Optimization of a New Inverse Method for Estimation of Individual Soil Hydraulic Parameters under Field Conditions

Abstract

In this research, the EISHP (Estimation of Individual Soil Hydraulic Parameters) model was developed for estimation of the individual soil hydraulic parameters of van Genuchten‘s equation (α, n, Ks, θs, and θr). This model was used to predict soil water content and pressure head at different soil depths by implicitly solving the Richards equation, applying the finite difference method. In this model, the objective function (OF) was computed by summation of squared differences between measured and estimated soil water content and pressure head up to ten days after wetting of soil columns. In the EISHP model, each parameter was optimized separately by achieving the minimum value of OF using the Newton-Raphson method and keeping the remaining parameters constant. The optimized parameter value was then used in optimizing the other parameters. Iteration was continued until the difference between the two computational steps of OF was negligibly small and a unique value had been estimated for each parameter. The main contribution of this research is presentation of a methodology for the optimization of individual parameters, which guarantees the uniqueness of parameter values. In contrast, simultaneous optimization of the entire parameter set would not necessarily lead to unique values.