Hybrid analytic network process (ANP)-Entropy model, time series analysis for predicting nitrate and fluoride in groundwater and cumulative health risk assessment

Abstract

The study focused on groundwater evaluation of the trans-Himalayan region using a novel concept of Hybrid Analytic Network Process (ANP) - entropy and Holt's exponential smoothing model for short-term prediction of groundwater quality. Hierarchy cluster analysis of groundwater was performed to identify the sources and spatial distribution of water quality. The PCA coupled with multivariate analysis dictated a high correlation of various ions i.e., NO3−, K+, F−, Cl−, SO42−, PO42−, and Na+ with the lithology of the area and helped in identification of various polluting sources i.e., excessive use of fertilizers, wastewater discharges and agrochemicals. The hybrid ANP-entropy technique offered a holistic approach to provide a more accurate and complete understanding of groundwater quality and revealed that groundwater quality in 21.43% (in winter) and 42.86% (in summer) of the locations fall into the poor category. The non-carcinogenic health risk analysis revealed that fluoride poses a major potential health risk followed by nitrate, primarily via the oral pathway while it was negligible through the dermal pathway. The total hazard index was observed highest in children (1.99) followed by infants (1.64). 95th percentile of the probable risk obtained through Monte Carlo simulation also dictated highest risk to infants than child. The predicted values of the Holt's exponential smoothing model for the year 2021 and 2022 was varied within ±5 % of the observed values. The proposed model effectively treats uncertainties in groundwater quality assessment and considers average and incremental increase both hence, can be used for short-term prediction and trend analysis of groundwater quality.