Assessment of Burned Area and Fire Severity Using Sentinel-2 Imagery: Case Studies from Türkiye’s Summer 2025 Wildfires

Authors : Çiğdem Şerifoğlu Yılmaz

Pages : 300-314

Doi:10.48123/rsgis.1751391

View : 117 | Download : 159

Publication Date : 2025-09-27

Article Type : Research Paper

Abstract :This study examines five major wildfires in Türkiye during the summer of 2025 (Aliaga-Foca, Seferihisar-Menderes, Akhisar, Tarakli-Geyve-Golpazari, and Antakya) using Sentinel-2 imagery. Burned areas and fire severity were assessed with Normalized Burn Ratio (NBR), differenced NBR (dNBR), and Normalized Difference Vegetation Index (NDVI). Burn severity mapping was complemented by analyses of pre-fire vegetation stress, wind dynamics, and topographic factors influencing fire spread. Results were validated against European Forest Fire Information System (EFFIS) data. Findings confirm that Sentinel-2 dNBR captured both the extent and severity of fire damage, showing consistency with EFFIS estimates. The most accurate results were achieved for Aliağa-Foça, Akhisar and Taraklı-Geyve-Gölpazarı. Seferihisar-Menderes (96.13 km²) and Tarakli-Geyve-Golpazari (61.31 km²) experienced the largest and most severe fires, marked by heterogeneous burn patterns. Notably, the strongest pre-fire drought signal did not align with the largest fire, underscoring the key role of real-time weather, fuel availability, and topography in shaping fire size and spread. For visual interpretation, False Colour Composite (FCC) imagery outperformed True Colour Composite (TCC) and NDVI. Overall, results highlight the effectiveness of Sentinel-2 and dNBR-based methods for post-fire monitoring and their importance for fire management under growing climate-related fire risks.
Keywords : Orman yangınları, Yanmış alanların tespiti, Yangın şiddeti, Sentinel-2, Uzaktan algılama

ORIGINAL ARTICLE URL