Diverse response of surface ozone to COVID-19 lockdown in China
Yiming Liu, Tao Wang, Trissevgeni Stavrakou, Nellie Elguindi, Thierno Doumbia, Claire Granier, Idir Bouarar, Benjamin Gaubert, Guy P. Brasseur
1 Diverse response of atmospheric ozone to COVID-19 lockdown in China
Yiming Liu , Tao Wang Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hong Kong, 999077, China Correspondence to: Tao Wang ([email protected]) and Yiming Liu ([email protected])
Abstract
Ozone (O ) is a key oxidant and pollutant in the lower atmosphere. Significant increases in surface O have been reported in many cities during the COVID-19 lockdown. Here we conduct comprehensive observation and modeling analyses of surface O across China for periods before and during the lockdown. We find that O decreased in the subtropical south, in contrast to increases in most other regions. Meteorological changes and emission reductions both contributed to the observed O changes. The larger reduction in NO x than volatile organic compounds induced O increases in the NO x -saturated north, whereas the comparably large decreases in both precursors resulted in O declines in the south. Our study highlights the importance of meteorology in the short-term O variability and the complex dependence of O on its precursors. Our findings are relevant to elucidating the underlying factors driving O changes during the COVID-19 lockdown in other parts of the world. The outbreak of coronavirus disease 2019 (COVID-19) has severely threatened public health worldwide, leading to millions of deaths . China, where the first case of COVID-19 was reported in the city of Wuhan, imposed country-wide measures from January 23 to February 13, 2020 to prevent the spread of the disease, including social distancing, teleworking, and closure of non-essential businesses
2, 3 . These restrictions drastically reduced transportation and industrial activities, resulting in a sharp decrease in emissions of air pollutants
4, 5 . The huge and large-scale emission reductions during the COVID-19 lockdown can be treated as a natural outdoor experiment to improve our understanding of the air pollutants' response to emission control. According to satellite and surface observations, NO concentrations decreased by over 50% in China during the lockdown period
6, 7 . The concentrations of other pollutants, including SO , particulate matter with an aerodynamic diameter less than 2.5 μm (PM ), particulate matter with an aerodynamic diameter less than 10 μm (PM ), and carbon monoxide (CO), also declined in a large area of China . However, surface ozone (O ) concentrations in northern China increased by factors of 1.5–2.0 . Similar O increases have been reported in southern Europe, India, and Brazil despite the large decrease in other pollutants
9, 10, 11 . However, the underlying factors driving the O changes during the city lockdowns remain unclear. This study analyzes surface O data across China before and during the country-wide lockdown. We find that O decreased in southern China while increasing in most other regions during the lockdown. Using a regional chemistry transport model, we isolate the impacts of meteorological changes and anthropogenic emission reductions on O . Our results highlight the importance of meteorological influences on the short-term O changes and the diverse response of O to its precursors’ emission reduction in different climatic and emission-mix regions. Results Observed O changes in different parts of China. The country-wide measures to control the spread of COVID-19 were implemented starting from January 23, 2020 (the exact date varies for different cities), just before the Chinese New Year. All enterprises remained closed until no earlier than February 13, except those required for essential public services, epidemic prevention and control, and residential life needs. We focused on the period during the COVID-19 lockdown from January 23 to February 12, 2020 (hereafter referred to as the CLD period), 3 weeks in total. We derived the changes in O and other pollutants by comparing the CLD period with the 3 weeks before the COVID-19 outbreak, from January 2 to 22, 2020 (hereafter referred to as the pre-CLD period). We focused on three typical regions in China (Fig. S1): north China (NC, including Beijing, Tianjin, Hebei, and western Shandong), central China (CC, including Hubei province, where Wuhan is situated, and the surrounding regions), and south China (SC, including the Pearl River Delta and the surrounding regions). Our NC region is situated in the North China Plain, which is known to suffer from severe haze in winter; CC was the original epicenter of the COVID-19 outbreak and is an important economic hub for the central regions of China; the Pearl River Delta, where the megacities of Guangzhou and Shenzhen are situated, is the most developed region in southern China. Figures 1 and S2 present the changes in observed concentrations of O and other pollutants during the CLD period compared to pre-CLD. The concentrations of CO, PM , and SO in China declined by approximately 20%, 30%, and 20%, respectively. As a precursor of O , NO decreased by about 50% across China as a whole, and by similar amounts in all regions (Fig. S2a). However, the O mixing ratio exhibited varying changes in different regions (Fig. 1a). In NC and CC, the whole-day average O increased significantly, by 112% and 73%, respectively (Fig. 1b); in contrast, it decreased slightly by 3% in SC. The O changes also varied between daytime and nighttime. During daytime, the O increase in most parts of China was smaller than the all-day average (Fig. 1c), 92% and 71% in the NC and CC regions (Fig. 1d). In the SC region, most stations saw a decrease in O during daytime, leading to a regional average O drop of 14%. During nighttime, the O mixing ratio increased significantly across China (Fig. 1e), by 154%, 77%, and 18% in NC, CC, and SC, respectively (Fig. 1f). These results reveal the diverse response of O during the lockdown in different regions and times of day. Contribution of meteorological changes and emission reductions to O . Ground-level O is produced by chemical reactions of O precursors, namely, NO x , volatile organic compounds (VOCs), and CO, under sunlight. Meteorology can modulate O concentrations by affecting the chemical reaction rates, dilution, deposition, transport, and natural emissions of precursors. In this study, we used a regional chemistry transport model (Weather Research and Forecasting model-Community Multiscale Air Quality model, WRF-CMAQ) to reproduce the changes in O during the lockdown in China. Figure S3 presents the observed and simulated O mixing ratios during the pre-CLD and CLD periods. Before the COVID-19 outbreak, the O mixing ratio was generally higher in southern China than in central and northern China (Fig. S3a, c, e). After the emission reductions associated with the lockdown, the O mixing ratio increased significantly in central and northern China, resulting in comparable levels to southern China (Fig. S3b, d, f). The WRF-CMAQ model performed reasonably well in capturing these O changes (Fig. 2a, d, g). Based on the modeling results, we separated the impacts of meteorological changes and emission reductions on the changes in O (Fig. 2). We found that the meteorological changes generally made a greater contribution to the O variation than did the emission reductions. For the whole-day average, the meteorological changes made a significantly positive contribution to O in central (by 5–10 ppbv) and northern (by 2–10 ppbv) China, but a negative contribution (by 1–2 ppbv) in southern China, Tibet, and eastern Neimenggu (Fig. 2b). The emission reduction contributed positively to O (by 2–5 ppbv) in the North China Plain, Yangtze River Delta, and some other urban areas, but negatively (by 0–4 ppbv) in other regions (Fig. 2c). During daytime, the effect of meteorological changes on O was similar to those for the whole-day average, but the positive contribution to O was slightly lower (Fig. 2e). The positive contribution of emission reduction to O was also lower in daytime than the all-day average, with positive contributions of 2–5 ppbv in the North China Plain and negative contributions in a broader rural area (Fig. 2f). During nighttime, the meteorological contributions to the O mixing ratio were less strongly positive (Fig. 2h), whereas those due to emission reductions were more strongly positive (Fig. 2i). Figure 3 shows the observed and simulated O changes in the NC, CC, and SC regions and the simulated impacts of meteorological changes and emission reductions. The observed O changes in these regions were generally captured by the simulation. However, the model underestimated the increase in NC and CC and over-predicted the decrease in SC, which can be attributed to the uncertainties of the meteorological simulations and estimated emission reductions across China (see Supplementary Information for detailed discussions). Overall, the results showed that the meteorological changes and emission reductions both played an important role in the O changes in these regions. For the whole-day average, the O increase in NC was attributed to the comparably sized contributions from both meteorological changes (3.1 ppbv) and emission reductions (3.3 ppbv) (Fig. 3a). In CC (Fig. 3b), the meteorological change (6.7 ppbv) was the primary cause of the O increase, whereas the contribution of emission reduction was much lower (0.8 ppbv). In SC (Fig. 3c), the meteorological changes (-1.7 ppbv) and emission reductions (-0.7 ppbv) both contributed to the O decrease. The trends in daytime were similar to the whole-day trends. During nighttime, the emission reduction contributed positively to O in all three regions (including SC), and its impact was stronger; the effect of meteorological changes weakened at night. Impacts of meteorological changes on O . The impacts of meteorological changes on O for the NC, CC, and SC regions can be explained by the changes in the overall weather pattern and specific meteorological factors. In winter, China’s land area is generally controlled by a cold high-pressure system (Fig. S4). During our study period, the center of this high-pressure system was located in northern China, moving southward from the pre-CLD to the CLD period, with weakening strength. During this time, the high-pressure system therefore became increasingly dominant in southern China, and the strengthened southward winds brought colder and drier air masses from the north (Fig. S5c), which decreased the temperature and specific humidity in local areas (Fig. S5a and b). In contrast, in central and northern China, the winds shifted to a more northward direction, transporting warmer and wetter air masses from the south (Fig. S5c), which increased the temperature and specific humidity (Fig. S5a and b). During daytime, the decrease in temperature in southern China, accompanied by increases in central and northern China, weakened and enhanced, respectively, the surface O chemical production
12, 13 , contributing to the respective decrease and increase in O . Biogenic emission is an important source of VOCs and thereby contributes to O formation in China . The temperature changes led to an increase and a decrease in biogenic emissions in central and southern China, respectively (Fig. S6). Thus, the temperature changes increased and decreased O in central and southern China, respectively, by influencing chemical reaction rates directly and altering biogenic emissions indirectly. The changes in the weather pattern also resulted in less cloud and precipitation in northern and central China, but more cloud and precipitation in southern China (Fig. S5e and f). Cloud can reduce the amount of solar radiation reaching the surface and thus the chemical production of O , while precipitation can also reduce O through the scavenging of its precursors
16, 17 . These opposite changes in cloud fraction and precipitation therefore helped to increase O in central and northern China and decrease O in southern China. Furthermore, in northern and central China, the significant increase in planetary boundary layer height during the lockdown promoted the transport of O from the upper air to the surface, contributing to the O increase in these regions
18, 19 . The increase in specific humidity in northern and central China and the decrease in southern China might also have contributed negatively and positively to the O mixing ratios in those regions, respectively
20, 21 (Fig. S5b). During nighttime, the changes in meteorological factors were similar to those in daytime (Fig. S7). However, the changes in temperature and cloud fraction exerted smaller impacts on O changes than those in daytime, considering the negligible biogenic emissions and solar radiation at nighttime. Thus, the effects of meteorological changes in the NC, CC, and SC regions during nighttime were generally smaller than during daytime (Fig. 3), and they became negligible in SC (Fig. 3c). Response of O to emission reductions. We further investigated the impact of multi-pollutant reductions on the O changes. Because transportation and industrial activities were reduced significantly during the lockdown and they were the major sources of NO x (>80%) and VOCs (>60%) (Fig. S8), the reductions of NO x and VOC emissions were more significant than those for CO, particulate matter (PM), and SO (Fig. 4a, c, e). The estimated NO x emission reductions were 42%, 48%, and 51% in the NC, CC, and SC regions, respectively, while the respective reductions for VOC emissions were 32%, 28%, and 45%. The relationship between O and its precursor emission is non-linear, especially for NO x . The reduction of NO x emission contributed positively to O in a large area of China due to the NO titration effect during nighttime (Fig. S9c), whereas the impact of NO x reduction was more complicated during daytime (Fig. S9b): NO x reduction contributed positively to O in NO x -saturated regions, including northern China and some urban areas in other regions, but negatively in NO x -limited regions, including southern China and other rural regions. The reduction of VOC emission contributed negatively to O across China (Fig. S9d-f). As an O precursor, the reduction of CO emission also negatively affected the O mixing ratio (Fig. S9g-i); in contrast, the PM and SO emissions reductions contributed positively to O (Fig. S9j-o) through the weakening of aerosol effects
