The Preservation of Englacial Structure Results from Passive Deposition, Which Occurs Under Restricted Conditions

SERVATION OF ENGLACIAL STRUCTURE:

A ANALYSIS OF BASAL ICE AND TILL

A thesis submitted in partial fulfilment of the requirements for the Degree of Master of Science in Geography in the University of Canterbury by K.E. Sinclair :::;;:~

University of Canterbury 1998 11

ABSTRACT

Englacial structure (sedimentary and glaciotectonic) found in the basal zone of many glaciers can, in some circumstances, be preserved in till deposits. This structure provides evidence that can be used to infer the basal zone characteristics of former glaciers, and may therefore contribute to the reconstruction of past glacial environments.

The preservation of englacial structure results from passive deposition, which occurs under restricted conditions. Some of these conditions relate to the characteristics of the basal ice, and others to the nature of the proglacial environment. Sublimation is an inherently passive process that occurs when ice is transformed directly to water vapour, without an intermediate liquid phase. This process formed the basis of an arid polar model of deposition (Shaw 1977a).

Laboratory experiments conducted on basal ice samples from the Taylor Valley, Antarctica, show that slow melt may also deposit till passively. Melt-out till was produced by allowing the debris to consolidate vertically, with little lateral dislocation during deposition. A comparative analysis of basal ice and laboratory-generated till deposits show that lamination, folding and banding were preserved. Two conceptual models of passive deposition illustrate that sublimation and slow melt may operate concurrently and in close association in arid polar environments. Both processes may lead to the preservation of englacial structure in arid polar environments, and it may be difficult to distinguish between sublimation and melt-out tills. 111

ACKNOWLEDGEMENTS

I wish to extend my thanks to the many people who have helped me complete this thesis.

My supervisors Prof. Bob Kirk, and Dr. Sean Fizsimons have both provided continued support and advice. Thank you Bob for your positivity and insights into thesis writing. Thank you Sean for an Antarctic adventure that I will never forget, and of course for your input throughout the year. I would also like to thank Dr. Wendy Lawson for inspiring lectures and field trips, and for the encouragement to start my project.

Assistance and companionship at the Suess Glacier from Marcus van der Goes (the chainsaw guru) and Prof. Regi Lorrain was greatly appreciated. Thank you Wendy and crew at the Taylor Glacier, for the charming hospitality, use of equipment and your chainsaw lessons Brian.

The laboratory facilities and technical support at Canterbury University were invaluable. I am especially grateful to Robert Spiers (Geology) for cheerfully allowing me access to cutting equipment, and for advice and help with thin-sections. Technical support in Geography, in the laboratory, in the photographic realm and from all the computer (6th floor) staff was appreciated. Dr. Henry Connor's thought-provoking advice and encouragement was gratefully received. I would also like to thank the Zoology Department for tolerating my unconventional use of their freezer, and the Civil Engineering Department for access to their laboratory. lowe a great deal to supportive and understanding friends. Gerard, thanks for developing such an interest in till and for asking difficult questions. Thank you Amanda, Victoria and Rebecca for your enthusiasm and friendship; Bids for your perspective and some adventurous diversions; and Nick for your entertaining e-mails and hospitality in Dunedin. Also, many thanks to the 2nd floor crowd for an enjoyable year.

Finally, my family have been an endless source of support in many ways. I am grateful for the understanding, advice (academic and otherwise) and your financial assistance throughout university. IV

CHAPTER 1: INTRODUCTION 1

CHAPTER 2: BASAL ICE FORMATION AND DEFORMATION 5 2.1 Temperate Glaciers 6 2.2 Polythermal Glaciers 8 2.3 Dry-based Glaciers 12

CHAPTER 3: GLACIGENIC DEPOSITION 18 3.1 Introduction 18 3.2 Nomenclature 19 3.2.1 Defining till 3.2.2 Genetic tenninology and classification systems 3.3 Passive Deposition 24 3.4 Conditions for the preservation of englacial structure 36

4: RESEARCH 4.1 Basal ice sample collection 41 4.2 Laboratory methodology 49

5: SAMPLE BLOCK CHARACTERISTICS 5.1 Sedimentary characteristics 55 5.2 Structural characteristics 56 5.3 Basal ice facies 63

CHAPTER 6: COMPARATIVE 6.1 Suess Glacier 72 6.2 Taylor Glacier 80

7: DISCUSSION AND CONCLUSIONS 89 7.1 Study implications and discussion 89 7.2 Future research 91 7.3 Conclusions 93

REFERENCE LIST 94 APPENDICES 105 v

LIST OF FIGURES

Chapter 2: Debris Entrainment

Figure 2.1 Regelation (Weertman 1961). 7

Figure 2.2 The shear hypothesis. 9

Figure 2.3 Upglacier-dipping flow lines produced experimentally with modeling clay (Hooke 1970). 10

Figure 2.4 Weertman's (1961) basal refreezing mechanism. 11

Figure 2.5 Debris entrainment by apron oveiTiding (Shaw 1977b). 13

Figure 2.6 Schematic representation of transpOlt and defOlmation at the margins of dry-based glaciers (Shaw 1977a). 15

Figure 2.7 Depositional model for the fonnation of thrust-block moraines at the margins of dIy-based glaciers that flow into lakes (Fitzsimons 1996). 17

Chapter 3: Glacigenic deposition

Figure 3.1 A. Model of deposition in arid polar environments (Shaw 1977a). B. Facies model of till deposition in arid polar environments Eyles et al. 1983). 27

Figure 3.2 Model of controlled moraine development (Type 2), (Rains & Shaw 1981). 28

Figure 3.3 Shaw's (1977b) interpretation of the glacigenic sequence at the LaCroix Glacier, Taylor Valley, Antarctica. 29

Figure 3.4 Possible till sequence from a body of debris-rich ice, showing spatial and temporal relationships between till-fonning processes in arid polar environments. 33

Figure 3.5 Genetic till prism (Dreimanis 1988). 34

Figure 3.6 Genetic till prism (Hickock 1990). 35 VI

Figure 3.7 Genetic till prism adapted to illustrate passive deposition in arid polar environments. 35

Figure 3.8 Sublimation till, Vega Island, Antarctic Peninsula (Shaw 1985). 37

Chapter Research Strategy

Figure 4.1 Location map - The D1Y Valleys (Fitzsimons 1996). 42

Figure 4.2 The Taylor Valley. 43

Figure 4.3 A. The Suess Glacier, right margin. B. Profile of the thrust block moraine ridge at the right margin of the Suess Glacier (Fitzsimons 1996). 44

Figure 4.4 The down-glacier branch of the tunnel, showing amber ice, overlying the basal zone. 46

Figure The telminus of the Taylor Glacier, with Lake Bonney in the foreground. 48

Figure 4.6 The Taylor Glacier basal zone, showing a sharp contact with overlying glacier ice. 48

Figure 4.7 Test block of ice made with altemating layers of Suess Glacier moraine sediment and black beach sand. 53

Figure 4.8 Melt-out trial sediment deposit. 53

Figure 4.9 Block 10 encased in a perspex box, prior to melt. 54

Figure 4.10 An assembled melt-out experiment. 54

Chapter 5: Sample Block Characteristics

Figure 5.1 Geometrical features of a fold profile. 58

Figure 5.2 The classification of fold profiles using dip isogon patterns (Ramsay 1967). 59

Figure 5.3 Field measurements of similar folds, produced by va

passive defOlmation (Donath & Parker 1964). 60

Figure 5.4 Geometrical features of a folded layer (Price & Cosgrove 1997). 61

Figure 5.5 A recumbent fold at the Taylor Glacier (Shaw 1985). 62

Figure 5.6 Isoclinal drag fold at the Taylor Glacier, Antarctica (Hamilton & Haynes 1961). 62

Figure 5.7 Ice facies at polythelmal glaciers. 65

Figure 5.8 - 5.11 Suess Glacier basal ice facies. 69 -71

Chapter 6: Comparative analysis

Figure 6.1 The relative in situ positions of blocks 9 and 10. 73

Figure 6.2 The effect of frictional drag at the sediment-plastic interface during melt. 79

Figure 6.3 Block 10. 81 Figure 6.4 Block 10 till.

Figure 6.5 Block 9. Figure 6.6 Block 9 till. 82

Figure 6.7 Block 1. Figure 6.8 Block 1 till. 83

Figure 6.9 -Block 7 & 8. 84

Figure 6.10 Cavities in block 7 till. Figure 6.11 Block 8 till. 85

Figure 6.12 Block 6. Figure 6.13 Block 6 till. 86

Figure 6.14 Laminated sequence of basal ice from the Taylor Glacier. 87

Figure 6.15 Taylor Glacier, block 1 tilL Figure 6.16 Taylor Glacier, block 2 till. 88 V1n

LIST OF TABLES

Table 2.1 Classification of lodgement, flow and melt-out tills. 23

Table 2.2 Genetic classification of terrestrial tills (Shaw 1977a). 23

Tables 6.1 - 6.5 Suess Glacier block dimensions and sedimentruy characteristics:

Block 10 74 Block 9 75 Block 1 76 Blocks 7 & 8 78 Block 6 79

Table 6.6 Taylor Glacier block dimensions and sedimentruy characteristics. 80 1

CHAPTER!

INTRODUCTION

There have been two principal approaches taken in attempts to understand the histoty of glaciated and fOlmerly glaciated areas. Some scientists have been concerned primarily with describing Pleistocene glacial sediments to make inferences about their genesis according to general principles (Boulton 1987). By analysing these sediments clues are found about the size of the former glacier, the ice-flow direction, its basal thermal regime, the mass balance, and therefore, the prevailing climate (Bennett & Glasser 1996). Another school attempts to understand contemporruy glacial erosion, transportation and deposition processes, which may then be applied to the examination and interpretation of their geological products.

To date, surprisingly few linkages have been conclusively made between sediments cunently in transpOltation by ice, and those that have been deposited by glaciers. This is considered a reflection of the tendency to delineate the glacial environment in order to simplify and bracket processes. Glacial geomorphlogists and geologists, according to the first approach, restrict their work to the interpretation of glacial sediments and landforms. -The second group, which Boulton (1987), terms "ditty-ice glaciologists," study contemporruy sedimentological processes. However, the entrainment, transport and deposition of sediment are not mutually exclusive, but are instead components of a process continuum. Some important links in this continuum may be found by more integrated study of the glacial environment, and by increased dialogue between both groups of scientists (Boulton 1987; Knight 1987).

Much of the 'work' done by a glacier in telms of the erOSIOn, entrainment and transport of sediment is concentrated at the basal zone. The basal zone lies at the intelface between glacier ice and the underlying substrate and exhibits different rheological and mechanical properties from 2 overlying ice. Glacier ice is a product of the fimification process, while basal ice is generated at, and interacts with the glacier bed (Paterson 1981). Chapter 2 reviews mechanisms that have been presented to account for the formation and defOlmation of englacial stmcture.

The basal zone is also stmcturally distinct, mainly by virtue of high debris concentrations relative to overlying ice. High strain rates yield structure that reflects debris entrainment and subsequent deformation (Kirkbride 1995). The sediment that is transported in the basal zone is termed 'englacial' for the purposes of this study, as this denotes transport in ice. The term also differentiates sediment transpOlied in the basal zone from structure that fOlms due to defOlmation beneath the glacier. 'Structure' refers here to sedimentruy (or primruy) structure; fabric and textural properties; as well as glaciotectonic (or secondruy) stmcture, such as folding, fissility and foliation.

The examination of englacial stmCture has been neglected by researchers despite the fact that its configuration can have a marked influence on the character of glacial deposits. The aim of this study is to docmnent englacial sedimentary stmcture, and to detelmine whether any of this structure can be preserved in glacial deposits, or till. If englacial structure can be identified in ancient glacial deposits, this may provide evidence to reconstruct former basal zones. As the basal zone is a major influence on glacier dynamics, its reconstmction will inevitably offer important insights into understanding past glacial environments.

The concept that glacial deposits could retain some characteristics of the parent basal ice was the comerstone of Shaw's (1977a) model of deposition in arid polar environments. This model rests on the assumption that sublimation (removal of ice by the direct transfOlmation to a vapour state, without an intermediate Fquid phase) is the dominant mode of deposition in arid polar environments and that sediment deposited in this way would retain englacial structure.

Shaw's model provoked debate that centered on the issues of preservation and fidelity (Shaw 1980). It has been questioned whether 3 englacial structure can be preserved and, if so, whether the assumed process (sublimation) is afaithful description of the actual process that leads to its preservation. Although the fonnation of some till exposures in Antarctica has been attributed to sublimation (Shaw 1977a, 1977b, 1985), this is not considered conclusive evidence that sublimation is the dominant ti11- forming process in arid polar environments (c.f. Eyles et al. 1986, Humphreys & Fitzsimons 1996).

The primary hypothesis of this study is that englacial structure can be preserved in till deposits. This occurs under restricted conditions, some of which relate to 'external factors,' such as the climate of the terminus area which affects, for example, the amount of melt-water available to redistribute sediment. Other conditions relate to the characteristics of the basal zone, such as the deblis concentration and dispersal.

The second hypothesis is that, along with sublimation, slow melt could also result in the preservation of englacial structure. In arid polar environments these processes may act together, and in close association. Unless the depositional process can be observed, sublimation till and melt out till deposits may be indistinguishable. For this reason they are classed together in Chapter 3, as passive deposition.

Chapter 3 reviews our current understanding of till as a product of glacigenic deposition. It outlines problems associated with defIning till and describes the development of a nomenclature that describes glacial deposition. The spatial and temporal relationships between sublimation and slow melt-out in arid polar environments are represented in two conceptual models of passive deposition, which illustrate the interrelationships between these till-forming processes. In Part 4 a set of conditions by which englacial structure can be preserved by passive deposition is developed.

It is diffIcult to observe and monitor glacial deposition. This is partly due to the inaccessibility of most basal zones and partly to the long time scale over which these processes occur. The use of a subglacial tunnel at the Suess Glacier in the Dry Valleys, Antarctica, allowed access to the basal zone. Chapter 4 describes the collection of basal ice samples from the 4 tunnel and from the Taylor Glacier, which exhibited very different basal zone characteristics. The second, temporal limitation was overcome by simulating passive deposition in the laboratory environment. The methodology employed to conduct these experiments is also described in Chapter 4, along with the sedimentary techniques used to analyse the concentration, size and sorting of the englacial debris.

The basal ice is broadly characterised through sedimentological, structural and facies descriptions in Chapter 5. In the subsequent comparative analysis (Chapter 6) each basal ice sample used in the laboratory experiments is described in detail. This provides a basis for a structural comparison with the resultant till deposit. The results and implications of this analysis are discussed in Chapter 7, and conclusions summarised. 5

CHAPTE

BASAL ICE FORMATION AND DEFO ATION

Debris enters the glacial system from supraglacial sources such as valley walls and nunataks, or subglacially by basal erosion and entrainment Debris that is transpOlted subglacially, in basal ice, can vary widely in concentration, and values between 0.01% to 70% by volume have been recorded (Boulton 1971; Lawson 1979a). It fOlms till upon the removal, during deposition, of sUlTounding (interstitial) ice. The entrainment of debris and its subsequent defOlmation in the basal zone fOlms englacial structure.

This chapter reviews mechanisms which may contribute debris to the basal zone and models that have been presented to account for the deformation of basal ice near the glacier margin. Where possible these are related to particular thelmal regimes (temperate, polythelmal or dry-based), as this controls the sedimentalY characteristics, spatial variability and configuration of the basal zone.

The three thelmal regimes or boundalY conditions were fIrst recognised by Weedman (1961) and Boulton (l972a). Each was summarised by Drewry (1986) as follows: 6

where 1\s is the heat due to sliding; 1\g is the heat due to geothennal heating; Ki is the thennal conductivity of the ice and (dT/dH) 1* is the temperature gradient in the lowennost layers of the ice mass.

In equation 2.1 the temperature gradient is not sufficient to conduct all the heat supplied at the basal zone (from ice movement and geothennal heat), and melting occurs at the ice-substrate interface. The glacier is 'wann-based', or 'temperate'. Equation 2.2 describes polythelmal boundary conditions, where the temperature gradient is just sufficient to conduct heat from the bed and there is an approximate balance between melting and freezing. When the temperature gradient can easily drain all heat from the bed, as shown in equation 2.3, the glacier is dIy-based. The ice is completely frozen to the substrate and any water draining along this interface will freeze.

