Predicting unsaturated soil hydraulic conductivity during drying and wetting processes with a statistical physical model.

We modified the original statistical-physical hydraulic conductivity model proposed by Usowicz, based on averaging water flow through soil pores as a net of capillary tubes. Three division zones of the retention curve were distinguished, and five capillaries with medium radii were assigned: one capillary in both the boundary effect zone and the residual zone, and three capillaries in the transition zone. The sixth capillary covered the entire retention curve with a constant radius up to the residual zone, where the radius decreased to the minimum value with decreasing water content. The equivalent capillary lengths were determined from the measured saturated hydraulic conductivity values under the drying process and the number of parallel capillary connections from those under the wetting. The sixth capillary is responsible for flow processes through vapor and water films. This approach allowed the prediction of unsaturated hydraulic conductivity at a full range of water content. The model was validated using measured data of variously textured soils with different compaction and disturbance. The statistical parameters showed good agreement between the measured and model-predicted data. Further studies are intended to improve the model predictability of the hydraulic conductivity by including an additional domain for preferential water flow through macropores.