Sir,

A large literature exists on superglacial sediment, particularly in relation to the development of dirt cones (Reference LewisLewis, 1940; Reference SharpSharp, 1949; Reference SwithinbankSwithinbank, 1950; Reference ListerLister, 1953; Reference PirritPirrit, 1953; Reference Streiff-BeckerStreiff-Becker, 1954; Reference McAllisterMcAllister, 1956; Reference KrenekKrenek, 1958; Reference DrewryDrewry, 1972) but rarely have the grain-size characteristics been described. Figure 1 illustrates the cumulative frequency curves for composite samples of superglacial sediment taken from three localities in the ablation zone of Østerdalsisen, an outlet glaciei of the ice cap Svartisen in Norway (lat. 66° 31′ N., long. 14° 07′ E.). The first sample is from a small dirt-cone field near the south-west margin of the glacier, where the average debris thickness on cone flanks was 10 mm. The largest cone had a basal area of 2.2 m² and a height of 0.45 m, and was elliptical in plan, suggesting that the sediment had been deposited originally in a stream channel. The second is from a melt-water channel near the centre of the glacier, where sediment had accumulated in small holes which pitted the stream bed, and the third from the bottom of a shallow depression which was probably part of a former stream course and in which a dirt cone was in the process of formation. The ice core beneath the sediment was only 0.12 m above the surrounding glacier surface.

Fig. 1.Cumulative frequency curves showing the size of material in a dirt-cone field (sample 1), a fluvial deposit (sample 2) and an incipient dirt cone (sample 3).

The similarity of the curves (Fig. 1) suggests that the dirt in the three localities had a common origin and, together with the morphological characteristics of the deposits, that the sites represented different stages in dirt-cone formation. On a glacier such as Østerdalsisen, where the amount of superglacial dirt is small and the dirt is widely distributed, a means of sediment concentration is required if cones are to develop at all. Although there is general agreement that differential ablation of clean and debriscovered ice is the formative mechanism, different modes of dirt concentration have been proposed. Reference LewisLewis (1940), Reference SharpSharp (1949) and Reference SwithinbankSwithinbank (1950) favoured deposition in melt-water streams and, although three samples cannot provide an adequate test of the hypothesis, the fact that the size-distribution curves for these sites of variable form and location are not significantly different does suggest a similar mode of concentration on Østerdalsisen. Highly localized deposition within superglacial streams could lead to the development of dirt cones even on glaciers with little surface dirt provided that sediment accumulates initially to a sufficient depth.

The material is moderately sorted according to the verbal scale suggested by Reference Folk and WardFolk and Ward (1957) for the inclusive graphic standard deviation, values of the parameter ranging from 0.72 to 0.80. Nevertheless, the distribution curves are skewed towards the finer fraction with approximately 50% of the sediment being in the narrow coarse silt range. Reference DrewryDrewry (1972), in one of the few detailed quantitative studies of dirt cones, has suggested that a threshold grain size of 0.2–0.6 mm exists for dirt-cone inception, below which no cones will develop because of the relatively low liquid limit of finer material. Less than 10% of the sample dirt from Østerdalsisen was coarser than 0.2 mm and yet dirt cones did form, albeit on a small scale. However, the small number of cones seemed to be related more to sediment availability than to sediment size. Although Drewry’s experiments were carried out largely with well-sorted unimodal material, the limited evidence presented here does suggest that dirt cones can develop in finer sediment provided other conditions are favourable. Of these, the amount of melting and the degree of sediment saturation may be particularly important. Experiments with a wide range of sediment mixtures in a variety of micro-meteorological environments are required before critical sediment conditions can be defined with assurance.

14 March 1973