Abstract: A petroleum-contaminated site was selected to determine the contamination boundary and volume of soil contaminated with dimethylbenzene. We compared the efficiency and precision of three 3D interpolation models (Kriging, IDW, and Nearest Neighbor) in determining the site contamination boundary. Employing a cross validation method, we found that the results of the IDW model were more accurate than the other two models in predicting the actual pollution conditions of the contaminated site. The mean error (ME) and root mean square error (RMSE) of IDW were 4.94, and 14.11, respectively. With a view towards remediation, the contaminated soil volume calculated by the IDW interpolation model was 666.2 m³. Different prediction models had uncertain influences on determining the contamination boundary. This study provides a contribution towards determining pollutant distribution and making decision for remediation of contaminated sites.

Key words: endophytic bacteria, salt stress, peroxidase, catalase., wheat seedling, 16S rDNA sequences