An Algorithm for Retrieving Soil Moisture Using L-Band H-Polarized Multiangular Brightness Temperature Data

Abstract

This letter presents an algorithm for retrieving soil moisture using only H-polarized multiangular brightness temperature at L-band. We developed a parameterized surface model based on a simple-empirical model, the $H_{p}$ model, for this retrieval algorithm. By analyzing a simulated database using the advanced integral equation model (AIEM), it was found that the roughness variable $H_{h}$ can be parameterized as a function of an effective roughness parameter $Sr=(k_{L}\cdot s)^{2-N}(s/l)^{N}$ . Influences of three surface roughness parameters (e.g., rms height, correlation length, and type of autocorrelation function) required to describe a rough surface on surface reflectivity were all considered in this parameterized model. Comparison with AIEM simulations over a wide range of soil conditions indicates a good performance of this model. Then, based on the $\omega-\tau$ model, this algorithm is applied on refined SMOS H-polarized multiangular brightness temperature. Retrieved soil moisture in Africa exhibits reasonable patterns and temporal changes. Validation using in situ soil moisture from Little Washita watershed and Yanco over 2010–2011 showed fine accuracy with root-mean-square errors of 0.031 and 0.045 m3/m3 for two areas, respectively.