Europe has few genuine natural ecosystems. Natural ecosystems generally are confined to small areas; agriculture and forestry occupy most soils. Many sites with intermediate soils are occupied by seminatural vegetation types.
Vegetation responds to climate change directly and indirectly. Direct effects include responses to temperature; indirect effects occur primarily as soil-mediated phenomena, such as the influence of precipitation on soil moisture regimes. Indirect effects also may occur as a result of the responses of herbivores and pests to climate changes, changes in soil fauna, and changes in the frequency and severity of disturbances such as fire. In addition, vegetation responds directly and indirectly to atmospheric CO2 concentrations. The responses are species-dependent; no two taxa respond to climatic change in exactly the same way (e.g., Stirling et al., 1997). The main response of individual taxa to climatic change consists of changes in distribution; adaptive evolutionary changes are very rare (Huntley, 1991). At the ecosystem level, the impact of CO2 on ecosystem processes remains very uncertain. Zaller and Arnone (1997) have documented an increase in surface-casting earthworm activity in Swiss grasslands exposed to 610 �l CO2/l, but such studies are extremely rare.
In a warmer climate, the pattern of species response will be extremely complex within Europe (Grime, 1996) because a variety of temperature and moisture gradients exist. Although there is a general temperature gradient from south to north and from low to high altitudes, east-west gradients in temperature and precipitation also exist; the latter are associated with increasing continentality toward central Europe. Further complications arise in the prediction of species responses to temperature changes because of the importance of the nature of the change. For example, a rise in late-summer temperatures will have different impacts than a rise in early-spring temperatures or an average rise in temperatures spread evenly throughout the year (Fitter et al., 1995). Similarly, the occurrence of late frost may play an important role in restricting the responsiveness of small-genome species to mean temperature changes (MacGillivray and Grime, 1995). This factor makes the prediction of species responses difficult; a general northward shift in species distributions is now recognized as too simplistic a hypothesis. Successful migration depends on a number of factors-in particular, the range of tolerance of a given plant or tree to heat and moisture stress, environmental conditions at the new location, the rate of migration, the presence of competing species, and natural and human barriers to migration (Thompson, 1994; Malanson and Cairns, 1997). Anthropogenic barriers are especially important for large portions of western and central Europe, where land use is dominated by direct human intervention.
Short-term experiments suggest that many types of plants will respond positively to increases in CO2 concentrations in the atmosphere (the so-called CO2 fertilization effect), whereby their photosynthetic rates increase if other factors remain constant; this is particularly the case for C3 plant types and all important European crops except maize (Semenov et al., 1996; Wolf et al., 1996). Most experiments have been undertaken using isolated plants with optimum nutrient supply. Such experimental conditions are relevant to horticultural and agro-industrial situations but are inapplicable to natural and seminatural plant communities (K�rner, 1996). A wide range of vegetation types may show little or no response to increasing CO2 concentrations under field conditions (K�rner, 1996). The responses, however, will be species-specific, especially when other factors such as enhanced nitrogen deposition are taken into account (H�ttenschwiler and K�rner, 1996a). The net primary productivity of other plants may increase (provided that they are not water limited), but there are many uncertainties regarding the long-term responses of plants to increased CO2; studies around natural sources of CO2 have not revealed any gradients in the growth rates or biomass of Mediterranean grassland species (K�rner and Miglietta, 1994). Studies of trees growing around natural sources of CO2 revealed no changes in stomatal density, but the guard cells were reduced in size (Miglietta and Raschi, 1993). Downy oak (Quercus pubescens Willd.) growing close to a CO2 source had lower stomatal conductance than those further away, but Holm oak (Quercus ilex L.) showed no such trend (Tognetti et al., 1996). In both species, the osmotic potential and apoplasmic fraction of water was elevated close to the CO2, indicating that these trees were more tolerant of drought conditions (Chaves et al., 1995; Tognetti et al., 1996).
There is increasing evidence that traits other than photosynthetic metabolism are more important in determining the response to elevated CO2 of different species under field conditions (e.g., K�rner, 1993; K�rner et al., 1995; Diaz, 1995, 1996). For example, increased levels of CO2 are likely to result in increased water-use efficiency in many species. Increased water-use efficiency may help many plants and trees resist the extremes of heat and drought that may occur more frequently in southern Europe and the Mediterranean region.
Studies of model ecosystems exposed to enhanced CO2 and nitrogen deposition suggest the presence of nonlinear system-level adjustments (H�ttenschwiler and K�rner, 1996b). These adjustments include physiological down-regulation of photosynthesis at the leaf level, reduced leaf area index, and increasing strength of below-ground carbon sinks. At the same time, no aboveground growth stimulation was observed. Changes were observed between CO2 concentrations of 280 and 420 �l/l; major changes in coniferous forest ecosystems may be underway already in response to increasing CO2 concentrations.
The impact of climate change on biodiversity and the composition of ecosystems in Europe is extremely difficult to predict. A great deal depends on the impacts on ecosystem processes, such as the rates and magnitudes of disturbance. The resilience of many ecosystems to change also is very uncertain; many Norway spruce forests, for example, are likely to persist for several hundred years in the absence of any major disturbance (Sykes and Prentice, 1996). This inertia, along with the possibility of species acclimation to changed environmental conditions (Kellom�ki and Wang, 1996), may delay the onset of many changes in natural ecosystems (Woodward, 1993). Models that currently are available for predicting such changes generally are restricted to the most important components of vegetation (e.g., trees in forests); little research has been done on possible future interactions between these and other ecosystem components under changed climatic and CO2 conditions. The models suggest that in some cases, species arrivals may compensate for species losses, whereas in others, human-induced changes in land use may delay the arrival of new species, causing a reduction in diversity. Consequently, there is a need to combine models that simulate changes in species distributions with models developed to look at species turnover at specific sites (compare van der Maarel and Sykes, 1993; �kland, 1995a,b; Fr�borg and Eriksson, 1997).
Other reports in this collection