Trends and health impacts of major urban air pollutants in Kazakhstan

ABSTRACT

The air quality in cities in Kazakhstan has been poorly investigated despite the worsening conditions. This study evaluates national air pollution monitoring network data (Total Suspended Particle-TSP, NO₂, SO₂, and O₃) from Kazakhstan cities and provides estimates of excess mortality rates associated with PM_2.5 exposure using the Global Exposure Mortality Model (GEMM) concentration-response function. Morbidity rates associated with PM₁₀ exposure were also estimated. Annual average (2015-2017) population-weighted concentrations were Kazakhstan cities was 157, 51, 29, and 41 μg m⁻³ for TSP, NO₂, SO₂, and O₃ respectively. We estimated a total of 8134 adult deaths per year attributable to PM_2.5 (average over 2015–2017) in the selected 21 cities of Kazakhstan. The leading causes of death were ischemic heart disease (4080), stroke (1613), lower respiratory infections (662), chronic obstructive pulmonary disease (434), lung cancer (332). The per capita mortality rate attributable to ambient air pollution (per 10⁵ adults per year) was less than 150 in nine cities, between 150 and 204 in nine cities, and between 276 and 373 in three industrial cities (Zhezkazgan, Temirtau, and Balkhash).

Implications: Quantitative information on the health impacts of air pollution can be useful for decision-makers in Kazakhstan to justify environmental policies and identify policy and funding priorities for addressing air pollution issues. This information can also be useful for policymakers by improving the quality of government-funded environmental reports and strategic documents, as they have many shortcomings in terms of the selection of air quality indicators, identification of priority pollutants, and identification of sources of pollution. This study has high significance due to the lack of data and knowledge in Central Asia, especially Kazakhstan.

Introduction

Air pollutants impact the climate system and hydrological cycle (Ramanathan et al. 2001), human health (Cohen et al. 2017), and agriculture (Chameides et al. 1999). In terms of the human health impact, exposure to ambient air pollution increases mortality and morbidity as a result of increased risk of ischemic heart disease (IHD), cerebrovascular disease, lung cancer (LC), chronic obstructive pulmonary disease (COPD), and lower respiratory infections (LRI) (Cohen et al. 2017).

The environmental situation in Kazakhstan has been suffering under the pressure of rapid economic development in recent decades due to the inadequacy of the existing environmental monitoring and control systems (Russell et al. 2018). In 2011, the concentrations of the major pollutants PM₁₀, NO₂, and SO₂ exceeded the European Union (EU) annual limit values in 10 out of 11 selected cities (World Bank 2013). The prevalence of COPD in Almaty (largest city of Kazakhstan) was 66.7 per 1000 people, which is 1.8- to 2.1-times higher than that in other post-Soviet countries (e.g., Ukraine and Azerbaijan) and is associated with the poor environmental situation (Nugmanova et al. 2018). Public environmental complaints have increased, and some court cases have arisen due to the local authorities’ failure to take necessary measures in the context of environmental safety (Radio Azattyk 2018).

Fossil fuel combustion is the main source of air pollution in Kazakhstan, as half of the total energy consumption of Kazakhstan is attributed to coal use (Kerimray et al. 2018). A major portion of electricity and heat production (66%) is based on coal burning (Karatayev et al. 2017), which emits large quantities of hazardous air pollutants (EIA 2019). Kazakhstan ranks second in the world, with per capita coal consumption equal to 157 kg per person in the household sector (Kerimray et al. 2017). Transportation is another major source of atmospheric emissions, as almost 63% of passenger cars do not satisfy the requirements of the Euro 4 emission standard (UNDP 2017). The quality of oil products is also an issue, with frequent cases of noncompliance with fuel quality standards and the late introduction of Euro 4 standard in 2018 (Kazenergy 2017).

Natural sources of air pollution, such as windblown dust and forest fires, can also contribute to the deterioration of the air quality in the cities of Kazakhstan. Dust and sandstorms are distributed unevenly in Kazakhstan, with the southern desert zone of Kazakhstan being more affected by dust storms from the Pre-Aral Karakum, Aralkum, Kyzylkum, and southern Pre-Balkhash deserts mainly in the spring and summer seasons (Issanova and Abuduwaili 2017). Rupakheti et al. (2019) suggested that clean continental aerosol is a major aerosol type over Central Asia and Kazakhstan, with a very low contribution of dust storms to the overall aerosol loading in Kazakhstan. This could be because Kazakhstan’s population density is among the lowest in the world (6 people per km² of land area), and most of its territory is unpopulated, especially in areas with sandy and sandy-loamy soils. The study of the NO₂ column in Kazakhstan revealed that NO₂ density was significantly lower in rural areas, which account for the majority of Kazakhstan, than in urban areas: by a factor of 4 in the summer and by a factor of 20 in the winter (Darynova et al. 2018).

Rupakheti et al. (2019) stated that the highest seasonal aerosol optical depth observed from satellites over Kazakhstan was in summer and that the lowest aerosol optical depth was in winter. The higher seasonal aerosol optical depth in the summer was associated with the lower amount of precipitation, high temperatures leading to photochemical reactions (Filonchyk et al. 2019), the formation of secondary aerosols, and the northerly winds transporting dust from the desert areas (Rupakheti et al. 2019).

To date, a very limited number of air quality assessment studies have been conducted in Kazakhstan cities. Previous studies have focused on transportation-related emissions in Almaty (Carlsen et al. 2013); tropospheric NO₂ concentrations from satellite measurements (Darynova et al. 2018); benzene, toluene, ethylbenzene and xylene (BTEX) concentrations and ratios in Almaty (Baimatova et al. 2016); availability and quality of air pollution data in Kazakhstan (Russell et al. 2018); and pollution characterization in Almaty (Nazhmetdinova et al. 2018).

The National Hydrometeorological Service of Kazakhstan (Kazhydromet) compares the measured concentrations to the local air quality guidelines, which differ from the international standards (e.g., World Health Organization (WHO), European Union). National air quality monitoring network (NAQMN) data are only rarely used for the analysis of trends, possible sources, and health impacts. In 2013, the World Bank (2013) conducted an assessment of air quality in eleven Kazakhstan cities using 1 year of data from 2011. It was estimated that PM₁₀ and PM_2.5 air pollution in the studied cities results in a 1.3 USD billion loss in the national economy due to additional health-care costs (0.9% of the GDP). It was concluded that PM pollution causes approximately 2800 premature deaths annually. However, the Global Burden of Disease Study estimated that 10,064 premature deaths occurred in 2010 as a result of ambient air pollution in Kazakhstan (WHO 2015). Ground measurements of PM were not available in many areas of the world (including Kazakhstan) in the Global Burden of Diseases Study 2010; therefore, annual mean PM_2.5 estimates were modeled using a combination of estimates from data from satellites, results from the global chemical transport model and ground measurements (Brauer et al. 2012). Global Ambient Air Quality Database (WHO 2018) does not contain data on PM₁₀ and PM_2.5 concentrations in Kazakhstan cities.

