Assessment of COVID-19 Impacts on Air Quality in Ulaanbaatar, Mongolia, Based on Terrestrial and Sentinel-5P TROPOMI Data

Received: May 14, 2022
Revised: July 10, 2022
Accepted: August 18, 2022

 Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Cite this article:

Ganbat, G., Lee, H., Jo, H.W., Jadamba, B., Karthe, D. (2022). Assessment of COVID-19 Impacts on Air Quality in Ulaanbaatar, Mongolia, Based on Terrestrial and Sentinel-5P TROPOMI Data. Aerosol Air Qual. Res. 22, 220196. https://doi.org/10.4209/aaqr.220196

HIGHLIGHTS

• Impacts of COVID-19 on air pollution in Ulaanbaatar are studied.
• Terrestrial and satellite data were used.
• NO₂ concentrations decreased by up to 45% during the strict-lockdowns.
• Ground measurements agree with satellite measurements on NO₂.

Keywords: Air pollution, Strict lockdown, COVID-19, Ulaanbaatar

1 INTRODUCTION

COVID-19 and air pollution are linked in at least two ways. On the one hand, lockdowns often led to reductions in traffic and industrial production, and thus a reduction in air pollution. On the other hand, exposure to air pollution has been identified as a factor that aggravates the course of COVID-19 infections.

 
1.1 Impacts of COVID-19 on Global Air Pollution

The lockdowns introduced in about 140 countries to slow down the spread of COVID-19 constitute the largest quarantine policy in the history of public health (Liu et al., 2021a), and led to unprecedented declines in land and air transportation and economic activities (Venter et al., 2021). The following lockdown measures potentially contributed to changes in air quality: (1) internal travel restrictions (domestic and within certain cities); (2) international travel restrictions; (3) closure of public transport systems; (4) (partial) shutdown of industry; (5) stay-at-home requirements and (6) contact limitations, e.g., through closure of shops, kindergartens, schools and universities, cancellation of events, restrictions on meetings and gatherings (Liu et al., 2021a). At the global level, daily PM_2.5, PM₁₀, SO₂, NO₂, and CO concentrations decreased during lockdown periods. The greatest reductions (23 to 60%) were observed for NO₂, mostly due to restrictions on intracity travel, and for PM (by 7% to 45%). Reductions were almost negligible for SO₂, which is typically emitted by industry and energy production, and O₃ concentrations were slightly higher during lockdowns as compared to reference periods (Liu et al., 2021a; Venter et al., 2021). In Italy, the worst-affected country during the first wave of the pandemic, strict lockdowns were imposed from March 2020 onwards. While NO₂ concentrations dropped remarkably in cities throughout Italy (by 25% to 59%), PM_2.5 concentrations mostly decreased but partially also increased, presumably due to increased domestic heating (Gualtieri et al., 2020). In the UK, the implementation of a lockdown at the end of March 2020 induced an abrupt reduction in NO₂, NO, and NO_x at urban roadside monitoring stations, but a gradual return of traffic offset more than half of the reductions by summer 2020 (Ropkins and Tate, 2021). Menut et al. (2020) cautioned that different meteorological conditions and a general decrease in air pollution in western Europe make direct comparisons difficult. Nevertheless, through a modelling approach that considered meteorology and general air pollution trends, the authors confirmed that NO₂ strongly and PM moderately decreased due to the lockdowns (Menut et al., 2020). In the US, Son et al. (2020) showed that in 10 states, regions with high baseline levels of air pollution experienced the largest air quality improvements following travel and other restrictions.

