key: cord-0703170-k6rxv8xm
authors: Pei, Zhipeng; Han, Ge; Ma, Xin; Su, Hang; Gong, Wei
title: Response of major air pollutants to COVID-19 lockdowns in China
date: 2020-07-11
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2020.140879
sha: c5a7f0db2f9edb762f26b8050ca0807da8780eac
doc_id: 703170
cord_uid: k6rxv8xm

Abstract COVID-19 suddenly struck Wuhan at the end of 2019 and soon spread to the whole country and the rest of world in 2020. To mitigate the pandemic, China authority has taken unprecedentedly strict measures across the country. That provides a precious window to study how the air quality response to quick decline of anthropogenic emissions in terms of national scale, which would be critical basis to make atmospheric governance policies in the future. In this work, we utilized observations from both remote sensing and in-situ measurements to investigate impacts of COVID-19 lockdown on different air pollutions in different regions of China. It is witnessed that the PM2.5 concentrations exhibited distinct trends in different regions, despite of plunges of NO2 concentrations over the whole country. The steady HCHO concentration in urban area provides sufficient fuels for generations of tropospheric O3, leading to high concentrations of O3, especially when there is not enough NO to consume O3 via the titration effect. Moreover, the SO2 concentration kept steady at a low level regardless of cities. As a conclusion, the COVID-19 lockdown indeed helped reduce NO2 concentration. However, the atmospheric quality in urban areas of China has not improved overall due to lockdown measures. It underscores the significance of comprehensive control of atmospheric pollutants in cleaning air. Reducing VOCs (volatile organic compounds) concentrations in urban areas would be a critical mission for better air quality in the future.

J o u r n a l P r e -p r o o f

Abstract：COVID-19 sudden struck Wuhan at the end of 2019 and soon spread to the whole country and the rest of world in 2020. To mitigate the pandemic, China authority has taken unprecedentedly strict measures across the country. That provides a precious window to study how the air quality response to quick decline of anthropogenic emissions in terms of national scale, which would be critical basis to make atmospheric governance policies in the future. In this work, we utilized observations from both remote sensing and in-situ measurements to investigate impacts of COVID-19 lockdown on different air pollutions in different regions of China. It is witnessed that the PM 2.5 concentrations exhibited distinct trends in different regions, despite of plunges of NO 2 concentrations over the whole country. The steady HCHO concentration in urban area provides sufficient fuels for generations of tropospheric O 3 , leading to high concentrations of O 3 , especially when there is not enough NO to consume O 3 via the titration effect. Moreover, the SO 2 concentration kept steady at a low level regardless of cities. As a conclusion, the COVID-19 lockdown indeed helped reduce NO 2 concentration. However, the atmospheric quality in urban areas of China has not improved overall due to lockdown measures. It underscores the significance of comprehensive control of atmospheric pollutants in cleaning air. Reducing VOCs (volatile organic compounds) concentrations in urban areas would be a critical mission for better air quality in the future.

In response to the COVID-19 (corona virus disease 2019) outbreak, the China central government has gradually implemented severe nationwide lockdown measures since the end of January 2020. These measures have put the society of China on hold, significantly reducing emissions of pollutants. The resultant dramatic changes in the atmospheric environments have been of great concern to scholars Kerimray et al., 2020; Kraemer et al., 2020) .

In the past decade, China has experienced several social pauses including routine ones such as the J o u r n a l P r e -p r o o f Lunar New Year and special events such as the Olympic game and G20. Whether spontaneous or obliged, these activities tend to lead to a weakening of production activities, providing special opportunities for atmospheric researches (Chan et al., 2015) . For example, Wang et al. (Wang et al., 2010) simulated daily mean concentrations of PM 2.5 and found that more than 60% of the PM 2.5 reductions during the Olympics were due to emissions controls. Yan et al. (Yan et al., 2012) found that near-surface SO 2 concentrations in the PRD region during the Asian Games were lower than in the same period in previous years using OMI data. Huang et al. (Huang et al., 2015) found that the tropospheric NO 2 concentration in North China Plain during APEC in 2014 was 47% lower than that in the same period of the previous 3 years; Li et al. (Li et al., 2016) found that the total PM 1 mass was reduced by 52 -57% during the 2015 China Victory Day Parade (V-day Parade) by excluding the effects of meteorological parameters. Compared with those events, the COVID-19 lockdown exerts much more evident influences on the society in terms of both persistence and strength. In order to ensure a good atmospheric environment during some major events, governments often required some factories to shut down and regulated traffic, but such initiatives are limited to a relatively small region, such as a city or a province, and usually last for from a few days to within half a month low level for over a month all across the country (Tian et al., 2020) .