As the basal ice of temperate glaciers is continually melting at the ice-bed interface, sliding is the most important component of glacier flow. Material eroded from the glacier bed is transported in a thin, generally less than 0.5 metre, layer of regelation ice immediately above the glacier bed (Kamb & LaChappelle 1964; Boulton 1972b, 1975). Regelation is induced by pressure-melting on the upstream side of small « 1 metre) bedrock obstacles (Weertman 1957, 1964). Water produced by pressure melting will flow around obstacles and freeze downstream in low-pressure cavities (Lliboutry 1964-65), Debris is cOnCutTently trapped in this regelation layer. Heat released during refreezing is transmitted back through the bedrock obstacle, which aids fut1her upstream melt (Figure 2.1). 7

ICE FLOW

high pressure - rrdt low pressure - refreezing

WA1ER ;)

Figure 2.1: Weertman's (1961) regelation mechanism.

Debris entrained by regelation originates from bedrock crests and consequently bands contain low debris volumes. The thickness of the basal ice is limited by the height of the bedrock topography in these circumstances, as entrainment of debris can not take place above this height (Paul & Eyles 1990). Bands are also limited in lateral extent because ice fonned in the lee of one obstacle can be destroyed downstream by fmther pressure-melting (Hubbard & Sharp 1989).

Robin (1976) presented a 'heat-pump' argument to explain regelation beneath some temperate glaciers. According to his theory, localised cold patches develop when melt water is lost from the system. The loss of latent heat causes refreezing at the scale of several square metres. The direct incursion or-ambient air temperatures into cavities may also contribute to the fonnation of basal ice by refreezing subglacial melt water (Anderson et al. 1982; Gemmell 1985). 8

POLYTHERMAL

The term 'polythermal' implies wet or dly-based zones which extend over several thicknesses of the glacier. These zones display a range of configurations, but commonly have a wet-based interior, constrained at the margin by a dry-based, relatively immobile zone. The dry-based margin acts as a obstacle to internal flow and the transition between these two zones has a dynamic influence on the entrainment and movement of debris. The dry based layer also prevents the lateral escape of melt-water, although some may percolate into underlying sediment (Boulton 1975; Menzies 1981).

Thick basal debris sequences which characterise polythermal glaciers are attributed to the vertical mobility of debris in the basal zone. Some mechanisms enable debris to be lifted to considerable distances above the glacier sole, for example 80 metres at the margin of the Greenland Ice Sheet, Thule (Bishop 1957; Swinzow 1962) and 50 metres at the Bames Ice Cap, Baffin Island (Goldthwait 1951). Three that are outlined below are the shear hypothesis; overriding of a wind-drift ice wedge and basal 'freezing on'ofsediment.

The shear hypothesis was proposed to account for distinct englacial debris bands exposed at the margins of the Bames Ice Cap (Baffin Island), (Goldthwait 1951; Ward 1952; Hooke 1973 a & b; Huddleston 1976), and the Thule Lobe of the Greenland Ice Sheet (Bishop 1957; Goldthwait 1960; Nobles 1960; Swinzow 1962; Hooke 1970). It was further investigated in Svalbard (Boulton 1970a), Ellesmere Island (Souchez 1971), Antarctica (Hollin & Cameron 1961, Souchez 1967, Tison et al. 1993) and Bylot Island, Canada (Dilabio & Shilts 1979). It presupposes the existence of a polythermal basal zone, where the interior is melting and the outer margins are frozen to the substrate. Active ice from the interior of the glacier moves outward to the margin where it is blocked by a zone of stagnant ice. The active ice is thrust over the stagnant ice in shear planes (Figure 2.2), that appear as upwarped flow lines at the glacier margin (Souchez 1967, Chinn & Di110n 1987). Locally derived debris can be scraped from the glacier sole 9

(Hutter & Olunloyo 1981), and moved to englacial positions along the shear planes, to be finally exposed at the glacier margin .

.. Stagnant Zone ...... +--- Active Zone

Ice flow <4---

Figure 2.2: The shear hypothesis.

The assumption that upwarped flow lines at the glacier margin result from shear was questioned by Hooke (I968, 1970). He suggested an alternative mechanism to describe these features in Greenland, whereby active ice is compressed as it ovenides a stagnant wind-drift ice wedge. This is an accumulation of superimposed ice, maintained by wind-drifted snow. It differs from the stagnant ice of the shear hypothesis because it is maintained by processes independent of the glacier. Simple experiments with modeling clay were used to demonstrate how upglacier-dipping flow lines can result from high compressive strains due to an obstruction of flow (Figure 2.3).

Weertman (1961) also objected to the shear hypothesis on the basis that the frequency and geometIy of the bands observed were incompatible with formation by shear. He suggested instead that the fluctuation of the pressure-melting point isotherm at the glacier bed causes the incorporation of sediment by repeated freezing-on of debris and melt-water (e.g. Hooke 1968, 1973a & b; Boulton 1972a; Hooke & Huddleston 1978; Moran et al. 1980; Rains & Shaw 1981), (Figure 2.4). Such fluctuations may be caused by a change in ice thickness, which increases the stress on the bed and causes pressure melting. The result is a net accretion of debris at the glacier sole (Sugden et al. 1987; Souchez et al. 1988), and by repetition of this process bands of basal debris are lifted into englacial positions and are separated by layers of refrozen meltwater. 10

Ablation zone em

Figure 2.3: Vpglacier dipping flow lines produced experimentally with modeling clay (Hooke 1970).

Freezing-on (also telmed 'net basal adfreezing') was accepted to explain basal sequences of debris bands and clean ice in areas such as Svalbard (Boulton 1970a), Antarctica (Gow et al. 1979) and Greenland (Herron & Langway 1979). The development of isotopic analysis has allowed refrozen ice to be conclusively distinguished from nOlmal firnified ice (e.g. Souchez & Lorrain 1978; louzel & Souchez 1982; Souchez & louzel 1984; Souchez & De Groote 1985). Analysis of basal ice in western Greenland, for example, now indicates that debris bands were formed by freezing-on, but intercalated clean-ice layers were not (Knight 1989). Thus, the origin of sequences of debris bands and clean-ice layers that corrunonly occur at the margins of polythelmal glaciers is the subject of further investigation. 11

A Frozen to bed Water freezing to bottom I Ice melting from bottom

Ice Flow ~ ICE

ICE Frozen to bed Water freezing to bottom

Figure 2.4: Weertman's (1961) basal refreezing mechanism, showing the incorporation of sediment by the fluctuation of the 0° isotherm. 12

2.3 DRY-BASED GLACIEHB

Dry-based glaciers are generally underlain by deep pennafrost and are dry-based throughout their entire length. Antarctic pennafrost, for example, can be up to 1 kilometre thick and geothennal heat flows have little effect on melt-water production (Humphreys 1993). Therefore, no basal sliding occurs and forward movement in such glaciers is produced by basal defonnation alone (Chinn 1980; Eyles et al. 1983). Consequently, internal friction is the primary heat source and the velocities of dry-based glaciers are generally an order of magnitude less than the velocities exhibited by temperate glaciers (Eyles et al. 1983).

Entrainment of debris beneath dIy-based glaciers can not easily be explained by subglacial processes, primarily due to the lack of melt water for refreezing (Holdsworth 1974). DIy-based glaciers may instead entrain debris by overriding and incorporating their frontal aprons (Shaw 1977a & b), (Figure 2.5). Aprons, composed of windblown sediment and snow, ice blocks, refrozen melt-water, supraglacial debris washed from the glacier surface and englacial debris, occur at the margins of many Antarctic glaciers. From their description they seem similar to Hooke's (1968, 1970a & b, 1973) wind-drift ice wedge (see Part 1.2.2), and usually rest against steep ice cliffs (usually 10-20 metres high). These cliffs are fOlmed when a semi-rigid zone of ice becomes grounded at the glacier margin and becomes an obstacle to flow (Holdsworth 1969). They are maintained by dry-calving and ablation on the cliff face (Chinn 1986, 1987, 1989, 1991).

Sediment on a frontal apron may be sorted and crudely stratified by mass-flow. Seasonal wind-drifted snow, intennittent ice-block falls from dry-calving and aeolian material result in alteinating layers of high and low debris concentrations (Shaw 1977b). Once incorporated, this material fonns a basal layer of dirty ice apparent at the margins of some dry-based glaciers. 13

frontal wall

1 frontal wall- former apron position

- velocity - vectors

deformed incorporated apron

, . I I 4 I I I clean glacier I I ..I /" A ,,/'1 " 1 ,," I '" I // I "

debris-rich basal

Figure Debris entrainment by apron oveniding. Shaded areas illustrate the sequential defOlmation of apron sediments (Shaw 1977b). 14

According to Shaw's (1977 a & b) model of deposition in arid polar environments (Figure 2.6), sediment incOlporated from the frontal apron during glacier advance becomes foliated and attenuated by shear (highly attenuated facies). The magnitude of shearing and attenuation increases with height above the base of the glacier. The result of ovelfolding and thrusting at the ice margin is the stratigraphic repetition of the basal debris zone, giving the appearance of multiple debris bands (Paul & Eyles 1990). Below the highly attenuated facies is a zone of massive debris where only minor, discontinuous foliation occurs. In Shaw's ( 1977 a) model this is the poorly attenuated facies. When the glacier retreats compressive flow occurs due to low flow rates at the thinning margins (Figure 2.6b). This produces folding and stacking of debris-rich ice.

Evans (1989a & b) applied Aber's (1982) two stage model for glaciotectonism in Canada to test Shaw's model. The first stage is proglacial or ice-marginal thrusting of defonnable sediment (Klassen 1982~ van der Wateren 1985), and staking of pennafrozen tenain. Thmsting was initiated in thick, defOlmable materials and is favoured in areas where a "suitable expanse of unlithified sediment is available, and where ice impinges on topographic highs (Evans 1989b, 119)." The second stage is subglacial shearing and penetrative defOlmation of the tluust masses (tluust block moraines) during apron oveniding.

Both Shaw (1977a) and Evans (1989a) discuss the incorporation of ice-marginal features by apron oveniding. At Ellesmere Island, where fluvial activity is the dominant transporting mechanism, melt-water deposits alluvium at the glacier margin. Following incOlporation, this results in coarse debris-rich bands and augen structures within the basal ice. Shaw (1977a) found evidence of sorted and stratified, water-laid sediment, and discrete bands and lenses in the basal ice at the Commonwealth Glacier and the Wright Lower Glacier, Antarctica. 15

A. ADVANCING GLACIER

highly attenuated, foliated ice and debris ice cliff

poorly attenuated ice and

B. RETREATING 'UI/LJ,r'a'L-'UI':"'"

I I I J I I I J I shear ,I /';' I •• ./ ;' ngld zone / / / / .....t'_ / -- .... - .... -.... -.....

zone of

Figure 2.6: Schematic representation of, A. transp0l1 and B. defonnation at the margins of dry-based glaciers. This fonns the first two stages of Shaw's (1977a) model of deposition in arid polar environments. 16

An altemative mechanism for the incorporation of ice-marginal features was investigated in the DIY Valleys by Fitzsimons (1996) and Humphreys & Fitzsimons (1996). Like Evans (1989a & b), they found that thrusting causes stacking of blocks of frozen, unconsolidated sediment, containing primary sedimentary stmctures. Horizontal, or cross-bedded sediment is interspersed with layers of ice of varying composition and algae. The well-sorted nature of the sediment implies that it has a fluvial or lacustrine origin. However, Fitzsimons (1996) rejects Shaw's (1977 a & b) apron entrainment model on the basis that it explains neither the primary sedimentary stmctures of fluvial or lacustrine origin, nor algae beds within the moraines. In addition the apron entrainment model does not adequately account for layers of clean ice which separate debris bands in the basal zone.

Instead, Fitzsimons (1996) suggested that dIy-based glaciers could experience transient wet-based conditions by flowing into saline lakes on the valley floors. Blocks of sediment may be incorporated when ice flows into sediment at the lake bed. The retention of primalY sedimentruy stmctures suggests that they were frozen when entrained. Rising flow vectors at the terminus (Chinn 1991) then thmst accreted ice and debris to the surface (Figure 2.7). The accretion of ice and debris in this way suggests that Weertman's (1961) refreezing model could be applicable to polar glaciers in circumstances where these glaciers reach the valley floor and become temporarily wet-based.

Fitzsimons' (1996) model also shows that under these conditions dry-based glaciers are able to entrain unconsolidated substrate. This accounts for a distinction found by Humphreys (1993) between Dry Valley glaciers that reach the valley floor and those that do not. The latter have much lower englacial debris concentrations, while glaciers that reach the valley floor can entrain much larger runounts of sediment. 17

+ + B + + + + + +

+ + + c + + + + + Thrust block moraines + + + + + + + + + +

Figure 2.7: Depositional model for the fOlmation of thrust-block moraines at the margins of dry-based glaciers that flow into lakes (Fitzsimons1996). 18

CHAPTER

GENIC OSITION

3.1 INTRODUCTION

Depositional processes in the glacial environment can be divided into glacigenic, glacifluvial and glacilacustrine (Boulton & Deynoux 1981). Glacigenic deposition is the direct removal of debris from interstitial ice to form till (,tillite' if lithified: Reading & Walker 1966). Till deposits can be genetically linked to their mode of deposition, and a definition of till, 'should identify the genetic processes responsible jor the deposition oj a material which is uniquely produced by glaCiers, and which should enable us to erect practical criteria by "which it should be recognised (Boulton 1976a). '

This statement introduces three issues that continue to challenge glacial geomorphologists. The first is the definition and scope of the term 'till'. The ~~cond is the identification of depositional processes from the stratigraphic and structural analysis of till. This is problematic because tills have a wide variety of depositional modes and intemal shucture. Sequences often result from a continuum of processes (Muller 1983) and till types often grade into one another (Shaw 1985). Finally, the vruiable, or polygenetic nature of till deposits has made the fOlmulation of practical criteria to recognise them difficult.

The following sections review the CUl1'ent understanding of till as a product of glacigenic deposition. Part 2 (nomenclature) outlines problems associated with defining till, and glacidepositional processes are described 19

ill VIa an explanation of genetic tenninolob'Y and genetic classification systems.

The most common till-fOlming processes in arid polar environments are slow melt and sublimation. The debate that has focused on the dominant mode of glacigenic deposition in arid polar environments is discussed in Part 3. This illustrates the difficulty associated with conclusively making genetic interpretations of till types. In addition, as sublimation and slow melt may occur concunently and in close association, they may grade into one another. For this reason they are classed together as passive deposition, which is conceptualised in two models, which illustrate the interrelationships between till-fOlming processes in arid polar environments. In Part 4 a set of conditions by which englacial structure can be preserved by passive deposition is developed.

3.2.1 Defining tin

The term till was originally applied to stiff, clay subsoil in Scotland, generally impervious and unstratified. Geikie (1863, 185), for example, described it as "a stiff clay full of stones vmying in size up to boulders, produced by abrasion carried on by the ice sheet as it moved over the land". The nineteenth century Scottish conception of till produced such pre eminent synonyms as boulder clay (Croll 1870), none of which adequately described the present understanding of till characteristics.

Debris and moraine have also been used synonomously with the term till (e.g. Chamberlain 1883, Bjorlykke 1967, Ellis & ChaIkin 1983, Bouchard 1989). It is now generally accepted that the tenn debris describes sediment transported in contact with glacier ice; that a moraine is a glacial landform, and that till is a sedimentalY product of glacigenic deposition. However, beyond this broad classification, the definition of till is still 20 evolving and remains contentious. This has been primarily ath'ibuted to the variability of till compared to other sediments (Flint 1971; Goldthwait 1971). However, the difficulty in al1'iving at a common definition also seems to reflect a disjointed approach to the study of the glacidepositional environment.

There is a lack of recognition in the literature that glacigenic deposition is part of a process continuum. Arbitrary boundaries are imposed in an attempt to define the point at which glacial debris in transport is deposited and becomes till. In reality this is a gradational transition, with no apparent natural boundruy. Similar divisions are made when defining the extent to which till can be SUbjected to syn- and post-depositional processes in the non-glacial environment before requiring reclassification. It remains contentious at what point the sedimentaty characteristics of glacigenic deposits can not longer be athibuted to glacial processes. This discussion reviews approaches taken to resolve these issues. It is, however, difficult to draw any conclusions, as few fOlmalised conventions for defining till have been widely accepted.

Some authors have distinguished till fOlmation from deposition (Shaw 1982; Lawson 1989; Lundqvist 1988). They consider that till is formed when the individual clasts and grains obtain their position relative to each other, and sedimentruy textures and stmctures are also fOlmed. Till can, therefore form on the surface of a glacier, although technically it is still in transport (Boulton & Deynoux 1981)' When there is no fm1her lateral movement and the only intel11al vettical movement is due to the reduction of pore space by the removal of the remaining ice and water till is considered deposited (Dreimanis 1988).

In practice, however, this is an impractical distinction. During deposition, vertical consolidation may alter sedimentaty stmctures, thereby changing the relative position of individual clasts and grains. Also, stagnant bodies of ice may remain in ice-cored moraines for considerable lengths of time without any lateral movement. However, the characteristics of sediment within this ice do not differ from those of the parent basal ice. 21

It can, therefore, only be classified as debris and not till. For these reasons till deposition refers here to the complete removal of interstitial ice.

