Breathing Cities: Air Quality, Population Exposure, and Sustainability Implications in 33 European Capitals

Abstract

This study investigates long-term urban air quality and population-level exposure in 33 European capital cities between 2010 and 2024. Using over 3.5 million hourly observations retrieved from official monitoring networks, city-level Air Quality Index (AQI) values were calculated and analyzed for exceedance frequency, seasonal dynamics, and spatial disparities. To account for public health relevance and sustainability implications, the analysis incorporated population-weighted exposure indicators reflecting both pollution severity and urban demographic scale. The results reveal substantial differences in air quality across the continent: cities such as Oslo and Reykjavik consistently maintain low AQI levels, while Sarajevo, Lisboa, and Madrid experience frequent exceedances. Notably, Paris shows the highest cumulative population exposure despite moderate pollution intensity, highlighting how urban density amplifies public health burdens. The use of harmonized AQI and exposure-adjusted metrics enables standardized comparisons across cities and supports sustainability-oriented urban planning. By quantifying unequal exposure burdens across Europe’s capitals, this study contributes to the evidence base for Sustainable Development Goal 11, emphasizing the need for data-informed air quality policies that address both environmental risks and population vulnerability.

1. Introduction

Urban air pollution remains a pressing environmental and public health challenge across Europe, with millions of urban residents exposed to pollutant levels exceeding health-based guidelines [1]. Air pollution contributes to a wide array of adverse health outcomes, including cardiovascular and respiratory diseases, premature mortality, and increased hospital admissions [2,3].

Urban air quality management is not only a technical challenge but also an institutional and governance issue. As air pollution is intricately tied to urbanization, mobility systems, and energy consumption, its effective mitigation requires integrated, adaptive policy frameworks that bridge environmental science with urban planning, governance, and innovation strategies.

The concept of sustainability in urban contexts involves more than environmental protection; it requires the coordination of environmental, economic, and social objectives through institutional innovation and participatory governance mechanisms [4]. The governance of air quality in the EU, particularly through instruments such as the Ambient Air Quality Directive (2008/50/EC), reflects a hierarchical yet increasingly multilevel and knowledge-based approach to environmental policy [5]. However, challenges persist in translating top-down regulation into effective local implementation, often due to policy fragmentation or limited institutional capacity [6].

Recent frameworks advocate the integration of urban transition management principles into air quality policy, emphasizing adaptive learning, stakeholder participation, and long-term visioning. This approach is consistent with broader sustainable development agendas, where innovation is leveraged to produce co-benefits for public health, equity, and climate mitigation [7]. For instance, the application of transition labs and citizen sensing, as demonstrated in cities such as Aachen, demonstrates how bottom-up engagement and experimental governance can improve both legitimacy and effectiveness of urban air quality strategies [6]. Furthermore, innovation-driven policy frameworks have shown success in contexts outside the EU. China’s National Innovation Demonstration Zones (NIDZs) have promoted cleaner energy consumption and encouraged industrial transformation to reduce urban air pollution, offering a model for how innovation policy can contribute directly to sustainable urban development [8].

This study builds on these conceptual frameworks by analyzing and comparing urban air quality conditions across 33 European capital cities using the composite Air Quality Index (AQI). The goal is to contribute to sustainability-oriented urban governance by demonstrating how standardized air quality metrics can support more transparent, comparable, and actionable insights for decision-makers. In doing so, the study aligns with the objectives of the EU Ambient Air Quality Directive (2008/50/EC), which emphasizes both environmental protection and the integration of scientific knowledge into multilevel governance frameworks.

The Air Quality Index (AQI) is a widely used tool for communicating the health implications of air quality in an accessible and standardized form, aggregating multiple pollutants into a single indicator [9]. While AQI thresholds are not perfect proxies for toxicological risk, they are effective for identifying periods of heightened exposure risk, particularly for vulnerable populations such as the elderly, children, and individuals with chronic health conditions [10].

European countries differ significantly in their air quality profiles due to a range of factors, including climatic conditions, geographic location, urban design, transportation systems, and regulatory enforcement [11]. Understanding these variations is crucial, as recent studies indicate that temporal and spatial variability in air pollution is critical for assessing population exposure and health risks, especially in urban environments [12,13]. For example, southern cities often experience higher ozone concentrations during spring as well as intrusions of Saharan dust, while eastern and central cities frequently record elevated particulate matter levels in winter and increased levels of ozone in summer [14,15,16]. Moreover, the interaction of pollution frequency, severity, and population size is essential for evaluating cumulative exposure burdens [17].

While the European context is the primary focus of this study, comparable challenges are evident across other rapidly urbanizing regions. In countries such as China and India, large-scale urban expansion and industrialization have resulted in critical environmental challenges. In China, for example, severe urban air pollution has prompted innovative policy responses, including the establishment of National Innovation Demonstration Zones, which have delivered measurable improvements in urban air quality through reduced energy consumption and emissions [8]. Similarly, integrated air quality management strategies in India stress the importance of linking emissions control to sustainable development goals, recognizing the role of institutional coordination and cost-effective policy measures [7]. Epidemiological evidence across East Asia also underscores the direct health consequences of air pollution, with increased risks of respiratory infections and mortality even from short-term exposure to particulate matter [18]. These findings collectively highlight the urgency of implementing effective mitigation measures within urban governance frameworks.

In the United States, urban air pollution remains a major public health concern, particularly due to the chronic effects of fine particulate matter (PM_2.5) and ground-level ozone. Epidemiological research has consistently shown strong associations between air pollution and increased risks of cardiovascular and respiratory mortality, especially among vulnerable populations such as the elderly, low-income residents, and those with pre-existing health conditions [19]. Urban form plays a key role in shaping air pollution levels, with sprawling and fragmented city layouts correlating with higher emissions from vehicular traffic and lower environmental efficiency [20]. Integrated assessments also underscore the compounded burden of air pollution, noting that socio-economic disparities often amplify both exposure levels and health vulnerability in marginalized communities [21]. Recent model-based analyses indicate that targeted reductions in specific emission sources—such as diesel engines—could yield substantial societal health benefits, with estimated avoided costs of hundreds of billions of dollars annually in the United States and Canada [22].

Beyond the Northern Hemisphere, many other regions are facing critical air quality challenges that underscore the global nature of urban air pollution. In Latin America, rapid urban growth combined with limited regulatory enforcement has resulted in persistent air pollution in cities such as Mexico City, Santiago, and São Paulo. Exposure to PM_2.5 in these regions has been linked to elevated rates of cardiovascular and respiratory mortality, particularly among vulnerable populations such as children and the elderly [23,24,25]. These health impacts underscore the need for better integrated governance and environmental management systems in rapidly urbanizing Latin American cities [26].

In Africa, where data availability is more limited, air pollution is an emerging crisis, exacerbated by unplanned urbanization, inefficient energy systems, and high reliance on biomass fuels. Studies using remote sensing data indicate that urbanization has strong spatial correlations with worsening air quality, particularly in Central Africa, where increased nightlight intensity—used as a proxy for urban growth—is associated with elevated aerosol optical depth (AOD) levels [27]. Health consequences are already substantial: recent estimates indicate that air pollution contributes to nearly 800,000 premature deaths annually across the African continent [28]. Yet, the lack of regulatory infrastructure and monitoring capacity continues to limit mitigation efforts.

