Systematic errors in Cryosat-2 swath elevations and their impacts on glacier mass balance estimates

Almost ten years ago, the first elevation estimates based on swath processing of interferometric CryoSat-2 altimeter observations were published, mapping the surface of Devon ice cap. The new method holds a great potential to provide dense data coverage, in space and time. Indeed, ESA recently started releasing digital elevation models at a 2 by 2 km resolution for a rolling 3 month data aggregation cycle. Such spatiotemporal resolutions are especially valuable in versatile and dynamic regions as mountain glaciers. In this presentation, we describe systematic errors on the order of 10 m with about yearly periodicity that arise in the proximity of hills and valleys. One error is caused by the superposition of multiple signals, the other is caused by the Fourier-transformation in the SAR beam-forming process. Both are intrinsic to the measuring concept, but their effect can potentially be limited by data filtering strategies. We report the influence of the commonly used coherence and power threshold based filtering on derived elevation change rates. For data users, awareness of these issues is especially important to interpret the observations correctly and to understand that there is a large systematic part in the overall uncertainty.