key: cord-0932145-k01yr961
authors: Zhu, Jia; Chen, Lei; Liao, Hong; Yang, Hao; Yang, Yang; Yue, Xu
title: Enhanced PM(2.5) Decreases and O(3) Increases in China During COVID‐19 Lockdown by Aerosol‐Radiation Feedback
date: 2021-01-18
journal: Geophys Res Lett
DOI: 10.1029/2020gl090260
sha: c8cd52f5c13d841d75bed4a7689c9166cdb6bfe5
doc_id: 932145
cord_uid: k01yr961

We apply an online‐coupled meteorology‐chemistry model (WRF‐Chem) embedded with an improved process analysis to examine aerosol‐radiation feedback (ARF) impacts on effectiveness of emission control due to Coronavirus Disease 2019 (COVID‐19) lockdown over North China Plain. Emission reduction alone induces PM(2.5) decrease by 16.3 μg m(−3) and O(3) increase by 10.2 ppbv during COVID‐19 lockdown. The ARF enhances PM(2.5) decrease by 2.7 μg m(−3) (16.6%) and O(3) increase by 0.8 ppbv (7.8%). The ARF‐induced enhancement of PM(2.5) decline is mostly attributed to aerosol chemistry process, while enhancement of O(3) rise is ascribed to physical advection and vertical mixing processes. A set of sensitivity experiments with emission reductions in different degrees indicate that the ARF‐induced enhancements of PM(2.5) declines (O(3) rises) follow a robust linear relationship with the emission‐reduction‐induced PM(2.5) decreases. The fitted relationship has an important implication for assessing the effectiveness of emission abatement at any extent.

wind speed and PBL height (PBLH) facilitate more stable atmosphere and in turn increase surface-layer PM 2.5 levels Qiu et al., 2017) . The decreased PBLH and surface temperature can also increase surface relative humidity, and therefore, accelerate formation of surface particulates via heterogeneous reactions and hygroscopic growth, exacerbating haze pollution (Liu et al., 2018) .

The manifestations of ARF impacts on another important air pollutant, ozone (O 3 ), include changes in photolysis rates and atmospheric dynamics (Tian et al., 2019; W. Wang et al., 2019; Xing et al., 2017) . The reductions in solar radiation owing to aerosols result in lower photolysis rates and less O 3 generation (Tian et al., 2019; W. Wang et al., 2019; Zhu et al., 2019) . The changes in aerosol-induced atmospheric ventilation and rainfall may also influence O 3 concentrations. The changes in atmospheric dynamics due to aerosols lead to O 3 decreases in winter but increases in summer (Xing et al., 2017) .

Since aerosols influence meteorological conditions and further air quality through ARF, it is of great interest to explore how ARF affects the responses of air quality to emission control, which is a tendency to improve air quality. Emission reduction leads to changes in aerosols, which impacts meteorological conditions and further air quality (PM 2.5 and O 3 ). Limited studies reported how aerosol-radiation interactions impacted the effectiveness of emission abatement for PM 2.5 pollution (M. Gao et al., 2017; W. Wen, Guo, et al., 2020; Xing et al., 2015; Zhou et al., 2019) . The effects of ARF on O 3 responses to emission mitigation, however, are totally unknown. What's more, the prominent physical or chemical processes responsible for ARF impacts remain largely elusive.

Efforts to inhibit the spread of Coronavirus Disease 2019 (COVID-19), e.g., the implements of nationwide restrictions on population movement (lockdown), have remarkably lowered social-economic activities and reduced anthropogenic emissions over China during January-February 2020, which happens to provide an opportunity to investigate how ARF impacts PM 2.5 and O 3 responses to emission control. Recent studies have reported significant NO 2 reductions, moderate PM 2.5 decreases, and undesirable O 3 increases during COVID-19 outbreak (Huang et al., 2020; Shi & Brasseur, 2020; P. Wang et al., 2020; .