Improvement in air quality due to lockdowns could also be noticed in Asian countries. Many countries in the region, particularly India and China, experienced a strong rise in air pollution as a consequence of their economic development that went along with increasing fossil fuel consumption in industry and transportation (Smith et al., 2001; Streets et al., 2003; Zhao et al., 2008; Smith et al., 2011; Chin et al., 2014). The core findings of case studies from East and Central Asia are summarized in Table 1. Changes in pollutant concentrations in the table have been rounded to whole numbers. It is important to note that the term “lockdown” refers to a wide range of restrictions, including those in mobility (up to curfews) and partial to full shutdowns of industrial production. Governments have also introduced different terminologies to refer to such periods, including COVID19 control policies, COVID-19 emergency responses, movement control orders, large-scale social restrictions and shutdowns. As the exact character of lockdowns and their implementation has often not been documented in detail, and papers are based on different numbers and durations of lockdown events and reference periods, the information provided should be considered as indicative values for the range of changes observed across Asia. In addition, it should be noted that a direct comparison between the different data is not meaningful since the studies did not only differ in methodologies, but also regarding the lockdown periods (timing, duration, and strength of the lockdown) and the reference periods against which the reductions have been calculated. Nevertheless, data from all countries, regions, and cities show lockdown-related reductions in air pollution.

Reductions in air pollution were observed not only over land but also in close proximity to pollution sources. Griffith et al. (2020) showed that long-range transport (LRT) of air pollutants from East Asia was approximately 50% less than during normal periods, especially from China. For South East Asia, Kanniah et al. (2020) reported a 27% to 30% reduction in NO₂ outflow over oceanic regions.

 
1.2 Impacts of Air Pollution on COVID-19 Pathogenesis

Even though the combined health impacts of air pollution and COVID-19 are not yet fully understood, several recent studies showed that air pollution appears to aggravate the risks related to COVID-19 infections (Srivastava, 2021). This is plausible since air pollution is known to be the cause of several respiratory illnesses such as chronic obstructive pulmonary disorder (COPD), cancer, and infections of the lung, but also to compromised immune systems (Gupta et al., 2021). A study in the Netherlands showed that PM_2.5, NO₂, and SO₂ concentrations in 355 municipalities correlated positively with registered COVID-19 cases, hospital admissions and deaths due to COVID-19 (Cole et al., 2020). Exposure to air pollution has in general been found to increase viral infections of the respiratory tract (Travaglio et al., 2021; Conticini et al., 2020). According to Setti et al. (2020), particulate matter can actually be a carrier of the COVID-19 virus. Research on urban air pollution in cities of China, India, Indonesia, and Pakistan revealed that long-term exposures to high air pollution, particularly to PM_2.5, negatively impacted the outcomes of COVID-19 infections. The mortality rate associated with COVID-19 is significantly correlated with PM_2.5 (p < 0.05), which is responsible for most of the air pollution-related deaths in the world (Gupta et al., 2021). Copat et al. (2020) found that PM_2.5 and NO₂ concentrations were more closely correlated to the spread and lethality of COVID-19 than PM₁₀. Filippini et al. (2021) showed for two regions of Italy that NO₂ concentrations correlated positively with COVID-19 severity. In a study on COVID-19 mortality in 66 administrative regions in Italy, Spain, France, and Germany, Ogen (2020) found 78% of the deaths to have occurred in those five regions that had the highest NO₂ concentrations and concluded that long-term exposure to this pollutant may be one of the most important contributors to COVID-19 fatality. According to Karuppasamy et al. (2020), improvements in air quality can therefore have the co-benefit to reduce mortality directly and indirectly (via reducing the risks associated with COVID-19 infections).

 
1.3 COVID-19 Measures Taken in Mongolia

According to the Law on Disaster Protection (Mongolian Parliament, 2003), the disaster preparedness regime in Mongolia is divided into three levels—daily preparedness, enhanced readiness, and public emergency readiness. Mongolia went into the state of “enhanced readiness” level on February 11, 2020. The precautionary measures of the “enhanced readiness” level include travel restrictions, partial remote work from home, school closings, and restrictions in public activities. The first positive case of coronavirus, who arrived in Mongolia via an international flight was reported on March 10, 2020 (Erkhembayar et al., 2020). On November 10, 2020, the first case of community transmission from an individual arriving from Russia was registered. After the patient was released from isolation for 21 days, Mongolia’s State Emergency Committee (SEC) announced the “public emergency readiness” level (or strict-lockdown) measures: all services and businesses except essential sectors were closed to work from home, the stay-at-home regime was activated, all international and domestic traffic beyond city boundaries was temporarily suspended, public gatherings suspended, and all levels of educational institutions are closed. In order to prevent the spread of coronavirus, three periods (a total of 63 days) of strict-lockdown periods were set from November 2020 to February 2021 (see Table 5).