The benefits and costs of air quality improvements are often “hidden” due to the lack of information linking air quality with health outcomes and consequent economic costs (Craig, Dragan, and Fimka 2015). By having a clear understanding of those links, policymakers can focus on a narrow set of air pollution sources to achieve better improvements in public health (Craig, Dragan, and Fimka 2015; Yang et al. 2017). Some studies assess the health impacts of ambient air pollution at the city level (Abe and Miraglia 2016), while others assess these impacts at the national level (Yang et al. 2017) or global level (Rao et al. 2012); nevertheless, there is a common agreement in terms of national legislation that the political agendas of countries should prioritize city-level impact assessments. As is the case in Kazakhstan, there is a clear lack of information linking air quality with health outcomes at the city level. Nazhmetdinova et al. (2018) also highlighted the need for further studies on the optimization of the environment and public health in Kazakhstan.

The global exposure mortality model (GEMM) was first introduced by Burnett et al. (2018) for the global estimate of mortality associated with long-term exposure to ambient fine particulate matter. The new GEMM is based on a larger number of cohort studies and a wider range of PM_2.5 exposure than the integrated exposure response (IER) functions used in the Global Burden of Disease (GBD) assessment in 2015 (Burnett et al. 2018). The IER used in GBD (2015) includes information from nonoutdoor sources (indoor air pollution, active/passive smoking) and assumes equal toxicity per unit dose. The new GEMM relied solely on studies of outdoor PM_2.5, including studies of Chinese men with PM_2.5 exposures up to 84 μg m⁻³ (Burnett et al. 2018). The global mortality attributable to ambient fine particulate air pollution was estimated to be 8.9 million people (95% CI: 7.5–10.3) using the GEMM, while GBD using IER obtained an estimate of 4 million deaths (in 2015) (Burnett et al. 2018). Burnett et al. (2018) concluded that health benefits due to reductions in PM_2.5 concentrations were much greater than those identified in previous studies (e.g., GBD 2015 Risk Factors Collaborators 2016), particularly in areas with high levels of exposure to PM_2.5. The GEMM was recently applied for the assessment of health impacts of ambient air pollution in Tehran (Bayat et al. 2019) and Europe (Lelieveld et al. 2019). Lelieveld et al. (2019) suggested that the health impacts attributable to ambient air pollution in Europe estimated with the GEMM are substantially higher than previous assessments. The previous study on the health impacts of particulate matter in Kazakhstan conducted by the World Bank (2013) employed exposure-response coefficients reported by Ostro (2004) and covered only four regions of Kazakhstan representing 30% of the total population of the country. This study evaluated the air quality in Kazakhstan cities, developed rankings of the most polluted cities of Kazakhstan and discussed potential sources of emissions contributing to the air pollution in those cities based on the available activity data. Our study aims to provide revised estimates of excess mortality rates associated with PM_2.5 exposure by using a novel concentration-response function of the GEMM. This study covers 21 major cities of Kazakhstan with 8.6 million population from the total 17.9 million population of Kazakhstan (in 2017). Large urban cities, administrative centers as well as smaller industrial cities with availability of air quality data were selected in this study. The most populated urban areas and smaller industrial cities were studied, where anthropogenic sources may play a dominant role in air quality deterioration, although no definite conclusion can be made on sources due to the lack of source apportionment studies.

Methodology

National air quality monitoring network (NAQNM)

NAQMN data on air quality were obtained from the Kazhydromet-operated national network of stations. Data are published on a monthly and annual basis in the form of information bulletins on the environmental situation in Kazakhstan. Kazhydromet monitors air quality in 49 cities/towns at 156 observation posts by operating 66 manual sampling, 90 automatic monitoring, and 14 mobile stations (Kazhydromet 2018). Kazhydromet monitors 34 different pollutants, including criteria and priority pollutants. Manual measurements of air quality are carried out in three types of “programs”: i) full (four times a day at 01:00, 07:00, 13:00 and 19:00), ii) incomplete (three times a day at 07:00, 13:00 and 19:00), and iii) shortened (two times a day at 07:00 and 13:00).

Kazhydromet used to monitor only TSP levels before the introduction of PM₁₀ and PM_2.5 monitoring in 2014. Currently, the number of PM₁₀ and PM_2.5 monitoring sites is still insufficient in many cities when compared to the EU standards of the minimum station number according to the population (World Bank 2013). There were no PM_2.5 measurements in five cities (Temirtau, Taraz, Ust-Kamenogorsk, Taldykorgan, and Kostanay) and in 10 cities there was only one PM_2.5 station (Supplementary File, Figure A1). In 21 cities total number of TSP stations was 53, while it was only 27 for PM_2.5 stations. There is a small correlation between the measured TSP values and measured PM_2.5 values, and cities with the highest TSP levels do not have the highest PM_2.5 values (Supplementary File, Fig A.2). This could be due to the different locations of the stations used to measure TSP and PM_2.5, as well as the low coverage of the cities with PM_2.5 stations, differences in frequency of measurements and applied measurement method. There are differences in the methodology for determining TSP and PM_2.5 because TSPs are determined by gravimetric method and PM_2.5 by optical method. Location of TSP and PM_2.5 stations are different in the cities. As an example, for Almaty, in 2017, the PM_2.5 stations were located far from the densely populated city center.

Due to the lack of complete PM₁₀ time series data and inconsistencies in the minimum required a number of measurement points in many cities, the incomplete PM₁₀ concentrations were determined using the PM₁₀/TSP ratio of 0.45 (World Bank 2013).

