Delhi's Most Polluted Diwali in Five Years: AQI 376, Blank Monitors and the Blame Game, Explained

Delhi's Diwali day was its dirtiest in five years. The Times of India reported an air quality index of 376 at 4 pm on Sunday, 8 November 2026, the worst Diwali-day reading since 382 in 2021. By Monday morning the city was still in the very poor band on the 24-hour average, while individual stations had already crossed into severe.

The same night, a large part of the monitoring network stopped publishing the dust reading that mattered most. Hindustan Times found a gap in PM2.5 at 43 of 59 stations between 10 pm and 6 am. What that gap means is now a technical argument and a political one.

The 60-second version

• Diwali day: AQI 376 at 4 pm on 8 November, the most polluted Diwali day in five years and the worst since 382 in 2021, The Times of India reported. Diwali eve, 7 November, was 369, which the paper called the most polluted since 2016.
• Monday: The Hindu, citing the CPCB live dashboard, put the city at 404 at 4 am (severe) and the 24-hour average at 370 at 9 am (very poor). The Times of India then put Monday at 370 at 11 am and 344 by 4 pm.
• Stations: On the CPCB list reported by The Hindu and NDTV, Burari Crossing was 441, Jahangirpuri 433 and Rohini 430. Mundka was 417. Anand Vihar was 335.
• The blanks: Hindustan Times counted PM2.5 gaps at 43 of 59 stations from 10 pm to 6 am. At 14 of those, the last reading was 900 or higher. CREA found 36 of 59 stations not reporting at 1 am, NDTV reported.
• The politics: AAP's Saurabh Bharadwaj alleged data fraud. BJP minister Kapil Mishra alleged a stubble-burning conspiracy and cited about 1,200 fire locations. The reported satellite count for Punjab on 8 November is 1,152.
• The rules still on: GRAP Stage 3, explained in GRAP Stage 3 in Delhi-NCR. Mumbai banned aerial fireworks until 25 November after 137 cracker fires.

Diwali day: 376, and how that compares

The Times of India (Kushagra Dixit and Priyangi Agarwal, timestamped 1:45 am IST on 10 November 2026) said Delhi saw its most polluted Diwali day in five years. The reading it used was 376 at 4 pm on 8 November. Hindustan Times, in a separate story, also said the city's average AQI on Diwali was 376, against 369 a day earlier.

The paper's Diwali-day series:

Diwali	Date	AQI
2026	8 November, 4 pm	376
2025	20 October	345
2024	31 October	328
2023	12 November	218
2022	24 October	312
2021	4 November	382

376 is below 2021. It is above every Diwali day in the four years after that. That is the sense in which the paper called 2026 the worst in five years, and the worst since 2021.

The day after is a different comparison. The Times of India said the city "narrowly escaped" its worst post-Diwali day in five years. Monday's index touched 370 at 11 am and improved to 344 by 4 pm, which the paper said was better than the 2025 post-Diwali reading. Its post-Diwali series: 351 in 2025, 339 in 2024, 358 in 2023, 303 in 2022 and 462 in 2021.

Two other Sunday figures from the same story: only 47 stations contributed to the Diwali-day average, and hourly data were initially missing for 35 stations. The paper said Delhi has 60 monitoring stations, run by the DPCC, the CPCB, IITM and the IMD, of which one had been defunct for a few months. Hindustan Times and CREA counted 59. This piece keeps both counts, with the source attached.

The Indian Express, updated early on 10 November, said the capital recorded its worst Diwali-day air since 2021, and that the average at 9:05 am on Monday was 370, with eight stations above 400.

Monday morning: severe at some stations, very poor as a city average

A single city number hides the spread. These are not one bulletin reprinted three ways. They are different hours, and in one case a different dashboard.

The Hindu, updated at 4:57 pm IST on 9 November and citing the CPCB live dashboard, said the index went up to 404 at 4 am, which is inside the severe band of 401-500, and that it had eased after that. The same report put the 24-hour average at 370 at 9 am, in very poor (301-400).

