American Alpine Jounrna and Accidents in North American Climbing

American Alpine Club Research Fund, 13. Dirt Bands

  • Feature Article
  • Climb Year:
  • Publication Year: 1949

American Alpine Club Research Fund


Joel E. Fisher

ALLEY glaciers exhibit, in occasional instances, unique cross- glacier concentric parabolic markings, termed “dirt bands.” Figures 1 to 9 show examples of these beautiful and striking phenomena. Early glaciologists studied them and advanced theories as to their origin.1 Best known of these earlier writers was Principal Forbes, from whose name all such concentric bands have been generally referred to in English as Forbes bands—the Ogives of Continental writers. It will be the attempt of this paper to show that there are two types of such bands, very similar to each other in superficial appearance, but totally different in origin. These types are called, in this paper, Forbes and Alaskan bands. Further, an attempt will be made to explain the origin of these two types.

Bradford Washburn has called attention to the fact that Forbes bands appear only on glaciers on which a substantial icefall exists farther up, and proposed the theory that such icefalls are prerequisites to banding.2 He suggests that, in such an icefall, as shown in Figures 11 and 12, substantial glacier-wide blocks of ice ride down through the icefall, unsplintered; and that accumulations of dust occur in the intervening hollows and eventually provide, at the outrun below the icefall, alternations of dark ice from dirty icefall areas and of light ice from clear ones.

The writer has for some years been making a study of such bands, in the field as well as by photographs. His basic conclusion is that the contrasting banding is noticeable primarily on account of certain alternating differences in the texture of the ice, rather than on account of local concentrations of dust. His conclusions agree with Washburn’s as to the icefall’s being the mechanical spacer of Forbes bands (in so far as true Forbes bands are con- corned), but not as to the surface dust’s concentrating in the hollows and becoming the coloring agent. With respect to the other type of bands, the Alaskan, the writer has quite different conclusions, conclusions that provide grounds for agreement with Vareschi’s studies of pollen distribution.3

Distinguishing Forbes from Alaskan bands, one may note that

In Forbes bands:

The dark band is wide—outstandingly wide.

Its upper and lower boundaries are clean-cut.

In cutting across a moraine, the light band may be clean of debris, while the dark area is loaded with debris.

The span of each pair of bands is great (as much, often, as 300 feet), representing one full year’s flow of the glacier, and remains constant, unless the glacier’s velocity is reduced by fanning out, etc.

The banding tends to disappear around the receding flanks of the glacier and, on occasion, on approaching the toe.

In Alaskan bands:

The dark band is much narrower than the white— always, a narrow-ribbon.

Its upper boundary will be clean-cut, but its down- glacier boundary is often hazy.

Medial moraines will be pretty constant in width, over both white and dark segments, if originating far up- glacier.

The span of each pair of bands may be as short as 50 feet, and may well become shorter and shorter, near the end of the glacier, even if the glacier does not fan out.

The banding becomes more distinct, if anything, around the receding flanks of a glacier, and near its toe.

A comparison of Figures 1 to 6 (Forbes bands) with Figures 7 to 9 (Alaskan bands) will clarify these comparisons.

Forbes Bands

Forbes bands will be examined first.

Washburn’s glacier-wide islands of ice, riding down through icefalls, are well shown in Figures 11 and 12. Figure 2, particularly, shows that the outflowing dark areas occupy hummocks of ice, the clean ice being in the hollows. This relationship, clean hollow vs. dirty hummock, is eventually reversed down-glacier, because clean ice suffers less surface melting by radiation than dirty ice. It is possible, in several illustrations, to identify a continuity of these hummocks of the outflow with the “islands” in the icefall, although surface ice debris on the lower part of every icefall partly smothers under its froth any true structural formations of ice of the glacier proper. Even so, the existence of multitudes of crevasses in every icefall makes it clear that the flow of ice down the icefall must be by simple gravitational flow, not by any extrusion continuing through the entire icefall from above—such that there can be no great pressure bearing down on the outrun of ice, as is imagined by some writers.4

The essence of this paper is the proposal that the dark bands are the lineal descendants of the above-mentioned glacier-wide islands of ice riding down through the icefalls, while the white bands are the descendants of the inlays of snow and of ice debris, in the crevasses separating those islands of ice in the icefalls.

Granted gravity flow, the looser “aerated” accumulation of powder snow and finer ice debris, within the glacier-wide crevasses separating glacier-wide islands in the icefall, is not subject to any great length-wise compression, especially near the top of the icefall, whence its firnification dates. Rather, such inlays of fresher snow and comminuted ice are years in arrears as to their state of firnification (a process accelerated by compression, as well as by other agencies), compared to the matrix of ice, in the outrun, which came down intact, from the upper névé, through the icefall, in the shape of those glacier-wide islands of ice. As a test of this inlay theory, the writer arranged in 1946 for a series of piles of bright- colored beads to be strewn above the icefall of the Mer de Glace.

