A Land-Surface Heterogeneity Index to Classify Continental Scale Near-Surface Soil Moisture Dynamics

Abstract:

Big data and global-scale estimation of hydrological state variables has enabled reliable quantification of soil moisture dynamics at global/continental scales. However, the mechanistic understanding of soil hydrological dynamics and land-surface heterogeneity is insufficiently studied. The biggest limitation in the mechanistic understanding of soil hydrological dynamics at the global/continental scale is that of variable land-surface heterogeneity. Land-surface heterogeneity enhances the hysteresis observed in unsaturated soil hydraulic parameters and the effect is compounded by the seasonal differences in land-surface heterogeneity and imposed meteorological conditions. In this study, we quantify seasonal land-surface heterogeneity based on a modified algorithm proposed by Gaur and Mohanty, 2019 and demonstrate its utility in classifying near surface soil hydrology for Contiguous U.S. (CONUS). Near-surface soil hydrology is quantified based on the effective soil water retention parameters (SWRPs) developed using SMAP data (Sehgal et al., 2020) at 36 km resolution. We cluster the SMAP-based SWRPs using a Gaussian Mixture Model based algorithm to identify ‘hydrologically similar’ regions, which are then studied against H-index for identical spatial patterns across CONUS. ANOVA was computed on the H-index values to identify inter-cluster differences in heterogeneity within the ‘hydrologically similar’ regions. Significant differences in the ANOVA analysis imply successful classification of near surface soil moisture dynamics by H-index and demonstrate the potential of replacing representative elementary volume driven soil hydrology classifiers like porosity to describe soil hydrology at the remote sensing scale.