Understanding the coupling effect of multiple urban features on land surface temperature in Europe

Abstract

Cities in Europe are facing significant challenges as rising temperatures exacerbate health risks, energy consumption, and environmental degradation. This study investigates the multiple features affecting land surface temperature (LST) across 780 European cities, categorized into eight macro-regions. The methodology involved utilizing SHAP values to interpret the results of a Random Forest model that evaluated the impact of individual urban features on LST. Additionally, the Generalized Additive Model (GAM) was employed to explore non-linear relationships between key urban features and LST, offering a deeper understanding of how specific features influence urban temperatures. The results highlight the important cooling effects of ecological attributes like tree height and evapotranspiration, particularly in warm regions (e.g., the Iberian Peninsula and Turkey). Conversely, urban structure elements like built-up volume were shown to increase LST. GAM analysis further revealed non-linear relationships, such as the diminishing returns of evapotranspiration (ET) on cooling in the Iberian Peninsula, where LST decreased sharply at ET < 3–4.5 mm/d but plateaued at higher values. These results emphasize the complexity of urban climate regulation, where ecological and built environments interact in diverse ways across different regions. This study underscores the importance of region-specific urban planning strategies that integrate both ecological and urban structure features to effectively mitigate the effects of urban heat. The findings offer actionable insights for policymakers seeking to enhance urban climate resilience and manage the growing challenges posed by global warming in European cities.