Thirty-six hours after this blog argued that Delhi's AQI was built on a hole, CPCB has confirmed the mechanism: readings above its 2015 calibration range of 0–1,000 µg/m³ are simply not used in the computation, and the validation methodology is under internal review. This sequel records what the board conceded, what it still owes — four specific fixes — and the two numbers that should end the argument: 91 of the world's 100 most polluted cities are Indian, and FICCI prices dirty air at $95 billion and 1.3 billion lost work days a year.
On 10 November, in “The Data Eclipse: Delhi's AQI Was Built on a Hole”, I said the blanks in the Diwali-night PM2.5 record were not dead sensors but an unexplained ceiling. I did not expect an answer within thirty-six hours. It came.
In an 11-point statement on X, reported by The Indian Economic Observer on 11 November, the Central Pollution Control Board has conceded the mechanism. The analysers deployed when the National Air Quality Index began in 2015 “were calibrated for a measurement range of 0–1,000 µg/m³”; automated validation therefore treats hourly readings beyond that range as “outside range”, and those readings “are not used in the computation”. Such hours, it adds, do not contribute to the station's 24-hour average.
A ministry source in the same report was blunter: “It is not that the data is missing. We have the data. However, that data is not used in the calculation.”
The concession is enough: those blanks were never failures. They were decisions.
The same source admitted the 2015 validation framework is outdated — instruments upgraded, methodology untouched, two or three scientists now reviewing it internally, no external expert group constituted the way one was in 2015. “This should have been revised two-three years ago,” he said.
My four asks stand where I left them:
- Publish a blank as ≥1,000, never as a null. Explaining a hole does not fill it.
- Time-stamp every revision, and leave the original bulletin standing. Not mentioned.
- Report an exposure index that does not stop at 500. CPCB says extrapolation “would not alter the assigned health impact category”. A worker at 3,700 and a cyclist at 400 share a category, not a risk.
- Write a station-siting rule. Silence.
The numbers have grown meanwhile. Amar Ujala reported on 11 November, citing aqi.in readings at 7 pm on Tuesday, that 91 of the world's 100 most polluted cities were Indian:
- Faridkot 605
- Bathinda 540
- Sri Ganganagar 535
- Sirsa 526
- Hisar 514
CPCB's own 4 pm bulletin that afternoon had Sri Ganganagar at 278, Sirsa at 322 and Hisar at 316. Two scales, one afternoon. The board calls that methodology, not under-reporting. I call it a reason to publish both.
And FICCI. NDTV reported on 11 November that its report with EY, “Clean Air, Healthy India”, prices dirty air at up to $95 billion a year — about 3% of GDP — with 1.3 billion working days lost and productivity falling 8–10% during bad periods. That is an employment figure, not an environmental one.
Except the trap inside it. If the ruler undercounts the peaks, it undercounts the cost. The economic case for clean air is being argued with a number that is too small.
I would rather CPCB's next statement carried four lines than a new committee. An internal review of a public instrument is not validation; it is housekeeping. Publish the lower bound, the time stamp, the number above 500, the siting rule. Then I will go back to writing about the air instead of the ruler.
With regards,
Hemen Parekh
www.Iam-Immortal.ai / www.IndiaAGI.ai / www.YourContentCreator.in
11 Nov 2026
No comments:
Post a Comment