Anthropogenic CH4 emissions arise from a variety of activities, dominated by biologic processes, each associated with considerable uncertainty. The future CH4 emissions in the scenarios depend in part on the consumption of fossil fuels, adjusted for assumed changes in technology and operational practices, but more strongly on scenario-specific, regional demographic and affluence developments, together with assumptions on preferred diets and agricultural practices. The writing team recommends further research into the sources and modeling approaches to capture large uncertainties surrounding future CH4 emissions.
The resultant CH4 emission trajectories for the four SRES scenario families portray complex patterns (as displayed in Figure 5-5 in Chapter 5). For example, the emissions in A2 and B2 marker scenarios increase throughout the whole time horizon to the year 2100. Increases are most pronounced in the high population A2 scenarios where emissions rise to between 549 and 1069 (A2 marker: 900) MtCH4 by 2100, compared to 310 MtCH4 in 1990. The emissions range by 2100 in the B2 scenarios is between 465 and 613 (B2 marker: 600) MtCH4 . In the A1B and B1 marker scenarios, the CH4 emissions level off and subsequently decline sooner or later in the 21 st century. This phenomenon is most pronounced in the A1B marker, in which the fastest growth in the first few decades is followed by the steepest decline; the 2100 level ends up slightly below the current emission of 310 MtCH4 . The range of emissions in Table TS-4 indicates that alternative developments in energy technologies and resources could yield a higher range in CH4 emissions compared to the "balanced" technology A1B scenario group that includes the A1B marker scenario discussed above. In the fossil-intensive A1FI group (combined from A1C and A1G groups, as in the SPM), CH4 emissions could reach some 735 MtCH4 by 2100, whereas in the post-fossil A1T scenario group emissions are correspondingly lower (some 300 MtCH4 by 2100). Interestingly, the A1 scenarios generally have comparatively low CH4 emissions from non-energy sources because of a combination of low population growth and rapid advances in agricultural productivity. Hence the SRES scenarios extend the uncertainty range of the IS92 scenario series somewhat toward lower emissions. However, both scenario sets indicate an upper bound of emissions of some 1000 MtCH4 by 2100.
Even more than for CH4 , the assumed future food supply will be a key determinant of future N2O emissions. Size, age structure, and regional spread of the global population will be reflected in the emission trajectories, together with assumptions on diets and improvements in agricultural practices. Other things being equal, N2O emissions are generally highest in the high population scenario family A2. Importantly, as the largest anthropogenic source of N2O (cultivated soils) is already very uncertain in the base year, all future emission trajectories are affected by large uncertainties, especially if calculated with different models as is the case in this SRES report. Therefore, the writing team recommends further research into the sources and modeling of long-term N2O emissions. Uncertainty ranges are correspondingly large, and are sometimes asymmetric. For example, while the range in 2100 reported in all A1 scenarios is between 5 and 10 MtN (7 MtN in the A1B marker), the A2 marker reports 17 MtN in 2100. Other A2 scenarios report emissions that fall within the range reported for A1 (from 8 to 19 MtN in 2100). Thus, different model representations of processes that lead to N2O emissions and uncertainties in source strength can outweigh easily any underlying differences between individual scenarios in terms of population growth, economic development, etc. Different assumptions with respect to future crop productivity, agricultural practices, and associated emission factors, especially in the very populous regions of the world, explain the very different global emission levels even for otherwise shared main scenario drivers. Hence, the SRES scenarios extend the uncertainty range of future emissions significantly toward higher emissions (4.8 to 20.2 MtN by 2100 in SRES compared to 5.4 to 10.8 MtN in the IS92 scenarios. (Note that natural sources are excluded in this comparison.)
The emissions of halocarbons (chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs), halons, methylbromide, and hydrofluorocarbons (HFCs)) and other halogenated compounds (polyfluorocarbons (PFCs) and sulfur hexafluoride (SF6 )) across the SRES scenarios are described in detail on a substance-by-substance basis in Chapter 5 and Fenhann (2000). However, none of the six SRES models has its own projections for emissions of ozone depleting substances (ODSs), their detailed driving forces, and their substitutes. Hence, a different approach for scenario generation was adopted.
