Till on a rock bed

In cases when the rock is of low permeability compared with the till, the melt flux must be discharged through the till to the nearest low-pressure channel. Even tills in old, hard Shield areas, when in a non-dilatant state, tend to have hydraulic conductivities of the order of 10-6-10-8ms-1 (e.g. Engquist et al., 1978). For a basal melt rate of the order of 10-9ms-1, a till 1m thick with a conductivity of 10-7ms-1 would be able to discharge the winter melt towards a low pressure channel with a head gradient low enough to ensure that pressures did not exceed ice pressures, provided there were efficient tunnel conduits spaced at about 2 km intervals. If the till was shearing and dilating to a depth of 0.5 m, with the conductivity increased by an order of magnitude in the dilatant horizon (see Piotrowski, this volume, Chapter 9), the spacing could be five times larger, and even larger if significant water fluxes were advected with the deforming till.

If significant quantities of summer meltwater were to reach the bed from the glacier surface, the necessary spacing of channels would be much reduced to the order of tens of metres. There is therefore a strong probability that an extreme oscillation of the system between winter and summer would produce an annual cycle of channel formation, from a winter condition of large, widely spaced channels, to a summer condition with a closely spaced network of channels. Boulton et al. (2001b) have suggested that beneath an ice sheet, the winter channels are stable, long lived R-channels that give rise to eskers. It is possible that the closely spaced network of summer channels could be the slowly flowing, high water pressure 'canals' of Walder & Fowler (1994), which would be short-lived, and thereby unable to integrate themselves and develop into an efficient low pressure channel system able to draw down water pressures over wide areas as well-developed R-channels. At locations remote from the glacier margin, they would be likely to be tributary to major R-channels.

The water pressure gradient in a till overlying low permeability bedrock will tend to be horizontal (Boulton & Dobbie, 1993), directed towards major low-pressure R-channels acting as sinks for groundwater flow. However, where there is a well connected fracture system in bedrock, such that water flow along interconnected fractures is able to draw down water pressures, there may be a strong, local downward water pressure gradient, sufficient to liquefy the till and force it into bedrock fractures. The results of such a process are commonly seen in bedrock areas, where till 'dykes' can penetrate many metres into bedrock (Fig. 2.2a). It is also suggested that this process of 'hydraulic intrusion' can lead to forcing apart of bedrock joint blocks by widening pre-existing or incipient fractures so as to break up the rock mass and make it more susceptible to erosion (Fig. 2.2b & c).

2.3.5 A thick unlithified sediment bed

Figure 2.3 shows an experimental site at the glacier Brei-damerkurjokull in Iceland which was overrun by a 'minisurge' of part of the glacier margin (Boulton et al., 2001a). The developing surge was anticipated some two months before it occurred, and prepared for by trenches dug in the forefield of the glacier to a depth of up to 2.5m. They revealed a till unit 1-3 m thick overlying the sands and gravels of an extensive aquifer of up to about 80-90 m in thickness that occurs over a wide area in the vicinity of the glacier margin. Water pressure transducers were emplaced in the till and in the underlying aquifer, and monitored during the glacier advance. Figure 2.4 shows results from the transducers at 65 m along the transect. They indicate the prevalence of a strong, downwardly directed water pressure gradient, reflecting downward flow of water from the glacier sole through the till into the underlying aquifer. The maximum pressure drop across the till is about 90kPam-1. The aquifer controls the water pressure at the base of the till, and the thickness and hydraulic conductivity of the till control the water pressure at the top of the till. It is clear that the greater the thickness of the till, the larger will be the water pressure and the lower the effective pressure in its topmost horizon for a given aquifer pressure. This has important implications for the evolution of mechanical coupling (see section 2.4).

In general, a thick bed of unlithified sediments, such as that described by Piotrowski (this volume, Chapter 9) from northern

Jointed Bedrock

Figure 2.2 Injection of till into jointed bedrock and incorporation of bedrock blocks into deforming till. (a) Till-filled bedrock joints at 6 m below the bedrock surface near Forsmark, Sweden. (b) Partially detached bedrock blocks in till at the bedrock surface, near Uig, Skye, Scotland. (c) Some detached bedrock blocks at the same locality are completely incorporated in the till as a result of till deformation. Others are partially detached.

Figure 2.3 Site at the margin of Breidamerkurjokull, Iceland, overrun by a minisurge (Boulton et al., 2001a). Water pressure transducers (marked by crosses) were emplaced in the walls of a trench dug prior to the minisurge. The glacier advance extended as far as the 0 m datum along the transect. Examples of the water pressure gradients at 65 m from datum are shown in Fig. 2.4.

