Air Quality and Environmental Policy in Kazakhstan: Challenges, Innovations, and Pathways to Cleaner Air

Abstract

Urban air pollution in Kazakhstan poses persistent risks; this study synthesizes measured concentrations, source evidence, and policy responses to inform mitigation in cold, inversion-prone cities. We compile national monitoring (Kazhydromet), community PM_2.5 sensors, emissions inventories and recent CEMS provisions, and appraise modeling approaches (Gaussian screening, Eulerian CTMs, and data-driven forecasting). Seasonal descriptive comparisons are performed for Astana using 56,944 observations (2023–2024), partitioned into heating and non-heating periods, and published receptor apportionment is integrated. Across major cities, annual PM_2.5 generally exceeds WHO guidelines and winter stagnation drives episodes. In Astana, the heating season means rose relative to non-heating equivalents—PM_2.5 12.3 vs. 10.6 μg m⁻³ (+16%) and SO₂ 21.9 vs. 14.8 μg m⁻³ (+23%)—while NO was unchanged; higher means but lower medians indicate episodic winter peaks. Receptor analyses attribute large shares of PM_2.5 to traffic (spark-ignition engines 30% and diesel 7%) and coal-related contributions including secondary nitrate (15%), consistent with power/heat and vehicle dominance. Evidence supports prioritizing clean heating (coal-to-gas and efficiency), transport emission controls, and dense monitoring to enable accountability within Kazakhstan’s Environmental Code and decarbonization strategy. A tiered modeling workflow can quantify intervention impacts and deweather trends; the near-term focus should be on reducing winter exposures.

1. Introduction

Air pollution is a global health emergency, linked to about seven million premature deaths each year, roughly eighty nine percent of which occur in low- and middle-income countries (LMICs). Mean ambient PM_2.5 exposure in LMICs is about 1.3 to 4 times higher than in high-income countries [1]. Many of the most affected countries lack robust monitoring and regulatory systems [2]. A 2021 United Nations assessment found that thirty-four percent of countries have no legal ambient air quality protections, and among those, eighty-six percent have no standards in place [3]. As of 2025, the World Health Organization noted that many nations still do not implement pollutant standards [4]. In many developing cities, limited and discontinuous records impede a clear view of trends [1]. For example, large African urban centers long had little or no routine air quality data, a situation that is only beginning to change with the introduction of low-cost sensor networks [5]. Even with these constraints, recent work shows progress through innovative monitoring. In Pakistan, where institutional and financial limits have constrained official networks, researchers demonstrated that a hybrid system combining low-cost IoT-based sensors with conventional stations can track particulate matter reliably [6]. A nationwide deployment of sixty-one low-cost PM_2.5 sensors showed strong agreement with reference monitors, with a Pearson correlation of 0.85, which supports the use of affordable sensors to supplement sparse regulatory coverage [1]. Similar efforts are emerging elsewhere. Open data platforms and sensor collaborations are expanding coverage in parts of Central Asia, including Kazakhstan’s neighbors, and use new technologies to fill gaps [7,8]. In Latin America, an analysis of Quito from 2019 to 2024 showed that overlapping crises, including energy shortages, political unrest, and drought influenced by climate variability, led to worsening air quality despite some control policies [1]. That case illustrates how vulnerable many developing cities remain to compounding stressors and highlights the need for resilient management underpinned by continuous monitoring. It also noted a lack of longitudinal studies in smaller cities across the Global South and the resulting evidence gaps [9].

In Kazakhstan, air pollution continues to affect public health, economic development, and environmental quality. Across many urban centers, annual concentrations of PM_2.5, NO₂, SO₂, and O₃ exceed health-based guidelines, particularly in winter [10]. In 2021, Almaty and Astana recorded annual PM_2.5 of 35.3 micrograms per cubic meter and 22.5 micrograms per cubic meter, with daily exceedances on 217 and 151 days, which far exceed recommended limits [11,12]. These city-level results mirror a broader national pattern documented in multi-city assessments and syntheses. While policy reports often cite more than ten thousand premature deaths and USD ten point five billion in annual economic losses, peer-reviewed estimates indicate 8134 adult premature deaths per year from 2015 to 2017 attributable to PM_2.5 across twenty-one Kazakhstani cities, which points to the magnitude of the health burden [3,13,14].

Climatic and topographic factors amplify these outcomes nationwide. Cities situated in valleys or basins (e.g., Ust-Kamenogorsk and Almaty) experience frequent temperature inversions and stagnant air that inhibit dispersion, producing severe winter smog episodes. Time series and boundary-layer studies across Central Asian urban settings further document the seasonal signal: high winter concentrations tied to shallow mixing heights and calm winds, with cleaner summer conditions except for episodic dust [11,15,16,17].

According to [17,18], Kazakhstan’s energy system and on-road transport are the principal drivers of urban PM_2.5 exposure. Coal-based district heating and residential stoves, together with motor vehicle emissions, dominate winter pollution in major cities. Receptor/source apportionment work (e.g., in Astana as a representative case) attributes the largest single share of PM_2.5 to spark-ignition engine vehicles (30%), with coal fly ash and secondary sulfate/nitrate accounting for substantial additional fractions—consistent with heating and power generation in the cold season. Urban VOC measurements in Almaty indicate traffic and fuel-combustion signatures, and fractal/time series analyses reveal robust seasonal/episodic dynamics with winter peaks [8,18].

Urban pollution profiles across Kazakhstan thus mirror patterns observed in other rapidly growing, cold-climate cities. PM_2.5 remains the central health concern because of deep lung penetration; PM₁₀ contributes to respiratory morbidity and reduced visibility; NO₂ (from traffic and boilers) often exceeds safe levels and, in city-level analyses, shows very strong statistical associations with adverse health outcomes; SO₂ (coal combustion) frequently breaches limits in winter; CO spikes during severe smog due to incomplete combustion; and O₃ tends to be a summer-season issue tied to photochemical formation [3,19].

Table 1. Major urban air pollutants: sources and health-based guideline values.

Pollutant	Predominant Urban Sources	Guideline Value *	Main Health/Environmental Impacts
PM2.5	Coal/power and residential heating; vehicle exhaust; secondary aerosol; dust resuspension	5 μg m−3 (annual); 15 μg m−3 (24 h)	Cardiovascular and respiratory disease; premature mortality; visibility/haze
PM10	Construction and road dust; combustion; resuspension	15 μg m−3 (annual); 45 μg m−3 (24 h)	Respiratory irritation; reduced visibility/soiling
NO2	Traffic, boilers, power plants (fuel; NOx)	10 μg m−3 (annual); 25 μg m−3 (24 h)	Airway inflammation; asthma; ozone/nitrate PM precursor
SO2	Coal/oil combustion in power and heat; industry	40 μg m−3 (24 h)	Respiratory effects; sulfate PM and acid deposition
CO	Incomplete combustion (vehicles, stoves, and open burning)	4 mg m−3 (24 h)	Reduced O2 delivery; cardiovascular stress; acute toxicity
O3	Secondary (NOx + VOCs under sunlight)	100 μg m−3 (8 h)	Respiratory symptoms; lung function decrements

* according to WHO Global Air Quality Guidelines, 2021 [20,21,22].

itemizes key urban sources and WHO targets for context.

Placed within the broader peer-reviewed literature, Kazakhstan’s urban air quality patterns echo well-documented global dynamics and policy responses. Long-term exposure to ambient PM_2.5 is among the leading worldwide health hazards, with multi-million premature deaths inferred by large cohort syntheses and integrated exposure–response models (Global Burden of Disease, GBD; Global Exposure Mortality Model, GEMM) [23,24,25,26]. According to Tang et al. (2016), cold-climate and basin-type cities elsewhere show analogous wintertime amplification via shallow mixing layers and persistent inversions (e.g., Beijing and other northern mid-latitude cities) [27]. At the same time, “accountability” studies demonstrate that targeted policy can causally improve air quality and health: Dublin’s coal sale ban yielded immediate reductions in particulates and all-cause, cardiovascular, and respiratory mortality; London’s low- and ultra-low-emission zones are associated with material roadside NO₂ declines and are being evaluated for child respiratory benefits; China’s post-2013 clean air and clean heating campaigns achieved substantial winter PM_2.5 reductions with measurable co-benefits [28,29,30,31,32,33]. Taken together, these findings motivate a Kazakhstan-wide synthesis to test alignment between observed patterns, source diagnostics, and the leverage of candidate interventions.

The high loads observed across Kazakhstan’s cities thus reflect an energy and mobility challenge shaped by harsh winters and, at times, unfavorable meteorology. In line with Kazakhstan’s climate strategy (carbon neutrality by 2060), authorities have begun to accelerate fuel switching, upgrade standards, densify monitoring, and foster public awareness [34]. Against this policy backdrop, a focused scientific synthesis is required to show whether observed patterns match source diagnostics and which levers yield the largest gains.

