Using Vegetation Continuous Fields to Monitor Change in Forest Cover

Using continous fields of percent forest cover from MODIS data, an approach to monitor change has been developed. The algorithm utilizes a regression tree to automatically generate annual global estimates of forest cover. The 500 meter MODIS data represent an intermediate resolution appropriate for monitoring many types of human-induced land cover change. Three years of MODIS data have been processed and initial results show that deforestation events can be detected on an annual basis. Results from Brazil highlight clearing for agroindustrial purposes, particularly in Mato Grosso state. Field work in this region has been performed and reveals a good correspondence between MODIS-estimated percent forest cover and in situ forest cover measurements. These initial results point the way of operational change detection mapping of global forest cover and other biophysical parameters.