Abstract:
Seasonal snow cover, characterized as either snow water equivalent (SWE) or snow depth (SD), defines water supply and has major impacts on climate. Satellite observations have a high potential for estimating the spatio-temporal variability of SWE. However, current satellite products, primarily based on microwave radiometer observations, are only available at coarse (e.g., 25 km) resolution. Moreover, these observations are only provided for dry snow conditions and are known to saturate at high values of SWE. Applications using high-resolution synthetic aperture radar (SAR) observations have so far been limited due to their generally low revisit frequency.

This study demonstrates the ability of the European Space Agency (ESA) C-band (5.4 GHz) Sentinel-1 mission to estimate SWE or SD at 1-km resolution with a repeat cycle of 6 days. The Sentinel-1 cross-polarization ratio (that is, the ratio of vertical-horizontal (VH) to vertical-vertical (VV) polarization backscatter) is shown to be sensitive to snow for both dry and wet snow conditions. Its potential for SWE and SD estimation is investigated over the European Alps and the Sierra Nevada in California (USA), two areas with challenging topographic conditions. Over both areas, the spatial distribution of SWE derived from Sentinel-1 reveals a clear dependence on topography and forest cover. For areas with less than 60% forest cover, the comparison of Sentinel-1 retrievals with in situ measurements results in correlations of 0.52-0.64 and normalized RMSE of 16%. Corresponding global model simulations and in situ measurements correlate to 0.27-0.71, with RMSE of 22%. The results highlight the potential of Sentinel-1 for area-wide snow estimates at unprecedented spatial (1 km) and temporal (weekly) resolution.