While prior works analyze single strands, e.g., health burdens [13], boundary-layer controls on winter peaks [14], or PM_2.5 source apportionment [35], a consolidated, Kazakhstan-centered account that juxtaposes municipal monitoring, receptor results, and policy-relevant modeling remains limited. Accordingly, this study addresses that gap by integrating “what is measured,” “where it comes from,” and “what works”.

This study introduces an integrative framework that connects what is measured, where it originates, and what works for policy. We refer to it as the Measurement, Attribution, and Intervention (MAI) framework. Measurement harmonizes municipal and community monitoring with CEMS records to produce comparable seasonal time series. Attribution then quantifies source contributions and meteorological controls through receptor modeling and process-aware diagnostics. Intervention uses these source fingerprints to construct policy scenarios and to rank options by expected concentration reductions, health benefits, and feasibility in Kazakhstan’s regulatory context.

Methodologically, the framework adapts urban air quality theory to cold, inversion-prone continental climates by linking observed winter exceedances to two mechanisms: limited vertical mixing, reflected in planetary boundary-layer constraints coupled with PM_2.5, and chemically consistent source signals, including secondary sulfate and nitrate associated with coal combustion and NO_x. This alignment between empirical diagnostics and intervention design ensures that measures are derived from observed patterns and evaluated within a coherent decision pipeline.

Thus, the article integrates pollutant characterization, monitoring and modeling evidence, and recent policy responses to provide a consolidated, country-wide account of urban air quality in Kazakhstan within its regional context. Accordingly, Section 2 details methods; Section 3 presents results for pollutant levels, seasonality, and sources; Section 4 discusses policy implications and expected gains under fuel switching and transport control; Section 5 concludes with research and implementation priorities.

2. Methodology

We operationalized the Measurement–Attribution–Intervention (MAI) framework using existing datasets and published evidence. For Measurement, we compiled national regulatory observations (Kazhydromet), complementary community PM_2.5 sensors, and information on CEMS; assembled 56,944 records for Astana (2023–2024); aligned them to a common resolution; partitioned the series into heating and non-heating seasons; produced descriptive summaries; and mapped monitoring sites against energy infrastructure. For Attribution, we interpreted source contributions using published receptor/source apportionment for Astana and boundary-layer diagnostics that link winter exceedances to shallow mixing and coal/traffic signatures. For Intervention, we conducted a desk review of policy instruments and modeling toolkits relevant to Kazakhstan (Gaussian screening, Eulerian CTMs, and data-driven forecasting) to identify leverage points and near-term priorities. Seasonal contrasts were evaluated with non-parametric tests (e.g., Mann–Whitney U), reported by pollutant, and no new process-based simulations were run.

2.1. Official Monitoring

Kazhydromet operates atmospheric air quality monitoring in about 170 observation posts across 70 urban settlements, including 15 mobile laboratories. These stations record particulate matter (PM_2.5, PM₁₀, and TSP), sulfur dioxide, nitrogen dioxide, carbon monoxide, ozone, and other parameters. The AirKZ mobile application and published bulletins provide access to these data. These observations serve as the foundation for annual analysis, pollution exceedance evaluation, and comparisons between cities.

2.2. Portfolio of Air Monitoring Methods Relevant to Kazakhstan

A reliable evidence base arises from complementary methods that differ in coverage, timing, and uncertainty. Reference grade fixed stations operated by Kazhydromet provide quality assured time series for PM_2.5, PM₁₀, SO₂, NO₂, CO, and O₃, which support compliance assessment with WHO values, trend analysis, and intercity comparisons [2,11]. Mobile laboratories add targeted campaigns along roads and near sources to map fine scale gradients and confirm hot spots during episodic events [2]. Low-cost sensor networks connected wirelessly supply dense spatial coverage at modest cost. With field calibration and routine quality control, they capture intraurban variability and short-lived peaks. Pilots in major Kazakhstani cities have shown practical value for communication and targeting [2,11]. Continuous emissions monitoring systems at large sources such as CHP plants provide near real-time stack concentrations and flows for SO₂, NO_x, and particulate matter, improving enforcement and linking ambient changes to operations. Remote sensing with ancillary data including meteorology and satellite AOD or NO₂ columns fills spatial gaps, informs model boundaries, and assists with attribution under regional transport [6,10,18]. Monitoring driven analytics combine receptor measurements such as black carbon and chemical tracers with statistical and machine learning forecasting to apportion sources and to generate short-term forecasts and meteorologically normalized indicators [8,24,26].

For inversion-prone cities in cold climates, the recommended approach is to maintain and modernize the reference network, deploy city-scale IoT sensor meshes with routine collocation for calibration, expand continuous emissions monitoring at major stacks, and operate forecasting and normalization workflows that couple observations with meteorology for alerts and for evaluation of policies during heating seasons [2,8,11,24,26]. This portfolio matches priorities in Astana and other cities where wintertime particulate matter and SO₂ are driven by coal-based heating and traffic emissions [2,10,26].

2.3. Supplementary Low-Cost Sensors

Nazarbayev University leads a regional initiative deploying over 120 field-calibrated PM_2.5 sensors across Kazakhstan. These sensors are installed indoors and outdoors in major cities (Almaty, Astana, Atyrau, Taraz, Shymkent, Kokshetau, and Semey) and provide high-resolution spatial data. The project emphasizes calibration, data protocols, and integration with satellite and meteorological records.

Low-cost PM_2.5 sensors were used for spatial context rather than headline statistics. Under high relative humidity and extreme cold typical of Astana winters, hygroscopic growth and condensation can bias optical readings upward and sub-zero temperatures may affect flow and electronics. We therefore interpret these data cautiously during such conditions, rely on regulatory monitors for exceedance metrics, and treat the sensors as qualitative corroboration despite basic colocation checks.

2.4. Emissions Monitoring and Inventories

As part of Kazakhstan’s 2060 carbon neutrality strategy, large stationary emitters are expected to use Continuous Emission Monitoring Systems (CEMSs). These systems provide real-time particulate and gas measurements directly from smokestacks and are incorporated into emissions inventories that support pollutant source analysis.

2.5. Modeling Approaches

We apply a tiered view of studies with modeling approaches used in Kazakhstan and comparable cold-climate cities:

(i) Gaussian dispersion for near-source screening, (ii) Eulerian chemical transport models (CTMs) for process-resolved simulation, and (iii) data-driven/statistical models for forecasting and post hoc inference.

• Gaussian dispersion (near-field screening)

Gaussian plume models assume steady flow over short periods and represent downwind dilution from an effective point source with reflections at the ground level. A standard elevated-source solution is

$$\mathit{C}(x,y,z)={\displaystyle \frac{Q}{2\pi {\sigma }_y{\sigma }_{z}U}}\mathrm{e}\mathrm{x}\mathrm{p}\left(-{\displaystyle \frac{y^2}{2{\sigma }_y^2}}\right)\left[\mathrm{e}\mathrm{x}\mathrm{p}\left(-{\displaystyle \frac{(z-H{)}^2}{2{\sigma }_z^2}}\right)+\mathrm{e}\mathrm{x}\mathrm{p}\left(-{\displaystyle \frac{(z+H{)}^2}{2{\sigma }_z^2}}\right)\right]$$

(1)

where $Q$ is the emission rate, $U$ the wind speed, $H$ the effective stack height, and ${\sigma }_y,{\sigma }_z$ the dispersion parameters that grow with distance and atmospheric stability. These models are fast and useful for stacks, road segments, and permitting, but they neglect chemistry, complex terrain, and highly unsteady flows, so they serve best as screening tools before process-based modeling.

2.

Eulerian CTMs (city to regional scales)

CTMs resolve advection, turbulent mixing, chemistry, emissions, and deposition on 3-D grids via the advection–diffusion–reaction balance as follows:

$${\displaystyle \frac{\mathsf{\partial }{c}_i}{\mathsf{\partial }t}}=-\nabla \cdot \left(u{c}_i\right)+\nabla \cdot \left(K\nabla c_i\right)+R_i\left(c_1,\dots ,c_N\right)+E_i-D_i.$$

(2)

where $c_i(x,t)$ is the concentration of species $i \left[\mu \mathrm{g}\cdot {\mathrm{m}}^{-3}\right];u(x,t)$ is the wind velocity $\left[\mathrm{m}\cdot {\mathrm{s}}^{-1}\right] ;K(x,t)$ is the eddy diffusivity tensor $\left[{\mathrm{m}}^2\cdot {\mathrm{s}}^{-1}\right]; R_i(\cdot )$ denotes gas–aerosol chemistry and secondary formation $\left[\mathrm{\mu }\mathrm{g}\cdot {\mathrm{m}}^{-3}\cdot {\mathrm{s}}^{-1}\right]; E_i(x,t)$ is emissions $\left[\mathrm{\mu }\mathrm{g}\cdot {\mathrm{m}}^{-3}\cdot {\mathrm{s}}^{-1}\right]$ mapped from inventories; and $D_i(x,t)$ is the dry/wet deposition sink [$\mu \mathrm{g}\cdot {\mathrm{m}}^{-3}\cdot {\mathrm{s}}^{-1}$].