The Chinese Lunar New Year holiday can lead to a weakening of the intensity of productive activity across the country . However, this social pause generally lasts only a week or less, while still allowing a significant portion of the factory to continue production and traffic tends not to decrease significantly, just to shift from large cities to small cities or rural areas (Guo et al., 2018) .

Very differently, the COVID-19 lockdown is believed to make both productive activities and traffic to a low level for over a month all across the country (Tian et al., 2020) . As the epicenter of J o u r n a l P r e -p r o o f COVID-19, Wuhan didn't completely reopen until May 2020. Thus, COVID-19 was an all-human catastrophe, but the social pause it caused provided a valuable and unprecedented opportunity to study the baseline of the atmospheric environment and the effects of human activities on it from the scale of the whole country. Now, several studies have reported different responses of the air quality to the lockdown in different regions (Dantas et al., 2020; Mahato et al., 2020; Otmani et al., 2020; Sharma et al., 2020; Wang et al., 2020a) . Those contradictory results underscored the importance of further studies into this area.

In this study, we focus on five representative indexes of air pollutions, namely, NO 2 , SO 2 , O 3 , HCHO, and PM 2.5 . The main aim of this work is to investigate how air pollutants response to the social pause caused by COVID-19 and to gain insight into the mechanism of generations of PM 2.5 . NO 2 is seen as a direct index for describing the intensity of human activities because there are hardly natural emissions of NO 2 (Lin and McElroy, 2011; Zhang et al., 2018) . Actually, soon after the COVID-19 lockdown started, researchers observed a rapid decrease of atmospheric concentration of NO 2 in both China and Italy (Bauwens et al., 2020) . SO 2 is a common index to describe industry emissions. Besides, during the pandemic, a rumor claimed that abnormal surge of SO 2 concentrations suggested extra death polls were concealed circulated on social media. The formation of tropospheric ozone is relatively complex, influenced on the one hand by solar radiation and on the other by the concentration of other atmospheric pollutants .

In recent years, there has been a general decrease in PM 2.5 in China's major cities Li et al., 2019; Zhang et al., 2019) , but a certain increase in tropospheric O 3 (Chan et al., 2017a; Chan et al., 2017b; Wang et al., 2020b) . PM 2.5 has a very large impact on visibility and is therefore more likely to be of public concern. However, O 3 concentrations are difficult to draw J o u r n a l P r e -p r o o f attention of public, but are extremely harmful and therefore a new hotspot for atmospheric pollution research. HCHO is an important proxy for volatile organic compounds (VOC), which are of more and more concern in recent years (Chan et al., 2019) .

Of those observations, NO 2 and HCHO concentrations were obtained by remote sensing to provide spatial distributions over the whole country. Other indices were obtained by in-situ measurements to provide more accurate and robust evidences. NO 2 is the only index which is monitored by both means in this work. Three representative cities located in different latitudes were selected as the study areas, namely, Beijing, Wuhan and Guangzhou. Beijing locates in the Beijing-Tianjin-Hebei region which is notorious for its bad air quality and large emissions of greenhouse gases (Qiu et al., 2020a; Shi et al., 2020b) . Wuhan is the first epicenter of COVID-19 and is located in central China. Guangzhou is located in the Pearl River Delta region which is the engine of China's economics in decades Shi et al., 2020a) . Those three cities are all centers of their regions and are separated from each other, enabling us to study possible differences on responses of atmospheric environments to the COVID-19 lockdown.

