| Sumario: | Limiting nitrogen pollution from crop production is essential for mitigating greenhouse gas emissions and protecting aquatic ecosystems while maintaining food security. Precision nitrogen management (PNM) provides a conceptual framework for achieving yield goals while maintaining nitrogen pollution within planetary boundaries by matching fertilizer rates to specific production conditions. Nevertheless, PNM strategies for smallholder contexts like India, a global nitrogen pollution hotspot, have proven costly to implement and are often ineffective. By combining survey data of production practices from 8705 wheat fields with digital soil mapping, we develop a novel PNM strategy that ‘learns from landscapes’ to generate and evaluate novel decision logic for nitrogen management. With this approach, ex-ante simulations indicate that reductions of 9% in nitrogen use and 16% in N2O emissions can be achieved without compromising yields, saving US$ 28 million per year in subsidies for the Indian state of Bihar alone. In contrast, conventional soil test-based recommendations may increase nitrogen use by 5% without corresponding yield gains. Our method that leverages large-n survey data and predictive modeling may provide a scalable pathway for PNM in similarly complex crop production environments where field and management heterogeneity is high.
|
|---|