Tsunami vertical evacuation (TVE) is increasingly recognized as a critical life-saving strategy in coastal areas where rapid access to natural high ground is not feasible. While prior research has examined TVE infrastructure and urban-scale evacuation, limited attention has been given to evacuee decision making and wayfinding inside high-rise buildings. This article presents an immersive virtual reality (VR) system to evaluate indoor vertical evacuation behavior across 12 existing buildings in four Chilean coastal cities. A total of 227 participants completed individual VR simulations, freely navigating realistic 3D building models during a hypothetical tsunami emergency. The system recorded evacuation trajectories and computed three indicators: total traveled distance, number of direction changes, and whether participants reached the eighth floor or higher. Pre- and post-experience questionnaires collected data on prior evacuation experience, perceived realism, and decision-making factors. Results indicate that participants generally selected effective evacuation routes. In 11 of 12 buildings, most participants reached at least the eighth floor (50–86% of total height), meeting national TVE recommendations. The average optimal route percentage was 98.2%, with no consistent relationship to building height or use, suggesting a stronger influence of internal layout characteristics. Participants reported high immersion and realism, low discomfort, and a strong reliance on floor height as the primary decision factor. Multivariate regression analysis identified age as the most consistent predictor of lower evacuation performance. These findings highlight the potential of existing buildings in Chile to support TVE and demonstrate the value of VR for analyzing indoor evacuation behavior and improving preparedness strategies.
No takes yet. Share an insight, caveat, or question.
León et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: