Chapter XVI: Glacial Features (2)

A third fact from regions once glaciated but now snow-free also combined with the two preceding facts in weakening the wholesale application of Johnson’s hypothesis. In many headwater basins the cirque whose wall at a distance seemed a unit was really broken into two unequal portions; a lower, much grooved and rounded portion and an upper unglaciated, steep-walled portion. This condition was most puzzling in view of the accepted explanation of cirque formation, and it was not until the two first-named facts and the applications of the curves of snow motion were noted that the meaning of the break on the cirque became clear. Referring to Fig. 198 we see at once that the break occurs at _y_ and means that under favorable topographic and geologic conditions sapping at _y_ takes place faster than at _x_ and that the retreat of _y-z_ is faster than _x-y_. It will be clear that when these conditions are reversed or sapping at _x_ and at _y_ are equal a single wall will result. On reference to the literature I find that Gilbert recently noted this feature and called it the _schrundline_.[63] He believes that it marks the base of the bergschrund _at a late stage in the excavation of the cirque basin_. He notes further that the lower less-steep slope is glacially scoured and that it forms “a sort of shoulder or terrace.”

If all the structural and topographic conditions were known in a great variety of gathering basins we should undoubtedly find in them, and not in special forms of ice erosion, an explanation of the various forms assumed by cirques. The limitations inherent in a high-altitude field and a limited snow cover prevented me from solving the problem, but it offered sufficient evidence at least to indicate the probable lines of approach to a solution. For example it is noteworthy that in _all_ the cases examined the schrundline was better developed the further glacial erosion had advanced. So constantly did this generalization check up, that if at a distance a short valley was observed to end in a cirque, I knew at once and long before I came to the valley head that a shoulder below the schrundline did not exist. At the time this observation was made its significance was a mystery, but it represents a condition so constant that it forms one of the striking features of the glacial forms in the headwater region.

The meaning of this feature is represented in Fig. 199, in which three successive stages in cirque development are shown. In _A_, as displayed in small valleys or mountainside alcoves which were but temporarily occupied by snow and ice, or as in all higher valleys during the earlier stages of the advancing hemicycle of glaciation, snow collects, a short glacier forms, and a bergschrund develops. As a result of the concentrated frost action at the base of the bergschrund a rapid deepening and steepening takes place at _a_. As long as the depth of snow (or snow and névé) is slight the bergschrund may remain open. But its existence at this particular point is endangered as the cirque grows, since the increasing steepness of the slope results in more rapid snow movement. Greater depth of snow goes hand in hand with increasing steepness and thus favors the formation of névé and even ice at the bottom of the moving mass and a constantly accelerated rate of motion. At the same time the bergschrund should appear higher up for an independent reason, namely, that it tends to form between a mass of slight movement and one of greater movement, which change of function, as already pointed out, would appear to be controlled by change from snow to névé or ice on the part of the bottom material.

The first stages in the upward migration of the bergschrund will not effect a marked change from the original profile, since the converging slopes, the great thickness of névé and ice at this point, and the steep gradient all favor powerful erosion. When, however, stage _C_ is reached, and the bergschrund has retreated to _c″_, a broader terrace results below the schrundline, the gradient is decreased, the ice and névé (since they represent a constant discharge) are spread over a greater area, hence are thinner, and we have the cirque taking on a compound character with a lower, less steep and an upper, precipitous section.

It is clear that a closely jointed and fragile rock might be quarried by moving ice at _c′-c″_ and the cirque wall extended unbroken to _x_; it is equally clear that a homogeneous, unjointed granite would offer no opportunities for glacial plucking and would powerfully resist the much slower process of abrasion. Thus Gilbert[64] observed the schrundline in the granites of the Sierra Nevada, which are “in large part structureless” and my own observations show the schrundline well developed in the open-jointed granites of the Cordillera Vilcapampa and wholly absent in the volcanoes of the Maritime Cordillera, where ashes and cinders, the late products of volcanic action, form the easily eroded walls of the steep cones. Somewhere between these extremes--lack of a variety of observations prevents our saying where--the resistance and the internal structure of the rock will just permit a cirque wall to extend from _x_ to _c′ ″_ of Fig. 199.

A common feature of cirques that finds an explanation in the proposed hypothesis is the notch that commonly occurs at some point where a convergence of slopes above the main cirque wall concentrates snow discharge. It is proposed to call this type the notched cirque. It is highly significant that these notches are commonly marked by even steeper descents at the point of discharge into the main cirque than the remaining portion of the cirque wall, even when the discharge was from a very small basin and in the form of snow or at the most névé. The excess of discharge at a point on the basin rim ought to produce the form we find there under the conditions of snow motion outlined in earlier paragraphs. It is also noteworthy that it is at such a point of concentrated discharge that crevasses no sooner open than they are closed by the advancing snow masses. To my mind the whole action is eminently representative of the action taking place elsewhere along the cirque wall on a smaller scale.

What seems a good test of the explanation of cirques here proposed was made in those localities in the Maritime Cordillera, where large snowbanks but not glaciers affect the form of the catchment basins. A typical case is shown in Fig. 201. As in many other cases we have here a great lava plateau broken frequently by volcanic cones of variable composition. Some are of lava, others consist of ashes, still others of tuff and lava and ashes. At lower elevations on the east, as at 16,000 feet between Antabamba and Huancarama, evidences of long and powerful glaciers are both numerous and convincing. But as we rise still higher the glaciated topography is buried progressively deeper under the varying products of volcanic action, until finally at the summit of the lava fields all evidences of glaciation disappear in the greater part of the country between Huancarama and the main divide. Nevertheless, the summit forms are in many cases as significantly altered as if they had been molded by ice. Precipitous cirque walls surround a snow-filled amphitheater, and the process of deepening goes forward under one’s eyes. No moraines block the basin outlets, no U-shaped valleys lead forward from them. We have here to do with post-glacial action pure and simple, the volcanoes having been formed since the close of the Pleistocene.

Likewise in the pass on the main divide, the perpetual snow has begun the recessing of the very recent volcanoes bordering the pass. The products of snow action, muds and sands up to very coarse gravel, glaciated in texture with an intermingling of blocks up to six inches in diameter in the steeper places, are collected into considerable masses at the snowline, where they form broad sheets of waste so boggy as to be impassable except by carefully selected routes. No ice action whatever is visible below the snowline and the snow itself, though wet and compact, is not underlain by ice. Yet the process of hollowing goes forward visibly and in time will produce serrate forms. In neither case is there the faintest sign of a bergschrund; the gradients seem so well adjusted to the thickness and rate of movement of the snow from point to point that the marginal crack found in many snowfields is absent.

The absence of bergschrunds is also noteworthy in many localities where formerly glaciation took place. This is notoriously the case in the summit zone of the Cordillera Vilcapampa, where the accumulating snows of the steep cirque walls tumble down hundreds of feet to gather into prodigious snowbanks or to form névé fields or glaciers. From the converging walls the snowfalls keep up an intermittent bombardment of the lower central snow masses. It is safe to say that if by magic a bergschrund could be opened on the instant, it would be closed almost immediately by the impetus supplied by the falling snow masses. The explanation appears to be that the thicker snow and névé concentrated at the bottom of the cirque results in a corresponding concentration of action and effect; and cirque development goes on without reference to a bergschrund. The chief attraction of the bergschrund hypothesis lies in the concentration of action at the foot of the cirque wall. But in the thickening of the snow far beyond the minimum thickness required for motion at the base of the cirque wall and its change of function with transformation into névé, we need invoke no other agent. If a bergschrund forms, its action may take place at the foot of the cirque wall or high up on the wall, and yet _sapping at the foot of the wall_ continue.

HIRAM BINGHAM DIRECTOR

LAMBRAMA QUADRANGLE]

From which we conclude (1) that where frost action occurs at the bottom of a bergschrund opening to the foot of the cirque wall it aids in the retreat of the wall; (2) that a sapping action takes place at this point whether or not a bergschrund exists and that bergschrund action is not a _necessary_ part of cirque formation; (3) that when a more or less persistent bergschrund opens on the cirque wall above its foot it tends to develop a schrundline with a marked terrace below it; (4) that schrundlines are best developed in the mature stages of topographic development in the glacial cycle; (5) that the varying rates of snow, névé, and ice motion at a valley head are the _persistent_ features to which we must look for topographic variations; (6) that the hypothesis here proposed is applicable to all cases whether they involve the presence of snow or névé or ice or any combination of these, and whether bergschrunds are present or not; and (7) at the same time affords a reasonable explanation for such variations in forms as the compound cirque with its schrundline and terrace, the unbroken cirque wall, the notched cirque, and the recessed, snow-covered mountain slopes unaffected by ice.

