Air Quality Improvement as a Co-Benefit of Clean Energy Transition in Megacities

Executive Summary

The energy transition is usually justified by what it prevents decades from now. In the world's largest cities, its first and most measurable return is the air people breathe this year.

The world has thirty-three megacities, up from eight in 1975, and nineteen of them are in Asia. They are where combustion is densest — traffic, power stations, industrial boilers, diesel generators and household stoves packed into a few hundred square kilometres — and where the most people breathe the result. The five most polluted major cities in 2024, home to more than 96 million people, recorded annual fine-particulate concentrations between nine and twenty-two times the World Health Organization's guideline.

The argument of this report is simple and, on the evidence, well supported. Most of what poisons megacity air comes out of the same exhaust pipes and chimneys as the carbon dioxide that warms the planet. Replacing combustion with clean electricity — rather than filtering it — removes both at once. The climate benefit is shared by everyone, everywhere, over centuries. The health benefit is collected by the city that acts, starting immediately. That asymmetry makes decarbonization one of the few climate policies whose most visible payoff lands inside a single mayor's term.

The report proceeds in seven parts. Section 1 sets out the scale and concentration of the urban burden. Section 2 shows why urban air pollution and greenhouse-gas emissions are, for the most part, the same problem. Section 3 reviews the evidence on how many lives decarbonization saves and what they are worth. Section 4 examines four megacities — Beijing, London, Shenzhen and Delhi — for what worked and what did not. Section 5 draws out design principles, Section 6 makes recommendations by actor, and Section 7 concludes.

1. The Urban Air Pollution Burden

Air pollution is the world's leading environmental risk to health, and it is concentrated where population and combustion are concentrated: in large cities.

1.1 A global burden, measured in millions

The Health Effects Institute's State of Global Air 2025 finds that air pollution contributed to 7.9 million deaths in 2023, making it the leading environmental risk factor for death worldwide. The character of that burden is often misunderstood. It is not principally a matter of acute respiratory attacks: 6.8 million of the deaths, 86 per cent, were from noncommunicable diseases — heart disease, stroke, chronic obstructive pulmonary disease, lung cancer, diabetes — and more than 625,000 were related to dementia. More than 675,000 were children under five. And 36 per cent of the world's population lives above 35 µg/m³ of PM2.5, the least stringent of the WHO's interim targets, seven times its 5 µg/m³ guideline.

A separate estimate isolates the part that energy policy can reach. Lelieveld and colleagues, writing in The BMJ, estimate all-cause excess deaths from fine particulates and ozone at 8.34 million a year, and attribute 5.13 million (95 per cent interval 3.63 to 6.32 million) specifically to ambient pollution from fossil-fuel use. Those deaths, in the authors' framing, could potentially be avoided by phasing out fossil fuels. Just over half of the burden (52 per cent) is cardiometabolic, led by ischaemic heart disease at 30 per cent; stroke and chronic obstructive pulmonary disease account for 16 per cent each.

1.2 Concentrated in cities

A study of 13,160 urban centres by Southerland and colleagues in The Lancet Planetary Health found that in 2019 approximately 86 per cent of urban inhabitants — 2.5 billion people — lived above the WHO's former 10 µg/m³ guideline, resulting in an excess of 1.8 million deaths (95 per cent interval 1.34 to 2.3 million). Two further findings in that study matter for policy. Attributable deaths rose in every region except Europe and the Americas between 2000 and 2019. And in some cities mortality rose even as concentrations fell, because ageing populations and rising rates of noncommunicable disease made each microgram more lethal. A megacity that holds its air quality steady while its population ages is losing ground.

That population is growing and concentrating. The UN's World Urbanization Prospects 2025 counts 33 megacities of more than ten million people, up from eight in 1975, with nineteen in Asia. Jakarta, with nearly 42 million residents, is now the largest, followed by Dhaka with almost 40 million and Tokyo with 33 million.

1.3 Deaths even at "moderate" levels

It would be a mistake to read the burden as confined to the handful of cities at the top of Figure 1. A causal multi-city study of ten Indian cities in The Lancet Planetary Health found that between 2008 and 2019, 7.2 per cent of all deaths — about 33,000 a year — could be linked to short-term PM2.5 exposure above the WHO's 24-hour guideline of 15 µg/m³. Each 10 µg/m³ increase in short-term exposure was associated with a 1.42 per cent increase in daily deaths. Delhi had the largest attributable fraction, but Mumbai, with around a third of Delhi's annual PM2.5, still recorded over 5,000 attributable deaths a year. There is no evident threshold below which cleaner air stops paying.