We examine the ARF effects on PM 2.5 and O 3 responses to emission reduction during COVID-19 lockdown, by using a fully online (two-way) coupled meteorology-chemistry model. An improved online integrated process rate (IPR) analysis scheme (i.e., process analysis) is developed in the model to explore how each physical/chemical process acts on the ARF impacts. This study focuses on PM 2.5 and O 3 air quality over North China Plain (NCP) of China from January 23 to February 29, 2020 when strict limitations on human activities to control COVID-19 spread were implemented. The study is believed to exert a novel contribution to understand the effectiveness of emission abatement.

The study region overlaid with meteorological and environmental monitoring sites is shown in Figure S1a . The analyzed region NCP in this study covers six provinces, including Beijing-Tianjin-Hebei-Shandong-Shanxi-Henan. Sources of meteorological and environmental measurements are described in Text S1. Figures S1b and S1c exhibit observed daily PM 2.5 and O 3 concentrations averaged over NCP before (January 1-22, 2020) and during (January 23-February 29, 2020) the COVID-19 lockdown period. The observed PM 2.5 concentration is 103.2 μg m −3 before COVID-19 lockdown and decreases to 69.8 μg m −3 (by 32.4%) during COVID-19 lockdown. The observed O 3 concentration, however, increases from 13.9 to 27.2 ppbv, almost doubling during COVID-19 lockdown.

We use a two-way coupled Weather Research and Forecasting with Chemistry model (WRF-Chem v3.7) to simulate meteorology, gas, and aerosol concentrations simultaneously (Grell et al., 2005) . The model configuration including natural emissions, and parameterization schemes are detailed in Text S2, and Table S1, respectively. Following P. Wang et al. (2020) , we use Multi-resolution Emission Inventory for China of 2016 (http://meicmodel.org/dataset-meic.html) as the basic anthropogenic emission inventory for the simulation period. It is noted that the emission change since 2016 will not affect the study significantly as the uncertainties of emission inventory usually surpass the emission changes over several year scales (Chen et al., 2014; Y. Zhao et al., 2011) . We estimate the emission reductions owing to COVID-19 lockdown following Huang et al. (2020) . The anthropogenic emissions of PM 2.5 and O 3 precursors summed over NCP are reduced by 8.3%-47.9% as a result of COVID-19 lockdown ( Figure S2 ). Among all precursors, nitrogen oxides (NO x ) emissions exhibit the most significant decrease of 47.9% since most transportation is prohibited during COVID-19 lockdown.

Four sensitivity simulations (ER_ARF, NoER_ARF, ER_NoARF, and NoER_NoARF) are conducted to examine how ARF affects the effectiveness of emission mitigation by conducting/no conducting emission reduction and with/without ARF during COVID-19 lockdown (Table S2 ). Experiment ER_ARF is designed to represent actual condition with emission reduction and ARF. The difference between ER_ARF and NoER_ARF shows the effects of emission reduction with ARF considered. The difference between ER_ NoARF and NoER_NoARF indicates the effects of emission mitigation alone which do not consider the feedbacks between aerosol and meteorology. Therefore, the difference between (ER_ARF-NoER_ARF) and (ER_NoARF-NoER_NoARF) reflects the impact of ARF on the effectiveness of emission reduction during COVID-19 lockdown, which is the aim of this study. To further test how the effects of ARF on the effectiveness of emission control vary with emission reduction degrees, we conduct another series of sensitivity experiments with emission reductions in different degrees (Table S2 ). The difference between (iER_ARF-NoER_ARF) and (iER_NoARF-NoER_NoARF) represents the impact of ARF on the effectiveness of emission reduction in different degrees. The comparisons between simulations (in Experiment ER_ARF) and measurements (meteorological and environmental) are shown in Text S3 and Figures S3-S6.

