Developing a SINTACS-based method to map groundwater multi-pollutant vulnerability using evolutionary algorithms.

In this study, the modified SINTACS method, a rating-based groundwater vulnerability approach, was applied to data from the Campanian Plain, southern Italy, to identify groundwater vulnerable areas accurately. To mitigate the subjectivity of SINTACS rating and weighting schemes, a modified SINTACS model was formulated by optimizing parameter ratings using the Wilcoxon rank-sum test, and the weight scores using the evolutionary algorithms including artificial bee colony (ABC) and genetic algorithm (GA) methods. The validity of the models was verified by analyzing the correlation coefficient between the vulnerability index and nitrate (NO) and sulfate (SO) concentrations found in the groundwater. The correlation coefficients between the pollutant concentrations and the relevant vulnerability index increased significantly from - 0.35 to 0.43 for NO and from - 0.28 to 0.33 for SO after modifying the ratings and weights of typical SINTACS. Besides, a multi-pollutant vulnerability map considering both NO and SO pollutants was produced by amalgamating the best calibrated vulnerability maps based on the obtained correlation values (i.e., the Wilcoxon-ABC-based SINTACS vulnerability map for NO and the Wilcoxon-GA-based SINTACS vulnerability map for SO). The resultant multi-pollutant vulnerability map coincided significantly with a land use map of the study area, where anthropogenic activities represented the main sources of pollution.