Urban green spaces (UGSs) are crucial for public health by supporting leisure-time physical activities (LTPAs), but the mechanisms by which micro-scale UGS features shape different LTPA types remain unclear. In this study, the relationship between the micro-scale features of UGSs and LTPAs was investigated in 63 sample plots of nine comprehensive parks in downtown Shanghai. Using the behavior annotation method and multiple linear regression analysis, we identified significant correlations between the UGS features and LTPA types. The results showed that sitting and chatting (SC) activities had the highest participation rate at 46. 84%, while sports and fitness (SF) activities had the lowest at 9. 82%. Walking and sightseeing (WS) activities and culture and entertainment (CE) activities accounted for 19. 99% and 23. 35% of participants, respectively. Spatial accessibility (SA) and canopy coverage ratios (CCRs) were significantly negatively correlated with SC, while seat number (SN), ground-cover density (DGNC), and three-dimensional green quantity (TGQ) were positively correlated. For WS, SN and tree density (DTREE) were positively correlated, while TGQ was negatively correlated. CE activities were positively associated with SN, DTREE, and Shannon’s diversity index of ground-cover (SHIGNC) but negatively associated with Shannon’s diversity index of trees (SHITREE). The regression models explained 65. 9%, 38. 3%, and 44. 3% of the variance in SC, WS, and CE, respectively, while the overall model was not significant for SF. These findings highlight the need to optimize rest facilities, vegetation diversity, and spatial layout in UGS design to accommodate diverse LTPA needs and foster health-oriented environments. The conclusions are mainly applicable to seasons with mild climates, and LTPA characteristics in different seasons require further verification.
Niu et al. (Tue,) studied this question.