Figure 5-3c: Global CO2 emissions from fossil fuels and industry in the B1 scenario family (standardized). The marker scenario is shown with a thick line without ticks, the globally harmonized scenarios with thin lines, and the non-harmonized scenarios with thin, dotted lines (see Table 4-3).
The strong trend toward ecologically more compatible consumption and production patterns in the B1 family is reflected by structural changes toward less energy- and material-intensive activities, which lead to a partial de-coupling of welfare and energy demands. In the B1 marker scenario (B1-IMAGE; de Vries et al., 2000) the rapid technological change toward resource saving and ecologically sound solutions is assumed to spread very quickly, facilitated by high capital stock turnover rates in currently less developed regions. As a result, energy requirements in B1-IMAGE increase slowly and a shift away from fossil fuels eventually breaks the already slow upward trend in carbon emissions (Figure 5-3c). Emissions peak around 2040 at 12 GtC, twice the 1990 level, and by 2100 the emissions fall below the base-year level to 5 GtC. Total cumulative carbon emissions in the B1-IMAGE scenario amount to 983 GtC by 2100. As for A1, the population projection adopted for this scenario family declines after 2050.
Rates of energy-intensity improvement in the first half of the 21st century range quite widely in B1 family scenarios, and lead to emission levels from 8.5 (B1T-MESSAGE) to 17.5 (B1- ASF) GtC in 2050. By 2100 the gap in annual emissions narrows again, with final emissions between 3 and 8 GtC in all the B1 scenarios (except B1-High-MiniCAM).
The B1 family also includes one scenario in which energy-related emissions continue to increase throughout the modeling period, B1-High-MiniCAM. In this scenario the final energy demand is assumed to rise more rapidly with increasing income than in the rest of the B1 scenarios. (The total carbon emissions in B1High-MiniCAM decline slightly later in the modeling period because of reduced land under management and associated carbon sequestration.)
Total cumulative carbon emissions in the B1 scenario group ranges between 770 and 1390 GtC by 2100. All but two B1 scenarios fall in the low cumulative emissions category (Table 5-2).
Figure 5-3d: Global CO2 emissions from fossil fuels and industry in the B2 scenario family (standardized). The marker scenario is shown with a thick line without ticks, the globally harmonized scenarios with thin lines, and the non-harmonized scenarios with thin, dotted lines (see Table 4-3).
In the B2 world, dynamics of technological change continue along historical trends ("dynamics as usual"). The exploitation of comparative regional advantages in energy resources and technologies leads to regionally different mixes of clean fossil and non-fossil supply. With the continued growth of population and of income per capita, a steady increase of CO2 emissions emerges in the B2 marker (Riahi and Roehrl, 2000), developed using the MESSAGE model (B2-MESSAGE). By 2050 emissions reach 11 GtC and by 2100 they reach 14 GtC (Figure 5-3d). Total cumulative CO2 emissions in the B2 marker scenario amount to 1160 GtC by 2100.
Emissions in the B2 scenarios with harmonized global input assumptions (population, GDP, final energy; B2-MESSAGE, B2-AIM, B2-MARIA) are very close in 2100. Differences in emissions are largest around 2050, which reflects the different patterns of structural change in the energy systems in anticipation of depletion of conventional oil and gas. The emissions range for non-harmonized scenarios in the B2 scenario group is larger - between 11.2 (B2-IMAGE6) and 15.4 (B2-ASF) GtC by 2050. Relatively high emissions in the B2- ASF scenario are explained by a large share of coal in the fuel mix, because of high oil and gas prices. In this scenario, coal is also widely used for synthetic liquid and gaseous fuel production. In 2100, B2-IMAGE emissions (9 GtC) drop below the B2 marker level, while in B2-ASF emissions continue to grow and reach 19 GtC by 2100. The B2 high scenario variant developed using MiniCAM assumes far less efficiency gains, smaller available resources of oil and gas, and less favorable development of solar power costs than in the B2 marker. As in B2-ASF, the near exhaustion of oil supplies in B2High-MiniCAM leads to a heavy reliance on synthetic fuels to supply the needs of the transportation sector. Emissions in this scenario increase to 23 GtC by 2100. An additional alternative with more coal use (B2C-MARIA) was explored using the MARIA model. Emissions in this scenario are considerably higher than in the original B2-MARIA case, but are very close to those of B2-ASF (19 GtC).