Meanwhile, in high-income countries such as Australia, air pollution remains a significant but more preventable risk, with modeling studies indicating that substantial health and economic benefits can be achieved through targeted reductions in PM_2.5 and ozone exposure [29]. Cost–benefit analyses also underscore the disproportionate public health returns of air quality interventions in urban areas, particularly when targeting transport and energy sectors [30]. Despite the availability of hourly pollution data across European monitoring networks, no studies have analyzed long-term temporal dynamics in the AQI across multiple capitals, nor have they combined AQI pollution data with population-scale exposure metrics. This study aims to fill that gap by analyzing over a decade of hourly AQI data (2010–2024) for 33 European capitals. The analysis explores both general trends and city-level case studies, with a focus on pollution intensity, exceedance patterns, seasonal cycles, and population-normalized exposure burdens. Four cities—Paris, Oslo, Lisboa, and Warsaw—were selected for in-depth comparison, representing different regions, population sizes, and air pollution regimes. By identifying spatial and temporal patterns in air quality and underscoring periods of high health risk, this work seeks to contribute to evidence-based strategies for urban air pollution mitigation and public health planning.

An important methodological innovation of this study lies in the unified calculation of the Air Quality Index (AQI) from raw hourly pollutant concentrations across all capital cities, rather than relying on nationally reported indices that often vary in thresholds and composition. This harmonized approach ensures cross-country comparability and enables consistent spatial and temporal analysis. Furthermore, by integrating population-weighted exposure estimates and multi-scale temporal diagnostics—from hourly to decadal—the study provides a comprehensive framework for interpreting urban air quality with direct relevance for sustainable urban planning and public health governance.

This study also contributes to sustainability science by integrating air quality monitoring, population-scale exposure metrics, and urban environmental disparities into a unified analytical framework. In doing so, it supports Sustainable Development Goal 11, which emphasizes the need for inclusive, safe, resilient, and sustainable cities. By identifying spatial and temporal patterns of air pollution alongside population vulnerability, the findings aim to inform equitable and evidence-based strategies for urban sustainability and public health protection.

2. Materials and Methods

2.1. Data Collection

Hourly air quality data were obtained using the “OpenAir” R package (version 2.18.0) in combination with the “saqgetr” library. This setup enables programmatic access to the European Environment Agency (EEA) Air Quality e-Reporting system via its open RESTful API, ensuring direct retrieval of validated and harmonized air quality measurements reported by EU member states under Directive 2008/50/EC [31,32]. All data were openly available and collected from officially designated monitoring stations.

The analysis focused on capital cities across Europe, excluding microstates such as Andorra, Monaco, Luxembourg, and San Marino due to their limited spatial coverage and data availability. Data from Ukraine, Belarus, Russia, and Armenia were excluded due to the lack of consistent records for the full study period.

For cities with multiple monitoring stations, a multi-criteria selection process was applied to ensure consistency and comparability. This selection prioritized:

• Temporal coverage: stations with data spanning from 2010 to 2024;
• Spatial context: exclusively urban-area stations;
• Monitoring type: preference for background stations (when available) to minimize local source influence;
• Proximity: locations situated close to the administrative center of the capital city.

This rigorous protocol ensured the selection of stations representative of general urban air quality rather than highly localized emissions.

Table 1. Characteristics of selected urban air quality monitoring stations in European capitals (2010–2024).

Site Code	Lat.	Lon.	Capital City	Population *	Country	Site Type	Missing Data
nl00014	52.359714	4.866208	Amsterdam	1,131,690	Netherlands	background	2%
gr0002a	37.978206	23.726842	Athens	3,155,600	Greece	traffic	6%
rs0032a	44.82112	20.459113	Belgrade	1,389,351	Serbia	background	9%
debe034	52.489451	13.430844	Berlin	3,552,123	Germany	background	8%
ch0031a	46.951	7.44087	Bern	429,920	Switzerland	traffic	1%
sk0004a	48.144672	17.113576	Bratislava	422,153	Slovakia	background	7%
betr001	50.849710	4.334103	Brussels	2,049,510	Belgium	background	1%
ro0070a	44.435044	26.098297	Bucharest	1,821,380	Romania	traffic	48%
hu0041a	47.508056	19.028067	Budapest	1,759,497	Hungary	traffic	3%
dk0045a	55.700279	12.561400	Copenhagen	1,320,826	Denmark	background	7%
ie0028a	53.353889	−6.278056	Dublin	1,201,426	Ireland	background	3%
fi00425	60.18739	24.9506	Helsinki	1,279,096	Finland	background	1%
pt03087	38.705	−9.210278	Lisboa	2,927,316	Portugal	background	1%
si0003a	46.065544	14.512720	Ljubljana	286,491	Slovenia	background	2%
gb0566a	51.52229	−0.125889	London	9,046,485	United Kingdom	background	2%
es1422a	40.419166	−3.703333	Madrid	3,174,624	Spain	background	3%
no0072a	59.91975	10.68973	Oslo	1,012,225	Norway	background	6%
fr04004	48.891666	2.346667	Paris	10,900,952	France	background	9%
me0002a	42.440973	19.256617	Podgorica	177,177	Montenegro	traffic	41%
cz0arep	50.088066	14.429220	Praha	1,291,552	Czechia	background	2%
is0006a	64.13886	−21.87341	Reykjavik	216,364	Iceland	background	35%
lv0026a	56.954847	24.104756	Riga	637,089	Latvia	background	87%
it0755a	44.215	12.056389	Rome	4,209,710	Italy	traffic	7%
ba0029a	43.867496	18.423061	Sarajevo	342,577	Bosnia And Herzegovina	background	41%
mk0043a	41.99917	21.44084	Skopje	584,208	North Macedonia	traffic	15%
bg0050a	42.680558	23.296786	Sofia	1,272,418	Bulgaria	background	2%
se0022a	59.316007	18.057807	Stockholm	1,582,968	Sweden	background	0%
ee0018a	59.414169	24.649457	Tallinn	437,027	Estonia	background	2%
al0201a	41.330269	19.821772	Tirana	475,577	Albania	traffic	80%
at9stef	48.20815	16.373254	Vienna	1,900,547	Austria	background	2%
lt00002	54.686111	25.21083	Vilnius	536,055	Lithuania	background	4%
pl0143a	52.290864	21.042458	Warsaw	1,767,798	Poland	background	1%
hr0007a	45.800338	15.974072	Zagreb	685,587	Croatia	traffic	3%

* According to the United Nations 2018 [33].

provides detailed metadata on the final set of monitoring sites used in the study, including the site code, city and country, site classification, station type, and percentage of missing data across the 15-year period. In most cases, urban background monitoring stations located near city centers were selected to ensure comparability. However, due to data availability constraints, traffic monitoring stations were used in several cities, which may introduce variation in local pollution levels due to differences in station surroundings and land-use types.

The final dataset included 33 stations across 33 European capital cities, covering a total period of 15 years (2010–2024). The dataset contains hourly observations of six key pollutants—PM₁₀, PM_2.5, CO, NO₂, O₃, and SO₂—all of which contribute to the composite Air Quality Index (AQI) as defined by the EEA. In total, the final dataset with AQI values comprises over 3.5 million individual data points.

2.2. Research Design

This study followed a temporal observational design based on the analysis of air quality data collected over a 15-year period (2010–2024) from monitoring stations located in the capital cities of 33 European countries. For each city, one representative air quality monitoring site (classified as urban-background or traffic-related) was selected based on data availability, completeness, and comparability, as described in the Section 2.1 “Data Collection.”

The analysis began with the calculation of hourly composite Air Quality Index (AQI) values from raw pollutant concentration data in RStudio. The AQI was computed following standardized procedures that normalize concentrations of key pollutants (e.g., PM_2.5, NO₂, and O₃) into a unified scale. These values served as the foundation for all subsequent analyses.

