Prediction and simulation of Malaysian forest fires by random spread
Abstract:This paper aims to study forest fires from risk maps and satellite data in order to predict the location of the burnt zones they provoke. The approach is based on the discrete stochastic model of a random spread, which describes both fire fronts and burnt areas. Under iteration, it provides the time evolution of the process. The probabilities of spontaneous extinction at each stage are derived. The random spread model is then applied to analyse the fires that occurred in the State of Selangor (Malaysia) from 2001 to 2004. It is used, first, to simulate forest fires, according to a law of intensity (fuel combustion) and another of extension (spread rate). But the next, and more important, result lies in the ability of the model to predict all places where burnt scars actually occurred, which turns out to be a significant verification.
Document Type: Research Article
Affiliations: 1: Earth Observation Centre, Faculty of Geography, Universiti Kebangsaan Malaysia, Bangi, Malaysia 2: Universite Paris-Est, Laboratoire d'Informatique Gaspard Monge, Equipe A3SI, ESIEE Paris, France
Publication date: 2010-07-01