Ten years hence, when this locality has travelled down through the icefall, relative absence of beads in the white bands would prove this proposal.5

It is this difference in degree of firnification which primarily accounts for the darker and lighter bands below the icefall. Not only is the less completely firnified ice, with its myriads of air bubbles, intrinsically whiter to the eye, but its surface appears more dust-free, to the observer, for this remarkable reason: It is a peculiarity of glacier ice that dust and fine dirt on its surface will bore dust holes quickly to a depth of several inches into “aerated” ice, but not at all into hard, non-aerated ice. Figure 10 illustrates this. Presumably, this selectivity would be due to the greater albedo of the whiter (aerated) ice, whereby ablation by radiant heat is minimized for the whiter ice itself—concentrated entirely underneath each particle of implanted absorbent dirt so that those dust particles rapidly melt the ice beneath, when it is “white” ice on which they lie.6

On darker ice, heat is absorbed just about as fast by the ice itself as by particles of dirt thereon, so no such fillip is given to any boring process. Result: all surface dust on darker ice remains on the surface, visible to the observer.

The final result is that those inlays of fresher snow and of slithers of ice, laid down in the intervening crevasses of the icefall, ride out below, not only whiter in themselves, but with any dust which may have accumulated on their candid surfaces now hidden, in so far as the eye of any observer is concerned, at the bottoms of the myriads of dust wells which honeycomb its surface. Result: these inlays appear exceptionally white to the eye.

The following diagram illustrates the process, as envisioned by the writer:

A, B, C, D, E represent ice debris and drifting snow which fill up the crevasses between the glacier-wide blocks of névé ice (1, 2, 3, 4) which calve off the upper plateau each year. A is coarse debris; B, the second year, becomes more pulverized and more compact; C, even more so; D and E are well on the way to refirnification, although each inlay always lags many years behind the adjacent unsplintered islands of original névé, in degree of firnification. Always, these inlays confine substantial air bubbles.

Bases of island blocks of ice, as at M and N, are under substantial compression, being on the inside of the vertical curve of the profile of the glacier, while the bases of the lower blocks, such as P, Q and R, while by no means under actual tension, are under relatively less compression, being here on the outside of the reversed vertical curve of the profile. It follows7 that, under isothermal conditions of the ice, ice at M, N and O will tend to be soft and yielding, while ice at P, Q and R will be hard, rigid, with greater cutting power—complying exactly with the condition demanded for scouring out the typical below-grade hollow at the base of every such step, in the profiles of glaciated valleys. In the meantime, being relieved of much compression at their bases, blocks 4, 5 and 6 allow the inlay between the intervening crevasses, E and F, to work down to the rock bed, so that from 6 on down every island block of ice is separated from its neighbor by an inlay of ice less completely firnified than in its own body and more aerated, extending right down to the bed, e.g. G, H, J, K, L, M. Along the surface of the icefall, ice blocks 5, 6 and 7 may well be smothered under a froth of excess ice debris, but once that froth vanishes, by surface ablation (as at 9), one will observe below such an icefall a succession of alternately whiter and darker bands on the surface of the glacier, each inlay, as J, K, L, M, being the whiter, more aerated ice, and each ancient block, 9, 10, 11, 12, the more completely firnified, and therefore the darker ice. Some dirt accumulated on the surface, above A, may even survive on surfaces 9, 10, 11 and 12, giving them a still darker appearance.

The mechanism of major crevassing in icefalls differs from that of crevassing elsewhere on a glacier, in that icefall crevasses remain open until filled in by surface snow and debris, while crevasses elsewhere on a glacier normally close up by contraction, long before surface debris fills them up. No scars of white inlays, therefore, survive to mark such other crevasses, in contrast with the white inlays marking icefall crevasses.

Alaskan Rands

As to Alaskan bands, any icefall theory cannot prevail. As Figures 7 and 9 will show, they are not preceded by such typical icefalls. Further, we have the interesting studies of Vareschi: on the Aletschgletscher in the Oberland, and on the Gepatschgletscher of the Ortler district, dark bands of his Ogives carry pollen preponderantly of summer production, while the light bands of the Ogives carry much less pollen and, at that, pollen characteristic of fall and winter. These Aletsch- and Gepatschgletscher Ogives are bands of the true Alaskan type. While Vareschi does not appear to recognize the basic differences between Forbes and Alaskan bands, his studies, actually restricted to the Alaskan type (Aletsch- and Gepatschgletscher bands), bear out the writer’s conclusion that dark bands of the Alaskan type are the outcrops of “fossilized” strata of snow laid down under suitable conditions in summer, when firnification proceeds fastest; and that the white areas are the outcrops of strata of winter snow, laid down in below-freezing conditions, with but little sun. It is quite obvious that snow deposited under summer skies will firnify faster than snow deposited under wintry conditions; and, when the contrast between the two seasons is sufficient (whether because of the typical differences of Arctic summer and winter in high latitudes, or because of physiographic shelter and exposure), then Alaskan bands develop. This relative difference in firnification of alternate strata, in Alaskan banding, produces exactly the same difference in apparent coloration that it produces in Forbes banding. It is notable that Alaskan bands occur chiefly on far northern glaciers where the difference between summer and winter, as to firnifying factors, is greatest, and on such occasional glaciers of temperate regions as happen to develop from névé catchment areas in which summer sun is exceptionally potent and, by the same token, in which winter snow accumulation is largely sunless. For example, the Jungfraufirn and the Ewig Schnee Feld above the Aletsch, and the cwm at the base of the north wall of the Breithorn, are areas outside of the Arctic in which contrasts between summer and winter would be particularly favorable to Alaskan banding, and glaciers flowing therefrom are among the few Alpine glaciers which do display banding of the Alaskan type.