Total cumulative carbon emissions across the B2 scenario group range between 1164 and 1686 GtC by 2100. Three of the eight B2 family scenarios fall into the "medium-high" cumulative emissions category, and the five others fall into the "medium-low" category (Table 5-2).
Table 5-2 suggests fairly strong contrasts in the level of cumulative emissions across the four scenario families (note that estimates in Table 5-2 also include emissions from land use). A variety of available energy supply and demand options means that cumulative emissions in the A1 family span the entire set of emission categories. The A2 family scenarios, characterized by a large population and a relatively carbon-intensive energy system, fall into either the "high" scenario category or in the upper part of the "medium-high" category. Only the A2-A1-MiniCAM scenario falls in the "medium-low" category, because of low cumulative emissions prior to 2050. As a consequence of the low energy consumption and non-fossil energy systems associated with its sustainable development theme, the B1 family scenarios are concentrated in the "low" emissions category. Finally, representatives of the B2 family are present in the two middle categories.
In addition to their cumulative emissions, the scenario families are characterized by very different emissions trajectories. The A1 fossil fuel scenarios have continuously increasing emissions, with rapid growth before 2050 and slower growth thereafter. The A1 scenarios with the "balanced" energy mix (e.g. the marker A1B-AIM) typically have emissions that decline after 2050, while the A1T technology scenarios have slower growth prior to 2050, and a steeper decline after 2050. As for the A1 fossil fuel scenarios, A2 family scenarios are characterized by high rates of growth in emissions prior to 2050 and subsequently continued growth but at lower rates. Unlike the rest of the A2 scenarios, the A2-A1-MiniCAM has more rapid growth in emissions after 2050 than before 2050, because per capita incomes in a number of developing regions do not reach a level at which per capita energy demands rise rapidly until the middle of the 21st century. The B1 family is characterized by lower growth in emissions prior to 2050, mostly because of lower rates of growth in energy demand, followed by declining emissions after 2050. Finally, the B2 family is characterized by relatively stable emissions post-2050, after roughly doubling emissions between 1990 and 2050.
Most changes in land use are induced by the demand for cropland and grassland, which is driven by the demand for food products, the extent of biomass energy use, and policies and practices associated with forest management. The 1990 land-use CO2 emissions remain fairly uncertain, estimated at 1.6 ± 1.0 GtC (Watson et al., 1996a); a similar level of uncertainty is attached to current land-use emissions. This uncertainty is reflected by the quantification of the SRES storylines - the 1990 emission estimates from different models range between 1.0 and 1.6 GtC, while the spread of estimates at the four-region level is even larger. For the sake of comparability, common, standardized (see Box 5-1 on standardization) emissions are established at 1.1 GtC in 1990 and 1.0 GtC in 2000, to reflect the net carbon flux from contemporary changes in forest cover.
Generally, the SRES models use different approaches to estimate land-use change emissions - in some cases the only source of emissions is tropical deforestation, while in other cases more sources and sinks (including natural) are included7 Moreover, methodologic differences and uncertainties in carbon content, carbon cycling, and land classification result in seemingly inconsistent results between models that cover the same land-use sources. These features complicate a straightforward comparison between land-use emissions in scenarios generated by different SRES models.
Figure 5-4: Standardized global CO2 emissions from land-use changes (net balance between anthropogenic changes in sources and sinks) for 40 SRES scenarios, classified into the four scenario families (each denoted by a different color code - A1, red; A2, brown; B1, green; B2, blue). Marker scenarios are shown with thick lines without ticks, globally harmonized scenarios with thin lines, and non-harmonized scenarios with thin, dotted lines (see Table 4-3). Black lines show percentiles, means, and medians for the 40 SRES scenarios. For numbers on the two additional illustrative scenarios A1FI and A1T see Appendix VII.