First, for ODSs, an external scenario, the Montreal Protocol scenario (A3, maximum allowed production) from WMO/UNEP (1998) is used as direct input to SRES. In this scenario corresponding emissions decline to zero by 2100 as a result of international environmental agreements, a development not yet anticipated in some of the IS92 scenarios (Pepper et al., 1992). For the other gas species, most notably for CFC and HCFC substitutes, a simple methodology of developing different emission trajectories consistent with aggregate SRES scenario driving force assumptions (population, GDP, etc.) was developed. Scenarios are further differentiated as to assumed future technological change and control rates for these gases, varied across the scenarios consistently with the interpretation of the SRES storylines presented in Chapter 4 as well as the most recent literature.
Second, different assumptions about CFC applications as well as substitute candidates were developed. These were initially based on Kroeze and Reijnders (1992) and information given in Midgley and McCulloch (1999), but updated with the most recent information from the Joint IPCC/TEAP Expert Meeting on Options for the Limitation of Emissions of HFCs and PFCs (WMO/UNEP, 1999). An important assumption, on the basis of the latest information from the industry, is that relatively few Montreal gases will be replaced fully by HFCs. Current indications are that substitution rates of CFCs by HFCs will be less than 50% (McCulloch and Midgley, 1998). In Fenhann (2000) a further technological development is assumed that would result in about 25% of the CFCs ultimately being substituted by HFCs (see Table 5-9 in Chapter 5). This low percentage not only reflects the introduction of non-HFC substitutes, but also the notion that smaller amounts of halocarbons will be used in many applications when changing to HFCs (efficiency gains with technological change). A general assumption is that the present trend, not to substitute with high GWP substances (including PFCs and SF6 ) , will continue. As a result of this assumption, the emissions reported here may be underestimates. This substitution approach is used in all four scenarios, and the technological options adopted are those known at present. Further substitution away from HFCs is assumed to require a climate policy and is therefore not considered in SRES scenarios. The range of emissions of HFCs in the SRES scenario is initially generally lower than in earlier IPCC scenarios because of new insights about the availability of alternatives to HFCs as replacements for substances controlled by the Montreal Protocol. In two of the four scenarios in the report, HFC emissions increase rapidly in the second half of the next century, while in two others the growth of emissions is significantly slowed down or reversed in that period.
Aggregating all the different halocarbons (CFCs, HCFCs, HFCs) as well as halogenated compounds (PFCs and SF6) into MtC-equivalents (using GWPs from IPCC SAR, notwithstanding the caveats given in footnote 6) indicates a range between 386 and 1096 MtC-equivalent by 2100 for the SRES scenarios. This compares with a range of 746 to 875 MtC-equivalent for IS92 (which, however, does not include PFCs and SF6). (The comparable SRES range excluding PFCs and SF6 is between 299 and 753 MtC-equivalent by 2100.) The scenarios presented here indicate a wider range of uncertainty compared to IS92, particularly toward lower emissions (because of the technological and substitution reasons discussed above).
The effect on climate of each of the substances aggregated to MtC-equivalents given in Table TS-4 varies greatly, because of differences in both atmospheric lifetime and the radiative effect per molecule of each gas. The net effect on climate of these substances is best determined by a calculation of their radiative forcing - which is the amount by which these gases enhance the anthropogenic greenhouse effect. The radiative forcing will be addressed in IPCC TAR and is thus not discussed in this report.
