4D Surface Velocity Fields of Rutford Ice Stream, West Antarctica, Inferred From Continuous Synthetic Aperture Radar Observations

Abstract:

The mechanics of West Antarctic Ice Sheet (WAIS) beds impose fundamental constraints on ice flow and help set the timescale of potential WAIS collapse. Understanding basal mechanics requires a variety of observational and modeling approaches, among them using synoptic-scale observations of surface velocities as constraints on numerical ice flow models. This approach has been developed and employed over the past few decades using available surface velocity data, which often amount to snapshots of ice flow at sporadic times. The next major step in the evolution of geodetic methods for glacier dynamics is capturing synoptic-scale, time-dependent ice-flow characteristics. Here, using spatially and temporally dense synthetic aperture radar (SAR) observations, we extend to the time domain existing methods for inferring velocity fields in three spatial dimensions from SAR and optical data. Our approach yields some of the first-ever truly 4D velocity fields. We briefly outline the underlying methodology before detailing the resulting velocity fields, which consist of a mean velocity vector, a user-defined set of temporal functions, and a DEM correction factor. Rutford Ice Stream (RIS), WAIS, provides an ideal proof-of-concept location because of its strong, well-documented, tidal-timescale ice flow variations. Existing GPS observations of RIS ice flow constrain the dominant periods of sinusoidal ice flow variability, which we use to define a set of sinusoidal functions used in the 4D velocity fields. COSMO-SkyMed, a 4-satellite constellation, collected SAR data for approximately 9 months beginning in late summer 2013 at sampling rates sufficient to distinguish the primary periods of tidally induced ice flow variability. The results, obtained using more than 1000 SAR-derived displacement fields, provide a map-view of tidally induced flow variations and tidal-timescale grounding line migration that indicate greater spatial heterogeneity in the response of the ice stream to tidal forcing than do GPS observations. Features in the 4D velocity fields are attributable to areas with known rigid and deformable beds. We compare the spatial characteristics of ice flow variability with the ice stream geometry and hydraulic potential, and discuss the implications of our results for mechanical models of the RIS bed.