Chapter 3 highlights the importance of technological change in long-run productivity
growth, but also for the historical transformations of energy end-use and supply
systems. The importance of technological change in explaining wide-ranging outcomes
in future emissions has been highlighted by Alcamo et al., (1995) and Gr�bler
and Nakicenovic (1996), among others. The latter reference also provides a
critical assessment of the previous IS92 scenario series and its comparison
to the literature. Prominent scenario studies of possible technological change
in future energy systems in the absence of climate policies include Ausubel
et al. (1988), Edmonds et al. (1994, 1996a), IIASA-WEC (1995), and Nakicenovic
et al. (1998). Future technology characteristics must therefore be treated as
dynamic, with future improvement rates subject to considerable uncertainty.
This is reflected in the SRES scenarios that adopt a wide range of improvement
rates for energy extraction, conversion, and end-use technologies (Table 4-11).
The actual representation of technological change in the six SRES models ranges
from exogenously prescribed availability, through cost and performance profiles
(which in some cases also include consumer or end-use costs for technology use),
to stylized representation of learning processes31.
Yet, as summarized in Chapter 3, model representations
of technological change are poorly developed, although evolving rapidly.
|Table 4-11: Summary of technology improvements for extraction, distribution, and conversion technologies assumed for the SRES scenarios. The classification reviews technology dynamics across the four marker scenarios and the four A1 scenario groups relative to each other. Illustrative, scenario-specific technology assumptions are discussed in the text. A1C and A1G have been combined into one fossil-intensive group A1FI in the SPM (see also footnote 1).|
Technology Improvement Rates
|A1G||Low||Very high||Very high||Medium|
A. Technology improvement rates in the A2 scenario are heterogeneous among the world regions.
B. B1: The assumed time-dependent learning coefficients range from 0.9 (i.e. a 10% reduction in the capital:output ratio on a doubling of cumulated production) for oil, 0.9-0.95 for gas, and 0.9-0.95 for surface coal mining to about 0.94-0.96 for non-fossil electric power generation options and 0.9-0.95 for commercial biofuels.
C. In the specific model implementations, "inconvenience costs" of energy-end use, including social externalities costs, are expected to be particularly important for traditional coal technologies (e.g., underground mining, cooking with coal stoves).
The A1B marker scenario represents the "balanced" technology development group of A1 scenarios; it assumes significant innovations in energy technologies, which improve energy efficiency and reduce the cost of energy supply. Consistent with the A1 scenario storyline, such improvements occur across the board and neither favor nor penalize particular groups of technologies. A1 assumes, in particular, drastic reductions in power-generation costs, through the use of solar, wind, and other modern renewable energies, and significant progress in gas exploration, production, and traansport. For a different view, alternative scenario groups embedded within the overall A1 scenario family explore pathways of cumulative technological change; that is, path-dependent scenarios in which technologies evolve on mutually largely exclusive development paths. In general this has been the historical experience, in which the success of particular energy technologies (the steam engine in the 19th century, or internal combustion in the 20th) have "locked out" other technological alternatives. These scenario groups explore alternative spectra of technology dynamics in the domains of unconventional oil and gas, coal, as well as post-fossil technologies. Salient technology assumptions are described below.
Keeping in mind the very different degrees of technological detail and the mechanisms for technology improvements represented in the different models, a consistent inter-scenario comparison of technology assumptions is best achieved within the framework of one particular model. An overview of different technology developments for the scenario groups of the A1 scenario is given in Box 4-8 for the AIM model, which was also used to develop the A1B marker scenario. (A comparison with the MARIA model indicated that technology cost assumptions and their dynamics are quite congruent.) To illustrate differences in technology characteristics that drive the four different SRES scenario families, corresponding scenario-specific data based on MESSAGE data are presented at the end of this Section.
As outlined above, besides the marker, three different groups of A1 scenarios were developed by the different modeling groups (combined into two in the SPM, see also footnote 1 ). In total, nine alternative runs are clustered in three scenario groups based on the AIM, MARIA, MESSAGE, and MiniCAM models.
