Changes of air pollutant emissions in China during two clean air action periods derived from the newly developed Inversed Emission Inventory for Chinese Air Quality (CAQIEI)

Abstract. A new long-term emission inventory called the Inversed Emission Inventory for Chinese Air Quality (CAQIEI) was developed in this study by assimilating surface observations from the China National Environmental Monitoring Centre (CNEMC) using the ensemble Kalman filter (EnKF) and the Nested Air Quality Prediction Modeling System (NAQPMS). This inventory contains the constrained monthly emissions of NO_x, SO₂, CO, primary PM_2.5, primary PM₁₀, and NMVOCs in China from 2013 to 2020, with a horizontal resolution of 15 km∗15 km. This paper documents detailed descriptions of the assimilation system and the evaluation results for the emission inventory. The results suggest that CAQIEI can effectively reduce the biases in the a priori emission inventory, with the normalized mean biases ranging from −9.1 % to 9.5 % in the a posteriori simulation, which are significantly reduced from the biases in the a priori simulations (−45.6 % to 93.8 %). The calculated RMSEs (0.3 mg/m³ for CO and 9.4–21.1 μg/m³ for other species, on the monthly scale) and correlation coefficients (0.76–0.94) were also improved from the a priori simulations, suggesting that CAQIEI can reasonably reproduce the magnitude and variation of emissions of different air pollutants in China. Based on CAQIEI, we estimated China’s total emissions (including both natural and anthropogenic emissions) of the 6 species in 2015 to be as follows: 25.2 Tg of NO_x, 17.8 Tg of SO₂, 465.4 Tg of CO, 15.0 Tg of PM_2.5, 40.1 Tg of PM₁₀, and 46.0 Tg of NMVOCs. From 2015 to 2020, the total emissions reduced by 54.1 % for SO₂, 44.4 % for PM_2.5, 33.6 % for PM₁₀, 35.7 % for CO, and 15.1 % for NO_x, but increased by 21.0 % for NMVOCs. Larger emission reductions were achieved during the 2018–2020 action plan than during the 2013–2017 action plan for most species. In particular, NO_x and NMVOC emissions were shown to increase during the 2013–2017 action plain, and there were obvious emission increases in the Fengwei Plain area over the Central China region. However, NO_x and NMVOC emissions declined during the 2018–2020 action plan, and the emissions over the Fengwei Plain area also decreased. This suggests that the emission control policies were improved in the 2018–2020 action plan. We also compared CAQIEI with previous inventories, which verified our inversion results in terms of total emissions of NO_x, SO₂ and NMVOCs, and more importantly identified the potential uncertainties in our current understanding of China’s air pollutant emissions. Firstly, CO emissions in China may be substantially underestimated by current inventories, with the CO emissions estimated by CAQIEI (426.8 Tg) being more than twice the amount in previous inventories (120.7–237.7 Tg). Significant underestimations for other air pollutant emissions may also exist over western and northeastern China. In addition, the NMVOC emissions were shown to be substantially underestimated over northern China but overestimated in southern China. Secondly, the emission reduction rates during 2015–2018 estimated by CAQIEI are generally smaller than those estimated by previous inventories, especially for NO_x, PM₁₀ and NMVOCs, suggesting that the mitigation effects of the air pollution control may be overestimated currently. In particular, China’s NMVOC emissions were shown to have increased by 26.6 % from 2015 to 2018, especially over the North China Plain (by 38.0 %), Northeast China (by 38.3 %), and Central China (60.0 %). In contrast, the emissions reduction rate of CO may be underestimated. Overall, our emissions inventory sheds new light on the complex variations of air pollutant emissions in China during its two recent clean air action periods, which could significantly improve our understanding of air pollutant emissions and related changes in air quality in China. The datasets are available at https://doi.org/10.57760/sciencedb.13151 (Kong et al., 2023).