Multivariate geostatistical analysis of soil contaminations

Soil is one of the most endangered compartments of our environment. The input of pollutants into the soil caused by industrial production, agriculture, and other human activities is a problem of high relevance. A contour analysis of soil contamination is the first step to characterize the size and magnitude of pollution and to detect emission sources of heavy metals. The evaluation and assessment of the large number of measured samples and pollutants require the use of efficient statistical methods which are able to discover both spatial and multivariate relationships. The evaluation of the presented case study - soil contamination by heavy metals - is carried out by means of multivariate geostatistical methods, also described as theory of Linear coregionalization. Multivariate geostatistics connects the advantages of common geostatistical methods (spatial correlation) and multivariate data analysis (multivariate relationships). In the given case study of soil pollution by heavy metal emissions it is excellently possible to detect multivariate spatial correlations between different heavy metals in the soil and to determine their common emission sources. These emission sources are located in the area under investigation.