Annual Energy Outlook 2006 with Projections to 2030 - Usinfo.org

DOE/EIA-0383(2006)February 2006Annual Energy Outlook 2006With Projections to 2030February 2006

For Further Information ...The Annual Energy Outlook 2006 (AEO2006) was prepared by the Energy Information Administration(EIA), under the direction of John J. Conti (john.conti@eia.doe.gov, 202-586-2222), Director, IntegratedAnalysis and Forecasting; Paul D. Holtberg (paul.holtberg@eia.doe.gov, 202/586-1284), Director, Demandand Integration Division; Joseph A. Beamon (jbeamon@eia.doe.gov, 202/586-2025), Director, Coal andElectric Power Division; Andy S. Kydes (akydes@eia.doe.gov, 202/586-2222), Acting Director, Oil and GasDivision and Senior Technical Advisor; and Glen E. Sweetnam (glen.sweetnam@eia.doe.gov, 202/586-2188),Director, International, Economic, and Greenhouse Gases Division.For ordering information and questions on other energy statistics available from EIA, please contact EIA’sNational Energy Information Center. Addresses, telephone numbers, and hours are as follows:National Energy Information Center, EI-30Energy Information AdministrationForrestal BuildingWashington, DC 20585Telephone: 202/586-8800E-mail: infoctr@eia.doe.govFAX: 202/586-0727World Wide Web Site: http://www.eia.doe.gov/TTY: 202/586-1181FTP Site: ftp://ftp.eia.doe.gov/9 a.m. to 5 p.m., eastern time, M-FSpecific questions about the information in this report may be directed to:Overview . . . . . . . . . . . . . . . Paul D. Holtberg (paul.holtberg@eia.doe.gov, 202/586-1284)Economic Activity . . . . . . . . . . Ronald F. Earley (rearley@eia.doe.gov, 202/586-1398)International Oil Markets . . . . . . G. Daniel Butler (gbutler@eia.doe.gov, 202/586-9503)Residential Demand . . . . . . . . . John H. Cymbalsky (jcymbals@eia.doe.gov, 202/586-4815)Commercial Demand . . . . . . . . . Erin E. Boedecker (eboedeck@eia.doe.gov, 202/586-4791)Industrial Demand . . . . . . . . . . T. Crawford Honeycutt (choneycu@eia.doe.gov, 202/586-1420)Transportation Demand . . . . . . . John D. Maples (jmaples@eia.doe.gov, 202/586-1757)Electricity Generation, Capacity . . Jeff Jones (jjones@eia.doe.gov, 202/586-2038)Electricity Generation Emissions . . Zia Haq (zhaq@eia.doe.gov, 202/586-2869)Electricity Prices . . . . . . . . . . . Lori Aniti (laniti@eia.doe.gov, 202/586-2867)Nuclear Energy. . . . . . . . . . . . Laura Martin (laura.martin@eia.doe.gov, 202/586-1494)Renewable Energy . . . . . . . . . . Chris Namovicz (chris.namovicz@eia.doe.gov, 202/586-7120)Oil and Gas Production . . . . . . . Ted E. McCallister (tmccalli@eia.doe.gov, 202/586-4820)Natural Gas Markets. . . . . . . . . Philip M. Budzik (pbudzik@eia.doe.gov, 202/586-2847)Oil Refining and Markets . . . . . . Anthony F. Radich (anthony.radich@eia.doe.gov, 202/586-0504)Coal Supply and Prices. . . . . . . . Michael L. Mellish (mmellish@eia.doe.gov, 202/586-2136)Greenhouse Gas Emissions . . . . . Daniel H. Skelly (dskelly@eia.doe.gov, 202/586-1722)AEO2006 will be available on the EIA web site at www.eia.doe.gov/oiaf/aeo/ in early February 2006. Assumptionsunderlying the projections and tables of regional and other detailed results will also be available in earlyFebruary 2006, at web sites www.eia.doe.gov/oiaf/assumption/ and /supplement/. Model documentationreports for the National Energy Modeling System (NEMS) are available at web site www.eia.doe.gov/bookshelf/docs.html and will be updated for AEO2006 during the first few months of 2006.Other contributors to the report include Joseph Benneche, Bill Brown, John Cochener, Margie Daymude,Linda Doman, Robert Eynon, James Hewlett, Elizabeth Hossner, Diane Kearney, James Kendell, Nasir Khilji,Paul Kondis, Han Lin Lee, Thomas Lee, Phyllis Martin, Tom Petersik, Chetha Phang, Eugene Reiser,Laurence Sanders, Bhima Sastri, Sharon Shears, Robert Smith, Yvonne Taylor, Brian Unruh, Jose Villar,Dana Van Wagener, Steven Wade, and Peggy Wells.

DOE/EIA-0383(2006)Annual Energy Outlook 2006With Projections to 2030February 2006Energy Information AdministrationOffice of Integrated Analysis and ForecastingU.S. Department of EnergyWashington, DC 20585This publication is on the WEB at:www.eia.doe.gov/oiaf/aeo/This report was prepared by the Energy Information Administration, the independent statistical andanalytical agency within the U.S. Department of Energy. The information contained herein should beattributed to the Energy Information Administration and should not be construed as advocating orreflecting any policy position of the Department of Energy or any other organization.

PrefaceThe Annual Energy Outlook 2006 (AEO2006), preparedby the Energy Information Administration(EIA), presents long-term forecasts of energy supply,demand, and prices through 2030. The projectionsare based on results from EIA’s National EnergyModeling System (NEMS).The report begins with an “Overview” summarizingthe AEO2006 reference case and comparing it withthe AEO2005 reference case. The next section, “Legislationand Regulations,” discusses evolving legislationand regulatory issues, including recently enactedlegislation and regulation, such as the Energy PolicyAct of 2005, and some that are proposed. “Issues inFocus” includes a discussion of the basis of EIA’s substantialrevision of the world oil price trend used inthe projections. It also examines the following topics:implications of higher oil price expectations for economicgrowth; differences among types of crude oilavailable on world markets; energy technologies onthe cusp of being introduced; nonconventional liquidstechnologies beginning to play a larger role in energymarkets; advanced vehicle technologies included inAEO2006; mercury emissions control technologies;and U.S. greenhouse gas intensity. “Issues in Focus”is followed by “Energy Market Trends,” which providesa summary of the AEO2006 projections forenergy markets.The analysis in AEO2006 focuses primarily on a referencecase, lower and higher economic growth cases,and lower and higher energy price cases. In addition,more than 30 alternative cases are included inAEO2006. Readers are encouraged to review the fullrange of cases, which address many of the uncertaintiesinherent in long-term forecasts. Complete tablesfor the five primary cases are provided in AppendixesThe projections in the Annual Energy Outlook 2006 arenot statements of what will happen but of what mighthappen, given the assumptions and methodologies used.The projections are business-as-usual trend estimates,given known technology, technological and demographictrends, and current laws and regulations. Thus, they providea policy-neutral reference case that can be used toanalyze policy initiatives. EIA does not propose, advocate,or speculate on future legislative and regulatory changes.All laws are assumed to remain as currently enacted; however,the impacts of emerging regulatory changes, whendefined, are reflected.Because energy markets are complex, models are simplifiedrepresentations of energy production and consumption,regulations, and producer and consumer behavior.Projections are highly dependent on the data, methodologies,model structures, and assumptions used in theirA through C. Major results from many of the alternativecases are provided in Appendix D. Appendix Ebriefly describes NEMS and the alternative cases.The AEO2006 projections are based on Federal,State, and local laws and regulations in effect on orbefore October 31, 2005. The potential impacts ofpending or proposed legislation, regulations, andstandards (and sections of existing legislation requiringfunds that have not been appropriated) are notreflected in the projections. For example, theAEO2006 reference case does not include implementationof the proposed, but not yet final, increase incorporate average fuel economy (CAFE) standardsbased on vehicle footprint for light trucks—includingpickups, sport utility vehicles, and minivans. In general,historical data used in the AEO2006 projectionsare based on EIA’s Annual Energy Review 2004, publishedin August 2005; however, data are taken frommultiple sources. In some cases, only partial or preliminary2004 data were available. Historical data arepresented in this report for comparative purposes;documents referenced in the source notes should beconsulted for official data values. The projections for2005 and 2006 incorporate the short-term projectionsfrom EIA’s September 2005 Short-Term Energy Outlookwhere the data are comparable.Federal, State and local governments, trade associations,and other planners and decisionmakers in thepublic and private sectors use the AEO2006 projections.They are published in accordance with Section205c of the Department of Energy Organization Actof 1977 (Public Law 95-91), which requires the EIAAdministrator to prepare annual reports on trendsand projections for energy use and supply.development. Behavioral characteristics are indicative ofreal-world tendencies rather than representations ofspecific outcomes.Energy market projections are subject to much uncertainty.Many of the events that shape energy markets arerandom and cannot be anticipated, including severeweather, political disruptions, strikes, and technologicalbreakthroughs. In addition, future developments in technologies,demographics, and resources cannot be foreseenwith certainty. Many key uncertainties in the AEO2006projections are addressed through alternative cases.EIA has endeavored to make these projections as objective,reliable, and useful as possible; however, they should serveas an adjunct to, not a substitute for, a complete andfocused analysis of public policy initiatives.ii Energy Information Administration / Annual Energy Outlook 2006

ContentsPageOverview .......................................................................... 1Legislation and Regulations .......................................................... 13Introduction ........................................................................... 14EPACT2005 Summary. .................................................................. 15California Greenhouse Gas Emissions Standards for Light-Duty Vehicles ......................... 22Proposed Revisions to Light Truck Fuel Economy Standards ................................... 23State Renewable Energy Requirements and Goals: Update Through 2005. ........................ 24State Air Emission Regulations That Affect Electric Power Producers. ........................... 27Federal Air Emissions Regulations. ........................................................ 28Update on Transition to Ultra-Low-Sulfur Diesel Fuel ........................................ 29State Restrictions on Methyl Tertiary Butyl Ether ............................................ 29Volumetric Excise Tax Credit for Alternative Fuels ........................................... 30Issues in Focus ..................................................................... 31Introduction ........................................................................... 32World Oil Prices in AEO2006 ............................................................. 32Economic Effects of High Oil Prices ........................................................ 33Changing Trends in the Refining Industry .................................................. 39Energy Technologies on the Horizon ....................................................... 41Advanced Technologies for Light-Duty Vehicles .............................................. 49Nonconventional Liquid Fuels ............................................................ 51Mercury Emissions Control Technologies ................................................... 58U.S. Greenhouse Gas Intensity and the Global Climate Change Initiative ......................... 59Market Trends. ..................................................................... 61Trends in Economic Activity .............................................................. 62International Oil Markets ................................................................ 64Energy Demand ........................................................................ 65Residential Sector Energy Demand ........................................................ 67Commercial Sector Energy Demand. ....................................................... 69Industrial Sector Energy Demand ......................................................... 71Transportation Sector Energy Demand ..................................................... 74Electricity Demand and Supply. ........................................................... 77Electricity Supply. ...................................................................... 78Electricity From Renewable Sources ....................................................... 81Electricity Fuel Costs and Prices .......................................................... 82Electricity Alternative Cases. ............................................................. 83Natural Gas Demand .................................................................... 85Natural Gas Supply ..................................................................... 86Natural Gas Prices. ..................................................................... 87Natural Gas Alternative Cases ............................................................ 88Oil Prices and Production ................................................................ 91Oil Price Cases ......................................................................... 92Oil Price and Technology Cases ........................................................... 93ANWR Alternative Oil Case .............................................................. 94Refined Petroleum Products .............................................................. 95Ethanol and Synthetic Fuels .............................................................. 97Coal Supply and Demand ................................................................ 98Coal Mine Employment and Coal Prices .................................................... 99Coal Transportation and Imports .......................................................... 100Coal Consumption ...................................................................... 101Coal Alternative Cases. .................................................................. 102Carbon Dioxide Emissions. ............................................................... 103Carbon Dioxide and Sulfur Dioxide Emissions ............................................... 104Nitrogen Oxide and Mercury Emissions. .................................................... 105Energy Information Administration / Annual Energy Outlook 2006iii

ContentsPageForecast Comparisons ...............................................................107List of Acronyms ....................................................................119Notes and Sources ..................................................................120AppendixesA. Reference Case ...................................................................... 133B. Economic Growth Case Comparisons .................................................... 165C. Price Case Comparisons. .............................................................. 173D. Results from Side Cases ............................................................... 184E. NEMS Overview and Brief Description of Cases ........................................... 199F. Regional Maps. ...................................................................... 213G. Conversion Factors. .................................................................. 221Tables1. Total energy supply and disposition in the AEO2006 reference case: summary, 2003-2030. ........ 112. CARB emissions standards for light-duty vehicles, model years 2009-2016 ...................... 233. Proposed light truck CAFE standards by model year and footprint category .................... 234. Key projections for light truck fuel economy in the alternative CAFE standards case, 2011-2030. ... 245. Basic features of State renewable energy requirements and goals enacted since 2003 ............. 256. Major changes in existing State renewable energy requirements and goals since 2003 ............ 257. New U.S. renewable energy capacity, 2004-2005 ........................................... 268. Estimates of national trends in annual emissions of sulfur dioxide and nitrogen oxides, 2003-2020. . 289. Macroeconomic model estimates of economic impacts from oil price increases ................... 3510. Time-series estimates of economic impacts from oil price increases ............................ 3611. Summary of U.S. oil price-GDP elasticities. ............................................... 3612. Economic indicators in the reference, high price, and low price cases, 2005-2030. ................ 3813. Technologies expected to have significant impacts on new light-duty vehicles ................... 5114. Nonconventional liquid fuels production in the AEO2006 reference and high price cases, 2030 ..... 5215. Projected changes in U.S. greenhouse gas emissions, gross domestic product,and greenhouse gas intensity, 2002-2020 ................................................. 6016. Costs of producing electricity from new plants, 2015 and 2030. ............................... 7817. Technically recoverable U.S. natural gas resources as of January 1, 2004. ...................... 8818. Technically recoverable U.S. crude oil resources as of January 1, 2004 ......................... 9119. Forecasts of annual average economic growth, 2004-2030. ................................... 10820. Forecasts of world oil prices, 2010-2030 .................................................. 10821. Forecasts of average annual growth rates for energy consumption, 2004-2030 ................... 10922. Comparison of electricity forecasts, 2015 and 2030 ......................................... 11123. Comparison of natural gas forecasts, 2015, 2025, and 2030 .................................. 11224. Comparison of petroleum forecasts, 2015 and 2030 ......................................... 11525. Comparison of coal forecasts, 2015, 2025, and 2030. ........................................ 116Figures1. Energy prices, 1980-2030 .............................................................. 42. Delivered energy consumption by sector, 1980-2030 ........................................ 53. Energy consumption by fuel, 1980-2030 .................................................. 64. Energy use per capita and per dollar of gross domestic product, 1980-2030 ..................... 65. Electricity generation by fuel, 1980-2030 ................................................. 76. Total energy production and consumption, 1980-2030 ...................................... 87. Energy production by fuel, 1980-2030. ................................................... 88. Projected U.S. carbon dioxide emissions by sector and fuel, 1990-2030 ......................... 109. Sulfur dioxide emissions in selected States, 1990-2003 ...................................... 2810. World oil prices in the AEO2005 and AEO2006 reference cases ............................... 32iv Energy Information Administration / Annual Energy Outlook 2006

ContentsFigures (Continued)Page11. World oil prices in three AEO2006 cases, 1980-2030 ........................................ 3312. Changes in world oil price and U.S. real GDP in the AEO2006 high and low price cases, 2004-2030 . 3713. GDP elasticities with respect to oil price changes in the high price case, 2006-2030. .............. 3814. Purvin & Gertz forecast for world oil production by crude oil quality, 1990-2020. ................ 4015. Sulfur content specifications for U.S. petroleum products, 1990-2014 .......................... 4016. U.S. hydrotreating capacity, 1990-2030 .................................................. 4017. Market penetration of advanced technologies in new cars, 2004 and 2030 ...................... 5118. Market penetration of advanced technologies in new light trucks, 2004 and 2030 ................ 5119. System elements for production of synthetic fuels from coal, natural gas, and biomass ............ 5420. Mercury emissions from the electricity generation sector, 2002-2030 .......................... 5921. Mercury allowance prices, 2010-2030 .................................................... 5922. Coal-fired generating capacity retrofitted with activated carbon injection systems, 2010-2030 ...... 5923. Projected change in U.S. greenhouse gas intensity in three cases, 2002-2020 .................... 6024. Average annual growth rates of real GDP, labor force, and productivity in three cases, 2004-2030 . . 6225. Average annual unemployment, interest, and inflation rates, 2004-2030 ....................... 6226. Sectoral composition of output growth rates, 2004-2030 ..................................... 6327. Sectoral composition of manufacturing output growth rates, 2004-2030 ........................ 6328. Energy expenditures as share of gross domestic product, 1970-2030 ........................... 6329. World oil prices in three cases, 1980-2030. ................................................ 6430. U.S. gross petroleum imports by source, 2004-2030. ........................................ 6431. Energy use per capita and per dollar of gross domestic product, 1980-2030 ..................... 6532. Primary energy use by fuel, 2004-2030 ................................................... 6533. Delivered energy use by fuel, 1980-2030 .................................................. 6634. Primary energy consumption by sector, 1980-2030 ......................................... 6635. Delivered residential energy consumption per capita by fuel, 1980-2030 ........................ 6736. Delivered residential energy consumption by end use, 2001, 2004, 2015, and 2030 ............... 6737. Efficiency indicators for selected residential appliances, 2004 and 2030 ........................ 6838. Variation from reference case delivered residential energy use in three alternative cases,2004-2030 .......................................................................... 6839. Delivered commercial energy consumption per capita by fuel, 1980-2030 ....................... 6940. Delivered commercial energy intensity by end use, 2004, 2015, and 2030 ....................... 6941. Efficiency indicators for selected commercial equipment, 2004 and 2030 ....................... 7042. Variation from reference case delivered commercial energy intensity in three alternative cases,2004-2030 .......................................................................... 7043. Buildings sector electricity generation from advanced technologies in two alternative cases, 2030. . . 7144. Industrial energy intensity by two measures, 1980-2030. .................................... 7145. Energy intensity in the industrial sector, 2004 ............................................ 7246. Average growth in industrial output and delivered energy consumption by sector, 2004-2030 ...... 7247. Projected energy intensity in 2030 relative to 2004, by industry .............................. 7348. Variation from reference case delivered industrial energy use in two alternative cases, 2004-2030 . . 7349. Transportation energy use per capita, 1980-2030 .......................................... 7450. Transportation travel demand by mode, 1980-2030. ........................................ 7451. Average fuel economy of new light-duty vehicles, 1980-2030 ................................. 7552. Sales of advanced technology light-duty vehicles by fuel type, 2015 and 2030. ................... 7553. Changes in projected transportation fuel use in two alternative cases, 2010 and 2030. ............ 7654. Changes in projected transportation fuel efficiency in two alternative cases, 2010 and 2030. ....... 7655. Annual electricity sales by sector, 1980-2030 .............................................. 7756. Electricity generation capacity additions by fuel type, including combined heat and power,2005-2030 .......................................................................... 7757. Electricity generation capacity additions, including combined heat and power, by region and fuel,2005-2030 .......................................................................... 7858. Levelized electricity costs for new plants, 2015 and 2030 .................................... 7859. Electricity generation from nuclear power, 1973-2030 ...................................... 79Energy Information Administration / Annual Energy Outlook 2006v

ContentsFigures (Continued)Page60. Additions of renewable generating capacity, 2004-2030 ..................................... 7961. Levelized and avoided costs for new renewable plants in the Northwest, 2015 and 2030 ........... 8062. Electricity generation by fuel, 2004 and 2030. ............................................. 8063. Grid-connected electricity generation from renewable energy sources, 1980-2030 ................ 8164. Nonhydroelectric renewable electricity generation by energy source, 2004-2030 ................. 8165. Fuel prices to electricity generators, 1995-2030 ............................................ 8266. Average U.S. retail electricity prices, 1970-2030 ........................................... 8267. Cumulative new generating capacity by technology type in three economic growth cases,2004-2030 .......................................................................... 8368. Cumulative new generating capacity by technology type in three fossil fuel technology cases,2004-2030 .......................................................................... 8369. Levelized electricity costs for new plants by fuel type in two nuclear cost cases, 2015 and 2030 ..... 8470. Nonhydroelectric renewable electricity generation by energy source in three cases, 2010 and 2030 . . 8471. Natural gas consumption by sector, 1990-2030 ............................................ 8572. Increases in natural gas consumption by Census division, 2004-2030 .......................... 8573. Natural gas production by source, 1990-2030. ............................................. 8674. Net U.S. imports of natural gas by source, 1990-2030 ....................................... 8675. Lower 48 natural gas wellhead prices, 1990-2030 .......................................... 8776. Natural gas prices by end-use sector, 1990-2030 ........................................... 8777. Lower 48 natural gas wellhead prices in three technology cases, 1990-2030 ..................... 8878. Natural gas production and net imports in three technology cases, 1990-2030 ................... 8879. Lower 48 natural gas wellhead prices in three price cases, 1990-2030 .......................... 8980. Net imports of liquefied natural gas in three price cases, 1990-2030 ........................... 8981. Net imports of liquefied natural gas in three LNG supply cases, 1990-2030 ..................... 9082. Lower 48 natural gas wellhead prices in three LNG supply cases, 1990-2030 .................... 9083. World oil prices in the reference case, 1990-2030 ........................................... 9184. Domestic crude oil production by source, 1990-2030 ........................................ 9185. World oil prices in three cases, 1990-2030. ................................................ 9286. Total U.S. petroleum production in three price cases, 1990-2030 .............................. 9287. U.S. syncrude production from oil shale in the high price case, 2004-2030 ...................... 9388. Total U.S. crude oil production in three technology cases, 1990-2030 .......................... 9389. Alaskan oil production in the reference and ANWR cases, 1990-2030 .......................... 9490. U.S. net imports of oil in the reference and ANWR cases, 1990-2030. .......................... 9491. Consumption of petroleum products, 1990-2030 ........................................... 9592. Domestic refinery distillation capacity, 1990-2030. ......................................... 9593. U.S. petroleum product demand and domestic petroleum supply, 1990-2030 .................... 9694. Components of retail gasoline prices, 2004, 2015, and 2030 .................................. 9695. U.S. ethanol fuel consumption in three price cases, 1995-2030. ............................... 9796. Coal-to-liquids and gas-to-liquids production in two price cases, 2004-2030 ..................... 9797. Coal production by region, 1970-2030 .................................................... 9898. Distribution of domestic coal by demand and supply region, 2003 and 2030 ..................... 9899. U.S. coal mine employment by region, 1970-2030 .......................................... 99100. Average minemouth price of coal by region, 1990-2030 ...................................... 99101. Changes in regional coal transportation rates, 2010, 2025, and 2030. .......................... 100102. U.S. coal exports and imports, 1970-2030 ................................................. 100103. Electricity and other coal consumption, 1970-2030 ......................................... 101104. Coal consumption in the industrial and buildings sectors and at coal-to-liquids plants,2004, 2015, and 2030 ................................................................. 101105. Projected variation from the reference case projection of U.S. total coal demand in four cases,2030 ............................................................................... 102106. Average delivered coal prices in three cost cases, 1990-2030. ................................. 102107. Carbon dioxide emissions by sector and fuel, 2004 and 2030 ................................. 103108. Carbon dioxide emissions in three economic growth cases, 1990-2030 .......................... 103vi Energy Information Administration / Annual Energy Outlook 2006

