Spatially Optimized Nutrient Management as a Climate-Resilient Strategy to Reduce Nitrogen Runoff from Global Croplands

Nitrogen fertilizer, as an indispensable input in crop production, has played a crucial role in enhancing crop yields. However, the associated cropland nitrogen runoff has significantly intensified surface water eutrophication, posing serious threats to aquatic ecosystems globally. Accurate estimation of cropland nitrogen runoff is essential for designing effective mitigation strategies, yet research in this area remains constrained by limited data availability. To address this gap, we developed a machine learning model based on a globally compiled data set of field observations. Using a global spatial interpolation approach driven by the similarity of environmental and management variables, we generated crop-specific, high-resolution global maps of nitrogen runoff emission factors. Our results showed that synthetic nitrogen fertilizer application in rice, wheat, and maize fields generated approximately 2.33 Tg N yr –1 (95% CI: 1.93–2.70 Tg N yr –1 ) of nitrogen runoff in 2020. Under future climate change scenarios (SSP1-2.6, SSP3-7.0, and SSP5-8.5), global nitrogen runoff was projected to increase by −0.2–1.4% by 2060 and 0.8–3.2% by 2100. We further evaluated 96 crop-nutrient management scenarios and identified region-specific, rather than universal, nitrogen runoff mitigation pathways. Spatially optimized nutrient management could reduce cropland nitrogen runoff by 22.2% (0.52 Tg N yr –1 ) without changing total nitrogen fertilizer input and showed sustained mitigation effects across future climate change scenarios. This study provides a climate-resilient and scalable spatial optimization strategy for mitigating cropland nitrogen runoff pollution and offers actionable guidance for sustainable cropland nutrient management.