ESTIMATING UNSATURATED HYDRAULIC CONDUCTIVITY FROM EASILY MEASURED SOIL PROPERTIES

Abstract

We measured hydraulic conductivity (saturated and unsaturated) on 127 soil cores, which were taken in different horizons of a wide variety of Belgian soil series. The hot air method (Arya et al. 1975) and the crust method (Bouma et al. 1983) were combined to obtain the complete range of hydraulic conductivity from saturation to air-dry. The textural composition in nine fractions, the organic carbon content, and the dry bulk density were determined for each of the sample horizons as well. Four different empirical models were evaluated on their performance in describing the measured hydraulic conductivity curves. The model parameters were estimated by linear and nonlinear regression techniques. It is concluded that the Gardner equation (1958) with three parameters best described the hydraulic conductivity for the given soils. Regression equations for estimating the Gardner parameters were established from simple soil properties, such as soil texture, carbon content, bulk density, and saturated hydraulic conductivity. We found that the three parameters can reasonably well be estimated from the textural composition and the saturated hydraulic conductivity. A one-dimensional sensitivity analysis indicates that the n parameter, representing the slope of the hydraulic conductivity pressure head relation in log-log scale, is most sensitive.