According to Shrestha et al. (2020), declines in air pollutant concentrations up to 43% related to COVID-19 lockdowns in Ulaanbaatar were reported. However, there was no mention of the combined reason for the declines which is indeed related to change in fuel type since winter of 2019-2020 and seasonal variations of the pollutants. The winter of 2019–2020 was characterized by substantial declines in PM_2.5 and PM₁₀ concentrations due to the transition from raw coal to briquette fuel in Ulaanbaatar (Ganbat et al., 2020). The current study investigates the effect of COVID-19 strict lockdowns on air pollutants in Ulaanbaatar, Mongolia, using terrestrial and satellite observations.

 
2 MATERIALS AND METHODS

 
2.1 Study Area

Ulaanbaatar, the capital of Mongolia, is located in a valley between the Bogd Khan mountain in the south and the extensions of the Khentii Mountains in the north (Fig. 1) at an approximate altitude of 1300 m above sea level. Ulaanbaatar has a population of nearly 1.5 million inhabitants, accounting for 47% of the total population of Mongolia. Ulaanbaatar is known as the ‘coldest capital’ in the world with winter temperatures often dropping below –20°C and the cold weather in winter is attributed to the Siberian high-pressure system, which causes the formation of a temperature inversion (Ganbat and Baik, 2016). In poor vertical mixing under the weather condition with temperature inversions, the majority of pollutant sources is crustal matter and coal combustion (Davy et al., 2011) from ger areas, where around half of the Ulaanbaatar’s population lives (Karthe et al., 2022). Heating in ger areas during the heating season is supplied by fuel-stoves. After a decade of the severe air pollution problem, a notable reduction in air pollution in Ulaanbaatar is seen after introducing upgraded briquette fuel since winter 2019–2020 (Ganbat et al., 2020; Soyol-Erdene et al., 2021). However, the PM concentrations in winter 2019–2020 still exceeded the national air quality standard levels.

Fig. 1. Air quality monitoring sites in Ulaanbaatar.

 
2.2 Terrestrial Data

Concentrations of air pollutants—NO₂, SO₂, PM₁₀, and PM_2.5, which are commonly used to assess air quality, are used in this study to investigate the effects of COVID-19 measures on air quality in Ulaanbaatar. Data from 1 January 2015 to 1 March 2021 were obtained from 12 air quality monitoring sites located at various points in Ulaanbaatar (Fig. 1 and Table 2). The sites are operated by the National Agency for Meteorology and Environmental Monitoring (NAMEM) and the Air Pollution Reduction Department (APRD) of the Municipality.

According to the current national air quality standard, MNS 4585:2016, which is amended in 2016, the national standard levels of 24-h NO₂, SO₂, PM₁₀, and PM_2.5 are set 50 µg m^–3, 50 µg m^–3, 100 µg m^–3, and 50 µg m^–3, respectively. The annual standard levels of NO₂, SO₂, PM₁₀, and PM_2.5 are 40 µg m^–3, 20 µg m^–3, 50 µg m^–3, and 25 µg m^–3, respectively.

