Predicting the Onset of Subglacial Drainage Channels

Abstract

The sliding speed of glaciers depends strongly on the water pressure at the ice-sediment interface, which is controlled by the efficiency of water transport through a subglacial hydrological system. The least efficient component of the system consists of “distributed” flow everywhere beneath the ice, whereas the “channelized” drainage through large, thermally eroded conduits is more efficient. To understand the conditions under which the subglacial network channelizes, we perform a linear stability analysis of distributed flow, considering competition between thermal erosion and viscous ice collapse. The calculated growth rate gives a stability criterion, describing the minimum subglacial meltwater flux needed for channels to form, but also indicates the tendency to generate infinitely narrow channels in existing models. We demonstrate the need to include lateral heat diffusion when modeling melt incision to resolve channel widths, which allows continuum models to recover Röthlisberger channel behavior. We also show that low numerical resolution can suppress channel formation and lead to overestimates of water pressure. Our derived channelization criterion can be used to predict the character of subglacial hydrological systems without recourse to numerical simulations, with practical implications for understanding changes in ice velocity due to changes in surface melt runoff.