal clay

sand lenses

foraminiferal sand

0

Foraminiferal mud (E)

minor glacial readvance (Figs 13.14 and 13.15), which results in a more varied lithostratigraphy than that found to the north. Based on the sediment and microfossil composition, the following scenario has been constructed (Fig. 13.15):

Deep bank >150m

Approx. scale

Foraminiferal sand (F)

407

FIG. 13.14. Sediments during the last deglaciation on the continental shelf off northern Norway (from Vorren et al., 1984; Henrich, 1990).

408

PAST GLACIOMARINE ENVIRONMENTS

current activity – indicating the intrusion of Atlantic water and withdrawal of polar water masses to higher latitudes (Stages D, E); 5 similarly, as in the Barents Sea, carbonate-rich sediments cap the diamictons in the bank areas (Freiwald et al., 1991). However, in contrast to the

accumulations on the Spitsbergenbanken, Barents Sea, the carbonate accumulations off northern Norway are not part of a glaciomarine environment. There are no sea ice or icebergs off northern Norway and the water masses all have temperatures well above 0°C (Stage F).

(a)

(d)

(b)

(e)

(c)

(f) Proglacial glaciomarine

Basal till

Iceberg turbate

pelite

Sandy pelite

Sandy diamicton diamicton (d);

Pebbly

Laminated clay

sediments

Gravelly, shelly

Lag deposits (e, f)

Pelitic calcareous sediments careous sand and silt

Cal-

Winnowing

FIG. 13.15. Sedimentary environments during the last deglaciation on the continental shelf off northern Norway (from Vorren et al., 1984, Henrich, 1990).

PAST GLACIOMARINE ENVIRONMENTS

13.4.3. The Continental Slope The slope system is the main depositional centre for glacially eroded sediments. Sequences 2–3 km-thick have been reported on the margins off Svalbard and the Barents Sea (Myhre and Eldholm, 1988; Vorren et al., 1991) (Fig. 13.10). Compared with lower latitude slopes, the polar slopes have been less thoroughly studied, and any understanding of the sediments and the sedimentary regime is still fragmentary. However, typical glacial/interglacial sediment cycles and settings have been identified (Vorren et al., 1998). Along the margins of the Norwegian-Greenland Sea, the slope outside the Barents Sea Trough has been the most studied (Vorren et al., 1998). A prominent feature of the continental slope is the presence of well-defined ‘fans’ or ‘aprons’, which debris flows, stacked together, form the primary building blocks (Fig. 13.16). The dimensions of the debris flows along various parts of the margin vary considerably but generally range from 0.5 to 40 km in width, 5–60 m thickness and <10–200 km in length (Vorren et al., 1998). The generation of debris flows has been related to peak glacial periods with ice resting on the shelf edge. It is postulated that ice advance leads to rapid deposition of sediments at the shelf break, probably as deformation till (Laberg and Vorren, 1995), resulting in the formation of ‘till deltas’ (Alley et al., 1989b) or ‘diamict aprons’ (Hambrey et al., 1992). The high sedimentation rate causes buildup of excess pore-pressure in these deposits and causes them to oversteepen, making them unstable

MORAINE RIDGE

O

U

MARGINAL MORAINE

E

TR G H MO FA UTH N

PALIMPSEST LAG

SLID

ICEBERG FURROWS

GU

LL

IES

TROUGH FILL STRATIFORM SHELF DIAMICTONS UPPER REGIONAL UNCONFORMITY

DEBRIS LOBES SLOPE SYSTEM

SHELF SYSTEM

FJORD SYSTEM

FIG. 13.16. Morphology, centres of deposition, and characteristic sedimentary features on the Bear Island Fan (Fig. 13.9) (from Vorren et al., 1989, reproduced with permission from Elsevier Science Publishers).

409

and ultimately leading to their downslope movement as debris flows (Dimakis et al., 2000). The sediments are found to be rich in clay, which favours a plug flow mechanism whereby there is no internal movement in sediments during flow. Many of these debris flows show large run-out distances on low gradient slopes. Vorren et al. (1998) suggest the run-out distance to be controlled by size of the feeding slide and/or slope gradient. Recent laboratory experiments have however shown that the long run-out distances could also be due to hydroplaning below the moving sediment layer. As a plug moves down, its head is lifted as a result of resistance of the standing body of water that allows a layer of water to penetrate below the moving mass detaching it from the sea floor. This reduces basal friction and allows mass flows to cover large distances, even on low gradient slopes (Elverhøi et al., 1990). During interglacials, the most pronounced cases of down slope transport, particularly may be caused by the cascading of cold, saline shelf water (Midttun, 1985). The margin is also characterized by huge slide scars. These sediment failures seem not to be directly correlated with interglacial/glacial cycles. However, as the main body of the sediments displaced during the slide is of glacigenic origin, such processes are of importance in understanding the facies/sediment distribution of glacially influenced margins. The Grand Banks slide of 1929 (Heezen and Ewing, 1952) was triggered by an earthquake, and a similar mechanism was proposed for one of the world’s largest submarine slides, the Storegga slide on the mid-Norwegian margin (Jansen et al., 1987; Bugge et al., 1988). On the basis of coring and acoustic profiling, it was possible, in the latter case, to identify three main slide events, covering 34 000 km2 (Fig. 13.17). The total volume of sediment involved is ~5500 km3. The average gradient of the whole slide scar is about 0.6°. The oldest and largest slide (~4000 km3 ), probably of a Middle Weichselian age, consisting mainly of a debris flow of uncompacted clay of Plio-Pleistocene age. The second and third slides, both from Holocene time, also involved Neogene and Palaeogene sediments. Huge sediment slabs, 200 m thick and 10–30 km wide, were transported up to 200 km downslope to 2000–2500 m water depth. Similar, although

410

PAST GLACIOMARINE ENVIRONMENTS

ICELAND

0

100

200

km NORWAY

LAST SLIDE SECOND SLIDE FIRST SLIDE

FIG. 13.17. (a) Bathymetric map of the Norwegian Sea showing the areal distribution of the Storegga slide (a) and a longitudinal section through the slide scar (b) and pre-slide reconstructions (c). The samples 49–21 and 49–23 (18B) contain debris flow deposits with lumps of Eocene–Oligocene sediments. Q, base of Pleistocene; P, base of Pliocene; MO, mid-Oligocene; Pal, Paleocene–Eocene; B, Tertiary basalt. Note that the slide deposits are shaded. In the cores, D, debris flow deposits; T, turbidite; Pe, pelagic post slide deposits (modified from Jansen et al., 1987; Bugge et al., 1988).

PAST GLACIOMARINE ENVIRONMENTS

somewhat minor, slides have also been recorded further north along the Norwegian continental margin (Vorren et al., 1998).

411

was deposited at the mid-Norwegian margin (e.g., Vøring Plateau) during the past 2.6 Ma. 13.4.4.2. Facies succession

13.4.4. The Deep Sea Environment 13.4.4.1. General background Rhythmically alternating continuous seismic reflection patterns characterize wide areas of the sediment cover on the sea floor of the Norwegian-Greenland Sea. Undisturbed sediment cores reveal typical glacial/interglacial sedimentation patterns that can be correlated over distances of several hundred kilometres. Individual glacial units can be traced continuously from the central sector of the NorwegianGreenland Sea to the adjacent continental margins, where most expand in thickness. As a result, a thick pile of glacial/interglacial sediments (250–500 m)

The sediments deposited predominantly from the interactions of pelagic sedimentation and glaciomarine processes record a spectrum of glacial and interglacial sediment facies. Each facies is characterized by a set of sedimentological, geochemical and micropaleontological features. Rhythmic and cyclic repetitions of sediment facies types dominate the stratigraphic record illustrative of glacial/interglacial climatic shifts (Figs 13.18 and 13.19). A noticeable feature of these cyclic variations is the pronounced colour changes between light and dark layers. The layers are of greatly varying thicknesses (1 m for light layers and centimetres to tens of centimetres for dark layers). The light sediment

Q 500

0

50

100 km

Previous sea floor

1500

2500 m

P MO B

Pal FSE

(b)

18 x vert. exaggeration

Q

P MO Pal

B FSE

(c) FIG. 13.17. (b) and (c)

412

PAST GLACIOMARINE ENVIRONMENTS

facies are comprised of two prominent lithologies (Fig. 13.19). Brownish, intensively bioturbated foraminiferal muds and foramnanno oozes (Facies A; Henrich et al., 1989a) have been deposited during interglacials at low sedimentation rates (1–2 cm ka–1 ). The second facies unit consists of brownish (Facies B) and grey (Facies C), moderately bioturbated, sandy to silty muds, with intermediate to low carbonate content and variable mixtures of coarse terrigenous debris and reworked organic carbon, deposited at moderately high sedimentation rates (2–5 cm ka–1 ). The scattered occurrence of dropstones and enrichment of sand-sized IRD along discrete layers point to episodic input of coarse terrigenous particles from icebergs or sea ice. Three types of dark diamictons (Facies D, E and F) (Henrich et al., 1989a) are intercalated with the light-

coloured glacial sediment packages. These dark diamictons occur at thicknesses of centimetres to tens of centimetres. Most layers reveal a sharp base and gradational top contact often modified by bioturbation. Variably sized dropstones and scattered mud clasts in a sandy to silty mud matrix are the predominant lithofacies. Occasionally, enrichment by coarse debris is observed along discrete bands. Also, these diamictons have suffered early diagenesis, probably a result of changing bottom water conditions. Most spectacular with respect to diagenetic overprint is the diamicton facies type F. An example of diagenetic overprint is sulphate reduction processes in porewaters during stages of incipient burial followed by downward protrusion of a secondary oxidation front during later stages, with an approximate time lag of a few thousand years. It forms a complex layer consisting of

MAJOR DEGLACIATION SEQUENCE Lithofaciestype A B

interglacial late deglaciation

F

late glacial/ early deglaciation

C

glacial

Mud - dropstone rich diamicton, dark grey transitional to dark olive grey

light coloured colored (grey, light (grey, brownish brownish grey) grey) foraminiferalmud mud foraminiferal light (grey, brownish) brownish) bioturbabioturbalight coloured colored (grey, ted ted sandy sandy mud mud (abundant (abundantdropstones) dropstones) dark olive grey bioturbated sandy mud dark olive grey laminated (Fe) sandy mud (abundant dropstones) very dark grey to black mud to sandy mud with abundant lithic and mud dropstones dark grey mud with few scattered dropstones O2-rich ICE RAFTING

CARBONATE PRODUCTIVITY

CARBONATE DISSOLUTION

O2- CO -rich 2 depleted

BOTTOM WATER CIRCULATION

FIG. 13.18. Characteristic facies succession formed in the Norwegian-Greenland Sea during major deglaciation events (from Henrich, 1990, reproduced by permission of the author). The sediment sequence is also illustrated by radiographs.

PAST GLACIOMARINE ENVIRONMENTS

413

Dark olive grey mud - dropstone rich diamicton

ICE RAFTING

C

glacial

E

minor deglaciation

C

glacial

CARBONATE PRODUCTIVITY

CARBONATE DISSOLUTION PEAK

BOTTOM WATER CIRCULATION

dark grey mud with few scattered dropstones dark olive grey laminated (Fe) sandy mud with abundant dropstones dark grey mud with few scattered dropstones O2-rich

O2depleted

CO2-rich

FIG. 13.19. Characteristic facies succession formed in the Norwegian-Greenland Sea during glacial periods, with the continental ice sheet close to the shelf break (from Henrich, 1990, reproduced by permission of the author). The sediment sequence is also illustrated by radiographs.

a basal very dark grey diamicton grading upwards into a dark olive grey diamicton. All three dark diamictons reveal very low carbonate contents (~0.3 per cent CaCO3 ), strong dissolution features on the planktonic foraminifera, high organic carbon values, and high content of sand-sized terrigenous debris, densely scattered dropstones and abundant sediment pellets. Rock fragments and large dropstones consist of various igneous, metamorphic, sedimentary lithologies and coal fragments. The diamictons may contain both ice-rafted chalk and coal particles. Since chalk is only exposed to glacial erosion along shallow subcrops of the southern Norwegian shelf (Bugge et al., 1984), or in the North Sea region and its adjacent continental margins in southern Scandinavia and Britain, a northward drift of icebergs is indicated.

13.4.4.3. Environmental interpretations Environmental parameters associated with the rhythmic facies successions are shown in Figure 13.20. The most prominent rhythm is the facies succession C–F– B–A, which represents a complete shift from full glacial to interglacial conditions. Facies C was characterized by a low carbonate shell production and some ice-rafting in the surface waters, and was deposited during an advance of the continental ice sheets onto the shelf. The offshore situation was characterized by a seasonally varying sea ice pack with a few drifting icebergs. Facies F represents a dramatic increase in icerafting and a strong decrease in carbonate shell production in surface waters. Facies F documents the maximum extension of the ice sheet close to, at, or

414

PAST GLACIOMARINE ENVIRONMENTS

(a) Facies C Advance of continental ice out to the shelf

Norwegian Shelf

Vøring Plateau

(b)

Facies F Ice sheet close to the shelf break during glacial maximum and its early retreat late glacial / early deglacial time

low salinity surface water lid resuspension of Corg-rich sediments rapid ice-berg drift, deposition of mud and lithic dropstones

(c) Facies B Rapid deglaciation on the shelf and intrusion of Atlantic water

13.5. CONCLUSION

low salinity surface lid (reduced) Norwegian Current

even below the shelf edge, in addition to an early period of ice-sheet retreat. During this period, largescale calving or frequent surges along the tidewater front of the grounded ice sheet delivered huge amounts of debris-laden icebergs into the open sea. In addition, large meltwater discharges may have formed a low-salinity lid on the surface and deposited finegrained sediment from sediment-laden plumes. Facies B is characterized by a gradual decrease in ice-rafting and an increase in carbonate shell production. Facies B represents the retreat of the ice sheet on the shelf and major deglaciation, characterized by a rapid retreat of the low-salinity surface water lid towards the coast. A rapid rise in sea level may have caused a sudden disintegration of the marine-based parts of the continental ice sheets. Facies A records a high interglacial carbonate productivity and completely oxygenated bottom waters (a rich benthic fauna). Other rhythmic facies successions commonly observed within certain glacial stages (C–D–C, or C–E–C, or C–F–B–C) indicate similar environmental changes but without a return to or initiation of full interglacial conditions. Such a situation might occur at any time when a grounded ice sheet on the shelf became unstable. The release of enormous masses of icebergs might also be triggered by glacio-isostatic processes (strong subsidence of the shelf responding to increased ice-loading on the continent might have destabilized grounded ice on the shelf).

high input of ice rafted debris occasional input of ice rafted debris

FIG. 13.20. Paleoenvironmental model of the Norwegian Sea and adjacent shelf. (a) Extensive glaciation, with the ice margin located at the shelf break, associated with offshore pack-ice drift. (b) Early or late glacial period, with intensive iceberg calving from grounded outlet glaciers. (c) Intrusion of the Norwegian Current and reduced iceberg drift, combined with increased carbonate production in open sea areas. The ice margin is now located within the present day coastal areas (from Henrich, 1990, reproduced by permission of the author).

The importance and widespread nature of glaciomarine sedimentation within the geological record has meant a re-evaluation of global glacial sedimentology. In the past, largely owing to lack of knowledge, glaciomarine sediments were relegated to a peripheral acknowledgement. With studies now from almost every continent of Pre-Pleistocene diamictites and associated facies, it is apparent that glaciomarine sedimentation is of critical importance if we are to understand global glaciation both today and in the past. The association of glacial sediments with carbonate-rich facies, the switches from facies types indicative of glaciation to cold water interglacial sediments,

PAST GLACIOMARINE ENVIRONMENTS

the increasing awareness of ice margin dynamics and how they impact upon and correlate land-based ice fluctuations with related changes offshore, and the fact that ocean basins are our greatest surviving depocentres, can only further underlie the significance of these glacial environments and sediments. Rather than a passive environment, glaciomarine environments experience enormous and, at times, sudden changes of great magnitude in the form of subaqueous debris slides, gully incision and related glaciogenic

415

debris transportation over vast areas of the ocean floor. It is only too apparent that knowledge of glaciomarine processes and associated sediments is still at an early stage of our understanding, with data collection and ocean exploration still to some degree in its infancy. With petroleum exploration and other human advances in the use of oceans, much applied work still needs to be done in understanding and utilizing those environments and sea-bed sediments that are glaciomarine in origin.

This Page Intentionally Left Blank

14

PROCESSES OF GLACIOTECTONISM F. M. Van der Wateren

In the vast range of processes operating in a glacial landscape, glaciotectonic processes produce structures and landforms that may be used to reconstruct environmental changes long after the ice has vanished. Push moraines can be reliable indicators of what went on at a glacial margin. Together, sedimentology, paleohydrology and structural geology can give information about the mass balance of the glacier or ice sheet. Subglacial structures can give an insight into the conditions at the glacier bed and indicate former ice flow directions. In this chapter current ideas on processes operating both beneath modern glaciers and at their margins are applied to explain quite distinct styles of glaciotectonic deformation that can be found in a glacial landscape. A description of the landforms alone will rarely be sufficient to identify the glacial regimes that produced them. Without a proper understanding of their internal structures the information gained from a geomorphological analysis may be quite ambiguous. Emphasis in this chapter lies on tectonic style and strain, starting from the assumption that different glacial regimes are reflected in certain characteristic glaciotectonic styles and widely varying amounts of finite strain. For comprehensive bibliographies on glaciotectonism the reader is referred to web sites: http://www.emporia.edu/earthsci/biblio/biblio1.htm, & /biblio2.htm, & /biblio3.htm. 417

14.1. INTRODUCTION A glacial environment offers many opportunities for deformation of unlithified and weakly lithified sediments. Deformation of the glacier bed, bulldozing of sediments at the margins, ploughing of the sea floor by calving icebergs, debris flows creeping down the glacier snout, these are but a few examples. An extensive literature on glaciotectonics demonstrates that these deformations are analogous to structures found in many orogenic belts, only different in scale. Strain rates in hard rocks are several orders of magnitude smaller than those in unlithified sediments. While fold and thrust belts in crustal collision zones require millions of years to form, the timescale for the formation of even the largest Quaternary push moraines may vary from tens to hundreds of years. Recent studies of modern environments suggest that quite large push moraines may form even within the few years of a glacier surge. Compared with orogenic mountain chains, glaciotectonic deformations are quite shallow. They rarely exceed 200 or 300 m in depth. Yet, although push moraines are several orders of magnitude smaller than mountain chains, they may form fold and thrust belts spanning half a continent. Moreover, shear zones in sediments beneath midlatitude ice sheets are comparable in size with crustal shear zones. The ice sheets themselves, deforming

418

PROCESSES OF GLACIOTECTONISM

under their own weight, are continent-size metamorphic tectonites. Since its introduction by Slater (1926) several definitions of the term ‘glaciotectonics’ have been proposed, most of which differ only in detail (Moran, 1971; Woldstedt and Duphorn, 1974; Banham, 1975; Rotnicki, 1976; Ruszczynska-Szenajch, 1980; Aber, 1985; Drewry, 1986). It is known under many names, varying from glacier tectonics and ‘Eistektonik’, to ice thrusting. Glaciotectonics involves structural deformations of the upper horizon of the lithosphere caused by glacial stresses. Glaciotectonism refers to the processes leading to these deformations. Their detachment from undeformed bedrock varies from a few centimetres to a few hundred metres. This definition excludes deformations of the entire crust owing to glacioisostatic movements and reactivation of crustal faults under ice loading. Contrary to Slater’s (1926) and Drewry’s (1986) usage, deformations within the ice itself are excluded, as are processes of glacial erosion, such as plucking of bedrock and the particle-byparticle removal and transport of sediment by glacier ice. Banham (1977) introduced the term ‘glaciotectonite’, meaning subglacially penetratively sheared sediments, analogous to mylonites (Lavrushin, 1971; Berthelsen, 1978; Schack Pedersen, 1988; Hart and Boulton, 1991). Here a wider definition is applied in which a glaciotectonite is any body of unlithified or weakly lithified sediment that is structurally deformed

TABLE 14.1. Glaciotectonites (numbers in brackets refer to relevant sections of this chapter) Active

Passive

Supraglacial

–

Collapse structures Debris flows

Subglacial

Shear zones

Diapirs and related structures Water escape structures

Streamlined bed Marginal

Compressive belts Fold and thrust belts

Collapse structures Diapirs and related structures Water escape structures

by glacial stresses. These stresses include those exerted by moving glacier ice, either subglacial or at the glacier margins, and those resulting from loading by (static) ice masses. Sometimes relatively undeformed bodies of well-lithified rock are incorporated in glaciotectonites in which penetrative deformation is restricted to incompetent strata (Kupsch, 1962; Bluemle and Clayton, 1984; Aber et al., 1989). Table 14.1 summarizes the structures resulting from different combinations of glacial stresses (active or passive) and position within the glacial system. Supraglacial deformations will not be discussed. They comprise structures in kames and flow tills and these are best treated in a sedimentological context (Menzies, 1996, chapters 2 and 3).

14.2. SOFT SEDIMENT DEFORMATION Deformation of unlithified sediments has been the subject of soil mechanics and engineering geology. Obviously, these sciences are mainly interested in systems experiencing quite small strains after failure. In glacial geology more interest centres on medium to high strains and a soil mechanical approach does not necessarily solve all problems. The most urgent problem is the present lack of knowledge of timedependent deformation processes in unlithified sediments. Whereas these processes are quite well understood for glacier ice, the rheologies of deforming tills and glacially pushed sediments are still largely unknown (Murray and Dowdeswell, 1992; Murray, 1994). In striking contrast to glaciological studies, very few glacial geological studies actually report measurements of stresses and strain rates in deforming glaciotectonites. Since one of the main goals of glaciotectonic studies is to reconstruct former ice sheet conditions, a proper understanding of soft sediment deformation processes is vital. For more detailed discussions the reader is referred to any soil mechanics textbook and for treatment of strain in geological materials Ramsay and Huber (1983, 1987), Passchier and Trouw (1996) and Maltman (1994). A sediment deforms when the shear stresses in the material exceed its shear strength. Shear strength is formulated in terms of effective stress, which is the

PROCESSES OF GLACIOTECTONISM

total stress minus the (neutral) pressure of the fluid in intergranular pores. All changes in the spatial arrangement of the particles (compaction, compression, shear deformation) are produced exclusively by changes in the effective stresses reflecting grain interaction. Sediment failure can be brittle or ductile. To mark off folds and shear zones from faults and other brittle structures it is necessary to define ductile deformation in unlithified sediments. Contrary to metamorphic rocks, where ductile deformation pertains to crystal lattice deformation, ductile deformation of unlithified sediments is a cataclastic flow process, leaving the grains essentially intact. Ductile deformation is defined here by structures that are produced by permanent continuous deformation without fracturing on the scale in which they are viewed. Brittle deformation of soft sediments typically produces discrete undeformed blocks separated by fault planes or narrow shear zones. On a microscale, ductile structures in soft sediments are built up by numerous small sliding planes, separating aggregates or single undeformed grains (Menzies and Maltman, 1992). The formation of fault planes is described by the Coulomb equation. Since failure is governed by the effective stresses, this law is formulated in its general form as: ␶ = C + ␮s (␴ – pw )

(14.1)

in which ␶ is the shear stress along a potential fault plane, equalling the sediment’s shear strength or plastic yield strength, C is cohesion, ␮ the coefficient of friction (a material constant, ␮s = tan ⌽, where ⌽ is the internal friction angle of the material), ␴ is the normal stress on the fault plane and pw the pressure of the pore fluid. The relationship (␴ – pw ) = ␴e thus represents the effective normal stress across a fault plane (intergranular contact pressure). Materials such as chalk, lignite, clay and loam are cohesive sediments whereas dry sand is cohesionless. There is no permanent strain for stresses below the yield strength. When the effective stress equals the yield strength the material deforms at a strain rate independent of the magnitude of the applied stress. This is called perfectly plastic deformation, which is time independent. A number of investigators have attempted to model glaciotectonic pro-

419

cesses using this principle (cf. Clayton and Moran, 1974). In reality, a certain amount of irreversible strain usually occurs well below the yield strength, while the material deforms by creep (Allen, 1985, p. 161). In the following, an orthogonal coordinate system with the z-axis perpendicular to the shear plane xy, and the x-axis as the direction of shearing will be used. The creep process may be modelled by linearly viscous flow of a Newtonian fluid: ␶zx = ␩s

ducs

(14.2)

dzcs

where ␶zx is the applied shear stress, ␩s the Newtonian viscosity (a material constant) and ducs/dzcs is the rate of change of the creep velocity with depth zcs across a layer of creeping sediment where ducs/dzcs = ␥s , the shear strain rate in the plane of shearing. In a more general form equation (Eq. (14.2)) may be written as: ␥˙ s =

1 ␩s

␶zxn

(14.3)

similar to Glen’s power flow law (Chapters 3 and 4). Here n is a constant that, in sediments, probably depends on the internal friction angle ⌽. For n = 1 flow is linearly viscous (Newton fluid), as in (Eq. (14.2)), and for other values of n flow is non-linearly viscous. The Bingham model is a useful concept describing materials that do not deform at stress levels below their yield strength and deform by creep at higher stresses (debris flows, Allen, 1985, p. 172). This combines plastic and viscous flow: ␥˙ s =

1 ␩s

(␶zx – ␶c )n,

␶zx ≥ ␶c

(14.4)

where for stresses above the yield stress (␶c ) deformation is either linearly or non-linearly viscous flow. Boulton’s subglacial strain experiments beneath Brei∂/ amerkurj¨okull in Iceland (Boulton, 1987a; Boulton and Jones, 1979; Boulton and Hindmarsh, 1987) have led to the formulation of sediment flow laws (Eqs (14.5) and (14.6)) for saturated till, relating the shear strain rate to the applied shear stress ␶zx (equal

420

PROCESSES OF GLACIOTECTONISM

to the basal shear stress ␶b by the glacier) and the effective stress ␴e : ␥˙ s = A(␴e )–m (␶zx – ␶c )n,

␶zx ≥ ␶c

(14.5)

in which the empirical constants A = 10.62, m = 1.25 and n =0.625. Stresses are in bars, strain rate in a–1 Alternatively: ␥˙ s = A(␴e )–m ␶n

(14.6)

where A = 3.99, m = 1.80 and n = 1.33. Equation (14.5) models a Bingham type flow and Eq. (14.6) non-linearly viscous flow. According to Boulton and Hindmarsh (1987) both flow laws fit the observed data very well. Equation (14.6) probably agrees best with the slow creep of clay-rich tills at low stress levels over long periods of time. Equations (14.5) and (14.6) illustrate the strong influence of porewater pressure on sediment viscosity through the effective stress, ␴e . When a sediment is subjected to stress, any reduction of the effective stress by increasing pore pressure may cause slow deformation by creep. For very high porewater pressures approaching the overburden (total stress) the effective stress approaches zero and the sediment liquefies and deforms at very high strain rates. Temperate ice sheets resting on deformable sediments behave in a fundamentally different way from those underlain by rigid bedrock. The latter correspond with the interior parts of Pleistocene ice sheets on the Canadian and Scandinavian shields and most present-day glaciers. The former are represented by large parts of ice sheets invading the lowland areas of North America and Northern Europe. Owing to the low basal shear stresses, soft bed glaciers flow at a much higher rate and develop lower profiles than their hard bed counterparts. Growth and decay of soft bed ice sheets may be expected to be quite rapid (Boulton and Jones, 1979). Effectively, the base of the deforming layer, a cataclastic shear zone, forms the glacier sole, as was already postulated by Banham (1975). The stability of soft bed glaciers and ice sheets is a function of the rate of deformation in the bed, which in turn depends on the amount of basal meltwater draining through the subglacial sediments and upon

their permeability (Murray and Dowdeswell, 1992). A positive effective stress leads to a stable state in which the bed does not deform, or deforms sufficiently slowly so as to allow steady-state ice flow. Unstable states, in which ice flow rates exceed the mass balance velocity, result from zero or negative effective stresses. The Boulton and Hindmarsh model identifies several feedback mechanisms that allow a soft-bed ice mass to return to stability. Recognition of the deformed sediments as the products of these processes will be a vital element of the reconstruction of ancient environments. 14.2.1. Frozen Sediments It is commonly believed that frozen saturated sediments, compared with dry sediments, are rigid under glacial stresses. The argument is that sheets of sand and gravel in push moraines or large sediment ‘rafts’ in till can only move relatively intact if they are bound by ice (Menzies, 1990c). Fault planes and folds in sediments are regarded as evidence of frozen conditions during glaciotectonic deformation. Sometimes the argument is taken even further and the depth to d´ecollement of glaciotectonic structures is believed to coincide with the depth of the permafrost (Richter et al., 1951). Theoretical and empirical arguments challenge this assumption (Van der Wateren, 1981, 1985, 1995). Considering a stratified sediment body containing layers of contrasting rheology subjected to glacial stresses, it can be expected that these layers react differently under the same stress field. A layer of low plastic yield strength or, alternatively, one of low viscosity may deform, while more competent layers remain undeformed or deform at a much slower rate. Outcrop and map scale structures demonstrate that sand and gravel are usually more competent than clay, loam and chalk, although Croot (1988b) described a Neoglacial push moraine from Iceland in which a gravel layer forms the base of the thrust sheets, while consolidated silts appear to be more competent. Porewater pressure plays the key role since sediments of low permeability, and particularly compressible sediments such as clays, can retain high pore pressures longer than permeable sediments (Hubbert and Rubey, 1959; Van der Wateren, 1985).

PROCESSES OF GLACIOTECTONISM

Soft sediments do not deform chaotically under stress. A large number of experiments confirm that faults may be generated in quite an orderly way in dry loose sand, leaving the intermediate blocks intact (Horsfield, 1977; Mandl et al., 1977). Slumps and landslides usually consist of a number of relatively undeformed blocks bound by fault surfaces. Submarine diapirs and compressive structures that are commonly found on large delta foreslopes further illustrate that freezing is not a precondition for faults and folds to be generated in soft sediments (Aber, 1988a). Observations from Iceland (Ing´olfsson, 1988) and Spitsbergen (Boulton et al., 1996) confirm that glaciotectonic deformations may occur in unfrozen saturated sediments. In both instances a glacier advanced across a fjord shearing muds from the bottom and pushing them at the ice margin (Boulton et al., 1996). Finally, evidence from North America (Quigley, 1983) and Europe (Van der Wateren, 1987, 1994, 1995) suggests that large lakes existed along mid-latitude ice sheets inhibiting the development of a frozen margin. In Germany, delta and lake sediments make up a large proportion of some of the largest push moraines (Van der Wateren, 1994, 1995) indicating these lakes were present prior to the formation of these push moraines. In situ observations of deforming frozen till indicate that ice-bound sediments may be much weaker than is commonly assumed. Strain-rate measurements in the deforming bed of a subpolar glacier in China led Echelmeyer and Wang (1987) to conclude that the effective viscosity of frozen silt is more than 100 times lower relative to ice under the same stress and temperature conditions. While in an Icelandic temperate glacier up to 90 per cent of the glacier motion was accomplished by deformation of the bed, 60–80 per cent of motion was accounted for by deformation in the Chinese cold glacier. Shear strain rates in the Brei∂/ amerkurj¨okull experiments are between two and five times higher than those in the Chinese subglacial experiments for a comparable ice overburden. The apparent viscosities of frozen silt and water-saturated till probably differ by less than one order of magnitude. No empirically tested criteria would appear to exist to distinguish glaciotectonic structures formed under frozen and unfrozen conditions. Therefore it would appear justifiable to assume

421

that subglacial shear zones, folds and faults may form in saturated frozen sediments just as well as in unfrozen sediments. 14.3. GLACIOTECTONIC REGIMES Since the 1970s it has become increasingly evident that deformation of the glacier bed plays a key role in the dynamics of mid-latitude ice sheets and ice streams (Chapters 3, 4 and 8). During glacial periods the continental ice sheets develop in the shield areas and then invade the mid-latitude sediment-covered lowlands. Large parts of these lowland areas are susceptible to subglacial shear. Conditions of the bed and the ice sheet control the production of structures and landforms by subglacial shear. It appears that the structures, which the ice left behind, can be grouped ice sheet centre equilibrium line (EL)

(a) Xm

Xe

X

u

(b) Xe

Xm

X

Xm

X

¶u ¶x

e Xe

(c)

c

FIG. 14.1. Theoretical velocity and strain rate distributions in a continental ice sheet. (a) Flow lines in a vertical cross section (xz plane). X is distance from ice sheet centre, Xe at equilibrium line, Xm at the margin. (b) Horizontal velocity u as a function of distance x. (c) Horizontal strain rate component ␥xx = ⭸u/⭸x is extensional (e) from the centre to the equilibrium line and compressive (c) from there to the margin.

422

PROCESSES OF GLACIOTECTONISM EL discontinuities in surface profile

(a) permafrost

basal equilibrium line warm based

cold based

discontinuities in surface profile

basal equilibrium line

(b)

low K

high K

u

(c) Xeb

Xe

Xm

X

¶u ¶x

e

e c

X

(d)

c

FIG. 14.2. Theoretical velocity and strain rate distributions in a temperate continental ice sheet (after Boulton, 1972, 1987a; Boulton and Jones, 1979). Discontinuities in surface profiles coincide with changes in basal shear stress. (a) Flowlines in the ablation area, related to basal thermal conditions. Ice is underlain by deformable sediments. Basal melting occurs beyond basal equilibrium line. Refreezing of basal meltwater occurs in a narrow zone near the margin. (b) Ice sheet resting on bed of low hydraulic conductivity (K) and low shearing resistance near the margin. (c) Horizontal velocity distribution for thermal and geological conditions in (a) and (b). (d) Horizontal strain rates for (a) and (b). High extensional strain rates near the margin may lead to formation of tunnel valleys and glacial basins. High compressive strain rates may lead to pushing at the margin.

according to two main glaciotectonic regimes, reflecting the characteristics of the ice sheet and the nature of the bed (Figs 14.l and 14.2). The mass balance velocity of an ice mass terminating on land gradually increases from the accumulation area to the equilibrium line, to decrease from there to the terminus. The accumulation area is thus characterized by extending flow (strain rate) (⭸u/⭸x > 0) and the ablation area by compressing flow (⭸u/⭸x < 0) (Fig. 14.1) (Boulton, 1987a; Hart et al., 1990). Where the ice moves from a rigid bed (the shield area) into an area of deformable sediments (the lowlands), the ice can be expected to accelerate. This produces extensional strains and a low surface profile in this part of the ice mass. A similar distribution of strain rates will be found in deforming beds; it is appropriate to distinguish two main glaciotectonic regimes, extensional and compressive regimes (Fig. 14.2). There are two more situations where flow in the ice sheet and within the deforming bed may change from extending to compressing flow. First, in subpolar ice sheets a zone near the margin is frozen to the bed, while further inward the ice is temperate and meltwater drains into the bed. Field experiments indicate that sediments beneath a frozen margin may deform more slowly than the saturated unfrozen interior parts of the bed (Fig. 14.2(a)). Secondly, if ice moves across a sediment of low yield strength (or viscosity) into an area of less deformable sediments, decreasing strain rates in the bed can be anticipated. This may occur where a temperate ice sheet moves from an impermeable bed to one that drains basal meltwater (Fig. 14.2(b)). 14.3.1. Subglacial Shear Zone and Marginal Compressive Belts Under the influence of the glacier’s basal shear stress subglacial sediments of low shear strength are subjected to horizontal simple shearing. Subglacial sediments are removed from this zone and transported towards the ice margin. Erosion of the substratum is strongest where ⭸u/⭸x >> 0. This regime is characterized by widespread diamictons and streamlined bedforms. Overdeepened glacial basins are commonly excavated on the proximal side of push moraines (Van der Wateren, 1995).

PROCESSES OF GLACIOTECTONISM

Sediments that are dumped at the ice terminus as subglacial, englacial and supraglacial diamictons and glaciofluvial deposits are subject to horizontal compression. This marginal compressive belt zone at former glacier margins includes pushed and stacked till sheets, and push moraines. Both subglacial shear zones and marginal compressive belts correspond, respectively, with the streamlined and glacial thrust terrains recognized in North America (Moran et al., 1980). The two regimes are characterized by quite different tectonic styles. Overprinting occurs, for instance, when an ice sheet, after a period of zero mass balance, allowing a compressive belt to develop, expands and overrides the compressive structures. Steeply inclined structures will be refolded and incorporated in a subglacial shear zone (a deformation till of parautochthonous material overlying pushed structures). Regional mapping of overprinting relations helps to identify periods of ice sheet expansion, retreat and readvance (cf. Van der Wateren,1995; who presents a distribution map of glaciotectonic styles in northwestern Europe).

14.4 SUBGLACIAL SHEAR ZONES Shear zones occur in several positions in a glacial environment. In the first place the ice itself principally moves by simple shear under the influence of gravity (Chapters 3 and 4). The bottom part of a glacier functions as a shear zone. When the bed deforms this becomes the lower part of the basal shear zone (Fig. 14.3(b)). Shear zones also form in the bottom layers of thrust sheets and nappes that are stacked near glacier margins (Fig. 14.3(a)). Finally, shear zones form the base of debris flows, flow tills, landslides and slumps occurring in deposits on and around glaciers (Fig. 14.3(c)). Subglacial simple shearing produces a tectonite known as deformation till. Strain measurements beneath Brei∂/ amerkurj¨okull (Boulton, 1987a) revealed that the glacier’s basal shear stress imposes a high amount of shear strain on the top few decimetres or metres of the bed. Strain rates in the saturated till varied from 10 to 55 a–1. Thus even after a few years the finite shear strain reaches extremely high values, of the order of ten up to a few hundred. To illustrate

423

this alteration in shear strain, an originally circular passive marker in the shear zone is transformed in this short time into an ellipse with a long axis more than a hundred times as long as the short axis and dipping less than 0.5°, for a shear strain ␥ = 10. For ␥ = 100 the aspect ratio of the finite strain ellipse equals nearly 10 000 and the circle virtually becomes a horizontal line. In view of the strong implications for the dynamic behaviour of former ice sheets, recognition of deformation tills in ancient deposits is vital. Where an ice sheet is underlain by soft sediments, subglacial shearing may be a more general mechanism of ice stream formation. Surging behaviour of glaciers may also be connected with a deforming till (Sharp, 1985; Clarke, 1987b). The large ice lobes that are characteristic of the Pleistocene ice sheet margins of Europe and North America may likewise be explained by fast ice flow or surging on a bed of very low shear strength (Boulton et al., 1985; Lagerlund, 1987). The next section on subglacial shear zones (or deformation tills) will be considered together with shear zones in push moraines, since they are both controlled by progressive simple shear and therefore their structural styles are similar (Van der Wateren, 1987, 1995). The two are distinguished in the field by their tectonic and sedimentary relations with overlying and underlying sediment masses. A distinction on the basis of their (micro-) structures alone will be rarely possible. To ease comparison most shear zones in this chapter are presented with dextral or right lateral shear sense, when looking at the plane of the cross section. 14.4.1. Tectonic Style of Soft Shear Zones The following ideas result from a comparative study of modern deformation tills and shear zones in push moraines of central Spitsbergen and their ancient counterparts in Germany and Norfolk (Hart, 1987, 1990; Van der Wateren, 1987, 1995, Van der Wateren et al., 2000; Hart et al., 1990; Boulton et al., 1999). Soft shear zones, or shear zones in unlithified materials, are in many ways similar to their hard rock counterparts. They share most of the macroscopic and microscopic structures known from mylonites and cataclastic shear zones, except those relating to

424

PROCESSES OF GLACIOTECTONISM

(a) Sr Sb Sh Sb

g (b)

sub nappe shear zone

Sr

ice

ice

A

Sh Sb

B1 Sr

g

B2

deformation till

(c) debris flow

g

substratum undeformed not to scale

FIG. 14.3. Shear zones in different parts of a glacial environment. The graphs show finite shear strain (␥) as a function of depth. (a) Shear zone in the basal layers of a soft sediment nappe as it is found in push moraines (after Van der Wateren, 1987). Sr is the minimal strain horizon containing rooted structures. Sb contains boudins, detached folds and a transposed foliation. Sh is the maximum strain, homogenized central horizon of the shear zone. Normal faulting is very common beneath nappes and thrust sheets in a push moraine. (b) Shear zone (deformation till) in subglacial sediments mirrors basal shear zone of the overlying glacier. A, B1 and B2 are horizons in Boulton’s (1987a) deformation till model. (c) Shear zone at the base of a debris flow. Note abrupt change of shear strain from debris flow to undeformed substratum.

recrystallization of minerals at higher temperatures and pressures (Maltman, 1987; Van der Wateren, 1995; Van der Wateren et al., 2000). Progressive simple shear is the simplest model describing the deformation history of subglacial shear zones with sufficient accuracy (Van der Wateren et al., 2000). It is a non-coaxial deformation generating asymmetric structures, in which the principal direction of finite extension (␭1 ) tends to become almost

parallel to the shear zone boundaries (Choukroune et al., 1987). At medium finite shear strains, a typical soft shear zone contains folded and strongly attenuated sediment layers. At higher strains these structures disintegrate in a completely homogenized matrix (Boulton, 1996b). A summary of progressive simple shear is presented in the following section (for a general discussion of progressive deformation in shear zones see Ramsay and Huber, 1983, pp. 15, 217,

PROCESSES OF GLACIOTECTONISM

and for a discussion of soft shear zones see Van der Wateren, 1995; Van der Wateren et al., 2000). In an ideal shear zone, strain varies continuously from one margin to the other, being highest in the centre where the shear strain rates are at maximum, with zero shear strain at the walls (Fig. 14.4). Finite strain trajectories (␭1 ) are sigmoidally shaped lines following the long axes of the finite strain ellipses (Fig. 14.5(a)). Heterogeneous volume change (dilation ⌬) inside the shear zone, perpendicular to its walls (Fig. 14.5(b)) and uniform homogeneous strain affecting the shear zone together with the surrounding sediments (Fig. 14.5(c)) will be ignored in this discussion.

g FIG. 14.4. Trajectories of finite elongation (␭1 ) in a shear zone.

(a)

425

According to Fig. 14.6, line A is the direction of shearing, the plane AB the shear plane and the plane AC the plane of shearing. In simple shear, all changes of length (longitudinal strain or extension, e) or changes of angles between material lines (shear strain, ␥) occur in the AC plane of shearing. An originally circular passive marker in the plane of shearing deforms to an ellipse after one increment of simple shear. If it is assumed that the strain increment is infinitesimally small, the long axis of the incremental strain ellipse will dip at 45° away from the shear direction (Fig. 14.7(a)). One of the lines of no incremental extension is parallel to the C axis. In progressive simple shear, the shear plane contains both the line of no incremental extension and the line of no finite extension. Finite extension is the sum of all previous incremental extensions. With each subsequent strain increment the elipticity of the finite strain ellipse increases and its long axis (1 + e1f ) rotates in the direction of shearing (Fig. 14.7(b)). Both the incremental and the finite strain ellipses are divided by lines of no longitudinal strain into four sectors, two in which material lines having orientations lying in this sector are shortened (ei and ef negative) and two in which material lines have become longer (ei and ef positive). The AC plane contains the longest and the shortest principal axis of the strain ellipsoid (␭1 and ␭3 ).

g (b)

D

plane of shearing

shear plane

(c)

FIG. 14.5. (a) ␭1 trajectories follow long axes of finite strain ellipses. (b) Shear zone with additional dilation (␦). (c) Homogeneous strain affecting both the shear zone and its walls.

B

C

A

FIG. 14.6. Definition of principal directions A, B and C in a shear zone.

426

PROCESSES OF GLACIOTECTONISM

g = -0.07 c

e1i

e1f

2 e1f

3

1

g = -3.75

C

1

A

2 A

e3i

e3f

(a)

(b)

3

e3f

(c)

(d)

FIG. 14.7. (a) Orientation of the longest and shortest principal axes of the incremental strain ellipse. Shaded areas denote extension. (b) Orientation of the longest and shortest principal axes of the finite strain ellipse. (c) Superposition of finite and incremental strains. See text for explanation of sectors 1, 2 and 3. (d) Finite strain ellipse and compressive and extensional sectors for a finite shear strain of –3.75.

Superposition of the incremental strain ellipse on the finite strain ellipse yields three sectors in progressive simple shear (Fig. 14.7(c)) and consideration, here, will be given to the effects on a competent layer in an incompetent matrix (Figs 14.8 and 14.10). 1 Finite longitudinal strain (ef ) and incremental longitudinal strain (ei ) are negative. Fabric elements, for example, sediment layers having this orientation, are shortened and continue to shorten during the next increment. Shortening is accomplished by thickening and folding of competent layers. 2 If ef is negative and ei is positive. Fabric elements in this sector have been initially shortened and will

T

R´ C

l1

be stretched during the next increment. This may lead to unfolding of previously folded competent layers, but thinning and boudinage of fold limbs will be more common. 3 Both ef and ei are positive. Fabric elements lying in this sector have been stretched and will undergo stretching during all subsequent strain increments. Competent layers suffer boudinage. Folded boudins are produced where pure shear strain overprints the simple shear fabric (i.e. when owing to non-uniform flow (⭸u/⭸x≠0) longitudinal compression alternates with longitudinal extension). This may be an indication of interference of different glaciotectonic regimes but is frequently a product of lithological inhomogeneities. A number of structures may be recognized that make it possible, if not to quantitatively determine the finite strain, at least to distinguish low strain and high strain fabrics and to assess the orientation of the finite strain axes and the tectonic transport direction

R M

GL

A

P

GL

62

310

FIG. 14.8. Brittle shear zone structures.

PS

1m

130

FIG. 14.9. Conjugate sets of Riedel shears beneath a nappe in the Dammer Berge push moraine (Van der Wateren, 1987; reprinted from: Till and Glaciotectonics (J. J. M. van der Meer, ed.), 1987a; courtesy of A.A. Balkema, Rotterdam). Shear is clockwise.

PROCESSES OF GLACIOTECTONISM

427

PLATE 14.1. Conjugate sets of Riedel shears in subglacially deformed rythmites, incorporated in a Saalian till in The Netherlands (van der Meer et al., 1985). Thin section is viewed under crossed nicols (vertical and horizontal). Bright bands are strongly oriented clay minerals as a result of layer parallel shear (M shear). Dark cross-cutting lines are oriented clays in the extinction orientation (R and R⬘ shears). Asymmetry indicates left lateral (anti-clockwise) shearing. Field width 6 mm from left to right.

428

PROCESSES OF GLACIOTECTONISM

(Platt, 1984; Malavieille, 1987; Petit, 1987; Chester and Logan, 1987). Riedel (1929), Tchalenko (1970) and Tchalenko and Ambraseys (1970) showed that brittle shear zones of all scales share characteristic combinations of faults and fractures. These are presented below in a model of a cataclastic shear zone. Where the material favours ductile deformation, the strain varies continuously from the shear zone walls to its centre. Materials containing inequant grains, such as clay minerals, may develop a distinct penetrative set of cleavage surfaces parallel to the ␭1 trajectories. This fabric results from preferred grain alignment in response to local strain. In soft shear zones, cleavages may be recognized, remarkably similar to S- and C-foliations in mylonites (Malavieille, 1987; O’Brien et al., 1987). Figure 14.8 summarizes brittle shear zone structures (after Petit, 1987). The extensional sector of the incremental strain ellipse is shaded. The inherent asymmetry is enhanced by a number of shear planes (Riedel shears (R, R⬘) and tension fractures (T)), creating extension in the shear direction. It further contains shear planes M and P with sense of shearing compatible with the bulk shear. The shear parallel (AC) section (containing the longest and the shortest principal axis of the strain ellipsoid) has a strong asymmetry, whereas the section normal to the shear vector (BC) shows a pattern of shear planes that is symmetric about the plane of shearing (Chester and Logan, 1987). These structures may be found on all scales. Riedel shears occur as conjugate sets of normal faults beneath tills or nappe shear zones (Fig. 14.9) or in thin sections of tills (Plate 14.1). Tension fractures (T) have been described as so-called till wedges (Åmark, 1986; Dreimanis and Rappol, 1997; Van der Wateren, 1999). Figure 14.10 summarizes structures typical of ductile shear zones. Buckle folds, kink bands and compressive crenulations are produced by finite shortening in the compressive sector of the strain ellipse. Boudins are produced by finite extension in the direction of shearing. Folds form in lithologies with sufficient interlayer viscosity contrast. In progressive simple shear they soon become asymmetric, overturned and attenuated in the shear direction. Fold axes initiating at high angles to the shear vector tend

C

A

FIG. 14.10. Ductile shear zone structures. Note that at higher finite strains the boudins will be oriented almost parallel to the shear plane (AB).

to rotate towards an orientation parallel to the bulk shear direction forming tube-shaped non-cylindrical folds or sheath folds. Sheath folds typically form at shear strains of ten or more (Cobbold and Quinquis, 1980). Although they are not diagnostic of simple shear regimes, in combination with other structures they may help to identify zones with very high shear strains. In shear parallel sections they appear as asymmetric isoclinal folds, while in the section perpendicular to the shear direction they show as eye or ring shapes (Kluiving et al., 1991; Van der Wateren, 1999). Figure 14.11 summarizes the structures that may be found in association with subglacial or sub-nappe shear zones. The layering, typical of many shear zones, is a tectonic lamination produced by repeated folding and attenuation at very high finite strains. In principle, the orientation of the finite strain ellipse can be determined from the relations between compressive and extensional structures in a shear zone. The relative asymmetry of AC cross-sections and symmetry of BC sections, both on a microscale and in outcrops, indicate the direction of shearing and therefore are good indicators of former ice flow directions. Kluiving et al. (1991) applied this principle to a section in the eastern Netherlands where three overlying tills belonging to quite different ice flow directions could be distinguished.

PROCESSES OF GLACIOTECTONISM

429

C B

A FIG. 14.11. Block diagram of typical shear zone structures. Note eye-shaped structures (cross-sections of sheath folds) and symmetric fault sets in section perpendicular to shearing (from Kluiving et al., 1991; reprinted from ‘Till stratigraphy and ice movements in eastern Overijssel, The Netherlands’, by S. J. Kluiving, M. Rappol and F. M. Van der Wateren, Boreas, 1991, 20, 193–205, by permission of the Scandinavian University Press).

In progressive simple shear the direction of finite extension is almost parallel to the direction of shearing. The very high finite strains produce a transposed foliation quite similar to some highly deformed metamorphites. Transposition is the process converting a primary (e.g., sedimentary) lamination into a tectonic lamination of highly attenuated and boudinaged fold limbs and detached, isolated intrafolial folds. At first sight it may resemble sedimentary layering but, in fact, it is a tectonic layering of extremely high finite strain. Repeated folding and attenuation transforms even simple two- or three-layered sedimentary sequences into complex multi-layered transposed units. Plates 14.2 and 14.3 are examples of chalk banding in a Danish clay-rich till and sandy laminae in a deformation till in Germany, respectively. Continued deformation eventually produces a completely homogenized tectonite of, at first sight, deceptively low strain (Boulton, 1996b). Only the rare occurrence of an isolated boudin or fold nose and the typical microscopic fabric of Riedel shear planes, cleavages and microboudins are witness to this deformation process. A typical soft shear zone can be divided into three units of different tectonic styles (Van der Wateren, 1987). In order of increasing finite shear strain and tectonic transport these units are (Fig. 14.3(a), (b)):

䊉

䊉

䊉

Sr – rooted recumbent structures at the lower boundary of the shear zone and low strain structures at the upper boundary; Sb – a transposed foliation of boudins, rootless intrafolial folds and highly attenuated bands of footwall material in a more allochthonous matrix; Sh – a completely homogenized mixture of exotic and more or less autochthonous elements, bearing witness to an extremely high finite shear strain in the central part of the shear zone.

These units may be distinguished wherever unlithified sediments have been subjected to shear stresses, whether this occurs beneath a sediment thrust sheet or beneath glacier ice. In the latter case the mirrored top half of the shear zone is in the basal ice layers and the tectonite would be called a deformation till. Units Sr and Sb correspond with the B horizon and unit Sh with the upper, A horizon of Boulton’s (1987a) deformation till. The Sbunit, of intermediate strain, may be banded if the lithological contrasts are sufficiently strong. Boudins may reach impressive dimensions. Large, so-called rafts, floes or mega-blocks of sediment or even lithified bedrock embedded in till have been described from North America and Europe. Banded tills have been interpreted as waterlain or meltout tills but in many cases are the product of

430

PROCESSES OF GLACIOTECTONISM

(a)

(b) PLATE 14.2. Transposed foliation of clay rich till and chalk laminae in Møns Klint, Denmark. Folding resulting from local compression in otherwise horizontally laminated deformation till.

PROCESSES OF GLACIOTECTONISM

431

PLATE 14.3. Transposed foliation of loamy till and sand laminae from the Dammer Berge, Germany.

subglacial shearing. The chalk banding in tills from Denmark (Berthelsen, 1978) and East Anglia (Hart, 1990) are clear illustrations of this process (Plate 14.2) (Hart, 1998). Rappol and Stoltenberg (1985) showed that the incorporation of footwall material had a distinct effect on the composition of the till even when banding and shearing were not immediately obvious.

14.5. MARGINAL COMPRESSIVE BELTS Push moraines have long been known by herdsmen and farmers in the Alps. The earliest references date from the sixteenth century (B¨ohm, 1901). Dramatic accounts of advancing valley glaciers shoving boulders, trees and even houses entered the early glacio-

432

PROCESSES OF GLACIOTECTONISM

logical literature and contributed to the glacial theories of Agassiz (1840), Geikie (1863) and Geikie (1884). During the Little Ice Age many glaciers advanced, producing vast outwash fans and push moraines. Particularly those in arctic and subarctic environments have been the subject of early structural and geomorphological analyses (von Drygalski, 1897; Gilbert, 1903; Tarr and Butler, 1909; Martin, 1913; Grant and Higgins, 1913; Tarr and Martin, 1914; Slater, 1926; Gripp, 1929; Todtmann, 1936, 1960). More recent studies of Neoglacial push moraines include (Croot, 1988b; Van der Wateren, 1995; Boulton et al., 1999). Only very few studies exist of modern, active push moraines. K¨aelin (1971) presented a geomorphological and structural analysis of the push moraine in front of the Thompson Glacier, Axel Heiberg Island, Canada, the development of which was observed over a number of years. Eybergen (1987) has carried out a similar study of a small push moraine in front of the Turtmann Glacier in the Swiss Alps. From these studies it can be concluded that push moraines form under the following conditions (Van der Wateren, 1987, 1995): (1) a suitable d´ecollement close to the surface; (2) an advancing or readvancing ice sheet; and (3) a coupling of the ice mass with its bed. Condition (1) introduces the influence of the geology of the substratum. The Rehburg line of push moraines in Germany and the Netherlands coincides with a zone in which Tertiary and Pleistocene clays come close to the surface (Van der Wateren, 1995). The push moraines along the escarpment of the Missouri Coteau in North America are another example of substratum-controlled glaciotectonic features (Kupsch, 1962; Moran et al., 1980; Bluemle and Clayton, 1984). Condition (2) involves glaciological conditions where an ice advance or readvance may have occurred on a continental scale owing to a positive mass balance, or may be local, owing to a surge or an ice stream. Condition (3) relates to the sediment substratum over which the ice sheet advances, since the sediment mechanical properties determine whether the ice is coupled with the underlying sediments, a prerequisite for pushing.

14.5.1. Tectonic Style of Push Moraines Push moraines are zones of strong horizontal compression. In Pleistocene glaciated areas, they rim vast till-covered plains dominated by strong extensional strains. On their upstream side depressions commonly occur from which the pushed masses have been removed. Typical examples are the ‘hill-depression forms’ in North Dakota, Alberta and Saskatchewan (Clayton and Moran, 1974; Bluemle and Clayton, 1984; Moran et al., 1980) and the push moraines of the Rehburg line in the Netherlands and Germany surrounding deep glacial basins (Meyer, 1987; Van der Wateren, 1987; van den Berg and Beets, 1987; de Gans et al., 1987). Push moraines, in the strict sense, are built of materials deposited in the foreland, before they are compressed by an advancing or re-advancing glacier or ice sheet. The sediments may consist of tills and debris accumulating at the snout during the summer that will be pushed when the glacier re-advances the next winter. The small annual push moraines in front of valley glaciers form in such a way. They may also include basal tills and glaciofluvial outwash from previous more extensive advances. Push moraines form in a large variety of unlithified or weakly lithified sediments, such as tills, fluvial and glaciofluvial sands and gravels, fine-grained delta and lake deposits, marine clay and silt, chalk, sandstone and shale. Where the moraines are well exposed, it can usually be shown that the d´ecollements coincide with incompetent strata. Figures 14.12 and 14.13 are synthetic block diagrams of two general types of push moraines. Such forms are, in many ways, analogous with accretionary wedge complexes at convergent plate margins. The models are mainly based on evidence from the Neoglacial Holmstr¨ombreen push moraine, Spitsbergen (Boulton et al., 1989, 1999; Van der Wateren, 1995) and the Pleistocene push moraine of the Dammer Berge, Germany (Van der Wateren, 1987, 1992, 1995). When fine-grained sediments form the bulk of a push moraine, the structures are dominantly ductile (i.e., folds and fold nappes). In push moraines dominated by coarse-grained sediments with a fine-

PROCESSES OF GLACIOTECTONISM

433

1 km

FIG. 14.12. Block diagram of the Holstr¨ombreen push moraine, Spitsbergen (after Boulton et al., 1989). Tectonic style changes from the foreland to the ice margin from Jura type concentric folds, through fold-thrust nappes to gravitational nappes.

5 km

FIG. 14.13. Block diagram of the Dammer Berge push moraine (after Van der Wateren, 1987). Push moraine comprises large (kilometre scale) nappes that are internally imbricated.

grained basal slip layer, the structures tend to be brittle and folds have longer wavelengths. Large subhorizontal nappes are the primary structural units. Folding is the dominant tectonic style in the Holmstr¨ombreen model (Fig. 14.12). In those parts of

the push moraine where thick fluvial gravel and sand sequences prevail, fold wavelengths tend to be larger than in more fine-grained and laminated lithologies. Concentric style folds near the push moraine front tend to become tighter and more asymmetric in the intermediate and internal zones of the wedge, where

434

PROCESSES OF GLACIOTECTONISM

they develop into recumbent folds, thrust and fold nappes as the strain increases. This change in tectonic style is quite similar to that in, for example, the bedrock forms of the Foothills and Front Ranges of the Canadian Rocky Mountains (Dahlstrom, 1970). In the internal zone, the finite strain is at maximum and thrust nappes and gravity-driven sliding nappes predominate at the surface. Figures 14.14(a), (b), (c) are examples of structures in the external, intermediate and internal zones, respectively, of the Holmstr¨ombreen push moraine. The central and proximal parts of the push moraine in the Dammer Berge model (Fig. 14.13) are built of

piles of large, more or less horizontal and internallydeformed nappes that have moved long distances on shear zones of Tertiary and Pleistocene clay and loam. The nappes measure up to several kilometres across and are up to ~100 m in thickness. As in the Holmstr¨ombreen push moraine, finite shortening increases in a hinterland direction. In the distal parts, steeply inclined folds and imbricate thrusts seem to take up most of the strain. The majority of the folds and imbricate structures appear to have been initiated at an early stage of movement of the nappe and remained active throughout subsequent stages of its development.

N

S 10 m

(c)

780

760

N

S 20 m

(b)

500

520

540 N

S 10 m

(a)

100

80

FIG. 14.14. Examples of structures typical of the external (a), intermediate (b) and internal (c) zone of the Holmstr¨ombreen push moraine. Horizontal axis shows distance in metres from the push moraine front. (a) Concentric fold style in external zone. Low strain. (b) Fold-thrust nappe in intermediate zone. Medium strain. (c) Thrust and gravitational gliding nappes in the internal zone near the ice margin. Highest amount of finite shortening. Some of the nappes develop into mud flows.

PROCESSES OF GLACIOTECTONISM

3

2

1

FIG. 14.15. Glaciofluvial deposition in relation to glaciotectonics (after Van der Wateren, 1987). Finite strain decreases from oldest to youngest sediments. 1, pre-tectonic sediments; 2, syn-tectonic fill of synclinal valley (delta and channel deposits); 3, post-tectonic sediments (channel and alluvial fan deposits). Reprinted from: Tills and Glaciotectonics (J. J. M. van der Meer, ed.), 1987a; courtesy of A.A. Balkema, Rotterdam).

The top layers of the proximal slopes were transformed into deformation till when the Scandinavian ice sheet advanced and overrode the push moraine, as occurred during the maximum advance of the Saalian glaciation in the Netherlands and Germany. It is thus bounded below and above by shear zones. Debris flows (flow tills) moved down the glacier front, slumps forming at the wedge front. Meltwater streams cutting into a push moraine while it is forming, generate syntectonic unconformity surfaces and sediment bodies. These sediments are more deformed themselves the earlier they have been deposited. Truncated anticlines with folded unconformity surfaces and growth folds (Fig. 14.15) are diagnostic of syntectonic sedimentation. These structures may help to estimate the chronosequence of tectonic events and to construct a stratigraphy of pre-, syn- and post-tectonic sediments (Van der Wateren, 1987, 1995). 14.5.2. Folds and Thrusts Fold styles associated with subglacial and nappe shear zones markedly differ from macroscopic folds in push moraines. The former are produced mainly

435

by layer-parallel simple shear and are dominated by recumbent similar and isoclinal fold styles. These folds are passive structures and the orientation of their axes bears no direct relation to the main stress field, but tends to align with the direction of maximum finite extension. Large-scale folds in push moraines are the result of horizontal compression producing steeply inclined thrusts and folds of concentric style. These active folds commonly have their axes oriented at high angles to the bulk shortening direction. Folds, like imbricate thrusts, seem to precede nappe movement in many of the larger push moraines. In the Holmstr¨ombreen push moraine they began as concentric or box folds, moving on a d´ecollement over an incompetent bed. When the folded and imbricated sediments are further compressed, the structures move as a whole on top of the foreland to form folded and imbricated nappes. In many push moraines, folds are found on a level that is too high for them to be rooted in the lowest d´ecollement of the push moraine. This implies that they have either formed at an early stage and been transported to their present high position as part of larger nappe structures or have formed at a late stage of compression.

