Parameter estimation of soil hydraulic characteristics by inverse modeling of the analytical equation for unsaturated subsurface water flow

Abstract

<jats:title>Abstract</jats:title> <jats:p>In drip-irrigated systems, the understanding of the soil wetting pattern is essential in defining the area effectively irrigated, the spacing between the emitters and their installation depth, and the irrigation rate. Thus, this study aims to estimate soil hydraulic characteristics through inverse modeling of an analytical equation used in wetting bulb simulation based on soil moisture measurements obtained in the field. The parameters of the Gardner model, which describes the unsaturated soil hydraulic conductivity, and the van Genuchten model that describes the soil moisture retention curve, were estimated by inverse modeling techniques. The following options were considered: (A) estimating parameter β while considering the other parameters, Ko, θr, θs, α, n, and m, as known and obtained experimentally; (B) estimating parameters Ko and β while considering the experimental retention curve as known; (C) estimating parameters Ko, β, α, n, and m while considering the values of the θr and θs volumetric moisture as known; (D) estimating all the parameters of Gardner and van Genuchten models (Ko, θr, θs, α, n, and m). The results indicate that option D showed better concordance between the estimated and observed moisture values. Thus, the inverse modeling of the analytical equation is an important tool for irrigation design and management.</jats:p>

Can Deep Learning Extract Useful Information about Energy Dissipation and Effective Hydraulic Conductivity from Gridded Conductivity Fields?

“…Non-local methods can be used to infer effective values for system parameters via inverse modeling, wherein the parameter field is constrained to be homogenous, and the corresponding best-fit equivalent upscaled parameter value is determined. Several recent studies [30][31][32][33] have used this technique for vadose zone parameter estimation. However, this approach requires solving the flow problem for specified boundary conditions.…”