key: cord-0986200-ww9tqvdx
authors: Wu, Shugang; Zhou, Weijian; Xiong, Xiaohu; Burr, G.S.; Cheng, Peng; Wang, Peng; Niu, Zhenchuan; Hou, Yaoyao
title: The impact of COVID-19 lockdown on atmospheric CO(2) in Xi’an, China
date: 2021-04-22
journal: Environ Res
DOI: 10.1016/j.envres.2021.111208
sha: 2d158c434fe6872376d0758f57188277471579ca
doc_id: 986200
cord_uid: ww9tqvdx

Lockdown measures to control the spread of the novel coronavirus disease (COVID-19) sharply limited energy consumption and carbon emissions. The lockdown effect on carbon emissions has been studied by many researchers using statistical approaches. However, the lockdown effect on atmospheric carbon dioxide (CO(2)) on an urban scale remains unclear. Here we present CO(2) concentration and carbon isotopic (δ(13)C) measurements to assess the impact of COVID-19 control measures on atmospheric CO(2) in Xi’an, China. We find that CO(2) concentrations during the lockdown period were 7.5% lower than during the normal period (prior to the Spring Festival, 25 Jan to 4 Feb 2020). The observed CO(2)(excess) (total CO(2) minus background CO(2)) during the lockdown period was 52.3% lower than that during the normal period, and 35.7% lower than the estimated CO(2)(excess) with the effect of weather removed. A Keeling plot shows that in contrast CO(2) concentrations and δ(13)C were weakly correlated (R(2)=0.18) during the lockdown period, reflecting a change in CO(2) sources imposed by the curtailment of traffic and industrial emissions. Our study also show that the sharp reduction in atmospheric CO(2) during lockdown were short-lived, and returned to normal levels within months after lockdown measures were lifted.

pollutants. Energy consumption (including fossil fuels) has increased sharply in recent years as a result of urbanization and modern lifestyle changes (Hosseini et al., 2019) .

Fossil fuels are responsible for 85% of CO 2 emissions and 64% of total greenhouse gas emissions (Razzaq et al., 2020) . As a consequence of increasing energy consumption, CO 2 -dominated greenhouse gases have also increased in recent years (Adeniyi et al., 2019) , and on a global scale are attributed to the following sectors: electricity and heat generation (44%), transportation (26%), and industry (19%), according to the International Energy Agency (IEA, 2020) . The consumption of fossil fuels causes ecological and environmental problems (Al-Juboori et al., 2020a) , such as climate warming and urban smog (Sher et al., 2020) . CO 2 emissions must be curtailed to mitigate global warming and enhance sustainable development (Lelieveld et al., 2019) . This strategy hinges on alternative energy technologies, which include: solar and wind power (Qazi et al., 2019) ; sustainable hydrocarbon fuel production from CO 2 (Al-Juboori et al., 2020b); hydrogen fuel cells (Al-Shara et al, 2019) ; and biofuels (Razzaq et al., 2020) . Additionally, carbon capture and sequestration techniques such as post-combustion CO 2 capture using fast adsorbents derived from biomass (Sher et al., 2020) can reduce CO 2 emissions. The transition from fossil fuels to renewable energy will play an essential role in CO 2 emissions reduction, but it is not happening fast enough (Gielen et al., 2019) . The COVID-19 pandemic provided an opportunity to test how fast carbon emissions can be reduced by sharply curtailing emissions.

Control measures to check the spread of COVID-19 led to a reduction in road and J o u r n a l P r e -p r o o f air traffic, a temporary closure of businesses, and a decrease in industrial productivity (Ding et al., 2020) . A number of factors that contributed to CO 2 emissions reductions (Wang et al., 2020b) include: 1) reduced power demands due to the delayed resumption of work after the Spring Festival; 2) a reduced demand for steel and blast furnace operation time; 3) a reduced energy demand from a variety of business enterprises; and 4) COVID-19 restrictions required people to stay at home. This latter factor markedly reduced road and air traffic, with immediate reductions in transportation-related CO 2 emissions. Several studies have performed statistical analyses to estimate national and global reductions in carbon emissions that resulted from pandemic prevention measures. For example, compared with data from the same time periods in 2019, daily global CO 2 emissions decreased by 17.0% by early April 2020 (Le Quéré et al., 2020) and decreased by 8.8% in the first half of 2020 (Liu et al., 2020a) . In China, carbon emissions fell by 11.0% over the first quarter of 2019 (Han et al., 2020) . There are also a number of observational studies that shows declines of atmospheric pollutants. Bauwens et al. (2020) found that NO 2 concentrations decreased rapidly following the COVID-19 lockdown both in China and Italy. Xu et al. (2020) reported that submicron aerosol mass concentrations were reduced by 50% during the COVID-19 lockdown in Lanzhou, China. Surface measurements made at more than 800 monitoring stations show that mean levels of PM 2.5 and NO 2 in northern China decreased by approximately 35% and 60%, respectively, after the COVID-19 lockdown (Shi and Brasseur, 2020) . Sharma et al. (2020) reported that a PM 2.5 concentration decrease of 43% in India during the COVID-19 lockdown J o u r n a l P r e -p r o o f period compared to the previous 4 years.

