key: cord-296618-aw4zm23z
authors: Huang, Guanyu; Sun, Kang
title: Non-negligible impacts of clean air regulations on the reduction of tropospheric NO2 over East China during the COVID-19 pandemic observed by OMI and TROPOMI
date: 2020-07-21
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2020.141023
sha: 
doc_id: 296618
cord_uid: aw4zm23z

Abstract We study the variation of tropospheric NO2 vertical column densities (TropNO2VCDs) over East China during the 2005–2020 lunar new year (LNY) holiday seasons to understand factors on the reduction of tropospheric NO2 during the outbreak of COVID-19 in East China using Ozone Monitoring Instrument (OMI) and TROPOspheric Monitoring Instrument (TROPOMI) observations. TropNO2VCDs from OMI and TROPOMI reveal sharp reductions of 33%–72% during 2020 LNY holiday season and the co-occurring outbreak of COVID-19 relative to the climatological mean of 2005–2019 LNY holiday seasons, and 22%–67% relative to the 2019 LNY holiday season. These reductions of TropNO2VCD occur majorly over highly polluted metropolitan areas with condensed industrial and transportation emission sources. COVID-19 control measures including lockdowns and shelter-in-place regulations are the primary reason for these tropospheric NO2 reductions over most areas of East China in 2020 LNY holiday season relative to the 2019 LNY holiday season, as COVID-19 control measures may explain ~87%–90% of tropospheric NO2 reduction in Wuhan as well as ~62%–89% in Beijing, Yangtze River Delta (YRD) and Pearl River Delta (PRD) areas. The clean air regulation of China also contributes significantly to reductions of tropospheric NO2 simultaneously and is the primary factor in the Sichuan Basin area, by explaining ~56%–62% of the tropospheric NO2 reduction there.

The COVID-19 disease caused by SARS-CoV-2, a novel coronavirus, was first reported in Wuhan, capital of Hubei Province in China, in December 2019 . This disease had spread quickly in Wuhan city and surrounding areas with 258 confirmed cases on January 21, 2020 (World Health Organization, 2020). By then, confirmed cases were also found in Japan, South Korea, Thailand and eventually worldwide, respectively (World Health Organization, 2020). The Chinese government had announced a strict lockdown and shelter-in-place law in Wuhan City on January 23, 2020 Tian et al., 2020) . These regulations required people to stay at home without using cars, suspended all public transportations and closed nonessential businesses and manufacturing factories. A day later, other cities in Hubei province also conducted similar lockdown and shelter-in-place regulations (Tian et al., 2020) . In the following week, the government of China had announced mandatory "shelter-in-place" orders. These COVID-19 control measures and consequently minimized human activities in China and worldwide are expected to result in substantially reduced NO x (NO x = NO+NO 2 ) emissions due to the dominancy of anthropogenic NO x sources (Bauwens et al., 2020; Collivignarelli et al., J o u r n a l P r e -p r o o f TROPspheric Monitoring Instrument (TROPOMI) with fine resolutions but only available in 2019-2020 LNY holiday seasons.

The LNY day is set according to the lunar calendar and varies from late January to early February on Julian calendars. Figure 1 shows the distribution of LNY days and the official national holidays during 2005-2020. Millions of people in mainland China return to their hometowns from working places in the beginning of the holiday season and travel back to work sites towards the end of the holiday season (Li et al., 2016) . The LNY holiday season with a large migration usually lasts approximately 40 days, 14 days before the LNY's day and 25 days after, as shown in Figure 1 . The industrial and transportation-related emissions decrease before the LNY's day, stay at a relatively low level during the LNY holiday, and then gradually increase back to normal after the holiday. This consequent impacts on air pollution levels is known as the "holiday effect" that shifts with LNY days (Feng et al., 2016; Huang et al., 2012; Ji et al., 2018; Tan et al., 2009; Yao et al., 2019) . As shown in Figure 1 , the variation of LNY dates on the Julian calendar is blurred since we averaged over a 40-day period. The natural emissions of NO x is considerably small during LNY periods (January to early March) because of low J o u r n a l P r e -p r o o f 6 2.2 OMI and TROPOMI observations Level 2 OMI tropospheric NO 2 product (OMNO2 version SPv3) (Krotkov et al., 2019) and TROPOMI Tropospheric NO 2 product (version 1.3.0) ( Copernicus Sentinel-5P, 2018) are used in this study.

OMI is a push room UV-VIS spectrometer that measures the Earth's backscattered sunlight since October 2004 aboard the Aura satellite (Levelt et al., 2006) . The Aura satellite is on a sunsynchronous polar orbit with an equator crossing time of approximately 1345 local time (Levelt et al., 2006) . The swath width of OMI is 2600 km, enabling global daily coverage with a resolution of 13km ×24 km (along × across-track) at nadir. The OMI level 2 NO 2 standard product is retrieved through a variation of differential optical absorption spectroscopy algorithm (Bucsela et al., 2013; Krotkov et al., 2017) . OMI has been proved with considerably stable performance since the launch in 2004, although OMI's global coverage and data sampling have been affected by the "row anomaly" that was first found in 2007 (Krotkov et al., 2017) . The impacts of row anomaly on data sampling are shown in Figure S1 . As a result, only OMI tropospheric NO 2 observations of rows 5-23 on clear sky scenes (with cloud coverage less than 0.3) are used, excluding those affected by the row anomaly, to avoid inconsistent sampling of data (Duncan et al., 2016; Krotkov et al., 2017; Krotkov et al., 2019) .

