Improved soil hydraulic conductivity function based on specific liquid–vapour interfacial area around the soil particles

Abstract

Improvement in the prediction of soil hydraulic conductivity function is a major concern in soil physics. Zand-Parsa and Sepaskhah (2004) [Zand-Parsa, Sh., Sepaskhah, A.R., 2004. Soil hydraulic conductivity function based on specific liquid–vapour interfacial area around the soil particles. Geoderma 119, 143–157] proposed a new method for the prediction of soil hydraulic conductivity function [ K( h) or K( θ) (where h and θ are soil water pressure head and water content, respectively] based on specific liquid–vapour interfacial area around the soil particles (SLVIA). In the present paper, the SLVIA method was improved as follows: (i) a more straightforward and efficient numerical technique for the prediction of K( θ) function was proposed instead of the analytical method in Zand-Parsa and Sepaskhah; (ii) saturated soil hydraulic conductivity ( K s) was predicted by assumption of a linear relationship between K( θ) and θ for very small θ ranges near saturated soil water content; (iii) the value of soil tortuosity (ratio of actual path length to the straight path length of flow) factor ( Tr) was estimated by using the Newton–Raphson method, by minimizing the difference between predicted and measured saturated soil hydraulic conductivity. The K( θ) curve was predicted by soil water characteristic curve and measured saturated soil hydraulic conductivity. Soil water characteristic curve was predicted by the van Genuchten method with four required soil parameters [ α, n, residual soil water content ( θ r), and saturated soil water content ( θ sat) [van Genuchten, M.Th., 1980. A closed-form equation for predicting the hydraulic conductivity of unsaturated soil. Soil Sci. Soc. Am. J. 44, 892–898]. A computer program (UNSATK model) written in VISUAL BASIC was used to predict the K( h) or K( θ) curve. The utility of the UNSATK model was tested for three soil types (sandy, loamy, and clayey textures) selected from the literature. The uniqueness of model outputs for each selected soil was checked by different input Tr values. Model predictions yielded reasonable agreement with measured soil hydraulic conductivity data and compared favorably with the widely applied Mualem–van Genuchen method.