Fluvial and lacustrine sediments that occur in glacial environments can be poorly sorted and resemble till in their lithologic, structural and textural characteristics. It is, therefore, common to use a sedimentological name without genetic implications, prior to classifying a poorly sorted sediment as till (Dreimanis 1982). Harland et a1. (1966) proposed diamict as a general term for all poorly sOlted sediments, both in their lithified and unlithified states. Similarly, diamicton and diamictite (if lithified) were proposed by Flint (1971) and Flint et a1. (1960) respectively, as non-genetic terms for terrigenous sediments which include, "any poorly sOlted clast sand-mud admixture regardless of the depositional environment (Eyles et a1. 1983, 394)."

The amount of syn- and post-depositional defOImation allowable before till becomes a diamict is unclear. Boulton & Deynoux (1981) suggested a "natural process boundmy," whereby a till becomes a diamict when intergranular stresses between sediment palticles reach zero. Reduced viscosity at this point is said to produce sedimentalY stmctures that differ from the stmcture of till deposits. However, they provide no examples of structures that may distinguish tills from diamicts, and therefore no real test to detelmine the position of this boundmy.

Many definitions of till have been proposed. Some authors take a very broad view in defining till. Harland et a1. (1966), for example, applied the term till to any diamictic sediment that contains glacially transported material. Boulton & Paul (1976) discuss allochthonous (far travelled from its source) flow till. This is categorised as till though it may have undergone frequent failure and remoulding.

The most restrictive definitions of till were provided by Boulton (1972a) and Lawson (1979a & b, 1981b). They describe till as a sediment formed by primary processes, releasing and depositing sediments directly from glacier ice, with no subsequent disaggregation and redeposition. The Till Work Group of the TI'l"QUA Commission on the Genesis and Lithology 22 of Quaternary Deposits adopted a similar definition of till as "a sediment that has been transp0l1ed by or from glacier ice with little or no sorting by water (Dreimanis 1982, 21)." This definition can be applied to tills deposited by passive deposition, where there is minimal water available for resedimentation and sorting. However, a strict application of these restrictive defmitions in other areas would exclude many glacial deposits from being tenned till (Lundqvist 1988, Dreimanis 1988).

3.2.2 Genetic Terminology and Classification Systems

Genetic tenns denoting till fOlming processes developed from the mid-twentieth century. Flint (1957) divided till deposits into ablation till, fonned due to surface melting, and lodgement till. Lodgement till is deposited, (lodged), beneath actively moving ice when frictional drag on pat1icles being moved over the bed equals the tractional force being exe11ed on it by the glacier ice (Boulton 1972a). This has proved a simplistic distinction which does not adequately account for the genesis of many tills.

Hartshorn (1958) proposed the telmjlow till for material modified by resedimentation in the supraglacial environment. Flow tills in Spitzbergen were investigated by Boulton (1968, 1972a), who described them as till released as a fluid mass from the englacial debris load when this is exposed by the downwasting of the glacier sUlface. The telm melt-out till was fonnally introduced by Boulton (1970b), although the concept was outlined nearly 100 years previously (Goodchild 1875). Melt-out till is released as; (1) subglacial melt-out till at the base of stagnant ice, or (2) supraglacial melt-out till at the top sUlface of an ice mass.

The classification that resulted from the initial tripa11ite division of lodgement, flow and melt-out tills (Boulton 1971, 1972a), (Table 3.1), had restricted application. Glacial geomorphologists responded by developing classification systems to account for the environment beyond the immediate glacier margin (e.g. Lawson 1979a, 1981b; Boulton & Deynoux 1981, Dreimanis 1988). Lawson (1981b), for example, proposed a classification based on the debris source and location as well as primalY and secondary 23 processes and their relative position. This classification recognIses secondary processes as being critical to environmental reconstruction, but they are classified as they would be in other sedimentary environments, and not as till.

Table 3.1: Tripartite division of flow till, melt-out till and lodgement till.

Supraglacial tills: flow till melt-out till

Subglacial tills melt-out till lodgement till

Polar environments have been the focus for significant modifications to previous classification systems. A non-hierarchical classification of till was proposed by Shaw (1977a), (Table 3.2). This includes, according to Shaw's model of deposition, the characteristics of till deposited in arid polar environments by sublimation. Sublimation till results from the slow removal of interstitial ice by its direct transformation to the vapour state, which is essentially a freeze-dlying process.

Table 3.2: Genetic classification of tenestrial tills (Shaw 1977a).

Position of Position of Process of Tectonic facies transportation deposition deposition

Supraglacial Proglacial Lowered Highly attenuated Englacial Lateral ice-contact Flow Poorly attenuated Basal Supraglacial Melt-out Subglacial Sublimation Lodgement 24

Numerous genetic classification systems have been proposed, which generally attempt to class till types according to their modes of formation, transportation and deposition. They are arbitrary to some extent because glacial deposits may contain tills that grade into one another, but have resulted from different glacidepositional processes. Classifications therefore seem useful in visually organising glacidepositional processes, but do not necessarily establish any practical criteria for their recognition.

DEPOSITION

In many environments melt-out till may be distinguishable from sublimation till because of the presence of substantial amounts of melt water during deposition. This tends to produce drainage structures, stratification, sorting and the readjustment of grain size characteristics. In arid polar environments, melt-water is released slowly, in limited quantities, and till is passively lowered on to the subsh'ate, It therefore displays characteristics that may not be distinguishable from those of sublimation till.

After extensive field investigations in the DIy Valleys, Humphreys (1993, 81) concluded that, "there is no sure way of distinguishing sublimation till from slow melt-out till and a genetic interpretation can not be made with any certainty." This is a question of equifinality, whereby two different processes may lead to the same result. For this reason, melt out and sublimation in arid polar environments are classed together as passive deposition.

This term avoids any genetic classification, which tend to bracket and simplify contemporary glacidepositional processes. For example, Shaw's (1977a) genetic classification of telTesttial tills (Table 3.2) excludes important intelTelationships between glacidepositional processes. In this 25 section the debate that evolved from Shaw's (1977a & b) work in the Dry Valleys is outlined, and important spatial and temporal relationships between sublimation and slow melt are illush'ated in two conceptual models of passive deposition,

Until recently sublimation has been accepted as the dominant deposition process in cold arid environments (Bull & Camein 1970; Shaw 1977a & b; Chinn 1981, 1991). Shaw's (1977a) arid polar model of deposition rests on the assumption that till is deposited primarily by sublimation. Intense folding in the basal zone and subsequent down wasting of the ice front results in the exposure of a hummocky topography (Figure 3.1a). A similar origin was proposed for transverse moraine ridges in Sweden (Minell 1977) and Scandinavia (Shaw 1979). This surface expression of the basal zone was evident in Rains & Shaw's (1981) model of controlled moraine development in Antarctica (Figure 3.2). This also included sublimation till as a significant component of the resultant landforms.

Shaw's (1977a) model of deposition fOlmed the basis for a distinct facies sequence for till deposited by polar glaciers (Shaw 1977b). This was described by Shaw 1977b, and subsequently incorporated into Eyles et aL's (1983) facies model for polar glaciers (Figure 3 .1 b).

Shaw (1977b) accumulated the majority of evidence to support this facies sequence at an exposure cut by a melt water sh'eam from the Lacroix Glacier, 200 metres from the glacier telminus (Figure 3.3). Supporting the glacially derived material is Facies A; a subsh'ate of undisturbed proglacial sediments. Facies B is 'a complex of poorly sOlted, crudely stratified flow deposits intercalated with discontinuous s0l1ed sediments (Shaw 1977b, 199), that are related to processes that occur on the frontal apron. According to Shaw (1977b), sediment on the apron slumps, leaving a flow scar which becomes a drainage zone. Fines are consequently winnowed by fluvial activity, leaving a relatively coarse, crudely sh'atified sediment. Small alluvial fans and sheet erosion deposits accumulate at the apron toe. Shaw (1977a) terms this the poorly attenuated facies. Facies C consists of 26 foliated layers and isolated clasts. This seems to conespond to the highly attenuated basal layer, discussed by Shaw (1977a).

The origin of the Lacroix exposure is unclear. Shaw (1977a) states that two of the tills were deposited by the LaCroix Glacier and the basal part of the sequence by the Taylor Glacier. He cites "evidence from fabric and the regional glacial history (Shaw 1977a , 244)," which was not presented. The LaCroix Glacier was visited in 1998 and the nature of the till within the exposure seems incompatible with the basal zone of the LaCroix Glacier which contains velY low sediment concentrations and few clasts. The glacier seems incapable of depositing a thick till sequence, as described by Shaw (1977b). It may instead have been deposited by the Taylor Glacier, which dominated the valley during the last glaciation.

It is insufficient to base a general model for glacial deposition on the examination of till deposits in the Taylor Valley alone, pa11icularly as the glaciers here exhibit wide-ranging depositional characteristics. The Taylor Glacier, for example, has been found to be polythennal (Robinson 1984). This finding was used by other authors (Dreimanis 1984; Kanow 1984; Kemmis & Hallberg 1984; Eyles et a1. 1986) to review the validity of a distinct arid polar model. They rejected sublimation as a dominant rill forming process and regarded the depositional sequences described by Shaw (1977a & b) as products of velY slow melt. This was supported by Humphreys (1993) and Humphreys and Fitzsimons (1996) who found no evidence of sublimation till in the DIY Valleys, the presence of preserved engiacial strdcture or a highly attenuated till facies. Shaw (1988) conceded that very slow melt occurs at the base of the Taylor Glacier but asserted that basal melt does not occur around the margins where till is being released supraglacially. A 27

Only minor reworking Moulded drumlin-like

Sublimation till Englacial structures preserved

Gm Interpretation May contain some evidence of till flow, sliding! Dms(r)/Sm slumping or free fall. Occasional signs of Oms sediment reworked by meltwater. Des Fold structures may be picked out by the Dms stratified till. Clast orientation reflects structure Stratified till, maybe picked out by alternating Dms patterns of clast size or density.

SYMBOliS

Diamictons ~ Size of symbol ~ p;oportionalto clast I"~:':.~~::' Gravel ~--'-:7'1 sIZe Sand

t_ ~ . Laminations (spacing proportional to Eriosional ;::;: ---:'1 Stratified Sheared thickness) - with silt and clay clasts. Confonnable Iii! - with dropstones. Loaded II (I 'I IJ Jointed • with loading structures. Interbedded

Figure 3.1 A. Model of deposition in add polar environments (Shaw 1977a). The fIrst two stages (transp011 and deposition) are shown in Figure 28

2.6. B. Facies model of till deposition in arid polar environments (Eyles et al. 1983), interpretation from Bennett & Glasser 1996.

