Abstract

Urban morphology and background climate are evolving drivers of the urban heat island (UHI) effect, yet the thermal influence of surrounding urban structure and its coupling with climate remain elusive. Here, we show that climate and morphology jointly shape global urban heat. Using a six-class urban typology, long-term climatology, and machine learning, we quantify the thermal influence of surrounding urban structure and project future urban heat across 2,213 cities. We define a city-level thermal impact of the surrounding built environment (TBE) as the area-weighting UHI change induced by specific built-up types. Climatically, cold regions most frequently exhibit high daytime TBE, while arid regions dominate high nighttime TBE. Structurally, a universal pattern persists; high TBE corresponds to denser and taller forms, whereas sparser and lower types dominate low TBE during day and night. Future projections indicate that climate change dominates TBE change in 69% of cities, whereas the Global South exhibits stronger tendencies toward morphology-driven and synergistic intensification than the Global North. Our results highlight the need for locally tailored adaptation strategies that target the dominant drivers—climatic, morphological, or both.

Urban Heat Islands (UHIs) are more than just a temperature difference—they emerge from the complex interplay between local climate and how cities are built. Recognizing this connection highlights the need for tailored solutions that address both environmental and structural factors.

Professor Jungho Im from the Department of Civil, Urban, Earth, and Environmental Engineering at UNIST, in collaboration with Professor    Cheolhee Yoo from the  Department of Urban Planning and Engineering at   Pusan National Universit (PNU), Seoul National University (SNU), and Pacific Northwest National Laboratory (PNNL) analyzed data from 2,213 cities worldwide. Using satellite imagery, machine learning, and climate models, they revealed that the intensity and patterns of UHIs depend heavily on regional climate and urban morphology.

The phenomenon occurs as buildings, roads, and pavement absorb heat, altering airflow and trapping warmth. The study shows that in colder regions, UHIs tend to be most prominent during the day, driven by building heat absorption. In dry climates, the effect shifts to nighttime, when heat stored during the day is released, amplifying nighttime heat. Within the same climate zones, urban density and building height also influence the strength of UHIs: dense, high-rise areas retain more heat, while sparser, lower structures tend to moderate it.

Using AI, the researchers divided cities into 1-kilometer grid cells. They analyzed how surrounding urban features and climate conditions contribute to local temperatures. Each cell was categorized into one of six urban types based on building density and height, considering both the immediate area and its surroundings.

Looking ahead, the team projected future UHI trends under different climate and development scenarios. Climate change emerges as the main driver, worsening heat in nearly 70% of cities. Meanwhile, rapid urban growth in developing nations is expected to intensify UHIs, especially where dense, tall buildings become more common.

Professor Im stresses that understanding urban heat requires integrating large-scale spatial data with AI tools. “Our findings show that effective strategies must be tailored to each city's climate and layout,” he said. “One-size-fits-all solutions don't work. We need approaches that reflect local conditions.”

He further added, “Advanced data analysis helps us grasp how cities generate heat. This research will support policies on heatwave response, urban redevelopment, and protecting vulnerable areas.”

The study was led by Siwoo Lee from UNIST and Professor Cheolhee Yoo of PNU, as co-first authors. Supported by the Ministry of Science and ICT (MSIT) and the National Research Foundation of Korea (NRF), the work was published in  Nature Communications  on May 11, 2026.

Journal Reference

Siwoo Lee, Cheolhee Yoo, Bokyung Son,   et al. , “Global patterns of urban heat shaped by climate and morphology,”       Nat. Commun. , (2026).