Soil moisture retrieval from ASAR measurements over natural surfaces with a large roughnessvariability

In this work, the accuracy of soil moisture retrieved from ASAR data over bare or sparsely vegetated surfaces is investigated by means of a simulation study. The soil moisture retrieval method is based on an optimization algorithm that appropriately inverts theoretical direct models by assimilating a priori information on surface parameters. In order to account for a large variability of roughness conditions, two complementary models have been used, namely the Integral Equation Method model and the Geometrical Optics model. The performance of the inversion method has been assessed on simulated noisy ASAR data, as a function of different a priori information quality level. © 2005 IEEE.