Probabilistic Assessment of Groundwater Quality and Fluoride Exposure Risk for North-Eastern Rajasthan, India

Abstract

Fluoride risk has been highlighted for many regions in India, particularly the north-western parts that include Rajasthan amongst the most severely affected. Comprehensive probabilistic analysis incorporating all the sources of uncertainties is essential for proper assessment of the exposure risk. In this study, a Monte Carlo simulation (MCS) based methodology is formulated for probabilistic assessment of water quality and contaminant risk. This is demonstrated with groundwater quality data from semiarid regions of Rajasthan. Measurement uncertainty is addressed and nonparametric method is implemented for the risk analysis. A method to ascertain the required number of MCS simulations for such application is proposed. Results indicate that the deterministic estimate could undervalue the water quality parameters as well as the actual fluoride exposure risk. This could be a reason for the increasing number of affected people in the region despite the remediation measures (based on deterministic estimates) being implemented. Therefore, water quality and contaminant risk assessment should be preferably conducted in a probabilistic approach and periodically performed with updated data for monitoring temporal variations. This article elucidates a comprehensive methodology for probabilistic risk assessment, with a case study. The body weight and water consumption parameters for different categories of populations in India are summarized from recent literature for ready reference in case of future studies. The results from the simulation analysis conducted for Rajasthan, India show the following. The deterministic estimates of the water quality parameters remain within the 95% confidence bounds of the probabilistic estimate. However, in probabilistic estimates, the central tendency (median) and the upper limits are higher than the deterministic estimates in general. Similar observations could be made about the fluoride exposure, with the upper limiting value in probabilistic analysis being up to 30% higher than the deterministic estimate (deformities in knee and hip bones). Higher variation between the deterministic and probabilistic estimates of fluoride exposure dose was observed, particularly due to the variations considered in the body weight and water intake. Therefore, it is suggested that wherever possible, probabilistic analysis should be conducted for estimating the quality of groundwater or the associated contaminant risk for better representation of the actual scenario.