Analysing spatial distribution and variability of vadose zone moisture in different subsidence areas using 3D empirical Bayesian kriging and pixel-by-pixel methods

Coal mining induces soil deformation that changes the dynamics of vadose zone moisture, affecting its movement and distribution. The spatial distribution and variability of vadose zone moisture in various subsidence areas are essential for efficiently managing and utilising water resources in coal mining regions. However, the current studies on the effects of coal mining subsidence on the movement and distribution of vadose zone moisture often neglect subsidence characteristics; consequently, the effect of these characteristics remains unclear. Therefore, this study aims to examine the vadose zone moisture in areas without disturbance, areas with uniform subsidence, and areas with non-uniform subsidence. Using 3D empirical Bayesian kriging in conjunction with pixel-by-pixel analysis of raster images, the spatial distribution and variability of vadose zone moisture were examined. The results show a significant reduction in the vadose zone moisture level due to subsidence, accompanied by increased spatial variability. This effect is particularly prominent in non-uniform subsidence areas, where preferential flow plays a crucial role. Utilizing the 3D empirical Bayesian kriging interpolation method, the three-dimensional spatial distribution of vadose zone moisture was effectively visualized. Coupled with pixel-by-pixel analysis, the continuous spatial variability in the vertical direction was successfully evaluated. Cluster analysis revealed distinct zones of low and high variability in vadose zone moisture across different areas, with non-uniform subsidence areas exhibiting a higher variability than that of undisturbed and uniform subsidence areas. These findings offer a theoretical foundation for understanding vadose zone moisture dynamics and water resource management in coal mine subsidence areas.