In the A1G scenario group, technological change enables a larger fraction of
the large occurrences of unconventional oil and gas, including oil shales, tar
sands, and especially methane hydrates (clathrates) to be tapped. High technological
learning and cost reduction effects could lower unconventional oil and gas extraction
costs by approximately 1% per year and conversion technology costs by about
factor of two (A1G-MESSAGE, see Roehrl and Riahi, 2000). As mentioned in Section
4.4.6, although these assumptions yield higher extractions of unconventional
oil and gas resources, they are not sufficient to tap significant fractions
of unconventional resources such as gas clathrates. Future scenario studies
might reassess the current state of knowledge on possible technology development
of these "exotic" fossil-fuel occurrences and the conditions under which they
could become a major future source of unconventional hydrocarbon supply (and
a massive source of carbon emissions). For the A1G scenario group, substantial
improvements and extensions of the present pipeline grids and entirely new natural
gas pipelines systems from Siberia and the Caspian to South East Asia, China,
Korea, and Japan after 2010/2020 would be needed. Since unconventional oil and
gas resources are distributed unevenly geographically, the scenario implies
both capital-intensive infrastructure investments and unprecedented large-scale
gas and oil trade flows. There is also little pressure to develop non-fossil
alternatives in such scenarios, so costs of non-fossil alternatives remain comparatively
high, even after significant technological improvements. For instance, solar
electricity costs could drop to US$0.05 per kWh (A1G-AIM).
Box 4-8: Technological Change in the AIM-based Quantifications for the A1 Scenario Family
The A1 storyline describes a world with rapid economic development. High economic growth results in pressures on resource availability, counterbalanced by technological progress, which is assumed to be highest among the four scenario families. In the AIM quantifications of the A1 storyline, rates of technological change are high both with respect to "supply push" factors (most notably RD&D) as well as with respect to "demand pull" factors (most notably high capital turnover rates). Since large resource availability and high incomes stimulate demand growth, technological change in energy supply receives a higher emphasis compared to changes in energy end-use technologies. Common technology assumptions in the A1 scenarios can be summarized as follows.
The supply of oil, gas, and biomass in the A1 scenario family is assumed to be very high and results from high rates of technological progress for fossil fuel and biomass exploitation technologies. Unconventional oil and gas, such as deep-sea methane hydrates, oil shale, etc., become available at relatively low cost. Also, large amounts of biomass are utilized through well-developed biomass farm plantations and harvest technologies, and biomass utilization technologies, such as biomass power generation and biofuel conversion technologies, become available at low costs through RD&D and other mechanisms of technology improvements (learning by doing and learning by using). High levels in the use of other renewable energy are reached when technologies for solar photovoltaics and thermal utilization, wind farms, geothermal energy utilization, and ocean energy are introduced at low cost. Energy end-use technologies are assumed to progress at medium rates compared with the fast rates of technological change in energy supply technologies.
The A1B marker and A1T scenarios assume drastic reductions in cost for
solar, wind, and other renewable energies. A1C assumes lower coal costs
and emphasizes coal exploitation technology progress and the introduction
of advanced coal-fired power generation technology, such as integrated
gasification combined cycle (IGCC). A1G assumes lower oil and gas costs
than other A1 scenarios. The cost of nuclear power is assumed to be the
lowest in A1G and A1T, and highest in A1C. The different cost assumptions
that drive and result from technological change in the A1 scenario family
are summarized in Table 4-12.
Technology progress is also assumed for land-use changes and sulfur emissions. Higher productivity increases in biomass and crop land (1.5% per year) in comparison to 0.5-1.0%) are assumed for the A1 world in the AIM quantification compared to those in the A2 and B2 scenario families. Desulfurization technologies could be introduced because of concerns of economic damage caused by acid rain and there would be strong financial support to install these technologies with the rapid income growth associated with the A1 world.