There is considerable temporal and spatial variability in PM_2.5/PM₁₀ ratios; for example, their median ranged from 0.4 to 0.8 with a median of 0.65 in the UK (Munir 2017), they ranged from 0.18 to 0.86 with a median of 0.58 in Turkey (Tecer et al. 2008). The ratio could also vary during the day depending on the source of the emissions, reaching minimum values (0.60– 0.65) in the traffic hours and higher values (0.75–0.85) during the remaining periods in Barcelona (Querol et al. 2001). The annual average PM_2.5/PM₁₀ values recorded for the entire study period were 0.45 at a roadside site and 0.51 at an urban site in Algeria (Talbi et al. 2018). In the urban area of Lanzhou, Northwest China, the annual average PM_2.5/PM₁₀ ratio was 0.425, and it was associated with dust emissions from deserts (Filonchyk and Yan 2018). Additionally, in Wuhan in Central China, the three-year average PM_2.5/PM₁₀ ratio was 0.62 ± 0.22 at the urban site and 0.68 ± 0.31 at the background site, with a maximum value in the winter: 0.75 ± 0.26 at the urban site and 0.80 ± 0.42 at the rural site (Xu et al. 2017). Ostro (2004) concluded that the PM_2.5/PM₁₀ ratio is 0.50 − 0.65 in urban areas of industrialized nations, and the ratios are likely to be much lower (0.35) in areas impacted by more crustal particles such as arid areas or cities with a significant number of unpaved roads or windy days. In Kazakhstan, there were no previous studies with PM_2.5 and PM₁₀ measurements that could be used to calculate the PM_2.5/PM₁₀ ratio. Some cities located in sandy areas of Kazakhstan (e.g., Aktau, Atyrau, Kyzylorda), as depicted by Issanova and Abuduwaili (2017), may have lower PM_2.5/PM₁₀ ratios, while other densely populated (e.g., Almaty, Nur-Sultan, Shymkent) and industrialized cities (Temirtau, Ust-Kamenogorsk, Karaganda, Pavlodar) may have higher PM_2.5/PM₁₀ ratios. Many cities in Kazakhstan have unpaved roads that can also contribute to the coarse fraction. Ostro (2004) suggests that “In the absence of a local measurement of the ratio, a value of 0.65 could be assumed for developed countries and 0.5 for developing countries.” The World Bank (2013) assumed PM_2.5/PM₁₀ ratios of 0.35, 0.45, and 0.55 for low, moderate, and high estimates, respectively. In this study, the midpoint estimate of the World Bank (2013) PM_2.5/PM₁₀ ratio of 0.45 was assumed with an error of ±22%. Future studies are needed to evaluate the spatial and temporal variations in PM_2.5/PM₁₀ in the cities of Kazakhstan.

Study area

In this study, 21 out of the 49 monitored cities/towns were evaluated. They include all administrative regional centers and represented 47% of the total population of Kazakhstan. Figure 1 presents the map with cities, towns, population numbers, location of heavy industries, and coal-fired power plants. Population numbers (from the beginning of 2018) were obtained from the Committee of Statistics of the Ministry of National Economy of the Republic of Kazakhstan (2018a).

Figure 1. Map with cities, population numbers (in thousands), location of heavy industries and coal power plants

The central, northeastern, and eastern regions of Kazakhstan are centers of heavy industry, accounting for the bulk of energy consumption of the country. In Ust-Kamenogorsk, Zhezkazgan, Balkhash mainly lead, zinc, and copper are produced by the pyrometallurgical method. Ferroalloy plants are located in Aktobe. In Pavlodar, there is aluminum and pipe rolling plant; in Temirtau the largest steel mill is located. In order to provide cheap electricity in those regions (central, northeastern and eastern regions of Kazakhstan), the largest coal-fired power plants are located with the predominant burning of Ekibastuz coal with an ash content of about 40%. Almaty and Nur-Sultan are the most densely populated cities (former and existing capital cities), with the coal power plants. Kostanay, Petropavlovsk and Kokshetau, Taraz are agricultural cities. The cities of Western Kazakhstan and Kyzylorda are oil-producing regions, with oil deposits located at the considerable distance from the selected cities.

Assessment of the health effects

As the greatest quantifiable share of the burden of disease from air pollution comes from PM_2.5, it is widely used as a common indicator (WHO 2015). The public health impact of PM_2.5 and PM₁₀ exposure was estimated using the following steps:

1. The annual average concentrations of TSP (average for 2015–2017) were taken from NAQNM. PM₁₀ and PM_2.5 are fractions of suspended solids, so their concentrations are estimated based on their ratios to TSP (PM₁₀/TSP = 0.45 and PM_2.5/PM₁₀ = 0.45) due to the lack of time series data and inconsistencies in the minimum required number of measurement points of PM_2.5 and PM₁₀ in the cities of Kazakhstan.

2. Data on the age-specific population exposed to PM according to the cities of Kazakhstan were taken from the statistical publication “Population of the Republic of Kazakhstan by sex and separate age groups at the beginning of 2018” of the Committee of Statistics of the Ministry of National Economy of the Republic of Kazakhstan (2018a). The age-specific population by regions and by cities were provided in the Supplementary File (Table A1).

3. Age-specific baseline mortality rates in Kazakhstan by region and city were obtained from the Demographic Yearbook of Kazakhstan of the Committee of Statistics of the Ministry of National Economy of the Republic of Kazakhstan (2018b) (Supplementary File, Table A2). The mortality rates associated with five causes LRI, COPD, LC, IHD, and stroke were obtained from Global Burden of Disease Collaborative Network (2018) (Table 1). Share of non-communicable diseases (NCD) was 86% of all deaths (WHO 2018). The WHO (2018), in its Noncommunicable Diseases Country Profile for Kazakhstan, states that mortality estimates for Kazakhstan contain a high degree of uncertainty since they are not based on national mortality data. A limitation of this study is the uncertainty of the national mortality data for Kazakhstan.

4. The concentration-response function of the GEMM (Burnett et al. 2018) (based on the Chinese Male Cohort Study) was used to estimate the number of deaths attributable to PM_2.5 exposure. Mortality numbers due to NCD and LRI were estimated. In addition, separate estimates of mortality numbers due to the following five causes were estimated: LRI, COPD, LC, IHD, and stroke. For IHD and stroke, age-specific (5-year increments) mortality rates (for adults >25 years) were estimated. For other causes (COPD, LC, and LRI), mortality rates for adults (>25 years) were estimated. The standard error for the θ parameter was obtained from Burnett et al. (2018) (Table A3).

5. PM₁₀ is associated with morbidity primarily due to chronic bronchitis, lower respiratory illness in children, and other respiratory symptoms. To estimate morbidity rates associated with PM₁₀ exposure, exposure-response coefficients (annual cases per 100,000 people) from international studies (Abbey et al. 1995; Ostro 1994; World Bank 2012) were employed (Table A4). A baseline level (natural background concentration) 15 µg m⁻³ for PM₁₀ was applied (World Bank 2012).

