Introduction

Sediments along a 60 m section of the north-western margins of Leirbreen, in south Norway, were studied in August 1981. Here the glacier flows against a reverse bedrock slope, producing a steep (15° to 16°) compressive lower section with no crevasses (Fig. 1). Shearing close to the ice margin had brought frozen current-bedded silts, sands, and gravels from a subglacial location to the glacier surface, where their gradual release during ablation was associated with only slight internal disturbance. Their incorporation into the glacier is considered to result from basal freezing close to the ice margin.

Fig. 1. Part of the snout of Leirbreen showing sediments brought to the glacier surface by shearing. Note figures, right centre.

Location and recent geomorphological history

Leirbreen lies on the northern flanks of Smørstabbtindein, 2 km east of Krossbu in western Jotunheimen (Fig. 2a). Its lower limits lie at an altitude of 1500 m, where the mean annual temperature is approximately −2.2°C with a mean January temperature of around −10°C and mean July temperature 6.5°C (based on altitude adjusted data from Fanaråken (2 062 m), Reference BruunBruun, 1967).

Fig. 2. (a) Location map.

(b) Flan of supraglacial and pro-glacial sediment units.

The glacier was visited by de Boer in 1947 (Reference de Boerde Boer in 1949), who described shoreline terraces composed of silty sediments occupying a broad shallow gully “recently deserted by the waters of a glacier-dammed lake” (Reference de Boerde Boer, 1949, p.332). The lake had apparently drained beneath the glacier as a result of thinning and recession. The terraces were still present in 1981, though in a subdued, degraded form, 300 m to the north of the present study area (Fig. 2a). In 1947 water was flowing from the former ice-dammed lake basin under the glacier, southwards towards an area of concentric crevasses surrounding a depression in the glacier surface, where subsidence into a cavity had apparently taken place. Approximately 140 m of retreat since 1947 had removed this crevassed area by 1981 and there was no sign of subsidence in the glacier surface. However, the bedrock slope continues to direct glacial melt water southwards beneath the glacier, parallel to the ice margin, as was evidenced by a small stream which emerged from beneath the ice within the present study area (Fig. 2b).

Site Description

Shear planes in the steep north-western ice front of Leirbreen showed a somewhat sinuous pattern of outcrop (Fig. 1) and within 30 m of the glacier edge small amounts of silt and sand were observed melting from these shear planes and washing down over the glacier snout in a layer a few centimetres thick. However, associated with the lowermost shear plane, along a 60 m section of the ice margin, were bedded silts, sands, and gravels forming an irregular layer up to about 1.3 m thick, extending from its line of outcrop on the glacier down the ice front and beyond the apparent ice limits. A detailed sedimentological analysis will be presented elsewhere (paper on glaciodeltaic sediments at Leirbreen, southern Norway, in preparation by the present authors) and only a brief description will be given here.

Below the ice cave, from which emerged the meltwater stream, the sediments comprised mainly silts and sands, and were divided into a pro-glacial sub-unit (unit 1a) and a supraglacial sub-unit (unit 1b) by the stream channel which followed the ice margin (Fig. 2 and 3). Above the ice cave, sediment unit 2 comprised gravel and sand and extended from the outcrop of the shear plane to beyond the ice limits, rising to a distinct mound some 1.5 m high close to the probable limit of glacier ice (Fig. 3).

Fig. 3. General view of ice margin looking north-east. Note spade, right centre. The sandy unit below the spade (unit la) is underlain by glacier ice. The pro-glacial ice-dammed lake terraces occupy a small valley beyond the glacier/bedrook skyline in the middle distance.

Along their line of outcrop on the glacier surface these sediments were frozen solid and displayed a sharply defined undulating upper boundary (Fig. 4). Down-slope from their outcrop the sediments were in process of thawing from the surface downwards.

Fig. 4. Englacial sediment band melting out on the glacier surface close to the ice margin.