Nor is PM2.5 the only urban pollutant that matters. Nitrogen dioxide, emitted overwhelmingly by road traffic and combustion, was associated with an estimated 4 million new cases of paediatric asthma each year between 2010 and 2015 across 194 countries and 125 major cities, according to Achakulwisut and colleagues. Because NO₂ falls off sharply with distance from the road, its burden is almost entirely urban, and its source is almost entirely combustion.

2. One Source, Two Problems

Urban air pollution and greenhouse-gas emissions are largely the same activity measured in two different units. That is what makes the co-benefit so large.

2.1 Most of the avoidable burden is fossil fuel

The most direct evidence comes from the same BMJ study. Its estimate of 5.13 million fossil-fuel-attributable deaths corresponds to 82 per cent of the maximum number of air pollution deaths that could be averted by controlling all anthropogenic emissions. Put the other way: of everything human beings could do about the air they have polluted, more than four-fifths of the mortality benefit lies in ending the combustion of coal, oil and gas. The remainder comes from sources such as agriculture, open burning and some industrial processes; natural sources such as desert dust sit outside human control altogether.

2.2 What megacity air is made of

Source apportionment — chemically fingerprinting the particles on a filter and tracing them back to their origins — tells the same story from the ground up. A systematic review of such studies compiled for the World Health Organization by Karagulian and colleagues estimated that, globally, 25 per cent of urban PM2.5 comes from traffic, 15 per cent from industry including power generation, 20 per cent from domestic fuel burning, 22 per cent from unspecified sources of human origin, and 18 per cent from natural dust and sea salt. The "unspecified" category consists mostly of secondary particles — sulphates, nitrates and organic aerosol formed in the atmosphere from sulphur dioxide, nitrogen oxides and volatile compounds that are themselves largely products of combustion.

The regional variation in Figure 3 is itself a guide to where the transition pays fastest. Traffic is the largest identified source in India (37 per cent), South-Eastern Asia (36 per cent) and Southern Asia (34 per cent) — regions containing many of the world's fastest-growing megacities. Domestic fuel burning dominates in Africa (34 per cent). Industry, including power generation, contributes 27 per cent in Southern Asia and the Middle East. In each case, the leading human source is a form of combustion that clean electricity can replace: vehicles with electric drives, household stoves with electric or gas cooking, boilers and generating plants with renewable power.

2.3 The deadliest and the safest electricity

The same arithmetic holds at the level of the power plant. Our World in Data's compilation of deaths per terawatt-hour of electricity, combining accidents and air pollution, puts coal at 24.6 and oil at 18.4, against 0.04 for wind, 0.03 for nuclear and 0.02 for solar. Coal is roughly 1,230 times deadlier than solar per unit of electricity. One terawatt-hour is about the annual electricity consumption of 150,000 people in the European Union — a small city.

2.4 Where the transition cuts the pathway

Figure 5 sets out the mechanism. Combustion in the city emits two streams. Carbon dioxide rises into the global atmosphere, where it accumulates for centuries and harms everyone slowly. The co-emitted pollutants — primary particles, nitrogen oxides, sulphur dioxide and ozone precursors — stay near the ground, within the city and its region, and are breathed within hours or days. A measure that replaces the combustion cuts both streams at their common source. A measure that only filters exhaust cuts the second stream and leaves the first untouched.

Table 1 — Transition measures and the pollution they remove

Compiled by the author from the source categories in Karagulian et al. (2015) and the pollutant pathways in Figure 5.

3. How Many Lives Decarbonization Saves

Across different models, scenarios and decades, the evidence converges: the air-quality benefit of cutting carbon is measured in millions of lives, and its value rivals or exceeds the cost of the cuts.

3.1 Global mitigation scenarios

The foundational estimate is West and colleagues in Nature Climate Change. Simulating global greenhouse-gas mitigation against a reference scenario, they find it avoids 0.5 ± 0.2 million premature deaths in 2030, 1.3 ± 0.5 million in 2050, and 2.2 ± 0.8 million in 2100. The annual toll avoided grows as the cleaner capital stock accumulates and as populations age and urbanise.

