摘要：Ice-flow fields, including the driving stress, provide important information on the current state and evolution of Antarctic and Greenland ice-sheet dynamics. However, computation of flow fields from continent-scale DEMs requires the use of smoothing functions and scales, the choice of which can be ad hoc. This study evaluates smoothing functions and scales for robust calculations of driving stress from Antarctic DEMs. Our approach compares a variety of filters and scales for their capacity to minimize the residual between predicted and observed flow direction fields. We find that a spatially varying triangular filter with a width of 8–10 ice thicknesses provides the closest match between the observed and predicted flow direction fields. We use the predicted flow direction fields to highlight artefacts in observed Antarctic velocities, demonstrating that comparison of multiple observational data sets has utility for quality control of continent-scale data sets.
