Lists of the “cleanest” and “most polluted” cities often combine incompatible measurements: a live AQI from one afternoon, a yearly PM2.5 average from another source, and an app-specific index from a third country. The ranking looks precise, but the method is not defensible.
The World Health Organization's current source is the WHO Ambient Air Quality Database V8.0, released on 30 June 2026. It contains annual PM2.5, PM10, and nitrogen-dioxide measurements for 8,024 human settlements and covers data years from 2010 through 2024. WHO explains the scope on its official V8.0 database page.
WHO also gives an unusually clear warning: the database should not be used to rank the most polluted cities or make simple country comparisons. Monitoring capacity, station placement, temporal coverage, and measurement methods differ. That warning should shape every city comparison.
Why “Average AQI” Is the Wrong Long-Term Metric
AQI is a daily communication index. In the United States, the reported AQI is the highest pollutant-specific index among ozone, particle pollution, carbon monoxide, sulfur dioxide, and nitrogen dioxide. Its value depends on the pollutant, averaging period, and national breakpoint system. AirNow defines the U.S. AQI as a daily index, not a universal concentration unit.
Three problems make “average AQI” unsuitable for international city comparisons:
- Countries use different AQI equations, pollutant sets, and category labels.
- The dominant pollutant may be PM2.5 on one day and ozone on another.
- Averaging index values discards the physical concentration and creates a statistic with no standard global interpretation.
For long-term particle pollution, annual mean PM2.5 in micrograms per cubic meter (µg/m³) is the clearer starting point. It still requires a year, geographic boundary, station type, and coverage statement.
Traceable Official Examples—Not a Ranking
The following figures illustrate the range of official reporting designs. They are intentionally not sorted and should not be treated as a league table. Each value links to the public agency that defines its period and method.
| Place | Completed period | Official PM2.5 result | Measurement scope | |---|---|---:|---| | New York City | 2023 | 5.8 µg/m³ | NYCCAS modeled citywide annual mean | | Stockholm | 2024 | 4.5 µg/m³ urban-background; 5.4–5.6 µg/m³ at two street sites | Individual official stations, not one city mean | | Paris | 2024 | Approximately 8–9 µg/m³ | Urban-background levels across Paris | | Beijing | 2025 | 27.0 µg/m³ | Municipal annual mean | | Delhi | FY 2025–26 | 102 µg/m³ | Government-reported city annual value | | Mumbai | FY 2025–26 | 33 µg/m³ | Government-reported city annual value |
New York's Health Department publishes the 5.8 µg/m³ 2023 citywide estimate in its Neighborhood Air Quality feature. The result is produced from NYCCAS street-level monitoring and a spatial model, so it is not identical to one federal regulatory monitor or a metropolitan-area average.
Stockholm's official 2024 report lists 4.5 µg/m³ at the Torkel Knutssonsgatan urban-background rooftop site and 5.6 and 5.4 µg/m³ at Hornsgatan and S:t Eriksgatan street sites. The tables are in Luften i Stockholm 2024. Presenting all three values preserves the spatial difference that a single “Stockholm average” would hide.
Airparif reports 2024 urban-background PM2.5 of about 8–9 µg/m³ in Paris, with somewhat higher concentrations near major roads. Its official Paris 2024 assessment says the current legal limit was met but the WHO annual guideline was exceeded across the city.
Beijing's municipal environmental authority reports a 2025 annual mean of 27.0 µg/m³, along with the monitoring trend since 2013, in its official 2025 air-quality briefing.
India's Ministry of Environment, Forest and Climate Change reported FY 2025–26 annual PM2.5 of 102 µg/m³ for Delhi and 33 µg/m³ for Mumbai in a July 2026 parliamentary reply.
These values are directly traceable, but the table still mixes monitoring designs and, in some cases, reporting years. It is a guide to reading official data, not proof that one city has a precise multiple of another city's personal exposure.
How to Use WHO V8.0 Responsibly
WHO V8.0 is the best current global directory for locating annual city measurements and their metadata. It compiles official national and subnational datasets; European Union data are sourced from the European Environment Agency. The database's technical download page identifies the workbook as the June 2026 release.
For a research comparison, first filter to:
- one pollutant, such as PM2.5;
- one completed data year;
- comparable station classifications;
- adequate temporal coverage;
- similar geographic definitions; and
- a documented minimum number of stations.
Then inspect the source and web-link fields for every row. Do not silently substitute an older year for one city, a metro area for another, or a traffic hotspot for an urban-background network. If a row lacks sufficient metadata, exclude it rather than filling the gap with an app estimate.
Even after those controls, describe the result as a dataset-specific comparison with limitations—not a definitive global rank. WHO notes that some countries have sparse coverage, data quality is heterogeneous, and partial-year measurements can deviate from a true annual mean.
The Benchmark Must Also Be Named
WHO's health-based annual PM2.5 guideline is 5 µg/m³, published in the 2021 Global Air Quality Guidelines. It is guidance, not a claim that 5 is a sharp boundary between harmless and harmful air.
Regulatory standards differ. The United States strengthened its annual primary PM2.5 standard to 9.0 µg/m³ in 2024; the EPA rule and fact sheet explain the change. The revised EU directive sets an annual PM2.5 limit of 10 µg/m³ from 2030, according to the European Commission's Zero Pollution dashboard.
A city may comply with its current law while exceeding the WHO guideline. “Below the legal limit” and “zero health risk” are not equivalent statements.
What Annual PM2.5 Does Not Tell You
It does not describe today's air
An annual mean cannot tell you whether a wildfire, dust storm, inversion, or ozone episode is affecting a city now. For today's outdoor activity, use the responsible local authority's current pollutant and forecast.
It does not capture every neighborhood
People near high-traffic roads, warehouses, ports, industrial sites, or combustion sources can experience concentrations different from the citywide mean. Stockholm's 2024 station results demonstrate that site type matters even within one city.
It does not cover every pollutant
PM2.5 is important for long-term health comparison, but ozone and nitrogen dioxide can remain serious problems where annual PM2.5 is relatively low. The metric must match the question.
It does not prove why concentrations changed
A before-and-after decline is not a causal policy evaluation. Weather, regional transport, economic activity, fleet turnover, fuel use, and monitoring changes can all affect the result. Strong policy studies disclose a counterfactual and uncertainty.
A Checklist for Any City Comparison
Before accepting a chart or article, ask:
- What pollutant is being compared?
- Is the number a concentration or an index?
- What is the unit and averaging period?
- Is the year complete and identical across cities?
- Does “city” mean municipality, urban area, or metro region?
- Are sites roadside, urban background, suburban, or mixed?
- How many monitors and what temporal coverage are included?
- Is the figure measured, modeled, or a combination?
- Which national AQI system applies?
- Does the linked source allow the exact value to be located?
If those questions cannot be answered, the ranking is not ready for health, relocation, or policy decisions.
Travel, Relocation, and Daily Decisions
Long-term and short-term data answer different questions. Annual PM2.5 helps describe background particle pollution. Seasonal history can show wildfire, winter heating, dust, or ozone patterns. A current AQI forecast helps plan today's activity.
For a home search, consider traffic proximity, commercial cooking, industry, wildfire exposure, ventilation, filtration, and official neighborhood maps. Those local factors may matter more than a small difference between two broad city averages.
A Note About Cigarette Equivalents
Do not enter an annual PM2.5 mean or an unknown international AQI directly into an AQI-to-cigarettes calculator. This site's estimate requires a U.S. PM2.5 AQI so the index can first be converted to a 24-hour PM2.5 concentration. If ozone, PM10, or another pollutant drives the index—or if the app uses Indian, Chinese, European, or another national AQI—the calculation is not valid. Review the methodology before interpreting any result.
Bottom Line
The current global reference is WHO Ambient Air Quality Database V8.0, not the older V6.1 release. WHO explicitly warns against using it as a simple city-ranking engine. A responsible comparison keeps the pollutant, unit, year, boundary, station type, coverage, and source attached to every value. For long-term particle pollution, annual PM2.5 is a useful starting point; “average AQI” is not.