Published: 9 August 2026 | Author: Team JanVayu | Reading time: 6 min
Everything else on this site is a number. Thirty-one photographs are not, and we should say why they are here.
You cannot photograph PM2.5. A PM2.5 particle is two and a half micrometres across or smaller, about a thirtieth the width of a human hair. Nothing you can see in a photograph is the thing that is killing people. A camera catches the visible fraction: the smoke at the moment it leaves the source, and the people standing in it.
That is enough to be worth doing, provided we are clear about what it shows and what it does not.
The gallery is an argument
The 31 photographs are ordered deliberately. Read start to finish, they go like this.
Update, 2 October 2026: when we wrote this in August the gallery held 32. One was a duplicate of another and has been removed, so it now holds 31. Four photographs taken outside India (Mexico, Essex 1986, Turkey, Taiwan) are now captioned as such, and the licence count is 29 Wikimedia Commons images plus two New York Times photographs used with permission.
First, the effect. Seven frames of what dirty air does to a view: a skyline dissolving at sunset, high-rises fading into winter smog, auto-rickshaws in a street where the far end has gone. Everyone has seen this. It is also the least informative part, and we put it first because it is what brings people in.
Then, the causes. A coal plant's cooling towers. A brick-kiln chimney on a city's edge. A diesel locomotive trailing black smoke. Paddy stubble alight after harvest, six different ways: flames sweeping a field, smoke drifting over a smoulder, farmers standing beside it, a worker inside it at harvest time. Waste burning at a roadside, and a wall of black smoke over water.
This part matters most, because the effect photographs are anonymous and the cause photographs are not. A grey skyline could be anywhere. A brick kiln is somebody's kiln, in somebody's district, running in a season you can name.
Then, the scale. Two NASA frames from orbit: stubble smoke streaming across the plains, and haze and farm fires blanketing north India. From the ground, smog looks like weather. From 700 kilometres up it shows what it is, a continental event with a shape and a direction.
Then, particular places. Gurgaon by day, with construction dust rising under the smog, and Gurgaon at night, thick enough to halo the streetlights. Mumbai's skyline over Chowpatty. Azadpur, where the Delhi Metro runs above CNG buses, two measures aimed at cutting emissions. We included them because a gallery of only failure would be a lie.
And last, Okhla and Ghazipur. Waste-pickers working the Ghazipur landfill in winter haze, beside one of Delhi's waste-to-energy plants. Residents of Sukhdev Vihar, a colony that shares a wall with the Okhla plant, demonstrating outside it. Cars coated in ash after a night of trucks delivering to the dumping site next door.
The sequence builds towards these last frames. This pollution is not ambient. It has an address.
What photographs get wrong
Two things, and both matter enough to state plainly.
Haze is not concentration. How thick the air looks depends on more than the mass of particles in it. We have not cited an optics source for the details, so we make only the narrow claim that looks are an unreliable guide to concentration. Winter smog scatters light beautifully and photographs like the end of the world. A summer dust event that reads similarly on a monitor can look like an ordinary hot afternoon. The most dramatic photograph is not the worst air, and anyone who says otherwise is reading a picture as if it were an instrument.
This is the same trap as leading with AQI instead of PM2.5: a proxy that feels informative stands in for a measurement that is.
The worst air is often invisible. The pollution that shortens lives most reliably is the chronic background, the annual mean that never makes a photograph because nothing about it looks like an event. A village at 60 µg/m³ every day of the year is a worse place to live than a city that hits 400 for a fortnight and gets photographed. Nobody photographs 60. It just looks like air.
So the map is the instrument, and the gallery is the reason to look at the map.
Where they come from
29 of the 31 are openly licensed. They are Creative Commons, CC0 or public domain images from Wikimedia Commons (the two NASA frames come by the same route), each carrying its photographer's name, its licence and a link to the original. Tap any image and the credit is there. We did not crop the credit off, and we used nothing whose licence we could not name.
Two are used with the photographer's permission: Bryan Denton's frames from The New York Times's Okhla investigation, of the residents and the ash. They are credited as such, without a licence URL, because there is no public licence to link to. Permission is not a licence, and we will not imply otherwise.
We do not generate images. Every frame here is a photograph of a real place on a real day. On a site whose claim is that its figures can be checked, an illustration that looks like documentary evidence would undo more than it decorates.
What is missing
Indoor air. Household air pollution from cooking fuel is a large part of India's burden (State of Global Air 2025 attributes about 0.5 million of roughly 2 million deaths to it), and the gallery has almost nothing of it. Outdoor smog is what gets photographed and uploaded under an open licence, and a woman's kitchen is not. This gap skews the sequence towards the street.
Most of the country is missing too. The frames cluster in Delhi–NCR because that is where the photographers are. The map covers 983,149 areas; the gallery covers perhaps a dozen.
If you have photographs
If you have documentary photographs of air pollution in India, especially of indoor exposure or from outside the north, and you would license them openly or give permission for use here, write to contribute@janvayu.in.
The credit stays on the picture. That is the whole arrangement.