Deforestation detection in Brazilian Amazon region in a near real time using Terra MODIS daily data

This work presents a methodology for detecting deforestation activities in a near real time using Terra MODIS daily data. This work is part of the operational PRODES Project of INPE (Brazil) that estimates the annual deforestation in Brazilian Amazon. The detection of deforested areas by MODIS daily data is a useful information for enforcement of forest protection laws. The proposed methodology, named "DETER" Project, will follow the dynamics of deforestation activities while the PRODES Digital project is evaluating the total area deforested during the study year using higher spatial resolution data. For the development of this methodology, it was used a temporal series of MODIS MOD09 product (surface reflectance) from June to October over an area covered by one Landsat ETM+ scene located in Mato Grosso State, Brazilian Amazon region. The obtained results were compared to the corresponding Landsat ETM+ images acquired in the same dates of MODIS images. The small deforestation areas (lower than 15 ha) are detected with less accuracy by MODIS. As the areas deforested increases, the MODIS results come near to the ETM+ results. The results showed that MODIS sensor daily data can be used for an operational deforestation detection in the Brazilian Amazon.