The impacts of COVID-19, meteorology, and emission control policies on PM2.5 drops in Northeast Asia

In January 2020, anthropogenic emissions in Northeast Asia reduced due to the COVID-19 outbreak. When outdoor activities of the public were limited, PM2.5 concentrations in China and South Korea between February and March 2020 reduced by − 16.8 μg/m3 and − 9.9 μg/m3 respectively, compared with the average over the previous three years. This study uses air quality modeling and observations over the past four years to separate the influence of reductions in anthropogenic emissions from meteorological changes and emission control policies on this PM2.5 concentration change. Here, we show that the impacts of anthropogenic pollution reduction on PM2.5 were found to be approximately − 16% in China and − 21% in South Korea, while those of meteorology and emission policies were − 7% and − 8% in China, and − 5% and − 4% in South Korea, respectively. These results show that the influence on PM2.5 concentration differs across time and region and according to meteorological conditions and emission control policies. Finally, the influence of reductions in anthropogenic emissions was greater than that of meteorological conditions and emission policies during COVID-19 period.

Since the first reported case of the novel coronavirus disease (COVID-19, a respiratory syndrome caused by SARS-CoV-2 infection) in Wuhan, China in December 2019 1,2 , COVID-19 has spread worldwide and as of June 21, 2020 the World Health Organization (WHO) reported that 8,708,008 people have been infected 3 . China, where the outbreak of COVID-19 was first reported, and South Korea, a neighboring country, declared national emergencies concerning the rapid spread of COVID-19 infections in mid-January 2020. Most economic activities were limited until restrictions were eased in China and South Korea in April and May 2020, respectively. The decrease in economic activities, including reduced traffic since January 2020, has led to the reduction of anthropogenic emissions caused by outdoor activities of the general public 4,5 . Over the course of the COVID-19 outbreak, reductions in air pollutant concentrations such as PM 2.5 (particles with a diameter of 2.5 μm or less), which posed serious threats with high environmental risks in Northeast Asia, have also been reported 6,7 .
In Northeast Asia, high concentration of PM 2.5 and related haze events frequently occurred in winter and early spring, mainly in China and South Korea where large amounts of air pollutant emissions are concentrated [8][9][10][11] . From December to March for the three years prior to the outbreak of COVID-19 (December 2016 to March 2017, December 2017 to March 2018, and December 2018 to March 2019, defined as the non-COVID-19 period), the PM 2.5 concentrations observed on the surface in China and South Korea were 62.2 μg/m 3 and 31.6 μg/m 3 , respectively. In contrast, the PM 2.5 concentrations observed from December 2019 to March 2020 (hereinafter defined as the COVID-19 period) were 50.0 μg/m 3 and 24.5 μg/m 3 respectively. They were reduced by 19.6% and 22.6% compared with the concentrations observed during the non-COVID-19 period.
When comparing the trend of PM 2.5 concentrations observed from December to March from 2016 to 2020 in China and South Korea (Fig. 1) We examined the changes in PM 2.5 concentration by into a first half (December and January) and second half (February and March), focusing on the period when the COVID-19 confirmed patients surged in China and South Korea (Fig. 1b,c). In the first half, prior to the COVID-19 outbreak, the concentration of PM 2.5 decreased by − 4.59 μg/m 3 /month in China and increased by + 1.17 μg/m 3 /month in South Korea. In the early stage of COVID-19 outbreak, the concentration even increased (+ 3.46 μg/m 3 /month) in China or showed a slight decrease (− 0.52 μg/m 3 /month) in South Korea. In the second half, when outdoor activities were limited, both China and South Korea showed a clearly declining trend of PM 2.5 concentration (China: − 7.60 μg/m 3 /month, South Korea: − 5.20 μg/m 3 /month). In the second half of the COVID-19 period, the regression slope for PM 2.5 concentration decreased by 5.0 µg/m 3 /month in China and by 6.6 µg/m 3 /month in South Korea compared with the slope for PM 2.5 concentration observed during the non-COVID-19 period.