ACTIVE STAlE -.f ,- ,---- -,.::.

~~~~~~~~~~~~~~:J sublimation and melt-out Ice core till

Figure 3.2: Model of controlled moraine development (Type 2), (Rains & Shaw 1981). 29

A Proglacial fluyial facies

B ApIOO flow and stralified sediment lac,es

C, Slightly attenuated basal till facies

C2 Highly atlenuated basal till facies

Diamicton. dispersed large clasts. intensely deyeloped horizontal jointing

Massive diamiction rare stratification, high concentration C1 of large clasts transillonal . contact wilh oyerlying bed Contorted Slratified~~s-an-d-- B Crudely stratified,large clasts, tenses of sorled~ules, contorted SlI!Ji!~$ Coarse ·anouiar gravels

Interbedded sands and gravels A paleo-channels wilh sand infill common. Small scale cross laminalion and horizontal stratificalion,

Section obscured by talus

Figure 3.3: Shaw's (1977b) interpretation of the glacigenic sequence at the LaCroix Glacier, Taylor Valley, Antarctica. 30

The occunence of sublimation till in mid polar environments has not been discounted. In fact, Shaw (1985) presented photographic evidence of sublimation till on Vega Island, Antarctica (Figure 3.8), and Lundqvist (1989) proposed that sublimation is a significant till-fonning process at the S0rsdal Glacier on the East Antarctic Coast. However, sufficient evidence has been presented subsequent to Shaw's (1977a & b) papers to doubt that sublimation is the dominant mechanism for deposition in all cases. However, there is no evidence to suggest that melt-out and sublimation do not occur concunently and in close association.

To illustrate this point, a stagnant body of deblis-rich ice that is isolated from active glacier ice will be considered. Supraglacial melt-out till will form on the smface of a body of ice at a rate dependent on the surface radiation balance and the deblis content of the ice. This overburden is susceptible to secondaty flow processes (Figure 3.7). It usually rests at a high angle of repose, in excess of 30° at the Taylor Glacier (Shaw 1977b), and dries out quickly in conditions of low relative humidity, producing a weakened till strQcture.

The supraglacial melt rate will initia11y be fast because a thin debris cover promotes ablation by reducing the albedo of the sUlface. When the overburden thickens, debris accumulation rates decrease because the downward heat transfer is limited by the thennal conductivity and thickness of the debris layer. Melt will cease when the overburden it is of sufficient thickness to insulate underlying ice from the sUlface heat flux. Chinn (1991) found that 80 centimetres of debris is sufficient to preserve ice cored moraines from melt-out in the DIy Valleys. Shaw (1977a) thought that this overburden must equal the thickness of the 'active layer: or the depth of seasonal melting, for supraglacial melt to cease.

Benn & Evans (1988) expressed the effect of smface debris layer thickening on ablation rates in the following equation:

(3.1) 31 where Abb' is the ablation rate for buried ice, kd is the thennal conductivity of the debris layer, hd is the debris layer thickness, Pi* is the density of ice with dispersed debris, L is the latent heat of fusion of ice, and To is the temperature of the debris smface. This relationship illustrates that if all other factors are held constant, the ablation (or melt) rate is inversely related to the overburden thickness.

Sublimation occurs below the insulating overburden, also depositing till supraglacially. Subglacially deposited sublimation till is precluded because sublimation will not occur given the absence of a vapour-pressure gradient between the debris-rich ice and the substrate (Shaw 1986).

The sublimation process is extremely slow. Bell (1966) calculated debris-release rates for the McMurdo Sound region, Antarctica. The time taken to remove interstitial ice, and to release debris, is given by

t Pi Pa / 2YJ /). R (3.2) where t is the time, Pa is the density of air, Pi is the density of ice, z is till thickness, YJ is the viscosity of air and /). R is the difference in vapour density between the top of the debris-rich ice and the atmosphere at the top of the debris layer. Calculations made by Bell (1966) indicate that it would take approximately 2000 years to release 1 metre of till but 7000 years to release 2 metres. In other words, it is expected that between one and two metres of s!lblimation till would be deposited over the time scale of Pleistocene glaciations (Shaw 1988).

Subglacial melt occurs at the base of the debris-rich ice. The rate of till deposition here is independent of the smface radiation balance and depends on the geothennal heat flux. Boulton (1970b) suggested that low geothennal heat fluxes in most areas, (relative to the subaerial thenna! gradient), cause top-melting to be dominant and Lawson (1979a) found that rates of subglacial deposition at the Matanuska Glacier were an order of magnitude slower than rates of supraglacial deposition. 32

Sediment deposition can, according to the above discussion, occur concurrently, at a number of structural levels within a stagnant body of ice. Till types will grade into each other in a vertical sequence and when interstitial ice has been completely removed, a downward facies sequence of supraglacial melt-out till (overburden), sublimation till and subglacial melt-out till can be envisaged. This facies sequence and the spatial and temporal relationships between sublimation and melt-out till are illustrated in Figure 3.4. Under restricted conditions, (as outlined in Part 4), and where it is not exposed to secondary redistribution, sublimation and melt out tills may display structural characteristics inherited from the parent ice. Solar Radiation

Surpaglacial melt-out

::: 0 r;;:::: >. ..0 v c:: 'N0 ~ c:: o Il) v s (/.l :.a (Il Il) v til ....; 2:l on g Il) "'C.1 til j:.J:..; til § Il) '"0 c:: eu a ,S:..... Q., ::l ...... 0 c:: 6 Il) ::: 't:..... til S o Il) Q :.a ;> 2:l 0 Subglacial S til ~ ~ "'C.1c:: 0 E u "'C.1 Il) Vl § t Time ~thennalbeatflux

Figure Model of passive deposition, spatial temporal relationships between processes in polar environments. 34

An alternative way to illustrate intenelationships between till-fonning processes is on a genetic till prism. This was conceptualised by Dreimanis (1988), (Figure 3.5) and developed by Hickock (1990), (Figure 3.6). Hickock (1990) suggested that this prism could be adapted to represent deposition in arid polar environments by either including sublimation with melt-out, or by adding a fourth side to the prism. This concept is illustrated in Figure 3.7 and, like Figure 3.4, represents deposition from a stagnant body of ice. The anow illustrates the redistribution, by flow, of supraglacial melt-out till. This facies is most susceptible to secondary redistribution, but it is recognised that other secondary processes may operate in arid polar environments. More complex till histories may also be represented on this prism (Hickock 1990).

Both 'models of passive deposition' (Figure 3.4 & 3.7) are considered useful in visualising the processes responsible for passive deposition, and the preservation of englacial structure in till. Unlike genetic classification systems, these models illustrate the intenelationships between till-forming processes, and recognise that these processes do not occur in isolation, but may instead occur in close association both spatially and temporally.

LODGEMENT TILL DEFORrvfA TION TILL

GRAVITY

Figure 3.5: Genetic till prism (Dreimanis 1988). 35

I I I

I I ,, I secondary tills PRIMARY TILLS --- MELIOUT OF LODGED SLABS OF MELTOUT DEBRIS, RICH ICE (M)

Genetic till prism (Hickock 1990).

Sublimation Deformation supra glacial (D) subglacial • PRIMARY TILLS SECONDARY TILLS

gravity flow

Lodgement Melt-out (L) (M)

Figure 3.7: Hickock's (1990) genetic till prism adapted to illustrate passive deposition in arid polar environments. Sublimation was added to the top left comer. The circles at the bottom comer represent melt-out till (both subglacial and supraglacial). The alTOW indicates the redistribution of supraglacial melt-out till by gravity flow. 36

CONDrrIONS FOR THE PRESERVA TION OF ENGLACIAL STRUCTURE

The hypotheses presented in Chapter 1 suggest that sublimation and melt-out till can be recognised primarily by the preselvation of englacial structure. This can include fabric and textural properties as well as, in more restricted cases, the preselVation of glaciotectonic stlUcture such as folding, fissility and foliation. Specific propel1ies of both the basal and proglacial environments that lead to the preselVation of englacial structure will be outlined. These are finally developed into a set of conditions under which englacial structure can be inherited by sublimation and melt-out till.

The extent to which englacial stlUcture is preselVed is initially controlled by characteristics of the debris-laden ice, in pat1icular the debris concentration. If there is a high debris concentration, that is well dispersed in the basal zone, structures are supported by interpru1icle contacts. Debris consolidates vertically with little lateral dislocation and the geometIy of glaciotectonic structures, including foliation, augen, laminae bent around clasts and attenuated folds, are retained in the till deposit (Dowdeswell & Sharp 1986; Benn & Evans 1998), (Figure 3.8). The extent to which these souctures are compacted depends on 'the amount of thaw-consolidation, which is cono'olled by the debris content.

Pebble fabrics (orientation of the long axis) in till may reflect the englacial fabric, which is generally aligned in the ice-flow direction (Lawson 1979a & b). Till fabrics do, however, generally s~ow a reduction in dip values I from rotation during thaw-consolidation. This also causes an increase in the dispersion from the mean orientation, relative to englacial fabrics (Benn & Evans 1998). Thaw-consolidation may affect supraglacial melt-out tills to a greater extent than subglacial melt-out tills as they are more frequently subjected to secondary processes, and are often not confined during deposition (Lawson 1979b). ==

(j 38

Low debris volumes allow lateral movement of particles during deposition and the consequent destmction of englacial stmcture (Fitzsimons 1990). Differential compaction may also occur, and sediment has been observed to drape over large clasts as a result of this during melt-out (Shaw 1979, 1982, 1985 & Bouchard et a1. 1984) and sublimation (Lundqvist 1989 & Dreimanis 1982). Sediment may collapse, causing a readjustment of grain contacts and particles become more closely packed as finer particles fall into pore spaces between larger ones (Lawson 1989).

As illustrated in Part 3, an overburden of supraglacial melt-out till is integral to the formation of sublimation till. If the ice contains low debris volumes there will be insufficient sediment to fOlm this overburden. The ice core then melts rapidly, leaving only a thin veneer of till. The overburden must also be thick enough to provide a confining load to allow the debris to consolidate without deformation during melt or sublimation. If till is released without a confining load it may move under the effect of gravity which causes flowing, sliding, spalling, slumping and dropping (Dreimanis 1982).

For till to be deposited passively there must be minimal sorting by water during its formation. As the release of melt-water is inherent to the melt-out process, excess melt water must be removed efficiently from the till matrix for passive deposition to occur. This is primarily controlled by the melt rate and the permeability of the surrounding material. Fast melt rates, in conjunction with low permeabilities results in the retention of excess melt water and elevated pore water pressures. This lowers the shear strength of the till so that structures are destroyed by secondruy flow processes (Boulton & Paul 1976). Loading and local debris flows may destroy englacial characteristics on slopes as low as 3° if the drainage is poor (Paul & Eyles, 1990).

The void ratio of melt-out till is a main control on its pelmeability. This is controlled by the effective stress on the smface of melting. Effective stress is the normal stress minus the pore water pressure. Morgenstem & Nixon (1971) found that the thaw consolidation ratio, R, is a significant factor in pore water pressures within melt-out till: 39

R (3.3)

where oc is the rate of advance of the thawing front and Cv is the coefficient of consolidation. The value of R is highest when the material is unconsolidated (low coefficient of consolidation) and melt is occUlTing rapidly. In these circumstances substantial pore water pressure and sediment deformations can be induced during melt-out (Paul & Eyles 1990).

Till may initially retain a high void ratio when deposited. Boulton & Paul (1976) demonstrate that newly released melt-out till has a high void ratio, which is reduced when till consolidates under high overburden pressures. For example, at Aavastmarkbreen, Oscar II Land, a void ratio of 0.86 was recorded near the surface of a till sequence, whereas at 1.4 metres the void ratio was 0.68. Lawson (J979a) documented a similar spatial variability in till porosity at the Matanuska Glacier, but Ronneli & Mickelson (1992) described newly deposited subglacial melt-out till with a high porosity.

Although high void ratios (and pelmeabilities) may facilitate passive deposition by allowing the free-drainage of melt-water, this also tends towards poor consolidation. This occurs in sublimation till because of its extremely slow rate of deposition (Be111966) and may occur in melt-out tills, pruticularly when the rate of melt-water release is slow, as in arid polar environments. Once deposited, poorly consolidated sublimation and melt-out tills have a low preservation potential when subjected to syn- and post-depositional processes.

In polar environments subhOlizontal jointing, which probably results from desiccation cracks due to drying at the smface after sublimation or melt out, weakens the surface structure of tilL Lundqvist (1989) noted that sublimation till found at the margin of S0rsdal Glacier, Antru'ctica, was structurally weak and could not survive any contact with water. The resultant mass flow deposits do not generally retain the characteristics of the Oliginal till. He concluded that the poor consolidation of sublimation till, in conjunction with the high probability that it will be exposed to water, may account for the infrequency of this till type, 40

Passive deposition is the prereqUIsIte for the retention of parent, englacial structnre in till deposits, and the above discussion can be summarised into a set of basic conditions for its occunence. They relate to the character of the basal ice, and to the nature of the pro glacial environment. There must initially be a high debris content, which is well dispersed within the basal zone; till must be released under a confining overburden so that it consolidates under its own weight with little lateral dislocation; excess melt-water produced during thaw must be removed efficiently and finally the till must be isolated from secondary redistribution processes.

Provided that the above conditions are met, the natnre of the depositional process will operate in conjunction with the climate of the terminus area to preserve englacial structnre. However, for any stlllcture to be recognised, it must have well defmed contacts with adjacent sediments. This may occur due to a different texture, colour or grain size which makes the stlllctnre identifiable within an otherwise homogenous till deposit. Differences may occur due to selective entrainment of a pa11icle-size range (Boulton 1975), conditions of transport and deformation. Thus, a fUI1her condition that adjacent englacial structnres must be lithologically variable must be included. This relates not to the preservation of englacial structnre per se, but for the ability of preserved structures to be recognised within till deposits. 41

CHAPTE 4

RESEARCH STRATEGY

1 BASAL ICE SAMPLE COLLECTION

Fieldwork this study was undeltaken in the Taylor Valley, South Victoria Land, Antarctica (Figure 4.1). The area, known as the 'Dry Valleys,' is located between 77°15'S and 78°S and is the focus of much current glaciological research. Six major valleys; the Mackay, Victoria, Wright, Taylor, Ferrar and Keottlitz have a combined area of around 2500 square kilometres. They traverse the Transantarctic Mountains from the ice-plateau margin to the coast of McMurdo Sound (Wheeler 1961).

The Taylor Valley is one of three valleys in this sequence (including the Wright and Victoria Valleys) that is not dominated by a major outlet glacier, and consequently has a large propOltion of ice-free land. Small, dry-based alpine glaciers extend down the valley sides (Figure 4.2). These glaciers initially form from accumulations of snow which develop in depressions at high elevations (Chinn 1990). The accumulation area of the Suess Glacier is Oil the Asgaard Range, and the glacier flows south-east into the Taylor Valley. It terminates on the valley floor, creating a narrow passage known as 'The Defile,' between the glacier margin and the valley side. The right margin extends up-valley, where it is flanked by Lake Popplewell (Figure 4.3a).

Exposures at the base of the right margin show layers of regelation ice, which is dark and bubble-free, interspersed with sediments from the valley floor and layers of bubbly ice. There is a single outer ridge (Figure 4.3b), interpreted by Fitzsimons (1996) to be a thrust block moraine. Distinct organic-rich bands dip up-glacier and can be traced along the 42

surface of the moraine. The left telminallobe extends down-valley and has two ice-cored moraine ridges about twenty metres high,

Glaciers where basal ice samples were collected ® Sampling sites

o 10 I ••

SOUTHERN OCEAN

RoSJ Sea

"- \ o 1000 " \ I I \ \ I \ \ I 180·E , 80·S

Figure 4.1: Location map - The DIy Valleys, after Fitzsimons (1996). 43

Figure 4.2: The Taylor Valley, with the Suess Glacier in the foreground. (Photo: C. Schliichter) 44

Figure 4.3 A: The Suess Glacier right margin, with Lake Popplewell in the foreground.

SW NE + ~ • + .. .. + .. • + 10m + .. t + • <- ~ + + L , • 10m

+ t + ; , .. <- + <- ~ • .. + Delta

Figure 4.3B: Profile of the Suess Glacier, right margm (Fitzsimons 1996). 45

The basal ice samples from the Suess Glacier were cut from the interior of the glacier. This was made possible by the use of a tunnel which was excavated in 1996-1997 by Dr. S. Fitzsimons (Otago University). Along with drilling and removing core samples in a vertical section through the glacier, tunnels that extend into the glacier margin are a common way to observe the basal zone. However, this has only been possible in areas where there is minimal subglacial melt-water such as the Greenland Ice Sheet (Swinzow 1962, Hooke 1970) and the Meserve Glacier, Antarctica (Holdsworth 1974).

