Unsaturated hydraulic properties of four soils of varying particle size distributions were evaluated by determining values of the five parameters in van Genuchten's (1980) hydraulic model. Saturated conductivities (Ks) and saturated water contents (θs) were directly measured and values of residual water content (θr) and the parameters α and n were evaluated by a nonlinear inversion method to minimize various objective functions. Method I uses an objective function involving sums of squared deviations between measured cumulative outflow with time Q(t) for one‐step pressure desorption and numerically simulated outflow from saturation to a final pressure head h = − 10 m. Method II supplements Q(t) data with the measured equilibrium water content θ at h = − 150 m, while Method III employs equilibrium θ(h) data only. Method I yields the most accurate description of Q(t) and the independently determined hydraulic diffusivity D(θ). A fair description of θ(h) is obtained within the range of the one‐step experiment but at lower θ predictions are less reliable especially for finer‐textured soil. Method II extends the range of validity of the predicted θ(h) to lower θ with generally small effects on predicted D(θ) and Q(t). Method III gives the best description of θ(h) but at the expense of accuracy in Q(t) and D(θ). Implications for routine evaluation of soil hydraulic properties are discussed.

No abstract

No abstract

No abstract