Try it
Set a PM2.5 concentration and see what each country's official index calls it.
At this concentration the four standards return 4 different verdicts for identical air.
Why they disagree
Every index in this table works the same way: measure a pollutant, find the band its concentration falls into, then interpolate linearly between that band's endpoints to get a number. The arithmetic is identical everywhere. What differs is where the bands are drawn.
Those choices reflect local conditions and policy, not disagreement about the science. India tolerates more PM10 before flagging concern, which is a reasonable response to natural dust. A country with cleaner baseline air can afford tighter thresholds.
Two practical consequences. First, an AQI number is meaningless without knowing whose scale produced it: a Thai AQI of 150 and a US AQI of 150 do not describe the same air. Second, the concentration in µg/m³ is the part that actually transfers between countries, which is why Atmos shows it.
United States · default worldwide · US AQI
EPA (pre-2024 breakpoints)The scale most apps show by default, anywhere in the world. Atmos uses these breakpoints today.
| Category | PM2.5 µg/m³ | PM10 µg/m³ | Index |
|---|---|---|---|
| Good | 0–12 | 0–54 | 0–50 |
| Moderate | 12.1–35.4 | 55–154 | 51–100 |
| Unhealthy for Sensitive Groups | 35.5–55.4 | 155–254 | 101–150 |
| Unhealthy | 55.5–150.4 | 255–354 | 151–200 |
| Very Unhealthy | 150.5–250.4 | 355–424 | 201–300 |
| Hazardous | 250.5–500.4 | 425–604 | 301–500 |
United States · US AQI (2024 revision)
EPA (current)In 2024 the EPA tightened the PM2.5 "Good" ceiling from 12.0 to 9.0 µg/m³. Atmos has not adopted these yet, so the same air can read higher here than in the app.
| Category | PM2.5 µg/m³ | PM10 µg/m³ | Index |
|---|---|---|---|
| Good | 0–9 | 0–54 | 0–50 |
| Moderate | 9.1–35.4 | 55–154 | 51–100 |
| Unhealthy for Sensitive Groups | 35.5–55.4 | 155–254 | 101–150 |
| Unhealthy | 55.5–125.4 | 255–354 | 151–200 |
| Very Unhealthy | 125.5–225.4 | 355–424 | 201–300 |
| Hazardous | 225.5–500.4 | 425–604 | 301–500 |
India · NAQI
Central Pollution Control BoardIndia tolerates more PM10 than the EPA before flagging concern, a reasonable response to natural dust. Its category names differ too, so "Moderate" in Delhi is not "Moderate" in Denver.
| Category | PM2.5 µg/m³ | PM10 µg/m³ | Index |
|---|---|---|---|
| Good | 0–30 | 0–50 | 0–50 |
| Satisfactory | 31–60 | 51–100 | 51–100 |
| Moderate | 61–90 | 101–250 | 101–200 |
| Poor | 91–120 | 251–350 | 201–300 |
| Very Poor | 121–250 | 351–430 | 301–400 |
| Severe | 251–500 | 431–600 | 401–500 |
China · AQI (HJ 633)
HJ 633-2012Numerically the category cutoffs match the EPA, but the concentrations behind them do not: 35 µg/m³ of PM2.5 is the top of "Good" here and already "Moderate" in the US.
| Category | PM2.5 µg/m³ | PM10 µg/m³ | Index |
|---|---|---|---|
| Good | 0–35 | 0–50 | 0–50 |
| Moderate | 36–75 | 51–150 | 51–100 |
| Unhealthy for Sensitive Groups | 76–115 | 151–250 | 101–150 |
| Unhealthy | 116–150 | 251–350 | 151–200 |
| Very Unhealthy | 151–250 | 351–420 | 201–300 |
| Hazardous | 251–500 | 421–600 | 301–500 |
Thailand · Thailand AQI
Pollution Control Department (2023)Thailand's public index is PM2.5-driven and tops out around 201 rather than 500, so its numbers are not comparable to a US AQI of the same value.
| Category | PM2.5 µg/m³ | Index |
|---|---|---|
| Very Good | 0–15 | 0–25 |
| Good | 15.1–25 | 26–50 |
| Moderate | 25.1–37.5 | 51–100 |
| Affecting Health | 37.6–75 | 101–200 |
| Hazardous | 75.1–500 | 201–500 |
WHO 2021 global guidelines
The World Health Organization publishes health-based targets rather than an index. They are not a national standard and no country's AQI is built on them directly, but they are the common yardstick behind the tables above and they are stricter than every one of them.
| Pollutant | Short-term | Window | Annual |
|---|---|---|---|
| PM2.5 | 15 µg/m³ | 24-hour | 5 µg/m³ |
| PM10 | 45 µg/m³ | 24-hour | 15 µg/m³ |
| NO₂ | 25 µg/m³ | 24-hour | 10 µg/m³ |
| O₃ | 100 µg/m³ | 8-hour peak | n/a |
| SO₂ | 40 µg/m³ | 24-hour | n/a |
| CO | 4000 µg/m³ | 24-hour | n/a |
How Atmos applies this
- Concentrations come first. Pollutants are read as µg/m³ from Open-Meteo, and the index is derived from them, not the other way round.
- Local standard where it matters. In India, China and Thailand the headline number is recomputed under CPCB, HJ 633 or PCD, so it matches what a local government app would show. Everywhere else defaults to the US scale.
- Absent is not zero. A pollutant a station does not report is shown as unavailable rather than as a reading of zero, which would look like clean air that was never measured.
- Known gap. Atmos still uses the EPA's pre-2024 PM2.5 breakpoints. The 2024 revision, listed above, is stricter, so the same air reads slightly better in the app than under the current US rule.