Validation of image-based earthquake damage maps
On 6 February 2023 a magnitude 7.8 earthquake affected the border region between Türkiye and Syria, killing an estimated 55,000 people and causing damage to more than 4 million buildings in an area exceeding 350,000 km2. Disaster response included many efforts at damage assessment, including with remote sensing data. Such rapid initial damage mapping is known to be of limited accuracy, as a function of image data type and quality, the method used (e.g., visual vs. automated analysis), but also factors such as building type and density, as well as weather conditions and landcover (e.g., tall vegetation). Much research has already been conducted on damages caused by the event, including with analysis of coarse Sentinel images (Yilmaz et al., 2023), medium-resolution but wide-coverage Planet imagery (Bansod, 2024), and with very detailed drone imagery (Schneibel et al., 2023).
However, validation of those maps, and creation of a better understanding of what conditions and factors affect map accuracy continues to be a challenge. Detailed damage data acquired on the ground in the city of Kahramanmaraş are now available, and collaborations with colleagues in Türkiye were established (e.g., Buznego Puerto et al., 2024), as well as such colleagues at the German Space Agency (DLR) that acquired images with a high-quality drone (Brauchle et al., 2024). This creates the possibility to deepen our understanding of how to optimise rapid damage mapping for events such as the one in Türkiye. The approach should include elements of established quantitative map accuracy assessment, but can also focus on more detailed analysis of how different properties of the urban configuration of Kahramanmaraş, but also seismic parameters such as aftershocks that triggered further collapse, have affected the quality of different damage maps. Lastly, it is also possible to investigate how to improve the performance of automated damage mapping approaches based on machine learning.
- BANSOD, S. D. 2024. Post-Disaster Damage and Response Assessment with PlanetScope Satellite Imagery. MSc Thesis, University of Twente.
- BRAUCHLE, J., GEßNER, M., KRAFT, T., HEIN, D., LESMEISTER, M., GONSCHOREK, J., BOCK, M. & BERGER, R. 2024. Regional Rapid Mapping for First Responders - Turkey 2023 Earthquake. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLVIII-3/W3-2024, 15-19.
- BUZNEGO PUERTO, C., VAN DEN HOMBERG, M. J. C., MARGUTTI, J., KUNDAK, S., GOKSU, C., YıLMAZ, A., AŞıCı, A. A. & KERLE, N. Complementarity of field and remote sensing based damage assessments in the early analysis of shelter loss for the February 2023 Türkiye earthquakes. Lecture Notes in Civil Engineering, 2024. 477-489.
- SCHNEIBEL, A., GÄHLER, M., HALBGEWACHS, M., BERGER, R., BRAUCHLE, J., GESSNER, M., GSTAIGER, V., HEIN, D., HENRY, C., MERKLE, N. & KLEIN, D. 2023. USING EARTH OBSERVATION TO SUPPORT FIRST AID RESPONSE IN CRISIS SITUATIONS- LESSONS LEARNED FROM THE EARTHQUAKE IN TURKIYE/SYRIA (2023). 39th International Symposium on Remote Sensing of Environment (ISRSE) - From Human needs to SDGs, Apr 24-28 2023 Antalya, TURKEY. 579-586.
- YILMAZ, C. S., YILMAZ, V., TANSEY, K. & ALJEHANI, N. S. O. 2023. Automated detection of damaged buildings in post-disaster scenarios: a case study of Kahramanmaraş (Türkiye) earthquakes on February 6, 2023. Natural Hazards, 119, 1247-1271.