Efficiency Evaluation of Dong Model for Determination of Fire Risk Potential in Zarrin Abad Forests

Continues occurrence of fire in northern forests of Iran during past years, necessitates conducting research, in order to predict future fire occurrence in these forests. For this reason a 1:25000 topographical map of Zarrin Abad Forests of Neka Township of Mazandaran province of I.R. Iran. was geo-referenced in GIS, firstly. Then digital model of elevation (DEM) with 30-meter pixel size was obtained from the geo-referenced map and slope, aspect and elevation maps were developed from the DEM. The vegetation cover density and residential areas (villages) maps were obtained from Nekachoob Corporation Company and digitized in GIS. All roads in the study area were routed using GPS and their map was developed. Then 200 and 500 meter buffers were considered on the maps, around the roads and the villages. All of the digital layers were classified according to Dong model. The fire potential map with five classes was developed by overlaying all of the maps, in GIS environment and allocating weight to each factor (map), regarding the Dong model. The past forest fires map was developed, using actual forest data and applying ground sampling with GPS. Then this map was overlaid on the forest fire potential map. Results showed that about 40 percent of the past forest fire areas are located on sites with very high or high fire risk which indicates the medium validity of the used model. Thus, Dong model was modified and weight of the layers was changed, based on their importance at the studied area and finally a new model was developed. A new fire potential map was generated, based on the modified model and was overlaid on the past forest fires map. Results showed that about 80 percent of the past forest fire areas are located on sites with very high or high fire risk which indicates a high validity of the used model for the studied area. It seems that the model might be applied for other forest sites at Caspian region of Iran, based on spatial locations of their watershed basins.