A new study by researchers from the University of East Anglia (UEA) and the Karlsruhe Institute of Technology (KIT) shows that many tropical and subtropical cities could warm up much faster than previously assumed.

The study originates from the doctoral thesis of Sarah Berk, who completed her doctorate at the UEA and was supervised by T.-T. Professor Peer Nowack at KIT, among others. It is published in the Proceedings of the National Academy of Sciences (PNAS):
https://www.pnas.org/doi/10.1073/pnas.2502873123

Classical global climate models do not have a high enough spatial resolution to provide data for individual cities. High-resolution urban climate models, on the other hand, are usually only simulated for a few cities or a few large metropolitan areas due to the high computing costs. To circumvent this bottleneck, the research team led by T. T. Professor Peer Nowack from the Institute for Theoretical Computer Science used machine learning models to analyze 104 medium-sized cities in the tropics and subtropics for the first time with regard to future changes in the urban heat island under climate change scenarios. The AI translates climate projections into more localized changes in urban surface temperatures based on previously observed relationships. Using this new data basis, previously overlooked intensifications of urban heat islands could be made visible.

The results show:

• In around 80% of the cities studied, temperatures are rising faster than in the surrounding areas.
• Normalized to scenarios with a global warming of 2˚C, the warming in around 15% of cities could even be 50-100% higher than in the surrounding rural regions.
• Cities in monsoon regions such as India, China and West Africa are particularly affected.

The AI-supported analysis shows, for example, that some cities in north-eastern China and northern India could experience up to 3°C of warming - significantly more than the 1.5-2°C predicted for surrounding areas.
For the first time, the use of AI makes it possible to see how badly such medium-sized cities could be affected in the future. According to the authors, this should help to respond more specifically to health risks and urban heat planning, as more extreme heat events will occur more frequently in the future anyway and could therefore have an even greater impact in cities.