ContentsFigures (Continued)Page109. Carbon dioxide emissions in three technology cases, 2004, 2020, and 2030 ...................... 104110. Sulfur dioxide emissions from electricity generation, 1990-2030 .............................. 104111. Nitrogen oxide emissions from electricity generation, 1990-2030. ............................. 105112. Mercury emissions from electricity generation, 1995-2030 ................................... 105Energy Information Administration / Annual Energy Outlook 2006vii

Overview

OverviewEnergy Trends to 2030The Energy Information Administration (EIA), inpreparing projections for the Annual Energy Outlook2006 (AEO2006), evaluated a wide range of trendsand issues that could have major implications for U.S.energy markets between today and 2030. AEO2006 isthe first edition of the Annual Energy Outlook (AEO)to provide projections through 2030. This overviewfocuses on one case, the reference case, which is presentedand compared with the Annual Energy Outlook2005 (AEO2005) reference case.Trends in energy supply and demand are affected by alarge number of factors that are difficult to predict,such as energy prices, U.S. economic growth,advances in technologies, changes in weather patterns,and future public policy decisions. In preparingAEO2006, EIA reevaluated its prior expectationsabout world oil prices in light of the current circumstancesin oil markets. Since 2000, world oil priceshave risen sharply as supply has tightened, first as aresult of strong demand growth in developing economiessuch as China and later as a result of supply constraintsresulting from disruptions and inadequateinvestment to meet demand growth. As a result ofthis review, the AEO2006 reference case includesmuch higher world oil prices than were projected inAEO2005. In the AEO2006 reference case, worldcrude oil prices, which are now expressed in terms ofthe average price of imported low-sulfur crude oil toU.S. refiners, are projected to increase from $40.49per barrel (2004 dollars) in 2004 to $54.08 per barrelin 2025 (about $21 per barrel higher than the projected2025 price in AEO2005) and to $56.97 per barrelin 2030.The higher world oil prices in the AEO2006 referencecase have important implications for energy markets.The most significant impact is on the outlook for U.S.petroleum imports. Net imports of petroleum are projectedto meet a growing share of total petroleumdemand in both AEO2006 and AEO2005; however,the higher world oil prices in the AEO2006 referencecase lead to more domestic crude oil production, lowerdemand for petroleum products, and consequentlylower levels of petroleum imports. Net petroleumimports are expected to account for 60 percent ofdemand (on the basis of barrels per day) in 2025 in theAEO2006 reference case, up from 58 percent in 2004.In the AEO2005 reference case, net petroleumimports were projected to account for 68 percent ofU.S. petroleum demand in 2025.Higher world oil prices are also projected to affectfuel choice and vehicle efficiency decisions in thetransportation sector. Higher oil prices increase thedemand for unconventional sources of transportationfuel, such as ethanol and biodiesel, and are projectedto stimulate coal-to-liquids (CTL) production in thereference case. In some of the alternative AEO2006cases, with even higher oil prices, domestic productionof liquid fuels from natural gas—“gas-to-liquids”(GTL)—is also stimulated. The production of alternativeliquid fuels is highly sensitive to oil price levels.The projected fuel economy of new light-duty vehiclesin the AEO2006 reference case in 2025 is higher thanwas projected in the AEO2005 reference case, primarilybecause of higher petroleum prices. The AEO2006reference case does not include implementation of theproposed, but not yet final, increase in fuel economystandards based on vehicle footprint for lighttrucks—including pickups, sport utility vehicles, andminivans—for model years 2008 through 2011.Much of the increase in new light-duty vehicle fueleconomy in the AEO2006 reference case reflectsgreater penetration by hybrid and diesel vehicles.Sales of “full hybrid” vehicles in 2025 are 31 percent(340,000 vehicles) higher in the AEO2006 referencecase, and diesel vehicle sales are 29 percent (290,000vehicles) higher, than projected in the AEO2005 referencecase. In spite of the higher projected sales ofhybrid (1.5 million) and diesel (1.3 million) vehicles in2025, each is expected to account for only 7 percent ofnew vehicle sales in the AEO2006 reference case,even though the projected hybrid sales are higherthan current industry expectations. The projectedWorld Oil Price Concept Used inAEO2006In previous AEOs, the world crude oil price wasdefined on the basis of the average importedrefiner acquisition cost of crude oil to the UnitedStates (IRAC), which represented the weightedaverage of all imported crude oil. Historically, theIRAC price has tended to be a few dollars less thanthe widely cited prices of premium crudes, such asWest Texas Intermediate (WTI) and Brent, whichrefiners generally prefer for their low viscosity andsulfur content. In the past 2 years, the price differencebetween premium crudes and IRAC has widened—inparticular, the price spread betweenpremium crudes and heavier, high-sulfur crudes.In an effort to provide a crude oil price that is moreconsistent with those generally reported in themedia, AEO2006 uses the average price ofimported low-sulfur crude oil to U.S. refiners.2 Energy Information Administration / Annual Energy Outlook 2006

Overviewsales figures for hybrids do not include sales of “mildhybrids,” which like full hybrids incorporate an integratedstarter generator, that allows for improvedefficiency by shutting the engine off when the vehicleis idling, but do not incorporate an electric motor thatprovides tractive power to the vehicle when it ismoving.The AEO2006 reference case includes minimalmarket penetration by hydrogen fuel cell vehicles,as a result of State mandates. Although significantresearch and development (R&D) is being conductedthrough the FreedomCAR Program, a co-funded partnershipbetween the Federal Government and privateindustry, those efforts are not expected to have a significantimpact on the market for fuel cell vehiclesbefore 2030.The AEO2006 reference case projection for U.S.imports of liquefied natural gas (LNG) is lower thanwas projected in the AEO2005 reference case. LNGimports are projected to grow from 0.6 trillion cubicfeet in 2004 to 4.1 trillion cubic feet in 2025, as comparedwith 6.4 trillion cubic feet in the AEO2005 referencecase. More rapid growth in worldwide demandfor natural gas in the AEO2006 reference casereduces the availability of LNG supplies to the UnitedStates and raises worldwide natural gas prices, makingLNG less economical in U.S. markets.AEO2006 includes consideration of the impacts of theEnergy Policy Act of 2005 (EPACT2005), signed intolaw on August 8, 2005. Consistent with the generalapproach adopted in the AEO, the reference case doesnot consider those sections of EPACT2005 thatrequire funding appropriations for implementationor sections with highly uncertain impacts on energymarkets. For example, EIA does not try to anticipatethe policy response to the many studies required byEPACT2005 or the impacts of the R&D fundingauthorizations included in the bill. The AEO2006 referencecase includes only those sections of EPACT-2005 that establish specific tax credits, incentives, orstandards—about 30 of the roughly 500 sections inthe legislation.Of the EPACT2005 provisions analyzed, incentivesintended to stimulate the development of advancednuclear and renewable plants have particularly noteworthyimpacts. A total of 6 gigawatts of newly constructednuclear capacity is projected to be added by2030 in the AEO2006 reference case as a result of theincentives in EPACT2005.EPACT2005 also has important implications forenergy consumption in the residential and commercialsectors. In the residential sector, EPACT2005sets efficiency standards for torchiere lamps, dehumidifiers,and ceiling fans and creates tax credits forenergy-efficient furnaces, water heaters, and air conditioners.It also allows home builders to claim taxcredits for energy-efficient new construction. In thecommercial sector, the legislation creates efficiencystandards that affect energy use in a number of commercialapplications. It also includes investment taxcredits for solar technologies, fuel cells, and microturbines.These policies are expected to help reduceenergy use for space conditioning and lighting in bothsectors.Economic GrowthThe projections for key interest rates—the Federalfunds rate, the nominal yield on the 10-year Treasurynote, and the AA utility bond rate—in the AEO2006reference case are slightly lower than those in theAEO2005 reference case. Also, the projected value ofindustrial shipments has been revised downward, inpart in response to the higher projected energy pricesin the AEO2006 reference case.Despite the higher forecast for energy prices, grossdomestic product (GDP) is projected to grow at anaverage annual rate of 3.0 percent from 2004 to 2030in AEO2006, identical to the projected growth ratefrom 2004 through 2025 in AEO2005. The ratio offinal energy expenditures to GDP has generally fallenover time and was only about 0.07 in 2004, down froma high of 0.14 during the 1970s. It is projected to fallto about 0.05 in 2030 as a result of continued declinesin energy use per unit of output and growth in otherareas of the economy. The main factors influencinglong-term economic growth are growth in the laborforce and sustained growth in labor productivity, notenergy prices.Energy PricesIn the reference case—one of several cases included inAEO2006—the average world crude oil price continuesto rise through 2006 and then declines to $46.90per barrel in 2014 (2004 dollars) as new supplies enterthe market. It then rises slowly to $54.08 per barrel in2025 (Figure 1), about $21 per barrel higher than theprice in AEO2005 ($32.95 per barrel). AlternativeAEO2006 cases address higher and lower world oilprices.The prices in the AEO2006 reference case reflect ashift in EIA’s thinking about long-term trends in oilmarkets. World oil markets have been extremely volatilefor the past several years, and EIA now believesthat the price path in AEO2005 did not fully reflectthe causes of that volatility and the implications forEnergy Information Administration / Annual Energy Outlook 2006 3

Overviewlong-term average oil prices. In the AEO2006 referencecase, the combined production capacity of membersof the Organization of the Petroleum ExportingCountries (OPEC) does not increase as much as previouslyprojected, and consequently world oil suppliesare assumed to remain tight. The United States andemerging Asia—notably, China— are expected to leadthe increase in demand for world oil supplies, keepingpressure on prices though 2030.In the AEO2006 reference case, world petroleumdemand is projected to increase from about 82 millionbarrels per day in 2004 to 111 million barrels per dayin 2025. The additional demand is expected to be metby increased oil production from both OPEC andnon-OPEC nations. In AEO2005, world petroleumdemand was projected to reach a higher level of 121million barrels per day in 2025. The AEO2006 referencecase projects OPEC oil production of 44 millionbarrels per day in 2025, 44 percent higher than the 31million barrels per day produced in 2004. In the AEO-2005 reference case, OPEC production was projectedto reach 55 million barrels per day in 2025, more than11 million barrels per day higher than in the AEO-2006 reference case. In the AEO2006 reference case,non-OPEC oil production increases from 52 millionbarrels per day in 2004 to 67 million in 2025, as comparedwith the AEO2005 reference case projection of65 million barrels per day.The average U.S. wellhead price for natural gas in theAEO2006 reference case declines gradually from thecurrent level as increased drilling brings on new suppliesand new import sources become available. Theaverage price falls to $4.46 per thousand cubic feet in2016 (2004 dollars), then rises gradually to more than$5.40 per thousand cubic feet in 2025 (equivalent toabout $10 per thousand cubic feet in nominal dollars)and more than $5.90 per thousand cubic feet in 2030.Figure 1. Energy prices, 1980-2030 (2004 dollars permillion Btu)3530252015105HistoryProjectionsElectricityPetroleum0Coal1980 1995 2004 2015 2030Natural gasLNG imports, Alaskan natural gas production, andlower 48 production from unconventional sources arenot expected to increase sufficiently to offset theimpacts of resource depletion and increased demand.The projected wellhead natural gas prices in theAEO2006 reference case from 2016 to 2025 are consistentlyhigher than the comparable prices in theAEO2005 reference case, by about 30 to 60 cents perthousand cubic feet, primarily as a result of higherexploration and development costs.In the AEO2006 reference case, the combination ofslow but continued improvements in expected mineproductivity and a continuing shift to low-cost coalfrom the Powder River Basin in Wyoming leads to agradual decline in the projected average minemouthcoal price, to approximately $20.00 per ton ($1.00 permillion British thermal units [Btu]) in 2021 (2004 dollars).Prices then increase slowly as rising natural gasprices and the need for baseload generating capacitylead to the construction of many new coal-fired generatingplants. In 2025, the average minemouth price inthe AEO2006 reference case is projected to be $20.63per ton ($1.03 per million Btu), an increase over theAEO2005 reference case projection of $18.64 per ton($0.93 per million Btu). Trends in coal prices measuredin terms of tonnage differ slightly from thetrends in prices measured in terms of energy content,because the average energy content per ton of coalconsumed falls over time as Western subbituminouscoal, which has a relatively low Btu content, claims alarger share of the market.Average delivered electricity prices are projected todecline from 7.6 cents per kilowatthour (2004 dollars)in 2004 to a low of 7.1 cents per kilowatthour in 2015as a result of declines in natural gas prices and, to alesser extent, coal prices. After 2015, average realelectricity prices are projected to increase, to 7.4 centsper kilowatthour in 2025 and 7.5 cents per kilowatthourin 2030. In the AEO2005 reference case,electricity prices were lower in the early years of theprojection but reached about the same level in 2025.The higher near-term electricity prices projected inthe AEO2006 reference case result primarily fromhigher expected fuel costs for natural-gas- and coalfiredelectric power plants.Energy ConsumptionTotal primary energy consumption in the AEO2006reference case is projected to increase at an averagerate of 1.2 percent per year, from 99.7 quadrillion Btuin 2004 to 127.0 quadrillion Btu in 2025—6.2 quadrillionBtu less than in AEO2005. In 2025, coal, nuclear,and renewable energy consumption are higher—4 Energy Information Administration / Annual Energy Outlook 2006

Overviewwhile petroleum and natural gas consumption arelower—in the AEO2006 reference case than in AEO-2005. Among the most important factors accountingfor the differences are higher energy prices, particularlyfor petroleum and natural gas; lower projectedgrowth rates in the manufacturing portion of theindustrial sector, which traditionally includes themost energy-intensive industries; greater penetrationby hybrid and diesel vehicles in the transportationsector as consumers focus more on fuelefficiency; and the impacts of the recently passedEPACT2005, which are projected to reduce energyconsumption in the residential and commercial sectorsand slow the growth of electricity demand.As a result of demographic trends and housing preferences,delivered residential energy consumption inthe AEO2006 reference case is projected to grow from11.4 quadrillion Btu in 2004 to 13.6 quadrillion Btu in2025 (Figure 2), 0.6 quadrillion Btu lower than inAEO2005. Higher projected energy prices in AEO-2006 and the impacts of EPACT2005 are expected tohelp reduce energy consumption for space conditioningand lighting.Consistent with projected growth in commercialfloorspace in the AEO2006 reference case, deliveredcommercial energy consumption is projected to reach11.5 quadrillion Btu in 2025. In comparison, theAEO2005 reference case projected 12.5 quadrillionBtu of commercial delivered energy consumption in2025. Three changes contribute to the lower projectionin AEO2006: significantly higher fossil fuelenergy prices, adoption of a revised projection of commercialfloorspace based on updated historical data,and the impacts of the EPACT2005 provisionsincluded in the reference case.After falling to relatively low levels in the early 1980s,industrial energy consumption recovered and peakedFigure 2. Delivered energy consumptionby sector, 1980-2030 (quadrillion Btu)40302010HistoryProjections01980 1995 2004 2015 2030TransportationIndustrialResidentialCommercialin 1997. In the 2000 to 2003 period, industrial sectoractivity was reduced by an economic recession. Theindustrial sector is projected to experience more typicaloutput growth rates over the AEO2006 projectionperiod, and industrial energy consumption isexpected to reflect this trend. The industrial value ofshipments in the AEO2006 reference case is projectedto grow by 2.0 percent per year from 2004 to 2025,more slowly than in AEO2005 (2.2 percent per year)due to a slight slowdown in projected investmentspending, higher energy prices, and increased competitionfrom imports. Delivered industrial energy consumptionin the AEO2006 reference case is projectedto reach 30.6 quadrillion Btu in 2025, slightly lowerthan the AEO2005 projection of 30.8 quadrillion Btu.The AEO2006 projection includes 1.2 quadrillion Btuof coal consumption in CTL plants, which was notincluded in AEO2005.Delivered energy consumption in the transportationsector in the AEO2006 reference case is projected tototal 37.3 quadrillion Btu in 2025, 2.7 quadrillion Btulower than the AEO2005 projection. The lower levelof consumption reflects both slower growth in milestraveled and higher vehicle efficiency. Over the past20 years, light-duty vehicle travel has grown by about3 percent annually. In the AEO2006 reference case itis projected to grow at a rate of 1.8 percent per yearthrough 2025 (as compared with 2.1 percent per yearin AEO2005), reflecting demographic factors (forexample, the leveling off of increases in the labor forceparticipation rate for women) and higher energyprices. The projected average fuel economy of newlight-duty vehicles in 2025 is also higher in the AEO-2006 reference case than was projected in AEO2005,primarily because the higher projected fuel prices inthe AEO2006 forecast are expected to lead consumersto demand better fuel economy, slowing the growth insales of new pickup trucks and sport utility vehicles.Total electricity consumption, including both purchasesfrom electric power producers and on-sitegeneration, is projected to grow from 3,729 billionkilowatthours in 2004 to 5,208 billion kilowatthoursin 2025, increasing at an average annual rate of 1.6percent in the AEO2006 reference case. In comparison,total electricity consumption of 5,467 billionkilowatthours in 2025 was projected in AEO2005.Growth in electricity use for computers, office equipment,and a variety of electrical appliances in theend-use sectors is partially offset in the AEO2006 referencecase by improved efficiency in these and other,more traditional, electrical applications.Total consumption of natural gas in the AEO2006 referencecase is projected to increase from 22.4 trillionEnergy Information Administration / Annual Energy Outlook 2006 5

Overviewcubic feet in 2004 to 27.0 trillion cubic feet in 2025(Figure 3), 3.7 trillion cubic feet lower than projectedin the AEO2005 reference case, mostly as a result ofhigher natural gas prices. After peaking at 27.0 trillioncubic feet in 2024, natural gas consumption isprojected to fall slightly by 2030, as higher natural gasprices result in a larger market share for coal in theelectric power sector in the later years of theprojection. The projected growth in natural gasdemand in AEO2006 results primarily from increaseduse of natural gas for electricity generation andindustrial applications, which together account for 62percent of the projected demand growth from 2004 to2025. In addition, demand for natural gas in the residentialand commercial sectors is projected to grow by1.5 trillion cubic feet in total from 2004 to 2025.In the AEO2006 reference case, total coal consumptionis projected to increase from 1,104 million shorttons in 2004 to 1,592 million short tons in 2025(Figure 3), 84 million short tons more than the 1,508million tons projected to be consumed in 2025 in theAEO2005 reference case. Coal consumption is projectedto grow at a faster rate in AEO2006 toward theend of the projection, particularly after 2020, as coalcaptures market share from natural gas, and as coaluse for CTL production grows. Coal was not projectedto be used for CTL production in the AEO2005 referencecase. In the AEO2006 reference case, coal consumptionin the electric power sector is projected toincrease from 1,235 million short tons in 2020 to1,502 million short tons in 2030, at an average rate of2.0 percent per year; and coal use at CTL plants isprojected to increase from 62 million short tons in2020 to 190 million short tons in 2030.Total petroleum consumption is projected to growfrom 20.8 million barrels per day in 2004 to 26.1 millionbarrels per day in 2025 (Figure 3) in the AEO-2006 reference case (1.9 million barrels per day lowerFigure 3. Energy consumption by fuel, 1980-2030(quadrillion Btu)6050403020HistoryProjectionsPetroleumCoalNatural gasNuclear10Nonhydrorenewables0Hydro1980 1995 2004 2015 2030than the AEO2005 projection). Petroleum demandgrowth in the AEO2006 reference case is lower in allsectors than was projected in AEO2005, due largely tothe impact of the much higher oil prices in AEO2006.Most of the difference—almost two-thirds—is in thetransportation sector.Total consumption of marketed renewable fuels inthe AEO2006 reference case (including ethanol forgasoline blending, of which 1.0 quadrillion Btu in2025 is included with “petroleum products” consumption)is projected to grow from 6.0 quadrillionBtu in 2004 to 9.6 quadrillion Btu in 2025 (Figure 3),as a result of State programs—renewable portfoliostandards (RPS), mandates, and goals—for renewableelectricity generation, technological advances,higher petroleum and natural gas prices, and theeffects of Federal tax credits, including those inEPACT2005. In AEO2005, total marketed renewablefuel consumption was projected to grow to 8.5 quadrillionBtu in 2025. In AEO2006, more than 60 percentof the projected demand for renewables in thereference case is for grid-related electricity generation,including combined heat and power (CHP), andthe rest is for dispersed heating and cooling, industrialuses, and fuel blending.Energy IntensityEnergy intensity, measured as energy use per dollarof GDP (2000 dollars), is projected to decline at anaverage annual rate of 1.8 percent from 2004 to 2030in the AEO2006 reference case (Figure 4), with efficiencygains and structural shifts in the economydampening growth in demand for energy services.The rate of decline in energy intensity is faster thanthe 1.6-percent annual rate of decline projected inAEO2005 between 2004 and 2025, largely because ofhigher energy prices in AEO2006, resulting in generallylower projected levels of energy consumption.Figure 4. Energy use per capita and per dollar ofgross domestic product, 1980-2030 (index, 1980 = 1)1.21.00.80.60.40.2HistoryProjections0.01980 1995 2004 2015 2030Energyuse percapitaEnergyuse perdollarof GDP6 Energy Information Administration / Annual Energy Outlook 2006

