Forests are among Earth's most productive and diverse terrestrial ecosystems.
Forest cover affects local, regional, and global climate; biogeochemical cycles;
and biodiversity. Changes in forest cover thus affect the provision of natural
resources and other services to humans and are of critical importance to the mutual
sustainability of ecosystems and economies worldwide.
Based on the GLCF tree-cover layers for 2000 and 2005,
the GLCF 2000-2005 Forest-Cover Change (FCC 2000-2005) layer represents global changes
in forest cover at 30-meter resolution between 2000 and 2005 epochs, as defined by
the NASA/USGS Global Land Survey collection of Landsat images. Following criteria
set by the United Nations Framework Convention on Climate Change (UNFCCC) and the
United Nations Food and Agriculture Organization (FAO), forest cover in each epoch
was defined as locations >1 hectare in area with >30% tree cover. Persistent forest
or non-forest cover and changes--i.e., loss and gain--over time were detected
probabilistically, incorporating the estimate and the uncertainty of tree-cover
in each pixel. These data thus represent the forest-cover change (FCC) class with
the highest probability in each 30-meter pixel from 2000 to 2005. Water, clouds, and
shadows are identified as well, and a probability layer is included to represent the
certainty of the classification.
The FCC product has been tiled using the
World Referencing 2 (WR2)
tiling system. Each tile has four files associated with it, the change map (CM) file,
the change probability (CP)file, browse and the preview image,
for example, p015r033_FCC_2000200 folder has the following files
Sexton, J.O., P. Noojipady, X.-P. Song, A. Anand, S. McMahon, C. Huang, M. Feng,
S. Channan, J.R. Townshend. in review. From continuous fields to categorical
change: modeling the propagation of uncertainty in land-cover change detection.
Remote Sensing of Environment