Integrated process rate (IPR) analysis, i.e., process analysis technique, is an advanced tool to quantitatively evaluate integrated rates of key processes simulated in the grid-based Eulerian models (J. Gao et al., 2018; Xing et al., 2017) . Chen et al. (2019) developed an improved IPR scheme in the WRF-Chem model to separate the processes influencing pollutant variations into different processes, i.e., emission source (EMIS), advection (TRAN), subgrid convection (SGCV), vertical mixing (VMIX), wet scavenging (WETP), gas-phase chemistry (GASC), aerosol chemistry (AERC), and cloud chemistry (CLDC). Traditionally, the IPR analysis is conducted over one time step (e.g., 60 min). Therefore, the contribution of each process to pollutant change is usually quantified compared with previous hour. Based on Chen et al. (2019) , we extend the use of IPR in this study to investigate the contribution of each physical/chemical process to pollutant change averaged over a period of time compared with another.

As shown in Figure S2 , the anthropogenic emissions over NCP exhibit significant reductions as a result of COVID-19 lockdown, which leads to decreases in aerosol concentrations ( Figure 2g ). The variations in aerosols perturb radiative balance and further change meteorological conditions. Figure 1 shows the changes in meteorological variables, including downward shortwave radiative flux at the surface (SW_SUR) and in the atmosphere (SW_ATM), 2m temperature (T 2 ), 2m relative humidity (RH 2 ), 10m wind speed (WS 10 ), and PBLH, due to emission control during COVID-19 lockdown. The changes are calculated by subtracting the model results of NoER_ARF from those of ER_ARF.

The SW_SUR exhibits overall increases when emission reduction is enforced, with the largest increases occurring in southern Hebei. Generally, the SW_SUR averaged over NCP is enhanced by 2.3 W m −2 during COVID-19 lockdown. X. Wen, Liu, et al. (2020) and Peters et al. (2020) provided observational evidences of increased solar radiation at the surface during COVID-19 outbreak over China and India, respectively. Contrary to the positive effect at the surface, the SW_ATM averaged over NCP decreases by 2.9 W m −2 as a result of emission control. It is well known that the existence of aerosol could reduce the SW_SUR but enhance SW_ATM due to aerosol scattering and absorption of solar radiation (Y. Gao et al., 2015; Qiu et al., 2017) . Therefore, the lower aerosol concentrations owing to emission mitigation lead to positive changes for SW_ SUR but negative changes for SW_ATM. Because the shortwave radiation reaching the ground is enhanced, near-surface temperature T 2 generally increases, with the average rise of 0.1 K over NCP and the maximum rise of 0.4 K in Shanxi. The RH 2 exhibits decreases over Beijing, Tianjin, Hebei, and Shanxi, resulting from the increase in saturation vapor pressure due to the increase in T 2 over these regions. The anomalous decreases in T 2 and increases in RH 2 occurring in Shandong result from the intensification of precipitation over the region (figure for precipitation is not shown). The warming due to increased SW_SUR and cooling due to decreased SW_ATM promote instability of atmosphere, which further accelerates near-surface wind speed and rises boundary layer. The WS 10 and PBLH averaged over NCP increase by 0.1 m/s and 10.5 m, respectively. The increased northwesterly is simulated over NCP, which may influence the transport of PM 2.5 and O 3 . ZHU ET AL. Figures 2a, 2b, 2e, and 2f show the spatial distributions of simulated PM 2.5 concentrations under four sensitivity simulations during COVID-19 lockdown. For each scenario, high PM 2.5 levels are all found over the analyzed region, with the largest concentrations in southern Hebei. Figures 2i and 2j exhibit the positive effects of ARF on PM 2.5 concentrations under two emission scenarios (shown by ER_ARF minus ER_NoARF and NoER_ARF minus NoER_NoARF, respectively). The radiative effects of aerosols lead to overall increases in PM 2.5 concentrations over NCP for each emission scenario. The aerosol-induced changes in meteorological variables (i.e., the cooling at the surface and the warming in the atmosphere, the more stable atmosphere, the increase in relative humidity at the surface, and the decrease in near-surface wind speed and PBLH) are beneficial for PM 2.5 production and accumulation in the lower atmosphere and therefore responsible for the significant increases of PM 2.5 levels. The positive feedback between aerosols and aerosol-induced meteorological conditions has been widely reported by previous studies (Huang et al., 2018; Liu et al., 2018; Su et al., 2018) .

