Fire Spread Prediction Using Probabilistic Cellular Automata: The Case of Urban Settlements in the Philippines

Abstract

Abstract. Fire disasters are common occurrences in the urban settlements of the Philippines. Concerned agencies like the Bureau of Fire Protection (BFP) and the Disaster and Risk Reduction Management Office (DRRMO) are constantly planning ways to prevent and mitigate fire disasters. The key to an effective plan against fire disaster is understanding how a potential fire can spread in a community. By combining both GIS and Probabilistic Cellular Automata (PCA), this paper solves the task of fire spread modeling and simulation. PCA is a model that consists of a regular grid of cells, whose cells are updated according to rules that take into account both the cell’s current state and the cell’s neighbors’ states. The model we developed factors in wind, building materials, and building density. The model was designed after several fires in major cities of Cebu, Philippines. An accuracy of 83.54% and a Cohen’s Kappa coefficient of 0.67 was achieved. Further, a web-based tool was developed to aid in fire disaster planning.

The influence of wind and the spatial layout of dwellings on fire spread in informal settlements in Cape Town

“…Where dynamic fire spread models have been applied they may have the advantage of producing relative risk distributions across individual buildings at the cost of being computationally heavy [43], or map risk at a cellular-, rather than individual building-, level [44]. Therefore, the approach laid out in this paper is comparably a novel and effective method to use as a quick evaluation of risk, relying on relatively fewer metrics.…”