A study published in Nature Climate Change covering more than 3,000 cities worldwide overturns a well-established assumption in climate adaptation policy. The urban heat island effect—the overheating of cities relative to their rural surroundings—reduces worldwide cold-related mortality at a level more than four times higher than the increase in heat-related mortality. Deploying cooling strategies without accounting for this double accounting risks worsening the annual health balance of cities in regions still subject to harsh winters.
The essentials
- The urban heat island effect reduces worldwide cold-related mortality at a level more than four times higher than the increase in heat-related mortality, according to Nature Climate Change (2025) on more than 3,000 cities.
- Widespread cooling strategies have an unfavorable net balance in high-latitude cities, where winter protection outweighs summer exposure.
- The success criterion for an urban adaptation policy can no longer be temperature reduction alone: it must integrate deaths avoided over the entire year.
- Solutions exist—seasonal shading, combined insulation, albedo adjustment—but their deployment requires local diagnosis that few cities currently possess.
- The real risk is uniformity: applying the same approaches to Seoul and Singapore, to Stockholm and Lagos, ignoring that the heat-cold trade-off varies radically according to latitude.
Study methodology and implications for urban policy
For about two decades, the urban heat island effect has been presented as a public health problem to solve. Cities absorb heat during the day and release it at night, drive temperatures up two to four degrees above surrounding rural areas, and expose residents to heat waves. The logical response: white roofs, vegetation, fountains, reflective materials. The stated objective is always the same: lower the thermometer.
The study published in Nature Climate Change in 2025 examines a dimension that these policies have generally neglected: the winter effects of the urban heat island. The same extra degrees that kill in August protect in January. On a sample of more than 3,000 cities, researchers calculated heat-related and cold-related mortality by comparing urban zones to their rural equivalents. Result: the reduction in cold-related deaths thanks to the urban heat island is more than four times greater than the increase in heat-related deaths it causes.
This figure does not say that urban heat waves are benign. It says that the net balance of the urban heat island is positive in many cities, and that any strategy aimed at eliminating it without verifying this balance risks killing more people in winter than it saves in summer.
The geography of thermal trade-off
The balance does not distribute uniformly. In Singapore or Miami, where winters are mild and summers are deadly, cooling the city is almost without drawback: cold does not kill, heat does. The classical logic of urban cooling applies without reservation.
In Stockholm, Moscow, Chicago, or Montreal, the calculation is different. Winters are long and cold, and the urban heat island plays a significant role for certain vulnerable populations. Eliminating this protection to face increasingly frequent summer heat waves, though not yet comparable to winter hardships in these latitudes, can worsen the annual health balance.
The study identifies this latitudinal gradient as the central variable. The closer a city is to the poles, the higher the winter benefit of the urban heat island, and the greater the risk posed by a uniform cooling policy. High-latitude cities are in a structurally different situation from those in the tropics or Mediterranean zones, and adaptation policies have not yet drawn all the consequences from this.
Who is exposed within cities
The geography between cities is only the first layer of the problem. Within a single city, exposure to cold and heat varies according to socioeconomic and infrastructural factors.
Elderly people living alone in poorly insulated apartments are both the most vulnerable to heat waves and the most dependent on ambient warmth in winter. Precarious populations, those who cannot afford heating or air conditioning, are directly exposed to both thermal extremes. Energy poverty—the inability to maintain a home at a sufficient temperature—is a mortality factor for both seasons, not just for summer.
A spatially differentiated cooling policy can produce contrasting effects by neighborhood and building characteristics. Consideration of social vulnerabilities appears in the recommendations of heat-health action plans, even if concrete implementation varies by territory. Heat action plans often measure their success by reductions in surface temperature rather than by reductions in mortality across the entire year.
The limits of surface metrics
The measurement problem is central to the issue. Dominant indicators in urban adaptation policies are temperature metrics: degrees gained through vegetation, reduction of the heat island effect measured by satellite, lowering of thermal peaks during heat waves. These indicators are visible, communicable, easy to integrate into a policy dashboard.
They do not say how many lives are saved over the entire year. An urban project can show a 1.5-degree decline during heat waves and simultaneously cause an increase in winter mortality, without any monitoring indicator flagging it.
