If a glacier or an ice sheet rests on a permeable bed, be it rock or soft sediment, a part of the subglacial meltwater will enter the bed and be evacuated as groundwater flow, governed by the same physical rules as groundwater flow in confined aquifers outside the glaciated areas. The major differences, however, are that (i) the flow is driven by hydraulic gradient imposed by ice overburden, and (ii) some groundwater may be advected within the sediment if it deforms in response to glacier stress.
The basal meltwater will enter the substratum if the pressure at the ice-bed interface is greater than the pressure within the bed. If the bed is a porous medium such as glacial deposits and most other sedimentary rocks, water flow will be governed by the Darcy law
where Q is the water flux, K is the hydraulic conductivity, A is flow cross-sectional area, pw is the density of water, g is the acceleration due to gravity, h is the hydraulic head and l is the flow length. The flow is driven by the hydraulic gradient (1/pwg)(dh/dl) determined largely by the ice-surface slope (Shreve, 1972; Fountain &
f dh )
1 Pw g J
I dl J
Walder, 1998). The potentiometric surface of groundwater confined by an ice sheet runs approximately parallel to the ice surface at a depth corresponding to the water pressure (Fig. 9.1A). At the ice margin the pressure in the aquifer is atmospheric, thus the groundwater changes from a confined to unconfined flow. Under conditions of high meltwater recharge and low drainage capacity of both the bed and the drainage systems at the ice-bed interface, the groundwater pressure may be elevated to the vicinity of the ice overburden pressure (glacier flotation point). Under such conditions groundwater recharge ceases because of equilibrated pressures and lack of head drop into the sediment.
Groundwater flow under polythermal glaciers and ice sheets will be affected by the distribution of frozen ground in ice-marginal areas (e.g. Haldorsen et al., 1996; Cutler et al., 2000) and under central parts of ice sheets where basal freezing may occur due to advection of cold ice (Fig. 9.1B). Because frozen soil is orders of magnitude less permeable than the same soil in unfrozen state (Williams & Smith, 1991), permafrost will act as a confining layer. Pressurized groundwater thus will be forced under the permafrost, and the confined drainage system will first terminate outside the permafrost zone or at large discontinuities in the frozen ground such as taliks under rivers, lakes and over salt domes.
Pressurized, often artesian (Haldorsen et al., 1996; Flowers & Clarke, 2002b; Flowers et al., 2003) groundwater at a glacier margin can cause hydrofracturing of sediment and rocks. Boulton & Caban (1995) suggested that such fractures may occur to a depth of about 400 m, and the zone affected by hydrofracturing can extend many tens of kilometres into the ice foreland. They gave examples of small-scale sediment dykes filling hydrofractures in Spitsbergen, England and Sweden (see also Larsen & Mangerud, 1992; van der Meer et al., 1999; Grasby et al., 2000). Plate 9.1 shows a hydrofracture of similar origin in northwest Germany.
Pressurized groundwater may also facilitate formation of push moraines, diapirs and large-scale extrusion moraines (Boulton et al., 1993; Boulton & Caban, 1995), typically by sediment liquefaction. Artesian groundwater behind an ice front is capable of lifting loosened, thick rock slabs (Bluemle & Clayton, 1984; Pusch et al., 1990), which may then be redeposited by ice as allochtho-nous rafts and megablocks within glacial sediments (Aber et al., 1989). Water pressure increase will also occur in subglacial aquifers that wedge out in the direction of glacier flow. Such groundwater traps were an important cause of glaciotectonism in northwestern Germany during the last glaciation (Piotrowski, 1993), and facilitated glaciotectonism along the Main Stationary Line in Denmark (Piotrowski et al., 2004; Fig. 9.2). Deformation also may be caused by fast ice retreat where a swath of land with pressurized groundwater is exposed, causing sediment blow-ups before pressure equilibrium is established.
Groundwater flow dynamics, especially the flow field, velocity and rate, strongly depend on the substratum porosity and hydraulic conductivity, which act as 'knobs that open and close subglacial drainage valves' (Flowers & Clarke, 2002a). In the case of sorted granular materials such as sand and gravel, these parameters do not differ much from those under non-glacial conditions. More difficult to estimate are hydrogeological parameters of subglacial tills, yet this sediment type is of particular impor ice surface ice surface
tance because it occurs immediately under most glaciers and is the first sediment on the pathway of the subglacial groundwater into the bed.
When a basal till is dilated due to shear deformation and high porewater pressure, its porosity is around 0.4 (Blankenship et al., 1986; Engelhardt et al., 1990b), whereas non-dilated tills appear to have porosities around 0.25-0.3 (Fountain & Walder, 1998). The difference corresponds to an increase in flow velocity of up to about 10 times. Hydraulic conductivity of tills from past glaciations varies over several orders of magnitude, typically between 10-1 and 10-7ms-1 (Freeze & Cherry, 1979), with an extreme effect on groundwater fluxes (see the Darcy equation).
Conductivity of a till under a glacier can be expected to vary accordingly, but its direct determination is complicated owing to limitations in the accessibility of glacier beds. Reliable estimates are, therefore, scarce. Examples of in situ estimates using different methods, and some laboratory tests on till samples collected from under modern glaciers and ice sheets, are given in Table 9.1 and show the extreme spread of values between some 10-2 and 10-12 ms-1. Worth noting is that in situ measurements tend to yield higher conductivities than laboratory tests, owing to large-scale non-Darcian flow pathways and large-scale till heterogeneities (e.g. Gerber et al., 2001), such as sand and gravel lenses, known to occur in nature but not accounted for in laboratory tests.
Besides the textural composition, also compression and shear deformation influence the conductivity of subglacial tills. Both result in highly anisotropic distribution of till properties with horizontal conductivity up to two orders of magnitude greater than the vertical one owing to particle alignment, as shown by laboratory experiments (e.g. Murray & Dowdeswell, 1992). Till deformation in response to glacier stress may increase the conductivity (in the case of dilation; Clarke, 1987b) or decrease it (in the case of compaction), as demonstrated experimentally by Hubbard & Maltman (2000). They also showed that hydraulic conductivity in tills is inversely related to the effective glacier pressure because pressurized porewater prevents sediment compaction and closure of drainage pathways.
If till undergoes pervasive deformation, substantial volumes of groundwater may be advected toward the ice margin within the high-porosity deforming layer (Alley et al., 1986a, 1987a). Such deformation may destroy drainage paths at the ice-bed interface (Clarke et al., 1984), but it will enhance drainage through the bed (Murray & Dowdeswell, 1992). Despite our still fragmentary knowledge about hydrogeological properties of tills under glaciers, one can safely conclude that these properties vary extremely both in space and time, influenced by the nature of the source material and stresses applied by the overriding glacier.
Was this article helpful?