Reductions in NO2 concentrations in Seoul, South Korea detected from space and ground-based monitors prior to and during the COVID-19 pandemic

Nitrogen oxides detected in urban regions are primarily emitted by transportation methods and are crucial precursors for air pollutants and climate forcers such as ozone and fine particulate matter. We investigate the trends of nitrogen dioxide (NO2) obtained from a satellite instrument and surface monitors over the megacity, Seoul, South Korea, from 2005 to 2019. Both satellite Ozone Monitoring Instrument NO2 and surface in situ concentrations decreased by up to 30% between 2015 and 2019 while significant trends were not observed between 2005 and 2015. Further analysis shows the continual reduction of NO2 concentrations prior to and during the COVID-19 pandemic in 2020. This study highlights the efficacy of South Korean pollution control policies targeting vehicular emissions. However, this study also found inconsistencies between trends observed in the official bottom-up emission inventory and data collected from space and surface sites. Further research will be urgently needed to understand the causes for the discrepancies.


Introduction
Nitrogen oxides (NO x as sum of NO and NO 2 ) in the low atmosphere are released primarily through fossil fuel combustions (Crutzen 1979, Ryerson et al 2001, Harley et al 2005, van Vuuren et al 2011, McDonald et al 2012 and can also be emitted through other modes of anthropogenic activities such as agriculture (Matson et al 1998, Almaraz et al 2018. NO x plays a central role in atmospheric chemistry and air quality, particularly in population-dense urban areas with heavy road traffic. NO x along with carbon monoxide (CO) and volatile organic compounds (VOC) is the main precursor of tropospheric ozone (O 3 ) (Haagen-Smit 1952, Crutzen 1979, Jacob 1999, Ryerson et al 2001, Marr and Harley 2002, Jacobson 2005, Pollack et al 2013, Kim et al 2016. Furthermore, NO x can react with OH radical or other cations to form secondary organic/inorganic aerosols (e.g., organic nitrate, ammonium nitrate) (Park et al 2004, Ayres et al 2015, Seinfeld and Pandis 2016. Therefore, NO x is an essential precursor for both O 3 and particulate matter (PM). Enhanced concentrations of both O 3 and PM are currently major environmental and public health issues in Seoul, South Korea. For example, a recent tropospheric ozone assessment report showed that ozone levels were increasing in South Korea between 2000 and 2014 (Gaudel et al 2018). High PM concentrations and severe haze with enhanced nitrates were observed in South Korea (Lee et al 2017, Jordan et al 2020. Therefore, it is imperative to understand NO x concentration trends in order to establish mitigation strategies and regulatory policies for O 3 and PM. There have been several studies concerning air quality in Seoul due to the extensive use of diesel engine passenger cars and trucks in Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
South Korea (Souri et al 2017, Goldberg et al 2019, Souri et al 2020. However, recent studies do not provide a consistent characterization of NO x emission estimations in the Seoul metropolitan area (Goldberg et al 2019, Kim et al 2020a. An accurate understanding of the atmospheric composition in Seoul will be beneficial in improving our overall understanding of emissions and atmospheric chemistry over continental Asia and the subsequent consequences of long-range transport and pollution sources. Seoul, the capital of South Korea, currently has a population of around 10 million people living in the urban core region-an area of 605 km 2 (http://data.si.re.kr/node/337). 40% of the national population resides in Seoul and the surrounding Gyeonggi-do region with a combined population of 22 million people over an 10,800 km 2 area (http://stat.gg.go.kr/statgg/tblInfo/mainStats.html).
Fuel sale statistics in Seoul indicated that the volume of diesel fuel sold was almost equal to that of gasoline fuel in 2010 (https://data.seoul.go.kr/dataList/datasetView.do?infId=10789&srvType=S&serviceKind=2; table 1). In the Gyeonggi region, the volume of diesel sold exceeded that of gasoline (table 1). In addition to the high-degree of anthropogenic activity, the greater use of diesel fuel for passenger cars in these regions can further contribute to NO x emissions and local air quality (see McDonald et al 2012 for detailed emission factors for motor vehicles).
In this study, we analyze the satellite-observed tropospheric NO 2 columns from 2005 to 2019 to derive the trends and enhancement levels of NO 2 in Seoul. Ground level air quality monitoring data are additionally utilized to confirm the satellite trends of NO 2 . Furthermore, to understand the effect of government response to

Satellite data
The Ozone Monitoring Instrument (OMI) is one of instruments onboard the EOS-Aura satellite. The satellite was launched into a sun-synchronous orbit in 2004 (Bucsela et al 2006, Levelt et al 2006, Boersma et al 2011, Levelt et al 2018, with a low-Earth (705 km) orbit and an Equator overpass time of approximately 13:45 local standard time (LST). The Differential Optical Absorption Spectroscopy (DOAS) technique is used to obtain NO 2 slant column densities (SCDs) from the OMI spectra. Next, air mass factors (AMF) are calculated for each ground pixels to convert SCDs to vertical column densities (VCDs) following the equation below: AMF is derived as a function of atmospheric scattering, vertical shape factor, and optical properties such as solar zenith angle, satellite viewing angle, and surface albedo (Palmer et al 2001). The final step is the separation of stratospheric and tropospheric components. Details on the NO 2 algorithm and data assessment can be found in Lamsal et al (2021). In this study, we utilize Version 4.0 daily OMI level 3 tropospheric VCD data available from    In this section, the declining trends in satellite NO 2 column in recent years is similar to the trends in the in situ surface observations. It confirms utility of space-based NO 2 observations in deriving trends of NO 2 abundances and NO x emissions (see also Richter et al 2005, Kim et al 2009, Russell et al 2010, Hilboll et al 2013, Lamsal et al 2015, Duncan et al 2016, Kim et al 2020a.

