A statistical-chemical characterization of four great soil groups in Southern New South Wales based on orthogonal polynomials

Abstract

Statistical procedures are described for the characterization and comparison of profiles within soil classification units. Profile trends in a measured characteristic are represented by the coefficients of orthogonal polynomials in depth, and these coefficients are used as observations for standard statistical procedures. The mean of the respective coefficients provides estimates of the mean profiles for the classification units, and the variance-covariance matrix of the coefficients can be used to calculate the confidence intervals of these means. Differences between the mean profiles can be tested by analyses of variance. The variability of a range of chemical properties in the profiles of four Great Soil Groups, viz. yellow podzolic, red earth, red-brown earth, and grey clay, is described. The characterization is based on the analysis of profiles sampled from areas of about 0.4 ha chosen to represent cultivated examples of the groups in southern New South Wales. Variability of most soil properties is greater between these areas than within them, notable exceptions being exchangeable calcium and magnesium within the grey clay Group. This high localized variability is attributed to the gilgaied nature of this group in this region. Also variation between the soil groups is generally greater than within the groups. The differences seem, however, of limited importance from the point of view of soil fertility.