Developing a fuzzy optimization model for groundwater risk assessment based on improved DRASTIC method

Groundwater pollution is a serious threat to water resources which attracts hydrologists attention for sustainable management. Comprehensive assessment of groundwater vulnerability, however, requires uncertainty analysis in conjunction with a proper method to determine prone areas to contaminants. This paper provides a new fuzzy optimization methodology using improved DRASTIC method to model groundwater vulnerability risk assessment which considers uncertainties embedded in the input parameters, optimizes the weights and modifies the rates of the model simultaneously. To better represent hydrogeological characteristics of an area, the rating scores of original DRASTIC method is modified using Wilcoxon test considering nitrate concentration (as the main interfering pollutant in the study area). Spearman correlation coefficient between vulnerability indices and nitrate concentration is used as a factor to measure how well improved DRASTIC performs for vulnerability assessment as compared to the original method. The results show that the correlation coefficient significantly increased from 0.573 to 0.789. To address uncertainties associated with the input and output of the model, reduced fuzzy transformation method (FTM) is empowered by genetic algorithm (GA) to consider uncertainties associated with input parameters as well as optimizing weights of improved DRASTIC model. The results show how correlation coefficient changes at different uncertainty levels. Considering uncertainties in the inputs, correlation coefficient changes from 0.746 to almost 0.758 at α-cut level equal to 0 in comparison with that of equal to 1. Comparison of the risk maps of improved DRASTIC and fuzzy model at different uncertainty levels reveals that the model performs robustly under uncertainties mainly since the vulnerability trend and more importantly the severity of vulnerability indices have not changed remarkably. Based on these maps, east and southeastern parts of the study area are highly susceptible to contamination where intense industrial and agricultural activities can be seen. This framework provides helpful information for decision-makers to consider risk assessment at different uncertainty levels as it offers a continuous range of vulnerability indices rather than fixed ones. Also, vulnerability risk assessment maps demonstrate vulnerability trend throughout the area for further controlling or remedying actions of groundwater network.