From heatwaves to rapid temperature drops: a Hong Kong study using annual performance metrics for resilient low-income subtropical housing

• Annual framework combing downscaled WRF, EnergyPlus and RTD detection • Integrated design increases summer comfort hours >60% under future climates • Integrated measures cut cooling load >70% and winter discomfort hours about 30% • Reduces RTD days by 44% and mean RTD magnitude around 20% by late century • Sobol GSA identifies shading, buffer zones and thermal zoning as high-leverage This paper develops a year-round, policy-ready framework that evaluates and optimises thermal resilience in subtropical low-income housing by explicitly incorporating rapid temperature-drop (RTD) events and novel cross-season performance metrics. The study employs WRF-downscaled hourly climates (historical, mid- and late-century SSPs), year-round EnergyPlus simulations, a sliding-window RTD detector and Sobol global sensitivity analysis to compare three design strategies (baseline, summer-optimised, fully integrated annual). Six annual metrics are introduced: May–Sep comfort hours (operative T ≤ 28°C), summer cooling-load reduction, Nov–March cold-discomfort hours (T < 18°C), winter heating-load change, RTD frequency (≥ 9°C drop / 6 h) and mean RTD magnitude. Results show integrated passive-adaptive interventions preserve summer cooling benefits while substantially improving winter resilience: large increases in summer comfort hours, >70% cooling savings in favourable cases, ∼30% fewer winter cold-discomfort hours, a 44% reduction in RTD days and meaningful reductions in RTD magnitude under late-century extremes. The framework exposes seasonal trade-offs (shading vs. winter solar gain; ventilation vs. airtightness) and yields ranked, high-leverage design levers for codes and retrofit policy. Framing design using annual RTD-aware metrics prevents single-season solutions that inadvertently shift risk between summer and winter.