We all know how bad Delhi’s air pollution is, but where does this pollution actually come from?

Kanpur city is separated from rural fields only by the Ganga. Cars can cross along the top of the Ganga Barrage, a dam and tourist attraction with ice cream. Air can cross anywhere. There is no kilometres-high boundary wall to divide the atmosphere at city borders. “Air flows like a fluid,” as the slogan reminds engineering students. But all that flow has to come from somewhere. I asked the engineering students’ teacher where.

In the Cafe Coffee Day at the centre of the IIT-Kanpur campus, the air conditioning made such a contrast to the warm, humid outside air that condensed water ran down the windows. The staff finally solved the problem by opening wide the two glass front doors, in a seeming attempt to air condition Kanpur itself. It was summer break. Around 10 or 15 students were working silently on their laptops across the large rooms of the coffee shop. Only in the section labelled “SILENT ZONE” were any of the students talking with one another.

I was there to talk with Professor Sachchida Nand Tripathi, or Sachi as he cheerfully introduced himself. We met in his office on a weekend: it seemed obvious to him that Saturday would be a work day. Perhaps he suggested the coffee shop to get away from the ringing of his office phone with requests for presentations. Sachi was once the PhD teacher of Sagnik – my collaborator who worked with Josh, Sangita, and me in chapter 2 to learn the effect of air pollution on child height. When I visited, Sachi was part of a team that was in the final process of submitting a new research paper to a journal. His paper is one of several studies seeking to sort out where north India’s air pollution comes from, based on what chemicals and particles it contains.

Many questions remain unanswered about the details of the source of north India’s air pollution, even to the experts.

In this, “source apportionment” (as the experts call it) is just like every other aspect of the air pollution problem: successive central and state governments have not collected enough data or tracked the challenge. A few months before we talked, Sachi had told a reporter from Reuters, “The state government does not have the mechanism to understand the sources of air pollution, how will they tackle it?” As Sachi explained to the reporter, the decision not to adequately investigate particulate matter cannot be excused by claiming that pollution is a surprise: “The state needs to act. This was very much coming.”

Because high quality data is so scarce, the science is at a stage where some differences among different studies’ conclusions have not been resolved. Sachi is part of one team working with the central government’s pollution control board. Others use different strategies and find conclusions that are a little different.

Each strategy for source apportionment is an approach to combining incomplete data with theory and models to fill in the gaps. Some studies are bottom-up: they start with estimates of emissions, and then use a computer model to fill in how different sources of pollution would contribute to the total concentration of particles.

Other studies are top-down. They look for specific chemical compounds in the final mix of particles that waft in the air. Top-down studies solve what amounts to a complex algebra problem to make the possible sources add up to the counts of fingerprint chemicals that are found in the mix. The results do not perfectly agree on the details, but both types of study find that north India’s pollution comes from many sources.

What is a concerned citizen to think about all this? What does it mean when different studies give different answers about the sources of the pollution?

Despite the confusion, three conclusions are clear. One is that much better monitoring is necessary. Pollution levels vary radically across places and times of year, so it would be no surprise if pollution’s sources did, too. It is not enough merely to track the fluctuations in the numerical levels of pollution. We also need details of the chemical make-up of the pollution that would let researchers learn more about its origins.

Sachi gave me an example: some properties of a bit of air pollution precisely reveal the process that released it. Josh Apte demonstrated a machine like this in his backyard garage. It sorts particles by giving a different electrical charge to particles depending on their mass. Grinning, Josh handed me his baby, walked around the corner, and lit a piece of paper on fire, so I could see his new apparatus blink with excitement as it cycled through its scan of particles.

Building such a system in India will require a major, sustained effort, but not an unprecedented one: other countries do it routinely. As Sachi lamented, the government bureaucrats appear to believe that they can commission a single study of source apportionment and resolve the matter, once and for all. “There was one apportionment study done in 2014-2015, and that’s enough,” officials seem to think.

He compared India with Paris and other foreign cities, where apportionment studies are continuous – “round the clock!” – and monitoring is constant, even though air pollution levels are much lower. “That is what they did in China,” where air pollution levels used to be almost as high as India’s, but have been recently brought under control.

Researchers understand this need. Almost everyone I talked with writing this book told me about some new study to figure out the sources of Delhi’s pollution.

The stories often turned out to be about different members of the same few research teams. The University of Birmingham recently paid to have an article in the Chronicle of Higher Education advertising that their top scientists are studying Indian source apportionment: “the university is working with partners to identify the different causes and effects of air pollution in China and India, which together account for half of all deaths linked to air pollution.”

“Great!” I thought, “and how timely that I am writing chapter three.” The word “India” is a link in that article. When I click it, the University of Birmingham website tells me that they are collaborating with engineers at IIT-Delhi, and partnering with my friend at the University of Texas. Of course, these are major research projects that need large teams. It is only good that big open questions are being tackled by international collaborations. My point is that we need many more of them, so I had hoped I would learn of something new. This was Sachi’s point, too. If in Paris – where PM2.5 is reported to be “good” as I type – the government monitors pollution sources round the clock at several sites, then India has a long way to go to build an adequate data collection network.

The second clear implication from the apportionment studies is that the air pollution comes from many sources. That means that no one quick fix will solve it all. Cars, trucks, industry, coal plants, trash burning, crop residue burning, construction, even traditional stoves and household fires to keep warm: air pollution is caused by a complex list of sources. This complexity makes it easy to appear to be taking aggressive action, without actually solving the problem. A government can highlight one small part of the problem while ignoring the rest. Even a government that wants to would be challenged to track down and plug each source of emissions – in the real world where bureaucratic resources and capacity are limited.

The third upshot of the source studies is that ‘Delhi’s’ air pollution is not only Delhi’s. So, policy cannot succeed if it only focuses there. Coal burnt in Kanpur, crops burnt in Haryana, cars idling on the Gurgaon highway – each of these contributes. Sachi implicated “all the northern Ganga belt” as a culprit: Haryana, Punjab, Delhi, UP, he listed. Getting to each source of the air pollution requires leaving Delhi.

Excerpted with permission from Air: Why India Must Solve Pollution and Climate Change Together, Dean Spears, Harper Collins.