Inland migration of fast-flowing outlet glaciers and ice streams: Model investigations of the transition from tributary to ice-stream flow

We have developed a two-dimensional full-stress model to (1) investigate the controlling processes in onset regions and (2) examine how those processes might affect the evolution of the onset region over time and space. We distinguish between processes that are "external" from those that are "linked" to the ice flow, and examine the relative importance of each in the inland- to ice-stream flow transition. External processes that could help enhance basal sliding and weaken the connection between the glacier and the bed (such as the morphology of the subglacial bed or the geothermal flux) are not likely to change over the short timescales of interest when studying onset-migration (i.e. 10's-100's of years). If external processes are the primary control on where ice streams can initiate, no migration or only a limited amount might be expected. Processes that are linked to the ice flow may also influence the strength of the ice bed connection, and help to facilitate the rapid-basal motion necessary for an ice stream to develop. Examples of some of these processes are the rate of frictional melting (linked to the basal drag and the basal sliding speed), the basal temperature gradient (which controls the rate at which heat at the bed is conducted upwards into the ice), and the hydraulic potential gradient (which controls the direction of basal water flow, and is dominated by the ice sheet surface slope). Non-linear feedback between these processes might allow for the position of the ice stream onset to migrate over time.