Abstract
This study compares the capabilities of soil water characteristic curve (SWCC) models by Brooks-Corey (BC) and van Genuchten (vG) in estimating the curve fitting parameters for lateritic soil – bentonite mixtures. The SWCCs of soil treated with 0–10% bentonite and compacted with British standard light (BSL) energy at compaction states representing dry of optimum, optimum, and wet of optimum conditions were measured by sequential desorption using pressure plate extractor. The fitting parameters of the two equations were determined by a non-linear fitting program. The fitting capabilities of the models on the measured data were compared by statistical indices namely the root mean square error (RMSE), linearity (R2) and index of agreement (d). It was observed that the RMSE values ranged from 0.0035 to 0.0150 for vG model which are somewhat lower than the values for BC equation. Similarly, R2 for the vG equation (≥ 0.99) are, on average, slightly higher than those of the BC fits. However, the d values associated with the BC model which varied between 0.788 and 0.971 are higher than those of the vG (0.784–0.968). Overall, the study established that the vG model proved marginally superior in respect of goodness of fit.