Sahelian-Grassland Parameter Estimation from Backscattered Radar Response

In recent years a special emphasis has been placed on the retrieval of physical parameters from polarimetric radar at microwave frequencies in many research programs. In this paper, we adapted a technique based on an empirical model and a genetic algorithm, and verify its applicability for a complex class of vegetation within a wide temporal interval. This complex class of vegetation is Sahelian grassland which is mainly composed of annual grass and shrubs. The proposed retrieval algorithm is conformed of 3 main steps: (1) Identification of sensitive parameters, (2) Development of the empirical model, and (3) Implementation of a genetic algorithm for the inverse process. For this class of vegetation the sensitive parameters are: the soil moisture content m_s, the grass density D, and the grass moisture content m_v. When applying the retrieval algorithm to simulated radar responses, a great agreement (an error of 6% when estimating the soil moisture content, 13% for the grass density, and 18% for the grass moisture content in the adult-plant stage) is observed between input parameters and estimated ones.