2019 Wildfire Conference Adressing the Challenges of Bushfire Management
Presentations Notes 2019: DAY 1 (Day 2 - Wednesday Nov. 13th, 2019).
A Bayesian Network Approach to Manage Fire Risk on a Network
(A. Rodríguez-Martínez, B. Vitoriano, I. Leguey / Group UCM-HUMLOG / Complutense University of Madrid )
Problem description. Measuring fire risk on a network is a difficult task since fire ignition and propagation is subject to many influencing factors, including the human factor. It is difficult even when a fire is declared, but it is much more difficult in advance when prevention and mitigation tasks are developed. Many scenarios can arise, and decisions must be made taking into account the uncertainty related to them. Among the prevention and mitigation tasks, in addition to the obvious tasks of education and awareness among the population, the most popular measures are prescribed burning and firebreaks. However, it is difficult to quantify how they can impact on fire risk to compare different landscape configurations. Therefore, a first step is to develop a model able to obtain a quantitative measure to approximate the global fire risk. It will be a Bayesian network model based on a landscape representation as a probabilistic network. This model will be used later to compare strategies choosing those ones with higher impact reducing risk with limited resources. Moreover, a decision model for locating fuel breaks over time subject to budget constraints will be shown.