Risk analysis and prediction of thallium pollution in natural water sources across China

Thallium (Tl) pollution of natural water sources is one of the serious challenges to the water environment and human health in the country. Despite the low environmental background level, localized threats still exist in polluted areas. In this study, we applied an extreme gradient boosting (XGBoost) classification model using 1986 water source Tl data and 12 relevant environmental variables to predict whether the Tl concentration in China's natural waters exceeds 0.1 μg/L, and to identify known and uncollected pollution areas. The results show that the hotspot areas, indicated by the natural water Tl risk map, include several regions in southern and southwestern China. We also explored the influence of important environmental factors related to Tl pollution, revealing that climate, lithology, and human activities combined to form Tl risk patterns. By combining the predicted results with population density maps, we found that densely populated areas facing the impacts of Tl pollution highly overlap with high-risk pollution zones. These findings emphasize the need to pay close attention to water sources at high risk of Tl pollution to ensure the safety of water for residents.