Qian Qu, Shuting Wang, Xiangang Hu
Riverine antibiotic pollution has gained significant attention owing to its negative impacts on the environment and human health. Although previous studies have revealed the spatial distribution of antibiotics in local regions, their vulnerability to antibiotic risks, environmental drivers, policy intervention effectiveness, and potential opportunities for mitigation remain unknown. Here, we constructed a random-forest-algorithm-assisted risk assessment framework and intelligent watershed management to capture the temporal and spatial variations in global river antibiotic risks. The interaction between precipitation and livestock density as a key driver explained ∼43% of the antibiotic risk variance. Rivers in Africa are highly vulnerable to antibiotic pollution. We provide empirical evidence supporting the effectiveness of antibiotic use control policies, identifying more than 15 successful interventions that significantly reduce risks. Notably, policy instruments demonstrate greater effects when implemented as a mix rather than in isolation. Optimization analysis demonstrated that reducing wastewater discharge (4.82 ± 0.04%) and livestock density (10.23 ± 0.05%) would alleviate antibiotic risk by 10.22 ± 0.08%, which would benefit ecological functions (e.g., water purification, gross primary productivity, and leaf litter decomposition) in rivers. These findings provide novel insights for advancing global antibiotic mitigation and strengthening integrated watershed management.