22, 23 , but their impacts were much smaller and were negligible (< 1 ppbv) due to the smaller reductions, compared with NO x and VOCs. The response of O to the emission reductions depended on the levels of NO x and VOC reductions in different regions. For the whole-day average, in NC, the positive contribution of the NO x reduction counteracted the negative contribution of the VOC reduction, leading to a substantial net O increase (Fig. 4b). In CC, the contributions of the NO x and VOC reductions were comparable in magnitude, and their opposite impacts resulted in only a slight change in O (Fig. 4d). In SC, the impact of the NO x reduction on O was smaller than that of the reduction of VOCs, leading to a net decrease in O (Fig. 4f). The impacts of emission reductions on O during daytime were similar to those for the whole-day average. During nighttime, the effect of the VOC reduction was weakened due to the lower rate of degradation of VOCs by radicals compared with daytime, and the O level increased in all three regions due to the NO x reduction. The above results show that the contribution of NO x reductions (with NO x falling by 42%–51%) to the rise in O decreased from NC to CC and to SC, indicating the decreasing level of NO x saturation from north to south. In contrast, the impact of the VOC reduction on O increased from north to south, which can in part be attributed to the regional variation of VOC reductions. In the SC region, transportation and industry are the predominant sources of VOCs (97%, compared with 85% and 60% in NC and CC, respectively) (Fig. S8). During the CLD period, the reduction of VOC emission in SC (45%) was significant and comparable with the NO x reduction (51%). In contrast, the VOC reductions in the NC (32%) and CC (28%) regions were much lower (Fig. 4a, c) and could not offset the positive impact of NO x reduction on O . The residential sector (mainly household coal burning) is an important source of VOC emission in the NC and CC regions, whereas its contribution is smaller in SC. The residential emissions in some northern rural areas increased during the CLD period because many migrant workers came back for the Chinese New Year holiday and were stranded there due to the lockdown
4, 8 . Discussion and implication
The drastic emission reductions during the COVID-19 outbreak offered an unprecedented real-world scenario to examine the human impacts on the air quality and obtain knowledge to help design more effective air pollution control policies. The responses of air quality to short-term emission controls have been studied during a number of public and political events in China, such as the Beijing Summer Olympic Games (August 2008)
24, 25 , the Nanjing Youth Olympic Games (August 2014)
26, 27 , the Asia-Pacific Economic Cooperation (APEC) meeting in Beijing (November 2014)
28, 29 , the Grand Military Parade in Beijing (September 2015)
28, 30 , and the G20 summit in Hangzhou (September 2016)
31, 32 . During these events, various emission-reducing measures were implemented in the cities concerned and their surrounding areas. Whereas atmospheric concentrations of primary air pollutants (NO x , CO, primary PM, and SO ) in the concerned cities generally decreased in response to the temporary control measures, the O concentrations showed mixed responses. O decreased during the Grand Military Parade and the Nanjing Youth Olympic Games but increased during the Beijing Olympic Games , G20 summit , and APEC meeting . The different O responses can be qualitatively attributed to differences in the meteorological conditions (including regional transport of air masses) and different control measures implemented by the local governments. Compared with the previously studied situations, the COVID-19 lockdown is unique in that emissions decreased across the whole country (and later the entire world) as opposed to a specific city or region, and the decreases were also much more drastic than those due to transportation restrictions alone. Moreover, the COVID-19 lockdown took place in winter, whereas the previous interventions occurred in summer and autumn, when meteorology and atmospheric chemistry are likely to be different from winter. The present study found differences in the responses of surface O to the emission decreases during the COVID-19 lockdown in northern (large increase) and southern (moderate decrease) China. Meteorology played a key role in the O changes during the lockdown, highlighting the necessity of taking meteorology into account when assessing the short-term response of O (and other air pollutants) to emission reductions . The COVID-19 lockdown led to larger reductions in NO x and VOCs than in SO and PM (Fig. 4). As NO x and VOCs are precursors of O , the changes in O are more sensitive to their reduction than to the reduction in PM (and SO ), which could increase in the photolysis rates and reduce uptake of reactive gases. Previous studies showed that the increase in summer O concentration from 2013 to 2017 in China was due to a larger reduction of PM and SO than of NO x35 . The findings of such studies on the response of O to different real-world emission reduction scenarios demonstrate the complex relationship of O with not only its precursors (NO x and VOC) but also particulates, highlighting the need for a holistic strategy to control both primary pollutants and secondary pollutants like O . In conclusion, the first country-wide lockdown during the COVID-19 outbreak in China drastically reduced transportation and industrial activities, leading to sharp declines in air pollutant emissions from these sectors. However, atmospheric O responded differently in the northern and southern regions, which can be explained by changes in meteorology and differences in the O chemistry regimes and the magnitudes of precursor reductions in these regions. Our study shows that the extent of VOC reduction during lockdown, which suppressed O formation, was insufficient to offset the NO titration effect in northern China, and that larger reductions of VOCs (e.g., from residential sectors) would have been needed to reduce the O concentration in the north. The rising O concentration in northern China during the COVID-19 lockdown and in recent winters should receive greater attention because O boosts the atmospheric oxidative capacity and therefore production of secondary aerosols
5, 36 , which are important components of winter haze in northern China. Our findings in China are relevant to untangling the underlying factors driving the O changes in other parts of the world during their COVID-19 lockdowns. Methods Surface measurement data.
