The Compounding Burden of Heat and Air Pollution

For each of 149 large cities, the analysis takes the three hottest months of 2024 and compares the days that were locally hot — at or above the city's own 90th percentile of daily maximum temperature — with the other days of the same season. In 131 cities the mean daily maximum 8-hour ozone level was higher on hot days, by a median of 13%; in 113 cities fine particles were higher too, also by a median of 13%. Ozone forms faster in hot sunshine, and the stagnant, high-pressure weather that brings heatwaves also traps pollution near the ground, so the pattern is expected; the data show how general it is.

Measured against the WHO's daily guidelines — 15 µg/m³ for PM2.5 and 100 µg/m³ for the 8-hour ozone maximum — 85% of hot days were polluted across the cities, weighted by population, against 75% of other hot-season days. The contrast is sharpest in high-income cities, where only half of ordinary summer days broke a guideline but 78% of hot days did: in Berlin 84% of hot days against 5% of other days, in Zurich 85% against 5%, in Madrid 81% against 35%. Heat is the trigger that turns clean air bad. Against the stricter interim targets of 37.5 µg/m³ for particles and 160 µg/m³ for ozone, 37% of hot days exceeded at least one, against 28% of other days.

In the most polluted cities the overlap is near-total: in Delhi, Lahore, Kano and Beijing every locally hot day exceeded a guideline, and Delhi had 91 days at or above 35 °C and 121 nights that never fell below 25 °C. There the problem is not that heat creates pollution but that heat lands on air that is already harmful. The CAMS model used here is a coarse global model whose levels can differ substantially from ground monitors, so absolute values in individual cities are uncertain; the comparison of hot and other days within each city is more robust. Choose a city below to see its season day by day.

Heat strains the body as it works to stay cool. Blood is pushed to the skin and the heart beats harder; sweating depletes fluids; breathing patterns change. The World Health Organization notes that this strain stresses the heart and kidneys, worsens cardiovascular, respiratory and other chronic conditions, and can cause acute kidney injury. A systematic review by Anenberg and colleagues adds that heat increases systemic and pulmonary inflammation as a consequence of thermoregulation and promotes blood clotting. Air pollution acts on the same systems from the other side. Fine particles trigger oxidative stress and inflammation in the lungs that can spread to the whole body, disturb the autonomic control of heart rhythm, and activate clotting; ozone, a powerful oxidant, inflames and damages the airways.

Where pathways converge, effects can compound. The largest study of the question, by Stafoggia and colleagues, analysed 22.6 million deaths in 620 cities in 36 countries during each city's six warmest months. Moving from the 75th to the 99th percentile of a city's temperature raised daily mortality by 8.9% on average — but by 5.3% when PM10 was at 10 µg/m³ and 12.8% when it was at 90, and by 2.9% when ozone was at 40 µg/m³ against 12.5% at 160. Read the other way, each 10 µg/m³ of PM10 raised mortality by 0.54% on cool days and 1.21% on the hottest; ozone had no detectable effect on cool days but 0.53% per 10 µg/m³ on the hottest. The authors call this suggestive evidence of effect modification; results for PM2.5 and nitrogen dioxide were similar but slightly weaker.

The wider literature points the same way. Anenberg and colleagues reviewed 56 studies, 39 of them on heat and air pollution together, and judged the evidence for synergistic effects of heat and air pollution to be of moderate quality and sufficient, while noting that nearly all studies risk exposure-measurement error. The groups most at risk overlap too: the WHO and the US Environmental Protection Agency both single out older adults, children, people with heart and lung disease, and people active outdoors, especially outdoor workers. Select a box in the diagram to trace its pathways and the evidence behind each link, and use the explorer to see how the heat effect grows with pollution.

At the scale of countries, double exposure can be defined by two thresholds: a hot climate, here at least 30 days at or above 35 °C in 2024 in the largest cities, and polluted air, here a national population-weighted PM2.5 of at least 25 µg/m³, the WHO's second interim target. On those definitions 25 countries with 2.6 billion people — a third of humanity — are doubly exposed. They include India, Pakistan and Bangladesh; Egypt, Iraq, Iran and Saudi Arabia; Niger, Mali, Chad, Burkina Faso and Sudan; and Myanmar and Vietnam. The sliders let readers set their own thresholds and watch the count change.

Heat adds a second, quieter dimension: the night. When minimum temperatures stay at or above 25 °C, the body gets no chance to recover from the heat of the day. Averaged over the population of each income group, the largest cities of lower-middle-income countries had 105 such nights in 2024, against 13 in high-income countries; Cotonou, Bangkok and Lagos each had more than 200. In the same cities, the days themselves were hotter — 56 days at or above 35 °C in lower-middle-income countries against 13 in high-income ones — and Khartoum, Juba and Niamey had more than 190.

The map can also show a compound burden index that adds vulnerability and capacity to exposure. It averages each country's percentile rank on hot days, particle pollution, the share of a chosen vulnerable group — older adults, children or people working in farms and industry — and low health spending. The index is a screening tool, not an estimate of deaths: it shows where the overlap of hazard, sensitivity and weak protection is greatest, and how that picture shifts depending on whose vulnerability is in view.

Vulnerability to heat and pollution has two broad profiles. In high-income countries it is mainly a matter of age: 20% of people are 65 or older, and older adults are the group most consistently identified as at risk from both heat and air pollution. In low- and lower-middle-income countries the population is young — 3% and 6% are over 65 — but far more people work outdoors or in hot workplaces: 56% of workers in low-income countries and 39% in lower-middle-income countries are in agriculture alone, against 3% in high-income countries. Children, whose share of the population is highest in poor countries, are a third group singled out by the WHO and the US Environmental Protection Agency.