Boundary conditions follow standard practice (zeroflux at surface for inert species with prescribed deposition velocities; lateral/top inflow from reanalysis/parent models). This term-resolved balance motivates CTMs for policy scenarios (secondary sulfate/nitrate; long-range transport).

Widely used systems include CMAQ for multipollutant air quality, WRF-Chem for online weather–chemistry coupling, and GEOS-Chem for global/hemispheric chemistry; they capture secondary aerosol and ozone formation, long-range transport, and scenario responses, at the cost of high data and computation requirements.

3.

Data-driven and statistical models (forecasting and inference)

Statistical and machine learning approaches learn mappings from observations (meteorology, emissions proxies, and past concentrations) to future concentrations and are valuable for high-frequency forecasting or for “deweathering” trends. A simple baseline is multiple linear regression,

$${\mathrm{P}\mathrm{M}}_{2.5,t+1}={\beta }_0+{\beta }_1{\mathrm{P}\mathrm{M}}_{2.5,t}+{\beta }_2{U}_t+{\beta }_3{T}_t+{\beta }_4{\mathrm{H}\mathrm{D}\mathrm{D}}_t+{\epsilon }_t,$$

(3)

where $\mathrm{P}{\mathrm{M}}_{2.5,t+1} \left[\mathrm{\mu }\mathrm{g}\cdot {\mathrm{m}}^{-3}\right]$ is the next-step concentration; $\mathrm{P}{\mathrm{M}}_{2.5,t}$ is persistence; $U_t$ is wind speed $\left[\mathrm{m}\cdot {\mathrm{s}}^{-1}\right];T_t$ is air temperature [° C]; $HD{D}_t$ is a heating degree proxy [°C day, base 18 °C] to capture seasonal demand; ${\epsilon }_t$ is the zero-mean error. Extensions (RF/GBM/LSTM) capture nonlinearity and temporal dependence; meteorological normalization yields deweathered trends for accountability.

These are often extended with random forests, gradient boosting, or LSTM networks to capture nonlinear and temporal structure. Random forest meteorological normalization is widely used to separate emission changes from weather variability, and recent surveys summarize the rapidly growing ML literature for air quality prediction and interpretation. Cautions include data shift sensitivity, limited physical interpretability, and the need for rigorous cross-validation and external testing.

For industrial stacks and road links, Gaussian modeling provides quick impact screening; for policy design and accountability (e.g., fuel switching or traffic measures), CTMs quantify secondary formation and regional transport; for operations, ML models deliver short-term forecasts and deweathered indicators that complement physics-based systems. A tiered workflow that combines these tools is now standard in urban air quality management.

Overall, the study combines multiple data streams: long-term and recent ambient concentration data from Kazhydromet and other studies, high-resolution temporal data for seasonal analysis, and emissions data from inventories and monitoring. All data are interpreted in light of their uncertainty and limitations (e.g., monitoring coverage or sensor accuracy). Where needed, statistical analyses (such as seasonal comparisons) were performed. For example, winter vs. summer differences in pollutant levels were assessed using descriptive statistics (means, medians, and interquartile ranges) and significance tests, as reported in the Results section. Such analyses help distinguish sustained seasonal changes from episodic spikes.

Process-based chemical transport models relevant to Kazakhstan include WRF Chem and CMAQ, while statistical baselines include random forest and gradient boosting regressors for forecasting and meteorological normalization. In this paper CTM or ML simulations were not run; these models are referenced as established options aligned with the MAI pipeline and proposed for targeted scenario evaluation in follow-up work.

3. Results

3.1. Ambient Air Quality in Kazakhstan Cities

Ambient air quality in many Kazakhstani cities is poor and, in some cases, worsening. Over the past two decades, pollution levels have remained persistently high despite various reform efforts. Kazakhstan ranked 30th out of 113 countries in urban air pollution in 2025 according to Numbeo’s index (scoring 73.2/100), only a modest improvement from a peak score of 79.6 in 2013 [27]. Indeed, annual PM_2.5 in Kazakhstan’s major cities remains well above the WHO health benchmark of 5 µg/m³ (see

Table 2. Major Kazakhstani cities: annual PM_2.5 in 2024 vs. WHO AQG-2024.

City	PM2.5 2024 (µg/m3)	×WHO	WHO Annual Category
Karaganda	104.8	21.0×	>50 (well above interim targets)
Temirtau	43.5	8.7×	35.1–50
Atyrau (Balykshi station)	32.1	6.4×	IT-1 (25.1–35)
Almaty	20.3	4.1×	IT-2 (15.1–25)
Oskemen (Ust-Kamenogorsk)	20.9	4.2×	IT-2 (15.1–25)
Shymkent	19.9	4.0×	IT-2 (15.1–25)
Aktau	18.0	3.6×	IT-2 (15.1–25)
Astana	15.4	3.1×	IT-2 (15.1–25)
Pavlodar	15.4	3.1×	IT-2 (15.1–25)
Kostanay	15.0	3.0×	IT-3 (10.1–15)
Aktobe	14.4	2.9×	IT-3 (10.1–15)
Kokshetau	11.6	2.3×	IT-3 (10.1–15)

Sources: [11,28].

).

The distribution across WHO categories underscores a broad urban burden: according to Table 2, a small group of industrial locations sits in the “well above interim targets” band (greater than 35 µg/m³), a wide middle tier occupies IT-2 (15.1–25 µg/m³), and only a few locations in the country meet or approach the guideline.

Seasonality explains much of these annual gaps. Multi-city analyses for Central Asia show that cold-season stagnation and shallow mixing layers amplify concentrations, placing cities such as Almaty and Astana among regional hotspots [5,29]. Episode-based trajectory work for 2021–2023 in Almaty and Astana confirms that air mass blocking and stagnation govern the worst winter peaks [3]. A Kazakhstan-specific multi-city assessment for Central Kazakhstan (Astana, Karaganda, and Zhezkazgan) independently reports sustained exceedances and strong winter–summer contrast, reinforcing the national pattern captured by the 2024 table [10].

Source evidence aligns with this picture. A recent receptor study in Astana resolves a large spark-ignition vehicle contribution to PM_2.5 alongside coal fly ash and secondary nitrate/sulfate, consistent with traffic and heating/power being the dominant urban sources [35]. Complementary measurements of black carbon show episodic surges during periods of blocking anticyclones and indicate contributions from CHP plumes, further clarifying the role of winter combustion [26]. Together, these findings explain why inland, coal-heated, and industrial cities tend to sit several-fold above the WHO annual guideline, whereas coastal or less industrial locations sometimes fare closer to IT-3 or IT-4 [6,11]. Beyond PM_2.5, concentrations of other pollutants in industrialized cities are also alarmingly high. According to national monitoring data, eight out of Kazakhstan’s fourteen largest cities had a “high” air pollution index (API ≥ 7) in 2019, indicating severe pollution levels [2].

Accordingly, the next step is to examine which sources dominate, and how their absolute and relative emissions have evolved, focusing on industry/power, residential heating, and on-road transport.

3.2. Key Sources of Urban Air Pollution in Kazakhstan

Key sources of urban air pollution are linked to the following drivers: coal-fired power and industrial emissions; residential heating with coal and biomass; transportation emissions; and others (

Table 3. Key sources of urban air pollution in Kazakhstan: indicators and recent trends (2018–2024).

Sector (Source Type)	What Predominates (Pollutants)	Absolute Indicators	Contribution to Population PM2.5 Exposure
Industry and power (stationary sources: CHP, metallurgy, and large coal boilers)	SO2, primary PM (ash/dust), NOx, CO, and NMVOC	2271.4 kt emitted from stationary sources in 2024, +0.6% vs. 2023; 2023: 2257.5 kt; and 2022: 2314.7 kt (BNS, 2022–2024 time series). Longer horizon shows values > 2.2–2.4 Mt (e.g., 2018 = 2446.7 kt).	Coal-fired CHPs significantly contribute to annual PM2.5 exposure; in Astana 7 μg/m3 (22%) from CHPs (World Bank modeling).
Residential heating (household stoves and small coal boilers)	Primary PM2.5, black carbon, SO2, CO	Absolute emissions are under-inventoried (many small, unpermitted sources). Household coal use remains widespread in peri-urban areas.	Largest single source of annual PM2.5 exposure in both capitals: Astana 10 μg/m3 (30%); Almaty 15 μg/m3 (25%) from household solid-fuel heating.
On-road transport (diesel and spark-ignition vehicles)	NO2/NOx, CO, primary PM, and VOCs → secondary nitrate PM2.5	Vehicle stock 5.16 million (October 2023) → 5.33 million (January 2024). Aging fleet; majority > 10 years. (BNS, 2023–2024).	Astana PM2.5 source apportionment: spark-ignition 30%, diesel 7%, and secondary nitrate 15% of PM2.5 (receptor PMF), confirming large traffic contribution to annual fine PM.

