Associations between perceived natural environmental features and physical activity intention: the mediating roles of perceived safety and perceived environmental restorativeness.

Background: Natural environments may support physical activity not only by providing opportunities for movement but also by shaping psychological appraisals of safety and restoration. Guided by prospect-refuge theory, this study distinguished prospect, threat-related refuge, and protective refuge. Objective: We examined whether perceived safety and perceived environmental restorativeness statistically mediated the associations between these perceived features and physical activity intention. Methods: A image-based randomized online survey experiment was conducted among 1,439 young adults. Participants were randomly assigned to evaluate one synthetic but realistic natural-landscape image and then completed measures of the three perceived features, perceived safety, the 11-item Perceived Restorativeness Scale (PRS-11), and physical activity intention. Spearman correlations and structural equation modeling with 10,000 bias-corrected bootstrap samples were used, adjusting for age, biological sex, household registration, educational level, and monthly family income. Results: Prospect and protective refuge were positively associated with physical activity intention, whereas threat-related refuge was negatively associated. The final model showed acceptable fit under commonly used lenient criteria (SRMR = 0.018, TLI = 0.900, CFI = 0.995, RMSEA = 0.064, 90% CI = 0.040 to 0.091). All nine specific indirect associations were statistically significant. For each environmental feature, the pathway through perceived safety alone was the largest specific indirect component in absolute magnitude (prospect: β = 0.151; threat-related refuge: β = -0.096; protective refuge: β = 0.129); the PRS-11-only and serial pathways were smaller. The direct paths to physical activity intention were not statistically significant after the mediators were included, a pattern consistent with indirect-only statistical mediation. Conclusions: Perceived safety and perceived environmental restorativeness may help explain why prospect, low threat-related concealment, and protective refuge are associated with stronger physical activity intention. Because the mediators and outcome were measured at the same time point, these findings do not establish temporal or causal mediation.