Introduction
Nitrogen (N) is essential for modern agriculture, yet crops like maize and wheat are notoriously "leaky." More than half of the applied synthetic fertilizer is often not absorbed, escaping into the environment as reactive nitrogen (Nr). These losses occur through gaseous emissions—such as ammonia ($NH_3$) and the potent greenhouse gas nitrous oxide ($N_2O$)—and hydrologic pathways like leaching and runoff. This "nitrogen cascade" contributes to air pollution, marine "dead zones," and accelerated global warming. Quantifying these losses at high resolution has historically been a challenge, leaving a significant gap in our understanding of the global nitrogen budget.
Methodology: Machine Learning for Global Mapping
To address this gap, researchers employed Random Forest (RF) machine learning algorithms to upscale localized field experiments to a global scale. The models were trained on a massive meta-analysis of over 6,700 observations from 560 peer-reviewed publications spanning 2000–2022. By integrating global gridded data on climate, soil properties, and crop-specific management, the team predicted "loss factors" (the percentage of applied N lost) at a high resolution of 5-arcmin (approximately 10km).
Key Drivers of Nitrogen Loss
A critical finding of the study is that soil properties often outweigh fertilizer application rates in determining nitrogen loss. While N input is a factor, variables such as Soil Organic Carbon (SOC), soil pH, and climate (e.g., Mean Annual Precipitation) are primary drivers. For instance, acidic soils were found to promote higher $N_2O$ emission ratios.
The analysis highlights that roughly three-fifths of all nitrogen loss occurs through hydrologic pathways (leaching and runoff). This underscores the importance of "indirect" $N_2O$ emissions, where nitrogen travels through water systems before being released as gas elsewhere—a factor often underestimated in agricultural carbon footprints.
Global Impact and Mitigation Potential
In 2020 alone, synthetic N fertilizer use in maize and wheat production resulted in the loss of approximately 8.15 million tonnes of nitrogen (4.46 million for maize and 3.69 million for wheat). Hotspots for these losses are concentrated in high-production regions like the US Midwest, China, and India.
The study identifies significant opportunities for mitigation:
- Improving Nitrogen Use Efficiency (NUE): Boosting NUE to 60% in underperforming regions could reduce total soil Nr losses by 4 million tonnes per year and slash indirect $N_2O$ emissions by 49%.
- Enhanced-Efficiency Fertilizers (EEFs): Substituting standard fertilizers with EEFs, which control nutrient release to match crop demand ("synchrony"), could reduce global losses to 70-75% of current levels.
Conclusion and Future Directions
While this high-resolution mapping provides a vital tool for regional management, limitations remain. Data gaps in tropical regions and the exclusion of fallow-season losses suggest areas for future refinement. Additionally, while technologies like EEFs show great promise, economic barriers for farmers must be addressed through policy and innovation to achieve the necessary reductions in global nitrogen pollution.
Summary updated