Box 51 Direct Observations And Measurements Of Glacial Abrasion

The direct observation of abrasion in action is extremely difficult because it involves digging tunnels through a glacier to access basal cavities. Some of the first observations are those of Boulton (1974) who described the movement of a basalt fragment over a large basalt roche moutonnee 20 m below the surface of Brei9amerkurjokull in southeast Iceland. The basalt fragment was removed from the base of the glacier and the surface that had been in contact with the bed was inspected. The fragment had been in contact with the bed at three points and between the points of contact crushed debris had been ploughed up in front. Striations produced by the fragment could be traced for 3 m. The largest striation was seen to deepen rapidly to 3 mm but then to gradually shallow to 1 mm. Boulton (1974) related this decrease in depth to the build up of crushed debris, which spreads the load at the interface over a wider area. The build up of a layer of crushed debris was thought to result in a change in the nature of motion from a jerky 'stick-slip' motion, to a relatively uniform sliding movement - the stick-slip motion produces a carpet of debris over which the particle subsequently slides. When this carpet is exhausted by comminution, the clast will again come into contact with the bed, thereby recutting the striation. This may explain the disappearance and reappearance of striae. These observations suggest that there are two abrasive processes: (i) the cutting of striae; and (ii) polishing of the bed by fine debris that is ploughed up when a striation is cut. Boulton (1974) went on to measure the rate of abrasion by cementing rock and metal plates to bedrock surfaces adjacent to basal cavities beneath Brei9amerkurjokull in Iceland and the Glacier d' Argentiere in the French Alps. These plates became quickly covered by basal ice and were later recovered for inspection (see Table below). Boulton's results have now been supplemented by measurements beneath 200 m of ice at Engabreen, Norway (Box 5.2). Here, Cohen et al. (2005) measured the friction between the debris in basal ice and a smooth tablet of rock inserted under the glacier. The ice contained 10% debris by volume and exerted local shear traction of up to 500 kPa. Calculations show that the shear traction due to the friction between the debris and the bed is around 100 kPa at Engabreen. These authors concluded that the friction between debris in basal ice and the bed is much higher than previously assumed and is sufficient to have a retarding effect on rates of glacier sliding.

Locality

Average abrasion rate

Ice thickness

Ice velocity

(mm yr )

(m)

(m yr1)

Marble plate

Basalt plate

[email protected] 1

3

1

40

9.6

[email protected] 2

3.4

0.9

15

19.5

[email protected] 3

3.75

32

15.4

Glacier d'Argentiere

36

100

250

Sources: Boulton, G.S. (1974) Processes and patterns of glacial erosion, in Glacial Geomorphology (ed. D.R. Coates), Proceedings of the Fifth Annual Geomorphology Symposia, Binghampton, Allen & Unwin, London, pp. 41-87. Cohen, D., Iverson, N.R., Hooyer, T.S., et al. (2005) Debrisbed friction of hard-bedded glaciers. Journal of Geophysical Research - Earth Surface, 110, F02007.

Sources: Boulton, G.S. (1974) Processes and patterns of glacial erosion, in Glacial Geomorphology (ed. D.R. Coates), Proceedings of the Fifth Annual Geomorphology Symposia, Binghampton, Allen & Unwin, London, pp. 41-87. Cohen, D., Iverson, N.R., Hooyer, T.S., et al. (2005) Debrisbed friction of hard-bedded glaciers. Journal of Geophysical Research - Earth Surface, 110, F02007.

5.1.1 Basal Contact Pressure

This is probably the most important variable that determines the rate of glacial abrasion. Using the analogy of a block of wood being sanded with a piece of sandpaper, it follows that the harder you press down on the surface of the wood the faster it is worn away. In a glacier this is the contact pressure between the clast in basal transport and the glacier bed beneath it: the greater the pressure the more abrasion will occur. There are, however, two alternative views on what controls this contact pressure and the movement of basal clasts. The first view was developed by Geoffrey Boulton and the second was suggested later by Bernard Hallet.

1. The Boulton model. This model assumes that the contact pressure between a particle in contact with the glacier bed is related to the effective normal pressure. As we saw in Section 4.6, this is a function of: (i) normal pressure, given by the weight of the overlying ice; and (ii) the basal water pressure, which acts in opposition to the weight of the overlying ice by buoying up the glacier, similar to the action of a hydraulic jack. Effective normal pressure therefore will be high when: (i) the ice is thick; and (ii) basal water pressure is low. This last point is of some importance because in a cold-based glacier, where there is little or no meltwater present at the glacier bed, effective normal pressure will be much higher than for a warm-based glacier of similar thickness. Similarly bedrock lithology beneath a warm-based glacier may also be important.

Effective normal pressures will be much higher on porous rocks because this will reduce the basal water pressure (see Section 4.6). Where the bed is not horizontal, for example where there is a bedrock obstacle, effective normal pressure is modified by an amount equal to the pressure of the ice flowing against the obstacle (Figure 4.6: see Section 4.6).