Figure 2.3 Site at the margin of Breidamerkurjokull, Iceland, overrun by a minisurge (Boulton et al., 2001a). Water pressure transducers (marked by crosses) were emplaced in the walls of a trench dug prior to the minisurge. The glacier advance extended as far as the 0 m datum along the transect. Examples of the water pressure gradients at 65 m from datum are shown in Fig. 2.4.

Figure 2.4 Water pressures in the till at 65 m along the transect (Fig. 2.3) at 6-h intervals on selected days. Early gradients (days 97,100) reflect upward flow of water through the till just beyond and at the advancing glacier margin as water flowing outwards from beneath the glacier in the subtill aquifer wells upwards towards the surface. Later gradients (days 104, 108, 115) reflect downward movement of water through the till from the glacier sole. Some reversals of gradient occur on day 115 as aquifer pressures exceed till pressures. Maximum gradients tend to occur in the late afternoon during periods of peak glacier surface melting.

Pore pressure - kPa

Figure 2.4 Water pressures in the till at 65 m along the transect (Fig. 2.3) at 6-h intervals on selected days. Early gradients (days 97,100) reflect upward flow of water through the till just beyond and at the advancing glacier margin as water flowing outwards from beneath the glacier in the subtill aquifer wells upwards towards the surface. Later gradients (days 104, 108, 115) reflect downward movement of water through the till from the glacier sole. Some reversals of gradient occur on day 115 as aquifer pressures exceed till pressures. Maximum gradients tend to occur in the late afternoon during periods of peak glacier surface melting.

Germany, will contain both more and less conductive strata. In most cases, irrespective of the precise ordering of the pre-glacial strata, we would expect a relatively low conductivity till to be created by the glacier to overlie the pre-existing sequence, and that this till would have a significantly lower conductivity than those on the shield (clay-silt tills from lowland areas typically have conductivities in the range 10-8-10-9ms-1 compared with 10-7ms-1for the Iceland till). If such a till were to be inserted into the Icelandic setting, we would expect a pressure drop across the till that was one to two orders of magnitude greater (1000-10,000 kPam-1).

These results have several important implications.

1 The hydrological properties of a glacier bed and the operation of the subglacial hydraulic system must play a fundamental role in governing the frictional drag offered by the bed to glacier movement. Much existing work has focused on the role of channels and a water film. The role of groundwater flow is at least of similar importance.

2 If water is forced from the glacier sole through a till of very low hydraulic conductivity, a very large water pressure gradient will develop. Such a gradient may be large enough to create hydrofracturing, resulting in an increase in conductivity. Under these circumstances, conductivity is not an intrinsic property of the till, but a product of its hydraulic setting and granulometry.

3 As a till thickens as a consequence of deposition, the pressure difference across it will tend to increase, so that the effective pressure at the top of a till accumulating above a subglacial aquifer with a constant water pressure will tend to reduce. The till will therefore offer progressively less shearing resistance to the overriding glacier, thereby facilitating faster flow.

4 A very large flux of meltwater can drain as groundwater through a subglacial aquifer. Low pressures in the aquifer (strong pressure drawdown) can co-exist with water pressures at the top of the overlying till that approach the ice pressure.

5 Local variations in the effective pressure in a till are primarily dependent upon the drainage pathway and the water flux rate. The drainage pathway can change as a consequence of reorganization of channels that act as groundwater sinks, and the flux rate can vary diurnally and seasonally.

6 The maximum effective pressure that a till has suffered, without subsequent remoulding, determines its consolidation state. We expect this to reflect the periods of strongest drainage and smallest meltwater flux.

7 It is important to establish how far from an ice-sheet margin there are strong fluctuations in pressure generated by diurnal and annual recharge changes in surface water recharge to the bed.

2.4 Tractional processes

Because of the influence of basal friction on the large-scale dynamics of ice sheets (section 2.1.3), the strength of coupling between the basal ice and its bed is a key issue for glaciology. The parallel issue for glacial geology is how the state of traction is reflected in the erosional or depositional products of glaciation.

The traction at the base of a glacier depends upon the temperature at the ice-bed interface, the effective pressure at the interface, which determines the degree of interlocking between the glacier sole and the bed, and the strength of subsole materials, whether lithified or unlithified. These vary according to the lithology of the substratum, in which three principal types can be identified.

Was this article helpful?

0 0

Post a comment