The remaining parts of this work are arranged as follow. The data and method we used are described in section 2. Main results are demonstrated in Section 3. Discussions on possible effect of meteorological factors and implications to future policy development regarding atmospheric governance are provided in Section 4. Finally, we conclude the whole study in Section 5.

Three metropolises located in different latitude zones in China are selected as the study area of this work, namely Beijing, Wuhan and Guangzhou. They locate in northern, central, and southern China respectively. Because they are not geographically adjacent, they have negligible J o u r n a l P r e -p r o o f atmospheric transportations to each other. Besides, the climatic characteristics of these metropolitan areas are quite different from each other. Beijing, Wuhan and Guangzhou have a temperate monsoonal climate, a subtropical monsoon humid climate and subtropical monsoonal climate, respectively (Han et al., 2018) . Based on the above considerations, these cities provide good conditions for studying the differences in responses of air qualities to COVID-19 lockdown in different regions. Locations of these three metropolises are demonstrated in Figure 1 .

Beijing is the core of the Beijing-Tianjin-Hebei region with highly developed heavy industry (Han et al., 2017) . Although the government has relocated heavy industry and highly polluting enterprises such as Shougang Group out of Beijing, most of these factories have been relocated in Hebei, around Beijing, the city's atmospheric environment thus remains very vulnerable to emissions from these heavy industry enterprises.

Wuhan is the first epicenter of COVID-19 around the world and is also the most important city in the central China in terms of economy, education, innovation and industry. Before the COVID-19 pandemic, Wuhan was famous for its multiple lakes and rivers in the city. The water-rich land cover feature leads to higher humidity and relatively high atmospheric stability in Wuhan. Once heavy pollution occurs, the rate at which pollution subsides is relatively slow unless heavy rainfall occurs (Han et al., 2014) .

Guangzhou is the capital of Guangdong province in southern China. Administratively, the city holds sub-provincial status and is located in the Pearl River Delta region. Guangdong's GDP has been ranked first in China for the past 30 years, and if Guangdong were a country, its GDP would be 13 th in the world. The developed manufacturing sector has also challenged the air quality in Guangzhou to a certain extent. 

The coronavirus first break out in Wuhan in the end of 2019 and was confirmed to be human-to-human transmission on 20 th January. Since then, several provincial administrative units in China have declared the highest level of public health emergency response within a week. To stop the spread of the coronavirus, China had cut off Wuhan from other regions since 23 rd , January, 2020. Guangzhou authority declared the lockdown on 23 rd , and one day later Beijing and Wuhan joined the lockdown. With the epidemic under control, Beijing and Guangzhou gradually lifted lockdown measures from 10 th , February, 2020. However, the level of recovery of industrial and manufacturing activity in these two regions remained very limited until March. As the epicenter of COVID-19, Wuhan had not lifted lockdown measures until 20 th , March, 2020 while production activities and traffic didn't largely resume until late April.

Agency to monitor air pollution. The onboard sensor is frequently referred to as TROPOMI (TROPOspheric Monitoring Instrument) . Compared to the OMI launched in 2004, which is still operating in orbit, TROPOMI's technical characteristics have improved significantly. OMI's nadir resolution is 13 km (along track) by 24 km (cross track), and TROPOMI's spatial resolution is significantly improved with a 1-5 times higher signal-to-noise ratio. For example, OMI NO 2 products have a resolution of 26 km by 48 km while TROPOMI NO 2 products have a resolution of 7 km by 3.5 km. Moreover, the retrieval-assimilation-modeling system for TROPOMI NO 2 products uses the 3-dimensional global TM5-MP chemistry transport model at a resolution of 1 x 1 degree as an essential element.

