Sensitivity of different scattering mechanisms to soil moisture and vegetation over corn fields in Argentina

Soil moisture is fundamental for several land applications that involve agricultural practices, such as irrigation management, crop growth monitoring, and the new precision farming approach. Thanks to the growing technological development in the field of Earth observation and remote sensing, Synthetic Aperture Radar (SAR) data can be used to estimate soil moisture with high accuracy and frequent temporal sampling. Unfortunately, the radar signal is influenced not only by changes in soil moisture, but also by the vegetation (i.e., plant height, growth stage, etc.) and other factors such as surface roughness. In this context, a possible solution could be given by the use of a polarimetric SAR decomposition, which could help disentangle the effects of vegetation from those of the soil by separating different scattering mechanisms, i.e., surface, double-bounce and volume. The objective of this study is to analyze which scattering mechanisms are mostly correlated to soil moisture or vegetation in order to evaluate their exploitation within a soil moisture retrieval algorithm. In fact, surface and double-bounce, which are known to be functions of the dielectric properties of the soil (surface) and soil-vegetation (double-bounce) are the contributions that could be used to estimate soil moisture after the vegetation effects has been removed (i.e., volume term) from the measured data [1]. A set of different polarimetric SAR decompositions, such as the Freeman-Durden three-components decomposition [2] or the Nonnegative Eigenvalue Decomposition [3], are applied to a time-series of geocoded covariance/coherency matrices (C3/T3). The measured C3/T3 matrices are derived from full-polarimetric L-band radar data collected by the SAOCOM-1A mission over five corn fields located in the Monte Buey site, Argentina, between October 2019 and February 2020. In particular, 9 descending SLC (Single Look Complex) images with a nominal spatial resolution of 10m x 6m in ground range and azimuth, respectively, are considered. The temporal evolution of the scattered powers related to each scattering contribution, along with the backscattering coefficients at different polarizations, will be evaluated with respect to in-situ data (i.e., soil moisture, plant height, crop growth stage) collected by the Argentinian Space Agency (CONAE) during the 2019-2020 season and concurrently with the SAOCOM acquisitions. Satellite-derived parameters, such as the Normalized Difference Vegetation Index (NDVI), will be also involved in the analysis as a proxy of the vegetation growth conditions. Then, the results obtained by using the polarimetric decomposition approach will be compared with the ones obtained by the application of the fully polarimetric model developed at Tor Vergata University. In fact, this electromagnetic model is able to simulate the radar backscatter along with the three different contributions (surface, soil-vegetation interaction, and volume) [4]. The in-situ data collected during the 2019-2020 field campaign will be used as a training dataset for the Tor Vergata model. Finally, a comparison will also be made with the results obtained by applying an empirical model, the Water Cloud Model (WCM), which is used to correct the vegetation effects on the radar backscatter [5].