Abstract
This project develops a spatial modeling framework to understand how wildfire risk evolves across formal and informal settlements in Valparaíso, Chile. Using thirty years of recorded fire event spread, population data, land use cadastre, and digital terrain models, it examines how physical and social variables interact to shape vulnerability in hillside communities. A cellular automata model is employed to simulate fire spread across heterogeneous landscapes, incorporating factors such as hill slope, wind exposure, vegetation type, built density, and access to water and evacuation routes. Each cell in the model represents a micro-unit of the urban-ecological interface, where ignition probability and propagation potential vary according to both environmental conditions and disaster-response status. Formal neighborhoods are typically well-served by suppression infrastructure and early warning systems, while informal areas lack such capacity, producing a differential pattern of spread and containment that the model makes spatially explicit. The simulation results are combined with a multi-criteria evaluation to quantify composite risk levels, translating dynamic fire behavior into cartographic layers of exposure, vulnerability, and adaptive capacity. These maps are validated against the 2014 and 2024 wildfire events to assess predictive performance. The framework is further tested in scenario-based disaster-response simulations to measure differential impacts on formal and informal settlements. The outcome is a hybrid methodology that combines computational modeling and participatory knowledge, advancing a more equitable approach to wildfire risk reduction and urban ecological resilience in contexts of informality.
Presenters
Yujie CaiStudent, Master in Design Studies (Ecologies), Harvard Graduate School of Design, Massachusetts, United States
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
KEYWORDS
Wildfire Modeling, Cellular Automata, Spatial Analytics, Informal Settlements, Risk Governance
