Global warming and urbanization have exacerbated urban heat island issues in cities, affecting the liveability and long-term health and well-being of its citizens. This study explores how landscape composition and configuration regulate diurnal air temperature variations and the cooling potential of urban green spaces.

Using a spatially-dense citizen science weather station network in Leuven (Belgium), mean night temperature, mean day temperature, and diurnal temperature range were recorded during summer heat periods of 2022. Urban land cover data were collected at two buffer scales (50m and 250m) and analysed using composition (relative cover) and configuration metrics (aggregation and shape indices). Multiple linear (mixed) models were used to investigate associations between the composition and configuration of various land covers and air temperature at different temporal and spatial scales.

Results reveal dynamic effects of urban landscape on air temperature. A 10% increase in tree cover mitigated daytime warming by 0.11°C, likely through shading and evapotranspiration. Aggregated trees further improved daytime cooling by 0.42°C. A 10% increase in grasses and shrubs offered significant night-time cooling up to 0.47°C, probably due to high albedo and ventilation potential. Additionally, increasing the aggregation of grasses and shrubs by 10% enhanced nocturnal cooling by 1.62°C, while increasing the shape complexity of trees promotes night-time cooling by 0.44°C.

This study further highlights the need for context-specific strategies. Conserving green locations as urban forests, while designing enclosed grey locations as urban savannas, characterized by interconnected grasses and shrubs interspersed with standalone trees, could mitigate urban heat stress and enhance urban resilience.