TROPOMI products can be downloaded for free on ESA's Sentinels Scientific Data Hub website (https://scihub.copernicus.eu). In this study, we obtained level 3 data from Google Earth Engine. Two kinds of products from TROPOMI, namely NO 2 and HCHO, were obtained to illustrate temporal-spatial distributions of those pollutants. In this work, we used tropospheric NO 2 and HCHO concentration products, namely, tropospheric_NO 2 _column_number_density and tropospheric_HCHO_column_number_density datasets, because they can better reflect anthropogenic emissions. The conversion to L3 is done by the harpconvert tool using the bin_spatial operation. The source data is filtered to remove pixels with QA values less than: 75% for the tropospheric_NO 2 _column_number_density and 50% for the tropospheric_HCHO_column_number_density.

Along with remotely sensed data, in-situ observations of atmospheric pollutants were also (Han et al., 2014) .

Daily measurements from sites were then converted to a daily average of a certain pollutant in terms of cities. Then, we calculated weekly average concentrations for different atmospheric pollutants.

This paper adopts a comparative approach to analyze the impact of the COVID-19 lockdown on the atmospheric environment in three regionally representative cities. Firstly, it is crucial to study the trends of the three pollutants using time series. Since the epidemic coincides with the Spring Festival, the two effects appear to overlap. Therefore, we have selected the data for the same period of 2019 in terms of the Chinese lunar calendar for comparison. It is worth noting that we used the lunar New Year to align two years of time series data. There is another advantage by doing so. Differences between meteorological factors can be partially reduced because the Chinese lunar calendar itself is highly correlated with meteorological characteristics.

J o u r n a l P r e -p r o o f

To map the spatial distribution of NO 2 concentrations, we utilized TROPOMI productions as described in section 2. Since there are evident blanks in daily map of NO 2 , we used the original daily NO 2 to produce the weekly average NO 2 and demonstrated them in Figure 2 . We have divided the whole period into 5 sub-periods according to the rule shown in Table 1 . Because 5 th , February, 2019 and 25 th , January, 2020 are the Lunar New Year for these two year, respectively, we aligned P3 for these two year first and then arranged other four sub-periods accordingly. As mentioned in section 2, lockdown measures came into effect in three cities on the 23 rd and 24 th January 2020. We can regard P3 of 2020 as the first period when anthropogenic emissions plunged.

Although, Guangzhou and Beijing authorities have nominally been gradually lifting the lockdown measure since 10 th , February, 2020, there was little sign of factory reopen and traffic resumption in the initial week, according to reports from state-own media and social media. Therefore, P3, P4

and P5 in 2020 can be all seen as the lockdown period for all three cities while the other two sub-periods are normal days. As the control group, only P3 can be seen as a social pause period owing to the Spring Festival holiday in 2019. Similarly, Figure 3 shows temporal-spatial distributions of HCHO during P1 to P5 in 2019 

The response of the PM 2.5 concentration to COVID-19 lockdown in Wuhan is different from that in Beijing. The middle panel of Figure 4 demonstrates that the PM 2.5 concentration was evidently higher in P1 of 2019 than that in P1 of 2020. In P2, the PM 2.5 concentrations for two years are quite similar. After the lockdown began, no evident decreasing trend of PM 2.5 was witnessed during P3 and P4. Compared with the mean PM 2.5 concentration during the same period in the last year, there are no significant reductions or increments either. As residents in Wuhan, we witnessed that people were still allowed to use their private vehicles in the first week after lockdown although the public transportation in Wuhan, along with cutting off connection with other cities, were suspended to mitigate spread of the coronavirus. Panic buying of goods and surging need for medical care due to suspected symptoms of pneumonia lead to increasing usages of private vehicles, partially offsetting the reduction in emissions resulting from the suspension of public transport. That could be one of reasons why the PM 2.5 concentration during P3 and P4 of 2020 did not exhibit a huge reduction compared with that in P2 of 2020, P3 and P4 of 2019.

