Delhi has had 51 'severe' air days this year. It had 40 in all of last year.

India's Graded Response Action Plan is a four-stage escalator. Stage I is triggered when the Air Quality Index runs 201–300, described as "poor". Stage II covers 301–400, "very poor". Stage III covers 401–450, "severe". Stage IV, "severe+", begins above 450 [s1].

On the evening of 13 December, Delhi and the National Capital Region went to Stage IV [s1]. The Press Information Bureau reported the city's daily average index at 431 at 4pm, rising to 441 by 6pm the following day [s1].

Stage IV brings the heaviest restrictions the plan contains: a ban on BS-IV trucks entering Delhi except for essential goods and services, a halt to construction and demolition, hybrid classes for Classes VI to IX and XI, and public and private offices running at 50% capacity with the remainder working remotely [s1].

The number that matters is not today's

By mid-December, Delhi had recorded 51 days in 2025 with an index above 401 — the "severe" threshold. In all of 2024 it recorded 40 [s2]. That is a 28% increase with sixteen days of the year still to run [s2]. Counting "very poor" and "severe" days together, the city logged 71 in 2024 and was on course for more than 80 in 2025 [s2].

Individual monitoring stations ran well above the citywide average. On 15 December, Jahangirpuri read 498, Akshardham 493, Barakhamba Road 474 and Barapulla Flyover 433; the previous day's citywide average was 461 [s2].

An emergency that recurs on this schedule is not an emergency. It is a climate.

Why this December

Two explanations are circulating, and they are not in conflict.

The meteorological one: wind speeds in Delhi averaged 4.2 km/h in December 2025 against about 6 km/h in December 2024, roughly a 30% reduction [s2]. Slow winter air over a basin traps whatever is emitted into it. Cold, still, stagnant conditions are the amplifier, not the source.

The source one is contested in its details. ETV Bharat's reporting attributes 37% of NCR pollution to sources inside Delhi proper and 63% to the surrounding region [s2]. A separate analysis puts vehicles and factories at about a third of the load, crop stubble and wood burning at about a fifth, and vehicles alone at 17%, with coal, household fuels and dust making up the remainder [s3].

Notably, the stubble-burning contribution to PM2.5 in 2025 has been described as negligible on Central Pollution Control Board data [s3] — which, if it holds, removes the seasonal scapegoat and points the analysis back at year-round local combustion.

There is also a measurement question. India's index is capped in a way that international indices are not; while the official figure sat above 450 for days, some international indices suggested a value closer to 700 [s3].

What that exposure does

The clinical consequences of fine particulate exposure are among the best-established findings in environmental epidemiology, and they are mostly cardiovascular rather than respiratory.

The World Health Organization estimates ambient air pollution caused 4.2 million premature deaths worldwide in 2019. Of those, 68% were attributed to ischaemic heart disease and stroke, 14% to chronic obstructive pulmonary disease, 14% to acute lower respiratory infections and 4% to lung cancer [s4]. Eighty-nine per cent occurred in low- and middle-income countries, with the highest burden in the South-East Asia and Western Pacific regions [s4].

That distribution is why the death toll from an air pollution episode is never visible in real time. A stroke during a smog week is recorded as a stroke.

The historical comparison usually reached for is London's Great Smog of December 1952, which killed at least 4,000 people in barely five days [s3]. Beijing's PM2.5 concentrations frequently exceeded 500 µg/m³ during its 2010–2013 crisis [s3] — a period that ended not with an emergency plan but with sustained structural changes to energy and industry.

What Stage IV can and cannot do

The restrictions are real: hybrid schooling and halved office occupancy reduce the number of people breathing outdoor air during the worst hours, and the construction ban removes a dust source [s1].

But an escalator triggered by an index is by construction a response to exposure that has already occurred. It cannot reduce cumulative annual dose, which is what the epidemiology is denominated in. A city that spends 51 days a year above 401 [s2] is not being protected by rules that switch on at 450.

The measure of whether anything has changed will not be how quickly Stage IV is imposed next winter. It will be the severe-day count.

What to watch

Whether the 2025 severe-day total closes above or below 55, and whether the CPCB source apportionment holds up under independent analysis. If local combustion really is the dominant year-round contributor [s3], the policy conversation has to move off stubble and onto transport, industry and household fuel — a much harder set of arguments.