The sustainability of groundwater resources is being threatened by overexploitation, anthropogenic influences and mismanagement. Considerable intensification of groundwater demand from negligible recharged aquifers in arid and semiarid regions, has underlined deterioration of groundwater level, thereby making aquifer systems more vulnerable. Groundwater vulnerability assessment, as a worldwide tool, deploys hydrogeologic conditions to predict vulnerable areas in order to support groundwater protection. This study aims to represent a comparative account of optimized DRASTIC-based groundwater vulnerability in Qazvin plain, Iran. The inherent drawback of DRASTIC framework approach is the subjectivity in assigning weights/ratings for each hydrogeological factors. Particle swarm optimization as one of the most popular nature-inspired metaheuristic optimization algorithms, is carried out for assigning relative weights to each DRASTIC parameter in order to improve objectivity and robustness of vulnerability index. The performance of generic and optimized DRASTIC indices is validated based on nitrate concentration in monitoring wells as accepted criteria in agricultural regions, using the Area under the Receiver Operating Characteristic (ROC) Curve (AUC). Results indicated that optimization of weights considerably improved the accuracy of vulnerability index from 0.58 for generic DRASTIC index to 0.75 for DRASTIC-PSO index. The proposed optimization process provides a feasible framework from integrating hydrogeological characteristics of the aquifer and major contamination load in the study area, thereby it can be versatile to be applied in different case studies.
Keywords: Groundwater, Vulnerability mapping, Optimized DRASTIC, Particle swarm optimization (PSO), GIS