*Landscape and Landform Research Group, Department of Geography, The University of Reading, Reading RG6 6AB, UK fDepartment of Geography, The University of Sheffield, Sheffield S10 2TN, UK
Rapidly flowing ice streams exert a profound influence on icesheet configuration. It is, therefore, essential to incorporate the spatial and temporal activity of ice streams in order to reconstruct accurately the evolution of an ice sheet through time. Recently, considerable progress has been made in identifying palaeo-ice stream imprints in formerly glaciated terrain and we now have a large (>50) population of palaeo-ice stream tracks to investigate (Stokes & Clark, 2001). Their identification has refined our reconstructions of ice-sheet dynamics but they also provide an unprecedented opportunity to advance our understanding of ice-stream behaviour and stability. For example, once a palaeo-ice stream track has been identified confidently, its 'footprint' may provide answers to several important glaciological questions such as:
1 How big was the ice stream and what was its likely catchment area?
2 What factors triggered its location within the ice sheet?
3 What was its likely flow mechanism?
4 For how long did it operate and did it operate more than once?
5 What factors led to its shut-down?
6 What was its wider impact on the ice sheet?
Given that the bedform record inscribed by an ice stream is intimately linked to its activity then it is possible to use inversion techniques to answer such questions. This case study demonstrates how the bedform record of the Dubawnt Lake Ice Stream in North Keewatin (Canada) can be used to glean pertinent information about the functioning of ice streams.
The Dubawnt Lake Ice Stream represents a >450-km-long bottleneck flow pattern of lineations, trending in a west-northwest direction northwest of Dubawnt Lake, Keewatin. In Kleman & Borgstrom's (1996) inversion model, it was highlighted as a surge fan generated rapidly during deglaciation of the Keewatin Sector (see also this chapter). More recently, detailed mapping of the flow-set and surrounding areas has established its prominence as one of the best-preserved terrestrial ice-stream imprints available for scrutiny (Stokes & Clark, 2003a).
The location of the flow-set is shown in Plate 41.1, which also illustrates typical examples of the size and pattern of subglacial bedforms in the onset zone, main ice-stream trunk and terminus. Detailed mapping reveals that bedforms get longer and more closely packed together in the main trunk of the ice stream and that the overall pattern of bedform elongation (Plate 41.1c) mimics the expected velocity field for a terrestrial ice stream. The sharp southern margin of the flow-set is another reliable indicator of an abrupt change in ice dynamics, marking the boundary between the rapidly flowing ice stream and the neighbouring slow-flowing ice (and the probable boundary between the fast warm-based ice and slower cold-based ice). The northern margin of the flow-set is less clear but suggests that the ice stream was around 140 km wide at its narrowest point, significantly wider than contemporary ice streams in the Antarctic and Greenland. The length is reconstructed at 450 km and the ice stream had an estimated catchment area of around 190,000 km2.
The location of the ice stream on the relatively flat Canadian Shield conflicts with the paradigm that ice streams require topographic funnelling or soft subglacial sediments for their initiation (Stokes & Clark, 2003a). Although softer sediments interrupt the characteristically hard crystalline bedrock on the Canadian Shield in this location, it is unlikely that they were in sufficient quantity to 'trigger' the ice stream. Rather, it is speculated that the evolution of large (>3000km2), deep (ca. 120m) proglacial lakes impounded at the ice sheet margin may have been important in initiating ice stream activity by inducing calving and taking the system beyond a threshold that was sufficient to trigger fast ice flow. Evidence of these lakes suggests that the only location where they existed for a significant length of time exactly matches the location of the ice-stream flow pattern (see Craig, 1964; Stokes & Clark, 2004).
Elucidating the flow mechanisms beneath ice streams is a major scientific challenge and we argue that the bedform record of palaeo-ice streams can shed some light on this problem. Rather than viewing the lineaments on the Dubawnt Lake Ice Stream bed as streamlined ridges, it is suggested that a better description may be that of a highly grooved till surface (see Plate 41.1b). On the basis of this observation and investigations of megascale glacial lineations that characterize other ice-stream beds (Clark et al., 2003b) our hypothesis is that the ice stream flowed by 'groove-ploughing'. Under this mechanism, large keels at the base of the ice stream plough through sediments, carving elongate grooves and deforming material up into intervening ridges. If correct, this
Was this article helpful?