Figure 1-3: Present observed distribution of global vegetation complexes redrawn from the database compiled by Olson et al. (1983). For comparison, the vegetation classes have been aggregated to the same classification system as that used for all other model results shown here. Figure 1-4: Past equilibrium vegetation according to BIOME (Prentice et al., 1992) during the last glacial maximum [i.e., 18,000 years BP (Prentice et al., 1994)]. The climate used for this simulation is derived from a GCM (not from paleoecological data, as this would lead to a circular argument). Note that according to this simulation, boreal forests cover only about 20% (Solomon et al., 1993) of today's potential Figure 1-5: Present (top) and future (bottom) potential natural vegetation according to the BIOME model (Prentice et al., 1992; Prentice et al., 1994). The present climate is given by the IIASA climate database (Leemans and Cramer, 1991). The projected shifts in the boundaries of the vegetation classes are due to climatic changes as projected by the difference between a GFDL GCM control run and a 2 x CO2 scenario (Wetherald and Manabe, 1986). They represent responses of plant functional types to cold tolerance, chilling requirements, and heat and Figure 1-6: Present (top) and future (bottom) equilibrium potential natural vegetation according to the MAPSS model (Neilson, 1993) The present climate is given by the IIASA climate database (Leemans and Cramer, 1991). The projected shifts in the boundaries of the vegetation classes are due to climatic changes as projected by the difference between a GFDL GCM control run and a 2 x CO2 scenario (Wetherald and Manabe, 1986). The model simulates climate responses due to a simulated steady-state leaf-area index, calculated from a site water and heat Climate Change Impacts on Forests gain in the potential area of forest distribution. Mainly IMAGETVM projects large net losses in the tropical rainforests. These losses are not due to climatic change, however, but are mainly caused by deforestation and other land-use changes (Zuidema et al., 1994). Although BIOME and MAPSS indicate relatively small net changes (Table 1-2, D*) in tropical dry forests, the actual changes encompass large gross losses (D), which are compensated by similarly large gross gains (not shown in Table 1-2). This is also partly true for the temperate zone, where MAPSS, for example, projects only relatively small net losses (D*); however, the associated gross losses (D) irrespective of gains (not shown) are more than six times larger. Mapped distributions indicate that temperate forests are likely to replace a large area of boreal forest, mainly due to the increase in winter temperatures (Figures 1-5 and 1-6). This allows temperate-zone vegetation to expand poleward into regions from which it is currently excluded by the -40°C threshold for the coldest month (Figures 1-5, 1-6, and 1-7). The boreal regions, especially, are expected to undergo large vegetation shifts, with both MAPSS and BIOME projecting large changing areas (Table 1-2, D). Both models show large losses in area for current boreal forests, despite their encroachment into current tundra. Shrinkage in total area due to the geographically limited poleward shift leads to a net loss between 379 Mha (25.9%) and 529 Mha (22.5%) 3 The IMAGE model is an exception in this case; it projects much smaller losses of only 33 Mha (2.9%) (Figures 1-5, 1-6, and 1-7). This is mainly due to the structure of the model, in which vegetation change is primarily driven by human lar.d use-which is not an important factor in the boreal zone. In summary, all models suggest that the world's forests are like ly to undergo major changes in the future, affecting more than a third of tropical dry (37.2%), temperate (35.7%), and boreal (40.4%) forests. Except in the temperate zone, the models suggest that there may be a net loss of forest area (Table 1-2, %D*). << Figure 1-7: Present (top), future (middle), and differing (bottom) potential natural vegetation influenced by human land use as generated by the terrestrial vegetation module (TVM; Leemans and van den Born, 1994) from the inicgrated climate change assessment model IMAGE 2.0 (Alcamo, 1994). The areas shown represent a projection starting with the year 1970 (top) and the internal, dynamic changes calculated by IMAGE 2.0 by the year 2050 (middle). The bottom graphic shows the new class for all areas that are predicted to change from one to another vegetation class. The simulation results generated by IMAGE 2.0 incorporate the effect of land-use changes (e.g.. deforestation) as driven by the dynamics of human populations and economic development. Note that IMAGE 2.0, unlike the other models, simulates climatic change independently from a GCM, since it generates its climate internally. The vegetational part (Leemans and van den Born, 1994) of the IMAGE model is similar to BIOME (Prentice et al., 1992) and represents a mixture between a transient and a static response of Averaged over all zones, this net loss amounts to 9.1% of the currently forested area. This is partly due to the fact that some regions are likely to lose forests for climatic reasons, while climatic gains in other regions might not be realized due to landuse pressures (IMAGE), and some models even predict a net loss (MAPSS). Averaged over all zones and all models, the projections indicate that 33% of the currently forested area is likely to change from the current vegetation class to a new one in response to climate forcing as generated by the GFDL 2 x CO2 scenario. Compared with other GCMs, the latter represents a medium global warming scenario. The change in annual mean temperature that occurs when one moves 100 km poleward may be as high as 0.7°C in mid- and high latitudes, but less at low latitudes. For summer temperatures and toward the interior of the continents, this value may be higher. With altitude, temperature changes of about 0.5-0.7°C per 100 m are also common (see also Chapter 5). With an expected warming between 0.1 and 0.35°C | per decade, this means that species would have to migrate 1.5-5.5 km toward the poles per year or increase elevation by 1.5-5.5 m per year in order to remain within similar climatic conditions. Many studies of past changes have estimated natural rates of migrations of trees ranging from 40 to 500 m per year (Davis, 1976, 1981, 1985; Huntley and Birks. 1983: Solomon et al., 1984; Gear and Huntley, 1991; Torrii, 1991). Similarly, Gear and Huntley (1991) calculated from several sites in Britain migration rates for Scots pine of only 40-80 m yr1. However, for other species, such as white spruce, much faster dispersal rates of up to 1-2 km yr1 also have been reported (e.g., Ritchie and MacDonald, 1986). It is not always clear whether the observed past rates were maximal rates of migration or whether they were limited by the rate at which the climate 3 The apparent lack of coincidence between relative and absolute changes is due to different definitions of the vegetation class "boreal forest." The MAPSS model has a greater area characterized as |