Key points are not available for this paper at this time.
The objective of this paper is to present a method for mapping burned areas in Brazilian Amazonia using Terra MODIS data. The proposed approach is based on INPEs PRODES Digital method that applies image segmentation to the fraction images derived from Landsat TM, using a nonsupervised classification per region algorithm followed by an image edition procedure for minimizing misclassifications. Acre State, located in the western region of Brazilian Amazonia, was considered for this study. The study area is primarily covered by tropical moist forests (Floresta Ombrofila Aberta) that has been deforested during the last decades. MODIS (MOD09) data used in this work were acquired on September 5 and 15, 2005 and on October 12, 2005 over the study area. The INPEs DETER (Detection of Deforested Areas in Real Time Project) 2004 dataset for Acre State was used to identify the land cover types affected by the fires. Landsat TM scene (path/row 002/67) acquired on October 13, 2005 and CBERS CCD (path/row 180/111) acquired on October 12, 2005 and field observations were also used as ground truth for supporting and validating the MODIS results. The results of multitemporal analysis with MODIS showed that about 6,500 km2 of land surface were burned in Acre State. From this, 3,700 km2 corresponded to the previously deforested areas and 2,800 km2 corresponded to the forested areas. This information is critical for regional and global environmental studies. The results showed that MODIS sensor daily data is a useful source of information for mapping burned areas and the proposed method can be used in an operational project in Brazilian Amazonia.
Shimabukuro et al. (Sat,) studied this question.