OverviewSince 1992, the energy intensity of the U.S. economyhas declined on average by 1.9 percent per year, andthe share of total industrial production accounted forby the energy-intensive industries has fallen sharply,by 1.3 percent per year on average from 1992 to 2004.In the AEO2006 reference case, the energy-intensiveindustries’ share of total industrial output is projectedto continue to decline, but at a slower rate of 0.8percent per year, leading to a slower rate of reductionin energy intensity.Historically, energy use per person has varied overtime with the level of economic growth, weather conditions,and energy prices, among many other factors.During the late 1970s and early 1980s, energy consumptionper capita fell in response to high energyprices and weak economic growth. Starting in the late1980s and lasting through 2000, energy consumptionper capita generally increased with declining energyprices and strong economic growth. Per capita energyuse is projected to increase in the AEO2006 referencecase, with growth in demand for energy services onlypartially offset by efficiency gains. Per capita energyuse increases by an average of 0.3 percent per yearbetween 2004 and 2030 in the AEO2006 referencecase, less than was projected in the AEO2005 referencecase, 0.5 percent per year between 2004 and2025, primarily because of the higher projectedenergy prices in AEO2006.Recently, as energy prices have risen, the potentialfor more energy conservation has received increasedattention. Although some additional energy conservationis induced by higher energy prices in the AEO-2006 reference case, no policy-induced conservationmeasures are assumed beyond those in existing legislationand regulation, nor does the reference caseassume behavioral changes beyond those observed inthe past.Electricity GenerationIn the AEO2006 reference case, the projected averageprices of natural gas and coal delivered to electricitygenerators in 2025 are, respectively, 31 cents and11 cents per million Btu higher than the comparableprices in AEO2005. Although the projected levels ofcoal consumption for electricity generation in 2025are similar in the two forecasts, higher natural gasprices and slower growth in electricity demand inAEO2006 lead to significantly lower levels of naturalgas consumption for electricity generation. As aresult, projected cumulative capacity additions andgeneration from natural-gas-fired power plants arelower in the AEO2006 reference case, and capacityadditions and generation from coal-fired power plantsthrough 2025 are similar to those in AEO2005.Inthelater years of the AEO2006 projection, natural-gasfiredgeneration is expected to decline, displaced bygeneration from new coal-fired plants (Figure 5). TheAEO2006 projection of 1,070 billion kilowatthours ofelectricity generation from natural gas in 2025 is 24percent lower than the AEO2005 projection of 1,406billion kilowatthours.In the AEO2006 reference case, the natural gas shareof electricity generation (including generation in theend-use sectors) is projected to increase from 18 percentin 2004 to 22 percent around 2020, before fallingto 17 percent in 2030. The coal share is projected todecline slightly, from 50 percent in 2004 to 49 percentin 2020, before increasing to 57 percent in 2030. Additionsto coal-fired generating capacity in the AEO-2006 reference case are projected to total 102gigawatts between 2004 and 2025, as compared with86 gigawatts in AEO2005. Over the entire periodfrom 2004 to 2030, 174 gigawatts of new coal-firedgenerating capacity is projected to be added in theAEO2006 reference case, including 19 gigawatts atCTL plants.Nuclear generating capacity in the AEO2006 referencecase is projected to increase from about 100gigawatts in 2004 to about 109 gigawatts in 2019 andto remain at that level (about 10 percent of total U.S.generating capacity) through 2030. The total projectedincrease in nuclear capacity between 2004 and2030 includes 3 gigawatts expected to come fromuprates of existing plants that continue operating and6 gigawatts of capacity at newly constructed powerplants, stimulated by the provisions in EPACT2005,that are expected to begin operation between 2014and 2020.Additional nuclear capacity is projected in some of thealternative AEO2006 cases. Total electricity generationfrom nuclear power plants is projected to growFigure 5. Electricity generation by fuel, 1980-2030(billion kilowatthours)4,0003,0002,0001,0002,094HistoryElectricity demand5,6191980 2030ProjectionsCoalNatural gasNuclearRenewables0Petroleum1980 1995 2004 2015 2030Energy Information Administration / Annual Energy Outlook 2006 7

Overviewfrom 789 billion kilowatthours in 2004 to 871 billionkilowatthours in 2030 in the AEO2006 reference case,accounting for about 15 percent of total generation in2030.The use of renewable technologies for electricity generationis projected to grow, stimulated by improvedtechnology, higher fossil fuel prices, and extended taxcredits in EPACT2005 and in State renewable energyprograms (RPS, mandates, and goals). The expectedimpacts of State RPS programs, which specify aminimum share of generation or sales from renewablesources, are included in the projection. TheAEO2006 reference case also includes the extensionand expansion of the Federal tax credit for renewablegeneration through December 31, 2007, as enactedin EPACT2005. Total renewable generation in theAEO2006 reference case, including CHP, is projectedto grow by 1.7 percent per year, from 358 billion kilowatthoursin 2004 to 559 billion kilowatthours in2030.The Clean Air Interstate Rule (CAIR) and the CleanAir Mercury Rule (CAMR), issued by the U.S. EnvironmentalProtection Agency (EPA) in March 2005,are expected to result in large reductions of pollutantemissions from power plants. In the AEO2006 referencecase, projected emissions of sulfur dioxide (SO 2 )from electric power plants in 2025 are 58 percentlower, emissions of nitrogen oxide 50 percent lower,and emissions of mercury 70 percent lower than projectedin the AEO2005 reference case.Energy Production and ImportsNet imports of energy on a Btu basis are projected tomeet a growing share of total U.S. energy demand(Figure 6). In the AEO2006 reference case, netimports are expected to constitute 32 percent and 33percent of total U.S. energy consumption in 2025 and2030, respectively, up from 29 percent in 2004. InFigure 6. Total energy production andconsumption, 1980-2030 (quadrillion Btu)150 History Projections12510075502501980 1995 2004 2015 2030ConsumptionNet importsProductioncomparison, the AEO2005 reference case projected a38-percent share for net imports in 2025. Higher projectionsfor crude oil and natural gas prices in AEO-2006 are expected to lead to increases in domesticenergy production (Figure 7) and reductions indemand, reducing the projected growth in imports ascompared with the AEO2005 projections.The projections for U.S. crude oil production, domesticpetroleum supply, and net petroleum imports inthe AEO2006 reference case are also significantly differentfrom those in AEO2005. U.S. crude oil productionin the AEO2006 reference case is projected toincrease from 5.4 million barrels per day in 2004 to apeak of 5.9 million barrels per day in 2014 as a resultof increased production offshore, predominantly fromthe deep waters of the Gulf of Mexico. Production isthen projected to fall to 4.6 million barrels per day in2030. In the AEO2005 reference case, U.S. crude oilproduction was projected to peak in 2009 at 6.2 millionbarrels per day and then fall to 4.7 million barrelsper day in 2025.Total domestic petroleum supply (crude oil, naturalgas plant liquids, refinery processing gains, and otherrefinery inputs) follows the same pattern as crude oilproduction in the AEO2006 reference case, increasingfrom 8.6 million barrels per day in 2004 to a peak of10.5 million barrels per day in 2021, then declining to10.4 million barrels per day in 2025 and remaining atabout that level through 2030. The AEO2005 projectionfor total domestic petroleum supply in 2025 waslower, at 8.8 million barrels per day.In 2025, net petroleum imports, including both crudeoil and refined products, are expected to account for60 percent of demand (on the basis of barrels per day)in the AEO2006 reference case, up from 58 percent in2004. In AEO2005, net petroleum imports accountedfor 68 percent of demand in 2025. The market shareFigure 7. Energy production by fuel, 1980-2030(quadrillion Btu)HistoryProjections35 Coal302520Natural gas15Petroleum10NuclearNonhydro5renewablesHydro01980 1995 2004 2015 20308 Energy Information Administration / Annual Energy Outlook 2006

Overviewof net petroleum imports grows to 62 percent ofdemand in 2030 in the AEO2006 reference case.Despite an expected increase in distillation capacityat domestic refineries in AEO2006, net imports ofrefined petroleum products account for a growingportion of total net imports, increasing from 17 percentin 2004 to 22 percent in 2030.Total domestic natural gas production, excluding supplementalnatural gas supplies, increases from 18.5trillion cubic feet in 2004 to 21.6 trillion cubic feet in2019, before declining to 20.8 trillion cubic feet in2030 in the AEO2006 reference case. In 2025, domesticnatural gas production is projected to be 21.2 trillioncubic feet, compared with 21.8 trillion cubic feetin the AEO2005 reference case. The lower level ofdomestic natural gas production in the AEO2006 referencecase is entirely attributable to lower levels ofoffshore production. Offshore natural gas productionin 2025 is lower in the AEO2006 reference case than itwas in AEO2005, due at least in part to the impacts ofHurricanes Katrina and Rita, which are expected todelay offshore drilling projects because of a lack ofrigs and to have a long-term effect on production levelsas a result of the slow recovery of production fromexisting fields.The incorporation of EIA data showing a lower levelof new reserve discoveries in 2004 than had beenanticipated also affects the long-term forecast for offshorenatural gas production. Lower 48 offshore productionis projected to fall slightly from the 2004 levelof 4.3 trillion cubic feet and then grow steadilythrough 2015, peaking at 5.1 trillion cubic feet as newresources come on line in the Gulf of Mexico. After2015, lower 48 offshore production declines to 4.3 trillioncubic feet in 2025 and 4.0 trillion cubic feet in2030. In the AEO2005 reference case, offshore naturalgas production was projected to increase morequickly and reach higher levels, peaking at 5.3 trillioncubic feet in 2014 before falling to 4.9 trillion cubicfeet in 2025. The projection for onshore production ofnatural gas is also generally lower for most of the projectionperiod in the AEO2006 reference case thanwas projected in AEO2005. In the later years of theAEO2006 reference case, however, with higher naturalgas prices, onshore production grows strongly, to14.7 trillion cubic feet in 2025—equal to the AEO-2005 projection. Projected onshore production inAEO2006 remains at the 2025 level through 2030.Lower 48 production of unconventional natural gas isexpected to be a major contributor to growth in U.S.natural gas supplies. Unconventional natural gas productionis projected to account for 45 percent ofdomestic U.S. natural gas production in 2030, ascompared with the AEO2005 reference case projectionof 39 percent in 2025. In AEO2006, however,unconventional natural gas production is lower in themid-term (between 2006 and 2020) than was projectedin AEO2005. The lower levels of production inAEO2006 before 2021 reflect a decline in overall naturalgas consumption in response to higher prices.Starting in 2021, the projected levels of unconventionalnatural gas production in the AEO2006 referencecase are higher than those in AEO2005, reaching9.5 trillion cubic feet in 2030.Construction planning for the Alaska natural gaspipeline is expected to start soon, and the new pipelineis expected to be completed by 2015. When thepipeline goes into operation, Alaska’s total naturalgas production is projected to increase to 2.2 trillioncubic feet in 2025 (from 0.4 trillion cubic feet in 2004),the same level as projected in the AEO2005 referencecase.The projection for net U.S. pipeline imports of naturalgas from Canada and Mexico (predominantly Canada)in the AEO2006 reference case in 2025 is 1.3trillion cubic feet lower than was projected in AEO-2005. AEO2006 projects a continued decline in netpipeline imports, to 1.2 trillion cubic feet in 2030, as aresult of depletion effects and growing domesticdemand in Canada. The AEO2006 reference casereflects an expectation that growth in Canada’sunconventional natural gas production (primarilyfrom coal seams) will not be adequate to offset adecline in conventional production in Alberta, basedin part on data and projections from Canada’sNational Energy Board and other sources.Growth in LNG imports is projected to meet much ofthe increased demand for natural gas in the AEO2006reference case, but the increase is less than was projectedin the AEO2005 reference case. The growth inLNG imports is moderated by three factors: highernatural gas prices reduce domestic consumption;higher world oil prices increase worldwide demandfor natural gas and LNG imports, which raises theprice of LNG; and, to a lesser extent, higher world oilprices lead to higher foreign demand for GTL production,which uses more natural gas as a feedstock,further increasing the price pressure on natural gasand LNG. Except for expansions of three of the fourexisting onshore U.S. LNG terminals (Cove Point,Maryland; Elba Island, Georgia; and Lake Charles,Louisiana), the completion of U.S. terminals currentlyunder construction, and the addition of newfacilities to serve the Gulf Coast, Southern California,Florida, and New England, no other new facilities areEnergy Information Administration / Annual Energy Outlook 2006 9

Overviewprojected to be built to serve U.S. markets in theAEO2006 reference case.Total net imports of LNG to the United States in theAEO2006 reference case are projected to increasefrom 0.6 trillion cubic feet in 2004 to 4.1 trillion cubicfeet in 2025 (about two-thirds of the import volumesprojected in the AEO2005 reference case) and to 4.4trillion cubic feet in 2030. In some of the AEO2006alternative cases, however, particularly those withrelatively higher natural gas prices, additional LNGimports and new terminals are projected.As domestic coal demand grows in the AEO2006 referencecase, U.S. coal production increases at an averagerate of 1.5 percent per year, from 1,125 milliontons in 2004 to 1,530 million tons in 2025 (higherthan the 2025 projection of 1,488 million tons in AEO-2005) and to 1,703 million tons in 2030. Productionfrom mines west of the Mississippi River is expectedto provide the largest share of the incremental coalproduction. In 2030, almost 63 percent of coal productionis projected to originate from the western Statesif coal transportation costs remain stable.Typically, U.S. coal production is driven by demandfor electricity generation; however, projected electricitydemand in 2025 is lower in AEO2006 than in AEO-2005, and the projected demand for coal in the electricpower sector in 2025 is also lower (1,354 million tonsin the AEO2006 reference case, compared with 1,425million tons in the AEO2005 reference case), despitegreater reliance on coal for electric power generationin the AEO2006 forecast. The projected increase incoal production in AEO2006 is the result of higherlevels of coal use in CTL production, projected to growto 62 million short tons in 2020 and 190 million shorttons in 2030. No coal use for CTL production was projectedin the AEO2005 reference case.Carbon Dioxide EmissionsCarbon dioxide (CO 2 ) emissions from energy use areprojected to increase from 5,900 million metric tonsin 2004 to 7,587 million metric tons in 2025 and 8,114million metric tons in 2030 in the AEO2006 referencecase (Figure 8), an average annual increase of 1.2 percentper year. The CO 2 emissions intensity of the U.S.economy is projected to fall from 549 metric tons permillion dollars of GDP in 2004 to 377 metric tons permillion dollars of GDP in 2025, an average decline of1.8 percent per year, and to 351 metric tons per milliondollars of GDP in 2030. In comparison, theAEO2005 reference case projected a 1.5-percent averageannual decline in emissions intensity between2004 and 2025 and 8,062 million metric tons of CO 2emissions in 2025.Projected CO 2 emissions in 2025 are lower in all sectorsin the AEO2006 reference case than they were inAEO2005, as higher energy prices slow energy consumptiongrowth in all sectors. Total primary energyconsumption in 2025 is more than 6 quadrillion Btulower in AEO2006 than was projected in AEO2005.Some of the effect of the lower projected consumptionon CO 2 emissions in the AEO2006 reference case after2020 is offset by a proportionately higher share of coaluse for electricity generation and the increased use ofcoal at CTL plants.Figure 8. Projected U.S. carbon dioxide emissions bysector and fuel, 1990-2030 (million metric tons)10,000TransportationIndustrialCommercial8,000Residential6,0004,0002,00001990 2004 2010 2030CoalNatural gasPetroleum10 Energy Information Administration / Annual Energy Outlook 2006

OverviewTable 1. Total energy supply and disposition in the AEO2006 reference case: summary, 2003-2030Energy and economic factors 2003 2004 2010 2015 2020 2025 2030Average annualchange, 2004-2030Primary energy production (quadrillion Btu)Petroleum ............................. 14.40 13.93 14.83 14.94 14.41 13.17 12.25 -0.5%Dry natural gas ......................... 19.63 19.02 19.13 20.97 22.09 21.80 21.45 0.5%Coal. ................................. 22.12 22.86 25.78 25.73 27.30 30.61 34.10 1.6%Nuclear power. ......................... 7.96 8.23 8.44 8.66 9.09 9.09 9.09 0.4%Renewable energy ...................... 5.69 5.74 7.08 7.43 8.00 8.61 9.02 1.8%Other. ................................ 0.72 0.64 2.16 2.85 3.16 3.32 3.44 6.7%Total ................................ 70.52 70.42 77.42 80.58 84.05 86.59 89.36 0.9%Net imports (quadrillion Btu)Petroleum ............................. 24.19 25.88 26.22 28.02 30.39 33.11 36.49 1.3%Natural gas ............................ 3.39 3.49 4.45 5.23 5.15 5.50 5.72 1.9%Coal/other (- indicates export). ............. -0.45 -0.42 -0.58 0.20 0.90 1.54 2.02 NATotal ................................ 27.13 28.95 30.09 33.44 36.44 40.15 44.23 1.6%Consumption (quadrillion Btu)Petroleum products. ..................... 38.96 40.08 43.14 45.69 48.14 50.57 53.58 1.1%Natural gas ............................ 23.04 23.07 24.04 26.67 27.70 27.78 27.66 0.7%Coal. ................................. 22.38 22.53 25.09 25.66 27.65 30.89 34.49 1.7%Nuclear power. ......................... 7.96 8.23 8.44 8.66 9.09 9.09 9.09 0.4%Renewable energy ...................... 5.70 5.74 7.08 7.43 8.00 8.61 9.02 1.8%Other. ................................ 0.02 0.04 0.07 0.08 0.05 0.05 0.05 0.9%Total ................................ 98.05 99.68 107.87 114.18 120.63 126.99 133.88 1.1%Petroleum (million barrels per day)Domestic crude production ................ 5.69 5.42 5.88 5.84 5.55 4.99 4.57 -0.7%Other domestic production ................ 3.10 3.21 3.99 4.50 4.90 5.45 5.84 2.3%Net imports ............................ 11.25 12.11 12.33 13.23 14.42 15.68 17.24 1.4%Consumption. .......................... 20.05 20.76 22.17 23.53 24.81 26.05 27.57 1.1%Natural gas (trillion cubic feet)Production. ............................ 19.11 18.52 18.65 20.44 21.52 21.24 20.90 0.5%Net imports ............................ 3.29 3.40 4.35 5.10 5.02 5.37 5.57 1.9%Consumption. .......................... 22.34 22.41 23.35 25.91 26.92 26.99 26.86 0.7%Coal (million short tons)Production. ............................ 1,083 1,125 1,261 1,272 1,355 1,530 1,703 1.6%Net imports ............................ -18 -21 -26 5 36 63 83 NAConsumption. .......................... 1,095 1,104 1,233 1,276 1,390 1,592 1,784 1.9%Prices (2004 dollars)Imported low-sulfur light crude oil(dollars per barrel). ...................... 31.72 40.49 47.29 47.79 50.70 54.08 56.97 1.3%Imported crude oil (dollars per barrel). ....... 28.46 35.99 43.99 43.00 44.99 47.99 49.99 1.3%Domestic natural gas at wellhead(dollars per thousand cubic feet). ........... 5.08 5.49 5.03 4.52 4.90 5.43 5.92 0.3%Domestic coal at minemouth(dollars per short ton) .................... 18.40 20.07 22.23 20.39 20.20 20.63 21.73 0.3%Average electricity price(cents per kilowatthour). .................. 7.6 7.6 7.3 7.1 7.2 7.4 7.5 0.0%Economic indicatorsReal gross domestic product(billion 2000 dollars) ..................... 10,321 10,756 13,043 15,082 17,541 20,123 23,112 3.0%GDP chain-type price index(index, 2000=1.000) ..................... 1.063 1.091 1.235 1.398 1.597 1.818 2.048 2.5%Real disposable personal income(billion 2000 dollars) ..................... 7,742 8,004 9,622 11,058 13,057 15,182 17,562 3.1%Value of manufacturing shipments(billion 2000 dollars) ..................... 5,378 5,643 6,355 7,036 7,778 8,589 9,578 2.1%Energy intensity(thousand Btu per 2000 dollar of GDP). ..... 9.51 9.27 8.28 7.58 6.88 6.32 5.80 -1.8%Carbon dioxide emissions(million metric tons) ..................... 5,785 5,900 6,365 6,718 7,119 7,587 8,114 1.2%Notes: Quantities are derived from historical volumes and assumed thermal conversion factors. Other production includes liquidhydrogen, methanol, supplemental natural gas, and some inputs to refineries. Net imports of petroleum include crude oil, petroleumproducts, unfinished oils, alcohols, ethers, and blending components. Other net imports include coal coke and electricity. Some refineryinputs appear as petroleum product consumption. Other consumption includes net electricity imports, liquid hydrogen, and methanol.Source: AEO2006 National Energy Modeling System, run AEO2006.D111905A.Energy Information Administration / Annual Energy Outlook 2006 11

