Groundwater quality assessment using entropy and CRITIC weights on water quality index methods and multiple linear regression in south 24 parganas, India

Abstract

An internationally recognized standard to determine the water quality of any groundwater resource is the water quality index or WQI. There is a decrease in the water levels of several coastal zones' existing groundwater sources. As a result, residents in coastal areas like South 24 Parganas, India, significantly rely on groundwater, which is gradually degrading. Therefore, it is imperative to identify the sources of groundwater contamination and oversee groundwater quality conditions. Consequently, the present research aims to address in detail the causes and influencing factors of groundwater contamination, as well as the status and monitoring of water quality. In order to do so, it makes of incredibly helpful cutting-edge technologies like multiple linear regression (MLR) and groundwater quality index (GWQI), which help to simplify complex tasks in place of more labor-intensive methods. Besides that, it also uses MLR to develop a new WQI equation for the South 24 Parganas district and assess its suitability. To assess whether water in coastal areas suitable for drinking, many weight-based water quality index (WQI) techniques were used. To ascertain WQI related to groundwater in particular coastal regions of eastern India, pre-existing weights such as entropy and criteria-importance through inter-criteria correlation (CRITIC) weights were employed. Though seawater intrusion has increasingly impacted groundwater quality, there have been few investigations on WQI in this case. Therefore, the goal of this study is to determine the efficacy of the WQI weighting strategy and to provide effective WQI approach by comparing the current weights for the assessment of water quality. Here, 100 locations from which the initial WQI results were derived were used to assess seven groundwater quality parameters: pH, TDS (total dissolved solids), Tur (turbidity), Mn (manganese), Fe (iron), TH (total hardness), and Cl (chloride). Based on these findings, MLR was used to create a new WQI equation that only included the following six variables: TH (total hardness), Cl (chloride), Mn (manganese), Fe (iron), and Tur (turbidity). Among the samples, it indicates that 41% and 5% are outstanding to good for drinking. Model 3 (dataset ratio 80:20, R20.994, RMSE-2.8007) has the best accuracy in WQI prediction out of the 12 best fitted models. It was feasible to show using the paired t-test that the new WQI's results did not differ substantially from the original WQI's results. The study found that the significant values greater than 0.05 were obtained for both the absolute WQI values and the quality ranges, indicating the suitability of the equation. One of the important benefits of lowering the number of variables for water quality monitoring is that it lowers the cost of monitoring, even when it comes to minimizing the eclipse effect. In conclusion, it is determined that developing an equation with fewer variables is crucial for keeping an eye on the water quality in the South 24 Parganas district.