The tunnel entrance at the Suess Glacier is located midway along the right margin, and extends approximately thilty metres beneath the glacier. The whole basal zone was visible in the down-glacier branch of the tunnel (Figure 4.4). Here, amber ice, which is discoloured by fine dispersed sediment, lies above a basal zone which is up to 5 metres thick. Sedimentary structures lie within a matrix of predominantly bubbly ice in the upper 3-4 metres. This is bounded at the upper surface by a thick band of sediment, up to 0.5 metres thick, that peters out towards the tunnel entrance. It contains coarse sediment and velY little interstitial ice (see Appendix 1 & 2, sample 12). In the lower basal zone clear, refrozen layers, and more complex deformational structures are common. Most of the blocks were cut from this lower zone, at heights vatying from 0.2 - 1.75 metres above the tunnel floor. The tunnel floor rested on a hard layer that prevented further downward excavation, and was therefore presumed to be the lower boundary of the basal zone.

Fomteen blocks of basal ice were cut from the tunnel walls, using a chainsaw. All the blocks were photographed and described in the field and their dimensions and orientations noted. They were sekcted to give a good representation of the basal zone. A smaller sample block, approximately 15 centimetres square, was cut adjacent to each larger block for later detelnllnation of the sediment concentration and for pmtic1e size analysis. 46

Figure 4.4: The down-glacier branch of the tunnel, showing the basal zone. 47

A similar sampling procedure was applied at the Taylor Glacier (Figure 4.5), although only the margin was accessible. This is a large polythermal glacier that drains from the East Antarctic ice sheet. It is approximately 90 kilometres long and played a major role in shaping the valley during late tertiary glaciations (Calkin 1973; Calkin & Bull 1974; Denton et al. 1984). Four blocks of basal ice (and small sample blocks), were cut from the exposed margin of the Taylor Glacier, which is accessible around four kilometres of the tenninus (Lawson 1996). The basal zone at the margin is up to 5 metres thick and has a sharp contact with the overlying ice (Figure 4.6). 48

Figure 4.5: The tenninus ofthe Taylor Glacier with Lake Bonney in the foreground.

Figure 4.6: The Taylor Glacier basal zone, showing a sharp contact with the overlying glacier ice. 49

4.2 LABORATORY METHODOLOGY

4.2.1 Sediment size and concentration

Sediment size and concentration were detelmined using the small sample blocks. Fifteen of these were analysed from the Suess Glacier, and four from the Taylor Glacier.

The blocks were first weighed, then melted in funnels lined with filter paper. The sediment was oven-dried and the sediment concentration by weight was determined by expressing the sediment as a percentage of the original sample weight. The sediment concentration can also be expressed by volume, which is detelmined by conveliing the water volume (collected after melt) to an ice equivalent volume. It was considered that there would be a greater error in volume calculations for the following reasons: water is lost to evaporation during melt; a specific gravity for the sediment is assumed in volume calculations (which is not necessalY when calculating concentration by weight) and the volume of the ice and water vary with temperature, which was not constant in these experiments.

The mean particle size and sOlting for the Suess Glacier samples was determined using a MacAtthur Rapid Sediment AtIalyser (Appendix 1). The sample was manually released into a settling tube, and the fall velocity of each 0.1 phi interval is recorded and convelied into weight data by the programme. An equivalent density of 2. 65grn/cc was used, as the sediment is quartz-based.

4.2.2 Melt-out experiments

These experiments were conducted over a tluee month period using the large sample blocks cut from the basal zones of the Suess and Taylor Glaciers. The methodology employed was designed to simulate passive deposition by slow melt. In this way the stmcture observed in the basal ice samples can be compared with the laboratOly-generated till deposits. 50

It was impossible within the time frame available to melt the ice at a rate comparable to the melH'ate in polar environments. However, other factors considered integral to passive deposition by melt (see Chapter 3.4) were controlled.

The first melt-out experiment was conducted in the field. A block of basal ice was cut from the Suess Glacier to fit a plastic box with drainage holes. It was weighted from the top and melted over a period of 13 days. This was unsuccessful, as differential melt at the sides caused sediment to slump, rather than consolidate veltically.

Techniques to allow free drainage, whilst insulating the sides of the block from melt (to prevent slumping), were trialed in the laboratory. Blocks of ice were made by freezing altemate layers of sediment and ice. Fluorescent tracer sand, used in coastal and fluvial studies, was initially used. However, the epoxy coating on the sediment pmticles prevented the sediment bands from freezing at an average temperature of _6°. Instead, alternating layers of black beach sand and Suess Glacier moraine sediment were used (Figure 4.7). The sediment-laden ice was frozen into a clear plastic box, and then transfened into an identical box that was drilled with 2 millimetre holes, at 15 millimetre intervals on the base and sides. This was supported on a stand which prevented the ponding of melt-water at the base of the block.

The slr1~s of the box were wrapped in Bidim A24, a geotextile filter. This was chosen because of its high pelmeability (Bidim guidelines: Geofabrics Australia PTY Ltd.). It is also reasonably thick (3 to 4 millimetres) which allowed water to drain out laterally, and then filter vertically through the fabric, rather than draining between the box and the filter. This would have caused edge melting, and may have allowed water to drain back into the sediment. Tenaml000 was used in the base of the box to prevent sediment escaping through drainage holes. It has good permeability (Terramdata: Ground Engineering New Zealand), but has a smoother surface than Bidim, so that when it is in direct contact with the sediment, it does not disturb the lower boundary of the block. Finally, the sides were insulated with 40 millimetre polystyrene to prevent edge melting. 51

Melt was observed over four days, by removing the insulation and rephotographing the block. There was some melt from the sides, or 'edge effect,' but sediment bands consolidated vel1ically without slumping. After complete melt, the boundaries between the bands were distinct, and some were observed to drape over clasts within the moraine sediment (Figure 4.8). This occurs as a result of differential compaction during melt-out (Bouchard et at 1984; Shaw 1979, 1982, 1985) and sublimation (Lundqvist 1989 & Dreimanis 1982).

This trial was considered successful in providing free drainage of melt-water and depositing till passively. It was, therefore, applied to the melt-out of samples from the Suess and Taylor Glaciers. Boxes were constructed from 3 millimetre plastic sheet. The sides were detachable and were adjustable by 100 millimetres at each comer. They rested on a plastic base-board, which along with the sides, was drilled with 4 millimetre holes at 3 centimetre intervals (Figure 4.9). Three different sized boxes were made so that concunent experiments could be run, and aU block sizes were accounted for. The geotextile filters were used in the same way as the melt out triaL Polystyrene insulation was replaced with 8 centimetre thick closed-cell foam, wrapped in reflective plastic. This has better insulating properties which, it was hoped, would reduce edge melt (Figure 4.10).

The ice was initially stored in a walk-in freezer, which allowed the structure on each face to be sketched, measured and labelled according to the facies codes described in Chapter 5. The blocks were trirrimed with a continuous blade, diamond-tipped circular saw and weighed. The saw produced flat surfaces that could fit neatly against the sides of the plastic boxes. For blocks containing folds, fine holes were drilled at hinge points and crests, and plastic··coated wire was inse11ed~ This was intended to give some indication, upon melt-out, of the position of these fold profile features in the till. Their position in relation to the edge of the block was measured before and after melt so that the amount of vel1ical consolidation could be quantified. 52

Seven blocks from the Suess Glacier and three from the Taylor Glacier were melted. The maximum and minimum air temperature was recorded daily. Once melted the sides of the box could be unscrewed and carefully removed. The sediment was then photographed and any of the structure measured and sketched. The melt-out till was then refrozen on the plastic base-board. The frozen sediment was cut (with the diamond tipped saw used for trimming the original samples), so that the cross-section could be examined for any internal structure that may have been preserved. The sediment was then oven-dried and reweighed to detelIDine the sediment concentration of the block. This is subject to a greater margin of error than the small sample blocks, which were melted specifically to detelIDine sediment size and concentration, as some sediment was lost during cutting. This was estimated to be 5 grams per block.

The till was oven-dried at low temperatures (20°) to prevent cracking. The Suess Glacier samples did not contain a significant silt fraction to bond pmticles, and the till collapsed easily upon the removal of a supporting container. The Taylor Glacier till samples retained their fOlID.

It was originally intended to make thin sections of all the till samples. This was trialed with two samples. A small section was cut from the frozen sediment. A cardboard and foil container was constructed to fit the sediment block and it was oven-dried until the weight stabilised, and all water was therefore evaporated. An epoxy resin with a fast curing time and good optical jJropelties was impregnated into the sample using a vacuum. A one millimetre thin section of this was made. Although the thin sections were more pelmanent and durable than the frozen samples, they did not show any structure that could not be observed in a frozen cross-section. 53

Figure 4.7: Test block of ice made from alternating layers of Suess Glacier moraine sediment and black beach sand.

Figure 4.8: Melt-out trial sediment deposit, showing a layer draped over a clast and distinct boundaries between sediment layers. 54

Figure 4.9: Block 10 encased in a perspex box prior to melt.

Figure 4.10: An assembled melt-out experiment. ss

CHAPTERS

SAMPLE BLOCK CHARACTERISTICS

This chapter describes the bulk characteristics of the basal ice samples used in this study. This is achieved via sedimentary, structural and facies descriptions of the ice samples. An accurate characterisation is considered integral to the subsequent comparative analysis of sedimentruy stmcture in basal ice and tilL The telminology inh'oduced here provides a basis for describing individual samples in detail in Chapter 6.

5.1 SEDIMENTARY CHARACTERISTICS

The sediment in the Taylor Glacier basal ice consists predominantly of fine sand (> 50% by weight), (Lawson 1996). Larger clasts, which are dispersed throughout this fine sediment, vary in size from coarse sand to boulders. The long axes orientation of clasts was measured at two sites (30 clasts per site) in the field, Of those measured, 89% had long-axes oriented in the ice-flow direction. Robil1son (1979) found that the debris content of the basal ice ranged from <1% to >500/0 by volume. The average concentration of sediment in the three sample blocks was 52.8% by weight (see Appendix 1).

The sediment concentration of the Suess Glacier sample blocks is shown in Appendix 1 and the size and sorting is presented, along with corresponding facies descriptions (see Part 3), in Appendix· 2. The sediment concentration of the Suess Glacier samples was highly variable and ranged from 3% to 73.4% by weight. This is considered to be in part a 56 function of the variability of the basal zone, but is also due to the sampling procedure, which was not random. A wide variety of basal ice types were deliberately selected. The samples would tend to have higher sediment concentrations than an average value for the Suess Glacier, as they were cut around sedimentary and glaciotectonic stmctures.

Particle sizes ranged from 1.77 to 0.49 phi (medium to coarse sand) and the sediment was moderately to well-sorted. Samples that contained folding had medium sized sand (average 1.66 phi). Blocks 7 and 8 were characterised by banding and foliation, and the sediment was notably coarser, with an average particle size of 0.5 phi. As the sediment contained in these samples was sand-sized, and larger clasts were rare, fabric measurements could not be taken.

CHARACTERISTICS

Previous documentation of the sedimentalY stmcture produced by entrainment and deformation in the basal zone is sparse. The terminology used to describe it is generally adapted from stmctural geology, primarily because debris-laden ice is viewed as a rock 'analogue.' An attempt is made here to describe the stmctural characteristics of the basal zones of the Suess and Taylor Glaciers within the context of previous work This is done to facilitate the subsequent stmctural comparison between this ice and lab-generated till deposits.

There are two broad classes of englacial stmcture within basal ice; sedimentary (or primary) and glaciotectonic (or secondalY). Sedimentary structure, such as foliation, lamination and banding, is produced by the initial entrainment of sediment. This may be defOlmed by shear and is repeatedly folded and attenuated at high finite strains to produce glacio tectonic structure (Huddleston & Hooke 1980; van def Wateren 1995). Extremely complex pattems of deformation in the basal zone are common. Boulton and Spling (1986) suggest that this is partly due to the 57 complex stress pattems produced by ice flow over an ilTegulaT substrate and partly from the strongly anistropic strain response of debris-rich ice.

It is difficult to distinguish between planar, sedimentary structure that has not been defonned, and highly attenuated glaciotectonic structure. For this reason structme is described here as either planar or folded. Planar structmes may be sedimentary or glaciotectonic, while folds have invariably been produced by glaciotectonic defonnation.

Planar structme occurs at a wide variety of scales. In this study banding is defined as being greater than 10 millimetres thick. Bands have well-defined contacts with sun-ounding ice, and contain little OT no interstitial ice. Finer planar shuctures that occur within a mah'ix of ice are classified as foliae. This is generally defined by variations in bubble or sediment content (Allen et al. 1960, Ragan 1969).

Lamination is initially a planar feature, and is used here in the sense of a layered sequence of lamina (a unit layer) that are less than 10 millimetres thick and are visually separable from layers above and below (Howell 1957). The Taylor Glacier basal ice is characterised by laminated sequences separated by bands of bubbly glacier ice. Lamination may be defonned, as at the Suess Glacier, where it is associated with glaciotectonism (see Part 3, Facies II).

Fold~g is a result of viscosity differences between two intemal variables or inhomogeneities in basal ice which act in association with local stress concentrations (Price & Cosgrove 1997). Inhomogeneities commonly occur where regelation has caused ice-crystal size variations (Lawson & Kulla 1978; 10uzel & Souchez 1982) and/or debris concenh'ations (Hooke et al. 1972). Local stress concenh'ations occur where there are changes in the bedrock topography, and the debris-laden ice becomes passively defonned by shear into folds (Huddleston 1977). During passive folding layer boundaries exercise little control on defOlmation and serve mainly as markers. Folds, therefore, show flow geometries that reflect the stress field and velocity gradients existing during defOlmation (Donath & Parker 1964). 58

Most fold descriptions are restricted to the fold profile. Some important geometrical features of a fold profile are shown in Figure 5.1. The hinge of a fold is the point of maximum curvature, and inflection points occur when the rate of change of curvature is zero. The median surface joins successive inflection points. The crest and trough are respectively the highest and lowest points on a fold above this horizontal datum. The amplitude is the perpendicular distance between the median surface and the fold hinge, and the wavelength is the distance between altemate inflection points (Price & Cosgrove 1997).

Median Surface

Figure : Geometrical features of a fold profile:

A - Amplitiude c - Crest t- Trough cc - Interlimb angle L - Wavelength i-Inflection point h Hinge 59

Most folds, however, occur as a folded layer, and not as a single surtace as shown above. Lines that join points of equal curvature on the inner and outer boundaries of a folded layer are called dip isogons (Elliott 1965). Dip isogon patterns were used by Ramsay (1967) to classify folds (Figure 5.2). Passive defOlmation in basal ice commonly produces similar folds which occur when fold boundaries have curvatures that are identical, or very similar. Consequently, the limbs of similar folds are generally thinner than the apices (Hills 1959) and they have constant dip isogon thicknesses around the hinge zone (Donath & Parker 1964). Figure 5.3 shows field measurements of similar folds, produced by passive deformation.

Strongly convergent Parallel

Weakly convergent Similar Divergent

Figure 5.2: Classification of folds according to dip isogon patterns (Ramsay 1967). 60

Figure 5.3: Field measurements of similar folds produced by passive deformation (Donath & Parker 1964).

When a fold is considered in three dimensions, the planes that intersect successive hinge points and crests are known as hinge and crest surfaces respectively (Price & Cosgrove 1997). The axial surface is equidistant from either limb of the fold (Hills 1959). A fold is symmetric when it has equal limb lengths, and an axial sUlface that is normal to the median surface. The converse applies for asymmetric folds (Figure 5.4).

If the axial plane becomes nearly horizontal the fold is recumbent (Figure 5.5). Folds illustrated in Shaw's (1977a) model (Figure 2.6) are recumbent in the ice flow direction. These are said to result from compressive flow, produced by low flow rates and thinning margins, during glacier retreat. Huddleston (1976) demonstrated that recumbent folds may also be 'seeded' by irregular bedrock topography. Isoclinal folds occur when the axial plane tends towards horizontal, and the two limbs are approximately paTallel. An isoclinal drag fold was described by Hamilton & Haynes (1961) at the Taylor Glacier (Figure 5.6). This was inferred to have fOlmed passively, and reflected differential flow in zones of shear. 61

Hinge lines of fold (1)

Figure 5.4: Geometrical features of a folded layer. Fold 1 IS symmetric, folds 2 & 3 are asymmetric (Price & Cosgrove 1997).

Because the ice samples used in this study needed to be transported a considerable distance for laboratOlY analysis the folds described in Chapter 6 are small. Some samples from the Suess Glacier contain folds that were cut from the inside hinge of larger folds (with limbs of up to 1.5 metres), or were parasitic folds. They were generally similar and recumbent in the ice flow direction. Figure 5.5: A recumbent fold at the Tayl.or Glacier (Sha\v 1985).

Figure 5.6: Isoclina'l drag fold at the TayllJr (Jlacier. .'\ntarctica (I fal11ilton &. Haynes 19(1) 63

BASAL ICE FACIES

A facies is an ice 'type' or an objectively described unit (Reading 1979). They can be distinguished on the basis of grain-size and texture, mineralogy, geochemistry (isotope analysis), sedimentary structures and ice-crystal characteristics. The range of ice characteristics present in the Suess and Taylor Glacier samples is described here by a visual facies classification. This is set within the context of facies analyses at sub-polar and alpine glaciers.

Basal ice observed in tunnels, boreholes or at the margin of a glacier can be used to establish relationships between facies and their processes of fonnation. This may indicate subglacial processes occulling at inaccessible bed zones (Knight 1993), or if a particular facies is linked to an entrainment process it may ideally be used as a proxy for subglacial conditions at other glaciers where it is observed (Hubbard et a1. 1996), Facies descriptions are used here simply as a tool to characterise basal ice samples from the Suess and Taylor Glaciers, prior to a stmctural comparison with lab-generated till deposits. A more comprehensive investigation would be required to relate these facies to any specific entrainment processes.

Lawson (1979a) produced an influential description of basal ice facies at the Matanuska Glacier, Alaska (Figure 5.7a). He documented an upper englacial zone, containing diffused and banded facies, and a basal zone containing dispe'3ed and stratified facies. The stratified facies was fUlther divided into three sub-facies: discontinuous, suspended and solid. The origin of each facies was primarily defmed by the physical characteristics of the ice and debris and isotope analysis (Lawson & Kulla 1978). Lawson (1979a) concluded that the dispersed facies lay at the intelface between supraglacial entrainment of debris and subglacial entrainment by localised pressure-melt and freeze-on. Debris in the stratified facies was derived entirely subglacially by the freezing of melt-water to the glacier sole (Weeltman 1961). Unaltered sedimentary stmctures suggest the incorporation, into this facies, of unlithified sediment as frozen blocks (Weertman 1961; Boulton 1970a). 64

Facies descriptions from the Greenland Ice Sheet introduced the telm 'clotted ice' (Figure 5.7b). Its formation was attributed to Llibouuy's (1993) 'regelation-squeezing' mechanism, where water migrates away from a pressurised interface at an obstacle and refreezes on the lee-side. This process may explain the selective entrainment of clots of silt-sized material in this facies (Knight 1987). Clotted ice has also been recognised in ice-core studies in Antarctica (Gow et al. 1979) and the Agassiz Ice Cap (Gemmell et al. 1986).

Apparent similarities between the facies descriptions at the Matanuska Glacier and at the Greenland Ice Sheet were recognised by Knight (1994) and unified into a single facies model for sub-polar glaciers (Figure 5.7c). Knight's (1994) reconciliation of these facies descriptions is a valuable attempt to introduce a common terminology into a rapidly evolving area of research.

The absence of any similar framework for temperate, alpine glaciers led to the identification of seven basal-ice facies in the westem European Alps (Hubbard & Sharp 1995). They were characterised on the basis of debris distribution, sedimentology and stable isotopic composition.

The clear facies is visually similar to the dispersed facies and clotted facies, and small debris agglomerations give it a u'anslucent appearance. It is devoid of internal layering, massive and debris-poor, often fOlming a maulx for other facies. There are clouds of defOlmed or flattened bubbles and debris 'smears,' which may form from tectonic defOlmation of a debI-is band. It is associated with Qervasive melting at grain boundaries on the stoss-side of bedrock hummocks, and refreezing with the consequent expulsion of water and gas. When healed fractures trap fine debris in discrete laminae, the planar facies is fOlmed. The laminae in this facies often cut across bubble-foliated glacier ice, and cross each other without displacement. Superglacial (Matanuska Glacier) Zone 0 -O.2m uniform distribution, clay-pebble sized. B A - facies has planar contact with englacial zone -:::t:~~~~:..... "!E.q' to increase hi debris concentration. - fewer bubbles. Diffused CLOTTED (Greenland) Suspended lenticular nodules of debris, up to ~~"'10m Banded Scmma. Size and concentration of debris increases Itm;v:lrlrl~ the top of the sequence. to H Range of values Clotted ice Englacial Diffused Zone (Matanuska Glacier) 300m debris concentration relative to Banded facies making an irregular, well

I U"llll'_U contact. - ice dark, bubble-free. .Finely laminated Diffused - intercalation of debris-rich! debris- poor layers, which may be discontinous; lenses and ---- -~=::::::: ~===~ . subfacies. i I Basal -Layers O.OOI-2m thick, limited lateral extent. SOLID '"-- - structure developed e.g. thrust faults and 3-15m compressional folds. I Zone lii~!!5ii'" I I I I I .001 .01 .1 10 100 ---*- Debris volume % c

DISCONTINUOUS SOLID ICE FLOW thickness: O.OS-2m Thickness: O.Ol-1.7m -irregular aggregates (lenses, Well defined layers with internally discontinuous layers, platelets) of SUSPENDED massive sediment. fine sediment, ice may be bubbly. thickness: O.OOl-L2m - undisturbed sedimentary structures - gradational, poorly defined Suspended particles and of fluvial and lacustrine origin. boundaries. aggregates of fine-grained - debris streaming. sediment.

Figure 5.7: Basal ice facies ofpolthem1a1 glaciers. A: Matanuska Glacier (Lawson 1979a); red arrows showing the correspondence with B: Knight's (1987a) facies description of the Russell Glacier, Greenland. C: A schematic section combining basal ice descriptions of Lawson (1979a) and Knight (1987a), after Knight 1994. 66

The laminated facies contains fine (0.1 - 1 millimetre) laminae that have have high concentrations of coarse debris and are isotopically eru'iched. The fonnation of these laminae is attributed by Hubbard & Sharp (1995) to regelation. The interfacial facies fonns in basal cavities as sheets (interfacial layered sub-facies); on sloping bedrock floors or in bedrock hollows (interfacial continuous subfacies) by freezing-on. When this ice is incorporated and transported to the margin it may fonn the dispersed faCies which, like the interfacial facies is bubble-free and contains coarse-grained debris, dispersed in crude layers. The debris may be distributed by enhanced defonnation and crudely stratified by simple shear.

Two basal ice facies identified by Hubbard & Sharp (1995) are comparable to descriptions from the Matanuska Glacier. The solid facies corresponds with Lawson's (1979a) basal stratified solid subfacies. It contains layers or pods of frozen, coarse-grained debris. It is probably frozen, unconsolidated sediment. The stratified faCies, is debris-rich, bubble free and interstratified with layers of glacier ice. Interstratification indicates a tectonic origin, as recumbent folding and repetition of these ice-types is associated with marginal compression, or simple shear over an iITegular substrate (Hubbard & Sharp 1995).

5.3.1 The Taylor Glacier

The basal ~ce at the Taylor Glacier is characterised by planar, laminated sequences. Within this laminated faCies, individual sand and silt laminae are distinguishable by variations in colour. They altemate frequently from dark green to light brown, and are intercalated with clear-ice laminae. They pinch and merge laterally and are frequently dissected by larger clasts (Figure 4.6). The lamination appears horizontal at the margin, however, the sides of the

0 blocks reveal that they are dip upglacier at between 40 0 and 46.

Laminated sequences are separated by bands of bubbly ice that have sharp upper and lower contacts with the laminae. This bubbly ice facies appears identical to overlying glacier ice. It may have been folded into the basal zone by compressive flow at the margin of the glacier. 67

A wide variety of ice types were observed in the Suess Glacier basal ice samples. They displayed a range of sediment concentrations, and recurring structural relationships. It is both difficult and beyond the scope of this work to make direct linkages between facies described at other glaciers and those identified here. However, some of the facies identified seem similar to Hubbard & Sharp's (1995) facies classification. This supports Fitzsimons' (1996) model (Figure 1.7), in which dry-based glaciers that flow into lakes become temporarily wet-based. rThey would thereby acquire basal ice characteristics normally associated with temperate glaciers.

Facies I is bubbly ice, that appears identical to overlying glacier ice. Holdsworth (1974) originally called this the amber facies, as it may contain fme, dispersed sediment which gives it an amber discolouration. It was rarely observed in the samples taken, as most of the basal ice appears to have been partially or completely melted and refrozen (see Facies V - VIII).

Basal ice with a high concentration of sediment was classed as either laminated or solid. Facies II, the laminated facies, consists of inter stratified laminae of clear ice and debris layers (2-5 millimetres thick), (Figure 5.8). The ice laminae are dark and bubble free, and debris laminae have a very high sediment concentration. They are often folded and remain conformable throughout the fold, suggesting that the lamination was formed prior to, and not as a consequence. of glaciotectonic deformation. Laminated sequences commonly grade into Facies III (Figure 5.8 & 5.9). This has a lower debris concentration, and the structure is less defined due to the dispersed nature of the sediment. Faint layering and lamination may be discontinuous and have poorly-defined boundaries with the ice matrix.

Any solid structures that contain very little interstitial ice are classed as Facies IV (Figure 5.8 & 5.10). This mainly occurs in association with banding and folding and would probably be termed 'stratified' under Hubbard & Sharp's (1995) classification. However, this term suggests simple planar features, but the solid facies observed at the Suess Glacier was frequently defomled by folding (Figure 5.7 & 5.8). When a structure contains very high 68

sediment concentrations, but some interstitial Ice IS visible between the sediment particles it is classified as Facies IVa.

The ice matrix in the basal zone of the Suess Glacier varies with the pattern of deformation and the degree of melting. Facies V is bubbly ice that may contain large, elongated bubbles and discontinuous foliation (Figure 5.10). It seems to correspond with the clear facies identified by Hubbard & Sharp (1995). If the bubbles are absent, this ice tends to be transparent, and individual foliae can be observed in three dimensions within a block. This foliation tends to consist of coarse sand, that occurs in discreet, single layers. When clear, dark, bubble-free ice occurs independently of the lamination in Facies II, or is greater than 5 millimetres thick, it is classified as Facies VI. This commonly occurs in isolated, crescent-shaped pockets on the inside of folds. Here, it tends to grade into ice that has not been subject to a complete melt-freeze cycle. This ice (Facies VII) is also sediment free, but is opaque and may have streaks of fme bubbles that give it a marbled effect (Figure 5.8). Clear (Facies VI) ice was also observed between lenses and bands (Figure 5.10) as thicker (20 to 60 millimetre) layers. 69

Figure 5.8: Basal ice in situ showing facies that display typical structural relationships within the ice. Facies VII grades into Facies VI at the fold hinge. Interstratified laminae of clear ice and debris are clearly defined in Facies II. This grades into less distinct laminae (Facies III) that, like Facies 11, are conformable with the fold. 70

Figure 5.9: Facies III.

Figure 5.10: Facies IV (solid) within a matrix of Facies V (bubbly ice) and Facies VI (clear ice). 71

Figure 5.11 (above) & 5.12 (right): complex folds produced by glaciotectonic deformation in the basal zone. 72

CHAPTER 6

COMPARATIVE ANALYSIS

The bulk characteristics of the basal ice at the Suess and Taylor Glaciers were detennined through sedimentological, structural and facies descriptions in the previous chapter. In this comparative analysis each block that was passively melted in the laboratOlY is described in detail. This provides a basis for comparing the structure in the ice samples, and any structure that may be inherited by till deposits.

6.2 GLACIER

These samples were cut from adjacent positions in the basal ice, from a macroscopic fold in the left branch of the tunnel. This fold had a limb length of 112 centimetres and was constructed from composite bands, which produced an outer and an inner hinge, separated by a zone of dispersed sediment. Figure 6.1 shows the fold and the relative in situ positions of the two blocks. The sedimentary characteristics and dimensions of the blocks are summarised in Table 6.1 and 6.2.

This fold could be traced around to the right branch of the tunnel and the hinge appeared to be located at the intersection of the two tunnel branches. The section shown in Figure 6.1 was not considered to be a true profile (nonnal to the hinge surface) because the tunnel branches were cut obliquely through the fold. It was, therefore, considered unwise to 73 classify the fold according to the dip isogon pattems described in Chapter 5.3. However, although dip isogons could not be reliably constructed, the fold could be visually classified as similar because the limbs tapered away from the hinge. Also, by examip.ing the limbs of the fold in both branches of the tunnel, the axial plane appeared horizonta1, and the fold could therefore be described as recumbent.

Figure 6.1: The relative in situ positions of blocks 9 & 10.

Block 10 (Figure 6.3) was cut from the hinge of the original fold. The resultant sample was composed of one folded, solid band (Facies IV), bounded at the upper and lower smfaces by interstratified laminae of clear ice and sediment (Facies II), which were 2 to 5 millimetres thick. The outer edge of the fold was Facies III (dispersed sediment), and this graded into Facies V, which is clear, but contains distinct bubbles, elongated in the ice flow direction. In this case the bubbles were up to 20 millimetl'es long. At the hinge clear, dark ice (Facies VI) graded into opaque ice (Facies VII). 74

TabJe 6.1: Block 10 - dimensions and sedimentaty characteristics.

Block Dimensions: large face (cm) 250x240 small face (cm) 200x240 Average particle size (Phi) 1.69 Size description Medium sand Sorting Well sorted Sediment concentration by weight (%) 20.4

After melt the till deposit was 6.5 centimetres high at the edge closest to the fold hinge, and tapered to approximately 1 centimetre at the opposite edge. The fold hinge was clearly distinguishable in the till deposit (Figure 6.4). Bands that were darker than the sun-ounding sediment followed the shape of the fold limbs and continued around the hinge zone. This banding was very fine with poorly-defined boundaries with the sun-ounding sediment. This made its measurement difficult, but it appeared to range from less than 1 millimetre to around 2 millimetres thick. Its configuration, colour and thickness suggest that it was originally Facies II, which bounded the solid (Facies IV), folded band. If so, then the sediment at the inside edge of this banding would be the Oliginal fold, and the outer sediment would originally have been Facies III, dispersed sediment.

Once the till was frozen and a cross section examined, however, this banding could not be identified. Similarly, when a thin section was made from a frozen block of till taken from the hinge zone, no structure was apparent.

Block 9 was cut from the inside hinge of the recumbent fold, and was composed of intensely deformed, solid sediment (Facies IV), (Figure 6.4). Small-scale parasitic folding can be seen at the upper edge of the fold. The inside of the hinge, like block 10, was clear ice (Facies VI), which graded into opaque ice (Facies VII). Surrounding the outer edge of the fold was Facies III. At the top this had a high concentration of dispersed sediment, but below the fold laminae were more distinct. 75

Table 6.2: Block 9 - dimensions and sedimentruy chru'acteristics

Block Dimensions: large face (cm) 240x260 small face (cm) 190x260 Average particle size (phi) 1.73 Size description Medium sand Sorting Well s011ed Sediment concentration by weight (%) 32.9

The melt-out of this block produced similar results, although the cross-sectional shape of the till deposit more clearly reflected the original shape of the fold (Figure 6.6). The small parasitic folds retained their shape on the surface of the deposit, and the hinge was identified from the position of the wire insel1ed before melt and from the orientation of pru1icles around the hinge zone, Again, once a cross section was cut through the frozen deposit, the patteln of particle imbrication was more difficult to identify. This may be because cutting the sample destroyed the pru1icle orientation at the outer boundary, which in the absence of any colour variations (like those observed in block 10), was the only means of identifying the fold shape in the deposit.

Block 1

This was a small, isoclinal fold, composed of a solid sediment lens in the centre, which was surrounded by laminae (Facies II), which graded into Facies III (Figure 6.7). The ice matrix was, at the top sUlface, bubbly ice (Facies I) and below the fold was Facies VII. The limb lengths were about 250 centimetres from the outer edge of the hinge to the opposite edge of the block, and it was cut 145 centimetres above the tunnel floor.

The melt-out till (Figure 6.8) from block 1 was 4 centimetres high at the thickest point, and like blocks 9 and 10, tapered away from the fold 76 hinge. The hinge could be identified in the deposit from the position of the wire inserted at the hinge point before melt. The inner, solid lens at the centre of the fold was distinguished from sUlTounding sediment (originally laminae) by fme cracks that marked its outer edge. There was no visible structure within the sU1Tounding till. The boundalY between the sediment lens and outer laminae was not visible in a frozen cross section, although, like block 10 this may have been due to disturbance at the cutting interface.

Table 6.3: Block 1 dimensions and sedimentary characteristics.

Block Dimensions: large face (cm) 260x240 small face (cm) 255x240 Average particle size (Phi) 1.67 Size description Medium sand Sorting Well s0I1ed Sediment concentration by weight (%) 21.9

Blocks 7 & 8

These blocks were cut from adjacent positions in a zone of banding towards the tunnel entrance, 20 centimetres above the tunnel floor. One solid sediment band ran through both blocks. This was bounded on its upper surface by laminae (Facies II), which passed abruptly into Facies V. This had distinct bubbles elongated in the ice-flow direction and occasionally fine foliae (2-4 millimetres thick), (Figure 6.9). Under natural light this ice matrix was extremely clear, to the extent that foliation could be seen in three dimensions within the block.

The plastic-sediment interface had a marked effect on the structure observed at the edge of the block 7 and 8 till deposits. Although the I sediment appeared homogenous at the edge, individual bands were separated by cavities that fonned during deposition (Figure 6.10). These were also visible in the block 1 till (Figure 6.8) and block 9 till (Figure 6.6). These cavities were initially thought to be drainage structures. This may be 77 partly correct because the Facies V ice matrix that separated bands in the parent basal ice would have been the dominant source of melt-water released during deposition. However, they all had a horizontal configuration, which would preclude this explanation alone, because vertical drainage structures would be expected to fonn during melt. It seems likely that imprinted on the pattern of melt, was the effect of the melt-water surface tension, which caused frictional drag against the plastic box. This retarded the deposition of individual layers, so that when they were deposited they did not compact sufficiently to be packed against the lower sediment, and the cavities remained.

This explanation accounts for the difficulty in observing the shucture beyond the edge of the till deposits in the block 1 and block 9 till. Examination of frozen cross-sections taken tlu-ough these samples revealed that the cavities disappeared within 4 centimetres of the till margin. It is, therefore, expected that the stmctures observed were upwarped at the sediment-plastic intelface, and more consolidated towards the centre. This is illustrated for a banded sample in Figure 6.2.

In the case of blocks 7 & 8, cross sections taken tlu'ough both wet and frozen till samples showed textural and slight colour variations, inherited from banding in the basal ice. Figure 6.11 shows the a frozen section of till cut from block 8, which was subsequently thin-sectioned. A layer of lighter, and coarser sediment can be faintly seen at the right-hand edge. The thin section did not reveal any fUl1her structure. 78

Table 6.4: Blocks 7 & 8 dimensions and sedimentary characteristics.

Block 7 Block 8 Block Dimensions: large face (em) 240x160 275x180 small face (em) 220x160 175x180 Average particle size (Phi) 0.49 0.51 Size description Coarse sand Coarse sand Sorting Well sorted Well s0l1ed Sediment concentration by weight (%) 24.5 22.3

Insulation

Perforated plastic box

Figure 6.2: The effect offrictional drag at the sediment-plastic interface during melt, shown here for a block containing sedimentary banding. 79

Block 6

This block was cut from a complex sequence of banding and folding in the right branch of the tunnel, 160 centimetres from the tunnel floor. It had a high sediment concentration (Table 6.5) and consisted predominantly of solid glaciotectonic structures within a matrix of Facies VII (opaque ice). The buckled layer seen at the base of the block in Figure 6.12 could be traced around the sides of the block, and was associated with an overlying clear layer of ice (Facies VI). Other solid structures included fine folds, as seen above the buckled layer, and lenses. The lens on the left edge of the block in Figure 6.12 may have been a fold hinge that has been fm1her defonned and compressed.

Table Block 6 dimensions and sedimentaty characteristics.

Block Dimensions: large face (cm) 265x265 small face (cm) 200x265 Average proticle size (Phi) 1.54 Size description Medium sand S01ting Well sorted Sediment concentration by weight (%) 39.6

The complexity of the stmctures exhibited by this sample, made the visual identification of any individual glaciotectonic stmcture in the till impossible (Figure 6.13). Both at the edge and in the interior of the frozen block of till, the sediment appeared homogenous. The smface of the till was very irregular, but this could not be con-elated with any structure observed in the basal ice sample. 80

6.2 TAYLOR GLACIER

The Taylor Glacier samples had similar sedimentary structure, and the resultant melt-out tills were indistinf:,TUishable. Figure 6.14 shows a typical laminated sequence of fine sand, that altemates between brown and green, and clear ice. The colour variation within the sediment made the identification of laminae in the till possible. These appeared to have buckled at the edge of the deposit (Figure 6.15), which, like the Suess Glacier samples described, could be due to frictional drag against the plastic box.

Laminae could be clearly distinguished within the frozen cross section. Most remained horizontal, but some draped over larger clasts in a similar manner to those in the melt-out trial (Figure 4.8). Individual laminae did not seem to consolidate a significant amount, suggesting that the sand laminae in the basal ice contained velY little interstitial ice. Larger clasts rotated at the edge of the till deposit, but those observed in the frozen cross-section remained preferentially orientated in the ice flow direction.

The till samples from the Taylor Glacier basal ice that were oven dried retained their original shape, but the colour variation was less pronounced and laminae were only faintly distinguishable within the deposit (Figure 6.16).

Table 6.6: Taylor Glacier block dimensions and sedimentmy characteristics.

Block 1 Block 2 Block 3 Block Dimensions: large face (cm) 21Ox290 250x200 240x230 small face (cm) 190x290 215x200 220x230 Size description Fine sand-silt Fine sand-silt Fine sand-silt Sediment concentration by weight (%) 48.1 54 56.3 81

Figure 6.3: Block 10 (true right). Yellow dot marks the hinge point, where plastic-coated wire was inserted prior to melt. Arrow shows the ice flow direction.

1 em Figure 6.4: Block 10 till, showing the corresponding hinge point. 82

Figure 6.5: Block 9 (true left). White dots mark the crests of two parasitic folds; the yellow dot marks the main hinge point; the arrow indicates the ice flow direction.

o 60 70 80 h'III,,, Ulillui""h Figure 6.6: Block 9 till, showing corresponding crests and hinge point. 83

Figure 6.7: Block 1. Arrow indicates the ice flow direction. Yellow dot marks the hinge point, where plastic-coated wire was inserted prior to melt.

I ern Figure 6.8 : Block 1 till. Hinge could be identified at the outer edge of the deposit only. 84

Figure 6.9: Block 7 (front). 85

Figure 6.10: Cavities in block 7 till, showing the effect of the sediment-perspex interface.

80 90 JOO 10 20 30 40 50 60 7C Figure 6.11: Block 8 till. 86

Figure 6.12: Block 6, (back). 1 em

Figure 6.13: Block 6 till. 87

Figure 6.14: Laminated sequence of basal ice from the Taylor Glacier. 88

Figure 6.15: Taylor Glacier, block 1 till. Ice flow was from right to left

Figure 6.16: Taylor Glacier, block 2 till (front), oven-dried. 89

CHAPTE 7

DISCUSSION ONCLUSIONS

7 .. 1 STUDY IMPLICATIONS & DISCUSSION

It is important to understand contemporaIY processes in order to make accurate interpretations and reconshuctions of fonnerly glaciated areas. Nevertheless, few attempts have been made to correlate contemporary sedimentological processes with their depositional products. This study takes an integrated approach to glacigenic deposition. It characterises and describes englacial shucture and compares this with its depositional product, till. In addition, it considers the effect that the depositional process has on the character of the till deposit. It is this 'process' link that has been neglected by glacial geologists, geomorphologists and glaciologists alike (Boulton 1987).

Glacigenic deposition was simulated in the laboratory so that a comparative aIlalysis could be made between englacial shucture in basal ice and any pres~rved in till deposits. The identification of englacial sUucture in laboratory-generated till provides conclusive evidence, in regard to the issue of preservation (Shaw 1980). It also supports, at a broad level, Shaw's (1977a) model of deposition in arid polar environments, which shows the preservation of highly attenuated glaciotectonic shuctures (Figure 3.1).

Shaw (1977a) atttibuted the preservation of englacial shucture in arid polar environments to sublimation. This process occurs passively, under confined conditions and leads to the preservation of quite delicate shuctures (Benn & Evans 1998). However, the fidelity (Shaw 1980) of Shaw's assumption was questioned (Dreimanis 1984; Karrow 1984; Kemmis & 90

Hallberg 1984; Eyles et al. 1986; Humphreys 1993; Humplu'eys & Fitzsimons 1996). This study found that, under restricted conditions, slow melt may also occur passively and preserve englacial stTucture. In arid polar environments sublimation and slow melt may act together and in close association. Both could retain englacial characteristics, and may grade into one another. As it would be difficult in many cases to distinguish between these till types, the term passive deposition was suggested, in Chapter 3.3, to avoid a genetic classification. The concept of passive deposition was illustrated ill two models that describe spatial and temporal interrelationships between slow melt and sublimation in arid polar environments.

Boulton (1976b, 182~183), in denying that stmcture could be preserved in till stated that, if basal ice is melted in a 'bucket with holes in the bottom, ... you will find the result, even if the ice looh very laminated, ... is massive till.' This description was confilIDed by a preliminary melt-out experiment conducted at the Suess Glacier. A block of basal ice was placed in a plastic container with drainage holes and allowed to melt. This basic experiment not simulate passive deposition but did fmID a basis to identify a set of conditions, in Chapter 3.4, by which englacial structure could be preserved. These conditions were applied, where possible, to the design of subsequent melt-out experiments (Chapter 4.2).

An initial requirement, as suggested by Boulton's 'bucket: was a supporting container that allowed free drainage so that melt-water could be removed efficiently from the till matrix during deposition. The use of adjustable plastic boxes, with drainage holes, in conjunction with a geotextile filter to prevent sediment escape prevented the accumulation of excess melt-water. The sediment particles seemed to retain their relative positions during deposition. Insulation prevented edge-melting and slumping, so that the sedimentaty stmcture was lowered with little lateral dislocation.

Conditions for the preservation of englacial structure that could not be controlled experimentally related to the characteristics of the basal ice samples. The ice must have a high concentration of debris that is well 91 dispersed, so that inter-particle contacts can supp011 englacial structure during deposition. The Suess and Taylor glacier samples had sufficient concentrations of dispersed sediment to meet this requirement. For example, in blocks containing folds from the Suess Glacier, the sediment concentration was sufficient to preserve the (consolidated) cross-sectional shape of the fold.

It was also stated that the sediment must display distinct lithological, textural or colour variations to make any preserved stmcture identifiable. This was not true for all Suess Glacier samples, as frictional drag against the supporting container during melt made the identification of some stlUcture possible at the edge of till deposits, regardless of the homogenous nature of the sediment. However, colour variations within the sediment, such as dark banding around the hinge zone in blocks 9 and 10 till, helped identify the shape of the folds in the sediment after the ice had melted. Also, colour and textural variation was the only means of identifying banding beyond the edge of till deposits from blocks 7 & 8, and laminae in the till deposits from the Taylor Glacier samples.

7.2 FUTURE RESEARCH

There js considerable scope for the development of techniques to describe the geomorphology of the basal zone. This would help to identify englacial structure preserved in till, and to distinguish it from other structure that forms from subglacial deformation, or from secondary redistribution (such as fluvial sorting). Some techniques to accurately describe the profiles of glaciotectonic structures, such as dip isogon analysis, could be adapted from structural geology. However, difficulty was encountered here in attempts to apply these techniques because some samples were cut obliquely from larger structures and folds observed at the edges of blocks may not have been true profiles. LaboratOlY mapping and measurement of these structures may have misrepresented the nature of the defonnation. 92

Plastic-coated wire, inselted at the hinge and crests points of folds prior to melt was useful in detelmining the approximate position of the hinge and crest in the till deposits. Altemative markers could be either colour-coated sediment, like that used in the melt-out trial described in Chapter 4.2, or synthetic particles with densities equivalent to the sediment in the ice sample. This could be inserted into holes drilled through the ice so that, although it may be disturbed at the edge, would be visible within cross-sections taken through frozen till deposits.

There is an obvious need to eventually apply the characteristics of laboratory-generated till to field investigations. The simulation of passive deposition could be developed fulther, with samples that display varied sedimentary and structural characteristics. In this way cIiteria for the identification of till types resulting from passive deposition could be developed. Comparisons could then be drawn between lab-generated till and true glacial till.

If these expeIiments were conducted on samples with wide-ranging sedimentary concentrations, then a threshold debris concentration for the preservation of englacial structure may be identified. This would indicate the debris concentration of former basal zones. Also, although a range of structures were analysed in these experiments, this was limited by the characteristics of the basal ice sampled. Other stmctural characteristics, in particular the preservation of proticle fabrics, were not investigated.

It would be interesting to simulate sublimation, perhaps through freeze-drying techniques, in order to compare the till propelties with those of melt-out till. As the frequency and OCCUlTence of sublimation till has been debated, this may help establish critelia for its recognition, and would also have useful applications to field investigations. 93

7.3 CONCLUSIONS

Two conclusions can be drawn from this study:

1. Englacial structure can be preserved in till deposits.

This was illustrated by a stmctural comparison between englacial structure, and lamination, folding and banding preserved in laboratOlY generated till deposits.

2. Passive melt-out may preserve englacial structure.

Sublimation is an inherently passive process and fOlmed the basis of a model of depostion in arid polar environments (Shaw 1977). The melt-out experiments conducted in this study showed that when provision is made for the removal of excess melt-water, debris can consolidate veliically with little lateral dislocation during melt. This may account, along with sublimation, for the preservation of englacial shucture in arid polar environments. 94 REFERENCE

Aber, J.S. 1982 Model for glaciotectonism. Bull. Geol. Soc. Denmark 30, 79-90.

Allen, C.R; Kamb, W.B.; Meir, M.F. & R.P. Sharp 1960 Structure of the lower Blue Glacier, Washington. J. Geol. 68(6),601-665.

Anderson, RI.; Hallet, B.; Walder, J.; & B.F. Aubry 1982 Observations in a cavity beneath Grinnell Glacier. Earth Smf Processes and Landforms 7, 63-70.

Bell, RA.I. 1966 A seismic reconnaissance in the McMurdo Sound regIOn, Antarctica. J. Glaciol. 6(44),209-221.

Benn, D.I. & J.A Evans 1988 Glaciers and Glaciation. John Wiley: New York.

Bennett, M.R. & N.F. Glasser 1996 Glacial geology: ice sheets and landforms. John Wiley: New York.

Bishop, B.C. 1957 Shear moraines in the Thule area, North-West Greenland. Snow, Ice and Permafrost Res. Est., Res. Rep. 17.

Bjorlykke, K. 1967 The Eocambrian "Reusch moraine" at Bigganjargga and the geology around Varangerfjord, Northern Norway. Norsk Geol. Unders. 18- 44.

Bouchard, M.A 1989 Subglacial landforms and deposits in central and northern Quebec, Canada, with emphasis on Rogen moraines. Sed. Geol. 62,293-308.

Bouchard, M.A; Cadieux, B. & F. Goutier 1984 L'Origine et les characteristiques des lithofacies du till dans Ie secteur nord du lac Albanel, Quebec: Une etude de-la dispersion glaciare clastiques. Chibougamoll - Stratigraphy and Mineralisation. Canadian Inst. Mining Sp. Pub. 34,244-260.

Boulton, G.S. 1968 Flow tills and related deposits on some Vestspitsbergen Glaciers. J. Glaciol. 7(51), 391-412.

1970a On the origin and transport of englacial debris in Svalbard Glaciers. J. Glaciol. 9(56),213-229.

1970b On the deposition of subglacial and melt-out tills at the margins of certain Svalbard Glaciers. J. Glacio!. 9(56), 231-245.

1971 Till genesis and fabric in Svalbard, Spitsbergen, in Goldthwait, RP. (ed.) Till: A Symposium. Ohio State University Press: Columbus, 41 - 72. 95

1972a The role of the thermal regime in glacial sedimentation, in Price, R.J. & D.E. Sugden (eds.) Polar Geom01phology. Inst. British Geographers. Sp. Pub. 4, 1-19.

1972b Modern arctic glaciers as depositional models for former ice sheets. J. Geol. Soc. London 128, 361-393.

1975 Processes and patterns of subglacial sedimentation: a theoretical approach, in Wright, A.E. & F. Moseley (eds.) Ice Ages Ancient and Modem. Seel House Press: Liverpool, 7-42.

1976a A genetic classification of tills and criteria for distinguishing tills of different origin, in Stankowski, W. (ed.) Till: its genesis and diagenesis. 10th International Union for Quaternary Research (INQUA) Congress: Poland.

1976b Criteria for the recognition of various types of till. Panel discussion in Stankowski, W. (ed.) Till: its genesis and diagenesis. 10th International Union for Quaternary Research (INQUA) Congress: Poland.

1987 Progress in glacial geology during the last fifty years. J. Glaciol. Sp. Issue, 25-32.

Boulton, G.S. & M. Deynoux 1981 Sedimentation in glacial environments and the identification of tills and tillites in ancient sedimentary sequences. Precambrian Res. 397 422.

Boulton, G.S. & M.A. Paul 1976 The influence of genetic processes on some geotechnical properties of glacial tills. Q.J. Eng. Geol. 9, 159-194.

