Most polluted cities in Asia
398 of 1,055 measured cities in 25 countries were above the WHO 24-hour guideline over the last 7 days. The highest checked peak: Multan, Pakistan, 630.6 µg/m³.
Nov 16 in Asia: 17,055 fire detections, ash warnings for 4 volcanoes.
Measured cities that day: 64 near many fires, 1 near a volcano with an ash warning. Left off the map: 3,273 detections at gas flares, steel works and other heat seen in the same place most days.
- Fires, seenHeat satellites detected that day (via NASA FIRMS); bigger and glowing means more detections in the area. Farm burning and other fires count too, not only wildfires. Satellites pass about twice a day and small fires often burn out between passes, so a lone small fire is drawn once it has been seen on two of the last three days, and faint marks are fires seen on both days before but not that day.
- Smoke, tracedOutlined by NOAA analysts from satellite images, light to heavy. Traced for North America only; it can be high above the ground.
- Smoke, modelWhere the Copernicus model put heavy smoke from burning that day, drawn only near fires satellites saw, and not over North America, where NOAA traces smoke instead.
- Dust, modelWhere the Copernicus model put thick desert dust in the air that day (dotted), denser dots for heavy dust. A model of the whole air column, not a measurement.
- Volcanic ash, warnedA volcano with an ash warning for aircraft that day; shaded where forecasters saw the ash cloud.
- Measured citiesEvery measured city, coloured by its PM2.5 that day (over the last 24 hours on Now); a ring where it had no reading. Larger and outlined where that day's evidence reached it: under smoke, 25+ fire detections within 100 km, an ash advisory from the Tokyo or Washington ash centres within 300 km, or thick dust in the Copernicus model (dashed: a model, not seen). Tap one of those to open it.
- 1–2 sensor citiesCities measured by only one or two low-cost sensors, as a hollow ring in the colour of their PM2.5. One cheap unit can read well off, so they take no rank and stay hidden until you switch them on.
Solid marks were seen, by satellites, NOAA's smoke analysts or aviation forecasters; dotted, hatched and dashed ones come from a model. Each day is a UTC day. Smoke outside North America isn't traced: there the model's heavy smoke near fires stands in.
Measured cities it reached
- Panipatfires211.0
- Multanfires176.8
- Kaithalfires163.9
- Gujranwalafires158.2
- Gwaliorfires151.0
- Sonīpatfires146.2
- Faisalabadfires134.1
- Khannafires128.7
- Karnālfires123.1
- Lahorefires115.6
- Butwālfires109.2
- Bāghpatfires104.8
And 53 more cities on the map.
A city's page names the likely cause under "Why it ranks here" when its own readings were bad that day; we don't model which way the wind blew.
Fires: NASA FIRMS (VIIRS, NOAA-20; by the hour on the live map, geostationary fire detections from GOES, Meteosat and Himawari via NASA FIRMS). Traced smoke: NOAA Hazard Mapping System. Volcanic ash: aviation ash warnings (SIGMETs) via NOAA's Aviation Weather Center, with the Tokyo and Washington Volcanic Ash Advisory Centres.
Ranked by checked peak: each city's worst 6 hours in a row, read at 2 or more monitoring locations 3 km or more apart (the second-highest location's average over those hours), so one faulty monitor or smoke over one corner of a city can't set it. 289 of these 1,055 cities are ranked: those measured at 2 or more such locations. A city needs a checked peak on 3 or more days, so some rest on fewer days than their average does (a network that went quiet on some days); their count is shown under the value. More days could only raise a peak, never lower it. A peak in a city measured by low-cost sensors comes from those sensors alone, and they can read well above government-grade monitors in smoke and humid air.
Countries
Each tick is a measured city on a log scale in µg/m³, coloured by its band; the tall bar is the median, the dashed line the WHO 24-hour guideline.
