Unsaturated hydraulic conductivity from transient multistep outflow and soil water pressure data

Soil water retention and unsaturated hydraulic conductivity functions [K(6)] estimated by the inverse solution technique through minimization of differences between measured and simulated transient outflow may be nonunique and differ from independently measured soil hydraulic data. Numerical and experimental studies have shown the benefit of using simultaneously measured soil water pressure head in the estimation of the soil water retention curve by the inverse technique. In this experimental study, soil water pressure head and transient cumulative outflow measured simultaneously are used to estimate AT(6). An alternative method for the direct measurement of Jf(8) from transient multistep outflow experiments was adopted. Desorption experiments were carried out for disturbed Yolo silt loam (fine-silty, mixed, nonacid, thermic Typic Xerorthent), Panoche loam (fine-loamy, mixed [calcareous], thermic Typic Torriorthent), Hanford sandy loam (coarse-loamy, mixed, nonacid, thermic Typic Xerorthent), and Oso Flaco fine sand columns. The optimized Af(9) values agreed well with the directly measured data for all soils, except the sand. Additionally, soil hydraulic functions so obtained for the Panoche loam agreed well with those determined using the evaporation method. Measured infiltration in a column of the Panoche loam matched numerical results using optimized parameters as determined from a sorption multistep experiment. The addition of soil water pressure head values in the optimization procedure provides unique parameters for the unsaturated hydraulic conductivity functions under our experimental conditions. N MODELS are extensively used in the modeling of water and solute transport in unsaturated porous media. The application of these models depends on knowledge of the soil hydraulic conductivity, K, as a function of water content, 0, or soil water pressure head, h, and the soil water retention function, 0(/z). With the interest in soil spatial variability, there is an increasing need for techniques to determine these functions fast and accurately. Over the years, numerous in situ and laboratory methods have been developed to determine these soil hydraulic properties. Although in situ methods generate results that are more representative of field conditions, laboratory experiments offer more flexibility in initial and boundary conditions. Soil hydraulic functions are also more accurately and more conveniently measured in the laboratory, and measurements are made faster across a wider range of soil water content. Consequently, research has been directed toward developing direct and indirect laboratory methods. A variety of direct methods have been developed to determine unsaturated hydraulic conductivity data under S.O. Eching and J.W. Hopmans, Hydrologic Science, Dep. of Land, Air, and Water Resources, Veihmeyer Hall, Univ. of California, Davis, CA 95616; and O. Wendroth, Institut fur Bodenforschung Zentrum fur Agrarlandschafts-und Landnutzungsforshung, Wilhelm-Pieck-straBe 72, 0-1278 Muncheberg, Germany. Received 19 Mar. 1993. *Corresponding author (jwhopmans@ucdavis.edu). Published in Soil Sci. Soc. Am. J. 58:687-695 (1994). steady-state conditions (Nielsen et al., 1960; Watson, 1967; Klute and Dirksen, 1986). These methods are time consuming since experiments have to reach several steady-state conditions. Faster non-steadyand quasisteady-state methods include the hot-air method (Arya et al., 1975), sorptivity method (Dirksen, 1979), the method of Ahuja and El-Swaify (1976), and several evaporation methods (Wind, 1968; Flocker et al., 1968; Plagge et al., 1990). Recently, Wendroth et al. (1993), reevaluated the evaporation method. In the past, indirect methods have relied on statistical pore-size distribution models to predict the hydraulic conductivity from soil water retention data (Childs and Collis-George, 1950; Burdine, 1953; Mualem, 1976). In recent years, however, the inverse solution technique applied to laboratory outflow data has become an attractive alternative for the indirect estimation of unsaturated hydraulic conductivity data. Outflow experiments are flexible in initial and boundary conditions, yield fast results across a wide range of water content, and are relatively cheap. The technique involves analytical (Gardner, 1956) or numerical solution of the Richards equation subject to imposed initial and boundary conditions. A transient-flow experiment is carried out on a saturated or near-saturated soil core with known initial and boundary conditions. Drainage is induced in one step (Kool et al., 1985) or by multistep (van Dam et al., 1990; Eching and Hopmans, 1993) pressure increments. These transient experiments are faster than steady-state experiments. For the numerical solution, the flow process is subsequently simulated using parameterized hydraulic functions. The unknown parameters are estimated by minimizing deviations between observed and predicted flow variables in the objective function (Dane and Hruska, 1983; Kool et al., 1985; Yeh, 1986; Kool et al., 1987; Kool and Parker, 1988; Eching and Hopmans, 1993). Laboratory outflow experiments used thus far have involved use of transient cumulative outflow supplemented with a limited number of soil water pressure head or soil water content data. However, problems have been reported on the nonuniqueness of the solution (Russo et al., 1991; Toorman et al., 1992; van Dam et al., 1992). Although the benefit of using soil water pressure data in combination with water content data derived from cumulative outflow is clear from theory (Kool and Parker, 1988; Toorman et al., 1992), few laboratory studies have been reported hi the literature. Eching and Hopmans (1993) confirmed the theoretical analysis of Toorman et al. (1992) and showed the improvement hi the estimation of 9(/i) from various one-step and multistep outflow experiments if cumulative outflow is combined with simultaneously measured soil water pressure head data. This study determined unsaturated hydraulic conductivity functions estimated by the inverse solution technique with cumulative outflow and soil water pressure head