Regionally representative selection of soil samples for an environmental specimen bank with particular reference to loess soils

Digital evalutation of a series of correlative geo-specific maps in different scales and subsequent multivariate data processing in the form of entropy and neighbourhood analyses are the basis of a comparative selection of soil samples for the environmental specimen bank of the Federal Republic of Germany. By means of regionalisation algorithms, partly developed for, or specifically adapted to the present purpose, the optimum location of sampling points was determined and the result corroborated on large-scale maps or by visual inspection, respectively. In addition an index of representativity was defined, grouping soil taxa in terms of acreage and spatial autocorrelation.Once the spatial structure of the soil associations in an area to be sampled is thus determined variogram analysis, in the second step, contributes to definitely select the most appropriate specimens on the basis of a representative sampling grid and with respect to relevant properties, for instance cation or anion exchange capacities and biodegradability potential.