We obtained the observed concentrations of surface O and other pollutants (PM , SO , CO, NO ) from the China National Environmental Monitoring Center (http://106.37.208.233:20035/). There are a total of 1643 environmental monitoring stations in China. Data quality control was conducted for the measurement data in accordance with previous studies
37, 38 . Fig. S1 shows the locations of the environmental monitoring sites.
Numerical modeling.
The CMAQ (v5.2.1) model was applied to simulate the O mixing ratios over China from January 2 to February 12, 2020. The WRF (v3.5.1) model was driven by the dataset of the National Center for Environmental Prediction (NCEP) FNL Operational Model Global Tropospheric Analyses with a horizontal resolution of 1° × 1° and provided meteorological inputs for the CMAQ model. The CMAQ model’s modeling domain covered the entire land area of China and its surrounding regions, with a horizontal resolution of 36 km × 36 km. SAPRC07TIC
39, 40 and AERO6i
41, 42 were adopted as the gas-phase chemical mechanism and aerosol mechanism, respectively. The CMAQ model has been improved with updated heterogeneous reactions to better predict the O concentration; details can be found in Liu and Wang . Emissions from the other countries were derived from the MIX emission inventory . International shipping emissions were taken from the Hemispheric Transport Atmospheric Pollution (HTAP) emission version 2.2 dataset based on the year 2010 . Biogenic emissions were calculated by the Model of Emissions of Gas and Aerosols from Nature (MEGAN) version 2.1 with meteorological inputs from the WRF model. Two experiments were conducted to investigate the impacts of meteorological changes and emission reduction on O during the CLD period. The baseline experiment was simulated with the same anthropogenic emissions for the pre-CLD and CLD periods. Based on the baseline run, the other experiment (Reduction Case) was performed with 70% and 40% emission reductions of transportation and industry, respectively, during the CLD period. Because there is no official emission inventory for the CLD period, these emission reductions were estimated according to the literature
4, 5 . Comparing these two experiments, the O changes during the CLD period relative to the pre-CLD period for the Reduction Case were considered to be entirely due to the meteorological changes and emission reductions. The impacts of the meteorological changes were quantified by subtracting the O mixing ratios of the CLD period from those of the pre-CLD period for the baseline experiment, while the impacts of emission reduction were estimated by comparing the O mixing ratio during the CLD period between the Reduction Case and the baseline experiment. Furthermore, we individually reduced the emissions of NO x , VOCs, CO, PM, and SO during the CLD period to elucidate the response of O to each pollutant’s reduction. The performance of the CMAQ model in simulating the O , NO , PM , SO , and CO concentrations for the Reduction Case was evaluated (Fig. S10 and Table S2), showing reasonable agreements with the respective surface observations. Details of the emission estimation and the model evaluation are presented in Supplementary Information. Data availability
The codes or data used in this study are available upon request from Yiming Liu ([email protected]) and Tao Wang ([email protected]).
Acknowledgements
This work was supported by the Hong Kong Research Grants Council (T24-504/17-N) and the National Natural Science Foundation of China (91844301). We would like to thank Prof. Qiang Zhang from Tsinghua University for providing the emission inventory.
Author contributions
T.W. initiated the research. Y.M.L. and T.W. designed the research framework. Y.M.L. performed model simulations and drew the figures. T.W. and Y.M.L. analyzed the results and wrote the paper.
Competing interests
The authors declare that they have no conflict of interest.
Materials & Correspondence
Correspondence and requests for materials should be addressed to T.W. or Y.M.L.
Additional information
Supplementary information is available for this paper.
Reference
1. WHO. Coronavirus Disease (COVID-2019) Situation Reports. World Health Organization (2020). 2. Chinazzi M , et al. The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak.
Science , 395-400 (2020). 3. Li R , et al.