14.5.3. Stacked Till Sheets Compressive zones near the margins of large ice sheets are comprised of more than just push moraines of deformed glaciofluvial deposits and other foreland materials. Compressing flow in the ablation zone produces a narrow terminal zone of thick folded and thrust till sequences (Boulton, 1996a,b). Apart from tills, steadily moving outward under basal shear, large blocks of footwall sediment or sedimentary rock may be piled up near the margin. These blocks or rafts are surrounded by till, which distinguishes them from structurally quite similar push moraines. Some of these blocks are known to have travelled very long distances. Viete (1960, p. 25, citing Moskvitin, 1938) mentions rafts of Lower Carboniferous rocks measuring 16–20 km2 that have travelled more than 200 km from their source to the Saalian end moraines northwest of Moscow. Moran (1971) reports stacked

436

PROCESSES OF GLACIOTECTONISM

till sheets and rafts of Cretaceous sediments from Saskatchewan and Alberta, Canada. One of these rafts measured ~1000 km2 (examples in Christiansen, 1971; Prange, 1978; Ringberg et al., 1984; Ruszynska-Szenajch, 1987; Aber et al., 1989). Stacked and folded till sheets are characteristic of the termination of the subglacial transport belt, although compressing flow may occur upstream from the terminus, where flow is locally non-uniform. The main problem in identifying these forms is to be certain that the tills overlying and underlying the rafts are of the same age. From a process point of view they are quite similar to push moraines since both are produced by the compressive strain regime near the ice margin. Stacked till sheets are typically produced at stable margins, whereas push moraines are products of glacial advance or a readvance.

14.5.4. Accretionary Wedge Model of Push Moraines It is difficult to overlook the similarities between push moraines and large mountain belts such as the Rocky Mountains or the Alps. Both are prismatic, in cross-section wedge-shaped, sediment bodies, formed by horizontal compression. Push moraines are produced wherever an advancing glacier becomes coupled to sediment masses at the margin. These are subsequently pushed to form sediment wedges, similar in shape to accretionary wedges (Stockmal, 1983). Coupling of advancing ice and marginal sediments may occur under one or more the following conditions: 1 when, after a stillstand, a glacier advances into outwash accumulated during this period; 2 when a glacier surges into outwash sediments, deposited when the margin was stationary; 3 when a glacier advance reaches an area in which a suitable d´ecollement layer, (e.g., a clay) comes nearer to the surface; and 4 when a glacier advances from an area in which the surface consists of fine-grained deformable sediments into an area of more coarse-grained, resistant sediments.

Only the first of these conditions is related to mass balance changes. It is, therefore, important to be able to distinguish them on the basis of sedimentary sequences and glaciotectonic structures. Many push moraines, in mountain areas, are almost entirely composed of proglacial outwash material (deltas and alluvial fans). An ice mass may be expected to be almost completely uncoupled if moving across a relatively smooth surface of fine-grained sediments, given a sufficient supply of subglacial meltwater. The bed deforms by simple shear, inhibiting large-scale thrusting and folding at the margin. This may also apply to frozen fine-grained sediments. A surface of coarse-grained sediments, not susceptible to subglacial shearing, may be expected to promote pushing, provided other conditions are satisfied. The mechanical aspects of pushing under the conditions noted above can be summarized based on an analysis by Van der Wateren (1995) following similar studies of accretionary wedges at convergent plate margins (Elliott, 1976; Chapple, 1978; Stockmal, 1983; Davis et al., 1983; Dahlen et al., 1984; Platt, 1986). In this analogy the ice sheet performs the role of the overriding plate advancing on a surface covered by sediments of varying shear strength, scraping these off their basement and accreting them to the advancing wedge. In longitudinal cross-section, push moraines are tapered sediment masses with their thick end abutting on the glacier. The folded and thrust masses exhibit a general spatial and temporal variation in style and intensity of deformation. Bulk finite strain increases from the thrust front towards the hinterland, with the tectonic style changing from d´ecollement folds to large overthrusts and nappes (Fig. 14.12). It has been observed in modern active push moraines, and confirmed by experiments in deforming wedges, that the highest longitudinal compressive strain rates occur in a rather narrow zone near the thrust front. Strain rates rapidly decrease towards the internal parts (Stockmal, 1983; Dahlen et al., 1984; Eybergen, 1987). Accretionary wedges are mechanically analogous to the sand wedges forming in front of a bulldozer. As the bulldozer moves forward, the sand wedge continuously grows and maintains a constant surface slope that is related to the rheological properties of the sediment. If a

PROCESSES OF GLACIOTECTONISM

glacier is compared to a bulldozer a problem immediately arises. How can ice, a fairly weak material, push relatively strong and brittle sediment masses to considerable elevations? In the past (e.g., Gripp, 1975) this question has been neglected and glacier profiles have been suggested that are too low to account for the push moraines at their fronts (Van der Wateren, 1985). Since very little is known about the rheologies of unlithified sediments the deformation mechanisms of the sediment wedge must be simplified. In this approximation only bulk deformation will be considered and the details of folding and faulting ignored. Figure 14.16 shows the geometry of a deforming wedge in a horizontal orthogonal coordinate system. The approximation is valid only for variations of stress and velocity in the wedge on longitudinal length scales several times the thickness. Thus the derived equations break down for areas close to the wedge front. The theory borrows concepts from other

437

approaches, such as ‘critical state’ (Davis et al., 1983) and ‘stability criterion’ (Platt, 1986). The following assumptions are utilized in the light of the available field evidence: 1 the wedge consists of two layers, a competent layer sliding on an incompetent basal layer. It is assumed to be mechanically continuous and macroscopically ductile (Price, 1973). In many push moraines the competent units are thicker than the soft basal layers; this geometry will be adopted here; 2 the wedge grows at the cost of the foreland by accretion of sediments to the front; and 3 the wedge is assumed to be in a critical state (Davis et al., 1983, i.e., ‘a critically tapered wedge that is not accreting fresh material is the thinnest body that can be thrust over its basal d´ecollement without any internal deformation; it is thus on the verge of shear failure everywhere. In contrast, a critically tapered

z

zs

dh

rs

a Fw + dFw

Fw

sz sx

h(x) tzx

zb

tb = tc dx

x

FIG. 14.16. Vertical cross-section of a deforming wedge in front of a glacier. Definition of parameters used are in the text.

438

PROCESSES OF GLACIOTECTONISM

wedge that is accreting material deforms internally while sliding, in order to accommodate the influx and to maintain its critical taper’ (p. 1154)). The compressive and gravitational forces are just adequate to overcome the resistance to sliding at the base (Dahlen et al., 1984). If it is assumed that the wedge is at the point of shear failure throughout (i.e., in a critical state), then at any one time the wedge surface strives for a condition of equilibrium between the forces propelling the wedge and those resisting its movement. This is similar to movement of a glacier where, over time, the surface profile is an expression of this equilibrium. Budd (1970) derived an equation for the basal shear stress in a glacier including the effect of longitudinal stress components. Compressive stresses, strains and strain rates are assumed to be negative. The force balance in a deforming wedge

can therefore be derived where, neglecting the longitudinal shear force gradient, the Coulomb shear strength ␶c in the wedge just balances the gravitational and longitudinal stresses (Fig. 14.17): ␶c = ␶ g +

⭸Fw

(14.7)

⭸x

in which the first term on the right-hand side is the gravitational driving stress. This can be envisaged to drive shear flow mainly in the weak basal layer, similar to glacier flow: ␶g = ␳s ghs (x)␣

(14.8)

This is the familiar Nye equation for small surface slopes ␣, where ␳s is the bulk density of the sediments in the wedge. The second term of Eq. 14.7 represents the gradient of the longitudinal normal force Fw

(a) stab ice

dFw =0 dx

le p ro

file

a , h too low

Fw

dFw <0 dx

F

after Platt 1986

(b) out-of-sequence thrust

SEDIMENT

FIG. 14.17. Deforming wedge grows when the potential energy supplied by the glacier increases owing to thickening of the ice under positive mass balance. The wedge develops from a stable profile (a) to an unstable profile with ␣ and h too high in the internal part and ␣ and h too low in the part that is to be added (b) to a stable profile again (after Platt, 1986).

PROCESSES OF GLACIOTECTONISM

(Fig. 14.16) driving compression and extension in the bulk of the wedge: Fw =

冕

zs

zb

冕

(␴x – ␴z )dz = 2

zs

zb

␴⬘xdz = 2hs␴˜

(14.9)

˜ x is the longitudinal deviatoric stress averaged where ␴⬘ over the wedge thickness. It can be shown that the following force balance (simplified from Eq. 14.7) is a reasonable approximation for a portion ␦x of which the length is of the order of hs , the thickness of the wedge: ˜x ␶g + 2␴⬘

⭸h ⭸x

– ␶c = 0

(14.10)

This balance is only valid away from the leading and trailing edges of the wedge. Assume that, at some time in segment ␦x, ␶g > ␶c . ˜ x > 0 (i.e., the Since ⭸hs/⭸x < 0 it follows that ␴⬘ longitudinal deviatoric stress in this part of the wedge) is tensile. If this stress is large enough to overcome the yield strength of the wedge it will extend. Conversely, if ␶g < ␶c then the wedge is under compression. Translating this into terms of changing hs , ␣ and ␶c it can be concluded that: (a) If either or both the surface slope ␣ and thickness hs attain such values that the gravitational driving stress ␶g is larger than the shear strength ␶c , the wedge will tend to extend thus reducing ␣ and hs until equilibrium is restored. If either ␣ or hs or both are too small the wedge will tend to shorten; (b) if the shear strength ␶c of some part of the wedge (e.g., in the basal layer) is reduced the wedge will extend and if ␶c is locally increased the wedge will shorten again. The stable wedge (Platt, 1986) is neither shortening nor extending, according to the stability criterion: ␶g – ␶c = 0

(14.11)

The gravitational force just balances the shear strength but, as the wedge is in a critical state, every

439

departure from equilibrium is followed by either extension or compression of that part of the wedge that has become unstable, which is where the Fw term enters. Horizontal compression by folding and thrusting, or horizontal extension by boudinage and normal faulting serve to maintain a stable surface profile. It will be clear that the configuration of a stable wedge is a convex parabolic surface profile since, like the profile of an ice sheet, it is essentially controlled by the gravitational driving stress, ␶g . The wedge (push moraine) in a way is an extension of the glacier, albeit of a lower profile. Since the density of the sediments is about twice as high as the ice density, the glacier surface near the interface with the push moraine must be twice as steep. Only if the glacier is thicker than the adjacent part of a sediment wedge will it be able to bulldoze the wedge. Thus the glacial bulldozer derives its strength from its thickness. The two most important variables determining slope and dimensions of a push moraine are the rate of advancement of the glacier margin and deformation rate of the push moraine (Van der Wateren, 1992). If the glacier advance is maintained over a sufficiently long period of time, and its rate has the same order of magnitude as that of the deforming sediments, the push moraine will be large. Strong, coarse-grained sediments produce thick and rather steep push moraines, whereas relatively weak, finegrained sediments produce extensive and flat push moraines. High advance rates produce small push moraines.

14.6. PASSIVE DEFORMATIONS Structures produced by static loading by ice on glaciogenic sediment masses are an indication that the ice has stopped moving (Banham, 1975; Brodzikowski and van Loon, 1980, 1983; Eissmann, 1987). These structures do not normally form in the relatively narrow compressive zone near the active ice sheet margin. Older load structures may be overridden and structurally modified by readvancing ice. Two mechanisms have been proposed to explain their origin. First, they may originate beyond the margin owing to loading of saturated muds by prograding sediment masses (proglacial deltas and

440

PROCESSES OF GLACIOTECTONISM

(a)

(b) PLATE 14.4. Thin section of a till from the Central Netherlands (van der Meer, 1987b; van der Meer et al., 1985). Crossed nicols, field width 6 mm from left to right. Rounding of fragments of a clay boudin in a sandy-silty matrix. Silt-clay lamination is a transposed foliation. Riedel shears in boudin cause it to fall apart and rotation of the small fragments produces rounded clay and till clasts testifying for progressive simple shear deformation at very high finite shear strain. (b) Two cleavage sets similar to mylonitic S-C fabric. One set parallel to the transposed foliation is cut by a steeply dipping cleavage. Cleavage asymmetry suggests right lateral (clockwise) shearing.

PROCESSES OF GLACIOTECTONISM

alluvial fans). In that case they are part of an advance or readvance sequence, the latter indicating a temporary glacial retreat (Hart, 1987; Hart et al., 1990). Secondly, they may be produced by loading of saturated tills by stationary ice masses. Surging glaciers are characterized by intensely crevassed snouts disintegrating into numerous blocks of ice, left behind after the surge has terminated. Water-saturated basal till may be squeezed up between the stagnant blocks, forming ‘mud walls’ as described by Gripp (1929) from several Spitsbergen glaciers (Solheim, 1988; Hart et al., 1990; Boulton et al., 1996).

14.7. CONCLUSIONS Several generalizations can be made about sediment structural styles and their distribution in glaciated terrains. Boulton and Hindmarsh (1987) demonstrated that for undrained conditions the apparent viscosity of deforming soft sediments strongly depends on pore fluid content (Chapter 5). Since this is a non-linear relationship, small variations on the porewater pressure have strong effects on the strain rate. The abundance of meltwater under a high hydraulic head in the bed of temperate ice sheets explains the extremely high strain rates in deforming till layers. Other critical components influencing basal debris deformation are sediment porosity and permeability (Menzies, 1989a).

Ee

Ec

Ee

D

441

In nappe shear zones high porewater pressures are produced by the expulsion of water from compressible fine-grained sediments, without replacement from other sources. Since the amount of porewater present will usually be rather low, strain rates will usually be lower than in deformation tills. These theoretical considerations are supported by the structural evidence. Shear zone geometries show great similarities, whether they are generated at the base of a nappe or in the bed of a temperate ice sheet. They differ in the abundance of water escape and liquefaction effects and the amount of finite shear strain. In thin section, deformation tills appear to be more intensely mixed, by alternate folding and attenuation, compared with nappe shear zones. These tills commonly contain fragments testifying to multiple deformation episodes, such as boudins composed of a conglomerate of rounded diamict clasts (Plate 14.4). On a microscale, boudins of clay and silt are more common in tills than in nappe shear zones (Van der Wateren, 1995). The deformed substratum is the ‘fingerprint’ of the various strain regimes of an ice sheet. Plotting glaciotectonic styles on a map of a glaciated area may thus reliably yield locations of ice sheet margins to be used for the reconstruction of former ice sheets. The distribution of these structural styles would give the location of regions of extending and compressing flow, of ice streams and other areas of rapidly flowing ice.

C

B

A

overdeepened glacial basin

FIG. 14.18. Conceptual cross-section of a push moraine, based on the Holmstr¨ombreen push moraine, showing glaciotectonic styles A, B, C, D and E. See text for explanation.

442

PROCESSES OF GLACIOTECTONISM

A synthetic sequence of structural styles in the Holmstr¨ombreen push moraine may serve as a basis for glaciotectonic mapping (Fig. 14.18): A Undeformed foreland: proglacial outwash directly related to the latest pushing event, overlying older sediments. B High-angle structures: ‘Jura-style’ folding of concentric and box folds with vertical and steeply dipping axial surfaces. Thrusts are rare and steeply dipping, with small throw. Minimal horizontal tectonic shortening. C Low-angle structures: strongly asymmetric, overturned and recumbent folds. Low-angle thrusts. Medium shortening. D Nappes: extensive horizontal, relatively thin thrust sheets and internally deformed owing to horizontal compression. Maximum shortening. E (Deformation) tills: boudinage, boudinaged folds and folded boudins, sub-horizontal shear planes, transposed foliation. Extremely high shear strain and horizontal extension. Subdivided into: Ec (compression), comprising stacked till sheets and other compressive structures; Ee (extension), with strong erosion of the substratum, comprising overdeepened (tunnel) valleys, drumlin fields and megaflutes. B, C and D represent terrains of horizontal compression, with increasing strain from B to D, whereas E is characterized by strong horizontal extension as a result of progressive simple shear. Since the flow rate in a deforming till layer decreases towards the margin, compressive structures can be expected to occur in tills that are deposited close to the margin: hence the division of E-style terrains in a purely extensional zone and a compressive zone, Ee and Ec, respectively. Each terrain type can be presumed to show different overprinting relationships if they are produced by either a glacial advance, or a glacial readvance during a general retreat. Two sequences may be distinguished: 1 Advance sequence (Fig. 14.19(a)), in which style E overprints compressive structures B, C, D and A (undeformed foreland). Boulton (1987a) argues that drumlins, cored by coarse-grained sediments, may be the result of overriding and streamlining of

glacial outwash (i.e., tectonic style E overprinting A (E/A)) (Kr¨uger, 1987). Some drumlins in northwest Germany originated as push moraines that have subsequently been overridden (Stephan, 1987). The Rehburg push moraines in western Germany (Van der Wateren, 1987, 1995) and those in the northern Netherlands (van den Berg and Beets, 1987) are push moraines overlain by deformation till. These are examples of style E/(B,C,D). 2 Re-advance sequence (Fig. 14.19(b)). When a readvance occurs during a general retreat, tills and outwash, dating from the previous glacial advance, may be incorporated into push moraines. Such moraines may be quite rare, as they will usually be eroded away by meltwater and overriding ice. Thus B, C and D overprint E style structures. The Lamstedt push moraine in Germany (van Gijssel, 1987) is an example of (C,D)/E. In view of the different amounts of finite strain in compressive structures and subglacial shear zones, observations of interference structures E/(B,C,D) will be quite rare. The former will usually be almost completely obliterated by subglacial shear. Interference structures of successive compressive phases have a better chance of preservation. In these cases overprinting relations provide a means of relatively dating successive deformation episodes.

(a)

Ee/A

Ee/D

Ee/C Ee/B

Ee/A

Ec/A

A

(b) D/E

C/E

B/E

FIG. 14.19. (a) Advance sequence. Style E overprints styles A to D. Triangles denote till belonging to the glacial advance. (b) Readvance sequence. Styles B, C and D overprint style E. Black triangles denote till belonging to older advance. Open triangles denote till belonging to latest advance.

PROCESSES OF GLACIOTECTONISM

Mapping glaciotectonic styles is a powerful tool to reconstruct former ice sheets, if the following reservation is made. Not all types of push moraines are equally useful for a reconstruction of ice sheet margins. If the basement geology is the main factor controlling the position of a line of push moraines,

443

individual ridges do not have to be synchronous and the use of this line for an ice sheet reconstruction is rather pointless. Stable margins are identified by stacked till sheets, dump end moraines, outwash fans and deltas, sometimes incorporated in a large push moraine.

This Page Intentionally Left Blank

15

GLACIAL STRATIGRAPHY J. Rose and J. Menzies Diluvium is chaos (Lossen, 1875)

15.1. INTRODUCTION Stratigraphy is the discipline by which we seek to understand ‘the history of how, when and why glacial deposits . . . came to be what and where they are today’ (Salvador, 1994, p xvii). In other words, glacial stratigraphy provides the methods and procedures by which we can reconstruct the history and patterns of past glaciations and associated environments. For many earth scientists, stratigraphy is the ultimate aim of their subject (Harland, 1992). For glacial geology it should include all the information about processes, geographical distributions, timing and environment of past glaciers and glaciation. Glacial stratigraphy is closely related to Quaternary stratigraphy, which is concerned with the history of the whole environment over the past 2.6 Ma, even though glaciations have occurred on many other occasions in Earth history (Deynoux et al., 1994) (Fig. 1.1). Quaternary stratigraphy includes the study of glaciations of the past. Many features of Quaternary stratigraphy such as ‘glacial’ and ‘interglacial’ subdivisions (Rose, 1989a) are derived from early studies of glaciation and from the fact that glaciers are such critical driving forces in bringing about changes on the Earth’s surface. However, Quaternary stratigraphy also includes the study of other physical,

chemical and biological processes and their interaction. Indeed it is through Quaternary stratigraphy that it is possible to reconstruct former landscapes at any scale and to understand what these landscapes were like, when they occurred and the rates and processes by which they changed. In an area such as Ontario, Canada, or northern England, these changes include glaciation, but in areas such as much of the southwestern USA or southern England glaciation plays no direct part in this history and the striking changes involve variations in river activity, vegetation cover or faunal types. The study of such a wide range of mutually interacting processes gives immense robustness to the scientific procedure and generates an inherent excitement in understanding the different processes, landscape patterns, landscape sensitivities and rates of change that may have occurred in a given area over a given period of time. Nevertheless, this chapter is concerned primarily with stratigraphic processes and procedures associated with glaciers and closely related processes, and reference to Quaternary stratigraphy will only be made when necessary. It is, however, exceedingly important to remember that glaciation and glacial processes drive so many other activities, such as the behaviour of meltwater rivers, and the rates of dust production and deposition, the presence of lakes and sea level, deformation of the Earth’s crust, the extent of permafrost, plants, and animals, and global atmospheric circulation.

445

446

GLACIAL STRATIGRAPHY

Yet, despite the obvious importance of glacial stratigraphy, many of the core aims have, in the past, been obscured by unnecessary detail. The result has been that the subject has, in many cases, become an interminable catalogue of till formations associated with glacial lobe advances and retreats, stillstands, periods of ice stagnation, and of sedimentary units caused by lowering and rising sea and lake levels, all bound up in complex network diagrams (Mickelson et ˘ al., 1983; Fulton, 1984; Sibrava et al., 1986). These complex diagrams constructed, at times, upon questionable dates or poorly understood glacial sedimentology or geomorphology form the basis of many glacial reconstructions and histories of glacial development. Traditionally these schemes have acquired an immutable status that has been defended at ‘all costs’, or have been modified reluctantly retaining an outmoded structure. The result is that too often revisions, instead of bringing greater lucidity and understanding, have wrought even more confusion and further mystification (Rose and Schl¨uchter, 1989). This chapter reviews the principles behind glacial stratigraphy, and will outline the appropriate stratigraphic methods available and current geochronological techniques used for correlation and dating. Emphasis will be placed on the methods and the associated complexities and stratigraphic problems. 15.2. RATIONALE Stratigraphy attempts to determine the concise chronological sequence of geological events over a wide area, as manifested in periods with a relatively distinctive or characteristic geological property. For glacial stratigraphy these properties typically relate to glaciogenic sediments or landforms. Historically the subject was dominated by morphostratigraphy that emphasized landform assemblage units and their spatial arrangement and diversity (Sissons, 1967, 1976; Lowe and Walker, 1984), but has since changed towards a more balanced approach based around glacio-sedimentology emphasizing lithofacies types and associations, landform/sediment assemblages and strictly controlled geochronometry (Rose, 1985). Stratigraphic relationships are considered within the context of Walther’s Law (Middleton, 1973).

Essentially, the law states that in a conformable succession of sediments only those lithofacies that exist in a vertical sequence occur adjacent to each other in nature. In other words, spatially contiguous lithofacies environments occurring across the Earth’s surface can be found in the same vertical succession as stacked lithofacies units. Individual depositional systems can be, therefore, separated within a stratigraphic package and are bounded by unconformities or sharp facies transitions into adjacent (spatial or temporal) unlike systems (Edwards, 1986; Brodzikowski and van Loon, 1991). Within a terrestrial glaciogenic sequence of sediments facies, transitions and unconformities are very common over short distances and over very short time spans, and glaciotectonic processes may transpose sedimentary units. It is these spatio-temporal variations, perhaps above all else, that complicate glacial stratigraphic interpretations. In glaciomarine sequences many of the problems inherent in land-based glacial stratigraphic differentiation are less acute. However, much of the stratigraphical evidence in the marine environment is circumstantial and rarely provides direct evidence of glacial action (Elverhøi, 1998). An additional problem in glacial stratigraphy is the tendency for only peripheral, marginal and distally linked sediments to record past evidence of glacial sedimentological systems. The preservation potential of glacial sediments tends to be limited, providing spatially fragmentary and disparate stratigraphic sequences at best. Much evidence of activity within the global glacial system is recorded in the form of proxy evidence such as isotopic signals in ocean sediments. There are no single lengthy records on land of direct glacial activity. This is in contrast to ocean records of almost uninterrupted successions of glacial presence on the globe (Sarnthein et al., 1984; Jansen and Sjøholm, 1991) (Fig. 15.1) (Chapter 2). Towards the centres of past ice sheets the stratigraphic record becomes increasingly limited, although whether this is a function of persistent glaciation or effective glacial erosion is not clear (Hirvas et al., 1988; Garcia Ambrosiani and Robertsson, 1992). The preservation potential of glacial sediments apparently reaches a maximum in low lying terrains on the peripheries of major ice sheets where possibly erosional activity by ice has been minimized, perhaps

GLACIAL STRATIGRAPHY

No. of IRD cm-2kyr-1 100,000

10,000

1,000

100

0 0.6 1.0 1.4

644A

1.8 2.2 2.6 3.0 3.4

Age (Ma)

HIATUS

447

3.8 4.2 642B

4.6

owing to low marginal ice profiles and higher depositional rates associated with highly deformable transported sediment. The other likely preservation sites are zones of minimal erosion close to low velocity ice at ice-shed locations (Hattestrand and Kleman, 1999) and chance locations where pockets of glacial sediment have been injected into or infill deep depressions such as weathering hollows or joint fissures. Those areas where saprolites have been detected in, for example, the eastern townships of Quebec and upper New Brunswick, or in the northeast of Scotland, are likely locations for well-preserved glacial stratigraphies. However even in these locations, the sedimentary units are discontinuous and often difficult to place within a stratigraphic framework. Where organic remains occur, biostratigraphy may be used to determine the Quaternary history of a region, or geochronometric methods such as radiocarbon, U-series, luminescence and/or exposure age dating may be applied to derive a timescale. But neither biostratigraphy nor geochronometry may tell anything directly about glacial processes or the associated glacial environment. These methods simply provide a biostratigraphic, geochronometric or chronostratigraphic framework into which glacial events may be slotted. Despite the obvious unsatisfactory nature of this approach it is these methods that have, hitherto formed the framework for the history of glaciation in regions such as Britain (Bowen et al., 1986a; Rose, 1989a; Ehlers et al., 1991) and Switzerland (Schl¨uchter, 1992) and northern Europe (Ehlers, 1983; Ehlers et al., 1995). In order that interpretations upon which glacial stratigraphies are based should have some credibility, it is essential that individual glacial processes

5.0 5.4 5.8

FIG. 15.1. A record of ice-rafted debris (IRD) deposited in the sediments of the Norwegian Sea over the last six million years. This is a continuous lithostratigraphic record of glaciation of Greenland over the period concerned. The results are taken from ocean cores 644A and 642B located, respectively, at 66°40.7⬘N, 4°34.6⬘E and 67°13.5⬘N, 2°55.7⬘E. Note that the y-axis is expressed as a logarithmic scale. The onset of intensive glacial activity around 2.6 Ma is very clear, but it is also apparent that Greenland was glaciated and IRD was discharged into the surrounding oceans in the preceding 3 Ma (from Jansen and Sjøholm, 1991; reprinted with permission from Macmillan Magazines Ltd.).

448

GLACIAL STRATIGRAPHY SCHEMATIC SECTION

20

DESCRIPTION

(CURRENT TERMINOLOGY)

Top soil Coversands

Devensian aeolian deposits Debris flow (derived from exposures of lower diamicton)

Mildenhall upper sands and gravels

Anglian glaciofluvial outwash deposits

High Lodge clayey-silts and High Lodge sands

Pre-Anglian fluvial and lacustrine sediments contorted, sheared and interbedded with lower diamicton as a result of glaciotectonic disturbance/transport

Lower sands

Mildenhall lower sands

Glaciofluvial outwash deposits

Lower diamicton

Mildenhall lower diamicton

Lodgement till deposited by ice during the Anglian glaciation

Chalk

Chalk

Sands

Sands and gravels 10

Contorted sands Silty sands and lag gravels Upper brown clayey-silt

5

0 Metres

INTERPRETATION

Breckland coversands Mildenhall upper diamicton

Upper diamicton

15

LITHOSTRATIGRAPHIC CLASSIFICATIONS

Grey clayey-silt Lower brown clayey-silt with interbedded sands

FIG. 15.2. An example of lithostratigraphic subdivision, High Lodge, eastern England. The site includes Middle Pleistocene glacial and non-glacial sediments, and is affected by glaciotectonism. Two formations are represented: the High Lodge Formation is the earlier, and is composed of ‘pre-glacial’ river sediments with a lithological assemblage determined by the rock within the river catchment. The Mildenhall Formation is the later and represents glacial, glaciofluvial and debris flow sediments with a lithological assemble including glacial debris transported to the site by glaciers from regions well beyond the ‘pre-glacial’ river catchment. Because of glaciotectonism the normal order of super-position has been greatly complicated and a kineto-stratigraphic interpretation is also appropriate with the units below the Mildenhall upper sand and gravel number being glaciotectonized, whereas this unit, along with the Mildenhall upper diamicton member, is unaffected by glaciotectonism, having been formed during and following ice wastage. It should be noted that the lithostratigraphy does not fully coincide with the kineto-stratigraphy (reproduced from Lewis, 1992, in Ashton, Cook, Lewis and Rose (eds), British Museum Press).

must be understood. It is also essential that stratigraphic subdivisions and classifications should be based on observed properties and not inferred genesis. Thus it is the body of sediment or the shape of the landform that is the critical factor, not the process by which it was formed. However, in reality this is impractical, as description alone without a genetic evaluation would be overwhelming and inhibit communication and scientific progress. The result is that an iterative process exists whereby observation is followed by interpretation that is then followed by a classification and subdivision of the observed evidence based on knowledge of the process. Thus, for instance, a succession of diamicton, sand and gravel, and laminated

silts followed by diamicton, would be analysed and interpreted and, depending upon the outcome of the interpretation, the stratigraphic succession could represent either: (1) one glacial sequence, with the lower diamicton being a till, the upper diamicton being a debris flow derived from the lower till separated by meltwater, and proglacial lake sedimentation (in which the units may be members of one formation); or (2) two glacial sequences in which the two diamictons represent two glacial episodes separated by non-glacial conditions (in which case the lower till, sand and gravel and laminated silts would be members of one formation, and the upper diamicton would represent a separate formation) (Ashton et al., 1992) (Fig. 15.2). Even

GLACIAL STRATIGRAPHY

in this simple case judgement is required, and alternative subdivisions may be proposed. This emphasizes the critical importance of accurate description and critical understanding of the processes of formation. 15.3 STRATIGRAPHY WITHIN GLACIAL ENVIRONMENTS The glacial environment is perhaps one of the most complex, dynamic and least well-understood sedimentological environments. Glacial events can occur with devastating impact, as in the case of surges or j¨okulhlaups that may cause profound deformation of underlying rocks and sediments and transport tonnes of sediments in a matter of a few hours or days. Conversely, other glacial processes such as distal lake or marine sedimentation, or subglacial erosion involve much slower rates and their effects can be enduring. Consequently, within any glacial environment a vast range of sediments, depositional mechanisms and sedimentation rates may be found juxtaposed. Imposed upon glacial sediments are the effects of subaerial non-glacial processes including pedogenesis and diagenesis. These processes may modify extensive or very restricted areas of glaciogenic sediments and landforms, over periods of time that range from hours to centuries and millennia. Subaerial processes will occur under differing climatic, vegetational and pedological regimes causing a wide variety of effects that can range from barely perceptible modification to total alteration or removal. Without adequate knowledge of the extent of erosion, subaerial postglacial sedimentation, pedogenesis or diagenesis, stratigraphic discrimination can be extremely difficult. A further problem for glacial stratigraphy is that many glaciated landscapes along the margins or in lowland central areas of Pleistocene ice sheets have been submerged by either freshwater or seawater over varying periods of time. In some places, such as the area submerged beneath Glacial Lake Agassiz in North America (Teller and Clayton, 1983), the lowlands around the Gulf of Bothnia in Scandinavia or the lowlands around Hudson Bay in Canada, sufficiently long periods of inundation have resulted in extensive and thick glaciolacustrine and glaciomar-

449

ine sedimentation or shoreline erosion, so that the original evidence of glaciation is obscured or removed. A persistent problem in understanding glacial stratigraphic relationships are the difficulties of correlating glacial events, both over small distances within the area of one glacial system, or over long distances between one glaciated area and another. Primarily these problems arise because of the heterogeneity of most glaciogenic sediments and the inherent lithological independence of any individual glacial system They also originate because different ice masses, and even different parts of the same ice mass respond to climate at different rates and with dissimilar geographical expressions. Indeed, glacial processes rarely operate with any degree of synchroneity even within temporal boundaries of millennia (Menzies, 1995, chapters 2, 3 and 4). The problems of correlation within a single or a number of closely linked glacial systems are illustrated in the Great Lakes basin of North America where an extensive and elaborate stratigraphic framework, originally based upon separate till sheet recognition has been developed over the past 50 years (Mickelson et al., 1983; Karrow, 1984a) (Fig. 15.3). Several methods of correlation have been made. In many cases the till sheets were recognized stratigraphically on a ‘layer cake’ stratigraphic position. It being assumed that the lowest till unit was stratigraphically equivalent over several thousands of square kilometres, and thenceforth in ascending order. Other tills were correlated on the basis of lithologic composition or the presence of distinctive suites of indicator lithologies that could be traced to particular sources. Thus, in extreme southern Ontario, the Catfish Creek till (Gibbard, 1980) was recognized on the basis of its stratigraphical position immediately overlying bedrock, its typically coarse-grained content and yellow colouration. However, as the complexity and vicissitudes of till depositional mechanics become better known, it has become understood that ‘counting-up’ or simple lithostratigraphical methods of stratigraphic correlation are inherently weak. It is now understood that many of the correlations are invalid and the fluctuations of the many ice lobes at the southern margin of the Laurentide ice sheet are essentially asynchronous.

450

GLACIAL STRATIGRAPHY

TWO CREEKS INT.

LATE WISCONSINAN

PT. HURON ST. MACKINAW INT.

HURON - GEORGIAN BAY LOBE

LAKE ERIE

TORONTO

ST. NARCISSE

12 ka HALTON T.

ST. JOSEPH T. 13 ka RANNOCK T.

WENTWORTH T.

ELMA T.

PORT BRUCE ST.

WARTBURG T. STRATFORD T. MORNINGTON T.

PORT STANLEY T.

TAVISTOCK T.

MARYHILL T.

STIRTON T. ERIE INT.

EARLY WISCONSINAN MIDDLE WISCONSINAN

NISSOURI ST. PLUM PT. INT.

ST. LAWRENCE VALLEY

CATFISH CREEK T.

CATFISH CREEK T.

25 ka

CHERRYTREE ST.

GENTILLY T.

TIME

THORNCLIFFE MEADOWCLIFFE T.

?

FM. SEMINARY T.

PORT TALBOT INT.

42-54 ka

GUILDWOOD ST. ST. PIERRE INT.

DUNWICH T.

?

BRADTVILLE T.

80 ka

SUNNYBROOK T. POTTERY RD. FM. SCARBOROUGH FM.

NICOLET ST.

ST. PIERRE FM. BECANCOUR T.

125 ka SANGAMONIAN

DON FM. PFK 1981

FIG. 15.3. Glacial lithostratigraphy units (T, tills) and ice fluctuations in the region of the southern Laurentide ice sheet from Lake Huron to the St Lawrence valley for the Last Interglacial (Sangamonian) and Last Glaciation (Wisconsinan). Note that the timescale is variable, and in radiocarbon years where possible. Note also the diachroneity of the lithostratigraphic boundaries based on the presence of tills. These are classical examples of diachronic units (reproduced from Karrow, 1984).

Similarly, in the western part of the Midland Valley of Scotland, till colour has been used to suggest that two different glacial episodes are represented stratigraphically (Menzies, 1981b). A similar method, but including till provenance indicators and particle size distribution, was used in separating three tills suggesting three glacial advances near Arthur, Ontario (Cowan et al., 1978) (Fig. 15.4). None of the above criteria can be used with any confidence unless there is a clear scientific rationale for the development of different glacial lithostratigraphic units, such as a reason for the introduction of new lithologies into separate till units.

A marine transgression occurred during the interstadial (ca. 13 500–11 000 14C years BP) before the Younger Dryas (ca. 11 000–10 000 14C years BP) in the Loch Lomond basin, western central Scotland. Here the till units of the Dimlington glaciation and the Loch Lomond glaciation can be clearly defined and distinguished on lithostratigraphic criteria (Rose, 1989a) (Fig. 15.5, Table 15.4) as only the younger glacial deposits incorporate the marine sediments. The problem of depositional hiatus is of critical importance. Hiatuses are not restricted to glacial environments but are particularly common within them. For example, within the subglacial terrestrial

GLACIAL STRATIGRAPHY

A

Overburden Outwash

451

B

Overburden Till 1

Till 1

Till 2

Till 2

Till 3

Base of Section

Base of Section

FIG. 15.4. Diamicton facies and associated sediments in the area around Arthur, Ontario (adapted from Cowan et al., 1978).

environment, periods of limited deposition, periods of equal deposition and erosion, or periods of net erosion may result in limited to zero sedimentation for varying periods of time over differing areas of the glacier bed. The recognition of localized or widespread depositional hiatuses remains difficult to substantiate without, for example, a distinct weathering horizon or a paleosol immediately below the suspected hiatus. Sediment reworking and re-sedimentation is a difficult and common problem in glacial stratigraphy. The term ‘re-sedimentation’ tries to convey the concept of sediment being eroded from a primary

depositional site, transported some varying distance from this initial source and again deposited (Ashton et al., 1992) (Fig. 15.2). The distinction between primary and secondary deposition remains vague but has been used to attempt stratigraphic division between different stages of deposition (Boulton and Deynoux, 1981) or to reduce the number of stages of deposition when individual diamicton units have been traditionally considered evidence for separate glacial events. It is a logical consequence of the above problems that even at a scale of a few kilometres, sediments from the same ice mass may defy stratigraphic

452

GLACIAL STRATIGRAPHY Loch Lomond

6

7

1

4

3 2

5

N0

of

DUMBARTON Inverleven

Clyde

km

moraine ridges extent of Loch Lomond Readvance sites

5

14C

yrs BP

organic detritus 10,560 ±160 14C yrs BP

Gartness4 & Drumbeg7

5

GLASGOW

Gartocharn Till

peat

Wilderness Till

Wilderness Till

11,700 ±170

Aber1 & Balloch2

Main Lateglacial Shoreline Clyde Beds 10,920 ±136

Blane Valley Silts

Gartocharn Till

Lake Blane

Kilpatrick Hills Firth

Muir Park Reservoir6

Croftamie3

Rhu Sands & Gravels

ENDRICK VALLEY

O.R.S.

LOCH LOMOND

14C

yrs BP

Blane Valley Silts Gartocharn Till

(deformation facies)

Clyde Beds Gartness Silts

gyttja pink clay gyttja pink clay 10,010 ±230 14C yrs BP 12,060 ±320 14C yrs BP 12,510 ±310 14C yrs BP

KILPATRICK HILLS/ MUIR PARK

Wilderness Till

BLANE VALLEY

FIG. 15.5. Diagrammatic representation of the stratigraphic evidence for the Loch Lomond glaciation in the southern part of the Loch Lomond basin. The inset map shows the location of the lithostratigraphic sites and the position of the Loch Lomond Readvance moraine, which is a morphostratigraphic unit. The stippled area shows the extent of ice cover at the maximum of the glacial event and the horizontal lines show the extent of the proglacial Lake Blane that formed at the same time (reproduced from Rose et al., 1988).

correlation unless there is continuity between sites. Processes of cut-and-fill, sediment reworking and glaciotectonic deformation all create highly complex stratigraphic sequences and arrangements. The result is that the relatively simple stratigraphic procedures of non-glacial stratigraphy are not applicable and stratigraphic correlation based upon single units must be approached with caution (Karrow, 1984b). The problem of correlation between different glacier systems on different parts of the globe is equally complex, but for different reasons. An example of this problem is the attempts to correlate the glacial response to the Younger Dryas climatic deterioration at the Last Glacial/present Interglacial transition (Bard and Broecker, 1992; Peteet, 1993; Lowe et al., 1994). In northwestern Europe the evidence for glacier expansion at this time is well

represented by glaciogenic sediments in the region close to the North Atlantic, whereas in more continental parts of Europe, evidence is mainly found in the form of proxy evidence such as lithological variations or oxygen isotope signals in lake sediments. In North America, evidence for glacier expansion is very variable. This range of evidence reflects the scale of climatic change at that time and the proximity of the region concerned to the energy source that amplified the change. It is evident that in attempting to develop glacial stratigraphies there are many major problems that have no simple solution. Obviously, with an understanding of glacial processes and environments, and resulting lithofacies, it will be possible to develop reasonable stratigraphic approximations that may subsequently be altered, revised or scrapped if new information appears.

GLACIAL STRATIGRAPHY

15.4 STRATIGRAPHIC NOMENCLATURE A problem that distinguishes glacial sediments from other sediments is their geologically short time-span and their enormous lithologic variability. Fossil assemblages are virtually absent from the glacial environment, and biostratigraphy is not part of glacial stratigraphy per se, although the presence of organic remains in adjacent non-glacial or proglacial sediments may provide proxy evidence of glacial conditions (Penny et al., 1969), and is generally the basis of a chronostratigraphic structure (Mitchell et al., 1973) (Figs. 15.3 and 15.5). Landform stratigraphy may play a very important role when subaerial modification of the glacial terrain is minimal. Likewise, soil stratigraphy may provide a range of evidence for reconstructing past environments and correlation of glacial episodes. In addition to the above stratigraphic methods, correlation is also provided by a range of geochronometric techniques (Menzies, 1996, chapter 14). The highly complex nature of glacial stratigraphy has meant that the use of the International Stratigraphic Guide (Salvador, 1994) and the North American Stratigraphic Code (NASC) (North American Commission on Stratigraphic Nomenclature (NACSN), 1983) are of restricted value. The result has been that Quaternary stratigraphy has adopted, as a matter of course, a much wider range of stratigraphic methods and a much more refined, rigorous ˘ and sensitive stratigraphic scheme (Sibrava et al., 1986; Fulton, 1989). Likewise, the need to deal with frequent hiatuses in sedimentary sequences has meant that much thought has had to be given to stratigraphic delimitation (Cowie et al., 1986; Rose, 1989a). Lithostratigraphic units form the critical building blocks of glacial stratigraphy. Where appropriate these units are combined with morphostratigraphic units. These units may be supported by proxy evidence of glaciation that can be provided by chemical stratigraphic units as indicators of enhanced (glacial) erosion and sedimentation, and isotopic stratigraphic units as indicators of ice volume. Similarly some biostratigraphic units (especially based on insect content) may provide proxy evidence for glacial conditions. A wide range of other lithostratigraphic evidence derived from loess, marine and

453

lake sediments may act as proxy indicators of glacial activity. These fundamental stratigraphic units have been used in glacial stratigraphy to provide evidence for glacial activity in the form of glacial advance, standstills and retreats, and essentially this, along with dating, is the critical aim of glacial stratigraphy. However, it is at this stage that glacial stratigraphy becomes intimately associated with Quaternary stratigraphy, which is based essentially on climatostratigraphy defined in terms of glacials and interglacials. In Quaternary stratigraphy these climatostratigraphic units form the main elements of the chronostratigraphic subdivision (Stages), and are used irrespective of whether the evidence suggests glacier cover or simply a significant change of climate without any evidence for the presence or absence of ice. The use of geochronometry to place ages on Quaternary events provides the basis of a timescale by which events may be dated and by which it is possible to determine the rates at which processes operate. For the purpose of glacial stratigraphy, a glacial stage is recognized as a major expansion of glaciers of long duration, and glacials are divided into stadials and interstadials. Stadials are defined as a limited expansion of glaciers or a subdivision of a glacial characterized by a relative deterioration of climate, and interstadials are defined as a period of relative improvement of climate but of lesser duration or lesser vegetational development than the present interglacial (Holocene) (Rose, 1989a, p. 47). Glacials are separated by interglacials, which have been defined in terms of the vegetational development (West, 1977; Karrow, 1989; Klassen, 1989). By this definition it is required that the vegetation of an interglacial should be equivalent to that which developed in a region at the thermal optimum of the present (Holocene) interglacial. These terms may apply in areas that are not glacierized and may never have been glacierized. These climatostratigraphic terms are the equivalent of ‘event stratigraphy’ used in other parts of the geological column (Salvador, 1994, p. 117). Altogether these terms create considerable confusion. First, although the terms are defined in relation to climate, some of the elements, such as interglacials, are based on climatic proxy such as vegetational

454

GLACIAL STRATIGRAPHY

development represented by pollen assemblage biozones that may be unreliable indicators of climate. For instance, vegetation will not only reflect climate but also the competition between the available species, soil productivity and time needed for colonization of different species. As an example it is known that summer temperatures in Britain about 13 000 14C years ago were warmer than now, but Britain was virtually treeless at the time because trees had not yet reached Britain from the refuges in which they survived the very harsh climate of the Last Glacial Maximum (LGM) (Walker et al., 1993). Secondly, some of these terms imply duration, which is of course not climatic. By definition, a stadial is shorter than a glacial of which it is part, and it is expected that an interstadial will be shorter than an interglacial (Lowe and Walker, 1984, p. 8). However, this is clearly not the case with the Upton Warren Interstadial Complex in Britain, which covers a period from about 45 000 to 26 000 14C years BP, roughly twice as long as the present interglacial. Indeed any interstadial recognized for the interval equivalent to Oxygen Isotope Stage (OIS) 3 is likely to be longer than Interglacial OIS Stage 5e. Thirdly, climatostratigraphic terms are used as the basis of chronostratigraphy (stages), yet, as they are climatically defined, it is highly improbable that the evidence for boundaries upon which these stages are defined represents different periods of time and is asynchronous (Lowe and Gray, 1980) (Fig. 15.3). For example, the sediments representing the beginning of the LGM (the Dimlington Stadial or Dimlington Chronozone – a sub-stage of the British Devensian Stage) is considered to have begun deposition about 26 000 14C years ago (Rose, 1985). Clearly, in areas where glaciation persisted throughout the Last Glaciation the evidence for the beginning of this climatic subdivision is likely to be much earlier, whereas in areas that were not so sensitive to climate deterioration, such as southeast France, the evidence for climate cooling would be later (Serat et al., 1990; Pons et al., 1992). Although a chronostratigraphic boundary can be defined at the stratotype (Figs 15.3 and 15.6), the likelihood of obtaining climatostratigraphic evidence from different climatic or geomorphic provinces to represent a synchronous boundary is very difficult. This is a problem for all chronostratigraphy that is not

controlled by geochronometry, especially that which is defined by pre-Quaternary biostratigraphy. However, the level of resolution at these earlier periods is poor and the problem is less acute. In Quaternary glacial stratigraphy a very high resolution is expected and this problem may have serious ramifications for understanding glacial activity. These problems were discussed by Watson and Wright (1980) and it is for the above reason that the NASC (NACSN, 1983) introduced the concept of diachronic units. A diachronic unit is defined as ‘a unit which comprises unequal spans of time represented either by a specific lithostratigraphic . . . or pedostratigraphic unit, or by an assemblage of such units’ (NACSN, 1983, p. 870). It is clear that these units are particularly appropriate to glacial stratigraphy. The purposes of diachronic units are clearly set out in the NASC (NACSN, 1983 pp. 870– 871). ‘Diachronic classification provides: (i) a means of comparing the spans of time represented by stratigraphic units with diachronic boundaries at different localities; (ii) a basis for broadly establishing in time the beginning and ending of deposition of diachronous stratigraphic units at different sites; (iii) a basis for inferring the rate of change in extent of depositional processes; (iv) a means of determining and comparing rates and durations of deposition at different localities, and (v) a means of comparing temporal and spatial relationships of diachronous stratigraphic units.’ The boundaries of the diachronic units are the times recorded by the beginning and end of deposition of the material evidence, and one or both of the boundaries may be time-transgressive. A diachron is the fundamental . . . unit and if a hierarchy of diachronic units is needed the terms episode, phase, span and cline, in order of decreasing rank, are recommended. The rank of the unit is determined by the scope of the unit . . . and not by the timespan represented by the unit at a particular place . . . An episode is the unit of highest rank and greatest scope in hierarchical classification. If the ‘Wisconsinan Age’ were to be redefined as a diachronic unit, it would have the rank of episode (NACSN, 1983, pp. 870–871). The concept and practical exercise of establishing stratotypes has been introduced in order that the

GLACIAL STRATIGRAPHY

455

MATERIAL UNIT C & DIACHRONIC UNIT C

Geochronologic & Chronostratigraphic Unit

MATERIAL UNIT B

MATERIAL UNIT A & DIACHRONIC UNIT A

FIG. 15.6. Schematic comparison of diachronic units with geochronological and chronostratigraphic units (reproduced from NACSN, 1983).

attribute of stratigraphic units can be defined unambiguously, and in order to facilitate communication between sites and between scientists. Stratotypes only apply to stratigraphic units that have been named formally (Salvador, 1994, p. 14), and a stratotype is defined as ‘a specific interval or point in a specific sequence of rock strata and constitutes the standard for the definition and characterization of the stratigraphic unit or boundary being defined’ (Salvador, 1994, p. 26). There are many varieties of stratotype (Salvador, 1994, p. 28) defined as follows: 䊉 䊉

holostratotype – the original stratotype; parastratotype – supplementary stratotype supplying additional information, especially about diversity of the unit;

䊉 䊉

䊉

lectostratotype – used in the absence of an adequately designated holostratotype; neostratotype – a replacement for the original holostratotype, which has been destroyed, covered or otherwise made inaccessible; hypostratotype – an additional stratotype used to extend knowledge of the holostratotype but it is always subsidiary to the holostratotype. 15.5 GLACIAL STRATIGRAPHIC PROCEDURES AND METHODS 15.5.1 Lithostratigraphy

In glacial geology lithostratigraphy provides the fundamental approach to any stratigraphic frame-

456

GLACIAL STRATIGRAPHY

TABLE 15.1. Lithostratigraphic terminology and facies units Group Formation Member Bed

Two or more formations Primary unit of lithostratigraphy Distinctive lithologic entity within a formation Distinctive lithologic layer within a member

work. It includes within it a number of stratigraphic variants that are concerned with other aspects of sediment lithology, but have become sufficiently distinctive or important to be known by a separate term. Included amongst these, of direct or indirect relevance to glacial stratigraphy, are kineto-stratigraphy, chemistratigraphy, isotope stratigraphy and magnetostratigraphy. Lithostratigraphy ‘deals with the description and systematic organisation of . . . rocks . . . into distinctive named units based on the lithological character of the rocks and their stratigraphic relations’ (Salvador, 1994, p. 31). A lithostratigraphic unit is a ‘body of rocks that is defined and recognised on the basis of its observable and distinctive lithological properties, or combination of properties and its stratigraphic relations’ (Salvador, 1994, pp. 31–32). A lithostratigraphic unit should be capable of being mapped and typically be a tabular entity. However, this is often difficult, but certainly not impossible, with glacial sediments because they are often of indeterminate lateral and vertical extent and may suffer intense deformation and intercalation (Maltman, 1994). Lithostratigraphic units are organized within a hierarchical system where the formation is the primary unit (Salvador, 1994, p. 33), and a group is a number of formations. Formations include members, which are composed of beds. These are outlined in Table 15.1 and reference is given to the appropriate lithofacies unit. A formation or facies association is far from easy to define, but should form a body of sediment or sediment facies that has distinctive lithological properties and represents a distinctive sedimentological environment that can be segregated from others on the basis of all or a combination of the following observable properties: particle size, particle shape and surface texture, sedimentary structures, bedding characteristics, geotechnical properties, lithology and

Facies group Facies association Facies unit Subfacies unit

mineralogy. Major unconformities would not be expected to occur within a single formation but smaller unconformities and disconformities are typical. Identification and designation of formations in glacial sediments is difficult because of rapid and complex lateral and vertical gradation. For example, a diamicton association may interfinger with a sand and gravel association. Separately, these associations may have sufficient importance in distinctiveness, thickness, extent and geomorphological significance to justify the designation as formations but, because of the interdigitation, it may be difficult to separate them in the field or on maps and it is apparent that they were associated during deposition. In this case it is more sensible to bring the two facies associations together and call them a single formation. This is the case with the Lowestoft Formation in eastern England, which represents a lithologically distinctive set of facies associations formed by glaciers and meltwater during the Anglian glaciation (Rose and Allen, 1977) (Plate 15.1). In this case the Lowestoft Formation includes tills, outwash sands and gravels, and proglacial lake sediments (Allen et al., 1991). A member is a ‘part of a formation. It is recognised . . . because it possesses lithologic properties distinguishing it from adjacent parts of the formation. No fixed standard is required for the extent or thickness of a member’ (Salvador, 1994, p. 34). Because of the highly complex nature of the glacial sedimentary environment, in comparison with the relatively simple sedimentary environments associated with other processes, members or facies units have become the most crucial elements in understanding glacial sedimentological environments and thus the development of a stratigraphic framework. It is at this level of stratigraphic nomenclature and decision making that most problems occur. If a member is mistakenly identified then any particular formation within which the

GLACIAL STRATIGRAPHY

457

PLATE 15.1. Example of glacial facies association in the Lowestoft Formation of eastern England. This sections shows examples of different till and sand and gravel facies determined by glacial transport path, transport position within, on, or below the ice and depositional process. The site is in the Chelmsford area of East Anglia.

member resides is potentially mis-identified and an inaccurate stratigraphic framework may be erected. The recognition of individual members or facies units may not be an easy task. There are no diagnostic characteristics that can specify whether a specific unit should be designated a member. Likewise, the status of a facies unit may change as new exposures become available, and the extent of particular lithological elements changes, or new relationships are found between the initial unit and newly discovered units. Consequently, it is critical that all characteristic features, parameters and properties of a unit should be identified and tabulated. Thus, even if the initial interpretation is proven to be mistaken, the data can be re-interpreted at a later date. In order to achieve this, a set of objective descriptive methods have been adopted (Rust, 1978; Collinson and Thompson, 1982; Eyles et al., 1983; Eyles and Miall, 1984; Tucker,

1986; Kemmis, 1996). In addition, no single section or site should ever be designated part of a stratigraphic framework unless all other contextual information (other sites, landforms) is mapped and appraised. Despite this, an evaluation of the stratigraphic status and process of formation of members requires individual judgement and this may lead to controversy and discussion, although such debate inevitably strengthens any final decision (Shaw, 1987). An example of this problem can be seen from studies of the large sections of the Scarborough Bluffs on the north shore of Lake Ontario, Canada. Here, the facies units have been described at different levels and explained by several widely different interpretations, with the result that a number of different members and formations have been proposed (Karrow, 1967, 1974; Eyles et al., 1983; Dreimanis, 1984). Similarly, the facies units revealed along the East Anglian coast of

458

GLACIAL STRATIGRAPHY

Norfolk, England, have led to long and detailed debates over a period of more than 100 years and a succession of different processes of formation have been suggested (Reid, 1882; Banham, 1975; Perrin et al., 1979; Boulton et al., 1984; Lunkka, 1988; Eyles et al., 1989; Hart and Boulton, 1991; Lee, 2001). However, in this latter case, although members have been re-interpreted on several occasions according to process of formation, description has been of a relatively high quality and the main observed properties have been sufficiently distinctive that the deposits continue to be allocated to the units that are known as the Lowestoft Formation and the North Sea Drift Formation (Hart and Boulton, 1991). Even this is now under re-examination as a result of detailed field mapping of lithostratigraphic units (Hamblin et al., 2000). Similar interpretative problems with reference to the Precambrian rocks of the Port Askaig Formation in western Scotland are discussed in Chapter 13. Figure 15.7 illustrates the types of problems that assail the stratigrapher in the field. If only Core I is examined, a conclusion may be reached that is different from that which would be reached when Cores I, II and III are examined together. Core I shows a section immediately above bedrock in which a dense clay-rich diamicton unit, exhibiting structures indicative of massive deformation, is conformably overlain by a dense clay-rich matrix-supported diamicton containing a well-oriented clast fabric; that is, in turn, conformably overlain by a coarser-grained diamicton containing evidence of occasional dropstones and structures associated with sediment rainout within a subaqueous environment. Core II, 78 m east of Core I, has a similar vertical profile except in the upper unit where a conformable but interfingering sandy unit of fine to medium massively bedded sands with occasional diamicton balls is seen to exist. Core III, 320 m east of Core II, reveals the presence of a deformed diamicton and matrix-supported diamicton overlain by well-stratified fine to medium sands that have evidence of being foreset and bottom-set beds of deltaic origin. If not taken in context such a sequence might be interpreted in at least two quite different ways: (1) Core I alone might have been seen as evidence of a land-based ice sheet with subglacial sediments

replaced at the top of the section by conformable sediments deposited in a localized small pond or lake possibly itself existing in a subglacial cavern; (2) if the cores are taken together the evidence would suggest ice marginal subaqueous sedimentation occurring in a large proglacial lake. In this case the deformed diamicton is interpreted as being extruded into the lake or deposited just up-ice of a groundingline prior to lake water incursion. In either case the diamicton would be conformable with a debris flow deposit that emanated from the grounding-line into a body of water, and that this debris flow was, in turn, buried by typical waterlain diamicton sediments. Such a scenario would be confirmed by the presence of deltaic deposits in the upper part of Cores II and III. All the main units would have sufficient status to be defined as members, although representing quite different processes, but in either case all the members would be part of a single formation. A bed or subfacies unit is the smallest distinct lithostratigraphic unit. ‘It is a unit layer in a stratified sequence of rocks which is lithologically distinguishable from other layers above and below from which it is separated by more or less well defined bedding surfaces’ (Salvador, 1994, p. 34). Usually a bed is indicative of a transient sedimentological environment. Intercalated units, tongues and lenses of sediment may be included, whereas rafts and intraclasts are exotics not indicative of the particular in situ sedimentological environment. Where distinctive visible changes occur in a member unit, each separate sub-unit can be defined as a bed, but only if a sedimentological basis exists for such segregation. Beds are exceptionally common in glacial environments reflecting the complexity of this system. They may vary from a multitude of units in a sand and gravel member formed by a highly variable iceproximal braided river system, to individual beds of laminated silt and clay or sands enclosed in massive diamictons formed by glaciogenic or debris flow processes. The term group is applied to a sequence of two or more contiguous or associated formations with significant and diagnostic lithological properties in common. Hitherto, the term group has not been widely used in glacial stratigraphy. However, as the need to define specific glacial environments increases

GLACIAL STRATIGRAPHY

459

West

East

Core I

Core II

Core III

50 45 40

?

35

Metres

30 25 20 15 10 5 0 78 metres

320 metres

Legend Diamicton

Sand

Contacts

intraclasts

bedded

graded

fractures

stratified

uncomformity

shear deformation

Gravel

dropstones

Bedrock

FIG. 15.7. Example of lithofacies and lithostratigraphic interpretation. See text for interpretation.

it is likely that the value of such a designation will increase. However, owing to the fragmentary nature of glacial sedimentological processes, the value of a group designation beyond local and regional levels will be limited. The term has been applied in southern England to the Kesgrave Group (Whiteman and Rose, 1992), which is a body of outwash sands and gravels deposited in the catchment of the ancestral river Thames over a period of about 1.2 Ma through much

of the Early and Middle Pleistocene. Previously, the Kesgrave sands and gravels had been given formation status (Rose and Allen, 1977), but additional work demonstrated the greater extent of the deposit, the greater complexity of the facies units, the presence of two distinctive facies associations, and the fact that these facies associations were deposited during nearly 50 per cent of Quaternary time. The lithostratigraphy of the Kesgrave sands and gravels is now based on the

460

GLACIAL STRATIGRAPHY

SW

NE

III

II

A

B

C

I

A

B

C

FIG. 15.8. Example of kineto-stratigraphic classification. At the left-hand side of the figure the stratigraphic successions are from three localities A, B, and C arranged along a SW–NE line. Directional elements derived from glaciotectonic deformation are indicated with the north arrow pointing upwards. The right-hand side of the figure shows the same deposits grouped into kineto-stratigraphic units I, II, and III (reproduced from Berthelsen, 1978).

Sudbury Formation and Colchester Formation, which represent the two distinctive facies associations. 15.5.2 Variants of Lithostratigraphy Relevant to Glacial Stratigraphy 15.5.2.1 Kineto-stratigraphy Kineto-stratigraphy stems from the work of Berthelsen (1973, 1978) on the glaciotectonically deformed sediments of Denmark. This stratigraphic approach is applied where conventional lithostratigraphic methods fail owing to intense sediment deformation, translation and penetrative movement into adjacent sediment resulting from active and subsequent passive stress application (Hirvas et al., 1976; Banham, 1977; Aber et al., 1989; Van der Wateren, 1992; Warren and Croot, 1994). Kinetostratigraphy attempts to differentiate sediments into units that have similar deformational characteristics and, in the case of glacial stratigraphy, these reflect

glacial deformational histories. The particular attributes that constitute kineto-stratigraphy are the forms of glacially induced geometric and directional indicators (Fig. 15.8). This method is especially appropriate when overthrust, folded and faulted sediments occur such that the Law of Superposition does not apply (Van der Wateren, 1992) (Fig. 15.2). In many cases subjacent pre-existing sediment has been so intensely altered that structurally the underlying sediment body becomes mixed with the overlying deformed sediment creating a single kineto-stratigraphic unit out of what were once two distinct lithofacies units (Rasmussen, 1975) (Fig. 15.9). Sediment deformation can occur during the process of deposition (syngenetic), immediately following deposition but within the same lithofacies environment (cogenetic) or some time after deposition and unrelated to the lithofacies environment that is generating the deformation process (epigenetic) (Chapter 12). Syngenetic deformation can be viewed as autokinetic, this means that deformation occurs

GLACIAL STRATIGRAPHY

461

10

9

12

11 8

2 7

6 1

5

4 3

FIG. 15.9. Important types of glacier-induced structures with their relative importance as directional indicators by the size of the dashed frame. 1, Boundary between an upper kineto-stratigraphic unit with domainal deformation and subjacent strata with extra-domainal deformation; 2, base of lodgement till; 3, way-up in glaciofluvial sediments; 4, field of combined glaciotectonics studies; 5, overthrusts – note how they converge in the direction opposite to the ice movement; 6, conjugate thrusts with opposite sense of movement; 7, sub-sole drag – shearing is intense in this zone; 8, striations in mylonitic till; 9, torpedo structure; 10, intrafolial folds of centimetre size; 11, site of till fabric analysis; 12, glaciotectonic structures of metre size. Other structures of less directional value are also shown, but not framed, for instance boudinage structures (reproduced from Berthelsen, 1978).

within the stratigraphic unit owing to internal stress conditions. It contrasts with cogenetic and epigenetic deformation, which are allokinetic deformation processes related to externally applied stress conditions. A final form of deformation occurs when various forms of deformed sediment are re-deformed (compound deformation) and this can be built into kinetostratigraphic schemes provided the individual deformational events have each produced structures that can be discriminated in the geological analysis (Ruszczynska-Szenajch, 1976, 1987; Menzies, 1990b). Kineto-stratigraphy provides a means by which it is possible to piece together various elements of a typically fragmented glacial sedimentary record and identify, or accommodate, depositional hiatuses of incalculable time periods (Berthelsen’s ‘empty’ locality) thus permitting a composite stratigraphic record to be developed (Fig. 15.10). Where an epigenetic imprint on a subjacent sediment has occurred followed by a depositional hiatus, it is possible that

elsewhere a sediment package with the same kinetic indicators can be found that can be then ‘added’ to form a synthesized stratigraphy. 15.5.2.2 Isotope stratigraphy Isotope stratigraphy, as relevant to glacial stratigraphy, is derived primarily from the study of the oxygen isotope signal stored in marine organisms (foraminifera) in ocean sediments. The shells of these microfossils preserve a record of the relative abundance of oxygen isotopes 18 and 16 in seawater, which have a ratio that is determined primarily by the volume of ice on the globe (Shackleton and Opdyke, 1973; Shackleton, 1987). This is due to the fact that the light isotope of oxygen (16O) is evaporated preferentially from the ocean and is stored in precipitation that contributes to the formation of glaciers and ice sheets. Thus relatively high ␦18O values indicate a relative abundance of ice cover on the globe, whereas relatively low ␦18O

462

GLACIAL STRATIGRAPHY

3

III II

1

2 0

I

5 metres

FIG. 15.10. Diagrammatic illustration of the use of an ‘empty’ locality. In the profile on the left there are deposits from only two kinetostratigraphic units (below and above the unconformity), but owing to the extra-domainal deformation (2) related to the ‘missing’ unit, the previous existence of the latter can be inferred as unit II on the right-hand profile (reproduced from Berthelsen, 1978).

values indicate conditions similar to the present, with even lower values if the Greenland and Antarctic ice sheets were to disappear. These figures only apply to the global ice volume values, and tell nothing about particular ice sheets. Also the relationship between the ␦18O value and ice volume is not constant as a growing ice sheet is less isotopically light than a steady-state ice sheet (Mix and Ruddiman, 1984). Nevertheless variations in the ␦18O values give an estimate of ice cover on the globe through time and, because these records are collected from long cores from continuous ocean bottom sediments, they provide a record of changes of ice cover over time and are therefore an excellent, if general, proxy of the history of glaciation. A large number of ocean cores have now been analysed and the results have been published for large parts and even the whole of the Quaternary and earlier and provide an indication of the relative volume of ice at different times (Fig. 15.11(a)).

These results show the inherent correlation between ice volume and orbital forcing parameters (Imbrie and Imbrie, 1979) calibrated at 65°N. Traditionally, ocean cores were dated by palaeomagnetism with particular reference to the Brunhes/Matuyama and Matuyama/Gauss magnetic reversals, assuming a constant rate of deposition between these chronological markers. Initially this provided a reasonable chronology for the record of glaciation, and it was within this framework that the oxygen isotope record was initially subdivided stratigraphically (Shackleton and Opdyke, 1973). Stratigraphic subdivision of the oxygen isotope record has been based on climatostratigraphic principles, with climatostratigraphic units defined by numbers from the most recent downwards. Each climatostratigraphic unit was in turn equated with a chronostratigraphic unit and defined as an Oxygen Isotope Stage (OIS). Warm episodes, similar to or almost as warm as the present, were given odd

GLACIAL STRATIGRAPHY

numbers with the present interglacial being designated OIS 1. This has been applied consistently, so that all interglacials have an odd number, except for the main interstadial of the Last Glaciation, which is designated OIS 3. Cold episodes, equivalent to a glaciation, were given even numbers, except for the Last Glaciation, which includes both OIS 3 and 4. The climatic deterioration of the Last Glacial Maximum is designated OIS 2. Using this scheme, stages can be subdivided to represent climatic oscillations and substages are defined by a lower case letter. Thus, the Last Interglacial as represented in the oceans, is subdivided into OI Sub-stages 5a, 5c and 5e, which are periods of relative warmth, and Sub-stages 5b and 5d, which are periods with glacier expansion (Shackleton, 1987; Bowen, 1994) (Fig. 15.12). With additional samples and the replication of evidence, confidence in the meaning and quality of the results has increased. The isotopic variations have been analysed mathematically in terms of the probable timing of orbitally forced climatic oscillations at latitude 65°N. Essentially, these variations occur at a predictable scale so that the timing of any change or climatic signal can be expressed with mathematical precision. Using this control the observed isotopic variations have been correlated with the predicted variation and a timescale has been derived. This process is known as orbital tuning (Berger and Loutre, 1988; Shackleton et al., 1990). In general, these results confirm those derived by palaeomagnetism and sedimentation rates, but add a level of refinement and confidence that has been, hitherto, not possible (Fig. 15.12(b)). Orbital tuning places the beginning of the Quaternary/Pleistocene at 2.60 Ma, and the Brunhes/Matuyama boundary at 0.78 Ma (Shackleton et al., 1990). The results for these studies show some very interesting patterns in the history of global Quaternary glaciation (Fig. 15.12) (Chapter 2). At the general scale, it can be seen that in the Early Pleistocene the pattern of climatic variation shows frequent oscillations of low magnitude (controlled by a 21 000-year periodicity determined by precession of the equinoxes) (Imbrie and Imbrie, 1979), and it seems that in the cold stages at this time glaciation was already established in Greenland and Antarctica, but was not important in temperate latitudes. In the period

463

between about 1.6 Ma and 0.9 Ma, the pattern of climatic variation shows moderate amplitude variations with moderate magnitude (controlled by a 42 000-year periodicity forced predominantly by the variation in the axial tilt of the Earth) (Ruddiman and Raymo, 1988). It has been suggested that in the cold stages at this time glaciation continued in Greenland and Antarctica but became an important process in the high altitude regions of the temperate latitudes and glaciogenic sediments were transported, for the first time, to the temperate lowlands (Rose et al., 1999). In the final 0.9 Ma of the later Middle Pleistocene and Late Pleistocene the pattern of climatic variation displays high amplitude oscillations with long duration (determined by a 96 000-year periodicity forced predominantly by the eccentricity of the Earth’s orbit), and it is this pattern that appears to account for the build-up of large-scale continental ice sheets that extended onto the lowland temperate regions of North America, Europe and Patagonia. There were 10 or 11 major expansions of glaciers over this period, and it was during OIS 12 and 16 that glaciers reached their greatest extent (Fig. 15.12). Isotope stratigraphy can also be derived from ice cores, although the ␦18O values are the inverse of those for ocean sediments. However, ice cores have a reliable record that goes back only as far as the Last Interglacial in Greenland (Johnsen et al., 1992; Dansgaard et al., 1993) (Fig. 2.12) and up to four glacial/interglacial cycles in Antarctica (Petit et al., 1999). Nevertheless, these records give much higher levels of precision and provide detailed evidence of climatic variation during a glaciation, providing a record of stadial and interstadial oscillations (Johnsen et al., 1995). Recently, an attempt has been made to use ice-core stratigraphy to provide a universal stratigraphy for the Late Pleistocene (Walker et al., 1999). The terms Greenland Stadial 1 (GS–1) is used for Younger Dryas, Greenland Interstadial 1 (GI–1 is used for the Windermere Interstadial, with the GRIP ice core as the stratotype for this scheme, Table 15.2). These data, also, provide a link between the dynamics of the major ice sheets, such as Greenland, and the oceans (Bond et al., 1993). Such data give evidence that is crucial to the understanding of how the Earth system works and the interactions between the hydrosphere, cryosphere and atmosphere.

464

GLACIAL STRATIGRAPHY –3

–2

–1

d 18O (‰)

0

1

2

3

9

5

21

35

63

4 78 82

5 6

12

96

100

16

6 0

20

40

60

80

120

100

Adjusted Depth (m)

–3

–2

–1

d 18O (‰)

0

1

2

3

4

5

6 0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Age (Ma)

1.6

1.8

2.0

2.2

2.4

2.6

GLACIAL STRATIGRAPHY

465

500 480

N / m

2

460 440 420 400 380 0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

More Ice

0

(c)

Age (Ma)

FIG. 15.11. (a) Oxygen isotope stratigraphy derived from ODP cores 677A and B down to 120 m, based on the planktonic (above) and benthonic (below) foraminifera. Particular OI stages are labelled. (b) As above, but tuned to the orbital timescale of Shackleton et al. (1990). (c) Insolation at 65°N in July for the past 1.8 Ma from Berger and Loutre (1991) and a model of ice volume run using the model of Imbrie and Imbrie (1979) (reproduced from Shackleton et al., 1990).

15.5.3 Morphostratigraphy Although morphostratigraphy is not unique to glacial stratigraphy it has played an important role in the development of concepts and models of glacial history and the dynamics of former glaciers. This is largely a function of the power of glaciers and glacial meltwater to form highly distinctive landforms, such as moraine ridges, kame terraces and meltwater channels, over very short periods of time. These landforms can, also, represent the positions of glacier margins as ice bodies expanded and retreated or down-wasted in response to changes in climate. Glacial landforms therefore can provide an indication of changes in glacial extent and consequently are a proxy for climate and a basis for climatostratigraphy.

A morphostratigraphic unit is defined as ‘comprising a body of rock that is identified primarily from the surface form it displays; it may or may not be distinctive in lithology from contiguous units; it may or may not transgress time throughout its extent’ (Frye and Willman, 1962). This definition was developed as a result of mapping glaciogenic sediments in Illinois, mid-west USA, where lithofacies are exceedingly complicated and often stratigraphically undiagnostic in terms of the contemporary sedimentological theory (Menzies, 1996, chapter 9). Conversely, the landforms could be resolved into a series of moraine ridges that were interpreted as forming at a glacier margin, and therefore gave a clear expression of the changing glacier configuration.

466

GLACIAL STRATIGRAPHY

(a)

(b) Stacked Oxygen Isotope Record

0

–1

0

Grand Pile Arboreal – Non-arboreal Pollen 1

0

1

Holocene

2

Late Devensian D

3

Middle Devensian

H

50 Time (103 BP)

100%

50

M 4 5a

Early Devensian

5b 100

St. Germain II Odderade St. Germain I Amersfoort / Brorup

5c 5d 5e

Ipswichian

Eemian

6 150 FIG. 15.12. Oxygen isotope record for the last 130 ka giving OI stages and sub-stages compared with the arboreal pollen record from Grande Pile southeast France, which is defined in terms of the Eemian Interglacial, the St Germain I/Amersfoort/Brorup, St Germain II/ Oddrade, Moershoofd (M), Hengelo (H) and Denekamp (D) Interstadials (reproduced from Bowen, 1994).

Moraine ridges are the most characteristic morphostratigraphic unit in glacial geomorphology and geology. They are indicators of glacier margin position. They are recognized by their form and can, when not modified by postglacial subaerial mass movement, be mapped with great precision. As emphasized by Frye and Willman (1962), these units should be identified independently of their internal composition, and indeed moraine ridges can be formed of a great variety of materials with a great diversity of forms. For example, the Younger Dryas

moraines in Scandinavia range from ice-contact deltas in western Norway and southern Finland, to push moraines formed of glaciomarine sediments in northwestern Norway, and boulder ridges in the Scandinavian north. Similarly, but at a smaller scale, the Loch Lomond Readvance moraine around the southern end of the Loch Lomond basin (Fig. 15.5, Table 15.3) ranges in composition from a till ridge on the hillside slopes in the west and south, a sand and gravel ridge in the valley bottoms of the west and south, an ice-contact sandur in part of the main outlet valley

GLACIAL STRATIGRAPHY

467

TABLE 15.2. Chronostratigraphic and geochronological hierarchical terminology Chronostratigraphic unit

Geochronologic unit

Example

Eonothem

Eon

Phanerozoic

Erathem

Era

Cenozoic

System

Period

Quaternary

Series

Epoch

Pleistocene

Stage

Age

Devensian/Wisconsinan

Chronozone

Chron

Dimlington Stadial/Younger Dryas

Km

1

51

30

901

53

500

0

28 0

2

40

53 908

49 20

0

Foinaven

3

0

50

0

806

47

0

30

DIONARD

867

20

0

40

protalus rampart

SRATH

778

787

5 600

4

Plàt Rèidh

Arkle

757

6

500

400 30

0

31 600 44 400

6

2

7

3

8

4 5

9 10

47

Meall Horn 777

7 0

1

60

and an ice-contact delta in parts of ice-dammed valleys and, in the southeast, where the ice margin grounded in a large proglacial lake (Lake Blane), the moraine forms a low, broad ridge of glaciotectonized glaciolacustrine silts (Rose, 1981). Although the sediments are so variable and would be difficult to correlate on many lithostratigraphic criteria, this landform indicates a roughly synchronous position for the maximal extent of the Loch Lomond glacier sometime just after 10 560 14C years BP (Rose et al., 1988). However, despite this lithological variability, it is essential to emphasize that there are elements of inherent sedimentological uniformity to the extent that all these sediments show evidence of ice-contact on the ice-proximal side of the ridge. This evidence takes the form of glaciotectonic structures in sorted sediments and a distinctive clast fabric in tills. This case emphasizes the need to link landform and sediment evidence whenever possible, and to integrate as many kinds of stratigraphic information as possible. An example of morphostratigraphic evidence is shown in Figure 15.13 (Sissons, 1977), which includes landforms such as glacial meltwater channels that are interpreted as forming at the ice margin, and kame terraces that are built-up against an ice margin. Although not strictly stratigraphical, the distribution of hummocky moraine and the distribution and arrangement of flutes and drumlins has been taken as evidence of glacier extent and behaviour, with the icedistal limit of these features giving a minimum position for the extent of ice cover. In a similar fashion the extent of ice cover is also given by the ice-

44 37

0 40 45 39

FIG. 15.13. An example of geomorphological mapping to derive a morphostratigraphy of Loch Lomond Stadial glaciers in the region of Foinaven, northwest Scotland. This is a simple scheme used to identify the limit of the glacier at the maximum of the Loch Lomond Readvance. The landform evidence used consists of: 1, moraine ridges; 2, fluted till; 3, hummocky moraine; 4, meltwater channels; 5, ice-distal limit of a ‘boulder spread’; 6, ice-proximal limit of periglacial features; 7, ice-distal limit of hummocky moraine or ‘drift’ sheet. 8 is the interpolated ice limit and 9 is very steep slopes of the glacier source area; 10 is contours at 100 m intervals (reproduced from Sissons, 1977, in Gray and Lowe (eds) with kind permission from Pergamon Press Ltd).

468

GLACIAL STRATIGRAPHY

proximal limit of the distribution of periglacial soils and slope landforms, on the assumption that these could not form beneath ice or would have been destroyed if the ice had covered the area. This does, of course, assume that these periglacial features did not form after the ice retreated. Innovative work by Boulton et al. (1985) has used the pattern of glacier flutes, drumlins, streamlined hills to reconstruct the activity of former glaciers and ice sheets across northern Britain, Scandinavia and North America. This technique uses this pattern of lineations to determine former ice-flow paths. Glacier marginal configuration is then reconstructed with a trend that is normal to the ice-flow vector. Recognition of superimposed patterns of lineations (Rose and Letzer, 1977; Rose, 1987a; Boulton and Clark, 1990a,b; Clark, 1994; Clark et al., 2000) has added a longer temporal dimension to this method of analysis and it has been possible to determine more than one stage of glacier development with different flow directions and ice dynamics. The great strength of morphostratigraphy is that the evidence is on the land surface and may, if it is well developed, be traced continuously across whole regions with a level of precision that is not possible with any other stratigraphic method. In addition the spatial extent of landforms means that the relationship between one landform and another, and hence one stratigraphic unit and another, can be investigated in detail. Using the example of moraine ridges, geomorphological mapping, such as illustrated in Figure 15.13, has the potential to demonstrate merging or overlapping ridges indicating different ice-streams within a glacier system. The ability to establish such temporal and spatial relationships is known as connectivity and can have a high level of stratigraphic precision equivalent to the superposition of stacked sediments. Examples could include graded sandur surfaces indicating contemporaneity of the meltwater deposition, or cross-cutting moraine ridges, which indicate that later ice extended across an earlier marginal position. Another advantage of continuity of surface expression is that glacial landforms can be readily dated by methods such as lichenometry or Schmidt hammer impact measurements (McCarroll, 1989; Matthews, 1992; Evans et al., 1999). Studies of this

kind allow the reconstruction of detailed morphostratigraphy with strict geochronology (Fig. 15.14). The level of precision and resolution provided by these methods across such wide areas of the Earth’s surface is rarely equalled by any other stratigraphic method. There are, of course, many problems associated with morphostratigraphy, not least the difficulty of measuring and analysing features of immense size and, in North America, over vast areas. Geomorphological mapping requires enormous time, great skill, effort and experience. Air photography and, today, the use of GIS has greatly helped in the process of geomorphological analysis, but imagery in the form of air photographs or digital terrain models (DTSs) is no substitute for field mapping, as the detailed form and internal composition of the landforms cannot be determined. A ubiquitous problem is that once glacial landforms such as moraine ridges or meltwater channels are formed, they are altered by subaerial processes and the original detailed form is lost. With this loss there is rarely any possibility of regaining the initial stratigraphic precision. At this stage morphostratigraphy fails to be of value as a stratigraphic technique. The use of this property, often known as the ‘freshness’ of landforms, has been used as a stratigraphic method. However, it is of little use in this role other than as a very unreliable ‘first order’ estimation, and this approach has created far more problems than it has solved. Hitherto, no satisfactory morphometric method has been devised to quantify this approach, or to overcome the intrinsic variation caused by the lithological variability of glaciogenic sediments and facies associations. The concept of ‘older’ and ‘newer’ drift is possibly as far as this approach can be taken (Wright, 1937) and even this approach can be shown to be fundamentally flawed, depending upon the original form of the glaciogenic landscape (Morgan, 1973). In mid-latitude regions, only glaciogenic landforms formed during or since the LGM can be studied by morphostratigraphic methods. Older features are likely to have been subject to considerable geomorphological modification making application of the technique unproductive and futile (Straw, 1960; Sparks and West, 1964).

GLACIAL STRATIGRAPHY

(a)

469

(b) glacier 1968

1971

1961 1 9 6 8

1500 m

1955 1 9 5 9 19 40 1936

194

19 9

51

1400 m

28 19

M8

1300 m

M7

1

91

1900

1870

52

18

M2

1200 m

25

18

M1

10

0

300 m

track

18

N

M4 M3

7

08

19

M5

M6

L e ir a

1750

1100 m

FIG. 15.14. Morphostratigraphy of the foreland region of Sorbreen Norway. The position of the ice margin is derived from detailed geomorphological mapping, and the geochronology has been constructed from a variety of techniques including lichenometry and air photo analysis (reproduced from Matthews, 1992).

15.5.4 Stratotypes Stratigraphic convention requires that in order to establish a stratigraphic sequence for a region it is necessary to designate a type section (stratotype) or set of type sections representing a type locality. A stratotype acts as an objective example or standard for a particular stratigraphic sequence and as a criterion against which other sections can be compared. Ideally, a stratotype should be a continuous record and contain the contact with the underlying stratigraphic unit. However, as noted by Bowen (1978) and Rose (1989a) such stratotypes are rare indeed within glacial

sediments, and in Britain, for instance, only for the Anglian and Devensian glacial stages have stratotypes been adequately defined and these are unsatisfactory with respect to the definition of their lower boundaries. The existence of the lower boundary in a stratotype is critical as it contains the change from one set of sedimentary characteristics, and hence from one depositional system or subsystem, to another. In glacial geology it could identify the onset of glaciogenic deposition at a site. At the stratotype for the Dimlington Stadial (LGM) in Britain the lower boundary is identified by the base of a laminated silt member that was deposited by snow melt or glacial

470

GLACIAL STRATIGRAPHY

meltwater in close proximity to glacial ice (Penny et al., 1969; Rose, 1985). This facies unit is succeeded by diamictons that represent the presence of glaciers within the region. The upper boundary is not required as it can be defined by the lower boundary of the succeeding unit (West, 1989) and the lack of requirement of a definition for the upper boundary means that new units can be inserted into a stratigraphic scheme without requiring re-definition of the upper boundary. For the Dimlington Stadial, for example, the upper boundary is defined by the base of the lacustrine clays that represent the onset of the Windermere Interstadial in Britain (Rose, 1985). It is known, for instance, that the transition between the stadial and interstadial conditions are missing in the type area of the Dimlington Stadial (Walker et al., 1993) and this can be accommodated by reference to the Interstadial type site without the need to redefine the stratotype for the Dimlington glaciation. These type sections, therefore, provide a record of sedimentological change that represent activity of importance to glaciation, such as an ice advance or retreat. The type section may include evidence for the approach of an ice front as indicated by increased coarsening of sediment or structures indicating progressively higher flow regimes or by the initial influx of glacially derived sand-size aeolian sediment (Rose et al., 1985). Stratotypes may also be defined on the basis of morphostratigraphy, but this is rare because morphostratigraphic evidence loses its precision as a consequence of non-glacial subaerial denudational processes and morphological boundaries are then very difficult to recognize and define. However, in the type area for the Loch Lomond Stadial in Britain, the Loch Lomond moraine ridge is included as part of the evidence for the stratotype (Rose, 1981, 1989a) (Fig. 15.5, Table 15.3). 15.5.5. Chronostratigraphy, Geochronology and Geochronometry Chronostratigraphy is ‘the organisation of rocks on the basis of their age or time of formation’ (Salvador, 1994, p. 113) and chronostratigraphic units are ‘a body of rocks that includes all rocks formed during a specific interval of geological time’ (Salvador, 1994,

p. 113). Geochronology is ‘the science of dating and determining the time sequence of events in the history of the earth’ (Salvador, 1994, p. 120) and Geochronological units are units of geological time (Salvador, 1994, p. 120). Geochronometry and geochronometric units deal with the quantification of geological time, that is, putting actual ages on the events or time periods. Reference to Salvador (1994, p. 120) for a definition of geochronometric terms shows an expected resolution of thousands or millions of years. This emphasizes the much higher level of precision required by glacial stratigraphy where resolution may be in decades, years or even seasons. There has been much confusion in the geological literature about the use of chronostratigraphy and geochronology. This is amplified in Salvador (1994, pp. 9–10) where it is pointed out that each chronostratigraphic unit is a rock body that has a corresponding interval of geological time (geochronological unit). This means that the chronostratigraphic units are the sediments, landforms and soils formed over a certain interval of time and that geochronological units are the interval of time during which these sediments are deposited, soils are formed or landforms are constructed and eroded. Geochronometric units also provide a quantitative value for the period of time that allows calculation of the rates at which processes operate. The hierarchy of chronostratigraphic and geochronological units is given in Table 15.3. Bearing in mind that lithostratigraphy and morphostratigraphy remain the building blocks of glacial stratigraphy, the aims of this science are most effectively expressed by using diachronic units with a climatostratigraphic basis. These methods reflect the climatic forcing of systematic glacier expansion, standstill and retreat, and the inherent variation that exists between sites as a function of local climates, variations in glacier catchment shape, size and steepness, and bed-material (Fig. 15.3). Nevertheless, chronostratigraphy remains the aim of most glacial stratigraphic studies, if only because it provides the most useful method of communication between scientists and with the layman. Table 15.3 provides an example of a local-scale stratigraphy, based on the type area for the Loch Lomond Stadial, in the southeastern part of the Loch Lomond Basin, west-central Scotland. This scheme is based on the

GLACIAL STRATIGRAPHY

471

TABLE 15.3. Stratigraphic subdivisions and terminology at the type area for the Loch Lomond Stadial, south east Loch Lomondside, western central Scotland (based on Rose, 1989a) Lithostratigraphic & Morphostratigraphic units: Within limit of Loch Lomond Readvance (Croftamie, Aber, Gartness, Carnock Burn, Loch Lomond 1 )

Beyond limit of Loch Lomond Readvance (Muir Park Reservoir 2 )

Brown silty clay (freshwater) Dark organic silty clay (marine) Grey & brown silty clay (freshwater)

Peat

Climatostratigraphy

Ice Core Stratigraphy

Chronostratigraphy

Stadial/Interstadial

Ice Core age K GRIP yr BP 7

Stage

Holocene Epoch

Flandrian Interglacial Stage

3

Gyttja Clay-gyttja

Blane Valley Silts Rhu Gravels Gartocharn Till/Loch Lomond Moraine Blane Valley Silts Main Lateglacial Shoreline

11,500

Pink clay

Loch Lomond Stadial

Greenland Stadial 1 GS-1

12,650 Greenland Interstadial 1 GI-1 14,700

Clyde Beds

Gyttja

Windermere Interstadial

Gartness Gravels, Sands and Clays Wilderness Till

Silty clay Wilderness Till

Dimlington Stadial

Greenland Stadial 2 GS-2 21,200

1

Dickson et al., 1978; Rose, 1980; Rose et al., 1988.

3

Site shown on Fig. 8.5.

5

Alley et al., 1993.

2

Donner, 1957; Vasari and Vasari, 1968; Vasari, 1977.

4

Rose, 1985.

6

Bard et al., 1991.

evidence summarized in Figure 15.5. The chronostratigraphy provides a basis for comparison with other sites in the British Isles, and indeed elsewhere. The geochronometry from the site provides dates that only approximate with the formally accepted ages (Mangerud et al., 1974) and with the newly defined ice core stratigraphy (Walker et al., 1999) for the boundaries between the stages and substages, although local dates do contribute to the understanding of the Earth system dynamics in the region. However, this site is accepted as a stratotype because of precedent (Simpson, 1933), the wide range of lithostratigraphic, morphostratigraphic and geochronometric evidence, and the persistence and accessibility of the sections (Rose, 1981, 1989a). For glacial behaviour specifically, the site also gives evidence for the behaviour and timing of the glacier that advanced down the Loch Lomond basin from the western Scottish Highlands during the Younger Dryas (Loch Lomond Stadial) climatic deterioration. Finally, Table 15.4 provides an example of a currently accepted chronostratigraphy for the Pleistocene of the British Isles and northwestern Europe (Bowen, 1994). At this level the scheme is subdivided into glacial and interglacial stages, and glaciation per se is subsumed under this climatic scheme, so that the

Late Devensian Glacial substage

7

Geochronology

Radiocarbon Years 4

Calendar Years 5

10,000

11,500

5

11,000

12,800

5

13,000

14,620

5

26,000

c.22,000 (Max) 6

Walker et al., 1999.

actual status of glacier behaviour hardly contributes to its derivation. Indeed the general extent of glaciation in some of the ‘glacial’ stages is not known. The timescale is based primarily on orbital tuning. The essential aim of this scheme is to provide a framework for a highly complex period of Earth history and a basis for communication between scientists from different major regions on the globe. 15.6. CONCLUSION Much of glacial stratigraphy has a solid scientific foundation, using an extended, and hence more powerful version, of the International Stratigraphic Code (Salvador, 1994). However, the complex nature of the glacial system means that interpretation is often far more complex than for the other depositional systems that dominate the preMiddle Quaternary parts of the Geological column. In these circumstances, and bearing in mind the immense difficulties of dealing with spatially extensive morphostratigraphic evidence, it is little wonder that in some instances stratigraphic frameworks that were constructed some decades ago now appear naive. Currently many past stratigraphies are subject to revision, as new descriptive methods and greater

472

GLACIAL STRATIGRAPHY

TABLE 15.4. Chronostratigraphy of the British Isles and north western Europe (based on Bowen, 1994) OI stages

Age (ka)

Chronostratigraphy British Isles

1

NW Europe HOLOCENE

11.5 Younger Dryas Bølling-Allerød

2 Late Devensian

Late Weichselian

25 3

Middle Devensian 50

4 70 5a 5b 5c 5d 5e

Upton Warren

Odderade

Chelford

Amersfoort-Brorup

Ipswichian

Eemian

(Ridgeacre)

Saale (Warthe)

Stanton Harcourt

Bantega/Hoogeven

130 6 186 7 245 8

Saale (Drenthe) 303

9

Hoxnian

Domnitz/Wacken

Swanscombe

Fuhne Holsteinian

Anglian

Elster II

Cromerian

Beestonian

Cromerian IV Glacial C Cromerian III Glacial B Cromerian I Glacial A Cromerian I Dorst (g) Leerdam (ig) Linge (g) Bavel (ig) Menapian Waalian (A,B,C) Eburonian

Pastonian to Ludhamian Pre-Ludhamian

Tiglian (C5 to A) Praetiglian

339 10 11 423 12 478 13 14 15 16 17 18 19 20 21 22 23

Waverley Wood

– 790 –

– 900 –

1.6 Ma

2.3 Ma Note: (g)-glacial, (ig)-interglacial.

GLACIAL STRATIGRAPHY

ARCTIC

473

OCEAN

CHRONOSTRATIGRAPHY EON

LA

ERA

UR

PERIOD

CORDILLERAN EPOCH

ICE

AGE/STAGE SHEET Chron

PA C I F I C OCEAN

CORDILLERAN ICE SHEET

EN

C L I M ATO S T R AT I G R A P H Y

TI

DE

IC

E SH

AGE/STAGE

E

SUBSTAGE STADIAL

E

T

LITHOSTRATIGRAPHY

LITHOSTRATIGRAPHIC UNIT - GROUP

LITHOSTRATIGRAPHIC UNIT - GROUP FORMATION MEMBER BED

ATLANTIC 0

500 km

OCEAN

FIG. 15.15. Illustration of stratigraphic terminology, superimposed on the Laurentide Ice Sheet at the Last Glacial maximum.

understanding of glacial processes are applied. It is essential, however that any new stratigraphy must not be cast in dogma as has been the case in the past (Rose, 1987b, 1989a, 1991; Gibbard and Turner, 1988; Bowen, 1999) and is subject to modification along with the discovery of new evidence and understanding. For a working framework, glacial stratigraphy can be comprehended at a range of scales from global to site. At the global, hemispheric, continental and continental/regional levels chronostratigraphy is the most appropriate level for assimilating and communicating stratigraphic information. At the continental/ regional, regional, regional/local and local levels climatostratigraphy is most effective, whereas at the

local and site levels lithostratigraphy is most appropriate (Fig. 15.15). There remain several continuing issues in glacial stratigraphy, in particular the need to establish better inter-regional and intercontinental stratigraphic correlations; to develop multiple dating systems applied to specific problem sites; to reconcile and validate the stratigraphic record between terrestrial and oceanic stratigraphies; to institute and expand objective methods of stratigraphic definition; finally, glacial stratigraphy must be utilized as a tool in obtaining insight into the processes and rates of operation of the glacial system, the climatic, geographical and geological conditions pertinent to and existing during ice sheet formation and decay.

This Page Intentionally Left Blank

16

PROBLEMS AND PERSPECTIVES J. Menzies 16.1. INTRODUCTION The intent of this closing chapter is not to sprinkle the text with references or to state dogmatically where the discipline as a whole should go to seek new answers but rather to give a perspective on the field by highlighting some of the persisting problems that dog the search for even greater understanding. It is apparent that although much information is shared and broadly disseminated, too often members studying one sub-environment take little heed of anything less than dazzling discoveries in another. Yet much of this science, like all others, is founded upon the slow and, in itself, uneventful accumulation of a myriad of minor details that, only when synthesized, move our understanding a minute step forward. However, too often these minor steps are not shared or their significance is not conveyed from one environment to another. A superb instance is the different approaches adopted by those interested in processes of diamicton and diamictite deposition. For example, the sum of the efforts of both Precambrian and Pleistocene researchers, both groups working on a similar problem, will undoubtedly add up to more than their separate endeavours; however, such collaboration is limited at present (Menzies, 2000). Perhaps only 20 years ago it was possible for a single researcher to have a moderate grasp of all the subdisciplines concerned with glacial environments, while today only an elementary understanding of

many of these subunits is possible. As with any form of specialization, breadth of knowledge is lost in pursuit of depth, thus an ever-increasing need arises for general perspectives to be expressed to avoid either needless repetition or pointless explorations. Likewise, the derivation of countless untested hypotheses has strewn the paths of glacial geology and geomorphology in almost careless abandon. The need to develop ‘pet’ hypotheses as de rigueur must be replaced by testing and re-testing of the few sound ideas that survive preliminary examination. Only by this means can a sound body of theories on processes, mechanics and corelationships within glacial environments be developed, against which new ideas can be checked and the science advanced. Linked to this sound footing is the requirement for reliability testing and validation of techniques, discriminant tools and characteristic signatures by as objective a means as possible. There are several areas within glacial geology that are less well studied and therefore leave gaps in our general understanding and may hold keys to existing problems that remain to be disclosed. An example has been the limited emphasis, beyond general description, that has, in the past, characterized glacial erosion. Likewise, the contribution of proglacial lakes to ice mass marginal instability, to proximal modes of deposition, to isostatic crustal movement and the link with ice mass balance gradient are profitable areas requiring greater research effort.

475

476

PROBLEMS AND PERSPECTIVES

New and exciting fields of research into glacial environments are being developed, often by crossfertilization from existing sciences. For example, the use of micromorphology in glacial sediment microstructure research shows enormous potential. Likewise, optically stimulated luminescence (OSL) as an added tool in dating methods should permit details of chronology and stratigraphy hitherto only inferred. By explaining aspects of the rheology of slurries much can be gleaned that is immediately pertinent to deformable bed conditions, porewater contents and stress field states. Similarly, from rock mechanics and knowledge of brittle fracture processes within tribology, a different perspective on the processes of glacial erosion can be acquired. Finally, there is a need to ‘step-back’ from the minutiae of research areas and examine glacial environments on regional and continental scales. With advances in GIS and increasingly discriminant remote sensing imagery combined with mega-geomorphology, an understanding of macroscale relationships generated by ice masses both in front of and at their beds may be gained. 16.2. PARADIGM SHIFTS Over the past two decades interest in glacial environments has grown immensely, as has the sophistication of techniques and methods used in their study. During this same time period a fundamental shift in the methodology of studying these environments has emerged. Where, in the past, a geographical/morphological approach was employed; today a movement to a sedimentological/glaciological method has gained ascendancy. This paradigm shift has been discussed previously (Chapter 1) and need not be repeated other than to note that in shifting from the former to the latter approach a geographical perspective has often been lost in the minutiae of detail utilized in processmechanistic studies. Where the geographical/morphological methodology emphasized the landform as its basic unit of study, the sedimentological/glaciological method begins with the individual particle or subfacies unit as a basic building unit. The methodological stages that have occurred in studying glacial environments are typical of any developing science – beginning with a descriptive

stage and moving to a more detailed ‘deconstructionist’ approach, where a process/mechanistic methodology is employed to examine in ever-increasing detail the fundamental components of study. This development from a descriptive to process-oriented stage has been triggered by the development of increasingly sophisticated and reliable tools (e.g., dating methods, SEM and sedimentological facies models). A third stage can now be anticipated in which the individual elements or facets of the glacial environment can be re-assembled within a paradigm that will emphasize both a geographical and mechanistic perspective. Such a paradigm will take into account the geographical conditions within which glacial and non-glacial sedimentary processes operate yet are constrained by glaciological, geotechnical and rheological parameters. 16.3. QUESTIONS ILLUSTRATIVE OF PROBLEMS As our understanding of glacial environments has increased, so novel and intriguing questions have arisen. The emergence of these questions is testimony to the vigour and intensity of future research in glacial environments. Some of the many questions that need to be answered have been categorized within individual sub-environments; however, many questions transgress these artificial boundaries and are pertinent to several or all glacial environments. To define these questions as problems is more than semantics since without adequate answers further scientific progress in understanding glacial environments will be limited and restricted. 16.3.1. General There are several questions that are relevant to all glacial environments and are perhaps of broader significance than those solely applicable to specific environments. (a) In general, recognition and differentiation of facies types persists as a continuing problem. The ability to discriminate between the varied environments within which diamictons can be deposited is perhaps the best example at present. It

PROBLEMS AND PERSPECTIVES

(b)

(c)

(d)

(e)

(f)

remains problematic to distinguish between diamictons from subglacial and subaqueous environments in many critical instances. Where significant environmental transformations caused by climatic change such as periglacial activity or diagenesis have influenced sediment types, recognition of these changes as signatures within the sediments remains elusive. Structures within glacial sediments may be developed as a result of intrinsic or extrinsic stress conditions or a combination of both; however, discrimination between these forms of structure generation are equivocal. In the last few years considerable interest has been engendered in trying to understand the many complex structures both at the micro- and macro-level found within all glacial diamictons. At present only limited explanations exist for these intriguing structures. A general understanding of the impact of porewater and porewater pressures on sedimentation mechanisms, rates of sedimentation and the structures imprinted by porewater in sediments in terms of physical and chemical effects remains limited. Porewater, in directly influencing effective stress levels, must have a major part to play in glaciotectonism and sediment deformation at all scales. In translocating clays and soluble minerals, porewater alters the chemical signature and, in some instances, the geotechnical properties of some sediments. The extent of this impact remains only inferred. Related to porewater are the many dewatering and syneresis structures that are attributable to the effects of high rates of saturated sediment deposition, overloading, glacial tectonism and neotectonism. It is typically stated that particles that have passed through the glacial system are distinctive in shape, surface morphology and size distribution, yet little research has closely investigated particle shape evolution, fracture mechanisms and surface morphological imprinting other than in a limited sense. The fields of tribology and rock mechanics is a potential rich source of new insights into these aspects of particle development that need to be considered. There is a determined need to relate glaciological parameters to those linked with glacial depositional

477

mechanics. For example, it is as yet unknown how ice streaming and surging grossly influences subglacial and proximal glacial sediments on land and subaqueously both in terms of previously deposited sediments and pencontemporaneous sedimentation processes. (g) In terms of stratigraphic correlation, there remains a need to better understand paleosol preservation and pedogenic formative processes under different interglacial and interstadial climatic conditions. The value of dating techniques in stratigraphic correlation and surface age dating continues to require improvements in reliability and validity. (h) As noted above under (d), the impact and varying likely response of glacial sediments to neoseismic activity is relatively little studied. Where, in the past, intense deformation or surface turbations have been regarded as pedogenic or glaciotectonic in origin, new evidence has linked many of these structures and intercalations of glacial sediments with recent seismic activity. From an applied aspect it is crucial that a clearer knowledge of this response by glacial sediments to seismic events be obtained. (i) As a final general concern, it remains relatively unknown how glacial sediments alter or change over time as a result of subaerial diagenesis. Some work in front of present Icelandic glaciers points to very rapid changes in clay and silt content with marked changes in geotechnical properties occurring but how penetrative these changes are over time has been little studied. Today, engineers deal with past glacial sediments that have undergone the effects of diagenesis over several thousands of years – thus such questions are imperative. 16.3.2. Subglacial Environments The following questions have not been arranged in any order of significance or pertinence but reflect many of the remaining issues that need to be tackled in understanding subglacial environments. Several questions or parts of questions are repeated in the subsections deliberately to re-emphasize their pertinence. Typically, many of these questions generate other related questions that require answers.

478

PROBLEMS AND PERSPECTIVES

(a) In recognizing subglacial sediment facies types a major problem remains in differentiating between sub-facies types and also sub-environments, not only within subglacial environments but between the subglacial and other environments. For example, the differentiation between subglacial debris flows and those generated in supraglacial or proglacial regimens remains conjectural. The key to recognition and discrimination lies in determining if diagnostic signatures can be found that would permit, for example, the distinction between sediments characteristic of soft bed deformation and those deposited under nondeforming subglacial bed conditions. (b) As yet relatively few studies have concentrated upon fully exploring and understanding the development of subglacial hydraulic systems both in the spatial context beneath ice masses of varying thermal and stress field regimens and in the temporal in terms of hydraulic system maturation, disruption and re-stabilization as must occur under surging conditions and other changing bed states. Similarly, the mechanisms of subglacial meltwater erosion and the influence of meltwater on overall ice sheet stability require considerable research. (c) As the importance and ubiquity of surging glacier conditions becomes more apparent it is critical that the signature, if any, be explored in detail, both within the subglacial sediments and on landscapes in terms of possible bedforms. (d) A related issue to both subglacial hydraulics and surging behaviour is the question of recognizing j¨okulhlaup events within the subglacial sedimentary record. At present evidence is beginning to accumulate that should permit recognition but still more work needs to be done both in modern glacial environments and on past glacial sediments. (e) Only limited work has been accomplished in recent years on the mechanics of subglacial glaciofluvial deposition whether in cavities, channels or sheets. Problems concerning when and how deposition occurs especially under full-pipe flow conditions remains enigmatic. Equally critical are the relationships between glaciofluvial sediments and diamicton deposition and the

(f)

(g)

(h)

(i)

(j)

general variations in subglacial bed conditions over time and space beneath an ice mass. A typical characteristic of most diamictons is a clast fabric of varying degrees of strength. In the past accepting that all subglacial diamictons were of a lodgement, melt-out or flow till type, clast fabric was used as a discriminant tool but, since depositional conditions beneath an ice mass are more complex than hitherto conceived, clast fabric and its significance must be re-examined, as must the mechanism(s) of fabric development given the complex rheological and thermal conditions found beneath ice masses. Within the subglacial environment, in particular, the enormous variation and transience in bed conditions must impact upon deposition mechanisms, structures and stratigraphy – yet this reflection of changing conditions remains undetected within existing subglacial sediments. Little is known concerning post-depositional diagenesis of subglacial sediments. The alteration of geotechnical properties is especially significant; however, the timing and extent of diagenetic change remains generally unknown. It is well known that on approaching an ice front exposed bedrock may exhibit evidence of iron oxide deposition in the lee side of boulders and bedrock protrusions. Likewise translocation of carbonates can be seen as calcite deposits and in many glaciofluvial gravels cementation by calcium carbonate is often reported. Both instances reveal the effectiveness of low temperature geochemistry in dissolution and re-precipitation; however, the details of the geochemical processes, the rates of chemical reaction and the influences of temperature and pressure are insufficiently understood. Questions such as ‘how are drumlins formed’ have been examined for over a century yet the origin of these forms remains enigmatic. The subglacial forms and possibly associated forms such as Rogen and fluted moraines encapsulate at the larger scale the problems of understanding subglacial depositional processes. It has been suggested that these forms are part of a bedform continuum and as such are intrinsically related in origin and development, however, this idea

PROBLEMS AND PERSPECTIVES

remains only one of several hypotheses on their genesis. If these forms are bedform assemblages what are the bedforming mechanisms, why do sets of these forms seem to co-exist, and if bedform transition occurs how is it manifest, and what are the controls on bedformation and maturation? (k) As a final example, the details of glaciotectonism and the relationship of this complex process to imprinting structures and other diagnostic signatures in subglacial sediments need further investigation. Likewise discrimination between neotectonic seismic events within subglacial sediments and contemporaneous glaciotectonism requires further detailed study. 16.3.3. Englacial Environments Englacial, like subglacial, environments are normally obscured from view. When such an environment is penetrated invariably the environment is altered, however slightly. Knowledge of englacial environments is therefore largely derived from indirect evidence. (a) The development of englacial hydraulic systems remains largely obscure. The stages and style of development are largely unknown as are the impact of changes in ice velocity, especially the effect of surging and subsequent englacial hydraulic system re-establishment. (b) The discrimination of englacial sediments from either supra- or subglacial ones remains in many cases an insurmountable problem. Only when the facies context can be ascertained is distinction possible. (c) The mechanics of sedimentation in englacial tunnels and conduit remains obscure. In many instances sedimentation would appear to occur under full-pipe flow meltwater regimes presumably under high hydraulic pressure gradients. Likewise knowledge of sedimentation processes in partially filled englacial drainage systems is insufficient. As a consequence both the mechanics and timing of the formation of eskers and kames and associated glaciofluvial englacial deposits is inadequately modelled and understood.

479

16.3.4. Supraglacial Environments As interest in valley glaciation as an analogue for past glaciations waned over the last 30 years, in many instances close attention to supraglacial environments declined. It was generally accepted, for example, that the impact and contribution from supraglacial sediment sources along the southern margins of the Laurentide Ice Sheet in southern Canada was limited; yet evidence from the mid-western states in the USA contradicted this assumption. As a result, in the past few years, a realization has increasingly dawned that a considerable volume of subglacial sediments were elevated over the last few kilometres towards the margins of the Laurentide Ice Sheet and appeared as supraglacial sediments sliding and flowing off the ice margins into morainal and other proximal proglacial environments both on land and into lakes. Thus the need to better comprehend these environments must again been viewed as essential. (a) Perhaps the greatest challenge is in distinguishing supraglacial sediments from all other subfacies. Where sediment originally was subglacially derived, as was the case virtually all along southern margins of the Laurentide Ice Sheet, facies differentiation is especially difficult. (b) It is likely that in the event of a surge, supraglacial sediments should reflect the impact of such a dynamic process but as yet recognition of such events has not been possible. (c) As our understanding of sediment delivery to the margins increases, it is apparent that there is a considerable need to be able to link changes in ice mass balance and thus balance gradient with changes in sediment delivery, especially in the supraglacial environmental system. If the variations in magnitude and frequency of changes in ablation/accumulation are sufficiently great, then thresholds that will influence sediment delivery to the margins will occur and be reflected in facies type, volume, possibly depositional mechanics and size and spatial/temporal distribution of moraines. 16.3.5. Proglacial Environments It is perhaps fair to say that interest in proglacial environments has been the realm of only a few

480

PROBLEMS AND PERSPECTIVES

dedicated researchers and that all too commonly the environment has been virtually ignored to the detriment of our understanding of the general glacial system. (a) The proglacial environment is invariably dominated by meltwater streams, thus the mechanics of stream development in relation to glacial hydrologic regimes is critical to an understanding of the dominating processes in the environment. (b) The impact of j¨okulhlaups is of such magnitude, both distally downstream and in the immediate proximal zone in any proglacial environment, that recognition of such infrequent but catastrophic events is an integral part in understanding ice mass behaviour and sedimentation within a glaciated basin. At present discrete characteristics that can be used to distinguish j¨okulhlaup events are still being sought both in proximal and distal locations. (c) Attempts to understand in-channel sedimentation processes such as braid development has become an increasingly important area of study with proglacial environments. Owing to the unique hydraulic regimes of glacial meltwater streams, the ability to understand the evolution of in-channel forms has the potential to permit recognition and differentiation of specific proglacial sedimentological processes and hydraulic regimens. (d) Within proglacial environments there are many forms and structures such as boulder rings that are indicative of both meltwater flow regimes and the impact of melting buried ice. At present our knowledge of many of these structures is limited, thus what these forms are symptomatic of remains hidden. (e) As is often the case ice masses advance over the proglacial zone resulting in sediment re-incorporation into the glacial system. The effect of ice advancement is multi-fold, causing ice shoving, pushing, squeezing and general deformation of proglacial sediments. In some cases sediments may be completely reworked back into the subglacial system losing all attributes of any previous depositional history, while in other cases sediments may be strongly tectonized, folded and faulted and/ or proglacial wedges may be constructed.

16.3.6. Glaciomarine/Glaciolacustrine Environments In the past decade advances in our knowledge of glaciomarine environments and past glaciomarine sediments has increased dramatically. As exploration of oceanic basins has progressed, data pertinent to glaciomarine sediments have proliferated at a immense rate. However, details of sedimentation processes and rates remain fragmentary at best. The extent and thickness of glaciomarine sediments, both from Pleistocene and pre-Pleistocene glaciations, has meant a major reassessment of the importance of subaqueous glacial sedimentation. Much of what has been learnt from glaciomarine investigations can be applied to the lacustrine environment but other aspects of lacustrine sedimentation, water column thermal stratification and debris release mechanisms differ. Of immense importance is the realization that the interaction between ice masses and lacustrine and marine environments is crucial in understanding many of the processes and styles of sedimentation that occur at the boundary between these environments. For example, in the Great Lakes region of North America and around the North Sea Basin in Europe the interaction of land ice and subaqueous environments is critical in understanding not only the forms of sedimentation that occurred where these different environments interacted but also in explaining the subsequent dynamics of the ice masses themselves. (a) Although much is known concerning subaqueous sedimentation under glacial conditions, there are still many details of sedimentation rates and the impact of ice stresses along groundinglines that remain to be fully explained. For example, the impact varying forms of basal ice movement and thermal conditions have on sediment facies types within a lake or ocean basin is poorly understood. (b) Where ice masses have surged into bodies of water sedimentation patterns and rates are not well known. The tectonic effects of surging on glaciomarine/lacustrine sediments is only known for a few locations. Detection of surging events as portrayed in these sediments is not possible at present.

PROBLEMS AND PERSPECTIVES

16.4. EPILOGUE Science has made enormous progress in understanding glacial environments and it is confirmation of the achievements and ongoing vitality of the discipline that so many new and intriguing questions remain to be solved. Today, the relevance of understanding glaciers and glacial sediments has never been greater, whether in the utilization of ice masses for irrigation purposes, or sediments for agriculture, toxic waste disposal and building materials uses.

481

Research in glacial environments has reached new and exciting phases while the relevance of these environments has become heightened as our awareness of global environmental change and potential long-term implications for the Earth have increased. As threats to the environment from numerous sources and activities increase, it is only too evident in those areas of the world underlain by glacial sediments and covered by glaciers and ice sheets that a vastly improved understanding of ice masses and associated sediments is indispensable.

This Page Intentionally Left Blank

REFERENCES

Aario, R. 1977a. Classification and terminology of morainic landforms in Finland. Boreas, 6: 87–100. Aario, R. 1977b. Associations of flutings, drumlins, hummocks and transverse ridges. Geojournal, 1: 65–72. Aario, R. 1977c. Flutings, drumlins and Rogen-landforms. Nordia, 2: 5–14. Aario, R. 1987. Drumlins of Kuusamo and Rogen-ridges of Ranua, northeast Finland. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 87–102. Aario, R. 1990. Morainic landforms in Northern Finland. In: R. Aario (ed.) Glacial Heritage of Northern Finland, Excursions Guide III Drumlin Symposium. Nordia Tiedonantoja Sarja A No. 1. Oulun Yliopisto Maantieteen Laitos publ., Oulu, pp. 13–28. Aber, J. S. 1985. Definition and model for Kansan glaciation. Institute for Tertiaryen–Quaternary Studies– TER–QUA Symposium Series, 1: 53–60 Aber, J. S. 1988. Ice-shoved hills of Saskatchewan compared with Mississippi delta mudlumps – Implications for glaciotectonic models. In: D. G. Croot (ed.) Glaciotectonics: Forms and Processes. A. A. Balkema, Rotterdam, pp. 1–9. Aber, J. S., Croot, D. G. and Fenton, M. M. 1989. Glaciotectonic Landforms and Structures. Kluwer Academic Publishers, Boston, Dordrecht, 200 pp. Addison, K. 1981. The contribution of discontinuous rockmass failure to glacier erosion. Annals of Glaciology, 2: 3–10. Adhemar, J. A. 1842. R´evolutions de la mer. Privately published, Paris. ´ Agassiz, L. 1840. Etudes sur les glaciers. Privately published, Neuchˆatel, 346 pp.

Ahlmann, H. W. 1948. Glaciological research on the North Atlantic coasts. Royal Geographical Society, Research Series No. 1, 83 pp. Aitken, A. E. 1990. Fossilization potential of Arctic fjord and continental shelf macrofaunas. In: J. A. Dowdeswell and J. D. Scourse (eds) Glacimarine Environments: Processes and Sediments. Geological Society of Canada, Special Publication, 53: 155–176. Albino, K. and Dreimanis, A. 1988. A time-transgressive kinetostratigraphic sequence spanning 180° in a single section at Bradtville, Ontario, Canada. In Croot, D. G. (ed.) Glaciotectonics. A. A. Balkema, Rotterdam, pp. 11–20. Alean, P., Braun, S., Iken, A. Schram, K. and Zwosta, G. 1986. Hydraulic Effects at the Glacier Bed and Related Phenomena. International Workshop 16–19 September, 1985, Interlaken, Switzerland. Mitteilungen der Versuchsanstalt f¨ur Wasserbau, Hydawlogie und Glaziologie, 90, 148 pp. Allen, J. R. L. 1982. Sedimentary Structures: Their Character and Physical Basis. 2 Volumes, Elsevier Science, Amsterdam, 611 pp. and 679 pp. Allen, J. R. L. 1985. Principles of Physical Sedimentology. George Allen & Unwin, London, 272 pp. Allen, P. 1984. Field Guide to the Gipping and Waveney Valleys, Suffolk, May, 1982. Quaternary Research Association, Cambridge, 116 pp. Allen, P., Cheshire, D. A. and Whiteman, C. A. 1991. The tills of southern East Anglia. In: J. Ehlers, P. L. Gibbard and J. Rose (eds) Glacial Deposits in Great Britain and Ireland. A. A. Balkema Publishing, Rotterdam, pp. 255–278. Allen, P. A. and Collinson, J. D. 1986. Lakes. In: H. G. 483

484

REFERENCES

Reading (ed.) Sedimentary Environments and Facies, 2nd edition. Blackwell Scientific Publishers, Oxford, pp. 63–94. Alley, R. B. 1989a. Water-pressure coupling of sliding and bed deformation: I. Water system. Journal of Glaciology, 35: 108–118. Alley, R. B. 1989b. Water-pressure coupling of sliding and bed deformation: II. Velocity-depth profiles. Journal of Glaciology, 35: 119–129. Alley, R. B., Blankenship, D. D., Bentley, C. R. and Rooney, S. T. 1986. Deformation of till beneath ice stream B, West Antarctica. Nature, 322: 57–59. Alley, R. B., Blankenship, D. D., Rooney, S. T. and Bentley, C. R. 1987a. Continuous till deformation beneath ice sheets. International Association of Hydrological Sciences, Publication No. 170: 81–91. Alley, R. B., Blankenship, D. D., Bentley, C. R. and Rooney, S. T. 1987b. Till beneath ice stream B. 3. Till deformation: evidence and implications. Journal of Geophysical Research, 92B: 8921–8929. Alley, R. B., Blankenship, D. D., Rooney, S. T. and Bentley, C. R. 1989a. Water-pressure coupling of sliding and bed deformation: III. Application to Ice Stream B, Antarctica. Journal of Glaciology, 35: 130–139. Alley, R. B., Blankenship, D. D., Rooney, S. T. and Bentley, C. R. 1989b. Sedimentation beneath ice shelves; the view from ice stream B. In: R. D. Powell and A. Elverhøi (eds) Modern Glaciomarine Environments; Glacial and Marine Controls of Modern Lithofacies and Biofacies. Marine Geology, 85: 101–120. Alley, R. B., Blankenship, D. D., Rooney, S. T. and Bentley, C. R. 1989c. (Abstr.) The sedimentary signature of deforming glacier beds in the Ross Embayment, Antarctica. In: N. W. Young (ed.) Proceedings of the Symposium on Ice Dynamics. Annals of Glaciology, 12: 210–211. Alley, R. B., Meese, D. A., Shuman, C. A., Gow, A. J., Taylor, K. C., Grootes, P. M., White, J. W. C., Ram, M., Waddington, E. D., Mayewski, P. A. and Zielinski, G. A. 1993. Abrupt increase in Greenland snow accumulation at the end of the Younger Dryas event. Nature, 362: 527–529. Alley, R. B., Clark, P. U., Keigwin, L. D. and Webb, R. S. 1999. Making sense of millennial-scale changes. In: P. U. Clark, R. S. Webb and L. D. Keigwin (eds) Mechanisms of Global Change at Millenial Time Scales. Geophysical Monograph 112. American Geophysical Union, Washington DC, pp. 385–393. Åmark, M. 1986. Clastic dikes formed beneath an active glacier. Geologiska F¨oreningens i Stockholm F¨orhandlingar, 108: 13–20.

Ambach, W., Ambach, E. and Tributsch, W. 1991. Unusual discoveries of corpses immersed in glacier ice after fatal accidents – glaciological aspects. Journal of Glaciology, 37: 185–186. Anderson, J. B. 1983. Ancient glacial-marine deposits: their spatial and temporal distribution. In: B. F. Molnia (ed.) Glacial-Marine Sedimentation. Plenum, New York, pp. 3–92. Anderson, J. B. and Ashley, G. M. 1991. Glacial Marine Sedimentation: Paleoclimatic Significance. Geological Society of America, Special Paper, 261, 232 pp. Anderson, J. B. and Domack, E. W. 1991. The glacial marine environment: subglacial and climatic setting. Foreword. In: J. B. Anderson and G. M. Ashley (eds) Glacial Marine Sedimentation: Paleoclimatic Significance. Geological Society of America, Special Paper, 261. Anderson, J. B., Kurtz, D. D., Domack, E. W. and Balshaw, K. M. 1980. Glacial and glacial marine sediments of the Antarctic continental shelf. Journal of Geology, 88: 399–414. Anderson, J. B., Kennedy, D. S., Smith, M. J. and Domack, E. W. 1991. Sedimentary facies associated with Antarctica’s floating ice masses. In: J. B. Anderson and G. M. Ashley (eds) Glacial Marine Sedimentation: Paleoclimatic Significance. Geological Society of America, Special Paper, 261, pp. 1–26. Anderson, L. W., 1978. Cirque glacier erosion rates and characteristics of Neoglacial tills, Pangnirtung Fjord area, Baffin Island, North West Territories, Canada. Arctic and Alpine Research, 10: 749–760. Anderson, R. S., Hallet, B., Walder, J. S. and Aubry, B. F. 1982. Observations in a cavity beneath Grinnel Glacier. Earth Surface Processes and Landforms, 7: 63–70. Andrews, J. T. 1971. Englacial debris in glaciers. Journal of Glaciology, 10: 410. Andrews, J. T. 1975. Glacial Systems: An Approach to Glaciers and their Environments. Duxbury Press, North Sciutate MA. Andrews, J. T. 1990. Fjord to deep sea sediment transfers along the northeastern Canadian continental margin: models and data. G´eographie Physique et Quaternaire, 44: 55–70. Andrews, J. T. 1997. Northern Hemisphere (Laurentide) deglaciation: processes and responses of the Ice Sheet/ Ocean Interaction. In I. P. Martini (ed.) Late Glacial and Postglacial Environmental Changes. Oxford University Press, pp. 9–27. Andrews, J. T. 1998. Abrupt changes (Heinrich events) in late Quaternary North Atlantic marine environments: a history and review of data and concepts. Journal of Quaternary Science, 13: 3–16.

