Surface ozone is a major air pollutant, and summertime ozone levels in the North China Plain (NCP) are significantly higher than other northern mid-latitude regions. Despite emission reduction efforts, summertime ozone pollution in the NCP has been increasing rapidly. While decreases in particulate matter (PM) and anthropogenic nitrogen oxides (NOx), along with meteorological changes, have been suggested as contributing factors, the role of soil NOx emissions remains largely unknown. The NCP, being a major anthropogenic emission region, also has extensive cropland using significant amounts of fertilizer. This intensive agriculture leads to substantial soil NOx emissions, reaching 20% of anthropogenic NOx emissions in summer. The complex interplay between anthropogenic and biogenic NOx sources complicates ozone formation, as the efficiency of ozone formation depends on whether the region is NOx-limited, VOC-limited, or in a transitional regime. Previous studies have shown the NCP to be in a transitional or NOx-saturated regime in urban/suburban areas and NOx-limited in rural areas. While some studies suggest ozone enhancements from agricultural soil NOx in NOx-limited regions, the interaction of soil NOx emissions with anthropogenic sources in NOx-rich regions like the NCP remains unexplored. Many air quality models used in China simplify or neglect soil NOx emissions, leaving their impact unknown. This study addresses this gap by using two atmospheric chemistry models to investigate the role of soil NOx emissions in ozone pollution regulation in the NCP.

Numerous studies have documented the increasing trend of surface ozone pollution in the North China Plain (NCP), highlighting its significant impact on human health and vegetation. These studies have investigated the role of meteorological factors and changes in anthropogenic emissions of precursors like NOx and VOCs in driving these trends. Several studies have focused on the source attribution of ozone pollution using atmospheric chemistry models and satellite observations. However, a critical gap existed in the understanding of the contribution of biogenic NOx emissions, especially from agricultural soils. Existing studies primarily focused on NOx-limited regions, while the NCP exhibits a complex mix of NOx-saturated and transitional regimes, making the interaction between anthropogenic and biogenic NOx sources particularly challenging to assess. This study builds upon this existing literature by explicitly incorporating a detailed parameterization of soil NOx emissions into sophisticated atmospheric chemistry models, enabling a more comprehensive analysis of the interplay between anthropogenic and biogenic NOx sources in ozone formation and the effectiveness of control strategies.

This study employed two atmospheric chemistry models, GEOS-Chem and WRF-Chem, to simulate ozone concentrations in the NCP under different emission scenarios. Soil NOx emissions were estimated using the Berkeley-Dalhousie Soil NOx Parameterization (BDSNP), a mechanistic model that accounts for available soil nitrogen (from fertilizer application and nitrogen deposition), soil temperature, and soil moisture. The BDSNP was implemented in GEOS-Chem, allowing for online calculation of hourly soil NOx emissions at each model grid. Anthropogenic NOx emissions were obtained from the Multi-resolution Emission Inventory for China (MEIC). The GEOS-Chem model simulations were conducted at a 0.25° × 0.3125° resolution over China, with a nested higher-resolution domain over East Asia. The WRF-Chem model was also used with a 27 km resolution over eastern China. The models were used to conduct several sensitivity simulations: one base simulation and various simulations where anthropogenic sources, soil NOx emissions, or lightning NOx emissions were turned off, allowing for the quantification of the individual and interactive contributions of these sources to ozone formation. The models also examined the effectiveness of various emission reduction scenarios (for NOx, VOCs, and CO) on ozone metrics such as MDA8, NDGT70, and AOT40, both in the presence and absence of soil NOx emissions. Satellite observations of tropospheric NO2 columns from OMI were used to validate the modeled NO2 concentrations. Furthermore, the study used surface ozone observations from the China National Environmental Monitoring Center (CNEMC) to evaluate the model performance and compare simulated and observed ozone levels. The ratio of surface H2O2 to HNO3 concentrations was used as an indicator of ozone formation regime.

The study found substantial anthropogenic and soil NOx emissions in the NCP. Soil NOx emissions were estimated to be 0.18 ± 0.01 Tg N a⁻¹ in the NCP, representing 11-20% of anthropogenic sources in July 2008-2017, with fertilizer-induced emissions accounting for 58% in July. These findings were supported by comparisons to satellite observations and field measurements. The models showed that the presence of soil NOx emissions significantly reduces the sensitivity of surface ozone to anthropogenic NOx emissions. In the GEOS-Chem model, the maximum ozone air quality improvements from controlling all domestic anthropogenic emissions decreased by 30% due to soil NOx. A similar reduction (around 30%) was observed in the WRF-Chem model. The effect of soil NOx on anthropogenic ozone was consistent across years. The study also demonstrated that the presence of soil NOx emissions increases the required reduction in anthropogenic NOx emissions to achieve target ozone levels, creating a soil NOx penalty. For example, to achieve a 5 ppbv reduction in MDA8 ozone, 41% of anthropogenic NOx reduction was needed without considering soil NOx emissions, but 54% was required with soil NOx emissions included (a 13% penalty). The penalty increased to 15% for a more ambitious 15 ppbv ozone reduction target. Joint reductions in anthropogenic NOx, VOCs and CO emissions mitigated this penalty to some degree. Analysis of ozone formation regime using H2O2/HNO3 ratio indicated that soil NOx emissions delayed the shift to a NOx-sensitive regime as anthropogenic NOx emissions decreased, further amplifying the soil NOx penalty. Different spatial patterns of ozone response were observed when comparing the effects of reducing anthropogenic vs. soil NOx emissions, reflecting the spatial heterogeneity of sources and ozone production efficiency.

The findings of this study highlight the critical underestimation of the role of agricultural soil NOx emissions in ozone pollution regulation in the NCP. The significant reduction in the effectiveness of anthropogenic emission control measures due to soil NOx emissions has major implications for air quality management in the region. The soil NOx penalty emphasizes the need to incorporate biogenic NOx sources into ozone mitigation strategies. Current emission control strategies focused primarily on anthropogenic sources underestimate the effort needed to achieve air quality goals. The complex interaction between anthropogenic and soil NOx emissions, varying across different ozone regimes, further underscores the importance of accurate source apportionment and targeted emission control strategies.

This study reveals the significant and previously underappreciated role of agricultural soil NOx emissions in ozone pollution regulation in the North China Plain. The presence of soil NOx emissions significantly suppresses the sensitivity of ozone concentrations to reductions in anthropogenic NOx, leading to a substantial emission control penalty. This penalty necessitates a reassessment of current emission control strategies. Future strategies should account for this soil NOx penalty by incorporating them into model predictions and emission reduction targets. Further research should focus on improving estimates of soil NOx emissions through enhanced direct measurements and refined model parameterizations, accounting for the role of other soil-emitted reactive nitrogen species like HONO. These insights are also relevant to other regions with high emissions of both anthropogenic and soil NOx, such as the Indo-Gangetic Plain.

The study relies on modeled estimations of soil NOx emissions, which are subject to uncertainties associated with the BDSNP parameterization and input data (fertilizer application, meteorological variables). Although the model performance was validated against observations, model uncertainties could influence the exact magnitude of the soil NOx penalty. The study focused primarily on summertime ozone pollution, and the role of soil NOx emissions might vary in other seasons. Future research should investigate seasonal variations of the soil NOx penalty.