advanced storage facilities would enable intermittent solar and wind resources to service a greater part of the electricity system's load. All these assumptions about technical progress explain why the FFES concludes that the costs of implementing the scenario are small compared with the baseline.

The high contribution of renewables in the FFES scenario is in contrast to the conclusion of an extensive analysis of renewable energy potential by the World Energy Council (1994). This analysis estimates that renewable energy resources would supply about 50% of total energy by 2100 in the so-called "ecologically driven scenario," which is the most far-reaching CO2 reduction case in the study. It is argued that the total potential for renewable energy could be larger at that time, but if renewables make a major contribution and energy growth is low over a long period as assumed in the ecologically driven scenario, supplies of petroleum and natural gas will still be available and it will be beneficial to mix different sources in the supply system.

An integrated global top-down/bottom-up analysis has been carried out in relation to the IPCC Working Group II Second Assessment Report (Volume 2 of this report) and has resulted in the construction of scenarios for a Low Emissions Supply System (LESS). The scenario results are shown in Figure 9.32.

The LESS scenarios provide estimates of the potential for greenhouse gas emission abatement using data developed from the detailed technology assessments in the Working Group IIa assessment process (IPCC, 1994). The energy supply systems were constructed from work by both bottom-up (Lashof and Tirpak, 1990; Johansson et al., 1993) and top-down modellers (Edmonds, 1994). In the bottom-up variant, the starting point is energy demand projections by world regions for 2025, 2050, 2075, and 2100 developed by the Response Strategies Working Group (RSWG, 1990) as part of the IPCC 1990 Assessment Report. A high economic growth version of this scenario demonstrated the importance of technology assumptions in the LESS analysis. Global GDP grows 8-fold by 2050 relative to 1985 and 28-fold by 2100, with primary energy consumption growing from 323 EJ in 1985 to 559 EJ in 2050 and 664 EJ in 2100. Biomass, mostly for fuels used directly, plays a major role. It accounts for 31% of primary energy in 2025 and rises to 50% in 2100 in one version of the scenario.

As a comparison with the LESS bottom-up analysis, a top-down analysis was carried out (Edmonds et al.[Edmonds, Wise, and MacCracken or Edmonds, Wise, Pitcher, et al.?] 1994), incorporating performance and cost parameters for some of the key energy technologies used in the construction of the base case. The following six technology cases were modelled:

1.

A reference scenario very similar to IPCC 1992a (IPCC, 1992; the exogenous enduse energy intensity improvement rate is 0.5% by 2005, rising to 1.0% by 2035 and reaching 1.5% by the year 2065);

2.

3.

4.

Similar to Case 1, but with an emphasis on energy-efficient power generation from fossil fuels (efficiency reaches 66% by 2095);

Similar to Case 1, but hydrogen, solar, and wind power become more competitive;

Similar to Case 3, but compressed hydrogen is used instead of liquefied hydrogen;

5.

Similar to Case 4, but biomass prices are more competitive;

6.

Similar to Case 5, but the autonomous rate of energy efficiency improvements increases to 2.0% per year in 2050.

The results of the scenario analysis are given in Figure 9.33, showing global annual fossil fuel CO2 emissions and energy production and use.

Cases 5 and 6 show that if the assumed technological characteristics are realized, significant reductions in CO2 emissions could be achieved without economic penalty, as the technologies embodied in the low emissions scenario become competitive under market conditions with traditional fossil fuel technologies.

These LESS scenario results support the suggestion that the disagreement in the numerical outcomes of the models, in the very long run, is due less to the modelling structure than to the exogenous hypotheses. From a decision-making point of view, however, the most sensitive issue that remains to be addressed is the plausibility of this long-term-transition, if one accounts for all the general equilibrium effects (for example, the implications of shifting agricultural activities to fuel production) and for all the transaction costs involved in the transition. It is, however, worth noting that the World Energy Council's ecologically driven case, which includes less voluntaristic assumptions than the LESS scenarios, points to a 60% reduction in global energy-related CO2 emissions from 1990 levels in 2100. This scenario is buttressed by a very comprehensive study of renewable energy prospects out to the year 2100.

9.3 Studies of the Costs of Carbon Sequestration

There are many difficulties inherent in developing and comparing estimates of carbon sequestration costs (see Chapter 8, Section 3.4). Those difficulties notwithstanding, this section attempts to summarize, compare, and critique the results of national, regional, and global carbon sequestration studies.

Many factors affect studies' estimates of carbon sequestration costs and potential, including assumptions, methods, and data used with respect to land area, land costs, treatment costs, discount rates, carbon capture rates and patterns, ecosystem components included in the analysis, and the treatment of forest products. The data and assumptions

employed by various studies are summarized in Tables 9.27 to 9.34.

As an illustration of the variation in data employed by sequestration studies, consider the estimates of land availability by region and practice listed in Table 9.27. For the tropics, Grainger (1988) has estimated that 621x10° ha of land may be suitable for establishing forest plantations. Dixon, Winjum, and Krankina (1991) appear to be in relatively close agreement with this figure, estimating that there are 551x106 ha suitable for forest plantations in Latin America, Africa, and Asia. In contrast, for the same three regions, Houghton et al. (1993) arrive at an estimate of less than half as much - 251x106 ha. The more optimistic figures are lent support by the studies of Ravindranath and Somashekhar (1994) and Xu (1994), which, taken together, suggest that there are more than 150x10 ha suitable for forestry plantations in India and China. The estimates for tropical agroforestry show similar disagreement. Dixon, Winjum, and Krankina (1991) estimate that there are 835x10° ha suitable for agroforestry, while Houghton et al. (1993) suggest that 1895x10° ha are available.

