Assessment of SMOS soil moisture retrieval parameters using tau-omega algorithms for soil moisture deficit estimation

•Soil moisture retrieval using tau-omega algorithm.•Evaluation of soil moisture retrieval parameters for SMOS soil moisture estimation.•Integration of WRF-NOAH Land Surface Model and MODIS based LST in soil moisture retrieval algorithms.•Assessment of SMOS retrieved soil moisture from different approaches for SMD prediction.•Highest performance is shown by WRF-NOAH Land Surface Model based LST.

SummarySoil Moisture and Ocean Salinity (SMOS) is the latest mission which provides flow of coarse resolution soil moisture data for land applications. However, the efficient retrieval of soil moisture for hydrological applications depends on optimally choosing the soil and vegetation parameters. The first stage of this work involves the evaluation of SMOS Level 2 products and then several approaches for soil moisture retrieval from SMOS brightness temperature are performed to estimate Soil Moisture Deficit (SMD). The most widely applied algorithm i.e. Single channel algorithm (SCA), based on τ-ω is used in this study for the soil moisture retrieval. In τ-ω, the soil moisture is retrieved using the Horizontal (H) polarisation following Hallikainen dielectric model, roughness parameters, Fresnel's equation and estimated Vegetation Optical Depth (τ). The roughness parameters are empirically calibrated using the numerical optimization techniques. Further to explore the improvement in retrieval models, modifications have been incorporated in the algorithms with respect to the sources of the parameters, which include effective temperatures derived from the European Center for Medium-Range Weather Forecasts (ECMWF) downscaled using the Weather Research and Forecasting (WRF)-NOAH Land Surface Model and Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature (LST) while the τ is derived from MODIS Leaf Area Index (LAI). All the evaluations are performed against SMD, which is estimated using the Probability Distributed Model following a careful calibration and validation integrated with sensitivity and uncertainty analysis. The performance obtained after all those changes indicate that SCA-H using WRF-NOAH LSM downscaled ECMWF LST produces an improved performance for SMD estimation at a catchment scale.