REFERENCES

Andrews, J. T. and King, C. A. M. 1968. Comparative till fabrics and till fabric variability in a till sheet and a drumlin; a small scale study. Proceedings of the Yorkshire Geological Society, 36: 435–461. Andrews, J. T and Syvitski, J. P. M 1991. Sediment fluxes along high latitude glaciated continental margins: northeast Canada and eastern Greenland. In Maybeck M. and Hay, W. (eds). Global sedimentary fluxes. NRC/NAS Press. Andrews, J. T. and Tedesco, K. 1992. Detrital carbonaterich sediments, northwestern Labrador Sea: Implications for ice-sheet dynamics and iceberg rafting (Heinrich) events in the North Atlantic. Geology, 20: 1087–1090. Antevs, E. 1922. The recession of the last ice sheet in New England. American Geographical Society Research Series, No. 11, 120 pp. Anthony, R. S. 1977. Iron-rich rhythmically laminated sediments in Lake of the Clouds, northeastern Minnesota. Limnology and Oceanography, 22: 45–54. Archard, J. F. 1953. Contact and rubbing of flat surfaces. Journal of Applied Physics, 24: 981–988. Arkhipov, S. A., Bespaly, V. G., Faustova, M. A., Glushkova, O. Y., Isaeva, L. I. and Velichko, A. A. 1986. Ice-sheet reconstruction. In: V. Œibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 475–483. Arnborg, L. 1955. Hydrology of the glacial river Austurflj´ot, Chapter 7, the Hoffellssandur – a glacial outwash plain. Geografiska Annaler, 37: 185–201. Ashley, G. M. 1979. Sedimentology of a tidal lake, Pitt Lake, British Columbia, Canada. In: Ch. Schl¨uchter (ed.) Moraines and Varves: Origin/Genesis/Classification. A. A. Balkema, Rotterdam, pp. 327–345. Ashley, G. M. 1988. Classification of glaciolacustrine sediments. In Goldthwait R. P. and Matsch, C. L. (eds). Genetic classification of Glacigenic Deposits. A. A. Balkema, Rotterdam, pp.243–260. Ashley, G. M., Southard, J. B. and Boothroyd, J. C. 1982. Deposition of climbing-ripple beds: a flume simulation. Sedimentology, 29: 67–79. Ashley, G. M., Boothroyd, J. C., and Borns, H. W., Jr. 1991. Sedimentology of late Pleistocene (Laurentide) deglacialphase deposits, eastern Maine; an example of a temperate marine grounded icesheet margin. In: J. B. Anderson and G. M. Ashley (eds) Glacial Marine Sedimentation; Paleoclimatic Significance. Geological Society of America Special Paper 261, pp. 107–125. Ashton, N. M., Cook, J., Lewis, S. G. and Rose, J. (eds) 1992. High Lodge: Excavations by G. de G. Sieveking 1962–68 and J. Cook 1988. British Museum Press, London, 192 pp.

485

Ashworth, P. J. and Ferguson, R. I. 1986. Interrelationships of channel processes, changes and sediments in a proglacial braided river. Geografiska Annaler, 68A: 361–371. Atkinson, B. K. 1984. Subcritical crack growth in geological materials. Journal of Geophysical Research, 89B: 4077–4114. Attig, J. W. and Clayton, L. 1993. Stratigraphy and origin of an area of hummocky glacial topography, northern Wisconsin, U.S.A. Quaternary International, 18: 61–68. Augustinus, P. C. 1992. The influence of rock mass strength on glacial valley cross-profile morphometry: a case study from the Southern Alps, New Zealand. Earth Surface Processes and Landforms, 17: 39–51. Aylsworth, J. M. and Shilts, W. W. 1985. Glacial features of the west-central Canadian Shield. Geological Survey of Canada, Current Research, part B, Paper 85–1B, pp. 375–381. Aylsworth, J. M. and Shilts, W. W. 1989. Bedforms of the Keewatin Ice Sheet, Canada. Sedimentary Geology, 62: 407–428. Bagnold, R. A. 1954. Experiments on a gravity-free dispersion of large solid spheres in a Newtonian fluid under shear. Proceedings of the Royal Society of London, 225A: 49–63. Baker, V. R., Kochel, R. C. and Patton, P. C. 1988. Flood Geomorphology. J. Wiley & Sons, New York, 503 pp. Banham, P. H. 1975. Glacitectonic structures: a general discussion with particular reference to the contorted drift of Norfolk. In: A. E. Wright and F. Moseley (eds) Ice Ages: Ancient and Modern. Geological Journal Special Issue 6. Seel House Press, Liverpool, pp. 69–94. Banham, P. H. 1977. Glacitectonites in till stratigraphy. Boreas, 6: 101–105. Banham, P. H. 1988. Polyphase glaciotectonic deformation in the Contorted Drift of Norfolk. In: D. G. Croot (ed.) Glaciotectonics: Forms and Processes. A. A. Balkema, Rotterdam, pp. 27–32. Baranowski, S. 1970. The origin of fluted moraine at the fronts of contemporary glaciers. Geografiska Annaler, 52A: 68–75. Baranowski, S. 1979. The origin of drumlins as an ice–rock interface problem. Journal of Glaciology, 23: 435–436. Bard, E., Fairbanks, R. G., Hamelin, B., Zindler, A. and Hoang, C. T. 1991. Uranium – 234 anomalies in corals older than 150, 000 years, Geochimica et Cosmochimica Acta, 55:2385–2390. Bard, E. and Broecker, W. S. 1992. The Last Deglaciation: Absolute and Radiocarbon Chronologies. Springer-Verlag, Berlin, 344 pp.

486

REFERENCES

Barnett, D. M. and Holdsworth, G. 1974. Origin, morphology, and chronology of sublacustrine moraines, Generator Lake, Baffin Island, Northwest Territories, Canada. Canadian Journal of Earth Sciences, 11: 380–408. Barnett, P. J. 1990. Tunnel valleys: evidence of catastrophic release of subglacial meltwater, central-southern Ontario, Canada. Geological Society of America, Abstracts with Programs, 22, 3. Barrett, P. J., Hambrey, M. J., Harwood, D. M. Pyne, A. R. and Webb, P. N. 1989. General Synthesis. In: P. J. Barrett (ed.) Antarctic Cenozoic History of the CIROS-I Drillhole. Bulletin in the Miscellaneous Series of the New Zealand Department of Scientific and Industrial Research, No. 245, pp. 241–251. Barrett, P. J., Adams, C. J., McIntosh, W. C., Swisher, III, C. C. and Wilson, G. S. 1992. Geochronological evidence supporting Antarctic deglaciation three million years ago. Nature, 359: 816–818. Bartek, L. R., Vail, R. R., Anderson, J. B., Emmet, P. A. and Wu, S. 1991. Effect of Cenozoic ice sheet fluctuations in Antarctica on the stratigraphic signature of the Neogene. Journal of Geophysical Research, 96: 6753–6778. Beecroft, I. 1983. Sediment transport during an outburst from Glacier de Tsidjiore Nouve, Switzerland, 16–19 June 1981. Journal of Glaciology, 29: 185–190. Beget, J. E. 1986. Modelling the influence of till rheology on the flow and profile of the Lake Michigan lobe, southern Laurentide ice sheet, USA. Journal of Glaciology, 32: 235–41. Behrendt, J. C. and Cooper, A. 1991. Evidence of rapid Cenozoic uplift of the shoulder escarpment of the Cenozoic West Antarctic rift system and a speculation on possible climatic forcing. Geology, 19: 315–319. Bell, M. and Walker, M. J. C. 1992. Late Quaternary Environmental Change: Physical and Human Perspectives. Longman, London, 273 pp. Benn, D. I. 1989. Controls on sedimentation in a late Devensian ice-dammed lake, Achnasheen, Scotland. Boreas, 18: 31–42. Benn, D. I. 1995. Fabric signature of subglacial till deformation, Breidamerkurjokull, Iceland. Sedimentology, 42:735–747. Benn, D. I. and Evans, D. J. A. 1996. The interpretation and classification of subglacially-deformed materials. Quaternary Science Reviews, 15: 23–52. Bennett, M. R. and Boulton, G. S. 1993. A reinterpretation of Scottish ‘hummocky moraine’ and its signficance for the deglaciation of the Scottish Highlands during the Younger Dryas or Loch Lomond Stadial. Geological Magazine, 130: 301–318.

Benson, C. S. 1962. Stratigraphic studies in the snow and firn of the Greenland ice sheet. U.S. Army CRREL Rept. 70, p. 107. Berger, A. and Loutre, M. F. 1988. New insolation values for the climate of the last 10 million years. Scientific Report 1988/13, Institut d’Astronomie et de Geophysique Georges Lemaitre, Universite Catholique de Louvain-laNeuve, Belgium. Berger, A. and Loutre, M. F. 1991. Insolation values for the climate of the last 10 million years. Quaternary Science Reviews, 10: 297–317. Berger, A. L. 1981. The astronomical theory of paleoclimates. In: A. L. Berger (ed.) Climatic Variations and Variability: Facts and Theories. D. Reidel, Dordrecht, pp. 501–525. Berger, W. H. 1990. The Younger Dryas cold spell – a quest for causes. Palaeogeography, Palaeoclimatology, Palaeoecology (Global and Planetary Change Section), 89: 219–237. Berner, R. A. 1990. Atmospheric carbon dioxide over Phanerozoic time. Science, 249: 1382–1386. Bernhardi, A. 1832. Wie Kamen die aus dem Norden Stammenden Felsbruchst¨ucke und Geschiebe, welche man in Norddeutschland und den benachbarten L¨andern findet, an ihre genenw¨artigen Fundorte? Jahrbuch f¨ur Mineralogie, Geognosie und Petrefaktenkunde, 3: 257–267. Berthelsen, A. 1973. Weichselian ice advance and drift succession in Denmark. Bulletin of the Geological Institute University of Uppsala, New Series 5: 21–29. Berthelsen, A. 1978. The methodology of kineto-stratigraphy as applied to glacial geology. Bulletin of the Geological Society of Denmark, Special Issue, 27: 25–38. Bezinge, A. 1987. Glacial meltwater streams, hydrology and sediment transport: the case of the Grande Dixence hydroelectricity scheme. In: A. M. Gurnell and M. J. Clark (eds) Glacio-Fluvial Sediment Transfer – an Alpine Perspective. J.Wiley and Sons Publishing, Chichester, pp. 473–498. Bindschadler, R., Harrison, W. D., Raymond, C. F. and Crosson, R. 1977. Geometry and dynamics of a surgetype glacier. Journal of Glaciology, 18: 181–194. Bishop, J. F. and Walton, J. W. L. 1981. Bottom melting under the George VI ice shelf, Antarctica. Journal of Glaciology, 27: 429–447. Bjørlykke, K., Bue, B. and Elverhøi, A. 1978. Quaternary sediments in the northwestern part of the Barents Sea and their relation to the underlying Mesozoic bedrock. Sedimentology, 25: 227–246. Bj¨ornsson, H. 1974. Explanation of j¨okulhlaups from Gr´ımsv¨otn, Vatnaj¨okull, Iceland. J¨okull, 24: 1–26.

REFERENCES

Bj¨ornsson, H. 1977. The cause of j¨okulhlaups in the Skafta River, Vatnaj¨okull, Iceland. J¨okull, 27: 71–78. Bj¨ornsson, H. 1988. Hydrology of Ice Caps in Volcanic Regions. V´ısindaf´elag ´Islendinga, Societas Scientarium Islandica, 45, Reykjav´ık. 139 pp. Bj¨ornsson, H. 1992. J¨okulhlaups in Iceland: prediction, characteristics and simulation. Annals of Glaciology, 16: 95–106. Bj¨ornsson, H. 1998. Hydrological characteristics of the drainage system beneath a surging glacier. Nature, 395: 771–774. Blankenship, D. D., Bentley, C. R., Rooney, S. T. and Alley, R. B. 1986. Seismic measurements reveal a saturated, porous layer beneath an active Antarctic Ice Stream. Nature, 322: 54–57. Blankenship, D. D., Bentley, C. R., Rooney, S. T. and Alley, R. B. 1987. Till beneath ice stream B, 1. Properties derived from seismic travel times. Journal of Geophysical Research, 92: 8903–8911. Blankenship, D. D., Rooney, S. T., Alley, R. B. and Bentley, C. R. 1989. Seismic evidence for a thin basal layer at a second location on ice stream B, Antarctica. (Abstract). Annals of Glaciology, 12: 200. Blatter, H. and Hutter, K. 1991. Polythermal conditions in Arctic glaciers. Journal of Glaciology, 37: 261–269. Blatter, H. and Kappenberger, G. 1988. Mass balance and thermal regime of Laika Ice Cap, Coburg Island. N.W.T. Canada. Journal of Glaciology, 34: 102–110. Blewett, W. L. and Rieck, R. L. 1987. Reinterpretation of a portion of the Munising moraine in northern Michigan. Bulletin of the Geological Society of America, 98: 169–175. Bluck, B.J. 1979. Structure of coarse grained braided stream alluvium. Transactions of the Royal Society of Edinburgh, 70: 181–221. Bluck, B. J. 1982. Texture of gravel bars in braided streams. In: R. D. Hey, J. C. Bathurst and C. R. Thorne (eds) GravelBed Rivers; Fluvial Processes, Engineering and Management. John Wiley and Sons, Chichester, pp. 339–355. Bluck, B. J. 1987. Bed forms and clast size changes in gravel-bed rivers. In: K. S. Richards (ed.) River Channels: Environment and Process. Institute of British Geography, Special Publication, 18, pp. 159–178. Bluemle, J. P. and Clayton, L. 1984. Large-scale glacial thrusting and related processes in North Dakota. Boreas, 13: 279–299. Bogacki, M. 1973. Geomorphological and geological analysis of the proglacial area of Skeidararj¨okull; central western and eastern sections, Scientific results of the Polish Geographical Expedition to Vatnaj¨okull (Iceland), 1968. Geographia Polonica, 26: 57–88.

487

Bond, G., Heinrich, H., Broecker, W., Labeyrie, L., McManus, J., Andrews, J., Huon, S., Jantschik, R., Clasen, S., Simet, C., Tedesco, K., Klas, M., Bonani, G. and Ivy, S. 1992. Evidence of massive discharges of icebergs into the North Atlantic Ocean during the last glacial period. Nature, 360: 245–249. Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L., Jouzel, J. and Bonani, G. 1993. Correlations between climate records from North Atlantic sediments and Greenland ice. Nature, 364: 407–412. Booth, D. B. and Hallet, B. 1993. Channel networks carved by subglacial water: Observations and reconstructions in the eastern Puget Lowland of Washington. Geological Society of America Bulletin, 105: 671–683. Boothroyd, J. C. and Ashley, G. M. 1975. Processes, bar morphology and sedimentary structures on braided outwash fans, northeastern gulf of Alaska. In: A. V. Jopling and B. C. MacDonald (eds) Glaciofluvial and Glaciolacustrine Sedimentation. Society of Economic Paleontologists and Mineralogists, Special Publication 23, pp. 193–222. Boothroyd, J. C. and Nummedal, D. 1978. Proglacial braided outwash: a model for humid alluvial-fan deposits. In A. D. Miall (ed.) Fluvial Sedimentology. Canadian Society of Petroleum Geologists, Memoir 5, Calgary, Alberta, Canada, pp. 641–668. Borns, H. W. 1973. Late Wisconsin fluctuations of the Laurentide Ice Sheet in southern and eastern New England. In: R. F. Black, R. P. Goldthwait and H. B. Willman (eds) The Wisconsinan stage. Geological Society of America Memoir, 136: 37–45. Bouchard, M. A. 1980. Late Quaternary geology of the T´emiskimie area, central Qu´ebec, Canada. Unpublished Ph.D. Thesis, McGill University, Montr´eal. 284 pp. Bouchard, M. A. 1989. Subglacial landforms and deposits in central and northern Qu´ebec, Canada, with emphasis on R¨ogen moraines. Sedimentary Geology, 62: 293–308. Boulton, G. S. 1970. On the deposition of subglacial and melt-out tills at the margins of certain Svalbard glaciers. Journal of Glaciology, 9: 231–245. Boulton, G. S. 1971. Till genesis and fabric in Svalbard, Spitsbergen. In R. P. Goldthwait (ed.) Till: a Symposium. Ohio State University Press, Columbus, pp. 41–72. Boulton, G. S. 1972. Modern Arctic glaciers as depositional models for former ice sheets. Quarterly Journal of the Geological Society of London, 128: 361–393. Boulton, G. S. 1974. Processes and patterns of glacial erosion. In: D. R. Coates (ed.) Glacial Geomorphology. Binghampton, New York, pp. 41–87. Boulton, G. S. 1975. Processes and patterns of subglacial sedimentation: a theoretical approach. In: A. E. Wright

488

REFERENCES

and F. Moseley (eds) Ice Ages: Ancient and Modern. Geological Journal, Special Issue, 6: 7–42. Boulton, G. S. 1976. The origin of glacially fluted surfaces: observations and theory. Journal of Glaciology, 17: 287–309. Boulton, G. S. 1978. Boulder shapes and grain-size distributions of debris as indicators of transport paths through a glacier and till genesis. Sedimentology, 25: 773–799. Boulton, G. S. 1979. Processes of glacier erosion on different substrata. Journal of Glaciology, 23: 15–37. Boulton, G. S. 1980. Genesis and classification of glacial sediments. In: W. Stankowski (ed.) Tills and Glacigene Deposits. Uniwersytet im. Adama Mickiewicza, Pozna˜n. Seria Geografia, No. 20, pp. 15–18. Boulton, G. S. 1983. Debris and isotopic sequences in the basal layers of polar ice sheets. In: G.deQ. Robin (ed.) The Climatic Record in Polar Ice Sheets. Cambridge University Press, Cambridge, pp. 83–89. Boulton, G. S. 1987a. A theory of drumlin formation by subglacial sediment deformation. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema, Rotterdam, pp. 25–80. Boulton, G. S. 1987b. Progress in glacial geology during the last fifty years. Journal of Glaciology, Special Issue: 25–32. Boulton, G. S. 1990. Sedimentary and sea level changes during glacial cycles and their control on glacimarine facies architecture. In: J. A. Dowdeswell and J. D. Scourse (eds) Glacimarine Environments: Processes and Sediments. Geological Society of London, Special Publication No. 53, pp. 15–52. Boulton, G. S. 1996a. The origin of till sequences by subglacial sediment deformation beneath mid-latitude ice sheets. Annals of Glaciology, 22: 75–84. Boulton, G. S. 1996b. Theory of glacial erosion, transport and deposition as a consequence of subglacial sediment deformation. Journal of Glaciology, 42: 43–62. Boulton, G. S. and Clark, C. D. 1990a. A highly mobile Laurentide ice sheet revealed by satellite images of glacial lineations. Nature, 346: 813–817. Boulton, G. S. and Clark, C. D. 1990b The Laurentide Ice Sheet through the last glacial cycle: drift lineations as a key to the dynamic behaviour of former ice sheets. Transactions of the Royal Society of Edinburgh, Earth Sciences, 81: 327–347. Boulton, G. S. and Dent, D. L. 1974. The nature and rates of post–depositional changes in recently deposited till from south east Iceland. Geografiska Annaler, 56A: 121–134. Boulton, G. S. and Deynoux, M. 1981. Sedimentation in

glacial environments and the identification of tills and tillites in ancient sedimentary sequences. Precambrian Research, 15: 397–422. Boulton, G. S. and Dobbie, K. E. 1993. Consolidation of sediments by glaciers; relations between sediment geotechnics, soft-bed glacier dynamics and subglacial ground-water flow. Journal of Glaciology, 39: 26–44. Boulton, G. S. and Dobbie, K. E. 1998. Slow flow of granular aggregates: The deformation of sediments beneath glaciers. Philosophical Transactions – Mathematical Physical and Engineering Sciences, 356: 2713–2747. Boulton, G. S. and Eyles, N. 1979. Sedimentation by valley glaciers; A model and genetic classification. In Ch. Schl¨uchter (ed.) Moraines and Varves: Origin/Genesis/ Classification. A.A. Balkema Publishing, Rotterdam, pp. 11–23. Boulton, G. S. and Hindmarsh, R. C. A. 1987. Sediment deformation beneath glaciers: rheology and geological consequences. Journal of Geophysical Research, 92B: 9059–9082. Boulton, G. S. and Jones, A. S. 1979. Stability of temperate ice caps and ice sheets resting on beds of deformable sediment. Journal of Glaciology, 24: 29–43. Boulton, G. S. and Spring, U. 1986. Isotopic fractionation at the base of polar and sub-polar glaciers. Journal of Glaciology, 32: 475–485. Boulton, G. S., Dent, D. L. and Morris, E. M. 1974. Subglacial shearing and crushing, and the role of water pressures in tills from south-east Iceland. Geografiska Annaler, 56A: 134–145. Boulton, G. S., Jones, A. S, Clayton, K. M. and Kenning, M. J. 1977. A British Ice-sheet model and patterns of glacial erosion and deposition in Britain. In Shotton, F.W. (ed.), British Quaternary Studies: Recent Advances, Clarendon Press, Oxford, pp. 231–246. Boulton, G. S., Morris, E. M., Armstrong, A. A. and Thomas, A. 1979. Direct measurement of stress at the base of glacier. Journal of Glaciology, 22: 3–24. Boulton, G. S., Cox, F. C., Hart, J. K. and Thornton, M. 1984. The glacial geology of Norfolk. Bulletin of the Geological Society of Norfolk, 34: 103–122. Boulton, G. S., Smith, G. S., Jones, A. S. and Newsome, J. 1985. Glacial geology and glaciology of the last midlatitude ice sheets. Journal of Geological Society of London, 142: 447–474. Boulton, G. S., van der Meer, J. J. M., Ruegg, G. H. J., Beets, D. J., Riezebos, P. A., Castel, I. I. M., Hart, J. K., Quinn, I., Thornton, M. and Van der Wateren, F. M. 1989. Preliminary report on the Glacitecs ’84 expedition to Spitsbergen. Internal Report of the Fysisch-Geografisch en Bodemkundig Laboratorium. University of Amsterdam.

REFERENCES

Boulton, G. S., van der Meer, J. J. M., Hart, J. K., Beets, D., Ruegg, G. H. J., van der Wateren, F. M. and Jarvis, J. 1996. Till and moraine emplacement in a deforming bed surge – an example from a marine environment. Quaternary Science Reviews, 15: 961–988. Boulton, G. S., van der Meer, J. J. M., Beets, D. J., Hart, J. K. and Ruegg, G. H. J. 1999. The sedimentary and structural evolution of a recent push moraine complex: Holmstrømbreen, Spitsbergen. Quaternary Science Reviews, 18: 339–371. Bowen, D. Q. 1978. Quaternary Geology; A Stratigraphic Framework for Multidisciplinary Work. Pergamon Press, Oxford, 221 pp. Bowen, D. Q. 1994. The Pleistocene of north west Europe. Science Progress, 76: 209–223. Bowen, D. Q. 1999. A revised correlation of Quaternary deposits in the British Isles. Geological Society Special Report No. 23, 174 pp. Bowen, D. Q., Rose, J. McCabe, A. M. and Sutherland, D. G. 1986a. Correlation of Quaternary glaciations in England, Ireland, Scotland and Wales. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 299–340. Bowen, D. Q., Richmond, G. M., Fullerton, D. S., Sibrava, V., Fulton, R. J. and Velichko, A. A. 1986b. Correlation of Quaternary Glaciations in the Northern Hemisphere. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 509–510 and chart 1. Bowen, D. Q., Sykes, G. A. and Turner, C. 1988. Correlation of marine events and glaciations on the northeast Atlantic margin; discussion. In: N. J. Shackleton, R. G. West and D. Q. Bowen (eds) The Past Three Million Years: Evolution of Climatic Variability in the North Atlantic Region. Cambridge University Press, Cambridge, pp. 209–225. Boyd, R., Scott, D. B. and Douma, M. 1988. Glacial tunnel valleys and Quaternary history of the outer Scotian shelf. Nature, 333: 61–64. Boyle, E. A. 1988. Cadmium: chemical tracer of deepwater paleoceanography. Paleoceanography, 3: 471–489. Bozhinskiy, A. N., Krass, M. S. and Popovnin, V. V. 1986. Role of debris cover in the thermal physics of glaciers. Journal of Glaciology, 32: 255–266. Bradley, W. C., Fahnestock, R. K. and Rowekamp, E. T. 1972. Coarse sediment transport by flood flows on Knik river, Alaska. Geological Society of America, Bulletin, 83: 1261–1284. Bradshaw, R. and Thompson, R. 1985. The use of magnetic measurements to investigate the mineralogy of Icelandic

489

lake sediments and to study catchment processes. Boreas, 14: 203–215. Brand, G., Pohjola, V. and Hooke, R. LeB. 1987. Evidence for a till layer beneath Storglaci¨aren, Sweden, based on electrical resistivity measurements. Journal of Glaciology, 33: 311–314. Breitkopf, J. H. 1988. Iron formations related to mafic volcanism and ensialic rifting in the southern margin zone of the Damara orogen, Namibia. Precambrian Research, 38: 111–130. Brennand, T. C. and Sharpe, D. R. 1993. Ice-sheet dynamics and subglacial meltwater regime inferred from form and sedimentology of glaciofluvial systems: Victoria Island, District of Franklin, Northwest Territories. Canadian Journal of Earth Sciences, 30: 928–944. Bretz, J. H. 1925. The Spokane flood beyond the channelled scablands. Journal of Geology, 33: 97–115. Brodzikowski, K. and van Loon, A. J. 1980. Sedimentary deformations in Saalian glaciolimnic deposits, near ´ Wl/ ost´ow (Zary area, Western Poland). Geologie en Mijnbouw, 59: 251–272. Brodzikowski, K. and van Loon, A. J. 1983. Sedimentology and deformational history of unconsolidated Quaternary sediments of the Jarosz´ow zone (Sudetic Foreland). Geologica Sudetica, 18:123–195. Brodzikowski, K. and van Loon, A. J. 1991. Glacigenic Sediments. Developments in Sedimentology, 49. Elsevier Science Publications, Amsterdam, 674 pp. Broecker, W. S. 1986. Oxygen isotope constraints on surface ocean temperatures. Quaternary Research, 26: 121–134. Broecker, W. S. 1992. Climate cycles; upset for Milankovitch theory. Nature, 359: 779–780. Broecker, W. S. and Denton, G. H. 1989. The role of ocean– atmosphere reorganizations in glacial cycles. Geochimica et Cosmochimica Acta, 53: 2465–2501. Broecker, W. S. and Peng, T.-H. 1982. Tracers in the Sea. Eldigio Press, Palisades, 690 pp. Broecker, W. S. and van Donk, J. 1970. Insolation changes, ice volumes and the O18 record in deep-sea cores. Review of Geophysics and Space Physics, 8: 169–197. Broecker, W. S., Bond, G., Klas, M., Bonani, G. and W¨olfli, W. 1990. A salt oscillator in the glacial Atlantic? 1. The Concept. Paleoceanography, 5: 469–477. Broecker, W. S., Bond, G., McManus, J., Klar, M. and Clark, E. 1992. Origin of the North Atlantic’s Heinrich events. Climate Dynamics, 6: 265–273. Brown, C. S., Meier, M. F. and Post, A. 1982. Calving speed of Alaska tidewater glaciers, with application to Columbia Glacier. United States Geological Survey, Professional Paper, 1258-C, 13 pp.

490

REFERENCES

Brown, N. E., Hallet, B. and Booth, D. B. 1987. Rapid soft bed sliding of the Puget Glacial Lobe. Journal of Geophysical Research, 92: 8985–8997. Budd, W. F. 1970. The longitudinal stress and strain-rate gradients in ice masses. Journal of Glaciology, 9: 19–27. Budd, W. F. 1975. A first simple model for periodically selfsurging glaciers. Journal of Glaciology, 14: 3–21. Budd, W. F. and Jenssen, D. 1975. Numerical modelling of glacier systems. International Association of Hydrological Sciences, 104: 257–291. Budd, W. F. and McInnes, B. J. 1974. Modelling Periodically Surging Glaciers. Science, 186: 925–927. Budd, W. F. and Smith, I. N. 1981. The growth and retreat of ice sheets in response to orbital radiation changes. International Association of Hydrological Sciences, 131: 369–409. Budd, W. F., McInnes, B. J., Jenssen, D. and Smith, I. N. 1987. Modelling the response of the West Antarctic Ice Sheet to a climatic warming. In: C. J. Van der Veen and J. Oerlemans (eds) Dynamics of the West Antarctic Ice Sheet. D. Reidel, Dordrecht, pp. 321–358. Bugge, T., Knarud, R. and Mørk, A. 1984. Bedrock geology on the mid-Norwegian continental shelf. In: A. M. Spencer, S. O. Johsen, A. Mørk, E. Nysaether, P. Songstad and A. Spinnangr (eds) Petroleum Geology of the North European Margin. London, Graham & Trotman, pp. 271–283. Bugge, T., Belderson, R. H. and Kenyon, N. H. 1988. The Storegga slide. Philosophical Transactions of the Royal Society of London, Series A, 325: 357–388. Burke, M. J. 1969. The Forth valley: an ice-moulded lowland. Transactions of the Institute of British Geographers, 48: 51–59.

Cai, J., Powell, R. D. and Loubere, P. 1992. Source components of glacimarine sediments defined by Fourier shapes of quartz silt particles. Geological Society of America, Abstracts with Programs, 24: 84. Calabrese, E. A. and Syvitski, J. P. M. 1987. Modelling the growth of a growth prograding delta: numerics, sensitivity, program code, and users guide. Geological Survey Canada, Open File, 1624, 64 pp. Calvert, S. E. 1983. Sedimentary geochemistry of silicon. In: S. R. Aston (ed.) Silicon Geochemistry and Biogeochemistry. Academic Press, New York, pp. 143–186. Campbell, W. J. 1973. Structure and inferred circulation of South Cascade Lake, Washington, USA. International Association of Scientific Hydrology, Publication 95: 259–262.

Cant, D. J. 1978. Development of a facies model for sandy braided river sedimentation: Comparison of the South Saskatchewan River and the Battery Point Formation. In A. D. Miall (ed.) Fluvial Sedimentology. Canadian Society of Petroleum Geologists, Memoir 5, Calgary, Alberta, Canada, pp. 627–639. Cant, D. J. and Walker, R. G. 1978. Fluvial processes and facies sequences in the sandy braided South Saskatchewan River, Canada. Sedimentology, 25: 625–648. Caputo, M. V. and Crowell, J. C. 1985. Migration of glacial centers across Gondwana during Paleozoic Era. Geological Society of America, Bulletin, 96: 1020–1036. Carol, H. 1947. The formation of roches moutonnees. Journal of Glaciology, 1: 57–59. Carruthers, R. G. 1953. Glacial Drifts and the Undermelt Theory. Harold Hill Publishers, Newcastle-upon-Tyne, 42 pp. Cepek, A. G. 1986. Quaternary stratigraphy of the German Democratic Republic. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 359–364. Chamberlin, T. C. 1888. The rock-scourings of the great ice invasions. US Geological Survey, Annual Report 7: 147–248. Chambers, R. 1853. On glacial phenomena in Scotland and parts of England. Edinburgh New Philosophical Journal, 54: 229–281. Chapple, W. M. 1978. Mechanics of thin-skinned fold-andthrust belts. Geological Society of America, Bulletin, 89: 1189–1198. Charlesworth, J. K. 1957. The Quaternary Era, with Special References to its Glaciation. Edward Arnold, London, 1700 pp. Cheel, R. J. and Rust, B. R. 1986. A sequence of softsediment deformation (dewatering) structures in late Quaternary subaqueous outwash near Ottawa, Canada. Sedimentary Geology, 47: 77–93. Chernova, L. P. 1981. Influence of mass balance and run-off on relief-forming activity of mountain glaciers. Annals of Glaciology, 2: 69–70. Chester, F. M. and Logan, J. M. 1987. Composite planar fabric of gouge from the Punchbowl Fault, California. Journal of Structural Geology, 9: 621–634. Chinn, T. J. H. 1989. Single folds at the margins of drybased glaciers as indicators of a glacial advance. Annals of Glaciology, 12: 23–30. Chorley, R. J., Dunn, A. J. and Beckinsale, R. P. 1964. The History of the Study of Landforms or the Development of Geomorphology, Volume 1. Geomorphology before Davis. John Wiley, New York, 678 pp.

REFERENCES

Choukroune, P., Gapais, D. and Merle, O. 1987. Shear criteria and structural symmetry. Journal of Structural Geology, 9: 525–530. Christiansen, E. A. 1971. Geology and Groundwater Resources of the Melville Area (62K, L) Saskatchewan. Saskatchewan Research Council, Geology Division, Map No. 12. Chumakov, N. M. and Elston, D. P. 1989. The paradox of Late Proterozoic glaciations at low latitudes. Episodes, 12: 115–120. Church, M. 1972. Baffin Island Sandurs: a study of Arctic fluvial processes. Geological Survey of Canada, Bulletin, 216, 280 pp. Church, M. and Gilbert, R. 1975. Proglacial fluvial and lacustrine environments. In: A. V. Jopling and B. C. McDonald (eds), Glaciofluvial and Glaciolacustrine Sedimentation. Society of Economical Paleontologists and Mineralogists, Special Publication No.23, Tulsa, pp. 22–100. Church, M. and Jones, D. 1982. Channel bars in gravel-bed rivers. In: R. D. Hey, J. C. Bathurst and C. R. Thorne (eds) Gravel-Bed Rivers; Fluvial Processes, Engineering and Management. John Wiley and Sons, Chichester, pp. 291–338. Church, M. and Ryder, J. M. 1972. Paraglacial sedimentation: a consideration of fluvial processes conditioned by glaciation. Geological Society of America, Bulletin, 83: 3059–3072. Churski, Z. 1973. Hydrographic features on the proglacial area of Skeidar´arj¨okull. Geographia Polonica, 26: 209–254. Clague, J. J. and Mathews, W. H. 1973. The magnitude of j¨okulhlaups. Journal of Glaciology, 12: 501–504. Clague, J. J. and Evans, S. G. 1997. The 1994 jokulhlaup at Farrow Creek, British Columbia, Canada. Geomorphology, 19: 77–87. Clapperton, C. M. 1975. The debris content of surging glaciers in Svålbard and Iceland. Journal of Glaciology, 14: 395–406. Clapperton, C. M. 1989. Asymmetrical drumlins in Patagonia, Chile. Sedimentary Geology, 62: 387–398. Clapperton, C. M. 1990. Quaternary glaciations in the southern hemisphere: An overview. Quaternary Science Reviews, 9: 299–304. Clapperton, C. M. 1993. Quaternary Geomorphology and Geomorphology of South America. Elsevier, Amsterdam, 796 pp. Clapperton, C. M. and Sudgen, D. E. 1990. Late Cenozoic glacial history of the Ross Embayment, Antarctica. Quaternary Science Reviews, 9: 253–272. Clark, C. D. 1993. Mega-scale glacial lineations and cross-

491

cutting ice-flow landforms. Earth Surface Processes and Landforms, 18: 1–29. Clark, C. D. 1994. Large-scale ice moulding: a discussion of the genesis and glaciological significance. Sedimentary Geology, 91: 253–268. Clark, C. D., Knight, J. K. and Gray, J. T. 2000. Geomorphological reconstruction of the Labrador Sector of the Laurentide Ice Sheet. Quaternary Science Reviews, 19: 1343–1366. Clark, P. U. 1997. Sediment deformation beneath the Laurentide Ice Sheet. In: I. P. Martini (ed.) Late Glacial and Postglacial Environmental Changes. Oxford University Press, pp. 81–97. Clark, P. U. and Walder, J. S. 1994. Sublacial drainage, eskers, and deforming beds beneath the Laurentide and Eurasian ice sheets. Bulletin of the Geological Society of America, 106: 304–316. Clark, P. U., Licciardi, C. M., MacAyeal, D. R. and Jenson, J. W. 1996. Numerical reconstruction of a soft-bedded Laurentide Ice Sheet during the last glacial maximum. Geology, 23: 679–682. Clarke, G. K. C. 1976. Thermal regulation of glacier surging. Journal of Glaciology, 16: 231–250. Clarke, G. K. C. 1982. Glacier outburst floods from ”Hazard Lake”, Yukon Territory, and the problems of flood magnitude prediction. Journal of Glaciology, 28: 3–21. Clarke, G. K. C. 1987a. A short history of scientific investigations on glaciers. Journal of Glaciology, Special Issue: 4–24. Clarke, G. K. C. 1987b. Subglacial till: a physical framework for its properties and processes. Journal of Geophysical Research, 92B: 9023–9036. Clarke, G. K. C. 1987c. Fast glacier flow: ice streams, surging, and tidewater glaciers. Journal of Geophysical Research, 92B: 8835–8841. Clarke, G. K. C. 1991. Length, width and slope influences on glacier surging. Journal of Glaciology, 37: 236–246. Clarke, G. K. C. and Blake, E. W. 1991. Geometric and thermal evolution of a surge-type glacier in its quiescent state: Trapridge Glacier, Yukon Territory, Canada, 1969–89. Journal of Glaciology, 37: 158–169. Clarke, G. K. C. and Collins, S. G. 1984. The 1981–1982 surge of Hazard Glacier, Yukon Territory. Canadian Journal of Earth Sciences, 21: 297–304. Clarke, G. K. C. and Waldron, D. A. 1984. Simulation of the August 1979 sudden discharge of glacier-dammed Flood Lake, British Columbia. Canadian Journal of Earth Sciences, 21: 502–504. Clarke, G. K. C., Nitsan, U. and Paterson, W. S. B. 1977. Strain heating and creep instability in glaciers and ice

492

REFERENCES

sheets. Reviews of Geophysics and Space Physics, 15: 235–247. Clarke, G. K. C., Collins, S. G. and Thompson, D. E. 1984. Flow, thermal structure, and subglacial conditions of a surge-type glacier. Canadian Journal of Earth Sciences, 21: 232–240. Clarke, G. K. C., Schmock, J. P., Ommanney, C. S. L. and Collins, S. G. 1986. Characteristics of surge-type glaciers. Journal of Geophysical Research, 91: 7165–7180. Clayton, L. and Moran S. R. 1974. A glacial process-form model. In: D. R. Coates (ed.) Glacial Geomorphology. SUNY at Binghamton, pp. 89–119. Clayton, L., Teller, J. T. and Attig, J. W. 1985. Surging of the southwestern part of the Laurentide Ice Sheet. Boreas, 14: 235–241. Clayton, L., Mickelson, D. M. and Attig, J. W. 1989. Evidence against pervasively deformed bed material beneath rapidly moving lobes of the southern Laurentide Ice Sheet. In: J. Menzies and J. Rose (eds) Subglacial bedforms – drumlins, Rogen moraine and associated subglacial bedforms. Sedimentary Geology, 62 (2/4): 203–208. Clement, P. 1984. The drainage of a marginal ice-dammed lake at Nordbogletscher, Johan Dahl Land, South Greenland. Arctic and Alpine Research, 16: 209–216. Clemmensen, L. B. and Houmark-Nielsen, M. 1981. Sedimentary features of a Weichselian glaciolacustrine delta. Boreas, 10: 229–245. CLIMAP Project Members 1981. Seasonal Reconstructions of the Earth’s Surface at the Last Glacial Maximum. Geological Society of America, Map and Chart Series MC-36. Coard, M. A., Cousen, S. M., Cuttler, A. H., Dean, H. J., Dearing, J. A., Eglinton, T. I., Greaves, A. M., Lacey, K. P., O’Sullivan, P. E. and Pickering, D. A. 1983. Paleolimnological studies of annually-laminated sediments in Loe Pool, Cornwall, U.K. Hydrobiologia, 103: 185–191. Coates, D. R. 1991. Glacial deposits. In: G. A. Kiersch (ed.) The Heritage of Engineering Geology; The First Hundred Years. Geological Society of America, centennial Special Volume 3, 299–322. Cobbold, P. R. and Quinquis, H. 1980. Development of sheath folds in shear regimes. Journal of Structural Geology, 2: 119–126. COHMAP Members 1988. Climatic changes of the last 18,000 years; observations and model simulations. Science, 241: 1043–1052. Colbeck, S. C. and Evans, R. J. 1973. A flow law for temperate glacier ice. Journal of Glaciology, 12: 71–86.

Collins, D.N. 1978. Hydrology of an alpine glacier as indicated by the chemical composition of meltwater. Zeitschrift f¨ur Gletscherkunde und Glazialgeologie, 13: 219–238. Collins, D. N. 1982. Water storage in an alpine glacier. In: J. W. Glen (ed.) Hydrological Aspects of Alpine and High-Mountain Areas. Proceedings of the Exeter Symposium, July 1982. International Association of Hydrological Sciences, Publication 138, pp. 113–122. Collinson, J. D. and Thompson, D. B. 1982. Sedimentary Structures. George Allen & Unwin Publishing, Boston, 194 pp. Collomb, E. 1847. Preuves de l’existence d’anciens glaciers dans les vall´es des Vosges. Victor Masson, Paris, 246 pp. Cowan, D. S. 1982. Deformation of partly dewatered and consolidated Franciscan sediments near Piedras Blancas Point, California. In: J. K. Leggett (ed.) Trench Forearc Geology. Geological Society of London Special Publication, No.10, pp. 439–457. Cowan, D. S. 1985. Structural styles in Mesozoic and Cenozoic m´elanges in the western Cordillera of North America. Geological Society of America Bulletin, 96: 451–462. Cowan, E. A. and Powell, R. D. 1990. Suspended sediment transport and deposition of cyclically interlaminated sediment in a temperate glacial fjord, Alaska, U.S.A. In: J. A. Dowdeswell and J. D. Scourse (eds) Glacimarine Environments: Processes and Sediments. Geological Society of London, Special Publication No. 53, pp. 75–89. Cowan, E. A. and Powell, R. D. 1991. Ice-proximal sediment accumulation rates in a temperate glacial fjord, southeastern Alaska. In: J. B. Anderson and G. M. Ashley (eds) Glacial Marine Sedimentation: Paleoclimatic Significance. Geological Society of America, Special Paper, 261, pp. 61–73. Cowan, E. A., Powell, R. D. and Smith, N. D. 1988. Rainstorm-induced event sedimentation at the tidewater front of a temperate glacier. Geology, 16: 409–412. Cowan, W. R., Sharpe, D. R., Feenstra, B. H. and Gwyn, Q. H. J. 1978. Glacial geology of the Toronto-Owen Sound area. In: A. L. Currie and W. O. Mackasey (eds) Toronto ’78 Field Trips Guidebook. Geological Association of Canada, pp. 1–16. Cowie, J. W., Zeigler, W., Boucot, A. J., Bassett, M. and Remane, E. 1986. Guidelines and statutes of the International Commission on Stratigraphy (ICS). Courier Forschungsinstitut Senckenberg, 83: 1–14. Croll, J. 1864. On the physical cause of the change of climate during geological epochs. Philosophical Magazine, 28: 121–137.

REFERENCES

Croot, D. G. 1988a. Glaciotectonics, Forms and Processes. A. A. Balkema, Rotterdam, 212 pp. Croot, D. G. 1988b. Morphological, structural and mechanical analysis of neoglacial ice-pushed ridges in Iceland. In: D. G. Croot (ed.) Glaciotectonics, Forms and Processes. A. A. Balkema, Rotterdam, pp. 33–47. Crosby, I. B. 1945. Glacial erosion and the buried Wyoming Valley of Pennsylvania. Geological Society of America, Bulletin, 56: 389–400. Crowell, J. C. 1981. Early Paleozoic glaciation and Gondwana drift. In: M. W. McElhinny and D. A. Valencio (eds) Paleoreconstruction of the Continents. American Geophysical Union and Geological Society of America, Geodynamics Series, 2, pp. 45–49. Crowell, J. C. 1982. Continental glaciation through geologic times. In: National Research Council, Climate in Earth History. National Academy Press, Washington, pp. 77–82. Crowell, J. C. 1983. The recognition of ancient glaciations. In: L. G. Medaris, Jr., C. W. Byers, D. M. Mickelson and W. C. Shanks (eds) Proterozoic Geology; Selected Papers from an International Proterozoic Symposium. Geological Society of America, Memoir 161, pp. 289–297. Crowley, T. J., Mengel, J. G. and Short, D. A. 1987. Gondwanaland’s seasonal cycle. Nature, 329: 803–807. Cunningham, F. 1990. James Forbes, Pioneer Glaciologist. Scottish Academic Press, Edinburgh. pp.329 Curry, B. B., Troost, K. C. and Berg, R. C. 1994. Quaternary Geology of the Martinsville Alternative Site, Clark County, Illinois. Illinois State Geological Survey, Circular 556, 85 pp. Dahlen, F. A., Suppe, J. and Davis, D. 1984. Mechanics of fold-and-thrust belts and accretionary wedges: cohesive Coulomb theory. Journal of Geophysical Research, 89(B12): 10087–10101. Dahl-Jensen, D. 1985. Determination of the flow properties at Dye 3, south Greenland, by bore-hole-tilting measurements and perturbation modelling. Journal of Glaciology, 31: 92–98. Dahlstrom, C. D. A. 1970. Structural geology in the eastern margin of the Canadian Rocky Mountains. Bulletin of Canadian Petroleum Geology, 18: 332–406. Dale, J. E., Aitken, A. E., Gilbert, R. and Risk, M. J. 1989. Macrofauna of Canadian Arctic fjords. Marine Geology, 85: 331–358. Dansgaard, W. 1987. Ice core evidence of abrupt climatic changes. In W. H. Berger and L. D. Labeyrie (eds), Abrupt Climatic Change; Evidence and Implications. D. Reidel, Dordrecht, pp. 223–233. Dansgaard, W. and Oeschger, H. 1989. Past environmental

493

long-term records from the Arctic. In: H. Oeschger and C.C. Langway, Jr. (Editors), The Environmental Record in Glaciers and Ice Sheets. J. Wiley, New York, pp. 287–318. Dansgaard, W., Clausen, H. B., Gundestrup, N., Hammer, C. U., Johnsen, S. F., Kristinsd´ottir, P. M. and Reeh, N. 1982. A new Greenland deep ice core. Science, 218: 1273–1277. Dansgaard, W., Johnsen, S. J., Clausen, A. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer, C. U., Avidberg, C. S., Steffensen, J. P., Sveinbj¨ornsd´ottir, A. E., Jouzel, J. and Bond, G. 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature, 364: 218–220. Dardis, G. F. 1987. Sedimentology of late-Pleistocene drumlins in south-central Ulster, Northern Ireland. In J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 215–224. Dardis, G. F. and McCabe, A. M. 1987. Subglacial sheetwash and debris flow deposits in Late-Pleistocene drumlins, Northern Ireland. In: J. Menzies and J. Rose (Editors), Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 225–240. Dardis, G. F., McCabe, A. M. and Mitchell, W. I. 1984. Characteristics and origins of lee-side stratification sequences in late-Pleistocene drumlins, Northern Ireland. Earth Surface Processes and Landforms, 9: 409–424. Davis, D., Suppe, J. and Dahlen, F. A. 1983. Mechanics of fold-and-thrust belts and accretionary wedges. Journal of Geophysical Research, 88: 1153–1172. de Charpentier, J. 1841. Essai sur les glaciers et sur le terrain erratique du bassin du Rhˆone. Marc Ducloix, Lausanne, 363 pp. de Gans, W., de Groot, T. and Zwaan, H. 1987. The Amsterdam basin, a case study of a glacial basin in The Netherlands. In: J. J. M. van der Meer (ed.) Tills and Glaciotectonics. A. A. Balkema, Rotterdam, pp. 205–216. ¨ De Geer, G. 1889. Andmor¨ aner i trakten mellan Spånga och Sundbyberg. Geologiska F¨oreningen i Stockholm F¨orhandlingar, 11: 395–397. De Geer, G. 1912. A geochronology of the last 12,000 years. 11th International Geological Congress, Stockholm, 1910. Compte Rendu, 1: 241–258. De Jong, C. 1992. Thresholds for channel change on two contrasting proglacial river fans, West Greenland. Geografiska Annaler, 74A: 1–12. De Mulder, F.J. and Hageman, B.P. (Editors) 1989. Applied Quaternary Research. A.A. Balkema Publishing, Rotterdam, 185 pp.

494

REFERENCES

Denton, G. H. and Hughes, T. J. 1981. The Last Great Ice Sheets. Wiley, New York, 484 pp. Denton, G. H., Bockheim, J. G., Wilson, S. C. and Stuiver, M. 1989. Late Wisconsin and early Holocene glacial history, inner Ross Embayment, Antarctica. Quaternary Research, 31: 151–182. Desinov, L. V. 1984. The surge of the Muzgazy Glacier. Mat. Glyatsiol. Issled. Khron. Obsuzhdeniya, 50: 145–146. Desnoyers, J. 1829. Observations sur un ensemble de d´epˆots marins plus r´ecens que les terrains tertiaires du bassin de la Seine, et constituant une formation g´eologique distincte; pr´eced´ees d’un aperçu de la non-simultan´eit´e des bassins tertiaires. Annales des Sciences Naturelles, 16: 171–214, 402–491. Deynoux, M,. Miller, J. M. G., Domack, E. W., Eyles, N., Fairchild, I. J. and Young, G. M. (eds) 1994. Earth’s Glacial Record. Cambridge University Press, Cambridge. 266 pp. Dickson, J. H., Stewart, D. A., Thompson, R., Turner, G., Baxter, M. S., Drndarsky, N. D, and Rose, J. 1978. Palynology, palaeomagnetism and radiometric dating of Flandrian marine and freshwater sediments of Loch Lomond, Nature, 274: 548–553. Dimakis, P., Elverhøi, A., Høeg, K., Solheim, A., Harbitz, C. B., Laberg, J. S., Vorren, T. O. and Marr, J. 2000. Submarine slope stability on high-latitude glaciated Svalbard–Barents Sea margin. Marine Geology, 162: 303–316. Dionne, J.-C. and Brodeur, D. 1988. Erosion des platesformes rocheuses littorales par affouillement glaciel. Zeitschrift fur Geomorphologie, 32: 101–115. Doake, C. S. M. 1987. Some aspects of the flow of the Ronne Ice Shelf. In: C. J. Van der Veen and J. Oerlemans (eds), Dynamics of the West Antarctic Ice Sheet. D. Reidel Publishing, Dordrecht, pp. 75–98. Dolgushin, L. D. and Osipova, G. B. 1975. Glacier surges and the problem of their forecasting. International Associations of Hydrological Sciences, Publication No.104: 292–304. Domack, E. W. 1988. Biogenic facies in the Antarctic glacimarine environment: basis for a polar glacimarine summary. Palaeogeography, Palaeoclimatology, Palaeoecology, 63: 357–372. Domack, E. W. 1990. Laminated terrigenous sediments from the Antarctic Peninsula: the role of subglacial and marine processes. In: J. A. Dowdeswell and J. D. Scourse (eds) Glaciomarine Environments: Processes and Sediments. Geological Society of London Special Publication. No. 53, pp. 91–103. Domack, E. W. and Lawson, D. E. 1985. Pebble fabric in

an ice-rafted diamicton. Journal of Geology, 93: 577–591. Domack, E. W. and Williams, C. R. 1990. Fine structure and suspended sediment transport in three Antarctic fjords. American Geophysical Union, Antarctic Research Series, 50: 71–89. Domack, E. W., Jull, A. J. T. and Nakao, S. 1991. Advance of East-Antarctic outlet glaciers during the Hypsithermal: Implications for the volume state of the Antarctic Ice Sheet under global warming. Geology, 19: 1059–1062. Domack, E. W., Foss, D. J. P., Syvitski, J. P. M. and McClennen, C. E. 1994. Transport of suspended particulate matter in an Antarctic fjord. Marine Geology, 121: 161–170. Dominik, J., Burrus, D. and Vernet, J.-P. 1983. A preliminary investigation of the Rhˆone River plume in eastern Lake Geneva. Journal of Sedimentary Petrology, 53: 159–163. Donnelly, R. and Harris, C. 1989. Sedimentology and origin of deposits from a small ice-dammed lake, Leirbreen, Norway. Sedimentology, 36: 581–600. Donner, J. J. 1957. The geology and vegetation of Lateglacial retreat in Scotland. Transaction of the Royal Society of Edinburgh, 63: 221–264. Dowdeswell, J. A. 1986. The distribution and character of sediments in a tidewater glacier, southern Baffin Island, NWT, Canada. Arctic and Alpine Research, 18: 45–56. Dowdeswell, J. A. and Scourse, J. D. (eds) 1990. Glaciomarine Environments: Processes and Sediments. Geological Society Special Publication No. 53, 423 pp. Dowdeswell, J. A. and Sharp, M. 1986. Characterization of pebble fabrics in modern terrestrial glacigenic sediments. Sedimentology, 33: 699–710. Dowdeswell, J. A. and Siegert, M. J. 1999. The dimensions and topographic setting of Antarctic subglacial lakes and implications for large-scale water storage beneath continental ice sheets. Geological Society of America, Bulletin, 111: 254–264. Dowdeswell, J. A., Hamilton, G. S. and Hagen, J. O. 1991. The duration of the active phase on surge-type glaciers: contrasts between Svalbard and other regions. Journals of Glaciology, 37: 388–400. Drake, L. D. 1972. Mechanisms of clast attrition in basal till. Geological Society of America, Bulletin, 83: 2159–2166. Drake, L. D. and Shreve, R. L. 1973. Pressure melting and regelation of ice by round wires. Proceedings of the Royal Society of London, Series A, 332: 51–83. Dredge, L. A. and Cowan, W. R. 1989. Quaternary geology of the southwestern Canadian Shield. In: R. G. Fulton (ed.) Quaternary Geology of Canada and Greenland.

REFERENCES

Geological Survey of Canada, Geology of Canada, No.1, pp. 214–235. Dreimanis, A. 1984. Comments on ‘Sedimentation in a large lake: an interpretation of the Late Pleistocene stratigraphy at Scarborough Bluffs, Ontario, Canada’. Geology, 12: 185–186. Dreimanis, A. 1988. Tills: their genetic terminology and classification. In: R. P. Goldthwait and C. L. Matsch (eds) Genetic Classification of Glacigenic Deposits. A. A. Balkema, Rotterdam, pp. 17–83. Dreimanis, A. 1991. The Laurentide Ice Sheet during the last glaciation: A review and some current reinterpretations along its southern margin. In: B. Frenzel (ed.) Klimageschichtliche Probleme der Letzten 130,000 Jahre. Pal¨aklima forschung, 1: 267–291. Dreimanis, A. 1993. Small to medium-sized glacitectonic structures in till and in its substratum and their comparison with mass movement structures. Quaternary International, 18: 69–80. Dreimanis, A. and Lundqvist, J. 1984. What should be called till? Striae, 20: 5–10. Dreimanis, A. and Rappol, M. 1997. Late Wisconsinan subglacial clastic intrusive sheets along Lake Erie bluffs, at Bradtville, Ontario, Canada. Sedimentary Geology, 111: 1–4. Dreimanis, A. and Vagners, U. J. 1971. Bimodal distribution of rock and mineral fragments in basal tills. In: R. P. Goldthwait (ed.) Till; A Symposium. Ohio State University Press, Columbus OH, pp. 237–250. Drewry, D. J. 1983. Antarctic ice sheet: aspects of current configuration and flow. In: R. Gardner and H. Scoging (eds) Mega-Geomorphology. Clarendon Press, Oxford, pp. 18–38. Drewry, D. J. 1986. Glacial Geologic Processes. Edward Arnold, London, 276 pp. Drewry, D. J. and Cooper, A. P. R. 1981. Processes and models of Antarctic glaciomarine sedimentation. Annals of Glaciology, 2: 117–122. Drewry, D. J., Jordan, S. R. and Jankowski, E. 1982. Measured properties of the Antarctic ice sheet: surface configuration, ice thickness, volume and bedrock characteristics. Annals of Glaciology, 3: 83–91. Dunbar, R. B. 1988. Holocene sedimentation along the coast of southern Victoria Land. Antarctic Journal of the United States, Annual Review Issue, 23: 110–112. Dunbar R. B., Leventer, A. R. and Stochton, W. L. 1989. Biogenic sedimentation in McMurdo Sound, Antarctica. Marine Geology, 85: 155–180. Dyke, A. S. and Morris, T. F. 1988. Drumlin fields, dispersal trains, and ice streams in Arctic Canada. Canadian Geographer, 32: 86–90.

495

Dyson, J. L. 1952. Ice-ridged moraines and their relation to glaciers. American Journal of Scienc, 250: 204–211. Dyurgerov, M. B. 1986. Computations of mass balance in glacier systems. (In Russian). Materialy Glyatsiologicheskikh Issledovaniy, 57: 8–15. Echelmeyer, K. A. and Harrison, W. D. 1989. (Abstr.) Dynamics of Jakobshavns ice stream, west Greenland. Annals of Glaciology, 12: 201. Echelmeyer, K. A. and Harrison, W. D. 1990. Jakobshavns Isbræ, West Greenland: seasonal variations in velocity – or lack thereof. Journal of Glaciology, 36: 82–88. Echelmeyer, K. A. and Wang Zhongxiang 1987. Direct observation of basal sliding and deformation of basal drift at sub-freezing temperatures. Journal of Glaciology, 33: 83–98. Echelmeyer, K. A., Butterfield, R. and Cuillard, D. 1987. Some observations on a recent surge of Peters Glacier, Alaska, USA. Journal of Glaciology, 33: 341–348. Edwards, M. 1986. Glacial environments. In: H. G. Reading (ed.) Sedimentary Environments and Facies. 2nd edition. Blackwell Scientific Publishers, Oxford, pp. 445–470. Ehlers, J. (ed.) 1983. Glacial Deposits in North-West Europe. A. A. Balkema, Rotterdam, 470 pp. Ehlers, J. 1996. Quaternary and Glacial Geology. Translated from Allgemeine und historische Quart¨argeologie; English version by P. L. Gibbard, J. Wiley & Sons, Chichester. p.578. Ehlers, J., Gibbard, P. L. and Rose, J. (eds) 1991. Glacial Deposits in Great Britain and Ireland. A. A. Balkema, Rotterdam, 580 pp. Ehlers, J., Kozarski, S. and Gibbard, P. L. 1995. Glacial Deposits in North-east Europe. A. A. Balkema, Rotterdam, 625 pp. Ehrlich, R. and Davies, D. 1968. Sedimentological indices of transport direction, distance, and process intensity in glacio-fluvial sediments. Journal of Sedimentary Petrology, 38: 1166–1170. Eidvin, T. and Riis, F. 1989. Nye dateringer av de tre vestligste borehullene i Barentshavet. Resultater og konsekvenser for den tertiære hevningen. Oljedirektoratet Stavanger, Norway, NPD-contribution 27, 43 pp. Eissmann, L. 1987. Lagerungsst¨orungen im Lockergebirge – Exogene und endogene Tektonik im Lockergebirge des n¨ordlichen Mitteleuropa. Geophysik und Geologie, Serie 3, Band 3: 7–77. Elfstr¨om, A. and Rossbacher, L. 1985. Erosional remnants in the Båldakatj area, Lapland, northern Sweden. Geografiska Annaler, 67A, 167–176. Elliott, D. 1976. The motion of thrust sheets. Journal of Geophysical Research, 81: 949–963.

496

REFERENCES

Elverhøi, A. 1984. Glacigenic and associated marine sediments in the Weddell Sea, fjords of Spitsbergen and the Barents Sea: a review. Marine Geology, 57: 53–88. Elverhøi, A. 1998. Glacial and oceanic history of the Polar North Atlantic margins. Quaternary Science Reviews, 17: 1–302. Elverhøi, A., Pfirman, S., Solheim, A. and Larsen, B. B. 1989. Glaciomarine sedimentation on epicontinental seas – exemplified by the northern Barents Sea. Marine Geology, 85: 225–250. Elverhøi, A., Nyland-Berg, M., Russwurm, L. and Solheim, A. 1990. Late Weichselian ice recession in the central Barents Sea. In U. Bleil and J. Thiede (eds) Geological History of the Polar Oceans; Arctic versus Antarctic. NATO ASI Series, C 308, Kluwer Academic Publishers, Dordrecht, pp. 289–307. Elverhøi, A., Harbitz, C. B., Dimakis, P., Marr, J., Mohrig, D. and Parker, G. In press On the dynamics of subaqueous debris flows. Oceanography, Embleton, B. J. J. and Williams, G. E. 1986. Low Palaeolatitude of deposition for late Precambrian periglacial varvites in South Australia: implications for palaeoclimatology. Earth and Planetary Science Letters, 79: 419–430. Embleton, C. and King, C. A. M. 1975. Glacial Geomorphology. Edward Arnold, 573 pp. Embleton, C. and Thornes, J. B. (eds) 1979. Processes in Geomorphology. Edward Arnold, 422 pp. Emiliani, C. 1955. Pleistocene temperatures. Journal of Geology, 63: 538–578. Engelhardt, H. F., Harrison, W. D. and Kamb, B. 1978. Basal sliding and conditions at the glacier bed as revealed by bore-hole photography. Journal of Glaciology, 20: 469–508. Evans, D. J. A. 1989. Apron entrainment at the margins of subpolar glaciers, north-west Ellesmere Island, Canadian High Arctic. Journal of Glaciology, 35: 317–324. Evans, D. J. A., Archer, S. and Wilson, D. J. H. 1999. A comparison of the lichenometric and Schmidt Hammer dating techniques based on data from the proglacial area of some Icelandic glaciers. Quaternary Science Reviews, 18: 13–41. Evans, I. S. 1987. A new approach to drumlin morphometry. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 119–130. Evenson, E. B. and Clinch, J. M. 1987. Debris transport mechanisms at active alpine glacier margins: Alaskan case studies. In: R. Kujansuu and M. Saarnisto (eds) INQUA Till Symposium, Finland 1985. Geological Survey of Finland Special Paper, 3, pp. 111–136.

