IPCC Working Group III (WGIII) appointed the SRES writing team in January 1997. After some adjustments, it eventually came to include more than 50 members from 18 countries. Together they represent a broad range of scientific disciplines, regional backgrounds, and non-governmental organizations. In particular, the team includes representatives of six scenario modeling groups and a number of lead authors from all three earlier IPCC scenario activities: the 1990 and 1992 scenarios and the 1995 scenario evaluation. Their expertise and familiarity with earlier IPCC emissions scenario work assured continuity and allowed the SRES effort to build efficiently upon prior work. The SRES team worked in close collaboration with colleagues on the IPCC Task Group on Climate Scenarios for Impact Assessment (TGCIA) and with colleagues from all three IPCC Working Groups (WGs) of the Third Assessment Report (TAR). Appendix II lists the members of the writing team and their affiliations and Chapter 1 gives a more detailed description of the SRES approach and process.
Taking the above audiences and purposes into account, the following more precise specifications for the new SRES scenarios were developed. The new scenarios should:
The writing team agreed that the scenario formulation process would consist of five major components:
As is evident from the components of the work program, there was agreement that the process be an open one with no single "official" model and no exclusive "expert teams." In 1997 the IPCC advertised in a number of relevant scientific journals and other publications to solicit wide participation in the process. To facilitate participation and improve the usefulness of the new scenarios, the SRES web site (http://sres.ciesin.org/) was created. In addition, members of the writing team published much of the background work used for formulating SRES scenarios in the peer-reviewed literature 3 and on web sites (see Appendix IV). Finally, the revised set of scenarios, the web sites, and the draft of this report have been evaluated through the IPCC expert and government review processes. This process resulted in numerous changes and revisions of the report. In particular, during the approval process of the Summary for Policymakers (SPM) in March 2000 at the 5th Session of the WG III in Katmandu changes in this SPM were agreed that necessitated some changes in the underlying document, including this Technical Summary. These changes have been implemented in agreement with the Lead Authors.
The first step in the formulation of the SRES scenarios was the review and the analysis of the published literature and the development of the database with more than 400 emissions scenarios that is accessible through the web site (www-cger.nies.go.jp/cger-e/db/ipcc.html); 190 of these extend to 2100 and are considered in the comparison with the SRES scenarios in the subsequent Figures. Chapters 2 and 3 give a more detailed description of the literature review and analysis.
Figure TS-1: Global energy-related and industrial CO2 emissions - historical development and future scenarios, shown as an index (1990 = 1). The median (50th), the 5th, and 95th percentiles of the frequency distribution are also shown. The statistics associated with scenarios from the literature do not imply probability of occurrence (e.g., the frequency distribution of the scenarios may be influenced by the use of IS92a as a reference for many subsequent studies). The emissions paths indicate a wide range of future emissions. The range is also large in the base year 1990 and is indicated by an "error" bar. To separate the variation due to base-year specification from different future paths, emissions are indexed for the year 1990, when actual global energy-related and industrial CO2 emissions were about 6 GtC. The coverage of CO2 emissions sources may vary across the 256 different scenarios from the database included in the figure. The scenario samples used vary across the time steps (for 1990 256 scenarios, for 2020 and 2030 247, for 2050 211, and for 2100 190 scenarios). Also shown, as vertical bars on the right of the figure, are the ranges of emissions in 2100 of IS92 scenarios and for scenarios from the literature that apparently include additional climate initiatives (designated as "intervention" scenarios emissions range), those that do not ("non-intervention"), and those that cannot be assigned to either of these two categories ("non-classified"). This classification is based on the subjective evaluation of the scenarios in the database by the members of the writing team and is explained in Chapter 2. Data sources: Morita and Lee, 1998a, 1998b; Nakic�enovic� et al., 1998.
Figure TS-1 shows the global energy-related and industrial CO2 emission paths from the database as "spaghetti" curves for the period to 2100 against the background of the historical emissions from 1900 to 1990. These curves are plotted against an index on the vertical axis rather than as absolute values because of the large differences and discrepancies for the values assumed for the base year 1990. These sometimes arise from genuine differences among the scenarios (e.g., different data sources, definitions) and sometimes from different base years assumed in the analysis or from alternative calibrations.4 The differences among the scenarios in the specification of the base year illustrate the large genuine scientific and data uncertainty that surrounds emissions and their main driving forces captured in the scenarios. The literature includes scenarios with additional climate polices, which are sometimes referred to as mitigation or intervention scenarios.
There are many ambiguities associated with the classification of emissions scenarios into those that include additional climate initiatives and those that do not. Many cannot be classified in this way on the basis of the information available from the database. Figure TS-1 indicates the ranges of emissions in 2100 from scenarios that apparently include additional climate initiatives (designated as "intervention" emissions range), those that do not ("non-intervention") and those that cannot be assigned to either of these two categories ("non-classified"). This classification is based on the subjective evaluation of the scenarios in the database by the members of the writing team and is explained in Chapter 2. The range of the whole sample of scenarios has significant overlap with the range of those that cannot be classified and they share virtually the same median (15.7 and 15.2 GtC in 2100, respectively), but the non-classified scenarios do not cover the high part of the range. Also, the range of the scenarios that apparently do not include climate polices (non-intervention) has considerable overlap with the other two ranges (the lower bound of non-intervention scenarios is higher than the lower bounds of the intervention and non-classified scenarios), but with a significantly higher median (of 21.3 GtC in 2100).
Historically, gross anthropogenic CO2 emissions have increased at an average rate of about 1.7% per year since 1900 (Nakicenovic et al., 1996); if that historical trend continues global emissions would double during the next three to four decades and increase more than sixfold by 2100. Many scenarios in the database describe such a development. However, the range is very large around this historical trend so that the highest scenarios envisage about a tenfold increase of global emissions by 2100 as compared with 1990, while the lowest have emissions lower than today. The median and the average of the scenarios lead to about a threefold emissions increase over the same time period or to about 16 GtC by 2100. This is lower than the median of the IS92 set and is lower than the IS92a scenario, often considered as the "central" scenario with respect to some of its tendencies. However, the distribution of emissions is asymmetric. The thin emissions "tail" that extends above the 95th percentile (i.e., between the six- and tenfold increase of emissions by 2100 compared to 1990) includes only a few scenarios. The range of other emissions and the main scenario driving forces (such as population growth, economic development and energy production, conversion and end use) for the scenarios documented in the database is also large and comparable to the variation of CO2 emissions. Statistics associated with scenarios from the literature do not imply probability of occurrence or likelihood of the scenarios. The frequency distribution of the database may be influenced by the use of IS92a as a reference for scenario studies and by the fact that many scenarios in the database share common assumptions prescribed for the purpose of model comparisons with similar scenario driving forces.
One of the recommendations of the writing team is that IPCC or a similar international institution should maintain such a database thereby ensuring continuity of knowledge and scientific progress in any future assessments of GHG scenarios. An equivalent database for documenting narrative and other qualitative scenarios is considered to be also very useful for future climate-change assessments. One difficulty encountered in the analysis of the emissions scenarios is that the distinction between climate policies and non-climate policy scenarios and other scenarios appeared to be to a degree arbitrary and often impossible to make. Therefore, the writing team recommends that an effort should be made in the future to develop an appropriate emissions scenario classification scheme.
Other reports in this collection