Beijing and New York have both reduced important air pollutants, but their stories are often told with incompatible “average AQI” numbers. That approach hides the pollutant, averaging period, monitoring boundary, and index system.
The more defensible comparison is to read each city's official monitoring record on its own terms. Beijing reports a citywide annual PM2.5 concentration from its regulatory network. New York City's Community Air Survey (NYCCAS) uses seasonal street-level sampling and spatial models to describe neighborhood patterns. Both are useful, but they are not the same measurement design.
The Verified Numbers
| City | Official metric | Period | Reported result | |---|---|---|---:| | Beijing | Citywide annual mean PM2.5 | 2013 | 89.5 µg/m³ | | Beijing | Citywide annual mean PM2.5 | 2025 | 27.0 µg/m³ | | New York City | Modeled citywide annual mean PM2.5 | 2023 | 5.8 µg/m³ | | New York City | Change in annual mean PM2.5 | 2009–2024 | 36% decrease |
Beijing's municipal government reports that annual PM2.5 fell from 89.5 µg/m³ in 2013 to 27.0 µg/m³ in 2025, a 69.8% reduction over the monitoring period. The same official release reports 311 days categorized as good air quality in 2025 and one heavy-pollution day. These figures come from the Beijing Municipal Ecology and Environment Bureau's 2025 briefing; an English municipal summary also reports the 27.0 µg/m³ annual mean.
For New York, the Health Department's neighborhood model reports a 2023 citywide average of 5.8 µg/m³ and identifies variation associated with traffic, buildings, industrial areas, and commercial cooking. The value and data year are shown on the city's Neighborhood Air Quality data feature. The newer NYCCAS annual report says citywide annual average PM2.5 declined 36% between 2009 and 2024, an average decline of 0.3 µg/m³ per year.
These are completed-period statistics. We do not use a partial 2026 average or claim to know either city's full-year 2026 result.
Beijing: A Large, Documented Reduction
Beijing's improvement since 2013 is substantial. The municipal record shows a long decline not only in PM2.5 but also in PM10, nitrogen dioxide, and sulfur dioxide. Heavy-pollution days fell from 58 in 2013 to one in 2025, according to the same official briefing.
The result should be described accurately: Beijing reduced its annual PM2.5 burden by roughly two-thirds, but its 2025 mean of 27.0 µg/m³ remained above the World Health Organization's 5 µg/m³ annual PM2.5 guideline. WHO presents that value as health-based guidance and provides interim targets for places starting at higher concentrations in its Global Air Quality Guidelines.
The municipal briefing attributes progress to sustained emissions reduction, structural changes, targeted projects, stronger monitoring, and regional coordination. It does not support simple claims that one measure or one source category produced a fixed percentage of the improvement. Weather also affects year-to-year concentrations, so policy evaluation requires more than comparing two endpoints.
Beijing's remaining challenge is therefore different from the situation in 2013. The city has moved far down the concentration range, while continued progress requires controlling regional transport, traffic and industrial emissions, combustion sources, secondary particle formation, and episodic meteorology. A lower annual mean also does not eliminate short pollution episodes.
New York: Lower Concentrations, Persistent Local Differences
NYCCAS was designed to show how pollution varies within New York City. Staff place samplers 10 to 12 feet above the ground on streetlight poles, sample each site for two weeks in each season, and combine measurements with land-use and emissions information. The program complements New York State's federal regulatory monitors, which are often located on rooftops. The NYCCAS methods description explains this design and its limitations.
The long-run decrease does not mean every neighborhood has the same air. The city's analysis associates higher recent PM2.5 with nearby commercial grills and charbroilers, traffic, and some industrial or warehouse areas. Building heating emissions have become less dominant than in the program's early years as heating fuels became cleaner.
The survey also demonstrates why one anomalous year should not be treated as a new baseline. Canadian wildfire smoke raised fine-particle measurements in 2023. A city can have a relatively low long-term mean and still experience severe short-term smoke episodes. Conversely, a clean day does not erase the annual exposure pattern.
NYC's 5.8 µg/m³ modeled average for 2023 was close to, but still above, WHO's 5 µg/m³ guideline. It was below the current U.S. annual primary PM2.5 standard of 9.0 µg/m³; EPA explains that revised regulatory standard in its 2024 particulate-matter rule. A federal legal standard and a WHO guideline serve different roles and should not be described as equivalent safety thresholds.
Why the City Values Are Not a Direct Scorecard
It is tempting to divide 27.0 by 5.8 and announce that one city is a precise multiple of the other. That would overstate what the data support for four reasons:
- Different years: Beijing's value is for 2025; New York's modeled level is for 2023.
- Different spatial designs: Beijing reports a municipal regulatory-network mean; NYCCAS models neighborhood-scale street-level conditions.
- Different geography: Administrative boundaries, population distribution, and regional transport differ.
- Different source and climate profiles: Coal and industrial history, building heat, traffic, cooking, wildfire smoke, humidity, and atmospheric chemistry are not interchangeable.
The correct conclusion is qualitative: Beijing recorded a very large reduction from a high starting point, while New York recorded a slower long-term decline from a much lower concentration range. Both still have work to do, especially for episodic pollution and neighborhood disparities.
What the Two Cities Teach
Measure long enough to distinguish trend from weather
Both records span more than a decade. That matters because a single year's wind, precipitation, wildfire smoke, or heating demand can obscure the emissions trend.
Report pollutant concentrations, not only AQI categories
Annual mean PM2.5 in µg/m³ can be compared with a stated guideline and tracked over time. “Average AQI” cannot do that reliably, particularly across countries with different index systems.
Pair citywide progress with neighborhood data
A city can meet a broad target while people near roads, warehouses, industrial activity, or combustion sources experience higher local concentrations. NYCCAS makes this spatial inequality visible; other cities benefit from comparable neighborhood monitoring.
Treat policy effects and observed trends as separate claims
An observed decline is not, by itself, proof of how much any one policy caused. Strong evaluations disclose their model, counterfactual, weather adjustment, and uncertainty.
Using These Numbers for Personal Decisions
Annual PM2.5 is useful for understanding long-term background pollution. It does not tell you whether to run outdoors today. For daily activity, consult the responsible local authority's current pollutant, forecast, and health guidance. During wildfire smoke or another episode, current PM2.5 can move far above the annual mean.
Do not enter either annual mean into the AQI-to-cigarettes calculator. The site's calculation starts with a U.S. PM2.5 AQI for a specific exposure period, not an annual concentration and not a Chinese AQI. If an app does not identify PM2.5 as the dominant pollutant and the index system as U.S. AQI, conversion is not supported. The methodology explains that boundary.
Bottom Line
Beijing's official annual PM2.5 fell from 89.5 µg/m³ in 2013 to 27.0 µg/m³ in 2025. New York City's official community survey reports a 36% decline from 2009 to 2024 and a modeled 2023 citywide mean of 5.8 µg/m³. Those are meaningful achievements measured with different systems. The honest comparison preserves the year, unit, monitoring design, and remaining uncertainty instead of reducing two complex urban records to an unsupported average-AQI ranking.