City trees cut the heat island by nearly half. The average resident gets 0.15°C

Urban tree planting is the most politically comfortable heat adaptation available. It is cheap relative to grid upgrades, it photographs well, and nobody opposes it. What has been missing is a global estimate of how much air temperature it actually removes, and for whom.

A study published on 6 May in Nature Communications supplies one, covering all 8,919 large urban areas on Earth [s1]. It contains two findings that sound contradictory and are not.

The two numbers

The first: current urban tree cover mitigates 41% to 49% of the maximum potential air-temperature urban heat island — the heat island that would exist if the canopy were removed entirely [s1]. Roughly half, which is the number the study's title carries.

The second: tree canopy reduces summer air temperature by a population-weighted mean of 0.15 ± 0.03°C [s1].

Both describe the same trees. They differ because they measure different things. The first is a share of a counterfactual — how much of a hypothetical worst case the existing canopy is holding back. The second is what an average urban resident's thermometer registers because of it.

The second number also conceals enormous spread. Cooling ranged from 0.0°C to 2.7°C across urban areas [s1]. An estimated 914 million people (95% CI 805–1,040 million) receive more than 0.25°C of cooling from existing canopy [s1] — a large population benefiting meaningfully, inside an average that looks negligible.

The distribution is backwards

The finding with the most direct public health consequence is where the cooling goes.

Cooling benefits are greater in already cooler areas — specifically, in high-income countries and in suburbs [s1]. The canopy is doing most of its work where the heat island is least severe and where residents have the most alternative protection: air conditioning, better-insulated housing, lower rates of outdoor work.

Densely settled low-income urban areas, which carry the highest heat-island intensity and the thinnest defences against it, have the least canopy to be credited for. The authors' conclusion is that tree canopy expansion in exactly those areas is necessary for equitable urban heat island mitigation and climate adaptation [s1].

This is a familiar pattern in adaptation research and a difficult one for policy. Tree cover is easiest to add where there is unpaved ground, water infrastructure and maintenance budget — that is, where it already exists. The places where a degree of shade would prevent the most illness are the places where planting is hardest and survival rates lowest.

What trees will not do

The study also tests trees against the warming ahead, and this is where the honest framing gets uncomfortable.

Current tree cover mitigates approximately 10% of the median mid-century climate-change warming under a moderate emissions scenario, with an estimated range of 6.7% to 18% [s1]. A plausible future expansion of tree cover performs almost identically, at 6.3% to 17% [s1].

That is the ceiling. Even substantially more canopy than cities currently have offsets under a fifth of projected mid-century warming under that scenario. Urban forestry is a real intervention with a bounded effect, and the bound is not close to the size of the problem it is frequently offered as an answer to.

What the estimate rests on

This is a modelled global estimate, not a measurement campaign. It derives air-temperature reductions attributable to canopy across nearly nine thousand cities — a scale no direct measurement could reach, and one that necessarily depends on assumptions about how canopy, land cover and local climate interact. The authors note that no global estimate had previously quantified air temperature reductions from contemporary or future tree cover [s1], which is the paper's contribution and also a reminder that there is no independent global estimate to check it against yet.

The wide reported ranges — 0.0 to 2.7°C of cooling, 6.7% to 18% of mid-century warming offset — are the study's own account of that uncertainty, and they should travel with the headline figures rather than be dropped from them.

What it means for a reader

For someone deciding whether a street tree matters, the answer in this paper is that it matters locally and specifically, not on average. A canopy that delivers 2.7°C on one block is doing something a public health department can measure. The 0.15°C global mean is not a description of any particular street; it is what happens when that block is averaged with every unplanted one.

The policy question the study poses is not whether to plant trees. It is whether planting continues to follow the path of least resistance, which the data say has been running toward cooler, wealthier, less dense neighbourhoods [s1].