Table 1. Share of mortality rate by cause

Cause of mortality	Share of total mortality by cause
COPD	3.54%
LC	2.4%
LRI	2.69%
Stroke	14.56%
IHD	28.65%

The GEMM hazard ratio function (Burnett et al. 2018), which describes the association between annual mean concentrations of ambient PM_2.5 (denoted as z) and mortality in each age group, is described by Equation 1(1) $GEMM\left(z\right)=exp\left\{\frac{\theta \mathrm{l}\mathrm{n}\left(\frac{z}{\alpha }+1\right)}{1+\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{z-\mu }{\nu }\right)}\right\}$(1) :

(1) $GEMM\left(z\right)=exp\left\{\frac{\theta \mathrm{l}\mathrm{n}\left(\frac{z}{\alpha }+1\right)}{1+\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{z-\mu }{\nu }\right)}\right\}$(1)

ϴ, µ, ν and α describe the shape of the mortality-PM_2.5 association and were obtained from Burnett et al. (2018).

The number of deaths (ΔY) attributable to air pollution for each city (and age group for IHD and stroke) was estimated using Equation 2(2) $\mathrm{\Delta }Y=\left(1-\frac{GEMM\left(z_0\right)}{GEMM\left(z\right)}\right)\times Y_0\times POP$(2) (Bayat et al. 2019):

(2) $\mathrm{\Delta }Y=\left(1-\frac{GEMM\left(z_0\right)}{GEMM\left(z\right)}\right)\times Y_0\times POP$(2)

where z is the current PM_2.5 concentration by city (from NAQNM), z₀ is the target PM_2.5 concentration (2.4 μg m⁻³), Y₀ is the age-specific baseline mortality rate by cause, and POP is the age-specific population exposed to air pollution.

Results and discussion

Air quality assessment in the major cities of Kazakhstan

NAQNM data trends of the average annual levels of PM₁₀, NO₂, SO₂, and O₃ concentrations of the cities were analyzed, and they were ranked for the period from 2015 to 2017. Average concentrations were compared with the coal consumption in 2017 according to the Committee of Statistics of the Ministry of National Economy of the Republic of Kazakhstan (2018c) and the number of vehicles in the corresponding region in 2017 according to the Committee of Statistics of the Ministry of National Economy of the Republic of Kazakhstan (2018d). The relationships between the possible sources of emissions and historical variation in each pollutant in the most polluted cities are discussed. NAQNM data were compared with WHO and EU annual limit values (European Commission 2019; WHO 2005). Annual (2015–2017) population-weighted concentrations in Kazakhstan cities were 157, 51, 29, and 41 μg m⁻³ for TSP, NO₂, SO₂ and O₃ respectively.

City ranking of air quality and pollution

The results indicate that the annual WHO limit value of 20 µg m⁻³ for PM₁₀ was not exceeded in only three cities - Kyzylorda, Aktobe, Kostanay (Figure 2), while the annual EU limit of 40 µg m⁻³ was not exceeded only in five cities - Kyzylorda, Aktobe, Kostanay, Petropavlovsk, and Kokshetau. The top three polluted cities were the capital, Nur-Sultan, and two industrial cities: Zhezkazgan and Temirtau. Zhezkazgan has one of the largest copper-processing plants in the country (Zhezkazgantsvetmet), while Temirtau is located near the largest steelmaking industrial complex “Arcelor Mittal Temirtau” JSC. There was no significant correlation (R² = 0.01) between PM₁₀ (2015–2017) and the total coal consumption in the region, which could be due to the effect of the distance to the source and meteorological conditions.

Figure 2. Average annual PM₁₀ concentration between 2015 and 2017 (based on TSP data) and coal consumption in the region

The highest PM₁₀ values are observed in cities (Nur-Sultan, Zhezkazgan, Temirtau, Balkhash, Almaty, etc.), where low-quality coal with high ash content (over 40%) is predominantly burned at the power plants. While in the cities at the East of Kazakhstan (Ust-Kamenogorsk and Semipalatinsk) where better quality coal is used with an ash content of 15–20%, PM₁₀ concentration was significantly lower. There is no coal power plant at Shymkent city, and relatively high PM₁₀ pollution could be caused by the construction industry and diesel vehicles (the city has the highest number of diesel trucks).

The WHO and EU limit values of 40 µg m⁻³ for NO₂ were exceeded in eight cities (Figure 3). The top five most polluted cities were Almaty (84 µg m⁻³), Nur-Sultan (78 µg m⁻³), Taraz (64 µg m⁻³), Kyzylorda (56 µg m⁻³), and Ust-Kamenogorsk (49 µg m⁻³). Almaty and Nur-Sultan are the most polluted cities, and they are also the most populated cities in the country. NO₂ is considered a marker of the number of combustion-related pollutants, particularly those from urban traffic (WHO 2005). The coefficient of determination R² between NO₂ and the number of registered passenger cars is 0.21. However, there is still a potentially strong relationship between these factors since the R² value may significantly increase if other dependent and independent factors are taken into account, such as fleet types, compositions of the registered vehicles, daily mileage, city cycles, and meteorological conditions.

Figure 3. Ranking of the cities by the average level of NO₂ (over 2015–2017) and the number of registered cars in the region

Coal-fired power plants could be another significant contributor of NO_x emissions in the cities of Kazakhstan. Cities such as Balkhash and Ekibastuz have a small population with operating coal-fired power plants. Kyzylorda and Taraz cities are characterized with relatively high NO₂ concentration (64 and 56 µg m⁻³, respectively), although there are no heavy industry enterprises and coal power plants in those cities. Transport, household coal combustion, power plants (gas and fuel oil) may contribute to NO₂ pollution in Kyzylorda and Taraz, and future studies for source identification are needed.

The five cities with the highest SO₂ pollution are Ust-Kamenogorsk (85 µg m⁻³), Kyzylorda (79 µg m⁻³), Almaty (50 µg m⁻³), Balkhash (48 µg m⁻³), and Temirtau (40 µg m⁻³) (Figure 4). The topographical and geographical conditions in Almaty and Ust-Kamenogorsk (location at the foothills of mountains) lead to formation of inversion layers, restricting the vertical dispersion of pollutants, which contribute to severe pollution. High concentrations of SO₂ in Ust-Kamenogorsk, Balkhash, Zhezkazgan could be due to the pyrometallurgical processing of sulfide ores. Due to high sulfur content in coal, its combustion is assumed to be the main SO₂ source in the country; however, no significant correlation (R² = 0.02) between SO₂ and total coal consumption parameters was identified. This needs further investigation by considering other factors that affect the dispersion of air pollutants.

