cost of the most energy-efficient equipment and technologies in the residential and commercial sectors will decline 35 percent from their reference case values by 2020 and that building shell efficiencies will improve 50 percent over reference case levels. In the transportation sector, the case assumes the cost and performance criteria from the Department of Energy's Office of Energy Efficiency and Renewable Energy for light-duty vehicles' and from the efficiency case of the recent interlaboratory study (the so-called "Five Lab Study") for air, freight, marine, and rail travel.' In the industrial sector, the assumed rates of decline in energy intensity for the energyintensive industries are increased, resulting in an average decline of 1.5 percent a year for the sector, compared with 1.1 percent a year in the reference case. These assumptions represent a plausible range in technology improvement, but not necessarily the extreme level that might occur if breakthroughs more significantly improve the costs and efficiencies of technologies relative to the reference case. Changes in the overall composition of the stock of energy-using equipment can be slow, however, and in general it is unlikely that equipment with remaining functional economic life would be prematurely retired.

In a 1998 end-use technology case, it is assumed that all future equipment purchases in the residential and commercial sectors will be made only from the equipment available in 1998 and that building shell efficiencies will be frozen at 1998 levels. Efficiencies for new transportation equipment are fixed at 1998 levels for all travel modes in the 1998 technology case, and plant and equipment efficiencies are fixed at 1998 levels for the industrial sector.

The high technology case projects an average annual decline of 1.2 percent in energy consumed per dollar of GDP through 2010, compared with 0.9 percent in the reference case. In the 1998 technology case, the average decline in intensity is only 0.8 percent a year. In the high technology

'Office of Energy Efficiency and Renewable Energy, Office of Transportation Technologies, OTT Program Analysis Methodology: Quality Metrics 98 (June 17, 1997).

2U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Scenarios of U.S. Carbon Reductions: Potential Impacts of Energy Technologies by 2010 and Beyond, ORNL/CON-444 (Washington, DC, September 1997).

case, energy consumption in 2010 is 4 percent lower than in the reference case. In the 1998 technology case, consumption increases by 2 percent in 2010 over reference case levels.

The lower energy consumption in the high technology case lowers carbon emissions 4 percent, or 79 million metric tons, in 2010, relative to the reference case (Figure 3). Approximately 42 percent, or 33 million metric tons, of the reduction in carbon emissions in 2010 in the high technology case is a result of lower electricity demand and generation. An additional 27 million metric tons of the reduction, or 34 percent, is in the transportation sector, as a result of shifts to more efficient or alternatively-fueled vehicles. In the 1998 technology case, emissions are 2 percent, or 31 million metric tons, higher in 2010 than in the reference case.

Figure 3

U. S. Carbon Emissions in Three End Use Technology Cases, 1990-2020 (million metric tons)

Macroeconomic Growth. The rate of technology improvement is only one of the assumptions that may impact the level of carbon emissions. The projected growth rate for the U.S. economy has a significant impact on both energy demand and carbon emissions. In the reference case, real GDP is projected to grow at an average rate of 2.2 percent a year through 2010. AEO98 includes two additional cases with higher and lower economic growth, incorporating higher and lower growth rates for population, labor force, and labor productivity.

Figure 4

U.S. Carbon Emissions in Three Economic Growth Cases, 1990-2020 (million metric tons)

With higher assumed real economic growth of 2.7 percent a year through 2010, energy consumption in 2010 is 4 percent higher than in the reference case, raising emissions 4 percent, or 80 million metric tons, over reference case levels (Figure 4). Lower real economic growth of

1.7 percent a year through 2010 reduces energy demand in 2010 by 4 percent, relative to the reference case, and carbon emissions by 5 percent, or 90 million metric tons.

Renewable Technologies. Increased generation by renewable technologies, which are carbon free, can reduce carbon emissions. AEO98 analyzed more rapid improvements in renewable generating technologies relative to the reference case. A high renewables case was defined by using the more optimistic Department of Energy's Office of Energy Efficiency and Renewable Energy assumptions for capital costs, operations and maintenance costs, and capacity factors for nonhydroelectric renewables. In addition, the high renewables case assumed that 305 megawatts of capacity at The Geysers geothermal station continue in operation after their projected retirement in the reference case, and that the share of landfill gas captured for energy production increases to 50 percent in 2020, compared with 40 percent in the reference case. The results of the high renewables case suggest that technology improvements would increase generation from nonhydroelectric renewables by 43 percent, or 46 billion kilowatthours, in 2010 compared to the reference case. The increment in generation is mostly from wind, biomass, and geothermal resources, which displace coal and natural gas. The share of total electricity generation from nonhydroelectric renewables increases to 3.5 percent in 2010, compared with 2.5 percent in the reference case. With the more optimistic renewable assumptions, carbon emissions are reduced by 11 million metric tons, or 1 percent, in 2010 (Figure 5).

Renewable Portfolio Standards. In addition to a case with more optimistic assumptions on renewable costs and performance, AEO98 also includes an analysis of the impacts of stimulating the development of renewable resources through the use of a renewable portfolio standard (RPS), which specifies a percentage of electricity generation to come from renewable sources. Several bills that include RPS provisions have been introduced in the U.S. Congress. AEO98 includes two scenarios which specify 2 percent of generation from nonhydroelectric renewable technologies in 2000, increasing to 5 and 10 percent in 2020. As a result of the RPS, generation from natural gas and coal plants is reduced and generation from biomass, wind, and geothermal plants increases. These technologies are the closest to being competitive with new natural gas

Figure 5

80

60

40

Changes in U.S. Carbon Emissions from the Reference Case in Five Alternative Cases, 2010 and 2020 (million metric tons)

200

20

0

-20

-40

-60

2020

fired technologies. Because of their relatively high cost, solar technologies are not likely to benefit from a RPS requirement. With a 5-percent RPS, total carbon emissions are reduced by 1 percent, or 14 million metric tons, in 2010, relative to the reference case. In the 10-percent RPS case, carbon emissions are reduced by 2 percent, or 31 million metric tons, in 2010 (Figure 5).

Nuclear Life Extension. As noted previously, nuclear generation declines significantly in the reference case projections. A high nuclear case in AEO98 looks at the impact of extending the life of nuclear generating plants 10 years beyond the retirement dates assumed in the reference case although no new nuclear plants are assumed. In 2010, operable nuclear capacity is 16 percent higher than in the reference case, reducing generation from fossil-fired plants and reducing carbon emissions by 1 percent, or 22 million metric tons. In a low nuclear case, nuclear