The mean PM 2.5 concentration in Guangzhou is much lower than those in Beijing and Wuhan as shown in Figure 3 . Only a very slight decline of PM 2.5 was observed due to COVID-19 lockdown in Guangzhou in terms of both year-over-year and month-over-month. Therefore, we can conclude that there is no significant impact of COVID-19 lockdown on the PM 2.5 concentration in Guangzhou. there is no guarantee that a complete observation product for a given area will be available within a week. Therefore, it is difficult to compare the data with ground-based observations over the same period one by one. However, by observing the general trends, it can be seen that the satellite products and ground-based observations can be well matched. Except for in P1 of Beijing and in J o u r n a l P r e -p r o o f P5 of Guangzhou, NO 2 concentrations over 3 cities exhibited a decline trend after COVID-19

lockdown. Interestingly, Figure 3 visually shows a large decline in NO 2 . Although ground stations in the three cities also observed a decline in the NO 2 concentration in response, the magnitude of the decline appears to be less dramatic than that shown in Figure 3 . When we look carefully back to Figure 3 , we can find that NO 2 concentrations decreased much less in major cities than in non-urban areas of the North China Plain. In the past decade, the SO 2 concentration has been significantly reduced over China. Figure 7 shows SO 

Meteorological factors, especially wind speed and precipitation, have an important impact on the atmospheric environment. In analyzing the lockdown's impact on the atmospheric environment, we chose the same time period on the lunar calendar, which has already somewhat eliminated the impact of meteorological factors. In this subsection, we further statistically compare wind speed and precipitation over the experimental period to analyze the potential contribution of differences in meteorological factors in two years to changes in the atmospheric environment. In addition, for the completeness of the weather data, we also counted the temperature. The weekly average wind speed decreased year-on-year in Beijing during the whole experimental period. We noticed that the weekly average wind speed dropped to 0.74 m/s in P3 of 2020. Meanwhile, there is no precipitation in that week. The temperature also kept steady.

Meteorological data indicated that Beijing was in typically calm weather conditions during P3 of 2020. The poor diffusion condition suggested that high concentrations of PM2.5 and O3 in P3 of 2020, namely the first week of the COVID-19 lockdown, are unlikely results of external input from surrounding provinces but are highly possible attributed to generations of secondary aerosols.

During P4 and P5, the wind speed in 2020 was just slightly lower than that in 2019 but the precipitation increased significantly year-on-year. Therefore, the meteorological factors are not the main reason for a higher PM 2.5 concentration in P4 and P5 of 2020, compared with that in 2019.

In Wuhan, the wind speed and the precipitation during P3 to P5 in 2020 is slightly lower than that in 2019 while the temperature is higher in 2020. We think the meteorological condition is similar in two years and is thus not the main reason for differences of atmospheric environments between two years. In Guangzhou, the meteorological conditions of two years during P3 and P4

are also similar. It is worth noting that there was an evident high wind and rainfall process in P5 of 2019. That explains why O 3 increased significantly in P5 year-on-year.

As a conclusion, meteorological factors indeed determined the difference in atmospheric quality between the two years in a few cases. However, the overall weather conditions were relatively similar in both years. Thus, differences in atmospheric quality should be attributed primarily to the COVID-19 lockdown.

The responses of the atmospheric environments to COVID-19 lockdown demonstrated significant differences among 3 cities. For Beijing, the most notable feature is that the concentration of PM 2.5 has increased rather than decreased compared to last year during P3 to P5 while concentrations of NO 2 and SO 2 declined at the same time. Figure 4 to 6 indicate that Beijing week-on-week, NO 2 and SO 2 concentrations decreased. Referring to Figure 3 , the concentration of HCHO, a critical proxy of VOCs which is, along with NO 2 , the precursor to tropospheric O 3 , declined from the high level of the last week but was still higher than that in P4 of 2019. 

As general a conclusion, we found the NO 2 concentrations decreased in all 3 cities. Besides, Figure 2 also demonstrates that NO 2 concentrations dropped significantly all over the country and exhibited no evident rebound during the lockdown period. Such a conclusion has been witnessed in many relevant researches in China, Italy and other countries.