Legislation andRegulations

Legislation and RegulationsIntroductionBecause analyses by EIA are required to be policy-neutral,the projections in AEO2006 generally arebased on Federal and State laws and regulations ineffect on or before October 31, 2005. The potentialimpacts of pending or proposed legislation,regulations, and standards—or of sections oflegislation that have been enacted but thatrequire implementing regulations or appropriationof funds that are not provided or specifiedin the legislation itself—are not reflectedin the projections.Selected examples of Federal and State legislationincorporated in the projections include the following:• EPACT2005, which, among other actions, includesmandatory energy conservation standards;creates numerous tax credits for businesses andindividuals, covering energy-efficient appliances,hybrid vehicles, small biodiesel producers, andnew nuclear power capacity; creates a renewablefuels standard (RFS); eliminates the oxygen contentrequirement for Federal reformulated gasoline(RFG); extends royalty relief for offshore oiland natural gas producers; and extends and expandsthe production tax credit (PTC) for electricitygenerated from renewable fuels• The Military Construction Appropriations Actof 2005, which contains provisions to supportconstruction of the Alaska natural gas pipeline,including Federal loan guarantees during construction• The Working Families Tax Relief Act of 2004,which includes an extension of the 1.8-cent PTCfor electricity generated from wind and closedloopbiomass to December 31, 2005; tax deductionsfor qualified clean-fuel and electric vehicles;and changes in the rules governing oil and naturalgas well depletion• The American Jobs Creation Act of 2004, whichincludes incentives and tax credits for biodieselfuels, a modified depreciation schedule for theAlaska natural gas pipeline, and an expansion ofthe 1.8-cent renewable energy PTC to include geothermaland solar generation technologies• The Maritime Security Act of 2002, whichamended the Deepwater Port Act of 1974 to includeoffshore natural gas facilities• State renewable portfolio standard (RPS) programs,including the California RPS passed onSeptember 12, 2002• The State of Alaska’s Right-of-Way Leasing ActAmendments of 2001, which prohibit leasesacross State land for a “northern” or “over-thetop”natural gas pipeline route running east fromthe North Slope to Canada’s MacKenzie RiverValley• The Outer Continental Shelf Deep Water RoyaltyRelief Act of 1995 and subsequent provisions onroyalty relief for new leases issued after November2000 on a lease-by-lease basis• The Omnibus Budget Reconciliation Act of 1993,which added 4.3 cents per gallon to the Federaltax on highway fuels• The Energy Policy Act of 1992 (EPACT1992)• The Clean Air Act Amendments of 1990(CAAA90), which included new standards for motorgasoline and diesel fuel and for heavy-duty vehicleemissions• The National Appliance Energy Conservation Actof 1987• State programs for restructuring of the electricityindustry.AEO2006 assumes that State taxes on gasoline, diesel,jet fuel, and E85 (fuel containing a blend of 70 to85 percent ethanol and 30 to 15 percent gasoline byvolume) will increase with inflation, and that Federaltaxes on those fuels will continue at 2003 levels (thelast time the Federal taxes were changed) in nominalterms. AEO2006 also assumes that the ethanol taxcredit as modified under the American Jobs CreationAct of 2004 will be extended when it expires in 2010and will remain in force indefinitely. Although thesetax and tax incentive provisions include “sunset”clauses that limit their duration, they have beenextended historically, and AEO2006 assumes theircontinuation throughout the forecast. AEO2006 alsoincludes the biodiesel tax credits created underEPACT2005, but they are not assumed to beextended, because they have no history of legislativeextension.Selected examples of Federal and State regulationsincorporated in AEO2006 include the following:• CAIR and CAMR—promulgated by the EPA inMarch 2005 and published in the Federal Registeras final rules in May 2005—which will limit emissionsfrom power plants in the United States• New boiler limits established by the EPA on February26, 2004, which limit emissions of hazardousair pollutants from industrial, commercial,14 Energy Information Administration / Annual Energy Outlook 2006

Legislation and Regulationsand institutional boilers and process heaters byrequiring that they comply with a MaximumAchievable Control Technology (MACT) floor• Corporate average fuel economy (CAFE) standardsfor light trucks promulgated by the NationalHighway Traffic Safety Administration(NHTSA) in 2003 (but not the new proposed increasein fuel economy standards for light trucksbased on vehicle footprint in model years 2008through 2011, which have not been promulgated)• The December 2002 Hackberry Decision, whichterminated open access requirements for new onshorereceiving terminals for LNGAEO2006 includes the CAAA90 requirement of aphased-in reduction in vehicle emissions of regulatedpollutants. It also reflects “Tier 2” Motor VehicleEmissions Standards and Gasoline Sulfur ControlRequirements finalized by the EPA in February 2000under CAAA90. The Tier 2 standards for RFG wererequired by 2004, but because they included allowancesfor small refineries, they will not be fully realizedfor conventional gasoline until 2008. AEO2006also incorporates the ultra-low-sulfur diesel fuel(ULSD) regulation finalized by the EPA in December2000, which requires the production of at least 80 percentULSD (15 parts sulfur per million) highway dieselbetween June 2006 and June 2010 and 100percent ULSD thereafter. It also includes the rulesfor nonroad diesel issued by the EPA on May 11, 2004,regulating nonroad diesel engine emissions and sulfurcontent in fuel.The AEO2006 projections reflect legislation that bansor limits the use of the gasoline blending componentmethyl tertiary butyl ether (MTBE) in the next severalyears in 25 States. It is assumed that MTBE willbe phased out completely by the end of 2008 as aresult of EPACT2005, which repealed the oxygenaterequirement for RFG.More detailed information on recent and proposedlegislative and regulatory developments is providedbelow.EPACT2005 SummaryThe U.S. House of Representatives passed H.R. 6 EH,the Energy Policy Act of 2005, on April 21, 2005, andthe Senate passed H.R. 6 EAS on June 28, 2005. Aconference committee was convened to resolve differencesbetween the two bills, and a report wasapproved and issued on July 27, 2005. The Houseapproved the conference report on July 28, 2005, andthe Senate followed on July 29, 2005. EPACT2005was signed into law by President Bush on August 8,2005, and became Public Law 109-058 [1].Consistent with the general approach adopted in theAEO, provisions in EPACT2005 that require fundingappropriations to implement, whose impact is highlyuncertain, or that require further specification byFederal agencies or Congress are not included inAEO2006. For example, EIA does not try to anticipatepolicy responses to the many studies required byEPACT2005, nor to predict the impact of R&D fundingauthorizations included in the bill. Moreover,AEO2006 does not include any provision thataddresses a level of detail beyond that modeled inEIA’s National Energy Modeling System (NEMS),which was used to develop the AEO2006 projections.AEO2006 includes only about 30 sections ofEPACT2005, which establish specific tax credits,incentives, or standards in the following areas:• Mandatory energy conservation standards fortorchiere lamps, dehumidifiers, and ceiling fanlight kits in the residential sector and for lightingequipment, packaged air conditioning and heatingequipment, refrigerator and freezer equipment,automatic icemakers, pre-rinse sprayvalves, exit signs, distribution transformers, andtraffic signals in the commercial sector• Tax credits for businesses and builders investingin energy efficiency and renewable energy properties;for purchasers of energy-efficient equipment,including water heaters, air conditioners, heatpumps, furnaces, boilers, windows, and otherenergy-efficient building shell products; for producersof energy-efficient clothes washers, dishwashers,and refrigerators; for purchasers of solarwater heaters, solar photovoltaic (PV) equipment,and fuel cells; for businesses investing in fuel cellsand microturbines; and for businesses investingin solar energy properties• Tax credits for the purchase of vehicles with leanburn engines or with hybrid or fuel cell propulsionsystems• An RFS that requires the production and use ofdefined amounts of renewable fuel by specificdates• Elimination of the oxygen content requirementfor RFG• Extension of tax credits for biodiesel producersand small ethanol producers• A tax credit for small agri-biodiesel producersEnergy Information Administration / Annual Energy Outlook 2006 15

Legislation and Regulations• Royalty relief for oil and natural gas production inwater depths greater than 400 meters in the Gulfof Mexico• Restrictions on new oil and natural gas drilling inor under the Great Lakes• Reduction of the existing capital recovery periodfor new electric transmission and distribution assetsfrom 20 years to 15 years• Expansion of the amortization period for pollutioncontrol equipment on coal-fired power plantsfrom 5 years to 7 years• A PTC of 1.8 cents per kilowatthour for up to6,000 megawatts of new nuclear capacity broughtonline before 2021• An investment tax credit for the construction anddevelopment of new or repowered coal-fired generatingprojects• Extension, modification, and expansion of thePTC for renewable electricity generation.The following discussion provides a summary of theprovisions in EPACT2005 that are included inAEO2006 and some of the provisions that could beincluded if more complete information were availableabout their funding and implementation. This discussionis not a complete summary of all the sections ofEPACT2005. More extensive summaries are availablefrom other sources [2].End-Use DemandThis section summarizes the provisions of EPACT-2005 that affect the end-use demand sectors.BuildingsEPACT2005 includes provisions with the potential toaffect energy demand in the residential and commercialbuildings sector. Many are included in Title I,“Energy Efficiency.” Others can be found in therenewable energy, R&D, and tax titles.Sections 101 through 105 and Section 109 addressFederal energy use, allowing for energy conservationmeasures in congressional buildings (Section 101);updating Executive Order mandates regarding Federalpurchasing requirements and energy intensityreductions (Sections 102 through 104); extending theuse of Energy Savings Performance Contracts tofinance projects through 2016 (Section 105); andupdating performance standards for Federal buildings(Section 109). The Federal purchasing requirementsand performance standards are represented inNEMS as a result of earlier Executive Orders. Otheraspects of these provisions address a level of detailthat is not modeled in NEMS.Sections 135 and 136 establish or tighten mandatoryenergy conservation standards for a number of residentialproducts and appliances and commercialequipment, affecting projected residential and commercialenergy use. Standards for torchiere lamps areexplicitly modeled in NEMS, allowing for a directaccounting of energy savings from a maximum wattallowance. Savings resulting from standards for residentialdehumidifiers and ceiling fan light kits, basedon shipment estimates, are phased in over theAEO2006 forecast period to account for capital stockturnover. Standards for explicitly modeled commercialequipment, including lighting equipment, packagedair conditioning and heating equipment,refrigerator and freezer equipment, and automaticicemakers, are directly represented in the AEO2006projections. Savings resulting from standards for exitsigns, traffic signals, distribution transformers, andpre-rinse spray valves are estimated and phased inover the AEO2006 forecast period to account for capitalstock turnover.Provisions under Title XIII provide tax credits tobusinesses and individuals for investment in energyefficiency and renewable energy properties. Section1332 provides a tax credit of $1,000 or $2,000 to buildersof homes that are 30 or 50 percent more efficientthan current code in 2006 and 2007. Section 1333allows tax credits for purchasers of energy-efficientequipment, including water heaters, air conditioners,heat pumps, furnaces, boilers, windows, and otherenergy-efficient building shell products. The credit isavailable in 2006 and 2007, and the amount varieswith the technology purchased. Section 1334 providesa tax credit for producers of energy-efficient clotheswashers, dishwashers, and refrigerators. Section1335 provides tax credits for purchasers of solarwater heaters, solar PV equipment, and fuel cells forthe years 2006 and 2007. All these tax credits are representedin AEO2006. For modeling purposes, it isassumed that the credits will be passed on to consumersin the form of lower first costs for purchases of theproducts specified.Section 1336 provides a business investment taxcredit of 30 percent for fuel cells and 10 percent formicroturbines, and Section 1337 increases the businessinvestment tax credit for solar property from thecurrent level of 10 percent to 30 percent. These provisions,which apply to property installed in 2006 or2007, are included in AEO2006.16 Energy Information Administration / Annual Energy Outlook 2006

Legislation and RegulationsIndustrialEPACT2005 includes few provisions that specificallyaffect industrial sector energy demand. Provisions inthe R&D titles that may affect industrial energy consumptionover the long term are not included inAEO2006.Section 108 requires that federally funded projectsinvolving cement or concrete increase the amount ofrecovered mineral component (e.g., fly ash or blastfurnace slag) used in the cement. Such use of mineralcomponents is a standard industry practice, andincreasing the amount could reduce both the quantityof energy used for cement clinker production and thelevel of process-related CO 2 emissions. Because theproportion of mineral component is not specified inthe legislation, this provision is not included inAEO2006. When regulations are promulgated, theirestimated impact could be modeled in NEMS.Section 1321 extends the Section 29 PTC for nonconventionalfuel to facilities producing coke or cokegas. The credit is available for plants placed in servicebefore 1993 and between 1998 and 2010. Each plantcan claim the credit for 4 years; however, the totalcredit is limited to an annual average of 4,000 barrelsof oil equivalent (BOE) per day. The value of thecredit is currently $3.00 per BOE, and it will beadjusted for inflation in the future indexed to 2004.Previously, the $3.00 credit had been indexed to 1979,and its value in 2004 was estimated at $6.56 per BOE[3]. Because the bulk of the credits will go to plantsalready operating or under construction, there islikely to be little impact on coke plant capacity.TransportationEPACT2005 includes many provisions with potentialeffects on energy demand, alternative fuel use, andvehicle emissions in the transportation sector. Theseprovisions provide for research, development, anddemonstration (RD&D) of technologies and alternativefuels. These provisions are not reflected inAEO2006 because of the uncertainty associated withthe impacts of RD&D programs. The act also calls forpolicy studies and tax incentives to promote improvedenergy efficiency and increase alternative fuel use.Provisions specific to the supply of alternative transportationfuels are discussed below, in the sections onpetroleum and renewable energy.EPACT2005 provides a tax credit for the purchaseof vehicles that have lean burn engines or employhybrid or fuel cell propulsion systems. The amount ofthe credit is based the vehicle’s inertia weight,improvement in city-tested fuel economy relative toan equivalent 2002 base year value, emissions classification,and type of propulsion system. The tax creditis also sales-limited, by manufacturer, for vehicleswith lean burn engines or hybrid propulsion systems.A phaseout period begins with the first calendar quarterafter December 31, 2005, in which a manufacturer’ssales of lean burn or hybrid vehicles reach60,000 units. Reduction of the credits begins in thefollowing quarter. For that quarter and the next, theapplicable tax credit will be reduced by 50 percent.For the next two quarters, the tax credit will bereduced to 25 percent of the original value. These taxcredits are included in AEO2006.Petroleum, Ethanol, and Biofuel ProvisionsThis section summarizes the numerous provisions ofEPACT2005 affecting the supply, composition, andrefining of petroleum and related products that areincluded in AEO2006.Renewable Fuels StandardSection 1501 includes an RFS that requires the productionand use of 4.0 billion gallons of renewablefuels in 2006, increasing to 7.5 billion gallons in 2012.For calendar year 2013 and each year thereafter, theminimum required volume of renewable fuels wouldbe an amount equal to the percentage of total gasolinesold in the Nation in that year that was representedby 7.5 billion gallons in 2012. In addition, starting in2013, the required amount of renewable fuels mustinclude a minimum of 250 million gallons derivedfrom cellulosic biomass. Small refineries with a capacitynot exceeding 75,000 barrels per calendar day areexempted from the RFS until 2011. NoncontiguousStates or territories (Alaska, Hawaii, Puerto Rico,Guam, etc.) are not covered but could petition to jointhe renewable fuels program. Both ethanol andbiodiesel are considered to be renewable fuels, and a2.5-gallon credit toward the RFS is provided for everygallon of cellulosic biomass ethanol produced. A programof renewable fuels credits would allow refiners,blenders, and importers flexibility to comply with theRFS across geographical regions and over successiveyears.The RFS is modeled in AEO2006, both for the minimumrequired volumes and for ethanol derived fromcellulosic biomass. Actual renewable fuel suppliesmay or may not exceed those minimum requirements,depending on the relative costs of renewable fuels andcompeting petroleum products. In the AEO2006 referencecase, ethanol consumption is projected toexceed the RFS, because it is projected to be availableEnergy Information Administration / Annual Energy Outlook 2006 17

Legislation and Regulationsat relatively low cost. AEO2006 implicitly reflects theethanol production and consumption behavior thatresembles the effect of a national RFS credit tradingsystem, resulting in ethanol blending in gasoline thatvaries by region.Elimination of Oxygen Requirement forReformulated GasolineSection 1504 eliminates the oxygen content requirementfor RFG. This provision takes effect immediatelyin California and 270 days after enactment ofEPACT2005 in the rest of the RFG regions. Withoutthe oxygen content requirement, refiners are likely tophase out MTBE in gasoline as soon as practical tominimize exposure to environmental liabilities in thefuture. Several refiners have announced plans to stopmaking MTBE when the oxygen content requirementexpires. Also in Section 1504, volatile organic compound(VOC) Control Regions 1 (southern) and 2(northern) for RFG would be consolidated by eliminatingthe less stringent requirements applicable togasoline designated for VOC Control Region 2.Elimination of the oxygen requirement for RFG isincluded in AEO2006. MTBE is assumed to be phasedout in all regions by the end of 2008. Ethanol is likelyto be favored in RFG blending in most regions, basedon economics and its other attractive blending characteristics,such as high octane value.Biofuel Tax CreditsCurrently, gasoline and highway diesel fuel excisetaxes are 18.4 and 24.4 cents per gallon, respectively.For each gallon of highway fuel, 0.1 cent is depositedin the Leaking Underground Storage Tank TrustFund, which is extended through 2011 under Section1362 of EPACT2005. The volumetric excise tax creditprogram, established in the American Jobs CreationAct of 2004, covers both ethanol and biodiesel. Itallows producers to claim the tax credit directly onbiofuels: 51 cents per gallon of ethanol, $1 per gallonof biodiesel made from agricultural commodities suchas soybean oil, and 50 cents per gallon of biodieselmade from recycled oil such as yellow grease. Thebiodiesel tax credit is extended through 2008 underSection 1344 of EPACT2005, and the ethanol taxcredit was previously extended through 2010 underthe American Jobs Creation Act of 2004. Historically,the ethanol tax credit has been extended when itexpired; AEO2006 assumes that it will remain inforce indefinitely. The biodiesel tax credits areincluded in AEO2006, but it is not assumed that theywill be extended indefinitely, because they are relativelynew and have only a short history of legislativeextension.Section 1345 provides for an additional credit up to 10cents per gallon for small agri-biodiesel producerswith annual production of 15 million gallons or less.Small ethanol producers currently cannot have productioncapacity above 30 million gallons per year toqualify for the special credit. Section 1347 raises thecapacity limit to 60 million gallons per year. AEO2006includes both the credit for small agri-biodiesel producersand the change in the application of the creditfor small ethanol producers.Tax Incentives Related to Petroleum RefiningSection 1323 provides temporary expensing for refineryinvestments, which would allow taxpayers todepreciate immediately 50 percent of the cost of allinvestment that increases the capacity of an existingrefinery by at least 5 percent or increases thethroughput of qualified fuels by at least 25 percent.Qualified fuels include oil from shale and tar sands.As a condition of eligibility, refiners of liquid fuelsmust report the details of refinery operations to theInternal Revenue Service. Section 392 also authorizesthe EPA, in a cooperative agreement with a State, tostreamline the review of a refinery permit application.Because NEMS does not model individual refineryinvestment decisions, this provision is notincluded in AEO2006.Natural Gas ProvisionsEPACT2005 contains several provisions intended toencourage or facilitate the development of domesticoil and natural gas resources and the domestic infrastructurefor importing LNG. Most are in Title III,“Oil and Gas.” Others, covering R&D and tax measures,are included in Titles IX and XIII.Section 311 clarifies the role of the Federal EnergyRegulatory Commission (FERC) as the finaldecisionmaking body on the construction, expansion,or operation of any facility that exports, imports, orprocesses LNG. Although it grants final authority toFERC, it directs the commission to consult with theStates on safety issues. Section 317 requires the U.S.Department of Energy (DOE), in cooperation withthe U.S. Departments of Transportation and HomelandSecurity, to conduct at least three forums onLNG, which are to be held in areas where LNG terminalsare being considered for construction and to bedesigned to promote public education and encouragecooperation between State and Federal officials.Because the AEO2006 reference case alreadyassumes that siting issues for LNG terminals are notinsurmountable, no changes were made in NEMS toaddress the LNG-related provisions in EPACT2005.18 Energy Information Administration / Annual Energy Outlook 2006

Legislation and RegulationsIn addition, it is unclear to what degree this provisionwill affect the siting of regasification terminals.Under Section 312, FERC is given the authority topermit a natural gas company to provide facilities fornatural gas storage at market-based rates if itbelieves the company will not exert market power.NEMS already assumes some market impact as aresult of incentive-based rates.Sections 321, 322, and 323 clarify provisions of theOuter Continental Shelf Lands Act, the Safe DrinkingWater Act, and the Federal Water Pollution ControlAct. Sections 341 and 342 provide clarifications ofexisting programs. Sections 343 through 347 addressroyalty relief. Specifically, Sections 343 and 344address incentives for natural gas production frommarginal wells and from deep wells in the shallowwaters of the Gulf of Mexico; Section 346 suspendsroyalties on offshore production in Alaska; and Section347 provides royalty relief for production fromthe National Petroleum Reserve, at the discretion ofthe Secretary of Energy. Sections 353 and 354 dealwith royalty relief for natural gas extracted frommethane hydrates and for enhanced oil and naturalgas production through CO 2 injection. None of theseprovisions is modeled in NEMS, and they are notincluded in AEO2006.Section 345, which provides royalty relief for oil andnatural gas production in water depths greater than400 meters in the Gulf of Mexico from any oil or naturalgas lease sale occurring within 5 years after enactment,is modeled in NEMS. The minimum productionvolumes for which royalty payments would be suspendedare as follows:• 5,000,000 BOE for each lease in water depths of400 to 800 meters• 9,000,000 BOE for each lease in water depths of800 to 1,600 meters• 12,000,000 BOE for each lease in water depths of1,600 to 2,000 meters• 16,000,000 BOE for each lease in water depthsgreater than 2,000 meters.For AEO2006, the water depth categories specified inSection 345 were adjusted to be consistent with thedepth categories in the Offshore Oil and Gas SupplySubmodule of NEMS. The suspension volumes are5,000,000 BOE for leases in water depths 200 to 800meters; 9,000,000 BOE for leases in water depths of800 to 1,600 meters; 12,000,000 BOE for leases inwater depth of 1,600 to 2,400 meters; and 16,000,000BOE for leases in water depths greater than 2,400meters. Examination of the resources available at 200to 400 and 2,000 to 2,400 meters showed that the differencesbetween the depths used in the model andthose specified in the act would not materially affectthe model results.Section 386, which prohibits new oil and natural gasdrilling in or under the Great Lakes, is included inAEO2006. Specifically, it states that no Federal orState permit or lease shall be issued for new oil or naturalgas slant, directional, or offshore drilling in orunder one or more of the Great Lakes. To reflect thisprovision, oil and natural gas resources underlyingthe Great Lakes were removed from the resource baseof the Oil and Gas Supply Module in NEMS.In Title XIII, Sections 1325 through 1327 provide taxincentives for the oil and natural gas industries thatinclude treatment of natural gas distribution lines as15-year property, treatment of natural gas gatheringlines as 7-year property, and exclusion of prepaymentson natural gas supply contracts with governmentutilities from arbitrage rules. NEMS does notinclude sufficient detail for modeling theseprovisions.Electricity ProvisionsEPACT2005 includes provisions to improve the reliabilityand operation of the electricity transmissiongrid, reduce regulatory uncertainty, and increase consumerprotection. These electricity provisions areincluded under Title XII, “Electricity ModernizationAct of 2005.” Most of them cannot be addressed at thelevel of detail included in NEMS or can be includedonly with additional specification not provided inEPACT2005. Title XIII, “Energy Tax Incentive Act of2005,” also includes tax incentives targeted towardelectricity generation or transmission properties.Section 1211 calls for the creation of mandatory reliabilitystandards for the electricity grid to replace thevoluntary standards in place today. The new standardswould be administered by “electric reliabilityorganizations” (EROs), which would be certified byFERC and would be responsible for developing andenforcing reliability standards for their regions. It isimplicitly assumed in AEO2006 that electricity will beprovided reliably.Several sections under Title XIII would affect theelectric power industry. Section 1308 shortens theexisting capital recovery period for new transmissionand distribution assets from 20 years to 15 years. Theproperty must have been placed in use after April 11,Energy Information Administration / Annual Energy Outlook 2006 19

