Ulaanbaatar, Mongolia is the coldest capital city in the world, with average winter low temperatures of −20˚ Celsius. Many families there live in gers, traditional Mongolian dwellings consisting of a wooden frame beneath several layers of wool felt. In the ger districts of Ulaanbaatar, cooking and heating energy is provided through indoor coal combustion in metal stoves with chimneys, and in wintertime, such stoves may be in use both day and night. The stationary and mobile sources of air pollution have polluted the neighboring areas and also the smokes from the ger areas surrounding the city have moved from uptown to downtown through air flows of the mountains and the rivers during wintertime and remained in the valley and became a cause of the cumulative air pollution. By 2018, more than 220,000 households are living in the ger area in Ulaanbaatar [5].

According to the ICD-X, all types of asthma coded as J45.0—Predominantly allergic asthma; J45.1—Non-allergic asthma; J45.8—Mixed Asthma and J45.9 - J46.0—Unspecified asthma were used for data collecting.

In Ulaanbaatar 15 air quality monitoring stations are measure air pollutant including PM 2.5 µg/m³, PM 10 mg/m³, SO₂ mg/m³, NO₂ mg/m³, CO mg/m³, O₃ μg/m³. As well from the meteorological factors were used max, min and average temperature, humidity and wind speed. We get data from all stations which measure those indicators. Health and air pollution data will be collected in a routine based on related agencies. In Ulaanbaatar, 15 air quality monitoring stations are operating and out of the 12 had an automatic operating system. Thus, we used those 12 station’s data for analyses of this article. All stations monitoring above mentioned indicators’ were measured every 15 - 20 minutes daily. There are located in multisite places over the Ulaanbaatar city based on mobile and static pollutant’s sources. The calibration of equipment is squared in accordance with the national standard MNS-17025.

Data were analyzed by using the program STATA-12. For testing the differences of the results were used appropriate non-parametric tests were used the Kruskal-Wallis one-way analysis of variance and Mann-Whitney U test as well. Furthermore, time-series cross-over and correlation analyses will be used for evaluating the health effects of air pollutants.

Furthermore, Spearman correlation analyses were used for evaluating the asthma hospital admission effect of air pollutants and meteorological parameters. Kolmogorov-Smirnoff test has been used to check the normality distribution of data.

To evaluate the lag effect in detail, we computed 24-h means starting from various times: the 0-h lag starts from 00:00 hours on a corresponding day, the 6-h lag starts from 18:00 hours on the previous day, and the 12-h lag starts from 12:00 hours on the previous day, …, etc., and they last for 24 h. The 6-, 12-, 18-, 72-h, 96-h lags were calculated and used to get the estimated effects. The 0-h lag is the pollutant value on the same day, the 24-h lag is the 1-day lag value, the 48-h lag is the 2-day lag value, etc. In order to allow for different lag structures for different causes of disease, the traditional method, and the Gausson regression analysis were supplied for model comparison. We compared the three time-series models for lags of 0, 1, 2 and 3 days.

Based on exposure-response functions from Ulaanbaatar and from the literature health impacts are derived using the equation below.

E = ( ( R R − 1 ) / R R ) * f p * P O P (1)

where E is the number of cases of each health endpoint attributed to air pollution (“excess cases”), RR is the relative risk of asthma admission between two levels of pollution (here the current level and a lower level obtained from an intervention or the lower threshold level), this is the current incidence rate of the asthma hospital admission, and POP is the exposed population considered. In the following, we assume POP is the total population in Ulaanbaatar, which was 1.5 million in 2017. For asthma hospital admissions, we replace f_p * POP with the actual annual number of hospital admissions.

RR is given by:

R R = exp ( β * ( C − C t ) ) (2)

where β is the exposure-response coefficient, C is the current pollution level and Ct is the target pollution level obtained from an intervention or the assumed threshold value.

The mean level of SO₂ during the cold season was 35.22 µg/m³ [CI 95%: 33.78 - 36.68] and during warm season it was 4.65 µg/m³ [CI 95%: 4.43 - 4.86 which is 7.57 times higher than the warm season. If compared to those levels with national standard and WHO recommended level, 1.8 times high during the cold period of time and 4.3 times less during the warm season (Figure 1(a), Table 1).

According to the NO₂ level, the daily mean was 45.39 µg/m³ [CI 95%: 44.56 - 46.19]. During the holiday (45.85 µg/m³ [CI 95%: 44.92 - 46.83]) mean concentration of NO₂ was significantly higher than the ordinary day mean concentration (44.34 µg/m³ [CI 95%: 42.93 - 45.73]). During the cold period (52.95 µg/m³ [CI 95%: 51.96 - 53.99]) the mean concentration of NO₂ was 1.66 times higher than the warm season (31.25 µg/m³ [CI 95%: 30.60 - 31.89) (Figure 1(b), Table 1).

The 24 hours average РМ 2.5 concentration of cold season was 114.10 µg/m³ [CI 95%: 109.16 - 119.04]. Regarding the 24 hours PM 10 concentration, during cold season daily average concentration was 226.77 µg/m³ [CI 95%: 218.30 - 235.25]. The 24 hours average РМ 2.5 concentration of cold season was higher than the warm season 3.9 times (Figure 1(c), Figure 1(d), Table 1).

Monitoring of carbon monoxide (since January 2010) and ozone (since January 2012) are not exceeded national standard permissible and WHO recommended level. The 8 hours average daily carbon monoxide concentration (1352.85 µg/m³ [CI 95%: 1313.07 - 1396.15]) was high during the cold season, ground-level Ozone concentration (39.10 µg/m³ [CI 95%: 37.95 - 40.35]) was high during the warm season (Figure 1(e), Figure 1(f), Table 1).

Daily mean temperatures ranged from −7.21˚C to 16.40˚C. The mean wind velocity was 2.59m/s [95% CI: 2.53 - 2.66] during the warm season and 1.89 m/s [95% CI: 1.84 - 1.94] during the cold season. For the whole year, average temperature was 1.02˚C [95% CI: 0.46 - 1.60], wind velocity was 2.14 m/s 95% CI: [2.09 - 2.18], relative humidity was 56.14% [95% CI: 55.54 - 56.75]. The relative humidity was varied from cold to warm-season 51.74% [95% CI: 50.65 - 52.88] to 51.48% [95% CI: 57.79 - 59.23] respectively (Table 1).

Table 1 shows the daily mean level of air pollutants, some meteorological factors and asthma hospital admission by different seasons and dows of the day.

There are 8283 asthma cases were recorded during study years in Ulaanbaatar. 68.2% of them were women, 80.0% of them were aged over 40 years old and 54.9% of them were registered during the winter and spring period of time. In one day 3.85 [95% CI: 3.70 - 4.00] people admitted due to asthma (Table 1). The admission cases of asthma were registered during a cold period of time was 2.6

times higher than the warmer time (Т = 19.46, p < 0.0001). Also during the weekdays’ admission cases of asthma were registered 4.6 times higher than the holidays (T = 22.76, p < 0.0001). Mean age of asthma patient were 51.10 [50.76 - 51.44] and was not observed significant difference in seasonal and daily difference on age (Table 2).

All correlation except ground-level ozone and asthma admission during the whole year was statistically significant. The general tendency of the whole year, all air pollutants apart from ground-level ozone had defined direct weak correlated with asthma hospital admission. The ground-level ozone had reversely correlated but it was not statistically significant with asthma admission. During the cold season, ambient air SO₂, mg/m³, PM 2.5 µg/m³, and CO, mg/m³ concentrations were directly weak correlated with asthma hospital admission. While, during warmer time ambient air PM 2.5 µg/m³, and O₃, µg/m³ concentrations were directly weak correlated with the asthma hospital admission.

In general, an influence on the asthma admission of meteoroidal factors is reversely correlated. Only relative humidity has directly correlated with asthma. During the cold and warmer season, strong wind decreasing asthma admission. When during warmer season humidity is directly correlated with asthma admission. But all of the correlations were weak (Table 3).

Significant associations were found for SO₂ in all lags. For NO₂ was less associated with asthma only the first day it has a significant association. For PM 10 it was significant only Lag 0 and Lag 3, for PM 2.5 all lags except Lag 1, for CO also in all lags except lag 2. According to the O₃, the first 2 Lags were observed in significant relation.

