Search

To understand dynamic of burned areas at Ilha Grande National Park (BINP) remote sensing historical data are excellent alternative. The analyze performance this data for detection burned areas is important to minimize omission and commission errors.Herein, this study analyzed the performance of MOD09 data and Fire_cci and MCD45 products to detect burned areas at Big Island National Park compared to Landsat data between 2007 and 2009. MOD09 data, Fire_cci and MCD45 products were used to map burned areas and compared with burned areas mapped by Landsat data. These investigation demonstrated, MOD09 data showed good spatial agreement in burned area mapping compared with Landsat data (Kappa = 0.88, quantity disagreement = 2.9% and allocation disagreement = 2.8%) as well as burned area values and distribution of burned area (mean overestimated of 10%). MOD09 data showed better performance in detect burned area and is suitable to historical reconstruction the dynamic of burned areas at park. The Fire_cci and MCD45 products do not presenter good performances to detect burned areas and are not suitable to understand the dynamic of burned areas at park. [ABSTRACT FROM AUTHOR]