The generation and dissipation of air pollutants undergo highly complex chemical/physical processes according to various atmospheric conditions such as meteorological conditions and emissions, and the final concentration of pollutants in the atmosphere is determined from these processes 12 . Therefore, it is difficult to identify that the reduction in the observed PM 2.5 concentrations during the COVID-19 period was directly influenced by the restrictions on anthropogenic activity. In response to the serious air pollution problems observed in the Northeast Asian region, governments in the region have been implementing policies to reduce emissions since early 2010, and as a result steady reductions in the concentrations of air pollutants, with priority given to primary pollutants, have been observed [13][14][15][16][17] . Moreover, owing to the recent climate change in Northeast Asia and changes in synoptic weather conditions, fluctuations in air pollutant concentrations have increased 18,19 . Because of this, it Alternatively, air quality modeling has been performed after assuming reduced emissions and the influences on the air pollutant concentration estimated 5,22,23 . Until now, there has been no study to quantitatively present the reduction of observed air pollutant concentrations influenced by COVID-19, separated from the influences of meteorological conditions and emission reduction policies. This study first aimed to distinguish the influences of meteorological conditions and emissions expected to be related to the PM 2.5 concentration reduction observed in the Northeast Asian region during the COVID-19 period. Secondly, the influences of emission reductions following emission control policies were separated and the results were presented quantitatively. To this end, meteorology-air quality modeling was performed for December to March for the past 4 years (2016-2020), including the COVID-19 period, and compared with the observed PM 2.5 concentrations. Figure 2 shows the normalized anomaly from PM 2.5 concentrations observed and simulated during the non-COVID-19 and COVID-19 periods in China and South Korea. Air quality modeling was performed by inputting  www.nature.com/scientificreports/ fixed anthropogenic emissions that do not account for annual changes in Northeast Asia. We used average PM 2.5 concentration during the non-COVID-19 period for each observation and simulation for the required average value for anomaly calculation. The observed PM 2.5 concentrations and the simulated PM 2.5 concentrations showed differences because of the uncertainty of the input data and the parametrization process in the model 24 , but the temporal change trends of the simulated and observed PM 2.5 concentration were similar, and the average prediction accuracy of the meteorology and air quality simulation results were reliable (R > 0.90 of air temperature at an altitude of 2 m and wind speed at an altitude of 10 m, R > 0.89 of PM 2.5 concentration simulated during the non-COVID-19 period, see Table S1 and Figs. S3 to 5 in Supplement for meteorology and air quality simulation results). For the non-COVID-19 period, the observed and simulated PM 2.5 showed a similar level of anomaly, and the trend of concentration change over time was also similar. However, for the COVID-19 period, the difference between the observed and the simulated anomaly increased (the average value of the difference between the non-COVID-19 period observation and the simulated anomaly: China 0.0% and South Korea 3.4%, the average value of the difference between the COVID-19 period observation and the simulated anomaly: China − 17.2%, and South Korea − 32.3%). In particular, the difference between observation-simulation anomaly for both China and South Korea were large, − 31.8% and − 57.1%, respectively in February. In other words, the observed concentration of PM 2.5 in the COVID-19 period decreased compared with that in the non-COVID-19 period in the Northeast Asia region, but the model predicted the PM 2.5 concentration to slightly decrease (China) or even increase (South Korea). The main reason for the large differences in observation-simulation anomaly during the COVID-19 period is the failure of the model to reflect reductions in anthropogenic emissions during the COVID-19 period. Figure 3 shows the trends in PM 2.5 concentration for the first half (December to January) and the second half . When comparing these to the non-COVID-19 period, the wind speed was 10.6% lower, and relative humidity 6.7% higher (see Fig. S6 of Supplement). During the first half of the COVID-19 period, the meteorological conditions (i.e., decreased surface wind speed and increased relative humidity) in the Northeast Asia region limited the dispersion of air pollutants, including PM 2.5 , leading to an increase in the air pollutant concentrations 14,25 .
