Footnotes for Chapter 4

1. During the approval process of the Summary for Policymakers at the 5th Session of WGIII of the IPCC from 8-11 March 2000 in Katmandu, Nepal, it was decided to combine two of thiese groups ( A1C and A1G) into one "fossil intensive" group A1FI, in contrast to the non-fossil group A1T, and to select two illustrative scenarios from these two A1 groups to facilitate use by modelers and policy makers. This leads to six scenario groups that constitute the four scenario families, three of which are in the A1 family. All scenarios are equally sound.

2. The harmonization criteria agreed by the writing team are indicated in Table 4-1. The classification of scenarios is quite robust against varying the percentage deviation harmonization criteria (see Section 4.4.1).

3. Deviations within each 10-year time period are not considered.

4. Additionally, three scenarios (A2-AIM, A2-MiniCAM, and B2- MiniCAM) deviate only slightly from the global harmonization criteria for between two to three time steps. Hence these scenarios can be considered as 'almost' harmonized and comparable with the other harmonized scenarios.

5. The coal and gas/oil intensive groups were merged into one fossil-intensive group in the Summary for Policymakers. More detailed information on these two groups is presented here, in Chapter 5 and in Appendix VII.

6. Different modeling teams provided different interpretations of what such a "balanced" resource-technology portfolio could be in the 21 st century. The assumed rapid technology dynamics that underlie the A1 scenario storyline necessitate that technologies and resource exploitation profiles change significantly. Hence, the concept of "balanced" development does not necessarily apply to any particular future date, but rather to the entirety of the scenario's development path throughout the 21 st century. As a consequence of the slow turnover rates in the capital stock of the energy sector, all scenarios necessarily rely more heavily on currently dominant resources and technologies in the near-term and project more radical departures only in the long-term.

7. Scenarios denoted by italics indicate scenario quantifications that share harmonized global population and GDP trajectories with the respective marker scenario (i.e. "globally harmonized" scenarios).

8. By design, the A1T scenario group explores the possibility of technological change in energy end-use technologies and hence lower energy demand compared to the A1 marker scenario. Considering that these scenarios provide similar levels of energy service, albeit at lower levels of final energy use, they are classified as "harmonized" based on population and GDP profiles only.

9. The scenario deviates only slightly from the global population and GDP assumptions of other "harmonized" scenarios within this scenario family.

10. The IMAGE results for the A2 and B2 scenarios are based on preliminary model experiments carried out in March 1998. As a result of limited resources it has not been possible to redo these experiments. Hence, the IMAGE team is not able to provide background data and details for these scenario calculations and the population and economic growth assumptions are not fully harmonized, as is the case for the IMAGE A1 and B1 scenarios.

11. The scenario deviates only slightly from the global population and GDP assumptions of other "harmonized" scenarios within this scenario family.

12. The B2-ASF shares global population and GDP assumptions with the B2 marker by 2100, but explores different dynamics of growth in the intervening time period.

13. The scenario deviates only slightly from the global population and GDP assumptions of other "harmonized" scenarios within this scenario family.

14. The dynamic profiles of land-use changes mean that scenario comparisons for any given year, such as 2100, are somewhat misleading. Hence, cumulative carbon emissions that result from land-use changes over the 1990 to 2100 period are used as a proxy indicator in Figure 4-4 (see Chapter 5 for a more detailed discussion).

15. To address this issue, during the approval process of the Summary for Policymakers at the 5th Session of WGIII of the IPCC from 8-11 March 2000 in Katmandu, Nepal, it was decided to select two illustrative scenarios in the SPM from two additional A1 groups, in addition to the marker scenarios.

16. The different population projection that underlies the "delayed development" scenario A2-A1-MiniCAM is described in Box 4-6.

17. The SRES writing team gratefully acknowledges the assistance of Anne Goujon of IIASA's Population Project in providing a numeric interpretation of the IIASA low population scenario that is consistent with the assumption of convergence of social and economic development underlying the B1 and A1 scenario storylines. Numeric scenario values and documentation is given in www.iiasa.ac.at/Research/POP/IPCC-special- report/.

18. Scenarios A1v1-MiniCAM and A1v2-MiniCAM adopt a slightly different population projection, generated endogenously in the MiniCAM model. Its main differences are an asymptotic fertility rate of 1.75 compared to the 1.5 of Lutz et al. (1996) and a slightly different temporal pattern of the demographic transition. The resultant demographic projection is about 5% lower in the first half of the 21 st century and about 10% higher toward 2100 compared to those of the other A1 family scenarios. As such, the demographic scenario is well within the uncertainty range characteristic of any long-term demographic projection.

19. For a detailed description and the scenario's underlying assumptions and numeric values see www.iiasa.ac.at/Research/POP/IPCC-special-report.

