The Role of Psychological-Physical-Physiological Factors in Decision-Making Behavior in Disasters: Insights from an Experimental Study in Gulangyu

Disaster perception is a multifaceted process involving the transition from visual perception to cognitive processing and, ultimately, to risk decision-making behavior (RDMB). While prior research has predominantly focused on simulation-based evacuation models, the comprehensive quantification of RDMB from a “human-centered perspective” remains underexplored. This study examines RDMB in heritage tourism destinations (HTDs), using Gulangyu as a case study. A “psychological-physical-physiological” framework is employed, incorporating contextual immersion to simulate disaster scenarios. The methodology integrates fully convolutional networks, eye-tracking technology, and street scene selection analysis, supplemented by mean squared error (MSE) tests and Shapley additive explanations (SHAP) interpretability models, to evaluate the influence of streetscape elements (SE) on RDMB. Key findings include: (1) video-based simulations effectively immerse participants in disaster scenarios; (2) an increase in the proportion of sky elements significantly correlates with higher click counts, with notable interaction effects among plant-building, building-sky, building-other, sky-sign, and sky-road elements; and (3) buildings and signs attract the most visual attention, while sky and other elements increase environmental complexity. This model provides a novel, quantitative approach to assessing the impact of streetscape features on RDMB, contributing to improved resilience and emergency response strategies in HTDs.