ASYMMETRICAL CREST LINES AND ABNORMAL VALLEY PROFILES IN THE CENTRAL ANDES

To prove that under similar conditions glacial erosion may be greater than subaërial denudation quantitative terms must be sought. Only these will carry conviction to the minds of many opponents of the theory that ice is a vigorous agent of erosion. Gilbert first showed in the Sierra Nevada that headwater glaciers eroded more rapidly than nonglacial agents under comparable topographic and structural conditions.[65] Oddly enough none of the supporters of opposing theories have replied to his arguments; instead they have sought evidence from other regions to show that ice cannot erode rock to an important degree. In this chapter evidence from the Central Andes, obtained in 1907 and 1911, will be given to show the correctness of Gilbert’s proposition.

The data will be more easily understood if Gilbert’s argument is first outlined. On the lower slopes of the glaciated Sierra Nevada asymmetry of form resulted from the presence of ice on one side of each ridge and its absence on the other (Fig. 200). The glaciers of these lower ridges were the feeblest in the entire region and were formed on slopes of small extent; they were also short-lived, since they could have existed only when glacial conditions had reached a maximum. Let the broken line in the upper part of the figure represent the preglacial surface and the solid line beneath it the present surface. It will not matter what value we give the space between the two lines on the left to express nonglacial erosion, since had there been no glaciers it would be the same on both sides of the ridge. The feeble glacier occupying the right-hand slope was able in a very brief period to erode a depression far deeper than the normal agents of denudation were able to erode in a much longer period, i.e., during all of interglacial and postglacial time. Gilbert concludes: “The visible ice-made hollows, therefore, represent the local excess of glacial over nonglacial conditions.”

In the Central Andes are many volcanic peaks and ridges formed since the last glacial epoch and upon them a remarkable asymmetry has been developed. Looking southward one may see a smoothly curved, snow-free, northward-facing slope rising to a crest line which appears as regular as the slope leading to it. Looking northward one may see by contrast (Fig. 194) sharp ridges, whose lower crests are serrate, separated by deeply recessed, snow-filled mountain hollows. Below this highly dissected zone the slopes are smooth. The smooth slope represents the work of water; the irregular slopes are the work of snow and ice. The relation of the north and south slopes is diagrammatically shown in Fig. 201.

To demonstrate the erosive effects of snow and ice it must be shown: (1) that the initial slopes of the volcanoes are of postglacial age; (2) that the asymmetry is not structural; (3) that the snow-free slopes have not had special protection, as through a more abundant plant cover, more favorable soil texture, or otherwise.

Proof of the postglacial origin of the volcanoes studied in this connection is afforded: (1) by the relation of the flows and the ash and cinder beds about the bases of the cones to the glacial topography; (2) by the complete absence of glacial phenomena below the present snowline. Ascending a marginal valley (Fig. 202), one comes to its head, where two tributaries, with hanging relations to the main valley, come down from a maze of lesser valleys and irregular slopes. Glacial features of a familiar sort are everywhere in evidence until we come to the valley heads. Cirques, reversed grades, lakes, and striæ are on every hand. But at altitudes above 17,200 feet, recent volcanic deposits have over large areas entirely obscured the older glacial topography. The glacier which occupied the valley of Fig. 202 was more than one-quarter of a mile wide, the visible portion of its valley is now over six miles long, but the extreme head of its left-hand tributary is so concealed by volcanic material that the original length of the glacier cannot be determined. It was at least ten miles long. From this point southward to the border of the Maritime Cordillera no evidence of past glaciation was observed, save at Solimana and Coropuna, where slight changes in the positions of the glaciers have resulted in the development of terminal moraines a little below the present limits of the ice.

From the wide distribution of glacial features along the northeastern border of the Maritime Cordillera and the general absence of such features in the higher country farther south, it is concluded that the last stages of volcanic activity were completed in postglacial time. It is equally certain, however, that the earlier and greater part of the volcanic material was ejected before glaciation set in, as shown by the great depth of the canyons (over 5,000 feet) cut into the lava flows, as contrasted with the relatively slight filling of coarse material which was accumulated on their floors in the glacial period and is now in process of dissection. Physiographic studies throughout the Central Andes demonstrate both the general distribution of this fill and its glacial origin.

So recent are some of the smaller peaks set upon the lava plateau that forms the greater part of the Maritime Cordillera, that the snows massed on their shadier slopes have not yet effected any important topographic changes. The symmetrical peaks of this class are in a few cases so very recent that they are entirely uneroded. Lava flows and beds of tuff appear to have originated but yesterday, and shallow lava-dammed lakes retain their original shore relations. In a few places an older topography, glacially modified, may still be seen showing through a veneer of recent ash and cinder deposits, clear evidence that the loftier parts of the lava plateau were glaciated before the last volcanic eruption.

