Groundwater quality index development using the ANN model of Delhi Metropolitan City, India

Abdul Gani; Mohit Singh; Shray Pathak; Athar Hussain

doi:10.1007/s11356-023-31584-4

Groundwater quality index development using the ANN model of Delhi Metropolitan City, India

Environ Sci Pollut Res Int. 2023 Dec 22. doi: 10.1007/s11356-023-31584-4. Online ahead of print.

Authors

Abdul Gani¹, Mohit Singh¹, Shray Pathak², Athar Hussain¹

Affiliations

¹ Department of Civil Engineering, Netaji Subhas University of Technology, New Delhi, 110073, India.
² Department of Civil Engineering, Indian Institute of Technology Ropar, Rupnagar, Punjab, 140001, India. shraypathak@gmail.com.

PMID: 38133760
DOI: 10.1007/s11356-023-31584-4

Abstract

Groundwater is widely recognized as a vital source of fresh drinking water worldwide. However, the rapid, unregulated population growth and increased industrialization, coupled with a rise in human activities, have significantly harmed the quality of groundwater. Changes in the local topography and drainage systems in an area have negative impacts on both the quality and quantity of groundwater. This underscores the critical need to assess the susceptibility of groundwater to pollution and implement measures to mitigate these risks. The water quality index (WQI) is an approach that simulates the water quality at peculiar locations for a particular period of time. The artificial neural network (ANN) model approach is such an idealistic methodology that can be utilized for WQI development and provides better results for specific locations in optimum time. Therefore, the goal of the current study is to provide a unique way for using artificial neural networks (ANN) to characterize the groundwater quality of Delhi Metropolitan City, India. In order to make the water fit for residential and drinking use, the research also pinpoints the geographical variability and spots where the contaminated region has to be sufficiently cleaned. A minimum WQI of 41.51 was obtained at the Jagatpur location while a maximum value of 779.01 was at the Peeragarhi location. During the training phase, the results obtained using the ANN model were highly favorable, demonstrating a strong association with an R-value of 98.10%, thus highlighting the program's exceptional efficiency. However, in accordance with the correlation regression findings, the prediction outcomes of the ANN model in testing are observed to be an R-value of 99.99-100%. This study confirms the promise and advantages of employing advanced artificial intelligence in managing groundwater quality in the studied area.

Keywords: ANN; Groundwater; Management; Spatial variability; Water quality index.