Boulton, G.S. & U. Spring 1986 Isotopic fractionation at the base of polar and sub-polar glaciers. J. Glaciol. 32 (112), 475-485.

Bull, C. & C:R. Carnein 1970 The mass balance of a cold glacier: Meserve Glacier, South Victoria Land, Antactica Int. Assn. Sci. Hydrol. Pub. 86 (IGASE), 429-446.

Calkin, P .E. 1973 Glacial processes in the ice tree valleys of south Victoria Land, Antarctica, in Fahey, B.D. & R.D. Thompson (eds.) Res. Polar & Alpine Geomorph. Geo Abstracts, Norwich, 167-186.

Chamberlin, T.c. 1883 Terminal moraines of the second glacial epoch. U.S. Geol. Surv. Annu. Rep. 1881-1882,291-402.

Chinn, T.J.H. 1980 Glacier balances in the Dry Valleys area, Victoria Land, Antarctica. World Glacier Inventory - Inventaire mondial des Glagiers, IAHS - AISH Publication, 126,237 - 247. 96

1981 Hydroloy and climate in the Ross Sea Area. Jl. R. Soc. NZ. 11(4),373-386.

1986 Structure and equilibrium of Dry Valley Glaciers. NZ. Antarctic Record, 6 (Sp. Supp.), 73-88.

1987 Accelerated ablation at a glacier ice-cliff margin, Dry Valleys, Antarctica. Arctic Alp. Res. 19(1), 71-88.

1989 Single folds at the margins of dry-based glaciers as indicators of a glacial advance. Ann. Glaciol. 12,23-30.

1990 The Dry Valleys, Chapter 10 in Hatherton, T. (ed.) Antarctica: the Ross Sea region. D.S.I.R Publishing: Wellington.

1991 Polar glacier margins and debris features. Mem. Soc. Geol. It. 6, 25-44.

Chinn, T.J.H. & A. Dillon 1987 Accelerated ablation at a glacier ice-cliff margin, Dry Valleys, Antarctica. Arctic & Alp. Res. 19(1), 71-88.

Croll, J. 1870 The boulder clay of Caithness, a product ofland ice. Geol. Mag. 7, 209-214 and 271-287.'

Denton, G.H.; Prentice, M.L.; Kellog, D.E. & T.B. Kellog 1984 Late Tertiary history of the Antarctic Ice Sheet: evidence from the Dry Valleys. Geol. 263- 267.

Dilabio, R.N.W. & W.W. Shilts 1979 Composition and dispersal of debris by modern glaciers, Bylot Island, Canada, in Schluchter, Ch. (ed.) Moraines and Varves: origin, genesis, classification. Balkema: Rotterdam.

Donath, F.A. & RB. Parker 1964 Folds and folding. Bull. Geol. Soc. Am. 75, 45-62.

Dowdeswell, J.A. & M.J. Sharp 1986 Characterisation of pebble fabrics in modern terrestrial glacigenic sediments. Sedimentol. 33, 699-710.

Dreimanis, J.A. 1976 Tills: their origin and properties. In Legget, RF. (ed.) Glacial till: an interdisciplinary study. Royal Soc. Canada Sp. Pub. 12, 11-49.

1982 Work Gp. 1 INQUA Commission on Geneis and Lithology of Quaternary Deposits. In Ch. Schltichter (ed.) Report oj activities 1977 - 1982, 12-3 L ETH Zurich.

1983 Precontemoraneous partial disaggredation and/or resedimentation during the formation and deposition of subglacial till. Acta Geol. Hispanica 18, 153 - 160. 97

1984 Discussion - lithofacies types and vertical profile models, interpretation of glacial diamict and diamictite sequences. Sedimentol. 31, 885- 886.

1988 Tills: their genetic terminology and classification, in Goldthwait, RP. & c.L. Matsch Genetic classification of glacigenic deposits. Balkema: Rotterdam, 17-85.

Elliott, D. 1965 The Quantitative Mapping of Directional Minor Structures. J. Geol. 73, 865-880.

Ellis, 1.M. & P.E. Calkin 1983 Environments and soils of Holocene moraines and rock glaciers, central Brooks Range, Alaska, in Evenson, E.B, & Schluchter, Ch. & 1. Rabassa (eds,) Tills and related deposits. Balkema: Rotterdam, 315-328.

Evans, D.1.A 1989a Apron entrainment at the margins of subpolar glaciers, north west Ellesmere Island, Canadian High Arctic. J. Glaciol. 35(121), 317-324,

1989b The nature of glaciotectonic structures and sediments at subpolar glacier margins, north-west Ellesmere Island, Canada. Geogr. Ann. 71A(3-4), 13-123.

Eyles, N.; Eyles, C,R. & AD, Miall 1983 Lithofacies types and vertical profile models; an alternative approach to the description and interpretation of glacier diarnict and diamictite sequences. Sedimentol, 30, 393 - 410,

1986 Lithofacies types and vertical profile models; an alternative approach to the description and interpretation of glacier diarnict and diamictite sequences, reply Sedimentol. 33, 151-155.

Fitzsimons, SJ, 1990 Ice-marginal depositional processes in a polar maritime environment, Vestfold Hills, Antarctica, J. Glacio!. 36(124),279-286,

1996 Formation of thrust-block moraines at the margins of dry based glaciers, South-Victoria Land, Antarctica, Ann. GI. 22, 68-74.

Flint, RF. 1957 Glacial and Pleistocene Geology. Wiley: London.

1971 Glacial and Quaternary Geology. Wiley: New York.

Flint, RF,; Sanders, lE. & 1. Rodgers 1960 Diamictite: a substitute term for symmictite. Geol. Soc. Am. Bull. 71, 1809-1810.

Geikie, A 1863 On the phenomena of the glacial drift of Scotland, Geol. Soc. Glasgow Trans. 1(2), 190,

Gemmell, lC. 1985 The ong1l1, composition and behavior of basal Ice at Nigardsbreen, Norway, Unpublished PhD thesis, University ofAberdeen. 98

Gemmell, AM.D., Sharp, MJ. & D.E. Sugden 1986 Debris from the basal ice of the Agassiz ice cap, Ellesmere Island, Arctic Canada. Earth Surf Processes and Landforms 11, 123-130.

Goldthwait, R.P. 1951 Development of end-moraines In east-central Baffin Island. J. Geol. 59(6), 567-577.

1960 Study of ice cliff in Nunatarsuaq, Greenland. CRREL Tech. Report 39.

1971 Introduction to till today, in Goldthwait, R.P. (ed.) Till a symposium. Ohio State University Press: Oxford, 3 - 36.

Goodchild, J.G. 1875 The glacial phenomena of the Eden Valley and the western part ofthe Yorkshire-Dale district. Geol. Soc. Lond. Q.J. 31, 55-99.

Gow, Al, Epstein, S. & W. Sheehy 1979 On the origin of stratified debris in ice cores from the bottom of the Antarctic ice sheet. J. Glaciol. 23(89), 185-192.

Hamilton, W. & P. T. Haynes 1961 Structure of the lower Taylor Glacier, South Victoria Land, Antarctica. us.G.s. Prof Papers 424-C, 206-20,

Harland, W.E.; Herod, K.N. & D ,H. Krinsley 1966 The definition and identification of tills and tillites. Earth Sci. Rev. 2, 225-226.

Hartshorn, J.H. 1958 Flow till in south-east Massachusetts. Geol. Soc. Am. Bull. 69,477 - 482.

Herron, S. & C.C. Langway 1979 The debris-laden ice at the bottom of the Greenland ice sheet. J. Glaciol. 23 (89), 193-207.

Hickock, S.R. 1990 Genetic till prism. Geol. 18, 517-519.

Hills, E.S. 1959 Outlines of Structural Geology. Methuen & Co: London.

Holdsworth, G. 1969 Structural glaciology of the Meserve Glacier. Antarc. J. U.S. 4(4), 126-128.

1974 Meserve Glacier, Wright Valley, Antarctica: Part 1. Basal processes. Ohio State Unt., Inst. Polar Studies, Rep. 37.

Hollin, IT. & R.L. Cameron 1961 I.G.Y. glaciological work at Wilkes Station, Antarctica. J. Glaciol. 3(29), 833-842.

Hooke, R. Le B. 1968 Correspondence. Comments on, the formation of shear moraines: an example from South Victoria Land, Antarctica. J. Glaciol. 7(50), 351-352. 99

1970 Morpholoy of the Ice sheet margm near Thule, Greenland. 1. Glaciol. 9(94), 303-325.

1973a Flow near the margin of the Barnes ice cap, and development ofice-cored moraines. Geol. Soc. Am. Bull. 84(12),3929-3948.

1973b Structure and flow in the margin of the Barnes ice cap, Baffin Island, N.W.T., Canada. 1. Glaciol. 12(66),423-438.

Hooke, R. Le B. & P.l Huddleston 1978 Origin of foliation in glaciers. 1. Glaciol. 20(83),285-299.

HowelJ, J.V. 1957 (ed.) Glossary of Geoogy and Related Sciences. American Geological Institute: Washington.

Hubbard, B. & M. Sharp 1989 Basal ice formation and deformation: a review. Prog. Phys. Geogr. 13, 529-558.

1995 Basal ice facies and their formation m the Western Alps. Arctic Alp. Res. 27(4), 301- 310.

Huddleston, P.l 1976 Recumbent folding in the base of the Barnes Ice Cap, Baffin Island, Northwest Territories, Canada. Geol. Soc. Am. Bull., 84(12), 1648- 1692.

1977 Similar folds, recumbent folds and gravity tectonics in ice and rocks. 1. Geol. 85, 113-122.

Huddleston, P.l & R. Le B. Hooke 1980 Cumulative deformation in the Barnes Ice Cap and implications for the development of foliation. Tectonophys. 66, 127- 146.

Hubbard, B; Sharp, M. & W. Lawson 1996 On the sedimentological character of Alpine basal ice facies. Ann. Glaciol. 22, 187 - 193.

Humphreys, K.A. 1993 Sediment and landform associations of arid polar glaciers. Unpublished Msc. project, University of Otago.

Humphreys, K.A. & S.J. Fitzsimons 1996 Landform and sediment associations of dry-based glaciers. Z. Geomorphol. 105,21-33.

Hutter, K. & V.O.S Olonyo 1981 Basal stress concentrations due to abrupt changes in boundary conditions: a cause for high till concentration at the bottom of a glacier. Ann. Gl. 2,29 - 33.

Jouzel, l& R.A. Souchez 1982 Melting-refreezing at the glacier sole and the isotopic composition of the ice. 1. Glaciol. 28(98), 35-42.

THE LIBRARY UNIVERSITY OF CI',NTERBURY CHRISTCliJh;.,H. N.l 100

Kamb, B. & E. LaChappeIle 1964 Direct observation of the mechanism of glacier sliding over bedrock. J. Glaciol. 5(38), 159-172.

Karrow, P.F. 1984 Discussion: Lithofacies types and vertical profile models; an alternative approach to the description and environmental interpretation of glacial diamict and diamictite sequences. Sedimentol. 31, 883-894.

Kemmis, T.J. & G.R Hallberg 1984 Discussion: Lithofacies types and vertical profile models, interpretation of glacial diamict and diamictite sequences. Sedimentol. 31, 886-890.

Kirkbride, M.P. 1995 Processes of transportation. Chapter 8 in Menzies, J. Modern Glacial Environments: processes, dynamics and sediments. Glacial Environments vol. 1. Butterworth-Heinemann: Oxford.

Klassen, RA. 1982 Glaciotectonic thrust plates, Byot Island, District of Franklin, Canada, in Currect Research. Geol. Surv. Can. Paper, 82-1A, 369-373.

Knight, P.G. 1987 Observations at the edge of the Greenland ice sheet: boundary implications for modellers, in The physical basis of lee Sheet Modelling. Proc. of the Vancouver symposium. 1AHS Pub. 170.

1989 Stacking of basal debris layers without bulk freezing-on: isotopic evidence from west Greenland. J. Glaciol. 35(120), 214-216.

1993 Glaciers: Prog. Phys. Geogr. 17(3),349-353.

1994 Two-facies interpretation of the basal layer of the Greenland ice sheet contributes to a unified model of basal ice formation. Geol. 22, 971-974.

Lawson, D.E. 1979a Sedimentological analysis of the western terminus region of the Matanuska Glacier, Alaska. C.R.R.E.L. Rep. 17-9.

1979b A comparison of the pebble orientations in ice and deposits of the Matanuska Glacier, Alaska. J. Geol. 87,629-645.

1981a Distinguishing characteristics of diamictons at the margins of the Matanuska Glacier, Alaska. Ann. Glaciol. 2, 78-84.

1981 b Sedimentological characteristics and classification of depositional processes and deposits in the glacial environment. C.R.R.E.L. Rep., 81-27.

1989 Glacigenic resedimentation: classification concepts and application to mass movement processes and deposits, in Goldthwait, RP. & c.L. Matsch Genetic classification of glacigenic deposits. Balkema: Rotterdam, 147- 173. 101

Lawson, W. 1996 The relative strengths of debris-laden basal ice and clean glacier ice: some evidence from Taylor Glacier, Antarctica. Ann. Glaciol. 23,270-276.

Lawson, D.E. & lB. Kulla 1978 An oxygen isotope investigation of the origin of the basal zone ofMatanuska Glacier, Alaska. J. Geol. 86, 193-203.

Lliboutry, L. 1964-65 Traite de glaiologe. Masson et Cie:Paris (2 vols.).

1993 Internal melting and ice accretion at the bottom of temperate glaciers. J. Glaiol. 39( 131), 50-64.

Lundqvist, l 1988 Glacigenic processes, deposits and landforms, in Goldthwait, RP. & c.L. Matsch (eds.) Genetic classification of glacigenic deposits. Balkema: Rotterdam.

1989 Till and glacial landforms In a dry, polar regIon. Z. GeomOlph N.F. 33(1),27-41.

Menzies, l 1981 Freezing fronts and their possible influence upon the processes of subglacial erosion and deposition. Ann. Glaciol. 2, 52-56.

1995 Modern Glacial Environments: processes, dynamics and sediments. Butterworth-Heinemann: Oxford.

Minell, H 1977 Hogs-back moraines in Harrejokko, northern Sweden. Geol. Faren 1 Stockh Faer. 99-3(570), 264-270.

Moran, S.R.; Clayton, L.; Hooke, R.Le B.; Fenton, M.M. & L.D. Andriashek 1980 Glacier bed landforms of the prairie region of North America. J. Glaciol. 25(93), 457 - 426.

Morgenstern, N.R. & IF. Nixon 1971 One-dimensional consolidation of thawing solids. Can. Geotech. J. 8, 558-65.

Muller, E.H. 1983 Till genesis and the glacier sole, in Evenson, E.B.; Schluchter, Ch. & 1. Rabassa (eds.) Tills and related deposits. Balkema: Rotterdam, 315-328.

Nobles, L. 1960 Investigations of structures and movement of the steep ice ramp near Red Rock Lake, Nunatarssuaq, Greenland. USSIPRE, report on contract DA-11-190-ENG-12.

Patterson, W.S.B. 1981 The Physics of Glaciers (2nd ed.). Pergamon Press: Oxford.

Paul, M.A. & N. Eyles 1990 Constraints on the preservation of diamict facies (melt-out) tills at the margins of stagnant glaciers. Quat. Sci. Rev. 9, 51-69.

Price, N.1. & J.W. Cosgrove 1997 Analysis of Geological Structures. Cambridge University Press. 102

Ragan, D.M. 1969 Structures at the base of an ice fall. J Geol. 77(6),647-667.

Rains, R.B. & l Shaw 1981 Some mechanisms of controlled moraine development, Antarctica. J Glaciol. 27(95), 113-128.

Ramsay, lG. 1967 Folding and Fracturing of Rocks. McGraw Hill: U.S.A.

Reading, HK. (ed.) 1979 Sedimentary environments and facies. Blackwell Scientific Publications: London.

Reading, H.G. & R.G. Walker 1966 Sedimentation in Eocambrian tillites and associated sediments in Finnmark, northern Norway. Palaeogeogr. Paleoclimatol. Paleocol. 2, 177-212.

Robin, G. de Q. 1976 Is the basal ice ofa temperate glacier at the pressure melting point? J Glacial. 16(74), 183-196.

Robinson, P.H 1984 Ice dynamics and the thermal regime of the Taylor Glacier, South Victoria Land, Antarctica. J Glaciol. 30(105), 153-160.

Ronnert, L & D.M Mickelson 1992 High porosity of basal till at Burroughs Glacier, southeastern Alaska. Geol. 20, 849 - 852.

Selby, M.l 1972 The terminus and moraines of glaciers in the McMurdo Dry Valleys, Antarctica. Proc.7th NZ. Geog. Conj, Hamilton 1972,247-257.

Shaw, l 1977a Till deposited in arid polar environments. Can. J Earth Sci. 14, 1239-1245.

1977b Till morphology and structure related to glacier flow. Boreas 6, 189-201.

1979 Genesis of Sveg tills and Rogen moraines of central Sweden: a model of basal melt-out. Boreas 8,409-426.

1980 Application of present-day glacial processes to the interpretation of ancient tills, in Stankowski, W. (ed.) Tills and Glacigene Deposits. UAM Geografta 20,51-55.

1982 Melt-out till in the Edmonton area, Alberta, Canada. Can. J Earth Sci. 19, 1548-1569.

1985 Subglacial and ice-marginal environments, chapter 2 in Ashley, G.M.; Shaw, J. & N.D. Smith (eds.) Glacial sedimentary environments. SEPM Short Course 16. Soc. of Paleontologists and Mineralogists: Oklahoma. 103

1986 Discussion: Lithofacies types and vertical profile models; an alternative approach to the description and interpretation of glacial diamict an diamictite sequences. Sedimentol. 33,151-155.

1988 Sublimation till, in Goldthwaith, RP. & Matsch, C.L. (eds.) Genetic classification ofglacigenic deposits. Balkema: Rotterdam.

Souchez, R.A. 1967 The formation of shear moraines, an example from South Victoria Land, Antarctica. J. Glaciol. 6, 837-844.

1968 Correspondence. Reply to Dr R LeB. Hooke's comments on, the formation of shear moraines: an example from South Victoria Land, Antarctica. J. Glacial. 7(50),352-353.

1971 Ice-cored moraines In south-western Ellesmere Island, N.W.I., Canada. J. Glacial. 10(59),245-254.

Souchez, RA. & J.M. De Groote 1985 aD - 0180 relationships in ice formed by subglacial freezing: paleoclimatic implications. J. Glaiol. 31( 109), 229-232.

Souchez, RA. & I. Iouzel 1984 On the isotopic composition in aD and of water and ice during freezing. J. Glacial. 30(106),369-372.

Souchez, RA. & RD. Lorrain 1978 Origin of the basal ice layer from Alpine glaciers as indicated by its chemistry. J. Glacial. 20(83), 319-328.

Souchez, RA., Lorrain, R, Tison, J.L., & J. Iouzel 1988 Co-isotopic signature of two mechanisms of basal ice formation in Arctic outlet glaciers. Ann. Glacial. 10, 163-166.

Sugden, D.E., Knight, P.G., Livesey, N., Lorrain R.D. Souchez, R.A., Tison, J.-L. & J. Iouzel 1987 Evidence for two zones of debris entrainment beneath the Greenland ice sheet. Nature 328 (16), 238-241.

Swinzow, G.K 1962 Investigations of shear zones in the ice sheet margin, Thule area, Greenland. J. Glacial. 4(32), 215-229.

Tison, l-L., Petit, J.-R., Barnola, I.-M., & W.e. Mahney 1993 Debris entrainment at the ice-bedrock interface in sub-freezing temperature conditions (Terre Ad6lie, Antarctica). J. Glacial. 33(113),47-50.

Ward, W.H. 1952 The glaciological studies of the Baffin Island expedition, 1950 Part IT: the physics of deglaciation in central Baffin Island. J. Glacial. 2(11) 9-23.

Weertman,1. 1957 On the sliding of glaciers. J. Glacial. 3(21),33-38.

1961 Mechanism for the formation of inner moraines found near the edge of cold ice caps and ice sheets. J. Glaciol. 30, 965-78. 104

1964 The theory of glacier sliding. J. Glaciol. 5(39), 287-303.

Wheeler, R.H. 1961 The ice-free areas of western McMurdo Sound. Proc. 3rd NZ. Geog. Conf. N.Z. Geographical Soc., 106-115. van der Wateren, P.F.M. 1985 A model of glacial tectonics applied to the ice pushed ridges in the central Netherlands. Bull. Geol. Soc. Denmark 34,55-74.

1995 Processes of Glaciotectonism. Chapter 10 in Menzies, J. 1995 Modern Glacial Environments: processes, dynamics and sediments. Butterworth-Heinemann: Oxford. 105

APPENDIX 1:

SEDIMENT CONTENT BY \VEIGHT - SAMPLE BLOCKS

SAMPLE ORIGINAL FINAL DRY SEDIMENT CONTENT WEIGHT (g) WEIGHT (g) BY WEIGHT (%)

Sl 571.3 125.2 21.9 S2A 472.6 14.15 3 S2B 670.2 227 33.9 S3 310.1 57,7 18,6 S4 367.] 65.2 19,9 S5 390.66 68 17.4 S6 697.2 276.1 39.6 S7 434.4 106.4 24.5 S8 465.8 103.9 22.3 S9 711.1 234.2 32.9 S10 537.8 109.8 20.4 Sl1 543.8 i 56.6 10.4 S12 331.9 243.7 73.4 S13 646.7 238.3 36.8 S14 497.1 164 33 T1 932,5 643 48.1 T2 765,2 413.2 54 T3 831.6 468.2 56.3 APPENDIX 2:

SEDIMENT SIZE, SORTING AND '-'JIlJL:.IU SAMPLE BLOCKS

SAMPLE I FACIES CODE MEAN VERBAL CLASSIFICATION (Part 1.3.4) (phi)

Sl IV (banding & folding), matrix of V 1.67 Medium sand, well sorted S2A III (indistinct lamination), matrix of VII 1.89 Medium sand, well sorted S2B (fold) 1 Medium sand, well sorted S3 IV & IVa (banding), matrix of V 0.57 Coarse sand, well sorted & IVa (banding), matrix of VII 1.57 Medium sand, moderately well sorted S5 IV & IVa (banding), matrix of V 1.77 Medium sand, well sorted S6 IV (folding, lenses, bands) & III (dispersed sediment) 1 Medium sand, well sorted S7 IV (banding and coarse foliation), matrix of VII 0.49 Coarse sand, well sorted S8 IV (banding and coarse foliation), matrix VII 0.51 Coarse sand, well sorted S9 II (fold) 1.73 Medium sand, well sorted S II (fold) 1.69 Medium sand, well sorted Sl1 II & III 1 Medium sand, moderately well sorted S12 IV 1.03 Coarse sand, moderately well sorted S13 II & III 1.18 Medium sand, moderately well sorted S14 II& 1.45 Medium sand, moderately well sorted T1, 2 & 3 lamination fme sand - silt