Remote sensing studies at the "Laboratoire de Glaciologie et de Géophysique de l'Environnement" are motivated by the possibility to measure from space important geophysical parameters such as the surface melt extent, the temperature and the snow accumulation rate of the polar ice sheets. Moreover, microwave data are weakly sensitive to atmospheric effects, and they provide information on a few centimeters or several meters of the snow cover. A semi-empirical model based on radiative transfer and a simple thermodynamic model, enable us to retrieve the measured brightness temperatures (SSM/I data) from the air temperatures measured by automatic weather stations; a constant emissivity is assumed and we fit the extinction coefficient of the medium. This method is tested on two Antarctic sites with different snow characteristics. Afterwards, the model is reversed in order to retrieve the snow surface temperature from the satellite data. At this point, the problem is still underdetermined, but the results are promising. In order to validate models, we measured radiometric data at different frequencies and incident angles, with the radiometer PORTOS (CNES) from a cable car, at Chamonix in the Alps. Simultaneously, the snow cover was characterized. These data are calibrated, corrected from local incident angle, the antenna pattern effect on the angular measurement is evaluated. Then, they are compared to computed data from a snow emissivity model for a uniform and a stratified medium and from a radiative transfer mode!. The problems of the snow structure characterization for microwaves and of the scattering by snow grains are raised.
