Estimating the wetting branch of the soil water retention curve from grain-size fractions

The development of a theoretical method for estimating the wetting branch of the soil water retention curve (SWRC) is required for knowledge of the soil infiltration process. The aim of this study was to derive the theoretical functions to represent the wetting branch of the SWRC based on the Jensen method, and then compare the Jensen method and Kool & Parker (KP87) model for estimating the wetting branch of the SWRC. Fifteen soil samples with varying basic properties (e.g., grain-size distribution and bulk density (BD)) were selected from the Unsaturated Soil Hydraulic Database (UNSODA) to test these two methods. Results showed that the Jensen method (root mean squared error (RMSE) = 0.057 cm(3) cm(-3)) produced a substantially better performance in predicting the wetting branch of the SWRC than the KP87 (RMSE = 0.089 cm(3) cm(-3)) for the 15 samples. The range of the scaled mean bias error (SMBE) between the Jensen method-predicted and measured soil water contents at all pressure heads was -0.529 to 0.402. A positive linear relationship between the SMBE and silt content was observed for the Jensen method. The findings of this study have practical significance for simulating the soil infiltration in the unsaturated zone.