What Actually Cools a City Down: The Evidence on Trees, Cool Roofs and Heat Action Plans — Analysis

Every summer produces the same headlines about record temperatures, and every summer the same confusion follows: is a hot city hot because of climate change, or because it is a city? The answer is both, and the distinction matters enormously for what you can do about it. Global warming raises the baseline. The urban heat island is a local surcharge on top of that baseline, created entirely by how a place is built — and unlike the baseline, it is something a city council can act on within a single budget cycle.

What the urban heat island actually is

The mechanism is unglamorous physics. Cities replace vegetation, soil and water with asphalt, brick, concrete and steel. Those materials absorb solar radiation efficiently and release it slowly. Meanwhile the plants that would have cooled the air through evapotranspiration — effectively sweating — are gone, and tall buildings along narrow streets block the wind that would otherwise flush warm air out. Add waste heat from vehicles, industry and, ironically, air conditioners, and you get a persistent thermal anomaly.

The US Environmental Protection Agency quantifies the effect at roughly 1-7°F higher daytime temperatures and 2-5°F higher nighttime temperatures than outlying areas. For a metro of a million or more, annual mean air temperature runs 1.8-5.4°F (1-3°C) above its surroundings, and on a clear, windless evening the gap can widen to as much as 22°F (12°C). Surface temperatures diverge far more dramatically than air temperatures: sunlit roofs and pavement can sit 50-90°F hotter than the air above them, which is why a parking lot radiates heat well past sunset.

Two details are routinely misunderstood. First, heat islands are not exclusive to megacities — they form in small towns and suburbs too, and they occur in spring and autumn, not just August. Second, the daytime number is the less important one. It is the compressed nighttime cooling that does the damage, because overnight recovery is how the human body discharges accumulated heat strain.

Why heat is the deadliest weather hazard

Heat kills quietly. There is no collapsed bridge or flooded street to photograph, so the toll is usually reconstructed statistically months later. The World Health Organization cites research finding approximately 489,000 heat-related deaths per year between 2000 and 2019 — with 45% in Asia and 36% in Europe — and describes heat stress as the leading cause of weather-related deaths. Heat-related mortality among people over 65 rose roughly 85% between 2000-2004 and 2017-2021. Individual events can be catastrophic: an estimated 61,672 heat-related excess deaths in Europe during the summer of 2022, and around 70,000 across the June-August 2003 European heatwave.

That toll is landing on a warmer baseline. The World Meteorological Organization confirmed in January 2026 that 2025 was among the three warmest years in the 176-year record, at about 1.44°C above the 1850-1900 average, and that the eleven years from 2015 to 2025 were the eleven warmest ever measured. Notably, 2025 both began and ended under cooling La Niña conditions and still ranked that high. Shift the whole distribution upward and a local surcharge that was merely uncomfortable starts pushing ordinary summer days across medical thresholds.

There is a grid dimension too. The EPA estimates that electricity demand for cooling rises 1.5-2.0% for every 1°F of air-temperature increase above roughly 68-77°F, implying that 5-10% of community-wide electricity demand exists purely to compensate for the heat island. Peak demand arrives on hot weekday afternoons — precisely when systems are most stressed — which links urban heat directly to blackout risk and to the emissions from marginal fossil generation.

What actually cools a city down

Here the evidence is more nuanced than the advocacy. A 2025 study of southern New England in Communications Earth & Environment estimated marginal afternoon air-temperature sensitivities of about -0.61°C per 0.1 increase in surface albedo and -0.07°C per 0.1 increase in tree-canopy fraction. But in the authors' Boston optimization — which layered feasibility and cost constraints onto those sensitivities — canopy expansion produced roughly 35% larger air-temperature reductions, while cool roofs delivered the greater total reduction in heat exposure, because nearly twice as much treatable area sat in the dense, vulnerable neighbourhoods that need relief most.

That apparent contradiction is the whole lesson: the best intervention is not the one with the strongest physics per unit area, but the one you can deploy across the most area where people actually live. Modelling elsewhere points the same way. A Melbourne study simulating ten years of summers found that combining maximum feasible ground vegetation, green roofs and cool roofs could cut air temperatures by up to 2.4°C — more than any single measure achieved alone. Work in Athens found reflective roofs and roads produced the largest daytime effect, with mean ambient reductions up to about 1°C but surface-temperature reductions of 9.5-11.5°C.

Trees do something the albedo numbers understate: they provide shade, which governs how hot a pedestrian actually feels regardless of the air temperature reading. Reflective pavement, by contrast, can worsen daytime pedestrian comfort by bouncing sunlight onto people while still improving conditions at night. Neither is a silver bullet, and the honest answer for most cities is a portfolio: canopy where roots and water allow, reflective roofing at scale on everything else.

The cheapest intervention is a plan

The highest-return measure is not a material at all. Ahmedabad, India, launched South Asia's first city-level heat action plan in 2013 after a deadly 2010 heatwave, combining a colour-coded alert system, cooling centres, health-worker training, public awareness campaigns and a cool-roofs programme. A World Bank ex-post assessment estimated roughly 2,380 deaths avoided in 2014-15 alone. India's National Disaster Management Authority issued national heat-action guidelines in 2016 and revised them in 2019, citing Ahmedabad as the template.

Governance follows. Phoenix appointed one of the world's first municipal chief heat officers in 2021, on the straightforward logic that a hazard which is nobody's explicit responsibility gets addressed by nobody. Warning infrastructure is spreading too: the WMO reports that the number of countries with multi-hazard early warning systems more than doubled from 56 to 119 by 2024 — though roughly 40% of countries still lack one. The WHO has estimated that scaling heat-health warning systems across 57 countries could save close to 100,000 lives a year. Compared with retrofitting a skyline, alerts and cooling centres are almost free.

The takeaway

Urban heat is the rare climate problem where the physics is settled, the interventions are cheap and locally controlled, and the main obstacle is attention. Cities cannot lower the global baseline on their own, but they can reliably shave several degrees off the local surcharge — and, critically, restore some overnight cooling — with canopy, albedo and a functioning alert system. For residents, the practical signal is to treat heat warnings with the seriousness given to storms, and to recognise that the dangerous hours are the ones after dark. For anyone planning infrastructure with a thirty-year life, the design temperature should be the one expected in the 2040s, not the average of the last decade.