The same study values that benefit. Global average marginal co-benefits of avoided mortality are US$50–380 per tonne of CO₂, which exceed marginal abatement costs in 2030 and 2050. In East Asia, the region containing the largest cluster of megacities, co-benefits in 2030 are ten to seventy times the marginal cost of abatement. Read plainly, that means a large share of near-term emission cuts in Asian economies would be justified on public-health grounds even if climate change did not exist.

3.2 The value of going faster

Speed matters because the health benefit arrives with the emission cut, not with the temperature outcome. Shindell and colleagues modelled the effect of cutting an additional 180 billion tonnes of carbon from 21st-century emissions — enough to shift a standard 2 °C pathway to 1.5 °C, or to reach 2 °C without relying on negative-emissions technology. The reduced air pollution yields 153 ± 43 million fewer premature deaths worldwide, with about 40 per cent occurring in the next forty years.

The study's geographic detail is what makes it a megacity finding. More than a million premature deaths would be prevented in each of many metropolitan areas in Asia and Africa, and more than 200,000 in individual urban areas on every inhabited continent except Australia. The benefits of accelerating climate policy, in other words, are not diffused thinly across the planet. They are concentrated in precisely the places with the most people and the dirtiest air.

3.3 The economic value

Two further estimates frame the scale in money. The World Bank's Global Health Cost of PM2.5 Air Pollution put the cost of health damage from fine-particulate exposure at US$8.1 trillion in 2019, equivalent to 6.1 per cent of global GDP, drawing on Global Burden of Disease estimates of 6.4 million deaths from PM2.5 that year. The International Monetary Fund's accounting of fossil-fuel subsidies reaches a related conclusion from the other side: most of the roughly US$7.4 trillion it counts is implicit — the unpriced cost of local air pollution and climate damage — borne by households' health rather than by treasuries.

The best retrospective evidence that clean-air regulation pays comes from the United States. The Environmental Protection Agency's second prospective study of the Clean Air Act Amendments estimated that their benefits exceed costs by a factor of more than thirty to one, and that in 2020 alone they would prevent over 230,000 early deaths. About 85 per cent of the monetised benefit came from reduced mortality associated with lower particulate matter.

4. Megacity Case Studies

Four cities show what cleaner air looks like in practice — how quickly it can come, which instruments deliver it, and what happens when the job is left half done.

4.1 Beijing: the fastest large-city clean-up on record

In 2013, a winter of extreme smog in northern China prompted the national Air Pollution Prevention and Control Action Plan and Beijing's own five-year clean air plan. The results are among the most dramatic ever recorded for a megacity. According to the State Council Information Office, Beijing's annual average PM2.5 fell from 89.5 µg/m³ in 2013 to 30.5 µg/m³ in 2024, a reduction of about two-thirds. Heavily polluted days fell from 58 to two, and days with good air quality rose by 114 to 290.

Nationally, Zhang and colleagues estimate in PNAS that the measures prevented about 0.41 million deaths in 2017 relative to 2013. Four measures did most of the work, together accounting for 0.37 million, or 92 per cent, of the avoided deaths.

Beijing's lesson cuts two ways, and both matter. It proves that megacity air can improve very fast when the political will is present. But most of China's early gains came from end-of-pipe controls and closing obsolete capacity rather than from decarbonization: tighter emission standards and upgraded boilers cut pollutants from plants that continued to burn coal. Residential fuel switching — replacing household coal with gas and electricity — was the measure closest to a transition, and even it contributed a substantial 2.2 µg/m³. As the cheap controls are exhausted, Beijing's remaining distance to the WHO guideline — still six times over — will have to come increasingly from replacing combustion itself.

4.2 London: a clean-air instrument, and its carbon limits

London's Ultra Low Emission Zone charges the most polluting vehicles to enter a zone that began in central London in 2019 and was extended to the whole of Greater London in 2023. The Greater London Authority's One Year Report estimates that, in 2024, roadside NO₂ concentrations were 27 per cent lower across London than they would have been without the scheme and its expansions, and 54 per cent lower in central London. The benefits were progressive: for some of the most deprived communities living near London's busiest roads, the report estimates an 80 per cent reduction in people exposed to illegal levels of pollution.

London also demonstrates the direction in which the co-benefit flows. Over 2019–2024, the scheme is estimated to have reduced London's nitrogen oxide emissions by 24 per cent and exhaust PM2.5 by 29 per cent — but carbon dioxide by only two per cent. That is not a failing of the ULEZ, which was designed as an air-quality measure and succeeded as one. It is evidence of the asymmetry at the heart of this report. Replacing an older, non-compliant car with a newer, compliant one cleans the air but still burns fuel. Replacing combustion with electricity does both.

