Introduction

The concentration and composition of PM2.5 or less than PM2.5, one of the air pollutants, can be toxic and can impact human health (Brunekreef and Holgate 2002). PM2.5 has numerous adverse effects such as cardiovascular disease, chronic pulmonary disease, oxidative stress, pneumonia (Miller and Xu 2018; Dominici et al. 2006; Feng et al. 2016) and the birth effect (Shang et al. 2019). According to the US study, long-term exposure to ambient PM2.5 is associated with increased risks of adult mortality (Franklin et al. 2008). In 2019, the average concentration of PM2.5 was 62 µg/m3; Mongolia ranked third in the world for air pollution after Bangladesh and Pakistan (IQAir. 2020). These countries belong to the developing countries of Asia, and a recent study shows that developing countries are more prone to air pollution than developed countries (Shisong et al. 2018).

Ulaanbaatar city is located in a valley with mountains to the north and south (Kottek et al. 2006). The topography, extensive pollution emissions and frequent temperature inversions have combined to cause high concentrations of PM2.5 in the winter months (Allen et al. 2013). The population growth caused mainly by rural to urban migration has led to the expansion of air pollution in the capital of Mongolia. The population in the early 1960s was 150,000, but it increased rapidly after democratization reached 580,000 in the 1990s and was recently reported to have exceeded 1.5 million (Mongolian statistical information service 2021).

Another contributing factor to increasing air pollution is a temperature inversion, frequently observed in mountainous urban areas and can cause severe air pollution problems, especially in wintertime (Ganbat and Baik 2016; Hou and Wu 2016). The study showed that the impact of coal combustion particles was high in winter, particularly condensed particles composed of organic carbon used by the PMF model. However, the impact of soil dust on the source of PM2.5 was dominant in the autumn and spring (APRD JICA, 2017). The result of coal-burning used for heating in the winter season accounts for a considerable proportion (60–70%) of air pollutant emissions of Ulaanbaatar (World Bank Sustainable and Alternative Energy Program 2009). As reported by a 2013 analysis, ambient PM2.5 provided 29% of cardiopulmonary deaths, 40% of lung cancer deaths and nearly 10% of all-cause mortality in Ulaanbaatar (Allen et al. 2013). The government of Mongolia has accomplished enormous actions to reduce air pollution in the last 13 years. One of the recent actions had banned the usage of raw coal in the capital city starting from 15 May 2019 (Minister of Nature, Environment and Tourism & Governor of the Capital City and the Mayor of Ulaanbaatar 2019). However, Khovd city, another study area, is still using raw coal.

Khovd city covers landscapes of the Great Lake Basin in the northeast, and it has a sheltered, extremely continental and dry climate. In winter, extensive temperature inversions involving warm air mass in lower mountain regions superimpose above cold air mass in the plains, leading to relatively warm conditions on the winter pastures at approx. 2000 m a.s.l (Zemmrich 2008). In 2019, more than 90,000 people lived in Khovd province, and 30,000 of them lived in Khovd city. In the city, there are 5200 households, and many organizations have low-pressure stoves that use raw coal (Mongolian statistical information service, 2021).

Ulaanbaatar city is a cultural, economic and political center of Mongolia. Likewise, Khovd is the central city of the western region of Mongolia. As mentioned earlier, there are several similarities between Ulaanbaatar and Khovd cities. Although several studies have been conducted on air pollution in Ulaanbaatar, there is a lack of research on air quality in other cities of Mongolia. Therefore, we aimed to compare the concentration of PM2.5 in Ulaanbaatar city and Khovd city. The overall objectives of the present study are (i) an analysis of PM2.5 variation in Ulaanbaatar in recent years; (ii) a comparison of PM2.5 of Ulaanbaatar and Khovd cities in 2021; (iii) factors affecting PM2.5 in Khovd city and approaches to reduce the concentration.

Materials and methods

Study area

The study area is located at NUM (106°55′6.30"E, 47°55′25.41"N) in the central part of Ulaanbaatar and at KhU (91°38′41.33"E, 48° 0′10.16"N) in central part of Khovd city. Both cities lie in a valley surrounded by mountain slopes. Sampling was conducted at Ulaanbaatar and Khovd cities (Fig. 1).

Fig. 1
figure 1

a Location of Khovd and Ulaanbaatar, Mongolia. b Air quality monitoring site in Khovd. c Air quality monitoring site in Ulaanbaatar

Full size image

Study instrument

The National University of Mongolia (NUM) was selected as the one of the study areas for the PM2.5 sensor to measure PM2.5 in Ulaanbaatar in 2017. The other one is at Khovd University (KhU), and its measurement data have been collected since 2020. To measure PM2.5, we used sensors of Panasonic model GA1 with a thermal resistor to induce an internal upward air flow to facilitate continuous sampling. This sensor unit has an accuracy of ± 10% from low to high concentrations (∼ 1,000 µg/m3) (Motlagh et al. 2020), and its average flow rate in the detection region was estimated to be 65 mm/sec (Nakayama et al. 2018). Briefly, the particle size distribution was estimated by measuring the scattering intensities of single particles. The output of the PM2.5 sensors was calibrated using nearly monodisperse polystyrene latex (PSL) particles (Ly et al. 2018). One of the main features of light-scattering particles relates to their low power consumption (Cheng et al. 2014; Chowdhury et al. 2018). The sensor devices are also equipped with a Wi-Fi module and mobile phone connectivity for data logging and visualization.

Methods

Sampling

Over the last four years, the PM2.5 concentration data had collected every 10-s interval through the PM2.5 sensor. Data consists of 120, 92 and 151 data points (daily average of PM2.5 concentration) from every heating season, semi-heating season and non-heating season, respectively. Monthly data of temperature and the wind speed were obtained from the web of University of Wyoming (University of Wyoming 2021).

Statistical analyses

Firstly, the R program which is free software environment for statistical computing, and graphics supported by the R Foundation for Statistical Computing was used to analyze the data. Also, we calculated the number of long-term data using OriginLab, scientific graphing and analysis software, in this study.

For this study, the following equation was used to construct the air quality index (AQI) of PM2.5 (U.S. Environmental Protection Agency 2018). The values of the break points for the AQI function below are given in Table 1.

$$Ip = \frac{IHi - IL}{{BPHi - BPLo}}\left( {Cp {-} BPLo} \right) + ILo$$
(1)

where \(Ip\) = the index for pollutant p.

Table 1 AQI levels and associated health impacts
Full size table

\(Cp\) = the truncated concentration of pollutant p or P pollutant of concentration. \(BP\) Hi = the concentration breakpoint that is greater than or equal to Cp. \(BP\) Lo = the concentration breakpoint that is less than or equal to Cp. \(I\) Hi = the AQI value corresponding to BPHi. \(I\) Lo = the AQI value corresponding to BPLo.

Results and discussion

Results

An analysis of PM2.5 in Ulaanbaatar from 2017 to 2020

We divided the PM2.5 monitoring period into; heating period, semi-heating period and non-heating period. Heating period, semi-heating period and non-heating period cover between November and February, September and October, and March–April and May–August, respectively. Long-term PM2.5 concentration, air temperature and wind speed values between 2017 and 2020 are given in Fig. 2. In 2017, January, the coldest month of the year, monthly average PM2.5 concentration was 162 µg/m3, with an average temperature of –25 °C and the wind speed of 2.1 m/s. In January of 2018, 2019 and 2020, PM2.5 concentration reached 88.5, 181.4, 88.9 µg/m3, whereas the average temperature was –26.4, –21.7 and –21.8 °C, and the wind speed was 2.4, 2.7 and 1.7 m/s, respectively. On the other hand, July, the warmest period of the year, had monthly average PM2.5 concentration 11 times higher than that of January in 2017, and the average air temperature and wind speed were 20.4 °C and 5.7 m/s, respectively. Whereas July 2018 had a PM2.5 concentration of 11.9 µg/m3, its average temperature was 16.7 °C and the wind speed was 4.9 m/s.

