The main components of the hydrological cycle are precipitation, evaporation, discharge, storage in reservoirs, groundwater, and soil. Because of many complex interactions between these factors in time and in space, changes in the hydrological cycle are more difficult to model and to analyze than temperature and precipitation data. A detailed discussion is beyond the scope of this paper (see IPCC 1996, WG I, Chapter 3; IPCC 1996, WG II, Chapter 10).
Potential changes in the hydrological cycle are easier to understand if some basic mechanisms are kept in mind. Generally, higher temperatures lead to higher potential evaporation and decreased discharge (which also is a function of precipitation, storage, and topography). The storage of water in the soil serves as a buffer; in winter and spring, increasing precipitation normally generates higher discharge because the buffer is full and evaporation is low. During the summer, storage is reduced by evapotranspiration and must be refilled before discharge begins. Changes in the hydrological cycle therefore are more variable than changes in other climatic factors. Seasonal-to-interannual variability in precipitation and temperature also accounts for some of the variability in hydrological characteristics in European river basins.
The hydrological cycle in Europe is strongly influenced by anthropogenic factors. During the past century, land-use patterns have changed, the drainage conditions of rivers and land have been altered, the proportion of impermeable (urban) areas has increased, and additional reservoirs have been built. Changes or trends in the discharge of rivers therefore are more difficult to analyze than temperature and precipitation trends, especially in densely populated areas (Forch et al., 1996; Holt and Jones, 1996; Giakoumakis and Baloutsos, 1997).
GCM-based analyses for the European continent (IPCC 1996, WG II, Table 10-1) give a range of possible responses of river runoff in a warmer global climate-from decreases in some regions (e.g., Hungary, Greece) to increases in other regions (United Kingdom, Finland, Ukraine); these estimates are a function of precipitation, evapotranspiration, and soil moisture projections in the different GCMs. The uncertainties of climate model results, however, remain very large in terms of hydrological forecasting, particularly at the regional scale. This limitation is particularly critical for water management practices in the future because water resource impacts occur at the local scale, not at regional or larger scales.
The river basin is the natural spatial unit for water resource management; because neighboring river basins can have quite different climate change impacts, making quantitative statements about regional-scale climate impacts on water resources is difficult given that many regions include a wide range of hydroclimatic regions. The results of catchment-scale simulations with conventional hydrological models driven by GCM data therefore are highly variable. Arnell and Reynard (1996), for example, simulated changes of �20% in annual runoff for 21 catchments in Great Britain-with a tendency toward lower amounts of discharge, especially in sensitive areas and during the summer months.
Another consequence of increasing temperatures is a change in the distribution of water in the landscape. Especially in flat regions (e.g., lowlands in The Netherlands and northern Germany), catchment areas depend on groundwater recharge. Changes in percolation therefore can change the size of catchments (Hoermann et al., 1995). On a local catchment scale, distribution of water in the landscape can change even if annual discharge remains unchanged: Whereas hilltops are severely stressed by drought, areas with high groundwater levels may remain largely unaffected.
The effects on different regions of Europe can be classified according to the climatic gradients from north to south and from west to east (the maritime-to-continental gradient). According to current and projected distributions of rainfall, there may be an increase in summer drought. Observed precipitation shows a decrease in summer and an increase in winter and spring. Such a change is likely to affect mainly areas that already are sensitive to drought: the southern and continental parts of Europe, especially the Mediterranean region. Additional consumption of water for irrigation may create additional depletion of water reservoirs and groundwater.
Although the debate about changes in the frequency of floods is still open, an increase in rainfall during periods when soils are saturated (i.e., winter and spring), along with earlier snowmelt, could increase the frequency and severity in floods. An increase in large-scale precipitation might lead to increased flood risks on large river basins in western Europe in winter. The increased temperatures expected in summer could lead to higher local precipitation extremes and associated flood risks in small catchment areas.
Another unknown factor is the effect of increased plant water-use efficiency (WUE) resulting from increased CO2 in the atmosphere. Under laboratory and greenhouse conditions, plants seem to use less water for constant yield or exhibit higher productivity. In the field, however, predictions of WUE at the catchment or even ecosystem level are not yet feasible. In water-limited systems (e.g., in the Mediterranean region), a higher WUE can increase or maintain productivity-whereas in the humid areas of northern Europe, it could theoretically decrease evaporation and therefore increase discharge.
Warmer average and extreme temperatures will enhance the demand for freshwater, particularly for agriculture and direct human consumption, although the water-quality requirements are different for these two types of consumption. Changes in precipitation patterns, particularly over regions that already are sensitive (e.g., the Mediterranean basin), may lead to increased demand for water for irrigation purposes, especially for soils with low water-storage capacities. If precipitation were to undergo a decline in the Mediterranean basin (which is not always the case suggested by GCM simulations), countries such as Spain, Italy, and Greece would face substantially increased risks of summer water shortages. In such a situation, increases in storage capacity would be needed to maintain existing water and energy supplies. The Netherlands could face the desiccation of most of its wetland areas or be forced increasingly to rely on the Rhine to maintain present water levels. Groundwater aquifers could be affected by increased saltwater intrusion as sea level rises. Given a possible 4�C temperature increase and a rise in the Alpine snow line of 500-700 m in the summer, the flow of the Rhine could decline by 10% during this season. A temperature increase of only 0.5�C and a daily rainfall decrease of a few percent would decrease runoff in Hungary by 25-30%. Despite projected wetter winters, drier summers and increased evaporation in southeastern England may reduce yields in impounding reservoirs by 8-15%. Aquifer yields, which are very important to the London area, may fall by 8%. Supplies of water during warmer, possibly drier, summers would need to be maintained through larger storage or transfers from wetter regions. Water quality may deteriorate because there would be less river flow to dilute contaminants.
In terms of water management, system yields of river basins are likely to decrease as a result of changes in flow patterns; these river basins certainly will be sensitive to expected increases in floods and droughts. There may be a potential for conflict over environmental and economic uses of freshwater, especially for river basins that are shared by several countries. The high population density in western Europe does not allow for much flexibility for major changes in flood-protection practices. Floodplains in most countries already are overpopulated; thus, it will be extremely difficult to find the necessary areas for an effective flood-protection system.
Other reports in this collection