Legislation and Regulations2005, to qualify for the new recovery period. Section1309 expands amortization of pollution control equipmenton coal-fired plants from 5 years to 7 years. Onlyplants that came online after January 1, 1976, wouldqualify for the new amortization period. These taxchanges are represented in AEO2006. Tax credits fornuclear and renewable energy production and for coalproduction and investment are discussed below.Nuclear Energy ProvisionsTitle VI of EPACT2005 includes several provisionsdesigned to ensure that nuclear energy will remain amajor component of the Nation’s energy supply. Sections601 through 610 update the Price-Anderson ActAmendment to the Atomic Energy Act of 1954, whichensures that adequate funds are available to the publicto satisfy liability claims in the event of a nuclearaccident, while limiting the liability of any individualreactor owner. EPACT2005 extends the coverage toall nuclear units brought on line through 2025,adjusts the maximum assessment and liability limit,and addresses incidents that might occur outside theUnited States. Section 608 allows small, modularreactors to be combined and treated as a single unitfor liability purposes. These provisions are notexplicitly modeled in NEMS, but AEO2006 implicitlyassumes that Price-Anderson coverage will beextended to any new nuclear units built in the UnitedStates.Under Title XIII, Section 1306 provides a PTC fornew nuclear reactors brought online through 2020.The PTC is worth 1.8 cents per kilowatthour for thefirst 8 years of operation, subject to an annual limit of$125 million per gigawatt of capacity. It is restrictedto a total of 6 gigawatts of new nuclear capacity. Thisprovision is included in AEO2006. Section 1310 modifiesthe rules for qualified decommissioning fundsand requires that a new ruling on the amounts fundedbe made whenever a plant receives a license renewal.Coal ProvisionsEPACT2005 includes numerous provisions thatauthorize funding for coal-related activities. Becausethey depend on future appropriations, they are notincluded in AEO2006.Sections 431 through 438, referred to as the CoalLeasing Act, ease or remove certain requirements forcoal leases on Federal lands. These provisions are notincluded in AEO2006, because specific lease requirementscannot be modeled directly in NEMS.Title XIII includes several provisions that alter thetax treatment of certain coal-related activities. Forexample, Section 1301 sets qualifications for receiptof a PTC of $1.50 per ton between 2006 and 2009 and$2.00 per ton through 2013 for coal produced onIndian lands. This provision is not included inAEO2006, because only limited data are available oncoal resources and production on Indian lands. (In2000, coal was mined from Indian lands in Arizona,New Mexico, and Montana.) One possible outcome ofthis provision would be to accelerate production ofcoal from Indian lands while the credit is available;however, given the relatively short time horizon ofthe provision (qualifying mines must be in servicebefore 2009) and the small share of total coal productionmade up by coal from Indian lands (3.6 percent in2004), the impact on national average minemouthprices for coal is likely to be small.Section 1307, Subsection 48A, establishes a $1.3 billioninvestment tax credit for the construction of newor repowered coal-fired generation projects, including$800 million for coal gasification projects and $500million for other projects that achieve certain targets,such as 99 percent SO 2 removal and 90 percent mercuryremoval from plant emissions. For integratedgasification combined-cycle (IGCC) technologies a20-percent investment tax credit may be applied toqualifying investments, and for other qualifyingadvanced technologies a 15-percent investment taxcredit is applicable. Repowering projects mustimprove the thermal design efficiency of coal-firedplants by 4 to 7 percent. This provision is modeled inNEMS by allowing up to 3 gigawatts of IGCC andanother 3 gigawatts of advanced coal-fired capacity totake advantage of the tax credit.Renewable Energy ProvisionsEPACT2005 contains several provisions intended toencourage or facilitate the use of renewable energyresources for electricity production. Most areincluded in Title II, “Renewable Energy.” Others arein the R&D, electricity, and tax titles. In addition, theact contains provisions to encourage the use of renewableenergy for transportation and in end-use applications,as described above.Section 203 requires the Federal Government, to theextent that it is “economically feasible and technicallypractical,” to purchase a minimum amount of electricitygenerated from renewable resources. The Federalpurchase requirement starts at 3 percent of thetotal amount of electricity consumed by the FederalGovernment in 2007 and increases stepwise to 7.5percent of the total in 2013 and thereafter. Renewableenergy used at a Federal facility that is producedon-site at the facility, on Federal lands, or on Indian20 Energy Information Administration / Annual Energy Outlook 2006

Legislation and Regulationsland will count double toward the requirement.Although the Federal Government is a major purchaserof electricity, the required purchases are notexpected to affect the projections of renewable electricitydemand in AEO2006.Several sections specifically address the production ofhydroelectricity at proposed or existing facilities. Section241 revises the appeals process for the licensingof hydropower projects by FERC. Appeals on licenseconditions and fishway rulings will now be heard in atrial-type hearing. Applicants may also propose alternativesto the conditions specified by FERC to achievethe purposes of the original license conditions. Theimpacts of these provisions on the cost of developingor relicensing hydroelectric projects are not clear, andthey are not included in AEO2006.Under Title XII, a number of electricity market provisionsdirectly address the use of renewable resourceswithin the Nation’s electricity grid. Section 1251requires utilities to offer net metering upon customerrequest. Net metering means that eligible customer-sitedgeneration resources may be used to offsetgross customer electricity purchases during thebilling cycle; that is, customer generation in excess ofinstantaneous demand will be fed back to the utilitydistribution system, causing the customer’s meter toeffectively “run backward.” Eligible resources andapplicable billing cycles are not defined in the provision,but credit will be given to States that haveadopted or voted on comparable standards. CurrentState net metering standards typically allow renewablegeneration (solar and sometimes wind or otherrenewable fuels) and sometimes allow other favoredtechnologies, such as fuel cells, to qualify. Generationis typically netted on a monthly basis, but nettingmay be allowed over longer periods. AEO2006assumes that excess generation from customer-sitedsources will offset purchased electricity at retail rates.Section 1253 eliminates the requirement for eligibleutilities to purchase electricity from qualified facilitiesunder the Public Utility Regulatory Policies Act(PURPA), which previously required utilities to purchasegeneration from small cogenerators and renewableplants at a rate equal to their avoided cost ofgeneration. Eligible utilities must have open electricitymarkets, including nondiscriminatory access towholesale generation markets and to transmissionand interconnection services. AEO2006 assumes thatall generation resources will compete in a nondiscriminatorymarket for generation, capacity, and transmissionservices.Several changes to the tax code, all involving the PTCfor renewable generation, are expected to have significantimpacts on the growth of renewable electricitymarkets. Section 1301 extends the eligibility date fornew renewable generation facilities to qualify for theinflation-adjusted tax credit for the first 10 years ofplant operation. Eligibility was set to expire afterDecember 31, 2005, but will now expire after December31, 2007. Although some eligible resources willcontinue to get the full, inflation-adjusted credit of1.5 cents per kilowatthour and others one-half of thatamount, all new eligible facilities—including efficiencyimprovements or additions of capacity at existingfacilities—will receive the full credit for the first10 years of their operation. AEO2006 specificallyaccounts for the extension of the eligibility period forrenewable resources and the expansion of the creditto hydroelectric facilities.In addition to the PTC modifications discussed above,Section 1302 will allow agricultural cooperatives toallocate renewable energy production tax credits totheir members, based on the “amount of business”done by each member with the cooperative. Eligiblecooperatives include those that are more than 50 percentowned by agricultural producers or entitiesowned by agricultural producers, thus allowing otherwisetax-exempt electricity cooperatives to takeadvantage of the PTC by transferring the benefitdirectly to their membership. Although this provisionis not specifically modeled, AEO2006 assumes that alleligible renewable capacity is built by tax-paying entitiesand thus is entitled to take the PTC.Incentives for Innovative TechnologiesEPACT2005 Title XVII, “Incentives for InnovativeTechnologies,” authorizes the Secretary of Energy,after consultation with the Secretary of the Treasuryand subject to budget appropriations, to provide Federalloan guarantees for a wide variety of projectsrelated to energy consumption and production technologies.Although EPACT2005 includes severalother technology incentives, the Title XVII programhas particular potential to influence the developmentof future energy technologies. The guarantees cancover up to 80 percent of the cost of a project over aperiod of up to 30 years (or 90 percent of a project’suseful life, whichever is less). To be eligible, projectsmust avoid, reduce, or sequester air pollutants oranthropogenic greenhouse gas (GHG) emissions andmust employ new or significantly improved technologies,as compared with those that are commerciallyavailable when the guarantee is issued. The eligibleproject categories include:Energy Information Administration / Annual Energy Outlook 2006 21

Legislation and Regulations• Renewable energy systems• Advanced fossil energy technologies, includingcoal gasification meeting certain requirements• Hydrogen fuel cell technologies for residential, industrial,or transportation applications• Advanced nuclear energy facilities• Carbon capture and sequestration practices andtechnologies• Technologies for efficient generation, transmission,and distribution of electric power• Efficient end-use energy technologies• Production facilities for fuel-efficient vehicles, includinghybrids and advanced diesel vehicles• Pollution control equipment• Refineries.Loan guarantees will also be available for gasificationfacilities that meet certain criteria. Eligible gasificationprojects include IGCC plants where 65 percent ofthe fuel used is a combination of coal, biomass, andpetroleum coke and 65 percent of the energy output isused to produce electricity. IGCC plants with a capacityof at least 100 megawatts using western coal arealso eligible. To receive loan guarantees, the IGCCprojects must emit no more than 0.05 pound of SO 2 ,0.08 pound of nitrogen oxides (NO x ), and 0.01 poundof particulates per million Btu of fuel input and mustremove 90 percent of the mercury in any coal that isused.Because funding levels and specific rules for this programare not yet known, its potential impacts are notrepresented in AEO2006. The program could providea flexible tool for stimulating investment in a widearray of promising technologies [4]. The leverageachieved by the program will depend on the risksassociated with the projects supported and theexpected loss that would occur if a loan defaultoccurred. For loans of the same size, riskier projectsrequire more Federal funding.California Greenhouse Gas EmissionsStandards for Light-Duty VehiclesThe State of California was given authority underCAAA90 to set emissions standards for light-dutyvehicles that exceed Federal standards. In addition,other States that do not comply with the NationalAmbient Air Quality Standards (NAAQS) set by theEPA under CAAA90 were given the option to adoptCalifornia’s light-duty vehicle emissions standards inorder to achieve air quality compliance. CAAA90 specificallyidentifies hydrocarbon, carbon monoxide,and NO x as vehicle-related air pollutants that can beregulated. California has led the Nation in developingstricter vehicle emissions standards, and other Stateshave adopted the California standards [5].California Assembly Bill 1493 (A.B. 1493), signed intolaw in July 2002, required the California AirResources Board (CARB) to develop and adopt GHGemissions standards for light-duty vehicles thatwould provide the maximum feasible reduction inemissions. In determining the maximum feasiblestandard, CARB was required to consider costeffectiveness,technological capability, economicimpacts, and flexibility for manufacturers in meetingthe standard. CARB was not allowed to consider thefollowing compliance options: mandatory trip reductions;land use restrictions; additional fees and/ortaxes on any motor vehicle, fuel, or vehicle miles traveled;a ban on any vehicle category; reduction in vehicleweight; or a limitation or reduction of speed limitson any street or highway in the State. Tailpipe emissionsof CO 2 , which are directly proportional to vehiclefuel consumption, account for the vast majority oftotal GHG emissions from vehicles. In August 2004,CARB released a report detailing its proposedGHG emissions standards for light-duty vehicles,which were approved by California’s Office of AdministrativeLaw on September 15, 2005.The standards approved in September 2005 coverGHG emissions associated with vehicle operation, airconditioning operation and maintenance, and productionof vehicle fuel. The standards apply to noncommerciallight-duty passenger vehicles manufacturedfor model years 2009 and beyond. The standards,specified in terms of CO 2 equivalent emissions, applyto vehicles in two size classes: passenger cars andsmall light-duty trucks with a loaded vehicle weightrating of 3,750 pounds or less; and heavy light-dutytrucks with a loaded vehicle weight rating greaterthan 3,750 pounds and a gross vehicle weight ratingless than 8,500 pounds. The CO 2 equivalent emissionstandard for heavy light trucks includes noncommercialpassenger trucks between 8,500 pounds and10,000 pounds. The regulations approved in September2005 set near-term standards, to be phased inbetween 2009 and 2012, and mid-term standards, tobe phased in between 2013 and 2016. After 2016, theemissions standards are assumed to remain constant.Table 2 summarizes the CO 2 equivalent standards.In October 2003, California, 11 other States, 3 cities,and several environmental groups filed a petition in22 Energy Information Administration / Annual Energy Outlook 2006

Legislation and RegulationsTable 2. CARB emissions standards for light-duty vehicles, model years 2009-2016CO 2 equivalent emissions standard (grams per mile)Tier Model YearPassenger cars and small light trucks(under 3,751 pounds)Heavy light trucks(3,751 to 8,500 pounds)Near term 2009 323 4392010 301 4202011 267 3902012 233 361Mid-term 2013 227 3552014 222 3502015 213 3412016 205 332the U.S. Court of Appeals, arguing that the EPAshould regulate GHG emissions from vehicles. In July2005, the court ruled that the EPA was not requiredto regulate GHG emissions under the Clean Air Act.Given the constraints on using other measures,improvements in fuel economy are the only practicalway to meet the standards. The automotive industry,which opposes A.B. 1493, has filed suit against CARB,arguing that California GHG emissions standards arein essence fuel economy standards and therefore arepreempted by a Federal statute that gives the U.S.Department of Transportation the only authority toregulate fuel economy [6]. CARB has not yet obtaineda Clean Air Act waiver from the EPA, which would berequired before it can implement its GHG emissionsstandards. For this reason and due to the uncertaintysurrounding the pending lawsuit, A.B. 1493 is notrepresented in the AEO2006 reference case. Potentialimpacts of the regulations were examined, however,in AEO2005, using the AEO2005 reference case as astarting point to estimate their likely effects on vehicleprices, GHG emissions, regional energy demand,and regional fuel prices [7].Proposed Revisions to Light TruckFuel Economy StandardsIn August 2005, NHTSA published proposed reformsto the structure of CAFE standards for light trucksand increases in light truck CAFE standards formodel years 2008 through 2011 [8]. Under the proposednew structure, NHTSA would establish minimumfuel economy levels for six size categoriesdefined by the vehicle footprint (wheelbase multipliedby track width), as summarized in Table 3. For modelyears 2008 through 2010, the new CAFE standardswould provide manufacturers the option of complyingwith either the standards defined for each individualfootprint category or a proposed average light truckTable 3. Proposed light truck CAFE standardsby model year and footprint category(miles per gallon)Modelyear1(43.0)Vehicle category andfootprint range (square feet)2(>43.0-47.0)3(>47.0-52.0)4(>52.0-56.5)5(>56.5-65.0)6(>65.0)2008 26.8 25.6 22.3 22.2 20.7 20.42009 27.4 26.4 23.5 22.7 21.0 21.02010 27.8 26.4 24.0 22.9 21.6 20.8*2011 28.4 27.1 24.5 23.3 21.9 21.3*Decrease due to changes in production plans provided to NHTSA and usedto establish an average that increases over time.fleet standard of 22.5 miles per gallon in 2008, 23.1miles per gallon in 2009, and 23.5 miles per gallon in2010. All light truck manufacturers would berequired to meet an overall standard based on saleswithin each individual footprint category after modelyear 2010.In determining the proposed light truck fuel economystandards, NHTSA addressed concerns related toenergy conservation, technology feasibility and economicpracticability, other regulations on fuel economy,and safety. In the evaluation of technology andeconomic practicability, NHTSA used gasoline priceprojections from the AEO2005 reference case, whichprojected that gasoline prices would range from $1.54to $1.61 per gallon (2004 dollars) over the 2004-2025forecast period. For the same period, the AEO2006reference case projects a range of $1.95 to $2.26 pergallon (2004 dollars). NHTSA, which will likelyreceive and address comments related to many issues,specifically asked for comments on the appropriategasoline price projection to use in defining the finalrule. Use of the AEO2006 reference case gasolineprices in the final rule could impact the final CAFEstandards. For example, using higher gasoline pricesin technology evaluations could lead to a finding thatEnergy Information Administration / Annual Energy Outlook 2006 23

Legislation and Regulationsadditional technologies are economically practical,with corresponding changes in fuel economy standardsfor some footprint categories.Because the new light truck fuel economy standardshave not been finalized, they are not included in theAEO2006 reference case. An alternative case wasdeveloped to examine the potential energy impacts ofthe proposed standards. Because NEMS does not currentlyrepresent the footprint-based standardsincluded in NHTSA’s recent proposal, the alternativecase assumes that manufacturers will adhere to theproposed increases in light truck fleet standards. Formodel year 2011, the alternative case applies afleet-wide standard of 24 miles per gallon, basedloosely on the change between 2010 and 2011 in theproposed footprint-based standards. Because no furtherchanges in fuel economy standards beyond 2011are assumed, the projected trends in light truck fueleconomy after 2011 reflect projected technologyadoption and market forces.New light truck fuel economy in the alternative case(Table 4) is projected to be 6 percent higher than thereference case projection in 2011 (24.9 miles pergallon, compared with 23.4 miles per gallon in thereference case). Consistent with the reference caseprojections, light truck fuel economy continues toimprove after 2011 in the alternative case, to 27.4miles per gallon in 2030, 4 percent higher than thereference case projection of 26.4 miles per gallon. Thehigher CAFE standards lead to higher prices for lighttrucks, resulting from increased use of lightweightmaterials, more complex valve trains, and advancedtransmissions. In the alternative case, the averageprice of a new light truck is projected to be 1.2 percent($350) higher than in the reference case in 2011 and0.5 percent ($170) higher in 2030. That increase is atleast partially offset, however, by the expectedTable 4. Key projections for light truck fuel economyin the alternative CAFE standards case, 2011-2030Projection 2011 2015 2020 2030Fuel economy of new light trucks(miles per gallon) 24.9 25.2 26.0 27.4Increase from reference caseprojection for purchase price of newlight trucks (2004 dollars) 350 250 210 170Annual reduction from referencecase projection for energy use by alllight-duty vehicles (quadrillion Btu) 0.13 0.26 0.35 0.44Cumulative reduction from referencecase projection for energy use by alllight-duty vehicles, 2004-2030(quadrillion Btu) 0.31 1.19 2.76 6.85reduction in fuel costs that would result from theincrease in average fuel efficiency.Total projected energy use by light-duty vehicles,including both cars and light trucks, in the alternativecase is projected to be 0.7 percent (0.13 quadrillionBtu) lower than the reference case projection in2011 and 1.8 percent (0.44 quadrillion Btu) lower in2030. Cumulative energy use by light-duty vehiclesfrom 2004 to 2030 is almost 7 quadrillion Btu lower inthe alternative case than projected in the referencecase.State Renewable Energy Requirementsand Goals: Update Through 2005AEO2005 provided a summary of 17 State renewableenergy programs in existence as of December 31,2003, in 15 States [9]. They included RPS programs in9 States, renewable energy mandates in 4 States, andrenewable energy goals in 4 States. Since 2003, 7more States and the District of Columbia have establishedrenewable energy programs (Table 5), including6 RPS programs and two renewable energy goals.No new mandates have been enacted since 2003,although a renewable goal instituted in Vermont willbecome mandatory if it has not been met by 2012. Inaddition, major changes and refinements have beenmade in a number of the State programs that were inexistence before 2004 (Table 6). No Federalrenewables requirement currently exists, although anationwide RPS was again considered in 2005.Although generally resembling earlier versions, someof the new programs and changes to existing programsinclude unique or unusual features:• Colorado’s new RPS is the first enacted through avoter initiative. The new RPS allows a coveredColorado utility (40,000 or more customers) to optout of the RPS, or an exempt utility to opt in, witha majority vote involving a minimum of 25 percentof the utility’s customers.• Connecticut’s RPS now includes energy conservation.• Delaware’s RPS includes municipal utilities andsome rural electric cooperatives, although theymay opt out.• Qualifying renewables under Hawaii’s RPS nowinclude electricity conservation measures, such asdistrict cooling systems using seawater air conditioning,solar and heat pump water heating,and ice storage, as well as reject heat in someinstances.24 Energy Information Administration / Annual Energy Outlook 2006

Legislation and Regulations• Under 2005 legislation, compliance with Minnesota’sobjective (goal) now becomes linked to theapplication for a certificate of need for new transmissionor generation facilities.• New York’s goals set in 2004 and the 2005changes in the Illinois program both resulted frompublic utility commission orders rather than fromlegislation.• In New York, the development of new generatingcapacity using renewable fuels is supportedthrough centralized procurement by the NewYork State Energy Research and DevelopmentAuthority, with funds collected through a chargeon investor-owned utilities.• The Illinois program allows imports of electricityonly from directly adjacent jurisdictions designatedas serious or severe NAAQS nonattainmentareas.• Vermont’s goal is to meet 100 percent of additionalelectricity demand through 2012 with allowedrenewable resources (up to 10 percent oftotal demand), and it broadly defines renewableTable 5. Basic features of State renewable energy requirements and goals enacted since 2003StateYearenactedRequirementsAcceptsexistingcapacityOut-of-StatesupplyCredittradingRenewable Portfolio StandardsColorado 2004 3-10% of generation, 2007-2015; 4% of requirement must be solar Yes Yes YesDelaware 2005 1-10% of retail sales, 2007-2019 Yes Yes YesDistrict of Columbia 2005 11% of sales by 2022; 3.5% of requirement must be solar Yes Yes YesMaryland 2004 3.5-7.5% of sales, 2006-2019 Yes Yes YesMontana 2005 5-15% of sales, 2008-2015 Yes Yes YesRhode Island 2004 3-16% of sales, 2007-2019 Yes Yes YesGoalsNew York 2004 25% of generation by 2013 Yes Yes NoVermont 2005 All growth, up to 10% of total sales, 2005-2012;goal becomes mandatory if not met by 2012Yes — —Table 6. Major changes in existing State renewable energy requirements and goals since 2003State Date of change New requirementsConnecticut July 2005 Effective January 1, 2006, Public Law 05-01 adds Class III renewables to the State RPS, to includenew customer-side combined heat and power systems and electricity savings from energy conservationand load management at commercial and industrial facilities, equal to 1% of generation in 2007,2% in 2008, 3% in 2009, and 4% in 2010.Hawaii June 2004 Senate Bill 2474 changes the goal of the State RPS, from 9% of sales by 2010 to 20% of sales by 2020,and includes ocean technologies, electricity conservation, and some cogeneration.Illinois July 2005 An Illinois Commerce Commission resolution adopts a sustainable energy plan that replaces the Staterenewable energy goal of 15% of sales by 2020 with an RPS requiring the State's largest electricutilities to begin supplying 2% renewable energy to Illinois customers by January 1, 2007, increasingby 1% annually to 8% by 2013; at least 75% of the requirement must be from wind power.Minnesota May 2005 Statute 216B.243 links compliance with the State’s renewable energy goal of 10.0% of electricity sales(by power producers other than Xcel Energy, see Statute 216B.1691) to obtaining a certificate of needfor new transmission or generation capacity.Nevada June 2005 Assembly Bill 03 increases overall renewables requirement from 5-15% of sales 2003-2013, to 6-20%,but (a) delays compliance by 2 years to 2005-2015, and (b) permits up to one-quarter of the requirementto be met by efficiency measures reducing electricity use.Pennsylvania November 2004 Senate Bill 1030 changes individual utility goals to RPS requiring 5.7% of sales in 2007, increasingto 18% in 2020 (with solar increasing to at least 0.5% of sales); RPS includes waste coal, coalgasification, and demand-side management and includes both credit trading and some capacity fromout-of-State suppliers in interconnected areas.Texas August 2005 Senate Bill 20 increases overall renewable energy requirement from 2,000 megawatts of new renewablecapacity by 2009 to 5,880 megawatts by 2015, including a non-mandatory target of at least 500megawatts from sources other than wind.Energy Information Administration / Annual Energy Outlook 2006 25