The Hindu's highest station readings, which match the CPCB list NDTV India also published:

Station	AQI
Burari Crossing	441
Vivek Vihar	437
Delhi Technological University	434
North Campus, Delhi University	433
Jahangirpuri	433
Rohini	430

NDTV India, citing CPCB, extended that Monday list. Other severe stations included Ghazipur 418, Mundka 417, Punjabi Bagh 409, NIH Narela 407 and ITO 407. Anand Vihar was 335, which is very poor, not severe. Akshardham was 337. NDTV said data were unavailable for IIT Delhi and Dwarka Sector 8.

Hindustan Times, in its report on the AAP allegation, gave a different morning snapshot from the CPCB Sameer app and the HT Pollution Dashboard: a city average of 366 at 6:15 am, and higher station scores than the list above. On that dashboard, Sonia Vihar was 463, Burari Crossing 452, Jahangirpuri and Vivek Vihar 447 each, Rohini 442, North Campus 440, DTU 436, Bawana 431, Chandni Chowk 423, Patparganj 419 and NIH Narela 410.

Use the label on the column. A station can be 433 on one CPCB pull and 447 on another dashboard later or earlier the same morning. Neither figure, on its own, is Monday's city AQI. The city figures in print from these reports are 404 at 4 am, 366 at 6:15 am, 370 at 9 am and at 9:05 am and at 11 am, and 344 by 4 pm.

The Times of India also said Monday brought the season's first dense fog, with visibility down to 50 metres at Safdarjung at 6.30 am.

The monitors went blank

The public argument starts from a narrower fact than "the stations were switched off."

Hindustan Times (Binayak Dasgupta, updated 3:19 pm IST on 9 November) read CPCB's Central Control Room records and found that PM2.5 at many stations climbed towards 1,000 micrograms per cubic metre and then went blank, while the same stations kept sending readings for gases. PM2.5 is fine dust, particles 2.5 micrometres across or smaller.

Across all 59 Delhi stations, 43 had a gap in PM2.5 between 10 pm and 6 am. At 14 of those 43, the last reading before the gap was 900 or more, within 100 of the 1,000 mark. Stations run by all three operators in that count went blank: CPCB, IITM and the DPCC.

HT compared 39 stations that also reported last Diwali:

• For the 1 am hour on 9 November, 13 of the 39 sent a PM2.5 reading and 38 sent nitrogen dioxide.
• Through the night, at least 37 of the 39 kept sending nitrogen dioxide, carbon monoxide and ozone.
• From 10 pm to 6 am those 39 stations should have sent 351 hourly PM2.5 readings. 148, or 42 percent, were blank. On ordinary nights in the days before Diwali, the same stations left 3.9 percent blank.
• Last Diwali, at least 31 of the 39 sent PM2.5 in every hour after 10 pm. This year the count was 13.
• No PM2.5 reading from any Delhi station went past 1,000 on the feed HT read. The highest values were exactly 1,000.0, at Sirifort and ITO, sent in the early morning after those stations had come back. Last Diwali, 29 of the 39 recorded at least one hour above 1,000. The highest then was 2,224 at NSIT Dwarka, from a copy of the CPCB feed kept by OpenAQ.

The climb, before the gaps, was steep. At North Campus, an IITM station, PM2.5 went from 326 micrograms per cubic metre at 8 pm to 999 at 10 pm, and the next readings were blank. At CRRI Mathura Road, also IITM, it went from 394 to 981 in the two hours to 10.30 pm. ITO, a CPCB station, last sent 994 at 10.45 pm.

PM10, which counts coarser dust as well, dropped out sooner. For the 1 am and 2 am hours, only 10 of the 39 stations sent a PM10 reading.

NDTV's account of a CREA analysis matches the shape of that night. The number of stations not reporting rose from seven at 10 pm on 8 November to 23 at 11 pm, 29 at midnight and 36 at 1 am on 9 November. Across 59 stations, CREA counted 227 missing hourly observations between midnight and 5 am. At RK Puram, PM2.5 rose to 924 before the station stopped displaying readings after 9 pm. NDTV also reported gaps at Nehru Nagar, Mundka, Sonia Vihar and Dwarka.

CREA analyst Manoj Kumar told NDTV that the timing of the missing data alone did not establish the cause. He said PM2.5 at Siri Fort reached 1,000, the maximum displayed on CPCB's public portal, and that there was a possibility the official AQI did not fully capture how severe the episode was.

