Using autologistic regression to simulate the forest fire development based on a cellular automata

The forest fire is one of the natural disasters that menaces the large parts of the forests. Forest fires proliferate increasingly due to oxygen resource depletion, greenhouse gases, and rising global temperatures. Due to the devastating effects of forest fires, research on modern methods for simulation forest fire is essential. In this research, Golestan forest fire development has been simulated based on a cellular automata and an autologistic regression was used for calibration. In the proposed method, 5 m spatial resolution has higher accuracy than the other spatial resolutions. Also, using Moore neighborhood filter is better than the Extended Moore neighborhood filter. At the best case in this research, the Kappa index, overall accuracy, and relative operating characteristic (ROC) obtained 79.7%, 87.1% and 84.4%, respectively

کلیدواژه ها:

Simulation ، Forest Fire ، Cellular Automata ، Autologistic Regression