Integrating land use dynamics and human health risk: A multi-scale assessment of future urban heat exposure using...

Urban heat islands (UHIs) pose a growing environmental challenge associated with rapid urbanization, leading to increased heat exposure risks for urban populations. The associations between land use configurations and UHIs have been established in many earlier works, however, limited studies have quantitatively assessed the differential impacts of multiple land use types across scales of space. To bridge this gap, we introduce an integrated approach that leverages the PLUS model with the SHAP method to predict future UHI patterns under multi-scale land use influences. The framework is implemented in Wuhan, China, which identifies the optimal spatial scale of influence for each land use type and projects future UHI distributions under three different land use scenarios. The results show that land use patterns at raster scales of 390 m and below have the greatest impact on UHI intensity, while fine-scale single-point effects are limited. Water bodies demonstrate strong cooling effects with rapid decline over distance, whereas the cooling contribution of vegetation is constrained by its fragmented spatial distribution. However, SHAP-based temporal analysis from 2005 to 2022 reveals that the mitigation effect of water has declined, with its relative importance decreasing by 7.6 %, while the intensifying influence of urban areas has grown by 10.6 %. Together, these spatial and temporal patterns suggest that enhancing vegetation connectivity and preserving the proximity between urban and water could effectively mitigate UHI effects. In addition, scenario simulations further indicate that optimizing land use patterns and improving how land use is organized has the potential to lower heat exposure risks for more than 1,000,000 people, which emphasize the potential for policy interventions to improve thermal environments in cities.