Retrieval of soil moisture using multi-temporal hybrid polarimetric RISAT-1 data

The capability of hybrid Polarimetric Synthetic Aperture Radar (PolSAR) to estimate soil moisture (under vegetation) was assessed within the agricultural region of Wallerfing in Lower Bavaria, Germany. A novel methodology based on a hybrid polarimetric decomposition together with a surface component inversion is developed to retrieve surface soil moisture. The model-based, hybrid decomposition technique is used assuming a randomly oriented vegetation volume to obtain the surface scattering component. After vegetation removal, the surface scattering component is inverted for soil moisture by comparison with the surface scattering component, modeled by the Extended Bragg (X-Bragg) model including a depolarization term. The developed algorithm is applied on multi-temporal hybrid polarimetric C-band RISAT-1 data acquired from April to October 2014. The estimated soil moisture values indicate an overall Root Mean Square (RMS) error of 6.2 to 8.5 vol.% accounting for the entire growing season and four different plant types.