Mapping and mitigating global perchlorate hazards in drinking water sources

Perchlorate exposure via drinking water has been extensively related to the incidence of thyroid diseases. However, escalating perchlorate contamination in drinking water sources remains underprioritized across much of the globe. Here, we report a geospatial model depicting the probability of perchlorate exceeding 70 μg/L in natural water under diverse climates and industrial contexts, encompassing nearly 200 countries/regions lacking robust monitoring data. We demonstrate that ∼145 million people are at risk of perchlorate exposure, with 93.6% residing in low- and middle-income countries. Upon geospatial mapping, we reveal that ∼80 countries/regions are facing potential public health threats caused by perchlorate exposure, which is previously unreported. Furthermore, by leveraging reinforcement learning, we propose spatially differentiated control strategies to achieve maximum reduction in perchlorate-affected populations with minimal cost. These key insights will contribute to the sustainable development of perchlorate-related industries and facilitate the achievement of Sustainable Development Goal 6 (Clean Water for All).