Temporal dynamics were then explored using the “timeVariation” function from the openair R package, which facilitated the decomposition of AQI patterns into diurnal, weekly, and monthly cycles. Also, seasonal patterns were analyzed for selected cities. In addition, to better understand the societal implications of urban air pollution, population-based indicators were calculated, incorporating both population size and the intensity of exposure to air pollution in each capital city. These indicators aimed to approximate the cumulative burden of air pollution on urban residents and provide a more policy-relevant assessment of air quality levels across different demographic contexts.

In the final step, long-term trends in annual average AQI were calculated using simple linear regression models for each capital city. The results included trend slopes, standard errors, and p-values, allowing for the assessment of whether air quality conditions were improving, deteriorating, or remaining stable over time.

2.3. AQI Calculation

The Air Quality Index (AQI) serves as a standardized indicator of air pollution, enabling the comparison of health-related air quality impacts across pollutants and locations. The AQI is calculated individually for each pollutant using the formula provided by the U.S. Environmental Protection Agency (EPA) methodology [9]:

$$\mathrm{A}\mathrm{Q}\mathrm{I}={\displaystyle \frac{{\mathrm{I}}_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}-{\mathrm{I}}_{\mathrm{l}\mathrm{o}\mathrm{w}}}{{\mathrm{C}}_{\mathrm{h}\mathrm{i}\mathrm{g}\mathrm{h}}-{\mathrm{C}}_{\mathrm{l}\mathrm{o}\mathrm{w}}}}\left(\mathrm{C}-{\mathrm{C}}_{\mathrm{l}\mathrm{o}\mathrm{w}}\right)+{\mathrm{I}}_{\mathrm{l}\mathrm{o}\mathrm{w}},$$

(1)

where

• AQI—the Air Quality Index;
• C—the pollutant concentration;
• C_low and C_high—the concentration breakpoints for C;
• I_low—the index breakpoint corresponding to C_low;
• I_high—the index breakpoint corresponding to C_high.

The breakpoint values for each pollutant—such as ozone (O₃), particulate matter (PM_2.5, PM₁₀), carbon monoxide (CO), sulfur dioxide (SO₂), and nitrogen dioxide (NO₂)—were adopted from the EPA-defined AQI categories and are summarized in

Table 2. AQI breakpoints and concentration thresholds for common air pollutants (based on U.S. EPA standards).

Ilow–Ihigh	O3, ppb *	O3, ppb **	PM2.5, μg/m3 ***	PM10, μg/m3 ***	CO, ppm **	SO2, ppb	NO2, ppb *
0–50	-	0–54	0.0–12.0	0–54	0.0–4.4	0–35 *	0–53
51–100	-	55–70	12.1–35.4	55–154	4.5–9.4	36–75 *	54–100
101–150	125–164	71–85	35.5–55.4	155–254	9.5–12.4	76–185 *	101–360
151–200	165–204	86–105	55.5–150.4	255–354	12.5–15.4	186–304 *	361–649
201–300	205–404	106–200	150.5–250.4	355–424	15.5–30.4	305–604 ***	650–1249
301–400	405–504	-	250.5–350.4	425–504	30.5–40.4	605–804 ***	1250–1649
401–500	505–604	-	350.5–500.4	505–604	40.5–50.4	805–1004 ***	1650–2049

* 1 h, ** 8 h, *** 24 h.

. The AQI categories and their associated health implications are presented in

Table 3. AQI categories and their associated health implications (based on U.S. EPA standards).

AQI Category	AQI Range	Health Implications
Good	0–50	Air quality is considered satisfactory, and air pollution poses little or no risk to health.
Moderate	51–100	Air quality is acceptable; however, there may be a risk for a small number of people who are unusually sensitive to air pollution.
Unhealthy for sensitive groups	101–150	Members of sensitive groups (e.g., people with respiratory or heart disease, older adults, and children) may experience health effects. The general public is unlikely to be affected.
Unhealthy	151–200	Everyone may begin to experience health effects; sensitive groups may experience more serious effects.
Very unhealthy	201–300	Health alert: everyone may experience more serious health effects.
Hazardous	≥301	Health warning of emergency conditions: the entire population is more likely to be affected.

.

2.4. Population Exposure Metrics and Threshold Analysis

To assess general population-level exposure to air quality exceedances, a series of indices derived from the hourly AQI values were computed, as defined in Section 2.3. Besides basic statistics like boxplots and annual and seasonal means, several required additional explanation.

It is important to note that this study uses AQI as an ambient, general population-level indicator and does not account for individual exposure durations or intra-population differences in vulnerability, such as age, health status, or socio-economic conditions.

2.4.1. AQI Class Frequency Distribution

Each hourly AQI observation was categorized according to U.S. EPA-defined thresholds (0–50, 51–100, 101–150, 151–200, 201–300, and ≥301). The frequency of hours falling into each category was calculated as a proportion of the total number of hours in the dataset, adjusted to account for missing data. This normalization approach ensured comparability across cities and timeframes with variable data completeness.

2.4.2. Exceedance Frequency (AQI > 100)

To focus on health-relevant exposures, the total number and percentage of hours with AQI exceeding 100 (the “Unhealthy for Sensitive Groups” threshold) were computed.

2.4.3. Mean AQI Excess Above Threshold

To quantify the severity of exposure, a new metric was defined—mean AQI excess above 100:

$$\mathrm{M}\mathrm{e}\mathrm{a}\mathrm{n} \mathrm{A}\mathrm{Q}\mathrm{I} \mathrm{e}\mathrm{x}\mathrm{c}\mathrm{e}\mathrm{s}\mathrm{s}>100={\displaystyle \frac{{\sum }_{\mathrm{A}\mathrm{Q}\mathrm{I}>100}(\mathrm{A}\mathrm{Q}\mathrm{I}-100)}{{\mathrm{N}}_{>100}}}$$

(2)

where N_>100 is the number of hours with AQI > 100. This metric captures the average magnitude of exceedance events and allows for comparison across cities with differing pollution profiles.

2.4.4. Population-Hours Above Threshold

To account for population size, a population exposure index was calculated as follows:

$${\mathrm{P}\mathrm{o}\mathrm{p}-\mathrm{h}\mathrm{o}\mathrm{u}\mathrm{r}\mathrm{s} \mathrm{a}\mathrm{b}\mathrm{o}\mathrm{v}\mathrm{e} \mathrm{t}\mathrm{h}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{h}\mathrm{o}\mathrm{l}\mathrm{d}}_{\mathrm{A}\mathrm{Q}\mathrm{I}>100}={\displaystyle \frac{{\mathrm{N}}_{>100}\times \mathrm{P}\mathrm{o}\mathrm{p}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}{\mathrm{1,000,000}}}$$

(3)

To keep the numbers reasonable, the value was divided by 1 million.

Similar indices (such as annual exceedance person-days and population-weighted intakes) have been used in other air pollution studies [10,34,35,36].

2.4.5. Temporal Trend Analysis

To examine long-term changes in air quality across European capitals, temporal trends in annual average AQI values were calculated for each city. Hourly AQI data were first aggregated into annual means for the period 2010–2024. The resulting dataset was analyzed using linear regression models, where AQI was modeled as a function of year, separately for each city. This analysis was conducted in R using the “lm” function from the “broom” package to extract slope coefficients, standard errors, and p-values. The slope of each model indicates the direction and magnitude of AQI change over time. Only cities with sufficient data coverage were included in the analysis; capitals with more than 80% missing values, such as Riga and Tirana, were excluded to maintain the robustness of the trend estimates.

