Field-saturated hydraulic conductivity estimated using class numbers of soil survey

Abstract

Readily available field data were used to estimate field-saturated hydraulic conductivity (Kfs). Kfs determinaion is laborious, however crucial for various applications, e.g. evaluation of soil functions or groundwater protection. Alluvial soils are close to the groundwater body and thus of special concern with respect to allochthonous contamination. A pedotransfer function (PTF) was developed for 11 alluvial soils. Response variable were 43 Kfs-values determined using a borehole permeameter. Qualitative and semi-quantitative data obtained in the field at 11 near-by soil pits were classified. These classifications comprise texture, humus content, stone content and soil structure as well as hydromorphic indicators such as iron molds, anoxic molds, concretions and soil color. 9 non-multicollinear input variables were selected via factor analysis. In 100 randomized regression realizations the class values were implemented applying 67% of the total data set for training to describe the Kfs-value. While the individual realizations showed considerable deviation to the measured Kfs-values, the averaged linear regression with mean values of intercepts and coefficients provided an acceptable accurracy. The quality criteria of the developed PTF consisted of the root mean square error (RMSE = 0,37), the Nash-Sutcliffe efficiency (NSE = 0,59), and the measure of determination (R2 = 0,59). Despite class thresholds of basic soil properties originally are not linked to soil hydraulic characteristics in the first instance, class values of soil survey can be utilized to estimate the field-saturated hydraulic conductivity in soils with homogeneous horizons.