We place emphasis, in this study, on the effects of emission mitigation on PM 2.5 air quality with and without ARF. With ARF considered, the emission reduction due to COVID-19 lockdown leads to overall declines of PM 2.5 concentrations (Figure 2c ). The averaged PM 2.5 level over the NCP decreases by 19.0 μg m −3 ; the largest PM 2.5 reduction exceeds 30.0 μg m −3 in Henan province. We further use the IPR analysis to examine the contribution of each physical/chemical process to PM 2.5 decrease. As shown in Figure 2d , the EMIS, AERC, and CLDC processes are responsible for the PM 2.5 decline. The primary emission of aerosol (EMIS) makes the largest contribution to the PM 2.5 decrease, followed by secondary transformation of aerosol (AERC and CLDC). On the contrary, the physical processes (e.g., VMIX and TRAN) exert an opposite effect on PM 2.5 changes. It's quite easy to understand the decreases induced by primary emission since strict emission control measures are implemented. The decreases in the chemical production of PM 2.5 (shown by AERC plus CLDC process) are mainly contributed by the declines in nitrate; instead, more sulfates are generated in response to the emission reduction ( Figure S7 ). Le et al. (2020) conducted a model sensitivity simulation and reported that NO x emission reduction would induce the reduction in nitrate aerosol but the increase in sulfate aerosol. The latter increase could be attributed to the promoted atmospheric oxidizing capacity. Huang et al. (2020) revealed that large decreases in NO x emissions during COVID-19 lockdown increased O 3 and nighttime NO 3 radical formation, and the intensification in atmospheric oxidizing capacity in turn promoted formation of secondary aerosol.

When ARF is not considered, the emission reduction alone also results in overall PM 2.5 declines over NCP ( Figure 2g ). However, the PM 2.5 decreases without ARF are much weaker than those with ARF ( Figure 2g vs. Figure 2c) . The PM 2.5 concentration averaged over the NCP (without ARF) decreases by 16.3 μg m −3 due to the COVID-19 restriction. The IPR analysis (Figure 2h ) also reveals that the PM 2.5 decline is contributed by EMIS, AERC, and CLDC processes.

The effect of ARF on the effectiveness of emission reduction for PM 2.5 air quality can be quantified by the difference between the emission-reduction-induced PM 2.5 changes with and without ARF (Figure 2k ). The consideration of ARF enhances PM 2.5 decreases all over the NCP. The largest enhancement of PM 2.5 decrease reaches 10.0 μg m −3 in southern Hebei. Generally, the ARF enhances the emission-reduction-induced PM 2.5 decline by 2.7 μg m −3 (16.6%) averaged over the NCP during COVID-19 lockdown. Further IPR analysis (Figure 2l ) suggests that the ARF-induced enhancement of PM 2.5 decline is attributed to AERC, TRAN, and CLDC processes. The AERC makes the largest contribution, indicating that fewer aerosols are generated through aerosol chemistry process, which leads to the enhancement of PM 2.5 decline with ARF considered. The increased northwesterly (Figure 1e ) brings low concentrations of PM 2.5 in northwestern China to NCP, accounting for the negative contribution from TRAN process.

