Delhi's 211 days of good-to-moderate air, checked day by day

Published: 19 October 2026 | Author: Team JanVayu | Reading time: 7 min

On 17 October Delhi's Environment Minister said the capital had recorded 211 days of good, satisfactory or moderate air so far in 2026, the best since 2016 if 2020 is set aside, and set it beside 200 such days in 2025, 209 in 2024, 206 in 2023, 159 in 2018 and 110 in 2016. Poor-to-severe days, he said, had fallen from 168 in 2016 to 79 (IANS report of the Delhi government's statement, 17 October 2026).

We hold CPCB's own daily AQI bulletin for Delhi for every day since May 2015, so we counted. The 2026 figure is right to within a day. The years it is compared with are counted over a different stretch of the calendar, and once that is fixed 2026 still leads, by a smaller margin and for a narrower reason than the list suggests.

The 2026 count, to 17 October

CPCB publishes one AQI figure for Delhi every afternoon, with a category: Good (0 to 50), Satisfactory (51 to 100), Moderate (101 to 200), Poor (201 to 300), Very Poor (301 to 400) or Severe (above 400). From 1 January to 17 October 2026 the bulletin gave Delhi:

That is 210 days at Moderate or better and 80 at Poor or worse. The government's figures are 211 and 79, so the totals differ by one day. The 18 October bulletin, which rated Delhi Moderate at 187, brings our count to 211.

We cannot say which day it is, because the government did not publish its day-by-day list. For 2023, 2024 and 2025 our January-to-September counts match the ones the Environment Ministry published on 30 September exactly (193, 188 and 187 days), which suggests both counts are read from the same CPCB bulletin. For 2026 we have 193 for those nine months against the Ministry's 194, and that one day carries through to October.

One practical note for anyone repeating this. On 1 to 9 July CPCB's bulletin spelled the city delhi, in lower case. A count that filters on the exact word Delhi loses those nine days.

The comparison years are whole years

Here is the same count for every year since 2016, done two ways: from 1 January to 17 October, the window the 2026 figure covers, and for the full calendar year.

The good-day figures quoted for 2025, 2024, 2023, 2018 and 2016 (200, 209, 206, 159 and 110) are the full-year column, exactly. The poor-day figure quoted for 2016, 168, sits in the year-to-date column, a day off ours. So the statement used one window for the poor days and another for the good days, and set nine and a half months of 2026 against twelve months of each earlier year.

Here the mismatch works against the claim, and the claim survives it. Counted over the same window, 2026 is ahead of every year since 2016 except 2020. The nearest is 2023, at 203, a lead of 7 days on our count. The list as quoted made 2024 the nearest, 2 days behind. Poor-or-worse days to 17 October are also the fewest outside 2020.

The Ministry's own release of 30 September did compare like with like, January to September, and on that basis 2026 led 2023 by one day, 194 to 193.

Our 2016 figures rest on 279 bulletins for the 291 days to 17 October; twelve days of that year have no Delhi bulletin in the archive. Every later year is complete or nearly so.

The lead over 2023 is October

The months show where the 7 days came from.

July to September is the monsoon, and every recent year is at or near the ceiling there, so those months cannot separate one year from another. The difference sits in the shoulder months. Over the first nine months 2026 and 2023 are level on our count, at 193 each. The whole of the lead comes from 1 to 17 October, when 2026 had 17 days at Moderate or better and not one Poor day; 2023 had 10 and 7. February 2026 was the worst February of the four years shown.

That matters for what comes next. In each of the ten years from 2016 to 2025, between 66 and 74 of the 74 days after 18 October were rated Poor or worse. The days still to come in 2026 are the ones Delhi usually loses. On that record the full-year count will probably finish between 211 and about 219, and whether 2026 ends ahead of 2023 and 2024 will be settled in November and December. The full-year comparison can be made fairly on 1 January 2027.

"Moderate" is above the national standard

The phrase "good-to-moderate" sounds like clean air. Most of it is not.

