Multi-variable approach to groundwater vulnerability elucidation: A risk-based multi-objective optimization model

Masoumeh Zare; Mohammad Reza Nikoo; Banafsheh Nematollahi; Amir H Gandomi; Raziyeh Farmani

doi:10.1016/j.jenvman.2023.117842

Multi-variable approach to groundwater vulnerability elucidation: A risk-based multi-objective optimization model

J Environ Manage. 2023 Jul 15:338:117842. doi: 10.1016/j.jenvman.2023.117842. Epub 2023 Mar 31.

Authors

Masoumeh Zare¹, Mohammad Reza Nikoo², Banafsheh Nematollahi³, Amir H Gandomi⁴, Raziyeh Farmani⁵

Affiliations

¹ Department of Civil and Environmental Engineering, Shiraz University, Shiraz, Iran. Electronic address: masoumezare86@gmail.com.
² Department of Civil and Architectural Engineering, Sultan Qaboos University, Muscat, Oman. Electronic address: m.reza@squ.edu.om.
³ Department of Environmental Sciences, University of California, Riverside, USA. Electronic address: bnema001@ucr.edu.
⁴ Faculty of Engineering and IT, University of Technology Sydney, NSW, 2007, Australia; University Research and Innovation Center (EKIK), Óbuda University, 1034, Budapest, Hungary. Electronic address: gandomi@uts.edu.au.
⁵ Centre for Water Systems, Department of Engineering, Mathematics and Physical Sciences, University of Exeter, Exeter, United Kingdom. Electronic address: R.Farmani@exeter.ac.uk.

PMID: 37004487
DOI: 10.1016/j.jenvman.2023.117842

Abstract

Groundwater vulnerability mapping is essential in environmental management since there is an increase in contamination caused by excessive population growth. However, to our knowledge, there is rare research dedicated to optimizing the groundwater vulnerability models, considering risk conditions, using a robust multi-objective optimization algorithm coupled with a multi-criteria decision-making model (MCDM). This study filled this knowledge gap by developing an innovative hybrid risk-based multi-objective optimization model using three distinguished models. The first model generated two series of scenarios for rate modifications associated with two common contaminations, Nitrate and Sulfate, based on susceptibility index (SI) and DRASTICA models. The second model was a multi-objective optimization framework using non-dominated sorting genetic algorithms- II and III (NSGA-II and NSGA-III), considering uncertainties in the input rates by the conditional value-at-risk (CVaR) technique. Finally, the third model was a well-known MCDM model, the COmplex PRoportional ASsessment (COPRAS), which identified the best compromise solution among Pareto-optimal solutions for weights of the contaminations. Regarding the Sulfate's results, although the optimized DRASTICA model led to the same correlation as the initial model, 0.7, the optimized SI model increased the correlation to 0.8 compared to the initial model as 0.58. For the Nitrate, both the optimized SI and the optimized DRASTICA models raised the correlation to 0.6 and 0.7 compared to the initial model with a correlation value of 0.36, respectively. Hence, the best and the lowest correlation among the optimized models were between SI and Sulfate concentration and SI and Nitrate concentration, respectively.

Keywords: Complex proportional assessment (COPRAS) technique; Groundwater vulnerability map; Land use-based index model; Multi-criteria decision-making (MCDM) model; Non-dominated sorting genetic algorithm- II and III (NSGA-II and NSGA-III).

MeSH terms

Algorithms
Groundwater*
Nitrates* / analysis
Uncertainty

Substances

Nitrates