Modeling, quality assessment, and Sobol sensitivity of water resources and distribution system in Shiraz: A probabilistic human health risk assessment

Abstract

Given water's vital role in supporting life and ecosystems, global climate change and human activities have significantly diminished its availability and quality. This study explores the health risks of drinking water consumption in the shiraz county water resources and distribution system. The result showed that the water was slightly alkaline. However, the average pH values during the study were within the permissible range. The area's abundance of total hardness and calcium was due to the high concentration of minerals in rocks and soils. The nitrate and fluoride concentrations in drinking groundwater varied from 0.02 to 116.70 mg/L and 0.10-1.85 mg/L, respectively. Although the water quality index indicated that 52.63, 45.03, and 20.3 percent of samples were of excellent, good, and poor quality in 2020, those percentages obtained 46.05, 52.09, and 14.0 percent in 2021. The regression values of training, testing, validation, and the proposed artificial neural network model were 0.93, 0.92, 0.85, and 0.92. The maximum levels of hazard quotient of nitrate and fluoride (except for adults) were higher than 1 in all age groups, indicating a high non-carcinogenic risk by exposure to nitrate. Furthermore, according to the Monte Carlo simulation, the 95th percentile hazard index in all groups was more than 1. Children and infants were more inclined towards risk than teens and adults based on the intake of nitrate and fluoride from drinking water. The Sobol sensitivity reflected that the nitrate concentration and ingestion rate are vital parameters that influence the outcome of the oral exposure model for all age groups. The interaction of ingestion rate with a concentration of nitrate and fluoride is an important parameter affecting the health risk assessment. In conclusion, these findings suggest that precise measures can reduce health risks and guarantee safe drinking water for residents of Shiraz County.

Keywords: Artificial neural network; Health risk assessment; Monte Carlo simulation; Positive matrix factorization; Sobol sensitivity; Water quality index.