Integrated Fuzzy Logic and Data Mining Approach for Assessing Groundwater Vulnerability to Sea-Water Intrusion

The GALDIT model is used for the assessment of the aquifer vulnerability of Groundwater (GW), but it relies on expert judgment that contains uncertainty and is one of its weaknesses. To tackle the challenge of managing uncertainty and identifying specific vulnerabilities, this study employed a combination of Mamdani Fuzzy Logic (MFL) and data mining techniques. Then the obtained results were compared with the traditional GALDIT method. According to the Electrical Conductivity (EC) values, the Heidke Skill Score (HSS) and Total Accuracy (TA) criteria were applied to assess the outcomes of the applied methods. The surrounding aquifers of Urmia Lake was case study in this paper. The findings indicated that the GALDIT index varied from 2 to 8. The TA and HSS values in the MFL modeling were determined to be 0.59 and 0.48, respectively, therefore MFL led to more accurate and reliable results for evaluating GW vulnerability compared to the traditional GALDIT method.