Legislation and Regulationsenergy as that which “relies on a resource that isbeing consumed at a harvest rate at or belowits natural regeneration rate” [10]. The Vermontlegislation is designed to encourage contracts fornew renewables capacity by allowing the new capacityto meet multiple States’ RPS requirements.New renewable capacity in Vermont canbe counted toward Vermont’s program, while itsrenewable energy credits may be marketed separatelyto renewables credit markets in neighboringStates.• Pennsylvania’s new Alternative Energy PortfolioStandard includes waste coal and coal gasification,which can contribute as much as 10 percentof the renewable generation requirement (set at18 percent of total generation in 2020).The 23 State renewable energy programs in effect in2005 generally are concentrated in three broad geographicareas, with 11 jurisdictions along the Northeasternand Mid-Atlantic seaboard (Connecticut,Delaware, the District of Columbia, Maine, Maryland,Massachusetts, New Jersey, New York, Pennsylvania,Rhode Island, and Vermont), 6 in the Southwest(Arizona, California, Colorado, Nevada, New Mexico,and Texas), 4 in the upper Midwest (Illinois, Iowa,Minnesota, and Wisconsin), and Hawaii and Montanaeach standing alone. No Southern, Southeastern, orNorthwestern State (except Montana) currently has arenewable energy program.Although efforts to coordinate renewable energy programsamong adjacent States have begun, no formalor informal coordination systems have been finalized.An example of efforts to establish such a systeminclude the newly formed Mid-Atlantic Organizationof PJM States, Inc. In order to prevent double counting,however, States in most interconnected regionsnow coordinate identification and tracking of the originsand contracted destinations of renewable energytransactions via power pools or other organizations.The New England States use the Generation InformationSystem of the New England Power Pool(NEPOOL), and the Mid-Atlantic States employPJM’s Generation Attribute Tracking System(GATS). Although the Midwestern Power Pool doesnot currently track the region’s renewable generation,a multi-State Midwestern effort is underway toestablish the Midwest Renewable Energy TrackingSystem (MRETS). Similarly, the California EnergyCommission and the Western Governors’ Associationare collaborating to establish a Western RenewableEnergy Generation Information Tracking System(WREGIS).New Renewable Energy Capacity, 2004-2005Table 7 summarizes EIA’s understanding of newrenewable energy capacity entering service in 2004and 2005. However, it is difficult to quantify the specificimpacts of State renewable programs. First, neitherthe individual States nor other sources identifyall the new renewable energy capacity that is built,and some new capacity may not be reported.Although large wind projects typically are recognized,smaller projects, such as landfill gas (LFG) or endusersited PV installations, may go unreported. Further,new capacity is not necessarily added inresponse to State renewable energy programs. Projectsmay be constructed for other reasons, and theymay or may not qualify for the State programs. Projectslocated in one State may serve the requirementsof another State or different States over time.Table 7. New U.S. renewable energy capacity, 2004-2005 (installed megawatts, nameplate capacity)Year Biomass GeothermalConventionalhydroelectricLandfillgasSolarphotovoltaics Wind Total2004Without standards 0.0 0.0 65.8 32.5 0.0 199.8 298.1With standards 19.9 0.0 4.5 30.0 3.0 281.6 339.12005Without standards 0.0 0.0 133.2 14.7 0.0 1,077.1 1,225.0With standards 34.1 37.0 26.1 24.6 3.6 1,716.7 1,842.12004 and 2005Without standards 0.0 0.0 199.0 47.2 0.0 1,276.9 1,523.1With standards 54.0 37.0 30.6 54.8 6.6 1,998.2 2,181.2Total 54.0 37.0 229.6 102.0 6.6 3,275.1 3,704.3PercentagesWithout standards 0.0 0.0 86.6 46.3 0.0 39.0 41.1With standards 100.0 100.0 13.3 53.7 100.0 61.0 58.926 Energy Information Administration / Annual Energy Outlook 2006

Legislation and RegulationsProjects located in States without renewable programsmay be explicitly or implicitly targeted to serveprograms in other States and, therefore, may be atleast partially “caused” by another State’s renewableprogram despite not being enumerated as such.New renewable energy capacity built today thatappears unsupported by a State renewable programmay result from an earlier favorable experience witha State program. For example, Table 7 does notinclude 362 megawatts of wind capacity from newprojects in Iowa in the “With Standards” category,because Iowa’s mandate was fully met by 2000; nordoes it include 62 megawatts of new wind capacitybuilt in North Dakota, which has no requirement,although the new capacity serves Minnesota’s RPS.Nevertheless, Table 7 provides some indication of theextent to which renewable programs are resulting inthe construction of new renewable energy capacityand also suggests the extent to which other key factors(for example, the Federal PTC) may promotegrowth in renewable capacity.Differences among renewable energy capacity additionsin different States can result from a range offactors separate from State renewable programs,including differences in natural endowments, electricityconsumption levels and rates of demandgrowth, the availability of alternatives, the presenceor absence of renewable energy proponents andchampions, and variations in consumer preferences.On the other hand, while States with renewableenergy requirements accounted for only 45 percent oftotal U.S. electricity supply, they accounted foralmost 60 percent of all new renewable energy capacityadded in 2004 and 2005.EIA’s analysis indicates that State-level requirementsprobably have led to somewhat more biomass,geothermal, LFG, and solar capacity than would otherwisehave been built, although the additionalamounts are small. Hydroelectric capacity does notappear to have been advanced by State-level renewablesrequirements. Expansion of wind power capacityappears to be strongly affected by the combinationof State requirements and the Federal PTC, asevidenced by the substantial construction of newwind capacity in 2005, particularly in States with RPSprograms.Among States with requirements and goals, theamount of renewable capacity added in 2004 and 2005varies significantly. Of the 23 States with renewablerequirements in 2004 and 2005, 4 have reported nonew renewable energy capacity (although requirementsin Delaware, the District of Columbia, andMaryland are new, and Connecticut is estimated tohave met its program requirements already). Inanother 7 States (Arizona, Hawaii, Massachusetts,New Jersey, Rhode Island, Vermont, and Wisconsin)15 megawatts or less has been added over the 2-yearperiod. In 3 States (Maine, Nevada, and Pennsylvania),between 25 and 35 megawatts has been added; in2 (Colorado and Illinois) between 65 and 75 megawattshas been added; and in 4 (Minnesota, Montana,New Mexico, and New York) between 100 and 200megawatts has been added in each State over the past2 years. California, with nearly 500 megawatts, andTexas, with more than 700 megawatts, togetheraccount for 55 percent of all new U.S. renewablecapacity attributed to State-level renewable energyrequirements and goals in 2004 and 2005.In contrast, Oklahoma and Washington, which haveno renewable energy requirements, each installedbetween 250 and 300 megawatts of new renewablecapacity in 2004 and 2005, and other States withoutprograms added smaller amounts. Most of the newcapacity in those States is wind power, suggestingthat good resources and the Federal PTC may be theprimary factors leading to new wind power installationsthere.Despite the expansion of State renewable energyprograms, new renewables capacity accounted for afairly small fraction of new U.S. electricity supplyadded in 2004 and 2005. Including conventionalhydroelectricity, all renewables currently accountfor 9.3 percent of total U.S. electricity generation,with nonhydroelectric renewables accounting for2.2 percent. The 3,700 megawatts of new renewablescapacity added during 2004 and 2005 accountedfor 12 percent of the 32,000 megawatts of newgenerating capacity that entered service during theperiod.State Air Emission RegulationsThat Affect Electric Power ProducersSeveral States have recently enacted air emission regulationsthat will affect the electricity generation sector.The regulations govern emissions of NO x ,SO 2 ,CO 2 , and mercury from power plants. Where firmcompliance plans have been announced, State regulationsare represented in AEO2006. For example,installations of SO 2 scrubbers and selective catalyticreduction (SCR) and selective noncatalytic reduction(SNCR) NO x removal technologies associated withthe largest State program, North Carolina’s CleanSmokestacks Initiative, are included. Figure 9 showshistorical trends in SO 2 emissions for selected States.Energy Information Administration / Annual Energy Outlook 2006 27

Legislation and RegulationsFederal Air Emissions RegulationsIn 2005, the EPA finalized two regulations, CAIR andCAMR, that would reduce emissions from coal-firedpower plants in the United States. Both CAIR andCAMR are included in the AEO2006 reference case.The EPA has received 11 petitions for reconsiderationof CAIR and has provided an opportunity forpublic comment on reconsidering certain aspects ofCAIR. Public comments were accepted until January13, 2006. The EPA has also received 14 petitions forreconsideration of CAMR and is willing to reconsidercertain aspects of the rule. Public comments wereaccepted for 45 days after publication of the reconsiderationnotice in the Federal Register. Several Statesand organizations have filed lawsuits against CAMR.The ultimate decision of the courts will have a significantimpact on the implementation of CAMR.Clean Air Interstate RuleThe final CAIR was promulgated by the EPA inMarch 2005 and published in the Federal Register asa final rule in May 2005 [11]. The rule is intended toreduce the atmospheric interstate transport of fineparticulate matter (PM 2.5 ) and ozone [12]. Both SO 2and NO x are precursors of PM 2.5 .NO x is also a precursorto the formation of ground-level ozone. CAIRwould require 28 States and the District of Columbiato reduce SO 2 and/or NO x emissions in a two-phaseprogram. The Phase I cap for NO x becomes effectivein 2009, and the Phase I cap for SO 2 starts in 2010[13]. The Phase II limits for both NO x and SO 2 startin 2015. The rule would apply to all fossil-fuel-firedboilers and turbines serving electrical generatorswith capacity greater than 25 megawatts that provideelectricity for sale. It would also apply to CHP unitslarger than 25 megawatts that sell at least one-thirdof their potential electrical output and supply moreFigure 9. Sulfur dioxide emissions in selectedStates, 1980-2003 (thousand short tons)1,2001,000800600400TexasNew YorkNorth Carolina200MissouriMassachusetts01980 1985 1990 1995 1999 2003than 219,000 megawatthours of electricity to the grid.Table 8 shows EPA estimates of CAIR’s impacts onSO 2 and NO x emissions. The AEO2006 reference caseprojections for SO 2 and NO x emissions are very closeto the EPA numbers.Under CAIR, the States would be responsible for allocatingNO x emissions allowances and taking the leadin pursuing enforcement actions, and they wouldhave flexibility in choosing the sources to be controlled.They could meet the emissions reductionrequirements either by joining the EPA-managed capand trade program for power plants or by achievingreductions through emissions control measures onsources in other sectors (industrial, transportation,residential, or commercial) or on a combination ofelectricity generating units and sources in other sectors.The 28 CAIR States are required to submit StateImplementation Plans (SIPs) to the EPA by September2006, showing how they intend to meet theirrespective caps.In order to participate in the cap and trade program,States would be required to regulate power plantemissions within their boundaries. The EPA would beresponsible for assigning State emissions budgets,reviewing and approving State plans, and administeringthe emissions and allowance tracking systems.Sources currently subject to the CAAA90 Title IVrules and to the NO x SIP Call trading program canuse allowances banked from those programs before2010 for compliance with CAIR. CAIR would requireadditional reductions in NO x emissions for Statesaffected by the NO x SIP Call. State NO x emissionscaps are based on each State’s share of region-wideheat input.The EPA plans to meet the SO 2 emission reductionrequirements by implementing a progressively morestringent retirement ratio on SO 2 allowances for electricitygenerating units of different vintages underthe CAAA90 Title IV Acid Rain Program. New SO 2allowances would not be issued under CAIR; powerTable 8. Estimates of national trends in annualemissions of sulfur dioxide and nitrogen oxides,2003-2020 (million short tons)ProjectionsEmissions 2003 2010 2015 2020EPASulfur dioxide 10.6 6.1 4.9 4.2Nitrogen oxides 4.2 2.4 2.1 2.1AEO2006Sulfur dioxide 10.6 5.9 4.6 4.0Nitrogen oxides 4.2 2.3 2.1 2.128 Energy Information Administration / Annual Energy Outlook 2006

Legislation and Regulationsplants would instead use the current pool of SO 2allowances issued under Title IV. Allowances issuedfor vintage years 2004 through 2009 could be retiredon a 1-to-1 basis, but allowances issued for vintageyears 2010 through 2014 would have to be retired on a2-to-1 basis, requiring two Title IV allowances to beretired for each ton of SO 2 emissions. Allowancesissued for vintage years 2015 and later would beretired on a basis of approximately 2.9 to 1. Thisretirement procedure is designed to integrate theCAIR rules with the existing Title IV SO 2 emissionsreduction program.Clean Air Mercury RuleCAMR (proposed as the Utility Mercury ReductionRule) for controlling mercury emissions from newand existing coal-fired power plants was promulgatedby the EPA in March 2005 and published as a finalrule in the Federal Register in May 2005 [14]. Powerplants with capacity greater than 25 megawatts andCHP units larger than 25 megawatts that sell at leastone-third of their electricity would be subject toCAMR.Under CAMR, Section 112 of the CAAA90 would bemodified to allow regulation of mercury emissionsunder a cap and trade program. The EPA estimatesthat CAMR, using the cap and trade approach, wouldreduce mercury emissions by nearly 70 percent whenfully implemented. The program would be implementedin two phases with a banking provision. ThePhase I cap, to be met in 2010, would be 38 short tons;the Phase II cap, to be met in 2018, would be 15 shorttons. In addition to these national caps, new powerplants would be subject to output-based limits onmercury emissions.Under the cap and trade approach, States would submitplans to the EPA to demonstrate that they wouldmeet their assigned State-wide mercury emissionsbudgets. With EPA approval, the States could thenparticipate in the cap and trade program. Allowanceswould be allocated by the States to power companies,which could either sell or bank any excess allowances.The final rule does not include a safety valve mechanismfor allowance prices.Update on Transition to Ultra-Low-SulfurDiesel FuelOn November 8, 2005, the EPA Administrator signeda direct final rule that will shift the retail compliancedate for offering ULSD for highway use from September1, 2006, to October 15, 2006. The change willallow more time for retail outlets and terminals tocomply with the new 15 parts per million (ppm) sulfurstandard, providing time for entities in the diesel fueldistribution system to flush higher sulfur fuel out ofthe system during the transition. Terminals will haveuntil September 1, 2006, to complete their transitionsto ULSD. The previous deadline was July 15, 2006.There is no change in the June 1, 2006, start date forrefiners to be producing ULSD. Also, during theextended transition period, diesel fuel meeting a22-ppm level can be temporarily marketed as ULSDat the retail pump. Finally, the EPA extended thebeginning date for the restriction on how much ULSDcan be downgraded to higher sulfur fuel by 15 days, toOctober 15, 2006, to be consistent with the end of thenew transition dates.The 45-day transition delay will help to ensurenationwide availability of 15-ppm ULSD before theintroduction of new model year 2007 diesel trucksand buses designed to operate on the improved fuel.These minor timing adjustments do not affect theAEO2006 projections.State Restrictions onMethyl Tertiary Butyl EtherBy the end of 2005, 25 States had barred, or passedlaws banning, any more than trace levels of MTBE intheir gasoline supplies, and legislation to ban MTBEwas pending in 4 others. Some State laws addressonly MTBE; others also address ethers such as ethyltertiary butyl ether (ETBE) and tertiary amyl methylether (TAME). AEO2006 assumes that all StateMTBE bans prohibit the use of all ethers for gasolineblending.Even with the removal of the oxygen content requirementfor RFG in EPACT2005, RFG is still expected tobe blended with ethanol, because it is not clear whereelse refiners could obtain the clean, high-octaneblending components needed to replace MTBE, whichsupplies 11 percent of the volume and a significantportion of the rated octane of RFG. Aromatic compoundsand olefins are high-octane blending components,but they are limited by the RFG requirementsand by the Federal Mobile Source Air Toxics program.Isooctane and alkylate are clean, high-octane blendingcomponents, but refinery capacity to producethem is limited, and it is often less expensive to useethanol at up to 10 percent by volume to offset part ofthe volume loss resulting from the removal of MTBE.As noted above, EPACT2005 also mandates the use of7.5 billion gallons of renewable motor fuels, such asEnergy Information Administration / Annual Energy Outlook 2006 29

Legislation and Regulationsethanol and biodiesel, by 2012 and requires renewablemotor fuel use to grow at the rate of overallmotor fuel use thereafter. In addition, some Stateshave their own renewable fuels programs. Minnesotacurrently requires all its gasoline supply to be blendedwith 10 percent ethanol, increasing to 20 percent ethanolif at least 50 percent of the new cars sold in theState can be guaranteed by their manufacturers to becompatible with the higher blend. Most currentautomobiles can use a maximum of only 10 percentethanol in gasoline, and automakers worry that widespreaduse of gasoline with 20 percent ethanol contentwill result in misfueling of vehicles not designedto use more than 10 percent ethanol.Several other State programs are contingent uponlocal ethanol supplies. Montana’s MTBE ban takeseffect only when 40 million gallons of ethanol productioncapacity is available in the State; and Hawaii hasa pending requirement for 85 percent of its gasolineto be blended with 10 percent ethanol if enough ethanolcan be produced in the State.Volumetric Excise Tax Creditfor Alternative FuelsOn August 10, 2005, President Bush signed intolaw the Safe, Accountable, Flexible, and EfficientTransportation Equity Act: A Legacy for Users(SAFETEA-LU) [15]. The act includes authorizationfor a multitude of transportation infrastructure projects,establishes highway safety provisions, providesfor R&D, and includes a large number of miscellaneousprovisions related to transportation, most ofwhich are not included in AEO2006 because theirenergy impacts are vague or undefined. Section11113, which provides a volumetric excise tax creditof 50 cents per gallon for alternative fuels, such as liquidfuels derived from the Fischer-Tropsch process, isincluded in AEO2006. This tax credit is expected tohave a small impact on transportation energy consumption,because it is scheduled to expire on September30, 2009, and only a small quantity ofalternative fuels will be produced in the pilot or demonstrationprojects that are expected to qualify for thecredit.30 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focus