Fig. 2
figure 2

The monthly average PM2.5 concentration in Ulaanbaatar between 2017 and 2020

Full size image

As can be seen from Fig. 2, the measurement of PM2.5 was illustrated from January 2017 to December 2020, as well as the temperature and wind speed for those periods. In 2017, in the coldest month, January, monthly average PM2.5 concentration was 162 µg/m3 with an average temperature of -25 °C and the wind speed of 2.1 m/s. In January of 2018, 2019 and 2020, the concentrations of PM2.5 were 88.5, 181.4 and 88.9 µg/m3, average temperatures were -26.4, -21.7 and -21.8 °C and the wind speed was 2.4, 2.7 and 1.7 m/s, respectively. However, in the hottest month, July 2017, monthly average PM2.5 concentration was 14.4 µg/m3 with an average temperature of 20.4 °C and the wind speed of 5.7 m/s, whereas July 2018 had a PM2.5 concentration of 11.9 µg/m3, its average temperature was 16.7 °C and the wind speed was 4.9 m/s. Moreover, in July 2019, the PM2.5 concentration was 11.8 µg/m3 with an average temperature of 18.9 °C and the wind speed of 5.3 m/s. Data collected for July 2020 show a very similar figure to that of July 2019, which had a monthly average PM2.5 concentration of 9.3 µg/m3 with an average temperature of 18.2 °C and the wind speed was 5.1 m/s. According to the chart, the regularity of the temperature and wind speed was approximately equal during study periods, while the PM2.5 concentration increased sharply which was 144.1 µg/m3 during the heating season of 2018–2019. However, concentrations of PM2.5 were relatively low in the heating season of 2020. Figures 3, 4 and 5show a boxplot of PM2.5 concentration and calculated AQI conveyed as a percentage of each season.

According to Fig. 3, the PM2.5 concentration during the heating season was 115 ± 55 µg/m3 in 2017, 125 ± 70 µg/m3 in 2018, 107 ± 70 µg/m3 in 2019 and 59 ± 29 µg / m3 in 2020. In the last four years until 2020, PM2.5 was a high level in the heating season of 2018, and in 2020 it was the lowest. During the study period, the maximum concentration was 401 µg/m3 in 2019, and the lowest concentration was 6.7 µg/m3 in 2019. However, the AQI of PM2.5 was 2%, 21% and 27%, 51% in “good”, “moderate” and “unhealthy for sensitive groups” and “unhealthy”, respectively, during the heating period of 2020. Nevertheless, in the 2017 heating season, 3% were “moderate”, 6% were “unhealthy for sensitive groups”, 70% were “unhealthy”, 21% were “very unhealthy”, and 1% were “hazardous”.

Fig. 3
figure 3

a Boxplot of PM2.5 concentration during the heating season; b AQI of PM2.5 in the heating season

Full size image

Figure 4 illustrates a distribution of numerical data, skewness and an average of PM2.5 for the semi-heating season from 2017 to 2020. The maximum value was in the 2018 semi-heating season and was 45.6 ± 27.8 µg/m3, and the minimum value was in 2020 and was 25.6 ± 9.9 µg/m3. In the semi-heating season of 2018, 1% of AQI was in the “very unhealthy” category, while this category was not in 2020. The calculated AQI categories were as follows: 2% “good” category, 42% “moderate” category, 30% “unhealthy for sensitive groups” category and 25% “unhealthy” category, in the semi-heating season of 2017. However, the AQI categories have been estimated in 8% as “good”, 76% as “moderate” and 16% as “unhealthy for sensitive groups” in 2020.

Fig. 4
figure 4

a Boxplot of PM2.5 concentration during the semi-heating season; b AQI of PM2.5 in the semi-heating season

Full size image

Figure 5 shows PM2.5 concentration of non-heating season from 2017 to 2020. In 2017, regarding PM2.5 parameter is 7.7 ± 14.6 µg/m3, 7.9 ± 14.7 µg/m3, 5.9 ± 13.9 µg/m3 and 5 ± 10.5 µg/m3 in 2017, 2018, 2019 and 2020, respectively, in non-heating period. During the non-heating season of 2017, 44% of the AQI was in the “good” category, but in 2020 it increases to 74%. However, AQI of PM2.5 was in the "moderate" category which is 54% and dropped to 26% in the non-heating season of 2020. The comparison of PM2.5 concentrations between 2019 and 2020 shown in Fig. 6 illustrates the difference in the usage of raw coal and improved fuel.

