Pore-Based Modeling of Hydraulic Conductivity Function of Unsaturated Rooted Soils

Abstract

Mualem's approach has been widely used to predict hydraulic conductivity functions (HCFs) of bare soils if a soil water retention curve (SWRC) model is available. The assumption that Mualem's approach holds is that the distribution of soil pores is spatially completely random. Under this assumption, relative hydraulic conductivity (K_r) is determined by the continuance probability of water-filled pores. However, this assumption is not valid for rooted soils, as root growth causes soil particle rearrangement, and thus soil pore rearrangement, altering the probability of pore connectivity. After reconsidering Mualem's assumption, this study attempts to develop a new approach for predicting HCF of rooted soils by modeling the root-induced pore rearrangement and the resultant change in the continuance probability of water-filled pores. Two approaches mentioned were incorporated with a root-dependent SWRC model to express HCF as a function of matric suction. The proposed model was validated against nine sets of measured HCFs from published studies. It was found that the proposed model reduced the root mean square error (RMSE) of K_r and lg K_r by 33% and 53%, respectively, as compared to traditional Mualem's model. Physically, the model's effectiveness depended on soil texture and root type. In fine-textured soils, roots were capable of displacing soil particles, thereby causing soil pore rearrangement. Also, coarse roots with high strength tend to alter pore distribution. After considering the effects of pore-level root-soil interaction on pore rearrangement, the proposed model provided a significant improvement in the prediction of HCF of unsaturated rooted soils.