In the Boulton model effective normal pressure controls the rate of abrasion. As effective normal pressure increases, abrasion will also increase as the clast in the base of the ice is being pushed harder into the bed. However, as effective normal pressure increases, the friction between the clast and the bed is also increased and this friction will ultimately begin to slow the movement of the particle and the basal ice that holds it will begin to flow around the clast. When this occurs, abrasion will start to decrease despite the fact that effective normal pressure is still increasing. Consequently, if all other variables are held constant (e.g., sliding velocity) then abrasion will first increase with effective normal pressure and then decrease until the friction between the clast and the bed is such that it will stop moving and will lodge.

2. The Hallet model. This model assumes that the contact pressure between a clast in basal transport is independent of the effective normal pressure. This theory is based on the premise that clasts are completely surrounded by ice and can be considered to be essentially floating within it. This occurs because ice will deform around a clast by creep due to the weight of the ice above it and therefore basal clasts are effectively surrounded by ice at all times. In this case the contact pressure between a clast and the glacier bed is a function of the rate at which ice flows towards the bed, forcing the clast into contact with the bed. This depends on: (i) the rate of basal melting; and (ii) the presence of extending glacier flow. In this model, abrasion is independent of variations in effective normal pressure and is primarily a function of basal melting. As we saw in Section 3.4 basal melting is favoured by: (i) rapid ice flow, which generates large amounts of frictional heat; (ii) thick ice; (iii) high ice surface temperatures; and (iv) the advection of warm ice towards the glacier bed.

We shall return to these two different views of basal contact pressure below because two very different models of glacial abrasion have been developed around them.

5.1.2 Basal Sliding

The rate of basal sliding controls the rate at which basal debris is physically dragged across the surface below, and consequently the greater the rate of basal sliding the greater the amount of abrasion. Using the sandpaper analogy, the faster you move the sandpaper back and forth then the faster the wood is worn away. Basal thermal regime is important (see Section 3.4), because sliding is not widespread beneath cold-based glaciers and these glaciers have less ability to abrade their beds (see Boxes 6.8 and 6.9).

5.1.3 The Concentration and Supply of Rock Fragments

The concentration of debris within basal ice also controls the rate of abrasion. Ice on its own cannot cause significant abrasion - it needs debris within it to do this. The rate of abrasion, however, is not increased simply by increasing the concentration of basal debris. In fact it has been suggested that abrasion is most effective where basal debris is relatively sparse. This is because basal debris increases the frictional drag between the ice and its bed and therefore reduces the sliding velocity. Glaciers with relatively clean basal ice are able to slide faster than those with large amounts of basal debris. There is a certain threshold of debris concentration above which the abrasion rate declines with increasing debris content, because of its adverse effect on the rate of basal sliding (Figure 5.1).

Figure 5.1 Schematic representation of the relationship between basal sliding velocity and the concentration of debris at the base of a glacier.

The type and shape of basal debris are also important. Some rocks are more durable than others, and a glacier armed with basal debris derived from a hard or resistant lithology will be more effective than one armed with a relatively soft lithology. The most effective combination occurs where a glacier armed with debris entrained from a hard substrate flows over a relatively soft lithology. If the debris is softer than the substrate little abrasion would occur, because erosion would preferentially reduce the size of the basal clasts first. The shape of the basal debris is also important because sharp fragments are able to make deeper incisions into the underlying bedrock than those with blunter or more rounded points or edges. Laboratory observations have shown that clasts in contact with the bed frequently rotate or flip, which helps to improve their life span as erosive tools beneath the glacier.

A continued supply of basal debris is also important because basal debris is quickly worn down and crushed. For abrasion to be effective, basal debris must therefore be continually replaced. This may occur either by: (i) the entrainment of fresh glacial debris at the glacier bed; or (ii) by basal melting, which progressively lowers debris down through a glacier towards the bed (see Section 7.2).

5.1.4 Abrasion Models

We have already seen that there are two alternative views concerning the nature and controls on the contact pressure between a basal clast and the bedrock beneath a glacier. Both Boulton and Hallet have developed numerical models with which to predict the patterns and amounts of glacial abrasion. The two models are very different.

