Why Is the City Still Hot After Sunset?
Urban heat islands do more than raise temperatures. They change how we sleep, when we can open a window, and how much of the day we can actually use the city.
At nine on a summer night in Kaohsiung, southern Taiwan, the sun has been down for hours.
The asphalt is still giving off heat. Exterior walls and balcony tiles are returning what they stored in the afternoon. Open a window and the air may not move; switch off the air conditioner and the room slowly warms again.
It is easy to say that Kaohsiung is simply hot. But if weather were the only explanation, why would one part of a city begin to cool while another kept releasing the day's heat?
The sun has gone down. Why is the city still giving off heat?
A City That Cannot Cool Down
Cooling Singapore compared today's built-up city with a hypothetical “all-green” version in which urban areas were replaced by vegetation. In the most affected modeled locations, the maximum nighttime urban heat island intensity reached roughly 4–7°C.1
That does not mean two existing neighborhoods in Singapore are routinely measured seven degrees apart. It does not mean every part of the city experiences the same difference every night, or that planting trees will automatically lower temperatures by seven degrees.
What the model shows is the possible scale of the effect. Roads, roofs, walls, traffic, heat released by human activity, and the loss of vegetation can change how much heat a place retains after dark.
Climate change and the urban heat island are not the same phenomenon. Climate change raises temperatures across a wider region. The urban heat island makes a built-up area warmer than less developed surroundings under the same broader weather conditions. When the two overlap, an already hot city has even less opportunity to recover at night.
Asphalt, concrete, metal roofs, and building façades absorb solar energy during the day and release it gradually after sunset. Trees and vegetation behave differently. A canopy blocks direct sunlight from reaching people and the ground, while plants and moisture-bearing soil move some heat away through evapotranspiration.
Wind matters as well.
Tzu-Ping Lin, a professor at National Cheng Kung University in nearby Tainan, has long studied thermal comfort, shade, and ventilation in subtropical cities. The university summed up one of his central ideas in the title of an English profile: “Create a pathway for the wind.”2
When broad building faces block the prevailing wind and the gaps between buildings are too narrow, air has difficulty moving through a neighborhood. Heat leaves the ground and building surfaces more slowly, while heat and air pollutants are more likely to remain trapped.
Building orientation, spacing, and continuous air paths therefore affect more than the view from a window. They help determine whether a neighborhood can release heat after sunset.
A city that absorbs more heat during the day and releases it more slowly at night may begin the next hot morning before the previous day's heat has fully dispersed.
The deeper problem is not one exceptionally hot afternoon. It is a city losing its ability to cool down again each night.
The Same 31°C Is Not the Same Heat
The air temperature in a weather forecast is not the same as the heat a person experiences while standing on a street.
One comparison presented at National Cheng Kung University showed an air temperature of 31.1°C in the sun and 31.0°C beneath a tree—almost no difference. Yet what the Chinese-language figure calls 體感溫度—translated here as “perceived temperature”—was about 45°C in the sun and 33°C in the shade, a difference of roughly 12°C.3
The tree did not lower the surrounding air temperature by twelve degrees.
Its canopy blocked solar radiation from reaching the body and the ground directly. Human heat exposure depends not only on air temperature, but also on sunlight, wind speed, humidity, and radiant heat from nearby roads, walls, and other surfaces.
The 12°C figure is one documented example, not a fixed effect that every tree will produce under every condition. Even so, it makes an important distinction visible: a weather station can record nearly identical air temperatures while the human body encounters two very different environments.
Trees are therefore not merely decorative. Shade can determine whether a sidewalk is usable in summer, whether people can wait at a bus stop, and whether an outdoor space works in daily life rather than only on a site plan.
What Hot Nights Take from the Body
When people think about heat-related illness, they often picture heatstroke, collapse, or an ambulance arriving. The body begins paying a price well before those visible emergencies occur.
To release heat, it increases blood flow to the skin and continues producing sweat. The heart and kidneys work harder to maintain body temperature and fluid balance. The World Health Organization warns that prolonged daytime and nighttime heat allows heat stress to accumulate, increasing the risk of heat exhaustion and heatstroke while also worsening cardiovascular, respiratory, kidney, and diabetes-related conditions.4
For a young and healthy person, the effects may appear as fatigue, poor concentration, or irritability. For older people and those already living with heart, kidney, or metabolic conditions, the same heat may add to an existing physical burden.
And heat exposure does not end at noon.
Night is normally when the body cools, sleeps, and recovers. When roofs, exterior walls, and streets continue radiating heat late into the night, bedrooms become harder to cool as well.