2.5. Gen AI

Generative artificial intelligence (GenAI) tools were used in a limited, supportive capacity during the development of this manuscript. Specifically, ChatGPT (o4-mini-high) was employed to assist in resolving coding errors and modifying R scripts used for AQI calculations and the generation of visualizations (e.g., boxplots and spatial maps). It is important to note that all data analysis and graphical outputs were created and executed independently by the author in R (version 4.3.1) within RStudio (version 2023.06.2 Build 561); GenAI did not generate any data or graphics.

Additionally, AI tools including ChatGPT (o3 and 4.o), Consensus 4.o, and ScholarGPT 4.o were used to

• Improve the clarity, structure, and cohesion of the Discussion and Conclusions (Section 4), particularly in organizing initial drafts and refining argument flow;
• Identify gaps in the literature and suggest a small number of additional, relevant references to support the interpretation of findings;
• Provide constructive feedback on potential limitations of the study, some of which were addressed and incorporated;
• Assist in the interpretation of AQI maps, though no novel insights beyond the author’s original analysis were ultimately adopted.

The use of these tools was limited to supporting analytical thinking and presentation; all final decisions regarding study design, data interpretation, and manuscript content were made solely by the author. No GenAI was used to generate original data, figures, or the conceptual basis for the research.

3. Results

3.1. Air Quality Index in European Capital Cities—An Overview

The comparative boxplot (Figure 1) illustrates the distribution of hourly Air Quality Index (AQI) values across selected European capitals from 2010 to 2024. The data reveal marked disparities in air quality, with Sarajevo exhibiting the highest median AQI (~62), a wide interquartile range (IQR ≈ 87), and extreme outlier values exceeding 250, indicating persistent and severe air pollution episodes. This finding aligns with numerous studies identifying Sarajevo as one of the most polluted cities in Europe [37,38,39].

Cities such as Lisboa, Brussels, Madrid, and Skopje also exhibited elevated median AQI values (52–58), with frequent outliers indicating episodic but intense pollution. For instance, particulate pollution in Skopje has been reported at critical levels, particularly during winter inversion periods [40,41]. In Lisboa, the concentrations of NO₂ and PM₁₀ (which are included in AQI calculations) exceed the recommended WHO levels, increasing health risks [42]. A similar situation was observed in Madrid, where high NO₂ concentrations were reported not only along traffic arteries but also near homes and primary schools [43]. In Brussels, high exposure to PM_2.5 and other pollutants has led to increased cardiovascular disease (CVD) and chronic obstructive pulmonary disease (COPD) mortality risks, especially in areas with limited communication accessibility [44,45].

London, Amsterdam, and Berlin display moderate median AQI values but with occasional spikes. Outliers beyond 300 AQI in cities such as London indicate acute pollution episodes, possibly due to weather inversions or localized emissions.

Conversely, Nordic capitals such as Reykjavik and Oslo demonstrate low median AQIs and compact IQRs, suggesting consistently good air quality.

The mean AQI across the 33 European capitals (Figure 2) shows a more than fourfold range, from a remarkably low 19.35 in Oslo to a high of 86.83 in Sarajevo. At the clean end of the spectrum are the two Nordic capitals—Oslo (19.35) and Reykjavik (22.49)—followed by Podgorica (23.07) and Rome (26.82), suggesting that Northern and some Central–Western cities benefit from comparatively pristine air. In contrast, the most polluted urban atmospheres are found in the Balkans, with Sarajevo (86.83) and Tirana (77.06) topping the list, and elevated mean AQI also being observed in Iberian capitals (Lisboa 73.92; Madrid 71.14) and Paris (61.48). Overall, there is a clear geographical pattern: lower background AQI in Northern and Central Europe, increasing progressively toward the southeast and (to a lesser extent) the southern periphery of the continent.

3.2. Seasonal Variations of the Air Quality Index in European Capital Cities

The mean winter AQI across Europe’s capitals (Figure 3) spans more than fourfold, from a very low 20.53 in Reykjavik to a striking 90.54 in Tirana. Northern cities again enjoy the cleanest air—Reykjavik (20.53) and Oslo (24.77) remain at the bottom of the winter pollution scale—while Central–Western capitals such as Prague (35.09) and Rome (36.54) register only moderate winter AQI levels. By contrast, the worst winter air quality is found in the Balkans and Southeastern Europe: Tirana tops the ranking at 90.54, followed by Skopje (80.42) and Sarajevo (75.97). In general, this seasonal snapshot reinforces a pronounced north–south and west–east gradient in winter pollution: air quality deteriorates markedly as one moves from the cool, maritime climates of Northern Europe toward the more continental and topographically enclosed basins of the Southeastern periphery.

Spring mean AQI across Europe’s capitals (Figure 4) varies nearly fivefold, from an exceptionally low 19.26 in Oslo and 19.37 in Podgorica to an extreme 100.52 in Sarajevo. Two other outliers—Reykjavik (21.26) and Rome (23.26)—also record some of the cleanest spring air, followed by Zagreb (31.62). At the opposite extreme, Sarajevo’s spring pollution soars above 100, with Lisboa (97.28) and Tallinn (91.88) not far behind; Berlin (83.35) and Vilnius (83.21) also exceed an AQI of 80. Whereas annual and winter means display a simple north–south gradient, spring AQI highlights intense pollution episodes in three distinct “hotspots”: the Balkan basin, the western Mediterranean (driven in part by Saharan dust and in part by increased ozone levels [14,15]), and the Baltic–northern plain. Central European capitals generally fall into a midrange AQI bracket (40–60), reflecting more moderate springtime air quality.

Summer (Figure 5) brings a markedly different AQI landscape to Europe’s capitals: Podgorica records the lowest summer mean AQI at just 15.33, with Oslo (17.04) and Rome (18.67) close behind. Sarajevo endures the season’s worst pollution, with a mean AQI of 106.56, followed by Madrid at 96.06, Berlin at 89.77, and Lisboa at 81.03. Mid-latitude Central European capitals such as Vienna (75.44) and Paris (75.56) fall into the moderate–unhealthy bracket, whereas the Baltic and northern Adriatic cities remain comparatively unscathed. This summer pattern highlights how regional meteorology—especially Saharan dust incursions in the western Mediterranean, increased ozone levels, and high-pressure-driven stagnation over continental basins—shapes seasonal air quality across southwestern Europe [3,14,15,46,47].

Mean autumn AQI (Figure 6) across the 33 European capitals spans from a remarkably low 15.98 (AQI units) in Oslo to a high of 70.61 in Tirana. Northern capitals—Oslo (15.98) and Reykjavik (18.79)—again enjoy the cleanest autumn air, while Central European cities cluster in a moderate band (Prague, 35.30; Vienna, 40.61; and Berlin, 46.35). Autumn pollution peaks in southeastern and some western cities: Tirana tops at 70.61, followed by Lisboa (62.37), Sarajevo (62.43), and Skopje (56.04).

