Empirical Removal of Tides and Inverse Barometer Effect on DInSAR From Double DInSAR and a Regional Climate Model

Abstract

Ice shelves—the floating extensions of the Antarctic ice sheet—regulate the Antarctic contribution to sea-level rise by restraining the grounded ice flowing from upstream. Therefore, ice-shelf change (e.g., ice-shelf thinning) results in accelerated ice discharge into the ocean, which has a direct effect on sea level. Studying ice-shelf velocity allows the monitoring of the ice shelves’ stability and evolution. Differential synthetic aperture radar interferometry (DInSAR) is a common technique from which highly accurate velocity maps can be inferred at high resolution. Because ice shelves are afloat, small sea-level changes—i.e., ocean tides and varying atmospheric pressure (aka inverse barometer effect) lead to vertical displacements. If not accounted for in the interferometric process, these effects will induce a strong bias in the horizontal velocity estimation. In this article, we present an empirical DInSAR correction technique from geophysical models and double DInSAR, with a study on its variance propagation. The method is developed to be used at large coverage on short timescales, essential for the near-continuous monitoring of rapidly changing areas on polar ice sheets. We used Sentinel-1 SAR acquisitions in interferometric wide and extra -wide swath modes. The vertical interferometric bias is estimated using a regional climate model (MAR) and a tide model (CATS2008). The study area is located on the Roi Baudouin Ice Shelf in Dronning Maud Land, East Antarctica. Results show a major decrease (67 m $\cdot$ a $^{-1}$ ) in the vertical-induced displacement bias.