India's AQI scale puts the top of Satisfactory, 100, at the national 24-hour standard for each pollutant. For PM2.5 the standard is 60 µg/m³, and a Moderate day covers 61 to 90 µg/m³. For PM10 the standard is 100 µg/m³ and Moderate runs to 250 (CPCB, National Air Quality Index, 2014). A Moderate day is therefore a day on which the pollutant that set the index was above the national 24-hour limit. The WHO's 24-hour guideline for PM2.5, set in 2021, is 15 µg/m³.

Of the 210 days to 17 October, 148 were Moderate. Days at Satisfactory or better, the days on which every reported pollutant was within the national standard, numbered 62 in 2026, against 79 in 2025, 66 in 2024 and 61 in 2023. So 2026 has fewer of those days than 2025. What changed is that many days which in another year would have been Poor came in as Moderate. That is a real improvement, and it is smaller than "211 good days" implies.

The pollutant that set the index on the 148 Moderate days was PM10 on 80, PM2.5 on 35 and ozone on 32. Coarse dust drove more of Delhi's Moderate days this year than fine particles did.

The fine-particle figure did not move

The same Ministry release of 30 September carries a number that the count of days leaves out. Delhi's average PM2.5 concentration from January to 28 September was 69 µg/m³ in 2026 and 69 µg/m³ in 2025, against 81 in 2024, 73 in 2023 and 80 in 2022. The release calls 2026 "joint-lowest", which is accurate: it is level with last year.

PM2.5 is the pollutant most closely tied to illness and early death, and its concentration is a quantity that can be averaged, unlike the AQI. On that measure 2026 is no better than 2025 so far. Both years sit well above India's annual standard of 40 µg/m³ and about fourteen times the WHO's annual guideline of 5 µg/m³.

So more days crossed into Moderate this year while the average exposure to fine particles stayed where it was. Both statements come from the Ministry's own figures, and both are true. The count of days is the one that made the headline.

The instrument has changed under the count

A city's AQI comes from the stations reporting that day, and Delhi's network has grown a great deal. The median number of Delhi stations reporting in each year's bulletin:

It is still growing. The bulletins of mid-October 2026 list 57 to 59 Delhi stations, of which 51 to 56 reported.

The comparison with 2016 sets an index built from five stations against one built from more than forty, and the count alone cannot say how much of the difference is the air and how much is the network. From 2019 the network is steadier, so the recent years are the fairer comparison. This is why our own bulletin layer publishes counts for each year and declines to draw a trend through them (no_trend in aqi-bulletins.json).

Weather can move a count like this on its own

A year's count of good days carries that year's weather: when the monsoon arrived and left, how much it rained, how often the wind blew. The India Meteorological Department declared the monsoon withdrawn from Delhi on 23 September, two days before the normal date, after a season wetter than normal (press reports of IMD's statement, 24 September). Ten months of one year cannot separate a policy effect from a kind season.

The way to separate them is to remove the weather first and look at what is left, which is what our weather-normalised trends do for 44 cities. That analysis currently runs to 2024. Until a weather-adjusted 2026 exists, the fair reading is that 2026 has so far been better than recent years, and that the cause is still open.

What would make a claim like this checkable

All of this CPCB already has:

• State the window. Say "1 January to 17 October" and compare the same dates in earlier years, as the 30 September release did.
• Count the days within the standard separately. Satisfactory or better is the count that means the air met the national limit; a Moderate day did not.
• Report PM2.5 beside the day count. The Ministry already does so in its own release.
• Say how many stations each year's figure rests on, so a reader can tell a change in the air from a change in the network.

Check it yourself

Every number in this post comes from CPCB's daily bulletin. The 2016 to 2025 bulletins are the UrbanEmissions.Info extraction of CPCB's PDFs (github.com/urbanemissionsinfo/AQI_bulletins); the 2026 bulletins we parse ourselves and publish as data/raw/cpcb-bulletins-2026.csv.gz, one row per city per day. The per-year counts are in data/aqi-bulletins.json, and the Delhi page carries the month-by-month picture.

The minister's statement is quoted from the IANS report of 17 October (via SocialNews). The Environment Ministry's January-to-September figures, including the PM2.5 averages, are from its PIB release of 30 September 2026, which we read on 19 October.

If you find a day we have miscounted, write to contribute@janvayu.in and we will correct it in public.