USING NEURAL NETWORKS TO PREDICT SOIL-WATER RETENTION AND SOIL HYDRAULIC CONDUCTIVITY

Direct measurement of hydraulic properties is time consuming, costly, and sometimes unreliable because of soil heterogeneity and experimenta l errors. Instead, hydraulic properties can be estimated from surrogat e data such as soil texture and bulk density with pedotransfer functio ns (PTFs). This paper describes neural network PTFs to predict soil wa ter retention, saturated and unsaturated hydraulic properties from lim ited or more extended sets of soil properties. Accuracy of prediction generally increased if more input data are used but there was always a considerable difference between predictions and measurements. The neu ral networks were combined with the bootstrap method to generate uncer tainty estimates of the predicted hydraulic properties. (C) 1998 Elsev ier Science B.V. All rights reserved.