Evenson, E. B., Dreimanis, A. and Newsome, J. W. 1977. Subaquatic flow tills: a new interpretation for genesis of some laminated till deposits. Boreas, 6: 115–133. Eybergen, F. A. 1987. Glacier snout dynamics and contemporary push moraine formation at the Turtmannglacier, Wallis, Switzerland. In: J. J. M. van der Meer (ed.) Tills and Glaciotectonics. A. A. Balkema, Rotterdam, pp. 217–231. Eyles, C. H. and Eyles, N. 1989. The upper Cenozoic White River ‘tillites’ of southern Alaska: Subaerial slope and fan-delta deposits in a strike-slip setting. Geological Society of America, Bulletin, 101: 1091–1102. Eyles, C. H., Eyles, N. and Miall, A. D. 1985. Models of glaciomarine sedimentation and their application to the interpretation of ancient glacial sequences. In: M. Deynoux (ed.) Glacial Record, Proceedings of the Till Mauretania ’83 Symposium. Palaeogeography, Palaeoclimatology, Palaeoecology, 51: 15–84. Eyles, N. 1979. Facies of supraglacial sedimentation on Icelandic and Alpine temperate glaciers. Canadian Journal of Earth Sciences, 16: 1341–1361. Eyles, N. 1983a. Glacial Geology: An Introduction for Engineers and Earth Scientists. Pergamon Press, Oxford, 409 pp. Eyles, N. 1983b. Modern Icelandic glaciers as depositional models for ‘hummocky moraine’ in the Scottish Highlands. In: E. B. Evenson, Ch. Schl¨uchter, and J. Rabassa (eds) Tills and Related Deposits. A. A. Balkema Publishers, Rotterdam, pp. 47–59. Eyles, N. 1993. Earth’s glacial record and its tectonic setting. Earth Science Reviews, 35 (1/2): 1–248. Eyles, N. and Menzies, J. 1983. The subglacial landsystem. In N. Eyles (ed.) Glacial Geology. Pergamon Press, Oxford, pp. 19–67. Eyles, N. and Miall, A. D. 1984. Glacial facies. In: R. G. Walker (ed.) Facies Models. Geoscience Canada Reprint Series 1, Second Edition, pp. 15–38. Eyles, N. and Rogerson, R. J. 1977. Glacier movement, ice structures, and medial moraine form at a glacier confluence, Berendon Glacier, British Columbia, Canada. Canadian Journal of Earth Sciences, 14: 2807–2816. Eyles, N., Sladen, J. A. and Gilroy, S. 1982. A depositional model for stratigraphic complexes and facies superimposition in lodgement tills. Boreas, 11: 317–333. Eyles, N., Eyles, C. H. and Miall, A. D. 1983. Lithofacies types and vertical profile models; an alternative approach to the description and environmental interpretation of glacial diamict and diamictite sequences. Sedimentology, 30: 393–410.

REFERENCES

Eyles, N., Clark, B. M. and Clague, J. J. 1987. Coarsegrained sediment gravity flow facies in a large supraglacial lake. Sedimentology, 34: 193–216. Eyles, N., Eyles, C. H. and McCabe, A. M. 1989. Sedimentation in an ice-contact subaqueous setting: The Mid-Pleistocene ‘North Sea Drifts’ of Norfolk, U.K. Quaternary Science Reviews, 8: 57–74. Fahnestock, R. K. 1963. Morphology and hydrology of a glacial stream; White River, Mount Rainier, Washington. U.S. Geological Survey Professional Paper, P 0422_A, pp. A1–A70. Fahnestock, M. and Humphrey, N. F. 1988. Borehole water level measurements, Columbia Glacier, AK. Ice, 86: 25–26. Fairbanks, R. G. 1989. A 17,000-year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature, 342: 637–642. Fastook, J. L. 1987. The finite-element method applied to a time-dependent flowband model. In: C. J. Van der Veen and J. Oerlemans (eds) Dynamics of the West Antarctic Ice Sheet. D. Reidel Publishing, Dordrecht, pp. 203–221. Fastook, J. L. and Hughes, T. J. 1990. The Changing Distribution of Crustal Ice Loads During the Last Global Glaciation Cycle. Kluwer Academic Publishers. The Netherlands. Fenn, C. R. and Gurnell, A. M. 1987. Proglacial channel processes. In: A. M. Gurnell and M. J. Clark (eds) Glacio-Fluvial Sediment Transfer – an Alpine Perspective. J.Wiley and Sons Publishing, Chichester, pp. 423–498. Ferguson, R. I. 1984. Sediment load of the Hunza River. In Miller, K. J. (ed). The International Karakoram Project. Cambridge University Press, Volume 2, pp. 581–598. Fernlund, J. M. R. 1994. Tills and non-till diamictons in glacial sequences. In: W. P. Warren and D. G. Croot (eds) Formation and Deformation of Glacial Deposits. A. A.Balkema Publishers, Rotterdam, pp. 29–38. Fischer, A. G. 1984. The two Phanerozoic suipercycles. In W. A. Berggren and J. A. Van Couvering (eds) Catastrophes in Earth History; The New Uniformitarianism. Princeton University Press, pp. 129–150. Fisher, D. A., Reeh, N. and Langley, K. 1985. Objective reconstructions of the Late Wisconsinan Laurentide Ice Sheet and the significance of deformable beds. G´eographie Physique et Quaternaire, 39: 229–238. Fisher, T. G. and Shaw, J. A. 1992. A depositional model for Rogen Moraine, with examples from the Avalon Peninsula, Newfoundland. Canadian Journal of Earth Scien-

497

ces, 29: 669–686. Flint, R. F. 1947. Glacial Geology and the Pleistocene Epoch. John Wiley & Sons, New York. 589 pp. Flint R. F. 1971. Glacial and Quaternary Geology. John Wiley & Sons, New York and London, 892 pp. Foldvik, A. and Gammelsrød, T. 1988. Notes on Southern Ocean hydrography, sea-ice and bottom water formation. Palaeogeography, Palaeoclimatology, Palaeoecology, 67: 3–17. Forbes, E. 1846. On the Connection Between the Distribution of the Existing Fauna and Flora of the British Isles, and the Geological Changes Which have Affected their Area, Especially During the Epoch of the Northern Drift. Great Britain Geological Survey, Memoir 1, pp. 336–432. Fowler, A. C. 1979. The use of a rational model in the mathematical analysis of a polythermal glacier. Journal of Glaciology, 24: 443–456. Freiwald, A., Henrich, R., Schaefer, P. and Willkomm, H. 1991. The significance of high-boreal to subarctic marl deposits in northern Norway to reconstruct Holocene climatic changes and sea level oscilliations. Facies, 25: 315–340. French, H. M. 1976. The Periglacial Environment. Longmans Publishing, London, 309 pp. French, H. M. and Gozdzik, J. S. 1988. Pleistocene epigenetic and syngenetic frost fissures, Be3chat´ow, Poland. Canadian Journal of Earth Sciences, 25: 2017–2027. Freysteinsson, S. 1972. Glacier burst in the K¨oldukvislarj¨okull area. J¨okull, 22: 83–88. Frye, J. C. and Willman, H. B. 1962. Morphostratigraphic units in Pleistocene stratigraphy. American Association of Petroleum Geologists, 46: 112–113. Frye, J. C. and Willman, H. B. 1973. Wisconsinan climatic history interpreted from Lake Michigan Lobe deposits and soils. In R. F. Black, R. P. Goldthwait and H. B. Willman (eds) The Wisconsian Stage. Geological Society of America, Memoir 136: 135–152. Fulton, R. J. (ed.) 1984. Quaternary Stratigraphy of Canada – A Canadian Contribution to IGCP Project 24. Geological Survey of Canada Paper, 84–10, 210 pp. Fulton, R. J. 1989. Quaternary Geology of Canada and Greenland. Geological Survey of Canada, Geology of Canada No.1, 839 pp. Fulton, R. J. and Pullen, M. J. L. T. 1969. Sedimentation in Upper Arrow Lake, British Columbia. Canadian Journal of Earth Sciences, 6: 785–791. Fulton, R. J., Fenton, M. M. and Rutter, N. W. 1986. Summary of Quaternary stratigraphy and history, Western Canada. In: V. Sibrava, D. Q. Bowen and G.

498

REFERENCES

M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 229–241. Funder, S. (co-ordinator) 1989. Chapter 13. Quaternary geology of the ice-free areas and adjacent shelves of Greenland. In: R. J. Fulton (ed.) Quaternary Geology of Canada and Greenland. Geological Survey of Canada, Geology of Canada, No.1 pp. 741–792. Furbish, D. J. and Andrews, J. T. 1984. The use of hypsometry to indicate long-term stability and response of valley glaciers to changes in mass transfer. Journal of Glaciology, 30: 199–211. Fyfe, G. 1990. The effect of water depth on ice-proximal glaciolacustrine sedimentation: Salpausselk¨a I, southern Finland. Boreas, 19: 147–164.

Gaffin, S. 1987. Ridge volume dependence on sea floor generation rate and inversion using long term sea-level change. American Journal of Science, 287: 596–611. Garcia Ambrosiani, K. and Robertsson, A. M. 1992. Early Weichselian interstadial sediments at H¨arn¨osand, Sweden. Boreas, 21: 305–317. Gard, G. 1988. Late Quaternary calcareous nannofossil biochronology and paleo-oceanography of Arctic and Subarctic seas. Meddelanden från Stockholms Universitets Geologiska Institution, 275: 8–45. Gard, G. and Backman, J. 1990. Synthesis of Arctic and Sub-Arctic coccolith biochronology and history of North Atlantic drift water influx during the last 500 000 years. In: U. Bleil and J. Thiede (eds) Geological History of the Polar Oceans: Arctic versus Antarctic. NATO ASI Series, C, Kluwer Academic Publishing, Dordrecht, pp. 417–436. Gardner, R. and Scoging, H. (eds) 1983. Mega-Geomorphology. Oxford University Press, Oxford. Geikie, A. 1863. On the phenomena of the glacial drift of Scotland. Transactions of the Geological Society of Glasgow, 1: 1–190. Geikie, J. 1884. The Great Ice Age. (3rd Edtion). Gerrard, J. A. F., Perutz, M. F. and Roch, A. 1952. Measurement of the velocity distribution along a vertical line through a glacier. Proceedings of the Royal Society of London, Series A, 213: 546–558. Gibbard, P. 1980. The origin of stratified Catfish Creek Till by basal melting. Boreas, 9: 71–85. Gibbard, P. L. and Dreimanis, A. 1978. Trace fossils from late Pleistocene glacial lake sediments in southwestern Ontario, Canada. Canadian Journal of Earth Sciences, 15: 1967–1976.

Gibbard, P. L. and Stuart, A. J. 1974. Trace fossils from proglacial lake sediments. Boreas, 3: 69–74. Gibbard, P. L. and Turner, C. 1988. In defence of the Wolstonian stage. Quaternary Newsletter, 54: 9–14. Gilbert, G. K. 1890. Lake Bonneville. United States Geological Survey Monograph, 1, 438 pp. Gilbert, G. K. 1903. Glaciers and Glaciation of Alaska. Harriman Alaska Series, Vol. 3. Doubleday, Page and Company, New York, 231 pp. Gilbert, G. K. 1906. Crescentic gouges on glaciated surfaces. Geological Society of America Bulletin, 17: 303–316. Gilbert R. 1975. Sedimentation in Lillooet Lake, British Columbia. Canadian Journal of Earth Sciences, 12: 1697–1711. Gilbert, R. 1990. Rafting in glaciomarine environments. In: J. A. Dowdeswell and J. D. Scourse (eds) Glaciomarine Environments: Processes and Sediments. Geological Society of London, Special Publication No. 53, pp. 105–120. Gilbert, R. and Desloges, J. R. 1987. Sediments of icedammed, self-draining Ape Lake, British Columbia. Canadian Journal of Earth Sciences, 24:1735–1747. Gilbert, R. and Shaw, J. 1981. Sedimentation in proglacial Sunwapta Lake, Alberta. Canadian Journal of Earth Sciences, 18: 81–93. Gilbert, R., Syvitski, J. P. M. and Taylor, R. B. 1985. Reconnaissance study of proglacial Stewart Lakes, Baffin Island, District of Franklin. In: Current research, part A. Geological Survey of Canada, Paper 85–1A: 505–510. Gilliland, R. L. 1989. Solar evolution. Palaeogeography, Palaeoclimatology, Palaeoecology (Global and Planetary Change Section), 75: 35–55. Gilpin, R. R. 1979. A model of the ‘liquid-like’ layer between ice and a substrate with applications to wire regelation and particle migration. Journal of Colloid and Interface Science, 68: 235–251. Gilpin, R. R. 1980. Wire regelation at low temperatures. Journal of Colloid and Interface Science, 77: 435–448. Giovinetti, M. B. and Bentley, C. R. 1985. Surface balance in ice drainage systems of Antarctica. Antarctic Journal of the United States, 20: 6–13. Gipp, M. R. 1993. The orientation of buried iceberg scours and other linear phenomena. Marine Geology, 114: 263–272. Gjessing, J. 1965. On ‘plastic scouring’ and ‘subglacial erosion’. Norsk Geografisk Tidsskrift, 20: 1–37. Glazyrin, G. Ye. and Sokolov, L. N. 1975. Calculations of the approximate number and area of glaciers in the nonstudied regions, Materialy Glyatsiologicheskikh Issledovaniy. Khronika Obsuzhdeniya, 26: 97–101.

REFERENCES

Glazyrin, G. Ye., Kamnyanskiy, G. M., Mzo, A. B., Nozdryukhin, V. K. and Salamatin, A. N. 1987. Mechanism of the Abramova Glacier surge in 1972–1975. Mat. Glyatsiol. Issled. Khron. Obsuzhdeniya, 60: 84–90. Glen, J. W. 1955. The creep of polycrystalline ice. Proceedings of the Royal Society of London, A228: 519–538. Gl¨uckert, G. 1973. Two large drumlin fields in central Finland. Fennia. 120: 37. Goldstein, B. 1989. Lithology, sedimentology, and genesis of the Wadena drumlin field, Minnesota, U.S.A. Sedimentary Geology, 62: 241–280. Goldthwait, R. P. 1971. Till: a Symposium. Ohio State University Press, Columbus OH, 402 pp. Goldthwait, R. P. 1979. Giant grooves made by concentrated basal ice streams. Journal of Glaciology, 23: 297–307. Goldthwait, R. P. 1988. Classification of glacial morphologic features. In: R. P. Goldthwait and C. L. Matsch (eds) Genetic classification of glacigenic deposits, pp. 267–277. Goldthwait, R. P. and Matsch, C. L. 1988. Genetic Classification of Glacigenic Deposits. A. A. Balkema, Rotterdam. 294 pp. Gomez, B. 1987. Bedload. In: A. M. Gurnell and M. J. Clark (eds) Glaciofluvial Sediment Transfer: An Alpine Perspective. John Wiley and Sons, Chichester, pp. 355–376. Gomez, B. and Small, R. J. 1986. Medial moraines of the Haut Glacier d’Arolla, Valais, Switzerland: debris supply and implications for moraine formation. Journal of Glaciology, 31: 303–307. Goodchild, J. G. 1875. The glacial phenomena of the Eden Valley and the western part of the Yorkshire Dale district. Quarterly Journal of the Geological Society of London, 31: 55–99. Goodman, D. J., King, G. C. P., Millar, D. H. M. and Robin, G. DeQ. 1979. Pressure-melting effects in basal ice of temperate glaciers: laboratory studies and field observations under Glacier d’Argenti‘ere. Journal of Glaciology, 23: 259–271. Goodwin, I. D. 1988. The nature and origin of a j¨okulhlaup near Casey Station, Antarctica. Journal of Glaciology, 34: 95–101. Gorlich, K. 1986. Glacimarine sedimentation of muds in Hornsund Fjord, Spitsbergen, Rocznik Polskiego Towarzystwa Geologicznego. Annales de la Societe Geologique de Pologne, 56: 433–477. Gow, A. J. 1963. Results of measurements in the 309 meter borehole at Byrd Station, Antarctica. Journal of Glaciology, 4: 771–784.

499

Gow, A. J., Ueda, H. T. and Garfield, D. E. 1968. Antarctic ice sheet: preliminary results of first core hole to bedrock. Science, 161: 1011–1013. Gow, A. J., Epstein, S. and Sheehy, W. 1979. On the origin of stratified debris in ice cores from the bottom of the Antarctic ice sheet. Journal of Glaciology, 23: 185–192. Grant, U. S. and Higgins, D. F. 1913. Coastal Glaciers of Prince William Sound and Kenai Peninsula. U.S. Geological Survey Bulletin, 526, 75 pp. Gravenor, G. P., von Brunn, V. and Dreimanis, A. 1984. Nature and classification of waterlain glaciogenic sediments, exemplified by Pleistocene, Late Paleozoic and Late Precambrian deposits. Earth Science Reviews, 20: 105–166. Gray, J. M. 1981. P-forms from the isle of Mull. Scottish Journal of Geology, 17: 39–47. Gray, J. M. 1991. Glaciofluvial landforms. In: J. Ehlers, P. L. Gibbard, and J. Rose (eds) Glacial Deposits of Britain and Ireland. A. A. Balkema Publishers, Rotterdam, pp. 443–453. Gray, J. M. and Lowe, J. J. (eds) 1977. Studies in the Scottish Late Glacial Environment. Pergamon Press, 197 pp. Greenland Ice-Core Project (GRIP) Members 1993. Climate instability during the last interglacial period recorded in the GRIP ice core. Nature, 364: 203–207. Greve, R. and MacAyeal, D. R. 1996. Dynamic/thermodynamic simulations of the Laurentide ice-sheet instability. Annals of Glaciology, 23: 328–335. Griffith, A. A. 1924. The theory of rupture. Proceedings of the 1st International Congress of Applied Mechanics (Delft), pp. 55–63. Gripp, K. 1929. Glaziologische und geologische Ergebnisse der Hamburgischen Spitsbergen Expedition 1927. Abhandlungen der Naturwissenschaftlichen Verein Hamburg, 22: 147–249. Gripp, K. 1975. Eisrandstudien ausgehend von Sermeq, SW-Gr¨onland. Meddelser om Gr¨onland, 195: 1–19. Grootes, P. M., Steig, E. J., Stuiver, M., Waddington, E. D. and Morse, D. in press. New 130,000-year long Taylor Dome, Antarctic 18O record indicates synchroniety of global glacial climatic fluctuations. (Submitted). Grove, J. M. 1988. The Little Ice Age. Methuen, London, 498 pp. Gurnell, A. M. 1987. Fluvial sediment yield from Alpine, glacierized catchments. In: A. M. Gurnell and M. J. Clark (eds) Glaciofluvial Sediment Transfer: An Alpine Perspective. John Wiley and Sons, Chichester, pp. 415–420. Gustavson, T. C. 1975a. Bathymetry and sediment distribution in proglacial Malaspina Lake, Alaska. Journal of Sedimentary Petrology, 45: 450–461.

500

REFERENCES

Gustavson, T. C. 1975b. Sedimentation and physical limnology in proglacial Malaspina Lake, southeastern Alaska. In: A. V. Jopling and B. C. McDonald (eds) Glaciofluvial and Glaciolacustrine Sedimentation. Society of Economic Paleontologists and Mineralogists Special Publication No. 23, pp. 249–263. Gustavson, T. C. and Boothroyd, J. C. 1987. A depositional model for outwash, sediment sources, and hydrologic characteristics, Malaspina Glacier, Alaska: A modern analog of the southeastern margin of the Laurentide Ice Sheet. Geological Society of America, Bulletin, 99: 187–200. Gustavson, T. C., Ashley, G. M. and Boothroyd, J. C. 1975. Depositional sequences in glaciolacustrine deltas. In: A. V. Jopling, and B. C. McDonald (eds) Glaciofluvial and Glaciolacustrine Sedimentation. Society of Economic Paleontologists and Mineralogists, Special Publication No. 23, Tulsa OK, pp. 264–280.

Haake, F-W. and Pflaumann, U. 1989. Late Pleistocene foraminiferal stratigraphy on the Vøring Plateau, Norwegian Sea. Boreas, 18: 343–356. Hagen, J. O., Wold, B., Liestol, O., Østrem, G. and Sollid, J.L. 1983. Subglacial processes at Bondhusbreen, Norway: preliminary results. Annals of Glaciology, 4: 91–98. Hald, M., Sættem, J. and Nesse, E. 1990. Middle and Late Weichselian stratigraphy in shallow drillings from the southwestern Barents Sea: foraminiferal, amino acid and radiocarbon evidence. Norsk Geologisk Tidsskrift, 70: 241–257. Haldorsen, S. 1981. Grain-size distribution of subglacial till and its relation to glacial crushing and abrasion. Boreas, 10: 91–105. Haldorsen, S. and Shaw, J. 1982. The problem of recognizing melt-out till. Boreas, 11: 261–277. Hallet, B. 1976. Deposits formed by subglacial precipitation of CaCO3. Geological Society of America, Bulletin, 87: 1003–1015. Hallet, B. 1979. A theoretical model of glacial abrasion. Journal of Glaciology, 23: 39–50. Hallet, B. 1981. Glacial abrasion and sliding: their dependence on the debris concentration in basal ice. Annals of Glaciology, 2: 23–28. Hallet, B. 1996 Glacial quarrying; a simple theoretical model. Annals of Glaciology, 22: 1–8. Hamblin, R., Moorlock, B. and Rose, J. 2000. A new glacial stratigraphy for Eastern England. Quaternary Newsletter, 92: 35–43. Hambrey, M. J. and Alean, J. 1992. Glaciers. Cambridge University Press, Cambridge, 208 pp.

Hambrey, M. J. and Harland, W. B. 1981. Earth’s PrePleistocene Glacial Record. Cambridge University Press, Cambridge, 1004 pp. Hambrey, M. J., Barrett, P. J., Ehrmann, W. U. and Larsen, B. 1992. Cenozoic sedimentary processes on the Antarctic continental margin and the record from deep drilling. Zeitschrift fur Geomorphologie, N.F., 86: 77–103. Hamilton, T. D. 1994. Late Cenozoic glaciation of Alaska. In: G. Plafker and H. C. Berg (eds) Geology of Alaska. Decade of North American Geology Project, Geological Society of America, Volume G-1, Chapter 27, 813–844. Hammer, C. U. 1989. Dating by physical and chemical seasonal variations and reference horizons. In: H. Oeschger and C. C. Langway, Jr. (eds) The Environmental Record in Glaciers and Ice Sheets. John Wiley and Sons, New York, pp. 99–121. Hammer, K. M. and Smith, N. D. 1983. Sediment production and transport in a proglacial stream: Hilda Glacier, Alberta, Canada. Boreas, 12: 91–106. Hance, J. H. 1937. The recent advance of the Black Rapids Glacier, Alaska. Journal of Geology, 45: 775–783. Hanebuth, T., Stattegger, K. and Grootes, P. M. 2000. Rapid flooding of the Sunda Shelf: A Late-GlaciaL sea-level record. Science, 288: 1033–1035. Hansel, A. K. and Johnson, W. H. 1987. Ice marginal sedimentation in a late Wisconsinan end moraine complex, Northeastern Illinois, USA. In J. J. M. van der Meer (ed.) Tills and Glaciotectonics. A. A. Balkema Publishers, Rotterdam, pp. 97–104 Hanvey, P. M. 1987. Sedimentology of lee-side stratification sequences in late-Pleistocene drumlins, north-west Ireland. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 241–253. Harbor, J. M. 1992. Numerical modeling of the development of U-shaped valleys by glacial erosion. Geological Society of America, Bulletin, 104: 1364–1375. Harbor, J. M. and Wheeler, D. A. 1992. On the mathematical description of glaciated valley cross sections. Earth Surface Processes and Landforms, 17: 477–485. Harbor, J. M., Hallet, B. and Raymond, C. F. 1988. A numerical model of landform development by glacial erosion. Nature, 333: 347–349. Harland, W. B. 1965. Critical evidence for a great InfraCambrian glaciation. Geologische Rundschau, 54: 45–61. Harland, W. B. 1992. Stratigraphic regulation and guidance: A critique of current tendancies in stratigraphic codes and guides. Geological Society of America Bulletin, 104: 1231–1235.

REFERENCES

Harris, P. W. V. 1976. The seasonal temperature-salinity structure of a glacial lake: J¨okuls´arl´on, south-east Iceland. Geografiska Annaler, 58A: 329–336. Harrison, A. E. 1964. Ice surges on the Muldrow Glacier, Alaska. Journal of Glaciology, 5: 365–368. Hart, J. K. 1987. The genesis of the North East Norfolk drift. Unpublished Ph.D. thesis, University of East Anglia, UK. Hart, J. K. 1990. Proglacial glaciotectonic deformation and the origin of the Cromer Ridge push moraine complex, North Norfolk, England. Boreas, 19: 165–180. Hart, J. K. 1994. Till fabric associated with deformable beds. Earth Surface Processes and Landforms, 19: 15–32. Hart, J. K. 1995. Subglacial erosion, deposition and deformation associated with deformable beds, Progress in Physical Geography, 19: 173–191. Hart, J. K. 1998. The deforming bed/debris-rich basal ice continuum and its implications for the formation of glacial landforms (flutes) and sediments (melt-out till). Quaternary Science Reviews, 17: 737–754. Hart, J. K. and Boulton, G. S. 1991. The interrelation of glaciotectonic and glaciodepositional processes within the glacial environment. Quaternary Science Reviews, 10: 335–350. Hart, J. K. and Roberts, D. H. 1994. Criteria to distinguish between subglacial glaciotectonic and glaciomarine sedimentation, I. Deformation styles and sedimentology. Sedimentary Geology, 91: 191–213. Hart, J. K., Hindmarsh, R. C. A. and Boulton, G. S. 1990. Styles of subglacial glaciotectonic deformation within the context of the Anglian ice-sheet. Earth Surface Processes and Landforms, 15: 227–241. Hattestrand, C. and Kleman, J. 1999. Ribbed Moraine Formation. Quaternary Science Reviews, 18: 43–61. Hays, J. D., Imbrie, J. and Shackleton, N. J. 1976. Variations in the earth’s orbit: Pacemaker of the ice ages. Science, 194: 1121–1132. Heezen, B. C. and Ewing, M. 1952. Turbidity currents and submarine slumps, and the 1929 Grand Banks earthquake. American Journal of Science, 250: 849–873. Heikkinen, O. and Tikkanen, M. 1989. Drumlins and flutings in Finland: their relationship to ice movement and to each other. Sedimentary Geology, 62: 349–356. Henderson, P. J. 1988. Sedimentation in an esker system influenced by bedrock topography near Kingston, Ontario, Canadian Journal of Earth Sciences, 25: 987–999. Heinrich, H. 1988. Origin and consequences of cyclic ice rafting in the Northeast Atlantic ocean during the past 130,000 years. Quaternary Research, 29: 142–152.

501

Henrich, R. 1986. A calcite dissolution pulse in the Norwegian-Greenland Sea during the last deglaciation. Geologische Rundschau, 75: 805–827. Henrich, R. 1989. Glacial/interglacial cycles in the Norwegian Sea: Sedimentology, paleoceanography and evolution of Late Pliocene to Quaternary northern hemisphere climate. In: O. Eldholm, J. Thiede and E. Taylor (eds) Proc. Final Reps. Ocean Drilling Program 104, pp. 189–232. Henrich, R. 1990. Cycles, rhythms and events in quaternary Arctic and Antarctic glaciomarine deposits. In: U. Bleil and J. Thiede (eds) Geological History of the Polar Oceans: Arctic versus Antarctic. NATO ASI Series, C 308, Kluwer Academic Publishers, Dordrecht, pp. 213–244. Henrich, R. 1991. Cycles, rhytms, and events on high input and low input glaciated continental margins. In: G. Einsele, W. Ricken and A. Seilacher (eds) Cycles and Events in Stratigraphy. Springer-Verlag, Berlin, pp. 751–772. Henrich, R., Kassens, H., Vogelsang, E. and Thiede, J. 1989a. Sedimentary facies of glacial/interglacial cycles in the Norwegian Sea during the last 350 ka. Marine Geology, 86: 283–319. Herman, Y., Osmond, J. K. and Somayajulu, B. L. K. 1989. Late Neogene Arctic paleoceanography, micropaleontology, stable isotopes and chronology. In Herman, Y. (ed), The Arctic Seas: Climatology, Oceanography, Geology and Biology. Van Nostrand Reinhold, New York, pp. 581–655. Herron, S. and Langway, Jr., C. C. 1979. The debris-laden ice at the bottom of the Greenland ice sheet. Journal of Glaciology, 23: 193–207. Hewitt, K. 1982. Natural dams and outburst floods of the Karakoram Himalaya. In: J. W. Glen (ed.) Hydrological Aspects of Alpine and High-Mountain Areas. Proceedings of the Exeter Symposium, July 1982, International Association of Hydrological Sciences, No. 138, pp. 259–269. Hicks, D. M., McSaveney, M. J. and Chinn, T. J. H. 1990. Sedimentation in proglacial Ivory Lake, Southern Alps, New Zealand. Arctic and Alpine Research, 22: 26–42. Hicock, S. R. 1990. Genetic till prism. Geology, 18: 517–519. Hicock, S. R. 1992. Lobal interactions and rheologic superposition in subglacial till near Bradtville, Ontario, Canada. Boreas, 21: 73–88. Hicock, S. R., Goff, J. R., Lian, O. B. and Little, E. C. 1996. On the interpretation of subglacial till fabric. Journal of Sedimentary Research, 66: 928–934.

502

REFERENCES

Hirano, M. and Aniya, M. 1988. A rational explanation of cross-profile morphology for glacial valleys and of glacial valley development. Earth Surface Processes and Landforms, 13: 707–716. Hirvas, H., Kujansuu, R. and Tynni, R. 1976. Till stratigraphy in Northern Finland. In D. J. Easterbrook and V. Sibrava (eds) Report nr. 3 IGCP Project 73/1/24 (Quaternary Glaciations in the Northern Hemisphere), pp. 256–273. Hirvas, H., Lagerb¨ack, R., M¨akinen, K., Nenonen, K., Olsen, L., Rodhe, L. and Thoresen, M. 1988. The Nordkalott Project: studies of Quaternary geology in northern Fennoscandia. Boreas, 17: 431–437. Hitchcock, E. 1841. First Anniversary Address before the Association of American Geologists. American Journal of Science, 41: 232–275. Hodgson, D. M. 1982. Hummocky and fluted moraine in part of North-West Scotland. Unpublished Ph.D. Thesis, University of Edinburgh. Hooke, R. LeB. 1984. On the role of mechanical energy in maintaining subglacial water conduits at atmospheric pressure. Journal of Glaciology, 30: 180–187. Hoffman, K. and Piotrowski, J. A. 2001. Till m´elange at Amsdorf, central Germany: sediment erosion, transport and deposition in a complex, soft-bedded subglacial system. Sedimentary Geology, 140: 215–234. Hoffman, P. F. 1989. Speculations on Laurentia’s first gigayear (2.0 to 1.0 Ga). Geology, 17: 135–138. Hoffman, P. F., Kaufman, A. J., Halverson, G. P. and Schrag, D. P. 1998. A Neoproterozoic snowball Earth. Science, 281: 1342–1346. Holdsworth, G. and Glynn, J. E. 1978. Iceberg calving from floating glaciers by a vibrating mechanism. Nature, 274: 464–466. Holmes, C. D. 1941. Till fabric. Geological Society of America, Bulletin, 52: 1299–1354. Holmes, C. D. 1960. Evolution of till-stone shapes, central New York. Geological Society of America, Bulletin, 71: 1645–1660. Holmlund, P. 1986. Geometry of the near-surface drainage system on Storglaci¨aren, Sweden. In: P. Alean, S. Braun, A. Iken, K. Schram and G. Zwosta (eds) Hydraulic Effects at the Glacier Bed and Related Phenomena. International Workshop 16–19 September 1985, Interlaken, Switzerland. Mitteilungen der Versuchsanstalt f¨ur Wasserbau, Hydrologie und Glaziologie, No. 90, pp. 57–58. Holmlund, P. 1988. Internal geometry and evolution of moulins, Storglaci¨aren, Sweden. Journal of Glaciology, 34: 242–248. Holtedahl, H. 1975. The geology of the Hardangerfjord,

West Norway, Bulletin – Norges Geologiske Undersokelse, 323, Bulletin 36, 87 pp.. Honjo, S., Manganini, J. and Wefer, G. 1988. Annual particle flux and a winter outburst of sedimentation in the northern Norwegian Sea. Deep Sea Research, 35: 1224–1234. Hooke, R. LeB. 1973. Structure and flow in the margin of the Barnes Ice Cap, Baffin Island, N.W.T., Canada, Journal of Glaciology, 12: 423–438. Hooke, R. LeB. 1984. On the role of mechanical energy in maintaining subglacial water conduits at atmospheric pressure. Journal of Glaciology, 30: 180–187. Hooke, R. LeB. 1991. Positive feedbacks associated with erosion of glacial cirques and overdeepenings. Geological Society of America, Bulletin, 103: 1104–1108. Hooke, R. LeB. 1998. Principles of Glacier Mechanics. Prentice Hall, Englewood Cliffs, NJ, 248 pp. Hooke, R. LeB., Brzozowski, J. and Bronge, C. 1983. Seasonal variations in surface velocity, Storglaci¨aren, Sweden. Geografiska Annaler, 65A: 263–277. Hooke, R. LeB., Wold, B. and Hagen, J. O. 1985. Subglacial hydrology and sediment transport at Bondhusbreen, southwest Norway. Geological Society of America, Bulletin, 96: 388–397. Hooke, R. LeB., Holmlund, P. and Iverson, N. R. 1987. Extrusion flow demonstrated by bore-hole deformation measurements over a riegel, Storglaci¨aren, Sweden. Journal of Glaciology, 33: 72–78. Hooke, R. LeB., Calla, P., Holmlund, P., Nilsson, M. and Stroeven, A. 1989. A 3 year record of seasonal variations in surface velocity, Storglaci¨aren, Sweden. Journal of Glaciology, 35: 235–247. Hooke, R. LeB., Laumann, T. and Kohler, J. 1990. Subglacial water pressures and the shape of subglacial conduits. Journal of Glaciology, 36: 67–71. Hoppe, G. and Schytt, V. 1953. Some observations on fluted moraine surfaces. Geografiska Annaler, 35: 105–115. Horsfield, W. T. 1977. An experimental approach to basement controlled faulting. Geologie en Mijnbouw, 56: 363–370. Houghton, J. T., Jenkins, G. J. and Ephraums, J. J. 1990. Climate Change, the I.P.C.C. Scientific Assessment. IPCC (Intergovernmental Panel on Climate Change), Sweden. 365 pp. Houmark-Nielsen, M. 1983. Glacial stratigraphy and morphology of the northern Bælthav region. In J. Ehlers (ed.) Glacial Deposits in North-West Europe. A. A. Balkema Publishers, Rotterdam, pp. 211–218. Houston, R. S., Lanthier, L. R., Karlstrom, K. K. and Sylvester, G. 1981. Early Proterozoic diamictite of

REFERENCES

southern Wyoming. In: M. J. Hambrey and W. B. Harland (eds) Earth’s pre-Pleistocene Glacial Record. Cambridge University Press, Cambridge, pp. 795–799. Howard, W. R. 1997. Palaeoclimatology; a warm future in the past. Nature, 388: 418–419. Hubbard, B. 1991. Freezing-rate effects on the physical characteristics of basal ice formed by net adfreezing. Journal of Glaciology, 37: 339–347. Hubbert, M. K. and Rubey, W. W. 1959. Role of fluid pressure in mechanics of overthrust faulting. Geological Society of America, Bulletin, 70: 115–166. Huddart, D. 1983, Flow tills and ice-walled lacustrine sediments, the Petteril Valley, Cumbria, England. In: E. B. Evenson, Ch. Schl¨uchter, and J. Rabassa (eds) Tills and Related Deposits. A. A. Balkema Publishers, Rotterdam, pp. 81–94. Hughes, T. J. 1981. Numerical reconstruction of paleo-ice sheets. In: G. H. Denton and T. J. Hughes (eds) The Last Great Ice Sheets. Wiley-Interscience, New York, pp. 221–261. Hughes, T. J. 1992. Theoretical calving rates from glaciers along ice walls grounded in water of variable depths. Journal of Glaciology, 38: 282–294. Hughes, T. J. 1998. Ice Sheets. Oxford University Press, Oxford, 343 p. Humlum, O. 1981. Observations on debris in the basal transport zone of Myrdalsj¨okull, Iceland. Annals of Glaciology, 2: 71–77. Humlum, O. 1985. Changes in texture and fabric of particles in glacial traction with distance from source, Myrdalsj¨okull, Iceland. Journal of Glaciology, 31: 150–156. Hutchinson, G. E. 1957. A Treatise on Limnology. Vol. I: Geography, Physics and Chemistry. Wiley, New York, 1015 pp. Hutter, K. 1983. Theoretical Glaciology; Material Science of Ice and the Mechanics of Glacier and Ice Sheets. D. Reidel Publishing, Rotterdam, 510 pp. Hutter, K., Blatter, J. and Funk, M. 1988. A model computation of moisture content in polythermal glaciers. Journal of Geophysical Research, 93: 12 205–12 214. Hutter, K. and Engelhardt, H. F. 1988. The use of continuum thermodytnamics in the formulation of ice-sheet dynamics. Annals of Glaciology, 11: 46–51. Hutton, J. 1795. Theory of the Earth, 2 vols. William Creech, Edinburgh. (Reprinted 1959 in facsimile, Hafner, New York. 203 pp.) Iken, A. and Bindschadler, R. A. 1986. Combined measurements of subglacial water pressure and surface velocity of the Findelengletscher, Switzerland. Conclusions about

503

drainage system and sliding mechanism. Journal of Glaciology, 32: 101–119. Imbrie, J. and Imbrie, K. P. 1979. Ice Ages: Solving the Mystery. Enslow, Hillside NJ, 224 pp. Imbrie, J., Hays, J. D., Martinson, D. G., McIntyre, A., Mix, A. C., Morley, J. J., Pisias, N. G., Prell, W. and Shackleton, N. J. 1984. The orbital theory of Pleistocene climate: Support from a revised chronology of the marine a¨ 18O record. In: A. Berger, J. Imbrie, J. Hays, G. Kukla and B. Saltzman (eds) Milankovitch and Climate, Part I. D. Reidel, Hingham MA, pp.269–305. Ing´olfsson, O. 1988. Large-scale glaciotectonic deformation of soft sediments: a case study of a late Weichselian sequence in Western Iceland. In: D. G. Croot (ed.) Glaciotectonics: Forms and Processes. A. A. Balkema Publishing, Rotterdam, pp. 101–107. Irwin, J. and Pickrill, R. A. 1982. Water temperature and turbidity in a glacial fed lake: Lake Tekapo. New Zealand Journal of Freshwater Research 16: 189–200. Iverson, N. R. 1990. Laboratory simulations of glacial abrasion: comparison with theory. Journal of Glaciology, 36: 304–314. Iverson, N. R. 1991a. Morphology of glacial striae: implications for abrasion of glacier beds and fault surfaces. Geological Society of America, Bulletin, 103: 1308–1316. Iverson, N. R. 1991b. Potential effects of subglacial waterpressure fluctuations on quarrying. Journal of Glaciology, 37: 27–36. Iverson, N. R. 1993. Regelation of ice through debris at glacier beds; implications for sediment transport. Geology, 21: 559–562. Iverson, N. R. In press. Regelation of ice through subglacial debris: implications for sediment transport. Geology. Jacobs, S. S. 1989. Marine controls on modern sedimentation on the Antarctic continental shelf. Marine Geology, 85: 121–153. Jaeger, J. C. and Cook, N. G. W. 1979 (3rd Edition, 593 pp.). Fundamentals of Rock Mechanics. 2nd Edition, Chapman and Hall Publishing, Halsted Press Book, London. Jahns, R. H. 1943. Sheet structure in granites: Its origin and use as measure of glacial erosion in New England. Journal of Geology, 51:71–98. Jansen, E. and Erlenkeuser, H. 1985. Ocean circulation in the Norwegian-Greenland sea during the last deglaciation: isotopic evidence. Palaeogeography, Palaeoclimatology, Palaeoecology, 49: 189–206. Jansen, E. and Sjøholm, J. 1991. Reconstruction of glaciation over the past 6 Myr from ice-borne deposits in the Norwegian Sea. Nature, 349: 600–603.

504

REFERENCES

Jansen, E., Befring, S., Bugge, T., Eidvin, T., Holtedahl, H. and Sejrup, H. P. 1987. Large submarine slides on the Norwegian continental margin: sediments, transport and timing. Marine Geology, 78: 77–107. Jellinek, H. H. G. 1959. Adhesive properties of ice. Journal of Colloid Science, 14: 268–280. Jenson, J. W., Clark, P. U., MacAyeal, D. R., Ho, C. L. and Vela, J. C. 1995. Numerical modeling of advective transport of saturated deforming sediment beneath the Lake Michigan Lobe, Laurentide Ice Sheet. Geomorphology, 14: 157–166. Jenson, J. W., MacAyeal, D. R., Clark, P. U., Ho, C. L. and Vela, J. C. 1996. Numerical modeling of subglacial sediment deformation: implications for the behaviour of the Lake Michigan Lobe, Laurentide Ice Sheet. Journal of Geophysical Research, 101(B4): 8717–8728. J´ohannesson, T., Raymond, C. F. and Waddington, E. 1989. Time-scale for adjustment of glaciers to changes in mass balance. Journal of Glaciology, 35: 355–369. Johnsen, S. J., Clausen, H. B., Dansgaard, W., Fuhrer, K., Gundestrup, N. S., Hammer, C. U., Iversen, P., Jouzel, J., Stauffer, B. and Steffensen, J. P. 1992. Irregular glacial interstadials recorded in a new Greenland Ice core. Nature, 359: 311–313. Johnsen, S. J., Clausen, H. B., Dansgaard, W., Gundestrup, N. S., Hammer, C. U. and Tauber, H. 1995. The Eem stable isotope record along the GRIP ice core and its interpretation, Quaternary Research, 43: 117–124. Johnson, A. M. 1970. Physical Processes in Geology. Freeman & Cooper, San Francisco, 577 pp. Johnson, C. B. Unpublished. Characteristics and mechanics of formation of glacial arcuate abrasion cracks. Ph.D. Thesis, Pennsylvania State University, 1975. Jones, G. A. and Keigwin, L. D. 1988. Evidence from Fram Strait (78°N) for early deglaciation. Nature, 336: 56–59. Jonsson, J. 1982. Notes on the Katla volcanoglacial debris flows. J¨okull, 32: 61–68. Jopling, A. V. and Walker, R. G. 1968. Morphology and origin of ripple-drift cross-lamination, with examples from the Pleistocene of Massachusetts. Journal of Sedimentary Petrology, 38: 971–984. Jouzel, J., Lorius, C., Petit, J. R., Genthon, C., Barkov, N. I., Kotlyakov, V. M and Petrov, V. M. 1987. Vostock ice core: a continuous isotope temperature record over the last climatic cycle 160,000 years. Nature, 329: 403–408. Jouzel, J., Barkov, N. I., Barnola, J. M., Bender, M., Chappellaz, J., Genthon, C., Kotlyakov, V. M., Lipenkov, V., Lorius, C., Petit, J. R., Raynaud, D., Raisbeck, G., Ritz, C., Sowers, T., Stievenard, M., Yiou, M. and Yiou,

P. 1993. Extending the Vostock ice-core record of palaeoclimate to the penultimate glacial period. Nature, 364: 407–412.

K¨alin, M. 1971. The active push-moraine of the Thompson glacier, Axel Heiberg Island, Candian Arctic Archipelago. Dissertation, Z¨urich, 68 pp. Kamb, B. 1970. Sliding motion of glaciers: theory and observation. Reviews of Geophysics and Space Physics, 8: 673–728. Kamb, B. 1987. Glacier surge mechanism based on linked cavity configuration of the basal water conduit system. Journal of Geophysical Research, 92: 9083–9100. Kamb, B. 1991. Rheological nonlinearity and flow instability in the deforming bed mechanism of ice stream motion. Journal of Geophysical Research, 96B: 16 585–16 595. Kamb, B. and Engelhardt, H. F. 1989. Flow mechanics of West Antarctic Ice Streams: observations by borehole geophysics. EOS, 70: 1081. Kamb, B. and LaChapelle, E. 1964. Direct observations of the mechanism of glacier sliding over bedrock. Journal of Glaciology, 5: 159–172. Kamb, B. and Shreve, R. L. 1966. Results of a new method for measuring internal deformation in glaciers. Transactions American Geophysics Union, 47: 190 (abstract). Kamb, B., Raymond, C. F., Harrison, W. D., Engelhardt, H. F., Echelmeyer, K. A., Humphrey, N. F., Brugman, M. M. and Pfeffer, T. 1985. Glacier surge mechanism: 1982–1983 surge of Variegated Glacier, Alaska. Science, 227: 469–479. Karrow, P. F. 1967. Pleistocene Geology of the Scarborough Area. Ontario Dept. of Mines, Geological Report 46. Karrow, P. F. 1974. Till straigraphy in parts of southwestern Ontario. Geological Society of America Bulletin, 85: 761–768. Karrow, P. F. 1984a. Comments on ‘Sedimentation in a large lake: an interpretation of the Late Pleistocene stratigraphy at Scarborough Bluffs, Ontario, Canada’. Geology, 12: 185. Karrow, P. F. 1984b. Lithofacies types and vertical profile models; an alternative approach to the description and environmental interpretation of glacial diamict and diamictite sequences – discussion. Sedimentology, 31: 883–884. Karrow, P. F. 1989. Quaternary geology of the Great Lakes subregion. In: R. J. Fulton (ed) Quaternary Geology of Canada and Greenland. Geological Survey of Canada No.1 also Geological Society of America, The Geology of North America v. K-1, Chapter 4, pp. 326–350. Kasting, J. F. 1989. Long-term stability of the Earth’s

REFERENCES

climate. Palaeogeography, Palaeoclimatology, Palaeoecology (Global and Planetary Change Section), 75: 83–95. Kasting, J. F. 1993. New spin on ancient climate. Nature, 364: 759–760. Kasting, J. F. and Toon, O. B. 1989. Climate evolution on the terrestrial planets. In: S. K. Atreya, J. B. Pollack and M. S. Matthews (eds) Origin and Evolution of Planetary and Satellite Atmospheres. University of Arizona Press, Tucson, pp. 423–449. Kasting, J. F., Pollack, J. B. and Crisp, D. 1984. Effects of high CO2 levels on surface temperature and atmospheric oxidation state of the early Earth. Journal of Atmospheric Chemistry, 1: 403–428. Kaszycki, C. A. 1987. A model for glacial and proglacial sedimentation in the shield terrane of southern Ontario. Canadian Journal of Earth Sciences, 24: 2373–2391. Kaszycki, C. A. 1989. Quaternary Geology and Glacial History of the Haliburton Region, South Central Ontario, Canada: A Model for Glacial and Proglacial Sedimentation. Unpublished Ph. D. Dissertation, University of Illinois at Urbana-Champaign, 219 pp. Kehew, A. E. and Lord, M. L. 1987. Glacial-lake outbursts along the mid-continent margins of the Laurentide icesheet. In: L. Mayer and D. Nash (eds) Catastrophic Flooding. The Binghampton Symposia in Geomorphology: International Series, No. 18, Allen & Unwin, Boston, pp. 95–121. Kelly, R. I. and Martini, I. P. 1986. Pleistocene glaciolacustrine deltaic deposits of the Scarborough Formation, Ontario, Canada. Sedimentary Geology, 47: 27–52. Kemmis, T. J. 1991. Glacial Landforms, Sedimentology, and Depositional Environments of the Des Moines Lobe, Northern Iowa. Unpublished Ph.D. Dissertation, University of Iowa, 384 pp. Kemmis, T. J. 1996. Lithofacies associations for terrestrial glacigenic succession. In Menzies, J. (ed) Past Glacial Environments – sediments, forms and techniques, vol. 2, Butterworth-Heineman. Kemmis, T. J., Hallberg, G. R., and Lutenegger, A. J. 1981. Depositional Environments of Glacial Sediments and Landforms on the Des Moines Lobe, Iowa. Iowa Geological Survey Guidebook Series, 6, 132 pp. Keys, H. J. D., Jacobs, S. S. and Barnett, D. 1990. The calving and drift of iceberg B-9 in the Ross Sea, Antarctica. Antarctic Science, 2: 243–257. Kick, W. 1989. The decline of the last Little Ice Age in High Asia compared with that in the Alps. In: J. Oerlemans (ed.) Glacier Fluctuations and Climatic Change. Kluwer Academic Publishers, Dordrecht, pp. 129–142. King, C. A. M. and Buckley, J. T. 1968. The analysis of

505

stone size and shape in arctic environments. Journal of Sedimentary Petrology, 38: 200–214. King, L. H. and Fader, G. B. J. 1986. Wisconsinan Glaciation of the Atlantic Continental Shelf of Southeast Canada. Geological Survey of Canada, Bulletin No. 363, 72 pp. King, L. H., Rokoengen, K. and Gunleiksrud, T. 1987. Quaternary Seismostratigraphy of the Mid-Norwegian Shelf, 65°–67°31⬘N – A Till Tongue Stratigraphy. IKU Publications, Trondheim, 114, 58 pp. Kirkbride, M. P. 1989. The influence of sediment budget on the geomorphic activity of the Tasman Glacier, Mount Cook National Park, New Zealand. Unpublished Ph.D thesis, University of Canterbury, New Zealand, 395 pp. Kirkbride, M. P. and Sugden, D. E. 1992. New Zealand Loses its Top. Geographical Magazine, 64: 30–34. Kjaer, K. H. and Kr¨uger, J. 1998. Does clast size influence fabric strength? Journal of Sedimentary Research, 68: 746–749. Klassen, R. W. 1989. Quaternary geology of the southern Canadian Interior Plains. In: R. J. Fulton (ed.) Quaternary Geology of Canada and Greenland. Geology of Canada No.1, Geological Survey of Canada, DNAG Volume K-1, Geological Society of America, pp. 138–166. Kluiving, S. J., Rappol, M. and Van der Wateren, F. M. 1991. Till stratigraphy and ice movements in eastern Overijssel, The Netherlands. Boreas, 20: 193–205. Knight, P. G. 1992. Glaciers. Progress in Physical Geography, 16: 85–89. Kobarg, W. 1988. The tide-dependent dynamics of the Ekstrom Ice Shelf, Antarctica. Reports on Polar Research, 50, 164 pp. Koc Karpuz, N. and Jansen, E. 1992. A high resolution diatom record of the last deglaciation from the SE Norwegian Sea: documentation of rapid climatic changes. Paleoceanography, 7: 499–520. Kochel, R. C. and Baker, V. R. 1988. Paleoflood analysis using slackwater deposits. In: V. R. Baker, R. C. Kochel and P. C. Patton (eds) Flood Geomorphology. Wiley-Interscience Publications, New York, pp. 357–376. Koerner, R. M. and Fisher, D. A. 1979. Discontinuous flow, ice texture, and dirt content in the basal layers of the Devon Island ice cap. Journal of Glaciology, 23: 209–221. Kor, P. S. G., Shaw, J. and Sharpe, D. R. 1991. Erosion of bedrock by subglacial meltwater, Georgian Bay, Ontario: a regional view. Canadian Journal of Earth Sciences, 28: 623–642. Koteff, C. and Pessl, Jr., F. 1981. Systematic Ice Retreat in New England. Geological Survey Professional Paper No. 1179, Washington DC, 20 pp.

506

REFERENCES

Kozarski, S. and Szupryczynski, J. 1973. Glacial forms and deposits in the Siduj¨okull deglaciation area, Scientific results of the Polish Geographical Expedition to Vatnaj¨okull (Iceland), 1968. Geographia Polonica, 26: 255–311. Krimmel, R. 1988. Cover photograph. Eos, 69:1. Krimmel, R. M. and Meier, M. F. 1989. Glacial Geology and Geomorphology of North America; Volume 1, Glaciers and Glaciology of Alaska. 28th International Geological Congress, Field Trip Guidebook T301, American Geophysical Union, 61 pp. Krimmel, R. M. and Vaughn, B. H. 1987. Columbia Glacier, Alaska: changes in velocity 1977–1986. Journal of Geophysical Research, 92: 8961–8968. Kristensen, M. 1983. Iceberg calving and deterioration in Antarctica. Progress in Physical Geography, 7: 313–328. Kr¨uger, J. 1979. Structures and textures in till indicating subglacial deposition. Boreas, 8: 323–340. Kr¨uger, J. 1987. Relationship of drumlin shape and distribution to drumlin stratigraphy and glacial history, Myrdalsj¨okull, Iceland. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema, Rotterdam, pp. 257–266. Kr¨uger, J. 1994. Glacial processes, sediments, landforms and stratigraphy in the terminus region of Myrdalsj¨okull, Iceland. Folia Geographica Danica, 21: 233. Kr¨uger, J. and Thomsen, H. H. 1984. Morphology, stratigraphy, and genesis of small drumlins in front of the glacier Myrdalsj¨okull, south Iceland. Journal of Glaciology, 30: 94–105. Kuhn, M. 1984. (Abstr.) Thirty years of mass balance and climate records on Hintereisferner: a basis for modeling. Annals of Glaciology, 5: 212. Kukla, G. J. 1977. Pleistocene land-sea correlations. EarthScience Reviews, 13: 307–374. Kukla, G. J. 1989. Long continental records of climate – an introduction. Palaeogeography, Palaeoclimatology, Palaeoecology, 72: 1–9. Kupsch, W. O. 1962. Ice-thrust ridges in Western Canada. Journal of Geology, 70: 582–594. Kutzbach, J. E. and Guetter, P. J. 1986. The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18 000 years. Journal of the Atmospheric Sciences, 43: 1726–1759. Laberg, J. S. and Vorren, T. O., 1995. Late Weichselian submarine debris flow deposits on the Bear Island Trough Mouth Fan. Marine Geology, 127: 45–72. Lagerb¨ack, R. 1988. The Veiki moraines in northern Sweden – widespread evidence of an Early Weichselian deglaciation. Boreas, 17: 469–486. Lagerlund, E. 1987. An alternative Weichselian glaciation

model, with special reference to the glacial history of Skåne, South Sweden. Boreas, 16: 433–459. LaSalle, P. and Shilts, W. W. 1993. Younger Dryas-age readvance of Laurentide ice into the Champlain Sea. Boreas, 22: 25–37. Laverdi‘ere, C., Guimont, P. and Dionne, J.-C. 1985. Les Formes et les marques de l’´erosion glaciaire du plancher rocheaux: signification, terminologie, illustration. Palaeogeography, Palaeoclimatology, Palaeoecology, 51: 365–387. Lavrushin, Y. A. 1971. Dynamische Fazies und Subfazies der Grundmor¨ane. Zeitschrift f¨ur Angewandte Geologie, 17: 337–343. Lawler, D. 1991. Sediment and solute yield from the Jokulsa a Solheimasandi glacierized river basin, southern Iceland. In: J. K. Maizels and C. Caseldine (eds) Environmental Change in Iceland: Past and Present. Kluwer Academic Publishers, Dordrecht, pp. 303–332. Lawn, B. R. 1967. Partial cone crack formation in a brittle material loaded with a sliding spherical indenter. Proceedings of the Royal Society of London, Series A, 299: 307–316. Lawn, B. R. and Wilshaw, R. 1975. Review of indentation fracture: principles and applications. Journal of Materials Science, 10: 1049–1081. Lawson, D. E. 1979a. Sedimentological Analysis of the Western Terminus Region of Matanuska Glacier, Alaska. CRREL Report, Hanover, New Hampshire 79–9: 122 pp. Lawson, D. E. 1979b. A comparison of the pebble orientations in ice and deposits of the Matanuska glacier, Alaska. Journal of Geology, 87: 629–645. Lawson, D. E. 1981a. Sedimentological Characteristics and Classification of Depositional Processes and Deposits in the Glacial Environment. Cold Regions Research and Engineering Laboratory, Report No. 81–27. Lawson, D. E. 1981b. Distinguishing characteristics of diamictons at the margin of the Matanuska glacier, Alaska. Annals of Glaciology, 2: 78–84. Lawson, D. E. 1982. Mobilization, movement and deposition of active subaerial sediment flows, Matanuska Glacier, Alaska. Journal of Geology, 90: 279–300. Leckie, D. A. and McCann, S. B. 1982. Glacio-lacustrine sedimentation on low slope prograding delta. In: R. Davidson-Arnott, W. Nickling, and B. D. Fahey (eds) Research in Glacial, Glacio-fluvial and Glacio-lacustrine Systems. 6th Guelph Symposium on Geomorphology. Geo Books Publishing, Norwich, pp. 261–278. Lee, J. R. 2001. Genesis and palaeogeographical significance of the Corton Diamicton (basal member of the North Sea Drift Formation), East Anglia, UK. Proceedings of the Geologists’ Association, 112, 29– 43.

REFERENCES

Leinen, M., Cwienk, D., Heath, G. R., Biscaye, P. E., Kolla, V., Thiede, J.and Dauphin, J. P. 1986. Distribution of biogenic silica and quartz in recent deep-sea sediments. Geology, 14: 199–203. Lemmen, D. S. 1984. Sedimentation in a glacially-fed arctic lake: Tasikutaak Lake Cumberland Peninsula, Baffin Island, NWT MSc thesis, Department of Geography, Queen’s University, Kingston, Ontario. Lemmen, D. S., Gilbert, R., Smol, J. P. and Hall, R. I. 1988. Holocene sedimentation in glacial Tasikutaaq Lake, Baffin Island. Canadian Journal of Earth Sciences, 25: 810–823. Leonard, E. M. 1986a. Varve studies at Hector Lake, Alberta, Canada and the relationship between glacial activity and sedimentation. Quaternary Research, 25: 199–214. Leonard, E. M. 1986b. Use of lacustrine sedimentary sequences as indicators of Holocene glacial history, Banff National Park, Alberta, Canada. Quaternary Research, 26: 218–231. Letr´eguilly, A. 1988. Relation between the mass balance of western Canadian mountain glaciers and meteorological data. Journal of Glaciology, 34: 11–18. Levson, V. M. and Rutter, N. W. 1988. A lithofacies analysis and interpretation of depositional environments of montane glacial diamictons, Jasper, Alberta, Canada. In: R. P. Goldthwait and C. L. Matsch (eds) Genetic Classification of Glacigenic Deposits. A. A. Balkema Pubishers, Rotterdam, pp. 117–142. Lewis, E. L. and Perkin, R. G. 1985. The winter oceanography of McMurdo Sound, Antarctica. In S. S. Jacobs (ed.) Oceanology of the Antarctic Continental Shelf. American Geophysical Union, Antarctic Research Series, 43, pp. 145–165. Libby, W. F. 1955. Radiocarbon Dating. 2nd Edition, University of Chicago Press, Chicago, 175 pp. Lindquist, P. A., Lai, H. H. and Alm, O., 1984. Indentation fracture development in rock continuously observed with a scanning electron microscope. International Journal of Rock Mechanics and Mining Sciences and Geochemical Abstracts, 21(4): 165–182. Lindsey, D. A. 1969. Glacial sedimentology of the Precambrian Gowganda Formation, Ontario, Canada. Geological Society of America Bulletin, 80: 1685–1702. Lingle, C. S. and Brown, T. J. 1987. A subglacial aquifer bed model and water pressure dependent basal sliding relationship for a West Antarctic ice stream. In Van der Veen, C. J. and Oerlemans, J. (eds.), Dynamics of the West Antarctic Ice Sheet. D. Reidel Publishing, Dordrecht, pp. 249–285.