TROPOMI is the single payload aboard the Sentinel 5 Precursor (S5P) satellite that has a sunsynchronous orbit with local overpass time of approximately 1330 with a near-daily global coverage since April 2018 (Veefkind et al., 2012) . The TROPOMI NO 2 retrieval algorithm is J o u r n a l P r e -p r o o f 7 × 3.5 km with a change in the S5P operation scenario since August 6, 2019 (orbit 9388) (Eskes and Eichmann, 2019) . Similar to OMI data, we only use TROPOMI observations with cloud coverage less than 0.3 (Eskes and Eichmann, 2019) . Although TROPOMI recommended pixels with quality assurance (QA) values greater than 0.75, we use TROPOMI pixels with QA values greater than 0.5, which includes good quality retrievals over clouds and over snow/ice scenes and is "useful for assimilation and model and model comparison studies" (Eskes and Eichmann, 2019) , to add more observations to this study in wintertime. In addition, since pixels with QA greater than 0.75 removes cloud scenes with cloud coverage greater than 0.5 (Eskes and Eichmann, 2019), we use a stricter threshold (<0.3) for cloud coverage to remove pixels that may be influenced by clouds and maintain sampling consistency. Furthermore, the difference is small between selecting 0.5 and 0.75 for QA value thresholds with both cloud threshold (<0.3) applied in this study as shown in Figure S2 .

Consequently, the combination of OMI and TROPOMI provides us a long temporal coverage of TropNO2VCD observations with a moderate resolution in 2005-2020 LNY holiday seasons and a fine spatial resolution in 2019-2020 LNY holiday seasons.

It is critical to ensure data consistency due to the use of long-term data from two spaceborne data sources. We produce level 3 data over LNY holiday seasons by averaging level 2 data to a common, regular 0.125°×0.125° grid and apply corrections to these regridded level 3 tropospheric NO 2 data. We use the physical oversampling method by Sun et al. (2018) to average temporally and spatially level 2 satellite observations to level 3 grid. Two-dimensional super Journal Pre-proof J o u r n a l P r e -p r o o f 8 Gaussian functions are used to represent the spatial response functions of satellite sensors. In comparison with conventional approaches that only consider the pixel corners, this method gives considerable advantages of visualizing the distribution and local gradients of trace gases and for scenarios with short temporal windows (Sun et al., 2018) . We directly regrid OMI level tropospheric NO 2 data to the 0.125°×0.125° grid, while we regrid TROPOMI level 2 data to 0.01°×0.01° first and then aggregate them to 0.125°×0.125° through area-weighted averaging (Sun et al., 2018; Zhu et al., 2017) . This is because the physical oversampling method assumes that the sensitivity of a satellite observation is represented by a continuous spatial response function, which is feasible when the grid size is not significantly larger than the satellite footprint.

The overpass times of OMI and TROPOMI are 1345 and 1330 local time, respectively (Levelt et al., 2006; Veefkind et al., 2012) . It is reasonable to assume that OMI and TROPOMI have similar observations on tropospheric NO 2 , since the lifetime of NO 2 in the troposphere is much longer than 15 minutes (Shah et al., 2020) . OMI TropNO 2 VCDs during 2019 and 2020 LNY holiday seasons (Figure 2 including spectral fitting algorithms, atmospheric mass factor (AMF) calculation, and spatial sampling discrepancies due to different pixel size and OMI's row anomaly (OMI and TROPOMI spatial sampling on selected days in 2020 are shown in Figure S3 ) (Cheng et al., 2019; Griffin et al., 2019; Krotkov et al., 2017) . As a result, a correction should be applied to OMI and TROPOMI level 3 data to reconcile the discrepancies.

Since we focus on spatially regridded and temporally averaged TropNO 2 VCDs over LNY holiday seasons in China instead of individual retrievals, we apply a simple correction to OMI and TROPOMI level 3 TropNO 2 VCDs. We obtain mean differences between OMI and TROPOMI level 3 data during 2019 and 2020 LNY holiday seasons as:

where i, j are latitude and longitude, respectively. OMI 2019 , OMI 2020 , TROPOMI 2019 and TROPOMI 2020 represent level 3 TropNO 2 VCDs measured by OMI or TROPOMI during 2019 and 2020 LNY holiday seasons, respectively. The correction term c(i,j) accounts for the potential sampling and algorithm biases between the OMI and TROPOMI NO 2 products, which are assumed to be steady over all LNY holiday seasons. Therefore, we apply the same corrections to both OMI and TROPOMI level 3 data as follows: corrected OMI TropNO 2 VCDs are defined as Wuhan had conducted the strictest lockdown and shelter-in-place regulations in the world (Tian et al., 2020) . In the Beijing area, ~11%-32% and ~68%-89% of the tropospheric NO 2 reduction is caused by the clean air regulations and COVID-19control measures, as Beijing city has conducted considerably strict control measures (Beijing Municipal Health Commission, 2020).

