Validation of CryoSat-2 SAR mode based lake levels

Here we evaluate the precision of the along-track mean lake levels derived from CryoSat-2 data for five lakes; Skanderborg sø, Mossø, and Arresø (Denmark), Okeechobee (US), and Vänern (Sweden) with a surface area ranging from 9 to 5600 km2. Three of the lakes are compared with gauge data. As a reference we compare our results with water levels obtained from Envisat. We find that the along-track precision of the mean based on CryoSat-2 is a few centimeter, even for the small lakes, which is a significant improvement compared to previous missions such as Envisat. When validating against gauge data we find RMS values of differences between the estimated lake levels and in-situ gauge measurements of only 8 cm or less, after bias correcting the estimated lake levels. To estimate the along-track mean values we apply a novel approach. Instead of attempting to identify and remove the polluted observations we use a mixture distribution to describe the observation noise, which prevents the polluted observations from biasing our final reconstructed time series. These results demonstrate the promising possibilities of the upcoming mission Sentinel-3, which potentially will be able to provide accurate time series even for small lakes.