Issues in FocusIntroductionThis section of the AEO provides in-depth discussionson topics of special interest that may affect the projections,including significant changes in assumptionsand recent developments in technologies for energyproduction, energy consumption, and emissions controls.With world oil prices escalating in recent years,this year’s discussions place special emphasis onworld oil prices, including a discussion of EIA’s worldoil price outlook, the impact of higher world oil priceson economic growth, and changing trends in the U.S.refinery industry.AEO2006 extends the AEO projections to 2030 for thefirst time. An important uncertainty with a longerprojection time horizon concerns the developmentand implementation of various technologies. Accordingly,this section includes a discussion of those technologiesthat, if successful, could affect the energysupply and demand projections in later years, focusingon energy technologies that could have theirgreatest impacts toward the end of the projectionperiod, those expected to have the greatest impact inthe automotive sector, and nonconventional liquidstechnologies that will play a growing role in meetingU.S. energy needs.World Oil Prices in AEO2006World oil prices in the AEO2006 reference case aresubstantially higher than those in the AEO2005 referencecase. In the AEO2006 reference case, worldcrude oil prices, in terms of the average price ofimported low-sulfur, light crude oil to U.S. refiners,decline from current levels to about $47 per barrel(2004 dollars) in 2014, then rise to $54 per barrel in2025 and $57 per barrel in 2030. The price in 2025 isapproximately $21 per barrel higher than the correspondingprice projection in the AEO2005 referencecase (Figure 10).The oil price path in the AEO2006 reference casereflects a reassessment of the willingness of oil-richcountries to expand production capacity as aggressivelyas envisioned last year. It does not represent achange in the assessment of the ultimate size of theworld’s petroleum resources but rather a lower levelof investment in oil development in key resource-richregions than was projected in AEO2005. Several factorscontribute to the expectation of lower investmentand oil production in key oil-rich producing regions,including continued strong worldwide economicgrowth despite high oil prices, and various restrictionson access and contracting that affect oil explorationand production companies.Although oil prices have stayed above $40 for the past2 years, world economies have continued to growstrongly: in 2004, global GDP registered the largestpercentage increase in 25 years. As a result, majoroil-exporting countries are likely to be less concernedthat oil prices will cause an economic downturn thatcould significantly reduce demand for their oil. Wheneconomies continue to grow despite higher oil prices,key suppliers have much less incentive to expand productionaggressively, because doing so could result insubstantially lower prices. Given the perceived lowresponsiveness of oil demand to price changes, suchan action could lower the revenues of oil exportersboth in the short term and over the long run.International oil companies, which normally areexpected to increase production in an environment ofhigh oil prices, lack access to resources in some keyoil-rich countries. There has been increased recognitionthat the situation is not likely to change over theprojection period. Furthermore, even in areas whereforeign investment by international oil companies ispermitted, the legal environment is often unreliableand complex and lacks clear and consistent rules ofoperation. For example, Venezuela is now attemptingto change existing contracts in ways that may makeoil company investments less attractive. In 2005, Russiaannounced a ban on majority foreign participationin many new natural resource projects and imposedhigh taxes on foreign oil companies. These changes,and others like them, make investment in oil explorationand development less attractive for foreign oilcompanies.The structure of many production-sharing agreementsalso increases the risk faced by major oil companiesin volatile oil price environments. Manycontracts guarantee a return to the host governmentat a fixed price, plus some percentage if the actualFigure 10. World oil prices in the AEO2005 andAEO2006 reference cases (2004 dollars per barrel)605040302010AEO2005HistoryProjections02000 2004 2010 2015 2020 2025 2030AEO200632 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focusworld oil price increases. The foreign company bearsthe full risk if the actual oil price falls below the guaranteedprice but does not reap significant rewards ifthe actual price is higher than the guaranteed price.This asymmetrical risk sharing discourages investmentwhen oil prices are likely to remain volatile. Itmay also hurt the oil-rich countries, if limited foreigninvestment prevents them from realizing the benefitsof the major technological advances that have beenmade in the oil sector over the past two decades.Because OPEC has less incentive to invest in expansionsof oil production capacity than was assumed inAEO2005, and because contracting provisions affectinginternational exploration and production companieshave shifted more risk to those companies, theAEO2006 reference case projects slower outputgrowth from key oil-rich countries after 2014 thanwas projected in the AEO2005 reference case.Energy market projections are subject to considerableuncertainty, and oil price projections are particularlyuncertain. Small shifts in either oil supply or demand,both of which are relatively insensitive to pricechanges in the short to mid-term, can necessitatelarge movements in oil prices to restore the balancebetween supply and demand. To address uncertaintyabout the oil price projections in the AEO2006 referencecase, two alternative cases posit world oil pricesthat are consistently higher or lower than those in thereference case. These high and low price cases shouldnot be construed as representing the potential rangeof future oil prices but only as plausible cases givenchanges in certain key assumptions.The high and low price cases in AEO2006 are basedon different assumptions about world oil supply. TheAEO2006 reference uses the mean oil and gasresource estimate published by the U.S. GeologicalSurvey (USGS) [16]. The high price case assumes thatthe worldwide crude oil resource is 15 percent smallerand is more costly to produce than assumed in the referencecase. The low price case assumes that theworldwide resource is 15 percent more plentiful andis cheaper to produce than assumed in the referencecase. Thus, the major price differences across thethree cases reflect uncertainty with regard to both thesupply of resources (primarily undiscovered andinferred) and the cost of producing them.Figure 11 shows the three price projections. As comparedwith the reference case, the world oil price in2030 is 68 percent higher in the high price case and 41percent lower in the low price case. As a result, worldoil consumption in 2030 is 13 percent lower in thehigh price case and 8 percent higher in the low pricecase than in the reference case. The high and lowprice cases illustrate that estimates of world oilresources that are lower and higher than the estimateused in the reference case can play a significant role indetermining future oil prices.The projections for world petroleum consumption in2030 are 102, 118, and 128 million barrels per day inthe high, reference, and low price cases, and the projectedmarket share of world petroleum liquids productionfrom OPEC in 2030 is about 31 percent in thehigh price case and 40 percent in the reference caseand low price cases. Because assumed productioncosts rise from the low price case to the reference caseto the high price case, the differences in net profitsamong the three cases are smaller than they mighthave been if the underlying supply curves for OPECand non-OPEC producers had remained unchanged.Although OPEC produces less output in the highprice case than in the reference case, its economicprofits are also less, because resources are assumed tobe tighter and exploration and production costshigher for conventional oil worldwide. In the absenceof tighter resources and higher costs, an OPEC strategythat attempted to pursue the output path in thehigh price case would subject OPEC to the risk of losingmarket share to other producers, as well as toalternatives to oil. Further discussions of the threeprice cases and their implications for energy marketsappear in other sections of AEO2006.Economic Effects of High Oil PricesThe AEO2006 projections of future energy marketconditions reflect the effects of oil prices on the macroeconomicvariables that affect oil demand, in particular,and energy demand in general. The variablesinclude real GDP growth, inflation, employment,exports and imports, and interest rates.Figure 11. World oil prices in three AEO2006 cases,1980-2030 (2004 dollars per barrel)10080604020HistoryProjections01980 1995 2004 2015 2030High priceReferenceLow priceEnergy Information Administration / Annual Energy Outlook 2006 33

Issues in FocusAlthough there is wide agreement that high oil priceshave negative effects on U.S. macroeconomic variables,the magnitude and duration of the effects areuncertain. For example, most of the major economicdownturns in the United States, Europe, and the AsiaPacific region since the 1970s have been preceded bysudden increases in crude oil prices. Although otherfactors were important, high oil prices played a criticalrole in substantially reducing economic growth inmost of these cases. Recent history, however, tells asomewhat different story. Average world crude oilprices have increased by more than $30 per barrelsince the end of 2001, yet U.S. economic activity hasremained robust, growing by approximately 2.8 percentper year from 2001 through 2004.This section describes the ways in which oil pricesaffect the U.S. economy [17], presents a brief surveyof the empirical literature on the economic impacts ofchanges in oil prices, and outlines the effects on theAEO2006 reference case projections of alternativeassumptions in the high and low price cases. Theresults of the alternative cases indicate how the U.S.economy is likely to be affected by different levels ofoil prices.Macroeconomic Impacts of High Oil PricesU.S. demand for crude oil arises from demand for theproducts that are made from it—especially gasoline,diesel fuel, heating oil, and jet fuel; and changes incrude oil prices are passed on to consumers in theprices of the final petroleum products. Increases incrude oil prices affect the U.S. economy in five ways:• When the prices of petroleum products increase,consumers use more of their income to pay foroil-derived products, and their spending on othergoods and services declines. The extra amountsspent on those products go to foreign and domesticoil producers and, if wholesale margins increase,to refiners. Domestic producers may payhigher dividends and/or spend more on oil discovery,production, and distribution. Foreign producersmay spend some or all of their extra revenueson U.S. goods and services, but the types of goodsand services they buy will be different from thosethat domestic consumers would buy. How quicklyand how much domestic and foreign oil producersspend on U.S. goods and services and financialand real assets will be critical in determining theeffects of higher oil prices on the aggregate economy[18].• Oil is also a vital input for the production of a widerange of goods and services, because it is used fortransportation in businesses of all types. Higheroil prices thus increase the cost of inputs; and ifthe cost increases cannot be passed on to consumers,economic inputs such as labor and capitalstock may be reallocated. Higher oil prices cancause worker layoffs and the idling of plants, reducingeconomic output in the short term.• Because the United States is a net importer of oil,higher oil prices affect the purchasing power ofU.S. national income through their impact on theinternational terms of trade. The increased priceof imported oil forces U.S. businesses to devotemore of their production to exports, as opposed tosatisfying domestic demand for goods and services,even if there is no change in the quantity offoreign oil consumed.• Changes in oil prices can also cause economiclosses when macroeconomic frictions preventrapid changes in nominal prices for final goods(due to the costs of changing “menu” prices) or forkey inputs, such as wages. Because there is resistanceon the part of workers to real declines inwages, oil price increases typically lead to upwardpressure on nominal wage levels. Moreover, nominalprice “stickiness” is asymmetric, in that firms,unions, and other organizations are much morereluctant to lower nominal prices and the wagesthey receive than they are to raise them. When anominal increase in oil prices threatens purchasingpower, the adjustment process is slowed, withmultiplier effects throughout the economy [19].• Finally, higher oil prices cause, to varying degrees,increases in other energy prices. Dependingon the ability to substitute other energy sourcesfor petroleum, the price increases can be large andcan cause macroeconomic effects similar to the effectsof oil price increases.The nature of the oil price increases, the state of theeconomy, and the macroeconomic policies undertakenat the time may accentuate or dampen theseverity of adverse macroeconomic effects. If priceincreases are large and sudden, their impacts onshort-term growth may be much larger than if theyare gradual, because sudden oil price shocks scarehouseholds and firms and prevent them from makingoptimal decisions in the near term.On the potential output side, sudden large priceincreases create widespread uncertainty about appropriateproduction techniques, purchases of newequipment and consumer durable goods like automobiles,and wage and price negotiations. As firms andhouseholds adjust to the new conditions, some plant34 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focusand equipment will remain idle, some workers will betemporarily unemployed, and the economy mayno longer operate along its long-run productionpossibilityfrontier. Although it is easy to differentiategradual from rapid price increases on a conceptualbasis, empirical differentiation is more difficult.In terms of the state of the economy, if the economyis already suffering from high inflation and unemployment,as in the late 1970s, then the oil priceincreases have the potential to cause severe damageby limiting economic policy options. Many analystsassert that it was the monetary policy undertaken inthe 1970s that really damaged the U.S. economy.The economic policies that are followed in response toa combination of higher inflation, higher unemployment,lower exchange rates, and lower real outputalso affect the overall economic impact of higher oilprices over the longer term. Sound economic policiesmay not completely eliminate the adverse impacts ofhigh oil prices described above, but they can moderatethem. Conversely, inappropriate economic policiescan exacerbate the adverse impacts. Overly contractionarymonetary and fiscal policies to contain inflationarypressures can worsen the recessionary effectson income and unemployment; expansionary monetaryand fiscal policies may simply delay the fall inreal income necessitated by the increase in oil prices,stoke inflationary pressures, and worsen the impactof higher prices in the long run.Empirical Studies of Oil Price EffectsThe mechanism by which oil prices affect economicperformance is generally well understood, but theprecise dynamics and magnitude of the effects areuncertain. Quantitative estimates of the overall macroeconomicdamage caused by oil price shocks in thepast and of the economic gains realized by oilimportingcountries as a result of the oil price collapsein 1986 vary substantially, in part because of differencesin the models used to examine the issue [20].Two different approaches have been used to estimatethe magnitude of oil price effects on the U.S. economy.One uses large, disaggregated macroeconomicmodels of the economy, and the other uses time-seriesanalysis of historical events to estimate directly themacroeconomic effects of oil price changes.In the first approach, macroeconomic models are usedin attempts to account for all the relationships amongthe major macroeconomic variables in the economy(as described by the National Income and Product,Balance of Payments, and Flow of Funds Accounts),and historical data are used to estimate statisticallythe parameters linking the variables. The advantagesof macroeconomic models are consistent accountingof macroeconomic relationships over time and theability to account for other events taking place.A recent Stanford University Energy Modeling Forum(EMF) study by Hillard Huntington found thatmost macroeconomic models report similar economiceffects of oil price increases [21]. Table 9 shows theresults for real GDP, the GDP price deflator, andunemployment obtained from three models and theiraverages [22]. The results are shown for a 33-percentincrease in the oil price, from $30 to $40. For example,the output results in Table 9 imply that a 33-percentincrease in the oil price sustained for 2 years reducesreal GDP relative to the baseline by 0.2 percent in thefirst year and 0.5 percent in the second year. In termsof an elasticity response of real GDP to oil price, thepercentage change in real GDP relative to the percentagechange in oil price is approximately 0.01 inthe first year and 0.02 in the second year.The second approach is simpler, focusing specificallyon the relationship between changes in crude oilprices and some measure of their economic impact,such as aggregate output, inflation, or unemployment.Time-series analyses of historical data are usedto estimate statistically an equation (or a system ofequations called “vector autoregressions”) thatexplains economic growth rates as a function of thepast growth in the economy and past changes in crudeoil prices. Many studies add the past values of additionalvariables to the system in order to incorporatetheir interactions with the oil price and GDPvariables.Table 9. Macroeconomic model estimates ofeconomic impacts from oil price increases(percent change from baseline GDPfor an increase of $10 per barrel)Estimate Year 1 Year 2Global Insight, Inc.Real GDP -0.3 -0.6GDP price deflator 0.2 0.5Unemployment 0.1 0.2U.S. Federal Reserve BankReal GDP -0.2 -0.4GDP price deflator 0.5 0.3Unemployment 0.1 0.2National Institute of Economic and Social ResearchReal GDP -0.2 -0.5GDP price deflator 0.3 0.5AverageReal GDP -0.2 -0.5GDP price deflator 0.3 0.4Unemployment 0.1 0.2Energy Information Administration / Annual Energy Outlook 2006 35

Issues in FocusTable 10 shows results for the U.S. economy from arecent study by Jimenez-Rodriguez and Sanchez [23],which are representative of the results obtained inthe time-series literature. Due to the nature of thereduced-form framework used, the results are directestimates of GDP elasticities with respect to oil pricechanges as of the given quarter after the permanentprice change. The asymmetric results allow separateestimates of GDP elasticity for oil price increases,decreases, and net increases (when oil prices exceedthe maximum over the previous 12 quarters). Whenthe six-quarter GDP elasticity estimated by Jimenez-Rodriguez and Sanchez (approximately 0.05) isapplied to a 33-percent price increase (to be comparablewith the average macroeconomic simulationresponse in Table 10), real GDP declines by 1.7 percent—morethan 3 times the effect on real GDP inmacroeconomic simulations.Generally, as indicated by the results in Table 10,time-series studies show larger impacts on outputand other variables than do macroeconomic simulations.Huntington offers four major reasons as to whythe empirical estimates are so different:• The larger impacts calculated from direct statisticalestimations often are attributed to a range ofmacroeconomic frictions that could make theeconomy’s response to an oil price shock fundamentallydifferent from its response to a smallerincrease in oil prices. Large macroeconomic modelsdo not differentiate between oil price increasesand decreases, or between surprise events andmore gradual price adjustments.• The larger estimates from time-series models mayalso reflect baseline economic conditions beforean oil price disruption that are fundamentally differentfrom today’s economic environment. Forexample, the oil price shocks of the 1970s hit theU.S. economy when it already was experiencinginflationary pressures.• Historical oil price shocks reduced not only aggregateoutput but also the country’s purchasingpower. Real national income fell as the costsof buying international goods (including oil)increased more than income from exports. Thehigher prices made the country poorer by requiringmore exports to balance each barrel of importedoil, leaving less aggregate output fordomestic consumption.• The oil price shocks of the 1970s completely surprisedfirms and households in many differentcountries at the same time. Firms and householdsmade decisions about production and prices thathad important consequences for the strategies ofother firms in the economy [24]. And yet, therewas little opportunity to coordinate strategies insuch an uncertain world. Now, after several differentoil price episodes, there has been significantlearning about how to cope with theuncertainties created by oil price shocks. It is unlikelythat firms and households will be surprisedin the same way or to the same degree as theywere by earlier shocks.If crude oil prices rise early in a particular year, whatwill be the impact on the economy at the end of thefollowing year? Huntington offers the following tentativeanswers, and Table 11 summarizes the impactson GDP, as well as the impacts on the GDP pricedeflator for all goods and services and the unemploymentrate. If the economy is operating at its potentialoutput level and inflation is constant, a reasonableestimate is that a 10-percent increase in the price ofoil that does not surprise households and firms(higher oil price in Table 11) will reduce potential output(GDP) by 0.2 percent. If the economy is operatingwell below its potential output level, the impact onGDP may be somewhat larger but is unlikely toexceed 0.2 percent after the first year. If the oil priceincrease comes as a complete surprise and the economyis already in a rising inflationary environment(oil price shock in Table 11), then it has the potentialto cause larger economic losses, which would be closerto those predicted by time-series models.Table 10. Time-series estimates of economic impacts Table 11. Summary of U.S. oil price-GDP elasticitiesfrom oil price increases (percent change frombaseline GDP for an increase of $10 per barrel) Price effect Year 1 Year 2AsymmetricQuarter Price increase Price decreaseNet priceincrease4 -0.048 -0.014 -0.0466 -0.051 0.002 -0.0588 -0.046 0.011 -0.05410 -0.044 0.010 -0.04812 -0.042 0.010 -0.043Higher oil priceReal GDP -0.011 -0.021GDP price deflator 0.007 0.017Unemployment rate 0.004 0.007Oil price shockReal GDP -0.024 -0.050GDP price deflator 0.019 0.034Unemployment rate 0.009 0.02036 Energy Information Administration / Annual Energy Outlook 2006

Issues in FocusAEO2006 Price CasesThe key feature of the AEO2006 high and low worldoil price paths is that they are not characterized bydisruption, but rather represent a gradual and sustainedmovement relative to the reference case path.Keeping this distinction in mind, the MacroeconomicActivity Module in NEMS, which contains the GlobalInsight Inc. (GII) Macroeconomic Model, is used toassess the economic impacts of the alternative pricepaths.Most of the results projected for the U.S. economy inthe high and low price cases relative to the referencecase are similar to the results for macroeconomicmodels discussed above. The AEO2006 high and lowprice cases are unique, however, in that they traceout, in a consistent manner, both the short-termimpacts of oil price increases and the longer termadjustments of the economy in response to sustainedhigh and low prices by employing a disaggregatedmacroeconomic model integrated with a very detailedenergy market model—NEMS.Figure 12 shows the percentage change from the referencecase projections for real GDP and oil prices inthe AEO2006 high and low price cases. In the highprice case, oil prices rise rapidly to 70 percent abovereference case prices within 10 years (2016), thenclimb more gradually to 80 percent above referencecase prices in 2030. In the low price case, oil prices donot change by as much relative to the reference case,declining to 34 percent below reference case prices in2016 and 44 percent below in 2030. Consequently, themacroeconomic effects in the two cases are notexpected to be symmetric.In each of the three cases, the U.S. economy growsat an average annual rate of 3.0 percent from 2004Figure 12. Changes in world oil price and U.S. realGDP in the AEO2006 high and low price cases,2004-2030 (percent difference from reference case)2.080.01.5Oil price01.0-44.02004 20300.50.0-0.5-1.0-1.5Real GDP-2.02004 2010 2015 2020 2025 2030Low priceHigh pricethrough 2030 (although the average growth rates inthe three cases do differ when calculated to two ormore decimal places). With such significant differencesin oil price paths in the three cases, why is theimpact on the long-term real GDP growth rate sosmall? The major reasons have to do with the natureof the oil price increases and decreases relative to thereference case and their short-term versus long-termimpacts on the economy.The oil price projections for 2005 and 2006 are thesame in the three cases. From 2007 to 2010, the realoil price increases by more than 2 percent annually inthe high price case, declines by 5 percent annually inthe reference case, and declines by 9.4 percent annuallyin the low price case. From 2010 to 2015, theannual changes in oil prices in the three cases average4 percent, -0.5 percent, and -5 percent, respectively.After 2015 the differences narrow considerably, andby 2030 the annual increases in oil prices average 1.1percent in the high price case, 0.8 percent in the referencecase, and zero in the low price case. With themaximum differences in growth rates among thethree cases occurring in 2010, the peak impacts onreal GDP and other economic variables occur approximately2 years later, in 2012.Over the 2006-2030 period, real GDP in the high priceand low price cases deviates from that in the referencecase for a considerable period. As the economyadjusts to the oil price changes, however, the differencesbecome smaller, and by 2030 real GDP isapproximately the same in the three cases, at $23,112billion in the reference case, $23,054 billion in thehigh price case, and $23,178 billion in the low pricecase.The discounted sum of changes in real GDP over theentire projection period provides a better indicator ofnet effects on the economy. In the low price case, thesum of the changes in real GDP, discounted at a7-percent annual rate, over the 2006-2030 period is$665 billion, and in the high price case the sum is-$869 billion. These sums represent approximately0.4 percent and -0.5 percent, respectively, of the totaldiscounted real GDP in the reference case over thesame period.The elasticity of real GDP with respect to oil pricechanges over the 2006-2030 period is -0.007 in boththe high price and low price cases. The year-by-year(marginal) and up-to-the-year (average) elasticities ofreal GDP with respect to oil price changes in the highprice case (Figure 13) shows that the short-termeffects of oil price increases are larger than theirlong-term effects.Energy Information Administration / Annual Energy Outlook 2006 37

