Modelling Forest Fires Using Complex Networks

Abstract

Forest fires have been a major threat to the environment throughout history. In order to mitigate its consequences, we present, in a first of a series of works, a mathematical model with the purpose of predicting fire spreading in a given land portion divided into patches, considering the area and the rate of spread of each patch as inputs. The rate of spread can be estimated from previous knowledge on fuel availability, weather and terrain conditions. We compute the time duration of the spreading process in a land patch in order to construct and parametrize a landscape network, using cellular automata simulations. We use the multilayer network model to propose a network of networks at the landscape scale, where the nodes are the local patches, each with their own spreading dynamics. We compute some respective network measures and aim, in further work, for the establishment of a fire-break structure according to increasing accuracy simulation results.

Preface to Numerical and Symbolic Computation: Developments and Applications—2021

“…"Modelling Forest Fires Using Complex Networks" constitutes the fifth article in this Special Issue and the authors are Perestrelo et al [5]. In this work, a mathematical model is proposed to predict fire spreading in each land portion divided into patches, considering the area and the rate of spread of each patch as inputs.…”

Modeling Fire Hazards Induced by Volcanic Eruptions: The Case of Stromboli (Italy)

“…Common propagation approaches include cellular automata, envelope models, semi-empirical models, and physical models. The more comprehensive full-physics-based models [16,17] are also computationally expensive (e.g., with representation of vegetation on the mesoscale as a porous medium and/or using balancing of the energy/heat equation), while others prefer a simplified physical model to achieve faster computations, for example solving 2D reaction/diffusion equations [15,18], or using even less physics, such as in estimate-based models using graphs [19], cellular automata [20][21][22][23] or envelope-based approaches [24]. Among the numerical models for simulating the propagation of wildfires, ignition mechanisms that utilise various input parameters are taken into consideration, including different types of trees, canopies, and the influence of wind [25], moisture content, radiant capacity [26], and/or topography [27].…”

A Multi-Scale Network with Percolation Model to Describe the Spreading of Forest Fires

“…For instance, work [27] applies multiplex networks to model fire propagation, simulating a 3-layer of possible fire development: ground, surface, and crown, where each node of the multilayer represents a group of trees. At a larger (landscape) scale, work [28] presents a multi-scale structure, where the nodes are local land patches, each with their own spreading dynamics and, as such, present different times for the fire spread.…”