Impact of COVID-19 on NO 2 abundances
The initial outbreak of coronavirus disease 2019 (COVID-19) was first identified in November 2019 according to World Health Organization (WHO 2020). In addition to more than 1 million deaths attributed to COVID-19, the pandemic has also caused global economic and social disruption (Leggett 2020) (~2.5 million deaths as of March 1, 2021, according to WHO website, https://covid19.who.int/). The first COVID-19 case in South Korea was identified in Daegu on 18 February, 2020. During this time the Korean government enforced social distancing protocol and various degrees of closures, which may have affected volumes of road-traffic due to remote-work/ virtual-learning and constrained commercial and industrial activities and travels. To understand the effects of COVID-19 on surface and tropospheric NO 2 concentrations, monthly climatology from 2005-2018 and the monthly averages from July 2019-May 2020 are compared in figure 4. Figure 4(a) shows values from the OMI L3 data, while figures 4(b) and (c) present the surface-observed concentrations for 07-09 LST and 12-14 LST, respectively. Both satellite and surface observational data indicate that the NO 2 concentrations from July 2019 to January 2020, preceding the COVID-19 outbreak in South Korea, are persistently lower than the climatological values by 6%-8%, suggesting that reductions of NO x occurred independently of the COVID-19 pandemic. NO 2 concentrations at surface strikingly decreased during the COVID-19 pandemic in South Korea (after February 2020) by up to~50% compared to the climatology, which is observed even in the morning commute time as well as the OMI overpass time around noon. The satellite data measured a smaller degree of NO 2 reduction. The NO 2 concentrations over SMA have remained at lower levels than climatological values after May 2020 (not shown). In China, NO 2 concentrations were substantially decreased during the initial lock-down period (24 January-February), but were recovered to climatological levels in early April towards the end of the mandated lock-down period (Bauwens et al 2020). In contrast, Seoul measured NO 2 concentrations consistently much lower than the climatology from 23 February to even May 2020 -with extended periods of social distancing having huge impacts on improving air quality. However, it is also important to note that NO 2 concentrations were lower than the climatological values prior to December 2019, before the COVID-19 outbreak in South Korea. From our observations, we speculate that the reduction of NO x emission and NO 2 levels over SMA is due to pollution Basic elements in the bottom-up inventory such as source speciation, activity, and emission factors will need to be critically evaluated for consistent trend estimations. Future studies adopting top-down approach that utilize the data from satellites such as OMI, TROPOMI, and GEMS (Kim et al 2020b) would be beneficial for improving the emission inventory.

Conclusions
The Seoul metropolitan area has the highest population density in South Korea. The air quality in this region is determined by a complex interaction of local emissions and foreign transportation (Jordan et al 2020). It is a prerequisite to quantify local emissions and their change to understand accurate source-receptor relationship in this region. NO x plays a crucial role in chemistry of O 3 and PM. Thus, in this study, we focus on detecting NO 2 abundances and their changes with time in SMA.
OMI tropospheric NO 2 columns and surface NO 2 concentrations over SMA were analyzed mainly from 2005 to 2019. The trends derived from surface monitoring sites are quite similar to those from OMI. The changes in NO 2 concentrations between 2005 and 2015 were negligible, while significant decreases were observed after 2015-in both surface and satellite observations. For the non-ozone season, reductions were detected, but are less significant than those of ozone season. While several factors such as long chemical lifetime due to weakened photochemical reactions (Lee et al 2013) and enhanced long-range transportation (Jeong et al 2011), and blocking via high pressure system (Yun and Yoo 2019) act on NO 2 concentrations during the nonozone season, local emissions have a dominant effect on air qualities in SMA during the ozone season. For this reason, we speculate that the local NO x emissions over the SMA region have decreased due to pollution controls led by Korean government. Meanwhile, NO x emission inventories estimated by NAPES do not agree with the recent decreasing trends of NO 2 concentrations, which requires further investigation on the bottom-up emission estimates as well as the top-down emission development utilizing environmental satellite retrievals and intensive measurement data sets based on aircrafts and surface monitors.
Despite the progress made by effective environmental policy and regulation, Seoul still has NO 2 concentrations almost double that of other megacities such as Los Angeles in the United States. Surface monitored NO 2 concentration in SMA and the Los Angeles Basin are 32 ppb and 15 ppb, respectively in 2016. OMI tropospheric NO 2 columns in SMA and the Los Angeles Basin are 15.6×10 15 molec. cm −2 and 7.7×10 15 molec. cm −2 , respectively for the same year. Thus, a continual effort must be made to reduce the NO 2 pollution in Seoul as striking decreases of the NO 2 concentration demonstrated during the COVID-19 pandemic. Preventive measures Korean government applied during the pandemic were important in protecting public health from the disease and from poor air quality. In the future, regional chemical transport modeling will be conducted to quantify the effect of Korean and Chinese NO x emission trends on NO 2 , O 3 , and PM concentrations in Seoul and the surrounding regions.