COUPLED ASSIMILATION OF SATELLITE AND GROUND-BASED SNOW DATA IN DISTRIBUTED PHYSICALLY BASED MODEL OF SNOW COVER AND RUNOFF GENERATION

A technique of constructing spatial fields of snow characteristics based on assimilation of satellite land surface monitoring data and available ground-based hydrometeorological measurements in a physical based snowpack model is presented. The constructed snow characteristics fields are used as inputs of distributed physically based model of runoff generation. The study area comprises the Vyatka River basin with the catchment area of 124,000 km2 and the Upper Don basin with the catchment area of 101800 km2. The snowpack model provides simulation of temporal changes of the snow depth, density and water equivalent (SWE). accounting for melt, sublimation, refreezing melt water and snow metamorphism processes with a special focus on forest cover effects .The model was first calibrated against available ground-based snow measurements and then applied to calculate the spatial distribution of snow characteristics in each 0.01o pixel of the spatial grid using satellite data and interpolated ground-based meteorological data. The satellite data include fractional snow-covered area (SCA), the NASA daily maps of SWE, albedo and snow surface temperature as well as the NOAA maps of land cover classification and tree cover fraction. The NASA maps of SWE were applied for assigning initial conditions in the snowpack model. Calculated values of SCA were compared to the SCA defined from the NASA MOD10_L2 maps. For sparsely forested area the calculated values of SCA turned up close to the values defined from the satellite data while for the area with dense forest there were significant differences between these values. The physically-based models of runoff generation for both river basins have been used to compare the hydrographs simulated by using the constructed SWE fields and the NASA daily maps of SWE. These models are based on a finite-element schematization of the catchment area and include the processes of snow cover formation and snowmelt, freezing and thawing of soil, vertical moisture transfer and evaporation, surface water detention, overland, subsurface and channel flow. The comparison of simulated hydrographs with the observed ones has shown that the suggested procedure of assimilation of satellite data gives a higher accuracy of representation of spatial distribution of snow cover and hydrograph simulations than the direct use of the NASA satellite SWE data.