Reconstruction of snowpack dynamics in the High Atlas by assimilation of data from satellites, ground-station, and reanalysis products

Snowpack melting in the High Atlas constitutes the major source of freshwater for the semi-arid agricultural plains of central Morocco. Despite its critical importance for the region, basic questions about the High Atlas snowpack remain largely unanswered. The scarcity of ground data has been a major obstacle to investigating snow processes in the High Atlas. Here, we demonstrate the potential of assimilating remotely-sensed meteorological variables and downscaled climate reanalysis data to gain important insights into snow water balance in a semi-arid region. We apply a distributed energy balance snow model based on SNOW17, constrained by topographic data, meteorological data from satellites and high-resolution dynamically-downscaled ERA-Interim data, to simulate snowpack dynamics within the Oum-Er-Rbia watershed, at the heart of Morocco's High Atlas. The simulations are compared to MODIS snow cover maps and observed snow depth at one field station. The snowpack is essentially concentrated above 2500m, extends over 500-6000 km² and holds 0.05-0.4 km³ at its peak in early February. Additionally, we find that losses by sublimation range from 0.06-0.14 km³ for an average of 0.09 km³ a year, about 10% of all snowfall. This study reveals the essential components of the snow water balance in the High Atlas and paves the way for better understanding of its sensitivity to climate change.