Where glaciers rest directly on aquifers or are separated from them only by a thin layer of low-permeability sediment, the groundwater is in contact with the water in channels at the ice-bed interface. Channel formation reduces the water pressure and generates a hydraulic gradient, which will drive groundwater from the surrounding sediment into the channel and create a catchment area along the channel. This was suggested independently by Shoemaker & Leung (1987) and Boulton & Hindmarsh (1987), and recently measured in piezometers nested around a channel under Breidamerkurjokull in Iceland (Boulton et al., 2001b).
Radial groundwater flow to the channel can fluidize the sediment, followed by its injection into the channel and subsequent removal by water flow. In a steady-state situation, this mechanism could produce large incisions resembling tunnel valleys by erosion in relatively narrow R-type channels serving as sediment sinks for the surrounding aquifers (Boulton & Hindmarsh, 1987). Possible capturing of groundwater by tunnel valleys was also suggested for one prominent tunnel valley tract in northern Germany (Piotrowski, 1994). However, it was postulated that an increase of effective pressure at the channel flanks would strengthen the sediment there and, in particular, prevent pervasive deformation of the glacier bed.
Modelling by Boulton et al. (2001b) and Fleming & Clark (2000) shows a progressively damped pressure wave propagating away from a channel into the sediment for tens of metres, consistent with some field measurements (e.g., Fountain, 1994; Hubbard et al., 1995; Murray & Clarke, 1995). Boulton et al. (2001b) predicted a bulb of low pressure in a shallow zone beneath the channel. Occurrence of this low-pressure zone is confirmed by Piotrowski et al. (1999), who documented a soft-sediment diapir under a subglacial channel of Saalian age in eastern Germany. The diapir demonstrates sediment creep into the channel from below, and yields support for theoretical considerations of Boulton & Hindmarsh (1987) and Shoemaker & Leung (1987) suggesting a sediment mobilization into a channel and erosion rates higher than could be estimated from channel dimensions alone.
Coupling of water flow in channels and in the bed also implies that, under specific circumstances such as a wave of ablation water reaching a subglacial channel or channel blockage by sediment injection or roof collapse, a water pressure increase will propagate into the aquifer. This will happen with delay, caused by hydraulic resistance of the sediment, particularly in low-conductivity, finegrained aquifers. The result will be a reversal of hydraulic gradient, now driving the water from the channel into the aquifer. Frequent shifts in groundwater flow directions close to channels would be expected where the subglacial hydrology is dominated by diurnal or seasonal ablation cycles, i.e. close to a glacier margin.
Was this article helpful?