Assessment of terrain factors on the pattern and extent of soil contamination surrounding a chemical industry in Chongqing, Southwest China

- PAHs soil surrounding a chemical plant with irregular terrain was investigated.
- PAHs ranked high concentration in soils in a contaminated site.
- The irregular terrain affected the pattern and contaminant distribution in the soil.
- Three-dimensional kriging considering terrain factors achieved the best accuracy.

In order to determine the pattern and spatial distribution of polycyclic aromatic hydrocarbons (PAHs) in contaminated soil at a site of an organic chemical plant with irregular terrain, we developed a method of soil sampling by subarea that considers topographic features. We divided the site into the west subarea, which had the highest elevation; the southeast subarea (second-highest elevation), and the northeast subarea (lowest elevation). Two- and three-dimensional kriging and principal component analysis (PCA) models were employed to interpret the contamination pattern and spatial distribution of PAH contamination and to calculate the volume of soil contaminated with PAHs. The results show that the total concentrations of 16 priority PAHs in each borehole ranged from 0.7 to 4889.5 mg/kg, with an average of 130 mg/kg. The pollutants mainly accumulated in the first layer of the site (0-1 m). Compared with the pollution levels at other chemical plants reported previously, the soil PAH concentrations measured in this study were higher. PCA results show that the explained variations of the first and second components were 88% and 0.9%, respectively. The samples had good separation characteristics between the different subareas and correlation and pertinence within the same subarea are strong. Volumes of soil contaminated with benzo[a]pyrene calculated using four different interpolation models were 131,019, 135,365, 142,741, 146,783 m3. Three-dimensional kriging that takes into consideration the surface-elevation change showed the lowest root-mean-square error and mean error, as well as higher accuracy compared with the other models. These results are important for defining remediation boundaries and the volume of contaminated soil for sites subjected to risk assessment and remediation.