Substantial undocumented infection facilitates the rapid dissemination of novel coronavirus (SARS-CoV-2).
Science , 489-493 (2020). 4. Wang P, Chen K, Zhu S, Wang P, Zhang H. Severe air pollution events not avoided by reduced anthropogenic activities during COVID-19 outbreak.
Resources, Conservation and Recycling , 104814 (2020). 5. Huang X , et al.
Enhanced secondary pollution offset reduction of primary emissions during COVID-19 lockdown in China.
National Science Review , (2020). 6. Shi X, Brasseur GP. The Response in Air Quality to the Reduction of Chinese Economic Activities During the COVID-19 Outbreak.
Geophys Res Lett , e2020GL088070 (2020). 7. Zhang R, Zhang Y, Lin H, Feng X, Fu T-M, Wang Y. NOx Emission Reduction and Recovery during COVID-19 in East China. Atmosphere , 433 (2020). 8. Wang Y, Yuan Y, Wang Q, Liu C, Zhi Q, Cao J. Changes in air quality related to the control of coronavirus in China: Implications for traffic and industrial emissions. Sci Total Environ , 139133 (2020). 9. Sicard P , et al.
Amplified ozone pollution in cities during the COVID-19 lockdown.
Sci Total Environ , 139542 (2020). 10. Siciliano B, Dantas G, da Silva CM, Arbilla G. Increased ozone levels during the COVID-19 lockdown: Analysis for the city of Rio de Janeiro, Brazil.
Sci Total Environ , 139765 (2020). 11. Sharma S, Zhang M, Anshika, Gao J, Zhang H, Kota SH. Effect of restricted emissions during COVID-19 on air quality in India.
Sci Total Environ , 138878 (2020). 12. Lee YC , et al.
Increase of ozone concentrations, its temperature sensitivity and the precursor factor in South China.
Tellus B: Chemical and Physical Meteorology , 23455 (2014). 13. Fu TM, Zheng YQ, Paulot F, Mao JQ, Yantosca RM. Positive but variable sensitivity of August surface ozone to large-scale warming in the southeast United States. Nat Clim Change , 454-458 (2015). 14. Wu K , et al. Estimation of biogenic VOC emissions and their corresponding impact on ozone and secondary organic aerosol formation in China.
Atmos Res , 104656 (2020). 15. Lelieveld J, Crutzen PJ. Influences of Cloud Photochemical Processes on Tropospheric Ozone.
Nature , 227-233 (1990). 16. Seinfeld JH, Pandis SN.
Atmospheric Chemistry and Physics-from Air Pollution to Climate Change . John Wiley & Sons (2006). 17. Shan W, Yin Y, Zhang J, Ding Y. Observational study of surface ozone at an urban site in East China.
Atmos Res , 252-261 (2008). 18. He J , et al. Air pollution characteristics and their relation to meteorological conditions during 2014–2015 in major Chinese cities.
Environ Pollut , 484-496 (2017). 19. Sun Y, Wang Y, Zhang C. Vertical observations and analysis of PM2.5, O3, and NOxat Beijing and Tianjin from towers during summer and Autumn 2006.
Advances in Atmospheric Sciences , 123 (2009). 20. Ma T , et al. Air pollution characteristics and their relationship with emissions and meteorology in the Yangtze River Delta region during 2014-2016.
J Environ Sci , 8-20 (2019). 21. Li R , et al. Air pollution characteristics in China during 2015–2016: Spatiotemporal variations and key meteorological factors. Sci Total Environ , 902-915 (2019). 22. Li K, Jacob DJ, Liao H, Shen L, Zhang Q, Bates KH. Anthropogenic drivers of 2013-2017 trends in summer surface ozone in China.
P Natl Acad Sci USA , 422-427 (2019). 23. Li K , et al.
A two-pollutant strategy for improving ozone and particulate air quality in China.
Nat Geosci , 906-910 (2019). 24. Wang T , et al. Air quality during the 2008 Beijing Olympics: secondary pollutants and regional impact.
Atmos Chem Phys , 7603-7615 (2010). 25. Wang S , et al. Quantifying the Air Pollutants Emission Reduction during the 2008 Olympic Games in Beijing.
Environ Sci Technol , 2490-2496 (2010). 26. Liang H, Chen ZM, Huang D, Zhao Y, Li ZY. Impacts of aerosols on the chemistry of atmospheric trace gases: a case study of peroxides and HO2 radicals. Atmos Chem Phys , 11259-11276 (2013). 27. Li SW , et al. Influence of pollution control on lead inhalation bioaccessibility in PM2.5: A case study of 2014 Youth Olympic Games in Nanjing.
Environment International , 69-75 (2016). 28. Liu H , et al. A paradox for air pollution controlling in China revealed by “APEC Blue” and “Parade Blue”.
Sci Rep-Uk , 34408 (2016). 29. Wang Z , et al. Changes in atmospheric composition during the 2014 APEC conference in Beijing. , 12695-12707 (2015). 30. Wang ZS , et al.
Analysis on air quality in Beijing during the military parade period in 2015.
China Environ Sci , 1628-1636 (2017). 31. Wu K , et al. Evolution and Assessment of the Atmospheric Composition in Hangzhou and its Surrounding Areas during the G20 Summit.