Excavation of unit 1a (Fig. 2b, trenches 1, 2, and 3) revealed a deltaic sequence (Reference Jopling and YatsuJopling, [^c1971]) of foreset and bottomset beds. The topset was, however, replaced by a layer between 30 cm and 60 cm thick (Fig. 5) of fluvial granule gravels with channel scour and fill structures interrupted by silt drapes. The underlying foreset layer ranged from 80 cm in thickness in trench 1 to 30 cm in trench 3. The planar cross beds consisted of fine sand, were between 5mm and 30 mm thick, and were cut by small-scale, Steeply inclined faults (Fig. 5).

Fig. 5. Trench 1 showing foreset beds overlain by gravels containing irregular layers of silt and sand. Note small-scale normal faults in the foreset beds.

The spade is resting on frozen sediment and the underlying ice mass is exposed in the right foreground.

The foreset beds graded through a concave toeset into a silt bottomset with ripple laminations. The bottomset was at most only a few centimetres in thickness, being truncated at its base by a buried mass of glacier ice with a sharply defined, smooth, slightly undulating surface (Fig. 6). In places this buried ice pinched out the bottomset completely, abruptly truncating the foresets.

Fig. 6. Base of trench 1 showing concave toeset grading into ripple-bedded bottomset silts which are pinched out on the left by the buried ice mass. Note sharp, smooth, undulating boundary between sediments and buried ice. The scale is approximately 25 cm in length.

The supraglacial sediments of unit 1b ranged in thickness from 20 cm to 70 cm and comprised silts and fine sands with well-developed climbing-ripple laminations (Fig. 7). The thickest part of the sequence showed a gradational change from ripple-drift silts at the base to sands above. These silts and overlying sands were texturally indistinguishable from the bottomset and foreset respectively of unit 1a, and apparently represented a lateral continuation of them. Thawing of these sediments had caused little disturbance of their sedimentary structures. It appeared that in their frozen state they contained no excess ice so that thaw-consolidation was minimal.

Fig. 7. Climbing ripple laminations in englacial sediment band, unit 1b. The hammer is resting on glacier ice. The ripple laminated silts grade upwards into medium sands.

Above the ice cave was a kame-like deposit of planar beds of gravels and coarse sand dipping at up to 35° towards the south-east (unit 2, Fig. 2b). Here the englacial sediment band was 60 cm thick at its outcrop on the glacier surface and sediment thickness increased down the ice margin to approximately 1.5 m (Fig. 3).

Pro-glacial topography in this area precludes the accumulation of these sediments in situ, as does the presence of the englacial sediments of unit 1b. Reference ShawShaw (1982) and Reference Haldorsen and ShawHaldorsen and Shaw (1982) have described englacial stratified sediments around the eastern margin of the stagnating glacier Omnsbreen, south Norway. Deposition within englacial passages is described, producing fluvial sediments interstratified with glacier ice. At Leirbreen, however, surface ice gradients, the lack of crevasses, and observations of recent retreat rates, indicate active marginal ice where compressive shearing rather than stagnation is the dominant process. It is unlikely that englacial passages as described at Omnsbreen would remain open in this compressive marginal zone.

It is suggested, therefore, that deposition of the delta originally took place in a subglacial pool ponded against the reverse bedrock slope. The correspondence between palaeocurrent directions in the sediments (Fig. 2b) and the direction of subglacial drainage from de Boer’s ice-dammed lake suggests that subglacial deposition resulted from drainage of the ice-dammed lake, which apparently occurred during the 1940’s.