The asymmetry of the peaks and ridges in the Maritime Cordillera cannot be ascribed to the manner of eruption, since the contrast in declivity and form is persistently between northern and southern slopes. Strong and persistent winds from a given direction undoubtedly influence the form of volcanoes to at least a perceptible degree. In the case in hand the ejectamenta are ashes, cinders, and the like, which are blown into the air and have at least a small component of motion down the wind during both their ascent and descent. The _prevailing_ winds of the high plateaus are, however, easterly and the strongest winds are from the west and blow daily, generally in the late afternoon. Both wind directions are at right angles to the line of asymmetry, and we must, therefore, rule out the winds as a factor in effecting the slope contrasts which these mountains display.

It remains to be seen what influence a covering of vegetation on the northern slopes might have in protecting them from erosion. The northern slopes in this latitude (14° S.) receive a much greater quantity of heat than the southern slopes. Above 18,000 feet (5,490 m.) snow occurs on the shady southern slopes, but is at least a thousand feet higher on the northern slopes. It is therefore absent from the northern side of all but the highest peaks. Thus vegetation on the northern slopes is not limited by snow. Bunch grass--the characteristic _ichu_ of the mountain shepherds--scattered spears of smaller grasses, large ground mosses called _yareta_, and lichens extend to the snowline. This vegetation, however, is so scattered and thin above 17,500 feet (5,330 m.) that it exercises no retarding influence on the run-off. Far more important is the porous nature of the volcanic material, which allows the rainfall to be absorbed rapidly and to appear in springs on the lower slopes, where sheets of lava direct it to the surface.

The asymmetry of the north and south slopes is not, then, the result of preglacial erosion, of structural conditions, or of special protection of the northern slopes from erosion. It must be concluded, therefore, that it is due to the only remaining factor--snow distribution. The southern slopes are snow-clad, the northern are snow-free--in harmony with the line of asymmetry. The distribution of the snow is due to the contrasts between shade and sun temperatures, which find their best expression in high altitudes and on single peaks of small extent. Frankland’s observations with a black-bulb thermometer _in vacuo_ show an increase in shade and sun temperatures contrasts of over 40° between sea level and an elevation of 10,000 feet. Violle’s experiments show an increase of 26 per cent in the intensity of solar radiation between 200 feet and 16,000 feet elevation. Many other observations up to 16,000 feet show a rapid increase in the difference between sun and shade temperatures with increasing elevation. In the region herein described where the snowline is between 18,000 and 19,000 feet (5,490 to 5,790 m.) these contrasts are still further heightened, especially since the semi-arid climate and the consequent long duration of sunshine and low relative humidity afford the fullest play to the contrasting forces. The coefficient of absorption of radiant energy by water vapor is 1,900 times that of air, hence the lower the humidity the more the radiant energy expended upon the exposed surface and the greater the sun and shade contrasts. The effect of these temperature contrasts is seen in a canting of the snowline on individual volcanoes amounting to 1,500 feet in extreme instances. The average may be placed at 1,000 feet.

The minimum conditions of snow motion and the bearing of the conclusions upon the formation of cirques have been described in the chapters immediately preceding. It is concluded that snow moves upon 20° slopes if the snow is at least forty feet deep, and that through its motion under more favorable conditions of greater depth and gradient and the indirect effects of border melting there is developed a hollow occupied by the snow. Actual ice is not considered to be a necessary condition of either movement or erosion. We may at once accept the conclusion that the invariable association of the cirques and steepened profiles with snowfields proves that snow is the predominant modifying agent.

An argument for glacial erosion based on profiles and steep cirque walls in a volcanic region has peculiar appropriateness in view of the well-known symmetrical form of the typical volcano. Instead of varied forms in a region of complex structure long eroded before the appearance of the ice, we have here simple forms which immediately after their development were occupied by snow. _Ever since their completion these cones have been eroded by snow on one side and by water on the other._ If snow cannot move and if it protects the surface it covers, then this surface should be uneroded. All such surfaces should stand higher than the slopes on the opposite aspect eroded by water. But these assumptions are contrary to fact. The slopes underneath the snow are deeply recessed; so deeply eroded indeed, that they are bordered by steep cliffs or cirque walls. The products of erosion also are to some extent displayed about the border of the snow cover. In strong contrast the snow-free slopes are so slightly modified that little of their original symmetry is lost--only a few low hills and shallow valleys have been formed.