Aerosol Air Qual Res , 2757-2769 (2019). 32. Yang WX, Yuan GH, Han JT. Is China's air pollution control policy effective? Evidence from Yangtze River Delta cities. J Clean Prod , 110-133 (2019). 33. Huang Q, Wang T, Chen P, Huang X, Zhu J, Zhuang B. Impacts of emission reduction and meteorological conditions on air quality improvement during the 2014 Youth Olympic Games in Nanjing, China.
Atmos Chem Phys , 13457-13471 (2017). 34. Liu Y, Wang T. Worsening urban ozone pollution in China from 2013 to 2017 – Part 1: The complex and varying roles of meteorology. Atmos Chem Phys , 6305-6321 (2020). 35. Liu Y, Wang T. Worsening urban ozone pollution in China from 2013 to 2017 – Part 2: The effects of emission changes and implications for multi-pollutant control. Atmos Chem Phys , 6323-6337 (2020). 36. Fu X , et al. Persistent Heavy Winter Nitrate Pollution Driven by Increased Photochemical Oxidants in Northern China.
Environ Sci Technol , 3881-3889 (2020). 37. Lu X , et al. Severe Surface Ozone Pollution in China: A Global Perspective.
Environ Sci Tech Let , 487-494 (2018). 38. Song CB , et al. Air pollution in China: Status and spatiotemporal variations.
Environ Pollut , 334-347 (2017). 39. Carter WPL. Development of the SAPRC-07 chemical mechanism.
Atmos Environ , 5324-5335 (2010). 40. Hutzell WT, Luecken DJ, Appel KW, Carter WPL. Interpreting predictions from the SAPRC07 mechanism based on regional and continental simulations. Atmos Environ , 417-429 (2012). 41. Murphy BN , et al. Semivolatile POA and parameterized total combustion SOA in CMAQv5.2: impacts on source strength and partitioning.
Atmos Chem Phys , 11107-11133 (2017).
42. Pye HOT , et al.
On the implications of aerosol liquid water and phase separation for organic aerosol mass.
Atmos Chem Phys , 343-369 (2017). 43. Zheng B , et al. Trends in China's anthropogenic emissions since 2010 as the consequence of clean air actions.
Atmos Chem Phys , 14095-14111 (2018). 44. Li M , et al. MIX: a mosaic Asian anthropogenic emission inventory under the international collaboration framework of the MICS-Asia and HTAP.
Atmos Chem Phys , 935-963 (2017). 45. Janssens-Maenhout G , et al. HTAP_v2.2: a mosaic of regional and global emission grid maps for 2008 and 2010 to study hemispheric transport of air pollution.
Atmos Chem Phys , 11411-11432 (2015). 46. Guenther AB , et al. The Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN2.1): an extended and updated framework for modeling biogenic emissions.
Geosci Model Dev , 1471-1492 (2012). Figures
Fig. 1 Observed changes in O mixing ratios across China before and during the COVID-19 lockdown period. ( a, c, e ) The spatial distribution of O changes for all-day average, daytime average, and nighttime average during the CLD period (January 2–22, 2020) compared with the pre-CLD period (January 23–February 12, 2020). The black boxes in ( a ) show the location of north China (NC, 184 sites), central China (CC, 108 sites), and south China (SC, 77 sites). ( b ) The variations of all-day average O mixing ratios from January 2 to February 12, 2020 for the NC, CC, and SC regions. ( d ) The same with ( b ) but for daytime average O . ( f ) The same with ( b ) but for nighttime average O . Fig. 2 Simulated changes in O mixing ratios across China during the COVID-19 lockdown period and contributions from meteorological changes and emission reductions. ( a, d, g ) The simulated total O changes for all-day average, daytime average, and nighttime average during the CLD period relative to the pre-CLD period. ( b, e, h ) Contribution of meteorological changes to O for all-day average, daytime average, and nighttime average. ( c, f, i ) The same with ( b, e, h ), respectively, but for contribution of emission reductions. Fig. 3 Changes in O mixing ratios during the COVID-19 lockdown period and contributions from meteorological changes and emission reductions for three typical regions. ( a ) Observed and simulated changes in O mixing ratios and the contributions from meteorological changes and emission reductions during the CLD period compared with the pre-CLD period in north China (NC). The O changes for the all-day average, daytime average, and nighttime average are presented. ( b ) The same with ( a ) but for central China (CC). ( c ) The same with ( a ) but for south China (SC). The locations of these three regions are shown in Fig. 1a. Note that the error bars mark the standard deviations within the region. Fig. 4 The estimated reductions of multi-pollutant emissions due the COVID-19 lockdown and their impacts on the O changes for three typical regions. ( a, c, e ) The estimated reductions of NO x , VOC, CO, PM, and SO emissions during the CLD period compared with the pre-CLD period for north China (NC), central China (CC), and south China (SC). The locations of these three regions are shown in Fig. 1a. ( b, d, f ) The impacts of different pollutant emission reductions due to the lockdown on O changes for the NC, CC, and SC regions. The O changes for the all-day average, daytime average, and nighttime average are presented. Note that the error bars mark the standard deviations within the region. Supplementary Information for “Diverse response of atmospheric ozone to COVID-19 lockdown in China”
Yiming Liu , Tao Wang Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hong Kong, 999077, China Correspondence to: Tao Wang ([email protected]) and Yiming Liu ([email protected]) This file includes Supplementary text, and Supplementary Figures S1-S10, Supplementary Table S1-S2 Supplementary text
Estimation of anthropogenic emissions in 2020
The most recent available data on China’s anthropogenic emissions are from 2017 (MEIC, , PM (particulate matter, including PM , PM , and its components), and SO in the power plant, industry, transportation, and residential sectors. The VOC emission was assumed to be unchanged from 2017 to 2020 because it increased by only 2% from 2013 to 2017 (Zheng et al., 2018). The NO , PM, and SO emissions from transportation were also assumed constant from 2017 to 2020, considering that they had changed little during 2015–2017. The NO x emission in the residential and industrial sectors was assumed to be the same in 2020 as in 2017 in view of its flat trend in recent years. Because the NO x emission from power plants decreased by ~47% from 2013 to 2017 (11.7% per year), we assumed it to further decrease by 35% from 2017 to 2020. The same approaches were applied to the reductions of the PM and SO emissions from 2017 to 2020. With these adjustments, we derived an estimated anthropogenic emission inventory for China in 2020. The model-simulated pollutant concentrations using this inventory showed a reasonable agreement with the surface measurement data (see the Model evaluation section below). Estimated reduction of anthropogenic emissions during the CLD period
We estimated the emission reductions during the COVID-19 lockdown period according to the recent literature (Wang et al., 2020; Huang et al., 2020). Transportation and industry were the sectors most affected by the COVID-19 lockdown in China. For the transportation sector, the decrease in national traffic volume was estimated at 70% during the lockdown according to transportation index data (Huang et al., 2020). The industry emissions were assumed to decrease by 40% across China in our study. We kept the emissions from power plants, residences, and agriculture unchanged because they were less affected by the city lockdowns (Wang et al., 2020; Huang et al., 2020). Fig. 4 presents the reductions of NO x , VOC, CO, PM, and SO emissions due to the 70% and 40% reductions of transportation and industry emissions during the lockdown, which generally match the estimation by Huang et al. (2020). Model evaluation
Statistical parameters were calculated to validate the model performance in simulating the air pollutant concentrations from January 2 to February 12, 2020, including the mean observation (OBS), mean simulation (SIM), mean bias (MB), mean absolute gross error (MAGE), root mean square error (RMSE), index of agreement (IOA), and correlation coefficient ( r ). The equations of these statistical parameters can be found in Fan et al. (2013). Table S2 shows the evaluation results for the simulated concentrations of SO , NO , CO, O and PM in China. Generally, the CMAQ model faithfully reproduced the observed pollutant concentrations with low biases. SO and PM were slightly overestimated with mean biases of 1.1 ppbv and 8.6 μg/m , respectively, which suggested that the emissions of SO and aerosols might have been more heavily reduced in 2020 than we estimated. CO was underestimated by the model with a mean bias of 0.33 ppmv. These biases of the simulated SO , PM , and CO concentrations have little impact on our results because the reductions of those pollutants exerted much smaller effects on O changes than did the reductions in NO x and VOCs. The model slightly underestimated the NO mixing ratios with a mean bias of -4.1 ppbv. This underprediction was partly attributed to the catalytic conversion method used to measure NO in the national network, which can overestimate NO (Zhang et al., 2017; Fu et al., 2019). The observed O mixing ratios were well captured by the model with a much smaller bias (2.6 ppbv). These evaluation results suggested reasonable estimations of the anthropogenic emissions for the year 2020 and during the lockdown. Fig. S10 shows the scatterplots of simulated and observed hourly O , SO , and PM concentrations. Most of the points fall within the range between 1:2 and 2:1 (the ratios between the simulated and observed concentrations). The general increase in O and decreases in SO and PM during the CLD period compared with the pre-CLD period were well captured by the model (see also Fig. S3 for the O changes across China), which indicated the faithful estimation of anthropogenic emission reductions during the lockdown. Fig. 3 shows that the model underestimated the O increase in the NC and CC regions and over-predicted the decrease in SC, which can be explained as follows. The simulated O mixing ratio and its increases were lower than the observed values in the NC region, possibly because of the overestimation of the NO titration effect due to the under-predicted planetary boundary layer height, especially during nighttime (Petersen et al., 2019). Moreover, the degree of emission reductions from transportation and industries might vary between regions with either positive or negative biases compared with our unified estimations across China, which would have affected the accuracy of the O simulations in these three regions in various ways. Nevertheless, the model was able to faithfully capture the observed O increases in NC and CC and decrease in SC. Overall, despite some uncertainties, the CMAQ model performance is acceptable and can support further analysis of O changes during the COVID-19 city lockdowns. Reference:
