Abstract
Wildfire is a natural element of many ecosystems as well as a natural disaster to be prevented. Climate and land usage changes have increased the number and size of wildfires in the last few decades. In this situation, governments must be able to manage wildfire, and a risk measure can be crucial to evaluate any preventive action and to support decision-making. In this paper, a risk measure based on ignition and spread probabilities is developed modeling a forest landscape as an interconnected system of homogeneous sectors. The measure is defined as the expected value of losses due to fire, based on the probabilities of each sector burning. An efficient method based on Bayesian networks to compute the probability of fire in each sector is provided. The risk measure is suitable to support decision-making to compare preventive actions and to choose the best alternatives reducing the risk of a network. The paper is divided into three parts. First, we present the theoretical framework on which the risk measure is based, outlining some necessary properties of the fire probabilistic model as well as discussing the definition of the event ‘fire’. In the second part, we show how to avoid topological restrictions in the network and produce a computable and comprehensible wildfire risk measure. Finally, an illustrative case example is included.
Highlights
Forest fires are an annual occurrence in many parts of the world, affecting the population and environment of the adjacent areas with significant economic and ecological losses, and often, human casualties
Ref. [10] is an analysis of operational research challenges in forestry where 33 open problems are formulated, being Problem 20 formulated as follows: ‘How can we develop tractable models that can be used to help determine when and where to implement fuel treatments on large flammable forest and wildland landscapes?’
The aim of this work is to develop a methodology to support decision-making in the fight against wildfires
Summary
Forest fires are an annual occurrence in many parts of the world, affecting the population and environment of the adjacent areas with significant economic and ecological losses, and often, human casualties. A considerable amount of literature has been published on the use of applied mathematics to tackle the problem of wildfires Many of these studies are focusing on subjects such as behavior and spread of fire [3], fire suppression [4], evacuation in case of risk [5,6,7], and location of firebreaks [8,9], for example. The purpose of this paper is to obtain a spatial probabilistic measure for wildfire risk, suitable to compare different landscape configurations after applying fuel treatments. Paper [17] entitled ‘Probability-based models for estimation of wildfire risk’, presents a statistical perspective to solve this problem They use a partition of the landscape for estimating the probability of fire in each 1 km pixel. We use a simulated case study provided by [21] to illustrate the application of the proposed method for decision-making
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call