The high-growth coal-intensive scenario group A1C assumes relatively large cost improvements in new and clean coal technologies, such as coal high-temperature fuel cells, IGCC power plants, and coal liquefaction. More modest assumptions are made for all the other technologies, except for nuclear technologies in A1C-MESSAGE, as this requires zero-carbon options to ease resource and environmental constraints. The relative costs between coal and oil- or gas-related technologies also shift in A1C-AIM. Progress in renewables is also assumed to be substantial. For instance, solar photovoltaic costs would decline to USCents3/kWh (A1C-AIM).
In the dynamic technology scenario group A1T, technological change, driven by market mechanisms and policies to promote innovation, favors non-fossil technologies and synfuels, especially hydrogen from non-fossil sources. Liquid fuels from coal, unconventional oil and gas sources, and renewables become available at less than US$30 per barrel, with costs that fall further, by about 1% per year, through exploitation of learning-curve effects (A1T-MESSAGE). A1T-MARIA also projects declining costs for biofuels, from about US$30 to US$20, after the 2020 period (and in comparison to the A1- MARIA scenario biofuels substitute coal-derived synfuels). Non-fossil electricity (e.g., photovoltaics) begin massive market penetration at costs of about USCents1 to 3 per kWh (A1T-MARIA, A1T-MESSAGE, A1T-AIM), and could continue to improve further (perhaps as low as USCents0.1/kWh in A1T-MESSAGE) as a result of learning-curve effects. An important difference between the marker scenario A1B and the A1T group is that in A1T additional end-use efficiency improvements are assumed to take place with the diffusion of new end-use devices for decentralized production of electricity (fuel cells, microturbines). As a result, final energy demand in the A1T scenario group is between 30% (A1T-AIM, A1T-MESSAGE) and 40% (A1T-MARIA) lower compared to the A1B marker scenario.
The A2 scenario family includes slow improvements in the energy supply efficiency and a relatively slow convergence of end-use energy efficiency in the industrial, commercial, residential, and transportation sectors between regions. A combination of slow technological progress, more limited environmental concerns, and low land availability because of high population growth means that the energy needs of the A2 world are satisfied primarily by fossil (mostly coal) and nuclear energy. However, in some cases regional energy shortages force investments into renewable alternatives, such as solar and biomass. For instance, intermittent renewable electricity supply options, such as solar and wind, are assumed to decline in costs to about USCents4/kWh and (because of storage requirements) to about twice that value when these intermittent sources are used for medium load applications (50% of electricity supply).
Consistent with the general environmentally conscious and resource-conservation thrust of the B1 scenario storyline, technological change is largely directed at improving conversion efficiency rather than costs for fossil technologies. Within the SRES Terms of Reference, no additional climate initiatives are assumed that could bar the application of certain technologies or yield forced diffusion of others. The thermal efficiency of centrally generated electricity is assumed to rise to 45% (conventional coal) or to 65% (gas combined cycles) by 2100, while specific investment costs decline slightly from 1990 levels. It is assumed that subsidies on coal for electricity generation are removed entirely. A specific feature of the IMAGE model used to generate the B1 marker scenario is that it treats non-fossil electricity generation technologies as highly generic; for instance, it does not distinguish between nuclear, solar, or wind-power generation technologies. The specific investment costs of generation options for non-fossil electricity and of the production and conversion of commercial biofuels are assumed to fall by 5-10% for every doubling of cumulated production. Cost decreases down to USCents2.5/kWh are anticipated once non-fossil options penetrate on a large scale. The costs of gaseous biofuels in the major producing regions (Latin America, Africa, NIS) are assumed to be in the order of US$3 to 5 per GJ from 2020 to 2030 onward. Liquid biofuels are produced in small amounts in almost all regions at costs in the order of US$3 to 6 per GJ. In all regions a gradual transition occurs from fossil fuels to non-fossil options in electric-power generation, because of rising fuel prices and declining specific investment costs for fossil alternatives. Learning rates were assumed, conservatively, to yield 2 to 6% cost reductions for every doubling of cumulative production. The shift would start in resource-poor industrialized regions such as Japan and Western Europe, but is somewhat tempered by rising conversion efficiencies of fossil-fueled power plants. One of the factors that constrains the use of natural gas in the scenario is the assumption that only a limited part of the transport market is open to competition from non-liquid fuels (between 50% around 2050 to 80% around 2100). Also, the market share of coal in industry is fixed exogenously at 10 to 15% in some regions, to reflect the decreasing environmental and social attractiveness of the more "dirty" coal.
