Riding under heat extremes: Exploring thermal and environmental interactions on bike-sharing

Abstract

The rise in extreme heat events has made urban thermal environments a significant constraint on Bike-Sharing Systems (BSS). While many studies have explored the Built Environment (BE) and climate change’s impact on Bike-Sharing Usage (BSU), urban thermal environments remain underexplored. This study uses the XGBoost model and SHAP analysis to examine how thermal and other urban environmental factors influence BSU under extreme heat, incorporating dynamic mean radiant temperature (T_mrt) to capture the spatiotemporal heterogeneity of perceived thermal environments. The findings highlight the crucial role of the urban thermal environment in predicting BSU during non-peak daytime hours. Interactions between T_mrt and factors like POI diversity, bus station distance, and other factors vary across time periods. Clustering and heat risk overlay analysis reveal differing resident decisions to “endure” or “avoid” heat risks, offering insights for designing more climate-resilient BSS and enhancing urban mobility strategies under extreme heat conditions.