Box 4-9: Dynamics of Technological Change in the MESSAGE-Based Quantifications for the Four SRES Marker Scenarios.

Technological change in energy supply and end-use technologies has historically been a main driver of structural changes in energy systems, efficiency improvements, and improved environmental compatibility. Yet, despite its crucial role, the mechanisms that underlie technological innovation and diffusion of new technologies remain poorly understood, so modeling technological change as an endogenous process to the economy and society is still in its infancy. Historically, the track record of technology forecasts has at best been mixed, with a number of notable failures particularly in the energy sector. In the 1960s, for instance, R&D in the US attempted to develop nuclear-propelled aircraft, and nuclear electricity was anticipated to become "too cheap to meter." Conversely, the dynamic technological changes in microprocessors, information technologies, and aeroderivative turbines (and their combination with the steam cycle in the form of combined cycle gas turbines) were largely underestimated. This is similar to the pessimistic market outlook for gasoline-powered cars at the end of the 19^th and start of the 20^th centuries.

In recognition of the considerable uncertainty in describing future technological trends, a scenario approach was adopted to vary technology-specific assumptions in the MESSAGE model runs of the SRES scenarios. Depending on the specific interpretation of the four SRES scenario storylines, alternative technologies and alternative ranges of their future characteristics were assumed as model inputs.

Two guiding principles determined the choice of particular technology assumptions in MESSAGE.

First, technologies not yet demonstrated to function on a prototype scale were excluded. Therefore, for instance, nuclear fusion is excluded from the technology portfolio of all SRES scenarios calculated with the MESSAGE model. However, production of hydrogen- or biomass-based synfuels (e.g. ethanol) or advanced nuclear and solar electricity generation technologies are included, as they have demonstrated their physical feasibility at least on a laboratory or prototype scale, or in some specific niche markets (even if they are uneconomic at currently prevailing energy prices). Second, the range of technology-specific assumptions is empirically derived. Statistical distributions of technology characteristics based on a large technology inventory (consisting of 1600 technologies) and developed at IIASA (Messner and Strubegger, 1991; Strubegger and Reitgruber, 1995) were used. Means, maxima, and minima from these distributions (e.g. of estimated future technology costs) guided which particular values to adopt across scenarios on the basis of the scenario taxonomy suggested by the scenario storylines (ranging from conservative to optimistic).

Tables 4-13a to 4-13e summarize the technology characteristics and resultant diffusion rates across the four SRES scenario families and their scenario groups. Table 4-13a presents a brief overview of a selection of major energy technologies represented in the MESSAGE model. (Being a detailed "bottom-up" model, MESSAGE literally contains hundreds of individual technologies, too many to summarize here; instead, only the most important technology groups, aggregated across many individual technologies, are presented.) Table 4-13b summarizes salient technology characteristics in terms of levelized costs (investment and operating costs levelized per unit energy output, excluding fuel costs) and Table 4-13c summarizes the resultant marker deployment (diffusion) of these technologies by 2050 and 2100 for the B2-MESSAGE marker scenario. This scenario is characterized by intermediate levels of growth in energy demand and conservative assumptions as to future technological change. The latter were adopted based on a literature survey (Strubegger and Reitgruber, 1995) as well as an expert opinion poll.

In particular, the B2-MESSAGE scenario adopted technology characteristics of the equally conservative IIASA-WEC Scenario B (Nakicenovic et al., 1998), which was based on the Strubegger and Reitgruber (1995) analysis, complemented by a review of some 100 energy experts assembled by WEC. Table 4-13d indicates how technology costs in the other MESSAGE scenarios differ from those of the B2 scenario. (The prevalence of negative values in Table 4-13d indicates that most scenarios are more optimistic concerning cost improvements of future technology than the MESSAGE B2 scenario.) Finally, Table 4-13e indicates the difference in market deployment (diffusion) of the other MESSAGE SRES scenarios compared to that of the B2 scenario. Positive values indicate higher market deployment, and negative ones show lower diffusion. However, differences across scenarios in terms of technology diffusion are not governed by technology costs alone. Other technology characteristics (such as efficiency and infrastructure availability) and market (demand) growth are also important in determining market deployment rates and diffusion potentials of energy technologies.

Figure 4-11: Global primary energy structure, shares (%) of oil and gas, coal, and non-fossil (zero-carbon) energy sources - historical development from 1850 to 1990 and in SRES scenarios. Each corner of the triangle corresponds to a hypothetical situation in which all primary energy is supplied by a single source - oil and gas, coal at the left, and non-fossil sources (renewables and nuclear) to the right. Constant market shares of these energies are denoted by their respective isoshare lines. Historical data from 1850 to 1990 are based on Nakicenovic et al. (1998). For 1990 to 2100, alternative trajectories show the changes in the energy systems structures across SRES scenarios. They are grouped by shaded areas for the scenario families A1, A2, B1, and B2 with respective markers shown as lines. In addition, the four scenario groups within the A1 family (A1, A1C, A1G, and A1T) that explore different technological developments in the energy systems are shaded individually. The A1C and A1G scenario groups have been merged into one fossil-intensive A1FI scenario group in the SPM (see footnote 1). For comparison the IS92 scenario series are also shown, clustering along two trajectories (IS92c,d and IS92a,b,e,f). For model results that do not include non-commercial energies, the corresponding estimates from the emulations of the various marker scenarios by the MESSAGE model were added to the original model outputs.