The assessment of soil contamination by heavy metals using geostatistical sequential Gaussian simulation method

Soil contamination by heavy metals is continuously increasing with a great attention in the world. Because, these elements negatively affect living life and the ecosystem. A total of 652 surface soil samples were collected at 100, 200, 400 m regular grid intervals from the study area in Turkey. The observed data does not show normal distribution. Cell declustering was done due to fact that data are not normal. Directional experimental semivariogram of the Cu and Ni showed that both geometric and zonal anisotropy exists in the data. Pb, Zn, Cr and Ba are qualified with omnidirectional experimental semivariogram models. The semivariograms characterized by spherical and Gaussian models of the elements were achieved. Geostatistical sequential Gaussian simulation (SGS) was applied to the study. A hundred simulated realizations depicted the spatial distribution and uncertainty of the elements in the site using a probabilistic approach which produced maps of the heavy metals. These maps showed contaminated and uncontaminated areas in the study site. The results revealed that 23%, 27%, and 24% of the study area at 60% probability were contaminated by the heavy metals including Cu, Cr, and Ni respectively. SGS results have been verified by a number of tests.