The Data Eclipse: Delhi's AQI Was Built on a Hole

Delhi's PM2.5 record went blank on Diwali night precisely as readings crossed about 1,000 µg/m³, the analysers' ceiling — and the blanks pulled the official AQI down to 344 before the missing hours were released. This post argues that an air-quality number is an audit trail, not a headline, and asks for four things: blanks published as ≥1,000 lower bounds, a time-stamped revision protocol, an exposure index that does not cap at 500, and a station-siting rule. Enforcement, ROTE and Trans-Tax, can only work on a number we can trust.

The highest PM2.5 figure in the Central Pollution Control Board's record for Diwali night is exactly 1,000.0 — at Sirifort for the quarter-hour from 4.45am, and again at ITO from 6.15am. Nothing in that file goes above 1,000. Hindustan Times reported this on 10 November, and that flat ceiling is the real story.

The instruments had not stopped. All night the same stations kept sending nitrogen dioxide and carbon monoxide. Only the pollutant that was turning inconvenient went blank, and it went blank as it neared 1,000 micrograms per cubic metre — what Manoj Kumar of the Centre for Research on Energy and Clean Air calls the instruments' upper measurement limit.

These are the figures HT drew from CPCB's records and from the Delhi Pollution Control Committee's release the following evening:

• 43 of 59 stations had a PM2.5 gap between 10pm and 6am;
• 148 of 351 station-hours were blank — 42%, against a 7% baseline for the season;
• every one of the 125 missing station-hours at DPCC's stations was above 1,000 on DPCC's own site;
• the 1am hour averaged 1,480 µg/m³ across 22 stations; in CPCB's record it averaged 547;
• Mundka touched 3,744 at 2am, JNU 3,023 at 4am, NSUT West 3,008 at 5am.

Meanwhile the number that became the day's permanent record: 344, Delhi's AQI in the 4pm bulletin. Counting the hours DPCC released later, HT estimates that bulletin should have read about 360, and the published 9am figure of 370 about 384. Both are floors — the blanks at CPCB's and IITM's stations were never filled. CPCB has not explained the outage; AAP's Saurabh Bharadwaj has called it data fraud.

"These observations should not simply be treated as missing when calculating hourly or daily averages or deriving AQI. Doing so could underestimate pollution levels." Kumar told HT that 1,000 should be read as a lower bound.

Anumita Roychowdhury of the Centre for Science and Environment called it a data eclipse. I would put it more plainly. An air-quality number is not a headline; it is an audit trail. A headline can be argued with. An audit trail either holds or it does not.

So, four things I would ask of CPCB, the Delhi government and the Commission for Air Quality Management:

1. Publish a blank as ≥1,000, never as a null. A missing hour is evidence, not absence.
2. Issue a revision protocol: a revised bulletin is reissued with a time stamp, and the original stays visible.
3. Report an exposure index that does not stop at 500. India's scale ends at 500; the particles do not.
4. Write a station-siting rule. Twenty new stations, thirteen added in October, many on the city's edge, will pull an average down if siting follows convenience rather than traffic and people.

None of this is exotic. In Project ROTE, in June 2025, I proposed NaVIC-tracked vehicles, a Harm Quotient and a Trans-Tax to price emissions in real time. But enforcement needs something to enforce against. Fix the instrument first, or the tax chases a number that lies.

I have written about Delhi's air for years, and the lesson has not changed. We do not lack data. We lack the discipline to publish the hours that embarrass us.

With regards,
Hemen Parekh
www.Iam-Immortal.ai / www.IndiaAGI.ai / www.YourContentCreator.in
10 Nov 2026