Climate Change 2001: Working Group II: Impacts, Adaptation and Vulnerability

	Other reports in this collection

3.4.6. Water Resource Scenarios

3.4.6.1. Reference Conditions

Water is a resource of fundamental importance for basic human survival, for ecosystems, and for many key economic activities, including agriculture, power generation, and various industries. The quantity and quality of water must be considered in assessing present-day and future resources. In many parts of the world, water already is a scarce resource, and this situation seems certain to worsen as demand increases and water quality deteriorates, even in the absence of climate change. Abundance of the resource at a given location can be quantified by water availability, which is a function of local supply, inflow, consumption, and population. The quality of water resources can be described by a range of indicators, including organic/fecal pollution, nutrients, heavy metals, pesticides, suspended sediments, total dissolved salts, dissolved oxygen, and pH.

Several recent global analyses of water resources have been published (Raskin et al., 1997; Gleick, 1998; Shiklomanov, 1998; Alcamo et al., 2000). Some estimates are shown in Table 3-3. For regional and local impact studies, reference conditions can be more difficult to specify because of large temporal variability in the levels of lakes, rivers, and groundwater and human interventions (e.g., flow regulation and impoundment, land-use changes, water abstraction, effluent return, and river diversions; Arnell et al., 1996).

Industrial wastes, urban sewage discharge, application of chemicals in agriculture, atmospheric deposition of pollutants, and salinization negatively affect the quality of surface and groundwaters. Problems are especially acute in newly industrialized countries (UNEP/GEMS, 1995). Fecal pollution of freshwater basins as a result of untreated sewage seriously threatens human health in some regions. Overall, 26% of the population (more than 1 billion people) in developing countries still do not have access to safe drinking water, and 66% do not have adequate environmental sanitation facilities—contributing to almost 15,000 deaths each day from water-related diseases, nearly two-thirds of which are diarrheal (WHO, 1995; Gleick, 1998; see Chapter 9).

Table 3-4: The role of various types of climate scenarios and an evaluation of their advantages and disadvantages according to the five criteria described in the text. Note that in some applications, a combination of methods may be used—for example, regional modeling and a weather generator (WGI TAR Chapter 13, Table 13.1).
Scenario Type or Tool		Description/Use	Advantagesa	Disadvantagesa
Incremental		Testing system sensitivity Identifying key climate thresholds	Easy to design and apply (5) Allows impact response surfaces to be created (3)	Potential for creating unrealistic scenarios (1,2) Not directly related to GHG forcing (1)
Analog
Palaeoclimatic		Characterizing warmer periods in past	Physically plausible changed climate that really did occur in the past of a magnitude similar to that predicted for ~2100 (2)	Variables may be poorly resolved in space and time (3,5) Not related to GHG forcing (1)
Instrumental		Exploring vulnerabilities and some adaptive capacities	Physically realistic changes (2) Can contain a rich mixture of well-resolved, internally consistent, variables (3) Data readily available (5)	Not necessarily related to GHG forcing (1) Magnitude of climate change usually quite small (1) No appropriate analogs may be available (5)
Spatial		Extrapolating climate/ecosystem relationships Pedagogic	May contain a rich mixture of well-resolved variables (3)	Not related to GHG forcing (1,4) Often physically implausible (2) No appropriate analogs may be available (5)
Climate Model-Based
Direct   AOGCM   outputs		Starting point for most climate scenarios Large-scale response to anthropogenic forcing	Information derived from the most comprehensive, physically based models (1,2) Long integrations (1) Data readily available (5) Many variables (potentially) available (3)	Spatial information poorly resolved (3) Daily characteristics may be unrealistic except for very large regions (3) Computationally expensive to derive multiple scenarios (4,5) L a rge control run biases may be a concern for use in certain regions (2)
High-   resolution/   stretched grid   (AGCM)		Providing high-resolution information at global/continental scales	Provides highly resolved information (3) Information derived from physically based models (2) Many variables available (3) Globally consistent and allows for feedbacks (1,2)	Computationally expensive to derive multiple scenarios (4,5) Problems in maintaining viable parameterizations across scales (1,2) High resolution dependent on SSTs and sea ice margins from driving model (AOGCM) (2) Dependent on (usually biased) inputs from driving AOGCM (2)
Regional   models		Providing high spatial/temporal resolution information	Provides very highly resolved information (spatial and temporal) (3) Information derived from physically based models (2) Many variables available (3) Better representation of some weather extremes than in GCMs (2,4)	Computationally expensive, thus few multiple scenarios (4,5) Lack of two-way nesting may raise concern regarding completeness (2) Dependent on (usually biased) inputs from driving AOGCM (2)
Climate Model-Based (cont.)
Statistical   downscaling		Providing point/ high spatial resolution information	Can generate information on high-resolution grids or nonuniform regions (3) Potential, for some techniques, to address a diverse range of variables (3) Variables are (probably) internally consistent (2) Computationally (relatively) inexpensive (5) Suitable for locations with limited computational resources (5) Rapid application to multiple GCMs (4)	Assumes constancy of empirical relationships in the future (1,2) Demands access to daily observational surface and/or upper air data that span range of variability (5) Not many variables produced for some techniques (3,5) Dependent on (usually biased) inputs from driving AOGCM (2)
Climate   scenario   generators		Integrated assessments Exploring uncertainties Pedagogic	May allow for sequential quantification of uncertainty (4) Provides "integrated" scenarios (1) Multiple scenarios easy to derive (4)	Usually rely on linear pattern-scaling methods (1) Poor representation of temporal variability (3) Low spatial resolution (3)
Weather Generators		Generating baseline climate time series Altering higher order moments of climate Statistical downscaling	Generates long sequences of daily or subdaily climate (2,3) Variables usually are internally consistent (2) Can incorporate altered frequency/intensity of ENSO events (3)	Poor representation of low-frequency climate variability (2,4) Limited representation of extremes (2,3,4) Requires access to long observational weather series (5) In absence of conditioning, assumes constant statistical characteristics (1,2)
Expert Judgment		Exploring probability and risk Integrating current thinking on changes in climate	May allow for "consensus" (4) Has potential to integrate very broad range of relevant information (1,3,4) Uncertainties can be readily represented (4)	Subjectivity may introduce bias (2) Representative survey of experts may be difficult to implement (5)
a Numbers in parentheses within the Advantages and Disadvantages columns indicate that they are relevant to the criteria described. The five criteria follow: 1) Consistency at regional level with global projections; 2) physical plausibility and realism, such that changes in different climatic variables are mutually consistent and credible and spatial and temporal patterns of change are realistic; 3) appropriateness of information for impact assessments (i.e., resolution, time horizon, variables); 4) representativeness of potential range of future regional climate change; and 5) accessibility for use in impact assessments.