The species composition of forests in Europe is determined more by past management activities than by natural factors (Ellenberg, 1986). Because there currently is a tendency toward a more ecological approach to forest management (e.g., Innes, 1993; L�m�s and Fries, 1995), it will be difficult to separate the influence of changing climates from the influence of changing management practices. The trend toward tree species compositions in forests that more closely mimic those occurring naturally has been regarded as making the forests more adaptable to climate change. The evidence to support this hypothesis, however, is very limited (see, e.g., Fries et al., 1997). Current species combinations are based on today's conditions; future communities are likely to be composed of species assemblages that do not necessarily occur today (Lindner et al., 1997). The successful establishment today of species mixes that will be appropriate for future climatic conditions represents a major challenge for modern silviculture.
The range of European forests is limited primarily by climate, either through moisture availability or through temperature (both absolute amounts and seasonal distributions) (Berninger, 1997). A change in temperature or precipitation will affect the current distributions and productivity of forests (Bugmann, 1996; Lindner et al., 1996; Kellomaki et al., 1997). The response can be predicted with greater ease at the northern boundaries of forests in Fennoscandia and northern Russia, where an expansion of Norway spruce and Scotch pine into tundra regions is likely to occur under warmer conditions (Sykes and Prentice, 1996). These changes would be accompanied by a northward movement of the southern limit of these two species in Fennoscandia. Similar conclusions were reached by Kr�uchi (1995), whose model predictions indicated the replacement of Norway spruce by beech and other broad-leaved species at a site in northern Germany (Solling), associated with an increase in mean annual temperature of 0.3�C/decade until the end of the 21st century.
The rate of northward extension of the forest limit and individual species is highly uncertain because it depends not only on the rate of climate change but also on associated rates of dispersal (Malanson and Cairns, 1997), soil development, species composition (e.g., spruce/pine/beech) (Sykes and Prentice, 1996), and age of the trees. Such changes in the distribution of species will be highly individualistic (Huntley and Webb, 1989; Huntley, 1991)-with some species expanding their ranges, some showing little or no change, and others contracting their ranges (Sykes and Prentice, 1995, 1996), possibly to the point of extinction. Climate projections suggest a displacement of climatic zones suitable for forests by 150-550 km over the next century (IPCC 1996, WG II, Chapter 1). This shift is faster than the estimated potential of many species to migrate (20-200 km/century) (Davis, 1981; Birks, 1989) or the capability of many soils to develop a new structure.
In mountain regions, certain species and communities could disappear altogether because the upward displacement of species living close to the upper reaches of mountains will be constrained by the lack of any place in which they can become established (Kienast, 1991; Kienast et al., 1996). Modeling studies in Switzerland (Kr�uchi and Kienast, 1993) suggest that a temperature increase of 3�C would result in deciduous broad-leaved trees invading the subalpine belt and coniferous trees invading the alpine zone. Within the deciduous zone, colline Carpinion forests would expand at the expense of sub-montane and low-montane beech forests (Brzeziecki et al., 1995). The species composition of future forests would depend on changes in the continentality of each site. Such results have not yet been confirmed by empirical studies. H�ttenschwiler and K�rner (1995) found no indication of an upward movement in Scotch pine in the Swiss central Alps in response to summer temperatures during the period 1982-1991-when temperatures were, on average, 0.8�C warmer than those of the period 30 years before. They argued that the primary control on the altitudinal distribution of Scotch pine and Arolla pine (Pinus cembra L.) was more likely to be interspecific competition than temperature. However, some studies (e.g., Hofer, 1992; Grabherr et al., 1994) have noted an upward shift in the distribution of some alpine species.
Changes in the forests of southern Europe are most likely to be driven primarily by changes in water availability, although changes in temperature (particularly reductions in frost occurrence) also may play a role in the expansion of some forest types (e.g., Quercus pyrenaica and Quercus rotundifolia). Changes in precipitation will determine the relative importance of sclerophyllous and deciduous species; water availability in the period April-June would be particularly important (Gavil�n and Fern�ndez-Gonz�lez, 1997).
Many changes in forests may occur as a consequence of subtle alterations in the competitive balance between species. For example, rising temperatures are likely to change the time interval between budburst and leaf fall, but the effects will differ among species. Kramer (1995) argues that the duration of the growing season is likely to decrease with increasing temperature for European larch (Larix decidua Miller) and pedunculate oak (Quercus robur L.), whereas it will increase for beech (Fagus sylvatica L.) and small-leaved lime (Tilia cordata Miller). There already is some evidence that a substantial (25-30%) proportion of the forests in Switzerland is poorly adapted to the expected natural species distribution under current climatic conditions-and that this proportion will increase with an increase in temperature (Kienast et al., 1996). Similar patterns are likely elsewhere in Europe.
(continues on next page...)
Other reports in this collection