A Vegetation Correction Methodology Applied for Soil Moisture Retrieval from C-Band Radar Observations

This research presents a methodology to correct backscatter (sigmadeg) observations for vegetation effects. The proposed methodology is based on the concept that the ratio between the surface scattering over the total amount of scattering (sigmadeg_surface/sigmadeg_soil) is affected only by the vegetation and can be described as a function of the vegetation water content. The data set used in this study was collected at USDA´s Optimizing Production Inputs for Economic and Environmental Enhancement (OPE³) experimental site in Beltsville, Maryland (USA) over a corn growth cycle in 2002 and includes C-band (4.75 GHz) HH- and VV-polarized observations acquired at incidence angles of 15, 35 and 55 degrees. During this period the corn crops reached peak biomass of 6.6 kg m^-2 and a soil moisture range varying from 0.02 to 0.26 cm³cm^-3. The results show that through application of the proposed vegetation correction methodology the soil moisture retrieval accuracy can be improved from 0.033 to 0.032 cm³cm^-3, 0.049 to 0.033 cm³cm^-3, and 0.079 to 0.047 cm³cm^-3 at incidence angles of 15, 35 and 55 degrees, respectively.