Future trajectories of land-use CO2 emissions are shown in Figure 5-4. Emissions in the 40 SRES scenarios range widely in the same year and change significantly over time. In scenarios with continued deforestation, emissions rise initially, then reach a maximum, and finally decline with depletion of forestland that can be cleared. At the other end of the scenario spectrum, emissions turn negative and land-use changes become an increasing CO2 sink through afforestation. By 2020, the resultant uncertainty ranges between 0 and 3 GtC across all 40 SRES scenarios with a median of 1.1 GtC. By 2050, the scenario range shifts to between -0.7 and 1.2 GtC (median, 0.5 GtC). By 2100 the scenario range lies between -2.8 and 2.2 GtC, with a median of 0.0 GtC. Interestingly, in specific years (e.g., in 2050) scenarios from all four SRES families fall within a relatively narrow emissions corridor (i.e. at least one scenario from each of the four SRES scenario families falls within the 25th and 75th percentiles of the emission range). This indicates that similar levels of carbon fluxes related to land use could arise from widely different socio-economic driving forces, depending on future trends in food demand and dietary patterns, agricultural productivity growth, forest practices, etc.
In general, the SRES CO2 land-use emissions follow the same pattern as found in the literature (see Chapter 3) - initially emissions increase because of continuing deforestation in developing regions and subsequently they decrease following a drop in population growth and increases in agricultural productivity. The main difference between the SRES scenarios and the literature reviewed in Chapter 3 lies in the maximum emission values, which are significantly lower in the SRES scenarios. Possibly this arises because the SRES models explicitly simulate land-use change as a function of pressure on the land (itself a function of agricultural productivity), while in the literature land-use CO2 emission scenarios are often based on trend extrapolation or statistical relationships (e.g. with population growth). The rapid economic development and technological advances assumed in the SRES storylines thus tend to mitigate CO2 emissions from deforestation, and in some cases lead to their reversal (e.g., turning deforested lands into carbon sinks).
As suggested by Figure 5-4, no simple relation exists between the scenario families and land-use emissions. As in the case of energy-related emissions, the A1 family scenarios cover the widest range of emissions and trajectories. In most cases, emissions in these scenarios decline after 2030 to zero or negative (carbon sinks) values. However, in two scenarios (A1B-IMAGE and A1B-MARIA), emissions increase later in the 21st century, reflecting assumptions on additional land pressures. As a rule, scenarios of the A2 family follow the same convex trajectory as A1 scenarios, also with two exceptions (A2-IMAGE and A2-AIM). In accordance with the sustainability emphasis of the B1 family storyline, scenarios from this family have the lowest initial emissions, which by 2080 or earlier drop to zero or negative levels. The B1 family marker scenario (B1-IMAGE) has negative land-use emissions for most of the modeling period, a property directly related to B1's "environmental conservation" emphasis.
As a result of different model specifications and detail, it is not possible to draw up consistent comparisons between sectoral emissions across different models. An overview of sectoral CO2 emissions by sector and source category is summarized in Box 5-2 on the basis of the results of the MESSAGE (for energy and industrial sources) and AIM (for land-use change sources) models.
Box 5-2: CO2 Emissions by Sector and Source for MESSAGE Scenarios
Table 5-4 gives an overview of CO2 emissions by sector and source category according to the IPCC reporting format given in Watson et al. (1996b). The differences in sectoral detail across models mean a consistent comparison and sectoral CO2 emission balances are only possible within one particular modeling framework. Table 5-4 presents the scenario results as calculated with the MESSAGE model for 1990, 2050, and 2100 and for the four scenario families and their scenario groups. Emissions related to land-use change were derived from consistent model runs with the AIM model.
As in Watson et al. (1996b), emissions are presented by sector, and emission
categories adopt both supply and demand perspectives for energy-related
CO2 emissions. The supply side CO2 balance accounts emissions at the point
of energy combustion, that is at a coal-fired power plant (electric generation)
or by burning coal in industrial boilers (direct fuel use by industry).
Conversely, the demand side CO2 balance accounts emissions per end-use
category, irrespective of whether emissions originate directly at the
point of end-use or upstream in the energy conversion sector. For example,
for residential and commercial energy uses, CO2 emissions include those
from direct fuel combustion as well as those emissions that originate
from the generation of electricity consumed by the residential and commercial
sectors. Finally, an emissions balance by source category is given, in
which emissions are accounted for at the level of primary energy (solids,
liquids, and gases), again after Watson et al. (1996b). Non-energy emissions
are included in a separate "others" category. Combined, these different
emission balances can serve as data input for subsequent mitigation analyses
at the sectoral level or at that of the entire economy.
Other reports in this collection