Robust Portfolio Prioritization of Fire-Risk Mitigation Actions in Vulnerable WUI Settlements: A Case-Informed Decision Framework for El Olivar, Chile
Cristóbal Ignacio Galleguillos Ketterer, Nicolás Norambuena Ortega, José Luis Valin Rivera, Wanderley Ferreira de Amorim, Leticia Arancibia Martínez, Lorena Jorquera Martínez, Adrián Klapp FritzFire-risk mitigation in vulnerable wildland–urban interface (WUI) settlements is a portfolio problem: warning, evacuation, fuel management, dwelling hardening, and spatial interventions are complementary yet compete for limited implementation resources. This study develops an auditable portfolio decision framework and applies it to public-document evidence for El Olivar, Viña del Mar, Chile. Six actions are represented by evidence-bounded ordinal envelopes for risk-reduction relevance, feasibility, time to initial benefit, and resilience to delayed external access; implementation-resource demand is imposed as a hard constraint rather than a compensatory criterion. A binary portfolio model is evaluated under equal, rank-order-centroid, unrestricted-simplex, and ordered-simplex preference structures using 200,000 uncertainty realizations without assuming equal spacing between ordinal categories. Under the baseline normalized implementation requirements (1–3 additive planning units, not monetary costs), routes/maps plus decentralized warning has a portfolio-acceptability fraction of 89.2–95.3% at B=2. At B=4, vegetation-fuel management joins that core, with a portfolio-acceptability fraction of 66.7–91.5%. These fractions quantify robustness of optimal selection across the defined admissible evidence and preference inputs; they are not empirical probabilities of intervention effectiveness, fire-loss reduction, or implementation success. Structural actions enter mainly as resources expand, while cluster-scale compartmentation remains limited by transferability to attached-housing morphology. The framework identifies staged bundles conditional on stated evidence, preference, and resource assumptions rather than a universal action ranking.