Heritage tourism destinations (HTDs) typically exhibit high building vulnerability, dense populations, and complex road networks, which increase evacuation risks under hazards. Conventional evacuation models neglect the effects of earthquake damage and agent perception. Gulangyu, inscribed on the UNESCO World Heritage List in 2017, has a dense stock of historic buildings, a fine-grained urban fabric, constrained connectivity, and low spatial legibility under high tourism pressure. Field investigations and local management records indicate that visitor disorientation and requests for assistance occur, and the island’s clear boundary facilitates study-area delineation and more complete data collection. A systematic assessment of seismic building vulnerability and spatial distribution for all 1,867 island buildings, combined with field evacuation experiments and agent-based simulation, was conducted to evaluate the effects of road damage, streetscape conditions, hesitation, panic responses, and companion behavior on evacuation efficiency. The results indicate that: (1) Under a seismic intensity VII scenario for Gulangyu, 25.6% of buildings remained largely intact, 70.6% were slightly damaged, and 3.9% were moderately damaged. (2) In field experiments, 73.53% of participants exhibited sustained tachycardia indicative of panic, 61.76% demonstrated hesitation, and 76.38% evacuated in groups. (3) Simulations showed that strengthening heritage buildings improved evacuation efficiency and that enhanced signage, together with permitting moderate hesitation, supported more orderly organization. Under these measures, the number of remaining evacuees decreased from 2,089 to 425, corresponding to a 79.6% increase in efficiency. These findings inform emergency evacuation management and spatial optimization in HTD settings. • Integrated building vulnerability and UAV-based 3D reality modeling for seismic risk mapping. • Developed agent behavior database using ELAN video and physiological data analysis. • Quantified impact of grouping, hesitation, and panic on evacuation efficiency. • Applied MassMotion to optimize multi-behavior parameters for historic district evacuation. • Incorporated visual attention analysis to refine group decision modeling in complex settings.
Du et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: