Surface Soil Moisture Retrieval Using the Improved Water-Cloud Model Based on Sentinel-1A and Sentinel-2 Data

Surface soil moisture is an essential component of the ecosystem, an important factor in agricultural development, and a primary environmental factor for crop growth. This paper presents an improved water-cloud model for retrieving surface soil moisture (SSM) in wheat fields, utilizing the combined C-band multi-polarimetric Sentinel-1 Synthetic Aperture Radar (SAR) and multispectral Sentinel-2 optical data. To eliminate the impact of vegetation on SSM retrieval, vegetation indices were obtained from Sentinel-2 data to estimate the vegetation water content (VWC). And then, these estimated results with vegetation coverage substituted into the water-cloud model. Additionally, support vector regression (SVR) algorithm was used to retrieve the SSM. In overall, the key results of our study were: (1) the normalized difference red edge Index (NDREI) obtained from red edge bands of Sentinel-2 was the most suitable for removing the effects of wheat cover on SSM retrieval; (2) the improved water-cloud model was an effective SSM retrieval method for wheat-covered regions, with the determination coefficient (R2) between the retrieved and measured SSM was 0.867 and that the root mean square error (RMSE) was 0.062 cm3/cm3. The results satisfied the requirement for SSM retrieval in the study region.