The measure of the excess of snow erosion over water erosion is therefore the difference between a northern or water-formed and a southern or snow-formed profile, Fig. 200. This difference is also shown in Fig. 201 and from it and the restored initial profiles we conclude that the rate of water erosion is to that of nivation as 1:3. This ratio has been derived from numerous observations on cones so recently formed that the interfluves without question are still intact.

Thus far only those volcanoes have been considered which have been modified by nivation. There are, however, many volcanoes which have been eroded by ice as well as by snow and water. It will be seen at once that where a great area of snow is tributary to a single valley, the snow becomes compacted into névé and ice, and that it then erodes at a much faster rate. Also a new force--plucking--is called into action when ice is present, and this greatly accelerates the rate of erosion. While it lies outside the limits of my subject to determine quantitatively the ratio between water and ice action, it is worth pointing out that by this method a ratio much in excess of 1:3 is determined, which even in this rough form is of considerable interest in view of the arguments based on the protecting influence of both ice and snow. I have, indeed, avoided the question of ice erosion up to this point and limited myself to those volcanoes which have been modified by nivation only, since the result is more striking in view of the all but general absence of data relating to this form of erosion.

a′-b′-c = preglacial profile.

a-b-d-c = present profile.

b′-d-c-b = total ice erosion in the glacial period.

a-b = surface of an alluvial valley fill due to
excessive erosion at valley head.

b-b′ = terminal moraine.

d-c = cirque wall.

e, e′ e″ = asymmetrical summits.]

If we now turn to the valley profiles of the glaciated portions of the Peruvian Andes, we shall see the excess of ice over water erosion expressed in a manner equally convincing. To a thoughtful person it is one of the most remarkable features of any glaciated region that the flattest profiles, the marshiest valley flats, and the most strongly meandering stretches of the streams should occur near the heads of the valleys. The mountain shepherds recognize this condition and drive their flocks up from the warmer valley into the mountain recesses, confident that both distance and elevation will be offset by the extensive pastures of the finest _ichu_ grass. Indeed, to be near the grazing grounds of sheep and llamas which are their principal means of subsistence, the Indians have built their huts at the extraordinarily lofty elevations of 16,000 to 17,000 feet.

An examination of a large number of these valleys and the plotting of their gradients discloses the striking fact that the heads of the valleys were deeply sunk into the mountains. It is thus possible by restoring the preglacial profiles to measure with considerable certainty the excess of ice over water erosion.

The results are graphically expressed in Fig. 202. It will be seen that until glacial conditions intervened the stream was flowing on a rock floor. During the whole of glacial time it was aggrading its rock floor below _b′_ and forming a deep valley fill. A return to warmer and drier conditions led to the dissection of the fill and this is now in progress. The stream has not yet reached its preglacial profile, but it has almost reached it. We may, therefore, say that the preglacial valley profile below _b′_ fixes the position of the present profile just as surely as if the stream had been magically halted in its work at the beginning of the period of glaciation. There, _b′-d-c-b_ represents the amount of ice erosion. To be sure the line _b-c_ is inference, but it is reasonable inference and, whatever position is assigned to it, it cannot be coincident with _b′-d_, nor can it be anywhere near it. The break in the valley profile at _b′_ is always marked by a terminal moraine, regardless of the character of the rock. This is not an accidental but a causal association. It proves the power of the ice to erode. In glacial times it eroded the quantity _b-c-d-b′_. This is not an excess of ice over water erosion, but an absolute measure of ice erosion, since _a′-b′_ has remained intact. The only possible error arises from the position assigned _b-c_, and even if we lower it to _b-c′_ (for which we have no warrant but extreme conservatism) we shall still have left _b′-c′-d-b_ as a striking value for rock erosion (plucking and abrasion) by a valley glacier.

A larger diagram, Fig. 203, represents in fuller detail the topographic history of the Andes of southern Peru and the relative importance of glaciation. The broad spurs with grass-covered tops that end in steep scarps are in wonderful contrast to the serrate profiles and truncated spurs that lie within the zone of past glaciation. In the one case we have minute irregularities on a canyon wall of great dimensions; in the other, more even walls that define a glacial trough with a flat floor. Before glaciation on a larger scale had set in the right-hand section of the diagram had a greater relief. It was a residual portion of the mountain and therefore had greater height also. Glaciers formed upon it in the Ice Age and glaciation intensified the contrast between it and the left-hand section; not so much by intensifying the relief as by diversifying the topographic forms.