Evaluation and comparison of Sentinel-2 MSI, Landsat 8 OLI, and EFFIS data for forest fires mapping. Illustrations from the summer 2017 fires in Tunisia
Hammadi Achour, Ahmed Toujani, Hichem Trabelsi, Wahbi Jaouadi
Abstract
This study aims to assess the performance of the Sentinel-2 and Landsat 8 sensors to map forest fires. We choose two fire events, the Haddad fire and the Sidi Ferdjani fire, in northwestern Tunisia in 2017. Several spectral indices were derived from each sensor and the performance of each spectral index was assessed. A validation exercise was undertaken for each fire to compare the spatial matching between the burned area retrieved from each spectral index and its homologue obtained from the Emergency Management Service (EMS). Our results indicate that ΔNBR and its relativized version RBR derived from both sensors exhibit the highest discrimination power (M-statistic values >2.5). The Sentinel sensor is slightly more efficient than the Landsat 8 in mapping burned scars, but both sensors produce acceptable results. We conclude that both sensors could be a good alternative to EFFIS data, particularly when there is a need to detect details inside the burned areas.