The net rate of carbon storage in forests varies as a result of interannual variability in rainfall, temperature, and disturbance regimes. For example, transient simulations with the Terrestrial Ecosystem Model (TEM) suggest that forests of the Amazon basin during hot, dry El Niño years were a net source of carbon (of as much as 0.2 Gt C yr-1), whereas in other years they were a net sink (as much as 0.7 Gt C yr-1) (Tian et al., 1998). Notably, source and sink strength varied across the basin, indicating regional variation. Similarly for the United States, the net rate of carbon storage varied from a net source of about 0.1 Gt C yr-1 to a net sink of 0.2 Gt C yr-1 (Schimel et al., 2000). The rate of carbon accumulation in undisturbed forests of the Amazon basin reported by Tian et al. (1998) was approximately equal to the annual source from deforestation in the same area. For the United States, Houghton et al. (1999)using inventory dataestimate a sink of 0.35 Gt yr-1 during the 1980s as a result of forest management and regrowth on abandoned agricultural lands, whereas Schimel et al. (2000) indicate that 0.08 Gt yr-1 stored from 1980 to 1993 could have been a result of carbon fertilization and climate effects. Thus, for the United States, the rate of carbon accumulation in forest regrowth on abandoned agricultural and harvested managed forest lands appears to be as large as or larger than the direct effects of CO2 and climate (Schimel et al., 2000). The SAR (Kirschbaum et al., 1996) suggests that modeled trends described above are unlikely to continue under projected climate change and high elevated CO2 concentrations. However, climatic variability (including ENSO events) is likely to increase under a changed climate (see Table 3-10), which may increase interannual variation in regional carbon uptakeas demonstrated by Tian et al. (1998) and Schimel et al. (2000).
In addition to fire and insect predation, other episodic losses may become increasingly important in response to locally extreme weather events. For example, in Europe, wind-throw damage has appeared to increase steadily from negligible values prior to about 1950; wind-throws exceeding 20 million m3 have occurred 10 times since then (UN-ECE/FAO, 2000b). Losses in 1990 and 1999 were estimated at 120 million and 193 million m3, respectivelythe latter equivalent to 2 years of harvest and the result of just three storms over a period of 3 days (UN-ECE/FAO, 2000b). The unusual 1998 icestorms in eastern North America caused heavy damage to infrastructure as well as large areas of forest, the extent of which is still being assessed (Irland, 1998).
Financial returns to forest landowners decline if these episodic events (including fire and insect predation) increase (Haight et al., 1995), although the impact on global timber supply of such episodic losses is unlikely to be significant. Local effects on timber and non-wood goods and services may be significant, although timber loss may be ameliorated through salvage logging. Moreover, the location of the loss is particularly important for non-wood products and services. For example, forest fires in recreational areas are known to have an impact (Englin et al., 1996).
Other reports in this collection