4.3 Shenzhen: electrifying a megacity's buses

By the end of 2017, Shenzhen had put 16,359 electric buses on its streets, becoming the first city in the world to electrify its entire bus fleet (World Resources Institute). The fleet was larger than the electric bus fleets of New York, Los Angeles, New Jersey, Chicago and Toronto combined.

The logic generalises to every megacity. The World Resources Institute notes that diesel buses make up about 0.5 per cent of a city's vehicle fleet yet account for around 20 per cent of its transport emissions, because they run far longer and further than private cars — and they do so on the busiest streets, at the height of pedestrians' lungs. A publicly procured, centrally depot-charged bus fleet is the most concentrated, most controllable piece of urban combustion there is. Shenzhen's case shows it can be replaced in a few years with a single procurement decision, backed by public finance.

4.4 Delhi: a transition begun and outgrown

Delhi was a pioneer. In 1998 India's Supreme Court ordered the city's entire bus, taxi and auto-rickshaw fleet to switch to compressed natural gas, and diesel buses had almost entirely left the streets by late 2002. Analysis by Resources for the Future found that converting the buses reduced concentrations of PM10, carbon monoxide and sulphur dioxide; converting three-wheelers from petrol yielded less, possibly because of poor technology.

Yet in 2024 Delhi was the most polluted major city in the world, at 108.3 µg/m³. The CNG programme was a fuel switch within fossil fuels, confined to public transport, while the city's other sources — among them growing private traffic, industry, construction and pollution carried in from beyond its boundaries — were not tackled on the same scale. Delhi's experience is the clearest warning in the megacity record: a single-sector switch, however bold, is eroded by growth everywhere else. The gains are durable only when the transition covers every major combustion source and extends across the airshed, not merely the municipality.

Table 2 — Four megacities compared

Outcomes as reported in the sources cited in Sections 4.1–4.4. Measures and time periods differ between cities and are not directly comparable.

5. Designing the Transition for Health

Every tonne of carbon abated is equal for the climate. It is not equal for health. A transition planned with that in mind saves far more lives for the same climate result.

5.1 Sequence by exposure, not only by tonnes

Climate planning treats a tonne of CO₂ as a tonne wherever it is emitted, and for the climate that is correct. For health, what matters is how much of the co-emitted pollution reaches human lungs, and that depends on where the combustion occurs. Exhaust from a bus on a crowded street, a diesel generator in a market or a stove in a kitchen is breathed at far higher rates than emissions from a tall stack outside the city. A transition sequenced by tonnes alone may retire a remote plant first and leave the street-level sources for later. A transition sequenced by exposure starts where the combustion is closest to people — buses, taxis, two- and three-wheelers, household fuels, backup generators and in-city plants — and saves more lives per tonne abated in its early years.

5.2 Retire, do not merely retrofit

Pollution controls on a coal plant can cut its particulate and sulphur emissions substantially, and where a plant must keep running they are worth installing. But research in Nature Climate Change finds that the health co-benefits of climate mitigation depend on strategic power plant retirements and pollution controls — which plants close first, and whether controls are in place on those that remain. Retirement removes every pollutant and all the carbon; retrofits remove some pollutants and no carbon, and, as Section 3 noted, remove the cooling aerosol along with them. Where a city's air is dominated by nearby coal, the plants whose plumes reach the most people should be first in the retirement queue.

5.3 Count the health dividend in the appraisal

If avoided deaths are worth US$50–380 per tonne of CO₂, a project appraisal or national climate plan that omits them understates the return of clean energy by a margin that can exceed the carbon price itself. That omission is routine. Ministries of finance appraise energy investments on energy costs, environment ministries count carbon, and health ministries — which bear the cost of pollution-related disease — are rarely at the table. Bringing the health value into appraisal would move a large share of urban clean-energy investment from "costly climate measure" to "high-return public-health investment," which is what, on the evidence, it is.

5.4 Plan at the scale of the airshed

Air does not respect municipal boundaries. A megacity's pollution includes secondary particles formed from gases emitted by power plants, industry and agricultural burning tens or hundreds of kilometres away, and its own emissions travel to its neighbours. London's scheme lowered NO₂ within five kilometres beyond the Greater London boundary by an estimated 14 per cent; Delhi's clean-up was undone in part by sources it did not control. City transitions need regional counterparts, and national grids decarbonizing at the same time, or the local dividend leaks away.

