Derivation of land cover and soil moisture from airborne polarimetric SAR-data

With its possibility to retrieve land cover, vegetation parameters and soil moisture, multifrequent and multipolarimetric SAR data have shown their potential to provide areal parameters for hydrological water balance and solute transport modelling. Over the drinking e I water reservoir catchment Zeulenroda (Germany) eperimental airborne X- and full polarimetric L-Band SAR data were acquired for land cover classification and soil moisture determination. The L-Band backscatter was decomposed into three eigenvalues and three eigenvectors providing information about the backscatter mechanism (alpha-angle) and the number of scattering processes (entropy H). Both parameters were plotted in the entropy-alpha-feature space to distinguish different land cover classes. The achieved overall classification accuracy was above 90%. The soil moisture information was derived from linear regression analysis of SAR data (intensity, polarimetry. principal components) with the ground truth TDR-measurements. The best conformity reached was expressed by a correlation coefficient of r = 0.85. The results could be seen as basic research on L-band polarimetry and backscatter signature information for future spaceborne missions like ALOS.