Analyzing and modeling the SMOS spatial variations in the East Antarctic Plateau

The SMOS brightness temperature (TB) collected on the East Antarctic Plateau revealed spatial signatures at L-band that have never before been observed when only higher-frequency passive microwave observations were available, and this has opened up a new field of research. Because of the much greater penetration depth, modeling the microwave ice sheet emission requires taking into account not only snow conditions on the surface, but should also include glaciological information. Even if the penetration depth of the L-band is not well known due to the uncertainty on the imaginary part of the ice permittivity, it is likely to be of the order of several hundreds of meters, which means that the temperature of the ice over a depth of nearly 1000 m influences the emission. Over such a depth, the temperature is related to both the surface conditions and to the ice sheet thickness, which in turn depends on the bedrock topography and on other glaciological variables. The present paper aims to provide a thorough theoretical explanation of the observed TB spatial variation close to the Brewster angle at vertical polarization, in order to limit the effect of surface and vertical density variability in the firn. In order to provide reliable inputs to the microwave emission models used for simulating TB data, an in-depth analysis of the temperature profiles was performed by means of glaciological models. The comparison between simulated and observed data over three transects totalling 2000 km in East Antarctica pointed out that, whereas the emission models are capable of explaining the TB spatial variations of several kelvins (0.7 and 2.9 K), they are unable to predict its absolute value correctly. This study also shows that the main limiting factor in simulating low-frequency microwave data is the uncertainty in the currently available imaginary part of the ice permittivity.