A study drawing on more than seven million sleep records from 47,628 adults in 68 countries found that on nights above 30°C, people slept about 14 minutes less on average than during the nights in the study's temperature range associated with the least temperature-related sleep loss. Older adults and people in lower-income countries showed greater effects.5
Fourteen minutes may not sound dramatic if it happens once. Hot nights, however, rarely arrive one at a time.
Repeated sleep loss follows people into the next day. Driving, riding a scooter, operating machinery, caring for children, attending meetings, and making decisions all depend on attention and reaction time.
A much smaller observational study followed 44 university students during a 12-day heat wave in the Boston area. Students living in buildings without air conditioning responded more slowly on morning tests of attention, processing speed, and calculation than students in air-conditioned buildings. The sample was small and cannot represent everyone, but it helps show that heat-related impairment is not confined to older or medically vulnerable people.6
When a city cannot cool at night, the body loses part of the time it would normally use to recover.
Air Conditioning Cools Rooms, Not Cities
Air conditioning is effective. During dangerous heat, it is also an essential form of health protection.
It would be neither reasonable nor safe to ask older adults, infants, people with chronic illness, or anyone else to endure hazardous indoor temperatures in the name of saving energy.
But air conditioning can cool only the spaces it reaches.
Homes, cars, offices, and some public buildings can be air-conditioned. Balconies, sidewalks, bus stops, school entrances, routes children take home, the minute spent waiting at a traffic light on a scooter, and the short walk from a parking space to a front door remain outside.
We can air-condition a room. We cannot air-condition an entire city.
If streets become difficult to walk after ten in the morning, if an older family member still cannot go outside after dinner, if children must remain indoors after school, and if a balcony stays hot late into the evening, the city has lost more than a few degrees.
It has lost part of the day in which ordinary life could have taken place.
Nor is urban heat distributed evenly. Some people can move from an air-conditioned home to an air-conditioned car and then into an air-conditioned workplace. Landscaping, construction, road maintenance, sanitation, logistics, and traffic control cannot all be moved indoors. The city continues to function because many people remain outside.
Taiwan's occupational safety rules provide one concrete example: when outdoor work reaches high levels of heat risk, they call for measures such as protection from direct sun, access to a cool rest area, and drinking water.7
Extreme heat is not an abstract environmental issue. It is a workplace safety issue. Those of us able to remain indoors often depend on someone else continuing to work outside.
A Kaohsiung Paradox: Lower PM2.5, Closed Windows
Kaohsiung presents another, less obvious contradiction.
The Kaohsiung–Pingtung Air Quality Zone is an official monitoring region covering Kaohsiung City and neighboring Pingtung County. In its 2024 annual review, Taiwan's Ministry of Environment reported an average PM2.5 concentration of 9.4 micrograms per cubic meter during the southwest monsoon period and 18.1 during the northeast monsoon period—nearly twice as high.8
This does not mean every summer day or every location in Kaohsiung is suitable for opening windows. The figures are regional seasonal averages, and they address PM2.5 rather than every form of air pollution. Ozone, traffic, construction, and other local sources still require attention to the day's air-quality readings and the conditions immediately around a home.
But when PM2.5 alone is considered, the warmer southwest-monsoon season can offer a better opportunity to bring in outdoor air. It is also the season when heat is most likely to make residents close their windows.
The split-system and window air conditioners common in Taiwanese homes usually cool and recirculate indoor air. On their own, they generally do not bring in outdoor air. Cooling and ventilation are different functions.9
Natural ventilation is not as simple as adding a window.
Che-Ming Chiang, a Taiwanese scholar whose research includes natural ventilation and indoor air quality, treats natural ventilation as a building-design problem. Wind pressure and temperature-driven buoyancy can move air through openings, but the result depends on the outdoor wind field, the position of inlets and outlets, and whether air has a continuous path through the interior.10
A window is not, by itself, a ventilation path.
A home can have windows without receiving useful airflow. Wind can reach one opening without crossing the room and leaving through another.
The urban heat island therefore does more than make the street hotter. It also shortens the period in which a home can comfortably use outdoor air.
A good home does not reject air conditioning. It preserves alternatives when conditions allow. When temperature, air quality, and safety are suitable, residents should still have the option to open a window and exchange indoor and outdoor air. Natural ventilation is not a substitute for cooling during dangerous heat, and opening windows is not advisable when outdoor air quality is poor. The value lies in retaining a real choice.
Some Heat Is Built into the City
The urban heat island is not an unavoidable consequence of summer.
