3.8 Key uncertainties and research priorities

There are major uncertainties in quantitative projections of changes in hydrological characteristics for a drainage basin. Precipitation, a principal input signal to water systems, is not reliably simulated in present climate models. However, it is well established that precipitation variability increases due to climate change, and projections of future temperatures, which affect snowmelt, are more consistent, such that useful conclusions are possible for snow-dominated basins.

Uncertainty has two implications. First, adaptation procedures need to be developed which do not rely on precise projections of changes in river discharge, groundwater, etc. Second, based on the studies completed so far, it is difficult to assess in a reliable way the water-related consequences of climate policies and emission pathways. Research on methods of adaptation in the face of these uncertainties is needed. Whereas it is difficult to make concrete projections, it is known that hydrological characteristics will change in the future. Water managers in some countries are already considering explicitly how to incorporate the potential effects of climate change into policies and specific designs.

Research into the water–climate interface is required:

• to improve understanding and estimation, in quantitative terms, of climate change impacts on freshwater resources and their management,
• to fulfil the pragmatic information needs of water managers who are responsible for adaptation.

Among the research issues related to the climate–water interface, developments are needed in the following.

• It is necessary to improve the understanding of sources of uncertainty in order to improve the credibility of projections.
• There is a scale mismatch between the large-scale climatic models and the catchment scale, which needs further resolution. Water is managed at the catchment scale and adaptation is local, while global climate models work on large spatial grids. Increasing the resolution of adequately validated regional climate models and statistical downscaling can produce information of more relevance to water management.
• Impacts of changes in climate variability need to be integrated into impact modelling efforts.
• Improvements in coupling climate models with the land-use change, including vegetation change and anthropogenic activity such as irrigation, are necessary.
• Climate change impacts on water quality are poorly understood. There is a strong need for enhancing research in this area, with particular reference to the impacts of extreme events, and covering the needs of both developed and developing countries.
• Relatively few results are available on the economic aspects of climate change impacts and adaptation options related to water resources, which are of great practical importance.
• Research into human-dimension indicators of climate change impacts on freshwater is in its infancy and vigorous expansion is necessary.
• Impacts of climate change on aquatic ecosystems (not only temperatures, but also altered flow regimes, water levels, and ice cover) are not adequately understood.
• Detection and attribution of observed changes in freshwater resources, with particular reference to characteristics of extremes, is a challenging research priority, and methods for attribution of causes of changes in water systems need refinement.
• There are challenges and opportunities posed by the advent of probabilistic climate change scenarios for water resources management.
• Despite its significance, groundwater has received little attention from climate change impact assessments, compared to surface water resources.
• Water resources management clearly impacts on many other policy areas (e.g., energy projections, nature conservation). Hence there is an opportunity to align adaptation measures across different sectors (Holman et al., 2005a, b). There is also a need to identify what additional tools are required to facilitate the appraisal of adaptation options across multiple water-dependent sectors.

Progress in research depends on improvements in data availability, calling for enhancement of monitoring endeavours worldwide, addressing the challenges posed by projected climate change to freshwater resources, and reversing the tendency of shrinking observation networks. Broadening access to available observation data is a prerequisite to improving understanding of the ongoing changes. Relatively short hydrometric records can underplay the full extent of natural variability and confound detection studies, while long-term river flow reconstruction can place recent trends and extremes in a broader context. Data on water use, water quality, and sediment transport are even less readily available.