Soil Moisture Retrieval Using Modified Vegetation Backscattering Model Based on Radarsat-2 Data

This article proposed a modified vegetation backscattering model to retrieve soil moisture using C-band RADARSAT-2 images and field measurements over agricultural study sites in Shaanxi province of China. The effects of vegetation and soil on radar signals were separated at pixel level by introducing vegetation coverage in the modified model. The direct scattering contribution from the underlying ground surface was considered as an important component in the total backscattering at agricultural sites. The results indicated that the modified model could be effectively applied to a variety of surface cover types ranging from sparse to full vegetation cover. The accuracy of soil moisture retrieval was significantly high with R2 and RMSE of 84.3% and 0.028 m 3 /m 3 , respectively. Therefore, the modified model was suitable for soil moisture retrieval at large scales by combining the advantages of SAR and optical remote sensing data.