 
2.3 Air Pollution Monitoring by Remote Sensing

Meteorological satellites have been able to monitor atmospheric water vapor content since the early 1970s, and the Advanced Very-High-Resolution Radiometer (AvHRR) sensors have provided global information on the occurrence and distribution of aerosols since 1981 (Stowe et al., 2002; Zhang et al., 2020). The introduction of the Global Ozone Measurement Experiment (GOME) spectrometer aboard the ERS-2 satellite in 1995, and the Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY) aboard the ENVISAT satellites since 2002 significantly widened the potential of satellite remote sensing for air pollution monitoring as it provided information on aerosols and a wide range of trace gases including NO₂, SO₂ and several others (Bovensmann et al., 1999; Noël et al., 2003). Further improvements in spectral, radiometric, spatial, and temporal resolution have been realized with the TROPOspheric Monitoring Instrument (TROPOMI) aboard the Sentinel 5 Precursor satellite. It allows the monitoring of temporal changes and assessment of long-term trends in atmospheric chemistry. It can quantify aerosols and several trace gases including NO₂, O₃, SO₂, CO, CH₄, and CH₂O. TROPOMI is based on a UVNS (UV–VIS–NIR–SWIR) spectrometer that scans an approximately 2600 km wide swath, along which it moves at a speed of 7 km s^–1. It offers a daily coverage at a spatial resolution of 7 km × 7 km in most bands, and 21 km × 28 km in the UV band (Veefkind et al., 2012). The accuracy and precision of the TROPOMI instrument for the parameters considered in this study are shown in Table 3.

China has recently developed into another major operator of satellites capable of air pollution monitoring, including aerosols, NO₂, SO₂, and several other pollutants (Zhang et al., 2020). Remote sensing-based approaches have contributed to air quality assessments at various scales ranging from local studies—often in affected cities— (e.g., Huang et al., 2021; Prunet et al., 2020) to regional and national-level investigations (e.g., Shikwambana et al., 2020; Stratoulias and Nuthammachot, 2020; Zheng et al., 2019) and global studies (e.g., Lim et al., 2020). Particularly, studies focusing on NO₂ to thoroughly understand the effects of the COVID-19 lockdown on the atmosphere are gaining popularity (e.g., Scheibenreif et al., 2021; Vîrghileanu et al., 2020; Wang et al., 2021). This study uses TROPOMI NO₂ spatial variations as an extension to illustrate changes in several pollutants during the COVID-19 strict and non-strict lockdown periods.

 
3 RESULTS AND DISCUSSION

The time series of daily mean NO₂, SO₂, PM₁₀, and PM_2.5 concentrations from January 2015 to March 2021 averaged over the air quality monitoring sites in Ulaanbaatar are shown in Fig. 2. In general, decreasing trends in NO₂, PM₁₀, and PM_2.5 after winter 2019–2020 can be clearly seen. The pollutant concentrations show that the main air pollutants in Ulaanbaatar were PM₁₀ and PM_2.5 before the winter of 2019–2020. The concentrations of all pollutants have strong seasonal patterns. The winter of 2020–2021 was the period with the lowest PM₁₀ (108.5 ± 36.2 µg m^–3) and PM_2.5 (76.0 ± 38.9 µg m^–3) since 2016. On the contrary, it is clearly seen that SO₂ concentrations significantly increased in the last two winters.

Fig. 2. Time series of daily mean NO₂, SO₂, PM₁₀, and PM_2.5 and concentrations from 1 January 2015 to 1 March 2021 averaged over the air quality monitoring sites in Ulaanbaatar.

A decrease in yearly-mean NO₂ concentration of 17% was observed in 2020 compared to the mean of 2015–2019. The decreasing trend is mild for the yearly values but more pronounced in winter. An increasing trend in SO₂ concentrations is observed in winters after 2019–2020. An increase in SO₂ concentration of 79% was observed in 2020 compared to 2015–2019 mean and 2.7 times higher than the national standard value of 20 µg m^–3.

A decrease in PM₁₀ concentration of 24% was observed in 2020 compared to the 2015–2019 mean. However, the annual PM₁₀ concentration still exceeds the national standard value of 50 µg m^–3. Decrease in PM_2.5 concentration of 51% was observed in 2020 compared to 2015–2019 mean. Comparisons of PM₁₀ and PM_2.5 concentrations between before and after the 2019–2021 winters show decreased trends while SO₂ concentration showed a notable increased trend. Specifically, SO₂ concentrations in winters 2019–2020 and 2020–2021 are steadily increased. The reason for the increased SO₂ concentration could be attributed to contents of the briquette fuel consumed in households in ger areas since 2019 and the increased demand in consumption of briquette fuel in households due to stay-at-home activity. In line with a previous study by Ganbat et al. (2020), the immediate reductions of PM₁₀ and PM_2.5 in winter 2019–2020 are related to the change in fuel type. Thus, the declines in PM₁₀ and PM_2.5 in winter 2020–2021 are likely attributed to the combined effects of measures of the transition from raw coal to upgraded briquette fuel and the effects of the COVID-19 strict-lockdowns.

