Assessing the relationship between urban park spatial features and physical activity levels in Residents: A spatial analysis Utilizing drone remote sensing

•Utilizing drone remote sensing and GIS for precise park and Physical Activity (PA) analysis.•Identifying 12 environmental attribute integration types in parks linked to different PA characteristics.•Highlighting paved areas and sports facilities’ impact on the number of PA participants and intensity.•Suggesting diversifying sports spaces to support various PA needs in parks. The park environment is crucial for promoting physical activity (PA). While numerous studies show that park environments influence PA behavior, inconsistencies remain, likely due to varing research methods and parks types. This study employs a fixed spatial grid method to systematically sample four representative parks in Tianjin, China. High-precision orthophoto map (DOM) data from drones provided detailed environmental attributes (like tree canopy area, lawn area, and paved area) and PA characteristics (number of participants, intensity, diversity). The results show: 1) Cluster analysis grouped 1839 park grids into 12 environmental attribute integrations, each correlating with different PA characteristics. “Tree-lined jogging corridors” and “Large sports field areas” exhibit the highest PA intensity, while “Entrance plazas”, “Central plazas,” and “Open sports spaces” have the highest number of participants and PA diversity. 2) Correlation analysis shows that various environmental attributes, including Lawn Area, and Paved Area, are significantly correlated with PA characteristics. 3）Random Forest analysis indicates the key attributes are the paved area for the number of PA participants and PA diversity, and specialized sports facilities area for PA intensity. These findings support urban green space planning and highlight the importance of better park environments for public health.