Cities play an important role in the effort to reduce carbon emissions (Xu et al., 2021) , as they account for about 70% of global carbon emissions (Churkina, 2016) . Xi'an is the largest city in northwestern China and all of its residential complexes were locked down from 5 February to 21 February 2020, due to COVID-19 control measures. Here we study the urban-scale effect of the lockdown in Xi'an on atmospheric CO 2 concentrations and stable carbon isotope compositions (δ 13 C) in the first quarter of 2020. The objective of this study is to detect the averaged CO 2 concentration change during the 2020 COVID-19 lockdown period relative to 2019 levels, using a correction for meteorological effects. Quantifying the impact of the COVID-19 lockdown on atmospheric CO 2 concentrations on a city scale is important to future carbon emission measures for sustainable development. It can also provide useful information for modeling studies.

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

Xi'an is currently the capital of Shaanxi Province and in historical times, was the capital of China during thirteen dynasties. Its population reached 10 million in 2018.

Xi'an is located in the south-central part of the Guanzhong Basin, bordered by the loess plateau to the north and the Qinling Mountains to the south. This basinal configuration, and the mild winds typical of Xi'an most of the year, inhibit the removal of air pollutants (Yang, 2003) . The observations made for this study (Fig 1) were carried out on the main building of the Institute of Earth Environment, Chinese Academy of Sciences (IEECAS) in southeast Xi'an from 1 Jan to 31 March 2019 and 2020. Fig.1 The location of the study site (IEECAS) in Xi'an

We divided the first quarter of 2020 into five stages according to the different measures taken in response to COVID-19. Stage 1 (1 January to 24 January 2020) represents a normal period before the Spring Festival holiday (the Chinese Lunar New

Year). Stage 2 (25 January to 4 February) is during the Spring Festival holiday of 2020. Spring Festival is the largest holiday in China. In normal years, people travel to visit their relatives and friends from the first day of the Chinese New Year. But in 2020, millions of people were asked to stay at home in an effort to stop the spread of the new coronavirus. Stage 3 (5 February to 21 February) was the lockdown period, with the strictest control measures enforced. Only one person per family was allowed to venture out to purchase daily necessities such as food and medicine, once every two days. The reopening of industries and schools in Xi'an was also delayed throughout this period. Transportation was largely restricted with few vehicles on the road during that time. Stage 4 (22 February to 28 February) was a transition period as the lockdown measures were relaxed. Businesses began to reopen and restrictions on residents began to be lifted. During stage 5 (29 February to 31 March) normal patterns were reestablished, approaching Stage 1 conditions. However, schools and cinemas did not reopen immediately, and group tours that crossed provincial borders were still restricted.

Atmospheric CO 2 concentration and its δ 13 C were measured using a Picarro G2131-i carbon isotopic analyzer (Picarro, Inc., USA). The Picarro analyzer measures CO 2 concentration, δ 13 C in CO 2 , CH 4 and H 2 O. The precision for CO 2 is better than J o u r n a l P r e -p r o o f 0.2 ppm and for δ 13 C is better than 0.2‰. Air samples were pumped directly into the Picarro analyzer at a flow rate of 25 ml/min. The CO 2 concentration was derived from the sum of dry air concentrations of 12 CO 2 and 13 CO 2 . The instrument was calibrated by two standard gases (cylinder 1 with CO 2 395.49±0.02 ppm, δ 13 C in CO 2 -8.980±0.008‰, CH 4 1993.5±0.2 ppb, and cylinder 2 with CO 2 491.43±0.02 ppm, δ 13 C in CO 2 -10.395±0.024‰, CH 4 3029.3±0.5 ppb) obtained from the Chinese Academy of Meteorological Sciences. Each standard gas is pressurized in a 29.5-L treated aluminum alloy cylinder (Scott-Marrin, Inc., California) fitted with a high-purity, two-stage gas regulator, and calibrated with cylinders assigned by the WMO/GAW CO 2 Central Calibration Laboratory operated by NOAA/ESRL.

To study the atmospheric CO 2 response to the lockdown, we took January to March, 2019, as a reference period. However, 6 days during the period (27 January to 1 February 2019) were missing due to instrument failure. A study in Shanghai, China found that atmospheric CO 2 and CO correlate well with each other (Wei et al., 2020) .

Daily averaged CO data for Xi'an were obtained from the Chinese Air Quality online Monitoring and Analysis Platform (CAQMAP, 2020), and we found that the daily averaged CO 2 and CO have a highly significant (p<0.001) linear relationship (Fig 2) .

Based on this, we reconstructed the daily average CO 2 concentrations to fill the gap from the 6 missing days in our record.

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

Although the comparison of the five stages in the first quarter of 2020, and the month-to-month comparison between 2020 and 2019 show an apparent influence of COVID-19 measures, these results include both natural carbon cycle variability and meteorological conditions. Quantifying and attributing changes in CO 2 concentrations requires accounting for meteorological effects in addition to direct emissions (Turner et al., 2020) . We did this by determining the difference between CO 2 excess-est and CO 2 excess-obs , as explained next. We adopted the method of Venter et al. (2020) , that predicts air pollution proxies (PM 2.5 , O 3 , and NO 2 ) from meteorological parameters, to estimate first quarter 2020 CO 2 concentrations (CO 2 est ). We first established the relationship between daily CO 2 concentration and weather parameters (temperature, relative humidity, wind speed concentration) is defined as the lockdown effect. In fact, the CO 2 excess includes fossil fuel CO 2 (CO 2 ff ) and biogenic CO 2 (CO 2 bio ) (Levin et al., 2003) . Former studies in

Xi'an showed that in winter months, the CO 2 excess comes predominately from fossil fuel emissions (Wang et al., 2018) , which can account for more than 90% (Zhou et al., 2020) . Thus the CO 2 excess in our study mainly reflects fossil fuel CO 2 variations.

Fig 4 shows that the CO 2 excess-obs is significantly (p=0.013) lower than the CO 2 excess-est by 35.7% (or 11.7 ppm) during the lockdown period, which is close to the reduction of fossil fuel emissions in China, by 32.0±12% in February 2020 (Tohjima et al., 2020) . However, during the normal period before the Spring Festival the CO 2 excess-obs is significantly (p=0.005) higher than the CO 2 excess-est by 24.0% (or 13.8 ppm).

In the other three stages there are no obvious differences between the observed and estimated values. This result indicates a clear COVID-19 lockdown effect on CO 2 .

The CO 2 excess-est during the lockdown period is 20.4% lower than that during the normal period before the Spring Festival. Note that the amplitude of this decline is significantly smaller than the uncorrected CO 2 excess-obs value, which declined 52.3%.

In the study of Liu et al. (2020b), a higher CO 2 concentration was observed during the lockdown period in 2020 than the same period in 2019. This higher CO 2 should be attributed to the weather conditions, rather than COVID-19 control measures since they observed a significant decline of on-road CO 2 concentration for the same period.

In the transition period the CO 2 excess-obs concentration was higher (p=0.738) than CO 2 J o u r n a l P r e -p r o o f pre-lockdown conditions. 

Stable isotopes (δ 13 C) in atmospheric CO 2 provide a valuable means to distinguish between different CO 2 sources in air because different sources can have very different J o u r n a l P r e -p r o o f δ 13 C values. In order to investigate whether CO 2 sources changed significantly during the lockdown period, the Keeling-plot method (Keeling, 1958) was used to determine δ 13 C values for each time period. The observed CO 2 can be divided into background CO 2 and source CO 2 . According to the mass balance of CO 2 and its stable carbon isotopes, we can write the following (Keeling, 1958):

CO 2 obs = CO 2 bg +CO 2 excess (2) δ 13 C obs ·CO 2 obs =δ 13 C bg ·CO 2 bg +δ 13 C excess ·CO 2 excess

Combining equations (2) and (3) we can obtain: δ 13 C obs =CO 2 bg /CO 2 obs ·(δ 13 C bg -δ 13 C excess ) +δ 13 C excess ) +δ 13 C excess (4) Fig.5 Keeling-plots morning rush hour in each stage of the first quarter of 2020.