Every 10 mg/m³ increase in SO₂ was associated with 0.9% [95% CI: 0.3% - 1.7%; p = 0.008] and 1.7% [95% CI: 1.0% - 2.4%; p < 0.0001] increase in estimated risk of asthma admission. For the NO_2, every 10 μg/m³ increase was associated with 2.3% [95% CI: 0.7% - 3.8%; p = 0.004] increase in estimated risk of asthma admission.

According to the PM 10 and PM 2.5, every 10 μg/m³ increase in of their concentration in ambient air associated with 3.6% [95% CI: 2.9% - 10.0%; p < 0.0001] to 6.1% [95% CI: 6.0% - 16.0%; p < 0.0001] and 3.1% [95% CI: 0.4% - 5.8%; p = 0.024] to 4.4% [95% CI: 2.0% - 6.8%; p < 0.0001] increase in estimated risk of asthma admission respectively.

**Correlation is significant at the 0.01 level (2-tailed). *Correlation is significant at the 0.05 level (2-tailed).

Every 10 μg/m³ increase in ozone was associated with 2.3% [95% CI: 0.7% - 3.9%; p < 0.0001] and 3.2% [95% CI: 1.5 - 5.1; p < 0.0001] increase in estimated risk of asthma admission.

While every 10 mg/m³ increase in CO was associated with 2.5% [95% CI: 0.4% - 7.0%; p.0.005] and 3.2% [95% CI: 1.5 - 5.1; p < 0.0001] increase in estimated risk of asthma admission (Table 4).

Nowadays, the anthropogenic impact on the atmosphere has been increasing with the enormous growth of the human population, the use of mineral resources, and the advanced technology and continues to be significant.

Poor air quality in Mongolian cities is mainly caused by coal combustion in ger areas, thermal power plants, and motor vehicle emissions. Mongolians are burning coal for heating and cooking purposes. Whereas, other countries such as tropical and subtropical countries burn only for cooking purposes at the household level. Besides major air pollutant sources related to human activities in urban areas, especially in Ulaanbaatar, ecological factors, seasonal characteristics, and socio-economic factors have influenced the increase of air pollution. This is similar to China and some other coal consuming countries. This is different from tropical countries like Malaysia, Thailand, and the Philippines, where their main sources come from the forest fire, dust, and combustion of fossil and other fuels by industries, households, and vehicles [6].

Many researchers noted that during the wintertime, inverse cyclone flows come to Mongolia from Central Asia, causing temperature inversion to form in the city’s atmosphere. The meteorological agency concluded that the temperature inversion was 500 to 800 meters of depth and 5 to 170 units of intensity. Due to this temperature inversion, dispersion of air pollutants and substances is slowed down and the smog over the city becomes stagnant. The geographical location and meteorological condition are contributing to the temperature inversion during the winter months. Extreme air pollution episodes were reported for the Muese Valley, Belgium, in 1930; Donora, PA, and the Monongehela River Valley in 1948; and London in 1952. In each of these cases, a persistent (3 to 6 days) temperature inversion combined with significant industrial and, in the case of London, domestic pollutant emissions, resulted in high ground-level concentrations that caused acute illness and, in some cases, death among the exposed populations [7].

The Air quality daily data compiled by this study indicated that the ambient air pollution levels in Ulaanbaatar was still relatively low, whereby the mean daily concentration being 24.56 mg/m³ for SO₂, 45.39 mg/m³ for NO₂, 182.73 µg/m³ for PM 10, 82.72 µg/m³ for PM 2.5, 1041.96 mg/m³ for CO, 29.42 mg/m³ for O₃ in Ulaanbaatar. However, during the peak burning time which is from mid-autumn to mid-spring concentrations of the pollutants increased greatly.

In the 1980s, a few air pollution episodes related to long-range transport from the east occurred in Western Europe. In 1985, increases in mortality and hospital admissions in Germany and temporary lung function changes in the Netherlands were seen during one such episode [7]. Attention shifted to weather-driven, day-to-day variations in air pollution over long periods of time as determinants of day-to-day variations in mortality, hospital admissions, and other public health indicators [8]. Such time-series studies were sporadic until about 1990, (Ostro 1984 [9]; Schwartz & Marcus 1990 [10]) but have increased exponentially since then.

