Abstract ID: 3.10982 | Accepted as Talk | Talk/Oral | TBA | TBA

Detailed information on the extent and state of the seasonal snow in high mountain regions is needed for applications in snow hydrology, water management and in the field of climate impact. Due to the high spatial variability of seasonal snow in space and time, high resolution satellites provide efficient means for comprehensive snow monitoring in high mountain terrain. We report on the development of an advanced method for monitoring snow extent from multiple optical satellite data optimized for scientific and operational application in mountain areas. Regarding snow extent, we developed a multi-spectral unmixing approach with locally adaptive endmember selection (LAMSU) that accounts for variations in illumination across mountainous terrain and offers flexibility regarding the optimum use of spectral sensor capabilities. Our approach separates regions illuminated by the sun from shaded regions using spectral classification rules for detecting different snow free and fully snow covered endmembers by applying adapted spectral band combinations. The algorithm is designed to provide consistent snow extent estimates from satellite sensors with different spatial resolution and spectral channels, such as sensors of the Copernicus Sentinel-2 and Sentinel-3 missions. By combining both satellite missions, we provide daily medium resolution snow products (300m) from Sentinel-3 SLSTR / OLCI together with high resolution snow products with 20m pixel size from Sentinel-2, acquired every few days over mountain regions. Maps of uncertainty are attached to the snow extent products. Snow extent from optical satellites can be combined with snow wetness products from Sentinel-1 SAR data to classify the snow state conditions (wet/dry). A change detection algorithm is applied exploiting the strong decrease of backscatter for wet snow in comparison to snow free conditions and dry snow. An uncertainty measure for the wet snow detection is provided, accounting for dual-frequency backscatter intensity and speckle statistics. In the presentation we outline the snow mapping procedure, show examples of snow products for different mountain regions worldwide, and report on the quality of the products in comparison with snow information from other sources.