The combined effects of this source change and weather condition may affect the changes in pollutant concentrations. To mention, according to official reports released by the NAMEM, the weather condition during the winter 2020–2021 was not peculiar (www.tsag-agaar.mn). The weather condition in the winter 2020–2021 showed the similar characteristics compared to the previous winters—low wind speed and temperature similar to previous years. To clearly see the impacts of strict-lockdowns to air pollution, the periods of strict-lockdowns and non-lockdowns are compared (please see Table 5 and Fig. 5 later).

It should be noted that even the concentrations are significantly reduced, the yearly-mean values of the pollutant concentrations remained persistently above the national standard values.

Table 4 supplements Fig. 2 providing a number of days with average pollutant concentration from November 1 to February 28, which covers three sequential strict-lockdowns, exceeding one, two, and three times the national air quality standard levels.

From November-February in 2014–2020 to 2020–2021, the days with an average NO₂ concentration above the national air quality standard level reduced from 101 to 45. The strict-lockdowns during November 2020–February 2021 have resulted in a 55% reduction in the number of days exceeding the national air quality standard level of NO₂ compared to 2014-2020. During November 2020–February 2021, there were no days exceeding two and three times the national air quality standard level of NO₂. A similar feature is seen in many cities in the world (Acharya et al., 2021). It is important to note that one of the substantial reasons for the decrease in NO₂ concentration was restrictions on city traffic during the strict-lockdown periods.

For winters 2019–2020 and 2020–2021, the number of days exceeding the national air quality standard level for PM₁₀ decreased by 30%. For winters before 2019, the number of days with average PM₁₀ concentrations exceeding three times the national air quality standard level ranged from 9 to 65, but for the past two winters, there were no such days.

For November 2017–February 2018, in 65 days the PM_2.5 concentration exceeded three times the national air quality standard level. For winters 2019–2020 and 2020–2021, the number of days exceeding the national air quality standard level for PM_2.5 decreased by 14% compared to 2014–2019. These sudden reductions of PM₁₀ and PM_2.5 concentrations during the last two winters could be linked to the transition from raw coal to briquette fuel. However, a greater declining trend is exhibited in winter 2020–2021, when the sequential strict-lockdowns have been imposed due to the COVID-19 pandemic.

Before winter 2019–2020, there were no days with average SO₂ concentrations exceeding three times the national air quality standard level, but for the last two winters, the days exceeding three times the national air quality standard level significantly increased—no days versus 11 and 56 days in winter 2019–2020 and 2020–2021, respectively.

Descriptive statistics of each strict-lockdown period with respect to the same periods in the same periods in the previous five years are done. The mean and standard deviation of pollutant concentrations during the strict-lockdown periods and their changes from the same periods in the previous five years are shown in Table 5. All pollutants except SO₂ present declines for the period from November 2020 to February 2021. Notable declines are apparent during the strict-lockdown periods (L1, L2, and L3). During the strict-lockdown periods, lower concentrations (from 39% to 72%) were observed. The concentration levels are likely to go up once the situation back to normal (N1 and N2) but remain still lower compared to the same periods in the previous five years.

On average, a decrease of 41%, 53%, and 55% in NO₂, PM₁₀, and PM_2.5 concentrations, respectively, is found, whereas a 229% increase in SO₂ concentrations during the strict-lockdown periods compared to the same periods in the previous five years. Such effect of lockdown on the reduction of NO₂ was observed in other cities of the world (Fu et al., 2020; Acharya et al., 2021) as a result of decreases in various anthropogenic activities.

