Exploring Sentinel-1 backscatter time series over the Atacama Desert (Chile) for seasonal dynamics of surface soil moisture

Recent research indicates an inverse relation between Synthetic Aperture Radar (SAR) signal and near-surface soil moisture (SM) over very dry sediments, arid to hyper-arid soils resp., caused by subsurface scattering effects. This phenomenon can lead to large errors when it comes to modelling and remote sensing-based estimation of SM. While the effect of subsurface scattering and its influence on SM estimates is well described and modelled in literature, its actual presence in recorded SAR data is largely unknown. Here we investigate the relation between C-Band SAR backscatter and SM in the hyper-arid environment of the Atacama Desert (Chile). Time series (2018–2020) of Sentinel-1 VV/VH intensities are compared to in situ SM, measured at 17 stations located across the Atacama Desert. Linear and non-linear regression modelling is applied to uncover the relationship between the SAR intensities and in situ SM, while, in addition, SM variations triggered by seasonal varying humidity (i.e., not by precipitation) are investigated. Results indicate (i) a very weak linear relationship between SM variations and SAR intensities (VV/VH) for most meteorological stations in the Atacama (R2