Subdistrict-Scale day–night assessment of urban heat vulnerability: A multi-dimensional framework and mitigation strategies for megacities – case of Guangzhou, China

Urban heat vulnerability, exacerbated by climate change and rapid urbanization, poses severe risks to human health, infrastructure, and ecosystems in subtropical megacities. Traditional assessments often overlook nighttime dynamics and microscale heterogeneities, limiting targeted interventions. To address these gaps, this study introduces a pioneering modular framework for subdistrict-scale day–night heat vulnerability evaluation, with three key objectives: (1) to construct a multidimensional framework that integrates day-night dynamics, (2) to assess the spatiotemporal heterogeneity and drivers of heat vulnerability at the subdistricts scale (township-level administrative units; Chinese: Jiedao/Zhen), and (3) to propose tailored mitigation strategies. Applied to Guangzhou, China—a subtropical megacity with 19 million residents—the framework leverages multi-source data and entropy-weighted indices to assess 176 subdistricts. Results reveal stark day–night contrasts: daytime vulnerability peaks in suburbs driven by high sensitivity and low adaptive capacity, while central districts exhibit low vulnerability via robust infrastructure. Nocturnally, central exposure dominance intensifie, amplifying urban heat island effects and health burdens, with suburban risks attenuating. Dominant factor analysis identifies heat exposure-driven nocturnal hotspots in cores, heat sensitivity-led daytime surges in peripheries, and adaptive capacity bottlenecks in transitions, informing tailored strategies: green retrofits for exposure, community health programs for sensitivity, and equitable infrastructure for adaptation. This framework advances vulnerability theory by embedding day–night micro-dynamics, offering a replicable tool for precision urban planning, climate justice, and Sustainable Development Goals 11 and 13 in global cities facing escalating heat extremes.