Fan, Q., Liu, Y. M., Wang, X. M., Fan, S. J., Chan, P. W., Lan, J., and Feng, Y. R.: Effect of different meteorological fields on the regional air quality modelling over Pearl River Delta, China, Int. J. Environ. Pollut., 53, 3-23, 10.1504/ijep.2013.058816, 2013. Fu, X., Wang, T., Zhang, L., Li, Q., Wang, Z., Xia, M., Yun, H., Wang, W., Yu, C., Yue, D., Zhou, Y., Zheng, J., and Han, R.: The significant contribution of HONO to secondary pollutants during a severe winter pollution event in southern China, Atmos. Chem. Phys., 19, 1-14, 10.5194/acp-19-1-2019, 2019. Huang, X., Ding, A., Gao, J., Zheng, B., Zhou, D., Qi, X., Tang, R., Wang, J., Ren, C., Nie, W., Chi, X., Xu, Z., Chen, L., Li, Y., Che, F., Pang, N., Wang, H., Tong, D., Qin, W., Cheng, W., Liu, W., Fu, Q., Liu, B., Chai, F., Davis, S. J., Zhang, Q., and He, K.: Enhanced secondary pollution offset reduction of primary emissions during COVID-19 lockdown in China, National Science Review, 10.1093/nsr/nwaa137, 2020. Petersen, A. K., Brasseur, G. P., Bouarar, I., Flemming, J., Gauss, M., Jiang, F., Kouznetsov, R., Kranenburg, R., Mijling, B., Peuch, V. H., Pommier, M., Segers, A., Sofiev, M., Timmermans, R., van der A, R., Walters, S., Xie, Y., Xu, J. M., and Zhou, G. Q.: Ensemble forecasts of air quality in eastern China - Part 2: Evaluation of the MarcoPolo-Panda prediction system, version 1, Geosci Model Dev, 12, 1241-1266, 10.5194/gmd-12-1241-2019, 2019. Wang, P., Chen, K., Zhu, S., Wang, P., and Zhang, H.: Severe air pollution events not avoided by reduced anthropogenic activities during COVID-19 outbreak, Resources, Conservation and Recycling, 158, 104814, https://doi.org/10.1016/j.resconrec.2020.104814, 2020. Zhang, L., Li, Q. Y., Wang, T., Ahmadov, R., Zhang, Q., Li, M., and Lv, M. Y.: Combined impacts of nitrous acid and nitryl chloride on lower-tropospheric ozone: new module development in WRF-Chem and application to China, Atmos Chem Phys, 17, 9733-9750, 10.5194/acp-17-9733-2017, 2017. Zheng, B., Tong, D., Li, M., Liu, F., Hong, C., Geng, G., Li, H., Li, X., Peng, L., Qi, J., Yan, L., Zhang, Y., Zhao, H., Zheng, Y., He, K., and Zhang, Q.: Trends in China's anthropogenic emissions since 2010 as the consequence of clean air actions, Atmos. Chem. Phys., 18, 14095-14111, 10.5194/acp-18-14095-2018, 2018. Figure S1: Location of 1643 environmental monitoring stations (red “+” symbols) in mainland China. The blue boxes denote the regions of north China, central China, and south China designated for further analysis. Figure S2: Percentage change of (a) NO , (b) CO, (c) PM , and (d) SO concentrations during the CLD period relative to the pre-CLD period. Figure S3: Observed (colored circles) and simulated (shaded color) O mixing ratios during the pre-CLD and CLD periods for the all-day average, daytime average, and nighttime average. Figure S4: Averaged sea-level pressure during the pre-CLD and CLD periods. Data are from the National Center for Environmental Prediction (NCEP) FNL Operational Model Global Tropospheric Analyses dataset. Figure S5: Changes in daytime temperature at 2 m height, specific humidity at 2 m height, wind field at 10 m height, planetary boundary layer (PBL) height, cloud fraction, and precipitation during CLD period relative to pre-CLD period. In panel (c), the shaded color and vector represent the wind speed and wind direction, respectively. Figure S6: Biogenic isoprene emissions during the pre-CLD and CLD periods and their difference (CLD minus pre-CLD). Figure S7: Changes in nighttime temperature at 2 m height, specific humidity at 2 m height, wind field at 10 m height, PBL height, cloud fraction, and precipitation during the CLD period relative to the pre-CLD period. In panel (c), the shaded color and vector represent the wind speed and wind direction, respectively. Figure S8: Proportions of NO x , VOCs, CO, PM, and SO Figure S9: Changes in O mixing ratios for the all-day average, daytime average, and nighttime average due to the reductions of NO x , VOC, CO, PM, and SO emissions during the CLD period compared with the pre-CLD period. Figure S10: Scatterplots of simulated and observed hourly (a) O , (b) NO , and (c) PM concentrations at environmental monitoring stations (~1500 sites) in China. The blue and red circles represent the observed mixing ratios or concentrations during the pre-CLD and CLD periods, respectively. Table S1: Scaling factors of different economic sectors to estimate the anthropogenic emissions of China for the year 2020 based on the 2017 MEIC emission inventory. Emitted species Power plant Industry Residence Transportation NO x -35% - - - SO -40% -40% -25% - PM - -20% -30% - Table S2: Evaluation results for the air pollutants across China from January 2 to February 12, 2020 (OBS is mean observation; SIM is mean simulation; MB is mean bias; MAGE is mean absolute gross error; RMSE is root mean square error; IOA is index of agreement; r is correlation coefficient; OBS, SIM, MB, MAGE, and RMSE have the same units as given in the first column, while IOA and r have no unit). Species OBS SIM MB MAGE RMSE IOA r SO (ppbv) 4.5 5.7 1.1 4.5 6.5 0.69 0.17 NO (ppbv) 14.2 10.1 -4.1 7.5 9.8 0.81 0.53 CO (ppmv) 0.85 0.52 -0.33 0.45 0.57 0.83 0.37 O (ppbv) 21.8 19.0 -2.8 10.5 13.2 0.90 0.57 PM (μg/m3