Model-based Inversion of Soil Parameters under Vegetation using Ground-to-Volume Ratios

An innovative model-based inversion of soil moisture and soil roughness under vegetation cover is developed using polarimetric ratios from the X-Bragg model and the modified three component decomposition. Normalized ground-to-volume ratios are incorporated for a substantial improvement of the estimation for soil parameters by splitting the cross-polarized scattering component SXX into a roughness and a vegetation contribution. Results for soil moisture and soil roughness estimation under a variety of vegetation types covering the vegetation growth period are presented for the multi-temporal L-band data set of the AgriSAR campaign conducted in 2006. For the soil roughness and the soil moisture a root mean square error for all investigated dates and crop types of 0.1 and 6.7Vol.% is achieved. Despite the promising results for inversion of the soil parameters the standard deviation of the inversion estimates report a significant uncertainty, which has to be further investigated.