Mechanism of Sediment Incorporation and Displacement by the Glacier

Weertman’s mechanism of basal freezing for subpolar glaciers (Reference WeertmanWeertman, 1961) whereby intact units of frozen subglacial sediment may be incorporated into the base at the transition from warm-based interior to cold-based margin, is well known. ( Reference BoultonBoulton 1970, Reference Boulton, Price and Sugden1972) elaborated on this theory, showing that when unconsolidated subglacial sediment freezes to the glacier, shearing is likely to occur at the boundary between frozen and unfrozen sediment, since this is where shear strength is lowest. Compression in the glacier at the thermal transition leads to the development of thrust planes which extend to the base of the frozen-on sediment, and shearing along the thrust planes brings frozen material to the glacier surface. Reference MenziesMenzies (1981) elaborated further on possible modes of subglacial sediment incorporation, and stressed our lack of knowledge regarding temperatures within subglacial debris. The presence of frozen stratified englacial sediments in sub-polar glaciers has been reported by Reference BishopBishop (1957), Reference GoldthwaitGoldthwait (1960), (Reference BoultonBoulton 1970, [^c Reference Boulton and Goldthwait1971]), and Reference HookeHooke (1970). (Reference ShawShaw (1971, Reference Shaw1972) described similar incorporation of frozen lacustrine sediments by a Pleistocene ice sheet in Shropshire, England.

In temperate glaciers, ice thrusting close to the margins has been explained in terms of compressive flow resulting from ice thinning (Reference KupschKupsch, 1962). Reference GunnGunn (1964) reported thrust planes near the fronts of the Fox and Franz Joseph Glaciers in New Zealand. Flow velocities in these glaciers decreased rapidly near the snout, resulting in transverse foliation dipping up-glacier and thrust planes parallel to the foliation. Morainic debris was incorporated in these thrust planes giving supraglacial ridges referred to by Gunn as “moraine ogives”. He did not state whether the debris associated with the thrust planes was frozen, nor did he mention the presence of stratified material. Apparently similar thrust planes with associated debris which formed dirt cones on the surface of Svartisen, north Norway, were described by Reference StokesStokes (1958).

Reference OkkoOkko (1955) mentioned water-lain sediments delivered by shearing to the snout of Hoffelsjökull, Iceland, and Reference FaegriFaegri (1934) photographed fluvio-glacial material associated with shear planes at the snout of Nigardsbreen, south Norway. At Midtdalsbreen, Hardangerjøkulen, south Norway, Reference Andersen and SollidAndersen and Sollid (1971) have described marginal ice-cored moraines derived from till-laden bands emerging from the glacier. Fluvio-glacial sediments were also delivered to the ice front in this way. Andersen and Sollid proposed that debris was frozen onto the base of the glacier close to the snout due to winter cold penetrating through the ice. Freezing of the marginal zone to the bed caused compressive flow and shearing which brought the frozen subglacial material to the surface.

Reference HookeHooke (1981), however, cast doubt on the effectiveness of the shear mechanism in bringing debris from the glacier bed to the surface. He cited evidence that debris concentrations in excess of 10% stiffen ice rather than soften it, making shear along a debris band unlikely. Reference IkenIken (1981) argued that shear fracture occurs in clear ice above the debris zone rather than within it, although how such shearing would bring debris to the surface was not explained. At Leirbreen the englacial sediment was frozen and its internal sedimentary structures were largely undisturbed. It is probable, therefore, that the sediments were frozen in a subglacial location close to the ice margin by the penetration of winter cold prior to their incorporation by the glacier. A thrust plane then developed through the marginal ice, extending below the frozen-on sediment. Thrusting along this plane brought the frozen debris to the Surface. In this example, therefore, it is argued that shearing along a thrust plane took place immediately below the debris band, probably along its lower boundary.

It is generally suggested that the maximum thickness of ice through which winter cooling may penetrate is approximately 15 m (Reference Boulton, Price and SugdenBoulton, 1972; Reference Derbyshire, Derbyshire, Derbyshire, Gregory and HailsDerbyshire, 1979; Reference PatersonPaterson, 1981). Should basal freezing occur, the resulting frozen subglacial sediments would be protected from summer melting by the presence of overlying ice, the temperature of which cannot rise above freezing. Indeed, near-surface temperature measurements by Reference PatersonPaterson (1972) in the ablation zone of the Athabasca Glacier showed ice temperatures remaining below the pressure-melting point throughout the summer. Paterson concluded that where summer ablation is less than the depth of winter cold penetration (c. 15 m) near-surface ice may remain permanently below its melting point in the ablation zone.