507

Lister, G. S. 1988. A 15,000-year isotopic record from Lake Z¨urich of deglaciation and climatic change in Switzerland. Quaternary Research, 29: 129–141. Liverman, D. G. E. 1987. Sedimentation in ice-dammed Hazard Lake, Yukon. Canadian Journal of Earth Sciences, 24: 1797–1806. Lliboutry, L. 1962. L’ erosion glaciaire. International Association of Hydrological Sciences Publication 59: 219–225. Lliboutry, L. 1968. General theory of subglacial cavitation and sliding of temperate glaciers. Journal of Glaciology, 7: 21–58. Lliboutry, L. 1978. Sliding of glacier on a plane scattered with hemispherical obstacles. Annales de Geophysique, 34: 147–162. Lliboutry, L. 1979. Local friction laws for glaciers: a critical review and new openings. Journal of Glaciology, 23: 67–95. Lliboutry, L. 1983. Modifications to the theory of intraglacial waterways for the case of subglacial ones. Journal of Glaciology, 29: 216–226. Lliboutry, L. and Ritz, C. 1978. Ecoulement permanent d’un fluide visqueux non lineaire (corps de Glen) autour d’une sphere parfaitement lisse. Annales de Geophysique, 34: 133–146. Lock, G. S. H. 1990. The Growth and Decay of Ice. Cambridge University Press, Cambridge, 434 pp. Loomis, S. R. 1970. Morphology and ablation processes on glacier ice. In: V. C. Bushnell and R. H. Ragle (eds.), Icefield Ranges Research Project Scientific Results Volume 2. American Geographical Society and Arctic Institute of North America, New York, Montreal, pp. 27–31. Lorius, C. 1999. The greenhouse effect; certainties and uncertainties, Comptes Rendus de l’Academie des Sciences, Serie II. Sciences de la Terre et des Planetes, 328 (4): 223. Lotter, A. F. 1989. Evidence of annual layering in Holocene sediments of Soppensee, Switzerland. Aquatic Sciences, 51: 19–30. Lowe, J. J. and Gray, J. M. 1980. The stratigraphic subdivision of the lateglacial in northwest Europe: a discussion. In: J. J. Lowe, J. M. Gray and J. E. Robinson (eds) Studies in the Lateglacial of Northwest Europe. Pergamon Press, Oxford, pp. 157–175. Lowe, J. J. and Walker, M. J. C. 1984. Reconstructing Quaternary Environments. Longman, London, 389 pp. Lowe, J. J., Ammann, B., Birks, H. H., Bjork, S., Coope, G. R., Cwynar, L., De Beaulieu, J. L., Mott, R. J., Peteet, D. M. and Walker, M. J. C. 1994. Climatic changes in areas adjacent to the North Atlantic during

508

REFERENCES

the last glacial–interglacial transition (14–9 ka BP): a Contribution to IGCP-253. Journal of Quaternary Science, 9: 185–198. Lowell, T. V., Heusser, C. J., Anderson, B. G., Moreno, P., Hauser, A., Heusser, L. E., Schl¨uchter, Ch., Merchant, D. R. and Denton, G. H. 1995. Interhemispheric correlation of Late Pleistocene events. Science, 269: 1541–1549. Lundqvist, J. 1969. Problems of the so-called Rogen Moraine. Sveriges Geologiska Unders¨okning, C 648: 32pp. Lundqvist, J. 1970. Studies of drumlin tracks in central Sweden. Acta Geographia Lodziensia, 24: 317–326. Lundqvist, J. 1981. Moraine morphology – terminological remarks and regional aspects. Geografiska Annaler, 63a: 127–138. Lundqvist, J. 1984. Indentation fracture development in rock continuously observed with a scanning electron microscope. International Journal of Rock Mechanics and Mineral Sciences and Geomechanics, Abstract, 21(4): 165–182. Lundqvist, J. 1989. Røgen (ribbed) moraine – identification and possible origin. Sedimentary Geology, 62: 281–292. Lundqvist, J. 1990. The Younger Dryas event in Scandinavia. In: J. Lundqvist and M. Saarnisto (eds) Termination of the Pleistocene, IGCP Project 253, Field Conference Guide 31. Geological Survey of Finland, pp. 5–24. Lundqvist, J., Clayton, L. and Mickelson, D. M., 1993. Deposition of the late Wisconsin Johnstown moraine, south-central Wisconsin. Quaternary International, 18: 53–59. Lunkka, J. P. 1988. Sedimentation and deformation of the North Sea Drift Formation in the Happisburgh area, North Norfolk. In: D. G. Croot (ed.) Glaciotectonics: Forms and Processes. A. A. Balkema Publishing, Rotterdam, pp. 109–122. Lunkka, J. P. 1994. Sedimentation and lithostratigraphy of the North Sea Drift and Lowestoft Till formations in the coastal cliffs of Northeast Norfolk, England, Journal of Quaternary Science, 9: 209–233. Lyell, C. 1839. Nouveaux e´ l´ements de g´eologie. PitoisLevrault, Paris, 648 pp. MacAyeal, D. R. 1989. Large-scale ice flow over a viscous basal sediment: theory and application to ice stream B, Antarctica. Journal of Geophysical Research, 94: 4071–4087. MacAyeal, D. R. 1992. Irregular oscillations of the West Antarctic ice sheet. Nature, 359: 29–32. MacAyeal, D. R., Bindschadler, R. A. and Scambos, T. A.,

1995. Basal friction of ice stream E, West Antarctica. Journal of Glaciology, 41: 247–262. MacClintock, P. and Dreimanis, A. 1964. Reorientation of till fabric by overriding glacier in the St. Lawrence Valley. American Journal of Science, 262: 133–142. MacCracken, M. C., Hecht, A. D., Budyko, M. I. and Izrael, Y. A. 1990. Prospects for Future Climate: A special US/ USSR Report on Climate and Climate Change. Lewis Publishers, Chelsea, Michigan, 270 pp. Mackiewicz, N. E., Powell, R. D., Carlson, P. R. and Molnia, B. F. 1984. Interlaminated ice-proximal glacimarine sediments in Muir Inlet, Alaska. Marine Geology, 57: 113–147. Magny, M. 1993. Solar influences on Holocene climatic changes illustrated by correlations between past lakelevel fluctuations and the atmospheric 14C record. Quaternary Research, 40: 1–9. Maher, L. J. and Mickelson, D. M. 1997. Palynological and radiocarbon evidence for deglaciation event in the Green Bay Lobe, Wisconsin. Quaternary Research, 46: 251–259. Maizels, J. K. 1976. A comparison of present-day and Pleistocene proglacial environments, with particular reference to geomorphology and sedimentology. Unpubl. Ph.D. Thesis, University of London, UK, 675 pp. Maizels, J. K. 1979. Proglacial aggradation and changes in braided channel patterns during a period of glacier advance: an Alpine example. Geografiska Annaler, 61A: 87–101. Maizels, J. K. 1989a. Sedimentology, paleoflow dynamics and flood history of j¨okulhlaup deposits: paleohydrology of Holocene sediment sequences in southern Iceland sandur deposits. Journal of Sedimentary Petrology, 59: 204–223. Maizels, J. K. 1989b. Sedimentology and palaeohydrology of Holocene flood deposits in front of a j¨okulhlaup glacier, South Iceland. In K. Beven and P. Carling (eds), Floods: Hydrological, Sedimentological, and Geomorphological Implications. J.Wiley & Sons, Chichester, pp. 239–253. Maizels, J. K. 1991. The origin and evolution of Holocene sandur deposits in areas of j¨okulhlaup drainage, south Iceland. In J. K. Maizels and C. Caseldine (eds) Environmental Change in Iceland: Past and Present. Kluwer Academic Publishers, Dordrecht, pp. 267–302. Maizels, J. K. 1992. Boulder ring structures produced during j¨okulhlaup flows – Origin and hydraulic significance. Geografiska Annaler, 74A: 21–33. Maizels, J. K. 1993. Lithofacies variations within sandur deposits: the role of runoff regime, flow dynamics and

REFERENCES

sediment supply characteristics. Sedimentary Geology, 85: 299–325. Maizels, J. K. and Russell, A. J. 1992. Quaternary perspectives on j¨okulhlaup prediction. In: J. M. Gray (ed.) Quaternary Proceedings, Applications of Quaternary Research. Quaternary Research Association, Cambridge, 2, pp. 133–152. Makinen, K. 1985. On the till-covered glaciofluvial formations in Finnish Lapland. Striae, 22: 33–40. Malavieille, J. 1987. Kinematics of compressional and extensional ductile shearing deformation in a metamorphic core complex of the northeastern Basin and Range. Journal of Structural Geology, 9(5/6): 541–554. Maltman, A. J. 1987. Shear zones in argillaceous sediments – an experimental study. In: M. E. Jones and R. M. F. Preston (eds) Deformation of Sediments and Sedimentary Rocks. Geological Society Special Publication No. 29, pp. 77–87. Maltman, A. J. 1988. The importance of shear zones in naturally deformed wet sediments. Tectonophysics, 145: 163–175. Maltman, A. J. 1994. Geological Deformation of Sediments. Chapman & Hall, London, 452 pp. Manabe, S. and Stouffer, R. J. 1993. Century-scale effects of increased atmospheric CO2 on the ocean–atmosphere system. Nature, 364: 215–218. Mandl, G., de Jong, L. N. J. and Maltha, A. 1977. Shear zones in granular material. Rock Mechanics, 9: 95–144. Mangerud, J., Andersen, S. T., Berglund, B. E. and Donner, J. J. 1974. Quaternary stratigraphy in Norden, a proposal for terminology and classification. Boreas, 3: 109–127. Marchant, D. R., Swisher III, C. C., Lux, D. R., West Jr., D. P. and Denton, G. H. 1993. Pliocene paleoclimate and East Antarctic Ice-Sheet history from surficial ash deposits. Science, 260: 667–670. Marcus, M. G. 1960. Periodic drainage of glacier-dammed Tulsequah Lake, British Columbia. Geographical Reviews, 50: 89–106. Marcussen, I. 1973. Studies on flow till in Denmark. Boreas, 2: 213–231. Marcussen, I. 1975. Distinguishing between lodgement till and flow till in Weichselian deposits. Boreas, 4: 113–123. Markgren, M. and Lassila, M. 1980. Problems of moraine morphology: Rogen moraine and Blattnick moraine. Boreas, 9: 271–274. Marshall, S. and Clarke, G. K. C. 1997a. A continuum mixture model of ice stream thermomechanics in the Laurentide Ice Sheet 1. Theory. Journal of Geophysics Research, 102(B9): 20 599–20 613. Marshall, S. and Clarke, G. K. C. 1997b. A continuum

509

mixture model of ice stream thermomechanics in the Laurentide Ice Sheet 2. Application to the Hudson Strait Ice Stream. Journal of Geophysics Research, 102(B9), 20 615–20 637. Marshall, S., Clarke, G. K. C., Dyke, A. S. and Fisher, D. A. 1996. Geological and topographic controls on fast flow in the Laurentide and Cordilleran Ice Sheets. Journal of Geophysics Research, 101(B8), 17 827–17 839. Martin, L. 1913. Some features of glaciers and glaciation in College Fjord, Prince William Sound, Alaska. Zeitschrift Gletscherkunde, 7: 289–333. Martini, I. P. 1990. Pleistocene glacial fan deltas in southern Ontario, Canada. In: A. Colella and D. B. Prior (eds) Coarse-Grained Deltas. International Association of Sedimentologists, Special Publication 10, pp. 281–295. Martinson, D. G., Pisias, N. G., Hays, J. D., Imbrie, J., Moore, T. C. and Shackleton, N. J. 1987. Age dating and the orbital theory of the ice ages: development of a highresolution 0 to 300,000-year chronostratigraphy. Quaternary Research, 27: 1–29. Mathews, W. H. 1956. Physical limnology and sedimentation in a glacial lake. Geological Society of America, Bulletin 67: 537–552. Mathews, W. H. 1959. Vertical distribution of velocity in Salmon Glacier, British Columbia. Journal of Glaciology, 3: 448–454. Mathews, W. H. 1965. Two self-dumping ice-dammed lakes in Britsh Columbia. Geographical Review, 55: 46–52. Matsch, C. L. and Schneider, A. F. 1986. Stratigraphy and correlation of the glacial deposits of the glacial lobe complex in Minnesota and northwestern Wisconsin. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 59–64. Matthes, F. E. 1930. Geological History of Yosemite Valley. U.S. Geological Survey Professional Paper, 160, 137 pp. Matthews, J. A. 1992. The Ecology of Recently-Deglaciated Terrain. Cambridge University Press, Cambridge, 386 pp. Matthews, R. K. 1984. Dynamic Stratigraphy: An Introduction to Sedimentology and Stratigraphy. 2nd Edition. Prentice-Hall, Englewood Cliffs, New Jersey, 489 pp. Mayo, L. R. 1984. Glacier mass balance and runoff research in the U.S.A.. Geografiska Annaler, Series A, 66: 215–227. Mazo, V. L. 1989. Waves on glacier beds. Journal of Glaciology, 35: 179–182. McCarroll, D. 1989. Potential and limitations of the Schmidt hammer for relative-age dating: field tests on

510

REFERENCES

Neoglacial moraines, Jotunheimen, southern Norway. Arctic and Alpine Research, 21: 268–275. McDonald, B. C. and Banerjee, I. 1971. Sediments and bedforms on a braided outwash plain. Canadian Journal of Earth Sciences, 8: 1282–1301. McDonald, V. C. and Vincent, J. S. 1972. Fluvial Sedimentary Structures Formed Experimentally in a Pipe and their Implications for Interpretation of Subglacial Sedimentary Environments. Geological Survey of Canada Paper 72–27, 30 pp. McDougall, I. and Wensink, H. 1966. Paleomagnetism and geochronology of the Pliocene-Pleistocene lavas in Iceland. Earth and Planetary Science Letters, 1: 232–236. McKenzie, G. D. and Peterson, D. N. 1975. Subglacial cavitation phenomena, Discussion and reply to paper by Vivian, R. and Bocquet, G. Journal of Glaciology, 14: 339–341. McManus, J. and Duck, R. W. 1988. Localised enhanced sedimentation from icebergs in a proglacial lake in Briksdal, Norway. Geografiska Annaler, 70A: 215–223. McSaveney, M. J. 1975. The Sherman Glacier rock avalanche of 1964; its emplacement and subsequent effects on the glacier beneath it, Doctoral Thesis, Ohio State University, Columbus OH, p. 426. McSaveney, M. J. 1978. Sherman Glacier rock avalanche, Alaska, U.S.A. In: B. Voight (ed.) Rockslides and Rock Avalanches 1, Natural Phenomena. Elsevier, New York, pp. 197–258. McWilliams, M. O. and McElhinny, M. W. 1980. Late Precambrian paleomagnetism of Australia: the Adelaide geosyncline. Journal of Geology, 88: 1–26. Meier, M. F. 1989. (Abstr.) Relation between water input, basal water pressure, and sliding of Columbia Glacier, Alaska, U.S.A.. Annals of Glaciology, 12: 214–215. Meier, M. F. and Post, A. 1969. What are glacier surges? Canadian Journal of Earth Sciences, 6: 807–817. Meier, M. F. and Post, A. 1987. Fast tidewater glaciers. Journal of Geophysical Research, 92: 9051–9058. Melhorn, W. N. and Kempton, J. P. 1991. Geology and Hydrogeology of the Teays-Mahomet Bedrock Valley System. Geological Society of America, Special Paper 258, 128 pp. Menzies, J. 1981a. Freezing fronts and their possible influence upon processes of subglacial erosion and deposition. Annals of Glaciology, 2: 52–56. Menzies, J. 1981b. Investigations into the Quaternary deposits and bedrock topography of central Glasgow. Scottish Journal of Geology, 17: 155–168. Menzies, J. 1984. Drumlins: A Bibliography. Geobooks, Norwich, 117 pp.

Menzies, J. 1986. Inverse-graded units within till in drumlins near Caledonia, Southern Ontario.Canadian Journal of Earth Sciences, 23: 774–786. Menzies, J. 1987. Towards a general hypothesis on the formation of drumlins. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema, Rotterdam, pp. 9–24. Menzies, J. 1989a. Subglacial hydraulic conditions and their possible impact upon subglacial bed formation. Sedimentary Geology, 62: 125–150. Menzies, J. 1989b. Drumlins – products of controlled or uncontrolled glaciodynamic response? Quaternary Science Reviews, 8: 151–158. Menzies, J. 1990a. Brecciated diamictons from Mohawk Bay, S. Ontario, Canada. Sedimentology, 37: 481–493. Menzies, J. 1990b. Evidence of cryostatic desiccation processes associated with sand, intraclasts, within diamictons, southern Ontario, Canada. Canadian Journal of Earth Sciences, 27: 684–693. Menzies, J. 1990c. Sand intraclasts within a diamicton m´elange, southern Niagara Peninsula, Ontario, Canada. Journal of Quaternary Science, 5: 189–206. Menzies, J. (ed.) 1995. Modern Glacial Environments – Processes, Dynamics and Sediments. Butterworth-Heineman, Oxford. Menzies, J. (ed.) 1996. Past Glacial Environments – Sediments, Forms and Techniques. Butterworth-Heineman, Oxford. Menzies, J. 2000. Microstructures in diamictites of the Lower Gowganda Formation (Huronian), near Elliot Lake, Ontario: Evidence for deforming-bed conditions at the grounding line? Journal of Sedimentary Research, 70: 210–216. Menzies, J. and Maltman, A. J. 1992. Microstructures in diamictons – evidence of subglacial bed conditions. Geomorphology, 6: 27–40. Menzies, J. and Rose, J. (eds) 1987. Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, 360 pp. Menzies, J. and Rose, J. (eds) 1989. Subglacial bedforms – drumlins, Rogen moraine and associated subglacial bedforms. Sedimentary Geology, 62: 117–407. Menzies, J. and Shilts, W. W. 1996. Subglacial Environments. In J. Menzies (ed.) Past Glacial Environments – Sediments, Forms and Techniques. Volume II, Butterworth-Heineman, Oxford. Chapter 2, pp. 15–136. Menzies, J. and Woodward, J. 1993. Preliminary study of subglacial diamicton microstructures as reflected in drumlin sediments at Chimney Bluffs, New York. In: J. S. Aber (ed.) Glaciotectonics and Mapping Glacial Deposits. Proceedings of the INQUA Commission on Formation and Properties of Glacial Deposits, Volume 1.

REFERENCES

Canadian Plains Research Centre, University of Regina, Regina, pp. 36–45. Menzies, J., Zaniewski, K. and Dreger, D. 1997. Evidence, from microstructures, of deformable bed conditions within drumlins, Chimney Bluffs, New York State. Sedimentary Geology, 111: 161–175. Mercer, J. H. 1978. West Antarctic ice sheet and CO2 greenhouse effect: a threat of disaster. Nature, 271: 321–325. Meyer, K.-D. 1987. Ground and end moraines in Lower Saxony. In: J. J. M. van der Meer (ed.) Tills and Glaciotectonics. A. A. Balkema, Rotterdam, pp. 197–204. Miall, A. D. 1977. A review of the braided-river depositional environment. Earth Science Reviews, 13: 1–62. Miall, A. D. 1978. Lithofacies types and vertical profile models in braided river deposits: A summary. In: A. D. Miall (ed.) Fluvial Sedimentology. Canadian Society of Petroleum Geologists, Memoir 5, Calgary, Alberta, Canada, pp. 597–604. Miall, A. D. 1983. Glaciomarine sedimentation in the Gowganda Formation (Huronian), northern Ontario. Journal of Sedimentary Petrology, 53: 477–491. Michel, B. 1978. Ice Mechanics. Les Presses de l’Universit´e Laval, Quebec City, P.Q., 499 pp. Mickelson, D. M. and Attig, J. W. (eds) 1999. Glacial Processes Past and Present. Geological Socity of America, Special publication No. 337, 200 pp. Mickelson, D. M., Clayton, L. Fullerton, D. S. and Borns, Jr., H. W. 1983. The late Wisconsin glacial record of the Laurentide ice sheet in the United States. In: S. C. Porter (ed.) The Late Pleistocene. Late Quaternary Environments of the United States, Volume I. University of Minnesota, Minneapolis, pp. 3–37. Middleton, G. V. 1970. Experimental studies related to problems of flysch sedimentation. In: J. Lajoie (ed.) Flysch Sedimentology in North America. Geological Association of Canada Special Paper, 7, pp. 253–272. Middleton, G. V. 1973. Johannes Walther’s law of correlation of facies. Geological Society of America Bulletin, 84: 979–988. Midttun, L. 1985. Formation of dense bottom water in the Barents Sea. Deep Sea Research, 32: 1233–1241. Milankovitch, M. M. 1930. Mathematische Klimalehre und Astronomische Theorie der Klimaschwankungen. In: W. K¨oppen and R. Geiger (eds) Handbuch der Klimatologie, I, teil A. Borntraeger, Berlin. Milankovitch, M. M. 1941. Kanon der Erdbestrahlung und seine Anwendung auf das Eiszeitenproblem. Royal Serbian Academy, Special Publication, 133, 633 pp. Miller, G. H. and de Vernal, A. 1992 Will greenhouse

511

warming lead to Northern Hemisphere ice-sheet growth? Nature, 355: 244–246. Mills, H. H. 1977. Textural characteristics of drift from some representative Cordilleran glaciers. Geological Society of America, Bulletin, 88: 1135–1143. Mills, H. H. 1987. Morphometry of drumlins in the northeastern and north-central USA. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 131–148. Mitchell, G. F., Penny, L. F., Shotten, F. W. and West, R. G. 1973. A Correlation of Quaternary Deposits in the British Isles. Geological Society of London Special Report, 4, 99 pp. Mitchum, Jr., R. M., Vail, P. R. and Sangree, J. B. 1977. Seismic stratigraphy and global changes of sea level; part 6. Stratigraphic interpretation of seismic reflection patterns in depositional sequences. American Association of Petroleum Geologists, Memoir, 26: 117–133. Mix, A. C. 1987. The oxygen-isotope record of glaciation. In: W. F. Ruddiman and H. E. Wright Jr. (eds) North America and Adjacent Oceans During the Last Deglaciation. The Geology of North America, Volume K-3, Geological Society of America, Boulder, Colorado, pp. 111–125. Mix, A. C. and Ruddiman, W. R. 1984. Oxygen-isotope analyses and Pleistocene ice volumes. Quaternary Research, 21: 1–20. Molnia, B. F., Atwood, T. J., Carlson, P. R., Post, A. and Vath, S. C. 1984. Map of Marine Geology of Upper Muir and Wachusett Inlets, Glacier Bay, Alaska; Sediment Distribution and Thickness, Bathymetry, and Interpreted Seismic Profiles. Open-File Report – U.S. Geological Survey, OF 84–0632. Mooers, H. D. 1989a. On the formation of tunnel valleys of the Superior lobe, central Minnesota. Quaternary Research, 32: 24–35. Mooers, H. D. 1989b. Drumlin formation: a time transgressive model. Boreas, 18: 99–107. Moran, S. R. 1971. Glacitectonic structures in drift. In: R. P. Goldthwait (ed.): Till, a Symposium. Ohio State University Press, pp. 127–148. Moran, S. R., Clayton, L., Hooke, R. LeB., Fenton, M. M. and Andriashek, L. D. 1980. Glacier-bed landforms of the prairie region of North America. Journal of Glaciology, 25: 457–476. Morawski, W. 1988. Watermorainic sediments: origin and classification. In R. P. Goldthwait and C. L. Matsch (eds) Genetic Classification of Glacigenic Deposits. A. A. Balkema Publishers, Rotterdam, pp. 143–144. Morgan, A. V. 1973. The Pleistocene geology of the area north and west of Wolverhampton, Staffordshire,

512

REFERENCES

England. Philosophical Transactions of the Royal Society of London, 265B: 233–297. Morland, L. W. and Boulton, G. S., 1975. Stress in an elastic hump: the effects of glacier flow over elastic bedrock. Proceedings of the Royal Society of London, Series A, 344: 157–173. Morland, L. W. and Morris, E. M. 1977. Stress fields in an elastic bedrock hump. Journal of Glaciology, 18(78), 67–76. Morris, E. M. 1979. The flow of ice, treated as a Newtonian viscous liquid, around a cylindrical obstacle near the bed of a glacier. Journal of Glaciology, 23: 117–129. Morris, E. M . and Morland, L. W. 1977. A theoretical analysis of the formation of glacial flutes. Journal of Glaciology, 17: 311–323. Morris T. H., Clark, D. and Blasco, S. M. 1985. Sediments of the Lomonosov Ridge and Makarov Basin: A Pliocene stratigraphy for the North Pole. Geological Society of America Bulletin, 96: 901–910. Mullins, H. T., Eyles, N. and Hinchey, E. J. 1990. Seismic reflection investigation of Kalamalka Lake: a ‘fiord lake’ on the Interior Plateau of southern British Columbia. Canadian Journal of Earth Sciences, 27: 1225–1235. Murray, T. 1994. Glacial deformation. In A. E. Maltman (ed.) The Deformation of Sediments. Chapman and Hall, pp. 74–93. Murray, T. and Dowdeswell, J. A. 1992. Water throughflow and the physical effects of deformation on sedimentary glacier beds. Journal of Geophysical Research, 97B: 8993–9002. Mustard, P. S. and Donaldson, J. A. 1987. Early Proterozoic ice-proximal deposition: The lower Gowganda Formation at Cobalt, Ontario, Canada. Geological Society of America Bulletin, 98: 373–387. Myhre, A. M. and Eldholm, O. 1988. The western Svalbard margin (74°–80°N). Marine and Petroleum Geology, 5: 134–156.

climates and plate motions inferred from major element chemistry of lutites. Nature, 299: 715–717. Nesje, A. and Whillans, I. M. 1994. Erosion of Sognefjord, Norway. Geomorphology, 9: 33–45. Newitt, D. M., Richardson, J. F., Abbott, M. and Turtle, R. B. 1955. Hydraulic conveying of solids in horizontal pipes. Transactions of the Institute of Chemical Engineers, 33: 93–110. Nickling, W. G. and Bennett, L. 1984. The shear strength characteristics of frozen coarse granular debris. Journal of Glaciology, 30: 348–357. Nilsson, T. 1983. The Pleistocene; Geology and Life in the Quaternary Ice Age. D. Reidel, Dordrecht, 651 pp. Nishiizumi, K., Elmore, D. and Kubik, P. W. 1989a. Update on terrestrial ages of Antarctic Meteorites. Earth and Planetary Science Letters, 93: 299–313. North American Commission on Stratigraphic Nomenclature (NACSN) 1983. North American Stratigraphic Code. American Association of Petroleum Geologists Bulletin, 67: 841–875. Nummedal, D., Hine, A. C. and Boothroyd, J. C. 1987. Holocene evolution of the south-central coast of Iceland. In: D. M. Fitzgerald and P. S. Rosen (eds) Glaciated Coasts. Academic Press, San Diego, pp. 115–150. Nye, J. F. 1953. The flow rate of ice from measurements in glacier tunnels, laboratory experiments, and the Junfraufirn borehole experiment. Proceedings of Royal Society, Series A, 219: 477–489. Nye, J. F. 1969. A calculation on the sliding of ice over a wavy surface using a Newtonian viscous approximation. Proceedings of the Royal Society of London, Series A, 311: 445–467. Nye, J. F. 1976. Water flow in glaciers: j¨okulhlaups, tunnels and veins. Journal of Glaciology, 17: 181–207. Nye, J. F. and Martin, P. C. S. 1968. Glacial erosion. International Association of Scientific Hydrology Publication, 79: 78–86.

Nakawo, M. and Takahashi, S. 1983. A simplified model for estimating glacier ablation under a debris layer. In: J. W. Glen (ed.) Hydrological Aspects of Alpine and Highmountain Areas. International Association of Hydrological Sciences, Publication 138, pp. 137–145. Nemec, W. 1990. Aspects of sediment movement on steep delta slopes. In: A. Colella, and D. B. Prior (eds) CoarseGrained Deltas. International Association of Sedimentologists, Special Publication 10, pp. 29–73. Nemec, W. and Steel, R. J. 1988. Fan Deltas: Sedimentology and Tectonic Settings: 1987, Bergen, Norway. Blackie and Son, Glasgow, 444 pp. Nesbitt, H. W. and Young, G. M. 1982. Early Proterozoic

O’Brien, D. K., Wenk, H.-R., Ratschbacher, L. and You, Z. 1987. Preferred orientation of phyllosilicates in phyllonites and ultramylonites. Journal of Structural Geology, 9: 719–730. O’Connor, J. E. and Baker, V. R. 1992. Magnitudes and implications of peak discharges from glacial Lake Missoula. Geological Society of America Bulletin, 104: 267–279. Oerlemans, J. 1989. On the response of valley glaciers to climatic change. In: J. Oerlemans (ed.), Glacier Fluctuations and Climatic Change. Kluwer Publications, Dordrecht, pp. 353–371.

REFERENCES

Oerlemans, J. 1991. The role of ice sheets in the Pleistocene climate. Norsk Geologisk Tidsskrift, 71: 155–161. Oescheger, H., Beer, J., Siegenthaler, U., Stauffer, B., Dansgaard, W. and Langway, C. C. 1984. Late glacial climate history from ice cores. In J. E. Hansen and T. Takahashi (eds) Climate Processes and Climate Sensitivity. American Geophysical Union, Geophysical Monograph No.29, Washington DC, pp. 299–306. Ørheim, O. 1985. Iceberg discharge and the mass balance of Antarctica. In: Glaciers, Ice Sheets and Sea Level: Effect of CO2-induced Climatic Change. Report of a workshop held in Seattle, Washington, September 1984. United States Department of Energy, Washington DC, pp. 210–215. Østrem, G. 1959. Ice melting under a thin layer of moraine, and the existence of ice cores in moraine ridges. Geografiska Annaler, 41A: 228–230. Østrem, G. 1975. Sediment transport in glacial meltwater streams. In: A. V. Jopling and B. C. McDonald (eds) Glaciofluvial and Glaciolacustrine Sedimentation. Society of Economic Paleontologists and Mineralogists, Special Publication, 23, pp. 101–122. Østrem, G. and Brugman, M. M. 1991. Glacier MassBalance Measurements: a Manual for Field Work. NHRI Science Report No.4, Ministry of Supply and Services Canada. 224 pp. Østrem, G. and Olsen, H. C. 1987. Sedimentation in a glacier lake. Geografiska Annaler, 69A: 123–138. Østrem, G., Bridge, C. W. and Rannie, W. F. 1967. Glaciohydrology, discharge and sediment transport in the Decade Glacier area, Baffin Island, N.W.T. Geografiska Annaler, 49A: 268–282. Owen, L. A. and Derbyshire, E. 1988. Glacially deformed diamictons in the Karakoram Mountains, northern Pakistan. In: D. G. Croot (ed.) Glaciotectonics: Forms and Processes. A. A. Balkema Publishers, Rotterdam, pp. 149–176. Page, N. J. 1981. The Precambrian diamictite below the base of the Stillwater Complex, Montana. In: M. J. Hambrey and W. B. Harland (eds) Earth’s pre-Pleistocene Glacial Record. Cambridge University Press, Cambridge, pp. 821–825. Palmer, A. C. 1972. A kinematic wave model of glacier surges. Journal of Glaciology, 11: 65–72. Palmer, A. C., Goodman, D. J., Ashby, M. F., Evans, A. G., Hutchinson, J. W. and Ponter, A. R. S. 1983. Fracture and its role in determining ice forces on offshore structures. Annals of Glaciology, 4: 216–221. Pardee, J. T. 1942. Unusual currents in glacial Lake Missoula, Montana. Geological Society of America Bulletin, 53: 1569–1599.

513

Parizek, E. J. 1959. Description and origin of stone lines (carpedoliths) near the Kansan drift border in northeast Kansas and northwest Missouri. Geological Society of America Bulletin, 70: 1654–1655. Parkin, G. W. and Hicock, S. R. 1988. Sedimentology of a Pleistocene glacigenic diamicton sequence near Campbell River, Vancouver Island, British Columbia. In: R. P. Goldthwait and C. L. Matsch (eds) Genetic Classification of Glacigenic Deposits. A. A. Balkema Publishers, Rotterdam, pp. 97–116. Pasini, G. and Colalongo, M. L. 1997. The Pliocene– Pleistocene boundary-stratotype at Vrica, Italy. In: J. A. Van Couvering (ed.) The Pleistocene Boundary and the beginning of the Quaternary. Cambridge University Press, Chapter 2, pp. 15–45. Passchier, C. W. and Trouw, R. A. J. 1996. Microtectonics. Springer, Berlin, 289 pp. Paterson, U., Nitsan, Yu., Klark, D. 1978. Neustoychivost’pri polzuchesti kak mekhanizm serdzhey lednikov – An investigation of creep instability as a mechanism for glacial surges, Materialy Glyatsiologicheskikh Issledovaniy. Khronika Obsuzhdeniya, 32: 82–92, 201–209. Paterson, W. S. B. 1980. Ice sheets and ice shelves. In: S. C. Colbeck (ed.) Dynamics of Snow and Ice Masses. Academic Press, New York, pp. 1–78. Paterson, W. S. B. 1981. Physics of Glaciers. Elsevier Science, Oxford. Paterson, W. S. B. 1994. The Physics of Glaciers. 3rd Edition, Pergamon Press, Oxford, 480 pp. Paterson, W. S. B. and Hammer, C. U. 1987. Ice core and other glaciological data. In: W. F. Ruddiman and H. E. Wright, Jr. (eds), North America and Adjacent Oceans During the Last Deglaciation. The Geology of North America, Volume K-3. Geological Society of America, Boulder, Colorado, pp. 91–109. Paterson, W. S. B. and Savage, J. C. 1963. Measurements on Athabasca Glacier relating to the flow law of ice. Journal of Geophysics Research, 68: 4537–4543. Paul, M. A. and Evans, H. 1974. Observations on the internal structure and origin of some flutes in glaciofluvial sediments, Blomstrandbreen, north-west Spitsbergen. Journal of Glaciology, 13: 393–400. Paul, M. A. and Eyles, N. 1990. Constraints on the preservation of diamict facies (melt-out tills) at the margins of stagnant glaciers. Quaternary Science Reviews, 9: 51–69. Pecker, J.-C. and Runcorn, S. K. 1990. The Earth’s climate and variability of the sun over recent millennia: geophysical, astronomical and archaeological aspects. Philosophical Transactions of the Royal Society of London, 330(A): 395–687.

514

REFERENCES

Penck, A. and Br¨uckner, E. 1909. Die Alpen im Eiszeitalter. Tauchnitz, Leipzig, Bd. 1–3, 1199 pp. Penny, L. F., Coope, G. R. and Catt, J. A. 1969. Age and insect fauna of the Dimlington Silts, East Yorkshire. Nature, 224: 65–67. Perrin, R. M. S., Rose, J. and Davies, H. 1979. The distribution, variation and origins of Pre-Devensian tills in easter England. Philosophical Transactions of the Royal Society of London, 287B: 535–570. Peteet, D. (ed.) 1993. Global Younger Dryas? Quaternary Science Reviews, 12: 277–355. Petit, J. P. 1987. Criteria for the sense of movement on fault surfaces in brittle rocks. Journal of Structural Geology, 9: 597–608. Petit, J. R., Jouzel, J. Raymaud, D., Barkov, N. I., Barnol, J.-M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, G., Delmotte, M., Kotlyakov, V. M., Legrand, M., Lipenkov, V. Y., Lorius, C., P´epin, L., Ritz, C., Saltzman, E. and Stievenard, M. 1999. Climate and atmospheric history of the past 420,000 years of climate and atmospheric history form the Vostok ice core, Antartica. Nature, 399: 429–436. Philip, J. R. 1980. Thermal fields during regelation. Cold Regions Science and Technology, 3: 193–203. Phillips, A. C., Smith, N. D. and Powell, R. D. 1991. Laminated sediments in prodeltaic deposits, Glacier Bay, Alaska. In: J. B. Anderson and G. M. Ashley (eds) Glacial Marine Sedimentation: Paleoclimatic Significance. Geological Society of America, Special Paper, 261, pp. 51–60. Pickrill, R. A. and Irwin, J. 1982. Predominant headwater inflow and its control of lake-river interactions in Lake Wakatipu. New Zealand Journal of Freshwater Research, 16: 201–213. Pickrill, R. A. and Irwin, J. 1983. Sedimentation in a deep glacier-fed lake – Lake Tekapo, New Zealand. Sedimentology, 30: 63–75. Piotrowski, J. A. 1994. Tunnel-valley formation in northwest Germany – geology, mechanisms of formation and subglacial bed conditions for the Bornnh¨oved tunnel valley. Sedimentary Geology, 89: 107–141. Piotrowski, J. A. and Smalley, I. J. 1987. The Woodstock drumlin field, southern Ontario, Canada. In J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 309–322. Platt, J. P. 1984. Secondary cleavages in ductile shear zones. Journal of Structural Geology, 6: 439–442. Platt, J. P. 1986. Dynamics of orogenic wedges and the uplift of high-pressure metamorphic rocks. Geological Society of America, Bulletin, 97: 1037–1053. Pons, A., Guiot, J., De Beaulieu, J. L. and Reille, M. 1992.

Recent contributions to the climatology of the last Glacial–Interglacial Cycle based on French pollen sequences. Quaternary Science Reviews, 11: 439–448. Porter, S. C. and Orombelli, G. 1981. Alpine rockfall hazards. American Scientist, 69: 67–75. Post, A. S. 1965, Alaskan Glaciers; recent observations in respect to the earthquake – advance theory, Science, 148: 366–368. Postma, G., Roep, T. B. and Ruegg, G. H. J. 1983. Sandygravelly mass-flow deposits in an ice marginal lake (Saalian, Leuvenumsche Beek Valley, Veluwe, The Netherlands), with emphasis on plug-flow deposits. Sedimentary Geology, 34: 59–82. Potter, J. R., Paren, J. G. and Loynes, J. 1984. Glaciological and oceanographic calculations of the mass balance and oxygen isotope ratio of a melting ice shelf. Journal of Glaciology, 30: 161–170. Powell, R. D. 1980. Holocene glacimarine sediment deposition by tidewater glaciers in Glacier Bay, Alaska. Unpublished Ph.D. Thesis, Ohio State University, Columbus, Ohio. Powell, R. D. 1983. Submarine flow tills at Victoria, British Columbia: discussion. Canadian Journal of Earth Sciences, 20: 509–510. Powell, R. D. 1984. Glacimarine processes and inductive lithofacies modelling of ice shelf and tidewater glacier sediments based on Quaternary examples. Marine Geology, 57: 1–52. Powell, R. D. 1990. Glacimarine processes at groundingline fans and their growth to ice-contact deltas. In: J. A. Dowdeswell and J. D. Scourse (eds) Glacimarine Environments: Processes and Sediments. Geological Society of London, Special Publication No.53, pp. 53–73. Powell, R. D. 1991. Grounding-line systems as secondorder controls on fluctuations of tidewater termini of temperate glaciers. In: J. B. Anderson and G. M. Ashley (eds) Glacial Marine Sedimentation; Palaeoclimatic Significance. Geological Society of America, Special Paper, 261, pp. 75–93. Powell, R. D. and Molnia, B. F. 1989. Glacimarine sedimentary processes, facies and morphology of the south-southeast Alaska shelf and fjords. Marine Geology, 85: 359–390. Powell, R.D., Pyne, A.R., Hunter, L.E. and Rynes, N.R. 1992. Polar versus temperate grounding-line sedimentary systems and marine glacier stability during sea level rise by global warming. Geological Society of America, Abstract with Programs, 24: 348. Prange, W. 1978. Der letzte weichselzeitliche Gletschervorstoß in Schleswig-Holstein – das Gef¨uge u¨ berfahrener

REFERENCES

Schmelzwassersande und die Entstehung der Morphologie. Meyniana, 30: 61–75. Prell, W. L., Imbrie, J., Martinson, D. G., Morley, J. J., Pisias, N. G., Shackleton, N. J. and Streeter, H. F. 1986. Graphic correlation of oxygen isotope stratigraphy application to the late Quaternary. Paleoceanography, 1: 137–162. Prentice, M. L., Denton, G. H., Lowell, T. V., Conway, H. C. and Heusser, L. E. 1986. Pre-Late Quaternary glaciation of the Beardmore Glacier region, Antarctica. Antarctic Journal of the United States, 21: 95–98. Prest, V. K. 1968. Nomenclature of moraines and ice-flow features as applied to the glacial map of Canada. Geological Survey of Canada, Monograph 67–57, p. 32. Prest, V. K., Grant, D. R. and Rampton, V. N. 1969. Glacial Map of Canada. Geological Survey of Canada Map 1253-A. Price, R. A. 1973. Large-scale gravitational flow of supracrustal rocks, Southern Canadian Rockies. In: K. A. De Jong and R. Scholten (eds) Gravity and Tectonics. John Wiley and Sons, New York, pp. 491–502. Punkari, M. 1984. The relations between glacial dynamics and tills in the eastern parts of the Baltic Shield. In: L.-K. K¨onigsson (ed.) Ten years of Nordic Till research. Striae, 20: 49–54. Quigley, R. M. 1983. Glaciolacustrine and glaciomarine clay deposition; a North American perspective. In: N. Eyles (ed.), Glacial Geology. Pergamon Press, Oxford, pp. 140–167. Quilty, P. G. 1985. Distribution of Foraminiferids in Sediments of Prydz Bay, Antarctica. South Australia Department of Mines and Energy, Special Publication, No. 5, pp. 329–340. Rabassa, J. 1987. Drumlins and drumlinoid forms in northern James Ross island, Antarctic Peninsula. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema Publishers, Rotterdam, pp. 267–290. Rabassa, J. and Clapperton, C. M. 1990. Quaternary glaciations in the southern Andes. In: C. M. Clapperton (ed.), Quaternary Glaciations in the Southern Hemisphere. Quaternary Science Reviews, 9: 153–174. Rabinowicz, E. 1965. Friction and Wear of Materials. John Wiley, New York, 244 pp. Radok, U., Barry, R. G., Jensen, D., Keen, R. A., Kiladis, G. N. and McInnes, B. J. 1982. Climatic and Physical Characteristics of the Greenland Ice Sheet, Parts I and II. Cooperative Institute for Research in Environmental

515

Sciences. University of Colorado, Boulder, 193 pp. and 4 appendices. Rains, B., Shaw, J., Skoye, R., Sjogren, D. and Kvill, D. 1993. Late Wisconsin subglacial megaflood paths in Alberta. Geology, 21: 323–326. Ramm, M. 1989. Late Quaternary carbonate sedimentation and paleoceanography in the eastern Norwegian Sea. Boreas, 18: 255–272. Ramsay, A. C. 1852. On the superficial accumulations and surface markings of North Wales. Quarterly Journal of the Geological Society, 8: 371–376. Ramsay, A. C. 1855. On the occurrence of angular, subangular, polished and striated fragments and boulders in the Permian breccia of Shropshire, Worcestershire, &c. Quarterly Journal of the Geological Society, 11: 185–205. Ramsay, J. G. and Huber, M. I. 1983. The Techniques of Modern Structural Geology. Volume 1, Academic Press, London. 307 pp. Ramsay, J. G. and Huber, M. I. 1987. The Techniques of Modern Structural Geology. Volume 2, Academic Press, London. 700 pp. Rappol, M. 1985. Clast-fabric strength in tills and debris flows compared for different environments. Geologie en Mijnbouw, 64: 327–332. Rappol, M. 1987. Saalian till in The Netherlands: a review. In: J. J. M. van der Meer (ed.) Tills and Glaciotectonics. A. A. Balkema Publishers, Rotterdam, pp. 3–21. Rappol, M. 1993. Ice Flow and Glacial Transport in Lower St. Lawrence, Quebec. Geological Survey of Canada Paper 90–19, 28 pp. Rappol, M. and Stoltenberg, H. M. P. 1985. Compositional variability of Saalian till in The Netherlands and its origin. Boreas, 14: 33–50. Rasmussen, L. Å. 1975. Kineto-stratigraphic glacial drift units on Hindsholm, Denmark. Boreas, 4: 209–217. Rastas, J. and Sepp¨al¨a, M. 1981. Rock jointing and abrasion forms on rouches moutonn´ees, SW Finland. Annals of Glaciology, 2: 159–163. Ray, L. R. 1935. Some minor features of valley glaciers and valley glaciation. Journal of Geology, 43: 297–322. Raymo, M. E. 1991. Geochemical evidence supporting T.C. Chamberlin’s theory of glaciation. Geology, 19: 344–347. Raymo, M. E., Ruddiman, W. F., Backman, J., Clement, B. M. and Martinson, D. G. 1989. Late Pliocene variation in Northern Hemisphere ice sheets and North Atlantic deep water circulation. Paleoceanography, 4: 413–446. Raymond, C. F. 1973. Inversion of flow measurements for stress and rheological parameters in a valley glacier. Journal of Glaciology, 12: 19–44.

516

REFERENCES

Raymond, C. F. 1978. Numerical calculation of glacier flow by finite element methods, final technical report for National Science Foundation Grant DPP74–19075, Unpublished Manuscript, 1978. Raymond, C. F. 1980. Temperate valley glaciers. In: S. C. Colbeck (ed.), Dynamics of Snow and Ice Masses. Academic Press, New York, pp. 79–139. Raymond, C. F. 1987. How do glaciers surge? A review. Journal of Geophysical Research, 92: 9121–9134. Raymond, C. F. and Harrison, W. D. 1988. Evolution of Variegated Glacier, Alaska, U.S.A., prior to its surge. Journal of Glaciology, 34: 154–169. Raynaud, D., Jouzel, J., Barnola, J. M., Chappellaz, J., Delmas, R. J. and Lorius, J. 1993. The ice record of greenhouse gases. Science, 259: 926–933. Rea, D. K. and Leinen, M. 1989. Quaternary Climates: The Ocean Sedimentary Record. Geological Society of America, Short Course, Annual Meeting, St. Louis, 245 pp. Reading, H. G. (ed) 1986. Sedimentary Environments and Facies. 2nd edition. Blackwell Scientific Publishers, Oxford, 557 pp. Reboul, H. 1833. G´eologie de la p´eriode quaternaire, et introduction ‘a l’histoire ancienne. Levrault, Paris, 222 pp. Reeh, N. 1968. On the calving of ice from floating glaciers and ice shelves. Journal of Glaciology, 7: 215–232. Reheis, M. J. 1975. Source, transportation and deposition of debris on Arapaho Glacier, Front Range, Colorado, U.S.A. Journal of Glaciology, 14: 407–420. Reid, C. 1882. The Geology of the Country around Cromer. Memoir of the Geological Survey, Great Britain. Reimnitz, E. and Kempema, E. W. 1988. Ice rafting: an indication of glaciation? Journal of Glaciology, 34: 254–255. Renberg, I. and Segerstrom, U. 1981. Application of varved lake sediments in paleoenvironmental studies. Wahlenbergia, 7: 125–133. Repo, R. 1954. Om forhallandet mellan rafflor och asar. Geologi, 6, 45 pp. Reynaud, L. 1987. The November 1986 survey of the Grand Moulin on the Mer de Glace, Mont Blanc massif, France. Journal of Glaciology, 33: 130–131. Richmond, G. M. and Fullerton, D. S. 1986a. Introduction to Quaternary glaciations in the United States of America. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 3–10. Richmond, G. M. and Fullerton, D. S. 1986b. Summation of Quaternary glaciations in the United States of America. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds),

Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 183–196. Richter, W., Schneider, H. and Wager, R. 1951. Die saaleeiszeitliche Stauchzone von Itterbeck-Uelsen (Grafschaft Bentheim). Zeitschrift der Deutschen Geologischen Gesellschaft, 102: 60–75. Ridge, J. C. and Larsen, F. D. 1990. Re-evaluation of Antevs’ New England varve chronology and new radiocarbon dates of sediments from glacial Lake Hitchcock. Geological Society of America, Bulletin, 102: 889–899. Ridge, J. C., Brennan, W. J. and Muller, E. H. 1990. The use of paleomagnetic declination to test correlations of late Wisconsinan glaciolacustrine sediments in central New York. Geological Society of America, Bulletin, 102: 26–44. Ridky, R. W. and Bindschadler, R. A. 1990. Reconstruction and dynamics of the Late Wisconsin ‘Ontario’ ice dome in the Finger Lakes region, New York. Geological Society of America, Bulletin, 102: 1055–1064. Riedel, W. 1929. Zur Mechanik geologischer Brucherscheinungen, ein Beitrag zum Problem der Fiederspalten. Zentralblatt Mineralogischer, Geologischer und Pala¨ ontologischer Abteilung v B: 354–368. Riley, N. W. 1982. Rock wear by sliding ice. Ph.D. thesis, University of Newcastle, UK. Rind, D. and Peteet, D. 1985. Terrestrial conditions at the last glacial maximum and CLIMAP sea-surface temperature estimates: Are they consistent? Quaternary Research, 24: 1–22. Ringberg, B., Holland, B. and Miller, U. 1984. Till stratigraphy and provenance of the glacial chalk rafts at ¨ Kvangby and Angdala, southern Sweden. Striae, 20: 79–90. Rist, S. 1955. Skeidar´arhlaup 1954. J¨okull, 5: 30–36. Rist, S., 1968. J¨okulhlaups from the cover of Myrdalsj¨okull on June 25, 1955 and January 20, 1956. J¨okull, 18: 243–248. Rist, S. 1983. Floods and flood danger in Iceland. J¨okull, 33: 119–132. Robe, R. Q. 1980. Iceberg drift and deterioration. In: S. C. Colbeck (ed.), Dynamics of Snow and Ice Masses. New York, Academic Press, pp. 211–259. Roberts, D. H. and Hart, J. K. 2000a. The glacial deposits of West Runton. In: S. G. Lewis, C. A. Whiteman and R. C. Preece (eds) The Quaternary of Norfolk and Suffolk Field Guide.. Quaternary Research Association, London, pp. 77–84. Roberts, D. H. and Hart, J. K. 2000b. The glacial deposits of Trimingham. In S. G. Lewis, C. A. Whiteman and R. C. Preece (eds) The Quaternary of Norfolk and Suffolk Field

REFERENCES

Guide. Quaternary Research Association, London, pp. 103–110. Roberts, M. C. and Rood, K. M. 1984. The role of the ice contributing area, the morphology of transvwerse fjords, British Columbia. Geografiska Annaler, 66A: 381–393. Robin, G. deQ. 1955. Ice movement and temperature distribution in glaciers and ice sheets, Journal of Glaciology, 2:523–532. Robin, G. deQ. 1969. Initiation of glacier surges, Canadian Journal of Earth Sciences, 6: 919–928. Robin, G. deQ. 1975. Ice shelves and ice flow. Nature, 253: 168–172. Robin, G. deQ. 1976. Is the basal ice of a temperate glacier at the pressure melting point? Journal of Glaciology, 16: 183–196. Robin, G. deQ. 1983. The Climatic Record In Polar Ice Sheets. Cambridge University Press, Cambridge, 212 pp. Robin, G. deQ. 1988. The Antarctic ice sheet, its history and response to sea level and climatic changes over the past 100 million years. Palaeogeography, Palaeoclimatology, Palaeoecology, 67: 31–50. Robin, G. deQ. and Weertman, J. 1973. Cyclic surging of glaciers. Journal of Glaciology, 12:3–18. Rocha-Campos, A. C. and Dos Santos, P. R. 1981. The Itarar‘e Subgroup, Aquidauana Group and San Greg´orio Formation, Paran´a basin, southeastern South America. In: M. J. Hambrey and W. B. Harland (eds) Earth’s prePleistocene Glacial Record. Cambridge University Press, Cambridge, pp. 842–852. Rodhe, L. 1988. Glaciofluvial channels formed prior tothe last deglaciation: examples from Swedish Lapland. Boreas, 17: 511–516. Ronnert, L. and Mickelson, D. M. 1992. High porosity of basal till at Burroughs glacier, southeastern Alaska. Geology, 20: 849–852. Rose, J. 1974. Small scale spatial variability of some sedimentary properties of lodgement and slumped till. Proceedings of the Geologists’ Association, 85: 223–237. Rose, J. 1980. Field Trip discussion. In Jardine, W. G. (ed) Glasgow Region. Quaternary Research Association Field Guide, 22–39 and 44–51. Rose, J. 1981. Field guide to the Quaternary geology of the Loch Lomond basin. Proceedings of the Geological Society of Glasgow, 1980/81: 3–19. Rose, J. 1985. The Dimlington Stadial/ Dimlington Chronozone: a proposal for naming the main glacial episode of the Late Devensian in Britain. Boreas, 14: 225–230. Rose, J. 1987a. Drumlins as part of a glacier bedform continuum. In: J. Menzies and J. Rose (eds) Drumlin

517

Symposium. A. A. Balkema Publishers, Rotterdam, pp. 103–116. Rose, J. 1987b. Status of the Wolstonian Glaciation in the British Quaternary. Quaternary Newsletter, 53: 1–9. Rose, J. 1989a. Stadial type sections in the British Quaternary. In: J. Rose and Ch. Schl¨uchter (eds) Quaternary Type Sections: Imagination or Reality? A. A. Balkema Publishers, Rotterdam, pp. 45–67. Rose, J. 1989b. Glacier stress patterns and sediment transfer associated with the formation of superimposed flutes. Sedimentary Geology, 62: 151–176. Rose, J. 1991. Stratigraphic basis of the ‘Wolstonian Glaciation’, and retention of the term ‘Wolstonian’ as a chronostratigraphic Stage name – a discussion. In: S. G. Lewis, C. A. Whiteman and D. R. Bridgland (eds) Central East Anglian and the Fen Basin, Field Guide. Quaternary Research Association, London, pp. 15–20. Rose, J. and Allen, P. 1977. Middle Pleistocene stratigraphy in south-east Suffolk. Journal Geological Society London, 133: 83–102. Rose, J. and Letzer, J. M. 1977. Superimposed drumlins. Journal of Glaciology, 18: 471–480. Rose, J. and Schl¨uchter, Ch. (eds) 1989. Quaternary Type Sections: Imagination or Reality? A. A. Balkema Publishers, Rotterdam, 208 pp. Rose, J., Allen, P., Kemp, R. A., Whiteman, C. A. and Owen, N. 1985. The early Anglian Barham Soil of Eastern England. In: J. Boardman (ed.) Soils and Quaternary Landscape Evolution. John Wiley & Sons, Chichester, pp. 197–229. Rose, J., Lowe J. J. and Switzer, R. 1988. A radiocarbon date on plant detritus beneath till from the type area of the Loch Lomond Readvance. Scottish Journal of Geology, 24: 113–124. Rose, J., Whiteman, C. A., Allen, P. and Kemp, R. A., 1999. The Kesgrave sands and gravels; ‘pre-glacial’ Quaternary deposits of the River Thames in East Anglia and the Thames Valley, Proceedings of the Geologists’ Association, 110: 93–116. R¨othlisberger, H. 1968. Erosive processes which are likely to accentuate or reduce the bottom relief of valley glaciers. International Association of Scientific Hydrology Publication, 79: 87–97. R¨othlisberger, H. 1972. Water pressure in intra- and subglacial channels. Journal of Glaciology, 11: 177–203. R¨othlisberger, H. and Iken, A. 1981. Plucking as an effect of water-pressure variations at the glacier bed. Annals of Glaciology, 2: 57–62. R¨othlisberger, H. and Lang, H. 1987. Glacial hydrology. In: A. M. Gurnell and M. J. Clark (eds) Glacio-Fluvial

518

REFERENCES

Sediment Transfer – an Alpine Perspective. J.Wiley Publishing, Chichester, pp. 207–284. Rotnicki, K. 1976. The theoretical basis for and a model of the origin of glaciotectonic deformations. Quaestiones Geographicae, 3: 103–139. Rotnicki, K. 1991. Retrodiction of palaeodischarges of meandering and sinuous alluvial rivers and its palaeohydroclimatic implications. In: L. Starkel, K. J. Gregory and J. B. Thornes (eds) Temperate Palaeohydrology. John Wiley, New York, pp. 431–471. Rototayev, K. P. 1983. Dynamics of the Byrs Glacier for the last 20 years. Mat. Glaytsiol. Issled. Khron. Obsuzhdeniya, 47: 147–156. Ruddiman, W. F. 1987. Northern oceans. In: W. F. Ruddiman and H. E. Wright Jr. (eds) North America and Adjacent Oceans During the Last Deglaciation. The Geology of North America, Volume K-3. Geological Society of America, Boulder, Colorado. pp. 137-–54. Ruddiman, W. F. and Raymo, M. E. 1988. Northern Hemisphere climate r´egimes during the past 3 Ma: possible tectonic connections. In: N. J. Shackleton, R. G. West and D. Q. Bowen (eds) The Past Three Million Years: Evolution of Climatic Variability in the North Atlantic Region. Philosophical Transactions Royal Society of London, B 318: 411–430. Ruddiman, W. F. and Wright, H. E., Jr. (eds) 1987. North America and Adjacent Oceans During the Last Deglaciation. The Geology of North America, Volume K-3, Geological Society of America, Boulder CO. Ruddiman, W. F., Raymo, M. E., Martinson, D. G., Clement, B. M. and Backman, J. 1989. Pleistocene evolution: Northern Hemisphere ice sheets and North Atlantic Ocean. Paleoceanography, 4: 353–412. Rundle, A. 1985. Braid morphology and the formation of multiple channels, the Rakaia, New Zealand. Zeitschrift f¨ur Geomorphologie, N.F., Supplement-Band, 55: 15–37. Russell, A. J. and Knudsen, O. 1999. An ice-contact rhythmite (turbidite) succession deposited during the November 1996 catastrophic outburst flood (j¨okulhlaup), Skeidararj¨okull, Iceland. Sedimentary Geology, 127: 1–10. Russell, A. J., Aitken, J. F. and De Jong, C. 1990. Observations on the drainage of an ice-dammed lake in West Greenland. Journal of Glaciology, 36: 72–74. Russell, H. A. J. et al. 1998. Sediment Thickness and the Greater Toronto and Oak Ridges Moraine Areas, Southern Ontario. Geological Survey of Canada, Open File 2892. Rust, B. R. 1972. Structure and process in a braided river. Sedimentology, 18: 221–245. Rust, B. R. 1977. Mass flow deposits in a Quaternary

succession near Ottawa, Canada: diagnostic criteria for subaqueous outwash. Canadian Journal of Earth Sciences, 14: 175–184. Rust, B. R. 1978. Depositional models for braided alluvium. In: A. D. Miall (ed.) Fluvial Sedimentology. Canadian Society of Petroleum Geologist, Memoir 5, Calgary, Alberta, Canada, pp. 605–625. Rust, B. R. and Koster, E. H. 1984. Coarse alluvial deposits. In: R. G. Walker (ed.) Facies Models, Second Edition. Geoscience Canada, Reprint Series 1, Toronto, pp. 53–69. Ruszczy´nska-Szenajch, H. 1976. Glacitectonic depressions and glacial rafts in mid-eastern Poland. Studia Geologica Polonica, 50: 1–106. Ruszczy´nska-Szenajch, H. 1980. Glacial erosion in contradistinction to glacial tectonics. In: W. Stankowski (ed.) Tills and Glacigene Deposits. Wydawnictwa Uniwersytetu Im. A. Mickiewicza, Pozna˜n, Seria Geografia, No. 20, pp. 71–76. Ruszczy´nska-Szenajch, H. 1987. The origin of glacial rafts: detachment, transport, deposition. Boreas, 16: 101–112. Rzechowski, J. 1986. Pleistocene till stratigraphy in Poland. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 365–372. Sakshaug, E. and Holm-Hansen, O. 1984. Factors governing pelagic production in polar oceans. In: O. HolmHansen, L. Bolis and R. Gilles (eds) Marine Phytoplankton and Productivity. Lecture notes on coastal and estuarine studies, 8. Springer, New York, pp. 1–17. Salvador, A. 1994. International Stratigraphic Guide, Second Edition. Geological Society of America, Boulder, 214 pp. Samtleben, C. and Bickert, T. 1990. Coccoliths in sediment traps from the Norwegian Sea. Marine Micropaleontology, 16: 39–64. Sanderson, T. J. O. 1979. Equilibrium profile of ice shelves. Journal of Glaciology, 22: 435–460. Sandgren, P., Bj¨orck, S., Brunnberg, L. and Kristiansson, J. 1988. Palaeomagnetic records from two varved clay sequences in the Middle Swedish ice marginal zone. Boreas, 17: 215–227. Sarmiento, J. L. 1993. Atmospheric CO2 stalled. Nature, 365: 697–698. Sarnthein, M., Winn, K. and Zahn, R. 1987. Paleoproductivity of ocean upwelling and the effect on atmospheric CO2 and climatic change during deglaciation times. In: W. H. Berger and L. D. Labeyrie (eds) Abrupt Climatic Change; Evidence and Implications. D. Reidel, Dordrecht, pp. 311–337.

REFERENCES

Sarnthein, M., Erlenkeuser, H, von Grafstein, R. and Schroeder, C. 1984. Stable-isotope stratigraphy for the last 750,000 years: ‘Meteor’ core 13519 from eastern equatorial Atlantic. Meteor Forschung-Ergembnisse C., 38: 9–24. Saunderson, H. C. 1982. Bed form diagrams and the interpretation of eskers. In: R. Davidson-Arnott, W. Nickling and B. D. Fahey (eds) Research in Glacial, Glacio-fluvial, and Glacio-lacustrine Systems. Proceedings of the 6th Guelph Symposium on Geomorphology, 1980. Geo Books, Norwich, pp. 139–150. Saunderson, H. C. and Jopling, A. V. 1980. Palaeohydraulics of a tabular, cross-stratified sand in the Brampton esker, Ontario. Sedimentary Geology, 25: 169–188. Schack Pedersen, S. A. 1988. Glacitectonite: brecciated sediments and cataclastic sedimentary rocks formed subglacially. In: R. P. Goldthwait and C. L. Matsch (eds) Genetic Classification of Glacigenic Deposits. A. A. Balkema, Rotterdam, pp. 89–91. Scherer, R. P. 1991. Quaternary and Tertiary microfossils from beneath Ice Stream B: Evidence for a dynamic West Antarctic Ice Sheet history. Palaeogeography, Palaeoclimatology, Palaeoecology (Global Planetary Change Section), 90: 395–412. Schermerhorn, L. J. G. 1974. Late Precambrian mixtites: glacial and/or nonglacial ? American Journal of Science, 274: 673–824. Schl¨ucter, Ch. 1992. Terrestrial Quaternary stratigraphy. Quaternary Science Reviews, 11: 603–607. Schmok, J. P. and Clarke, G. K. C. 1989. Lacustrine sedimentary record of ice-dammed Neoglacial Lake Alsek. Canadian Journal of Earth Sciences, 26: 2092–2105. Scholz, C. H. 1987. Wear and gouge formation in brittle faulting. Geology, 15: 493–495. Scholz, C. H. and Engelder, J. T. 1976. The role of asperity indentation and ploughing in rock friction – I. International Journal of Rock Mechanics and Geomechanics Abstracts, 13: 149–154. Schubert, G. and Yuen, D. A. 1982. Initiation of ice ages by creep instability and surging of the East Antarctic ice sheet. Nature, 296: 127–130. Schweizer, J. and Iken, A. 1992. The role of bed separation and friction in sliding over an undeformable bed. Journal of Glaciology, 38: 77–92. Schytt, V. 1969. Some comments on glacier surges in eastern Svalbard. Canadian Journal of Earth Sciences, 6: 867–873. Serat, G., Dricot, E. and Wansard, G.. 1990. Evidence for an early glacial maximum in the French Vosges during the last glacial cycle, Nature, 346: 453–456.

