Optimum Sentinel-1 Pixel Spacing for Burned Area Mapping

This study evaluated the relationship between mapping accuracy and computing time when detecting burned areas from Sentinel-1 C-band backscatter coefficient images processed at different pixel spacing (i.e., 20, 30, 40 and 50 m). The analysis was carried out in six study areas distributed worldwide. An existing burned area algorithm, based on Sentinel-1 C-band post-to pre-fire backscatter coefficient differences and the Reed-Xiaoli anomaly detector was used to detect and map burned areas. Marginal differences in accuracy were observed when varying the spatial spacing with slightly more accurate products being obtained at higher spacing as speckle was reduced. However, the computing time was considerably higher when processing at the lowest spacing (20 m) in comparison to the remaining pixel spacing tested. The results suggest that detecting burned areas at 30 to 50 m pixel spacing from C-band backscatter provides significant advantages over large areas (i.e., continental to global scales) while also improving the detection accuracy.