According to this survey air pollution has been increasing intensively during the last decade in big cities of Mongolia especially in Ulaanbaatar due to urbanization, poor urban plan, migration from rural areas to urban areas, poverty, unemployment, economic crises, exercise with unclean technology mainly coal for heating and cooking, raise of number of motor vehicle and other air pollution sources; as result of this phenomena, population health has been worsening. The modest contribution of stationary sources is caused by the high concentration due to the emissions from large power plants, which have substantial stack heights, within the study area [11].

Traffic-related air pollution is currently a growing concern, and NO₂ may serve as a surrogate for traffic-related combustion products [12]. In Ulaanbaatar, NO₂ level, daily mean was 45.39 µg/m³ [CI 95%: 44.56 - 46.19] which was 4.56 µg/m³ higher from the permissible level of national standard and WHO recommendation level. Also during the holiday (45.85 µg/m³ [CI 95%: 44.92 - 46.83]) mean concentration of NO₂ was significantly higher than the ordinary day mean concentration (44.34 µg/m³ [CI 95%: 42.93 - 45.73]). During the warm season, 0.2% of SO₂ concentration and 12.0% of NO₂ were exceeded from the permissible level. Whereas during cold season 58.8% of SO₂ concentration and 55.0% of NO₂ were exceeded from the permissible level.

The 24 hours average РМ 2.5 concentration of cold season was 114.10 µg/m³ [CI 95%: 109.16 - 119.04]. Regarding the 24 hours PM 10 concentration, during cold season daily average concentration was 226.77 µg/m³ [CI 95%: 218.30 - 235.25]. The 24 hours average РМ 2.5 concentration of cold season was higher than the warm season 3.9 times, 2.3 times higher than the national standard level and 8.6 times higher than the WHO recommended level. Whereas, 24 hours PM 10 concentration was 2.4 times higher than the warm season, 3.7 times higher than the national standard level and 9.1 times higher than the WHO recommended level.

Monitoring of carbon monoxide (since January 2010) and ozone (since January 2012) are relatively new indicators of air quality monitoring in Mongolia. Those two pollutants are not exceeded the national standard permissible and WHO recommended level. The 8 hours average daily carbon monoxide concentration (1352.85 µg/m³ [CI 95%: 1313.07 - 1396.15]) was high during the cold season, ground-level Ozone concentration (39.10 µg/m³ [CI 95%: 37.95 - 40.35]) was high during the warm season.

Mongolian weather condition is very harsh. The causes of air pollution episodes depend on several factors including local and regional emissions, topography and meteorological parameters including wind speed, wind direction, temperature, and solar radiation. The relative importance of such factors obviously depends on the climatic region and the season of the year. Examples have been cited from London, Oslo, Helsinki, and Milan. For instance, for London, the frequency of low wind speeds (<2 m/s) can be typically 25%. Such conditions can lead to local episodes of air pollution [13].

In northern Europe, the lowest ambient temperature is in most cases encountered during meteorological situations with a stable atmospheric stratification. A stable atmospheric stratification with low wind speeds or calm conditions is most unfavorable in terms of air pollution. Regarding air pollution episodes of PM 10 and NO₂, elevated atmospheric pressure is commonly necessary, but not a sufficient condition, for the occurrence of an episode [14].

The general tendency of the whole year, all air pollutants apart from ground-level ozone had defined direct weak correlated with asthma hospital admission. During the cold season, ambient air SO₂, mg/m³, PM 2.5 µg/m³, and CO, mg/m³ concentrations were directly weak correlated with asthma hospital admission. Low wind speed is statistically significantly related to stable atmospheric stratification. On the other hand, the re-suspension of particles from road and street surfaces increases with increasing wind speed. Hollander et al. (1999) [15] observed qualitatively the same dependency for particle number concentrations in the city of Kuopio in Finland the seasonal variations of air pollution indexes and the relevant statistical properties from autumn 1999 to summer 2003 were seen in Hong Kong. Such phenomena are mainly affected by the Hong Kong weather cycle with heavy rainfall and southeast wind dominant during the summer, dry and cool air and northwest wind dominant in winter [16].