Overall, the seasonal cycle of urban air quality in Europe reaches its minimum in autumn and peaks in spring and summer, driven by a combination of natural and meteorological processes as well as anthropogenic contributions. In winter and autumn, residential heating—particularly the combustion of wood, coal, and other solid fuels—contributes approximately 16–21% of PM_2.5 mass in many European cities, often exceeding vehicle traffic at residential monitoring sites and driving marked winter peaks in particulate pollution [45,48]. Source-apportionment studies further reveal that the share of residential heating in ambient PM₁₀ increases by roughly 20–40% during winter compared to summer in Central European basins, reflecting the prevalence of biomass and fossil-fuel heat production [49,50]. Road transport remains a principal source of fine particles and NO_x year-round—accounting for about 40% of Europe’s total NO_x emissions—and diesel vehicles are responsible for a significant fraction of urban NO₂ and PM₁₀ burdens [51,52]. Cold-start effects amplify these vehicular emissions: NO_x release can be over 3.3 times higher at sub-freezing temperatures, while PM_2.5 emission factors rise by more than 60% during winter cold starts relative to warm-weather operation [53,54]. Non-exhaust traffic sources—most notably road dust—add an additional 20–35% to urban PM₁₀ levels, with contributions peaking in summer under dry conditions and diminishing in winter when increased precipitation scavenges particles from the atmosphere [55,56]. This combined interplay of meteorology, Saharan dust, biomass burning, residential heating, and vehicular emissions underlies the pronounced spring–summer maximum and autumn minimum in AQI across Europe’s capitals.

3.3. Health Impact and Population Indices

The distribution of hourly air quality in European capitals (

Table 4. Percentage * of hourly observations by Air Quality Index (AQI) category in 33 European capitals.

City	0–50	51–100	101–150	151–200	201–300	>300
Amsterdam	52%	33%	9%	3%	1%	0%
Athens	59%	28%	5%	2%	0%	0%
Belgrade	49%	28%	9%	4%	1%	0%
Berlin	48%	28%	9%	5%	3%	0%
Bern	58%	29%	7%	4%	1%	0%
Bratislava	70%	22%	1%	0%	0%	0%
Brussels	39%	47%	9%	3%	1%	0%
Bucharest	33%	16%	3%	0%	0%	0%
Budapest	71%	24%	2%	0%	0%	0%
Copenhagen	56%	19%	13%	5%	1%	0%
Dublin	67%	22%	7%	1%	0%	0%
Helsinki	54%	32%	11%	3%	0%	0%
Lisboa	44%	30%	16%	7%	2%	0%
Ljubljana	63%	20%	7%	5%	3%	0%
London	48%	43%	6%	1%	0%	0%
Madrid	42%	37%	11%	6%	2%	0%
Oslo	89%	4%	0%	0%	0%	0%
Paris	49%	32%	7%	3%	1%	0%
Podgorica	53%	5%	0%	0%	0%	0%
Praha	86%	11%	1%	0%	0%	0%
Reykjavik	59%	5%	0%	0%	0%	0%
Riga	8%	5%	1%	0%	0%	0%
Rome	84%	9%	0%	0%	0%	0%
Sarajevo	25%	15%	8%	6%	5%	0%
Skopje	40%	34%	7%	3%	0%	0%
Sofia	61%	27%	7%	2%	0%	0%
Stockholm	62%	24%	10%	3%	1%	0%
Talin	54%	26%	13%	4%	1%	0%
Tirana	4%	12%	3%	1%	0%	0%
Vienna	64%	18%	8%	5%	2%	0%
Vilnius	64%	18%	9%	4%	1%	0%
Warsaw	54%	30%	9%	4%	1%	0%
Zagreb	72%	21%	2%	0%	0%	0%

* Remaining percentages indicate missing data.

), as measured by the Air Quality Index (AQI), exhibits substantial variation in atmospheric cleanliness. Among the AQI categories, the range of 0–50—representing good air quality—had the highest proportion of hourly observations in Oslo, where 89% of the recorded hours fell into this category, indicating consistently favorable air conditions. Furthermore, with only 4% of hours in the 51–100 (moderate) category, Oslo stands out as the cleanest capital in Europe. Similarly, Prague and Rome demonstrated high air quality, with 86% and 84% of hours, respectively, classified within the 0–50 range over the last 15 years, highlighting a sustained trend of clean and breathable urban air.

Encouragingly, no European capital recorded AQI values exceeding 300 in more than 1% of observed hours. Nevertheless, instances of extremely poor air quality were documented, with 318 such hours in Skopje and 175 in Sarajevo. Among the 16 capitals where AQI values in the 201–300 (very unhealthy) range were recorded, Sarajevo showed the highest proportion, at approximately 5% of all measured hours. Berlin and Ljubljana followed with 3%, while Madrid, Lisboa, and Vienna each exhibited 2%. In cities including Brussels, Belgrade, Paris, Amsterdam, Tallinn, Warsaw, Copenhagen, Bern, Stockholm, and Vilnius, the 201–300 category occurred in approximately 1% of recorded hours, indicating sporadic yet noteworthy exposure to elevated pollution levels.

From a public health perspective, the share of time a city experiences AQI levels above 100 serves as a critical marker of chronic exposure risk, especially for vulnerable populations such as children, the elderly, and individuals with pre-existing respiratory or cardiovascular conditions. Even brief episodes of elevated air pollution are associated with increased hospital admissions and all-cause mortality in susceptible groups.

The distribution of AQI exceedances across European capitals between 2010 and 2024 (Figure 7) reveals considerable heterogeneity, with no clear spatial pattern corresponding to geographical or climatic zones. The highest proportion of hourly exceedances was recorded in Lisboa (24.9%), followed by Sarajevo (19.5%), Helsinki (19.0%), Tallinn (18.9%), and Madrid (18.9%), indicating substantial long-term exposure to potentially unhealthy air conditions. Several other major capitals, including Berlin (16.2%), Warsaw (15.2%), London (14.6%), and Belgrade (14.5%), also reported exceedance levels above 14%, suggesting moderately elevated air quality concerns.

At the lower end of the distribution, cities such as Rome (0.2%), Oslo (0.3%), Podgorica (0.4%), and Reykjavik (0.4%) reported fewer than 0.5% of hours with AQI > 100. Notably, Prague (1.1%), Riga (1.3%), and Bern (1.4%) also exhibited very low exceedance levels, reflecting relatively stable air quality conditions. The observed variation across the continent underscores the influence of localized factors—such as emission sources, topography, regulatory enforcement, and meteorological patterns—over broad regional trends. This finding highlights the need for city-specific strategies in air quality management and health risk mitigation, rather than one-size-fits-all regional approaches.

While the frequency of AQI exceedances reveals how often air quality deteriorates, the mean excess over AQI 100 provides a deeper measure of pollution severity during those episodes (Figure 8). This metric quantifies the average intensity of exceedances, capturing potential cumulative health burdens from short-term spikes in pollutant concentrations. Higher values indicate a combination of both elevated pollutant levels and the persistence of harmful exposure.

Across European capitals from 2010 to 2024, Sarajevo stands out with the highest mean excess value of 66.8, suggesting frequent and intense pollution episodes. Ljubljana (58.7) and Berlin (52.2) follow, along with Vienna (51.0) and Reykjavik (48.6)—the latter being a notable outlier given its generally low number of exceedance hours. Other cities exhibiting high average exceedance intensity include Bern (45.7), Madrid (45.4), Vilnius (45.0), and Paris (44.4). These cities, despite varying in total hours above AQI 100, experienced substantial pollution levels when exceedances did occur.

At the opposite end, Rome (3.0), Budapest (4.4), Podgorica (4.9), and Zagreb (4.9) recorded the lowest mean excess values, indicating that air quality breaches were generally mild in both frequency and severity. Bucharest (6.6) also fell within this lower range, reflecting relatively moderate pollution levels during exceedance events. These findings suggest that lower exceedance intensity is not necessarily correlated with city size, as both large and mid-sized capitals appear among the least impacted by high-magnitude AQI episodes.

While the frequency and intensity of air quality exceedances are crucial, the population-exposure burden is best captured through person-hours affected by AQI > 100, standardized per 1 million inhabitants (Figure 9). This indicator integrates pollution data with urban population size to estimate the relative impact of degraded air quality on public health.

From 2010 to 2024, Paris reported the highest exposure burden, with over 150,749 person-hours per million residents spent in air quality conditions above the health threshold. It is the highest person-hour burden among all European capitals. This is a striking result given that Paris did not rank at the top in any of the previous analyses: its AQI exceedance frequency (11.1%) and mean excess over 100 (44.4) were moderately high but not exceptional. However, due to its large and dense population, even relatively moderate pollution levels result in a disproportionately high cumulative exposure. These results underscore the crucial role of demographic scaling in air quality assessment: in populous cities, modest exceedances can translate into substantial public health burdens.

Other major Western capitals such as Lisboa (90,439), London (80,279), Madrid (74,343), and Berlin (71,391) also recorded high exposure values, underscoring how large populations compound the health impact of moderate-to-high air pollution episodes. These cities may not always rank highest in exceedance severity or frequency, but their large urban populations amplify the net public health burden.

By contrast, Podgorica (96), Reykjavik (121), Oslo (406), and Bratislava (750) reported the lowest person-hour burdens, reflecting both lower AQI exceedance levels and smaller urban populations. Rome (985), despite its relatively large population, ranked as the fifth lowest, suggesting limited combined exposure despite occasional exceedances. This metric reveals that even cities with modest pollution levels can accumulate a high public health burden if population exposure is widespread, emphasizing the importance of population-normalized indicators in environmental health assessments.

3.4. Temporal Trend Analysis of AQI (2010–2024)

The linear trend analysis revealed divergent patterns of air quality change across European capitals between 2010 and 2024 (

Table 5. Linear trends * in annual Air Quality Index (AQI) (2010–2024).

Capital City	Slope	Standard Error	p-Value
Bucharest	−1.9721	1.0188	0.0889
Skopje	−1.7655	0.3399	<0.001
Prague	−1.7023	0.3001	<0.001
Lisboa	−1.6503	0.4395	0.0027
Athens	−1.611	0.6089	0.0214
Brussels	−1.3832	0.3634	0.0025
London	−0.9397	0.2798	0.0057
Reykjavik	−0.8153	0.5649	0.1768
Talin	−0.7526	1.2816	0.5679
Rome	−0.7217	0.2557	0.0154
Budapest	−0.6888	0.4969	0.1909
Zagreb	−0.5086	0.2121	0.0337
Podgorica	−0.2258	0.5288	0.6806
Helsinki	−0.0583	0.2245	0.7995
Sofia	−0.0123	0.6614	0.9854
Vienna	0.2158	0.3199	0.5127
Stockholm	0.2868	1.1418	0.8059
Warsaw	0.2876	0.832	0.7356
Copenhagen	0.3499	0.2086	0.1193
Oslo	0.5488	0.3654	0.159
Brno	0.5514	0.3139	0.1045
Madrid	0.6359	0.5681	0.2849
Ljubljana	0.8115	0.5783	0.1858
Paris	1.1007	0.3098	0.0045
Vilnius	1.1616	0.5885	0.0719
Belgrade	1.3042	1.7171	0.4635
Berlin	1.4493	0.3012	<0.001
Dublin	1.5973	0.7639	0.0585
Bratislava	2.3789	0.6843	0.0046
Amsterdam	2.4149	0.6897	0.0044
Sarajevo	2.8129	2.3072	0.2575

* Estimated linear trends for annual average AQI values across selected European capitals, based on ordinary least squares regression. Cities with more than 80% missing data (e.g., Tirana and Riga) were excluded from the analysis.

). Cities such as Sarajevo, Amsterdam, Bratislava, and Berlin exhibited statistically significant increasing trends in annual AQI values, suggesting a deterioration in urban air quality despite ongoing mitigation efforts. Conversely, several cities demonstrated improving air quality over time, with Prague, Skopje, Lisboa, and Athens showing significantly negative slopes, indicating decreasing AQI levels. Notably, London and Brussels also recorded downward trends with statistical significance (p < 0.01), highlighting the impact of sustained policy interventions. While some cities, such as Dublin and Vilnius, showed positive trends, their p-values suggest weaker evidence for consistent change. These regional differences underscore the uneven effectiveness of air quality management across Europe and support calls for more harmonized and locally tailored interventions.

Caution is advised in interpreting these trends, as the COVID-19 pandemic (2020–2021) introduced temporary emission reductions unrelated to structural policy change. This period may distort long-term AQI trajectories, and future assessments should consider excluding or adjusting for these years.

3.5. Case Studies from Oslo, Lisboa, Warsaw, and Paris

To reflect the diversity of air quality conditions, demographic exposure, and geographic representation across Europe, four capital cities were selected as case studies: Paris, Oslo, Lisboa, and Warsaw. These cities were chosen to capture a broad spectrum of air pollution profiles and regional contexts, as well as to contrast pollution burden against population dynamics:

• Paris, the most populous capital in the dataset (10.9 million residents), had only moderate exceedance frequency (11.1% of hours) and mean excess severity (44.4 AQI units), but emerged as the city with the highest population exposure burden, at 150,749 person-hours with AQI > 100 per million residents. This selection underscores the importance of accounting for population scale in assessing environmental health risks, as even moderate pollution in a dense metropolis can generate an outsized burden.
• Oslo represents the northern European region and serves as a reference case for low-exposure urban environments. It ranked among the lowest in every major metric: only 0.3% of hours exceeded AQI 100, the mean excess was just 22.1, and population exposure was minimal at 405 person-hours per million. As a consistently clean capital, Oslo offers a valuable counterpoint for examining best practices in air quality management.
• Lisboa, representing southern Europe, had the highest frequency of AQI exceedances (24.9%) among all capitals and a high mean excess of 43.2. It also had one of the highest population-weighted exposure values (90,439 person-hours per million), highlighting persistent pollution challenges despite a moderate population size. Lisboa exemplifies cities where geographic or meteorological factors may intensify pollutant accumulation.
• Warsaw was selected as a representative of Central and Eastern Europe. Its metrics reflect a moderate pollution profile: exceedances occurred during 15.2% of observed hours, with a mean excess of 43.5 and a population-weighted exposure of 31.044 person-hours per million. Warsaw provides a valuable midpoint case for examining regional dynamics and transitional air quality regimes.

Together, these cities form a comparative framework for exploring how air quality, health vulnerability, and urban form intersect across varied European contexts.

Each city exhibits a distinct annual pattern (Figure 10). Oslo maintains the lowest and most stable AQI profile, with daily values rarely exceeding 100 and only minor seasonal variability. In contrast, Lisboa shows prominent seasonal cycles, with summer peaks frequently reaching or exceeding the AQI 100 threshold, which is consistent with elevated ozone levels common in warmer climates. Warsaw and Paris also display seasonal patterns, with pollution spikes in summer. In these three cities, there are several episodes of AQI reaching or exceeding 250, which are classified as “health alerts”.

