Assessment of EnMAP spectral bands for burned area detection using reference data from globally distributed regions

Catastrophic wildfire seasons have highlighted the need for large-scale, high-temporal-resolution monitoring. The Center for Satellite-based Crisis Information (ZKI) of the German Aerospace Center therefore operates an automatic satellite-based wildfire monitoring system using mid-resolution optical sensor data to provide burnt area perimeters with high update frequency, later enhanced with high-resolution imagery. This system was recently updated to support EnMAP, the German hyperspectral spaceborne imaging spectrometer. Given EnMAP's 94 bands in the infrared (NIR and S-SWIR) spectrum, determining the most suitable wavelengths for burnt area derivation is essential. This study examines all 94 bands across 20 globally distributed areas of interest. Results show the optimal wavelength depends on land cover type: approximately 800 nm for forested areas and 1070 nm for grasslands. This is the first broad-scale hyperspectral wildfire analysis study, illustrating the suitability of hyperspectral data for wildfire research. The accompanying reference dataset represents the first publicly available source of globally distributed burnt areas derived from hyperspectral data.