(a) Plateau Ice Overflow

Wright and PriestleyReference Wright and Priestley ⁷ introduced the term wall-sided glaciers to describe, “streams of ice originating in, and fed by, upland ice of any description, and flowing down towards sea level unconfined by any marked valley wall”. These could originate either by local advance of highland ice or by retreat of the highland ice. Any depressions occupied by these glaciers are regarded as unconnected with the erosive action of the ice. Wall-sided glaciers were recognized in Spitsbergen by Ahlmann Reference Ahlmann ⁸ but he did not make it clear whether these glaciers which he defines elsewhereReference Ahlmann ⁹ as “covering the side of a valley or some part of it” conformed to Wright and Priestley’s definition. A. HambergReference Hamberg ¹⁰ describing niche glaciers of northern Sweden regarded them as genetically related to hanging glaciers, from which at times, he found them difficult to differentiate.

On the basis of an overflow of plateau ice the development of niche glaciers could be reconstructed as follows:

• (1) The growth of small plateau ice caps which move out radially to spill over the plateau edges.

• (2) Great thicknesses of ice in the overspill avalanche down the slope leaving a thin ice tongue clinging to the steep slope and maintaining its connexion with the plateau ice.

• (3) Melt water from both the plateau and overspill ice will gully the slope below the ice lobe and form a fan at its foot.

• (4) With time nivation processes will hollow out the cliff beneath the ice tongue, and furthermore, differential weathering in a region of near horizontal strata will etch out small resistance differences in the rock strata. The more resistant bands will stand out as less weathered and steeper slope zones, whereas the less resistant bands will have less steep slopes upon which névé and ice will accumulate more readily. Where slopes are less steep the accumulated snow and ice patches will sap at the cliff behind and will enable the niche glacier ice to accumulate, quite apart from the plateau ice and so to develop a convex profile.

• (5) If nivation continues and the niche glacier grows, the hollowing out of the slope will become more pronounced, and a small moraine will develop at the toe of the glacier when ice movement begins to take place.

• (6) Development will diminish with, for example, increased ablation thinning the plateau ice. If this continues sufficiently far the glacier may become detached from its parent ice, in which case its subsequent life will be short unless it is able to maintain itself by direct precipitation.

• (7) With the final melting of the niche glacier the mark of their occupance is left in funnel-shaped hollows in the hillside, with gullies and fans below them. Melt water from seasonal snow accumulation in the hollows and also from the plateau may cause subsequent modification of the funnel shape and by erosion convert them to fluvial gully forms.

It should be noted that the distinctive funnel shape of the hollows would be a function of the combined action of longitudinal erosion by snow, ice and melt water, with lateral widening from weathering along less resistant strata.

This hypothesis of plateau ice overflow makes the initiation of ice masses on slopes as steep as 42 degrees conceivable. Once an initial mass of ice is established it is suggested that processes associated with the niche glacier ice lead to the development of a hollow in the hillside. In this way these glaciers could not be considered true wall-sided glaciers in the sense of Wright and Priestley, but this difference could be due to the more active melt water processes in Bünsow Land.

In testing this hypothesis the niche glaciers of Bünsow Land were examined and it was seen that in many cases the upper margin of the niche glacier was connected to the plateau ice above. Where there was no such connexion, or where the connexion was along only part of the upper ice margin the niche glacier was waning and the impression gained was that the supply of ice had been largely cut off with severance from the ice cap. However, the serious objection to this hypothesis is that niche glaciers were observed on some slopes above which there was a ridge too narrow to support plateau ice. Thus, although some of these glaciers may originate from the overflow of plateau ice this is not likely to be a universal process.

(b) Development of Snow patches

Alternative hypotheses for the development of niche glaciers may be based upon the accumulation of snow, but the problem involved here is whether sufficient snow could accumulate on slopes as steep as 42 degrees. R. D. Salisbury, an early worker in this field, described glaciers developed upon the face of cliffs in Greenland as cliff glaciers.Reference Salsbury ¹¹ His photographs show these to be the same as the Bünsow Land niche glaciers lying in similar funnel-shaped hollows, and he suggests that they originated as snow drifts lodging against the cliff at the junction of the steep cliff face with the scree slope below. Salisbury also stated that there was no connexion between the top of the cliff glacier and the plateau ice. Salisbury made no mention of the processes that might be involved in the development or modification of the hollows in which the cliff glaciers lay as he was concerned primarily with the form of the glaciers.

