Snow depth estimation at country-scale with high spatial and temporal resolution

Monitoring snow depth is important for applications such as hydrology, energy planning, ecology, and safety evaluation outdoor winter activities. Most methods able to estimate large regions can only do so in a spatial resolution of up 1 km ground sampling distance (GSD). This limits their usage high alpine areas, where this fails capture local distribution patterns caused by the pronounced topographical features. In work we use recurrent convolutional neural network at (10 m GSD), weekly, scale based on satellite data sources elevation maps, without need measurement stations ground. The proposed method achieves unprecedented results large-scale, high-resolution mapping. resulting maps are evaluated over period three years against high-fidelity obtained with airborne photogrammetry. Finally, also produce well-calibrated uncertainty estimates every individual via probabilistic regression framework.