Author: İbrahim Halil Saylan, Resul Çömert
Publishing Date: 2019
Volume: 1 Issue: 1
Forests, which can be defined as the lungs of the world, cover one third of the world’s land surface and are considered as the main source of biodiversity. One of the biggest factors threatening this biodiversity is forest fires. Mapping of burned areas is a very important factor in determining the damage assessment and planning the interventions in the fire zone. Because it is a fast and cost-effective method, remote sensing is an ideal tool for mapping burned areas. In this study, the success of different band indices produced from Level-1C and Level-2A images of Sentinel-2A satellite in change detection of burned area was examined. The research area was selected as the burned forest area in Kineta, Greece in 2018. In this context, DVI (Difference Vegetation Index), NDVI (Normalized Difference Vegetation Index), NBR-1 (Normalized Burned Area Index 1) and NBR-2 (Normalized Burned Area Index 2) were used as band indices. Within the scope of the study, firstly produced indexes NBR-1 and NBR-2 were applied to post-event images and mapped the burned areas, but the results were found to be insufficient. Therefore, pre-event and post-event difference indices were obtained by providing satellite images before fire. When Level-1C and Level-2A indexes are compared; It was observed that d-NBR-1 and d-NBR-2 indices gave higher accuracy values for Level-1C and d-DVI and d-NDVI indices for Level-2A.
Key Words: Burnt Area, Change Detection, Sentinel-2A, Band İndex