Delhi overtakes Lahore as the world's most polluted city with an AQI of about 334

The reading and the official bulletin are also measured differently. A private tracker usually reports an hourly or live reading from its sensors, while the Central Pollution Control Board reports a 24-hour average at a fixed time each day, so the two can differ even when both are accurate.

India's own index has different labels. The Central Pollution Control Board calls 201 to 300 poor, 301 to 400 very poor and 401 to 500 severe, which is why the same number can be described as hazardous in one scale and very poor in the other.

Reports on the day pointed to three causes: winter weather that traps emissions from vehicles close to the ground, dust from construction and smoke from burning of crop residue in neighbouring states. The Diwali fireworks, which fell earlier this month, added to the load. On the night of Diwali, the board's 24-hour index for Delhi was 376, in the very poor band, with Ballabgarh at 398, and a later check found that up to 36 of the 59 monitoring stations in the city were offline at one point, so the picture of the worst days is incomplete. The average on the Monday after was 344.

The 13 November reading is a single day's number. It is lower than the Diwali night figures, so what has changed is the ranking, and not necessarily a sharp rise in the air's toxicity compared with the days around Diwali.

What does it mean for health?

Fine particles, known as PM2.5 because they are smaller than 2.5 micrometres, can travel deep into the lungs and cause effects beyond the respiratory system. A study published ahead of World Diabetes Day on 14 November, an umbrella review of 13 systematic reviews covering over 200 million cumulative participant observations, found that long-term exposure to PM2.5 is associated with a 10% to 16% higher risk of type 2 diabetes for every 10 micrograms per cubic metre increase. One of its authors, Dr Anoop Misra of Fortis C-DOC Hospital in New Delhi, is among the researchers. The association appeared in Asian and Western populations and followed a dose-response pattern.

The researchers are careful about what it shows. They rate the certainty of the evidence on new cases as low, and the evidence on prevalence, diabetes-related deaths and blood glucose as very low. The findings show an association, not a direct cause, and they do not prove that a specific cut in PM2.5 would cut diabetes cases in proportion. The authors recommend treating air pollution as part of the prevention of heart, kidney and metabolic disease, expanding air-quality monitoring and running longer studies. Possible pathways include inflammation, oxidative stress and impaired insulin signalling, though these are not fully understood.

What happens next?

The official response is staged. The Commission for Air Quality Management orders restrictions on construction, vehicles and industry under the Graded Response Action Plan, which is tightened as the index rises. The accounts seen do not say which stage is in force on 13 November, and readers should check the commission's orders. The weather is the swing factor: winds and rain can clear the air within a day or two, while still, cold nights hold it down.

For people in the city, the practical points are those on the national health advisories: limit time outdoors when the index is in the very poor or severe bands, keep windows closed during the worst hours and use masks rated for fine particles where exposure cannot be avoided. The board's daily bulletin and the station readings, not a ranking, are the numbers to follow.