Snow water equivalent in high-mountain regions

Climate change has significantly impacted the high-mountain cryosphere, and understanding precipitation patterns is essential for assessing these changes but precipitation trends with elevation remain unclear. A major source of uncertainty is the limited spatial and temporal coverage of ground-based observations in high-mountain regions due to their remoteness and the harsh conditions that make it challenging to maintain weather stations. Snow water equivalent (SWE) observations are also sparse but can be obtained from glacier monitoring. SWE measurements on glaciers are thus typically the only few direct ground observations that can represent precipitation at high elevations. Thus, the presented research aims at (1) quantifying precipitation biases at high elevation by exploiting SWE observations on glaciers and (2) developing data-driven methods (multiple linear regression, gradient boosting, neural network) to provide continuous SWE estimates from local to global scale, by fusing different data sources (remote sensing observations, reanalyses, topographical descriptors). These methods hold great promise for improving SWE estimates in various regions worldwide and across different spatial scales, thus potentially providing crucial information for the development of effective water management strategies in the face of climate change.