Findings of the survey by Enkhjargal A. et al [17] shows the percentage of combined effects on respiratory diseases in the winter season is 52.9 for average temperature, humidity, nitrogen dioxide, and РМ 10; the percentage of combined effects on cardiovascular diseases in the winter season is 37.2 for average temperature, humidity, nitrogen dioxide, and РМ 10; the percentage in the autumn is 5.4 for РМ 10 only, and the percentage in the summertime is 17.4 for wind speed and average temperature, while in the autumn the percentage of combined effects of sulfur dioxide, nitrogen dioxide, and PM 10 is 25.4.

Epidemiological data do not show any evidence of excess asthma mortality associated with SO₂ exposure in urban areas. Exposure data indicate that peak SO₂ concentrations greater than the 1.34 µg/m³ may cause mild to moderate asthmatic individuals residing in the vicinity of certain sources to be at increased risk [18].

Wong et al. (2002) [19] conclude that for respiratory admissions (≥65 years of age), significant positive associations were observed PM 10, nitrogen dioxide, sulfur dioxide, and ozone in 2 cities. These associations tended to be stronger at shorter lags in Hong Kong and at longer lags in London. Associations were stronger in the cool season in Hong Kong and in the warm season in London, periods during which levels of humidity are at their lowest in each city. For cardiac admissions (all ages) in both cities, significant positive associations were observed for PM 10, NO₂, and SO₂ with similar lag patterns.

The association between increased emergency department visits for asthma and air pollution was investigated in Anchorage, Alaska [20], Reno, Nev [21], and Seattle, Wash [22]. All 3 studies found significant associations between various measures of PM 10 and increases in visits for asthma. In Seattle, a change of 11 µg/m³ in PM 2.5 was associated with a 15% increase in asthma visits. The Anchorage and Reno studies did not find a significant association between emergency department visits for asthma and carbon monoxide (CO); however, CO was associated with asthma visits in Seattle. In the Seattle study, there was a high correlation between 24-hour CO concentrations and PM 10 values (0.74). Other investigators have reported that the peak hourly NO₂ concentration was significantly associated with increased emergency department visits for asthma during winter months in Northern California [23] and with increased emergency department visits and respiratory admissions in Athens, Greece [24]. Peak hourly NO₂ was also associated with increased emergency department visits for asthma during both the winter and summer months in Barcelona, Spain [25].

In this study, every 10 mg/m³ increase in SO₂ was associated with 0.9% - 1.7%, for the NO_2, it was 2.3%, for PM2.5 3.1% - 4.4%, for ozone 2.3% - 3.2%, and for CO it was 2.5% - 3.2% increase of asthma admission.

In Ulaanbaatar, it has been estimated for PM 10 that every 10 mg/m³ increase of air pollutants 3.6% - 6.1% in the estimated risk of asthma admission. The most health-hazardous pollutant in Mongolian is PM 10 level which shows its chemical source is different from other countries, mainly natural characteristics. It was similar to the above mentioned 3 cities (Anchorage, Reno, and Seattle).

The correlation mainly between asthma admissions cases with meteorological parameters is because the cold winter conditions in Ulaanbaatar result in the accumulation of pollutants in the atmosphere. Thus, population exposure to air pollution is increasing significantly during the winter months.

In conclusion, in recent years, air pollution is increasing sharply in Mongolia. The main cause of poor air quality is coal combustion in the ger areas, thermal power plants, and motor vehicles. The associations between air pollution and asthma hospital admission have been well established by many researchers throughout the world ever since the fateful incident of the London Fog of 1952. Thus, as expected this study demonstrated significant correlations between asthma hospital admission with PM 2.5, PM 10, SO₂, NO₂, CO, O₃, and some meteorological parameters. The correlation mainly between asthma admission cases with meteorological parameters is because of the cold winter condition. Thus population exposure to air pollution increases significantly in the winter months.

Our study has several limitations. We estimated only hospital admission data instead of including emergency room visits and outpatient asthma data.