Evaluation of Hydrus program in predicting soil water retention curve

Abstract

The purpose of this study was to evaluate the Hydrus program in predicting soil water retention curve by comparing the predicted curves with those obtained empirically. A set of 8 disturbed soil samples differ in texture and bulk density (ρ_b) was used in this study. The results showed that a big difference between the predicted and empirical curves was found when only the soil texture or sand, silt and clay content (SSC) were used as inputs. The best predicted curves were observed when taking both the SSC and bulk density as inputs. This last simulation produced a Root Mean Square Error (RMSE) of 0.015 cm³/cm³. In addition, an underestimation of 12% at saturation (θ_s) in the predicted curves was also observed. More generally, the results pointed out that Hydrus could be a useful tool for predicting and estimating soil water retention curve if the RMSE value and deviation at saturation were taken into account.