Forest fire susceptibility mapping in the Minudasht forests, Golestan province, Iran

Forests are important natural resources having the role of supporting economic activity which plays a significant role in regulating the climate and the carbon cycle. Forest ecosystems are increasingly threatened by fires which caused by a range of natural and anthropogenic factors. Hence, spatial assessment of fire risk is very important to reduce the impacts of wild land fires. In current research, evaluation of forest fire susceptibility is performed using remote sensing and geographic information system data of Minudasht forests, Golestan province, Iran. Factors affecting the fire occurrence, such as normalized difference vegetation index (NDVI) and land use were extracted from classified Landsat-7 ETM⁺ imagery. Slope degree, slope aspect, topographic wetness index, topographic position index, and plan curvature were computed using topographical database. Other factors affecting on the forest fires are distance to roads, distance to rivers, distance to villages, soil texture, wind effect, annual temperature, and annual rain. To delineate forest fire susceptibility mapping in the study area, the Shannon’s entropy (SE) and frequency ratio (FR) models has been applied. Forest fire locations were specified in the study area from MODIS data and extensive field surveys. 106 (≈70 %) locations, out of 151 forest fire identified, were used for the forest fire susceptibility maps, while the remaining 45 (≈30 %) cases were used for the model validation. The findings revealed that the most important conditioning factors were the NDVI, land use, soil and annual temperature. Therefore, preventive measures need to be applied in the ecological conditions. Ultimately, the receiver operating characteristic curve for forest fire susceptibility maps was depicted and the area under the curve was computed. The validation results showed that the area under the curve for SE and FR is equal of 83.16 and 79.85 % with standard errors of 0.044 and 0.047, respectively. The mentioned results can be applied to early warning, fire suppression resources planning and allocation works.