Climate Change 2001:
Working Group II: Impacts, Adaptation and Vulnerability
Other reports in this collection

3.8.2. Implications of SRES Scenarios for Atmospheric Composition and Global Climate

Estimates of atmospheric composition resulting from the SRES emissions scenarios are presented in TAR WGI Chapters 3-5. Information on CO2 and ground-level O3 concentrations is given in Tables 3-2 and 3-9. More detailed regional estimates of pollutant concentrations and deposition of acidifying compounds based on these scenarios also are beginning to emerge (e.g., Mayerhofer et al., 2000; see Section 3.4).

To interpret the possible range of global temperature and sea-level response to the SRES scenarios, estimates have been made with simple models for all 35 of the quantified SRES scenarios (Table 3-9; see also TAR WGI Chapters 9 and 11). Estimates of global warming from 1990 to 2100 give a range of 1.4-5.8°C—somewhat higher than the 0.7-3.5°C of the SAR. The main reason for this increase is that the levels of radiative forcing in the SRES scenarios are higher than in the IS92a-f scenarios, primarily because of lower sulfate aerosol emissions, especially after 2050. The temperature response also is calculated differently; rather than using the conventional idealized, equilibrium climate sensitivity range of 1.5-4.5°C (IPCC, 1996a), the simple model is tuned to the effective climate sensitivities of a sample of individual AOGCMs (see TAR WGI Chapter 9 for details). Sea-level rise between 1990 and 2100 is estimated to be 9-88 cm, which also accounts for uncertainties in ice-melt parameters (see TAR WGI Chapter 11).

3.8.3. Implications of SRES Scenarios for Regional Mean Climate Regional Information from AOGCMs

Estimates of regional climate change to 2100 based on AOGCM experiments are described in TAR WGI Chapters 9 and 10. The results of nine AOGCMs run with the A2 and B2 SRES scenarios5 display many similarities with previous runs that assume IS92a-type emissions, although there also are some regional differences (see below). Overall, rates of warming are expected to be greater than the global average over most land areas and most pronounced at high latitudes in winter. As warming proceeds, northern hemisphere snow cover and sea-ice extent will be reduced. Models indicate warming below the global average in the North Atlantic and circumpolar southern ocean regions, as well as in southern and southeast Asia and southern South America in June-August. Globally there will be increases in average water vapor and precipitation. Regionally, December-February precipitation is expected to increase over the northern extratropics and Antarctica and over tropical Africa. Models also agree on a decrease in precipitation over Central America and little change in southeast Asia. Precipitation in June-August is projected to increase in high northern latitudes, Antarctica, and south Asia; change little in southeast Asia; and decrease in Central America, Australia, southern Africa, and the Mediterranean region.

The main differences between the SRES-based and IS92-based runs concern greater disagreement in the SRES runs on the magnitude of warming in some tropical and southern hemisphere regions and differing intermodel agreement on the magnitude of precipitation change in a few regions, possibly as a result of aerosol effects. However, there are no cases in which the SRES and IS92a results indicate precipitation changes of opposite direction (see TAR WGI Chapter 10). Regional Climate Characterizations

Only a limited number of AOGCM results based on the SRES emissions scenarios have been released and analyzed to date (i.e., results for the A2 and B2 scenarios), and none were available for the impact studies assessed in this report. In the meantime, alternative approaches have been used to gain an impression of possible regional changes in climate across a wider range of emissions scenarios. One method uses results from existing AOGCM simulations and scales the pattern of modeled regional climate change up or down according to the range of global temperature changes estimated by simple climate models for different emissions scenarios or assumptions about climate sensitivity (Santer et al., 1990; Mitchell et al., 1999; see also detailed discussion in TAR WGI Chapter 13). This "pattern-scaling" method has been employed by Carter et al. (2000), using results from simulations with seven AOGCMs, all assuming a radiative forcing approximating the IS92a emissions scenario (for GHGs but excluding aerosols) scaled across a range of global temperature changes estimated by using a simple climate model for the four preliminary marker SRES emissions scenarios.

Regional-scale summary graphs of scaled temperature and precipitation changes were constructed for 32 world regions, at subcontinental scale, chosen to represent the regions being assessed by Working Group II (Carter et al., 2000). Examples of individual plots are shown in Figure 3-2. Changes are plotted alongside estimates of "natural" multi-decadal variability of temperature and precipitation, extracted from two multi-century unforced AOGCM simulations. The graphs thus provide a quick assessment of the likely uncertainty range and significance of each AOGCM projection; they also show the extent to which different AOGCMs agree or disagree with regard to regional response to a given magnitude of global warming. Although a preliminary comparison of these results with SRES AOGCM runs (which also include aerosol forcing) suggests broad agreement on regional temperature and precipitation changes, more rigorous comparison remains to be carried out, offering a useful test of the pattern-scaling method.

Other reports in this collection