|Pachepsky, Yakov - DUKE UNIVERSITY|
Submitted to: Soil Science Society of America Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: June 17, 1999
Publication Date: November 1, 1999
Interpretive Summary: The soil controls the downward movement and storage of water which is critical to crops, flood prediction, and engineering design of conservation practices. Laboratory and field methods for determining soil hydraulic properties are time consuming and expensive especially on a regional scale; thus, the value of standard soil survey qualitative descriptions need to be evaluated to describe soil effects on the water storage and movements in soils. The usefulness of using soil taxonomy, soil water regime, soil temperature regime, and soil texture classes in describing soil effects on the water storage and movement were found to be useful for soils in Oklahoma. The customers of such research are federal agencies such as EPA, NRC, DOD, DOE, USGS, NASA, NRCS, NOAA; state agencies; and consulting and private firms doing environmental work.
Technical Abstract: Pedotransfer functions, dependencies of soil water retention and soil hydraulic conductivity on basic parameters available from soil survey, are widely used to predict soil functioning in agricultural and environmental systems. Our objective was to compare accuracy and reliability of pedotransfer functions developed for data groups selected according to soil taxonomy, soil water regime, soil temperature regime, and textural classes. We developed pedotransfer functions to estimate water contents at -33 cm and -1500 peons. Clay, sand, coarse fragments, and organic matter contents; the bulk density; and the ratio of caton exchange capacity to clay content were used as predictors. To assess accuracy and reliability of pedotransfer functions, were randomly split each data set into development and testing subsets. The group method of data handling prosed pedotransfer functions that were more accurate than multiple linear regressions, selected the most significant predictors and produced explicit equations so that the relative importance of the predictors could be assessed. Grouping improved the accuracy of pedotransfer functions in most cases, most probably because of the similarity in pedotransfer functions relations within groups. None of the grouping criteria could be the best one.