Fig. 5
figure 5

a Boxplot of PM2.5 concentration during the non-heating season; b AQI of PM2.5 in the non-heating season

Full size image
Fig. 6
figure 6

PM2.5 concentration of Ulaanbaatar in January 2019 and 2020

Full size image

Figure 6 shows the total measurements of PM2.5 in the most polluted month (January) of 2019 and 2020. According to the study, PM2.5 concentration reached 1100 µg/m3 several times in 2019, which means 73 times more than the suggested level by the WHO air quality guidelines (15 µg/m3 24-h mean) (World Health Organization 2021). However, in 2019, the maximum concentration of PM2.5 decreased by 2 times, reaching 540 µg/m3. The mean concentrations of PM2.5 were 184.6 ± 172.3 µg/m3 and 88.9 ± 72.8 µg/m3 in January 2019 and 2020, respectively.

Comparison of PM2.5 between Ulaanbaatar city and Khovd city

The result was based on the measurement data of PM2.5 concentrations in Ulaanbaatar and Khovd from January to September 2020. The PM2.5 data was interrupted due to technical conditions from October to December. To show the comparison of PM2.5 in Ulaanbaatar and Khovd city, the mean concentration, minimum and maximum measurement, and standard deviation were calculated (Table 2).

Table 2 Comparison of PM2.5 monthly concentrations between Ulaanbaatar city and Khovd city, in 2020
Full size table

Table 2 compares the PM2.5 concentrations of Ulaanbaatar and Khovd city from January to September in 2020. During the heating season from January to February, the average PM2.5 concentration exceed the WHO air quality guidelines, which is 129 ± 57 µg/m3, 71.1 ± 32.9 µg/m3, respectively, in Ulaanbaatar city and 175.6 ± 59 µg/m3, 101.2 ± 47.3 µg/m3, respectively in Khovd city. In comparison with the non-heating period, the PM2.5 concentration was 10.1 ± 3.9, 7.3 ± 1.7, 7.6 ± 3.4, 9 ± 5.1 and 12.1 ± 4.5 µg/m3 of May, June, July, August and September in Ulaanbaatar city. However, the measurement of PM2.5 was 3 ± 5.9, 1.1 ± 1.1, 0.6 ± 0.9, 2.1 ± 6.8 and 12.7 ± 13.2 µg/m3 in May through September, respectively in Khovd city.

In Fig. 7, the measurement of PM2.5 is illustrated day by day from January to March in 2020. The outdoor PM2.5 concentration was extremely high, and every day of January and February exceeded the WHO air quality guidelines. Air quality has improved dramatically since March 19 so that March 11 and March 14 did not exceed WHO air quality recommendations.

Fig. 7
figure 7

a Khovd city, b Ulaanbaatar city, the daily mean concentrations of PM2.5 from January to March, 2020

Full size image

On the other hand, the total measured 26 days in January 2020 exceeded the 50 µg/m3 in Ulaanbaatar, while in February, only one day out of 29 days was below 15 µg/m3. The PM2.5 concentrations of March dropped sharply; however, 2nd, 25th and 30th of March met the air quality recommendations of WHO (Fig. 7).

Factors affecting PM2.5 concentrations of Khovd and Ulaanbaatar city

The highest PM2.5 concentration of Khovd and Ulaanbaatar city occurred in January, which needs to be analyzed further to determine the contributing factors. Therefore, the working day (Monday, 6th of January) and weekend day (Sunday, 5th of January) were selected from the January data of Ulaanbaatar and Khovd city in order to explore differences between weekends and weekdays (Figs. 7, 8). The colored area of the figure shows a peak of the PM2.5 concentration, and it can conduct the frequency with which it increases and decreases over time.

