Effects of enhanced efficiency fertilizers on soil nitrous oxide emissions in corn agroecosystems: Integrating machine learning and meta-analysis

Nitrous oxide (N2O) emissions from corn (Zea mays) production are a major source of agricultural greenhouse gases. Enhanced efficiency fertilizers (EEFs) are designed to reduce nitrogen losses, yet their effectiveness varies across soils and management systems. We conducted a global meta-analysis of 61 field studies (644 observations) to quantify the effects of urease inhibitors, nitrification inhibitors, mixed inhibitors, and controlled-release fertilizers on soil N2O emissions and corn yield. All EEF types significantly reduced N2O emissions relative to conventional fertilizers, with nitrification and mixed inhibitors providing the greatest mitigation. Yield responses were consistently positive, indicating no production penalty. Subgroup analyses revealed stronger mitigation in alkaline soils, irrigated systems, deep fertilizer placement, and split applications. A machine learning Random Forest model further identified EEF type, soil organic carbon, and soil texture as the dominant predictors of mitigation, highlighting the nonlinear and context-dependent controls on EEF performance. Together, the meta-analysis and machine learning results support targeted, site-specific EEF application strategies to reduce N2O emissions while maintaining crop productivity.