Urban warming reshapes the safe outdoor activity opportunity window: dual-metric evidence from Heat Risk Hours and PAO.

Background: Against the backdrop of climate warming and intensifying extreme heat, urban heat stress not only elevates health risks but also disrupts the continuity of health promotion and public service provision by compressing the Meteorological feasibility of time resources for safely conducting outdoor physical activity. However, prior research has largely focused on linear changes in single heat indicators, with limited efforts to characterize-within a unified and comparable framework-the dual-pathway process of hazardous exposure-time expansion and safe activity-window contraction. Existing studies also rarely provide systematic quantification and validation of intra-warm-season risk reallocation, such as increasing late-season concentration and shifts in average peak timing. To address these gaps, we propose an Outdoor activity opportunity window framework that shifts the object of analysis away from observed physical activity behaviors-which are difficult to obtain continuously over long periods-toward a reproducible measure of Meteorological feasibility of time resources. Methods: Using Universal Thermal Climate Index (UTCI) statistics for 31 provincial capital cities and municipalities in China from 2010 to 2024, we developed a dual-metric system comprising Heat Risk Hours (hours under high-risk heat-stress conditions) and Physical Activity Opportunity (PAO) Rate (the share of time within the safe activity window). We applied robust trend estimation and standardized composite metrics to quantify the rate of deterioration in the opportunity window and to produce a tiered ranking of cities. We further conducted an external consistency validation of September UTCI risk hours using station-based observations of daily maximum temperature hot days (≥33°C/≥35°C). In addition, monthly data from May to September were used to test intra-warm-season risk reallocation and average peak timing: at the national scale, the late-season peak share (August-September) increased over time (≥33°C: 0.0076/year; ≥35°C: 0.0055/year), and the mean timing of peak risk shifted later overall (≥33°C: 0.0096 months/year; ≥35°C: 0.0071 months/year). In this study, PAO refers to the meteorology-constrained feasibility opportunity for outdoor physical activity and is intended to represent constraints on time resources; it is not equivalent to the amount of activity that actually occurs, nor is PAO Rate derived from survey-based participation data. Results: Our results show that: (1) a pronounced Dual-pathway squeeze exists at the national scale, with rising heat-risk exposure and shrinking safe activity windows occurring in parallel; (2) cities exhibit substantial heterogeneity in the "speed of worsening," forming an identifiable tiered structure; and (3) warm-season risk undergoes structural reallocation along the time axis, with many cities showing signs of greater late-season concentration, although city-specific trajectories remain heterogeneous. As an application-oriented translation, we incorporated the rate of change in the share of the population aged 65 years and older as a dynamic vulnerability indicator, and coupled opportunity-window deterioration with aging dynamics to identify priority scenario cities. We quantified the magnitude of between-group differences using Cliff's delta (δ), a distribution-free effect size. Priority scenario cities displayed systematic differences in late-season structural indicators, suggesting that the temporal focus and resource prioritization of adaptation-oriented governance can be further refined. Conclusion: This study offers a reproducible dual-metric diagnostic framework linking urban heat risk and the Outdoor activity opportunity window, enabling the identification of priority contexts and supporting targeted investments in age-friendly cooling infrastructure, indoor substitute activity spaces, and late-warm-season warning-scheduling management for more precise resource deployment.