5.5 Put the benefit where the burden is

Pollution exposure falls hardest on the poor — those living beside major roads, in informal settlements near industry, and in homes that still cook with solid fuels. London's ULEZ delivered its largest proportional benefits to its most deprived roadside communities. A transition that electrifies private cars for the affluent before it cleans public buses, or retires suburban plants before it replaces the stoves in low-income neighbourhoods, saves fewer lives and deepens inequity. The pollution–poverty–climate nexus discussed in an earlier report applies with particular force inside megacities.

6. Policy Recommendations

What each of the principal actors should do to collect the health dividend of the transition in the world's largest cities.

6.1 National governments

• Put health in the national climate plan. Quantify and publish the avoided deaths and health-cost savings of each nationally determined contribution and sectoral decarbonization pathway, using WHO and Global Burden of Disease methods, so that finance ministries see the full return.
• Prioritise retirement of coal and oil plants that affect megacities. Rank plants for early retirement by the population exposed to their plumes, not only by age or cost, and fit pollution controls on any plant that must continue to operate in the interim.
• Adopt the WHO air-quality guidelines as the long-run target, with the interim targets as binding milestones, and tie them to the pace of the energy transition rather than to end-of-pipe controls alone.
• Reform fossil-fuel subsidies that underprice the air pollution borne by city residents, and redirect a share of the savings to clean public transport and clean cooking for low-income households.

6.2 City and metropolitan authorities

• Electrify what the city controls first. Municipal buses, taxis, refuse trucks and service fleets are the most concentrated and most controllable urban combustion; Shenzhen shows they can be replaced within a few years.
• Pair low-emission zones with electrification. Zones like London's deliver fast air-quality gains; coupling their charges and exemptions with zero-emission requirements over time converts them from clean-air policies into decarbonization policies as well.
• Replace diesel backup generators with reliable grid supply, rooftop solar and battery storage, beginning with hospitals, schools and markets.
• Monitor densely and publish openly. Low-cost sensor networks and open data make the health dividend visible to residents — which is what sustains the political will to continue.

6.3 Development finance and investors

• Price health into project appraisal. Multilateral development banks should include the monetised value of avoided air-pollution mortality in the economic analysis of energy and transport projects, as a standard line rather than an optional annex.
• Finance urban electric fleets at scale, through aggregated procurement, leasing structures that separate battery and vehicle ownership, and concessional finance for cities in low- and middle-income countries where the cost of capital is highest.
• Blend climate and health finance. The same investment delivers both returns; financing structures that let climate funds and health funds co-invest would reduce the cost to each.

6.4 Health agencies and researchers

• Make health ministries a party to energy planning. They bear the cost of pollution-related disease and hold the data to count the benefit of reducing it.
• Track avoided deaths, not only concentrations. Because ageing populations make each microgram more lethal, a city can reduce concentrations and still see attributable deaths rise; mortality-based indicators capture what matters.
• Close the evidence gap in the megacities of South Asia and Africa, where exposure is highest and ground monitoring and local epidemiological studies remain thinnest.

7. Conclusions

The transition is the largest public-health opportunity available to the world's megacities, and its dividend can be collected now.

Air pollution contributed to 7.9 million deaths in 2023, and more than five million a year are attributable to the burning of fossil fuels. That toll falls most heavily on the world's thirty-three megacities and the fast-growing cities that will join them, where combustion is concentrated and billions of people breathe the result. It is not an unavoidable cost of urban life. It is, for the most part, the local face of the same activity that is warming the planet.

That shared source is what makes the co-benefit so powerful. Replacing combustion with clean electricity removes the carbon and the co-emitted pollutants together, and the evidence on the value of doing so is remarkably consistent: hundreds of thousands to millions of deaths avoided each year, 153 million over the century from a faster pathway, and health benefits worth US$50–380 per tonne of carbon — enough, in much of Asia, to justify near-term emission cuts on public-health grounds alone.

The megacity record shows how quickly air can improve and what determines whether the gains last. Beijing cut its fine-particulate pollution by two-thirds in eleven years. London cleaned its roadside air while barely touching its carbon, proving that clean-air policy is not climate policy. Shenzhen electrified an entire bus fleet within a few years. Delhi showed that a bold switch in one sector is overtaken by growth unless the transition reaches every major source across the airshed.

