Z-number based assessment of groundwater vulnerability to seawater intrusion

Abstract

Groundwater (GW) quality assessment is an essential issue, especially in arid and semi-arid regions, due to its critical role in obtaining requirements for freshwater. The GALDIT method is widely recognized as a practical approach for assessing GW susceptibility, which considers six key factors, including the occurrence of groundwater (G), hydraulic conductivity of the aquifer (A), elevation of groundwater above sea level (L), distance from the shoreline (D), impact of existing seawater intrusion (I), and thickness of the aquifer (T). The GALDIT method, being an unsupervised model, has a drawback in that it depends on expert opinion for parameter ranking, leading to an increase in uncertainty. To address this uncertainty and evaluate the susceptibility of the aquifers of Urmia lake basin to saltwater intrusion, this study employed the Z-number which is a novel adaptation of Fuzzy Logic (FL). In contrast to classic FL that does not account for data reliability, the Z-number approach considers both constraints and data reliability, making it a more effective method to manage data uncertainty compared to classic fuzzy models. To illustrate this methodology, the GALDIT parameters (inputs) and Electrical Conductivity (EC) values (outputs) were employed to determine the vulnerability of aquifers. Additionally, the GALDIT method was utilized as a benchmark model to evaluate the inherent vulnerability of aquifers, and its outcomes were compared with those of the proposed model. Upon analyzing the results, it was found that the Z-number Based Modeling (ZBM) not only outperformed the GALDIT method but also enhanced the quality of outcomes by 85% and 35% based on Heidke Skill Score (HSS) and Total Accuracy (TA) criteria compared to the classic FL.