A statistical model for basal sliding

Abstract:

Basal sliding, rather than ice deformation, is responsible for nearly all of an ice-sheet's mass flux into its surrounding oceans. The importance of basal sliding has lead many to speculate on a so-called `sliding law', or a relation between observable quantities and the sliding speed of flowing ice. To date, a universal sliding law has been elusive, primarily due to lack of observations of the processes occurring at the ice-bed interface. In recent years, Interferometric Synthetic Aperture Radar (InSAR) surface velocity observations have grown from a few outlet glaciers to nearly complete coverage of both Antarctica and Greenland. We use these data to constrain an ice-sheet model, and employ control methods to determine the basal traction field. While this method of estimating basal traction incorporates errors in the geometry and rheology of ice, it does provide an estimate of basal processes that can not be directly observed at the desired resolution. With a comprehensive estimate of the basal traction, we derive parameter estimates which linearly relate ice-sheet observations to the estimated traction fields via a generalized linear model. Here, we present the methodology, linear parameter estimates, traction field estimates, and corresponding surface speed. To conclude, an experiment featuring the dynamical evolution of an ice sheet's basal traction is conducted. In it, we see that the dynamically evolving field provides a sliding law that is at least as plausible as static prescriptions of basal traction or heuristic relations involving sliding.

One of the statistically-generated-basal-traction fields for Antarctica is shown below.