Soil Moisture Profile Estimation by Combining P-band SAR Polarimetry with Soil Hydrological Modeling

Kurzfassung

Soil moisture is a key hydrological variable with great influence on various processes of land-atmosphere interactions, like infiltration or subsurface flow [1]. Depending on the frequency, microwave remote sensing can be used to estimate soil moisture near the surface (~0-5 cm, L-band) [2] or at deeper layers (~20-30 cm, P-band) [3]. However, typical soil moisture retrieval methods can only provide one value within the vertical integral from the land surface to the penetration depth of the microwaves. However, since radar backscatter is able to provide information about soil moisture discontinuities, an advanced method is needed to acquire evidence about the moisture distribution across the vertical soil column. The knowledge about the continuous soil moisture profile is important for many environmental applications, like land surface modeling or weather and climate monitoring [4]. Thus, a joint approach of remote sensing and hydrological modeling is proposed to estimate soil moisture profiles. First, the hybrid decomposition method of [3] is used to separate polarimetric P-band SAR observations from the AirMOSS campaign [5] into the individual scattering mechanisms (soil, dihedral, vegetation). That way the polarimetric scattering angle α_s^Data is estimated from the soil scattering component. Second, soil moisture profiles are simulated with the HYDRUS-1D, which models water flow in variably saturated, layered media by numerically solving the Richard’s equation [6]. In this study, an ensemble of soil moisture profiles with varying assumptions on, e.g. the number of soil layers or the initial soil matrix potential, are simulated. Every profile is then used to estimate the backscatter coefficient σ_PQ^o using the multi-layer small perturbation method, which models the backscattering from multiple soil layers with depth [7]. Based on modeled σ_PQ^o, the soil scattering angle α_s^Model [8] is calculated for every simulated profile. Lastly, the soil moisture profile is estimated from the best fit between observed α_s^Data and simulated α_s^Model, leading to the most representative continuous soil moisture profile. First profile estimates are conducted at single locations within three AirMOSS sites (Walnut Gulch, AZ, MOISST, OK, and Harvard Forest, MA) and correlate with precipitation events. An in-depth validation with in situ measured soil moisture is currently underway.