Figure 4. Ranking of the cities by the average concentration of SO₂ (over 2015–2017) and coal consumption in the region

Despite the international practice requiring O₃ monitoring in large cities with more than 300 thousand inhabitants (World Bank 2013), O₃ was measured in only 11 out of the 21 cities considered. The top five most polluted cities were Aktobe (65 µg m⁻³), Ekibastuz (46 µg m⁻³), Aktau (45 µg m⁻³), Ust-Kamenogorsk (44 µg m⁻³) and Shymkent (43 µg m⁻³) (Figure 5). Correlation analysis between O₃ data and the number of registered passenger cars revealed no significant correlation (R² = 0.01).

Figure 5. Ranking of the cities by the average annual concentration of O₃ (over 2015–2017) and the number of registered cars in the region

No significant correlation between any data pairs of PM₁₀, NO₂, and SO₂ concentration levels in Kazakhstan cities was obtained (R² ≤ 0.1) (Figures A.3–A.5). These results are not very surprising since all cities have significantly different urban dynamics and source compositions.

There are cities located in East Kazakhstan, Pavlodar, and Karaganda regions with metallurgical plants and coal power plants, resulting in complex structure of emissions sources. One example is Ust-Kamenogorsk, where non-ferrous metallurgy enterprises, coal power plants, and coal-fired boiler houses are located.

In Kazakhstan, enterprises are required to receive environmental permit (license for pollution). Applications for environmental permits from enterprises undergo an environmental review at the regional departments of the Ministry of Ecology, Geology and Natural Resources of the Republic of Kazakhstan. Permits for emissions are given with the condition that the content of pollutants at the border of the residential zone and the sanitary protection zone is not higher than the values of the maximum one-time national standards on concentration of air pollutants (Ministry of National Economy of the Republic of Kazakhstan, 2015). At the same time the air quality is generally accessed using daily or annual average limit values, which are substantially lower than maximum one-time limits. There are no formal procedures to determine the effect of emissions from an individual enterprise to the average daily pollutant concentration. As an example, according to the dispersion model, the concentration of SO₂ did not exceed maximum one-time limit at the border of the sanitary protection zone (formally meeting the requirements of the legislation), while the average daily concentration still may exceed daily limit in the city. Using maximum one-time national limit allows enterprises to increase their quantities of permitted emissions, since formally there is no responsibility for high air pollution in the city. Coal-fired power plants regularly increase permitted levels of air emissions (Department of Ecology of Almaty 2015; Department of Ecology of Almaty region 2015), which may be one of the main reasons for the high pollution in Almaty, Ust-Kamenogorsk, Nur-Sultan, Pavlodar, etc. Additionally, for power plants as subjects of natural monopolies, discount is applied to their emission charge rates. In 2020, in Kazakhstan new environmental code is expected to be adopted with the new principle of Best Available Technologies (BAT) system. Although it is not yet evident that with the new legislation discounts and easing of some requirements for polluting enterprises will be eliminated.

Pollution trends and potential sources in the most polluted cities

Nur-Sultan is the most polluted city in terms of PM₁₀ (138 µg m⁻³) and the second most polluted city in terms of NO₂ (78 µg m⁻³). Nur-Sultan has been experiencing dramatic increases in the population and the number of transport vehicles with a very fast urbanization rate after it became the capital in 1997. In 2015, the total primary energy consumption in Nur-Sultan was 22 TWh, of which 70% was attributed to coal that was primarily used for heating or CHP facilities (World Bank 2017). The city is still lacking a connection to the gas network and, therefore, has limited gas supplies. According to a household survey, approximately 10% of the total-detached houses use coal for heating (Kerimray et al. 2018), which also contributes to the reduction in air quality. Benchmarking of energy consumption in Nur-Sultan revealed that there is a 30–40% energy saving potential; per capita energy consumption in Nur-Sultan in 2015 was 91 GJ (25,200 kWh/capita), which is high compared to that of similar cities (World Bank 2017). Its annual average growth rate in PM₁₀ concentration was 14% for the period from 2011 to 2017 (Figure 6). The annual average PM₁₀ concentration was 135 µg m⁻³ in 2017, exceeding the limit set by the WHO (20 µg m⁻³) by 6.8 times and the EU standards (40 µg m⁻³) by 3.4 times. The average annual NO₂ concentration increased from 55 in 2011 to 80 µg m⁻³ in 2017, exceeding the limit offset by the WHO and the EU standards (40 µg m⁻³) by two times in 2017. The increasing trends of PM₁₀ and NO₂ can be attributed to the rapidly increasing population from 697,000 in 2000 to 973,000 in 2017, which has consequently led to higher energy demand that has been met by unsustainable energy sources (coal, low-quality oil products). The number of registered transport vehicles in the city increased rapidly, with an average annual growth rate of 7.2% for the period from 2011 to 2017.

Figure 6. Trends in average annual concentrations of PM₁₀ (left) and NO₂ (right) in Nur-Sultan

Almaty is the most polluted city in terms of NO₂ (annual average concentration 84 µg m⁻³), the third most polluted city in terms of SO₂, is the largest city in the country (former capital) with a population of 1.8 million people and the largest vehicle stock of 463 thousand cars in 2017. Due to the location of the city, the wind speed is slow, which leads to a high level of air pollution. In 2015, the total primary energy consumed in Almaty was 42.4 billion kWh, of which coal, natural gas, and oil products each accounted for approximately 30% (World Bank 2017). The efficiency of energy use remains at a low level, as the primary energy consumption per capita in Almaty of 90 GJ (24,907 kWh) in 2015 is high among similar cities in terms of specific annual energy consumption per inhabitant (World Bank 2017). The level of losses in the energy transformation and distribution system is nearly the same as the final energy consumption for the entire residential sector in the city (World Bank 2017). Household coal consumption is another important source of pollutant emissions, as approximately 7% of the total number of households in Almaty use coal (Kerimray et al. 2018).

In Almaty, the annual average growth rate in PM₁₀ concentration was 7% for the period from 2011 to 2017, while the average annual NO₂ concentration decreased by an average rate of 4% during the same period (Figure 7). Decreasing levels of NO₂ concentration may be associated with the gradual renewal of vehicle stocks, improved fuel quality, and the few developments in public transport (UNDP 2017). The stock of vehicles in Almaty is growing at a significantly lower rate than that in Nur-Sultan, with an annual growth rate of 0.7% (compared to 7.2% in Nur-Sultan). There is a gradual renewal of the stock of vehicles, as the share of vehicles, which do not satisfy the requirements of the Euro 2 standard, has decreased from 38–40% in 2009 to 10–13% in 2017, while the share of vehicles following the Euro 5 class requirements has rapidly increased from 0% in 2009 to 25% in 2017 (The Municipality of Almaty City 2018). Additionally, electricity generation at the power plants in Almaty has increased at an annual average rate of 6.3% from 2015 to 2017 (The Municipality of Almaty City 2018). The coverage of the city with central heating has increased gradually from 88% in 2015 to 92.5% in 2016–2017 (The Municipality of Almaty City 2018).