Aerosol-induced changes in meteorological conditions can also influence surface-layer O 3 concentrations by altering physical and chemical process. Figures 3i and 3j present the negative effects of ARF on O 3 concentrations under two emission scenarios. The radiative effects of aerosols result in overall decreases in O 3 concentrations over NCP under any emission scenario. Compared to the surface O 3 concentration without aerosol feedback, the surface O 3 concentration with ARF averaged over NCP declines by 2.1 and 2.9 ppbv under two emission scenarios. Xing et al. (2017) also reported that aerosol-radiation effects reduced surface daily maxima 1 h O 3 over China by up to 39 μg m −3 through the combination of changes in photolysis rates and changes in atmospheric dynamics in January of 2013.

We then focus intensively on the effects of emission reduction on O 3 air quality with and without ARF. With ARF considered, the emission control due to COVID-19 lockdown leads to overall increases of O 3 concentrations (Figure 3c ). The averaged O 3 level over the NCP increases by 11.0 ppbv; the largest O 3 increase exceeds 16.0 ppbv in Hebei province. We further use the IPR analysis to quantify the contribution of each physical/chemical process to O 3 increase. As shown in Figure 3d , the GASC process accounts for the O 3 increase. On the contrary, the physical process (e.g., VMIX) exerts an opposite effect on O 3 change. During winter, the NCP is a VOC-limited region due to higher NO x and lower biogenic VOC emissions (He, Zhang, et al., 2017; Leung et al., 2020) . Under VOC-limited regime, NO x reductions can relax OH depletion by NO x and in turn produce more O 3 ; in addition, NO x reductions can also increase O 3 by alleviating NO x titration (Le et al., 2020; Leung et al., 2020) . The PM 2.5 decrease could also be a factor for O 3 increase via the aerosol-photolysis interaction (G. Li, Bei, et al., 2017; Wu et al., 2020) When ARF is excluded, the emission reduction alone also leads to overall O 3 increases over NCP (Figure 3g ). However, the O 3 increases without ARF are weaker than those with ARF ( Figure 3g vs. Figure 3c ). The O 3 concentration averaged over the NCP (without ARF) rises by 10.2 ppbv in response to the COVID-19 restriction. Further IPR analysis (Figure 3h ) also suggests that the GASC process contributes to the O 3 increase.

The impact of ARF on the effectiveness of emission reduction for O 3 air quality can be quantified by the difference between the emission-reduction-induced O 3 changes with and without ARF. The consideration of ARF enhances O 3 increases over most of NCP and weakens O 3 increases over a small fraction of the region (Figure 3k ). The largest enhancement of O 3 increase exceeds 2.0 ppbv in Shanxi. On average, the ARF enhances the emission-reduction-induced O 3 increase by 0.8 ppbv (7.8%) over the NCP region during COVID-19 lockdown. We conduct further IPR analysis in Figure 3l and find that the ARF-induced enhancement of O 3 increase is attributed to TRAN and VMIX processes. The increased northwesterly (Figure 1e ) brings high concentrations of O 3 in northwestern China to NCP, accounting for the positive contribution from TRAN process. With the development of PBL (Figure 1f ), more O 3 are transported downward from the upper atmosphere to the near surface, leading to the increases in surface-layer O 3 levels He, Gong, et al., 2017) contributed by VMIX process.

As described above, the emission reduction alone during COVID-19 lockdown induces PM 2.5 decrease by 16.3 μg m −3 averaged over NCP and the consideration of ARF enhances PM 2.5 decrease by 2.7 μg m −3 , leading to the net PM 2.5 decrease by 19.0 μg m −3 . That is, a model without ARF (i.e., offline model) will considerably underestimate emission-reduction-induced PM 2.5 decline by 16.6%. We further conduct a set of sensitivity experiments with emission reductions in different degrees to test the relationship between the PM 2.5 decrease induced by emission control alone and the enhancement of PM 2.5 decrease contributed by ARF. As presented by Figure 4a , the ARF-induced enhancement of PM 2.5 decrease shows a robust linear relationship (R 2 = 0.96, statistically significant at 99% confidence level) with the emission-reduction-induced PM 2.5 decrease without ARF. The stronger the abatement actions are, the greater the PM 2.5 improvement enhanced by ARF is.

