Modelling Forest Fire Spread using Cellular Automata, GIS and PSO

Every year, millions hectares of forests are victims of fires that caused by natural and human. Happened wild fires through the world show that this damaging phenomenon, impact on the ecosystem of incompatibility region's vegetation and wildlife. Forests are among the capitals of each country, thus providing effective solutions in order to prevent and manage the complex process environment is ideal for managers and planners. CellularAutomata is a simple method with high efficiency that can simulate dynamics phenomenon like urban expanding, forest fire and so on. This method used in forestfire simulation and has high grade in these studies. But the important thing about thecellular automaton is. In this study swarm intelligent are used for calibration of cellular automata in forest fire. Result of simulation based on cellular automata, swarm intelligent and Geographic Information System shows the ability of this method. Proposed model implemented in South East forest of Minoodasht city of Golestan province. After implementing primitive model in this region, kappa coefficient was 58%, after calibratingmodel by swarm intelligent, this coefficient raised to 92% and overall accuracy was promoted to 95%. Total time of real fire was 23 h and total burned area was 807.5 ha and simulation results showed 22.8 h and 760.8 ha. According to these figures and kappa coefficient, this model can used to simulate forest fires in other regions and would be used as a Support Decision System.