Abstract:

Fighting fires is an extremely important issue to prevent the loss of forests. With the developing technology, satellite images have become one of the basic tools for fighting fire over time. By using various indexes, it is possible to detect the burning area from satellite images with image processing analysis and surface temperature analysis. Unlike the methods made by ground measurements, it is very fast and relatively less costly. In this study, the Deliömerli forest fire that took place in Menderes district of Izmir province on July 1, 2017, and lasted about 67 hours was analyzed using Landsat-8 satellite images. The surface temperature maps were created from the images obtained before and during the fire, and surface temperature differences were examined for different land uses. Accordingly, it has been observed that the surface temperature of the area burning during the fire increased compared to other land use areas. In determining the fire area, three different methods were used and the spatial size of the fire area was compared. In this context, 1,235 hectares with Normalized Difference Vegetation Index (NDVI), 1,221 hectares with Normalized Burn Ratio (NBR) indices, and 1,296 hectares with the classification study were determined to be burned. It has been observed that the results are consistent with the damage assessment results announced by the General Directorate of Forestry. The burning areas were overlapped with the CORINE land use map and the extent of the impact of the fire on each land use class was examined. As a result of the overlap analysis, it was determined that 75.7% of the burning area was coniferous forests, 19.3% plant areas, 4.2% mixed forests and 0.7% agricultural areas.

Keywords: