Mapping forest fires in Antalya using open-access satellite data together with deep learning and machine learning methods, monitoring post-fire recovery of vegetation and land surface temperature, and relating this process to climate variables.
Antalya is one of the provinces in Türkiye most affected by forest fires. This project takes a holistic approach to the forest fires that have occurred in the province over the last decade, using satellite imagery and artificial intelligence methods.
The first component of the project focuses on identifying burned areas from satellite imagery quickly and reliably. In this context, the performance of deep learning models is compared with classical remote sensing methods.
The second component examines how burned forests recover over the years. In addition to changes in vegetation, the effects of fire on land surface temperature are monitored, and the recovery process is related to climate variables such as temperature, precipitation and drought. The factors influencing recovery are revealed using explainable machine learning methods.
The project aims to build an information base that will help identify priority areas for post-fire rehabilitation and afforestation.
