Assessing the Potential of Space‐borne C‐band SAR Data for Spatial Soil Moisture Information over a Large Area

Abstract

The backscattering coefficient σ0 from a time series of ERS‐2 SAR data showed high correlation (r2 =0.94) and high sensitivity to the observed surface soil moisture at a field scale. The observations agreed well with the simulations from the two surface scattering models tested in the study area. A single coincidence of the ERS‐2 and RADARSAT missions, with two different polarizations, provided an opportunity to estimate soil moisture over a large area by inverting the Dubois model. The estimates were validated with the field observations over a small area and are in good agreement with standard error of estimate (SEE) equal to 1.01 cm3cm−3. Furthermore, different soil moisture estimation methods were tested and compared with the observations to better understand the temporal and spatial variability of this important parameter. Soil moisture at a point was simulated using three hydrological models and compared with the radar backscattering and the point observations. There is a large variation in the responses among them. Simple statistical approaches viz., coefficient of variation and principal component analysis were used for mapping the spatial patterns of soil moisture over a large area using a time series of ERS‐2 data. These simple methods helped to map areas where the moisture availability is not varying much. This information about soil moisture distribution is of vital importance to the hydrological processes especially the variable source area concept.