Ablation at Leirbreen has not been measured, but at the adjacent glacier, Storbreen, at an altitude of 1 400 m (somewhat lower than the margins of Leirbreen), net ablation has averaged around 2.2 m during the 1970’s (Reference LiestølLiestøl, 1972–80). At Leirbreen it therefore appears that the surface few metres of ice in the ablation area may not by warmed to the freezing point during the summer.

In the absence of heat conducted from above, thawing of the frozen submarginal sediments can only occur through basal melting by geothermal heat, which would reduce the thickness of the frozen layer only slightly, and by sensible heat carried beneath the glacier by melt water. Clearly, the latter may be locally significant. There are, however, no crevasses at the margins of Leirbreen to allow penetration of melt water and warm air from the ice surface to the glacier bed. In addition, with a negative thermal gradient near the bed of the glacier, heat from laterally penetrating melt water is likely to be conducted upwards to warm and melt the glacier base, as well as downwards to warm and thaw frozen subglacial sediments.

Beneath the marginal zone of Leirbreen, therefore, sediments frozen by winter cooling are likely to remain frozen throughout the following summer, protected from thawing by the overlying ice. It follows that if the ice thickness above a particular location remains less than approximately 15 m for a number of years, that is the glacier is neither advancing nor retreating rapidly, the thickness of the submarginal frozen sediments would increase annually as the cold wave of successive winters penetrated to the bed. Such incremental freezing over a number of years might well be responsible for producing a submarginal layer of frozen sediments of the extent and thickness observed melting out of the ice surface at Leirbreen in 1981, and described in this paper.

Conclusions

A band of deltaic sediments was observed melting out of the glacier surface close to the margins of Leirbreen. Along its outcrop on the glacier surface the sediment band was frozen solid, and ripple laminations in silts and sands were well preserved. On the basis of palaeocurrent directions in the deltaic sequence, it was concluded that deposition occurred subglacially during drainage of an ice dammed lake to the north, during the 1940’s (Reference de Boerde Boer, 1949). At that time the glacier was larger than today, the snout extending some 150 m beyond its position in 1981.

The lack of disturbance in the sediment band exposed in the glacier surface suggested that it was frozen prior to incorporation by the glacier, and that it was not affected by shearing during subsequent glacial transport. Initial subglacial freezing of the sediments, and their freezing onto the sole of the glacier, probably resulted from winter cooling through thin marginal ice. Consideration of the thermal regime suggested that if winter cooling led to freezing of unconsolidated sediments beneath the glacier, they would be protected from subsequent summer thawing by the overlying ice. Further cooling during subsequent winters might then lead to incremental increases in the thickness of the subglacial frozen sediments. Annual winter cooling of glacier ice below its freezing point can be shown to occur to some 15 m depth, and subglacial frozen sediments are likely, therefore, in a zone within approximately 41 m of the snout of Leirbreen, where ice thickness is less than 15 m. Such a frontal zone of cold ice, with subglacial frozen sediments, is most likely to occur in higher, more continental glaciers where winter snow accumulation and summer ablation rates are relatively low. It may be an extremely important factor in explaining variations in moraine morphology and composition between different glaciers in Jotunheimen.

At Leirbreen, compressive flow against the bedrock slope, or against cold-based ice frozen to the bed, led to the formation of thrust planes which penetrated through the ice and into the frozen subglacial sediment. Transport through the glacier to the surface resulted from shearing along a thrust plane which lay immediately below the lower boundary of the frozen sediment layer. Internal deformation within the layer itself therefore did not occur. The preservation potential of the resulting ice-cored deltaic sediments is however low, melting of the buried ice will inevitably cause slumping and faulting, and pro-glacial drainage during spring periods of nival thaw is likely to erode the soft sediments rapidly.