Pedotransfer function application for estimation of soil hydrophysical properties using parametric methods

Abstract

Soil hydraulic properties are needed as input data to describe and simulate the transport of water and solutes in the soil profile. The most important characteristics are the soil moisture retention curve (SMRC) θ(<i>h</i>) and the hydraulic conductivity function <i>k</i>(θ) or <i>k</i>(<i>h</i>), where θ is the soil moisture content, <i>h</i> is the pressure head and <i>k</i> is the hydraulic conductivity. SMRC represents the amount of water remaining in the soil under equilibrium conditions and is unique for each soil. The measurement of SMRC is laborious and time-consuming and so there are not enough data available sometimes. Various SMRC estimation models have been proposed and used extensively to overcome this problem. Other more easily available soil properties, such as particle size distribution, organic matter content, soil structure and bulk density, were used for the estimation of SMRC. Bouma and van Lanen (1987) called these models “transfer functions”, and later on they were called “pedotransfer functions”. This study is based on European works by Wösten et al. (1998, 1999), and others. The pedotransfer functions derived by Wösten et al. (1998) were used in the first part of the study. In the second part, the authors derived their own pedotransfer functions for the sites where all necessary data were available. The methodology of data processing was similar to that used by Wösten et al. (1998) for continuous pedotransfer functions. The use of continuous pedotransfer functions was tested on data sets from several sites in the Czech Republic (Cerhovice, Černičí, Brozany, Ovesná Lhota, Tupadly, Džbánov, Podlesí and Žichlínek). Unfortunately, the available Czech data sets are not as large as the data sets used in Wösten’s work. Quite good new estimates of SMRC (expressed as pF curves) were found e.g. for the Cerhovice and Černičí sites; the estimates for a man-made soil profile in Brozany and for natural soils in Ovesná Lhota, Tupadly, Džbánov, Podlesí and Žichlínek were less successful, partly because of insufficient input data. The applications of continuous pedotransfer functions derived by Wösten et al. (1998) for the Czech data sets were not very successful, either. The quality and size of the input data sets are critical factors for a successful use of pedotransfer functions.