High resolution mapping of soil moisture by SAR: Data integration and exploitation of prior information

Abstract

Two different approaches to deal with the problem of estimating soil moisture content from SAR data in the presence of vegetation are presented. They exploit also the information about the biomass provided by ancillary optical data. The first method is suitable for sparse vegetation and is founded on the application of the well-known water cloud model. As for dense vegetation canopy, we have designed a model that expresses the variation of the component of the backscattering coefficient due to the soil characteristics as a function of the variations of the measured backscattering coefficient and of the biomass, assuming the availability of a time series of radar and optical data. To carry out the soil moisture retrieval, a multi-temporal inversion algorithm, based on the Bayesian MAP criterion, has been developed. It integrates all the samples of the time series of SAR data corrected for the vegetation effects. The approaches were evaluated on two case studies; the first one concerning an ENVISAT/ASAR observation of an agricultural site located in Northern Italy. The second test was performed on the AirSAR data collected during the SMEX02 experiment. The comparison between the estimated soil moisture contents and the in situ measurements has given encouraging results.