Reevaluating\nthe\nDrivers of Fertilizer-Induced N<sub>2</sub>O Emission: Insights from\nInterpretable Machine Learning

Direct nitrous oxide (N<sub>2</sub>O) emissions from\nfertilizer\napplication are the largest anthropogenic source of global N<sub>2</sub>O, but the factors influencing these emissions remain debated. Here,\nwe compile 1134 observations of fertilizer-induced N<sub>2</sub>O\nemission factor (EF) from 229 publications, covering various regions\nand crops globally. We then employ an interpretable machine learning\nmodel to investigate the driving factors of fertilizer-induced N<sub>2</sub>O emissions. Our results reveal that pH, soil organic carbon,\nprecipitation, and temperature are the most influential factors, overweighing\nthe impacts of management practices. Nitrogen application rate has\na positive impact on the EF, but the effect diminishes as nitrogen\napplication rate increases, which has been overestimated in previous\nstudies. Soil pH has three-stage influence on EF: positive when 7.3\n≤ pH ≤ 8.7, significantly negative between 6.8 and 7.3,\nand insignificant at lower pH levels (4.7 ≤ pH ≤ 6.8).\nMoreover, we confirm the nonlinear contributions of temperature and\nprecipitation to EF, which may cause an unexpected increase in N<sub>2</sub>O emission under climate change. Our research provides crucial\ninsights for global N<sub>2</sub>O modeling and mitigation strategies.