SmogRank / History / Country race

Countries, worst air first, day by day

Each country's typical measured city over the 30 days to every day since January 2023, worst first. Press Play to watch the order change, or drag the slider to any day.

The 30 days to29 May 2025 That day's board

The worst countries for PM2.5 over the 30 days to 29 May 2025: Chile 40.7 µg/m³, Peru 35.7 µg/m³ and India 31.3 µg/m³.

  1. 1Chile47 cities40.7
  2. 2Peru3 cities35.7
  3. 3India198 cities31.3
  4. 4Mexico49 cities21.9
  5. 5South Korea140 cities15.4
  6. 6Hong Kong13 cities13.9
  7. 7Laos3 cities13.1
  8. 8Thailand67 cities12.9
  9. 9New Zealand6 cities11.3
  10. 10Malta4 cities10.9
  11. 11Taiwan45 cities10.6
  12. 12Serbia27 cities10.2
  13. 13Greece4 cities10.1
  14. 14Japan556 cities10.0
  15. 15Poland70 cities8.5
  16. 16Latvia4 cities8.2
  17. 17Hungary11 cities, was 37.9
  18. 18United Kingdom67 cities7.8
  19. 19Bosnia and Herzegovina5 cities7.7
  20. 20Slovakia41 cities7.7

+ 19 more countries, 7.2 µg/m³ and below

  1. 31United States836 cities5.4

Not counted on this day, too few cities measured: Lithuania since 25 May 2025, Estonia since 20 May 2025, Martinique since 3 May 2025 and 4 more countries.

GoodModerateSensitive groupsUnhealthyVery unhealthyHazardousWHO 24-hour guideline, 15 µg/m³Fades where it stood: not counted that day

A country's number is its typical measured city: the median of its cities' means over the 30 days (a city needs 21 of them measured; 5 of 7 on the 7-day race), shown only with 3 such cities. A country whose monitors go quiet fades out where it stood rather than sliding down, so a gap never reads as cleaner air; "was" beside a city count marks a month after many cities joined or left. Some large countries are thin or missing because their monitors' licences don't allow reuse. How we rank.

Not covered: the government monitoring networks of China and Turkey, which publish only on their own websites; we are building the readers for them. Most cities there are missing from this board, not clean. Why.