The purpose of this section is to provide a summary of major methodological issues in the science of climate scenario development and to relate these developments to applications of scenarios in this report. We distinguish between a climate scenario, which refers to a plausible future climate, and a climate change scenario, which implies the difference between some plausible future climate and the present-day climate, though the terms are used interchangeably in the scientific literature. This brief overview is distilled largely from material presented in TAR WGI Chapter 13. See that chapter, as well as TAR WGI Chapters 8, 9, and 10, for more complete coverage of this subject.
Methods of climate scenario development largely have been ignored in earlier IPCC assessments, although some aspects of scenario development have been alluded to (e.g., palaeoclimatic analogs in Folland et al., 1990; downscaling methods in Kattenburg et al., 1996). Table 3-4 provides an overview of the main methods, which also are discussed in TAR WGI Chapter 13. Thus, we present only a very brief summary of three major methods. A fourth method, expert judgement, that also has been used in developing climate scenarios (NDU, 1978; Morgan and Keith, 1995), is discussed further in Section 3.5.5.
In this approach, particular climatic (or related) elements are changed by realistic but arbitrary amounts. They are commonly applied to study the sensitivity of an exposure unit to a wide range of variations in climate and to construct impact response surfaces over multivariate climate space. Most studies have adopted incremental scenarios of constant changes throughout the year (e.g., Terjung et al., 1984; Rosenzweig et al., 1996), but some have introduced seasonal and spatial variations in the changes (e.g., Rosenthal et al., 1995); others have examined arbitrary changes in interannual, within-month, and diurnal variability, as well as changes in the mean (e.g., Williams et al., 1988; Mearns et al., 1992, 1996; Semenov and Porter, 1995). Some of these studies are discussed in Chapter 5.
Temporal and spatial analogs also have been used in constructing climate scenarios. Temporal analogs make use of climatic information from the past as an analog of possible future climate (Pittock, 1993). They are of two types: palaeoclimatic analogs and instrumentally based analogs.
Palaeoclimatic analogs: Palaeoclimatic analogs are based on reconstructions
of past climate from fossil evidence, such as plant or animal remains and sedimentary
deposits. Two periods have received particular attention: the mid-Holocene (~5-6
ky BP3), when northern hemisphere temperatures are
estimated to have been about 1°C warmer than today, and the Last (Eemian)
Interglacial (~120-130 ky BP), when temperatures were about 2°C warmer.
The major disadvantages of this method are the causal differences between past changes in climate and posited future changes (Crowley, 1990; Mitchell, 1990) and the large uncertainties about the quality of palaeoclimatic reconstructions (Covey, 1995; Kneshgi and Lapenis, 1996; Borzenkova, 1998). However, these scenarios continue to be used occasionally in impact assessments (Anisimov and Nelson, 1996; Budyko and Menzhulin, 1996) and are useful for providing insights about system vulnerability to climate change.
Instrumentally based analogs: Periods of observed regional or global-scale warmth during the historical period also have been used as an analog of a GHG-induced warmer world. Scenarios are constructed by estimating the difference between the regional climate during the warm period and that of the long-term average or that of a similarly selected cold period (Lough et al., 1983; Rosenberg et al., 1993). Major objections to the use of these analogs include the relatively minor changes in climate involved (although small changes could be adequate for examining near-term climate change) and, again, differences between the causes of historical fluctuations and those of posited larger future climate changes (Glantz, 1988; Pittock, 1989).
Spatial analogs: These are regions that today have a climate analogous to that anticipated in the study region in future. For example, Bergthórsson et al. (1988) used temperatures in northern Britain as a spatial analog for the potential future temperatures over Iceland. The approach is severely restricted, however, by the frequent lack of correspondence between other important features (climatic and nonclimatic) of the two regions. Nevertheless, spatial analogs are still adopted in a few studiesfor example, to assess potential effects of climate change on human health (see Chapter 9).
Other reports in this collection