Carbon emissions from energy use are projected to increase by an average of 1.3 percent a year from 1997 to 2020, reaching 1,975 million metric tons (Figure 118). This projection is slightly higher than the AEO98 projection of 1,956 million metric tons, due to higher energy consumption-particularly, coal for electricity generation and petroleum for transportation.

Increasing concentrations of carbon dioxide, methane, nitrous oxide, and other greenhouse gases may increase the Earth's temperature and affect the climate. The AEO99 projections include analysis of the Climate Change Action Plan (CCAP), developed by the Clinton Administration in 1993 to stabilize U.S. greenhouse gas emissions by 2000 at 1990 levels. Carbon emissions from fuel combustion, the primary source of greenhouse gas emissions, were about 1,346 million metric tons in 1990. The analysis does not account for carbon-absorbing sinks, the 13 CCAP actions related to non-energy programs or gases other than carbon dioxide, nor any future mitigation actions that may be considered to meet the reductions proposed in the Kyoto Protocol.

Emissions in the 1990s have grown more rapidly than projected at the time CCAP was formulated, partly due to lower energy prices and higher economic growth than projected, which have led to higher energy demand. In addition, some CCAP programs have been curtailed. Additional carbon mitigation programs, technology improvements, or more rapid adoption of voluntary programs could result in lower emissions levels than projected here.

1995 2000 2005 2010 2015

U.S. carbon emissions from energy use are projected to grow at an average annual rate of 1.3 percent; however, per capita emissions grow by only 0.4 percent a year (Figure 119). To stabilize or reduce total emissions, population growth would need to be offset by reductions in per capita emissions.

Emissions in the residential sector, including emissions from the generation of electricity used in the sector, are projected to increase by 1.2 percent a year, reflecting the ongoing trends of electrification and penetration of new appliances and services. Significant growth in office equipment and other uses is also projected in the commercial sector, but growth in consumption-and in emissions, which also increase by 1.2 percent a year-is likely to be moderated by slowing growth in floorspace, coupled with efficiency standards, voluntary efficiency programs, and technology improvements.

Transportation emissions grow at an average annual rate of 1.7 percent as a result of increases in vehicle-miles traveled and freight and air travel, combined with slow growth in the average lightduty fleet efficiency. Industrial emissions are projected to grow by only 0.9 percent a year, as shifts to less energy-intensive industries and efficiency gains moderate growth in energy use.

Petroleum products are the leading source of carbon emissions from energy use. In 2020, petroleum is projected to contribute 823 million metric tons of carbon to the total 1,975 million tons, a 42-percent share (Figure 120). About 81 percent (665 million metric tons) of the petroleum emissions result from transportation use, which could be lower with less travel or more rapid development and adoption of higher efficiency or alternative-fuel vehicles.

Coal is the second leading source of carbon emissions, projected to produce 676 million metric tons in 2020, or 34 percent of the total. The share declines from 36 percent in 1997 because coal consumption increases at a slower rate through 2020 than consumption of petroleum and natural gas, the sources of virtually all other energy-related carbon emissions. Most of the increases in coal emissions result from electricity generation. A slight increase in emissions from industrial steam coal use is partially offset by a decline in emissions from coking coal.

In 2020, natural gas use is projected to produce 475 million metric tons of carbon emissions, a 24percent share. Of the fossil fuels, natural gas consumption and emissions increase most rapidly through 2020, at average annual rates of 1.7 percent; however, natural gas produces only half the carbon emissions of coal per unit of input. Average emissions from petroleum use are between those for coal and natural gas. The use of renewable fuels and nuclear generation, which emit little or no carbon, mitigates the growth of emissions.

Electricity use is a major cause of carbon emissions. Although electricity produces no emissions at the point of use, its generation currently accounts for 36 percent of total carbon emissions, and that share is expected to increase to 38 percent in 2020. Coal, which accounts for about 52 percent of electricity generation in 2020 (excluding cogeneration), produces 81 percent of electricity-related carbon emissions (Figure 121). In 2020, natural gas accounts for 30 percent of electricity generation but only 18 percent of electricity-related carbon emissions. Between 1997 and 2020, 50 gigawatts of nuclear capacity are expected to be retired, resulting in a 43percent decline in nuclear generation. To compensate for the loss of nuclear capacity and meet rising demand, 345 gigawatts of new fossil-fueled capacity (excluding cogeneration) will be needed. Increased generation from fossil fuels will raise electricityrelated carbon emissions by 213 million metric tons, or 40 percent, from 1997 levels. Generation from renewable technologies increases by 53 billion kilowatthours, or 12 percent, between 1997 and 2020 but is insufficient to offset the projected increase in generation from fossil fuels.

The projections include announced activities under the Climate Challenge program, such as fuel switching, repowering, life extension, and demand-side management, but they do not include offset activities, such as reforestation. Additional use of lower carbon fuels, reduced electricity demand growth, or improved technologies all could contribute to lower emissions than are projected here.

1990 1995 2000 2005 2010 2015 2020 CAAA90 called for annual emissions of sulfur dioxide (SO2) by electricity generators to be reduced to approximately 12 million short tons in 1996, 9.48 million tons between 2000 and 2009, and 8.95 million tons a year thereafter. More than 95 percent of the SO2 produced by generators results from coal combustion, with the rest from residual oil.

In Phase 1, 261 generating units at 110 plants were issued tradable emissions allowances permitting SO2 emissions to reach a fixed amount per yeargenerally less than the plant's historical emissions. Allowances may also be banked for use in future years. Switching to lower sulfur, subbituminous coal was the option chosen by more than half of the generators. In Phase 2, beginning in 2000, emissions constraints on Phase 1 plants will be tightened, and limits will be set for the remaining 2,500 boilers at 1,000 plants. With allowance banking, emissions are expected to decline from 11.9 million tons in 1995 to 11.4 million in 2000 (Figure 122). Since allowance prices are projected to increase after 2000, it is expected that 26.4 gigawatts of capacityabout 88 300-megawatt plants-will be retrofitted with scrubbers to meet the Phase 2 goal (Table 11).

Table 11. Scrubber retrofits and allowance costs, 2000-2020

Forecast

Cumulative retrofits

2000 2005 2010 2015 2020

13.6 13.6 25.8 26.4 26.4

90 240 293 182 130

2005

Nitrogen oxide (NO) emissions in the United States will fall significantly over the next 5 years as new legislation takes effect (Figure 123). First will be the second phase of the NOx reduction program from CAAA90, which calls for NO, reductions at electric power plants in two phases-the first in 1995 and the second in 2000. It is expected that the second phase of CAAA90 will result in NO, reductions of 1.5 million tons between 1999 and 2000.

A second piece of legislation, the ozone transport rule (OTR), will take effect in 2003. After studying the ozone transport problem, the U.S. Environmental Protection Agency (EPA) issued the OTR in September 1997. The OTR sets caps on NO. emissions in each of 22 midwestern and eastern States during the 5-month summer season (May through September). The EPA wants to establish a cap and trade program with tradable emission permits. Holders of the permits would be free to use them themselves or sell them to someone whose NOx emission reduction options are more costly.

The OTR is expected lead to a total NOx emissions reduction of 0.7 million tons between 2002 and 2003 as control technologies are installed on utility boilers. By 2020, 10 gigawatts of capacity is expected to be retrofitted with advanced combustion controls, selective noncatalytic reduction units (SNCR) are expected to be added to 96 gigawatts, and selective catalytic reduction units (SCR) are expected to be added to 111 gigawatts. The annualized cost is estimated to be $2 billion, relative to about $200 billion in annual consumer expenditures for electricity.

Differences in long-run economic forecasts can be traced primarily to different views of the major supply-side determinants of growth: labor force and productivity change. Other forecasts are presented in Table 12. The WEFA forecast shows the highest economic growth compared to the AEO99 and DRI reference cases, including higher growth rates for the labor force. The AEO99 long-run forecast of economic growth is higher than the AEO98 forecast by 0.2 percent, when compared on a similar basis, with a projected annual growth rate for GDP of 1.8 percent from 1997 to 2020.

(BTAB) are shown in Table 13 (IEA, 1996; PEL, February 1998; PIRA, October 1997; NRCan, April 1997; BTAB, September 1998). With the exception of IEA in 2005 and PEL, the range between the AEO99 low and high world oil price cases spans the range of other published forecasts beyond 2005.

Table 13. Forecasts of world oil prices, 2000-2020 1997 dollars per barrel

Total Energy Consumption

The AEO99 forecast of end-use sector energy consumption over the next two decades shows far less volatility than has occurred historically. Between 1974 and 1984, volatile world oil markets dampened domestic oil consumption. Consumers switched to electricity-based technologies in the buildings sector, while in the transportation sector new car fuel efficiency nearly doubled. Natural gas use declined as a result of high prices and limitations on new gas hookups. Between 1984 and 1995, however, both petroleum and natural gas consumption rebounded, bolstered by plentiful supplies and declining real energy prices. As a consequence, new car fuel efficiency in 1995 was less than 2 miles per gallon higher than in 1984, and natural gas use (residential, commercial, and industrial) was almost 25 percent higher than it was in 1984.

Given potentially different assumptions about, for example, technological developments over the next 20 years, the forecasts from DRI, GRI, and WEFA have remarkable similarities to those in AEO99. Electricity is expected to remain the fastest growing source of delivered energy (Table 14), although its rate of growth is down sharply from historical rates in each of the forecasts, because many traditional