Visible reductions in PM₁₀ and PM_2.5 remain to be consistently seen for the period from November 2020 to February 2021. The substantial reductions in PM₁₀ and PM_2.5 concentrations from November 2020 to February 2021 compared to the previous five years could be due to the combination of fuel change and COVID-19 mitigation measures. The maximum decrease in NO₂, PM₁₀, and PM_2.5 concentrations was observed in L3. An opposite trend is observed for SO₂ concentration. SO₂ concentration increased from twofold (212%) to threefold (331%) during the periods. Time series of daily mean NO₂, SO₂, PM₁₀, and PM_2.5 concentrations from November to March are shown in Fig. 3. The figure reveals significant decreases in NO₂, PM₁₀, and PM_2.5 concentrations during the whole period covering three sequential strict-lockdowns.

Fig. 3. Time series of NO₂, SO₂, PM₁₀, and PM_2.5 concentrations from 1 November to 28 February.

The daily-mean (highest) NO₂ concentration reduction ranged from 27 (25) % to 41 (39) % between November 2020–March 2021 and the same periods in the previous five years. NO₂ concentration for the period between November 2020 and March 2021 decreased by 35% on average compared to the previous five years. During the three sequential strict-lockdowns, the daily NO₂ concentration drops to below 20 µg m^–3 first ever since 2015.

To take a further look at the variations of pollutant concentrations between strict-lockdowns, we compare the pollutants concentrations averaged over seven days before and after the start of the strict-lockdown. The daily mean concentrations of NO₂ were 41.8, 63.6, and 53.6 µg m^–3 before the first, second, and third strict-lockdown periods, which reduced to 33.5, 47.5, and 35.3 µg m^–3, respectively. The NO₂ concentrations were reduced by 20%, 25%, and 34%, respectively.

The daily mean concentrations of SO₂ were 46.8, 233.6, and 167.5 µg m^–3 before the first, second, and third strict-lockdown periods, which reduced to 49.9, 156.9, and 139.5 µg m^–3, respectively. The SO₂ concentrations changed by –7%, 33%, and 17%, respectively.

The daily mean concentrations of PM₁₀ were 118.6, 142.0, and 108.9 µg m^–3 before the first, second, and third strict-lockdown periods, which reduced to 87.6, 103.5, and 78.4 µg m^–3, respectively. The PM₁₀ concentrations were reduced by 26–28%.

The daily mean concentrations of PM_2.5 were 45.9, 100.4, and 85.9 µg m^–3 before the first, second, and third strict-lockdown periods, which reduced to 34.9, 77.2, and 57.4 µg m^–3, respectively. The PM_2.5 concentrations are reduced by 24%, 23%, and 33%, respectively.

Among NO₂, PM₁₀, and PM_2.5, the NO₂ concentrations significantly reduced by 34% after the third strict-lockdown. The third strict-lockdown exhibited the greatest reductions in pollutant concentrations.

Fig. 4 shows that the probability distribution functions (PDFs) of concentrations of NO₂, PM₁₀, and PM_2.5 were consistently lower during the three sequential strict-lockdown periods.

Fig. 4. Probability distribution functions of (a) NO₂, (b) SO₂, (c) PM₁₀, and (d) PM_2.5 concentrations for averaged over strict-lockdown periods (red) and the same periods in previous five years (blue).

The peak occurrence of NO₂, PM₁₀, and PM_2.5 concentrations during strict-lockdown periods are 45.3, 76.7, and 52.5 µg m^–3, while they were 70.3, 128.6, and 113.1 µg m^–3, respectively, during the same periods in the previous five years. The percentage exceeding the national standard (MNS 4585:2016) level for NO₂, PM₁₀, and PM_2.5 constituted 23%, 50%, and 67% during the lockdown periods while it was 89%, 84%, and 91%, respectively, during the same periods in previous five years. NO₂ experienced the greatest benefit, with a –66% reduction in percentage exceeding the national standard value.

For SO₂, high concentrations of SO₂ became more frequent during the strict-lockdown periods compared to the same periods in the previous five years. The percentage exceeding the national standard level for SO₂ increased from 54% to 89%. The peak occurrence of SO₂ concentration during strict-lockdown periods was 46.8 µg m^–3, while it is increased to 123.5 µg m^–3 during the same periods in the previous five years.

The hourly patterns of the pollutant concentrations are illustrated in Fig. 5. The pollutant concentrations showed two peaks during the day.

Fig. 5. Hourly variations in (a) NO₂, (b) SO₂, (c) PM₁₀, and (d) PM_2.5 concentrations averaged over strict-lockdown periods (blue) and the same periods in the previous five years (red).

The concentrations of NO₂, PM₁₀, and PM_2.5 are recorded as being substantially less during the strict-lockdown periods. For example, daily mean NO₂, PM₁₀, and PM_2.5 concentrations during strict-lockdowns were 63.3, 158.7, and 120.2 µg m^–3, which were reduced by 33, 39, and 46 %, respectively, when compared with those of the same periods in the previous five years. SO₂ concentration exhibited an increasing trend against previous years. The changes in concentrations were observed, however, the time of maximum and minimum concentrations are not changed much. These results clearly indicate that measures taken during the strict-lockdowns have substantially affected the air pollution in Ulaanbaatar.

As an extension of the measurement data to illustrate the spatial variations of NO₂ for the study period, tropospheric NO₂ columns from TROPOMI during COVID-19 strict-lockdown and no strict-lockdown periods (L1, N1, L2, N2, and L3) in winter 2020-2021 are shown in Fig. 6. The tropospheric NO₂ columns over the Ulaanbaatar are shown in Fig. 6(a) for the first COVID-19 strict-lockdown period from November 11, 2020 to December 14, 2020 (L1). Fig. 6(b) illustrates that the tropospheric NO₂ concentration has remained high (> 0.00021 mol m^–2) from December 14, 2020 to December 22, 2020 during no strict-lockdown (N1).

Fig. 6. Tropospheric NO₂ column densities over Ulaanbaatar observed by Sentinel-5P TROPOMI during the strict-lockdowns (left panel) and periods between them (right panel). Details of the abbreviations (L1, N1, L2, N2, and L3) are given in Table 5.

TROPOMI data for NO₂ at the city scale showed a reduction for lockdown periods as compared to non-lockdown periods. The observed reduction between N1 and L2 based on TROPOMI data was 48.6% for the whole city, as compared to an average reduction of 28.5% for the 11 monitoring stations equipped with NO₂ sensors. Between N2 and L3, a reduction of 35.2% was observed for the city area based on TROPOMI data, as compared to an average reduction of 34.6% for the 11 monitoring stations. In this context, it should be mentioned that the monitoring stations were set up at locations that represent different pollution pattern (including sites with a high density of heating stoves and sites with significant traffic), but that averages may not be fully representative for the entire city. Moreover, one of the monitoring stations (Bayankhoshuu) was not equipped with an NO₂ sensor, thus leading to an information gap for ground-based data in one of the city’s largest and most highly polluted ger areas. As standard deviations for NO₂ concentrations across the city were high for TROPOMI data (between ±30% and ±59% depending on the time period), averages were also calculated for 1 km buffer zones around the monitoring stations. In this case, standard deviations were far smaller (between ±8% and ±11%) for the time periods considered, and the observed reductions were 53.4% (between N1 and L2) and 35.7% (between N2 and L3). All in all, TROPOMI-based observations allowed a realistic assessment of general trends of NO₂ pollution (see Fig. 7) while also providing coverage of areas without monitoring stations.

Fig. 7. The correlation analysis between terrestrial data and TROPOMI NO₂ tropospheric concentration.

