On the synergy of SMAP, AMSR2 AND SENTINEL-1 for retrieving soil moisture

An algorithm for retrieving soil moisture content (SMC) from synergic use of both active and passive microwave acquisitions is presented. The algorithm takes advantage of the integration of microwave data from SMAP, Sentinel-1 and AMSR2 for overcoming the SMAP radar failure and obtaining a SMC product at enhanced resolution (0.1 degrees x 0.1 degrees) and improved accuracy with respect to the original SMAP radiometric SMC product. A disaggregation technique based on the Smoothing filter based intensity modulation (SFIM) allows combining the radiometric and SAR data. Disaggregated microwave data are used as inputs of an Artificial Neural Networks (ANN) based algorithm, which is able to exploit the synergy between active and passive acquisitions. The algorithm is defined, trained and tested using the SMEX02 experimental dataset and data simulated by forward electromagnetic models based on the Radiative Transfer Theory. Then the algorithm is adapted to satellite data and tested using one year of SMAP, AMSR2 and Sentinel-1 co-located data on a flat agricultural area located in the Po Valley, in northern Italy. Spatially distributed SMC values at 0.1 degrees x 0.1 degrees resolution generated by the Soil Water Balance Model (SWBM) are considered as reference for this purpose. The synergy of SMAP, Sentinel-1 and AMSR2 allowed increasing the correlation between estimated and reference SMC from R congruent to 0.68 of the SMAP based retrieval up to R congruent to 0.86 of the combination SMAP + Sentinel-1 + AMSR2. The corresponding Root Mean Square Error (RMSE) decreased from RMSE at 0.04 m(3)/m(3) to RMSE congruent to 0.024 m(3)/m(3).