Mapping biomass with remote sensing: a comparison of methods for the case study of Uganda

Table 1 Main characteristics of the biomass and C maps used for the comparative study

Map	Coverage	Format	Spatial resolution/VMMU	Number of classes	Pool	Variable	Biomass data	Spatial data	Ancillary data	Period	Approach
NBS [20]	Uganda	Vector	4 - 50 ha	48	Aboveground	Biomass	NBS Field data	NBS LC map	NBS Ecozone map	circa-2000	CA
Avitabile [21]	Uganda	Raster	30 × 30 m	376	Aboveground	Biomass	subset of NBS Field data	Landsat	None	1999-2003	DR+
Baccini [15]	Tropical Africa	Raster	1 × 1 Km	343	Aboveground	Biomass	Field data	MODIS	None	2000-2003	DR
Drigo [12]	Eastern Africa	Vector	40 - 120 ha	143	Aboveground	Biomass	Biome average	Africover	FAO GEZ	circa-2000	CA
Gibbs & Brown [14]	Tropical Africa	Raster	5 × 5 Km	176	Above and belowground	Biomass	Biome average	GLC2000	CIESIN Population	2000	CA
Henry [4]	sub-Saharan Africa	Raster	300 × 300 m	346	Aboveground	Carbon	Field data	Globcover	FAO GEZ, MODIS VCF	2005	CA+
Reusch & Gibbs [17]	Global	Raster	1 × 1 Km	14	Above and belowground	Carbon	IPCC Biome average	GLC2000	FAO GEZ, Frontier forest map	2000	CA

The spatial resolution of vector-based maps is given by a Variable Minimum Mapping Unit (VMMU), which is class-specific. The NBS map used in this study was derived from the NBS LC map by applying to each stratum its respective average biomass density reported in Drichi [21]. The spatial datasets used by the biomass/C maps included the Africover LC map [36], the Global Land Cover map for the year 2000 (GLC2000) [37], the Globcover 2005 map [38], the Food and Agriculture Organization (FAO) Global Ecological Zone (GEZ) map [30], the MODIS Vegetation Continuous Field (VCF) tree cover products [39], the CIESIN's Gridded Population of the World dataset [40] and the Frontier Forest map [41]. The number of biomass classes of the maps with continuous values was calculated after their conversion to integer values (i.e. classes of 1 Mg ha^-1 interval).

ISSN: 1750-0680