Using relief parameters in a discriminant analysis to stratify geological areas with different spatial variability of soil properties

The spatial variation of soil properties within landscapes is controlled by the soil forming factors relief, parent material, climate, organisms, and time. Although this relation is a paradigm in soil survey, it is rarely con sidered in the analysis of spatial variability of soil properties. Homogene ous soil units are mostly mapped according to the soil properties found clo se to the soil surface. Nevertheless, a large heterogeneity may occur at gr eater depths for soils developed in two different geological strata. This i s often the case in former periglacial areas where pleistocene sediments co ver older strata. The aim of this paper is to show how discriminant analysi s can be used to objectively determine the soil depth at which geology chan ges. This information may then help to develop better soil surveys and impr ove the geostatistical regionalization of soil properties. The study area i s a 1.5 km(2) soilscape 50 km north of Munich with large variability in rel ief and parent material. At 450 nodes of a rectangular 50 x 50 m grid; fund amental soil properties were measured for each soil horizon. Relief paramet ers were calculated using a Digital Elevation Model (DEM) derived from more than 4000 elevation measurements. Parent material of the soils may be sedi ments of the tertiary period (TS) or quaternary sediments (QS). Altogether, 86 soil horizons were classified with confidence to TS and 496 to QS. They were the training data set for a discriminant analysis to distinguish hori zons of TS from horizons of QS. In addition to several relief parameters, t he horizons' depth position within the soils was used as a discriminant var iable. The discriminant functions classified 87% of TS and 85% of QS traini ng data set horizons correctly by using elevation above sea level, depth, s lope and upslope watershed area as independent variables. Solving the discr iminant functions with respect to the boundary depth between QS and TS and applying the result to the DEM yielded a map of boundary depth for each loc ation of the study area. At 19 locations within the study area, the predict ions were validated with an independent data set. The root of the mean squa red differences between the measurement and the prediction was 7.5 cm for t his second data set. This is within the uncertainty of measurement. Boundar y depth was finally used to divide the study area into separate areas of th e two strata dependent on the depth of interest within the soil. This allow ed separate variogram calculations for each stratum and depth. The resultin g variograms for soil texture showed a larger spatial variability for strat um TS than for QS. Consequently, four times as many locations must be measu red for stratum TS than for QS to obtain the same precision of spatial inte rpolation. (C) 1999 Elsevier Science B.V. All rights reserved.