Investigation on small-scale vertical spatial heterogeneity of soil water retention properties in Chinese Loess Plateau

Quantifying spatial heterogeneity in soil water retention properties (SWRP) is crucial for enhancing the accuracy of hydrogeological simulations. However, studies on the spatial heterogeneity of SWRP in the Chinese Loess Plateau (CLP) remain scarce, especially at the vertical scale. We conducted laboratory tests on undisturbed loess cores collected from boreholes in CLP to analyze soil physical parameters (SPPs) and SWRP. Measured soil water characteristic curves (SWCCs) were fitted to the Brooks-Corey (BC), Fredlund-Xing (FX), and van Genuchten (vG) models. It was revealed that the FX and vG models outperformed the BC model. The geostatistical analysis identified the Gaussian model as optimal for describing the semivariograms of both SPPs and SWCC fitting parameters (FPs). Strikingly, over 90% of these parameters exhibited strong vertical spatial dependence, with an average autocorrelation length of 213.878 cm for SPPs and 320.678 cm for FPs. Moreover, SWRP was found to be significantly influenced by both SPPs and the vertical position relative to the loess ridge slope surface. Parameters near the ridge slope surface showed significantly degraded spatial dependence. These findings provide valuable insights for parameterizing the spatial heterogeneity of soil water retention properties, which are beneficial for hydrogeological modelling in shallow CLP loess strata.