Quaternary Science Reviews 18 (1999) 1213}1246 The landform and sediment assemblage produced by a tidewater glacier surge in Kongsfjorden, Svalbard Matthew R. Bennett!, Michael J. Hambrey", David Huddart#, Neil F. Glasser$, Kevin Crawford% !School of Earth and Environmental Sciences, University of Greenwich, Medway University Campus, Chatham Maritime, Kent, ME4 4TB, UK "Centre for Glaciology, Institute of Geography and Earth Sciences, University of Wales, Aberystwyth, Ceredigion, SY23 3DB, UK #School of Education and Community Studies, Liverpool John Moores University, I. M. Marsh Campus, Barkhill Road, Liverpool, L17 6BD, UK $School of Biological and Earth Sciences, Liverpool John Moores University, Byrom Street, Liverpool, L33AF, UK %Department of Environmental and Biological Studies, Liverpool Hope University College, Hope Park, Liverpool, L16 9JD, UK Abstract This paper describes the landform and sediment assemblage produced by a surge (in 1948) of the Kongsvegen/Kronebreen tidewater glacier complex in northwest Spitsbergen. The main geomorphological products of this advance are two large thrustmoraine complexes on opposite sides of the fjord, and a system of geometrical ridges revealed on glacier decay. The thrust-moraines are composed largely of diamicton, sandy and muddy gravel, gravelly sand, sand and mud, with minor laminites. All of these appear to be derived from the fjord #oor and represent both "ne fjord basin sediments and coarse grounding-line fan deposits. Thrusting was the principal mode of emplacement of the sediment onto the adjacent land areas during the 1948 advance. However, the geomorphology of the thrust-moraine complexes on either side of the fjord is quite di!erent, re#ecting a transpressive regime on the southwest side (mainly long ridges) and a normal compressive regime on the northeast side (short ridges and pinnacles of a &hummocky' nature). The advance which produced the moraine complex has previously been attributed to a surge of Kongsvegen, but the glaciological and geomorphological evidence suggests that the advance involved both Kongsvegen and Kronebreen. Comparison of the landform assemblage produced by this event with that produced by other tidewater glacier surges demonstrates the diverse range of landform assemblages associated with glacier surges, or other episodes of rapid #ow, within glaciomarine environments.( 1999 Elsevier Science Ltd. All rights reserved 1. Introduction The landform/sediment assemblages associated with rapidly advancing tidewater glaciers are poorly understood, yet are of considerable signi"cance for our understanding of the geomorphology and sedimentology of low-lying coastal and continental shelf areas. The need for modern analogues with which to interpret the Pleistocene record is well illustrated by the recent debate over the landform/sediment assemblage along the margins of the Irish Sea (e.g. Eyles and McCabe, 1989, 1991; McCar- *Corresponding author. E-mail: m.r.bennett@greenwich.ac.uk. roll and Harris, 1992; Huddart, 1993; Huddart and Clark, 1993; McCarroll, 1995). In recent years there has been a proliferation of literature describing the processes and landforms associated with glaciomarine sedimentation (e.g. Dowdeswell, 1987; Eyles et al., 1985; Powell and Molnia, 1989; Dowdeswell and Scourse, 1990; Powell and Domack, 1995). However, little attention has been paid to the landforms and sediments which result from the rapid advance of tidewater glaciers either due to surges, or simply as a result of the non-climatic icemarginal #uctuations common to tidewater glaciers (Warren, 1992). Solheim (1991) provides an exception in describing the landforms and sediments produced by a surge of the tidewater glacier Bra> svellbreen in Svalbard (Fig. 1A). He 0142-9612/99/$ - see front matter ( 1999 Elsevier Science Ltd. All rights reserved. PII: 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 0 4 1 - 9 1214 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 described a subaquatic landform assemblage consisting of an outer push-moraine inside which there is a rhomboidal ridge network (crevasse-"ll ridges; Solheim and P"rman, 1985; Solheim, 1991). In many respects, this is very similar to the broad landform model proposed by Sharp (1988b) for surge-type glaciers. More recently, Boulton et al. (1996) described the landform assemblage produced by a tidewater glacier surging onto the island of Coraholmen in Svalbard (Fig. 1A). They emphasised the importance of subglacial deformation beneath the advancing glacier in reworking glaciomarine muds to form deformation till and identi"ed a network of crevasse-"lled ridges formed behind moraines or &till tongues' produced by the extrusion of deformation till in front of the glacier. This type of work illustrates the potential of landbased data in assisting with the interpretation of o!shore seismic records and in developing landform models with which to interpret the Pleistocene record. In this paper we review evidence for an alternative landform/sediment assemblage to that described by Boulton et al. (1996), in this case produced by a surge of the Kongsvegen/Kronebreen tidewater glacier complex at the head of Kongsfjorden, Svalbard, in 1948 (Fig. 1A). This work adds to the range of landform models avail- Fig. 1. (Continued) able with which to interpret the landform and sedimentary facies of low-lying coastal areas. 2. The glacial history of Kongsfjorden Fig. 1. Location maps for Kongsfjorden and the Kongsvegen/ Kronebreen tidewater glacier complex. [ In (A) 1"Engelskbukta (Uve( rsbreen and Comfortlessbreen); 2" Coraholmen, Ekmanfjorden and Sefstr+mbreen; 3" Bra> svellbreen] At the head of Kongsfjorden there are two tidewater glaciers, Kongsvegen and Kronebreen, which collectively drain an area of 1013 km2*part of the Holtedahlfonna ice-"eld. Today Kronebreen is split by both Colletth+gda and Ossian Sarsfjellet to form three tidewater margins M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 referred to as Kronebreen North, Central and South (Fig. 1B). The Kongsvegen/Kronebreen glacier complex is sometimes referred to as Kongsbreen, a term used when the glacier complex extended well into the main fjord (Hagen, pers. com. 1997). At present Kronebreen South and Kongsvegen ice front is #owing and retreating at a varying rate; Kongsvegen is almost stationary and receding slowly relative to Kronebreen South, which has a maximum ice velocity of 785 m a~1, an annual calving rate of over 0.25 km3 and a rate of recession exceeding 50 m a~1 (Lefauconnier et al., 1994). Three advances of this combined tidewater glacier complex are recorded in historical documents, the earliest being some time around 1800 (Liest+l, 1988). Two subsequent advances, one in 1869 and the other in 1948, are better constrained both in space and time. The ice front at the height of the 1869 advance is shown in a contemporary "eld sketch reproduced by Lamont (1876), while the extent of the 1948 advance can be "xed using aerial photographs taken in 1936 and 1948. The recession of the glacier complex following the 1948 advance is shown in aerial photographs from 1970, 1977, 1990, and 1995 (Fig. 1C). Additional evidence is available in the form of oblique photographs taken during "eld investigations in the 1960s (Wilhelm, 1961; Voigt, 1965; Meier, 1969). These glacier #uctuations have traditionally been interpreted as the product of glacier surges, with both Kongsvegen and Kronebreen being identi"ed as surge-type glaciers (Liest+l, 1988; Hagen et al., 1993). There is, however, no direct evidence for this conclusion within the Kronebreen basin; looped medial moraines (Meir and Post, 1969), elevated trimlines, indications of quiescent phase #ow, such as well-developed supraglacial stream networks and potholes (Strum, 1987), and a reduction in crevasse density (Lawson et al., 1994) were not observed. Kronebreen has all the attributes of a fast #owing tidewater*as opposed to a surge-type*glacier. In contrast, Kongsvegen does have attributes indicative of a surge-type glacier, such as looped medial moraines, supraglacial drainage networks including potholes, and is currently experiencing a period of quiescence (Glasser et al., 1998). Both the 1869 and 1948 advances are associated with moraine assemblages on the southwest side of Kongsfjorden and on the southern tip of Ossian Sarsfjellet. The 1948 complex is, however, better developed and linked to the contemporary ice margin (Fig. 1C), and therefore forms the focus of this paper. 3. Methodology The geomorphological components were mapped in the "eld using enlarged 1 : 15,000 aerial photographs (1990 and 1995; Norsk Polarinstitutt). The cross-sectional morphology of prominent landforms was surveyed 1215 using an Abney level, while detailed plans were constructed using a Leica TC600 Total Station. Thrusts within the moraine complexes were recorded by measuring the dip and direction of dip of well-de"ned rectilinear thrust surfaces on individual ridges/mounds and within a limited number of sections. The results are plotted on Schmidt-equal area stereographic projections. The glacier structures within Kongsvegen were recorded from aerial photographs, and detailed site plans were prepared using a network of cairns (Glasser et al. 1998). Interpretations were also made using the 1948, 1970, and 1995 aerial photographs. Sedimentary facies were described using a combination of grain size, fabric and clast shape analysis. Sediments were classi"ed on the basis of the Hambrey (1994) modi"cation of the Moncrie! (1989) classi"cation for poorly-sorted sediments. Particle size was determined using a combination of sieving techniques (sand and gravel fractions) and a SediGraph 5000 D Particle Size Analyzer (silt and clay fractions). Three-dimensional fabric data were collected for samples of 50 prolate clasts and are plotted on Schmidt-equal area stereographic projections. Clast shape was analysed for samples of 50 clasts following the method of Benn and Ballantyne (1994), which has been found to give good discrimination of glacial facies in Arctic environments (Bennett et al., 1997). All shape samples were of mixed lithology, unless otherwise stated. 4. Glaciology of the Kongsvegen/Kronebreen complex The structural evolution of the Kongsvegen /Kronebreen glacier complex was examined using both historical aerial photographs and contemporary observations at the margin of Kongsvegen. 4.1. Aerial photograph analysis Vertical aerial photographs exist from 1948, 1970, 1977, 1990 and 1995, and are complemented by a range of oblique aerial photographs from 1936 and 1956, although complete coverage of the entire glacier basin is limited to the later sorties ('1977). Oblique aerial photographs from 1936 show the Kongsvegen/Kronebreen ice margin located at the southern tip of Ossian Sarsfjellet, running straight across the fjord (Fig. 1C). The "rst vertical aerial photographs of the Kongsvegen/Kronebreen glacier complex were taken in 1948 during or just after the surge (Liest+l, 1988). Unfortunately the glacier complex is shown on just one of the photographs, at the end of a east-west sortie along Kongsfjorden. Consequently the Kongsvegen/Kronebreen South glacier con#uence is not visible and the relative dominance of these two glaciers during the 1948 event cannot be identi"ed. However, later observations made by Voigt (1965) 1216 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 suggests that both #ow units were of equal importance. The 1948 aerial photographs show a heavily crevassed ice-lobe extending into the fjord (Fig. 1C). No glacier structures are visible on the glacier surface because of the intensity of crevassing, although the density of crevasse is no greater than that on Kronebreen South today. On the southwest side of Kongsfjorden, the ice margin surmounts a deeply dissected drift slope, separated from the valley side by a river system. The whole of the moraine complex on the southwest side of the fjord was covered by active glacier ice in 1948 except for this narrow (50 m wide), steep (c. 333) and dissected outer drift slope. On Ossian Sarsfjellet, the glacier margin in 1948 was located along a prominent meltwater system running parallel to the modern shoreline. Signi"cantly, not all the moraine complex was ice-covered and part of this complex may therefore have been proglacial. Oblique aerial photographs in 1956 show a reduced glacier lobe, but one still in contact with the coastline of Ossian Sarsfjellet. By 1970 Ossian Sarsfjellet was ice-free, although on the southwest side of the fjord ice had not retreated very far (Fig. 1C). A heavily crevassed, calving glacier in the centre of the fjord is separated from a crevasse-free, grounded and partly debris-covered area to the southwest. These two #ow units are separated by a transpressive fault system. This consists of at least 3 major longitudinal fractures between which there are a series of sigmodial tension fractures showing evidence of extension. A series of prominent moraine ridges can be seen emerging from these fractures close to the ice margin. By 1977 the ice margin was very similar to that of today (Fig. 1C), with an active calving ice cli! in front of Kronebreen South and a grounded ice cli! in front of the crevasse-free ice of Kongsvegen. On the 1995 air photographs there is evidence of a small fan/delta emerging from beneath the glacier in front of the southwest side of the Kronebreen South ice margin, which could be observed in the "eld at low tide in 1996, indicating that water depths are very shallow at this point. 4.2. Structural glaciology of Kongsvegen Today the margin of Kongsvegen is deeply embayed. Recession of the main calving front (Kronebreen South) has occurred more rapidly than the grounded ice on the southwest side of the ice front (Kongsvegen). As a consequence an ice cli! oriented approximately parallel to the direction of #ow occurs behind the beach on the southwest side of the fjord (Fig. 1C). The glacier structures within this cli! and on the surface of Kongsvegen have been described in detail elsewhere (Bennett et al., 1996a; Glasser et al., 1998), and are only summarised brie#y here. Four types of structure can be recognised in the terminal area of Kongsvegen. These structures are termed S }S in order of formation according to normal conven0 3 tions of structural geology. The main structures are: (1) strati"cation (S ); (2) longitudinal foliation (S ); (3) 0 1 thrusts (S ); and (4) crevasse traces (S ). The earliest 2 3 structure is primary strati"cation (S ) inherited from 0 snowfall in the accumulation area. This structure is not readily visible on the glacier surface in the "eld, but is evident on aerial photographs where it shows up in the middle and upper reaches of the glacier as a di!use pattern of light and dark ice. It can also sometimes be seen in the ice cli! as a series of sub-horizontal and folded ice layers. The second structure is longitudinal foliation (S ). As on other glaciers, this is visible on the glacier 1 surface as intercalated layers of coarse-bubbly and coarse-clear ice (Allen et al., 1960). The foliation has a consistent strike (060}2403) sub-parallel to glacier #ow. The foliation is not linear but has a low-amplitude sinuosity. The structure dips steeply at angles of between 703 and 853. The foliation is o!set on the glacier surface by a series of planar fractures (S ), often associated with 2 regelation ice and basal debris layers, which have orientations that are either transverse or diagonal to the direction of glacier #ow. In the ice cli! these fractures are seen to o!set the strati"cation and are interpreted as listric thrusts. They rise from the glacier bed at between 303 and 503 and terminate either just below the glacier surface or on it. These thrusts are o!set by crevasse traces (S ) which consist of sharp fractures picked out by vari3 ation in the bubble content of the ice. The crevasse traces have an average strike of 175}3553, dips which vary from 503 to 903, and concave down-glacier outcrop traces. This type of relationship is typical of surge-type glaciers (Lawson et al., 1994), since a compressive wave and associated thrust formation precedes the surge-front, which is itself associated with extension, and therefore crevasse formation. Three of the four structures described above are associated with debris. The greatest volume is found within the thrusts (Fig. 2A), which contain layers of basal debris up to 3 m thick at the base of some thrusts. Most of the debris-rich thrusts do not crop out on the glacier surface. Angular supraglacial debris occurs within the folded strati"cation and can be seen emerging along the fold traces. On the glacier surface the supraglacial debris cover rarely exceeds a thickness of one clast and forms distinct longitudinal bands along the foliation. Dyke-like structures of basally derived sediment occur parallel to the foliation (Fig. 2B and C). These are formed by the incorporation of basal sediment into the foliation during its formation which is then elevated by continued folding and revealed on the glacier surface by surface lowering in the terminal zone (Glasser et al., 1998). 4.3. Summary The aerial photograph evidence does not resolve conclusively the question of which of the two glaciers was M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1217 Fig. 2. Debris-rich structures on the Kongsvegen ice margin. (A) Two debris-rich thrusts exposed within the Kongsvegen ice cli!. Ice #ow is from left to right. (B, C) Foliation parallel debris ridges formed by folding of subglacial debris parallel to the foliation during lateral compression of the ice tongue (Glasser et al., 1998). The boundinaged nature of the debris dyke is visible in (B). dominant during the 1948 advance. According to Hagen et al. (1993) the dominant glacier in 1948 was Kongsvegen in contrast to the earlier 1869 event which he assigns to a surge of Kronebreen. The evidence for this statement is not, however, presented. Kongsvegen has some of the characteristics of a surge-type glacier and is currently experiencing a quiescent phase with a #ow velocity near zero (Lefauconnier et al., 1994), a phenomenon common to surge-type glaciers (Sharp, 1988a; Lawson et al., 1994). The presence of debris-rich thrusts, of 1218 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 a similar angle to those typical of surge-type glaciers (Bennett et al., 1996a; Glasser et al., 1998) in the margin of Kongsvegen today may also be suggestive. Observations made by Voigt (1965) suggest that both #ow units contributed equally to the 1948 advance. It is, however, unlikely that both glaciers surged simultaneously. More importantly, given that Kronebreen is still active, has not undergone a post-surge quiescent phase, and has none of the features associated with surge-type glaciers in its basin, it seems unlikely that it is a surge-type glacier (cf. Hagen et al., 1993). It is possible, however, that a surge of Kongsvegen could have precipitated a simultaneous advance of both Kronebreen South and Kongsvegen. This idea is outlined schematically in Fig. 3, and develops the ideas of Mercer (1961). Mercer emphasised the importance of fjord geometry and the size of the calving glacier front on the location of the ice margin within a fjord. He assumed that all mass was lost via calving and that the rate of calving was a function of the width of the calving front. In a fjord with a uniform width, a small decrease in the equilibrium-line altitude would cause a tidewater glacier to Fig. 3. Schematic model of the interaction of a fast-#owing tidewater glacier with a surge-type glacier. This model is an extension of the ideas outlined by Mercer (1961) and is explained in detail within the text. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 advance down the fjord until it widens allowing the width of the calving front to increase (Fig. 3A). If we apply this type of hypothetical model to Kongsfjorden two points become apparent. Firstly, the glacier con#uence between Kronebreen Central and Kronebreen South (Fig. 1C) is a point of unstable behaviour, since glacier con#uence would reduce the size of the joint calving front and could therefore precipitate an advance (Fig. 3B). Secondly, a surge of Kongsvegen would cause a change in the relative size of the calving front of both Kongsvegen and Kronebreen South and could cause both glaciers to advance. Consider the two hypothetical examples C and D in Fig. 3, which show two con#uent tidewater glaciers. The left-hand glacier is a surge-type tidewater glacier, while the right is a fast-#owing tidewater glacier. Each example is taken in turn below. 1. Example C. Initially the calving front of both glaciers is in equilibrium with the #ux of ice to the margin and therefore remains stationary. A surge of the left-hand glacier causes a readjustment of the calving front; an increase for the surge-type glacier and a decrease for the right-hand glacier. If the width of the calving ice cli! in front of the surge-type glacier is in equilibrium with the ice #ux during the surge the glacier will remain stationary. However, the right-hand glacier is no longer in equilibrium and must advance to increase the width of its calving front (Fig. 3C). Consequently, a surge in one glacier causes an advance in another. 2. Example D. Here the #ux of ice during the glacier surge is not in equilibrium with the width of the new ice cli! and both glaciers advance down the fjord until the calving front of both glaciers widen su$ciently to restore equilibrium. If one applies this type of model to Kongsfjorden, then a surge of Kongsvegen could precipitate an advance of Kronebreen South. Critical in this type of model is the speed at which the surge-wave passes through Kongsvegen relative to the response time of Kronebreen South. If the passage of the surge wave is rapid then Kongsvegen would simply advance into the fjord and pinch o! the snout of Kronebreen South. In practice, however, glacier surges often occur relatively slowly and may last several years in Svalbard (Dowdeswell et al., 1991; Hamilton and Dowdeswell, 1996; Hambrey et al., 1996). Equally, Kronebreen South has a very high velocity, with a maximum of 785 m a~1 (Lefauconnier et al., 1994), and would have a rapid reaction time. The essential point here is that we do not know the cause of the 1948 advance, simply that both glaciers were involved and that Kronebreen South does not have the characteristics of a surge-type glacier, although Kongsvegen does. The type of model outlined above, however, illustrates how the interaction of a surge-type glacier with a fast-#owing tidewater glacier can result in a range of complex glacier #uctuations such as those 1219 observed in Kongsfjorden and is a model which needs further exploration. 5. The landform and sediment assemblage The 1948 advance is recorded on land on the southwest side of Kongsfjorden and on the southern tip of Ossian Sarsfjellet (Fig. 1C). The assemblage of landforms and their associated sediments is shown in Fig. 4 for the southwest side of Kongsfjorden and in Fig. 5 for Ossian Sarsfjellet. On the southwest side of Kongsfjorden an ice-cored lateral moraine extends northwards from the current ice margin (Fig. 4) and is separated from the valley side by a small #uvial system running the length of the moraine. This lateral moraine is actively back-wasting towards the southwest as a result of debris #owage and ablation of the buried ice, which is revealed in the #ow-scars and channels (Fig. 6). Between the lateral moraine and the fjord shore there is a series of low ridges forming geometrical ridge networks (Fig. 4; Bennett et al., 1996a). At the northern end of this lateral moraine there is an in#exion towards the east in the landform assemblage and the lateral moraine is replaced by a large moraine complex consisting of an imbricate stack of wedge-shaped sediment sheets composed of sandy gravel and diamicton. These sediments appear to be derived from the fjord #oor, as indicated by the presence of foraminifera, shell fragments and salt e%orescence. This moraine complex is being periodically reworked during wet weather by debris #ows. On Ossian Sarsfjellet the landform assemblage consists of an extensive moraine complex composed of diamictons, sands, gravels and muds which form a series of stacked mounds and pinnacles (Fig. 5). This moraine complex is bisected by a prominent meltwater system which starts as a meltwater channel and kame terrace system in the east, and ends in the west in a series of rock-cut meltwater channels. Diamicton is draped and &spilled' over a prominent north}south escarpment at the eastern side of Ossian Sarsfjellet (Fig. 5). In the following section, we describe and interpret the range of sedimentary facies present within this landform complex, and then focus on the geomorphology of the principal landform elements, which are the two moraine complexes and the geometrical ridge networks. 6. Sedimentary facies 6.1. Description Eight broad facies can be recognised both on Ossian Sarsfjellet and on the southwest side of the fjord. 1. Diamicton-1. This diamicton is characterised by a high proportion of subangular and subrounded 1220 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1221 Fig. 5. Landform and sediment facies on Ossian Sarsfjellet. (A) Facies distribution map, also shown in the location of surveyed pro"les in Fig. 18. (B) Particle size for matrix samples of some of the facies in (A). b Fig. 4. Landform and sediment facies on the southwest side of Kongsfjorden. (A) Facies distribution map, also shown in the location of surveyed pro"les in Figs. 6 and 9. (B) Particle size for matrix samples of some of the facies in (A). (C) Shaped data for key diamicton facies. [VA"very angular and WR"well rounded are the end member of the roundness scale: very angular, angular, subangular, subrounded, rounded and well rounded] clasts and is generally massive with a wide range of clast sizes. Gravel content is variable, but typically between 20 and 40%, although a signi"cant proportion of the diamictons recorded on Ossian Sarsfjellet were clast-poor ((5%). Clasts are typically subangular or subrounded, and "ne-grained sedimentary lithologies are commonly striated. Typical 1222 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 6. Two pro"les across the lateral moraine on the southwest side of Kongsfjorden, showing the relationship between the moraine and the rectilinear ridge networks. See Fig. 4 for pro"le locations. shapes are shown in Figs. 7 and 8. The matrix varies in its silt content, but is usually quite sand-rich; the clay content is small (Figs. 4B and 5B). Foraminifera, shell fragments and salt e%orescence occur in some of the diamictons, particularly the "negrained clast-poor examples. 2. Diamicton-2. In contrast this diamicton contains a high proportion of angular and very angular clasts (Figs. 7 and 8). It normally consists of a massive sandy diamicton, with a high proportion of angular and very angular clasts, although there is usually a wide range of shapes (Fig. 7). It can be distinguished on the basis of roundness, but not particle form from diamicton-1 (Fig. 4C). Only a few clasts are striated. It occurs extensively on the icecored lateral moraine on the southwest side of c Fig. 7. Clast shape data for the principal facies within the moraine complex on the southwest side of Kongsfjorden. The clast shape matrix consists of four uni-lithology samples taken from the same location within each facies. The in#uence of clast lithology within a single environment can be seen by reading the matrix horizontally, while the contrast between di!erent facies is obtained by reading the matrix vertically. [VA"very angular and WR"well rounded are the end member of the roundness scale: very angular, angular, subangular, subrounded, rounded and well rounded] M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1223 1224 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 8. Co-variant plots of the RA index (percentage of Very Angular and Angular clasts) versus the C index (percentage of clasts with c : a axial ratio 40 of )0.4) for all the clast shape data obtained from the moraine complex on the southwest side of Kongsfjorden and from Ossian Sarsfjellet. The subglacial, supraglacial and glacio#uvial facies envelopes are obtained from Bennett et al. (1997). Kongsfjorden and is being actively reworked and sorted by debris-#ow activity. 3. Sandy gravel and gravelly sand facies-1. This facies consists of interbedded sandy gravel, sand and silt units either in subhorizontal sheets or in large channels. Matrix supported gravels are also present in some of the sequences. Foraminifera and shell fragments occur abundantly within the matrix. This gravel facies only occurs in the moraine complexes and can be distinguished on the basis of clast roundness from diamicton-1, which is more angular. This clast shape distinction is most marked for soft limestone clasts (Fig. 7). These gravels appear to have a higher silt content on Ossian Sarsfjellet than that recorded within the moraine complex on the southwest side of fjord (Figs. 4B and 5B). 4. Sandy gravel and gravelly sand facies-2. This is a diverse facies ranging from openwork pebble gravel to cobble gravels. Channels and tabular cross sets are often evident, as are isolated units of laminated or rippled coarse sand. Shell fragments and foraminifera are absent. It is found within the glacio#uvial systems of the moraine system and occasionally within moraine-mounds. It can be distinguished from the "rst gravel facies on the basis of clast roundness (Figs. 7 and 8). 5. Muddy gravel. This facies is characterised by moderately well-sorted pebble/cobble gravel, with clasts coated with mud. Clast shapes are similar to those of the "rst sandy gravel subfacies and are usually more rounded than those of diamicton-1 and the second sandy gravel subfacies. This facies is usually associated with the sandy gravel facies. 6. Sand. A range of moderately well-sorted sand units are present. These may be laminated or rippled, but are more commonly massive. 7. Mud. Sandy mud with occasional pebble gravel clasts, shell fragments and foraminifera occur within the moraine complexes, particularly on Ossian Sarsfjellet. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Table 1 The abundance of di!erent facies in the thrust moraines of Ossian Sarsfjellet and on the southwestern side of Kongsfjorden Facies Southwest side Kongsfjorden Ossian Sarsfjellet Diamicton-1 Sandy gravel/gravelly sand Muddy gravel Sand Mud Sand-mud laminites ***** **** * * * * **** *** * *** *** ** 4. 5. * occasional ***** abundant 8. Sand-mud laminites. Rhythmically laminated mud and sand, occasionally with isolated gravel clasts, occur within both the moraine complexes, although they are more evident on Ossian Sarsfjellet. Thicker sand-mud layers are also present in minor amounts. The frequency with which these eight facies are encountered within the moraine-complex of Ossian Sarsfjellet and that of the southwest side of Kongsfjorden is indicated in Table 1. The moraine-mounds of Ossian Sarsfjellet contain a much "ner-grained facies assemblage than the moraine complex on the southwest side of Kongsfjorden which is dominated by diamictons and sandy gravels. Supraglacial debris (e.g. diamicton-2) does not form a signi"cant part of the moraine complex and is only recorded as a thin veneer on some moraine-mounds. 6.2. Interpretation Foraminifera, shell fragments and salt e%orescence are common to many of the facies described above suggesting that much of the sediment is derived originally from the fjord #oor. Each facies is interpreted below. 1. Diamicton}1. These are interpreted in general as basal diamictons on the basis of their clast shape (Figs. 7 and 8) and the presence of abundant striations. The clast poor examples, rich in foraminifera and shell fragments, appear to be reworked or derived from the fjord #oor where they may have been originally deposited by iceberg dumping, accounting for the basal a$nity in their clast shapes. All the diamictons have been sheared to varying amounts and are perhaps best described as deformation till. 2. Diamicton}2. This diamicton is interpreted as a combination of supraglacial debris and basal debris. The basal debris is elevated to the glacier surface along debris-rich thrusts within the glacier ice and by foliation-parallel folding (Glasser et al., 1998). 3. Sandy gravel and gravelly sand facies-1. This facies is interpreted as the product of sedimentation within an 6. 7. 8. 1225 ice-proximal glaciomarine grounding-line fan. The presence of broad channels, shell fragments, foraminifera and matrix supported gravels, indicative of subaquatic deposition in an environment rich in suspended sediment, all support this interpretation (Powell, 1981, 1984, 1990). Sandy gravel and gravelly sand facies}2. This is found within the glacio#uvial systems of both moraine complexes and is typical of sandur environments in general (Maizels, 1995). Muddy gravel. This occurs in association with the sandy gravel facies. The presence of a silt matrix coating each clast is indicative of rapid sedimentation within an environment in which suspension settling is occurring. Such conditions are common within some types of grounding-line fans (Powell, 1990; Powell and Domack, 1995). Sand. These sediments are of variable origin and interpretation is usually dependant on the facies association. Laminated examples can be seen forming today in shallow silting ponds in front of dewatering debris #ows, with which they are normally interbedded. Massive sand units can also be seen within part of the active glacio#uvial systems. In addition massive sand units are sometimes a feature of glaciomarine environments (Hambrey, 1994). Mud. Due to the presence of foraminifera and shell fragments, this facies is interpreted as the product of suspension settling within a glaciomarine environment. The presence of laminated sequences with occasional isolated lonestones, interpreted as dropstones, supports this interpretation. Sand-mud laminites. These are probably tidal rhythmites and are common features of glaciomarine sequences (Mackiewicz et al., 1984; Cowan and Powell, 1990). In summary, therefore, the sedimentary facies of the moraine complexes studied re#ect for the most part either "ne fjord basin sediments or coarse grounding-line fan deposits, and give some insight into the sedimentology of the fjord #oor (cf. Elverh+i et al., 1980, 1983; Boulton, 1990). 7. Moraine complexes The moraine complexes on the southwest side of the Kongsfjorden and on Ossian Sarsfjellet are very di!erent in size and character and are therefore dealt with separately. 7.1. Southwest side of Kongsfjorden The morphology and internal geometry of this moraine complex is shown in the transect in Fig. 9. The 1226 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 9. Pro"le across the moraine complex on the southwest side of Kongsfjorden. See Fig. 4 for its location. The matrix grain size "eld is de"ned by a minimum of two bulk samples. [VA"very angular and WR"well rounded are the end member of the roundness scale: very angular, angular, subangular, subrounded, rounded and well rounded] internal structure is based on exposures and more importantly the 3-dimensional outcrop geometry of the individual facies recorded along the transect. No evidence was found for large-scale folding in the sediments, although associated glacier ice was strongly folded. The moraine complex consists of a series of stacked sediment rafts, separated by listric thrust faults. Each slab of sediment is bounded by thrusts and forms a distinct ridge often composed of a di!erent facies from adjacent ridges. The sediment slabs are composed of in situ sediment which has not been subjected to resedimentation. The upglacier face of each ridge has a rectilinear form de"ned by a thrust surface, and dips between 303 and 403. Individual ridges or sediment slabs can be traced along the strike of the moraine complex for distances ranging between 5 and 550 m (Fig. 10A}C). The moraine complex can be divided into three sections separated by two areas of debris #ow activity (Fig. 9), working from the southeast or fjord side they are: 1. Section 1 (0}200 m; Fig. 9) consists of a series of sediment slabs/ridges composed of sandy and mud- rich diamictons. These ridges are separated by an area of debris #ows from the second section. 2. Section 2 (300}600 m; Fig. 9) consists of a system of sharp-crested gravel ridges which form the largest and most prominent part of the moraine complex (Figs. 10 and 11). Each ridge is between 20 and 50 m high and can exceed 800 m in length. At its most complex it consists of 6 sub-parallel sediment slabs each forming a ridge (Fig. 10A). The upglacier face of each ridge or sediment slab forms a rectilinear slope, with dips varying from 303 to 543. This system of gravel ridges contains slabs of coarsely crystalline, strongly folded glacier ice (Fig. 10D), interleaved between gravel units and separated from them by thrusts which dip upglacier at between 323 and 543. The gravel ridge system is dominated by sands, sandy gravels and muddy gravels, which were probably entrained or thrust up from a grounding-line fan within the fjord. These prominent gravel ridges are separated by an area of debris #ows (Fig. 10B) from the third section. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1227 Fig. 10. Gravel thrust-moraines on the southwest side of Kongsfjorden. (A) The main ridge system consists of a series of stacked or imbricate slabs of gravelly sand, sandy gravel and muddy gravel. Ice #ow was from right to left. (B) This shows the relationship of the gravel ridges to seasonally active debris #ows which are in"lling the area to the left of the gravel ridges. Ice #ow was from right to left. (C) Individual gravel ridges are continuous for over 500 m and have a "sh plate-like crest. Ice #ow was from left to right. (D) Slabs of glacier ice occur within the gravel ridge complex as shown in this photograph. Note the person for scale. 3. Section 3 (750}950 m; Fig. 9) consists of a series of irregular ridges which show evidence of signi"cant sediment-#ow both on the southwest, or ice-distal face, of the complex and back towards the gravel ridges (Fig. 9). However, in situ sediment does occur revealing sequences of diamicton, laminates of mudsand, and sand units. The ice-distal face of this moraine complex is the only part of it visible on the 1948 aerial photographs (Fig. 4A). The rest of the complex can be seen melting out of the glacier surface on the 1970 aerial photographs; glacier ice is visible between the three main ridge systems which form the moraine complex. Each of the three ridge systems is orientated along a prominent strike-slip fracture sub-parallel to the direction of #ow. The moraine complex is interpreted as a thrust complex formed both by proglacial and englacial thrusting, incorporating glaciomarine sediment and transporting it onshore (Fig. 12) in a fashion similar to the moraines described in front of Uve( rsbreen by Hambrey and Hud- dart (1995) and at Comfortlessbreen by Huddart and Hambrey (1996), both of which are located in Engelskbukta, the next fjord to the south of Kongsfjorden (Fig. 1A). In this type of moraine complex thrusts propagate both within the glacier lobe and proglacially rising from a common deH collement surface, or sole thrust, beneath the glacier. Debris-rich basal ice or slabs of subglacial sediment are entrained along thrusts within the glacier lobe, thickened by folding and bed-parallel thrusting, and then melt out to form a moraine complex consisting of imbricate sediment slabs each forming a ridge. Glacier ice may be incorporated into the moraine complex between individual thrust slabs (Fig. 9), although this need not occur. Sediment within the upper part of a thrust is draped over the glacier fore"eld as the ice melts. However, it is the sediment at the base of the thrust which forms the mound/ridge, the upglacier face of which is de"ned by the thrust surface. Consequently thrust-moraine mounds need not necessarily incorporate buried ice, since it is only the basal part of the thrust which forms the mound or ridge. The morphology of the resultant 1228 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 11. The gravel ridges system within the moraine complex of the southwest side of Kongsfjorden. (A) Map of the principal ridge crests and thrusts, with their dip and orientation of dip. Ice #ow from right to left. (B) A series of serial pro"les across the ridge system. All the ridges are composed of sandy gravel, sands, muddy gravel or glacier ice. mound or ridge re#ects the geometry of the thrust and its inclination. The melt out of debris from thrusts, to form ice-free moraine-mounds, can be seen taking place at the margins of many of the glaciers in the Kongsfjorden area (Bennett et al., 1996b; Hambrey et al., 1997), including Kongsvegen where the process can be observed within the terminal ice cli! (Bennett et al., 1996a). Due to the location of the ice front in 1948, close to the outside edge of the moraine complex, it would appear that the moraine system formed predominately by thrusting within the glacier lobe (Figs. 4A and 12). On the 1970 aerial photographs the three main ridge systems of the moraine complex can be seen melting out of the glacier, each separated by ice, and occupying a prominent longitudinal strike-slip fracture. This coupled with the orientation of the moraine complex (sub-parallel to the fjord axis, the inferred #ow direction, and therefore the direction of maximum compression) suggests that the thrust system was transpressive, involving both strike-slip as well as normal compression. The linear nature of the moraine complex, dominated by well-de"ned thrust ridges, which contrasts with those formed by this process elsewhere (Hambrey and Huddart, 1995; Huddart and Hambrey, 1996), probably re#ects this transpressive regime. The geometry of the thrust-slabs in relation to the former fjord #oor is indicated by the preservation of primary sedimentary structures within the main gravel ridge system, and were exposed in a series of sections excavated in the crest of this system (Fig. 13). At the southeast end of the gravel ridge system, a channel sequence was exposed in a section cut into the downglacier face and running parallel to the ridge crest. The long axis of the channel dips out of the face at 403 while the transverse axis is parallel to the trend of the ridge crest. Only the right-hand side of the channel is visible (Fig. 13A). It is "lled by a series of conformable gravel and sand units, the most distinctive (Unit A, Fig. 13), being composed of open work gravels coated with silty clay indicative of rapid sedimentation in some form of subaquatic #ow. This is consistent with the interpretation of the gravel ridge system being derived from a groundingline fan. The channel indicates that the sea #oor is discordant (i.e. perpendicular) to the plane of thrusting. A second exposure on the ridge crest shows units of pebble/granule gravel and sand dipping conformably M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1229 Fig. 12. Model for the formation of the moraine system on the southwest side of Kongsfjorden involving thrusting both within the glacier ice and proglacially. Fjord bottom sediments is thereby transferred onshore. with the upglacier rectilinear face of the ridge (Fig. 13B). A thrust-fault associated with a #ame structure of "ne sand also occurs within this sequence. A similar situation is also indicated in Fig. 13C. In both these cases, the upglacier face of the ridge appears to be parallel to the former fjord #oor. A more common occurrence, however, is subhorizontal sand and gravel units in which the bedding is truncated by both the upglacier and downglacier face of the ridge (Fig. 13D). These subhorizontal units often show a wave-like undulation along the length of the ridge crest, suggesting either lateral compression or primary deposition in a series of domes. The range of geometrical relationships between the primary bedding and the thrust surface is consistent with that typically found within imbricate or duplex thrust systems (Butler, 1983), formed within a bedded sequence in which there is a contrast in competence between each layer. The type of geometries observed in the ridge complex (Fig. 13) are shown schematically in Fig. 14. The ridge morphology we see today is eroded/degraded from this type of complex. Clast fabrics were obtained from all the diamicton thrust blocks, both on the southwest side of Kongsfjorden and on Ossian Sarsfjellet. A total of 34 fabrics were obtained and their eigenvalues are plotted in Fig. 15 1230 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 13. The sedimentology with the crests of the gravel ridges showing the relationship of primary sedimentary structure to the ridge or thrust geometry. The particle size of prominent units is also shown Phi (+) units. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1231 Fig. 14. Illustration of the type of imbricate or duplex thrust system which could explain the range of tectonic geometries indicated by the ridge crest sediments shown in Fig. 13. (following: Dowdeswell et al. 1985; Dowdeswell and Sharp, 1986; Hart, 1994). Two distinct groups of fabric are evident on this plot (Group A and C) with a few stray points between (Group B), examples of each fabric type are also shown in Fig. 15. Two of the fabric clusters correspond to process "elds identi"ed by Dowdeswell et al. (1985) and Dowdeswell and Sharp (1986); Group C occurs in a similar location to lodgement tills, while Group A clusters around glacigenic debris #ows (Fig. 15). The Group A fabrics may also be consistent with some deformation tills (Benn, 1994). The strong lodgement till-like fabric within some of the diamictons (Group C) could either be the result of sediment deformation or re#ects a pre-deformation subglacial fabric. Since much of the diamicton is derived from the fjord #oor where strong fabrics are normally absent (Domack and Lawson, 1985), a tectonic origin is suggested for the fabric. The strong fabric may be indicative of either, a thin deforming layer and moderately high strain rates (Hart, 1994), or alternatively brittle, or brittle-ductile, strain within the thrust slab (Benn and Evens, 1996). Group A fabrics have eigenvalues which are either indicative of re-sedimentation (Dowdeswell and Sharp, 1986) or subglacial deformation (Benn, 1994, 1995). However, due to small scale fabric variability the former interpretation appears more likely. They also occur in locations likely to have experienced some re-sedimentation. This is well illustrated by a coastal section through diamicton at the northeastern end of the transect shown in Fig. 9. Here there are two superimposed diamicton units, distinguishable only on the basis of a slight variation in gravel content and colour. The lower diamicton has a Group C fabric, while the upper diamicton has a Group A fabric. This suggests that the surface of the thrust mound has undergone some degree of sediment #owage, probably post-depositionally. In general, this is a pattern which repeats itself with shallow, near-surface, fabrics plotting within Group A, while deeper, undisturbed, fabrics are typical of Group C. 7.2. Ossian Sarsfjellet The moraine complex of Ossian Sarsfjellet is composed of numerous individual mounds/ridges with a more chaotic disposition than that found in the moraine complex on the southwest side of Kongsfjorden (Fig. 16A). However, the individual mounds and ridges have the same key characteristics as those on the southwest side of the fjord as well as those typical of both englacial and proglacial thrust complexes described elsewhere by the authors (Hambrey and Huddart, 1995; Bennett et al., 1996b; Huddart and Hambrey, 1996; Hambrey et al., 1997), namely: (1) the mounds have wellde"ned upglacier rectilinear faces; (2) each mound is composed of undisturbed basally derived sediment; (3) individual mounds and ridges are separated from one another by thrusts with adjacent mounds and ridges commonly being composed of di!erent facies; and (4) the mounds/ridges have a stacked or imbricate form. The Ossian Sarsfjellet moraine complex is therefore also interpreted as the product of glacial thrusting both proglacially and within the glacier lobe itself, although the range of morphological products contrast with those on the southwest side of the fjord. The sedimentary facies within the moraine complex are "ner-grained than on the southwest side of the fjord and are dominated by glaciomarine sands, muds and diamictons. The moraine complex on Ossian Sarsfjellet can be divided into two morphological facies separated by the main #uvial system which runs parallel to the coast and bisects the moraine system (Fig. 5). The area to the north of the #uvial system is composed of large, linear, stacked moraine ridges, while the area to the south of the #uvial system is more chaotic in form, being composed of 1232 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 15. Eigenvalue plot for fabrics taken from diamictons within the moraine complexes of both Ossian Sarsfjellet and the southwest side of Kongsfjorden. The data is grouped into two main areas (Group A and C) with a few stray points (Group B). Typical fabrics for each cluster of data points is also shown using Schmidt-equal area stereographic projections. Contours are at 1, 3, 5, 7% per 1% area. The process "elds shown in the inset are based on Dowdeswell et al. (1985). numerous small mounds, ridges and pinnacles producing a more &&hummocky'' appearance (Fig. 16A). On the 1948 aerial photographs the ice margin is located along the meltwater system (Fig. 5A). Both areas of moraine com- plex, the subglacial south and proglacial north, contain similar sedimentary facies, vegetation and soil development, and are consequently believed to both date from 1948. The thrust orientations within the two parts of the M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 16. The Ossian Sarsfjellet moraine complex. (A) The complex consists of numerous small ridges/mounds of varying geometry. However, all have pronounced upglacier rectilinear or thrust surfaces and have an imbricate form. This is illustrated by the morainemound in the foreground. (B) Moraine-mound pinnacle within the Ossian Sarsfjellet moraine complex, composed of glaciomarine diamicton. The pinnacle is approximately 5 m high. moraine system are however di!erent, in the proglacial system thrusts dip towards the southeast, while in the subglacial part thrusts dip more commonly towards the southwest (Fig. 17). The moraine complex as a whole does not appear to contain much buried ice, since the complex is dry and there is no evidence of contemporary debris-#ow activity. The morphological di!erences between the two moraine systems*proglacial and englacial*are summarised in Table 2. The proglacial moraine system consists of an imbricate stack of ridges, typically 20 m high and 150 m long. Each 1233 ridge has a well-developed rectilinear face or thrust surface which dips upglacier at between 203 and 403 (Fig. 18A and B). The inclination of these thrust surfaces appears to decrease towards the front face (i.e. the deformation front) of the moraine complex. Thrusting occurred against a gently rising bedrock slope. The front face of this proglacial thrust system cross-cuts earlier thrust mounds and a prominent moraine ridge formed during the 1869 advance, from which it can be distinguished on the basis of vegetation, degree of degradation and soil development (Fig. 17). In the rest of the moraine complex, to the south of the #uvial system (Figs. 5 and 17), the individual morainemounds are morphologically more diverse (Fig. 16). Four morphological components can be identi"ed within this part of the moraine complex: (1) a dissected drift sheet, up to 5 m thick, with an undulating surface; (2) individual moraine-mounds with a convex-upward crestline trace, usually 5 to 10 m high and 10 to 15 m long (Fig. 16A); (3) linear ridges usually between 10 and 20 m high and up to 100 m long composed of sandy gravels; and (4) pinnacles, between 1 and 5 m high, composed of diamicton which, although degrading rapidly, have not been formed by erosion (Fig. 16B). Mound or ridge morphologies are not facies-speci"c, although ridges are more common in sandy gravels and pinnacles are restricted to diamictons and muddy sands. The range of mound and ridge morphologies present re#ects the geometry of the thrusts from which they were formed within the former ice margin (Bennett et al., 1998). There are two main variables to be considered: "rstly, the angle of the thrusts and secondly the geometry of its strike trace. The steeper a thrust the more prominent a ridge is likely to be once the ice has melted away, since more sediment is left in the mound at the base of the thrust and less is draped over the glacier fore"eld as the ice melts. This is shown schematically in Fig. 19A, and is based on observations within the ice cli! of Kongsvegen (Bennett et al., 1996a) where tall prominent ridges and mounds can be seen melting out of steep thrusts. However, there is a critical angle of approximately 453, above which the rectilinear slope is not maintained after the ice melts, because of collapse. The second variable is the type of thrust. The outcrop trace of a thrust on glacier surfaces within Spitsbergen tend to vary from linear to arcuate or lobate forms (Hambrey et al., 1996). Short arcuate or lobate thrusts will result in mounds with a crestline-trace which is convex in the direction of thrusting; the resultant mound is highest where the fault angle was highest at the thrust apex (Fig. 19B). Thrusts with a more linear strike trace will result in ridges. The pinnacles are less easily explained and there are two possible models. Firstly, a spire may form in the apex of a tight lobate thrust. The thrust angle is greatest at the thrust apex and consequently any sediment within it would tend to melt out to form a steep sided mound at this point. This would be an 1234 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 17. Structural map of both the 1948 and 1869 thrust moraine complexes on Ossian Sarsfjellet. Prominent rectilinear (thrust) surfaces are shown in the map along with Schmidt-equal area stereographic projections of the dip and direction of dip of these surfaces. The 1948 ice and deformation limits are also shown. extreme example of that illustrated in Fig. 19B. Alternatively, a pinnacle may form at the intersection of two thrusts, sediment being elevated higher at the point of intersection (Fig. 19C). Interference patterns between lobate thrusts can be seen in their outcrop trace on contemporary ice surfaces (Hambrey et al., 1996). Both mechanisms could generate the diamicton pinnacles observed, although the second model seems more plausible since it would be di$cult to produce a lobate thrust which was su$ciently tight to produce a pinnacle as opposed to simply a conically shaped mound. A general model for the formation of the moraine complex on Ossian Sarsfjellet is shown in Fig. 20. Advancing ice generated a large proglacial thrust complex on which a series of moraine-mounds were formed by debris melting out of thrusts within the glacier lobe itself. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Table 2 Morphological and structural di!erences between the proglacial and englacial thrust-moraine systems on Ossian Sarsfjellet. Proglacial thrust-moraines Englacial thrust-moraines Angle of thrusting varies between Angle of thrusting varies between 153 and 303, with a strike trace of 253 and 403, with a strike trace of 1203 to 1803 0803 to 2603 Linear thrusts, relatively uniform Linear and lobate thrusts of varyin size and geometry ing size and geometry Imbricate thrust ridges, typically Ridges, mounds and pinnacles, involving 10}40 m thick sediment typically involving 5}15 m thick sediment slabs. May not always slabs appear imbricate and individual ridges/mounds may be widely spaced Cross-cuts earlier moraine ridges No cross-cutting relationships, all moraines of similar age Supraglacial drape absent Supraglacial material may drape individual moraine mounds/ridges or forms garlands around the base No buried ice Buried ice may occur locally, greater evidence of re-sedimentation by debris #ow and slumping Again this involves the entrainment and onshore transfer of glaciomarine sediment. The range of sedimentary facies within the Ossian Sarsfjellet moraine complex is more diverse than that on the southwest side of the fjord and there is a higher proportion of mud and sand (Table 1). The "ne-grained facies are heavily fractured and deformed. Two types of deformation facies were observed: (1) a meH lange of sand, silt and clay; and (2) faults, thrusts and boudins. The most common deformation facies, occurring in all parts of the moraine Ossian Sarsfjellet is a meH lange of sand, silt and clay (Fig. 21). The sediments are all "ne grained, contain isolated clasts and are partially mixed. Irregular pods of sand, with di!erent particle sizes and degrees of sorting occur, with di!use and irregular boundaries (Fig. 21D}F). Occasionally irregular fold hinges can be identi"ed (Fig. 21F), although this is the exception rather than the norm. The "ne-grained muds are typically heavily fractured and break up into blocks of a centimetre in size (Fig. 21C and E). Fractures are often open, show evidence of shear and may be lined with "ne sand. The degree of mixing and deformation is variable, ranging from areas in which deformed sand layers can be followed as irregular stringers to areas with a chaotic mixture of sand and mud, in which irregular pods ((1}5 mm in size) of di!erent grain sizes occur in close juxtaposition. The original sediment appears to have consisted of mud-sand laminites and thicker mud-sand layers. The deformation of this sequence does not show strong primary shear, but is better described as a crude 1235 mixing of the sediment. We suggest that this facies represents three phases of deformation. First, soft sediment deformation caused by loading of saturated sediment by ice occurred. An element of hydrofracturing, caused by impeded sediment drainage on loading, may be responsible for some of the sand-"lled fractures and the chaotic mixing of the sediment layers (Clark, 1949; Hubbart and Willis, 1957; Boulton and Caban, 1995; Boulton et al., 1996). The second phase of deformation involved lateral displacement of the sediment to form thrust ridges or mounds and is responsible for the development of shear/tension fractures within the "ne-grained muds. The "nal phase of deformation appears to have been extensional and is represented by the open fractures and probably resulted from the withdrawal of ice support around the thrust ridge or mound. The second deformation facies was observed in an excavation through the crest of a short ridge, transverse to the strike of the rectilinear face (Fig. 22). It is characterised by multiple units and complex deformation structures that indicate deformation events prior to and after thrusting. The facies includes layers of mud, silt, "nemedium sand and sandy gravel. Undisrupted primary bedding shows signs of soft-sediment deformation in the form of load structures, but most bedding is attenuated and boundinaged in the "ne sand and silt, or has been subjected to small-scale normal faulting in medium sand (Fig. 21A and B). These structures suggest bed-parallel extension, which can only have come from the loading of soft sediment by ice. The second phase of deformation is that which has the most pronounced morphological expression, namely thrusting. Although bedding remains essentially parallel to thrusting, it is truncated by the morphologically constrained over- and under-lying thrusts and small-scale reverse thrusts as well as chevron and recumbent folding. These structures represent the main compressive phase in the sediment. The last deformation event was the formation of near-vertical fractures some of which were open to widths of 1 cm. These may have formed once the supporting ice had melted away. These deformation structures have implications for the mechanism of sediment entrainment along thrust which form within glacier ice. Traditional explanations for sediment entrainment along thrusts involve the &freezing-on' of basal sediment which is then elevated with basal ice as thrust displacement occurs (Boulton, 1967, 1972). For this type of model to be consistent with the observed deformation structures it would require ice to "rst advance over unfrozen sediment, before basal freezing occurs. In addition, thrust propagation involves ice displacement and therefore warm sliding ice. For &freezingon' to occur a thermal transition within the glacier prior to or associated with the phase of compressive #ow causing thrust displacement must occur. The alternative is that the sediment is entrained without basal freezing. The simplest mechanism for this would be sediment #ow 1236 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1237 Fig. 19. Schematic model of the relationship of moraine-mound morphology to the geometry of the thrusts within a glacier. (A) Shows the relationship of mound height and de"nition to the angle of thrusting; steep thrusts result in high well de"ned moraine-mounds. (B) Moraine-mound morphology is also controlled by thrust geometry; ridges result from linear thrusts, while lobate thrusts result in arcuate mounds. (C) One hypothesis for the formation of moraine-mound pinnacles involving interference between two thrusts. along a thrust, perhaps assisting in its initial propagation. However, this is inconsistent with the preservation of primary sedimentary structures at some locations (Fig. 13). The evidence presented here suggests that unfrozen slices of basal sediment are incorporated into the thrust. The degree of sediment deformation during this process is probably a function of grain size and therefore of the porewater pressures generated. Preservation of primary bedding, within thrust slabs (Fig. 13) is most marked within free draining sands and gravels, while deformation appears greatest in "ne grained mud-sand sequences in which high porewater pressure can be generated. A full discussion of the entrainment process within glacial thrusts must await a greater understanding of the rheological behaviour of this type of basal sediment and of the ice-bed interface. 8. Geometrical ridge network This landform assemblage occurs inside the main thrust-moraine complex on the southwest side of the fjord between the shore and the ice-cored lateral moraine. It has been described previously by Bennett et al. (1996a) and is therefore only brie#y summarised here. The ridge network consists of two intersecting ridge components which form a rectilinear pattern when viewed from above (Fig. 23). The two components are: (1) transverse ridges; and (2) longitudinal ridges. Transverse ridges are typically straight, sharp-crested, and asymmetrical in cross-section, with an upglacier face of between 103 and 203 with a steeper downglacier face ('203) showing evidence of slumping. The ridges are generally of uniform height (4}8 m), although some have conical b Fig. 18. Two pro"les across the 1948 thrust moraine complex on Ossian Sarsfjellet the locations of the two pro"les are shown in Fig. 5. The matrix grain size "eld is de"ned by a minimum of two bulk samples. [VA"very angular and WR"well rounded are the end member of the roundness scale: very angular, angular, subangular, subrounded, rounded and well rounded] 1238 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 20. Model for the formation of the Ossian Sarsfjellet thrust-moraine complex. highs, associated with intersections with longitudinal ridges. The strike orientation of the transverse ridges is variable, but most are orientated perpendicular to the fjord axis and therefore the direction of glacier #ow. The ridges are composed of homogeneous, well-consolidated, sandy diamicton in which subrounded particles dominate. Striated, pebble to boulder-sized clasts are common and large #at-iron or bullet-shaped clasts, typical of lodgement till are also present (Sharp, 1982; KruK ger, 1984). Clast fabrics taken from the upglacier face plot in Group C, while those from the slumped downglacier face plot in Group A (Fig. 15). In contrast, the longitudinal ridges are more varied. They are typically low ((1 m), poorly de"ned, symmet- rical, beaded and slightly sinuous in morphology. They vary in length from 10 to 60 m. In all cases, the longitudinal ridges are cross-cut by the transverse ones. Sedimentologically the longitudinal ridges are also more varied ranging from sandy diamicton to a gravel-rich sand. The sandy diamicton has the same characteristics as that in the transverse ridges, while the gravel-rich sand has similar clast shapes but is simply a better sorted sediment. The ridge network can be seen melting out of Kongsvegen ice margin today and each element traced to a distinct glacial structure (Bennett et al., 1996a; Glasser et al., 1998). The transverse ridges are emerging from the base of thrusts and form in the same manner as the thrust moraine described above (Figs. 2A and 19A). The c Fig. 21. Deformation structures within the moraine mounds of Ossian Sarsfjellet. (A) Boudins within "ne sands and silts. (B) Soft sediment deformation structures associated with loading and dewatering. The photograph is approximately 30 cm wide. (C) The muds are heavily fractured and break up into cube shaped blocks (c. 1 cm). These fractures may contain sand, show evidence of shear and are often open to several millimetres. (D}F). A meH lange of sand, silt and mud with occasional lonestones. Irregular fold hinges are some times evident (F) but for the most part the components simply appear to be mixed with little sense of the direction of deformation. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1239 1240 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Fig. 22. Field sketch of the deformation structures observed in a trench dug perpendicular to the mound crest and parallel to the direction of thrusting on Ossian Sarsfjellet with the 1948 ice limit. Three phases of deformation structure can be identi"ed. Some of the boudins are illustrated in Fig. 21A. longitudinal ridges are foliation-parallel and represent linear pods of sediment emerging from the foliation either basally or supraglacially. They have been described as foliation-parallel ridges (Glasser et al., 1998) and result from the incorporation of basal sediment into the foliation either at the glacier bed or where it has been elevated to the surface along thrusts and reworked by supraglacial streams. Lateral #ow compression enhances the foliation by tight folding, squeezing the sediment to form linear pods parallel to the foliation (Figs. 2B and C). In this case there is a direct link between debris-rich glacier structures and the landforms emerging in front of the glacier. The rectilinear ridge network is similar in planform to the crevasse squeeze ridge systems described by several authors (Sharp, 1985a; Solheim, 1991; Boulton et al., 1996), although in cross-section they are morphologically distinct. True crevasse-"ll ridges usually M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 1241 Fig. 23. The morphology and planform geometry of the geometrical ridge network on the southwest side of Kongsfjorden. These landforms are described in greater detail by Bennett et al. (1996a). 1242 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 have a more dyke-like form (Sharp, 1985a; Boulton et al., 1996). 9. Discussion The above morphological, sedimentological and structural data provide a framework for developing a model to explain the development of the moraine complex at Kongsfjorden. This model is applicable to similar situations where tidewater glaciers are able to transfer large volumes of sediment from a fjord onto low relief terrain. This investigation also raises a more general issue about the variability of landform assemblages produced by glacier surges in glaciomarine environments. 9.1. Landform-sediment model for Kongsfjorden The landform assemblage produced by the 1948 advance of the Kongsvegen/Kronebreen tidewater glacier complex at the head of Kongsfjorden consists of a thrustmoraine complex composed primarily of glaciomarine sediments transported onshore by the advancing glacier. Flow compression and thrusting appear to be facilitated by reverse bedrock slopes. Glaciomarine sediments, particularly those from grounding-line fans, form the principal facies within the moraine complex on the southwest side, while on the northeast side fjord bottom muds dominate. As the glacier complex retreated, melting out of debris-rich ice structures resulted in a rectilinear ridge network, similar in planform, although not in ridge crosssection, to that frequently described as crevasse-"ll ridges (Sharp, 1985a; Boulton et al., 1996). The contrast in morphology between the thrust-moraine complex on the southwest side of the fjord and that on Ossian Sarsfjellet, re#ects the di!ering nature of the tectonic regime in these two locations. On the southwest side of the fjord there appears to have been a transpressive regime between fast #owing ice in the centre of the fjord and grounded ice on the southern side of the fjord (Fig. 24). This resulted in a more linear thrust system and consequently an extensive ridge complex. The presence of bedded sediments in the form of a large grounding-line fan allowed the development of an imbricate thrust system during the transpressive episode, in which the style of deformation is similar to that within a classic duplex thrust system (cf. Butler, 1983). Elements of this transpressive system are visible on the 1970 aerial photograph in which the three main ridge systems of the moraine complex (Fig. 9) can be seen melting out of prominent longitudinal fractures sub-parallel to the #ow direction. The process of transpression between the two #ow units was assisted by the presence of a steep bedrock slope along the southwest side of the fjord. In contrast, on Ossian Sarsfjellet there were two directions of compression. One from the southwest produced by the advancing front of Kronebreen South and a second from the southeast from Kronebreen Central (Fig. 24). The interference pattern between these two directions of compression resulted in the more complex and varied thrust-moraine morphology. Proglacial thrusting appears to have been the result of compression from the southeast, that is from Kronebreen Central, and may predate compression from the southwest caused by the advance of Kronebreen South. The intersection of thrusts caused by the interference between these two Fig. 24. Palaeogeographical map of Kongfjorden in 1936 and 1948. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 directions of compression may have resulted in the distinctive diamicton pinnacles in the Ossian Sarsfjellet moraine complex. 9.2. Surges and the glaciomarine record: landform variability Sharp (1988b) proposed a landform model for landbased glacier surges which consists of some form of large push-moraine, behind which a rectilinear network of ridges is produced by the melting out of debris within basal crevasses and from other debris-rich structures, including thrusts. This type of genetic model has been widely accepted as typical of surge-type glaciers, and was applied to the o!shore record by Solheim (1991) in describing a surge of the tidewater glacier Bra> svellbreen in eastern Svalbard (Fig. 1A). However, perhaps the most detailed description of the landform record associated with a tidewater glacier surge is provided by Boulton et al. (1996) who described the landform/sediment assemblage produced by a surge of the Svalbard tidewater glacier Sefstr+mbreen (Fig. 1A). This assemblage is dominated by evidence for subglacial deformation of glaciomarine mud and its deposition through sediment #ow into basal crevasses and by extrusion in front of the advancing glacier. The landform assemblage described by Boulton et al. (1996) is very di!erent from that reported here for Kongsfjorden. Although the principal morphological components are similar*large moraine behind which there is a network of rectilinear ridges*the origin of these landforms is very di!erent. Subglacial deformation and sediment #ow into basal crevasses was not recorded in Kongsfjorden, instead the assemblage is dominated by landforms formed by thrusting both in front and within the body of the advancing glacier. The contrast between these two landform assemblages, both formed by tidewater glacier surges, re#ects two signi"cant di!erences. Firstly, part of Kongsvegen/Kronebreen advanced over coarse free draining gravels in which porewater pressures would not have been suitable for rapid deformation, whereas Sefstr+mbreen advanced over "ne-grained glaciomarine muds ideal for subglacial deformation. Secondly, the landform assemblage produced by Sefstr+mbreen is dominated by landforms associated with glacial structures indicative of #ow extension (crevasse-"ll ridges), while that in Kongsfjorden is dominated by landforms formed by structures indicative of #ow compression (thrust ridges). This re#ects a di!erence in the overall tectonic regime of the two glacier surges; a function of both a contrast in coastal/fjordal geometry and the composition of the two glacier lobes. In the case of Sefstr+mbreen the glacier lobe was composed of a single surging #ow unit, which expanded into a broader fjord basin (Ekmanfjorden) and would therefore have had a predominantly extensional regime, except in the 1243 terminal zone. In contrast, Kongsfjorden is more topographicaly constrained by steep fjord sides and the advancing glacier lobe was composed of two competing #ow units (Kongsvegen and Kronebreen South), only one of which was surging. In such conditions a compressive regime is more likely to dominate, and is re#ected in the resultant landform assemblage. The contrast between Kongsvegen/Kronebreen and Sefstr+mbreen emphasises, therefore, the importance of the overall tectonic regime or setting in determining the landform/sediment assemblage produced by a tidewater glacier surge. This is determined by such variables as: (1) the availability of deformable sediment; (2) coastal/fjordal geometry; and (3) the composition of the advancing glacier lobe. These two contrasting landform assemblages may re#ect end member of a continuum; one re#ecting compressive situations and the other extensional ones. This contrast also emphasises the need for a range of modern analogue studies if we are to develop comprehensive landform/sediment models for rapid tidewater glacier advances, which are essential in the interpretation of the glaciomarine Pleistocene record. Finally, it is also interesting to note that the Kongsfjorden landform complex shows a close resemblance to the nonsurge-induced partially glaciomarine complex of Comfortlessbreen in Engelskbukta (Fig. 1A), the next fjord to the south of Kongsfjorden (Huddart and Hambrey, 1996). A similar suite of facies is present including glaciomarine muds, gravel and diamicton. Furthermore the moraine complex is morphologically similar, with thrust-ridges of a variety of forms present. The main di!erence is that, whereas the bulk of the Kongsfjorden sediment is derived from the fjord, at Comfortlessbreen probably less than half is. This has important implications for our ability to discriminate between landform assemblages produced by surge-type glaciers and those produced by nonsurge-type glaciers undergoing compressive #ow and in practice it may not be possible. 10. Conclusions From a detailed investigation of the morphology, sedimentology and tectonic structures of a moraine complex produced by a tidewater glacier, linked to an investigation of the ice structures and debris content in the ice margin, the following conclusions can be drawn. 1. Two contrasting moraine complexes were produced by the 1948 advance of Kongsvegen/Kronebreen on either side of Kongsfjorden. Structural glaciological and geomorphological evidence suggests that a transpressive glaciotectonic regime was responsible for the prominent ridge system on the southwest side and a normal compressive regime for the short-crested 1244 2. 3. 4. 5. 6. 7. 8. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 ridges and pinnacles of a &hummocky' nature on the northeastern side (Ossian Sarsfjellet). The glacier advance involved both Kronebreen South and Kongsvegen. We suggest that this advance was precipitated by a surge of Kongsvegen and have presented a simple model based on the work of Mercer (1961) to explain this observation. This type of model needs further investigation. The sediments within the moraine complexes comprise diamicton, muddy and sandy gravel, gravelly sand, sand, mud and sand-mud laminite. Broken shell fragments and foraminifera are common and many exposures show salt e%orescence. These facies have been reworked from the fjord #oor and include coarse ice-proximal grounding-line fan deposits and "ne basinal muds. Emplacement of these sediments on land by thrusting is indicated by rectilinear upglacier-facing slopes, imbricate stacking of often contrasting sediment slabs, evidence for deformation within sheets of sediment (including shear and recumbent folding) and a clear linkage to thrusting in the contemporary ice margin. Reworking of sediment is achieved by debris-#ow processes, aided by the melting of thrust-blocks of ice preserved in the moraine complexes. The contrast between the landform/sediment assemblage documented here for Kongsvegen/Kronebreen and that reported by Boulton et al. (1996) for Sefstr+mbreen is attributed to a di!erence in the overall tectonic regime during the surge; a contrast between compression at Kongsvegen/Kronebreen and extension at Sefstr+mbreen. This is determined by such variables as: (1) the availability of deformable sediment; (2) coastal/fjordal geometry; and (3) the composition of the advancing glacier lobe. Discrimination between landform/sediment assemblages produced by nonsurge-type and surge-type tidewater glaciers is problematic as surge-type moraine complexes, of the Kongsvegen/Kronebreen type, in which thrusting is the dominant mechanism, are also found at the margins of nonsurge-type glaciers in Svalbard. Understanding the interaction between sedimentation and glaciotectonics in marginal-marine settings is highly relevant to interpreting the late Pleistocene record in more southerly latitudes. This investigation o!ers one highly relevant model, but a range of contrasting glaciomarine-glacioterrestial settings need to be investigated before older glacigenic sequences and landforms can be fully understood. Acknowledgements Fieldwork was funded in 1995 by the University of Greenwich and Liverpool John Moores University, and in 1996 by the UK Natural Environment Research Council (Grant No. GR9/02185). The logistical support of Nick Cox and the UK Natural Environment Research Council Arctic Base at Ny-A_ lesund is gratefully acknowledged as is Jean-Francois Ghienne for "eld assistance in 1995. We would also like to thank the Sysselmann of Svalbard for permission to work and camp on the Ossian Sarsfjellet plant reserve. Reviews by Doug Benn and Jaap van der Meer are gratefully acknowledged. References Allen, C. R., Kamb, W. B., Meier, M. F., & Sharp, R. P. (1960). Structure of the Lower Blue glacier, Washington. Journal of Glaciology, 68, 601}625. Benn, D. I. (1994). Fabric shape and intepretation of sedimentary fabric