AN ANALYSIS OF SATURATED HYDRAULIC CONDUCTIVITY IN A FORESTED GLACIAL TILL SLOPE

We made a statistical analysis of 100 soil core samples taken from six soil pits in a forested glacial till slope in south central Sweden to determine the spatial variabilty of the saturated hydraulic conductivity, Ks. We also analyzed correlations to other soil properties. Analysis of variance (AN-OVA) determined whether the Ks values were significantly dependent upon the position in the slope and the sampling depth below soil surface. A multifactor analysis of variance (MANOVA) tested the interaction effect between position in the slope (elevation level) and soil depth (horizon level) on Ks. The highest correlation of Ks, 0.77, was established with drainable porosity (the difference between the total porosity and the field capacity value at a water tension of −10 kPa). The correlations of Ks with total porosity and bulk density were lower, 0.64 and −0.62, respectively. However, correlations of Ks with all of these properties increased with all of these properties increased with depth. Even though the correlation coefficients are moderate, the relationship between Ks and drainable porosity provides a quick method of estimating the relative permeabilities when only the drainable porosities are known. The correlation between Ks and depth in centimenters below ground surface was very low, 0.59. Sample groups of different core lengths (50 and 100 mm long) could not be distingusihed statistically, and therefore the main analysis included all Ks values irrespective of core length. As expected, a first ANOVA indicated that the Ks values differed significantly with soil depth, with a strong decreasing trend with increasing depth. The average Ks value of the undisturbed basal till was 2 orders of magnitude lower than the two surface horizons. A second ANOVA of the Ks values from the different pits, treated as individual sample groups, gave also statistically isolated sample groups. By excluding the samples from one sample group (i.e., the Ks values from one pit) from the ANOVA analysis, however, it was possible to obtain a statistical significance for five pits, i.e., homogeneous conditions concerning the saturated hydraulic conductivity in the slope. For the interaction effect between pit location and horizon level on the Ks valuese, the MANOVA showed that there was no singnificant difference at the selected significance level of 5%. A multiple range test using a 95% confidence limit, however, separated the differenet pits into four different categories based on the average Ks values. These categories also followed the natural siting of the pit locations, i.e., the highest Ks values were found in the middle region of the glacial till slope, followed by the pits from the uppermost soil-covered area in the slope. The multipal range analysis also separated the different horozons into three possible categories, with the highest conductivities in minearl horiozn no. 2, a coarse-textured wave-washed horizon. © 1990 Williams and Wilkins.