ESTIMATING THE UNSATURATED SOIL HYDRAULIC PROPERTIES FROM A REDISTRIBUTION EXPERIMENT: APPLICATION TO SYNTHETIC DATA

Abstract

In this study, we used HYDRUS-1D to simulate redistribution of water flow from a wet soil sample on the top into the dry soil sample at the bottom and to estimate soil hydraulic properties by inverse solution. Real-time surface contact tensiometers were assumed to monitor soil pressure head at two depths (one at a depth of 0 cm, located on the wet sample, the other at a depth of 12 cm, located at the bottom of the dry sample) during a 5-hour-long laboratory water redistribution experiment. Soil water retention curve, soil hydraulic parameters (saturated water content θa, soil bulk density, saturated hydraulic conductivity Ka), and soil organic matter were assumed as known. The soil hydraulic parameters (empirical shape parameters α and n) were estimated from time-varying pressure head measurements and soil water retention data by solving a one-dimensional, inverse water redistribution problem using HYDRUS-1D. The simulated pressure head and soil water retention curve during the redistribution experiment compared favorably with their corresponding observed values. To find answers to the questions of uniqueness, identifiability, and stability of different experimental setups, a numerical example of redistribution was carried out. The uniqueness of the inverse solution was analyzed using numerical experiments to estimate two soil hydraulic parameters (α, n) of the Mualem−van Genuchten model. To study the shape of the objective function near its minimum, response surfaces for the estimated parameters were generated. The response surfaces for both cases (with and without error in observations) indicated that the performance of the redistribution experiment is reasonable for inverse parameter estimation.