The effects of uncontrolled fires on natural ecosystems and the factors affecting their occurrence are widely studied worldwide. Fire occurrence modeling is important to prioritize the fire risk in a given area and identify practices to prevent it from happening. The purpose of the present study was to simulate the prediction of fire and identify the most important factors in fire occurrence using Bayesian Belief network in Chaharmahal and Bakhtiari Province. Data from 205 fire sites and 205 sites with ‘no-fire’ experience were recorded. Climatic factors (e.g. annual precipitation and annual mean temperature), topography (elevation, slope, direction), land cover, and human factors (e.g. distance from the residential area, distance from agricultural land and distance from the road) of the sites were selected and embedded into the BBN model. The results indicated that the ability of the BBN model to predict the occurrence of fire was excellent (AUC= 0.923). Since there are many fire events in this province, the results of this study can be used as a fundamental and powerful tool for decision-makers to reduce the incidence of fire and its hazard.

Keywords: Central Zagros Region, Wildfire, Modelling, Prediction.

Full-Text [PDF 1039 kb]   (1347 Downloads)