Soil moisture retrieval from L-band measurements over a variety of agricultural crops

The objective of this work is to evaluate soil moisture retrieval methods over different canopy, with the view of developing an operational algorithm (for the SMOS project for instance). We tested different inversion process of the τ-ω model, using or not using ancillary information, to estimate soil moisture and τ-ω parameters (vegetation parameters and soil roughness) simultaneously under different crop types. The best general method (same for all data sets) provides a RMSE=0.059 m³/ m³ using no ancillary information and a RMSE=0.048 m³/ m³ if we use soil and vegetation a priori information. Moreover, using specific methods to each data set provide RMSE values between 0.04 and 0.05m³/ m³.