However, as a very important index for the air quality, the PM 2.5 concentration exhibited a relatively more complicated response to the COVID-19 lockdown. A high PM 2.5 concentration can be visually perceived by people, and is thus of more public concern than other atmospheric pollutants. Secondary aerosols are the main component of PM 2.5 in China. Unlike NO 2 , the generation of secondary aerosols involves complex atmospheric chemical processes, including both emissions of precursors and meteorological conditions.

There is an effect of COVID-19 lockdown on SO 2 concentration in Beijing. However, such an effect is negligible in Wuhan and Guangzhou. The atmospheric SO 2 concentration in most cities of China has kept declining in the past decade and is now at a low level. Therefore, we think the interaction of atmospheric SO 2 with other pollutants is very limited. An outbreak of SO 2 emissions could lead to evident degradation of atmospheric environments as shown in the Beijing case. However, decline of SO 2 seems to exert little influences on atmospheric environments.

In recent years, O 3 and VOCs become emerging factors influencing air quality after SO 2 and 

Impact of Coronavirus Outbreak on NO2 Pollution Assessed Using TROPOMI and OMI Observations

Observations of tropospheric NO2 using ground based MAX-DOAS and OMIn measurements during the Shanghai World Expo

On the summertime air quality and related photochemical processes in the megacity Shanghai

MAX-DOAS measurements of tropospheric NO2 and HCHO in Nanjing and a comparison to ozone monitoring instrument observations

Observations of tropospheric aerosols and NO 2 in Hong Kong over 5 years using ground based MAX-DOAS

Air pollution reduction and mortality benefit during the COVID-19 outbreak in China. The Lancet Planetary Health

The impact of COVID-19 partial lockdown on the air quality of the city of Rio de Janeiro, Brazil

Significant reduction of PM2.5 in eastern China due to regional-scale emission control: evidence from SORPES in

Data mining based framework for exploring household electricity consumption patterns: A case study in China context

Study on spatial and temporal distributions of contaminants emitted by Chinese New Year'Eve celebrations in Wuhan

Performance Evaluation for China's Planned CO2-IPDA

Feasibility Study on Measuring Atmospheric CO2 in Urban Areas Using Spaceborne CO2-IPDA LIDAR

APEC Blue" phenomenon: Regional emission control effects observed from space

Assessing air quality changes in large cities during COVID-19 lockdowns: The impacts of traffic-free urban conditions in Almaty

The effect of human mobility and control measures on the COVID-19 epidemic in China

Rapid transition in winter aerosol composition in Beijing from 2014 to 2017: response to clean air actions

Parade Blue": effects of short-term emission control on aerosol chemistry

Detection from space of a reduction in anthropogenic emissions of nitrogen oxides during the Chinese economic downturn

Abrupt declines in tropospheric nitrogen dioxide over China after the outbreak of COVID-19

Effect of lockdown amid COVID-19 pandemic on air quality of the megacity Delhi

Impact of Covid-19 lockdown on PM10, SO2 and NO2 concentrations in Salé City (Morocco)

CO2 Concentration, A Critical Factor Influencing the Relationship between Solar-induced Chlorophyll Fluorescence and Gross Primary Productivity

A comparison of oco-2 sif, modis gpp, and gosif data from gross primary production (GPP) estimation and seasonal cycles in North America

Effect of restricted emissions during COVID-19 on air quality in India

An inversion method for estimating strong point carbon dioxide emissions using a differential absorption Lidar

Measurement of CO2 rectifier effect during summer and winter using ground-based differential absorption LiDAR

An investigation of transmission control measures during the first 50 days of the COVID-19 epidemic in China

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

Quantifying the Air Pollutants Emission Reduction during the

Ozone pollution in China: A review of concentrations, meteorological influences, chemical precursors, and effects

Contrasting trends of PM2.5 and surface-ozone concentrations in China from

SO_2 long-term monitoring by satellite in the Pearl River Delta

Drivers of improved PM2.5 air quality in China from

Comparing OMI-based and EPA AQS in situ NO2 trends: towards understanding surface NOx emission changes