By plotting δ 13 C obs and 1/CO 2 obs , the mean isotopic signature of the CO 2 excess can be obtained as the y intercept of the Keeling-plot curve. We applied this method to the morning rush hour for each period we divided in 2020 to study the effect of lockdown on vehicle emissions (Fig 5) .

The results show that the δ 13 C excess values for the normal period before Spring (Sturm et al., 2006) , and energy usage patterns (Górka and Lewicka-Szczebak, 2013) due to the adjustment of the epidemic prevention policies from strict to loose during the lockdown period. Another possible reason for the low R 2 might result from the low CO 2 range (Zobitz et al., 2006) , which is only 41.6 ppm in the lockdown period. The δ 13 C excess values in the other four periods are close to those observed in Wroclaw, Poland, which are -25.7‰ and -27.3‰, during two heating seasons (Górka and Lewicka-Szczebak, 2013) , indicating fossil fuel combustion CO 2 sources during the winter season.

The impact of the COVID-19 lockdown on atmospheric CO 2 concentrations in Xi'an was assessed using ground observations corrected for the influence of weather.

The results show that during the lockdown period, observed CO 2 concentrations were J o u r n a l P r e -p r o o f 7.5% lower than normal (before the Spring Festival). Daily CO 2 sources changed during the lockdown, as reflected by the low correlation (R 2 value) observed using the Keeling-plot method applied before, during, and after the lockdown period. Although the impact of the lockdown on atmospheric CO 2 concentration in Xi'an was large, its impact was short-lived. Following the relaxation of the pandemic prevention measures, CO 2 concentrations increased again to similar levels as observed in 2019.

This study quantifies to some extent the rate and magnitude of changes that can occur by sharply curtailing anthropogenic CO 2 emissions in an urban environment. In practice, we expect that such reductions can be achieved through the implementation of green technologies. Our monitoring approach can be used in the future to assess the efficacy of such technologies.

The role of urbanization in the global carbon cycle

NO x emissions reduction and rebound in China due to the COVID-19 crisis

The role of renewable energy in the global energy transformation

One-year spatial and temporal monitoring of concentration and carbon isotopic composition of atmospheric CO 2 in a Wroclaw (SW Poland) city area

Assessing the recent impact of COVID-19 on carbon emissions from China using domestic economic data. Science of the Total NOAA

Towards real-time verification of CO 2 emissions

Investigation of CO 2 variation and mapping through wearable sensing techniques for measuring pedestrians' exposure in urban areas

Towards sustainable energy: a systematic review of renewable energy sources, technologies, and public opinions

Modeling viscosity and density of ethanol-diesel-biodiesel ternary blends for sustainable environment

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

Development of biomass derived highly porous fast adsorbents for post-combustion CO 2 capture

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

Measurements of CO 2 , its stable isotopes, O 2 /N 2 , and 222 Rn at Bern, Switzerland. Atmospheric Chemistry and Physics

Detection of fossil-fuel CO 2 plummet in China due to COVID-19 by observation at Hateruma

Observed Impacts of COVID-19 on Urban CO 2 Emissions

COVID-19 lockdowns cause global air pollution declines

Emission characteristics of atmospheric carbon dioxide in Xi'an, China based on the measurements of CO 2 concentration, Δ 14 C and δ 13 C

Temporal characteristics of greenhouse gases (CO 2 and CH 4 ) in the megacity Shanghai, China: Association with air pollutants and meteorological conditions

COVID-19 impact on the concentration and composition of submicron particulate matter in a typical city of Northwest China

Investigating the impacts of three-dimensional spatial structures on CO 2 emissions at the urban scale

Analysis of Pollution Meteorological Conditions in Xi'an

Fossil fuel CO 2 traced by radiocarbon in fifteen Chinese cities

Sensitivity analysis and quantification of uncertainty for isotopic mixing relationships in carbon cycle research

This work was jointly supported by National Natural Science Foundation of China (41730108), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA23010302, XDB 40000000), Natural Science Basic Research