However, a closer examination of monthly AQI variations (Figure 11) shows that Oslo retains the lowest and most stable monthly profile with an elevated winter peak, but overall AQI values are consistently below 30, indicating minimal seasonal fluctuation and reflecting the generally clean air environment characteristics of Northern Europe. In contrast, Lisboa exhibits a pronounced summer peak, with AQI levels rising above 100 in April and May, likely driven by photochemical smog and elevated ozone concentrations during warmer months [15]. Paris and Warsaw both exhibit a pronounced unimodal seasonal pattern, with AQI levels peaking during the summer months, particularly from June to August, and remaining comparatively lower during winter. This pattern differs from the commonly observed winter peaks in air pollution associated with heating and traffic emissions and instead reflects the growing dominance of secondary pollutants, notably tropospheric ozone, formed through photochemical reactions involving NO_x and volatile organic compounds (VOCs) under intense solar radiation [13,57]. In both cities, ozone has become the primary summer pollutant, contributing to AQI elevations even in the absence of local primary emissions. This seasonal shift is well documented in regional studies that highlight the increasing importance of ozone-related pollution episodes in urban centers across Central and Western Europe [2]. Notably, Paris maintains consistently higher AQI levels than Warsaw throughout the year, likely due to its larger urban scale and the influence of long-range transported secondary pollutants [17]. This phenomenon aligns with broader European trends documented in air quality studies, which underscore ozone’s growing role in summer air quality degradation, particularly in cities with declining wintertime NO₂ and PM levels [2].

During further analysis of AQI variation in the hour–week chart (Figure 12) for Oslo, Lisboa, Warsaw, and Paris, there is a clear daily pattern in all cities except Oslo. This bimodal daily cycle is evident in Lisboa, Paris, and Warsaw on weekdays, with AQI values typically showing a small peak during the morning (6–10 AM) and a pronounced peak in the evening (5–9 PM) hours. Notably, in all three cities, the evening peak is significantly higher than the morning peak, indicating greater pollutant accumulation in the latter part of the day, possibly due to higher traffic density combined with reduced atmospheric dispersion at night.

In Lisboa, this evening elevation becomes especially critical: AQI values frequently exceed 100 during Saturday and Sunday evenings, crossing into the “unhealthy for sensitive groups” category. This weekend rise—unusual compared with typical weekday commuter trends—may be attributed to the “weekend effect” [58], which is explained by the fact that the ozone (O₃) concentrations in Lisboa tend to rise on weekends due to lower nitrogen oxides (NO_x) emissions from reduced traffic. This reduces ozone destruction (the NO_x-titration effect), particularly in the summer and tourist-heavy period.

By contrast, Oslo displays an exceptionally stable and low AQI profile, with minimal daily or weekly variability, consistent with earlier findings of its low overall pollution burden and effective emissions control. This could be explained by the implementation of air quality policies, such as low emission zones, parking fees, and a ban on diesel vehicles, which have resulted in decreased pollution [59].

4. Discussion and Conclusions

4.1. Overview of the Main Findings

This study offers a comprehensive assessment of urban air quality across 33 European capitals from 2010 to 2024, providing an integrated view of pollution severity, temporal dynamics, and population exposure burdens. The analysis revealed profound spatial disparities in AQI distributions across the continent, with Sarajevo, Lisboa, and Madrid emerging as pollution hotspots, while Oslo, Reykjavik, and Podgorica consistently maintained low AQI values (Figure 2). Sarajevo recorded the highest mean AQI (86.83, Figure 2), the highest seasonal means (e.g., 106.56 in summer, Figure 5), and one of the highest average exceedance severities (66.8 AQI units, Figure 8), pointing to chronic and intense pollution episodes in this Balkan basin city.

In contrast, Nordic capitals such as Oslo and Reykjavik demonstrated exceptionally low median AQI values (Figure 1), minimal seasonal fluctuation (Figure 3, Figure 4, Figure 5, Figure 6 and Figure 11), and almost negligible exceedance frequency (Table 4). Oslo, in particular, recorded only 0.3% of hours above the AQI 100 threshold (Table 4) and a population-weighted exposure burden of just 405 person-hours per million inhabitants (Figure 9).

Analysis of annual AQI trends between 2010 and 2024 revealed divergent trajectories across European capitals. While cities such as Sarajevo, Berlin, and Amsterdam exhibited statistically significant increases in AQI, others—including Prague, Skopje, and Lisboa—showed meaningful improvements (Table 5). However, caution is advised in interpreting these trends, as temporary emission reductions during the COVID-19 pandemic may have introduced short-term anomalies into otherwise consistent patterns.

Seasonal analyses underscored a consistent pattern of elevated spring and summer AQI values (Figure 4 and Figure 5), largely attributable to ozone formation under photochemically active conditions and Saharan dust intrusions [3,15] A striking example is Lisboa (Figure 11), which exhibited pronounced spring peaks and the highest AQI exceedance frequency (24.9%, Figure 9, Table 4). The presence of secondary pollutants like tropospheric ozone dominated warm-season AQI levels in cities such as Paris and Warsaw, superseding the more expected winter particulate pollution [13].

The seasonal peaks in AQI identified in this study likely reflect different dominant pollutants and atmospheric processes across European capitals. In many Central and Eastern European cities, high winter AQI values reported in the literature are often associated with elevated PM₁₀ levels from residential heating using solid fuels, combined with temperature inversions that limit vertical mixing and favor pollutant accumulation [48,49,50]. In contrast, spring and summer AQI peaks observed in several Mediterranean and Western European cities are frequently linked to episodes of Saharan dust intrusion [3,14] and enhanced photochemical activity leading to secondary aerosol and ozone formation under warm, sunny conditions [13,15]. Meteorological factors strongly modulate these events: low wind speeds and stable atmospheric stratification promote pollutant buildup, while precipitation can rapidly reduce concentrations through wet deposition [55,56]. Seasonal changes in prevailing winds also contribute to the long-range transport of particulate matter and other pollutants, producing high-AQI episodes even in locations far from the original sources [14,17].

Perhaps most notably, when population size was accounted for, the relative air pollution burden shifted. Paris, despite only moderate exceedance frequency and severity, emerged as the city with the highest population-weighted exposure (150,749 person-hours/million, Figure 9), highlighting the disproportionate health risks borne by large urban populations. This underscores the importance of integrating demographic factors into air pollution risk assessments.

4.2. Comparison with the Previous Studies

The results of this study are consistent with previous pan-European air quality assessments. These findings reflect the patterns observed by the WHO and EEA, which emphasize the worsening air quality gradients from Northern to Southeastern Europe [1] and highlight the persistent burden of particulate matter and ozone in southern cities. For example, Saharan dust intrusions were identified as key contributors to springtime and summer pollution episodes, as well as ozone episodes in the Western Mediterranean [14,15]. These atmospheric phenomena correspond well with the seasonal peaks observed in Lisboa and Madrid in the dataset (Figure 3, Figure 4, Figure 5, Figure 6 and Figure 11).

The growing dominance of ozone as a summertime pollutant in urban environments such as Paris and Warsaw is also supported by broader trends reported in literature. Studies have shown that ozone levels are rising in urban areas across Europe, even as NO₂ and PM levels decline [16,57]. The results of the present study confirm this seasonal shift in dominant pollutant types and underscore the necessity of adapting policy responses to emerging ozone threats.

In addition, the high health burden revealed in populous cities such as Paris (Figure 9) aligns with recent findings [10] that emphasize the compounding effects of pollutant co-occurrence and population density on public exposure. Similarly, other studies [17] have found that regional transport significantly contributes to fine PM concentrations in megacities like Paris, amplifying the pollution impact even when local emissions are moderated.

The findings of this study align with broader global patterns of urban air pollution while highlighting distinct regional specificities. For example, cities such as Warsaw and Paris exhibited long-term positive AQI trends, consistent with other European research documenting gradual declines in pollutant emissions, especially from road traffic and industrial sources due to regulatory measures [51,52]. In contrast, data from Latin American cities such as Mexico City and Santiago indicate more persistent pollution, where PM_2.5 levels remain elevated despite localized interventions [23,25]. Similarly, research from China has confirmed a historical worsening of air quality during rapid urbanization, followed by subsequent improvement due to policy reforms like the National Innovation Demonstration Zones [8].

