Retracking Cryosat data in the SARIn mode and robust lake level extraction

Lake levels are an indicator for the hydrological state of an area. Lake levels are directly and indirectly affected by meteorological variations and are both indicators for and caused by water management. Therefore, it is important to monitor the changes of water levels in lakes. As an alternative to water level gauges, satellite radar altimetry is an established method. The problem of radar altimetry so far is that it only gives reliable lake levels for relatively large lakes. The recent Cryosat mission passes many lakes in the so-called SARIn mode at improved along-track resolution. Still, the accuracy and robustness of standard level 2 is limited due to pollution of the waveform by surrounding land topography. We propose to actively extract the water surface component by a novel retracking method based on cross-correlation of observed level 1b waveforms with a generic simulated waveform. Different scatterers will result in different cross-correlation solutions, but, as the water level is contributing to each consecutive waveform, it can efficiently be extracted by a majority voting scheme. To demonstrate the performance of the proposed retracker, we compared the estimated lake levels with accurate Jason-2 estimates over Lake Nasser in Egypt. The proposed retracking algorithm performs better than the level 2 product, and has a RMS of differences with Jason-2 lake levels of 0.30 m. The remaining differences with Jason-2 data are correlated to the amount of topographic pollution, but also show some remaining surface slope artifact. By using the proposed method, one can optimally profit from Cryosat measurements for lake level measurements. Not only can we retrieve lake levels closer to the shore, but due to its dense ground track spacing (7–8 km at the Equator) it is possible to estimate lake levels over many more lakes than with other altimeters. Cryosat lake level estimates are therefore a promising extension of existing time series as measured by notably Envisat, Jason-1 & 2 and GFO.