Nonlinear response of soil nitric oxide emissions to fertilizer nitrogen across croplands

Abstract

Nitric oxide (NO), as a short-lived climate forcer, has direct and indirect detrimental impacts on environmental quality and human health. The amount of nitrogen (N) fertilizer application to agricultural soils is considered a robust predictor of total NO emissions, but the estimates of cropland NO emissions have large uncertainties due to the widely used constant emission factors (EF) as e.g., default values recommended by Intergovernmental Panel on Climate Change (IPCC) methodologies. By compiling 223 field experiments with at least three N-input levels across various croplands, we performed a meta-analysis to determine how soil NO emissions respond to N inputs. Our results showed for the first time that the mean change in EF per unit of additional N input (∆EF) across all available data was significantly higher as compared to zero, indicating that the NO response to N additions increased significantly faster than the assumed linear. On average, upland grain crops showed significantly higher ∆EF than that of horticultural crops or lowland rice. A higher ∆EF was also appeared in sites with mean annual precipitation < 600 mm, mean annual temperature ≥ 15 °C, soil organic carbon ≥ 14 g C kg^− 1 or total N ≥ 1.4 g N kg^− 1, and where synthetic N fertilizers were usually applied. By assuming various N application rates, the IPCC default (0.7% or 1.1%) EF model would have overestimated or underestimated NO emissions compared to our ∆EF model. Overall, our meta-analysis results exert high potential to improve estimates of cropland NO inventories, and help address disparities in global NO budgets and develop more targeted mitigation efforts.