Sources: [30,31,32,33,34,35].

).

3.2.1. Coal-Fired Power Plants and Industrial Emissions

Heavy industry and power generation dominate ambient pollution in cities with major facilities. Kazakhstan’s economy remains energy-intensive and coal-dependent, and coal-fired combined heat and power (CHP) plants supply electricity and district heating in most large cities, emitting substantial PM, SO₂, NO_x, and CO [2]. For example, Almaty CHP-2 alone consumes in the order of millions of tonnes of coal per year; EBRD’s modernization documents report 2.5 Mt annually, underscoring its role as a large point source [36,37]. Metallurgical hubs (e.g., Temirtau and Ust-Kamenogorsk) and coal-based power lead to SO₂ levels that can reach multiples of the WHO’s guideline values, consistent with the high city Air Pollution Index (API) classes documented nationally. Longer-run permitting and emissions analyses indicate mixed trends over the past decade, with no decisive reduction in the overall industrial load prior to recent regulatory reforms [38,39,40,41,42].

3.2.2. Residential Heating with Coal and Biomass

Dispersed, small-scale heating is a leading source of fine particulate pollution in winter. In peri-urban areas not connected to district heating, households burn raw coal in stoves and small boilers, releasing pollution at low heights and directly in breathing zones during inversions. World Bank analyses attribute ≈10 µg m⁻³ (30%) of annual population-weighted PM_2.5 in Astana and 15 µg m⁻³ (25%) in Almaty to household solid-fuel heating, making clean heating measures the single most cost-effective lever for winter PM_2.5 reduction [43,44].

3.2.3. Transportation Emissions

On-road exhaust adds substantial NO₂, CO, primary PM, and precursors of secondary nitrate. Source apportionment in Astana resolves spark-ignition 30%, diesel 7%, and secondary nitrate 15% of annual PM_2.5, confirming traffic as a major contributor alongside coal-related factors [26]. Episode-based work across the two largest cities further links high-pollution days to stagnant synoptic regimes, with traffic and heating components co-varying under winter meteorology [3]. Policy analyses prioritize fuel/vehicle standards, inspection, fleet renewal, and area controls given documented benefits in comparable settings (e.g., London LEZ) [20].

3.2.4. Natural Factors and Transboundary Influences

Dust storms and natural emissions can episodically raise particulates, especially in arid regions; however, cold-season combustion overwhelmingly drives the worst urban episodes. Transboundary transport occasionally modulates events, but local sources dominate urban metrics in Kazakhstan’s large cities [2].

3.2.5. Seasonal Patterns

A sharp winter–summer contrast characterizes Kazakhstan’s urban air. Heating peaks and stable stratification lead to accumulation under shallow planetary boundary layers, while summers generally show better dispersion and lower coal use. For Almaty, boundary-layer analyses explicitly document the inverse coupling between PBL height and PM_2.5, and episode studies for Almaty and Astana show wintertime surges during stagnant conditions [3,5].

3.3. Air Quality Status and Trends in Astana

3.3.1. Observed Concentrations, Exceedances, and Wintertime Intensification

Ambient air quality in Astana is characterized by high concentrations of particulate and gaseous pollutants, especially in comparison to health-based guidelines. Other pollutants in Astana show a similarly concerning status. For instance, the annual average nitrogen dioxide (NO₂) concentration in urban Astana has been reported at levels greatly above the WHO’s guideline of 10 μg/m³.

Reported values of 1.74 mg·m⁻³ (1740 µg·m⁻³) for NO₂ do not represent an annual average and are implausible as such for urban backgrounds. We therefore treat them as episodic/short-term maxima at roadside locations and do not use them as annual means. Consistent with national monitoring, typical urban NO₂ in winter often exceeds 40–50 µg·m⁻³ at busy sites, indicating high traffic contributions; our seasonal lens emphasizes these winter elevations without over-interpreting episodic outliers.

Sulfur dioxide (SO₂) levels in Astana also frequently breach limits during winter. National standards for SO₂ (50 μg/m³ 24 h mean, roughly vs. the WHO guideline of 40 μg/m³) have been exceeded in multiple years. In 2017, Astana’s annual average SO₂ was around 0.93 mg/m³ (930 μg/m³), nearly twice the allowable limit. While subsequent years saw some reductions, winter spikes of SO₂ continue to coincide with the heating season, confirming that coal burning is a principal source.

Carbon monoxide (CO), though typically a localized pollutant, reaches high concentrations on polluted days; reports note that during severe winter smog episodes in Astana, CO levels have surged to 45–72 mg/m³ (over 10 times the typical limit of 4–6 mg/m³ for 24 h). These extreme CO levels occur in cold, calm conditions when emissions from vehicles and coal stoves accumulate, and they pose acute risks to cardiovascular health.

Ground-level ozone (O₃) in Astana is generally below the 8 h guideline of 100 μg/m³ for most of the year due to limited sunlight in winter and fewer VOC emissions relative to more industrialized or traffic-heavy cities. However, during summer heat waves or stagnant air, ozone episodes have been observed; Astana can experience afternoon ozone peaks that approach or occasionally exceed the guideline on hot sunny days. Ozone is expected to be a growing concern if temperatures rise with climate change and if traffic volumes increase, because it is a secondary pollutant formed from NO_x and volatile organic compounds.

Over the past decade, Astana’s air quality trends indicate that pollution remains a severe problem and, according to some measures, may have worsened before recent interventions. Kazakhstan’s national urban air pollution index (API), which aggregates multiple pollutants, showed Astana’s API nearly doubled from about 3.7 in 2005 to 7.0 in 2023. (Higher API signifies worse pollution; an API above 7 is considered “high” pollution by Kazakh standards.) This rise suggests that any improvements from cleaner technology were outpaced by increases in emissions due to urban growth. By the early 2020s, Astana at times registered among the highest pollution levels in the country, rivaling the traditionally more polluted industrial city of Almaty. For example, in the winters of 2021–2023, Kazhydromet noted that Astana’s average air pollution levels (especially PM_2.5) were at the top among monitored cities in Kazakhstan, reflecting the city’s rapid growth in emissions [29].

On a positive note, very recent data (2022–2023) hint at a plateau or slight improvement in some pollutant concentrations, potentially attributable to the early impacts of policy measures like partial coal-to-gas conversion (see Discussion). For instance, Astana’s average PM_2.5 in 2021 (22.5 μg/m³) was marginally lower than some earlier estimates for 2018–2019 (which were around 25–30 μg/m³). However, it is too soon to declare a definitive improving trend; year-to-year variability in weather (e.g., a warmer winter vs. a colder one) also influences these annual averages. Continued monitoring in coming years will be needed to confirm if Astana’s air pollution is decisively trending downward as mitigation efforts intensify. Figure 1 below maps energy infrastructure and air quality monitoring stations in Astana.

In Figure 1, red markers indicate combined heat and power plants (CHPPs), blue markers indicate air quality monitoring stations, and green markers indicate local boilers. The spatial distribution of Astana’s energy infrastructure illustrates a clear overlap between major emission sources and the monitoring network. Large combined heat and power plants are concentrated in the northern and northeastern sectors, while numerous local boilers are dispersed across residential peripheries, creating a cumulative pollution load that is systematically captured by the city’s air quality stations.

To conduct a more detailed analysis of air quality variations throughout the year, a new column was added to the dataset, categorizing data into heating (2 October 2023, to 19 April 2024) and non-heating (20 April 2024, to 1 October 2024) seasons. This approach enabled the assessment of seasonal influences on pollution levels, identifying characteristic trends for each period, and a more precise evaluation of air pollution dynamics based on seasonal variations in pollutant concentrations. As a result, the final dataset comprises 56,944 records (

Table 4. Descriptive Statistics of Air Pollutants During Heating and Non-Heating Periods (units: µg/m³).