It is not difficult to see how such annual stratification, even where there is only moderate extrusion flow, will intersect the surface, lower down the glacier, spaced much closer together than one single year’s flow of ice. Where there is pronounced extrusion flow, not only will this closer spacing be even more marked, but also the faster flow of the lower levels of a glacier will result in a vertical thinning of those same lower strata; and when those thinner strata eventually intersect the surface, far down glacier, they will be necessarily very close together because, once thinned, they can never thicken again. This process explains why Alaskan banding is often more closely spaced near the toe than farther up glacier. In judging from some vantage point an apparent change in the spacing of bands, one must make careful allowance for perspective, if looking down-glacier. The following sketch shows the mechanism of the formation of Alaskan banding:


Both Forbes and Alaskan bands are then to be considered as owing their appearance to slight differences in the degree of firnification of the ice in the alternate bands, differences which not only produce ice of white or gray hue, but also favor the hiding of dirt in shallow surface dust wells on the ice which is already the whiter.

The processes which produce the two types are, in Forbes banding, a vertical splitting up of the ice, by great cross-glacier crevasses in the icefall, such crevasses being immediately back-filled by fresh surface snow, or ice debris; and, in Alaskan banding, a condition of accumulation conducive to extreme variation between summer and winter texture of snow, resulting in decidedly noticeable annual stratification. There being no pronounced extrusion flow possible below any icefall, Forbes bands must always appear under near-gravity flow, so that the vertical structure (Forbes banding) induced in the icefall retains its near-vertical orientation, all the travel below. There being more or less extrusion flow where there is no icefall, the horizontal strata (Alaskan type) from high up tend to be rotated into vertical strata, deep within the glacier, as the glacier flows; and, when at length these strata intersect the surface of the glacier far down its course, they appear to be nearly vertical, strongly simulating the Forbes design.

1 Louis Agassiz, Études sur 'les glaciers (Neuchâtel, 1840), pp. 97-126; J. D. Forbes, Occasional Papers on the Theory of Glaciers (Edinburgh, 1859), p. 39; William Huber, Les Glaciers (Paris, 1867), pp. 197-200 and 206; I. C. Russell, Glaciers of North America (New York, 1897), p. 43; W. H. Sherzer, “Glacial Studies in the Canadian Rockies and Selkirks,” Smithsonian Miscellaneous Collections (Quarterly Issue), XLVII (1905), 465 ff.

2 “Morainic Bandings of Malaspina and Other Alaskan Glaciers,” Bulletin of the Geological Society of America, 46: 1879-89, 1935.

3 Volkmar Vareschi, Die Pollenanalytische Untersuchung der Gletscherbewegung (Bern, 1942). Since the principles of gravitational and extrusion flow of ice are at times involved, reference is made also to M. Demorest and R. F. Flint, American Journal of Science, 240: 31 and 113, 1942.

4 R. Streiff-Becker, “Beitrag zur Gletscherkunde: Forschungen am Claridenfirn im Kanton Glarus,” Denkschriften der Schweizerischen Naturforschenden Gesellschaft, 75: 111-32, 1943.

5 For map showing exactly where these beads have been planted, see A.A.J., VI (1947), 331. Note, however, that the other experiment described in that place—cleaning the névé surface of the Trift Glacier (Gotthardmassif) —has since been found valueless, inasmuch as surface dust plays an unimportant part in determining the contrasts between dark and light areas.

6 Inasmuch as “aerated” ice is about 10% less dense than normal ice, the same quantity of heat, furthermore, will melt it faster than normal ice. My guide, Felix Julen, under my direction, has measured the density of “aerated” glacier ice as about 0.83, which compares with a density of 0.92 for glacier ice containing relatively few air bubbles.

7 J. E. Fisher, “The Pressure Melting Point of Ice and the Excavation of Cirques and Valley Steps by Glaciers,” A.A.J., VII (April 1048), 67-72.

This AAJ article has been reformatted into HTML. Please contact us if you spot an error.