A critical assessment of the role of measured hydraulic properties in the simulation of absorption, infiltration and redistribution of soil water

Abstract

Free-water absorption and infiltration were measured in a repacked fine sand. Redistribution following constant-flux infiltration was also measured. These processes were then computer-simulated using independently measured soil water retentivity and conductivity data. Neither free-water absorption, nor ponded infiltration could be accurately similated, as they depended so critically on the soil water retentivity near saturation, where errors of observation prohibit accurate measurement. For simulating redistribution, useful answers were obtained even when hysteresis was ignored. But redistribution was grossly over-predicted when the Campell-Brooks and Corey method was used for inferring the unsaturated hydraulic conductivity from saturated conductivity and retentivity data. Campbell has produced a simple yet elegant, computer simulation package for soil water movement. But to use it successfully required alternative ways of evaluating the hydraulic parameters. Ponded infiltration was successfully simulated only when the sorptivity was used as an independent input. To simulate redistribution required accurate determination of the unsaturated hydraulic conductivity. Software such as Campbell's is now readily available, and likely to come into increasing use. For it to be used effectively more effort and physical insight needs to be directed towards appropriate characterization of the critical hydraulic properties. This physical emphasis will lessen the likelihood of such useful software generating erroneous and misleading results.