Assessing the Probability of Forest Fire Occurring using Dong Model, Artificial Neural Network and K Nearest Neighbors in Babolrood Basin, Mazandaran

Forest fire is recognized as a significant threat to the safety of human life, infrastructure and the environment. One of the most important steps in reducing the risk of forest fires is determination of the areas with the high probability of forest fire occurrence. Choosing the appropriate methods for modelling of the forest fires is very important. Due to the importance of the issue in this study, first using library studies and expert advices, the most important variables affecting the occurrence of fire in Babolrood basin-Mazandaran province were determined and then the results of three models of dong, artificial neural network and K nearest neighbors were compared in determining the probability of fire occurrence. The results showed that the most important variables affecting the fire occurrence are temperature, rainfall and distance from residential areas. The results of artificial neural network are more reliable than the other two models. According to the results, about 35% of the study area has very high and high potential for forest fire.