What the blanks do to the city number

HT applied CPCB's method to the 39 stations for the 24 hours to 9 am on 9 November. Counting only the hours that were recorded, HT's estimate of Delhi's AQI was 437. Filling each blank with a straight line between the last reading before it and the first reading after, the estimate was 471. CPCB's official AQI at 9 am, for all Delhi stations, was 370. On the 39 stations alone, CPCB's figure was 381.

The blanks pull a 24-hour average down because the missing hours are the worst hours, and because a station needs 16 hours of readings to count. HT said seven stations, including Nehru Nagar and Pusa (IITM), sent no PM2.5 at all from 10 pm to 6 am and so dropped out of the city average. A straight line also cannot rise above the readings on either side. If the air went past 1,000 in the gap, HT said, the estimate would be higher than 471, though not higher than 500, where the scale stops.

In the hour from 2 am, 14 of the 39 stations were still sending PM2.5, and their median was 693 micrograms per cubic metre. HT said that hour sat past the top of the AQI scale. It is not the day's AQI, which is a 24-hour average.

A ceiling, a glitch, or something else

Anil Gupta, described by NDTV as a member of the DPCC and the CPCB, attributed the gaps to the size of the spike. "When there is a sudden spike in PM2.5 and it crosses 999, the monitoring station stops giving a reading," he said. "There is no deliberate move. There is no malfunction. It is a technical glitch." Asked whether the missing figures were stored inside the station, he said: "There is no internal mechanism."

NDTV also noted that CPCB's 2019 technical specifications prescribe a measurement range of up to 1,000 micrograms per cubic metre for certain particulate analysers, and that the same specifications provide for station-level data storage, diagnostic records and error logs. Reaching a ceiling does not, by itself, prove why a particular station went blank. Storage on paper does not prove that a valid measurement was kept after an instrument went out of range.

Dipankar Saha, a former CPCB additional director, told NDTV he did not accept the high-value explanation. "In my opinion, a glitch has nothing to do with high or low values," he said. Technical trouble, he said, can sit at the analyser or at the display. He called for the monitoring system to be investigated regularly if the public is to trust it.

HT's own reading was that the pattern fits a limit at about 1,000 on the dust channel, either in the monitors or in CPCB's data system. A station that loses power or connection stops sending everything. These kept sending gases. HT said it had sent queries to CPCB. The story, as updated on Monday afternoon, did not carry a CPCB reply.

The numbers that appeared later

The public feed HT read on Monday morning did not show a PM2.5 value above 1,000. The Times of India, in its early Tuesday story, said the DPCC later updated data at some of its stations, and it published peaks far above that line.

At midnight on 8 November, IITM recorded its highest average PM2.5 concentration, at 671.1 micrograms per cubic metre. Among stations, the paper reported a Jahangirpuri peak of 4,985 on Sunday night and Monday morning, and said the DPCC later updated a PM2.5 level of 776 at 11 am on Sunday. Mundka was 3,744 at 2 am, which the paper said remained unchanged, and JNU was 3,023.2. The prescribed 24-hour national limit for PM2.5 is 60.

Those station peaks and the 24-hour limit answer different questions. A reading at 2 am is not a 24-hour average. The city's average PM2.5 concentration on Diwali 2026, The Times of India said, was 257 micrograms per cubic metre, against 228 last year, 201 in 2024, 134 in 2023, 176 in 2022 and 318 in 2021.

CREA's Manoj Kumar told NDTV that the same 257 made this the fourth most polluted Diwali since 2016. That ranking is not the same claim as "the worst Diwali day in five years." One is an average concentration of fine dust, set against Diwalis back to 2016. The other is the AQI on the festival day, set against 2021 to 2025. Both can be reported. They should not be collapsed into one sentence.

The Times of India also said firecrackers were the biggest contributor a day before Diwali, accounting for 41.1 percent of the total load in IITM data up to 9 am on 8 November. Traces of lead, potassium and sulphur were confirmed. Officials told the paper they could confirm mercury only in a longer study.

Sunil Dahiya, founder of Envirocatalysts, estimated that crackers added roughly 450 micrograms per cubic metre to peak-hour PM2.5 and about 270 to 400, more than half of what people breathed, between 9 pm and 2 am. He said the pollution built from about 8 pm, peaked between 1 am and 3 am, and stayed above 500 for four hours. That is his analysis, as the paper reported it.

Anumita Roychowdhury of the Centre for Science and Environment called the morning after "the direct result of a toxic convergence of adverse winter meteorology, escalating pre-Diwali traffic gridlocks and a rise in contribution of farm fires." Firecrackers, she said, were "an uncontrolled tipping point."

What the court allowed, and what people did

The Supreme Court had allowed only certified green crackers on 7 and 8 November, and the complete ban returned on 9 November. The hours and the no-cracker zones are in Green crackers allowed in Delhi-NCR for Diwali 2026.

The Hindu reported that people defied the deadline well past 10 pm on Diwali night, and that traditional crackers, not only the certified green ones, were also available in Delhi markets. The Indian Express quoted an 81-year-old Moti Nagar resident, Baboo Lal Gupta, who said crackers in his colony went on until 3 am.

Farm fires: 1,152 on the satellite count, and a conspiracy allegation

Punjab recorded 1,152 active farm fires on 8 November, the highest single-day count in the state this season, The Times of India and The Indian Express reported. The Indian Express said the state had 3,307 fires up to 7 November, which was 77 fewer than the 3,384 in the same stretch last year, a decline of 2.3 percent. The Diwali-day spike took the cumulative count to 4,459 on 8 November, against 3,622 at the same point last year, an increase of 23.1 percent.

That count is not, by itself, a finding about who lit the fires.

Delhi minister Kapil Mishra alleged on Monday that the fires were set on purpose. The Indian Express said he did not blame farmers. He said people entered fields and lit them while farmers were celebrating at home, and that the Delhi government had started receiving such calls on Diwali evening. He said satellite data showed roughly 100 to 125 locations a day from 5 to 7 November, and about 125 on 7 November, then 1,200 locations on Diwali evening.

"How did the number suddenly rise to 1,200 locations on Diwali? How did the figures increase by more than 10 times?" he said. "This is a conspiracy against farmers, a conspiracy against Diwali, a conspiracy against Delhi and a conspiracy against Punjab."

He asked Punjab chief minister Bhagwant Mann to register FIRs wherever fields were set alight on the night of 8 November, and to examine phone records of AAP MLAs and leaders in those areas, including people who had travelled from Delhi to Punjab. He said that if the inquiry were honest, "it is possible that Kejriwal himself could end up in jail." He also claimed Monday-morning AQI readings of 1,238 in Malout, 1,307 in Abohar and 1,651 in Sangrur. Those Punjab figures are his claims, as The Indian Express reported them.

1,200 and 1,152 are not the same number. The first is the minister's description of the satellite jump. The second is the active-fire count The Times of India and The Indian Express reported for 8 November. The count shows how many fires were recorded. It does not, in the reporting above, name who lit them. A festival-night jump is compatible with more than one explanation, including farmers burning residue that evening. The conspiracy claim is his.

Baltej Pannu, AAP's media in-charge in Punjab, said the BJP was targeting the state's farmers. "When paddy harvesting had not even started in Punjab, BJP leaders in Delhi had started saying that smoke coming from Punjab had covered Delhi," The Indian Express quoted him. He said a BJP leader was producing "a new way to defame Punjab's farmers" every day, and that the state had brought stubble burning largely under control.

Gufran Beig, founder of SAFAR and chair professor at the National Institute of Advanced Studies, told The Times of India that an unusual turn to easterly winds, linked to an approaching cyclone and a western disturbance, kept stubble smoke from Punjab and Haryana away from Delhi on Monday. He put those winds at about 6.7 metres per second from the surface to nearly 400 metres, against a usual 3 to 4 metres per second. He said Delhi was "very lucky" this year, not cleaner. The Times of India said those easterlies were expected to persist until Tuesday morning, and that the wind could then turn back toward the north-west. The paper said the post-Diwali easing was more meteorological than the result of pollution-control measures.

What AAP alleged about the data

AAP leader Saurabh Bharadwaj alleged "AQI data fraud" and said readings were being manipulated to pull the city's average down. Hindustan Times reported the allegation on Monday.