Season	Statistics	CO	NO	PM10	PM2.5	SO2
Annual	Mean	543.56	8.21	21.99	11.54	18.66
	Std	556.41	27.90	33.30	18.43	16.88
	75th percentile	578.82	4.72	24.88	13.14	24.06
	25th percentile	291.17	0.00	5.54	1.64	8.51
Non-Heating	Mean	481.43	7.48	20.67	10.59	14.85
	Std	471.00	29.74	24.21	15.18	11.43
	75th percentile	513.51	4.83	24.53	11.44	18.69
	25th percentile	260.02	0.00	7.45	2.07	7.12
Heating	Mean	590.75	8.77	22.99	12.28	21.91
	Std	609.21	26.40	38.76	20.57	19.84
	75th percentile	624.91	4.58	25.11	14.78	28.61
	25th percentile	315.35	0.00	3.22	1.38	8.94

.).

The second dataset used in this study consists of meteorological open data obtained from the [39] website. The meteorological sensor is located in the capital of Kazakhstan at a latitude of 51°11′ N and a longitude of 71°26′ E, positioned at an altitude of 339 m above sea level. The dataset includes meteorological parameters recorded at three-hour intervals between 1 April 2023, and 31 March 2024:

• T: Air temperature (°C) at 2 m above the surface.
• RH: Relative humidity (%) at 2 m.
• WS: Mean wind speed (m/s) at a height of 10–12 m.
• WD: Mean wind direction (compass points) at a height of 10–12 m.

The analysis shows that particulate and gaseous pollutants in Astana rise markedly during the winter heating season, with the sole exception of nitric oxide (NO). Mean PM_2.5 increases from 10.6 µg m⁻³ in the non-heating period to 12.3 µg m⁻³ in winter, a 16% rise. PM₁₀ follows the same pattern: 23.0 µg m⁻³ in winter versus 20.7 µg m⁻³ in summer. Despite this, median values for both particulate fractions are lower in winter (PM_2.5 median 4.4 µg m⁻³; PM₁₀ median 9.6 µg m⁻³) than in summer (5.1 µg m⁻³ and 12.0 µg m⁻³, respectively). The discrepancy between the higher means and lower medians indicates that winter air quality is characterized by sporadic but severe pollution episodes, whereas background PM levels outside those episodes remain comparable to—or even below—summer baselines. In summer, the absence of extreme peaks is offset by persistent sources such as road dust and regional dust storms, which keep the median slightly elevated.

Sulfur dioxide (SO₂) and carbon monoxide (CO) show consistent and pronounced winter increases. Average SO₂ rises from 14.8 µg m⁻³ in the warm season to 21.9 µg m⁻³ in winter, an escalation of almost 48%. Median SO₂ climbs from roughly 15.8 µg m⁻³ to 21.1 µg m⁻³ over the same transition, reflecting the broad influence of coal combustion for district heating and small boilers. CO displays a 23% seasonal uplift: 481 µg m⁻³ in summer against 591 µg m⁻³ in winter. The entire interquartile range for winter CO shifts upward, meaning even the lowest quartile of winter concentrations surpasses typical summer values. This points to additional incomplete combustion sources in cold months—domestic stoves, low-efficiency boilers, and idling vehicles.

NO behaves differently. Its winter mean (8.8 µg m⁻³) and summer mean (7.5 µg m⁻³) are statistically indistinguishable (p ≈ 0.29). The median hovers at 1.3–1.5 µg m⁻³ year-round. Urban NO usually stems from traffic and power generation. Traffic volumes may fall slightly during extreme cold, counterbalancing higher cold-start emissions. Moreover, power plant NO_x is quickly oxidized to NO₂, which was not measured directly in this dataset. Other Kazakh cities that monitor NO₂ do observe clear winter increases, suggesting that the total NO_x burden does in fact rise in heating months even if NO alone does not register the change (

Table 5. Pollutants by type and by seasonal variations, in (µg m⁻³).

Pollutant	Heating Season Mean ± SD (µg m−3)	Non-Heating Season Mean ± SD (µg m−3)	Heating Season Median [IQR] (µg m−3)	Non-Heating Season Median [IQR] (µg m−3)
PM2.5	12.3 ± 20.6	10.6 ± 15.2	4.4 [1.4–14.8]	5.1 [2.1–11.4]
PM10	23.0 ± 38.8	20.7 ± 24.2	9.6 [3.2–25.1]	12.0 [7.4–24.5]
SO2	21.9 ± 19.8	14.8 ± 11.4	21.1 [8.9–28.6]	15.8 [7.1–18.7]
CO	590.7 ± 609.2	481.4 ± 471.0	433 [315–625]	363 [260–514]
NO	8.8 ± 26.4	7.5 ± 29.7	1.3 [0.0–4.6]	1.5 [0.0–4.8]

).

All seasonal contrasts are significant at p < 0.001 except for NO, confirming that NO remains stable while PM, SO₂, and CO intensify in winter. Extreme winter pollution events explain the elevated means for PM, whereas elevated SO₂ and CO medians indicate a sustained seasonal shift tied to coal combustion (

Table 6. Winter vs. summer contrasts for Astana (2023–2024): tests and p-values.

Pollutant	Test for Medians (Mann–Whitney U)	p-Value	Interpretation
PM2.5	U-test	<0.001	Winter intensification with episodic peaks (mean goes up, median goes down).
PM10	U-test	<0.001	Winter intensification; coarse dust modulates summer medians.
SO2	U-test	<0.001	Winter elevation consistent with coal combustion.
CO	U-test	<0.001	Winter elevation consistent with incomplete combustion.
NO	U-test	0.29	No significant seasonal change in NO alone.

.).

3.3.2. Automated Emissions Monitoring by Industries

On the emissions side, the government is pushing for better monitoring at the source. The national Strategy for Achieving Carbon Neutrality to 2060 (approved in 2023) explicitly calls for large emitters—power plants, factories, and boiler houses—to install automatic continuous emission monitoring systems (CEMSs). This is a significant regulatory step. With real-time smokestack data for SO₂, NO_x, particulate matter, and related pollutants, authorities can verify permit compliance and detect violations promptly. A range of monitoring technologies is under consideration or already in use, including in situ sensors, opacity monitors, and infrared or ultraviolet spectroscopy for gas analysis. Accuracy and reliability are essential, which requires careful method selection and calibration for Kazakhstan’s operating conditions. The shift to continuous emissions monitoring should strengthen enforcement of industrial limits and, if the data are published, increase public transparency. In practical terms, oversight moves from infrequent self-reported measurements to continuous observation [45].

3.3.3. Seasonal Variations and Pollution Episodes

Astana shows a marked seasonal cycle in air quality that reflects the contrast between winter and summer in a continental climate. Winter, which coincides with the heating season, brings substantially higher pollution, whereas summer is comparatively cleaner. Two drivers explain this pattern. First, emissions vary sharply by season. Second, meteorological conditions in winter inhibit dispersion. During November through March, temperatures frequently fall well below freezing, with midwinter averages around minus 10 to minus 20 °C. The cold requires extensive fuel use for space heating. The two major coal-fired combined heat and power plants in Astana, commonly referred to as CHP 1 and CHP 2, together with numerous boiler houses, increase combustion to supply district heating, and many suburban households burn coal or wood in stoves. As a result, particulate and gaseous emissions surge in winter and recede in summer [46].

Observations indicate that PM_2.5 concentrations are roughly three times higher in winter than in summer on average. In 2021 the winter mean in Astana was on the order of 30 to 35 μg m⁻³, compared with approximately 10 to 15 μg m⁻³ in summer. A statistical analysis of 2023 to 2024 data confirms the same pattern. The winter mean PM_2.5 was 12.3 μg m⁻³, which is 16 percent higher than the summer mean of 10.6 μg m⁻³. PM₁₀ shows a similar increase, with a winter mean of 23.0 μg m⁻³ and a summer mean of 20.7 μg m⁻³. Sulfur dioxide and carbon monoxide exhibit even larger winter surges. Average SO₂ rises by about 48 percent in winter, from 14.8 to 21.9 μg m⁻³, largely due to coal burning, and mean CO rises by about 23 percent, from 481 to 591 μg m⁻³, as small stoves and vehicle engines operate under cold conditions. These seasonal differences are highly statistically significant for all major pollutants except NO, with p less than 0.001, which indicates a genuine seasonal effect rather than random variability.

Interestingly, the median concentrations for particulates do not increase as much as the mean in winter, and in some cases are even lower than summer medians. For instance, median PM_2.5 in winter was ~4.4 μg/m³ compared to 5.1 μg/m³ in summer in one dataset. This counterintuitive result indicates that, outside of extreme episodes, baseline winter air can be relatively clear (especially on windy or mild days), whereas summer has steady moderate sources (dust, etc.) keeping the median up. The much higher winter mean vs. median implies that winter air quality is characterized by sporadic but severe pollution episodes rather than uniformly high pollution all the time. In practical terms, there are many winter days with fairly clean air (particularly when winds are strong or during thaws), but also some days with extremely high pollution that skew the average upward. These episodes correspond to cold, stagnant weather when emissions build up (e.g., multi-day cold snaps with little wind).