Figure 7. Trends in average annual concentrations of PM₁₀ (left) and NO₂ (right) in Almaty

Ust-Kamenogorsk As the center for a heavy industry employing outdated technology coupled with extreme weather conditions, Ust-Kamenogorsk is the most polluted city in terms of SO₂ (annual average 85 µg m⁻³). Ust-Kamenogorsk is a major mining and metallurgical center of the country, producing nonferrous metals, especially zinc, lead, silver, copper, gold, uranium, beryllium, tantalum, and titanium. The SO₂ concentration fluctuated widely during the period of 2011–2017 (Figure 8). Understanding the reasons behind such fluctuations requires further research. The NO₂ concentration decreased from 89 μg m⁻³ in 2011 to 40 µg m⁻³ in 2016, followed by an increase to 60 µg m⁻³ in 2017 (exceeding the cutoff values of the WHO and EU by 1.5 times). The Municipality of East Kazakhstan Region (2018) also reported that the air pollution index (API) of cities with high levels of air pollution was “9.” The API is calculated as the average concentration of various pollutants divided by the maximum permissible concentration and adjusted for the harmfulness of sulfur dioxide. Its high levels of air pollution can be explained by i) metallurgy industry facilities and CHP plants; ii) low levels of wind speed; iii) transport relying mainly on gasoline and aged vehicles stock (78% of cars more than 10 years old).

Figure 8. Average annual concentrations of PM₁₀ (left) and NO₂ (right) in Ust-Kamenogorsk

Aktobe is the most polluted city in terms of O₃ (65 µg m⁻³). Aktobe is an industrial city with factories making ferroalloys and chromium compounds. Ozone is produced as a result of a complex set of reactions that involve volatile organic compounds (VOCs) and nitrogen oxides (NO_x). Heavy industry facilities could be potential sources of VOCs. The concentration of O₃ increased from 51 µg m⁻³ in 2015 to 83 µg m⁻³ in 2017 (Figure 9). The concentration of NO₂ in Aktobe is below the WHO and EU cutoff values. The Municipality of Aktobe Region (2017) reported that flaring of petroleum gas is a major environmental challenge in the region, as it accounted for 23% of the total emissions from stationary sources. Electricity consumption in the Aktobe region increased at an annual rate of 7% from 2012 to 2016 (The Municipality of Aktobe Region 2017). The city is provided with gas via the gas network; therefore, heat and power generation rely on gas. The only source of district heating is “Aktobe CHP,” which relies on natural gas and petroleum gas. The annual growth rate in the number of registered transport vehicles in the Aktobe region was 3.5% during the period from 2011 to 2017, reaching 149 thousand in 2017.

Figure 9. Average annual concentration level of O₃ (left) and NO₂ (right) in Aktobe

Estimated health impacts of exposure to PM_2.5 and PM₁₀ in the major cities of Kazakhstan

The results presented here demonstrate that exposure to PM₁₀ in the major cities of Kazakhstan caused 3249 new cases of chronic bronchitis, 5991 hospital admissions, and 117,317 emergency room visits annually in the period from 2015 to 2017 (Table 2). The World Bank (2013) estimated three-times lower morbidity levels in Kazakhstan in 2011 in four selected regions: 1016 cases of chronic bronchitis, 2040 cases of hospital admissions, 40,400 cases of emergency room visits, 6,712,000 restricted activity days, 92,100 lower respiratory illnesses in children and 21,364,000 respiratory symptoms. Annually, the number of per-capita (adults) restricted activity days was 3.8 and the number of per-capita respiratory symptoms was 11.9.

Table 2. Additional morbidity cases due to exposure to PM₁₀ in the major cities (annually over the 2015–2017)

City	Chronic bronchitis (>16 years)	Hospital admissions (all population)	Emergency room visits (all population)	Restricted activity days (>16 years)	Lower respiratory illness in children (0–15 years)	Respiratory symptoms (>16 years)
Nur-Sultan	807	1521	29793	5153341	62789	16401067
Zhezkazgan	71	132	2577	452445	5231	1439955
Temirtau	151	259	5070	966808	8047	3076972
Shymkent	494	1009	19759	3154451	49384	10039382
Balkhash	38	68	1325	241700	2427	769236
Almaty	843	1458	28552	5383190	47115	17132587
Aktau	67	132	2593	430098	6003	1368832
Atyrau	94	192	3758	600359	9382	1910706
Taraz	100	200	3907	635531	9419	2022645
Pavlodar	117	200	3922	745702	6286	2373278
Ekibastuz	46	81	1578	295787	2656	941376
Karaganda	133	232	4550	849225	7758	2702752
Semipalatinsk	88	154	3005	562036	5094	1788740
Ust-Kamenogorsk	73	122	2391	464605	3541	1478,657
Taldykorgan	30	56	1095	192104	2229	611392
Uralsk	52	95	1857	334775	3517	1065459
Kokshetau	27	48	939	169441	1772	539264
Petropavlovsk	20	33	646	125906	948	400710
Kyzylorda	0	0	0	0	0	0
Aktobe	0	0	0	0	0	0
Kostanay	0	0	0	0	0	0
Total	3249	5991	117317	20757503	233597	66063010

It was estimated a total of 8134 adult (age ≥25 years) deaths attributable to PM_2.5 per year (average over 2015–2017) (Table 3). The leading causes of death were ischemic heart disease (4080), stroke (1613), lower respiratory infections (662), chronic obstructive pulmonary disease (434), lung cancer (332) (Table 4; Supplementary File: Table A6, Table A7). In average across the cities of Kazakhstan, per capita mortality rate attributable to ambient air pollution is 165 per 10⁵ adults per year, which is higher than in Europe (133 per 10⁵ for all ages) (Lelieveld et al. 2019) and Tehran (118 deaths per 10⁵ adults) (Bayat et al. 2019). In Tehran, percentage of deaths attributable to PM_2.5 was 15.3% of total deaths, with pollution level of 31 μg m⁻³. In Atyrau (Kazakhstan) having similar pollution (29 μg m⁻³), this indicator was estimated at 15%.

