Wildfire Behaviour
Several modelling approaches have been developed to simulate wildfire behaviour (i.e. spread) ranging in sophistication and representation of the processes controlling wildfire spread. Perry (1998) provides a good review of these different options and distinguishes three types of model; physical, semi-physical and empirical.

My research interest and experience is oriented around what Perry classes as 'empirical' models - specifically the class of models known as ‘cellular automata models’. These grid-based fire simulation models divide a landscape into finite elements (cells) and then, by assuming the internal state of a cell is uniform, treats spread within and between cells on an individual basis.

The Forest Fire Cellular Automata (FFCA) model is one good example of a cellular automata. Below is my NetLogo implementation of the model in NetLogo. In this model a two-dimensional grid of cells is considered. Grid cells may be unoccupied (depicted below by brown) or occupied by trees (depicted by green). Trees are randomly dropped on the grid at each time step, and if a cell is unoccupied, then a tree is ‘planted’. If a cell is already occupied, then nothing happens when a tree is dropped on it. At every 1/f time step a ‘match’ is randomly dropped on the grid, where f is the sparking frequency. If the match falls on an unoccupied grid cell, a wildfire is not ignited and the model moves to the next time step where tree drops are again attempted. However, if a match falls on an occupied site the tree burns and the model fire spreads to all neighbouring non-diagonal occupied sites (depicted below by red). You can experiment with the model below by changing the sparking frequency of the model by moving the slider in the top right.