As for O 3 air quality, the emission reduction alone during COVID-19 lockdown leads to a significant O 3 increase of 10.2 ppbv averaged over NCP and the consideration of ARF enhances O 3 increase by 0.8 ppbv, leading to the net O 3 increase by 11.0 ppbv. This indicates offline models without ARF will underestimate emission-control-induced O 3 rise by 7.8%. The sensitivity experiments with varying emission reductions indicate a statistically significant linear relationship (R 2 = 0.96) between the emission-reduction-induced PM 2.5 decrease without ARF and the enhancement of O 3 increase contributed by ARF (Figure 4b ).

ZHU ET AL. The fitted linear relationships have an important implication for assessing the effectiveness of emission abatement at any extent, and also provide offline models in the absence of ARF with an enforceable scheme to quantify the influence of ARF on the effectiveness of emission abatement and further estimate the net PM 2.5 or O 3 changes with ARF considered in response to emission reduction in any degree.

We use an online-coupled model WRF-Chem to examine ARF effects on PM 2.5 and O 3 responses to emission reduction during COVID-19 lockdown over NCP of China. The emission reduction alone induces PM 2.5 decrease by 16.3 μg m −3 ; the ARF enhances PM 2.5 decrease by 2.7 μg m −3 (16.6%), which is mainly attributed to aerosol chemistry process. For O 3 , the increase of 10.2 ppbv caused by emission reduction alone is enhanced by 0.8 ppbv (7.8%) through ARF, which is ascribed to physical advection and vertical mixing processes. Beyond this, we extend our result for the COVID-19 case study to consider a set of emission reduction scenarios, and find that the ARF-induced enhancement of PM 2.5 decline (O 3 rise) linearly responses to emission-reduction-induced PM 2.5 decrease. However, whether the fitted relationship for wintertime applies to summer condition remains unknown, which needs to be verified through further sensitivity experiments for summertime in future studies. In addition, the weakened ARF due to improved PM 2.5 air quality since China's clean air actions would contribute to worsening O 3 pollution; quantitatively evaluating the contributions from weakened ARF will have an important implication for understanding rising O 3 levels over China since 2013, which is another following topic of great interest.

Unit-based emission inventory and uncertainty assessment of coal-fired power plants

Assessing the formation and evolution mechanisms of severe haze pollution in the Beijing-Tianjin-Hebei region using process analysis

Meteorological influences on PM 2.5 and O 3 trends and associated health burden since China's clean air actions

Distinguishing the roles of meteorology, emission control measures, regional transport, and co-benefits of reduced aerosol feedbacks in

Modeling the feedback between aerosol and meteorological variables in the atmospheric boundary layer during a severe fog-haze event over the North China Plain

Effects of black carbon and boundary layer interaction on surface ozone in Nanjing

Fully coupled "online" chemistry within the WRF model

Air pollution characteristics and their relation to meteorological conditions during 2014-2015 in major Chinese cities

Multi-year application of WRF-CAM5 over East Asia-Part I: Comprehensive evaluation and formation regimes of O 3 and PM 2.5

Enhanced secondary pollution offset reduction of primary emissions during COVID-19 lockdown in China

Impact of aerosol-PBL interaction on haze pollution: Multiyear observational evidences in North China

Wintertime particulate matter decrease buffered by unfavorable chemical processes despite emissions reductions in China

Unexpected air pollution with marked emission reductions during the COVID-19 outbreak in China

Widespread and persistent ozone pollution in eastern China during the non-winter season of 2015: Observations and source attributions

Aerosol and boundary-layer interactions and impact on air quality

Aerosol radiative effects and feedbacks on boundary layer meteorology and PM 2.5 chemical components during winter haze events over the Beijing-Tianjin-Hebei region

The distributions and direct radiative effects of marine aerosols over East Asia in springtime