Boulton's abrasion model assumes that the contact pressure on a rock particle at the base of a glacier is a function of the normal effective pressure. As a consequence his model predicts that abrasion will be controlled by: (i) the effective normal pressure; and (ii) the ice velocity. Effective normal pressure is controlled by ice thickness and basal water pressure (see Section 4.6). The relationship between abrasion and these two variables within Boulton's model is illustrated in Figure 5.2. This graph shows that for a given ice velocity abrasion increases to a peak as effective normal pressure increases, and then falls rapidly to zero as the friction between debris and bed becomes sufficient to retard the movement of the

Figure 5.2 Graphic representation of Boulton's abrasion model. The graph shows theoretical abrasion rates plotted against effective normal pressure for different ice velocities. In Zone A, abrasion rates increase with increasing pressure whereas in Zone B abrasion rates decline with increasing pressure. Zone C, located to the right of the higher x-axis intercept for any one ice velocity, is an area of no abrasion and basal debris is deposited as lodgement till. [Modified from: Boulton (1974) in Glacial Geomorphology (ed. D.R. Coates), George Allen and Unwin, figure 7, p. 52]

Figure 5.2 Graphic representation of Boulton's abrasion model. The graph shows theoretical abrasion rates plotted against effective normal pressure for different ice velocities. In Zone A, abrasion rates increase with increasing pressure whereas in Zone B abrasion rates decline with increasing pressure. Zone C, located to the right of the higher x-axis intercept for any one ice velocity, is an area of no abrasion and basal debris is deposited as lodgement till. [Modified from: Boulton (1974) in Glacial Geomorphology (ed. D.R. Coates), George Allen and Unwin, figure 7, p. 52]

particle. At effective normal pressures above a critical level no abrasion occurs, but instead debris hitherto transported is deposited. Erosion and deposition appear therefore to be two parts of a continuum.

Boulton has used this model to predict the evolution of bedrock bumps by glacial abrasion (Figure 5.3). In Section 4.6 we saw how effective normal pressure varied across an obstacle (Figure 4.6). Given this pattern of variation Boulton used his abrasion model to predict how the shape of a two-dimensional obstacle would change with erosion. He assumed that the bump had a sinusoidal shape, that the ice velocity over the bump was 50 m per year and that the pressure fluctuation over the bump was 130 kpa (Figure 4.6). Given these values he charted the evolution of two bedrock bumps under a glacier: one that had an effective normal pressure of 70 kpa and one which experienced 240 kpa. The two patterns of evolution are quite different and are shown in Figure 5.3. The bump with an

Figure 5.3 Patterns of abrasion across a sinusoidal bedrock bump for a constant ice velocity of 50 m per year for two different values of normal pressure using Boulton's abrasion model. [Modified from: Boulton (1974) in Glacial Geomorphology (ed. D.R. Coates), George Allen and Unwin, figure 9, p. 56]

Figure 5.3 Patterns of abrasion across a sinusoidal bedrock bump for a constant ice velocity of 50 m per year for two different values of normal pressure using Boulton's abrasion model. [Modified from: Boulton (1974) in Glacial Geomorphology (ed. D.R. Coates), George Allen and Unwin, figure 9, p. 56]

effective normal pressure of only 70 kpa evolved into a stoss-and lee-form (see Section 6.2.2). The normal pressures all fall within the zone of rising abrasion and pressure (Zone A: Figure 5.2). The rate of abrasion was therefore highest on the up-glacier flank, low on the crest of the bump, zero on its lee flank where a cavity forms and high at the foot of the lee flank. A roche moutonnee shape is therefore produced with time (Figure 5.3A). The morphology of the bump under 240 kpa of effective normal pressure evolves very differently. Here the normal pressures fall within zones of falling or zero abrasion with increasing pressure (Zones B and C: Figure 5.2). Lodgement occurs on the up-glacier flank, slight abrasion occurs on the crest of the bump, maximum abrasion occurs on the down-glacier flank and slight abrasion at the foot of this flank. The effect is to produce a form that migrates in an up-glacier direction, with a steep up-glacier flank similar to a crag and tail (Figure 5.3B). In summary, therefore, the key implications of Boulton's abrasion model are:

1. Variations in ice thickness control abrasion and lodgement via effective normal pressure.

2. Variations in basal water pressure, controlled by such factors as bed permeability (geology) control abrasion via effective normal pressure.

3. Ice velocity controls the rate of abrasion.

4. Abrasion and lodgement form part of a continuum.

Hallet used an alternative approach to Boulton in formulating his abrasion model. Basal rock particles are envisaged as essentially floating hydrostatically in the ice and are therefore independent of effective normal pressure. In this model the rate of basal melting and ice velocity are the key controls on abrasion. As we will see in the next chapter Hallet's model has been used widely in numerical models designed to study the evolution of large glacial landforms. The main implications of Hallet's model are:

1. Abrasion is highest where basal melting is greatest.

2. Abrasion is independent of effective normal pressure and therefore of basal water pressure, although not glacier thickness because this controls the rate of basal melting.

3. Lodgement and abrasion are independent processes.

These two models contain very different predictions and at first sight these two theories seem to conflict. It is, however, possible that each model represents different but equally valid subglacial conditions. Boulton's model and predictions may apply where the basal ice is particularly dirty and debris-rich and therefore likely to behave as a solid slab. The rigid nature of this slab prevents the ice deforming around each clast. In contrast, Hallet's model and predictions may be more appropriate in areas where basal debris is sparse and the ice is consequently less rigid.

0 0

Post a comment