Soil moisture prediction of bare soil profiles using diffuse spectral reflectance information and vadose zone flow modeling

Soil hydraulic property information of the vadose zone is key to quantifying the temporal and spatial variability of soil moisture, and for modeling water flow and contaminant transport processes in the near surface. This study deals with exploring the feasibility of using diffuse soil spectral information in the visible, near-infrared and shortwave infrared range (350–2500 nm) to estimate coarse-scale soil hydraulic parameters and predict soil moisture profiles using a topography-based aggregation scheme in conjunction with a 1D mechanistic water flow model. Three different types of parametric transfer functions (so-called spectrotransfer functions, STFs; pedotransfer functions, PTFs; and spectral pedotransfer functions, SPTFs) were aggregated from the point scale to 1 km² pixel size. to provide coarse scale estimates of van Genuchten-Mualem (VGM) hydraulic parameters. The coarse scale hydraulic parameters were evaluated by simulating soil water dynamics of the 1 km² pixels across the Zanjanrood River sub-watershed (ZRS) in northwest Iran. Resultant soil water states were compared with ground-truth measurements and advanced synthetic aperture radar (ASAR) estimates of soil water content. The topography-based aggregation scheme was found to provide effective values of the VGM hydraulic parameters across the ZRS study site. The coarse scale STFs performed best in terms of simulating surface, near-surface and subsurface soil water dynamics, followed by the coarse scale SPTFs and PTFs, which performed similarly. The average simulated soil water contents of the surface layer closely correlated with ASAR estimates during relatively wet periods. Simulated subsurface soil water dynamics matched well with the ground-truth measurements. These findings indicate the feasibility of using spectral data to predict VGM hydraulic parameters and, ultimately, to predict soil water dynamics at the larger scales.