Researchers from the Nature Climate Change study advocate for a different metric: the number of deaths avoided annually, aggregating both seasons. This shift may seem technical. It is actually political: it changes the success criteria for a policy, and thus the trade-offs it must make. A city adopting this metric will be led to prefer solutions that protect in winter as much as they cool in summer, to diagnose its local thermal risk profile before deploying standardized solutions, and to integrate mortality into its adaptation plans rather than rely solely on satellite data.
This type of reasoning is not without precedent in other sectors of public policy. The Andean energy transition showed how a climate objective formulated without accounting for local secondary effects—here water, there seasonal mortality—can produce results opposite to those intended.
Under what conditions cooling becomes a solid health strategy
The study’s result does not argue against cooling cities. It argues against uniform, undiagnosed cooling, measured by thermometer rather than by mortality balance. Solutions exist, and their logic is known even if their deployment remains marginal.
Seasonal shading—adjustable shading devices that protect from the sun in summer but allow heat in winter—is a documented technical path. Vegetation with deciduous species naturally produces this effect: the tree provides shade in summer, loses its leaves in winter, and allows solar heat to pass through. These solutions require finer design work than reflective roofs or high-emissivity materials, deployed identically year-round.
Building thermal insulation is the other structural lever. Well-insulated housing reduces both exposure to summer heat and dependence on ambient warmth in winter. Energy renovations that combine these two objectives, still a minority in renovation programs in France and elsewhere in Europe, are both the most costly in the short term and the most effective on annual balance.
Modulation of urban albedo—the capacity of surfaces to reflect solar radiation—could eventually become seasonal, with materials that adapt their reflectivity based on temperature or season. These technologies exist as prototypes; their scaling up is still far off.
Finally, heat action plans must be rethought as annual thermal risk plans. This means mapping, street by street and neighborhood by neighborhood, both summer heat exposure and winter cold vulnerability, and identifying populations facing both. Several Nordic cities have begun this work; it remains extremely rare in Mediterranean or continental cities.
Two trajectories for urban adaptation by 2050
Climate adaptation plans in major cities are written today for horizons extending to 2050. The Nature Climate Change study implicitly sketches two possible trajectories.
The first is uniform diffusion. Cooling solutions—white roofs, reflective materials, fountains, vegetation—are deployed at large scale in cities facing increasingly hot summers. Surface temperature indicators improve. But in high-latitude cities, the annual health balance deteriorates: heat-related deaths avoided in summer are offset, or even surpassed, by excess cold-related deaths in winter. Precarious populations, exposed to both extremes and unable to respond individually, pay the price of policies designed for only the warm season.
The second trajectory is targeted adaptation. Cities begin with a diagnosis of their annual thermal risk profile, integrating both seasons and intra-urban variations. They deploy differentiated solutions by neighborhood, building, and population: seasonal shading where summer is deadly, reinforced insulation where cold is still dangerous, renovation plans that address both summer overheating and winter protection. They evaluate their projects on deaths avoided annually, not on degrees of cooling.
This second trajectory is technically accessible. It is politically harder: it requires local diagnoses that few cities currently have the tools to produce, mortality metrics that local elected officials do not yet master, and coordination between health, urban planning, and energy policies that remains the exception. It also requires accepting that the right answer for Vilnius is different from that for Marseille, and that uniformity of solutions is a risk as much as a convenience.
The signals to watch to know which trajectory prevails are fairly precise. The joint evolution of heat-related and cold-related deaths by city, if statistical systems commit the means to measure it. The share of renovations combining summer overheating protection and winter insulation in national programs. The adoption of annual mortality metrics in local adaptation plans, today nearly absent from reference documents. These three signals will tell, by 2030, whether the lesson of this study has been heard or whether urban adaptation has remained trapped in a single-season logic.
Energy poverty remains the filter that reveals most quickly the blind spots of climate policies, because the least solvent populations always combine the exposures that aggregated indicators erase.
The precise condition that urban planners and epidemiologists must establish together in the coming years is the following: reduce exposure to extreme heat without increasing vulnerability to cold episodes. The Nature Climate Change study shows that the answer depends on location and season, and that an adaptation policy that ignores these variables can produce unexpected effects.