519

Shabtaie, S. and Bentley, C. R. 1987. West Antarctic ice streams draining into Ross Ice Shelf: configuration and mass balance. Journal of Geophysical Research, 92: 1311–1336. Shackleton, N. J. 1969. The last interglacial in the marine and terrestrial record. Proceedings of the Royal Society of London, 174(B): 135–154. Shackleton, N. J. 1987. Oxygen isotopes, ice volume and sea level. Quaternary Science Reviews, 6: 183–190. Shackleton, N. J. and Opdyke, N. D. 1973. Oxygen isotope and palaeomagnetic stratigraphy of equatorial Pacific core V28–238: oxygen isotope temperatures and ice volumes on a 105 year and a 106 year scale. Quaternary Research, 3: 39–55. Shackleton, N. J., Backman, J., Zimmerman, H., Kent, D. V., Hall, M. A., Roberts, D. G., Schnitker, D., Baldauf, J. G., Desprairies, A., Homrighausen, R., Huddlestun, P., Keene, J. B., Kaltenback, A. J. Krumsiek, K. A. O., Morton, A. C., Murray, J. W. and Westberg-Smith, J. 1984. Oxygen isotope calibration of the onset of icerafting and history of glaciation in the North Atlantic region. Nature, 307: 620–623. Shackleton, N. J., Berger, A. and Peltier, W. A. 1990. An alternative astronomical calibration of the lower Pleistocene timescale based on ODP Site 667. Transactions of the Royal Society of Edinburgh, Earth Sciences, 81: 251–261. Shakesby, R. A. 1985. Geomorphological effects of j¨okulhlaups and ice-dammed lakes, Jotunheimen, Norway. Norsk Geografisk Tidsskrift, 39: 1–16. Sharp, M. J. 1985. Sedimentation and stratigraphy at Eyjabakkaj¨okull – an Icelandic surging glacier. Quaternary Research, 24: 268–284. Sharp, M. J. and Gomez, B. 1986. Processes of debris comminution in the glacial environment and implications for quartz sand-grain micromorphology. Sedimentary Geology, 46: 33–47. Sharp, M., Richards, K. S. and Tranter, M. 1998. Glacier Hydrology and Hydrochemistry. J. Wiley, Chicester, UK, 342 pp. Sharpe, D. R. 1988a. The internal struture of glacial landforms: an example from the Halton till plain, Scarborough Bluffs, Ontario. Boreas, 17: 15–26. Sharpe, D. R., 1988b. Late glacial landforms of Wollaston Peninsula, Victoria Island, Northwest Territories: product of ice-marginal retreat, surge, and mass stagnation. Canadian Journal of Earth Sciences, 25: 262–279. Sharpe, D. R. and Barnett, P. J. 1985. Significance of sedimentological studies on the Wisconsinan stratigraphy of southern Ontario. G´eographie Physique et Quaternaire, 39: 255–273.

520

REFERENCES

Sharpe, D. R. and Cowan, W. R. 1990. Moraine formation in northwestern Ontario: product of subglacial fluvial and glaciolacustrine sedimentation. Canadian Journal of Earth Sciences, 27: 1478–1486. Sharpe, D. R. and Shaw, J. 1989. Erosion of bedrock by subglacial meltwater, Cantley, Quebec. Geological Society of America, Bulletin, 101: 1011–1020. Shaw, J. 1979. Genesis of the Sveg tills and R¨ogen moraines of central Sweden: a model of basal melt out. Boreas, 8: 409–426. Shaw, J. 1982. Melt-out till in the Edmonton area, Alberta, Canada. Canadian Journal of Earth Sciences, 19: 1548–1569. Shaw, J. 1985. Subglacial and ice marginal environments. In: G. M. Ashley, J. Shaw and N. D. Smith (eds) Glacial Sedimentary Environments. Society of Economic Palaeontologists and Mineralogists, Short Course 16, Tulsa, OK, pp. 7–84. Shaw, J. 1987. Glacial sedimentary processes and environmental reconstruction based on lithofacies. Sedimentology, 34: 103–116. Shaw, J. 1988a. Coarse-grained sediment flow facies in a large supraglacial lake, Discussion. Sedimentology, 35: 527–530. Shaw, J. 1988b. Models of terrestrial glacial sedimentation. In: J. Shaw and G. Ashley (eds) Glacial Facies Models, Continental Terrestrial Environments. Geological Society of America Short Course, pp. 1–78. Shaw, J. 1988c. Subglacial erosion marks, Wilton Creek, Ontario. Canadian Journal of Earth Sciences, 25: 1256–1267. Shaw, J. and Freschauf, R. C. 1973. A kinematic discussion of the formation of glacial flutings. Canadian Geographer, 17: 19–35. Shaw, J. and Sharpe, D. R. 1987. Drumlin formation by subglacial meltwater erosion. Canadian Journal of Earth Sciences, 24: 2316–2322. Shaw, J., Gilbert, R. and Archer, J. J. J. 1978. Proglacial lacustrine sedimentation during winter. Arctic and Alpine Research, 10: 689–699. Shaw, J., Kvill, D. and Rains, B. 1989. Drumlins and catastrophic subglacial floods. Sedimentary Geology, 62: 177–202. Shaw, J., Faragini, D. M., Kvill, D. R. and Rains, R. B. 2000. The Athabasca fluting field, Alberta, Canada: implications for the formation of large-scale fluting (erosional lineations). Quaternary Science Reviews, 19: 959–980. Shilts, W. W., Aylsworth, J. M., Kaszycki, C. A. and Klassen, R. A. 1987. Canadian Shield. In W. L. Graf (ed.) Geomorphic Systems of North America. Geological

Society of America Centennial Special Volume 2, pp. 119–161. Shoemaker, E. M. 1986a. The formation of fjord thresholds. Journal of Glaciology, 32: 65–71. Shoemaker, E. M. 1986b. Debris-influenced sliding laws and basal debris balance. Journal of Glaciology, 32: 224–231. Shoemaker, E. M. 1986c. Subglacial hydrology for an ice sheet resting on a deformable aquifer. Journal of Glaciology, 32: 20–30. Shoemaker, E. M. 1988. On the formulation of basal debris drag for the case of sparse debris. Journal of Glaciology, 34: 259–264. Shoemaker, E. M. 1992a. Subglacial floods and the origin of low-relief ice-sheet lobes. Journal of Glaciology, 38: 105–112. Shoemaker, E. M. 1992b. Water sheet outburst floods from the Laurentide ice sheet. Canadian Journal of Earth Sciences, 29: 1250–1264. Shoji, H. and Langway Jr., C. C. 1985. Comparison of mechanical tests on the Dye-3, Greenland ice core and artificial laboratory ice. Annals of Glaciology, 6: 305. Shreve, R. L. 1984. Glacier sliding at subfreezing temperatures. Journal of Glaciology, 30: 341–347. Shreve, R. L. and Sharp, R. P. 1970. Internal deformation and thermal anomalies in lower Blue Glacier, Mount Olympus, Washington, U.S.A. Journal of Glaciology, 9: 65–86. Sibrava, V. 1986. Correlation of European glaciations and their relation to the deep-sea record. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 433–441. Sibrava, V., Bowen, D. Q. and Richmond, G. M. 1986. Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5, 510 pp. Siegert, M. J. 2000. Antarctic subglacial lakes. EarthScience Reviews, 50: 29–50. Siegert, M. J., Dowdeswell, J. A., Gorman, M. R. and McIntyre, N. F. 1996. An inventory of Antarctic subglacial lakes. Antarctic Science, 8(3): 281–286. Sigenthaler C. and Sturm, M. 1991. Slump induced surges and sediment transport in Lake Uri, Switzerland. International Association for Theoretical and Applied Limnology Congress Proceedings 1989, 241: 955–958. Sikonia, W.G . and Post, A. 1980. Columbia Glacier, Alaska: Recent Ice Loss and its Relationship to Seasonal Terminal Embayments, Thinning and Glacier Flow. United States Geological Survey Hydrologic Investigations Atlas, HA-619, 3 sheets.

REFERENCES

Simpson, J. B. 1933. The late-glacial readvance moraines of the Highland border west of the river Tay. Transactions of the Royal Society of Edinburgh, 57: 633–645. Sissons, J. B. 1958. Supposed ice-dammed lakes in Britain with particular reference to the Eddleston valley, southern Scotland. Geografiska Annaler, 40: 159–187. Sissons, J. B. 1967 The Evolution of Scotland’s Scenery. Oliver & Boyd, Edinburgh, 259 pp. Sissons, J. B. 1976. Scotland. Methuen & Co., London, 150 pp. Sissons, J. B. 1977. The Loch Lomond Readvance in the northern mainland of Scotland. In: J. M. Gray, and J. J. Lowe (eds) Studies in the Scottish Lateglacial Environment. Pergamon, Oxford, pp. 45–59. Sissons, J. B. 1982. A former ice-dammed lake and associated glacier limits in the Achnasheen area, central Ross-shire. Transactions of the Institute of British Geographers, New Series 7: 205–216. Slater, G. 1926. Glacial tectonics as reflected in disturbed drift deposits. Proceedings of the Geologists Association, 37: 392–400. Small, R. J. 1987. Englacial and supraglacial sediment: transport and deposition. In Gurnell, A. M. and Clark, M. J. (eds), Glacio–Fluvial Sediment Transfer – an Alpine Perspective, J. Wiley Publishing, Chichester, pp. 111–145. Small, R. J. and Clark, M. J. 1976. Morphology and development of medial moraines: reply to comments by N. Eyles. Journal of Glaciology, 17: 162–164. Small, R. J., Clark, M. J. and Cawse, T. J. P. 1979. The formation of medial moraines on alpine glaciers. Journal of Glaciology, 22: 43–52. Smalley, I. J. 1966. Drumlin formation; a rheological model. Science, 151: 1379–1380. Smalley, I. J. and Unwin, D. J. 1968. Formation and shape of drumlins and their orientation and distribution. Journal of Glaciology, 7: 377–480. Smart, C. C. 1990. Comments on: ‘Character of the englacial and subglacial drainage system in the lower part of the ablation area of Storglaci¨aren, Sweden, as revealed by dye-trace studies’. Journal of Glaciology, 36: 126–128. Smith, D. G. 1983. Anastomosed fluvial deposits: modern examples from Western Canada. In: J. D. Collinson and J. Lewin (eds) Modern and Ancient Fluvial Systems. International Association of Sedimentologists, Special Publication, No. 6, pp. 155–168. Smith, N. D. 1990. The effects of glacial surging on sedimentation in a modern ice-contact lake, Alaska. Geological Society of America, Bulletin, 102: 1393–1403.

521

Smith, N. D. and Ashley, G. M. 1985. Proglacial Lacustrine Environment. In: G. M. Ashley, J. Shaw and N. D. Smith (eds) Glacial Sedimentary Environments. Society of Economic Paleontologists and Mineralogists, Short Course 16, Tulsa, OK, pp. 135–216. Smith, N. D. and Syvitski, J. P. M. 1982. Sedimentation in a glacier-fed lake: The role of pelletization on deposition of fine-grained suspensates. Journal of Sedimentary Petrology, 52: 503–513. Smith, N. D., Vendl, M. A. and Kennedy, S. K. 1982. Comparison of sedimentation regimes in four glacier-fed lakes of western Alberta. In: R. Davidson-Arnott, W. Nickling, and B. D. Fahey (eds) Research in Glacial, Glacio-fluvial, and Glacio-lacustrine Systems. 6th Guelph Symposium on Geomorphology. Geo Books Publishing, Norwich, pp. 203–238. Sneed, E. D. and Folk, R. L. 1958. Pebbles in the lower Colorado River, Texas; A study in particle morphogenesis. Journal of Geology, 66: 114–150. Sohl, L. E., Christie-Blick, N. and Kent, D. V. 1999 Palleomagnetic polarity reversals in Marinoan (ca. 600 a) glacial deposits in Australia; Implications for the duration of low-latitude glaciation in Neoproterpzoic time. Geological Society of America Bulletin, 111: 1120–1139. Solheim, A. 1988. Glacial Geology of the Northern Barents Sea, with Emphasis on the Surge-related, Ice-proximal Depositional Environment. Norsk Polarinstitut Rapportserie, 47. Solheim, A. and Kristoffersen, Y. 1984. Sediments above the upper regional unconformity; thickness, seismic stratigraphy and outline of the glacial history. Skrifter – Norsk Polarinstitutt, 179B, p. 26. Solheim, A. and Pfirman, S. L. 1985. Sea-floor morphology outside a grounded surging glacier, Bråsvellbreen, Svalbard. Marine Geology, 65: 127–143. Sollid, J. L. and Sørbel, L. 1984. Distribution and genesis of moraines in Central Norway. In: L.-K. K¨onigsson (ed.) Ten years of Nordic Till Research. Striae, 20: 63–67. Souchez, R. A. and Lorrain, R. D. 1991. Ice Composition and Glacier Dynamics. Springer-Verlag, Berlin, 207 pp. Sparks, B. W. and West, R. G. 1964. The drift landforms around Holt, Norfolk. Transactions of the Institute of British Geographers, 35: 27–35. Spindler, M. 1990. A comparison of Arctic and Antarctic sea ice and the effects of different properties on sea ice biota. In: U. Bleil and J. Thiede (eds) Geological History of the Polar Oceans; Arctic versus Antarctic. NATO ASI Series, C 308, Kluwer Academic Publishers, Dordrecht, pp. 173–186. Stauffer, B. 1999. Cornucopia of ice core results. Nature, 399: 412–413.

522

REFERENCES

Steig, E. J., Brook, E. J., White, J. W. C., Sucher, C. M., Bender, M. L., Lehman, S. J., Morse, D. L., Waddington, E. D. and Clow, G. D. 1998a. Synchronous climate changes in Antarctica and the North Atlantic. Science, 282: 92–94. Steig, E. J., Wolfe, A. P., Miller, G. H. 1998b. Wisconsinan refugia and the glacial history of eastern Baffin Island, Arctic Canada; coupled evidence from cosmogenic isotopes and lake sediments Geology, 26 (9): 835–838 Steiner, J. and Grillmair, E. 1973. Possible galactic causes for periodic and episodic glaciations. Geological Society of America, Bulletin, 84: 1003–1018. Stenborg, T. 1969. Studies on the internal drainage of glaciers. Geografiska Annaler, 51A: 13–41. Stephan, H.-J. 1987. Form, composition and origin of drumlins in Schleswig-Holstein. In: J. Menzies and J. Rose (eds) Drumlin Symposium. A. A. Balkema, Rotterdam, pp. 335–345. Stephenson, S. N. and Doake, C. S. M. 1982. Dynamic behaviour of Rutford ice stream. Annals of Glaciology, 3: 295–299. Stevens, R. L. 1990. Proximal and distal glacimarine deposits in southwestern Sweden: contrasts in sedimentation. In: J. A. Dowdeswell and J. D. Scourse (eds) Glacimarine Environments: Processes and Sediments. Geological Society Special Publication No.53, pp. 307–316. Stockmal, G. S. 1983. Modeling of large-scale accretionary wedge deformation. Journal of Geophysical Research, 88: 8271–8287. Stone, B. D. and Force, E. R. 1983. Coarse-grained glaciolacustrine deltaic sedimentation in ice marginal and ice-channel environments. Geological Society of America, Abstracts with programs, 15: 126. Stone, K. H. 1963. Alaskan ice-dammed lakes. Annals of the Association of American Geographers. 53: 332–349. Straw, A. 1960. The limit of the ‘Last’ Glaciation in north Norfolk. Proceedings of the Geologists’ Association, 71: 379–390. Streeter, S. S., Belanger, P. E., Kellogg, T. B. and Duplessy, J. C. 1982. Late Pleistocene paleo-oceanography of the Norwegian-Greenland Sea: Benthic foraminiferal evidence. Quaternary Research, 18: 72–90. Strom, K. M. 1938. The catastrophic emptying of a glacierdammed lake in Norway, 1937. Geologie der Meere und Binnengew¨asser, 2: 443–444. Stuiver, M. and Braziunas, T. F. 1989. Atmospheric 14C and century-scale solar oscillations. Nature, 338: 405–408. Stump, E. and Fitzgerald, P. G. 1992. Episodic uplift of the Transantarctic Mountains. Geology, 20: 161–164. Sturm, M. 1986. Formation of a strandline during the 1984 j¨okulhlaup of Strandline Lake, Alaska. Arctic, 39: 267–269.

Sturm, M., Beget, J. and Benson, C. 1987. Observations of j¨okulhlaups from ice-dammed Strandline Lake, Alaska: implications for paleohydrology. In: L. Mayer and D. Nash (eds) Catastrophic Flooding. Allen and Unwin, Boston, pp. 79–94. Sturm, M. and Benson, C. S. 1985. A history of j¨okulhlaups from Strandline Lake, Alaska, U.S.A. Journal of Glaciology, 31: 272–280. Sturm, M. and Cosgrove, D. M. 1990. An unusual j¨okulhlaup involving potholes on Black Rapids Glacier, Alaska Range, Alaska, U.S.A. Journal of Glaciology, 36: 125–126. Sturm, M. and Matter, A. 1978. Turbidites and varves in Lake Brienz (Switzerland): deposition of clastic detritus by density currents. In: A. Matter, and M. E. Tucker (eds) Modern and Ancient Lake Sediments. International Association of Sedimentologists, Special Publication No. 2, pp. 147–168. Sturm, M., Hall, D. K., Benson, C. S. and Field, W. O. 1991. Non-climatic control of glacier-terminus fluctuations in the Wrangell and Chugach Mountains, Alaska, U.S.A. Journal of Glaciology, 37: 348–356. Sugden, D. E. 1977. Reconstruction of the morphology, dynamics, and thermal characteristics of the Laurentide Ice Sheet at its maximum. Arctic and Alpine Research, 9: 21–47. Sugden, D. E. 1978. Glacial erosion by the Laurentide Ice Sheet. Journal of Glaciology, 20: 367–391. Sugden, D. E. and John, B. S. 1976. Glaciers and Landscape – A Geomorphological Approach. Edward Arnold, London, 320 pp. Sugden, D. E., Clapperton, C. M., Gemmell, J. C. and Knight, P. G. 1987. Stable isotopes and debris in basal glacier ice, South Georgia, Southern Ocean. Journal of Glaciology, 33: 324–329. Sutinen, R. 1985. On the subglacial sedimentation of hummocky moraines and eskers in Northern Finland. Striae, 22: 21–26. Sveian, H., Aa, A. R., Kjaernes, P. A., 1979. Ice movements and deglaciation of the central parts of Saltfjellet, Nordland, North Norway. Bulletin Norges Geologiske Undersokelse, 49: 1–20. Sverdrup, H. V. 1935. The tempertature of the firn on Isachsen’s Plateau and general conclusions regarding the tempertature of the glaciers on West Spitsbergen. Geografiska Annaler, 17: 53–88. Swithinbank, C. 1988. Antarctica. In: R. S. Williams and J. G. Ferrigno (eds), Satellite Image Atlas of Glaciers of the World. United States Geological Survey, Professional Paper 1386-B, 278 pp. Swithinbank, C. T., Chinn, T. J., Williams, R. S. and Ferrigno,

REFERENCES

J. G. 1988. Satellite Image Atlas of Glaciers of the World. Antarctica United States Geological Survey, Professional Paper, 1386-B. Syvitski, J. P. M. 1989. On the deposition of sediment within glacier-influenced fjords: oceanographic controls. Marine Geology, 85: 301–329. Syvitski, J. P. M. 1991. Towards an understanding of sediment deposition on glaciated continental shelves. Continental Shelf Research, 11: 897–937. Syvitski, J. P. M. and Praeg, D. B. 1989. Quaternary sedimentation in the St. Lawrence estuary and adjoining areas, eastern Canada: An overview based on high resolution seismo-stratigraphy. G´eographie Physique et Quaternaire, 43: 291–310. Syvitski, J. P. M., Burrell, D. C. and Skei, J. M. 1987. Fjords: Processes and Products. Springer-Verlag, New York, 379 pp. Syvitski, J. P. M., LeBlanc, K., William, G. and Cranston, R. E. 1990. The flux and preservation of organic carbon in Baffin Island Fjords. In: J. A. Dowdeswell and J. D. Scourse (eds) Glacimarine Environments: Processes and Sediments. Geological Society of London Special Publication No.53, pp. 177–199. Talbot, C. J. and Von Brunn, V. 1987. Intrusive and extrusive (micro)melange couplets as distal effects of tidal pumping by a marine ice sheet. Geological Magazine, 124: 513–525. Tarling, D. H. 1974. A palaeomagnetic study of Eocambrian tillites in Scotland. Journal of the Geological Society of London, 130: 163–170. Tarr, R. S. and Butler, B. S. 1909. The Yakutat Bay Region, Alaska. U.S. Geological Survey, Professional Paper No. 64. Tarr, R. S. and Martin, L. 1914. Alaskan Glacier Studies. National Geographical Society, Washington, 498 pp. Taviani, M., Reid, D. E. and Anderson, J. B. 1993 Skeletal and isotopic composition and paleoclimatic significance of late Pleistocene carbonates, Ross Sea, Antarctica. Journal of Sedimentary Petrology, 63: 84–90. Taylor, K. C., Lomorey, G. W., Doyle, G. A., Alley, R. B., Grootes, P. M., Mayewski, P. A., White, J. W. C. and Barlow, L. K. 1993. The ‘flickering’ switch of late Pleistocene climate change. Nature, 361: 432–436. Taylor, S. R. and McLennan, S. M., 1985. The Continental Crust: its Composition and Evolution. Blackwell Scientific, Oxford, 312 pp. Tchalenko, J. S. 1970. Similarities between shear zones of different magnitudes. Geological Society of America, Bulletin, 81: 1625–1640. Tchalenko, J. S. and Ambraseys, N. N. 1970. Structural

523

analysis of the Dasht-e-Bayaz (Iran) earthquake fractures. Geological Society of America, Bulletin, 81: 41–60. Teller, J. T. 1987. Proglacial lakes and the southern margin of the Laurentide Ice Sheet. In: W. F. Ruddiman, and H. E. Wright, Jr. (eds) North America and Adjacent Oceans During the Last Deglaciation. The Decade of North American Geology, Boulder, CO, Geological Society of America, K3, pp. 39–69. Teller, J. T. 1990. Volume and routing of Late-Glacial runoff from the southern Laurentide ice sheet. Quaternary Research, 34: 12–23. Teller, J. and Clayton, L. 1983 Glacial Lake Agassiz. Geological Association of Canada Special Paper, 26, 451 pp. Teller, J. T. and Thorliefson, L. H. 1987. Catastrophic flooding into the Great Lakes from Lake Agassiz. In: L. Mayer and D. Nash (eds), Catastrophic Flooding. The Binghampton Symposia in Geomorphology: International Series, No. 18, Allen & Unwin, Boston, pp. 121–138. Theakstone, W. H. 1967. Basal sliding and movement near the margin of the glacier Østerdalsisen, Norway. Journal of Glaciology, 6: 805–816. Theakstone, W. H. 1979. Observations within cavities at the bed of the glacier Østerdalsisen, Norway. Journal of Glaciology, 23: 273–281. Theodorsson, T. 1980. The Hagafell glacier surge, J¨okull, 30: 75–77. Thomas, G. S. P. 1984. A late Devensian glaciolacustrine fan-delta at Rhosesmor, Clwyd, North Wales. Geological Journal, 19: 125–141. Thomas, G. S. P. and Connell, R. J. 1985. Iceberg drop, dump and grounding structures from Pleistocene glaciolacustrine sediments, Scotland. Journal of Sedimentary Petrology, 55: 243–249. Thomas, R. H. 1979. The dynamics of marine ice sheets. Journal of Glaciology, 24: 167–177. Thompson, A. 1988. Historical development of the proglacial landforms of Svinafellsj¨okull and Skaftafellsj¨okull, Southeast Iceland. J¨okull, 38: 17–31. Thompson, L. G. and Mosley-Thompson, E. 1982. Microparticle concentration and size-distribution determinations from the J-9 Core, Ross Ice Shelf. Antarctic Journal of the United States, 17: 83–85. Thompson, L. G., Mosley-Thompson, E., Dansgaard, W. and Grootes, P. M. 1986. The Little Ice Age as recorded in the stratigraphy of the tropical Quelccaya Ice Cap. Science, 234: 361–364. Thompson, R. and Kelts, K. 1974. Holocene sediments and magnetic stratigraphy from Lakes Zug and Zurich, Switzerland. Sedimentology, 21: 577–596.

524

REFERENCES

Thomsen E. and Vorren, T. O. 1986. Macrofaunal palaeoecology and stratigtaphy in late Quaternary shelf sediments off northern Norway. Palaeogeography, Palaeoclimatology, Palaeoecology, 56: 103–150. Thorarinsson, S. 1939. Hoffellsjokull, its movement and drainage. Geografiska Annaler, 21: 189–215. Thorarinsson, S. 1957. The j¨okulhlaups from the Katla area in 1955 compared with other j¨okulhlaups in Iceland. J¨okull, 7: 21–35. Thorarinsson, S. 1969. Glacier surges in Iceland, with special reference to the surges of Br´uarj¨okull. Canadian Journal of Earth Sciences, 6: 875–882. Thorarinsson, S. 1974. V¨otnin Str´ı∂/ . Saga Skei∂/ ar´arhlaupa og Gr´ımsvatnagosa. Bka´utg´afa Menningarsj´o∂/ s, Reykjav´ık, 255pp. Thorp, P. W. 1991. Surface profiles and basal shear stresses of outlet glaciers from a Late-glacial mountain ice field in western Scotland. Journal of Glaciology, 37: 77–88. Todtmann, E. 1936. Einige Ergebnisse von glazialgeologischen Untersuchungen am S¨udrand des Vatnaj¨okull auf Island (1931 und 1934). Zeitschrift der Deutsche Geologische Gesellschaft, 88: 77–87. Todtmann, E. 1952. Im Gletscherrueckzugsgebiet des Vatna Joekull auf Island, 1951. Neues Jahrbuch fuer Geologie und Palaeontologie, Mon. 9: 401–411. Todtmann, E. M. 1960. Gletscherformen auf Island (Vatnaj¨okull). Universit¨at Hamburg. Abhandlungen aus dem Gebiet der Auslandkunde, Band 65, Reihe C (Naturwissenschaften), Band 19. T´omasson, H. 1974. Gr´ımsvatnahlaup 1972. Mechanism and sediment discharge. J¨okull, 24: 27–39. Trenhaile, A. S. 1975. The morphology of a drumlin field. Annals of the Association of American Geographers, 65: 297–312. Trewartha, G.T. 1968. An Introduction to Weather Climate. McGraw-Hill. Tucker, M. E. 1986. The Field Description of Sedimentary Rocks. Geological Society of London Handbook Series/ Open University Press. 112 pp. Tufnell, L. 1984. Glacier Hazards. Longman Publishing, London, 97 pp. Tweed, F. S. and Russell, A. J. 1999. Controls on the formation and sudden drainage of glacier-impounded lakes: Implications for j¨okulhlaup characteristics. Progress in Physical Geography, 23: 79–111. Tyndall, J. 1864. On the conformation of the Alps. Philosophical Magazine, 28: 225–271. Uskov, Yu. S. and Dil’muradov, N. 1983. On the glacier surges in the Pamirs. Mat. Glyatsiol. Issled. Khron. Obsuzhdeniya, 47: 142–146.

Vail, P. R., Mitchum, R. M. andThompson, S. 1977. Seismic stratigraphy and global changes in sea-level. American Association of Petroleum Geologists Memoir, 26: 83–97. Van Campo, E., Duplessy, J. C., Prell, W. L., Barratt, N. and Sabatier, R. 1990. Comparison of terrestrial and marine temperature estimates for the past 135 kyr off southeast Africa: a test for GCM Simulations of palaeoclimate. Nature, 348: 209–212. Van Couvering, J. A. (ed.) 1997. The Pleistocene Boundary and the Beginning of the Quaternary. Cambridge University Press, 269 pp. Van den Berg, M. W. and Beets, D. J. 1987. Saalian glacial deposits and morphology in The Netherlands. In: J. J. M. van der Meer (ed.) Tills and Glaciotectonics. Balkema, Rotterdam, pp. 235–251. Van der Meer, J. J. M. 1982. The Fribourg area, Switzerland: A study in Quaternary geology and soil development. Ph.D. thesis, University of Amsterdam, Publicaties van het Fysisch Geografisch en Bodemkundig Laboratorium 32, Amsterdam, 203 pp. Van der Meer, J. J. M. 1983. A recent drumlin with fluted surface in the Swiss Alps. In: E. B. Evenson, Ch. Schl¨uchter and J. Rabassa (eds) Tills and Related Deposits: Genesis, Petrology, Application, Stratigraphy. A. A. Balkema Publishers, Rotterdam, pp. 105–110. Van der Meer, J. J. M. (ed.) 1987a. Tills and Glaciotectonics. A. A. Balkema Publishers, Rotterdam, 270 pp. Van der Meer, J. J. M. 1987b. Micromorphology of glacial sediments as a tool in genetic varieties of till. In: R. Kujansuu and M. Saarnisto (eds) INQUA Till Symposium, Finland 1985. Geological Survey of Finland Special Paper, 3, pp. 77–89. Van der Meer, J. J. M. 1993. Microscopic evidence of subglacial deformation. Quaternary Science Reviews, 12: 553–587. Van der Meer, J. J. M. 1996. Micromorphology. In J. Menzies (ed.) Past Glacial Environments – Sediments, Forms and Techniques. Volume II, Butterworth-Heineman, Oxford. Chapter 12, pp. 335–356. Van der Meer, J. J. M. 1997. Particle and aggregate mobility in till: microscopic evidence of subglacial processes: Quaternary Science Reviews, 16: 827–831. Van der Meer, J. J. M., Rappol, M. and Semeijn, J. N. 1985. Sedimentology and genesis of glacial deposits in the Goudsberg, Central Netherlands. Mededelingen van de Rijks Geologische Dienst, 39: 1–29. Van der Meer, J. J. M., Verbers, A. L. L. M. and Warren, W. P. 1994. The micromorphological character of the Ballycroneen Formation (Irish Sea Till): A first assessment. In: W. P. Warren and D. G.Croot (eds) Formation

REFERENCES

and Deformation of Glacial Deposits. A. A.Balkema Publishers, Rotterdam, pp. 39–49. Van der Veen, C. J. 1999. Fundamentals of Glacier Dynamics. A. A. Balkema Publishing, Rotterdam, 462 pp. Van der Veen, C. J. and Oerlemans, J. 1987. Dynamics of the West Antarctic Ice Sheet. D. Reidel Publishing, Dordrecht, 368 pp. Van der Wateren, F. M. 1981. Glacial tectonics at the Kwintelooijen sandpit, Rhenen, the Netherlands. Mededelingen Rijks Geologische Dienst, 35: 252–268. Van der Wateren, F. M. 1985. A model of glacial tectonics, applied to the ice-pushed ridges in the Central Netherlands. Bulletin of the Geological Society of Denmark, 34: 55–74. Van der Wateren, F. M. 1987. Structural geology and sedimentology of the Dammer Berge push moraine, FRG. In: J. J. M. van der Meer (ed.) Tills and Glaciotectonics. A. A. Balkema, Rotterdam: pp. 157–182. Van der Wateren, F. M. 1992. Structural Geology and Sedimentology of Push Moraines. – Processes of soft sediment deformation in a glacial environment and the distribution of glaciotectonic styles. Amsterdam: Ph.D. Thesis, 238 pp. Van der Wateren, F. M. 1994. Proglacial subaquatic outwash fan and delta sediments in push moraines; indicators of subglacial meltwater activity, Sedimentary Geology, 91: 1–4. Van der Wateren, F. M. 1995. Structural geology and sedimentology of push moraines. – Processes of soft sediment deformation in a glacial environment and the distribution of glaciotectonic styles. Mededelingen Rijks Geologische Dienst, 54: 1–168. Van der Wateren, F. M. 1999. Structural geology and sedimentology of Saalian tills near Heiligenhafen, Germany. Quaternary Science Reviews, 18: 1625–1639. Van der Wateren, F. M., Kluiving, S. J. and Bartek, L. R. 2000. Kinematic indicators of subglacial shearing. In: A. J. Maltman, M. J. Hambrey and B.P. Hubbard (eds) Deformation of Glacial Materials. Geological Society, Special Publication 176, The Geological Society, pp. 259–278. Van Gijssel, K. 1987. A lithistratigraphic and glaciotectonic reconstruction of the Lamstedt Moraine, Lower Saxony (FRG). In: J. J. M. van der Meer (ed.), Tills and Glaciotectonics. A. A. Balkema, Rotterdam, pp. 145–155. Vasari, Y. 1977. Radiocarbon dating of the Lateglacial and Early Flandrian vegetation succession in the Scottish Highlands and the Isle of Skye. In Gray, J. M. and Lowe,

525

J. J. (eds.) Studies in the Scottish Lateglacial Environment, 143–162. Vasari, Y. and Vasari, A. 1968. Late and post-glacial macrophyttic vegetation in the lochs of Northern Scotland. Acta Botanica Fenna, 80:1–120. Vaughan, D. G. and Doake, C. S. M. 1996. Recent atmospheric warming and retreat of ice shelves on the Antarctic Peninsula. Nature, 379: 328–331. Veevers, J. J. and Powell, C. McA. 1987. Late Paleozoic glacial episodes in Gondwanaland reflected in transgressive–regressive depositional sequences in Euramerica. Geological Society of America, Bulletin, 98: 475–487. Velichko, A. A. and Faustova, M. A. 1986. Glaciations in the East European region of the USSR. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 447–461. Vernekar, A. D. 1968. Long-period Global Variations of Incoming Solar Radiation. Final Rept., ESSA Contract 22–137–67(N), The Travelers Research Center, Inc., 289 pp. Viete, G. 1960. Zur Entstehung der glazigenen Lagerst¨orungen unter besonderer Ber¨ucksichtigung der Fl¨ozdeformationen im mitteldeutschen Raum. Freiberger Forschungshefte, C-78: 1–257. Visser, J. N. J. 1989. The Permo-Carboniferous Dwyka Formation of Southern Africa: Deposition by a predominantly subpolar marine ice sheet. Palaeography, Palaeoclimatology, Palaeoecology, 70: 377–391. Vivian, R. and Bocquet, G. 1973. Subglacial cavitation phenomena under the Glacier d’Argenti‘ere, Mont Blanc, France. Journal of Glaciology, 12: 439–451. Von B¨ohm, A. E. 1901. Gesichte der Mor¨anenkunde. Ab. Der Geogr. Gesell. Wien, III Band, 4, 334 pp. Von Drygalski, E. 1897. Gr¨onland-Expedition der Gesellschaft f¨ur Erdkunde zu Berlin. Berlin. Vorren, T. O., Hald, M. and Thomsen, E. T. 1984. Quaternary sediments and environments on the continental shelf off Northern Norway. Marine Geology, 57: 229–257. Vorren, T. O., Hald, M. and Lebesbye, E. 1988. Late Cenozoic environments in the Barents Sea. Paleoceanography, 3: 601–612. Vorren, T. O., Lebesbye, E., Andereasson, K. and Larsen, K. B. 1989. Glacigenic sediments on a passive continental margin as exemplified by the Barents Sea. Marine Geology, 85: 251–272. Vorren, T. O., Richardsen, G., Knutsen, S. M. and Henriksen, E. 1991. Cenozoic erosion and sedimentation in the western Barents Sea. Marine and Petroleum Geology, 8: 317–340.

526

REFERENCES

Vorren, T. O., Laberg, J. S., Blaume, F., Dowdeswell, J. A., Kenyon, N. H., Miernert, J., Rumohr, J. and Werner, F., 1998. The Norwegian-Greenland Sea Continental Margins: morphology and late Quaternary sedimentary processes and environment. Quaternary Science Reviews, 17: 273–302. Waitt Jr., R. B. 1985. Case for periodic, colossal j¨okulhlaups from Pleistocene glacial Lake Missoula. Geological Society of America, Bulletin, 96: 1271–1286. Walder, J. S. 1982. Stability of sheet flow of water beneath temperate glaciers and implications for glacier surging. Journal of Glaciology, 28: 273–293. Walder, J. S. 1986. Hydraulics of subglacial cavities. Journal of Glaciology, 32: 439–445. Walder, J. S. and Hallet, B. 1979. Geometry of former subglacial water channels and cavities. Journal of Glaciology, 23: 335–346. Walder, J. S. and Hallet, B. 1985. A theoretical model of the fracture of rock during freezing. Geological Society of America, Bulletin, 96: 336–346. Walder, J. S. and Hallet, B. 1986. The physical basis of frost weathering: Toward a more fundamental and unified perspective. Arctic and Alpine Research, 18: 27–32. Walker, M. J. C. 1973. The nature and origin of a series of elongated ridges in the Morley Flats area of the Bow Valley, Alberta. Canadian Journal of Earth Sciences, 10: 1340–1346. Walker, M. J. C., Coope, G. R. and Lowe, J. J. 1993. The Devensian (Weichselian) Lateglacial palaeoenvironmental record from Gransmoor, East Yorkshire, England. Quaternary Science Reviews, 12: 659–680. Walker, M. J. C., Bj¨orck, S., Lowe, J. J., Cwynar, L. C., Johnsen, S., Knudsen, K-L., Wohlfarth, B., and members of the INTIMATE group 1999. Isotopic ‘events’ in the GRIP ice core: a stratotype for the Late Pleistocene. Quaternary Science Reviews, 18: 1143–1150. Walling, D. E. 1984. Dissolved loads and their measurement. In Hadley R. F. and Walling, D. E. (eds), Erosion and Sediment Yield: Some Methods of Measurement and Modelling. Geo Books, Norwich, pp. 111–177. Warren, C. R. and Glasser, N. F. 1992. Contrasting response of south Greenland glaciers to recent climatic change. Arctic and Alpine Research, 24: 124–132. Warren, C. R. and Hulton, N. R. J. 1990. Topographic and glaciological controls on Holocene ice-sheet margin dynamics, Central West Greenland. Annals of Glaciology, 14: 307–310. Warren, W. P. and Ashley, G. M. 1994. Origins of the icecontact stratified ridges (eskers) of Ireland. Journal of Sedimentary Research, 64: 433–449.

Warren, W. P. and Ashley, G. M. (eds) 1997 Ice-contact sedimentation; processes and deposits. Quaternary Science Reviews, 16: 627–819. Warren, W. P. and Croot, D. G. (eds) 1994. Formation and Deformation of Glacial Deposits. A. A. Balkema Publishers, Rotterdam, 223 pp. Watson, R. A. and Wright Jr., H. E. 1980. The end of the Pleistocene: A general critique of chronostratigraphic classification. Boreas, 9: 153–163. Webb, P. N. 1990. The Cenozoic history of Antarctica and its global impact. Antarctic Science, 2: 3–21. Weertman, J. 1957. On the sliding of glaciers. Journal of Glaciology, 3: 33–38. Weertman, J. 1962. Stability of Ice-age Ice Caps. U.S. Cold Regions Research and Engineering Laboratory Research Report, 97, 12 pp. Weertman, J. 1964. The theory of glacier sliding. Journal of Glaciology, 5: 287–303. Weertman, J. 1966. Effect of a basal water layer on the dimensions of ice sheets. Journal of Glaciology, 6: 191–207. Weertman, J. 1969. Water lubrication mechanism of glacier surges. Canadian Journal of Earth Sciences, 6: 929–942. Weertman, J. 1972. General theory of water flow at the base of a glacier or ice sheet. Reviews of Geophysics and Space Physics, 10: 287–333. Weertman, J. 1979. The unsolved general glacier sliding problem. Journal of Glaciology, 23: 97–115. Weertman, J. 1980. Bottom crevasses. Journal of Glaciology, 25: 185–188. Weertman, J. 1986. Basal water and high-pressure basal ice. Journal of Glaciology, 32: 455–463. Weertman, J. and Birchfield, G. E. 1983. Stability of sheet water flow under a glacier. Journal of Glaciology, 29: 374–382. Weirich, F. H. 1984. Turbidity currents: monitoring their occurrence and movement with a three-dimensional sensor network. Science, 224: 384–387. Weirich, F. H. 1985. Sediment budget for a high energy glacial lake. Geografiska Annaler, 67A: 83–99. West, R. G. 1977. Pleistocene Geology and Biology, Second Edition. Longman, London, 440 pp. West, R. G. 1989. The use of type localities and type sections in the Quaternary, with especial reference to East Anglia. In: J. Rose and Ch. Schl¨uchter (eds) Quaternary Type Sections: Imagination or Reality? A. A. Balkema Publishers. Rotterdam, pp. 3–10. Whalley, W. B. 1971. Observations of the drainage of an ice-dammed lake – Strupvatnet, Troms, Norway. Norsk Geografisk Tidsskrift, 25: 165–174.

REFERENCES

Whalley, W. B. and Langway Jr., C. C. 1980. A scanning electron microscope examination of subglacial quartz grains from Camp Century core, Greenland – a preliminary study. Journal of Glaciology, 25: 125–131. Whiteman, C. A. 1983. Great Waltham. In: J. Rose (ed.) Diversion of the Thames Field Guide. Quaternary Research Association, Cambridge, pp. 163–169. Whiteman, C. A. 1987. Till lithology and genesis near the southern margin of the Anglian ice sheet in Essex, England. In: J. J. M. van der Meer (ed.) Tills and Glaciotectonics. A. A. Balkema, Rotterdam, pp. 55–66. Whiteman, C. A. and Rose, J. 1992. Thames river sediments of the British Early and Middle Pleistocene. Quaternary Science Reviews, 11: 363–375. Whittecar, G. R. and Mickelson, D. M. 1979. Composition, internal structures, and an hypothesis of formation for drumlins, Waukesha County, Wisconsin, U.S.A. Journal of Glaciology, 22: 357–370. Wickham, S. S., Johnson, W. H., and Glass, H. D. 1988. Regional Geology of the Tiskilwa Till Member, Wedron Formation, Northeastern Illinois. Illinois State Geological Survey Circular 543, 35 pp. Williams, D. M., Kasting, J. F. and Frakes, L. A. 1998. Low-latitude glaciation and rapid changes in the Earth’s obliquity explained by obliquity-oblateness feedback. Nature, 396: 453–455. Williams, G. E. 1975. Late Precambrian glacial climate and the Earth’s obliquity. Geological Magazine, 112: 441–544. Williams, G. E. and Schmidt, P. W. 1997. Paleomagnetism of the paleoproterozoic Gowganda and Lorrain Formations, Ontario: low paleolatitude for Huronian glaciation. Earth and Planetary Science Letters, 153: 157–169. Wilson, K. C. 1969. Comments on ‘Sediment transport in conveyance systems’. Bulletin of the International Association of Scientific Hydrology, 14: 131–134. Wilson, R. c. l., Drury, S. A. and Chapman, J. L. 2000. The Great Ice Age. Routledge, London, 267 pp. Winograd, I. J., Coplen, T. B., Landwehr, J. M., Riggs, A. C., Ludwig, K. R., Szabo, B. J., Kolesar, P. T. and Revesz, K. M. 1992. Continuous 500,000-year climate record from vein calcite in Devils Hole, Nevada. Science, 258: 255–259. Winograd, I. J., Landwehr, J. M, Coplen, T. B. and Riggs, H. C. 1997. Duration and structure of the four interglaciations. Quaternary Research, 48: 141–154. Wold, B. and Østrem, G. 1979. Subglacial constructions and investigations at Bondhusbreen, Norway. Journal of Glaciology, 23: 363–379.

527

Woldstedt, P. and Duphorn, K. 1974. Norddeutschland und angrenzende Gebiete im Eiszeitalter. Koehler, Stuttgart, 500 pp. Wolf, T. C. W. and Thiede, J. 1991. History of terrigenous sedimentation during the past 10 My in the North Atlantic (ODP Legs 104, 105 and DSDP Leg 81). Marine Geology, 101: 83–102. Worsley, T. R. and Kidder, D. L. 1991. First-order coupling of paleogeography and CO2, with global surface temperature and its latitudinal contrast. Geology, 19: 1161–1164. Worsley, T. R. and Nance, R. D. 1989. Carbon redox and climate control through Earth history, a speculative reconstruction. Palaeogeography, Palaeoclimatology, Palaeoecology (Global and Planetary Change Section), 75: 259–282. Wright, W. B. 1937 The Quaternary Ice Age. 2nd Edition, MacMillan, London, 478 pp. Yevteyev, S.A. 1959. Opredeleniye kolichevstva morenngogo materiala, perenosimogo lednikami vostchnogog poberezh’ya Antarktidy. Informatsiooyy Byulleten’ Sovetskoy Anarkticheskoy Ekspeditsii, 11: 14–16. Young, G. M. 1981. The Early Proterozoic Gowganda Formation, Ontario, Canada. In: M. J. Hambrey and W. B Harland (eds) Earth’s pre-Pleistocene Glacial Record. Cambridge University Press, Cambridge, pp. 807–812. Young, G. M. 1991. The geologic record of glaciation: relevance to the climatic history of Earth. Geoscience Canada, 18: 100–108. Young, G. M. 1994. Eustatic and Plate Tectonic Controls on Post-glacial Successions in the Proterozoic. Geological Association of Canada and Mineralogical Association of Canada; Annual meeting; program with abstracts, Waterloo, ON, Canada, 16–18 May 1994. Young, G. M. 1995. Are Neoproterozoic glacial deposits preserved on the margins of Laurentia related to the fragments of two supercontinents? Geology, 23: 153–156. Young, G. M. 1997. Tectonic and glaciolacustrine controls on postglacial stratigraphy. In: I. P. Martini (ed.) Late Glacial and Postglacial Environmental Changes. Oxford University Press, pp. 29–267. Young, G. M. and Gostin, V. A. 1989a. An exceptionally thick upper Proterozoic (Sturtian) glacial succession in the Mount painter area, South Australia. Geological Society of America, Bulletin, 101: 834–845. Young, G. M. and Gostin, V. A. 1989b. Depositional environment and regional stratigraphic significance of the Serle Conglomerate: a Late Proterozoic submarine

528

REFERENCES

fan complex, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology, 71: 237–252. Young, G. M. and Nesbitt, H. W. 1985. The Gowganda Formation in the southern part of the Huronian outcrop belt, Ontario, Canada: stratigraphy, depositional environments and regional tectonic significance. 27 May 2000. Precambrian Research, 29: 265–301. Young, G. M., von Brunn, V., Gold, D. J. C. and Minter, W. E. L. 1998. Earth’s oldest reported glaciation: Physical and chemical evidence from the Archean Mozaan Group (~2.9 Ga) of South Africa. Journal of Geology, 106: 523–538. Young, J. A. T. 1975. Ice wastage in Glen Feshie, Inverness-shire. Scottish Geographical Magazine, 91: 91–101.

Zagwijn, W. H. 1986. The Pleistocene in the Netherlands with special reference to glaciation and terrace formation. In: V. Sibrava, D. Q. Bowen and G. M. Richmond (eds) Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews, 5: 341–345. Zielinski, G. A., Mayewski, P. A., Meeker, L. D., Whitlow, S., Twickler, M. S., Morrison, M., Taylor, K., Alley, R., Gow, A. J. and Meese, D. 1995. A continuous record of volcanism (Present – 7000 B.C.) and climatic implications. Unpublished manuscript. Zilliacus, H. 1989. Genesis of De Geer Moraines in Finland. Sedimentary Geology, 62: 309–317. Zumberge, J. H. 1955. Glacial erosion in tilted rock layers. Journal of Geology, 63: 149–158.

INDEX

Ablation, 53, 58, 61, 88, 103, 149, 276, 279, 317–18, 337, 362–3, 479 area (zone), 36, 56–60, 68, 83–4, 94, 97, 110, 112, 150–3, 156–9, 161,163, 167–9, 267, 273, 281, 317, 388, 422, 435 rates, 55, 63–4, 388 season, 59, 114, 329 Abrasion, 112, 131–4, 136–9, 142, 147, 155, 164, 245, 300–1 rates, 134, 142–5 stoss surfaces, 134, 136–9, 143, 145 Abrasion models, 134, 138, 145 Accretionary tectonics, 437 Accretionary wedge, 432–9 Accumulation, 48, 53–9, 61–7, 158–9, 362, 479 Accumulation area ratios (AAR), 169 Accumulation rates, 37–9, 368, 388 Accumulation zone (area), 59, 68, 83–4, 96, 112, 150–3, 159, 161, 163, 422 Acoustic sediment tracing, 409 Active ice flow, 193 Activity index, 65 Adfreezing, 69, 150–4, 159, 161 Advance sequence of glaciotectonic overprinting (style/model), 423, 442 Aeolian debris/dust, 167 Aeolian sediments, 161, 470 Aerosols, 50, 161 see also Atmospheric gases

Afghanistan, 21 Africa, 1, 4, 17–9, 21, 48 Aftonian interglacial, 26 Air bubbles (in glacier ice), 54 Alaska, 5, 23, 39, 58, 86, 91,96, 98, 151, 154, 168, 176, 185, 210, 274, 288, 309, 311, 326, 362, 363, 373, 385 Albedo, 47, 48, 57, 59 Alberta, 56, 87, 432, 436 Algae (red), 396 Aller¨od, 32, 38, 60, 282 Allokinetic discontinuities, 205, 210, 217, 461 Alluvial fans, 298, 436, 461 Altithermal, see Hypsithermal Amber ice, 69, 72, 155 Amorphous silica, 392 see also Silica Anchor ice, 383 Andes Mountains, 23, 34, 42, 44, 47, 151, 168 Anglian Glaciation, 179, 456, 469 Antarctica, 4, 5, 10, 21, 24, 34, 36, 38, 42, 44–5, 47, 52, 60, 62, 69, 72, 82, 93, 86, 89, 91, 93–4, 96, 113, 125, 159, 161, 167, 185, 195, 226, 362, 365, 366, 379, 381, 384, 393, 396, 463 Peninsula, 366 Ross Sea, 44, 366, 379 sub-Antarctic Islands, 44, 362 Antarctic diatom ooze belt, 397 529

Antarctic Ice Sheet, 42, 44, 60, 61, 88, 91, 112, 152, 238, 268, 379, 387, 389 East Antarctic Ice Sheet, 22, 60, 93, 125, 336, 379, 389 West Antarctic Ice Sheet, 23, 72, 82, 86, 90, 91, 93, 94, 185, 366, 387, 389 Anthropogenic influences, 16, 52 Anti-greenhouse effect, 21, 46 Ape Lake, British Columbia, 342 Appalachian Mountains, 26, 270 Aquifers, 103, 115, 124, 183, 188, 248, 278 Arapaho Glacier, Colorado, 163 Arboreal pollen, 34 Archean, 16, 17–8, 22, 401 Arctic diatom ooze belt, 397, 412 Arctic Ocean, 362, 393 Argillites, 400 Arid polar conditions, 203 Artesian conditions, 121, 263 see also Quasi-artesian conditions Arthur, Ontario, 450 Ash layers (volcanic), 297 Asia, 32, 42, 45, 62, 163, 288, 319 Asperities on glacier bed, 81, 82, 132 Astronomical theory of glaciation, 16–17, 28, 47–8, 50 Asymmetric isoclinal folds, 428, 435 Athabasca Glacier, Alberta, 87 Atlantic Ocean, 4, 19, 22, 32, 38, 48, 282, 380, 381, 389, 396, 403, 452

530

Atmospheric dust, 37, 38, 46, 72, 301 Atmospheric gases, 37, 52, 482 Carbon dioxide, 10, 17, 18, 19, 20, 21, 22, 37, 38, 46, 47, 48, 50, 52, 396, 399 Methane, 47, 52 Atmospheric transparency (transmissivity), 50 Australia, 1, 4, 18, 19, 21, 22, 61 Authigenic minerals, 13, 342 Autochthonous sediments, 384, 423, 429 Autokinetic discontinuities, 195, 210, 217, 460 Avalanches, 58, 91, 150, 157, 163 Axel Heiberg Island, Canada, 432

Back-pressure (in ice shelves), 89, 121, 246 Backwater zones, 292, 294 see also Slackwater Bacterial metabolism, 384 Badlands (Lake Missoula), USA, 293 Baffin Island, 72, 164, 243, 284, 286, 291, 309 Balance gradient, 57, 59, 60, 64, 65, 475, 479 Ball and pillow structures, 180, 209, 210, 347 Banding (structures) in sediment, 429, 431 Banham’s model of glaciotectonic deformation, 174, 181, 420 Barbados, 32 Barents Sea, 393, 402–14 Barents Sea Ice Sheet, 405 Barents Sea/Northern Norwegian Shelf, 404–8 Barents Sea Trough, 409 Barnacles, 396 Barnes Ice Cap, Baffin Island, 72 Bar & channel facies models for braided rivers and glacial outwash, 310–1 Barotropic flow, 372, 378 Bars, 286, 288, 292, 294–5, 298, 300, 302, 305–6, 308, 310–11, 313

INDEX

Basal cavities, 80, 85, 140 Basal crevasse, 101, 226, 233, 234, 236, 274, 369 Basal debris, 53, 69, 147, 149, 154, 157, 161–4, 180, 188, 223, 249, 267, 324, 327, 331, 363, 369, 373, 380, 385, 441 Basal hydraulic regime, 85 Basal ice-bed separation, 135, 140, 141, 145, 265 cavitation, 85, 94, 97, 256 no cavitation, 85 Basal ice flow, 75, 77, 222 thermal zonation/conditions, 53, 190, 193 velocity, 67, 82, 85, 92, 94, 118, 136, 190, 195 Basal meltwater, 97, 99, 114, 123, 231, 420, 422 Basal sliding (slip), 53, 62, 83, 85, 86, 87, 91, 93, 94, 97, 98, 101, 112, 154, 155, 389, 433 sliding with no separation (cavities), 85 sliding with separation (no cavities), 85 velocity, 112 Basal stress, 100, 113, 190 local stress concentrations, 77, 81, 147 shear stress, 58, 60, 80, 81, 82, 85, 87, 91, 92, 93, 125, 144, 308, 328, 420, 422, 423, 438 Basal thermal conditions, 53, 62, 66–9, 72, 94, 113, 185, 187, 219, 247, 480 see also Subglacial thermal conditions Basal till, 143, 402, 407, 432, 441 Basal water layer, 82 Basal water pressure, 92, 94, 97, 99, 101, 114, 116, 123, 130, 135, 136, 140, 141, 142, 143, 145, 150, 185, 363, 420 Basin(s) (glaciated), 62, 143, 241, 246, 293, 308, 323, 359, 384, 385, 422, 432, 480 Bavaria, Germany, 5, 25 Beaches, 180, 355, 356, 383 Bear Island Trough, N. Atlantic, 407 Beaufort Sea, Arctic Canada, 60

Bed/ice interface, 11,12, 63, 66, 67, 72, 82, 83, 85, 86, 87, 91, 112–18, 122–5, 150, 154, 185–90, 195, 210, 218–19, 223, 231, 232, 237, 238, 245, 278 Bed roughness, 85, 93, 118, 122, 142, 145 Bedding planes in sediment, 142, 145, 195, 210, 217, 306, 351 Bedform development, see Subglacial bedform development Bedload, 179, 281, 284, 291, 300, 302, 315, 343, 374, 377, 385 Bedrock forms, see Plucked bedrock forms & streamlined bedrock Bedrock fracture, see Rock mechanics Benthic fauna, 414 Benthonic formanifera, 28, 396, 397, 400, 412, 461 Cd/Ca ratios, 400 Berg dumps, 344 Berging , see Calving Bergschr¨und, 140, 143 Bergstone, 382 Bering Glacier, Alaska, 98 Berthelsen’s ‘empty’ locality, 461–2 Beryllium concentrations (10Be), 37, 50 Biferten Gletscher, Switzerland, 226 Bijou Creek type facies model, 313 Billingen moraine, Sweden, 328 Bimodal grain sizes, 164, 178, 300, 306 Bingham viscoplastic material, 86, 189, 419, 420 Bioclastic detritus (carbonate), 366, 379, 383 Biogenic opal production, 397–9 Biogenic sediment, 150, 379, 383, 396 Biogenic silica, 399 Biological limnology, 342–3 Biomass, 10, 400 Biosiliceous muds & oozes, 383, 384, 385 Biostratigraphy, 447, 453, 454 Bioturbated sediments, 412 Black Rapids Glacier, Alaska, 98 Blairgowrie, Scotland, 260 Blattnick moraines, 222, 333

INDEX

‘Block-in-matrix’ m´elange, 209 Bloomington moraine, Illinois, 326 Blown snow, 58 Bluck’s model of braided stream alluvial deposition, 298, 300, 309 Bluck’s model of lateral bar sedimentology, 309, 310 Blue Glacier, Washington State, 82, 92 B¨olling, 38, 472 Bolivia, 42 Bondhusbreen, Norway, 157 Bossons Glacier, French Alps, 297 Bottom waters, 342, 399, 412 Boudinage, 73, 210, 426, 428, 429, 439, 441, 442 Boulder beds, 16, 21, 200 Boulder berms, 294 Boulder lags, see Pavements Boulder lines, 383 Boulder rings, 480 Boulton’s Model of Abrasion, 138 Brackish water, 217, 379, 383 Braiding, 178, 246, 286, 288, 292, 294, 295, 306, 308–13, 458, 480 Brazil, 18, 21 Breccia, 205, 210 Brei–amerkurj¨okull, Iceland, 86, 174, 419, 421, 423 Brenva Glacier, Italy, 169 Brine formation, 399 Britain (UK), 5, 8, 16, 26, 47, 179, 246, 413, 447, 454, 468, 469, 470 British Columbia (BC), 56, 243, 363 Brittle deformation in rock, 132 Brittle failure (fracture) in sediment, 419, 428, 433, 437, 476 Br´uarj¨okull, Iceland, 98 Brunhes Chron, 25, 30, 462, 463 Brunhes/Matuyama reversal, 462, 463 Buckle folds, 428 ‘Bulldozing’ of sediment, 417, 436, 437, 439 Buried channels, see Subglacial meltwater channels Buried ice, 177, 263, 323, 480 Buried soils, see Paleosols Buried valleys, see Tunnel valleys Bylot Island, NWT, 270

Byrd Glacier, Antarctica, 89 Byrd Ice Core (Byrd Station), 10, 69, 72, 73

C-Channels, 116, 119 C-foliations, 428 Cadmium/Calcium ratios, 400 see also Cd/Ca ratios Cailleux-Tricart roundness index (RI), 301 Calabria, Italy, 24 Calcareous foraminifera, 28, 397, 399, 400, 402, 412, 413, 461 Calcareous tests, 28, 396 Calcium carbonate dissolution in sea water, 399 Calcite, 34, 396, 478 see also Vein calcite California, 34, 58 Calving, 44, 52, 58, 159, 235, 343, 348, 366, 368, 369, 371, 373, 377, 379, 381–3, 388, 389, 392, 402, 414, 417 Camp Century Core, 36, 77 Canada, 13, 17, 36, 39, 42, 47, 72, 88, 96, 119, 129, 151, 185, 190, 193, 223, 226, 237, 238, 248, 270, 274, 327, 328, 362, 400, 432, 436, 445, 449, 457, 479 Canadian Eastern continental shelf, 248 Canadian Shield, 62, 238, 241, 323, 470 Cant’s facies model of a sandy braided river, 311 Canterbury Plains, New Zealand, 288 Carbon dioxide, see Atmospheric gases Carbonate-diamicton association, 396 Carbonate flux, 397 Carbonate production & dissolution, 396, 397, 399 Carbonate sedimentation, 396–7 Carboniferous Period, 17, 21, 47, 435 ‘Card-house’ structures, 200, 217 Caribbean Sea, 28 Cascade Mountains, Washington State, 39 Cataclastic shear zone, 428 Cataclasis, 419, 420, 423, 428

531

Catastrophic discharge, 16, 232, 252, 282, 336 Catchment denudation rate, 161, 163, 283, 284, 286 Catfish Creek Till, 449 Causes of Glaciation, 15–52 Cavetto, 257 Cavitation, 94, 163, 256 Cd/Ca ratios, 400 Cenozoic, 15, 16, 21, 22, 24, 26, 28, 32, 34, 39, 42, 46, 47, 50, 361, 391, 392, 400, 402 Chalk banding in till, 429, 431 Chalk fragments, 180, 413 Channel fill, 313, 322, 400, 401 Chaotic bedding, 323, 347, 377 Chattermarks, 251, 278 Chemical weathering, see Corrasion Chibougamau Formation, Canada, 18 Chile, 223 China, 34, 86, 185, 187, 283, 421 Chlorine/beryllium ratios (36Cl/10Be), 37 Chronosequence, 25 Chronostratigraphy, 10, 24, 392, 447, 453, 454, 462, 470–3 Chute channels, 292 Circular disintegration ridges, 330, 333 Cirques and cirque glaciers, 58, 143, 147, 150, 153, 156, 163, 317 Clarke-channels, see C-Channels Clast fabric, 176, 178, 181, 195, 205, 209, 212, 217, 224, 228, 233, 235, 274, 302, 323, 392, 458, 467, 478 dip, 176, 200, 302 eigenvalues/eigenvectors, 203, 327 ‘herring bone’ patterns, 228 inherited, 200 micro-fabric, 199, 200 Clast imbrication, 306, 308, 311, 314, 347, 435 Clast ‘ploughing’, 86, 101, 173, 200 Clay minerals, 278, 428 CLIMAP (Climate: Long-range Investigation, Mapping And Prediction), 44 Climatic Optimum, 39, 45, 51, 453 Climatic warming, 4, 10, 13, 38, 45, 47, 52, 88, 359, 387

