Direct measurements and computer simulations of modern ice sheets show that the ice-bed interface tends to be at temperatures well below the pressure melting point in the central, divide regions, in the near terminal zone, where this terminates on land, in the summit regions of hills and mountains beneath the ice sheet where the ice is locally thin, and beneath slowly flowing icesheet zones between ice streams. Christoffersen & Tulaczyk (2003) have shown how the ice-bed interface can freeze when ice streams stagnate, potentially leading to processes of sediment incorporation by freezing in, as discussed by Weertman (1957) and Boulton (1972). Goldthwait (1960) produced evidence from northwest Greenland which shows that a frozen sediment bed can be largely uneroded by the glacier sole, and Holdsworth (1974) drew similar conclusions from study of a subglacial rock bed in Antarctica.
The rationale for non-erosion is provided by the results of Jellinek (1959), who showed that the adhesive strength of an ice-rock interface at temperatures well below the melting point was significantly greater than the maximum average shear stress that is normally generated at the base of an ice sheet. Under these conditions, we expect sliding to be inhibited at the ice-bed interface and the movement of ice sheets to be dominated by internal flow, with basal shear stresses of the order of 100 kPa. However, sliding can occur at temperatures lower than the ambient pressure melting point (Shreve, 1984; Echelmayer & Wang, 1987), and can locally abrade bedrock (Atkins et al., 2002), leading to debris incorporation in basal ice (Holdsworth, 1974), and deformation of any frozen subglacial sediments either by brecciation and block incorporation (Boulton, 1979) or by ductile deformation (Davies & Fitzsimons, 2004). Although local stress-strength relationships may permit significant debris or sediment blocks to be torn away from the substratum, in general cold-bed erosion rates are likely to be very small compared with those achieved by a rapidly sliding glacier sole.
Very large frozen sediment and rock masses can be transported by ice if impeded, inefficient drainage beneath a subglacial frozen horizon creates high water pressures beneath a subglacial horizon. Provided that there is strong adhesion between the glacier sole and underlying frozen materials, high water pressures at the base of the frozen horizon can reduce the effective pressure and the friction to such a low value that sliding can occur along the frozen-unfrozen interface. It has been suggested that this process could permit very large sediment or rock masses to move with the glacier (Weertman, 1957; Mackay & Matthews, 1964; Moran, 1971; Boulton, 1972) and may explain some of the large, relatively undisturbed, pre-existing rock or sediment masses that are found in glacial till sequences (e.g. Christiansen, 1971) and many very large glacier tectonic structures. Permafrost effectively becomes part of the glacier, with the basal décollement beneath rather than above it.
Was this article helpful?