Investigating Various Spectral Resolution Scenarios on Predicting Soil Hydraulic Properties

Pedotransfer functions (PTFs) have been developed to indirectly predict soil hydraulic properties (SHPs) from easily measurable soil properties mainly including textural properties, soil organic matter and bulk density. In the last few decades, several studies have addressed the potential of soil spectral information in visible, near-infrared (350-2500 nm), to provide predictors to estimate elementary soil properties. Predicting SHPs by soil spectral data is a new approach that has not yet been explored. In this study, the feasibility to estimate the Mualem-van Genuchten (MvG) hydraulic parameters was investigated using Spectro Transfer Functions (STFs). Four scenarios of data affrication namely: ASD full spectrum (scenario I), EnMAP (scenario II), Sentinel-2 (scenario III) satellite-based spectral resolution and laboratory and soil map-based Rosetta and HYPRESPTFs (scenario IV) were investigated. A Stepwise Multiple Linear Regression (SMLR) coupled with bootstrap method was employed to derive STFs. The most appropriate results for predicting MvG parameters were obtained for scenarios I and II. Compared with scenario IV, all the other three spectral scenarios performed reasonably well in terms of predicting soil water retention characteristics and unsaturated hydraulic conductivity. These findings suggest that spectral reflectance data at various spectral resolution levels is a promising indirect and quick method for large scale soil hydraulic parameter estimations.