Modeling and Zoning of Fire Prone Areas in Zagros Forests Using Geographic Information System Based on Logistic Regression

Forests play a vital role in the sustainability of ecosystems as one of the most important natural renewable resources. One of the most important disturbances affecting Zagros forest ecosystems is forest fires. Therefore, identifying critical fire areas to reduce potential damage is necessary. The purpose of this study is to investigate the effect of effective variables in causing fire and to prepare a fire risk map. For this purpose, the variables affecting the fire including altitude, slope, direction, Distance from residential areas, distance from waterways and distance from the road were determined to determine the impact of each on the fire. Elevation map of sea level, slope and geographical direction was prepared with the help of digital elevation model. Distance maps of residential areas and distance from the road were prepared from digital maps of 1.25000. Also, the areas where fires occurred during the years 90-94 were harvested by GPS. In this study, logistic regression method was used to investigate the effect of various factors on fire. The results showed that altitude, distance from waterway and slope percentage were the most important factors influencing forest fires in the region. Modeling was performed based on three variables that had a significant relationship with forest fires in the region and the coefficients obtained from the logistic regression method. The validation results of the model with a negligence coefficient of about 0.500 and a rock curve coefficient of 0.701 showed the accuracy, fit and validity of the obtained model. The results also showed that 81% of the area is located in critical and dangerous areas.