Table 3. Additional NCD+LRI mortality cases by age ranges due to exposure to PM_2.5 in the major cities of Kazakhstan (annually over the 2015–2017) (SE-standard error; SE for $\theta$ parameter obtained from Burnett et al. 2018)

	25–29		30–34		35–39		40–44		45–49		50–54		55–59		60–64		65–69		70–74		75–79		80–84		85+		Total
City/Age range	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Mean	SE
Nur-Sultan	21	3	30	3	36	4	51	6	55	6	86	9	94	10	103	11	107	12	70	8	121	13	75	8	90	10	939
Zhezkazgan	4	0	5	0	8	0	9	0	11	0	15	0	21	0	23	0	25	0	18	0	34	1	21	0	24	0	216
Temirtau	7	0	11	0	16	0	18	0	21	0	29	1	41	1	46	1	50	1	35	1	68	1	41	1	48	1	433
Shymkent	20	0	25	0	34	1	39	1	53	1	69	1	89	2	94	2	111	2	78	1	126	2	73	1	96	2	905
Balkhash	2	0	3	0	5	0	6	0	7	0	9	0	13	0	15	0	16	0	11	0	22	0	13	0	15	0	138
Almaty	27	1	43	1	53	1	71	1	82	1	119	2	156	3	183	3	216	4	125	2	270	5	190	3	297	5	1831
Aktau	3	0	4	0	6	0	8	0	9	0	11	0	12	0	14	0	15	0	10	0	12	0	8	0	9	0	120
Atyrau	6	0	8	0	11	0	13	0	16	0	25	0	30	0	26	0	28	0	19	0	35	1	30	0	31	1	277
Taraz	5	0	7	0	10	0	16	0	16	0	21	0	27	0	30	0	33	1	25	0	41	1	21	0	21	0	275
Pavlodar	8	0	12	0	18	0	20	0	24	0	36	1	46	1	51	1	56	1	33	1	80	1	45	1	54	1	483
Ekibastuz	3	0	5	0	7	0	8	0	10	0	15	0	19	0	21	0	23	0	14	0	33	1	19	0	22	0	199
Karaganda	10	0	15	0	24	0	26	0	31	1	42	1	59	1	66	1	72	1	50	1	96	2	59	1	67	1	617
Semipalatinsk	8	0	11	0	15	0	18	0	19	0	29	0	40	1	47	1	50	1	31	1	64	1	43	1	59	1	435
Ust-Kamonogorsk	7	0	10	0	13	0	16	0	17	0	25	0	35	1	41	1	44	1	28	0	56	1	38	1	52	1	381
Taldykorgan	3	0	3	0	4	0	6	0	7	0	9	0	12	0	13	0	14	0	10	0	15	0	11	0	12	0	118
Uralsk	5	0	5	0	9	0	11	0	14	0	21	0	26	0	26	0	27	0	16	0	42	1	29	0	35	1	267
Kokshetau	3	0	3	0	5	0	5	0	7	0	11	0	14	0	16	0	17	0	12	0	24	0	15	0	17	0	151
Petropavlovsk	3	0	5	0	6	0	7	0	8	0	12	0	18	0	21	0	21	0	13	0	31	0	21	0	25	0	191
Kyzylorda	1	0	2	0	2	0	4	0	4	0	5	0	6	0	7	0	8	0	6	0	9	0	5	0	5	0	64
Aktobe	2	0	3	0	4	0	5	0	6	0	8	0	9	0	10	0	9	0	6	0	14	0	9	0	11	0	95
Kostanay	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Total	148		212		286		356		415		596		768		854		940		611		1193		763		992		8134

Table 4. Additional COPD, LC, LRI, IHD, and stroke mortality cases due to exposure to PM_2.5 in the major cities of Kazakhstan (annually over the 2015–2017)

City	COPD		Lung cancer		LRI		IHD	Stroke	City total (by 5 causes)
	Mean	Standard error	Mean	Standard error	Mean	Standard error	Mean	Mean
Nur-Sultan	49	16	38	10	63	23	491	202	149
Zhezkazgan	11	4	9	2	15	5	111	45	35
Temirtau	23	7	18	5	30	11	223	91	70
Shymkent	48	15	37	9	68	24	483	191	153
Balkhash	7	2	6	1	11	4	75	28	24
Almaty	100	30	77	18	151	52	972	367	328
Aktau	6	2	5	1	10	3	69	25	21
Atyrau	15	4	11	3	24	8	156	55	50
Taraz	15	4	11	3	24	8	157	55	50
Pavlodar	26	8	20	5	43	14	270	95	89
Ekibastuz	11	3	8	2	18	6	111	39	37
Karaganda	33	10	25	6	56	18	349	118	114
Semipalatinsk	23	7	18	4	40	13	156	82	81
Ust-Kamonogorsk	20	6	15	3	35	11	138	70	71
Taldykorgan	6	2	5	1	11	3	42	22	22
Uralsk	14	4	11	2	24	7	97	48	49
Kokshetau	8	2	6	1	14	4	54	27	28
Petropavlovsk	10	3	7	2	17	5	70	31	34
Kyzylorda	3	1	2	0	4	1	23	9	10
Aktobe	4	1	3	1	6	2	34	13	14
Kostanay	0	0	0	0	0	0	0	0	0
Total	434		332		662		4080	1613	7121

Previously World Bank (2013) estimated 2800 premature deaths due to particulate matter pollution in 2011 in four selected regions Karaganda region, East Kazakhstan region, Almaty region, and Pavlodar region (representing 30% of total population). The WHO GBD study estimated that 10,064 premature deaths occurred due to particulate matter pollution in 2010 in Kazakhstan (WHO 2015) (62 per 10⁵ population) using integrated exposure-response functions. Estimated deaths were not presented by cause in the WHO (2015) data. These differences in the estimates could be attributed to worsening air quality levels with elapsed time as well as differences in target concentration, applied exposure-response methodology, and hazard ratios (this study applied a novel GEMM), as well as the coverage of the exposed population.

The per-capita mortality rate attributable to ambient air pollution varied across the cities, from 34 to 373 per 10⁵ adults (Figure 10), due to the differences in the baseline mortality rates and PM_2.5 concentrations. Differences in the per-capita baseline mortality rate across cities may be attributed to many factors, such as air pollution, economy, lifestyle, poverty rates, and the availability of health-care services. Attributable to ambient air pollution mortality rate (per 10⁵ adults per year) was less than 150 in nine cities, between 150 and 204 in nine cities, and between 276 and 373 in three cities. Three cities with heavy industries and coal-fired power generation, namely Zhezkazgan, Temirtau, and Balkhash had highest PM_2.5 mortality rate: 276–373 per 10⁵ adults (Figure 10 and Table 5). These industrial cities also had baseline mortality rate among adults of 1494 per 10⁵ which is higher than the average baseline mortality rate among adults across all studied cities (1129 per 10⁵ adults). This could be a consequence of air pollution with heavy metals (lead, zinc, iron) and requires future investigation.

