USE OF FOREST INVENTORIES AND GEOGRAPHIC INFORMATION-SYSTEMS TO ESTIMATE BIOMASS DENSITY OF TROPICAL FORESTS - APPLICATION TO TROPICAL AFRICA

One of the most important databases needed for estimating emissions of carbon dioxide resulting from changes in the cover, use, and manageme nt of tropical forests is the total quantity of biomass per unit area, referred to as biomass density. Forest inventories have been shown to be valuable sources of data for estimating biomass density, but inven tories for the tropics are few in number and their quality is poor. Th is lack of reliable data has been overcome by use of a promising appro ach that produces geographically referenced estimates by modeling in a geographic information system (GIS). This approach has been used to p roduce geographically referenced, spatial distributions of potential a nd actual (circa 1980) aboveground biomass density of all forests type s in tropical Africa. Potential and actual biomass density estimates r anged from 33 to 412 Mg ha(-1) (10(6)g ha(-1)) and 20 to 299 Mg ha(-1) , respectively, for very dry lowland to moist lowland forests and from 78 to 197 Mg ha(-1) and 37 to 105 Mg ha(-1), respectively, for montan e-seasonal to montane-moist forests. Of the 37 countries included in t his study, more than half (51%) contained forests that had less than 6 0% of their potential biomass. Actual biomass density for forest veget ation was lowest in Botswana, Niger, Somalia, and Zimbabwe (about 10 t o 15 Mg ha(-1)). Highest estimates for actual biomass density were fou nd in Congo, Equatorial Guinea, Gabon, and Liberia (305 to 344 Mg ha(- 1)). Results from this research effort can contribute to reducing unce rtainty in the inventory of country-level emission by providing consis tent estimates of biomass density at subnational scales that can be us ed with other similarly scaled databases on change in land cover and u se.