532

Climatostratigraphy, 453, 454, 462, 465, 470, 473 Climbing ripples, 347, 353, 354, 356 Clotted ice facies (or dispersed), 69 CO2 (Carbon dioxide), 10, 17, 18, 19, 20, 21, 22, 37, 38, 46, 47, 48, 50, 52, 396, 399 Coccoliths, 397, 402 Coccolithus pelagicus, 397 Colchester Formation, England, 460 Cold-based subglacial conditions (dry-based), 66, 68, 69, 72, 74, 93, 152, 154, 156, 157, 164, 180, 194 ‘Cold patches’, 67, 149 ‘Cold wave’, 66 Colorado, 163, 164 Columbia Glacier, Alaska, 86, 382 Columbia Plateau, Washington State, 293 Comma forms, see Sichelwannen Comminution, see Abrasion Compressional ridges, 344 Compressive flow (ice), 82–3, 94, 154, 270, 318, 331 Concentric (box) folds, 435, 442 Jura style, 442 Conchoidal fractures, 251, 278 Conjugate faults, 428 ‘Contorted Drifts’, 179, 180 Coquina (shells), 379 Coral, 28, 32 Cordilleran Ice Sheet, 39, 42, 44 Coriolis Forces, 4 Corrasion, 256, 257 Cosmic dust, 150 Coulomb failure criterion, 140, 172, 419, 438 Couplet, 305, 351, 356, 357 see also Varve Coupling line, 376 Cowan’s m´elange concept, 209 Crack growth in rock, 132–3, 139, 140–2, 145 Crag & tail, 234, 236–7, 240, 254 Crary Trough, Antarctica, 393 Crescentic furrows (gouges), 131, 133, 138, 237, 251 see also Sichelwannen Cretaceous, 22, 47, 436

INDEX

Crevasses, 74–6, 79, 83, 96, 97, 101, 104, 107, 110–12, 114, 143, 150, 156, 225, 226, 233, 234, 236, 246, 259, 267, 270, 274, 332, 366, 368, 369, 441 Critical lodgement index, 195 Critical particle size, 164 see also Terminal grade concept Cromer interglacial, 24, 26 Cromer Tills, 179, 180 Cross beds, 306, 311, 313, 314, 349 Cross-cutting lineations, 222, 249, 259, 262, 278 Cross laminations, 311 Cross-stratification, 309, 313 Cross-valley moraines, 234, 328, 377 Cryoconites, 168 Cryostatic pressures, 102, 172, 206 Crystalline slip (dislocation climb), 54, 80 Cumbria, England, 324, 325 Currents (in lakes), 216, 305, 335, 343, 344, 347, 349, 350, 351, 354, 355, 357 Currents (in oceans), 4, 47, 48, 88, 89, 216, 363, 366, 371, 377, 378, 380, 381, 382, 383, 384, 393, 396, 397, 408 Cyclopels, 372 Cyclopsams, 372

Dammer Berge push moraine, 432, 434 Dansgaard-Oeschger cycles, 38, 48 Darcy’s Law, 189 ‘Dead ice’, 97, 275, 331, 332, 346, 349 Dead Sea, Israel/Jordan, 34 Debris aprons, 348 Debris flows, 108, 125, 150, 152, 173, 177–8, 189, 210, 216, 303, 308, 313, 315, 319, 320, 322, 323, 326, 327, 347, 349, 369, 400, 402, 409, 417, 419, 423, 435, 448, 458 Debris grainfall, 369 Debris/ice concentrations, 68, 70, 73, 154, 155, 161, 162–3, 167, 251, 317, 325 see also Ice/debris concentrations Debris meltout, see Meltout

Debris (mud) ‘pockets’ or ‘clots’, 72 Debris-release mechanism, 83, 173, 365, 366, 381, 385, 480 Debris-rich ice bands, 69, 70, 72, 74, 77, 135, 136, 150, 154, 267, 315, 325 D´ecollement, 174, 420, 432, 435, 436, 437 Decoupling (basal ice/bed interface), 82, 94, 100, 189, 192 Deep sea cores, 15, 28, 30, 34, 37, 39, 42, 389, 396 Deep sea environment (conditions), 246, 399, 411–5 Deep sea records, 32, 34, 37, 44, 51, 392 Deep Sea Drilling Project (DSDP), 402 Deep water formation, 392–3, 396 Deformable beds (deforming sediment), 12, 63, 82, 83, 86–8, 92, 93–4, 123–5, 130, 166, 174, 185, 187, 189, 190, 193, 209, 226, 246, 476 Deformation structures, 174, 179, 180, 189, 322, 366 Deformation till, 409, 423, 429, 435, 441, 442 De Geer moraine, 218, 234–6, 328 Deglaciation (nature & style), 23, 32, 38, 42, 62, 106, 274, 277, 286, 316, 323, 325, 327, 356, 388, 392, 402, 407, 414 De Kalb mounds, 265 Deltaic foreset beds, 310, 313, 328, 347, 353, 354, 356, 458 Deltaic sediments, 215, 259, 262, 288, 327–8, 344–7, 353–6, 439, 458, 467 Deltas & estuaries, 215, 259, 262, 267, 288, 293, 327, 328, 330, 344–7, 353, 356, 366, 371, 374, 378, 383, 384, 402, 409, 421, 432, 436, 439, 443, 458, 466, 467 see also Glaciolacustrine Denmark, 32, 246, 322, 431, 460 Density mixing (homopycnal), 343 Deserts, 21, 362 Des Moines Ice Lobe, Laurentide Ice Sheet, 319

INDEX

Devensian glaciation (UK), 8, 24, 179, 454, 469 Devensian Ice Sheet, 100 Devils Hole, Nevada, 34–5, 38 Devon Island Ice Cap, Canada, 36, 72 Devonian, 17, 21 Dewatering structures, 195, 205, 210, 217, 347, 477 ball & pillow, 180, 209, 210, 347 flame, 210, 323, 347, 349 pipes & dish, 209, 347 Diachronic unit, 454, 470 Diagenesis, 12, 185, 278, 412, 449, 477, 478 Diagenesis of snow, 54, 74 see also Snow transformation Diamictite (tillite)1, 18, 42, 400–2, 475 Diamicton (diamict, till), 12, 67, 101, 123, 158, 165, 166, 171–80, 185, 194–205, 209–10, 215–17, 246, 263, 267, 272, 276, 278, 292, 318–30, 347, 349, 358, 366, 377–80, 382–4, 392, 396, 402, 405–9, 412, 413, 422–3, 441, 448, 451, 456, 458, 470, 475, 476, 477, 478 Diapirism, 205, 206, 217, 273, 421 Diatom microfossils, 351, 357, 379, 397, 399 Dike, 101, 228, 237, 323 see also Dyke structures Dilatant behaviour in sediments, 87, 139, 231 Diluvium, 16, 445 Dimlington Stadial (chronozone), 450, 454, 469, 470 Dirt cones, 167 Discharge: (ice) flux, 59, 65, 94, 153, 243, 245, 246 meltwater, 67, 82, 90, 91, 92, 97, 101, 104, 106, 107, 109, 112–16, 118–23, 125–9, 141, 176, 188, 206, 216, 323, 240, 248, 252, 262, 279, 281, 284, 286, 291, 328, 337, 340, 344, 347, 351, 353, 372, 374, 378, 379, 382, 385, 414 sediment discharge, 142, 157, 163, 379, 385

Dish & pillar structures, 209, 347 Disintegration ridges, 333 Dislocation climb, 54, 80, 84 see also Crystalline slip Distal lakes, 337–40, 342, 351, 352, 353, 449 Divergent ice flow, 226 Dolostones, 20 Dome C (Circe) Ice Core (Greenland), 36 Donau glaciation, 25 Don River, CIS, 42 Donjek type facies model, 313 Drag coefficient, 134 Drake Passage, 22, 23, 47 Draped laminations, 354 ‘Draw-down’ effects, 388 Dropstones, 180, 217, 344, 349, 358, 380, 383, 392, 400, 402, 407, 412, 413 Drumlins, 13, 62, 193, 195, 218, 222–3, 224, 225–32, 233, 247, 254, 257, 261, 267, 275, 278, 288, 442, 467, 468 Drumlinoids, 224, 228, 231 Dryas octapetala, 32 Dryas Stadial (Younger ), 32, 38, 45, 50, 450, 452, 463, 466 Dry Valleys, Antarctica, 195 Ductile extrusion, 215 Ductile shear, 428, 437 Ductile structures, 428, 423, 437 Dunes (fluvial), 178, 295, 305, 306, 313, 314, 315, 347 Dwyka Formation, South Africa, 21 Dye, 3 Ice Core, Greenland, 36 Dye tracers, 123 Dyke structures, 101, 228, 237, 323 Dynamic balance line (DBL), 97, 98

Earthquakes, 151, 384, 409 East Anglia, 26, 179, 181, 431, 437 Edinburgh Castle, Scotland, 236, 237 Eemian Interglacial (Eem) (n.b. Ipswichian), 26, 32 Effective stress, 12, 80, 82, 87, 108, 124, 188, 210, 231, 246, 251, 277, 418, 419, 420 Eigenvalues (-vectors), 203, 322

533

Eistektonik, 418 see also Glaciotectonics Ekalugad River, Canada, 284 ELA (equilibrium line altitude), 56–60, 84 Ellesmere Island, Canada, 185 Elliot Lake, Ontario, 401 Elster glaciation, 8, 25, 42 End moraine, 25, 26, 270, 322, 325–9, 332, 377, 435, 443 Enderby Land, E. Antarctica, 60 ‘Energy of glacierization’, 65 Englacial: environments, 479 channels (conduits), 157, 178, 262–3, 279, 281, 317, 346, 372 debris, 68, 69, 73, 74, 82, 152, 153, 154, 156, 157, 163, 164, 167, 171, 176, 203, 205, 210, 236, 317, 363, 365, 369, 373, 380, 423 debris meltout, 200 hydraulic systems, 101, 102, 105, 106, 108–12, 113, 114, 116, 120, 121, 122, 143, 158, 166, 281 lakes, 336 temperatures, 65 England, 174, 179, 445, 458, 459 Eocene, 22 Epigenetic deformation, 460, 461 Epilimnion, 341, 343, 350, 353, 355, 357 Equilibrium line, 56–60, 82, 84, 153, 156, 163, 361, 362, 363, 388, 422 Equipotential planes of water pressure, 119, 121 see also Potentiometric surface Erdalsbreen, Norway, 286 Erosion, rates of, 142, 143, 144, 162, 245 Esker(s), 62, 112, 193, 206, 265–73, 288, 328, 374, 378, 479 Estonia, 226 Europe, 1, 4, 5, 8, 21, 22, 24, 25, 26, 28, 32, 39, 42, 45, 46, 60, 83, 169, 246, 319, 420, 421, 423, 429

534

European Alps, 4, 5, 15, 24, 25, 39, 42, 773, 83, 125, 282, 297, 309, 325, 431, 432, 436 European Pollen Zones, 32, 38 Eustatic changes in sea and ocean level, 13, 90, 91, 316, 388, 389 ‘Event stratigraphy’, 453 Extending flow (ice), 83, 156, 223, 422 Extraterrestrial sediment, 152

Fabric, 176, 178, 181, 195, 200, 203, 205, 209, 212, 217, 224, 228, 233, 235, 236, 274, 302, 322, 323, 392, 458, 467, 478 see also Clast fabric Facies coding system classification, 303, 305 Miall’s, 303, 305, 306, 312, 313 Rust’s, 311, 313 Facies Models, 311, 313 ‘Faint’ young sun, 16, 17, 19, 22 Fan deltas, 262, 286, 328, 374 Faults: bedrock, 34, 241, 242, 246, 249, 418 sediment, 179, 180, 205, 206, 210, 212, 217, 233, 322, 323, 330, 419, 420, 421, 428, 437, 439, 460, 480 Fennoscandian Ice Sheet, 13, 60, 88, 152, 270 Fennoscandian Shield, 238 Filchner Ice Shelf, Antarctica, 89, 393 Film lubrication layer, 82, 85, 91, 93, 114, 116, 118, 135, 142, 150, 231, 366 see also Basal water layer Finland, 18, 222, 223, 226, 252, 274, 328, 466 Fiords, see Fjords Firn, 54, 65, 107 Firn line, 156 Fissile structures, 195, 407 Fission track-tephrochronology, 34, 39 Fjord, 62, 235, 237, 241–6, 378, 379, 384, 385, 387, 421 Flame structures, 210, 323, 347, 349

INDEX

Flickers (dust fluctuations), 38 Floating glacier termini, 56, 60, 88–90, 119, 120, 121, 149, 150, 158, 185, 213, 215, 216, 361, 363–78, 379, 380, 381, 382, 388, 402 Flocculation, 217, 379 Floodplain deposition, 301 Floods, see J¨okulhlaups Florida, 21 Flow lenses, 217 Flow till, 12, 108, 173, 177–8, 210–15, 219, 233, 319, 418, 423, 435, 478 Fluid stressing, 256 Fluidization, 177, 206, 210, 256, 262, 323, 400 Fluted moraine, 147, 195, 222, 223, 226, 230, 233–4, 276, 478 Fluvioglacial sediment, see Glaciofluvial sediment Folds (in ice), see Ice tectonics Folds (in sediment), 154, 155, 180, 181, 206, 212, 217, 224, 233, 322, 323, 332, 400, 407, 417, 419, 420, 421, 423, 424, 426, 428, 429, 432, 433, 434, 435, 436, 437, 439, 441, 442, 460, 480 Foraminifera, 28, 396, 397, 399, 400, 402, 407, 412, 413, 461 Forbes bands (ogives), 77 Formation of glacier ice, 54–5 Fossil content of sediment, 9, 24, 26, 343, 351, 357, 392, 399, 407, 453, 461 Fossil fuels, 52 Fractures (in ice), see Ice tectonics Fram Strait, 397 France, 73, 297, 325, 454 Frazil ice, 383 Freeze-thaw processes, 13, 246 French Alps, 73, 297, 325 Friction cracks, 254 Frictional heat, 116, 190, 192, 195, 267 Frictional wear, 134 Frost cracking (shattered debris), 139 Frozen bed conditions, see Subglacial thermal conditions

Frozen sediments, 13, 61, 62, 63, 67, 83, 101, 115, 148, 150, 154, 188, 233, 277, 380, 420–1 ‘Full-pipe flow’, 262, 478 Furrow, 120, 131, 257–8, 384

Gastropods, 343 Gauss-Matuyama paleomagnetic boundary, 42, 462 Geochemical processes (subglacial), 278, 478 Geochronometry, 446, 447, 453, 454, 470–1 George VI Ice Shelf, Antarctica, 363 Geotechnical properties of sediment, 12, 85, 148, 209, 210, 456, 476, 477, 478 Geothermal heat (flux), 59, 64, 65, 67, 82, 91, 103, 114, 116, 136, 137, 176 German Alpine Foreland, 25 Germany, 5, 16, 25, 226, 246, 327, 421, 423, 429, 432, 442 ‘Gilbert’-type delta, 328, 347, 353 Gipping Glaciation, 8 Glacial chutes, 262–3 Glacial Lake Agassiz, see Lake Agassiz Glacial mill, see Moulin Glacial outburst floods, see J¨okulhlaups Glacial outwash, 5, 25, 26, 108, 158, 161, 248, 275, 286, 288–94, 298, 300, 301, 302, 303, 306, 308, 309, 310–5, 327, 329, 340, 347, 352, 374, 378, 402, 405, 432, 436, 442, 443, 456, 459 Glacial overdeepening, 119, 120, 143 Glacial overprinting, 62, 190, 230, 262, 277, 278, 412, 423, 426, 442 Glacial polish, 16, 138, 142 Glacial thrust terrains, 423 Glaciated basin(s), 62, 143, 241, 246, 293, 308, 384, 422, 432, 480 Glacial hydrology, 102–30 Glacial isostasy (glacio-isostasy), see Isostatic readjustment Glacial stratigraphy, 445–73

INDEX

Glacial surface profile, 63, 80, 100, 209, 368, 422, 438, 439 Glaciated Terrain-Ice Sheet Model, 62–3 see also GT-IS Model Glaciated trough(s), 62, 143, 241, 242, 243, 244, 245, 246, 293, 308, 384, 405, 422, 432, 480 bedrock sills, 246 cross profile, 87, 88, 143, 243, 244 longitudinal profile (thalweg), 143, 241, 254, 388 riegels, 241, 246 Glacier Bay, Alaska, 5, 58, 373 ‘Glacier burst’, see J¨okulhlaups Glacier sensitivity, 55, 57, 168, 282 Glacier (ice sheet) surges, see Surges Glacier tectonics, see Glaciotectonics Glacier d’Argenti´ere, French Alps, 73 Glacio-eustasy, 13, 90, 91, 316, 388, 389 Glaciofluvial deposition/ sediments, 25, 171, 178–9, 205, 209, 233, 237, 247, 259, 262, 263–73, 277, 288, 301, 321, 323, 330–3, 347, 380, 423, 432, 435, 478, 479 Glaciogenic sediment flows, see Flow tills Glacio-isostasy, 12, 13, 60, 62, 63, 90, 129, 267, 288, 316, 322, 335, 337, 388, 405, 414, 418, 475 Glaciolacustrine environments, 335–59 Glaciomarine environments, 391–415 Glaciotectonics, 417–43 Glen’s Law, 75, 80, 84, 94, 419 Global climates, 4, 9, 10, 34, 50, 68, 389 Global cooling, 47, 48, 52, 388 Gondwanaland, 17, 21, 22 Gowganda, Huronian, 1, 400–2 Grain size (sediment), 123, 138, 155, 164, 178, 179, 190, 194, 195, 210, 216, 233, 263, 286, 291, 295–6, 297, 300, 301, 302, 306, 321, 330, 392, 450, 456 Grain size (snow), 54 Grampian Mountains, Scotland, 276 Grand Banks, N. Atlantic, 26

Grand Banks Slide, N. Atlantic, 409 Gravel lag, 233, 313 Great Basin, western USA, 34 Great Britain, see Britain Great Lakes (Canada/USA), 39, 54, 327, 449, 480 Greece, 34 Green Bay Lobe, Wisconsin, 325, 326 Greenhouse effect, 17, 47, 52 Greenhouse gases, 16, 38, 47, 52 Greenland, 4, 5, 10, 22, 23, 36, 38, 39, 44, 48, 53, 59, 61, 72, 73, 91, 94, 102, 113, 167, 226, 238, 288, 309, 362, 381, 389, 462, 463 Greenland Ice Core Project (GRIP), 10, 36, 37, 38, 463 Greenland Ice Sheet, 4, 23, 36, 38, 44, 59, 61, 72, 73, 94, 113, 167, 462, 463 Greenland Ice Sheet Project, 2 (GISP2), 36, 37, 38, 39, 45 Greenland-Norwegian Sea, 48, 393, 396, 397, 402–14 Grimsv¨otn, Iceland, 125 GRIP, see Greenland Ice Core Project Grooves, 62, 131, 138, 165, 240, 249, 293 Grounding line, 13, 60, 62, 81, 88, 89, 90, 91, 121, 150, 158, 159, 161, 183, 185, 193, 215, 216, 217, 234, 236, 246, 263, 363, 365, 366, 369, 371, 373, 374–8, 380, 381, 382, 383, 384, 385, 387, 388, 389, 402, 458, 480 Ground moraines (non-linear unoriented forms), 220, 326–9, 332 Blattnick moraines, 222, 333 hummocky moraine, 273, 275–7, 332, 333, 467 Pulju moraines, 222, 273–5 Veiki moraines, 265, 273, 275 Groundwater, 4, 183, 278, 291, 333, 340 GT-IS Model (Glaciated Terrain-Ice Sheet Model), 62–3 Guelph-Paris moraine, Ontario, 325 Gulf of Bothnia, 449 G¨unz glaciation, 5, 25

535

H-beds (Hard or Rigid subglacial beds), 83, 92, 94, 99, 114–6, 123, 188–9, 190, 192, 193, 231, 238, 252 Hallet’s Model of Abrasion, 134 Happisburgh, England, 179 Hazard Lake, Yukon Terr., 129 Headwall, 143, 156 ‘Heat pump’ effect, 149 Heinrich ‘events’, 38 Helicoidal flow (in ice) (cork-screw), 231 Hemipelagic sediments, 374, 377, 381, 382, 385 Hemipelagic suspension, 378–9, 385 Hertzian contact, 133 Himalayan Mountains, 47, 168, 169, 309 Holmstr¨ombreen moraine, Svalbard, 432–5, 442 model, 433 Holocene, 24, 28, 32, 37, 38, 39, 45–6, 48, 50, 51, 52, 167, 357, 396, 399, 404, 409, 453 Holostratotype, 455 Holstein interglacial (n.b. Hoxnian), 26, 42 Homopycnal flow (density mixing), 343 Hoxnian interglacial, 26, 42 Hudson Bay, Canada, 18, 42, 62, 193, 449 Hummocky moraine, 273, 275–7, 332, 333, 467 Huronian Supergroup, Canada, 17, 18 Hurwitz Group, Canada, 18 Hydraulic conductivity, 124, 188, 189 Hydraulic gradient, 178, 188, 273 Hydraulic pumping, 217 Hydraulic seal, 125 Hydro-electric power (HEP), 4 Hydrographic processes, 371–2, 383–4, 385 Hydrographs, 126, 281 Hydrology of glaciers, see Glacial hydrology Hydrolysis, 141 Hydrostatic pressure, 113, 178, 189, 247, 248, 252, 262, 263, 270, 273 Hyperconcentrated flow, 288

536

Hypolimnion, 341, 343 Hypopycnal, 343, 372, 377 Hypsithermal, 39

Ice (debris-rich), 70, 72, 135, 136, 150, 315 stagnant, 97, 101, 110, 129, 163, 169, 171, 172, 176, 177, 187, 194, 225, 263, 276, 323, 331, 332, 333, 441 Ice/bed decoupling, 94, 189 Ice/bed interface, see Subglacial Icebergs, 16, 38, 159, 216, 348, 351, 353, 366, 371, 373, 378, 379, 380, 381, 382, 383, 392, 399, 404, 407, 408, 412, 413, 414, 417 Iceberg calving, see Calving Iceberg rafted debris (IBRD), 373, 379, 380, 381, 382, 383 Ice-cliff margin, 77, 156, 185, 344, 348–9, 366, 373 Ice-contact delta, 267, 327, 344–7, 353, 371, 374, 389, 466, 467 Ice cores, 10, 34–9, 44, 45, 48, 50, 72, 73, 167, 389, 463, 471 Ice-cored landforms, 168, 349 Ice creep (glide), 79, 88, 167 Ice crystal(s) (shape, size), 36, 54–5, 66, 69, 72, 77, 79, 80, 84, 154, 167, 383 Ice-dammed lakes, 97, 122, 125, 129, 130, 351, 467 Ice/debris concentrations, 55, 66, 68, 74, 119, 154, 167, 262, 385 Ice density, 54, 80, 149 Ice disintegration forms, 265, 273, 274–7, 330–3 Pulju moraines, 222, 273–5 Ice divide, 60, 222, 270, 273 Ice fabric, 55, 66, 77 Ice facies types (zones), 66, 72, 154, 163 Ice falls (seracs), 63, 77, 107, 373 Ice flow dynamics, 37, 64, 76 ice sheets, 79–81, 82–3 ice shelves, 88–90 valley glaciers, 87–8 Ice flux (discharge), 59, 65, 94, 153, 243, 245, 246

INDEX

Icehouse Climatic State, 17 Ice keel scouring, 384 see also Keel marks Iceland, 1, 5, 23, 86, 100, 122, 125, 163, 165, 174, 185, 226, 283, 288, 294, 297, 309, 313, 362, 419, 420, 421, 427 Ice loading, see Isostatic readjustments Ice marginal lakes, 335, 337, 342 Ice proximal deposits, 344–9, 379, 396, 401–2 Ice rafting, 28, 32, 42, 180, 308, 347, 349, 374, 379–82, 383, 389, 392, 413, 414 ice rafted debris (IRD), 379, 380, 392, 396, 400, 401, 402, 403, 404, 407, 412 ice shelf rafted debris (ISRD), 379, 380 sea ice rafted debris (SIRD), 379, 383 Ice rise(s) (rumples), 81, 88, 89 Ice sheet modelling, 60–3 Ice shelves, 13, 23, 60, 81, 82, 88–90, 91, 121, 149, 150, 153, 158–61, 213, 246, 363, 365, 366, 378, 379, 380, 381, 382, 383, 391, 392, 393, 402 Ice side-wall friction, 81, 82, 130 Ice stratigraphy, 72 Ice Stream B, W. Antarctica, 82, 86, 91, 185 Ice Stream C, W. Antarctica, 91 Ice streams, 60, 61, 62, 63, 74, 81, 82, 86, 87, 91, 92, 93, 94, 96, 107, 153, 156, 163, 185, 192, 226, 240, 242, 270, 366, 441, 468, 477 Ice velocity: balance, 91–4, 103, 420, 422 basal, 91–4, 119, 144 fast, 63, 82, 91, 92, 93, 94 surging, 94–6 surface, 67, 87, 91 Ice-walled lakes, 333 Ice wedges, 292 Illinoian glaciation, 8, 26, 39, 42 Illinois, 39, 248, 326, 327, 465 Imbrication (of clasts), 178, 284, 286, 302, 306, 308, 311, 314, 435

Immobile sediments (within a deforming bed state), 85, 87, 124, 155, 166, 174, 231 Incompetent strata, 418, 432 India, 4, 16, 18, 21 Talchir boulder beds, 16, 21 Indiana, 248 Injection structures, 200, 206 Inner channel, see Chute channel Insolation curves, 30, 43, 51 Insulating effect of supraglacial debris on ablation rate, 168 see also Supraglacial environments Interflow, 343, 356, 366, 379 Interglacial periods, 8, 9, 16, 24, 25, 26, 28, 32, 34, 36, 37, 38, 42, 44, 47, 50, 51, 60, 391, 392, 393, 396, 397, 399, 400, 402, 405, 409, 411, 412, 413, 414, 445, 452, 453, 454, 463, 471, 477 Interlobate moraines, 325 Internal deformation of ice, 54, 75, 83, 84–5, 90, 93, 94, 112, 176 International Decade of Ocean Exploration, see CLIMAP project International Stratigraphic Guide, 453, 471 Intraclasts, 101, 205, 217, 228, 235, 247 Inverse grading, 308, 323 Iowa, 26, 39, 319, 322 Ipswichian Interglacial, 26 Iran, 21 IRD, see Ice rafted detritus Ireland, 13, 226 Isostatic readjustments (rebound, depression), 12, 13, 60, 62, 63, 129, 267, 288, 316, 322, 335, 337, 388, 405, 414, 418, 475 Isotope Stages, 28, 32, 34, 37, 38, 39, 42, 44, 179, 454, 462, 463 Italy, 24, 169 Itarar´e Formation, South America, 21 Ivory Glacier, New Zealand, 163

Jakobshavn Isbræ, west Greenland, 91, 92 Japan, 34

INDEX

Johnstown End Moraine, Wisconsin, 325, 326 Joints (rock), 80, 131, 139, 140, 142, 145, 166, 252, 259, 447 Joints (sediments), 195, 200, 210 J¨okulhlaups, 12, 101, 125–30, 252, 281, 286, 293, 294, 313, 314, 315, 336, 351–2, 449, 478, 480 J¨okulhlaup initiation mechanisms, 129–30 Jurassic, 22 Juvenile water, 103

Kalix Till, Sweden, 218 Kame(s), 62, 112, 265, 275, 276, 288, 322, 330, 418, 465, 479 Kame terraces, 330, 467 ‘Kame and Kettle’ Topography, 257, 276–7 Kansan glaciation, 8, 26 Karakoram Mts., 168, 169 Katabatic winds, 343 Keewatin District, NWT, 223, 237, 267, 270 Keewatin Ice Divide, 270 Kemp Coast, E. Antarctica, 60 Kenai Peninsula, Alaska, 362 Kenya, 58 Kesgrave Group, England, 459 Kettle holes, 206, 275, 276, 292, 293, 305, 313, 315, 326, 328, 329, 330, 337 Kinematic wave, 53, 60, 63, 78, 96 Kinetic sieving, 155, 178 Kineto-stratigraphy, 456, 460–1 Kink bands, 212, 428 Kirkham moraine, Cumbria, 325 ‘Knock and lochan’ topography, 238

Labrador, 60, 88, 152, 223 Labrador Current, 381 Labrador Sea, 48 Lacustrine sediments, 209, 263–5, 322, 326, 359, 480 Lake Agassiz (glacial), Manitoba, 282, 449 Lake Blane (glacial), Scotland, 467 Lake Bonneville, Utah, 356

Lake bottom environments & sediments, 343, 344, 347, 349–50, 356 Lake Michigan, Mich./Ont., 327 Lake Missoula, Montana, 293 Lake Ontario, Ont./NY, 457 Lake Ontario Ice Lobe, 325 Lake Superior, Canada/USA, 18, 400 Laminated clays, 203, 358, 407 ‘Laminated Diamicton’, East Anglia, 179, 180 ‘Laminated ice’, 69 Laminated silts, 448, 458, 469 Laminae, 180, 195, 200, 203, 205, 265, 305, 311, 313, 315, 321, 322, 342, 347, 350, 354, 356, 358, 366, 372, 392, 400, 401, 407, 428, 448, 458, 469 varve, 305, 351, 357, 400 Laminites, 358 Lamstedt push moraine, Germany, 441 Landsat Imagery, 10 Landsliding and surging, 421, 423 Lapse rates, 57 Larsen Ice Shelf, Antarctica, 365 Last Glacial Maximum (LGM), 32, 44, 62, 454, 463, 468, 469 Late Cenozoic glaciation, 16, 21, 24, 26, 32, 46, 47, 361 Late summer runoff, 279 Latent heat: freezing, 65 melting, 142 Lateral moraines, 107, 147, 155, 156, 325, 329 Lattice defects (ice crystals), 84 Laurasia, 21, 22 Laurentide Ice Sheet, 13, 26, 38, 42, 60, 62, 67, 68, 100, 174, 190, 270, 282, 293, 319, 325, 479 Law of Superposition, 460 Lawson Type Flows (flow tills), 178, 213–4, 322, 323 Lead 210 chronologies, 37, 351, 357 Lebenspuren, 351 Lectostratotype, 455 Levene moraine, Sweden, 328 Lewis River, Canada, 284 Lichenometry, 468 Limnology, 337, 340–3

537

Linked-cavity model/system (Kamb), 104, 113, 114, 116, 122–3 Liquefaction of sediments, 177, 206, 210, 369, 441 Lithofacies associations, 11, 13, 54, 73, 130, 148, 193, 210, 224, 303, 312, 313, 315, 349, 374, 377, 412, 446, 452, 456, 460, 465 Lithostratigraphy, 407, 449, 450, 453, 454, 455–61, 467, 470, 471, 473 Little Ice Age, 39, 45–6, 432 Loch Lomond Basin, Scotland, 450, 470, 471 Loch Lomond Stadial (Readvance), 450, 466, 467, 470, 471 Lodgement process of till deposition, see Till critical lodgement index, 195 Lodgement till, 140, 174, 179, 181, 195, 196–200, 205, 206, 209 Lodgement tillite, 400 Loess, 32–4, 42, 283, 288, 297, 359, 453 Lonestones, 379 Longitudinal compression, 82–3, 97, 425, 426, 436, 438 Longitudinal extension, 77, 82–3, 97, 425, 426, 436, 438 Longitudinal strain rates: in ice, 82–3, 97, 98, 112, 125 in sediment, 425, 426 Lowestoft Formation, Britain, 456, 458 Lowestoft Glaciation, Britain, 8 Lowestoft Till, 174, 179, 181 Lunate fractures, 251 Luochuan, China, 34 Lyngsdalselva, Norway, 281 Lysocline, 403

M-beds (Mobile or Soft subglacial beds), 83, 84, 87, 92, 94, 114, 115, 123, 188, 189, 190, 193, 231, 249 McMurdo Ice Shelf, 161 Macrofauna, 379 Magnetic signature, 20 see also Paleomagnetism

538

Magnetostratigraphy, 392, 456 Maine, 328 Maine School of ice sheet modelling, 61–2 Malaspina Glacier, Alaska, 363 Malaspina Lake, Alaska, 342 Manganese nodules, 342 Manitoba, 193 Marengo moraines, Illinois, 326 Marine-ending glacier termini, 363–78, 385, 387–9 Marine fan sedimentation, 328 Mass balance, 2, 10, 55–7, 58–9, 60, 62, 63, 64–5, 67, 82, 88, 90, 91, 92, 94, 103, 114, 151, 163, 168, 169, 279, 388, 417, 422, 423, 475, 479 Massachusetts, 143 Mass wasting, 178, 233, 319, 322 Matanuska Glacier, Alaska, 86, 154, 156, 176, 178, 210, 326, 344 Matuyama Chron, 30, 32, 42, 462, 463 Matuyama/Gauss reversal, 42, 462 Matuyama Paleomagnetic Chron, 30, 32 Maunder Minimum, 50 Medial moraine, 96, 154, 156, 157, 325, 329 see also Moraine Medial moraine septa, 154, 156, 163 see also Debris septa Medieval warm interval, 50 Medvezhiy Glacier, Tadzhikstan, 91, 97–8 Mega-blocks of sediment, 429 Megadrumlins, 230 Megaflutes, 442 Megaripples, 311, 313, 315 M´elange, see Diamicton m´elange Melbourne School of ice sheet modelling, 61 Melting glacial bed conditions, see Subglacial thermal bed conditions Meltout till, 171, 176–7, 180, 195, 196–200, 200–5, 209, 210, 219, 233, 262, 263, 273, 278, 320, 369, 429 sublimation till, 195

INDEX

Meltwater, see Subglacial meltwater channels discharge, 67, 82, 90, 91, 92, 97, 101, 104, 106, 107, 109, 112–16, 118–23, 125–9, 141, 176, 188, 206, 216, 323, 240, 248, 252, 262, 279, 281, 284, 286, 291, 328, 337, 340, 344, 347, 351, 353, 372, 374, 378, 379, 382, 385, 414 erosion marks, 193, 254 jets (buoyant), 121, 263, 372, 378 see also Plume P-forms, 240, 247, 251, 252, 256 pipe flow, 101, 115, 123, 178, 203, 262, 479 potholes, 5, 131, 252, 257, 259, 262, 293 S-forms, 256, 257, 262, 273 Meltwater channels, 124, 193, 238, 252, 259–63, 273, 281, 291, 294–5, 311, 329–30, 402, 465, 467, 468 ‘C’ channels (subglacial), 119 closure rates, 114, 121, 122, 262 englacial, 263, 279, 346 gradient, 263 longitudinal profile (thalweg), 254, 263 ‘N’ channels (subglacial), 101, 119 plumes, 4, 121, 180, 216, 343, 372, 377, 378, 379, 383, 397, 405, 414 ‘R’ channels (subglacial), 119, 122, 124 till channels, 124 tunnels, 67, 68, 77, 101, 121, 150, 171, 178, 189, 205, 241, 246–9, 262, 263, 267, 270, 273, 281, 346, 347, 369, 442, 479 Meromixis, 342 Mesozoic, 22, 47 Metalimnion, 341, 353 Metamorphoses of snow, see Snow transformation Meteorites, 150 Miall’s facies model types, 303 Michigan, 18, 42, 100 Midland Valley, Scotland, 450

Milankovitch Theory (effect, forcing), 16, 25, 28, 46, 47–8 Mindel glaciation, 5, 25 Mini-surges, 98, 113 Miocene, 22, 23, 39, 42 Missouri, 26 Missouri Coteau, 432 Mixing (water turnover in lakes), see Water body turnover Mobile (soft) beds, see M-beds Molluscs, 24, 396 Monomictic lakes, 342 Montana, 17, 100 Moraine(s): ablation, 276 annual, 125 Blattnick, 222, 333 Cross-valley, 234, 328, 377 De Geer, 218, 234–6, 328 End, 25, 26, 270, 322, 325–9, 332, 377, 435, 443 Fluted, 147, 195, 222, 223, 226, 230, 233–4, 276, 478 see also Fluted moraine frontal-dump, 377 hummocky, see Hummocky moraine interlobate, 325 lateral, 107, 147, 156, 325, 329 medial, 96, 154, 156, 157, 325, 329 Pulju, 222, 273–5 push, 101, 327, 417, 420, 421, 422, 423, 431, 432–6, 439, 442, 443 Ra, 328 ramp-type, 377 ribbed, 223 Rogen, 195, 218, 222, 223–5, 226, 230, 233, 234, 349, 377, 478 Sevetti, 265 Terminal, 62, 248, 288, 294, 325 Veiki, 265, 273, 275 Washboard, 234, 328 Morphostratigraphy, 25, 446, 453, 465–8, 470, 471 Moscow, Russia, 435 Moulin (glacial mill), 75 Mount Kenya, 58 Mud flows, 303, 306 Mud pellets, 380

INDEX

Mudstones, 400, 401 Mueller Glacier, New Zealand, 158 Muir Glacier, Alaska, 373, 385 Muldrow Glacier, Alaska, 98 Munich, Germany, 25 Muschelbr¨uch, 257 Myrdalsj¨okull, Iceland, 226, 293, 314

N-channels, 101, 116, 119, 189 see also Nye channels Nappes (in soft sediment), 101, 423, 428, 432, 433, 434, 435, 436, 441, 442 Nebraska, 26 Nebraskan glaciation, 8, 26 Neogene, 47, 409 Neoglacial, see Little Ice Age Neogloboquadrina pachyderma, 397, 399 Neotetonics, 477, 479 marine slide triggers, 409 Net mass balance, 58–9, 65, 67, 91 Netherlands, 428, 432, 435, 442 Nevada, 34 N´ev´e, 54, 65 New Brunswick, Canada, 447 New England, USA, 42 New Guinea, 44 New Hampshire, 165 New York State, 42, 60 New Zealand, 21, 42, 44, 151, 158, 163, 166, 168, 169, 238, 283, 288, 309 Non-coaxial deformation, 424 Non-laminar flow in ice, 75–7 Non-pervasive deformation, 189 Norfolk, England, 179, 423, 458 North America, 5, 8, 16, 17, 18, 19, 21, 23, 24, 26, 28, 32, 39, 47, 60, 61, 62, 97, 174, 270, 319, 323, 327, 333, 357, 420, 421, 423, 429, 432, 449, 452, 453, 468 North American glacial chronology, 39–42 North American Stratigraphic Code (NASC), 453 North Atlantic Ocean, 19, 22, 32, 38, 48, 282, 380, 381, 389, 403, 452 North Atlantic Deep Water Flow (NADW) (Current), 48

North Dakota, 223, 228, 432 North European glacial chronology, 25–6 North Sea, 26, 246, 480 North Sea Drifts, 179, 180, 458 North West Territories (NWT), 119, 193, 237 Northern Ireland, 13, 226 Norway, 16, 45, 69, 152, 157, 223, 281, 286, 328, 407, 408, 466 Norwegian Current, 397 Norwegian-Greenland Sea, 393, 396, 402–14 Nova Semlya, Russia, 362 Nunatak, 150, 152, 213, 317 Nye channels (N-channels), 101, 116, 119, 189 NWT, see North West Territories

Obliquity, 16, 17, 20, 22, 30, 48 Ocean circulation, 9, 11, 46, 47, 52, 366 ODP Leg, 104, 402 Ogives, 74, 77 see also Forbes bands and wave ogives Oligocene, 22, 42 Ontario, Canada, 12, 18, 42, 60, 223, 228, 246, 323, 325, 328, 400, 445, 449, 450, 457 Opal (biogenic) production, 397–9 Orbital periods, 28, 48 Ordovician, 1, 21, 22 Organic carbon, 384, 392, 412 Orifice, 122, 123 see also Linked-cavity system Osar, see Esker Østerdalisen, Norway, 69 Outwash (sandur), 5, 25, 26, 108, 158, 161, 163, 248, 270, 275, 286, 288, 291–5, 298, 300–11, 313–15, 327, 329, 337, 340, 347, 352, 374, 378, 402, 405, 432, 436, 442, 443, 456, 459 Overconsolidation, 174, 200, 407 Overdeepened glacial basins, 119, 120, 143, 262, 422, 442 Overflow (hypopycnal), 343, 350, 356, 372, 378 Overprinting, see Glacial overprinting

539

Overriding by ice, 101, 150, 154, 185, 206, 288, 294, 319, 326, 349, 436, 442 Oxygen Isotope Ratios, 28, 32, 34, 37, 38, 39, 42, 44, 179, 454, 462, 463 Oxygen Isotope Stratigraphy, 26–31, 391, 456, 461–5

P-forms, 240, 247, 251, 252, 256 Pacific Ocean, 4 Pack ice, 399 Pakitsaq, Greenland, 36 Paleolatitudes (low), 17, 20, 22 Paleomagnetism, 17, 20, 22,25, 26, 34, 39, 42, 357, 359 Paleosols, 26, 39, 451, 477 Paleozoic, 17, 21, 391 Panama Isthmus, 47 Parastratotype, 455 Particle size distribution, see Grain size distribution Particle trajectory, 149, 152–4, 159, 161 Passive ice flow, 173, 203, 259, 263–5, 273–7 landforms, 263–5, 273–7 Patagonia, 23, 42, 44, 223, 238, 288, 363, 463 Patagonian Andes Mountains, 23, 42 Pedological processes, 449, 477 Pedostratigraphy, 454 Pelagic calcareous tests, 28, 396 Pelagic carbonates, 397, 399, 402 Pelagic sedimentation rates, 32, 385 Perched aquifer, 248, 356 Percolation zone, 65, 66 Percussion cracks & erosional forms, 249–51 Periglacial environments/processes, 276, 468, 477 Permafrost, see Frozen sediments Permafrost ice wedges, 210, 292 Permian, 16, 21 Permo-Carboniferous, 17, 21, 47 Peru, 42 Pervasive deformation, 139 Peyto Glacier, Alberta, 56 Phanerozoic, 17, 21, 22 Phytoplankton, 383, 397, 399

540

Piedmont glaciers, 25, 82, 163 Pingo formation, 292 Pipe channels, 178, 179, 210, 347 Pipe-flow of porewater in subglacial sediment, 101, 115, 123, 178, 203, 262, 478, 479 Pipes & dish structures, 209, 347 Pitted erosion zone, 63 Pitted moraine, see Hummocky moraine Place Glacier, British Columbia, 56 Plane of d´ecollement, see D´ecollement Plankton, 28, 383, 396, 397, 399, 402, 413 Plate tectonics, 17, 20, 46, 47, 62 Platte type facies model, 313 Pleistocene, 1, 4, 8, 9, 13, 24, 25, 26, 28, 32, 34, 36, 37, 39, 42, 44, 45, 47, 48, 58, 59, 60, 73, 87, 88, 90, 167, 179, 183, 185, 187, 288, 317, 318, 319, 322, 323, 324, 325, 327, 328, 329, 330, 336, 388, 407, 420, 423, 432, 434, 449, 459, 463, 471, 475, 480 Pleistocene/Holocene boundary, 24, 38, 48 Pliocene, 23, 24, 28, 30, 39, 42, 47 Pliocene/Pleistocene boundary, 24 Ploughing (ice bergs), 404, 407, 417 Ploughing of clasts, 86, 101, 173, 200 Plucked bedrock forms, 220, 249–52 Plucking, see Quarrying Plumes, 4, 121, 180, 216, 343, 372, 377, 378, 379, 383, 397, 405, 414 Plunge pools, 293 Poland, 26, 42, 226, 246, 282 Pollen records, 26, 32–4, 38, 42, 278, 357, 454 Polycrystalline ice, 54, 80, 84 Polymictic lake, 342, 344 Polythermal bed conditions, 66, 67, 68, 69, 72–4, 83, 85, 113, 140, 149, 150, 152, 154, 155, 156, 157, 163, 164, 187, 189, 190 Porewater, 12, 82, 86, 102, 124, 188, 189, 203, 217, 231, 263, 399, 412, 441, 476, 477

INDEX

Porewater pressure, 80, 85, 87, 108, 139, 172, 173, 177, 185, 189, 206, 231, 420, 441, 477 Port Askaig Tillite, Scotland, 12, 458 Potassium/argon dating, 39 Potholes, 5, 131, 252, 257, 259, 262, 293 Precambrian, 12, 17, 21, 22, 62, 396, 458, 475 Precambrian Shield, see Canadian or Fennoscandian Shields Pre-Cenozoic glaciations, 16–23, 400–2 Precession, 16, 30, 48, 52, 463 Pressure melting, 36, 53, 69, 72, 85, 90, 93, 134, 142, 148, 149, 150, 152, 155, 164, 174, 176, 187, 195, 240, 245, 252, 267 Proterozoic, 17, 19, 20, 391, 400, 401 Provenance (clast), 69, 157, 195, 209, 210, 226, 380, 450 Proxy evidence, 39, 50, 149, 335, 446, 452, 453, 462, 465 Prydz Bay, Antarctica, 379 Puget Ice Lobe (Pleistocene), Washington State, 39 Pulju moraine, 222, 273–5 Push moraines, 101, 327, 417, 420, 421, 422, 423, 432–5, 436, 437, 439, 442, 443 Push ridges, 219 Q-beds, 84, 94, 116, 123, 188, 189–90, 193, 231 Quarrying, 131, 132, 139–42, 143, 145, 401 Quasi-artesian conditions (hydraulic systems), 121 Quasi-hard/mobile beds, see Q-beds Quaternary, 24, 26, 32, 39, 44, 45, 53, 60, 61, 93, 100, 126, 189, 190, 417, 445, 447, 453, 454, 459, 462, 463, 471 Queb´ec, 18, 42, 143, 223, 224, 447 Quelccaya Ice Cap, Peru, 39 Quick sands, 347 Ra moraine, Norway, 328 R-channels, 119, 122, 124 see also R¨othlisberger channels

Radiocarbon dating, 26, 37, 351, 357, 391 Rafting, 32, 38, 42, 180, 374, 379–82, 383, 392, 402, 407, 413, 414 Rafts, see Sediment rafts (intraclasts or ice rafts) Rain-out (of sediments), 215, 216, 217, 263, 366, 458 Raised plateaux, 333 Rapid ice flow, see Ice streams or surges Rannoch Moor, Scotland, 276 Readvance sequence of glaciotectonic overprinting (style/model), 442 Recessional moraines, 325 Red algae, 396 Red beds, 20 Reefs (coral), 32, 237 Regional areal erosion (subglacial), 238–40 Regional linear erosional forms (subglacial), 240–9 Release mechanisms (surges), 100 Regelation, 63, 69, 72, 83, 87, 93, 135, 136, 138, 149, 164, 166, 195 Weertman regelation, 69, 149, 154 Rehburg line, 432, 442 Relaxation time, see Response time Renland Ice Core, Greenland, 36 Resedimentation processes (reworking), 203, 210, 277, 319, 326, 327, 358, 401, 402 Response (relaxation) time (ice mass adjustment), 57, 60, 82, 114, 282 Rheology of sediments, see Soft sediment deformation Rhythmites, 351, 356, 357, 401 Ribbed moraine, 223 Riedel shears, 212 Riegels, see Rock steps Rigid subglacial bed, 114, 123, 174, 188–9, 420, 422 Rinnen, see Tunnel valleys Rinnentaler, see Tunnel valleys Ripples: climbing, 347, 353, 354, 356 fluvial, 179, 180, 295, 305, 306, 311, 313, 347, 354, 356 megaripples, 311, 313, 315

INDEX

Riss glaciation, 5, 25 Robin ‘Heat Pump’ effect, 149 Roche moutonn´ee, 138, 237, 238, 240, 249, 251–2, 254, 257 Rock drumlins (roc drumlins), 254, 257 Rock flour, 132, 383 Rock fracture, 102, 131, 132, 133, 134, 138, 139–42, 153, 155, 164, 165, 166, 167, 251, 278 Rock steps, 140, 241, 246 Rock strength, 141 Rocky Mountains, Canada/USA, 5, 26, 39, 56, 283, 434, 436 Rogen moraines, 195, 218, 222, 223–5, 226, 230, 233, 234, 349, 377, 478 Ronne Ice Shelf, Antarctica, 89, 366 Ross Ice Shelf, Antarctica, 60, 89, 365, 366, 378, 393 Ross Sea, 44, 366, 379 Rossby waves, 47 R¨othlisberger channels, see R-channels Roundness of clasts/particles, 165, 166, 167, 301–2 Royal Mile, Edinburgh, Scotland, 236 Russia, 42, 226 Rust’s facies model types, 311, 313 Rutford Ice Stream, Antarctica, 91

S-forms, 256, 257, 262, 273 Saalian end moraines, 435 Saale glaciation, 8, 25, 42, 435 Sable Island, Nova Scotia, 248 Saddle, see Ice saddle St. Elias Mountains, Canada/USA, 44 St. Lawrence River, Canada/USA, 282 Salinity (sea water), 28, 47, 48, 159, 340, 342, 363, 377, 393, 396, 399, 403, 407, 409, 414 Salpausselk¨a moraines, Finland, 328 Saltation, 179 Sand dykes (or dikes), 323 Sand laminae, 322 Sand lenses, 205, 347, 407 Sand shadows, 310 Sandurs, see Glacial outwash

Sangamon interglacial, 26, 32, 39, 42 Saprolites, 447 Saskatchewan, 223, 228, 432, 436 Scablands, 293 Scandinavia, 5, 16, 25, 45, 179, 256, 309, 327, 413, 420, 449, 466, 468 Scandinavian Ice Sheet, 25, 26, 42, 45, 179, 402, 435 Scandinavian Shield, see Fennoscandian Shield Scarborough Bluffs, Ontario, 12, 457 Sch¨otter, see Glacial outwash plains Scotian Shelf, Nova Scotia, 248 Scotland, 5, 12, 16, 17, 179, 236, 238, 276, 447, 450, 458, 470, 471 Scott type facies model, 313 Scottish erratics, 179 Scour marks, 247 Scoured terrain, 238 Sea ice, 48, 158, 361, 362, 379, 381, 383, 384, 385, 392, 393, 396, 397, 399, 402, 404, 405, 407, 408, 412, 413 Sea ice rafted debris (SIRD), 379, 383 Sea level change, see Eustatic Change Sea surface temperatures (SST), 28, 30, 44 Sea water, 28, 121, 372, 385, 396, 399 Sediment budgets, 351 Seismic events, see Neotectonics SEM (scanning electron microscopy), 399, 476 Sentinel Glacier, British Columbia, 56 Seracs, 63, 107, 373 Sevetti moraine, 265 Shelbyville moraine, Illinois, 326 Shell fragments in glacigenic sediments, 16, 180, 407, 461 Sherman Glacier, Alaska, 151 Shorelines (raised), see Raised beaches Shropshire, England, 16 Siberia, 21, 22 Sichelwannen, 120, 257 Sierra Nevada, California, 58

541

Siliceous ooze, 379, 385, 396 Silurian, 21 Sintering (rounding of snow particles), 54 Siple Coast, W. Antarctica, 366 Sirius Formation, Antarctica, 42 Skaftafellsj¨okull, Iceland, 294 Skovde moraine, Sweden, 328 Slackwater deposits, 294 Slickensides, 210 Slumping, 125, 217, 262, 330, 354, 371 Slump terraces, 344 Snow: densification (consolidation), 54, 67 firnification, 150, 154 transformation, 54, 67, 74, 84 Snowlines, 44, 52 Snowy Pass Supergroup, 18 Soft sediment deformation or mobilization, see Deformation Soils, see Paleosols Solar causes of glaciation, 16–18, 46–50 faint young sun theory, 16, 17, 19, 22 Solar radiation, 46, 47, 48–50, 51, 55, 59, 340, 341 Solheimaj¨okull, Iceland, 294 Solheimasandur, Iceland, 314 Slifluction, 276, 306 Sorge’s Law, 54 South Africa, 17, 18 South America, 21, 22, 39, 44, 47, 363 South Island, NZ, 288 Southern Alps, NZ, 44, 168 Spitsbergen, see Svalbard Spitsbergenbanken, 405, 408 Spring meltwater flows (‘spring event’), 14 Squeeze moraines, 210, 236, 369 ‘Squeeze’ processes, 69, 82, 177, 210, 234, 236, 274, 348, 369, 377, 385, 441 Stages (isotope), see Isotope stages Stagnant ice, 97, 101, 110, 129, 163, 169, 171, 172, 176, 177, 187, 194, 225, 263, 276, 323, 331, 332, 333, 441

542

‘Sticky spots’ or ‘stick-slip’, 67, 85 Stillstand, 217, 267, 325, 407, 436, 446 Stillwater complex, USA, 17 Storegga Slide, North Atlantic Ocean, 409 Storglaci¨aren, Sweden, 86, 143 Strain rate of ice, 74, 154, 185 longitudinal strain rate, 82–3, 97, 112, 125, 185 Stratigraphic resolution, 391 Stratotype, 24, 25, 454–5, 463, 469–70, 471 Streamlined bed(rock)forms, see Crag & tail; Erosion marks, Flutes, Grooves, P-forms, Roches mouton´ees, Rock drumlins, S-forms, Scallops, Scoured terrain Streamlined landforms/bedforms, see Drumlins, Drumlinoids, Fluted moraine, Megadrumlins, Radial moraine Stress corrosion, 132, 141 Striae, 1, 62, 131, 132, 133, 138, 166, 249–51, 270, 273 Striation(s), see Striae Structural discontinuities in sediment, 101, 195, 205, 206, 209, 210 classification, 207 Subaquatic diamictons, 215, 217 Subaquatic diamicton m´elanges, 217 Subglacial bedform development, 12, 53, 78, 130, 138, 179, 183, 190, 193, 195, 209, 218–24, 230, 233–4, 275, 277, 422, 478, 479 Subglacial glaciofluvial sediments, 205–9, 263–5 Subglacial glaciolacustrine sediments, 205–9, 263–5 Subglacial hydraulic systems, 2, 91, 99, 102, 106, 109, 112–14, 115, 116–25, 135, 185, 219, 223, 247, 478 Subglacial lakes, 125, 252, 336 Subglacial meltwater channels, 124, 193, 238, 252, 259–63, 273, 274 Subglacial meltwater floods, 12, 82, 125, 225, 252 see also J¨okulhlaups Subglacial till, see Till

INDEX

Sublimation till, 173, 174, 176, 195, 203 Submarine diapirs, 217, 421 Sudbury Formation, England, 460 Summit Ice Core, Greenland, 36 Sun, see Solar causes of glaciation Supercontinent cycle, 17, 20, 22 Supercooled meltwater, 113 Supraglacial lakes, 125, 129, 323, 337 Surges, 13, 63, 82, 94–102, 105, 109, 113, 116, 118, 122, 123, 150, 163, 319, 331, 369, 414, 417, 432, 436, 441, 449, 479, 480 comparison between the Variegated & Medvezhiy Glaciers, 113 hypotheses of surge behaviour, 99–100 mini-surges, 113 sedimentology associated with surging, 100–2 surface morphology, 96–7 velocity, 94 Svalbard (Spitsbergen), 5, 100, 362, 409, 421, 423, 432, 441 Sveg tills, 176, 224 Sweden, 86, 143, 177, 185, 222, 224, 274, 275, 327, 328, 357 Swiss Alps, 4, 15, 432 Switzerland, 5, 226, 447 Syndepositional deformation, 194 Syneresis, 477 Syntectonic sedimentation, 435

Tadzhikistan, 91, 96 Talchir Boulder Beds, India, 16, 21 Tasman Glacier, NZ, 151, 166, 169 Tasmania, 44 Taylor Dome, Antarctica, 36, 44 Tectonic lamination, 428, 429 Tectonic styles, see Glaciotectonics Temperate subglacial bed types, see Hard or Rigid (‘H’), Soft or Mobile (‘M’), Quasi-Rigid/Soft (‘Q’) beds Tenaghi-Philippan, Macedonia, 34 Tension fractures, 428 Tephra, 39, 297 Tephrochronology, 34, 357

Terminal grade concept, 164 Terminal moraine, 62, 248, 288, 294, 325 Tertiary, 22, 24, 432, 434 Tests, 28, 396 Thames River, England, 459 Thawed bed conditions, see Subglacial thermal bed conditions Thaw consolidation, 203, 205 Thermal classification of ice mass types, 54, 65–8, 73–4 Thermal stratification of glacial lakes, 341–2, 343, 350, 480 Thermocline, 341, 353, 356 Thermohaline circulation (conveyor belt), 48 Thermokarst, 106, 125, 292 Thermoluminescence, 34 Thin film, see Basal water layer Thixotropic deformation, 313 Thompson Glacier, Axel Heiberg Island, 432 Through-valleys, 94 Thrust planes (in ice), 75–7 Thrust block ridges, 219 Thrust folds, 72, 460 Tidal flats, 313, 383 Tibetan Plateau (Qinghai-Xizang), 47 Tibet, 21, 47 Tidewater glaciers, 13, 54, 56, 91, 121, 213, 407 Tidewater termini, 363, 368–78, 379, 381, 382, 383, 387, 388, 391, 399, 414 Till (deposition), see Diamictons Till balls, 206 Till channels (Alley’s), 124 Till deltas, 217, 378, 402, 409 Till fabric, see Clast fabric Till prism, 173, 194 Till sheets, 26, 181, 267, 363, 423, 435–6, 442, 443, 449 Till tongue, 173, 194, 378 Till wedges, 428 Tillite, 18, 22, 400, 402 Time lag, see Response time Toronto, Canada, 12 Total Organic Carbon (TOC), see Organic Carbon Trace fossils, 343, 351

INDEX

Transantarctic Mountains, Antarctica, 21, 23, 44 Transport pathways, 53, 152–61, 164, 343, 350 Transverse subglacial landforms, see Subglacial transverse moraines Trapridge Glacier, Yukon Terr., 96, 98 Trollheim type facies model, 311 Trough(s) (glaciated), 62, 241–6, 288, 318 Trough cross-stratification, 311, 313, 314 Tunnel valleys, 189, 241, 246–9, 262, 270, 442 Tunneldalen, see Tunnel valleys Tunneltaler, see Tunnel valleys Turbidite, 401 Turbulent jets, 372 Turkey, 21 Turtmann Glacier, Switzerland, 432 ‘Two Component Mixing’ glaciomarine environments, 382–3 Type ‘A’ ripple-drift cross lamination, 354 Type ‘B’ ripple-drift crosslamination, 354 Type ‘S’ sinusoidal ripple lamination, 304

UK, see Britain U-shaped valleys, 88, 143, 144, 243, 244 Ukraine, 42 Underflow, see Hyperpcynal Ungava, Canada, 60, 152, 193 United Kingdom, see Britain United States of America, see USA Upstream B, see Ice Stream B

Upton Warren Interstadial Complex, England, 454 Upwelling of ocean water, 263, 372, 397 Uranium-series dating, 28 Uruguay, 21 Urumqi Glacier, China, 86, 187 USA, 26, 34, 39, 42, 61, 163, 185, 193, 223, 226, 246, 248, 293, 297, 325, 328, 445, 465, 479

Valley train deposits, 288, 297, 330 Valley wall frictional drag, 87, 130 Variegated Glacier, Alaska, 86, 91, 96, 97, 98 Varves, 305, 351, 357, 400 Varved argillites, 400 Varvites, 358 Veiki moraine, 265, 273, 275 Vein calcite, 34 Velva, North Dakota, 228 Volcanoes, see Subglacial volcanoes Vøring Plateau, N. Atlantic Ocean, 402, 411 Vostok Ice Core, 34, 36, 37, 38, 44 Vostok Station, 36, 125 Vrica section, Clabria, Italy, 24

Wales, 16 Wallows (glaciomarine), 383 Warthe moraine stage, 25 Washboard moraine, 234, 328 Washington State, 39, 92, 185, 293, 297 Water body turnover (lakes): dimictic, 341 meromixis, 342 monmictic, 342 polymictic, 342, 344

543

Waterlain till, 180, 215, 216–7, 263, 429, 458 Water layer, see Basal water layer Wave ogives, see Ogives Wave trains (in ice), see Ogives Weathering, 13, 17, 18, 20, 47, 139, 140, 151, 152, 320, 447, 451 Weddell Sea, 393 Wedge, see Accretionary wedge or till deltas or grounding line wedge ‘Weertman regelation’ film, 69, 149, 154 Weichsel glaciation, 25, 42, 282, 396, 405, 407, 409 Wind drift of snow, 57 Wind mixing in glacial lakes (wind stress), 343, 353 Windermere Interstadial, 463, 470 Wisconsin, 100, 193, 325, 326 Wisconsinan glaciation, 8, 26, 38, 39, 42, 52, 54, 61, 248, 282, 396, 454 Witwatersrand succession, South Africa, 17 Worcestershire, 16 Wright Dry Valley, Antarctica, 195 W¨urm glaciation, 5, 25 Wyoming, 18, 39

Xian, China, 34 Xifeng, China, 34

Yana outwash, Alaska, 311 Yarmouth interglacial, 26 Young Sun theory, see Solar causes of glaciation Younger Dryas, 32, 38, 45, 50, 450, 452, 463, 466 Yukon Territory, 44, 96, 129, 297

This Page Intentionally Left Blank