Figure 10. Mortality rate attributable to ambient air pollution (per 105 adults) by cities of Kazakhstan

The percentage of PM_2.5 mortality from total adults deaths was varying from the lowest 3% to 26%, with more than 20% in four most polluted cities: Nur-Sultan (25%), Zhezkazgan (25%), Temirtau (24%), and Shymkent (21%) (Table 5).

Table 5. Characteristics of the cities and estimated PM_2.5 mortality

	Population (>25 years)	Number of Registered vehicles in 2017 (region), thousands	Coal consumption in 2017, thousand tons (region)	Annual average PM2.5 concentration (2015–2017), µg m^−3	Baseline mortality rate (>25 years), per 100000 people, 2017	PM2.5 mortality rate (NCD+LRI), per 100000 people (>25 years)	PM2.5 mortality (NCD+LRI) (>25 years), total	Share of PM2.5 mortality to baseline mortality
Large, urban cities with coal power plants
Nur-Sultan	577689	251	3743	62	636	163	939	26%
Almaty	1134335	462	2348	37	917	161	1831	18%
Karaganda and Pavlodar regions: Cities with heavy industry and coal power plants
Zhezkazgan	57975	283	38394	61	1494	373	216	25%
Temirtau	118442			59	1494	365	433	24%
Balkhash	50128			39	1494	276	138	18%
Karaganda	319292			24	1494	193	617	13%
Pavlodar	236962	156	68166	28	1425	204	483	14%
Ekibastuz	100599			27	1425	198	199	14%
Cities in the East of Kazakhstan: Coal power plants and heavy industry
Ust-Kamenogorsk	224147	301	7588	20	1519	170	381	11%
Semipalatinsk	228235			23	1519	190	435	13%
Cities in the North of Kazakhstan (coal power plant in Petropavlovsk)
Petropavlovsk	145833	147	317	12	1749	131	191	7%
Kokshetau	101317	178	2258	18	1518	149	151	10%
Kostanay	161659	172	1893	0	1511
Cities in the West of Kazakhstan: gas network available, oil and gas industries
Atyrau	175506	114	1	29	1065	158	277	15%
Uralsk	186334	117	10	18	1376	143	267	10%
Aktobe	279396	149	226	6	1074	34	95	3%
Aktau	95885	138	107	33	741	125	120	17%
Cities in the South of Kazakhstan: gas network available
Taldykorgan	97784	474	1529	19	1119	121	118	11%
Kyzylorda	152530	110	98	7	1023	42	64	4%
Taraz	184345	194	645	28	1023	149	275	15%
Shymkent	458999	482	1087	46	938	197	905	21%

Conclusion

Using the data from monitoring stations, this study revealed that air quality is poor in the majority of the large cities in Kazakhstan. In 18 (out of 21) of the cities studied, the annual limit value of the WHO for PM₁₀ was exceeded. In terms of NO₂ concentration, the WHO limit was exceeded in eight cities. Nur-Sultan, Almaty, Ust-Kamenogorsk, and Aktobe were the most polluted cities in terms of PM₁₀, NO₂, SO₂ and O_3, respectively. The rapidly growing Nur-Sultan (capital of Kazakhstan) had increasing trends of PM₁₀ and NO₂, with annual average PM₁₀ concentrations (2015–2017) reaching 135 µg m⁻³ (exceeding the WHO annual limit by a factor of 6.9). The largest city of Kazakhstan and former capital, Almaty, had decreasing levels of NO₂, possibly because of developments in the transport sector; however, the values still exceeded the WHO annual limit by a factor of 2.1 (on average from 2015 to 2017). Ust-Kamenogorsk, with heavy industry and coal power plants, was the most polluted city in terms of SO₂ (85 µg m⁻³), with fluctuating trends from 2011 to 2017. The concentration of O₃ increased in the industrial Aktobe from 51 µg m⁻³ in 2015 to 83 µg m⁻³ in 2017. The linear correlation of the average annual concentration of pollutants with total coal consumption or the number of registered vehicles in the region were small, which can be associated with many factors, including the resolution and inaccuracies of the data used, the influence of meteorological and terrain conditions, and distribution and transportation factors between the source and receptor. Understanding sources of pollution requires future research with source apportionment and air quality modeling in the polluted cities of Kazakhstan.

It is estimated that exposure to inhalable particles in the major cities of Kazakhstan causes 8134 adult premature deaths annually. Differences in the cities were vast as per capita mortality rate attributable to ambient air pollution was varying from 34 to 373 per 10⁵ adults per year due to the differences in the PM_2.5 concentration and baseline mortality rates. Three industrial cities Zhezkazgan, Temirtau, and Balkhash had higher per capita PM_2.5 mortality rate compared to the other cities.

Government intervention measures to control or reduce atmospheric pollution in Kazakhstan can lead to significant benefits to society in terms of reducing negative health impacts.

The National Hydrometeorological Service of Kazakhstan is still in its development, as the number of measurement stations and priority pollutant coverage is gradually improving. In this study, to estimate health impacts, data on total suspended particles were used due to the low coverage of the PM_2.5 network for the historical period from 2015 to 2017. Health impacts need to be recalculated with measured PM_2.5 data when the PM_2.5 data become available with sufficient coverage in the cities of Kazakhstan.

Acknowledgment

The authors thank Reza Bayat for providing valuable comments to the paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.

Additional information

Funding

This work was supported by Postdoctoral Fellowship Program of Al-Farabi Kazakh National University, and Haller Lomax LLP.

Notes on contributors

Aiymgul Kerimray

Aiymgul Kerimray is a Postdoctoral Researcher at Al-Farabi Kazakh National University, Center of Physical-Chemical Methods of Research and Analysis, Almaty, Kazakhstan.

Daulet Assanov

Daulet Assanov is a Senior Researcher Excellence Center “Veritas,” D. Serikbayev East Kazakhstan State Technical University, Ust-Kamenogorsk, Kazakhstan.

Bulat Kenessov

Bulat Kenessov is a Professor, Director of the Center of Physical-Chemical Methods of Research and Analysis at Al-Farabi Kazakh National University, Almaty, Kazakhstan.

Ferhat Karaca

Ferhat Karaca is Associate Professor at the Department of Civil and Environmental Engineering at Nazarbayev University, Nur-Sultan, Kazakhstan.