Blue-green spaces, heat stress, and health-supportive public-space use under urban stress: machine learning and SEM evidence from jogging flow in Fuzhou, China.

Urban blue-green spaces (BGS) are increasingly discussed in relation to health-oriented and resilience-oriented urban planning, particularly in high-density cities exposed to heat-related environmental stress. However, empirical evidence remains limited on how BGS and heat stress are associated with everyday public-space use. Using Fuzhou, China, as a case study, we integrated crowdsourced jogging data, BGS indicators, land surface temperature (LST), and urban development variables to analyze jogging flow as a behavioral proxy for health-supportive public-space use. Machine learning (ML) and structural equation modeling (SEM) were combined to identify key environmental correlates and examine their direct and indirect pathways. Nine ML models were compared, and SHAP analysis was used to interpret feature importance and nonlinear relationships. The results show that distance to vegetation (DTP), green view index (GVI), sky view index (SVI), and LST were key factors associated with jogging flow. BGS-related variables showed positive associations with jogging flow, whereas higher LST was associated with lower predicted jogging flow. SEM further indicated that green space, blue space, and public activity opportunity proxies were positively associated with jogging flow, while heat stress showed a negative direct association. These findings suggest that BGS may be linked to health-supportive public-space use by providing favorable environmental and activity-related conditions. The study provides planning-relevant evidence for understanding how blue-green environments may contribute to public-space vitality and resilience-oriented urban planning under heat-related urban stress.