SUITABILITY OF PARAMETRIC MODELS TO DESCRIBE THE HYDRAULIC-PROPERTIESOF AN UNSATURATED COARSE SAND AND GRAVEL

The performance of parametric models used to describe soil water reten tion (SWR) properties and predict unsaturated hydraulic conductivity ( K) as a function of volumetric water content (theta) is examined using SWR and K(theta) data for coarse sand and gravel sediments, Six 70 cm long, 10 cm diameter cores of glacial outwash were instrumented at ei ght depths with porous cup tensiometers and time domain reflectometry probes to measure soil water pressure head (h) and theta, respectively , for seven unsaturated and one saturated steady-state floww condition s, Forty-two theta(h) and K(theta) relationships were measured from th e infiltration tests on the cores. Of the four SWR models compared in the analysis, the van Genuchten (1980) equation with parameters m and n restricted according to the Mualem (m = 1 -1/n) criterion is best su ited to describe the theta(h) relationships. The accuracy of two model s that predict K(theta) using parameter values derived from the SWR mo dels was also evaluated. The model developed by van Genuchten (1980) b ased on the theoretical expression of Mualem (1976) predicted K(theta) more accurately than the van Genuchten (1980) model based on the theo ry of Burdine (1953). A sensitivity analysis shows that more accurate predictions of K(theta) are achieved using SWR model parameters derive d,vith residual mater content (theta(r)) specified according to indepe ndent measurements of theta at values of h where partial derivative th eta/partial derivative h approximate to 0 rather than model-fit theta( r) values. The accuracy of the model K(theta) function improves marked ly when at least one value of unsaturated K is used to scale the K(the ta) function predicted using the saturated K, The results of this inve stigation indicate that the hydraulic properties of coarse-grained sed iments can be accurately described using the parametric models. In add ition, data collection efforts should focus on measuring at least one value of unsaturated hydraulic conductivity and as complete a set of S WR data as possible, particularly in the dry range.