In contrast, the emission control policies in China and South Korea contributed to a reduction of PM 2.5 concentration in both the first and the second half of the COVID-19 period. The influences of emission control policies on changes in PM 2.5 concentrations were calculated to be − 15% (− 10.5 μg/m 3 ) and − 8% (− 4.2 μg/m 3 ) in China, and − 9% (− 2.9 μg/m 3 ) and − 4% (− 1.3 μg/m 3 ) in South Korea in the first and second half, respectively. Compared with the first half, the reduction of PM 2.5 concentration due to policy in the second half fell to 40% in China and 44% in South Korea. Interestingly, in northern China, emission reduction policy contributed to a decrease of − 8% for PM 2.5 concentration in the first half but an increase of 2% in the second half. This is due to the trend of change in the simulated PM 2.5 concentration is larger than the observed PM 2.5 concentration with a relatively large prediction error (RMSE 26.3 μg/m 3 , R 0.78, see Table S1 in Supplement) during the second half of northern China compared with other regions.
Lastly, the influence of anthropogenic emission reductions due to COVID-19 on PM 2.5 concentrations calculated for the second half of COVID-19 were − 16% (− 9.0 μg/m 3 ) for China (− 29% in the northern region, − 16% in the central region, and − 12% in the southern region) and − 21% (− 7.0 μg/m 3 ) for South Korea (− 22% in Seoul Metropolitan Area (SMA)). In February to March 2020, when a distinct reduction was expected, these figures show that the influences of anthropogenic emissions reduction due to social distancing during the COVID-19 period in Northeast Asia were greater than that of meteorological conditions or emission control policies on the PM 2.5 concentration reduction.
Our study used the observed and simulated PM 2.5 concentrations from December to March for the past four years, including the COVID-19 period, to distinguish the influences of meteorological conditions and emissions on PM 2.5 concentrations in the Northeast Asian region, and to separate the influences of anthropogenic emission reduction due to COVID-19 from national emission control policies. Note that it is assumed that influences of meteorology and emissions on the PM 2.5 concentrations are independent to each other for the simple estimates in this study. Nevertheless, we believe that the method of separating the impacts of meteorology, emission control policy and COVID-19 presented in this study may be an effective way to evaluate the effectiveness of air quality regulations.
In Northeast Asia, especially China and South Korea, the severe PM 2.5 pollution frequently occurred in winter and early spring, which draws a high public interest in air quality. In these countries, various emission reduction policies are continuously implemented to reduce the national PM 2.5 levels. For example, the Chinese government actively implements the Air Pollution Prevention and Control Action Plan (2013-2017). In South CMAQ modeling for non-COVID-19 and COVID-19 periods was performed by entering fixed anthropogenic emissions with no annual change to determine the impact of meteorological conditions on PM 2.5 concentration changes during the COVID-19 period in the Northeast Asia region (Fig. 4). The difference between the simulated PM 2.5 concentrations in the non-COVID-19 period and the COVID-19 period ( C sim ) did not take into account the emission changes from 2016 to 2020 (e.g., changes in emissions due to emission control policies for each country, or emission changes from issues such as . Therefore, C sim can be seen as PM 2.5 concentration change according to the meteorological changes ( C met ).
In the next step, PM 2.5 change due to the total change in emissions during the COVID-19 period ( C emis ) was calculated by subtracting C met from the observed PM 2.5 concentration change ( C obs ). C emis includes not only the PM 2.5 concentration change due to emission control policy ( C policy ), but also anthropogenic emission reduction effects from social distancing during the COVID-19 period ( C COVID ). To distinguish the two effects, C policy was calculated from the difference between the observed PM 2.5 concentration change slope ( Slope O ) and the simulated PM 2.5 concentration change slope ( Slope M ) during the non-COVID-19 period. Slope M is the corrected slope calculated by multiplying the ratio of observed concentration and simulated concentration (α obs = C obs /C mod ) to the simulated PM 2.5 concentration change slope to take into account the difference between