The meteorological conditions in winter strongly exacerbate pollution. Astana frequently experiences temperature inversions in winter, where a cold, dense layer of air near the ground is trapped under a layer of warmer air aloft. Under inversion conditions, pollutants emitted at ground level (from vehicles, house chimneys, etc.) cannot rise and disperse; instead, they accumulate in the shallow boundary layer over the city. Astana’s flat topography provides no natural ventilation (unlike cities in complex terrain where cold air can drain away), and high-pressure weather systems often bring very light winds. Studies have documented that, during strong Siberian anticyclone events, Astana can have essentially stagnant air for days, leading to multi-day smog episodes. Residents report visible haze or a “coal smoke” smell on these winter days. Measurements bear this out: on still, frigid mornings, PM_2.5 concentrations can spike well into the hundreds of μg/m³. In an extreme case reported in January 2020, Astana experienced PM_2.5 levels over 3000 μg/m³ (on an hourly basis) during a severe pollution episode—an extraordinarily high concentration, nearly 20 times the national permissible level, illustrating what can happen under worst-case inversion conditions. Although such extremes are rare, winter daily PM_2.5 peaks of 100–150 μg/m³ are not uncommon in Astana. Correspondingly, NO₂ and CO also accumulate to high levels during these episodes, occasionally prompting local authorities to issue pollution alerts when short-term indices become “hazardous”.

By contrast, summer (June through August) offers much cleaner air in Astana. There is virtually no need for heating, so coal combustion is minimal. The CHPs operate at reduced loads (mainly for electricity generation), and some are taken offline for maintenance in summer. Emissions from traffic also tend to be lower per vehicle, as engines operate more efficiently in warm temperatures (and there are no cold starts). Moreover, summer weather features more mixing and ventilation: Astana is known for being windy, especially in spring and summer, with frequent breezes that disperse pollutants. Convective turbulence during daytime (due to surface heating) helps break up any inversions. As a result, baseline PM_2.5 levels in mid-summer often drop to single-digit μg/m³, approaching clean continental background levels. For instance, on some of the cleanest days in July PM_2.5 emissions are around 3–5 μg/m³, which is near the WHO guideline value. Ozone is one pollutant that can be higher in summer due to stronger sunlight, but Astana’s ozone has remained moderate in most measurements, partly because the city’s relatively lower VOC emissions and its latitude limit peak ozone formation compared to cities in subtropical climates. One environmental concern in summer is dust: the steppe region can produce dust storms during dry spells (late spring or early summer), and construction activity in the growing city adds to particulate matter. These dust events can temporarily raise PM₁₀ (and to a lesser extent PM_2.5) even in the absence of combustion sources. However, such events are usually short-lived and occur sporadically. Overall, summertime air quality in Astana is often within Kazakhstan’s national standards and occasionally even within WHO guidelines, which starkly contrasts with the heavily polluted winter air.

Transitional seasons show intermediate conditions. In autumn, as temperatures fall in October and heating begins, concentrations rise from summer lows. Time series analyses identify a secondary peak in late October that reflects the start of the heating season combined with frequent inversions before very cold weather sets in. In spring, concentrations decline by April as heating demand ends and winds strengthen, although episodic events such as agricultural burning or long-range dust winds can produce short spikes. By May, air quality is substantially better than in winter.

Overall, Astana’s pattern is one of winter-dominated exposure: concentrations in the cold season are several times higher than in summer, and acute smog episodes are more common. Hospital admissions for asthma and bronchitis increase in winter, consistent with the higher exposure to irritants. The effectiveness of the city’s air quality measures will therefore be evaluated mainly by improvements in winter conditions, through cleaner heating and reductions in cold season traffic emissions.

3.3.4. Source Contributions to Pollution in Astana

Identifying the major sources of Astana’s air pollution is crucial for effective mitigation. As outlined earlier, the city’s emissions are dominated by fuel combustion in two sectors—(1) coal-based energy and heating, and (2) motor vehicle traffic—with additional contributions from dust and minor industrial sources. Here we present quantitative insights from emissions inventories and source apportionment studies to apportion Astana’s pollution to these sources.

Recent emission inventories (circa 2019) for Astana indicate that the city emits in the order of hundreds of thousands of tons of pollutants per year (including CO₂) from all sources combined. Focusing on air pollutants, the largest share comes from power generation and residential heating. A World Bank analysis identified coal combustion for heat and power as the top contributor to Astana’s PM_2.5, accounting for roughly half or more of the particulate pollution in winter. This includes emissions from the large CHP plants (e.g., Astana TPP-1 and TPP-2), which burn local coal and emit substantial fly ash, SO₂, and NO_x, as well as numerous smaller coal-fired boilers and household stoves especially in the city outskirts. The dominance of coal sources is evident in the chemical composition of particulate matter: coal combustion produces ash rich in minerals and sulfate. In a 2019–2021 receptor modeling study (PM_2.5 source apportionment in Astana), a factor interpreted as “Coal Fly Ash” contributed 17% of PM_2.5, and a “Secondary Sulfate” factor (sulfate particles derived from SO₂ emissions, mainly from coal) contributed another 8.5%. Additionally, a factor labeled “Local Power Plant(s)” with heavy ash content added 4%. Summing these coal-related factors, we see that in the order of 30% of Astana’s year-round PM_2.5 (and a much higher fraction in winter) comes from coal-burning sources (both direct particles and secondary formation). This aligns with winter observations that SO₂ and PM rise together, implicating coal as the common culprit. Coal’s share on the worst winter days is even larger—during high-smog episodes, one can attribute the majority of PM_2.5 to heating emissions. For example, one analysis found that in a January pollution episode, coal smoke markers (sulfate, potassium from coal, etc.) dominated the PM_2.5 mass, consistent with the fact that Astana’s winter smog is heavily a “coal smoke” smog.

The stationary-source emission totals reported here inherit uncertainties from their underlying activity data and emission factors. Data completeness varies by facility and season, and the seasonal allocation introduces additional uncertainty. We therefore present these values as indicative aggregates for contextualization and source apportionment discussion rather than as precise mass balances, and we cross-checked facility-level entries across public releases for internal consistency.

The other major contributor is motor vehicle emissions. Astana’s growth has led to a sharp increase in vehicles on the road (personal cars, buses, and trucks). Many of these are older or diesel vehicles lacking modern emission controls, and traffic congestion has worsened in the city. Vehicular emissions are rich in NO_x, CO, particulates (especially from diesel), and volatile organic compounds (VOCs). Source apportionment results showed a “spark-ignition vehicle” factor—essentially gasoline vehicle exhaust—contributing about 30.3% of PM_2.5 in Astana. A smaller “diesel” exhaust factor contributed 7.1%. Together, these indicate that roughly 37% of fine particulate pollution could be traced to motor vehicles (this includes primary carbonaceous particles and secondary nitrate formed from NO_x). Indeed, a “secondary nitrate” factor (15.1%) was identified, which is largely from NO_x emissions (vehicles and power plants) forming nitrate aerosols. A substantial portion of that nitrate likely originates from vehicle NO_x as well, given traffic’s dominance in city NO_x emissions. Thus, if we combine direct vehicle particles and vehicle-derived secondary nitrate, vehicles emerge as the single largest sector for PM_2.5 in Astana (on an annual basis), on par with or exceeding coal heating in some analyses. This is consistent with findings in Almaty, where traffic is the dominant NO₂ and a major PM source, and Astana shows a similar pattern albeit with a relatively greater coal impact than Almaty. Vehicles also emit most of the urban CO and a significant share of VOCs (like benzene or formaldehyde) that contribute to ozone formation. The heavy traffic contribution underscores that air quality management must address transport via cleaner fuels, vehicles, and mobility policies alongside cleaning up the heating sector.

Beyond combustion, dust plays a non-negligible role, especially in coarse particulate (PM₁₀). Astana’s expansion has involved continuous construction (new buildings and roadworks), and its dry steppe surroundings provide ample loose soil that can be lofted by wind. The source apportionment analysis identified a “soil/road dust” factor contributing about 7.9% of PM_2.5. This likely underestimates dust’s role in PM₁₀, because coarse particles were not fully captured in the PM_2.5 fraction. On particularly windy days with less combustion pollution (e.g., in spring), dust can become the dominant pollutant, though such days usually still have moderate PM levels overall. Dust particles are generally less toxic than combustion particles, but they can carry irritants and degrade visibility. There are also industrial and miscellaneous sources; Astana is not heavily industrial, but there are small factories and an oil refinery in the industrial outskirts that emit some SO₂, NO_x, and VOCs. These did not emerge strongly as separate factors in the PM_2.5 analysis, likely because their emissions are dwarfed by the major sectors or their signatures overlap (e.g., the refinery’s emissions might appear in the general fuel combustion factors). However, industrial emissions in nearby regions (outside the city limits) can occasionally influence Astana via regional transport—for example, a cement plant or mining operation tens of kilometers away might contribute a few μg/m³ under certain wind conditions. Seasonal open burning (such as agricultural field burning or illegal trash burning) is another intermittent source; while officially such burning is restricted, in practice some smoke from field fires or waste burning does occur, but it is hard to quantify and likely a minor contributor on an annual basis. Lastly, transboundary influences like forest fires or dust storms from other regions can occasionally raise pollution levels. For instance, smoke from large wildfires in Siberia or steppe fires in Kazakhstan/Russia can drift into Astana in summer, causing short-term haze and elevated PM; similarly, dust from the Karakum or other deserts can be transported during rare dust storm events. These episodes are infrequent compared to daily local sources, but they highlight that not all of Astana’s pollution is home-grown.

Receptor modeling (PMF) identified eight factors contributing to fine particulate matter. The largest contributors are vehicle emissions (spark-ignition gasoline vehicles 30% and diesel 7%) and coal combustion (fly ash 17%, secondary sulfate 8.5%, plus a power plant factor of 4%). Nitrate aerosols (~15%) derive from NO_x (largely traffic and some power plant). Soil/road dust contributes ~8%. These percentages (which sum to 100%) are approximate annual averages—in winter, the coal fraction would be higher, while in summer dust might be more prominent [35].

The implication of this source breakdown is clear: Astana’s air pollution is fundamentally an energy and transportation problem. Combustion of coal (for heat/power) and oil (fuel for vehicles) dominates the emissions. This is in line with the national picture: Kazakhstan’s air pollution is largely driven by legacy coal-fired infrastructure and an aging vehicle fleet, rather than, say, unique industrial processes. It also means the solutions lie in transforming those sectors—cleaner energy and cleaner transport (discussed in the next section). It is noteworthy that Astana’s profile, while severe, is somewhat less industrial than some other Kazakh cities; e.g., Karaganda or Temirtau have large steel or mining industries that contribute heavy metal pollution and additional SO₂ beyond just coal heating. Astana being a newer capital city has fewer heavy industries, but its rapid urbanization has made traffic and residential coal use the focal issues.

3.3.5. Comparisons and Health Implications

Placing Astana in context, we find its air quality is among the worst for capital cities of comparable size and income level, though not as extreme as the most polluted cities globally. Within Kazakhstan, Astana’s annual PM_2.5 (22–25 μg/m³ in recent years) is slightly lower than Almaty’s (35 μg/m³), which suffers from a basin geography trapping pollution. However, Astana has at times overtaken other cities to become the country’s most polluted city on certain metrics. Smaller industrial cities like Temirtau, Karaganda, or Ust-Kamenogorsk historically had higher peak pollutant levels (due to heavy industry), but Astana’s growth has narrowed the gap. Regionally, Astana’s winter PM_2.5 (50–100 μg/m³ on bad days) is similar to or somewhat better than Bishkek, Kyrgyzstan, and Dushanbe, Tajikistan, which also experience coal-smog winters. Ulaanbaatar, Mongolia, is a more extreme case (often >200 μg/m³ in January) and represents a worst-case coal heating scenario. In contrast, cold-climate cities that use cleaner energy, such as many in Europe (e.g., Helsinki or Oslo), maintain PM_2.5 well under 10 μg/m³ annually. Even large Eastern European cities like Warsaw or Moscow, after reducing coal use, have achieved 10–15 μg/m³ annual PM_2.5, showing that Astana’s 22 μg/m³ is not an inevitability of climate, but rather of fuel choices.

For NO₂, Astana and Almaty stand out with very high urban concentrations—in busy districts, NO₂ levels approach those seen in megacities known for traffic smog (on par with parts of Shanghai or Los Angeles). Almaty’s worst NO₂ hotspots (near traffic and coal boiler clusters) are extreme due to stagnation, while Astana’s citywide NO₂ is slightly lower thanks to more wind and fewer cars. Ozone is generally lower in Astana than in sunnier cities, as noted, which is a small relief in summer.

In global rankings (such as IQAir’s World Air Quality Report), Astana has periodically appeared among the top 30–50 capital cities in terms of poor air quality, especially when winter averages are considered. However, on an annual basis it fares better than the notorious mega-polluted capitals (Delhi, Dhaka, etc.), which have 80–100+ μg/m³ PM_2.5. Astana’s pollution is more episodic and seasonal, whereas those cities have high pollution year-round. If Astana succeeds in its current clean air initiatives, it could move closer to the ranks of cleaner cold capitals in the next decade [28].

The health implications of Astana’s pollution are significant. As noted, fine particulates and NO₂ contribute to respiratory illnesses, cardiovascular disease, and premature mortality [47]. Using epidemiological exposure–response functions, the World Bank estimated that achieving interim air quality gains such as a one third reduction in PM_2.5 could save thousands of lives each year across Kazakhstan, with Astana contributing a substantial share given its population size. Local health statistics indicate that hospital admissions for asthma, chronic bronchitis, and myocardial infarction rise during polluted winter months in Astana, although detailed city-specific studies are still emerging. The reported correlation between NO₂ and selected diseases in Astana (r = 0.95) suggests that lowering emissions from traffic and heating could yield measurable public health benefits through reduced incidence. Chronic exposure to PM_2.5 is associated with lung cancer; given current concentrations in Astana, elevated long-term risk would be expected, consistent with national concerns about high rates of lung cancer and chronic obstructive pulmonary disease that likely reflect both pollution and smoking. Economic effects are also material. Illness related to pollution increases healthcare expenditure and reduces productive workdays. For a growing urban economy, cleaner air supports a healthier workforce and sustains the city’s appeal to investors and event organizers. The next section considers how policymakers are responding and what progress has been made.

4. Discussions and Analyses

4.1. Policy Responses and Mitigation Measures

Improving air quality in Astana (and Kazakhstan more broadly) requires coordinated actions addressing the main pollution sources identified: coal combustion and vehicle emissions. In recent years, the Kazakh government, along with city authorities and international partners, has initiated a suite of policy measures and technological interventions aimed at cleaner air. These efforts recognize that pollution control aligns with other national goals, including climate change mitigation (since cutting coal use also lowers CO₂) and public health improvement.

Transition from Coal to Cleaner Fuels: The most significant ongoing effort is the gasification of Astana’s energy supply. Historically, Astana’s heat and power came from burning coal, but the government invested in the Saryarka natural gas pipeline (completed in 2019) to bring gas from western Kazakhstan to the capital. Since then, a major campaign has been underway to connect city districts and individual households to the new gas network. By mid-2024, around 12,000 households in Astana had switched from coal stoves to gas heating, and new gas boiler plants were constructed to replace some coal-fired units [37]. The plan aims for the full gasification of Astana by 2025, meaning most coal-fired boilers and many stoves would be phased out, and the large CHPs would either be converted to burn gas or supplemented by new gas power plants [38]. The impact of this transition cannot be overstated: burning natural gas emits negligible particulate matter and far less SO₂ than coal (gas contains very little sulfur and produces no ash). NO_x emissions also drop per unit heat, and gas combustion can be more easily controlled for NO_x (through low-NO_x burners, etc.).

Model projections by the World Bank found that if Astana achieves full residential and district heating gasification, winter PM_2.5 concentrations could be reduced by over 50% [33,34]. Early signs in the winter of 2022–2023 were promising: neighborhoods that switched to gas reported noticeably cleaner air and less smoke smell on cold days, and citywide average PM_2.5 ticked downward slightly compared to prior winters. However, the full benefit awaits the near-complete elimination of coal after 2025. The gasification policy is also popular due to its climate co-benefits—Kazakhstan’s CO₂ emissions per GDP are among the highest in the world from coal dependence, so this helps on that front as well. Challenges remain in implementation. Authorities must ensure that the gas network can meet peak winter demand, that households can afford conversion, and that contingency plans are in place. The government has provided subsidies for low-income households to install gas boilers, and initial delays and occasional gas shortages have been addressed [39]. To secure equitable benefits from clean heating, Astana pairs the gas network rollout with targeted subsidies for low-income households that cover connection and appliance conversion. This design lowers the risk of energy poverty and raises uptake, which is essential for delivering winter PM_2.5 reductions where the health burden is greatest [35].