The practical conclusion is to plan the transition for health as well as for carbon: to count the lives it saves in every appraisal, to sequence it by the exposure it removes, to retire the plants whose plumes reach the most people, to electrify the fleets cities control, and to put the benefit where the burden is heaviest. The climate return of decarbonization will be measured over generations. In the megacities, its health return can be measured this year, in lives.

References

Every quantitative claim above is attributed inline. The principal sources are collected here.

• Health Effects InstituteState of Global Air 2025 — 7.9 million deaths in 2023, 86 per cent from noncommunicable diseases, 36 per cent of the world above 35 µg/m³. See also the summary of findings.
• Lelieveld et al.Air pollution deaths attributable to fossil fuels: observational and modelling study, The BMJ 383 (2023) — 8.34 million all-cause excess deaths; 5.13 million attributable to fossil fuels, 82 per cent of the anthropogenic maximum. See also the Max Planck Institute for Chemistry summary.
• Southerland et al.Global urban temporal trends in fine particulate matter (PM2·5) and attributable health burdens, The Lancet Planetary Health (2022) — 13,160 urban centres; 86 per cent of urban inhabitants above 10 µg/m³; 1.8 million deaths in 2019.
• The Lancet Planetary HealthAmbient air pollution and daily mortality in ten cities of India: a causal modelling study (2024) — 7.2 per cent of deaths, about 33,000 a year.
• Achakulwisut et al.Global, national, and urban burdens of paediatric asthma incidence attributable to ambient NO₂ pollution, The Lancet Planetary Health (2019) — about 4 million new paediatric asthma cases a year.
• United NationsWorld Urbanization Prospects 2025 — 33 megacities, 19 in Asia; Jakarta, Dhaka and Tokyo the largest.
• IQAir5 Most Polluted Major Cities in the World in 2024 — Delhi, Lahore, Dhaka, Kinshasa and Karachi. See also the 2024 World Air Quality Report.
• Karagulian et al.Contributions to cities' ambient particulate matter (PM): a systematic review of local source contributions at global level, Atmospheric Environment 120 (2015) — urban PM2.5 source shares by region.
• Our World in DataWhat are the safest and cleanest sources of energy? — deaths per terawatt-hour of electricity by source.
• West et al.Co-benefits of mitigating global greenhouse gas emissions for future air quality and human health, Nature Climate Change 3 (2013) — 0.5, 1.3 and 2.2 million deaths avoided in 2030, 2050 and 2100; US$50–380 per tonne of CO₂.
• Shindell et al.Quantified, localized health benefits of accelerated carbon dioxide emissions reductions, Nature Climate Change 8 (2018) — 153 ± 43 million premature deaths avoided.
• Nature Climate ChangeHealth co-benefits of climate change mitigation depend on strategic power plant retirements and pollution controls (2021).
• World BankThe Global Health Cost of PM2.5 Air Pollution: A Case for Action Beyond 2021 (2022) — US$8.1 trillion, 6.1 per cent of global GDP in 2019.
• International Monetary FundFossil Fuel Subsidies — about US$7.4 trillion in explicit and implicit subsidy (roughly US$725 billion explicit, US$6.7 trillion implicit), most of it the unpriced cost of air pollution and climate damage.
• US Environmental Protection AgencyBenefits and Costs of the Clean Air Act 1990–2020: the Second Prospective Study — benefits exceeding costs more than thirty to one; over 230,000 early deaths prevented in 2020.
• Zhang et al.Drivers of improved PM2.5 air quality in China from 2013 to 2017, PNAS 116 (2019) — 0.41 million deaths avoided; contributions of individual measures.
• State Council Information Office of ChinaFrom smog to sunshine: Beijing's decade-long clean air campaign pays off (2025) — PM2.5 89.5 to 30.5 µg/m³.
• Greater London AuthorityLondon-wide ULEZ One Year Report (March 2025) — NO₂, NO_x, PM2.5 and CO₂ reductions against a no-ULEZ scenario.
• World Resources InstituteHow Did Shenzhen, China Build World's Largest Electric Bus Fleet? — 16,359 electric buses by the end of 2017.
• Resources for the FutureThe Impact of Delhi's CNG Program on Air Quality, Discussion Paper 07-06 (2007).
• Carbon BriefExplainer: How human-caused aerosols are 'masking' global warming — the IPCC AR6 assessment of aerosol cooling.
• H HeuristicsRelated reports in this series: the clean transition as mitigation infrastructure; one source, two signs, on why cleaning the air and cooling the planet are not the same intervention; and the pollution–poverty–climate nexus.