Snowmelt dynamics observed by dense X-band time series acquired by COSMO-SkyMed constellation

The seasonal snow is one of the largest water reservoirs in nature, storing water during winter, and gradually releasing it in spring during the melt. This guarantees freshwater supply for the lowlands even in the long term, making the mountains the “water towers” of the downstream regions. In fact, the delayed water release from the head watersheds to the forelands is essential for a large number of human activities such as irrigation, drinking water supply and hydropower production. On the other hand, snowmelt may cause natural disasters such as wet-snow avalanches, gliding or release of highly enriched accumulated contaminants able to cause severe impact on water quality.

In recent years, Synthetic Aperture Radar (SAR) has demonstrated capable to provide information about the melting process. In particular, with the launch of the European Commission (EC) Copernicus Programme Sentinel-1 mission, C-band SAR images are regularly acquired every 6 days and delivered free of charge. This opened the possibility to observe a phenomenological relationship between the snow melting process of high altitude snowpacks and the multi-temporal radar backscattering acquired by Sentinel-1. The identification of the temporal signature for each pixel of a Sentinel-1 time series allowed us to detect the onset of the three phases that made up the snowmelt i.e., moistening, ripening and runoff, with a good reliability. However, the mechanisms that drive the snowpack response at microwaves depend on frequency; therefore, different snowpack signatures are expected if using different frequency bands, as the X band available onboard the Italian Space Agency (ASI)’s COSMO-SkyMed (CSK) constellation.

In this work, we analyze a dense X-band time series acquired by the CSK over the Schnalstal catchment in Italy during the snowmelt season. This allows us to point out the similarities and the differences between the electromagnetic interactions using C- and X-band SAR during the snowmelt. Depending on the shorter wavelength, the X-band is more sensitive than C-band to small quantities of liquid water inside the snowpack. Therefore, X band shows an earlier response than C band to the moistening of the surface snow layer (especially for steep local incidence angles), and a more pronounced loss of interaction with deeper layers. X-band is also more sensitive to the increase in the superficial roughness with the consequence of possibly anticipating the runoff onset. However, by comparing the runoff time in the Schnalstal catchment during the melting season 2020-2021, a general agreement between C- and X-band is found even though the characteristic shape of the signature exhibits more variations at X-band than C-band.

This research is part of the 2019-2022 project ‘Development of algorithms for estimation and monitoring of hydrological parameters from satellite and drone’, funded by ASI under grant agreement n. 2018-37-HH.0.