Detection of forest degradation caused by fires in Amazonia from time series of MODIS fraction images

摘要

A new method is presented to detect and assess the extent of burned forests in a tropical ecosystem. Our study area is located in Mato Grosso state southern flank of the Brazilian Amazon region. MODIS images are used over the dry season of year 2010. The proposed method is based on (i) linear spectral mixing model applied to MODIS imagery to derive soil and shade fraction images and (ii) image segmentation and classification applied to a multi-temporal dataset of MODIS-derived images. In a first step, deforested areas are identified and mapped from the soil fraction images while burned areas are identified and mapped from the shade fraction images. Then, burned forest areas are mapped by combining a forest/non forest mask with the resulting burned area map. Our results show that 14,220 km2 of forests were degraded by fire in Mato Grosso during year 2010. Our approach can be potentially used operationally for detecting forest degradation due to fires. The proposed method can also be applied to time series of medium and high spatial resolution images for regional and local analysis.