Abstract
A tortuosity parameter of 0.5 suggested by Mualem (1976) is generally used to predict unsaturated hydraulic conductivity. However, the validity of using a parameter of 0.5 for forest soils needed to be confirmed. We evaluated the tortuosity parameter for forest soils under different site conditions. Undisturbed soil samples were taken from the forest soils which derived from the granite group and Mesozoic-Paleozoic and Quaternary sedimentary rocks. Retention parameters were first fitted, then the tortuosity parameter was optimized using the lognormal distribution conductivity model (the M-LN model) combined with Mualem's (1976) model. Most optimized tortuosity parameter values were less than 0.5; they varied from - 7 to 10 with a high frequency in the range of about - 1 to 0. This means that, for most forest soils, the predicted hydraulic conductivity tends to be underestimated when the most common value (0.5) of the tortuosity parameter is used. Modifying this to -0.72, as presented in this study, corrected the underestimation tendency and reduced prediction errors. When the unsaturated hydraulic conductivity for forest soils is predicted from water retention and the M-LN model, the tortuosity parameter should thus be modified to an appropriate value for more accurate prediction.