The range of estimates of land availability for the temperate areas is only slightly narrower. Adams et al. (1990), Moulton and Richards (1990), and Richards et al. (1993) suggest that there may be 100x10° to 115x10 ha of marginal agricultural land suitable for afforestation in the United States alone. However, Dixon, Winjum, and Krankina (1991) identify only 42x10° ha in all of North America as suitable for forest plantations. Parks and Hardie (1995) consider only 10.4x10° ha in the United States for their analysis, though the land area constraint in their study is determined by an assumed programmatic budget and not by land availability.

Studies have also shown a wide range of estimates of land costs. As discussed in Chapter 8, this factor has proved particularly difficult because of the many nonmarket considerations associated with the social cost of converting between land uses. Because of the difficulty of determining the appropriate figures, some studies simply have not included land costs as an element of the cost analysis (e.g., Dixon, Schroeder, and Winjum, 1991). Others have apparently assumed that the use of land is costless because it is either public land (New York State, 1991) or because the wood products will eventually pay for the land (van Kooten et al., 1992; Xu, 1994). As might be expected, those that have included land costs have arrived at a wide range of estimates for that variable. Table 9.28 provides a summary of the land cost data employed in the various studies. Note that several of the studies that provide estimates of land availability do not include cost figures.

Initial treatment costs (see Table 9.29) are generally expressed as a capital outlay, while the maintenance costs, if included, are expressed as annual costs. Land costs may be expressed as either annual costs (rent) or capital costs. The cost analysis is facilitated by summarizing these costs as either a net present value equivalent or an equivalent annual cost. The key factor for this operation is the discount rate applied to these costs. Table 9.30 summarizes the discount rates used by various sequestration cost studies.

The importance of the choice of discount rate depends critically upon the specific structure of the analysis. For example, Moulton and Richards (1990) defined the cost per tonne of carbon as a ratio of land rent plus annualized establishment costs to average annual carbon capture. Since establishment costs are such a small part of the total costs in that analysis, the difference between applying a 4% and a 10% discount rate was minor. However, in Richards et al. (1993), which used land purchase costs and time-dependent carbon yield curves, raising the discount rate from 3% to 7% nearly doubled the unit cost of carbon sequestration.

One of the most significant differences among studies is how they have addressed the irregular flows of carbon inherent in carbon sequestration and the differences in - patterns among activities. For example, as Figure 9.34 illustrates, planting Loblolly pine on agricultural land in the Southeast Region of the United States leads to carbon uptake rates that peak during the second decade after planting and taper off during the next four decades. In contrast, the carbon uptake rates associated with planting Ponderosa pine in the Mountain States do not peak until the sixth decade after planting. Other carbon sequestration activities have similar variations in their flows over time.

Carbon sequestration studies have dealt with carbon flows in one of several ways. Some, such as Adams et al. (1993) and Moulton and Richards (1990), have used average carbon yields, expressed in tonnes per acre per year over the first forty years after tree stand establishment. Others have used yield curves such as those illustrated in Figure 9.34 to describe expected carbon flows on a year-by-year basis (Nordhaus, 1991[a or b?]; Richards et al., 1993). A third approach, introduced by Schroeder (1992), expresses programme effects on carbon in terms of storage rather than flows. The method, called mean carbon storage (MCS), assumes that once a practice is implemented, the forest system is sustained in the same use over time. Thus the carbon changes are expressed in terms of the change in the amount of carbon storage on site, averaged over one full rotation. This is expressed as

where C, is the standing carbon (tonnes) in year i, and n is the rotation length. Finally, the standing carbon method, a variation on the MCS method, expresses accomplishments in terms of the carbon standing at the end of the analysis period, say fifty years. This is the method used by Ravindranath and Somashekhar (1994). Table 9.2.5 provides a summary of the methods employed by sequestration cost studies.

[EQUATION NEEDS TO BE TIDIED UP. PLEASE PROVIDE CLEAN VERSION. WHAT IS THE TERM A?]

Carbon flows into forests can also be reversed by harvesting. Those studies that have concentrated on plantation establishment have dealt with this issue in one of three ways (Table 9.32). The group of studies that employ the MCS method assume that all carbon is released upon harvest, but that the forestry practices are repeated in continuous rotations (Dixon, Schroeder, and Winjum, 1991; Dixon, Winjum, and Krankina, 1991; Dixon et al., 1994). Hence, the concept of carbon release is built into the analysis. Another group of studies assumes that the land planted with trees is

permanently withdrawn from other uses, including harvest of wood products, so that there is no release of carbon (e.g., Nordhaus, 1991[a or b?]; Richards et al., 1993). This assumption must be reflected in the calculation of land costs. Finally, some studies simply do not address the release of carbon upon harvest, implicitly assuming that either the forest area will not be harvested, or that the harvest will occur so far in the future as not to be a concern (e.g., Moulton and Richards, 1990).

Several components of a forest ecosystem store carbon, including tree trunks, branches, leaves, and coarse and fine roots; soils; litter; and understory. Studies have varied significantly with respect to how they address these various components. Some have included all components in their carbon accounting (e.g., Moulton and Richards, 1990). Others have limited their analysis to above-ground carbon (Dixon et al. [Dixon, Winjum, and Krankina or Dixon, Schroeder, and Winjum], 1991). Table 9.33 provides a summary of which carbon components are included in each of the studies reviewed.

Box 8.2 in Chapter 8 [DON'T HAVE BOX 8.2. PLEASE SUPPLY.] provided a discussion and sample calculations of the various summary statistics used in carbon sequestration cost studies to capture the concept of "dollars per tonne of carbon sequestration." Van Kooten et al. (1992) demonstrate the importance of the choice of summary statistics in their analysis of the cost-effectiveness of carbon sequestration in