The added value of global km-scale simulations for representing heat extremes at the city scale

Heat extremes are intensifying with climate change. However, coarsely resolved global climate models (like those from CMIP6) underestimate peak intensities and cannot capture spatial details of heat extremes. These limitations are particularly important at city scales, where population exposure and vulnerability are highest and local feedbacks can further intensify extremes. To support adaptation to heat extremes in cities, it is crucial to provide accurate local information. Therefore, we evaluate how heat extremes are represented in a global km-scale model, with a focus on the city scale. The latest km-scale climate models allow better resolution of small-scale processes and feedbacks, and improve the representation of land-surface heterogeneity at global scale. Therefore, they have the potential to improve the representation of urban heat extremes at scales important for impacts compared to established CMIP6 models. Here, we use the so-called “storyline simulations” from the IFS-FESOM model that reconstruct historical events by nudging the large-scale circulation to observations across three climate states: pre-industrial, present-day, and a future with 2°C warming. First, we evaluate heat extremes in km-scale storyline simulations against observations and reanalysis to establish their applicability. Then, we show the improved representation of heat extremes in km-scale resolution compared to the CMIP6 model resolution. We further quantify changes in the frequency, duration, and spatial extent of heat extremes on a global scale and for selected cities. We find that at the city scale, the added value of km-scale resolution is prominent. For example, during the 2021 heatwave in Athens, peak temperatures at CMIP6-like resolution are underestimated by 8.2°C compared to km-scale representation of the city center. At km-scale, the city is resolved into the city center and surrounding suburban areas, capturing higher local temperatures in the city center and showing spatial variability across the urban area. Our results demonstrate that global km-scale models improve the representation of heat extremes at city scales and allow for consistent analysis of urban heat extremes worldwide. This is relevant for city planners, showing that km-scale models have the potential to inform adaptation strategies under climate change.