~77%-86% and ~62% of the NO 2 reduction in YRD and PRD is caused by the COVID-19

control measures, as cities of YRD (e.g., Shanghai City) and PRD (e.g., Guangzhou City) had conducted similar control measures as Beijing city (Health Commission of Guangdong Province, 2020; Shanghai People's Government, 2020). In the Sichuan Basin area, however, only 38%-44% of the reduction is caused by the COVID-19 measures with the remaining 56%-62% explained by the clean air regulations, due to less strict control measures (Health Commission of Sichuan Province, 2020). Overall, the ratio of NO 2 reduction caused by clean air regulations and COVID-19 control measures is associated with how strict the local COVID-19 control measures were conducted.

We studied the variation of TropNO 2 VCDs in East China during the 2005-2020 LNY holiday seasons and the outbreak of COVID-19 by using OMI and TROPOMI tropospheric NO 2 observations. We have converted level 2 data of OMI and TROPOMI to 0.125°×0.125° grid level 3 data by using the physical oversampling method (Sun et al., 2018) and applied corrections to OMI and TROPOMI NO 2 level 3 data to obtain a consistent dataset for this study.

The corrections were obtained through the mean difference between OMI and TROPOMI tropospheric NO 2 in the 2019 and 2020 LNY holiday seasons. 

Impact of coronavirus outbreak on NO 2 pollution assessed using TROPOMI and OMI observations

Beijing City COVID-19 Prevention Information

A new stratospheric and tropospheric NO 2 retrieval algorithm for nadir-viewing satellite instruments: applications to OMI

Tropospheric NO 2 vertical column densities retrieved from ground-based MAX-DOAS measurements at Shangdianzi regional atmospheric background station in China

Lockdown for CoViD-2019 in Milan: What are the effects on air quality?

Satellite NO2 retrievals suggest China has exceeded its NOx reduction goals from the twelfth Five-Year Plan

A space-based, high-resolution view of notable changes in urban NOx pollution around the world

Airborne Nitrogen Dioxide Plummets Over China

S5P Mission Performance Centre Nitrogen Dioxide

Flawed estimates of the effects of lockdown measures on air quality derived from satellite observations | Copernicus

Operational global reanalysis: progress, future directions and synergies with NWP

Impact of anthropogenic emission on air quality over a megacity -revealed from an intensive atmospheric campaign during the Chinese Spring Festival

Characterization and source identification of fine particulate matter in urban Beijing during the 2015 Spring Festival

Team, the O. core, 2019. OMI/Aura Nitrogen Dioxide (NO2) Total and Tropospheric Column 1-orbit L2 Swath 13x24 km V003

Science Objectives of the Ozone Monitoring Instrument

Spatial-Temporal Analysis on Spring Festival Travel Rush in China Based on Multisource Big Data

Satellite constraint for emissions of nitrogen oxides from anthropogenic, lightning and soil sources over East China on a high-resolution grid

A large decline of tropospheric NO2 in China observed from space by SNPP OMPS

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

Recent reduction in NO x emissions over China: synthesis of satellite observations and emission inventories

Increase in tropospheric nitrogen dioxide over China observed from space

Atmospheric Chemistry and Physics: From Air Pollution to Climate Change

Effect of changing NO&amp

lifetime on the seasonality and long-term trends of satellite-observed tropospheric NO&lt;sub&gt;2&lt;/sub&gt; columns over China

Shanghai People's Government, 2020. Shanghai COVID-19 Prevention News

A physics-based approach to oversample multi-satellite, multispecies observations to a common grid

Air pollution "holiday effect" resulting from the Chinese New Year

An investigation of transmission control measures during the first 50 days of the COVID-19 epidemic in China. Science (80-. ). eabb6105

Changes in air quality during the lockdown in Barcelona (Spain) one month into the SARS-CoV-2 epidemic

TROPOMI on the ESA Sentinel-5 Precursor: A GMES mission for global observations of the atmospheric composition for climate, air quality and ozone layer applications

Novel Coronavirus (2019-nCoV) Situationa Report-1

The effects of firework regulation on air quality and public health during the Chinese Spring Festival from 2013 to 2017 in a Chinese megacity

Characterization of lead-containing atmospheric particles in a typical basin city of China: Seasonal variations, potential source areas, and responses to fireworks

Long-term trends in NO2 columns related to economic developments and air quality policies from

Formaldehyde (HCHO) As a Hazardous Air Pollutant: Mapping Surface Air Concentrations from Satellite and Inferring Cancer Risks in the United States

Kang Sun acknowledges support from NASA ACMAP and RRNES programs.

The authors declare no conflict of interest.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.Journal Pre-proof J o u r n a l P r e -p r o o f