Estimating hydraulic conductivity of a fine-textured soil using tension infiltrometry

Tension infiltrometry has become quite popular to determine near saturated hydraulic properties of soils and to examine the effects of macropores on infiltration. The unsaturated hydraulic conductivity has traditionally been determined from steady-state tension infiltrometer data using Wooding's analytical solution. Recently, hydraulic parameters are increasingly being estimated from transient data using a numerical solution of the flow problem and nonlinear parameter optimisation. This inverse approach is convenient but its accuracy will vary with soil type. In this study, tension-disc infiltration data obtained on a tilled clay soil. located in the Basilicata region in Southern Italy, were analysed using both the traditional analytical and the numerical Inversion approach. Three different parametric models were used for the unsaturated hydraulic conductivity for Wooding's solution: the two-line exponential and the bimodal piecewise-continuous models yielded similar results while the Mualem-van Genuchten (MvG) model provided a somewhat poorer description of the cumulative infiltration. The estimates for K were one to two orders larger than those reported for core samples in soil databases. The optimised hydraulic functions revealed the presence of a bimodal pore system with a break-point in pressure head at -30 min. Excellent agreement between measured and fitted infiltration curves was obtained for the numerical inversion procedure. The hydraulic conductivities obtained with Wooding's solution were systematically hi.-her than those obtained from the inverse solution. The difference is presumably clue to a back of steady-state conditions and was greater than found in previous studies on coarser soils: the greatest difference (about 50%) occurred at lower pressure heads. The inverse solution may produce better results if another parametric model is used for the hydraulic properties instead of the Mualem-van Genuchten function. (C) 2004 Elsevier B.V. All rights reserved.