Evaluation of Pedotransfer Functions for Predicting Saturated Hydraulic Conductivity for U.S. Soils

Abstract

Saturated hydraulic conductivity (Ksat) is a key input to models that simulate groundwater flow at different spatial scales. The direct measurement of Ksat is feasible with field-scale hydrologic models, which require a small number of measurements. For larger scale (watershed or basin) hydrologic models, the required number of Ksat values to capture the spatial variability is too large for the direct measurement to be feasible. Alternatively, Pedotransfer functions (PTFs) have been developed to estimate Ksat in terms of readily available soil properties such as particle size distribution, bulk density, and organic matter content. Over the past two decades, many PTFs have been developed. The goal of this research is to evaluate the accuracy of twenty four PTFs for predicting Ksat. The functions were divided into three groups according to their input requirements: group 1(six functions) requires effective porosity only as an input, group 2(nine functions) requires inputs of particle size distribution, bulk density and total porosity, and group 3(nine functions) requires inputs of particle size distribution, bulk density and organic matter content. The three groups of PTFs were evaluated using three databases of U.S. soils that have 1911, 956, and 678 samples of measured Ksat values. The soil databases of groups 2 and 3 were divided according to the USDA texture classification into four classes and PTFs were evaluated and ranked according to their performance in predicting Ksat for the entire soil’s data base and for each soil textural class. The results showed that the PTFs developed by Suleiman et al. (2001) and Minasny and McBratney (2000) are the best models to estimate Ksat when the available measurement is the effective porosity only. If more measurements like particle size distribution, bulk density and porosity are available, the PTF developed by Cosby et al. (1984) is the best model for predicting Ksat. If more soil properties measurements are available like organic matter in addition to particle size distribution and bulk density, the PTF developed by Nemes et al. (2005) is the best model for predicting Ksat. Finally if all soil properties measurements are available in the soil database, the PTFs developed by Cosby et al. (1984), Nemes et al. (2005), Saxton et al. (1986), Saxton and Rawls (2006), and Julia et al. (2004) are the best models for predicting Ksat respectively. Results of this study can be applied to automate the process of predicting Ksat for large scale hydrologic modeling. The GIS interfaces of hydrologic models could be programmed to estimate Ksat using the most accurate PTF for each soil texture class.