The approach that underlies the B2 scenario storyline translates into important future improvements of technologies, albeit at more conservative rates than in scenarios A1 or B1, but with higher rates than in scenario A2. Compared to A1 and B1, cost improvements are more modest, because of the regionally fragmented technology policies assumed to characterize a B2 world. Hence, technology-spillover effects and benefits from shared development expenditures are more limited in the scenario. The high emphasis of environmental protection at the local and regional levels is reflected in faster development and diffusion of energy technologies with lower emissions, including advanced coal technologies, nuclear, and renewables. For instance, solar and wind electricity-generating costs are assumed to decline to USCents3/kWh, that is, a similar level as assumed for the long-term costs of advanced, clean coal technologies (such as IGCCs). As conventional oil supplies dwindle, initially high-cost synfuels from coal and also biofuels are introduced as substitutes. With increasing production volume, costs are assumed to decline from initial levels of some US$7/GJ to US$2.6/GJ. Conventional coal technologies undergo the lowest aggregate rates of improvement in the scenario and are also subject to increasing controls of social and environmental externalities (mining safety, particulates, and sulfur emissions). Increasingly, therefore, only advanced coal technologies are deployed. Nonetheless, extraction and conversion costs increase, especially in regions with a large share of deep-mined coal and in high population density agglomerations. In regions with abundant surface minable coal reserves (e.g., North America and Australia), coal extraction costs remain relatively low.
As a consequence of the "multi-model approach" used in SRES, detailed improvement assumptions and scenario implementations for individual technologies vary greatly from one model to another, although the same storyline characteristics were used as guiding principles and many scenarios share similar assumptions on improvement potentials for different technologies. Detailed quantitative comparisons are difficult because of different time profiles of technology improvements assumed in the different models, different representations of regional technology, and the modeling of the international diffusion of technology. For instance, many models assume aggregate regional rates of technological change (e.g., MARIA, MiniCAM, ASF), whereas others attempt to represent spatial and temporal diffusion patterns more explicitly (e.g., MESSAGE, AIM).
It is difficult to quantify the influence of varying technology-specific scenario assumptions on scenario outcomes, because in most model simulations the technology assumptions were varied in conjunction with other salient scenario characteristics, such as economic growth and resource availability (e.g. in the MiniCAM simulations). Therefore, the impact of alternative assumptions with respect to technological change can be best quantified within a particular scenario family and with "fully harmonized" scenario quantifications (i.e. with comparable energy demand), as discussed for the A1 scenario groups above. In some scenarios within other scenario families, technology-specific sensitivity analyses were performed, such as in the B2C-MARIA scenario variant of the B2-MARIA quantification. The main differences between the two scenarios are the respective costs of coal and nuclear power. In B2C-MARIA, the price of coal was assumed to be US$1.4/GJ, while that in B2-MARIA is US$1.8/GJ. In contrast, the capital costs of nuclear power stations are US$1400/kW in B2-MARIA, while those in B2C-MARIA are assumed to remain at US$1800/kW. Thus, even comparatively small variations in relative technology characteristics such as costs and efficiencies can lead to wide differences in scenario outcomes. As discussed in Chapter 5, for instance, changing the relative economics between coal and nuclear in the two MARIA scenarios results in a difference of more than 200 GtC cumulative emissions32 over the 21st century.
An illustration of inter-scenario variability in technology costs and diffusion is given in Box 4-9 for the MESSAGE model simulations for one representative scenario of each scenario family and scenario group. As stated above, differences in technology diffusion across scenarios are influenced by many more factors than just alternative technology characteristics and cost assumptions. Growth of energy demand, resource availability and costs, and local circumstances (local air-quality regulations that require desulfurization of fuels or stack gases, or land availability and prices that influence biomass costs) are also important determinants of speed and potentials for the diffusion of new energy technologies.
Other reports in this collection