The accurate monitoring of snow water equivalent (SWE) is critical to assess global water resources. Despite the importance, we still lack a basic understanding of how much snow is seasonally stored on Earth, especially in mountains where a major part of the terrestrial snow is located. The central hypothesis of Snowtrane is that Machine Learning Algorithms (MLA), more specifically Artificial Neural Networks, can be trained to address physical snow backscattering mechanisms that are too complex to embed in an empirical change detection algorithm or even in physics-based radiative transfer models that still remain a simplification of reality through imposed model assumptions and boundary conditions. In this project, we hypothesize that snow retrievals based on MLA outperform those from empirical change detection techniques, especially when also auxiliary EO data sets are included as input to further explain the scattering processes at hand.