Glacial Erosion

Although we can easily identify individual landforms of glacial erosion it is often much more difficult to quantify the extent to which an entire landscape has been modified or affected by glacial erosion. Haynes (1977) and Riedel et al. (2007) quantified the effects of glacial erosion on the Scottish and the USA/Canadian Cordilleran landscapes respectively by examining the degree of valley connectivity in these areas. Fluvial landscapes tend to produce dendritic drainage patterns with low valley connectivity, whereas erosion by ice sheets will modify drainage patterns, breach watersheds and cut new troughs into the landscape. This will tend to increase the interconnectivity of the valley systems in a landscape. By using topological measures of connectivity, originally developed to study the connectivity of transport networks, these authors compiled maps to show the valley connectivity across the two areas. Connectivity is defined by two indices, a and p. The a index is defined by:

where V is the number of stream junctions, E the number of stream segments and G the number of separate sub-basins. The p index is defined by:

Alpha and beta values are determined from analysis of map data. A perfectly dendritic drainage network has an a value of 0 and a p value < 1. Diagram A shows the proportion of the terrain occupied by glacial valleys in Scotland, and Diagram B shows the connectivity of these valleys indicated by their a values. The maps show that valley connectivity, and therefore ice sheet erosion, is highest in north and west Scotland and lowest in the east and south. This pattern is to be expected since the high rates of accumulation and ablation on the more maritime west coast produce the steep mass balance gradients that favour intense glacial erosion. The lower mass balance gradients of the more continental eastern and southern parts of Scotland mean that these areas would be less likely to experience such intense erosion. This work provides a good example of a simple quantitative method by which we can determine the effects of glacial erosion on a landscape.

Sources: Haynes, V.M. (1977) The modification of valley patterns by ice-sheet activity. Geografiska Annaler, 59 A, 195-207. Riedel, J.L., Haugerud, R.A. and Clague, J.J. (2007) Geomorphology of a cordilleran ice sheet drainage network through breached divides in the North Cascades Mountains of Washington and British Columbia. Geomorphology, 91, 1-18. [Modified from: Haynes (1977) Geografiska Annaler, 59A, figures 3 and 4, p. 109]

guide to the intensity of glacial erosion. If the valley pattern is predominantly dendritic it retains its preglacial fluvial form and the intensity of erosion or duration of glacial erosion must have been low. On the other hand, if the dendritic pattern has been replaced by interconnected troughs then the intensity of glacial erosion has probably been much greater. The modification of valley patterns has been used to estimate the intensity of glacial erosion in Scotland and the North American Cordilleran (Box 6.7).

Fjords are drowned glacial troughs. These deep, often-linear features carved into bedrock represent the effects of glacial erosion in situations where ice flow is confined by topography and channelled along a trough or valley. Fjord landscapes are common in the Arctic and Antarctic, as well as along the maritime fringes of southern South America, northwest Europe and Norway, Canada, Alaska and New Zealand. Individual fjords are generally recognised to be palimpsest features, developed over successive glaciations. Fjords erode rapidly under glacial conditions and their considerable dimensions indicate that they represent significant volumes of rock removal by glacial erosion. Mean Quaternary glacial erosion rates of between 1 and 2 mm yr-1 have been calculated for glacial troughs in western Norway and Scotland. Valley patterns in fjord landscapes have been used to quantify the effects of glacial erosion and it has been suggested that the dimensions and longitudinal profiles of fjords are adjusted to the discharge of ice (Box 6.6).

6.3.3 Cirques

Cirques are large bowls that open downslope and are bounded upslope by a cliff or steep slope known as a headwall. The headwall is usually arcuate in plan and is much steeper than the cirque floor. The floor of the cirque may contain an enclosed rock basin and show evidence of glacial erosion, while the headwall is predominantly formed by glacial quarrying and periglacial freeze-thaw weathering. Cirques are usually created by individual glaciers, although they may drain collectively into larger valley glaciers.

Glacial cirques are found in mountainous terrain subject to local glaciation. They may also occur as part of landscapes of ice-sheet erosion, such as those in Britain (Figure 6.14). Here phases of local glaciation allowed cirques to develop, although their morphology has been modified during periods of more intense ice-sheet glaciation. Most cirques are the product of cumulative erosion during several phases of glaciation.

