Bengaluru: New Delhi registered very high levels of pollution during Diwali. Among the several questions that arose during this time were concerns around missing data, as news reports flagged that air quality data from several monitoring stations across the city disappeared during crucial hours at night when firecracker bursting was at its peak.However, the Central Pollution Control Board (CPCB), in a statement on Tuesday (November 10) night, denied this. It said that the data was not ‘missing’, and that it had been omitted from analysis as part of methodology; and that capping the air quality index (AQI, a single measure of several concentrations of pollutants in the air) at 500 was a “design feature” in the way India’s National Air Quality Index (NAQI) is calculated.Analyses by experts including researchers at the Centre for Research on Clean Air and Energy (CREA) suggest that data did indeed go missing. And worse, the missing data may have led to undercounting air pollution levels in the city, primarily that of fine particulate matter (or PM2.5, a major air pollutant that consists of fine particles below 2.5 micrometres in diameter).Flagging unusually high readings rather than excluding them altogether, greater transparency in how this is done, and expanding measurement capabilities are crucial to take concrete steps to tackle air pollution in the city, the scientists say.‘Data not missing’: CPCBThe night of November 8, Diwali, brought not just noise and light through firecrackers but also some astonishing numbers to New Delhi: 3,744 in Mundka at 2 am 3,023 at 4 am at Jawaharlal Nehru University. More than 2,000 at Mandir Marg and several other places.These are the PM2.5 concentrations (in µg/m³) that several air monitoring stations clocked as per data released by the Delhi Pollution Control Committee (DPCC) on Monday, per news reports.These numbers were meant to reflect online in real-time. They did not. They disappeared from open-access portals of the CPCB, a statutory organisation under the Ministry of Environment, Forest and Climate Change whose mandates include monitoring air quality.According to one estimate, more than 20 monitoring stations did not report data between midnight and 2 am during peak smog hours that night. Delhi has 59 air quality monitoring stations across the city. The DPCC released the missing numbers only a day after Diwali.But late on Tuesday night, the CPCB claimed that the data was not really ‘missing’. It said that PM2.5 observations exceeding 1,000 µg/m³ are treated as out-of-range and excluded from AQI calculations, as particulate matter analysers have a measurement range of 0-1,000 µg/m³:“During extreme episodes, which are occasional like Diwali firecracker bursting, this may appear as a data gap at the particular station concerned, and those hours do not contribute to the 24-hour average of relevant pollutant for concerned station.”So did the PM2.5 concentrations at peak firecracker bursting time on Diwali night not matter?The missing data matteredIt did, found Manoj Kumar, an analyst at CREA. He studied hourly PM2.5 data from Delhi’s 59 monitoring stations between 4 pm on November 8 and 4 pm on November 9, and identified 255 missing station-hour observations, accounting for 18% of the 1,416 expected observations.Considering it was Diwali night, what if all these observations may have clocked PM2.5 concentrations of at least 1,000 µg/m³, the upper measurement limit cited by CPCB? Kumar found that under such a scenario, the pooled average across the expected station-hour observations increased from 269 to 400 µg/m³, a 49% increase.Of course, there’s a caveat in this analysis. It’s based on assumptions, as all modeling studies are.“The 400 µg/m³ figure is therefore a sensitivity estimate, not an observed concentration or a revised official average. It illustrates the potential implications of how missing observations are handled,” Kumar wrote in his analysis published Wednesday.‘A design feature’ says CPCBIn its statement on Tuesday, the CPCB also provided justifications for why India’s AQI is capped at 500. Per the CPCB, an AQI of 500 or above (categorised as “Severe”, with listed health implications including respiratory effects on healthy people and serious health impacts on people with lung diseases) “indicates almost the same health risk”.“Beyond this level, the AQI remains in the ‘Severe’ category and extrapolation of the index would not alter the assigned health impact category. The capping at 500 is thus a design feature of the index, as has been evolved in 2015 methodology of NAQI…”, the CPCB’s statement said.But this “design feature” does not distinguish between substantially different pollution concentrations, Kumar told The Wire.“For example, PM2.5 levels of 260, 500 and 800 µg/m3 all fall under the Severe category. Adding severity categories or extending the AQI beyond 500 would better communicate the magnitude of extreme pollution and help assess the impact of interventions. However, this must be accompanied by monitoring instruments that can reliably measure PM2.5 beyond 1,000 µg/m³ and revised data-processing protocols to ensure extreme readings are not excluded,” he said.The CPCB statement also suggested that short-lived episodic high-pollution events do not matter as much as long-term exposure does:“The occasional short-lived spikes of pollutant concentrations in ambient air, such as those during the bursting of firecrackers, are episodic and short lived in nature. The NAQI recognises that the prolonged exposure, rather than a momentary reading is the more relevant indicator of ill effects on human health.”While long-term exposure causes most harm, it does not mean short spikes are harmless, said Sunil Dahiya, founder and lead analyst at EnviroCatalysts. “Deaths and hospital admissions rise within a day or two of a jump in PM2.5, and the risk of heart attacks rises within hours.”On Diwali night, Delhi’s official average stayed above 500 µg/m³ for about four hours, and the DPCC’s own monitors recorded more than 1,000 µg/m³ at 20 stations, he noted.“An index built on 24-hour averages and broad categories plays down hazardous spikes like these, and leaves people without the warning they need to protect themselves. These episodes are predictable and preventable, and they should be prevented. But cutting them is no substitute for cutting emissions year-round across the airshed.”No updates in methodology yetThe CPCB also said in its statement that it keeps the methodology to calculate AQI “under review in light of advancements in monitoring technology and international practice”, and that “internal deliberations and consultations with respect to processing of data beyond system design range are underway”.But given all the “technological advancements” that the CPCB itself is admitting has occurred over the years, has the methodology for AQI calculation ever been updated since it was launched in 2015?“As far as the public record shows, no,” Dahiya told The Wire.“The breakpoints are the ones set in 2014, and the standards behind them date from 2009. The original report itself said the CPCB should consider reviewing the breakpoints every three years, so three or four reviews have been missed. Meanwhile the WHO tightened its guidelines in 2021, the US revised its index in 2024 and the CPCB now admits the instruments have advanced. The review under way is internal and narrow, limited to readings beyond the instruments’ range. India needs an independent, time-bound review of the whole index and of the standards it rests on.”Kumar also told The Wire that he has not found evidence of a comprehensive revision to its underlying calculation methodology since 2015 when the Index was launched.Ways forwardBoth Kumar and Dahiya are clear: the government can take several steps to deal with these concerns surrounding air quality data.Gaps in data are a serious problem; something that should be addressed, Dahiya said.“Monitors [on Sunday] went silent for five to eleven hours, right through the peak, and by 1 am only 23 of Delhi’s 59 stations were reporting. The DPCC’s own portal shows that 20 of its stations stayed above 1,000 µg/m³ for a combined 126 hours, with readings up to 3,744. So the longest gaps hide the worst air. The official average is then built from the cleaner stations that are left, which shows lower levels and gives a false sense of improvement. Nor is this occasional: the same thing happened on Diwali 2025. Readings beyond the scale should be recorded and flagged, not deleted,” Dahiya recommended.Two other important measures are improving transparency by disclosing how missing and out-of-range observations are handled in 24-hour PM2.5 averages and AQI calculations, and expanding measurement capabilities by reviewing and upgrading monitoring instruments to better capture extreme pollution concentrations, Kumar wrote in his analysis.The “CPCB’s clarification also mentions that internal deliberations on processing such readings are underway. Greater clarity on the proposed approach and steps to address the measurement limitation would strengthen confidence in air quality reporting,” his analysis recommended.