Environmental Health and Water Pollution Nexus: Insights into the Vulnerability of Ecosystem Health

Environmental health monitoring is a valuable tool for investigating ecological and environmental changes related to water pollution risks. Sensitive communications between ecological health, environmental parameters, and water quality (WQ) are desired to maintain sustainable water resources (WRs). By implementing new technical scheme, this study will effectively explore the water pollution behavior and environment-health nexus through a effective framework. The current methodology combines deep learning (DL) approaches, long short-term memory (LSTM) with three modern metaheuristic optimizers, Heat Transfer Search (HTS), Flow Direction (FD), and Equilibrium Optimization (EQ), respectively. This study applied explainable artificial intelligence (XAI) by implementing Shapely Additive Explanation (SHAP), and simultaneously uncertainty analysis through Bayesian Inference Uncertainty Analysis (BIUA), feature importance analysis (FIA) through SHAP analysis, to capturing the optimum input features. The South Platte River basin has been chosen as the study area in United States, by using the United States Geological Survey (USGS) archive for fundamental data. As the result showed, the highest precision of the EQ-LSTM model was noticed during the estimation of dissolved oxygen (R²= ۰.۷۸), while the highest precision of the FD-LSTM model was observed during the estimations of pH (R²= ۰.۷۹). The FD and EQ optimizers have been proven to be robust with low error margins of ۰.۰۰۲ and ۰.۰۲۲, respectively, after considering BIUA, which increases the model's confidence during fluctuating environmental conditions. Furthermore, some of the pH instabilities could emerge as a threat to the ecosystems in the future. The present work can contribute more toward the achievement of the United Nations Sustainable Development Goals (SDGs) through the robustness of the WRS pollution extremes prediction and related hazards.

کلیدواژه ها:

Water resources ، Sustainable Development Goals ، Heat transfer search ، Flow direction ، Equilibrium optimization ، explainable artificial intelligence