|Vereecken, H -|
|Weynants, M -|
|Javaux, M -|
|Schaap, M -|
|Van Genuchten, M -|
Submitted to: Vadose Zone Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: October 14, 2010
Publication Date: December 1, 2010
Citation: Vereecken, H., Weynants, M., Javaux, M., Pachepsky, Y.A., Schaap, M., Van Genuchten, M. 2010. Using pedotransfer functions to estimate the van Genuchten-Mualem soil hydraulic properties: A review. Vadose Zone Journal. 9:795-820. Interpretive Summary: Models of water flow in variably saturated soils are commonly used in many engineering applications. The ability of soils to retain and transmit water is characterized by parameters incorporated into these models. Measurement of soil water parameters is extremely laborious and is impractical for many applications. Therefore, model parameters are routinely estimated from readily available soil properties. Formulas for estimating these parameters have been proposed by van Genuchten and Mualem, which are frequently used. This paper is a comprehensive review of the accuracy, reliability, and utility of such estimations. The role of different soil properties as predictors is discussed and illustrated. The need for development of a new generation of predictive methods is emphasized where soil structure, i.e. arrangement of soil particles in space and in time, is included in such predictions.
Technical Abstract: In this paper, we review the use of the van Genuchten Mualem (VGM) model to parameterize the soil moisture retention characteristic (MRC) and the nsaturated hydraulic conductivity curve (HCC), as well as its use in developing pedotransfer functions (PFTs). Analysis of literature data showed that MRC parameterization using the classical van Genuchten model with m=1-1/n produced the largest deviations, expressed in terms of the corrected root mean squared residual (RMSRA), for pressure head values between 330 (pF 2.5) and 2500 cm (pF 3.4). While MSRA values in the wet and dry ranges were much smaller, the deviations here are still important in terms of predicting the HCC because of strong nonlinearities in the hydraulic conductivity function. The model of Schaap and van Genuchten (2006) performed best in describing the unsaturated hydraulic conductivity, K. The VGM model with the saturated conductivity Ks fixed at its measured value and the pore connectivity parameter l at 0.5 produced increasingly higher RMSR values when the soil became drier, with RMSR values in log10(K) being up to 3. The smallest deviations between measured and predicted unsaturated hydraulic conductivity values were found for pressure head values between 30 (pF 1.5) and 2500 cm (pF 3.4). The most accurate PTFs for estimating the MRC were obtained when using soil organic matter and soil moisture values at specific pressure heads, in addition to textural properties and bulk density. The root mean squared residual (RMSR) for the PTFs approached in this case the RMSR values of the direct fit, thus suggesting a need to improve both PTF development as well as the MRC parameterization. Inclusion of the soil water content at different pressure heads in the PTFs for the unsaturated hydraulic conductivity only marginally improved their prediction compared to the PTF using textural properties and bulk density. Including soil organic matter as a predictor was to have a larger effect on the prediction than information on soil moisture. In further advancing the development of PTFs, we advocate the establishment of databases of soil hydraulic properties that 1) are derived from standardized and harmonized measurement procedures that also consider the support scales, 2) contain new predictors such as soil structural properties, in addition to soil information typically available, and 3) allow for the development of time-dependent PTFs. Successful use of structural properties in PTFs will require a parameterization that accounts for the effect of structural properties on the soil hydraulic functions.