Figure TS-10: Global anthropogenic SO2 emissions (MtS/yr) - historical development from 1930 to 1990 and (standardized) in the SRES scenarios. The dashed colored time-paths depict individual SRES scenarios, the solid colored lines the four marker scenarios, the solid thin curves the six IS92 scenarios, the shaded areas the range of 81 scenarios from the literature, the gray shaded area the sulfur-control and the blue shaded area the range of sulfur-non-control scenarios or "non-classified" scenarios from the literature that exceeds the range of sulfur control scenarios. The colored vertical bars indicate the range of the SRES scenario families in 2100. For details of the two additional illustrative A1 scenarios see Appendix VII. Database source: Gr�bler (1998).
Emissions of sulfur portray even more dynamic patterns in time and space than the CO2 emissions shown in Figures TS-7 and TS-8. Factors other than climate change (namely regional and local air quality, and transformations in the structure of the energy system and end use) intervene to limit future emissions. Figure TS-10 shows the range of global sulfur emissions for all SRES scenarios and the four markers against the emissions range of the IS92 scenarios, more than 80 scenarios from the literature, and the historical development.
A detailed review of long-term global and regional sulfur emission scenarios is given in Gr�bler (1998) and summarized in Chapter 3. The most important new finding from the scenario literature is recognition of the significant adverse impacts of sulfur emissions on human health, food production, and ecosystems. As a result, scenarios published since 1995 generally assume various degrees of sulfur controls to be implemented in the future, and thus have projections substantially lower than previous ones, including the IS92 scenario series. Of these, only the two low-demand scenarios IS92c and IS92d fall within the range of more recent long-term sulfur emission scenarios. A related reason for lower sulfur emission projections is the recent tightening of sulfur-control policies in the OECD countries, such as the Amendments of the Clean Air Act in the USA and the implementation of the Second European Sulfur Protocol. Such legislative changes were not reflected in previous long-term emission scenarios, as noted in Alcamo et al. (1995) and Houghton et al. (1995). Similar sulfur control initiatives due to local air quality concerns are beginning to impact sulfur emissions also in a number of developing countries in Asia and Latin America (see IEA, 1999; La Rovere and Americano, 1998; Streets and Waldhoff, 2000; for a more detailed discussion see Chapter 3). As a result, even the highest range of recent sulfur-control scenarios is significantly below that of comparable, high-demand IS92 scenarios (IS92a, IS92b, IS92e, and IS92f). The scenarios with the lowest ranges project stringent sulfur-control levels that lead to a substantial decline in long-term emissions and a return to emission levels that prevailed at the beginnings of the 20 th century. The SRES scenario set brackets global anthropogenic sulfur emissions of between 27 and 169 MtS by 2050 and between 11 and 93 MtS by 2100 (see Table TS-4). In contrast, the range of the IS92 scenarios (Pepper et al., 1992) is substantially higher starting at 80 MtS and extending all the way to 200 MtS by 2050 and from 55 to 230 MtS by 2100.
Reflecting recent developments and the literature, it is assumed that sulfur emissions in the SRES scenarios will also be controlled increasingly outside the OECD. As a result, both long-term trends and regional patterns of sulfur emissions evolve differently from carbon emissions in the SRES scenarios. As a general pattern, global sulfur emissions do not rise substantially, and eventually decline even in absolute terms during the second half of the 21st century, as indicated by the median of all scenarios in Figure TS-10 (see also Chapters 2 and 3). The spatial distribution of emissions changes markedly.
Emissions in the OECD countries continue their recent declining trend (reflecting the tightening of control measures). Emissions outside the OECD rise initially, most notably in Asia, which compensates for the declining OECD emissions. Over the long term, however, sulfur emissions decline throughout the world, but the timing and magnitude vary across the scenarios. It should be noted that SRES scenarios assume sulfur controls only and do not assume any additional climate policy measures. Nevertheless, one important implication of this varying pattern of sulfur emissions is that the historically important, but uncertain, negative radiative forcing of sulfate aerosols may decline in the very long run. This view is also confirmed by the model calculations reported in Subak et al. (1997) and Nakicenovic et al. (1998) based on recent long-term GHG and sulfur emission scenarios.
Other reports in this collection