Spatial estimation of the thickness of low permeability topsoil materials by using a combined ordinary-indicator kriging approach with multiple thresholds

Accurately estimating a spatial pattern of low permeability topsoil materials, such as clay, silt, and mud, is critical because these materials are a natural barrier to groundwater recharge and flow in subsurface engineering applications. This study determined the spatial distributions and thicknesses of low permeability topsoil materials in the Choushui River alluvial fan, Taiwan by using a combined ordinary-indicator kriging (OIK) approach with multiple thresholds. The thicknesses were first estimated using ordinary kriging (OK). To reduce the overestimation of low values and underestimation of high values, indicator kriging (IK) was then adopted to probabilistically categorize the thickness of the materials. The maximum occurrence probability among the categories was selected for determining the most suitable thickness category of low permeability topsoil materials. Finally, the thicknesses of the materials estimated using OK were amended according to the upper and lower limits of the most suitable category determined using multi-threshold IK. The research results reveal that the combined OIK approach with multiple thresholds can reduce the overestimation of low values and underestimation of high values for various topsoil thicknesses and characterize a reliable hydrogeological pattern in topsoil. The approach facilitates evaluating groundwater recharge and constructing a numerical model of coupled surface water and groundwater flow.