In a video and in remarks to ANI, he claimed the government had put 20 new stations in open, green areas, and that water was being sprinkled around the Jahangirpuri monitor to lower its reading. "I, for one, do not trust it," he said of the government AQI. "Last night, several areas in Delhi had an AQI of more than 3000." He said he had written repeatedly to CAQM and the CPCB. "Everyone is involved in this fraud," he said, and he warned that wrong numbers put "the sick, the elderly, and children" at risk.

Two limits on that claim follow from the reporting, not from a finding that he is wrong about the gaps.

First, the gaps are real on the record HT and CREA read. A missing hour at the peak is a real problem for a 24-hour average, which is the point of HT's 437 and 471 against CPCB's 370. That is a data problem. It is not, by itself, a finding of fraud.

Second, NDTV noted that India's AQI turns pollution into a score capped at 500. A claim of an AQI above 3,000 is not a CPCB index. It can be a description of a concentration, or of a private scale. Bharadwaj stated it as an AQI. Gupta's account, that the monitors stop displaying PM2.5 above 999, is the explanation on offer from a member of the two pollution bodies. Saha's account is that a glitch is not explained by the value. Neither account is a court finding, and Bharadwaj's fraud charge remains an allegation.

GRAP, and what the IMD has actually said

The Hindu reported on Monday that GRAP Stage 3 remained in force across Delhi and the adjoining districts of Haryana, Uttar Pradesh and Rajasthan. It described GRAP as a step that can be taken on forecasts from the IMD and IITM before the citywide index itself is in the severe band. What Stage 3 restricts, including construction, the BS-III and BS-IV vehicle limits, and hybrid classes up to Class V, is set out in GRAP Stage 3 in Delhi-NCR.

The weather outlook is not one sentence.

The IMD's all-India bulletin issued at 8:21 am IST on 8 November, before Diwali night, said a fresh western disturbance was likely to affect the northwestern Himalaya from 10 November, and that isolated to scattered rain was likely over Haryana, Chandigarh and Delhi, and over Punjab, during 11 to 13 November.

After Diwali, IMD scientist Krishna Mishra told NDTV India that, for the moment, Delhi-NCR should not expect rain or a change in the wind. He said a western disturbance was expected over Jammu and Kashmir from Monday night, with heavy rain and strong winds there, and that it had not been forecast to bring rain or strong winds to the other northern states unless it strengthened. NDTV's report of that interview said relief was not expected for the next 48 hours.

Beig's easterly, described above, is a scientist's account of why Monday did not get worse, not an IMD press note. The 8 November bulletin's rain window for Delhi is 11 to 13 November. Mishra's comment, given after the festival, was that the next two days did not look like relief. Both are in the record. Tuesday morning's wind is the first test of which description the city actually gets.

Mumbai banned aerial fireworks until 25 November

The same Diwali night produced a different order in Mumbai. The Hindu reported that the city banned rockets and other aerial firecrackers until 11.59 pm on 25 November, after 137 firecracker-related fires on 8 November alone. The BMC's count, as The Hindu and The Indian Express reported it, was 142 fire calls that day, of which 137 were attributed to firecrackers. Two women were injured.

The ban took effect on 9 November. It covers public and private places inside BMC limits, including homes, terraces, balconies, courtyards, parking areas and slums. The Hindu said the Urban Disaster Management Authority of Greater Mumbai issued it under the Disaster Management Act, 2005, because aerial crackers can land on roofs, balconies and combustible material in a city of high-rises, narrow lanes and slums. Violations draw action under that Act and other laws.

The Indian Express said the order is unusual. Section 51 of the Act, as the paper set it out, provides imprisonment of up to six months, or a fine, or both, for refusing to comply with a direction without reasonable cause, and up to one year if that refusal causes loss of life or an imminent danger of it.

Hours earlier, Thane district banned rockets and similar aerial crackers until 8 January 2027, The Hindu and The Indian Express reported. The Indian Express said that order covers the six municipal corporations of Thane, Navi Mumbai, Kalyan-Dombivli, Bhiwandi, Mira-Bhayandar and Ulhasnagar. Together with Mumbai, the paper said, seven of the nine municipal corporations in the Mumbai Metropolitan Region are covered. Vasai-Virar and Panvel are the two left out.