| Country | Cities | Typical city (median) | Every city, µg/m³ | Worst city |
|---|---|---|---|---|
| Nepal | 3 | 124.0 | Worst: Butwāl 126.3 µg/m³ | |
| Pakistan | 17 | 120.4 | Worst: Multan 271.2 µg/m³ | |
| Bangladesh | 1 | 96.5 | Worst: Dhaka 96.5 µg/m³ | |
| India | 225 | 58.9 | Worst: Delhi 178.9 µg/m³ | |
| Kyrgyzstan | 1 | 41.9 | Worst: Bishkek 41.9 µg/m³ | |
| Kazakhstan | 4 | 35.5 | Worst: Talghar 39.3 µg/m³ | |
| Indonesia | 6 | 34.6 | Worst: Jakarta 60.1 µg/m³ | |
| Vietnam | 1 | 33.0 | Worst: Hanoi 33.0 µg/m³ | |
| Uzbekistan | 9 | 30.9 | Worst: Nukus 47.4 µg/m³ | |
| United Arab Emirates | 1 | 24.8 | Worst: Dubai 24.8 µg/m³ | |
| Bhutan | 1 | 24.8 | Worst: Phuntsholing 24.8 µg/m³ | |
| Armenia | 2 | 24.3 | Worst: Yerevan 40.2 µg/m³ | |
| Laos | 1 | 24.2 | Worst: Vientiane 24.2 µg/m³ | |
| Turkey | 1 | 21.3 | Worst: Istanbul 21.3 µg/m³ | |
| Azerbaijan | 1 | 20.5 | Worst: Baku 20.5 µg/m³ | |
| South Korea | 138 | 15.8 | Worst: Pubal 33.3 µg/m³ | |
| Thailand | 69 | 15.8 | Worst: Samut Sakhon 28.9 µg/m³ | |
| Singapore | 1 | 14.4 | Worst: Singapore 14.4 µg/m³ | |
| Philippines | 2 | 13.4 | Worst: Quezon City 17.1 µg/m³ | |
| Kuwait | 6 | 12.7 | Worst: Şabāḩ as Sālim 15.0 µg/m³ | |
| Hong Kong | 1 | 12.1 | Worst: Hong Kong 12.1 µg/m³ | |
| Taiwan | 43 | 11.9 | Worst: Erlun 19.6 µg/m³ | |
| Lebanon | 2 | 11.2 | Worst: Beirut 15.4 µg/m³ | |
| Japan | 518 | 8.2 | Worst: Kamisu 22.2 µg/m³ | |
| Tajikistan | 1 | 4.9 | Worst: Dushanbe 4.9 µg/m³ |
Highest checked peaks
- 1MultanPunjab, Pakistan 500+ US AQI630.6 µg/m³
- 2DelhiDelhi, India 500+ US AQI403.4 µg/m³
- 3SargodhaPunjab, Pakistan 500+ US AQI384.2 µg/m³
- 4MeerutUttar Pradesh, India 500+ US AQI382.6 µg/m³
- 5LahorePunjab, Pakistan 500+ US AQI362.2 µg/m³
- 6PeshawarKhyber Pakhtunkhwa, Pakistan 500+ US AQI330.6 µg/m³
- 7GhāziābādUttar Pradesh, India 494
- 8GujranwalaPunjab, Pakistan 448
- 9DhakaDhaka Division, Bangladesh 406
- 10BhiwadiRajasthan, India 401
- 11GwaliorMadhya Pradesh, India 400
- 12Dera Ghazi KhanPunjab, Pakistan 390 US AQI4 of 7 days
- 13FaisalabadPunjab, Pakistan 388 US AQI4 of 7 days
- 14FaridabadHaryana, India 359
- 15PuneMaharashtra, India 346
- 16ButwālLumbini Province, Nepal 343
- 17MorādābādUttar Pradesh, India 340
- 18SialkotPunjab, Pakistan 337
- 19GurugramHaryana, India 328
- 20GujratPunjab, Pakistan 316
- 21Otegen BatyraAlmaty Region, Kazakhstan 313
- 22AhmedabadGujarat, India 310
- 23BishkekBishkek, Kyrgyzstan 279
- 24NoidaUttar Pradesh, India 269
- 25JamrudKhyber Pakhtunkhwa, Pakistan 266
What stands behind each figure: Monitors 3 km+ apart One monitor location Low-cost sensors About these marks
Low-cost sensors: consumer units, counted only in cities with no government-grade monitor; they read close to reference monitors on the whole, though AirGradient's read 15 to 25% high in humid air, and Sensor.Community's read well below. How we use them. How accurate low-cost sensors are, from our own measurements.
All cities in the worldNot covered: the government monitoring networks of China, Turkey, Iran and Georgia, which publish live readings only on their own websites. The networks of Turkey and Iran don't answer requests from outside their countries, so we can't read them. A public mirror of China's release platform gives us its national stations' hourly readings a day late, too late for live boards: past boards from January 2023 on rank its cities on them. Past boards from January 2023 to December 2025 rank Turkey's cities on its verified hourly readings, which the European Environment Agency publishes. Georgia's network publishes running 24-hour means, not hours, so we take its daily means as its agency publishes them, a day late, too late for live boards: past boards from January 2023 on rank its cities on them. Most cities there are missing from this board, not clean. Why.
Updated 2026-11-17 11:07 UTC. Values are PM2.5 (fine particles), the median of each city's monitors, as the US AQI PM2.5 sub-index. Some low-cost readings: Sensor.Community, ODbL (data and licence). Some readings: Uzhydromet (Uzbekistan's Hydrometeorological Service); Vietnam's government monitors; Environment Public Authority, Kuwait, no reuse terms published (why we use them).