Land Use, Land-Use Change and Forestry

	Other reports in this collection

3.5.4. Global Assessment

Policymakers require information on potential global carbon amounts that may result from implementation of Article 3.3. Scientific data to estimate the potential amount of carbon involved in ARD activities on a global scale since 1990 are unsatisfactorily sparse. At best, with the few available estimates, one can put coarse bounds on those amounts and refine the bounds with results from isolated regional studies and with mathematical models.

3.5.4.1. How Much Carbon is Global Deforestation Likely to Release and How is It Distributed Spatially?

The effects of deforestation will depend in large part on the amount of land deforested and the amount of carbon in forests at the time of deforestation. Published averages are about 400 t C ha^-1 above and below ground in boreal forests, 150 t C ha^-1 in temperate forests, and 250 t C ha^-1 in tropical forests (Dixon et al., 1994). There is considerable variability to these values. Olson et al. (1983) estimate that aboveground biomass of all forest and woodland varies between 40 and 250 t C ha^-1, with boreal forest and woodland ranging from 20 to 80 t C ha^-1 in the north and 60 to 140 t C ha^-1 in the south. These values for boreal forest do not include the large quantities of soil organic carbon reported from Russian forests (Dixon et al., 1994). Temperate forests also vary widely: from 60 to 140 t C ha^-1 for closed forest. Tropical regions may contain as little as 20-50 t C ha^-1 in savannas and dry forests but up to 250 t C ha^-1 in wet tropical forests (Olson et al., 1983). The global area of forests is about 4.2 billion ha, including 1.4 billion ha in the boreal zone, 1.0 billion ha in temperate forests, and 1.8 billion ha in tropical forests (Dixon et al., 1994). The estimate of global forest area is much more sensitive than that of biomass to minimum tree cover on which the forest definition is based.

The partitioning of organic carbon into living biomass and dead organic matter pools differs greatly between ecological zones (Olson et al., 1983). The proportion of total ecosystem carbon in below-ground biomass and soil organic matter is greatest in boreal forests (84 percent) and smallest in the tropics (50 percent), with temperate forest values falling between those two (63 percent). Therefore, the inclusion of soil carbon into the definition of forest carbon will produce considerable liability for highest-latitude countries when deforestation emissions are considered, with increasingly less liability nearer the equator. Moreover, the decision about which carbon pools to include under the Kyoto Protocol will have differential effects on the timing of the carbon release in all three regions. On average, deforestation activities are likely to release less carbon than the true carbon stock values because not all carbon will be oxidized or exported as wood products. In addition, oxidation of organic residues resulting from deforestation declines as a negative exponential process requiring many years (Harmon et al., 1996; Micales and Skog, 1997). Thus, the emission of carbon from soils in the first decade or so following deforestation will be much more rapid than carbon sequestration during the first decade or so following afforestation and reforestation (Harmon et al., 1986; Kirilenko and Solomon, 1998).

The total amount of carbon that may be released annually to the atmosphere by global deforestation because of agriculture and fuel wood extraction was projected with a spatially explicit model (IMAGE 2.1) by Alcamo et al. (1996). Their projections at 5-year intervals (Table 3-16) include carbon releases of 0.94 Gt C yr^-1 in 1990. Linear interpolation of their 5-year interval data produces values of 1.85 Gt C yr^-1 in 2008 and 2.09 Gt C yr^-1 in 2012. More recent simulations with a more detailed version of the model (Leemans et al., 1998) are lower overall, with estimated carbon releases of 1.02 Gt C yr^-1 in 1990, 1.31 Gt C yr^-1 in 2008, and 1.31 Gt C yr^-1 in 2012. Yamagata and Alexandrov (1999) obtained a value of 1.0 Gt C yr^-1 for 2005-2010. These latter values are considerably lower than the most recent estimates of carbon emissions from deforestation during 1980-1989: 1.7 � 0.8 Gt C yr^-1 (Houghton, 1999; see Section 1.2.1.2).

These estimates are sensitive to the forest definitions used, however. For example, if savannas are considered to be forests (normally they are not)-and conversion of savannas is considered as deforestation-the estimate of Yamagata and Alexandrov rises to 1.8 Gt C yr^-1. In contrast, inclusion of a 40 percent tree cover threshold (excluding both savanna and some open woodland) reduces the estimate from 1.8 to 0.9 Gt C yr^-1.