Anthropogenic drivers of 2013-2017 trends in summer surface ozone in China

A two-pollutant strategy for improving ozone and particulate air quality in China

Aerosols and their impact on radiation, clouds, precipitation, and severe weather events

New positive feedback mechanism between boundary layer meteorology and secondary aerosol formation during severe haze events

Impacts of interactive dust and its direct radiative forcing on interannual variations of temperature and precipitation in winter over East Asia

Black carbon amplifies haze over the North China Plain by weakening the East Asian winter monsoon

Enhanced air pollution via aerosol-boundary layer feedback in China

The impact of COVID-19 related measures on the solar resource in areas with high levels of air pollution

Simulated impacts of direct radiative effects of scattering and absorbing aerosols on surface layer aerosol concentrations in China during a heavily polluted event in

The response in air quality to the reduction of Chinese economic activities during the COVID-19 outbreak

Relationships between the planetary boundary layer height and surface pollutants derived from lidar observations over China: Regional pattern and in fluencing factors

Aerosol radiative effects on tropospheric photochemistry with GEOS-Chem simulations

Severe air pollution events not avoided by reduced anthropogenic activities during COVID-19 outbreak. Resources, Conservation and Recycling

The impact of aerosols on photolysis frequencies and ozone production in Beijing during the 4-year period 2012-2015

Impact of emission reduction on aerosol-radiation interaction during heavy pollution periods over Beijing-Tianjin-Hebei region in China

Relationship between the COVID-19 outbreak and temperature, humidity, and solar radiation across China

Aerosol-photolysis interaction reduces particulate matter during wintertime haze events

Air pollution and climate response to aerosol direct radiative effects: A modeling study of decadal trends across the northern hemisphere

Impacts of aerosol direct effects on tropospheric ozone through changes in atmospheric dynamics and photolysis rates

Dust-wind interactions can intensify aerosol pollution over eastern China

Enhanced PM2.5 pollution in China due to aerosol-cloud interactions

Quantifying the uncertainties of a bottom-up emission inventory of anthropogenic atmospheric pollutants in China

Substantial changes in nitrogen dioxide and ozone after excluding meteorological impacts during the COVID-19 outbreak in mainland

The impact of aerosol-radiation interactions on the effectiveness of emission control measures

Correlations between PM 2.5 and ozone over China and associated underlying reasons

References From the Supporting Information

Modeling impacts of urbanization and urban heat island mitigation on boundary layer meteorology and air quality in Beijing under different weather conditions

China's emission control strategies have suppressed unfavorable influences of climate on wintertime PM 2.5 concentrations in Beijing since 2002

Modeling sea-salt aerosols in the atmosphere: 1. Model development

Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from Nature)

Climate change from 1850 to 2005 simulated in CESM1 (WACCM)

A modeling study of the peroxyacetyl nitrate (PAN) during a wintertime haze event in Beijing

Simplification of a dust emission scheme and comparison with data

Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires

Simulating aerosol-radiation-cloud feedbacks on meteorology and air quality over eastern China under severe haze conditions in winter

Meteorological measurements from NOAA's National Climatic Data Center are publicly available at https:// www.ncei.noaa.gov/data/global-hourly/access/2020/. Observed PM 2.5 and O 3 concentrations from China National Environmental Monitoring Centre can be obtained from https://met.iap.ac.cn/data/openaq/CN/. Multi-resolution Emission Inventory for China can be accessed publicly from http://meicmodel.org/dataset-meic.html, and the emission reduction ratios due to COVID-19 lockdown are available from Table S1 at https://academic.oup.com/nsr/advance-article/doi/10.1093/nsr/nwaa137/5859289. Model results are available at https://zenodo.org/record/4059188#.X3Q5EWgzabh. The authors declare no conflict of interest.

This work is supported by the National Key R&D Program of China (2019YFA0606804), the National Natural Science Foundation of China (42007195), and the University Natural Science Research Foundation of Jiangsu Province (18KJB170012).