Das Projekt "Bewertung von geostatistischen Methoden zur Beurteilung regionaler Bodenbelastungen mit Schwermetall" wird vom Umweltbundesamt gefördert und von Eidgenössische Technische Hochschule Zürich (ETH), Institut für Terrestrische Ökologie, Arbeitsgruppe Bodenschutz durchgeführt. The objective of this project was to investigate areal patterns of heavy metal distribution in soils, to evaluate the usefulness and applicability of geostatistical techniques for assessing such patterns and in designing appropriate sampling strategies for regional soil contamination surveys. Two test regions of comparable size (20-25 km2) were selected for the sampling of field data, one in the Swiss Jura close to the city of La Chaux-de-Fonds, the other in the flat valley of the Furtbach close to Zurich. Similar sampling strategies were used in both regions, combining a aregular grid with nested schemes at a subset of grid points. In the Furttal area, sampling was performed in two stages. The majority of grid points without nesting were sampled in a second stage and not included in the analysis performed within this project. For external reasons of time constraint, all points were sampled in one campaign in the La Chaux-de-Fonds area. Geostatistical as well as classical statistical analysis showed that the spatial distribution of heavy metals in the top soil to the two test regions were quite different. Bounded and well-structured semivariograms were obtained from the la Chaux-de-Fonds data, while the Furttal data of the first sampling campaign showed to much scatter at larger lag distances to allow a meaningful estimation of semivariograms. Geological substrate, land use and soil type explained some, but for none of the analyzed metals and none of the two regions more than one third of the total variance. Kriging was found to be applicable and useful to map the areal distribution of heavy metals in the topsoil of the La Chaux-de-Fonds area, while additional sampling was found to be required for the Furttal area nine order to adequately cover gaps of more than 250 m between support points.