Fig. 8
figure 8

Sunday, 5th January of PM2.5 concentration in Khovd and Ulaanbaatar city (the colored area indicates an increase of PM2.5)

Full size image

Figure 8 illustrates the PM2.5 concentration of Ulaanbaatar and Khovd as of January 5, 2020. On Sunday or 5th of January, the PM2.5 rose between 00:00 and 03:20, 07:37—12:15, and 17:50—00:00 in Khovd city. At 9:30 in Khovd, the PM2.5 reached 1065.4 μg/m3, or the maximum concentration, 70.9 times higher than the WHO air quality guidelines. The increase of PM2.5 was painted orange in the figure during 01:40—02:45 and 22:40—00:00 in Ulaanbaatar. The maximum concentration is 232.4 μg/m3 in Ulaanbaatar at 23:50, which is 15.5 times higher than the WHO air quality guidelines.

Figure 9 presents the PM2.5 concentration in Ulaanbaatar and Khovd as of January 6, 2020. On a working day, Monday, the colored area showed that the PM2.5 concentration was excessive between 00:00—05:50 and 07:30—12:30 in Khovd city. On this day, the maximum concentration was 1290.8 μg/m3, which is 86 times higher than the WHO air quality guidelines. In Ulaanbaatar, PM2.5 increased from 00:00 to 03:00 and from 19:10 to 00:00, reaching a maximum of 302.4 μg/m3, or 20.2 times higher than the WHO air quality guidelines.

Fig. 9
figure 9

Monday, 6th January of PM2.5 concentration in Khovd and Ulaanbaatar city (the colored area indicates an increase of PM2.5)

Full size image

The firing hours of ger residential areas coincide with the increased period of PM2.5 concentrations in Khovd; therefore, those residential households are considered to be a major source of the pollution.

Discussion

The concentration of PM2.5 in Ulaanbaatar and Khovd city

In Ulaanbaatar from 2017 to 2020, the average annual concentrations of PM2.5 were 54.8 ± 49.3, 56 ± 51.8, 49.5 ± 52.4 and 30.2 ± 24.7 µg/m3, respectively. Both natural (meteorological and topographical condition) and anthropogenic (excessive amounts of biomass burning, transportation, etc.,) factors have influence on PM2.5 pollution. Ganbat and Baik (2016) suggested that the persistent temperature inversion layer is deep in the valleys and shallow over the mountain slopes in Ulaanbaatar. Local winds, which include urban breezes, mountain slope winds and up-and down-valley winds, were found weak in the presence of the temperature inversion using the mesoscale model coupled with the advanced urban canopy model. By calculating the temperature inversion layer thickness and intensity of Ulaanbaatar for 2008–2016 every winter, a recent study has found a strong correlation between the temperature inversion intensity and PM2.5 concentrations (Wang et al. 2018). According to the study, the concentrations of PM10 and PM2.5 were positively correlated with temperature and relative humidity, respectively, and strongly but negatively correlated with wind speed; wind speed and relative humidity were two key factors affecting the distributions of PM2.5 and PM10 concentration (Zhao et al. 2014). Considering the relationship between PM2.5 level and meteorological factors, geographic conditions need to be studied further to provide more accurate scientific evidence in order to support effective policy actions for reducing air pollution levels.

PM2.5 concentration levels compared between Khovd and Ulaanbaatar city from January through September of 2020. During the heating season (months of January and February), high concentrations of PM2.5 occurred and continued for a considerably long period. The government of Mongolia has banned fossil fuels in Ulaanbaatar since May 2019, whereas raw coal consumption is still acceptable without any restrictions in Khovd city. The PM2.5 concentrations of Ulaanbaatar during the heating season of 2019–2020 have declined since the beginning of this regulation.

Comparison of the PM2.5 levels in Beijing and Ulaanbaatar cities and policy actions

Among developing countries in Asia, China is one of the countries struggling with air pollution, and the city of Beijing faces similar conditions with Ulaanbaatar city. Table 2 shows the PM2.5 levels of Beijing and Ulaanbaatar from 2011 to 2020. Both cities have been pursuing policies to reduce air pollution since 2013. According to Table 2, PM2.5 concentrations of Beijing were lower than Ulaanbaatar from 2011 to 2020, while PM2.5 concentrations of Ulaanbaatar were lower in 2020.

According to Table 3, PM2.5 concentration in Beijing was lower than in Ulaanbaatar in 2011–2012, 2016, and 2018–2019, while PM2.5 concentration in Ulaanbaatar was lower in 2013–2015, 2017 and 2020.