The precise definition of a cirque varies: for example in three separate studies of Scottish cirques the total number identified in the landscape varied between 347 and 876. Part of the reason for this is that cirques tend to grow in clusters, and individual features may become amalgamated over time to become composite features. For example, one large cirque may actually be a composite of several smaller feeder cirques. This variety of form makes the identification of individual cirques somewhat subjective. However, some of this subjectivity may be removed by using mathematical formulae to describe and classify cirque morphology.

Figure 6.14 Photograph of Cwm Cau, a glacial cirque cut into the mountain Cadair Idris in Snowdonia National Park, North Wales. [Photograph: N.F. Glasser]

In longitudinal profile, the headwall of a cirque is normally much steeper than its floor. The basic profile shape of a cirque therefore can be described by a logarithmic curve:

where Y is the vertical distance from the valley floor, X is the horizontal distance from the centre of the valley, e is a constant and k is a shape constant.

This type of curve is known as a k-curve and the value of k is known as the k-number. For most cirques the k-number is between 0.5 and 2. The greater the value of the k-number the steeper the headwall of the cirque (Figure 6.15). A cirque fitted by a k-curve with a k-number of 2 will have a steep headwall and will be overdeepened to such an extent that its floor will mostly likely contain a lake. The type of long profile that a cirque possesses is a function of its bedrock lithology and of the structural weakness within it. For example, in areas where bedrock dips into a cirque then the floor of the cirque will also tend to dip inwards and the cirque may contain a small lake. Where bedrock dips out of the cirque it will have a floor that slopes outwards and will be best described by a low k-number (Figure 6.15). Jointing within the bedrock also determines the nature of the head-wall. Closely spaced joints tend to produce blocky headwalls, whereas widely spaced joints produce smoother headwall profiles.

The evolution of cirque morphology through time can be examined using ergodic reasoning. Ergodic reasoning suggests that under certain circumstances sampling in space can be equivalent to sampling through time, and consequently, space-time transformation is permissible as a working tool. This works on the assumption that

6.1 Microscale Features of Glacial Erosion 171

Figure 6.15 Empirical curves (k-curves) fitted to the longitudinal profiles of four Scottish cirques. The higher the value of k, the more enclosed the cirque basin. [Modified from: Haynes (1968) Geografiska Annaler, 50A, figure 3, p. 223]

Figure 6.15 Empirical curves (k-curves) fitted to the longitudinal profiles of four Scottish cirques. The higher the value of k, the more enclosed the cirque basin. [Modified from: Haynes (1968) Geografiska Annaler, 50A, figure 3, p. 223]

within a population of evolving landforms there will be a range of different examples at different stages of development. By comparing each example within a given area (i.e. in space) a model of landform evolution through time may be established. For example, within a mountainous area there may be a range of cirques each of which may have been initiated at a different time or developed at a different rate. Consequently the spatial variation in cirque morphometry within the area will provide insight into how cirques evolve through time.

This type of methodology has been used to establish the following models for the evolution of cirques in different areas. As cirques increase in size they appear to become more enclosed both in plan and in profile. Figure 6.16 shows a simple model of cirque evolution based on morphometric observation in the Scottish Highlands in which a well-defined cirque basin develops with time. In practice this type of model may hold only in areas of relatively uniform bedrock geology in which the structural weaknesses are present in all directions. In most rock masses this is not case and structural or lithological weaknesses tend to be orientated in one direction. In these situations the morphology of the cirque and its evolution may be strongly guided by orientations of these structural and lithological weaknesses. For example in the mountains of North Wales, cirques cut along the strike of the outcropping geology tend to be more elongated than those cut across strike. Work on these cirques also suggests that cirques which are orientated parallel to geological structures may evolve in a different fashion than those that are

172 Landforms of Glacial Erosion A: Cirque in plan B: Cirque in profile



Cirque Up

Figure 6.16 A model for the evolution of a cirque, based on observations of cirque morphology in the Scottish Highlands. (A) Cirque evolution in plan view. (B) Cirque evolution in profile view.

[Modified from: Gordon (1977) Geografiska Annaler, 59A, figure 7, p. 192]

Cirque Up

Figure 6.16 A model for the evolution of a cirque, based on observations of cirque morphology in the Scottish Highlands. (A) Cirque evolution in plan view. (B) Cirque evolution in profile view.

[Modified from: Gordon (1977) Geografiska Annaler, 59A, figure 7, p. 192]

orientated at right angles to the geological structure. Those cirques oriented along the geological structure appear to have experienced faster rates of headwall retreat than downcutting and consequently have evolved an elongated, flat-floored morphology. In contrast those that have developed transverse to the geological and lithological trend appear to have experienced similar rates of downcutting and headwall retreat. Consequently, they have evolved into more compact and enclosed basins. Once a basin has developed in the floor of a cirque it will continue to grow due the same positive feedback mechanism outlined for rock basins (see Section 6.2.3).

