Evaluation of Pedo‐Transfer Functions

Abstract Measuring unsaturated hydraulic properties, including water retention vs. tension relationships, is expensive, time consuming, and labor intensive. Several attempts have been made to estimate these relationships from available soil data such as particle‐size distribution, organic C content, and bulk density. An estimation method that describes the soil water retention relationship based on other soil characteristics is referred to as pedo‐transfer function (PTF). The PTF approach will supply data for water and transport models for the unsaturated zone. Pedo‐transfer functions were evaluated by comparing estimated with measured water retention values using test data sets with a broad range of soils. Accuracy of the retention prediction is quantified by the root of the mean squared difference (RMSD) between the measured and predicted retention function. After the application of several PTFs to many soils in a data set, the average 〈RMSD〉 was found in the range of 0.04 to 0.13 m 3 m −3 depending on the PTF. When the RMSD was calculated in the range from 0.1 J kg −1 to wilting point, the lowest 〈RMSD〉 was found to be about 0.05 m 3 m −3 . Using a PTF to supply the water retention for unsaturated soil water flux models (Richards equation), the most practical (and the most accurate) method is to predict retention function parameters (e.g., the parameters of the van Genuchten equation). But, the prediction of the van Genuchten retention function parameters α and θ r (residual water content) using a PTF was found to be inaccurate, even when the PTF is based on nonlinear regression analysis and principle factor analysis.