Studies have proved these tropospheric NO₂ hotspots in downtown and midtown areas are associated with human activities including ground traffic and industrial activities (Hashim et al., 2021; Huang et al., 2021; Rissman et al., 2013). Apparently, just after lockdown measurements were imposed from December 23, 2020 to January 10, 2021 (L2), the NO₂ concentrations dropped sharply as shown in Fig. 6(c). However, during no strict-lockdown period from January 11, 2021 to February 10, 2021 (N2), NO₂ concentrations once again increased as exhibited in Fig. 6(d). Another significant reduction in NO₂ concentrations was observed during the third lockdown from February 11, 2021 to February 23, 2021 (L3) as documented in Fig. 6(e).

Previous studies have also shown a reduction of tropospheric NO₂ columns in 2020 lockdown period over various regions in the world (Goldberg et al., 2020; Hashim et al., 2021; Huang et al., 2021; Wang et al., 2021). Likewise, similar results have been reported in Asia regions, particularly Wang and Su (2020) and Bauwens et al. (2020) investigated a sharp decline in NO₂ concentrations over China and South Korea respectively based on Sentinel-5P data. Our results are comparable with these studies, indicating that the COVID-19 lockdown played an important role in the NO₂ reduction.

The results of correlation analysis between terrestrial data over the ground station and TROPOMI NO₂ tropospheric concentration from Sentinel-5P are shown in Fig. 7. The results show that the NO₂ concentration retrieved by TROPOMI is highly correlated with the surface monitoring concentration of NO₂ in Ulaanbaatar (correlation coefficient = 0.91, R² = 0.85).

 
4 SUMMARY

Following the novel COVID-19 (Coronavirus) case that was spread worldwide, prevention measures were implemented by the Government of Mongolia. The spread of COVID-19 disease to the community was first reported on 10 November 2020 in Mongolia. Starting 11 November 2020, the Government of Mongolia and the State Emergency Commission announced a series of intermittent strict-lockdowns to prevent the COVID-19 spread. This study focuses on three sequential strict-lockdowns from November 2020 to February 2021 which were announced to discontinue the novel coronavirus spread. A significant effect on air quality has been seen in Ulaanbaatar, Mongolia.

The impact of strict-lockdown on air pollution in Ulaanbaatar was assessed by comparing air pollutant concentrations before and during strict-lockdowns and during non-strict lockdown periods. Changes in Ulaanbaatar’s air quality showed significant declines in NO₂ (up to 45%), PM₁₀ (up to 72%), and PM_2.5 (up to 59%) concentrations between November 2020 and March 2021 compared to the previous five years. These reductions were among the greatest observed across Asia, which is plausible considering the strict character of the lockdowns and the transition to cleaner coal briquettes. However, the concentrations were still above the national standard values, and SO₂ concentrations showed an increasing trend in the last two years, especially in winter. The reason could be attributed to the contents of the briquette fuel consumed in households in ger areas after the city-wide transfer from the raw coal to briquette fuel and the increased demand in consumption of briquette fuel in households due to long hours spent at home. The measures taken during the strict-lockdown periods clearly influenced the values of daily patterns of NO₂, PM₁₀, and PM_2.5 concentrations. The maximum concentration peaks appreciably decreased. In contrast, it is important to note that SO₂ concentration increased during the last two winter months after 2019. Furthermore, Sentinel-5P retrieved NO₂ tropospheric concentrations which were employed as an extension to illustrate changes of several pollutants during the COVID-19 strict and non-strict lockdown periods, are in agreement with the reductions observed at ground stations. Our current study underlines the findings of studies from other parts of the world, revealing both positive and negative impacts of lockdowns on air quality for Ulaanbaatar, Mongolia. However, as Ulaanbaatar has been developing very dynamically in recent years, and a major campaign to substitute the fuel for heating for about half of the city’s households was implemented just prior to the pandemic, it is difficult to fully disentangle the impacts of COVID-19 from other developments. As only preliminary findings on the fuel substitution exist so far, the related uncertainties need to be addressed once the pandemic situation has improved and consolidated findings on the impacts of fuel substitution become possible.

 
ACKNOWLEDGEMENTS

Daniel Karthe would like to express his gratitude to German Academic Exchange Service (DAAD) for funding a long-term lectureship at the German-Mongolian Institute for Resources and Technology (GMIT).