The elevation of cirques and their aspect can be used to provide general palaeoclimatic information. The altitude at which an abandoned cirque lies may be used as a measure of the former regional snowline within the area on the assumption that cirques are formed by discrete glaciers. In most cases this reflects a composite snowline averaged over numerous periods of glaciation during which the cirques were occupied. The closer the altitude of a cirque to sea level the lower the snowline. Cirques close to sea level provide evidence of relatively harsh climates, whereas those at high altitudes may form when the regional snowline is much higher. For example, cirque-floor altitudes within the Snowdon mountains of North Wales rise in elevation from the southwest to the northeast, a trend that has been interpreted as reflecting the direction of prevailing snow-winds during the later part of the Cenozoic Ice Age. The aspect of a cirque or direction in which they open is also indicative of palaeoclimate, because the location of the cirque glacier that cut the cirque is controlled by the direction of snow-bearing winds and the direction of incoming solar radiation. In the northern hemisphere most cirques face towards the northeast, although this aspect may be modified by strong snow-winds such that they form in the lee of mountain slopes crossed by the winds.

6.3 Macroscale Features of Glacial Erosion 173

Cirques provide another good example of a positive feedback system because the deeper the cirque becomes the more efficient its shape is for trapping snow and shading the accumulation areas. The cirque glacier will consequently be bigger. A bigger glacier is likely to achieve more efficient erosion and therefore to continue to excavate the cirque, a cycle that will continue while the glacier survives.

6.3.4 Giant Stoss and Lee Features

Preglacial valley spurs and other hills may be eroded by ice into giant stoss and lee forms. These are given the collective term of giant roches moutonnées and range in size from hundreds of metres to several kilometres across. These features are carved in bedrock and may appear as either quarried valley spurs or as free-standing and isolated bedrock protrusions. Their morphology is similar to that of smaller roches moutonnees, with ice-smoothed proximal surfaces and quarried distal surfaces. This suggests that they are formed by a similar process but simply on a much larger scale. Smaller roches moutonnees and other erosional landforms are usually found superimposed on these asymmetrical spurs and hills.

Good examples of giant stoss and lee features are to be found in Glen Dee, Scotland, where large streamlined hills with lee-side cliffs up to 160 m high are to be found (Figure 6.17). These large-scale landforms formed under thin ice near the

Figure 6.17 Cross-sections through two giant stoss and lee forms in Glen Dee, Scotland. [Modified from: Sugden et al. (1992) Geografiska Annaler, 74A, figure 7, p. 259]

end of the last glaciation, during a time of high ice velocity and abundant melt-water. Other large stoss and lee features are to be found in south Greenland. These large composite landforms are ice-polished on their upstream side, whereas their downstream face is quarried. The upstream sides of these landforms contain smaller whalebacks superimposed upon the main landform.

6.3.5 Tunnel Valleys

Tunnel valleys are elongate depressions with overdeepened areas along their floors cut into bedrock or unconsolidated sediment. They are frequently sinuous in planform and may occur in anastomosing networks, although they also exist as independent, straight valleys. At their largest they may be 2 km wide, over 100 m deep and extend for between 6 and 30 km in length. In general, however, they tend to be smaller, perhaps only 600-800 m wide and less than 60 m deep. In long profile they contain enclosed hollows or isolated, often elongated, basins. They are commonly infilled to varying depths with a variety of different types of sediment, including sediment gravity flows and thick units of glaciofluvial sands. Tunnel valleys occur extensively on the continental shelf of northern Europe and North America, but buried examples have also been recorded on land. The morphological variation within tunnel valleys indicates three main theories of tunnel valley formation.

1. Tunnel valleys are cut into unconsolidated sediment where deformable subglacial sediment creeps into subglacial conduits from the sides and below, followed by removal of this material through the conduit by meltwater flow. Tunnel valleys are created, therefore, by lowering of the sediment surface on either side of the conduit.

2. Tunnel valleys form during deglaciation, at or close to the ice margin, by subglacial meltwater erosion and consequently the valleys are time-transgressive.

3. They form by subglacial meltwater erosion during high-discharge catastrophic channelised flood events. In this scenario, tunnel valleys within anastomosing networks would have formed synchronously.

0 0

Post a comment