Urban areas in the United States present a somewhat different trajectory. Long-term monitoring has shown significant reductions in ozone and PM_2.5 over the past two decades as a result of regulatory interventions, particularly the implementation of the Clean Air Act [19], with additional improvements projected through integrated emissions models [21]. However, exposure inequality persists among marginalized populations. In African cities, increasing urbanization has been linked to worsening aerosol levels, though data limitations remain a significant challenge [28]. Compared to these contexts, the cities in this study generally exhibit moderate AQI levels with seasonal and spatial fluctuations, but also evidence of long-term increases in several cases (e.g., Amsterdam and Zagreb), suggesting that further mitigation effort is needed.

4.3. Health Relevance and Population Burden

The public health significance of this study lies not only in identifying when and where AQI exceeds the “Unhealthy for Sensitive Groups” threshold (>100), but also in assessing how these exceedances intersect with population exposure. It is well established that even short-term elevations in air pollution are associated with acute cardiovascular and respiratory health outcomes, including increased hospital admissions and mortality [2].

Findings from this study demonstrate that AQI exceedances are widespread but highly variable, both in frequency and severity (Table 4). Cities like Lisboa, Madrid, and Sarajevo experience high rates of exceedances and considerable severity (Figure 7 and Figure 8), while others, such as Paris, accumulate substantial cumulative health burdens due to population scale (Figure 9). These results imply that health risks are not solely a function of pollution intensity but also of the population size exposed during pollution episodes.

This is particularly critical for cities like Paris, where moderately frequent and moderately severe exceedances translate into the highest cumulative exposure burden (150,749 person-hours per million, Figure 9), surpassing cities with worse pollution metrics. By contrast, Oslo’s near-zero exceedance frequency and low exposure burden highlight the effectiveness of sustained air quality governance [59].

4.4. Policy Implications

From a policy standpoint, the use of population-weighted exposure indicators alongside traditional AQI summaries is strongly recommended. Policy-makers should not rely solely on pollution averages to guide interventions but should also consider who is being exposed, how often, and under what conditions. Failure to do so risks underestimating the cumulative health burden in large cities with “moderate” pollution.

The policy relevance of these findings lies in the observed mismatch between pollution frequency and public health burden. While cities like Sarajevo and Madrid experience more intense pollution episodes (Figure 1), large capitals like Paris, due to their population size, contribute most to the cumulative exposure risk (Figure 9). This supports the need for tailored air quality strategies—focusing not only on peak levels but also accounting for urban demographics.

Given the diversity of pollution sources, seasonal patterns, and demographic profiles, no single mitigation strategy will be effective across all European capitals. Rather, the evidence supports tailored, context-specific approaches:

• Northern Cities (e.g., Oslo and Helsinki): These cities exhibit the lowest AQI levels and exposure burdens, likely due to stringent air quality policies, cleaner energy use, and effective emissions control (including low-emission zones, diesel bans, and high public transit use). Oslo’s example demonstrates that continued investment in transport regulation and urban design can yield sustained air quality benefits [59].
• Southern and Western Cities (e.g., Lisboa, Madrid, and Paris): These cities face major challenges from ozone and PM levels, exacerbated by Saharan dust events and stagnant summer conditions. They also have large, dense populations, which increase public health vulnerability. Policy responses in these cities should focus on reducing ozone precursors—especially NO_x and VOCs—through stricter vehicle emission standards, improved public transit, and regional coordination to address transboundary pollution.
• Central and Eastern Cities (e.g., Warsaw, Budapest, and Sarajevo): These regions often face wintertime peaks from residential heating and solid fuel combustion. As source-apportionment studies suggest, biomass and coal contribute significantly to PM levels in winter [48]. Transitioning to cleaner heating sources, incentivizing energy-efficient retrofits, and enforcing building emissions standards could substantially reduce cold-season pollution.
• High-Burden Metropolises (e.g., Paris, London, and Berlin): These cities require hybrid strategies that account for both local emissions and imported pollution. Measures could include enhanced regional air quality monitoring, traffic restrictions and controls, and increased public health advisories during high-AQI episodes. Public communication strategies are especially crucial in these cities to inform and protect vulnerable groups during episodic spikes.

In all regions, evidence-based planning and AQI-based early warning systems can serve as vital tools for protecting population health. The inclusion of metrics like mean excess severity and person-hours exposed provides urban policy-makers with practical insights to prioritize interventions and allocate resources.

4.5. Limitations

However, several limitations of this study should be acknowledged. First, despite efforts to harmonize station selection, variability in monitoring infrastructure across European capitals may introduce bias, particularly in under-monitored regions such as the Balkans. Second, the exclusion of several cities with more than 80% of missing data (e.g., Riga and Tirana) reduces the geographic comprehensiveness of the study. Third, the AQI, while a useful communication tool, aggregates multiple pollutants and may mask pollutant-specific risks, especially for ozone or fine particulate matter (PM_2.5), which have distinct health profiles. Fourth, this study relies on AQI data derived from a single monitoring station per capital city, selected based on strict criteria to ensure comparability. However, this approach inevitably generalizes urban air quality and does not capture intra-city spatial variability or localized pollution hotspots. Moreover, the analysis does not incorporate personalized exposure data or stratification by demographic characteristics such as age, health status, or socio-economic background. As a result, the findings reflect average ambient conditions rather than individual-level exposure or vulnerability, limiting the ability to assess environmental justice implications or health impacts for specific subpopulations.

Moreover, this analysis does not directly link pollution to health outcomes (e.g., morbidity and mortality), which would require access to health records and cohort-level exposure estimates. Finally, the population estimates are static and do not reflect seasonal fluctuations (e.g., tourism in Lisboa), which may lead to an underestimation of actual exposure burdens during peak periods.

4.6. Contributions and Future Directions

Despite the limitations, this study contributes both methodologically and substantively to the field of urban environmental assessment. By applying a unified AQI calculation method across all European capitals, it enables cross-national comparability that is rarely achieved in existing studies. Furthermore, the integration of temporal patterns, exposure-adjusted metrics, and long-term trends provides a comprehensive framework for interpreting urban air pollution that supports sustainable policy development. These methodological choices help fill an important gap in air quality research—particularly by facilitating comparisons between cities and informing locally relevant yet regionally coordinated actions aligned with the goals of sustainable urban development. By combining standardized AQI indicators with population-weighted exposure metrics, this study provides a comparative dataset that can support sustainability-related urban diagnostics. While it does not propose direct policy interventions, the analysis offers an evidence base for understanding how air quality burdens vary across European cities and how these burdens intersect with population vulnerability—both of which are central concerns for advancing SDG 11.

Building on the findings of this study, future research should focus on directly linking AQI exceedances to health outcomes, such as hospital admissions, mortality, or respiratory-related morbidities, using health data where available. The integration of satellite-based remote sensing could enhance coverage in data-scarce regions and improve spatial resolution. Further, scenario-based modeling incorporating projected changes in climate, land use, and transportation could help assess future pollution burdens under different policy pathways. Separating AQI into pollutant-specific indices—especially for ozone and PM_2.5—would allow for more targeted mitigation strategies. Finally, integrating socio-demographic layers into exposure models could illuminate intra-urban disparities and support environmental justice analyses, thereby ensuring that pollution mitigation efforts are equitable as well as effective.