Abstract. We introduce a variational data assimilation scheme to assimilate X-band Synthetic Aperture Radar (SAR) data into a snowpack evolution model. The structure properties of a snowpack, such as snow density and grain optical diameter of each layer, are simulated over a period of time by the snow metamorphism model Crocus, fed by the local reanalysis SAFRAN at a French alpine location. These parameters are used as inputs of an Electromagnetic Backscattering Model (EBM) based on Dense Media Radiative Transfer (DMRT) theory, which calculates the simulated total backscattering coefficient. Next, 3D-VAR data assimilation is implemented in order to minimize the discrepancies between model simulations and observations obtained from SAR acquisitions, by modifying the parameters of a multilayer snowpack calculated by Crocus. The algorithm then reinitializes Crocus with the optimized snowpack structure properties, and therefore allows it to continue the simulation of snowpack evolution where adjustments based on remote sensing data has been taken into account. Results obtained using TerraSAR-X acquisitions on Argentière Glacier (Mont-Blanc massif, French Alps) show the high potential of this method for improving snow cover simulation.
3D-VAR multilayer assimilation of X-band SAR data into a detailed snowpack model