data. Journal of Sedimentary Research, A64, 910}915. Benn, D. I. (1995). Fabric signiature of subglacial till deformation, BreidamerkurjoK kull, Iceland. Sedimentology, 42, 735}747. Benn, D. I., & Ballantyne, C. K. (1994). Reconstructing the transport history of glacigenic sediments: a new approach based on the co-variance of clast form indices. Sedimentary Geology, 25, 215}227. Benn, D. I., & Evans, D. J. A. (1996). The intepretation and classi"cation of subglacially-deformed materials. Quaternary Science Reviews, 15, 23}52. Bennett, M. R., Hambrey, M. J., Huddart, D., & Ghienne, J. F. (1996a). The formation of geometrical ridge networks (&crevasse-"ll' ridges), Kongsvegen, Svalbard. Journal of Quaternary Science, 11, 430}438. Bennett, M. R., Huddart, D., Hambrey, M. J., & Ghienne, J. F. (1996b). Moraine development at the high-arctic valley glacier, Pedersenbreen, Svalbard. Geogra,ska Annaler, 78A, 209}222. Bennett, M. R., Hambrey, M. J., & Huddart, D. (1997). Modi"cation of clast shape in high-arctic glacial environments. Journal of Sedimentary Petrology, A67, 550}559. Bennett, M. R., Hambrey, M. J., Huddart, D., & Glasser, N. F. (1998). Glacial thrusting and moarine-mound formation in Svalbard and Britain: the example of Coire a' Cheud-chnoic (Valley of Hundred Hills), Torridon Scotland. Quaternary Proceedings, in press. Boulton, G. S. (1967). The development of a complex supraglacial moraine at the margin of S+rbreen, Ny Friesland, Vestspitsbergen. Journal of Glaciology, 6, 717}735. Boulton, G. S. (1972). The role of thermal reH gime in glacial sedimentation. In: Price, R. J., & Sugden, D. E. (Eds.), Polar geomorphology (vol 4 pp. 1}19). Institute of British Geographers Special Publication. Boulton, G. S. (1986). Push-moraines and glacier-contact fans in marine and terrestrial environments. Sedimentology, 33, 677}698. Boulton, G. S. (1990). Sedimentary and sea level changes during glacial cycles and their control on glacimarine facies architecture. In: J. A. Dowdeswell, & J. D. Scourse, (Eds.), Glacimarine environments: processes and sediments (vol 53, pp. 15}52). Geological Society Special Publication. Boulton, G. S., & Caban, P. (1995). Groundwater #ow beneath ice sheets: Part II-its impact on glacier tectonic structures and moraine formation. Quaternary Science Reviews, 14, 563}587. Boulton, G. S., Van der Meer, J. J., Hart, J., Beets, D., Ruegg, G. H. J., Van der Wateren, F. M., & Jarvis, J. (1996). Till and moraine emplacement in a deforming bed surge } an example from a marine environment. Quaternary Science Reviews, 15, 961}987. Butler, R. W. H. (1983). The terminology of structures in thrust belts. Journal of Structural Geology, 4, 239}245. Clark, J. B. (1949). A hydraulic process for increasing the productivity of wells. ¹ransactions of the American Institute of Mechanical Engineers, 186, 1}11. M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Cowan, E. A., & Powell, R. D. (1990). Suspended sediment transport and deposition of cyclically interlaminated sediment in a temperate glacial fjord, Alaska, U.S.A. In: J. A. Dowdeswell & J. D. Scourse, (Eds.), Glacimarine environments: processes and sediments (vol. 53, pp. 75}89). Geological Society Special Publication. Domack, E. W., & Lawson, D. E. (1985). Pebble fabric in an ice-rafted diamicton. Journal of Geology, 93, 577}591. Dowdeswell, J. A. (1987). Processes of glacimarine sedimentation. Progress in Physical Geography, 11, 52}90. Dowdeswell, J. A., & Sharp, M. (1986). Characterisation of pebble fabrics in modern terrestrial glacigenic sediments. Sedimentology, 33, 699}710. Dowdeswell, J. A., & and Scourse, J. D. (1990). Glacimarine environments: processes and sediments (vol. 53) Geological Society Special Publication. Dowdeswell, J. A., Hambrey, M. J., & Wu, R. (1985). A comparison of clast fabric and shape in Late Precambrian and modern glacigenic sediments. Journal of Sedimentary Petrology, 55, 691}704. Dowdeswell, J. A., Hamilton, G. S., & Hagen, J. O. (1991). The duration of the active phase on surge-type glaciers: contrasts between Svalbard and other regions. Journal of Glaciology, 37, 388}400. Elverh+i, A., Liest+l, O., & Nagy, J. (1980). Glacial erosion, sedimentation and microfauna in the inner part of Kongsfjorden, Spitsbergen. Norsk Polarinstitutt Skrifter, 172. Elverh+i, A. L+nne, O., & Seland, R. (1983). Glaciomarine sedimentation in a modern-fjord environment, Spitsbergen. Polar Research, 1, 127}149. Eyles, C. H., Eyles, N., & Miall, A. D. (1985). Models of glaciomarine sedimentation and their application to the interpretation of ancient glacial sequences. Palaeogeography, Palaeoclimatology, Palaeoecology, 51, 15}84. Eyles, N., & McCabe, A. D. (1989). The Late Devensian ((22, 000 BP) Irish Sea Basin: the sedimentary record of a collapsed ice sheet margin. Quaternary Science Review, 8, 307}351. Eyles, N., & McCabe, A. D. (1991). Glaciomarine deposits of the Irish Sea basin: the role of glacioisostatic disequilibrium. In: J. Ehlers, P. Gibbard, and J. Rose, (Eds.), Glacial deposits of Great Britain and Ireland, (pp. 311}331). Rotterdam: Balkema. Glasser, N. F., Hambrey, M. J., Crawford, K, Bennett, M. R., & Huddart, D. (1998). The structural glaciology of Kongsvegen, Svalbard and its role in landform genesis. Journal of Glaciology, in press. Hagen, J. O., Liest+l, O., Roland, E., & J+rgensen, T. (1993). Glacier atlas of Svalbard and Jan Mayen. Norsk Polarinstitute Medd 129. Hambrey, M. J. (1994) Glacial environments. London: University College Press. Hambrey, M. J., & Huddart, D. (1995). Englacial and proglacial glaciotectonic processes at the snout of a thermally complex glacier in Svalbard. Journal of Quaternary Science, 10, 313}326. Hambrey, M. J., Dowdeswell, J. A., Murray, T., & Porter, P. R. (1996). Thrusting and debris-entrainment in a surging glacier: Bakaninbreen, Svalbard. Annals of Glaciology, 22, 241}248. Hambrey, M. J., Huddart, D., Bennett, M. R., & Glasser, N. F. (1997). Genesis of &hummocky moraine' by thrusting in glacier ice: evidence from Svalbard and Britain. Journal of the Geological Society of ¸ondon, 154, 623}632. Hamilton, G. S., & Dowdeswell, J. A. (1996). Controls on glacier surging in Svalbard. Journal of Glaciology, 42, 157}168. Hart, J. K. (1994). Till fabric associated with deformable beds. Earth Surface Processes and ¸andforms, 19, 15}32. Hubbart, M. K., & Willis, D. G. (1957). Mechanics of hydraulic fracturing. Petroleum ¹ransactions of the American Institute of Mechanical Engineers, 210, 153}169. Huddart, D. (1993). Controversial Irish Sea basin glacial models: some answers from the Cumbrian lowlands. Proceedings of the Cumberland Geological Society, 5, 476}480. 1245 Huddart, D., & Clark, R. (1993). Con#icting interpretations from the glacial sediments and landforms in Cumbria. Proceedings of the Cumberland Geological Society, 5, 419}436. Huddart, D., & Hambrey, M. J. (1996). Sedimentary and tectonic development of a high-arctic thrust moraine complex, Comfortlessbreen, Svalbard. Boreas, 6, 227}243. KruK ger, J. (1984). Clasts with stoss-lee forms in lodgement tills: a discussion. Journal of Glaciology, 30, 241}243. Lamont, J. (1876). >achting in the Arctic Seas. London: Murray. Lawson, W. J., Sharp, M. J., & Hambrey, M. J. (1994). The structural geology of a surge-type glacier. Journal of Structural Geology, 16, 1447}1462. Lefauconnier, B., Hagen, J. O., & Rudant, J. P. (1994). Flow speed and calving rate of Kongsbreen glacier, Svalbard, using SPOT images. Polar Research, 11, 59}65. Liest+l, O. (1988). The glaciers in Kongsfjord area, Spitsbergen. Norsk Geogra,ska ¹iddskrift, 42, 231}238. Mackiewicz, N. E., Powell, R. D., Carlson, P. R., & Molina, B. F. (1984). Interlaminated ice-proximal glacimarine sediments in Muir Inlet, Alaska. Marine Geology, 57, 113}147. Maizels, J. (1995). Sediments and landforms of modern proglacial terrestrial environments. In: J. Menzies, (Ed.), Modern glacial environments: processes, dyncamics and sediments, (pp. 365}416). Oxford: Butterworth-Heinemann. McCarroll, D. (1995). Geomorphological evidence from the Llyn Peninsula constraining models of the magnitude and rate of isostatic rebound during deglaciation of the Irish Sea Basin. Geological Journal, 30, 157}163. McCarroll, D., & Harris, C. (1992). The glacigenic deposits of western Llyn, North Wales: terrestrial or marine? Journal of Quaternary Science, 7, 19}29. Meier, S. (1969). Terrestrial photogrammetry on an arctic glacier during the polar night. <ermessungsinformatiionen aus Jena, 21, 45}49. Meir, M. F., & Post, A. (1969). What are glacier surges? Canadian Journal of Earth Sciences, 6, 807}817. Mercer, J. H. (1961). The response of fjord glaciers to changes in the "rn limit. Journal of Glaciology, 3, 850}858. Moncrie!, A. C. M. (1989). Classi"cation of poorly sorted sedimentary rocks. Sedimentary Geology, 65, 191}194. Powell, R. D. (1981). A model for sedimentation by tidewater glaciers. Annals of Glaciology, 2, 129}134. Powell, R. D. (1984). Glaciomarine processes and inductive lithofacies modelling of ice shelf and tidewater glacier sediments based on Quaternary examples. Marine Geology, 7, 1}52. Powell, R. D. (1990). Glacimarine processes at grounding-line fans and their growth to ice-contact deltas. In: J. A. Dowdeswell, & J. D. Scourse, (Eds.), Glacimarine environments: processes and sediments (vol. 53, pp. 53}73). Geological Society Special Publication. Powell, R. D., & Molnia, B. F. (1989). Glacimarine sedimentary processes, facies and morphology of the south-southeast Alaska shelf and fjords. Marine Geology 85, 359}390. Powell, R. D., & Domack, E. (1995). Modern glaciomarine environments. In: J. Menzies, (Ed.), Modern glacial environments: processes, dyncamics and sediments (pp. 445}486). Oxford: ButterworthHeinemann. Sharp, M. (1982). Modi"cation of clasts in lodgement tills by glacial erosion. Journal of Glaciology, 28, 475}481. Sharp, M. (1985a). &&Crevasse-"ll'' ridges-a landform type characteristic of surging glaciers? Geogra,ska Annaler, 67A, 213}220. Sharp, M. (1985b). Sedimentation and stratigraphy at EyjabakkajoK kull-an Icelandic surging glacier. Quaternary Research, 24, 268}284. Sharp, M. (1988a). Surging glaciers: mechanisms. Progress in Physical Geography, 12, 349}370. Sharp, M. (1988b). Surging glaciers: geomorphic e!ects. Progress in Physical Geography, 12, 533}559. 1246 M.R. Bennett et al. / Quaternary Science Reviews 18 (1999) 1213}1246 Solheim, A. (1991). The depositional environment of surging sub-polar tidewater glaciers: a case study of the morphology, sedimentation and sediment properties in a surge a!ected marine basin outside Nordaustlandet, the Northern Barents Sea. Norsk Polarinstitutt Skrifter, 194. Solheim, A., & P"rman, S. L. (1985). Sea-#oor morphology outside a grounded, surging glacier: Bra> svellbreen, Svalbard. Marine Geology, 65, 127}143. Strum, M. 1987. Observations on the distribution and characteristics of potholes on surging glaciers. Journal of Geophysical Research, 92 (B9), 9015}9022. Voigt, U. (1965). Die Bewegung der Gletscherzunge des Kongsvegen (Kingsbay, Westspitzbergen). Petermanns Mitteilungen, 109, 1}8. Warren, C. R. (1992). Iceberg calving and the glacioclimatic record. Progress in Physical Geography, 16, 253}282. Wilhelm, F. (1966). Die glaziologischen Ergebnisse der Spitzbergenkundfahrt der Sektion Amberg des Deutschen Alpenvereins. Mitteilungen der Geographischen Gesellschaft in Mu( nchen, 46, 151}183.