Building-Stock Age Composition and Surface-Heat Persistence in Seoul: Landsat and Building-Geodata Evidence for Heat-Resilience Screening

High-density cities require sustainable urban planning approaches that can screen recurrent surface heat while accounting for existing building-stock conditions. This study examines whether building-stock age composition, interpreted as a proxy for urban-renewal conditions and building-stock management needs, is associated with surface-heat persistence in Seoul, South Korea. Using 19 cloud-filtered summer Landsat Collection 2 Level-2 scenes from 2019 to 2025 and 485,473 building-age polygons, we construct a 250 m grid-cell evidence matrix. q75/q90 persistence is defined as the proportion of valid scenes in which a grid cell falls within the upper-tail scene-normalized LST distribution. Area-weighted mean building age is positively associated with q75 and q90 persistence; a p25-to-p75 increase in mean age corresponds to a 5.17 percentage-point increase in q75 persistence (95% CI: 4.54–5.81). The association is retained under spatial-filtering, spatial-lag-covariate, valid-pixel coverage-proxy, year-balanced, and threshold-universe checks. A planning-priority classification identifies 238 very-high-priority cells and 1166 high-priority cells for field review, roof/material surveys, vegetation/shading audits, and retrofit-feasibility assessment. The framework is a screening and prioritization tool, not evidence of causal redevelopment effects or direct human thermal exposure.