The idea that erosion could take place beneath snow-patches was developed by W. V. Lewis in 1939.Reference Lewis ¹² The mechanism he invoked was the action of freeze and thaw and the removal of material from beneath the snow patches by water trickling underneath the snow, resulting in the formation of nivation hollows. Lewis considered that circular snow-patches gave rise to corries, but he also recognized longitudinal and transverse snow-patches. His longitudinal patches occupied hollows in the hillside which are very like the funnel-shaped hollows associated with the niche glaciers in Bünsow Land. Lewis did not account for the development of different shaped snow-patches, but a reasonable assumption is that the shape of initial niches in the hillside would be the controlling factor. The question then arises as to whether the longitudinal snow-patches are lying in funnel-shaped hollows which were initiated by niche glacier development following the overspill of the plateau ice; or whether the hollows were formed by niche glaciers which had developed from earlier snow-patches.

Following Lewis, L. H. McCabeReference McCabe ¹³ studied snow-patches in Bünsow Land but his conclusions do nothing to resolve the problem of the origin of niche glaciers. He regarded the funnel-shaped hollows as incipient corries, “linking the original snow gullies with a fully matured corrie”, and suggested that the longitudinal snow gullies were cut back above resistant bands of rock. The coalescence of several such gullies would then give rise to incipient corries in which banks of ice would accumulate to rest on the bench formed by more resistant strata, and so develop into small corrie glaciers. McCabe does not appear to recognize the development of the niche glacier either as a form distinct from the corrie glacier, or as an intermediate stage, neither does he make any observations concerning the initiation of the niche glacier hollows; nevertheless, these ideas can form the basis of a good working hypothesis for their formation.

(c) Niche Glaciers and Hillside Gullies

The very gentle dip of the rock bedding planes in Bünsow Land together with the marked lithological variations gives rise to structural benches and distinct changes of slope which run nearly horizontally along the sides of valleys and fjords. A free rock face is commonly exposed on the upper part of the slope, while scree accumulations of varying extents mantle the lower slopes. These features make the general aspect of the landscape very like tropical arid and semi-arid landscapes of similar structures. Like their tropical counterparts the upper valley slopes of Bünsow Land are notched by marked incisions or gullies. It is suggested here that the funnel-shaped hollows of Biinsow Land are developed from water-worn gullies.

Water-worn gullies in structurally controlled slopes are narrow and very steep-sided where they cut through resistant rock strata and widen out with less steep sides on the less resistant rocks. This gives rise to many ledges in restricted gully channels on which snow may accumulate, and a névé patch thus established will, as it develops to a glacier, bring about the rounding of the gully hollow by nivation processes. This hollow will be enlarged both by the growth of the glacier and also by the coalescence of adjacent hollows as they develop. Growth will also lead to the extension of the ice higher up the slope above the initial ledge, while continued nivation may cause the steepening of the upper parts of the slopes upon which the niche glacier lies. In this way the very steep slopes of some niche glaciers no longer present a problem which makes their initiation from snow-patch accumulation difficult as it is seen to be a secondary feature arising as the glacier cuts back into the hillside.

The resulting funnel-shaped hollow, if exposed by the melting of the glacier, becomes subject to incision by running water which usually derives from snow melt. Several such gullies may dissect a large hollow and may in their turn serve as sites for subsequent snow accumulation. In this way the initially large, well-developed niche glacier hollows may experience a second cycle of development.

While variation in lithology is necessary for the development of the breaks of slope on which niche glaciers originate, lithology also controls the shape of the hollow which subsequently develops. Weaker rocks give rise to open funnel-shaped hollows (Fig. 5a) and the slopes between neighbouring glaciers are not prominent. The upper slopes of the niche glacier hollow in this type of rock are usually greater than the slopes of the hillside into which the hollow is cut, and it is here that the advance of niche glacier development leads more quickly to the coalescence of neighbouring hollows. Strong rocks which support steep cliff faces give rise to steep-sided funnel-shaped hollows (Fig. 5b). The nivation processes cut back more rapidly than the processes of weathering that attack the cliff face and so steep buttresses are maintained between neighbouring hollows. These buttresses have a gradient that is frequently greater than that of the rear wall of the hollows.

Fig. 5 (a).

The characteristic form of niche glacier hollows developed in weak rocks

Fig. 5 (b).

The characteristic form of niche glacier hollows developed in more resistant rocks

Niche glacier hollows therefore show variations in form which reflect both their development and the rock type, but despite this their form is quite distinctive. The rounded halffunnel-shaped hollow may be considered a landform characteristic of niche glacier development.