20. It was noted in the government review process that A2's population projection by 2050 is 11.3 billion people, higher than the highest UN projection (10.7 billion) published in the report The State of World Population 1999 (UNFPA, 1999). As this UN projection extends only to 2050, it is necessary to consider the corresponding UN long-range population projections that extend to 2150 (UN, 1998). In these, the corresponding high/medium scenario is projected to have 10.8 billion people by 2050, which increases to 14.6 billion by 2100. The UN presents two additional scenarios that result in yet higher population levels: the UN high scenario projects 11.2 billion people by 2050 and 17.5 billion by 2100, and the UN constant fertility scenario projects 14.9 and 57.2 billion people by 2050 and 2100, respectively. For comparison, the A2 population scenario adopted from Lutz et al. (1996) indicates population levels of 11.3 billion by 2050 and 15.1 billion by 2100. Thus, A2 population levels by 2050 are comparable to those of the UN high scenario but remain significantly below the UN constant fertility scenario. By 2100, A2's population assumption is comparable to that of the UN high/medium scenario (14.6 billion) and significantly below that of the other two UN scenarios (high and constant fertility, with 17.5 and 57.2 billion, respectively). Hence, the adopted values for the A2 scenario are well within the range of the UN long-range population projections. As discussed in Chapter 3, the reason for the comparatively higher 2050 population in the Lutz et al. (1996) scenario is lower assumed mortality rates compared to the UN projections in the medium/high scenario. The lower mortality rates characteristic of A2's demographic scenario are judged to be more consistent with the A2 scenario storyline than alternative UN projections, such as the UN high/medium variant, that results in lower population by 2050, albeit at the expense of higher mortality. Yet by 2100 the differences in global population between the two scenarios is rather small (15.1 versus 14.6 billion, or 3%).

21. The SRES writing team gratefully acknowledges the assistance of Thomas B�ttner of the UN Population Division, New York, in developing more detailed regional population projections based on the UN 1998 medium projection, and in making these data available to the SRES writing team in electronic form.

22. The A1v2-MinCAM scenario reaches only about 64% of the A1B marker global GDP in 2100.

23. In B2-IMAGE, economic output is higher in OECD90 and lower in the other regions compared to the B2 marker.

24. In particular, the consistency of a continued fast demographic transition to 2050 combined with a scenario of stagnating per capita income growth for as much as five decades is questioned by a number of members from the writing team.

25. Note that this statement only indicates the relative position of the A2 scenario compared to other SRES scenario families. In absolute terms the scenario's decline in energy intensity is very substantial - on average, energy use per unit of GDP declines by a factor of more than two as a result of the compounding effect of an improvement rate of final energy intensity of 0.8% per year. Comparison of this improvement rate with the SRES scenario range calculated by the ASF model indicates that A2's energy intensity improvement rates are one-third lower compared to the B2 scenario and less than half compared to the B1 scenario. By 2100, A2's final energy intensity is calculated by the ASF model at 5.9 MJ/$, which compares to the literature range of up to 7 MJ/$, and a value of 7.3 MJ/$ in the A2-A1- MiniCAM scenario, which contains the highest energy-intensity trajectory within the 40 SRES scenarios. Thus, the A2 scenario's energy-intensity improvement rates are well within the uncertainty range as indicated by the scenario literature and are not considered overly pessimistic by the writing team. During the government review process, comment was made on the fact that energy intensities in A2 are one-third higher than those in B2. This figure is classified as "reasonable" for an inter-family scenario variation by the writing team because it is consistent both with the underlying differences in per capita GDP (i.e. productivity) growth between the two scenario families and with the relationship between energy intensity improvements and macro-economic productivity growth identified in the literature assessment in Chapters 2 and 3.

26. 1 mill is 0.1 USCents (US$0.001).

27. c.i.f., cost, insurance, freight (included in price); f.o.b., free on board (i.e. insurance and transport costs not included in fuel price delivered "free on board" transport vessel only). These different cost-accounting methods for international energy trade are particular important for transport and infrastructure intensive fuels such as natural gas.

28. Corresponding data were not available for all scenarios developed with other models, and hence a detailed comparison across the entire SRES scenario set was not possible.

29. Even with this "conservative" assumption cumulative oil extraction in the A2 marker scenario totals 16 ZJ, or 2.7 times currently identified, recoverable oil reserves (6 ZJ or 143.3 billion tons; BP, 1999). Thus, the A2 scenario also assumes that in future it will be possible to continue the historical trend in which large quantities of (undiscovered or presently uneconomic) oil resources are transferred into recoverable reserves. Some analysts consider such a future trend as definitely optimistic (see the literature review in Chapter 3).

30. Resource availability assumptions also appear to be rather model specific in this scenario family. For instance, in many scenarios patterns of resource availability resemble the hypotheses retained by a particular model used for quantification of a marker scenario in one of the other three scenario families.

31. Roehrl and Riahi (2000) provide a description of the methodology of representing technological change in MESSAGE as used here in the SRES scenarios.

32. Cumulative carbon emissions (all sources) are 1359 GtC for B2- MARIA and 1573 GtC for B2C-MARIA (see Chapter 5).

33. Adopting a common accounting convention avoids misrepresentation of the contribution of renewable and other new energy forms, which can be both under- or over-represented by inconsistent accounting conventions, as continues to be the case in energy statistics and scenario studies.

34. Nonetheless, cumulative oil extraction to 2100 equals 2.7 times the currently identified reserves of conventional oil in the scenario.

35. Except B1-MARIA and B1High-MiniCAM.

36. For comparison, the corresponding productivity growth rates in the other scenarios range from 1% per year (B2-AIM) to 2% per year (B1- AIM), consistent with their respective storylines. This emphasizes productivity and efficiency (B1) and more fragmented technology and productivity growth (B2). For A2-AIM, crop productivity growth rates range between 1% and 1.5 % per year in the DEV and IND regions, respectively; the difference is explained by only slowly closing productivity gaps (approximated by GDP/capita), characteristic of this scenario storyline. Similar differences also characterize other salient scenario assumptions of importance to land-use changes (like biomass yields, animal productivity, or the distribution of grain- versus range-fed cattle). For instance, feed and protein yields from pasture land are assumed to grow at 1.5 % per year and biomass yields at 0.5 % per year in the A1 scenario.