Figure 13.2: The cascade of uncertainties in projections to be considered in developing climate and related scenarios for climate change impact, adaptation and mitigation assessment.
The specific climate scenario needs of the impacts community vary, depending on the geographic region considered, the type of impact, and the purpose of the study. For example, distinctions can be made between scenario needs for research in climate scenario development and in the methods of conducting impact assessment (e.g., Woo, 1992; Mearns et al., 1997) and scenario needs for direct application in policy relevant impact and integrated assessments (e.g., Carter et al., 1996a; Smith et al., 1996; Hulme and Jenkins, 1998).
The types of climate variables needed for quantitative impacts studies vary widely (e.g., White, 1985). However, six "cardinal" variables can be identified as the most commonly requested: maximum and minimum temperature, precipitation, incident solar radiation, relative humidity, and wind speed. Nevertheless, this list is far from exhaustive. Other climate or climate-related variables of importance may include CO2 concentration, sea-ice extent, mean sea level pressure, sea level, and storm surge frequencies. A central issue regarding any climate variable of importance for impact assessment is determining at what spatial and temporal scales the variable in question can sensibly be provided, in comparison to the scales most desired by the impacts community. From an impacts perspective, it is usually desirable to have a fair amount of regional detail of future climate and to have a sense of how climate variability (from short to long time-scales) may change. But the need for this sort of detail is very much a function of the scale and purpose of the particular impact assessment. Moreover, the availability of the output from climate models and the advisability of using climate model results at particular scales, from the point of view of the climate modellers, ultimately determines what scales can and should be used.
Scenarios should also provide adequate quantitative measures of uncertainty. The sources of uncertainty are many, including the trajectory of greenhouse gas emissions in the future, their conversion into atmospheric concentrations, the range of responses of various climate models to a given radiative forcing and the method of constructing high resolution information from global climate model outputs (Pittock, 1995; see Figure 13.2). For many purposes, simply defining a single climate future is insufficient and unsatisfactory. Multiple climate scenarios that address at least one, or preferably several sources of uncertainty allow these uncertainties to be quantified and explicitly accounted for in impact assessments. Moreover, a further important requirement for impact assessments is to ensure consistency is achieved among various scenario components, such as between climate change, sea level rise and the concentration of actual (as opposed to equivalent) CO2 implied by a particular emissions scenario.
As mentioned above, climate scenarios that are developed for impacts applications usually require that some estimate of climate change be combined with baseline observational climate data, and the demand for more complete and sophisticated observational data sets of climate has grown in recent years. The important considerations for the baseline include the time period adopted as well as the spatial and temporal resolution of the baseline data.
Much of this chapter is devoted to assessing how and how successfully these needs and requirements are currently met.
Other reports in this collection