Cities are getting hotter, putting some neighborhoods – and especially vulnerable residents – at serious health risk. As extreme heat becomes more frequent, cities urgently need better ways to understand where heat builds up and how it can be reduced.

Maps that show only urban heat intensity offer limited guidance for action. More useful are approaches that link heat patterns to the physical form of neighborhoods, such as building density, height, and vegetation. The widely used Local Climate Zone (LCZ) framework does this, but its fixed categories often miss important local variations within the same zone.

To address this gap, a recent study by PDPC Frontrunner 4: Towards Social and Urban Resilience researchers uses a data-driven clustering approach to identify residential urban forms across 99 cities in the Netherlands. Using just five key characteristics – floor space index, ground space index, open space ratio, average building height, and fraction of vegetation cover – the method distinguishes nine residential ‘typo-morphologies’ at a neighborhood-level scale, revealing underrepresented LCZ types and new sub-classes. The findings show that recently densified areas experience some of the highest heat levels, while differences in vegetation within low-rise neighborhoods strongly influence local temperatures. Overall, the newly developed approach preserves LCZ’s global comparability while better capturing local conditions, providing a stronger foundation for targeted heat adaptation and greening strategies in residential areas.

The study was performed by researchers from the Pandemic and Disaster Preparedness Center (PDPC), and was first published on February 1, 2026 in Sustainable Cities and Society:

Habib, M., Zendeli, D., Van Esch, M., Timmermans, W. J., & Van Ham, M. (2026). Unsupervised clustering approach to residential typo-morphologies across multiple cities for urban heat vulnerability assessment. Sustainable Cities and Society, 107107.

Read the full publication here.