Table 3 PM2.5 concentration of Beijing and Ulaanbaatar from 2011 to 2020. Source: (Statista 2021) PM2.5 of Beijing, China, (Enkhjargal et al. 2020) PM2.5 of Ulaanbaatar, Mongolia
Full size table

Air pollution has now become an increasing source of environmental degradation in the developing countries of East Asia. China, in particular, is experiencing dramatic levels of aerosol pollution over a large portion of the country due to the rapid push to industrialize (Alles 2009). The government of China has been taking major actions to control PM2.5 since 2013. These actions include phasing out small and polluting factories, upgrades on industrial boilers, promoting clean fuels in the residential sector, and so on (Zhang et al. 2019). Since 2013, urban and rural households in Beijing have been provided with more subsidies to expedite the replacement of coal with electricity or natural gas. As of 2017, the plurality of Beijing households has become coal-free except for some in remote rural areas. Although enormous actions were taken, approximately 53.4% of all 338 municipal cities in China still did not meet the National Ambient Air Quality Standard in 2017 (Report on the State of the Ecology and Environment in China 2019). Particularly, PM2.5 concentrations decreased by an average of 23% during this period for all municipal cities in China (Lu et al. 2020).

In order to combat air pollution in Ulaanbaatar, numerous energy and policy actions have been introduced to ger residential districts, such as replacing traditional heating stoves with advanced models, substituting raw coal-fueled boilers with centralized heating stations, implementing affordable low-cost housing projects and renovating old roadways and intersections with much better technology and designs. Stove replacement was not successful which unfortunately required a special fuel which the marketplace could not fully provide. The electricity charges and zero-cost policy for electricity from 21:00 to 06:00. is not practical (Koo et al., 2020). According to the National Program for Reducing Air and Environmental Pollution (NPRAEP), 20,000 households had to be reinstated into apartments in 2019 (Resolution of the Government of Mongolia 2017). However, a total of 10,287 apartments were commissioned in 2013–2020 (Implementing agency of the Capital City Governor 2020). Since 2017, the legal and policy environment on reducing air pollution has been further strengthened, but outcomes are still unacceptable.

To reduce air pollution in Khovd city, a briquette plant was planned to be built in 2021, but construction has not begun yet (The Government of Mongolia 2020). On the other hand, infrastructure improvements and redevelopment of ger areas were planned (Asian Development Bank 2021). Byambajav Ganzorig et al. (2021) studied that in November through February of 2019–2020, PM2.5 concentration rapidly decreased by 44% compared to the same period of previous years in Ulaanbaatar (Byambajav et al. 2021). This study shows a positive result of the use of improved fuel instead of raw coal in Ulaanbaatar city. Hence, this quality of improved fuel needs to be implemented in other large cities, such as Khovd. Although progress is underway, there is still a long way to go for Mongolia to optimize its economic and energy structure.

Conclusion

Khovd does not have a PM2.5 measurement station, in another words, there is no government oversight. Therefore, this study will be the first and baseline data for long-term measurement of PM2.5 in Khovd. Based on the results, a comparative study of PM2.5 in Ulaanbaatar for 2017–2020 shows that PM2.5 concentration in 2020 is the lowest yet exceeded the WHO air quality standards. According to a comparison of Khovd and Ulaanbaatar city, average PM2.5 concentration was 129 ± 57 µg/m3 in Ulaanbaatar and 175.6 ± 59 µg/m3 in Khovd in January, the coldest month. In June, the warmest month, the PM2.5 concentration in Ulaanbaatar was 7.3 ± 1.7 µg/m3 and 1.1 ± 1.1 µg/m3 in Khovd. However, the PM2.5 concentration of Khovd is higher than in Ulaanbaatar. Also, it continues to be high concentration from 8 o’clock to 12 o’clock, which is the period of active burning of raw coal in ger areas.

This study was conducted for monitoring purposes, and further research is required. In terms of population concentration, climatic factors and geological formations, Ulaanbaatar should be more polluted, but Khovd's coldest month, January, is more polluted than Ulaanbaatar. Thus, it is suggested that similar monitoring surveys need to be conducted in other cities as well and prompt actions need to be taken in the near future.