Exposure, meanwhile, runs the other way from age. The hottest and most polluted cities are concentrated in lower-middle-income countries, and so are the hot nights. The capacity to respond follows income: median health spending per person is about $130 a year in low-income countries, $270 in lower-middle-income countries, $1,180 in upper-middle-income countries and $4,130 in high-income countries, a 32-fold gap. Cooling at home, filtered and air-conditioned clinics, ambulance response and treatment for heart and kidney emergencies all depend on that spending.

The compound burden index combines these strands, and its ranking depends on the lens. Through the lens of older adults, South Asia leads — Bangladesh, Pakistan and India — followed by Niger, Sudan, Syria and Myanmar: the elderly share there is modest, but heat, pollution and thin health spending are extreme. Through the lens of children or outdoor workers, the Sahel rises to the top — Niger, Chad and Mali — followed by Sudan, the Central African Republic, Burkina Faso and Pakistan. The point is not the precise order, which depends on the chosen indicators, but that whichever group is in view, the heaviest combined burden falls on countries that are hot, polluted and under-resourced — and that these are seldom the countries where most research on heat and air pollution has been done.

Taking each country's three hottest months in the ERA5 record and weighting by population, the average person's hot season in 2015–2024 was about 1.2 °C warmer than in 1961–1990, and in 2024 it was 1.8 °C warmer. The rise has not been even. High-income countries, led by central and south-eastern Europe, warmed most, by 1.6 °C over the decade on average; low-income countries by 1.5 °C; upper-middle-income countries by 1.3 °C; and lower-middle-income countries, dominated by South Asia, least, by 0.9 °C on this national, area-weighted measure.

Rates of warming are not the same as danger, however. A degree added to a mild European summer and a degree added to a pre-monsoon summer already near 40 °C have very different consequences, and the second also lands on populations with less cooling, more outdoor work and weaker health services. The Lancet Countdown estimates that in 2024 the average person experienced a record 16 extra health-threatening hot days because of climate change, and that infants and people over 65 experienced an average of 20 heatwave days, increases of 389% and 304% over the 1986–2005 average.

Warming also feeds the pollution side of the burden. Higher temperatures speed the chemistry that makes ozone — the ozone–climate penalty — and hotter, drier conditions fuel the fires whose smoke caused a record 154,000 deaths in 2024. The hot season is therefore lengthening and intensifying at the same time as the pollution that accompanies it, which is why the overlap documented in this report should be expected to grow unless both hazards are reduced.

Most health systems warn about heat and air quality through separate channels, with separate thresholds, often run by different agencies. Yet the evidence above shows that the risk on a hot day depends on the air, and the risk on a polluted day depends on the temperature. The first step is therefore joined-up warning: alerts that combine forecast temperature, ozone and particles, and advice that covers both — staying cool and limiting exertion, but also keeping windows closed during smoke or ozone peaks and using filtered, cooled spaces. The WHO stresses that the health impacts of heat are predictable and largely preventable with public health measures; the same is true of their interaction.

The second step is protecting the groups who sit at the intersection. For older adults and people with heart, lung or kidney disease, that means outreach during combined episodes and access to cooled, filtered rooms in homes, clinics and community centres. For outdoor workers — 2.41 billion people are exposed to excessive heat at work, according to the ILO, which attributes 18,970 deaths and 22.85 million injuries a year to it — it means enforceable rules on rest, shade, water and working hours, extended to bad-air days. For children, it means schools that stay cool and clean when the season peaks.

The third step is cutting the hazards at source, together. Reducing methane and the nitrogen oxides and volatile organic compounds that form ozone lowers the peak that hot days bring; cleaner cooking and heating cut the particles that make poor cities' hot seasons so dangerous; fire management reduces smoke. Cooling needs to grow too, but efficiently and powered cleanly, so that the response to heat does not add to the air pollution that makes heat deadlier. Urban trees and shade cool streets, although some species emit ozone-forming compounds, a reminder that measures should be judged on both hazards at once.

Cities: for every country of 5 million people or more, the largest city in the GeoNames database plus, for the most populous countries, up to seven further cities of over a million people at least one degree apart — 149 cities in 126 countries. Daily maximum and minimum 2-metre temperature for 2024 come from the ERA5 reanalysis via the Open-Meteo archive API. Hourly PM2.5 and ozone for each city's hottest three consecutive months of 2024 come from the Copernicus Atmosphere Monitoring Service global model via the Open-Meteo air-quality API; daily PM2.5 is the 24-hour mean and daily ozone the maximum 8-hour running mean. A day is locally hot if its maximum temperature is at or above the city's 90th percentile for 2024, and polluted if PM2.5 exceeds the WHO 24-hour guideline of 15 µg/m³ or ozone exceeds the 8-hour guideline of 100 µg/m³.

National measures: population-weighted PM2.5 from SatPM (Washington University in St. Louis) via Our World in Data; the warm season as each country's three hottest months in ERA5 national monthly means (area-weighted) via Our World in Data, compared with 1961–1990; and World Bank indicators for population, age structure, employment in agriculture and industry, current health expenditure per person, income and electricity access. Country heat values are the population-weighted averages of their cities. The burden index averages four percentile ranks — days at or above 35 °C, PM2.5, the chosen vulnerable group's share of the population, and low health spending — and is a screening tool, not a risk estimate.

Caveats: city grid cells do not capture street-level or indoor exposure, urban heat islands, or humidity; CAMS levels can differ substantially from ground monitors, and the satellite and model products are not directly comparable; 2024 was an exceptionally hot year; and national averages hide the large differences in heat and pollution exposure within cities by income, housing and work. The joint-effect curve is interpolated between published estimates. Related reports: the unequal geography of air pollution exposure, the pollution–climate feedback system and human exposure across the polycrisis century.