Road materials, the continuity of tree shade, the amount of soil and vegetation left uncovered, the placement of buildings, the paths available to wind, and the protection of windows from direct sun all affect the heat people eventually experience.
Singapore has extensive air conditioning, but it has not treated air conditioning as a complete solution to urban heat.
Cooling Singapore is developing a Digital Urban Climate Twin that brings together buildings, ground surfaces, vegetation, traffic, heat released by human activity, solar radiation, and wind. It is intended to help planners compare how different proposals might affect urban temperatures and outdoor thermal comfort before construction takes place.11
The significance of the model is not simply its technical complexity. It reflects a more basic decision:
A city should not wait until construction is finished and then hand all of its heat to air conditioning.
If the way a city is built can contribute several degrees to nighttime heat under particular conditions, then the value of shade, ventilation, and large green spaces cannot be measured only by scenery, recreation, or the number of people who enter them.
That leads to a more practical question. If a large park stores less heat and releases cooler air after sunset, does its effect stop at the gate?
The next question is how far that cooling can travel—from nighttime measurements at Shinjuku Gyoen to research on sixty parks in Taipei.
Sources and Notes
The figures in this essay belong to specific studies, locations, periods, measurements, or modeled scenarios. They should not be read as fixed temperature differences that occur in every city, neighborhood, or night.
- Cooling Singapore / Singapore-ETH Centre, “Research”. The 4–7°C range comes from a model comparing current urbanized conditions with a hypothetical “all-green” condition in which urban areas were replaced by vegetation. It represents maximum nighttime urban heat island intensity at particular modeled locations, not a routine measured difference between two existing neighborhoods.
- National Cheng Kung University, “Create a Pathway for the Wind: Prof. Tzu-Ping Lin's Solution to the Urban Heat Island Effect”, and the university magazine interview “讓路給風走——專訪建築學系林子平老師”. On shade, ventilation, building orientation, spacing, and the accumulation of heat and air pollutants.
- National Cheng Kung University magazine, “讓路給風走——專訪建築學系林子平老師”. The comparison of 31.1°C in the sun and 31.0°C beneath a tree, with what the Chinese-language source calls 體感溫度—translated here as “perceived temperature”—at about 45°C and 33°C, is one example presented on the university campus. It does not mean trees always lower air temperature by 12°C.
- World Health Organization, “Heat and Health”. On cumulative heat stress, heat illness, and the worsening of cardiovascular, respiratory, kidney, and diabetes-related conditions.
- Minor, K. et al., “Rising Temperatures Erode Human Sleep Globally”, One Earth, 2022. The observational study analyzed more than seven million sleep records from 47,628 adults in 68 countries. The average reduction of about 14.08 minutes on nights above 30°C is relative to the study's reference conditions, not a fixed loss experienced by every person on every hot night.
- Cedeño Laurent, J. G. et al., “Reduced Cognitive Function During a Heat Wave Among Residents of Non-Air-Conditioned Buildings”, PLOS Medicine, 2018. This small observational study followed 44 university students in the Boston metropolitan area over 12 days; 24 lived in air-conditioned dormitories and 20 in non-air-conditioned dormitories.
- Taiwan Occupational Safety and Health Administration, “Occupational Safety and Health Facilities Rules—Prevention of Hazards in High-Temperature Outdoor Work”. Article 303-1 applies when outdoor work reaches Level 4 or higher under Table 3, with exceptions including some short-duration work. The English title is a descriptive translation by Tang-Yu Life.
- Taiwan Ministry of Environment, 2024 Annual Review of Air Pollution Prevention and Control, fig. 28. On average PM2.5 concentrations in the Kaohsiung–Pingtung Air Quality Zone during the southwest and northeast monsoon periods. The English title is a descriptive translation by Tang-Yu Life.
- Taiwan Ministry of Environment, Indoor Air Quality Information Network, “室內空品迷思解答懶人包”. On the distinction between recirculating cooled indoor air and bringing in outdoor air.
- Che-Ming Chiang, “綠色居住生活科技之研發及改善現況之調查與策略研究—子計畫二:綠色通風科技之策略與研發(I)”, National Cheng Kung University Repository. On natural ventilation, openings, outdoor wind conditions, and indoor airflow paths.
- Cooling Singapore / Singapore-ETH Centre, “Cooling Singapore”. On the Digital Urban Climate Twin under development and its intended use in comparing urban-planning scenarios.
Video background: Bloomberg Originals, “How Singapore Uses Science to Stay Cool”.