• biodiversity;
  • carbon sequestration;
  • climate change;
  • land-use change;
  • plantation;
  • REDD;
  • savanna;
  • scrub


While research continues on the causes, consequences, and rates of deforestation and forest degradation in the tropics, there is little agreement about what exactly is being lost, what we want back, and to whom the ‘we’ refers. Particularly unsettling is that many analyses and well-intended actions are implemented in fogs of ambiguity surrounding definitions of the term ‘forest’—a problem that is not solely semantic; with development of markets for biomass carbon, vegetation classification exercises take on new relevance. For example, according to the basic implementation guidelines of the Kyoto Protocol, closed canopy natural forest could be replaced by monoclonal plantations of genetically engineered exotic tree species and no deforestation would have occurred. Following these same guidelines, carbon credits for afforestation could be available for planting trees in species-rich savannas; these new plantations would count towards a country moving towards the ‘forest transition,’ the point at which there is no net ‘forest’ loss. Such obvious conflicts between biodiversity conservation and carbon sequestration might be avoided if ‘forest’ was clearly defined and if other vegetation types and other ecosystem values were explicitly recognized. While acknowledging that no one approach to vegetation classification is likely to satisfy all users at all scales, we present an approach that recognizes the importance of species composition, reflects the utility of land-cover characteristics that are identifiable via remote sensing, and acknowledges that many sorts of forest degradation do not reduce carbon stocks (e.g., defaunation) or canopy cover (e.g., over-harvesting of understory nontimber forest products).