What to watch next

1. Tuesday's wind. The Times of India, after Beig's account of the easterlies, said those winds were expected to last until Tuesday morning and then possibly turn back toward the north-west, which would reopen the path for stubble smoke. Mishra's IMD comment was that Delhi-NCR should not expect a cleaning rain or a useful change in wind over the next 48 hours.
2. The 11 to 13 November rain window in the IMD bulletin of 8 November, which included Delhi. A later comment from an IMD scientist narrowed what the city should expect before that window.
3. Whether Stage 4 is invoked. Stage 3 is the step The Hindu still described as in force on Monday. The city averages through Monday afternoon, down to 344 by 4 pm in The Times of India's account, were in very poor. The severe band starts at 401. Several stations were already there on Monday morning.
4. The missing hours. CPCB had not, in the stories above, explained why PM2.5 stopped while gases continued, or whether the later peaks at Mundka and elsewhere are the recovered record of the blank hours.
5. The political claims. Bharadwaj's fraud allegation and Mishra's conspiracy allegation are both unresolved. The satellite count of 1,152 fires is the number in the news reports. The 1,200 is the minister's.

The Diwali-day index was 376. The night's dust was high enough, on the public feed, to hit 1,000 and then disappear from 43 stations. The argument now is whether that average of 370 the next morning describes the air people breathed, or only the hours the monitors agreed to print.

---

This explainer draws on The Times of India (Kushagra Dixit and Priyangi Agarwal, 1:45 am IST on 10 November 2026) for the Diwali-day AQI of 376, the comparison with 2021 to 2025, the Diwali-eve AQI of 369, Monday's 370 at 11 am and 344 by 4 pm, the post-Diwali series, the 47 stations in the Diwali-day average, the 35 stations with hourly data initially missing, the count of 60 stations, the IITM midnight PM2.5 average of 671.1, the Jahangirpuri, Mundka and JNU peaks, the 24-hour PM2.5 limit of 60, the Diwali PM2.5 average of 257 and the yearly comparison, the 41.1 percent firecracker share, the 1,152 farm fires, the Safdarjung visibility of 50 metres, and the comments from Gufran Beig, Anumita Roychowdhury and Sunil Dahiya. Hindustan Times (Binayak Dasgupta, updated 3:19 pm IST on 9 November 2026) is the source for the 43 of 59 PM2.5 gaps, the 14 readings of 900 or more, the 39-station comparison, the 1,000 ceiling, the 2,224 reading at NSIT Dwarka last Diwali, the 437 and 471 estimates, CPCB's 370 and 381, and the station traces at North Campus, CRRI Mathura Road and ITO. NDTV is the source for the CREA timeline (7, 23, 29 and 36 stations, and 227 missing hours), Manoj Kumar's comments, the RK Puram reading of 924, Anil Gupta's explanation, Dipankar Saha's disagreement, the CPCB 2019 range of 1,000, and the AQI cap at 500. The Hindu (updated 4:57 pm IST on 9 November 2026) is the source for the CPCB live-dashboard reading of 404 at 4 am, the 24-hour average of 370 at 9 am, the station readings from Burari Crossing at 441 through Rohini at 430, GRAP Stage 3 remaining in force, the past-10 pm cracker deadline, and the Mumbai ban. NDTV India is the source for the longer CPCB station list, including Mundka at 417 and Anand Vihar at 335, and for IMD scientist Krishna Mishra's comment that relief was not expected for 48 hours. The Indian Express is the source for the Monday average of 370 at 9:05 am, eight stations above 400, the worst Diwali-day air since 2021, Kapil Mishra's allegation and the 1,200 figure, Baltej Pannu's reply, the farm-fire totals of 3,307 and 4,459, and the Mumbai and Thane ban details including Section 51. A further Hindustan Times report (9 November 2026) is the source for Saurabh Bharadwaj's allegations, the 6:15 am average of 366, the Diwali average of 376 against 369, and the HT Pollution Dashboard station list. The IMD all-India bulletin issued at 8:21 am IST on 8 November 2026 is the source for the western disturbance from 10 November and the rain window for Delhi on 11 to 13 November. AQI readings move through the day. Check the CPCB bulletin for the latest figure.