A Bayesian network framework for quantitative analysis of education continuity under earthquake risk: integrating physical and social vulnerabilities

Uninterrupted access to quality education is a cornerstone of community resilience and sustainable development. Yet, earthquakes often pose threats to education continuity by damaging school buildings or reallocating educational spaces for emergency use. This study evaluates education disruption risks, which inherently depend on both physical and social vulnerability. Given the large number of parameters influencing this problem and their associated aleatory and epistemic uncertainties, we propose a Bayesian network (BN) framework. The framework incorporates advanced system reliability tools to handle the complex topology of school compounds, as well as a Delphi survey to quantify the intersectional effects of social vulnerability on shelter demand. The framework is demonstrated by investigating seven schools in Cagayan de Oro, Philippines. The results show the varying risks of education disruption across the schools, given a uniform hazard level, providing insights on disaster equity of the city. To further support decision-making, we perform sensitivity analysis and policy simulations, which reveal varying impacts of vulnerability factors on education continuity. In response to the special issue call, our findings contribute to the quantitative evaluation of disaster risks across communities and vulnerability factors, particularly for equity-informed risk assessment where interactions between physical and social vulnerability shape overall risk.