A multi-model study for understanding the contamination mechanisms, toxicity and health risks of hardness, sulfate, and nitrate in natural water resources.

Several water quality contaminants have attracted the attention of numerous researchers globally, in recent times. Although the toxicity and health risk assessments of sulfate and water hardness have not received obvious attention, nitrate contamination has gained peculiar research interest globally. In the present paper, multiple data-driven indexical, graphical, and soft computational models were integrated for a detailed assessment and predictive modeling of the contamination mechanisms, toxicity, and human health risks of natural waters in Southeast Nigeria. Majority of the tested physicochemical parameters were within their satisfactory limits for drinking and other purposes. However, total hardness (TH), SO, and NO were above stipulated limits in some locations. A nitrate health risk assessment revealed that certain areas present a chronic health risk to children, females, and males due to water intake. However, the dermal absorption route was found to have negligible health risks. SO in some locations was above the 100 mg/L Nigerian limit; thus, heightening the potential health effects due to intake of the contaminated water resources. Most samples had low TH values, which exposes users to health defects. There are mixed contamination mechanisms in the area, according to graphical plots, R-mode hierarchical dendrogram, factor analysis, and stoichiometry. However, geogenic mechanisms predominate over human-related mechanisms. Based on the results, a composite diagrammatic model was developed. Furthermore, predictive radial basis function (RBF) and multiple linear regression (MLR) models accurately predicted the TH, SO, and NO, with the RBF outperforming the MLR models. Insights from the RBF and MLR models were useful in validating the results of the hierarchical dendrogram, factor, stoichiometric, and graphical analyses.