Time series investigation of soil moisture estimation using compact polarimetry at L-band

Abstract

The applicability of a recently developed compact polarimetric decomposition and inversion algorithm for C-band to estimate soil moisture under growing agricultural vegetation cover is investigated using simulated L-band compact Polarimetric Synthetic Aperture Radar (PolSAR) data. The surface scattering component is separated from the volume scattering component through a model-based compact polarimetric decomposition under the assumption of a randomly oriented vegetation volume and reflection symmetry. The extracted surface scattering component is compared with two physically-based, low frequency surface scattering models such as Extended Bragg (X-Bragg) and Polarimetric Two Scale Model (PTSM). The algorithm is applied on a time series of simulated L-band compact polarimetric E-SAR data from the AgriSAR 2006 campaign over the Gormin test site in Germany. The compact PolSAR derived soil moisture is validated against in situ measurements. Including the entire growing season and three different crop types, the estimated soil moisture values indicate an overall RMSE of 9–12 vol.% and 9–15 vol.% using the X-Bragg and PTSM, respectively.