In parallel, the city is diversifying its energy mix and improving efficiency. Astana has substantial solar potential, so installation of solar panels and investment in renewable energy are on the agenda. Energy efficiency measures, including the better insulation of buildings, are being promoted because a well-insulated home may use half the fuel of a poorly insulated one, which directly reduces emissions [40]. The Environmental Code adopted in 2021 requires large plants to modernize with cleaner technology or face higher emission fees. For Astana this has meant, for example, that coal-fired combined heat and power plants must retrofit filters or electrostatic precipitators to capture more fly ash while they remain in service [41]. Officials note that provisions on Best Available Techniques encourage upgrades and, where upgrades are not feasible, eventual retirement of obsolete units. Air quality and climate objectives are aligned. Kazakhstan’s carbon neutrality strategy for 2060 implies a phase out of unabated coal in power and heating, which is consistent with the need to eliminate coal related air pollution [38]. In sum, the shift to cleaner energy remains the cornerstone of Astana’s mitigation strategy, with natural gas in the short term and renewables and efficiency in the longer horizon, and progress is already visible on these fronts.

4.2. Transportation Emissions Control

The second pillar of clean air policy concerns vehicle emissions. Astana is modernizing public transport by renewing the bus fleet and introducing buses that run on compressed natural gas, with a smaller cohort of electric buses. This lowers emissions from one of the city’s most persistent sources, namely transit buses [42]. The city is also extending cycling infrastructure and pedestrian areas to provide viable alternatives to private car travel. At the national level, fuel quality standards have been strengthened. Low-sulfur gasoline and diesel that meet Euro 5 specifications are now available, which reduces SO₂ from vehicles and improves combustion efficiency [43].

Electric mobility is expanding. Kazakhstan adopted incentives for electric vehicle imports, and between 2022 and 2023 the number of electric cars reportedly tripled, albeit from a small base [44,45]. As the capital, Astana has attracted a large share of this growth, and charging infrastructure is being installed across the city. Urban planning measures such as new bypass roads and adaptive traffic signal control are used to ease congestion and cut idling emissions. A priority measure under review is the reintroduction of routine vehicle emission inspections, together with possible Low-Emission Zones in the city center where high-emitting vehicles would be restricted. Almaty is piloting related steps, and Astana is assessing feasibility [46]. International advisors, including the World Bank, recommend these actions and note that curbing vehicle emissions yields immediate local benefits, especially at NO₂ hotspots [48,49,50,51].

Transport policy remains demanding because it requires behavioral change and sustained investment. Progress is gradual, the vehicle fleet continues to grow, and many older cars remain on the road for economic reasons. Even so, stricter standards and a shift to cleaner modes can directly lower chronic NO₂ and CO levels. Over the next decade, an expansion of electric mobility together with the enforcement of Euro 5 and Euro 6 standards would be expected to produce a marked decline in roadside NO₂.

4.3. Industrial Emissions and Regulatory Framework

Astana is not a major industrial center, yet recent reforms of Kazakhstan’s regulatory framework have indirect benefits for the capital. The Environmental Code adopted in 2021 brought national rules closer to European and international practice. Key provisions require large cities to prepare local air quality action plans, mandate detailed emissions inventories that capture all sources, and introduce Best Available Techniques for industrial facilities [41].

The Code also tightened the system of environmental permits and emission limit values for industry, replacing relatively lenient standards inherited from the Soviet period [42]. Under the earlier approach, a facility could receive a limit that just met a hygiene norm at the fence line, which often resulted in weak controls. The new approach seeks alignment with health-based benchmarks such as WHO guideline values and requires upgrades to pollution control in order to meet those levels. For Astana this implies that any significant industrial site, including the small refinery and municipal waste incineration, must adopt better abatement.

The law further strengthens public access to environmental information and formal participation by stakeholders. This transparency has enabled local organizations in Almaty and Astana to request stronger enforcement and to monitor commitments. Implementation remains the central challenge. As of 2023, the government was still developing the secondary regulations that put the Code into effect, including definitions of Best Available Techniques by branch, city-specific air quality targets, and capacity for enforcement [52,53,54,55,56,57]. Effective enforcement requires reliable data. Expansion of monitoring through continuous emissions monitoring at stacks and denser ambient networks is necessary to make the new standards operational. In parallel, Astana’s administration and the Ministry of Ecology are preparing an Air Quality Management Plan that consolidates measures such as gasification and transport control with explicit targets for PM_2.5 and other pollutants for 2025 and 2030.

4.4. Public Awareness and Civil Society

Public awareness and engagement on air quality in Kazakhstan have grown and now influence policy. Evidence from Turin and Dublin shows that perceived air pollution affects self-rated health and mobility choices, and that denser monitoring networks correlate with higher institutional trust, which helps explain why transparency and citizen-facing data tools support policy uptake [48]. For many years air pollution was a largely hidden problem because data were not easily accessible and the health impacts were poorly understood. This pattern is changing. The release of real-time information through the AirKZ application and public websites allows citizens to identify periods of poor air quality. Widely reported pollution episodes have prompted national discussion, including the decision in November 2023 by authorities in Ust-Kamenogorsk to move schools to remote learning because of smog [47,49]. In Astana, civic groups and residents organize awareness campaigns and protests that call for cleaner air, and physicians increasingly frame air pollution as a public health problem that requires prevention rather than treatment alone. This growing pressure from the public accelerates governmental action and supports enforcement as elected officials respond to constituent concerns. Government bodies now involve non-governmental organizations in consultations; Kazakhstan has joined international initiatives such as the UNECE Batumi Action for Clean Air and has hosted policy dialogs where officials from Astana and Almaty exchange experience with peers from Europe and Asia. As monitoring expands and transparency improves, citizens are more likely to support difficult measures such as restricting high-emitting vehicles or adjusting utility prices to finance cleaner fuels, because the health rationale is clearer. While central policy remains essential, sustained demand from the public helps maintain momentum and oversight.

Astana’s air quality, although still poor, is on a path where substantial improvement is plausible within the next decade. The city has identified the main causes in outdated energy and transport systems and has begun to implement measures that have proved effective elsewhere, including a shift to cleaner fuels, cleaner vehicles, and stronger environmental regulation. Early signs are cautiously encouraging, with stabilization or slight declines in pollution and a visible rise in public engagement. If current policies are carried through and maintained, Astana could move toward levels observed in cleaner peer capitals, with large health and economic gains for residents. The case also shows that even in demanding climatic and institutional settings, air pollution can be reduced with political commitment, public support, and policies grounded in documented effects. Ongoing monitoring and research in Astana will guide adjustments and help ensure that expected benefits are delivered for the population of Kazakhstan’s capital [58,59,60,61,62].

5. Conclusions

A synthesis of published studies and observational records indicates that PM_2.5, NO₂, SO₂, CO, and O₃ are the predominant pollutants in Astana, and that their emissions arise chiefly from the energy and transport sectors; over several decades, researchers have developed a suite of models, ranging from Gaussian plume formulations to three dimensional chemical transport simulations, that together allow concentrations to be estimated and forecast under a variety of meteorological and emissions regimes. Drawing on data from regulatory ground stations and complemented by newer sensor networks, the literature documents not only the overall scale of pollution but also the recurrent winter smog episodes that exceed health guidelines by considerable margins, while Astana-focused analyses consistently attribute the bulk of the burden to coal combustion for heat and power and to vehicle exhausts, with smaller contributions from dust and a limited set of ancillary activities; taken together, these findings provide a coherent empirical basis for mitigation targeted at the dominant sources.

Against this backdrop, policy measures at both municipal and national levels are moving forward, as the fuel mix shifts from coal toward natural gas and renewables, transport policy is reinforced through investment in public transit, stricter fuel standards, and tighter vehicle controls, and environmental regulation and enforcement are strengthened to support implementation; critically, source apportionment studies in combination with health assessments converge on the conclusion that reducing coal smoke and traffic emissions would yield substantial public health gains, thereby aligning scientific diagnosis with the current direction of policy.

For researchers and policymakers, Astana illustrates that long-known urban air pollution problems persist in parts of Central Asia and can be addressed through appropriate technology and well-designed policy. Continued evaluation is needed to confirm, for example, whether PM_2.5 declines as gas infrastructure expands, and to assess how warmer winters and changing wind patterns may alter dispersion and secondary formation. Further studies would be valuable on indoor air quality given household use of coal and biomass, on the influence of urban form on dispersion including possible street canyon effects, and on the utility of dense sensor networks with statistical forecasting for operational management. Future appraisals should incorporate health impact assessment alongside standard air quality metrics to quantify avoided illness and premature mortality from cleaner heating and transport.

Astana is moving toward cleaner air and offers lessons that may transfer to cities with similar conditions. The experience to date suggests that strong data, informed public engagement, and policy grounded in evidence can place even heavily polluted urban areas on a credible path to healthier air. The next decade will be crucial to turning plans into measurable improvements and to positioning Astana as a reference case for effective air quality management in cold-climate, coal-reliant settings.