Issues in FocusTo portray the short-term dynamics of the economyas it reacts to oil price changes, Table 12 shows 5-yearaverage annual growth rates for U.S. oil prices (theimported refiners acquisition cost of crude oil), realGDP, potential GDP, and the consumer price index(CPI), as well as 5-year averages for the Federal fundsrate and unemployment rate, over the 2005-2030period. Higher oil prices in the short term feedthrough the economy and reduce aggregate expenditureson goods and services. As aggregate demand isless than aggregate supply, unemployment increases.With higher prices there would also be a tendency forinterest rates to rise. In the high price case, real GDPgrowth averages 3 percent per year over the 2005-2010 period, CPI inflation averages 2.3 percent peryear, and the average unemployment rate for the5-year period is 5 percent. In the reference case, thecomparable rates are 3.2 percent (average annual realGDP growth), 2 percent (average annual CPI inflation),and 4.8 percent (unemployment). PotentialGDP growth and the Federal funds rate are not significantlydifferent in the two cases over the 2005-2010 period. The impacts of high prices on real GDPshown in Table 12 are in agreement with the averageresults shown in Table 9.In the high price case, as unemployment increases,the Federal Reserve lowers the Federal funds ratefrom its projected level in the reference case. At thesame time, total employment costs are lower, whichtends to slow price growth in the economy. Over the2010-2015 period, even though oil prices continue togrow by 4.1 percent annually in the high price case (asopposed to declining by 0.5 percent annually in thereference case), real GDP growth is about the same inthe two cases, although it is increasing from a lowerFigure 13. GDP elasticities with respect to oil pricechanges in the high price case, 2006-20300.000-0.010-0.020-0.0302005 2010 2015 2020 2025 2030Year-by-yearPeriod averageTable 12. Economic indicators in the reference, high price, and low price cases, 2005-2030 (percent)Indicator 2005-2010 2010-2015 2015-2020 2020-2025 2025-2030 2005-2030Reference caseAverage annual growth ratesOil price -2.3 -0.5 0.9 1.3 0.8 0.0Real GDP 3.2 2.9 3.1 2.8 2.8 3.0Potential GDP 3.3 2.4 2.6 2.8 2.8 2.8Consumer price index 2.0 2.7 3.0 3.0 2.8 2.75-year averagesFederal funds rate 4.6 5.4 5.4 5.1 5.0 5.1Unemployment rate 4.8 4.7 4.4 4.6 4.9 4.7High price caseAverage annual growth ratesOil price 3.6 4.1 2.1 1.2 1.2 2.4Real GDP 3.0 2.9 3.2 2.8 2.8 2.9Potential GDP 3.2 2.4 2.7 2.8 2.8 2.8Consumer price index 2.3 2.9 2.8 2.7 2.7 2.75-year averagesFederal funds rate 4.6 5.2 4.9 4.7 4.7 4.8Unemployment rate 5.0 5.2 4.7 4.7 4.9 4.9Low price caseAverage annual growth ratesOil price -5.6 -4.8 -0.7 0.0 0.0 -2.3Real GDP 3.3 3.0 3.0 2.8 2.8 3.0Potential GDP 3.3 2.4 2.6 2.9 2.9 2.8Consumer price index 1.9 2.6 3.1 3.0 2.9 2.75-year averagesFederal funds rate 4.5 5.5 5.6 5.3 5.3 5.2Unemployment rate 4.8 4.5 4.2 4.5 4.8 4.638 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focusbase in the high price case. The Federal funds rate islower in the high price case than in the reference case,and the unemployment and CPI inflation rates arehigher.After 2015, as the differential in the oil price growthrates between the high price and reference casesshrinks, rebound effects from the lower employmentcosts and lower Federal funds rate in the high pricecase are stronger than the contractionary impacts ofhigher oil prices, leading to higher real GDP growthand lower CPI inflation than in the reference case. Asa result, in 2030, the real GDP growth rate and unemploymentrate in the high price case are nearly thesame as in the reference case, but the Federal fundsrate is lower.The assumptions behind the oil price cases are that:the price changes do not come as a shock and come tobe expected over time; the Federal Reserve is able tocarry out an activist monetary policy effectively,because core inflation remains low; exchange rates donot change from those in the reference case; andother countries experience impacts similar to those inthe United States. Changes in any of these assumptionscould increase the projected impacts on the U.S.economy.The economic impact of oil price changes is an issuethat continues to attract considerable attention, especiallyat this time, when oil prices have continued torise over the past 3 years. Over the past 30 years,much has been learned about the nature of the economicimpacts and the extent of damage possible.Empirical estimates based on history provide two setsof results. In the 1970s and 1980s the damages weresubstantial, and it is believed that recession followed—andmay have been caused by—the oil priceincreases. Current literature suggests that, in today’sU.S. economy, sustained higher oil prices can slowshort-term growth but are not likely to cause a recessionunless other factors are present that shock economicdecisionmakers or lead to inappropriateeconomic policies. The AEO2006 high and low pricecases provide estimates of the economic impacts onsuch an economy, and the projections in the pricecases are within the range that other macroeconomicmodels predict.Changing Trends in the Refining IndustryThere have been some major changes in the U.S.refining industry recently, prompted in part by a significantdecline in the quality of imported crude oiland by increasing restrictions on the quality offinished products. As a result, high-quality crudes,such as the WTI crude that serves as a benchmark foroil futures on the New York Mercantile Exchange(NYMEX), have been trading at record premiums tothe OPEC Basket price.WTI is a “light, sweet” crude: light because of its lowdensity and sweet because it has less than 0.5 percentsulfur content by weight. This combination of characteristicsmakes it an ideal crude oil to be refined in theUnited States, yielding a greater portion of its volumeas “light products,” including both gasoline and dieselfuel. Premium crudes like WTI yield almost 70 percentof their volume as light, high-value products,whereas heavier crudes like Mars (from the deepwaterGulf of Mexico) yield only about 50 percent oftheir volume as light products. The AEO2006 projectionsuse the average price of imported light, sweetcrudes as the benchmark world oil price [25].The average sulfur content of U.S. crude oil importsincreased from 0.9 percent in 1985 to 1.4 percent in2005 [26], and the slate of imports is expected to continue“souring” in coming years. Crude oils are alsobecoming heavier and more corrosive than they werein the past, largely because fields with higher qualityvarieties were the first to be developed, and refiners’preference for quality crudes has led to the depletionof those reserves over the past 100 years and reducedthe market share of the light, sweet crude thatremains.The industry standard measure for oil density is APIgravity; a lower gravity indicates higher density(heavy viscous oil), and a higher gravity indicateslower density (lighter, thinner oil). Over the past 20years, the API gravity of imported crude oil hassteadily declined, from 32.5 degrees to 30.2 degrees[27]. The standard measure for corrosiveness is thetotal acid number (TAN), indicating the number ofmilligrams of potassium hydroxide needed to neutralizethe acid in 1 gram of oil. The most corrosivecrudes, with TANs greater than 1, require significantaccommodation to be processed. Usually, their corrosivenessis mitigated by the addition of basic compoundsto neutralize the acid; however, some refinershave chosen instead to upgrade all their piping andunit materials to stainless steel. Whereas there werevirtually no high-TAN crudes processed in 1990, theynow make up about 2 percent of the crude oil slate,and a Purvin & Gertz forecast indicates that they willincrease to 5 percent or more in 2020 [28] (Figure 14).As refining inputs have declined in quality, demandfor high-quality refined products has increased. TheEnergy Information Administration / Annual Energy Outlook 2006 39

Issues in FocusEPA has developed new environmental rules that willrequire refineries to reduce the amount of sulfur inmost gasoline to 30 ppm by 2006, from over 400 ppmin the early 1990s, and the sulfur content of highwaydiesel fuel to 15 ppm by October 2006, from over 2,000ppm before 1993. By 2014, virtually all diesel fuelmust be below 15 ppm [29] (Figure 15). To meet thesespecifications at the pump, refiners must produce dieselcontaining one-half that amount of sulfur before itenters the distribution system, because the lowsulfurproduct is expected to pick up trace amounts ofsulfur as it moves through pipelines and other distributionchannels.To meet higher quality standards with poorer qualityfeedstocks will require significant investment by U.S.refiners. The principal method for reducing sulfurcontent in fuels is hydrotreating, a chemical processin which hydrogen reacts with the sulfur in crude oilto create hydrogen sulfide gas that can easily beremoved from the oil. Hydrotreaters are specializedfor the refinery streams they process. In aggregate,the dramatically lower sulfur specifications for petroleumfuels will necessitate a doubling of U.S.hydrotreating capacity by 2030, to 27 million barrelsa day, from 14 million barrels a day in 2004. Most ofthe new capacity (23.4 million barrels a day) isexpected to be installed by 2015 (Figure 16).the refinery. Extra heavy oils, like those from theOrinoco region in Venezuela or the Alberta tar sandsin Canada, are typically upgraded in a process that isboth capital- and energy-intensive but can yield ahighly desirable product. Canada’s Syncrude SweetBlend produced from tar sands is a high-quality syntheticcrude (syncrude) that trades at near paritywith WTI; however, the cost of the upgrades is almost$15 a barrel, in addition to the cost of tar sandsrecovery.The second approach is to “convert” heavy oil at therefinery directly to light products, in a process moretypical of the refining process for conventional oils.Chief among methods of conversion is thermal coking,in which heavy oil from a vacuum distillation unitis fed to a heating unit (coker) that splits off lighterhydrocarbon chains and routes them to the traditionalrefinery units. The almost pure carbon remainingis a coal-like substance known as petroleum coke.The accumulated coke can be removed from thecoking vessels during an off cycle and either sold,primarily as a fuel for electricity generation, or usedFigure 14. Purvin & Gertz forecast for world oilproduction by crude oil quality, 1990-2020(million barrels per day)100HistoryProjectionsLow maximum sulfur specifications may also haveimplications for products not directly affected by thepending EPA rules. Suppliers of such high-sulfurproducts as jet fuel, home heating oil, and residualfuel may have to find alternative distribution channelsif pipeline operators concerned about contaminationstop accepting high-sulfur fuels.As for adapting to heavier crude slates, there are twobasic approaches. The first is to “upgrade” the oil to alighter oil in the producing region, before it is sent to8060402001990 1995 2000 2005 2010 2015 2020High TANHeavy sourLight sourLight sweetFigure 15. Sulfur content specifications for U.S.petroleum products, 1990-2014 (parts per million)3,000HistoryNonroad dieselProjectionsFigure 16. U.S. hydrotreating capacity, 1990-2030(million barrels per day)30 History Projections252,000Highway diesel20151,000Traditional gasoline10Reformulated gasoline01990 1995 2000 2005 2010 2014501990 1995 2000 2004 2010 2015 2020 2025 203040 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focusin gasification units to provide power, steam, and/orhydrogen for the refinery.U.S. refineries are among the most advanced in theworld, and their technological lead will undoubtedlyleave U.S. refiners uniquely prepared to adapt andtake advantage of discounts available for processinginferior crudes. Adaptation will require extensivefuture investments, however, and may take sometime to achieve.Energy Technologies on the HorizonA key issue in mid-term forecasting is the representationof changing and developing technologies. Howexisting technologies will evolve, and what new technologiesmight emerge, cannot be known with certainty.The issue is of particular importance inAEO2006, the first AEO with projections out to 2030.For each of the energy supply and demand sectorsrepresented in NEMS, there are key technologiesthat, while they may not be important in the markettoday, could play a role in the U.S. energy economy by2030 if their cost and/or performance characteristicsimprove with successful R&D. Moreover, it is possible,if not likely, that technologies not yet conceivedcould be important 20 to 30 years from now. Althoughthe direction and pace of change are unpredictable,technological progress is certain to continue.Buildings SectorA variety of new technologies could influence futureenergy use in residential and commercial buildingsbeyond the levels projected in AEO2006. Two suchtechnologies are solid-state lighting and “zeroenergy” homes.Solid-state lighting. Solid-state lighting (SSL) is anemerging technology for general lighting applicationsin buildings. Two types of SSL currently under developmentare semiconductor-based light-emitting diode(LED) and organic light-emitting diode (OLED) technologies.Both are commercially available for specializedlighting applications. Consumers are likely to befamiliar with the use of LEDs in traffic signals, exitsigns and similar displays, vehicle tail lights, andflashlights. They are less likely to be familiar withOLEDs, used in high-resolution display panels forcomputers and other electronic devices.Lighting accounted for 16 percent of total primaryenergy consumption in buildings in 2004, second onlyto space heating at 20 percent. Thus, changes in theassumptions made about development and enhancementof SSL technologies could have a significantimpact on projected total energy consumption in residentialand commercial buildings through 2030.Beginning with AEO2005, SSL based on LED technologyhas been included as an option in the NEMSCommercial Module, based on currently availableproducts. Those products are more than four times asexpensive as comparable incandescent lighting, withonly slightly greater efficiency (called “efficacy” andmeasured in lumens per watt), and so have virtuallyno impact in the AEO2006 projections. In order forLEDs and OLEDs to compete successfully in generallighting applications, several R&D hurdles must beovercome: costs must be reduced, efficacy must beincreased, and improved techniques must be developedfor generating light with a high color renderingindex (CRI) that more closely approximates the spectrumof natural light and is needed for many buildingapplications.DOE’s R&D goals call for SSL costs to fall dramaticallyby 2030. The real promise for LED lighting isthat efficacies could approach 150 to 200 lumens perwatt—more than twice the efficacy of current fluorescenttechnologies and roughly 10 times the efficacy ofincandescent lighting [30]. An additional goal is toincrease LED operating lifetimes from 30,000 hoursto 100,000 hours or more, which would far exceed theuseful lifetimes of conventional technologies (generally,between 1,000 and 20,000 hours). Longer usefuloperating lives are particularly valuable in commercialapplications where lamp replacement representsa major element of lighting costs.For general illumination applications, OLED technologylags behind LED technology. If research goals arerealized, the advantages of OLED technology will belower production costs than LEDs, similar theoreticalefficacies (200 lumens per watt for white light), andthe flexibility to serve as a source of distributed lighting,as is currently provided by fluorescent lamps.Zero energy homes. DOE’s Zero Energy Homes (ZEH)program encompasses several existing technologiesrather than a single emerging technology. The ZEHprogram takes a “whole house” approach to reducingnonrenewable energy consumption in residentialbuildings by integrating energy-efficient technologiesfor building shells and appliances with solar waterheating and PV technologies to reduce annual netconsumption of energy from nonrenewable sources tozero [31]. This is an emerging integrated technology;the ZEH concept is novel for conventional housingunits [32]. ZEH prototypes have been shown to generatemore electric energy than they consume duringEnergy Information Administration / Annual Energy Outlook 2006 41

Issues in Focusperiods of peak demand for air conditioning, whileapproaching the goal of zero net annual energy purchases.The technological hurdle is to make ZEHhomes without subsidies both cost-competitive andattractive as alternatives to conventional homes.ZEH homes currently are not characterized or identifiedas an integrated technology in the NEMS ResidentialModule; however, most of the constituentZEH technologies are characterized as separateoptions. Several whole-house options are modeled,characterized according to their efficiencies relativeto current residential energy codes, with the followingoptions:• Current residential code• 30 percent more efficient than current code (modeledto meet ENERGY STAR requirements)• 40 percent more efficient than current code• 50 percent more efficient than current code (modeledalong the lines of PATH concepts [33])• Solar PV and solar water heating technologies.In addition to ZEH, a long list of emerging buildingstechnologies has been compiled by the AmericanCouncil for an Energy-Efficient Economy. Theyincluded six identified as high-priority technologieson the basis of such criteria as the cost of conservedenergy, savings potential, and likelihood of success:• For residential and small commercial buildings:1-watt standby power for consumer appliances,aerosol-based duct sealing, and leak-proof ducts• For commercial buildings: integrated building design,computerized building diagnostics, and“retro-commissioning” [34].Because they are still in the early stages of development,the information needed to characterize thesesix high-priority technologies or programs is not yetavailable, and they are not included in AEO2006;however, they do hold promise if they can be successfullycommercialized.Industrial SectorThe industrial sector is diverse, and there are manypotential technological innovations that could affectindustrial energy use over the next 25 years. Twotechnologies, fuel gasification and nanotechnologies,could have impacts across a broad array of industries.Gasification could be especially important to thepaper business; successful nanotechnologies couldhave very broad impacts.Black liquor gasification. Black liquor is a waste productfrom papermaking. It contains inorganic chemicalsthat are recovered for reuse in the papermakingprocesses and lignin from the initial pulpwood inputsthat is also recovered and used as a fuel for boilers andfor cogeneration. Current practice uses Tomlinsonboilers to recover the inorganic chemicals andcombust the organics to produce steam [35]. Blackliquor gasification coupled with a combined-cyclepower plant (BLGCC) has been proposed as a way tomake better use of the lignin and recover a larger portionof the inorganic chemicals from the liquor.R&D on BLGCC technology has been underway forseveral years. The American Forest and Paper Association’sAgenda 2020: Technology Vision and ResearchAgenda for America’s Forest, Wood and PaperIndustry, first published in 1994, has been revisedseveral times over the years. A recent progress reportindicates that successful industry-wide implementationof BLGCC could provide an additional 30gigawatts of on-site electricity generation capacitybeyond the 8 gigawatts operating in 2004 [36].DOE-sponsored R&D activities in support of BLGCCwere evaluated by the National Academy of Sciences(NAS) in a 2001 report [37], in which it was indicatedthat DOE’s expectation that Tomlinson boilers wouldbe replaced in a 10- to 20-year time frame probablywas optimistic. The report also noted that “movingfrom the existing black liquor gasification units tosystems suitable for use with combined cycle requiresbench-scale research as well as demonstration.” Thetechnology is not explicitly represented in AEO2006and is not expected to have an impact on the industrialsector in the reference case. In the high technologycase, the potential impact of BLGCC isrepresented as an increasing amount of biomassbasedCHP capacity, up to 3 gigawatts (43 percent)more than in the reference case in 2030.Nanotechnology. Nanotechnology refers to a widerange of scientific or technological projects that focuson phenomena at the nanometer (nm) scale (around0.1 to 100 nm) [38]. While not as far along as BLGCC,nanotechnologies have much larger potential impactsif they are successfully developed. Indeed, it has beensuggested that nanotechnology applications in theindustrial sector could yield a new industrial revolution[39]. Possible applications include, for example,very thin solar silicon panels that could be embeddedin paint [40]; very thin video screens with about thesame thickness and flexibility as newspapers, whichcould be updated continuously with current news[41]; and very strong, very light materials that could42 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focusrevolutionize transportation systems and dramaticallyreduce per capita energy consumption [42].While the potential applications of nanotechnologiesare diverse, many issues, including potential impactson human health, remain to be studied. AEO2006does not include potential energy applications ofnanotechnology, because they still are speculative.Transportation SectorThe transportation module in NEMS addresses technologiesspecific to light-duty vehicles, heavy trucks,and aircraft. The majority of the advanced technologiesrepresented reflect improvements to conventionalpower train components, including suchtechnologies as variable valve timing and lift, camlessvalve actuation, advanced light-weight materials,six-speed and continuously variable transmissions,cylinder deactivation, and electronically driven parasiticdevices (power steering pumps, water pumps,etc.). Vehicles powered by batteries or fuel cells arealso explicitly represented in AEO2006, but their penetrationresults largely from legislatively mandatedsales.Transportation technologies not currently includedin NEMS that could potentially become viable marketoptions include homogeneous charge compressionignition (HCCI), grid-connected hybrid vehicles, andhydraulic hybrid vehicles. HCCI—which combinesfeatures of both spark-ignited (gasoline) and compression-ignited(diesel) engines—can operate on avariety of fuels. In the HCCI engine, an extremelylean mixture of fuel and air is autoignited in the cylindervia compression. Autoignition can damage thepistons in spark-ignited engines, but the extremelyhigh air-to-fuel ratio in HCCI engines prevents flamepropagation and results in a much cooler burn. As aresult, HCCI engines are very efficient, with low levelsof emissions that do not require expensiveafter-treatment devices. The fuel properties and cylinderconditions needed for HCCI combustion arewell understood; however, it is extremely difficult tocontrol ignition in multiple-cylinder engines across awide range of load conditions, as needed for vehicleapplications.Grid-connected hybrid vehicles are similar to thehybrid vehicles sold today, except that the batteriesprovide an all-electric range of about 50 miles, and anexternal source to charge the batteries is required.Unlike current hybrid vehicles that use high-powerbatteries to supplement the power of gasolineengines, grid-connected hybrid vehicles are alsodesigned to operate as all-electric vehicles and, assuch, require a much larger battery pack for energystorage, a larger electric motor, and related componentsthat enable them to function over a much widerrange of driving conditions. Although all-electric drivinggreatly reduces the vehicles’ gasoline consumption,the costs of the battery pack and othercomponents are significant. Marketing studies haveindicated that there is a lack of consumer interest in“plug-in” vehicles but that a limited market wouldexist if their incremental costs relative to conventionalvehicles could be reduced to at most $5,000.Hydraulic hybrid vehicles use hydraulic and mechanicalcomponents to store and deliver energy. In ahydraulic hybrid, the gear-driven transmission isreplaced by a hydraulic pump/motor that is also usedto store and recoup energy through the transferof fluid between hydraulic accumulators. Recenthydraulic hybrid prototypes are designed to providelaunch assist in heavy vehicle applications, allowingacceleration with less engine power. The hydraulichybrid system has been shown to provide a 50-percentimprovement in fuel economy at a cost of about $600.Current hydraulic systems are large and heavy, however,and the EPA is funding R&D to reduce their sizeand weight while improving their efficiency.Oil and Natural Gas SupplyIn the oil and natural gas supply area, new technologiesfor the economical development of unconventionalresources could grow in importance. One of themost plentiful unconventional resources is naturalgas hydrates—ice-like solids composed of light hydrocarbonmolecules, primarily methane, trapped in acage-like crystalline lattice of water and ice.The 1995 National Oil and Gas Resource Assessment,conducted by the USGS and the Minerals ManagementService, produced the first systematic appraisalof in-place natural gas hydrate resources in U.S.onshore and offshore regions [43]. Its mean (expectedvalue) estimate of in-place natural gas hydrates offshorein U.S. deepwater areas was 320,000 trillioncubic feet, and its mean estimate of in-place naturalgas hydrate resources onshore in Alaska’s NorthSlope was 590 trillion cubic feet. In comparison, totalU.S. natural gas production in 2003 was 19 trillioncubic feet, and year-end 2003 reserves were 193 trillioncubic feet. According to these estimates, ifnatural gas hydrate resources could be developed economically,they could supply U.S. natural gas needsfor many years.Commercial production of natural gas hydrates hasnot yet been attempted. Short-term production testsEnergy Information Administration / Annual Energy Outlook 2006 43

Issues in Focushave been conducted in Canada’s MacKenzie Deltaregion, however, and natural gas hydrates may havebeen produced unintentionally at the MessoyakhaField in Russia’s West Siberian Basin.Commercial production of natural gas hydrates isexpected to use one or more of three techniques: pressurereduction, heat injection, and solvent phasechange. The techniques used will depend on the characteristicsof the natural gas hydrate formation beingdeveloped. Each has advantages and disadvantages.The pressure reduction technique has the lowest cost,but it requires a free-gas (non-hydrate) zone belowthe hydrate deposit, and the production rate would belimited by heat transfer rates within the formation.The heat injection technique, using steam or hotwater, does not require a free-gas zone, and it wouldachieve higher production rates than are possiblewith pressure reduction. On the other hand, it is morecomplex and more costly, requiring large amounts ofwater and energy to heat it. The solvent phase changetechnology is the most expensive, and it could lead towater contamination problems, but it does notrequire energy for water heating and is not subject tothe formation of ice dams, which can be a problem forthe heat injection technique.In the United States, the existence of large conventionalnatural gas deposits in the Prudhoe Bay andPoint Thomson Fields on Alaska’s North Slope isexpected to preclude any significant production fromhydrates on the North Slope for many years to come.For example, if the Alaska natural gas pipelinebecame operational in 2015, it would take about 21years (until 2036) to deplete the 35 trillion cubic feetof proven North Slope conventional natural gasresources at a pipeline capacity of 4.5 billion cubic feetper day, or 17 years (until 2032) at a pipeline capacityof 5.6 billion cubic feet per day. Moreover, the NorthSlope has a large undiscovered base of conventionalnatural gas resources beyond the volumes estimatedto be recoverable in currently known fields. Therefore,any significant commercial production of NorthSlope natural gas hydrates could be 30 years or moreinto the future.Production of oceanic natural gas hydrates is atleast as problematic, because the deposits are not aswell mapped and characterized, and because no productionof oceanic hydrates has yet occurred. Moreover,akin to the situation on the Alaska North Slope,there are considerable conventional natural gasdeposits yet to be found and developed in the deepwaterGulf of Mexico. Considerable R&D will also berequired before any exploitation of oceanic naturalgas hydrates can be considered. Research on oceanichydrates is almost certain to continue, given the vastsize of the potential resource.BiorefineriesRising world oil prices in recent years have heightenedinterest in alternative sources of liquid fuels,including biofuels. Currently, two biologically derivedfuels, biodiesel and ethanol, are used in the UnitedStates to augment and improve supplies of gasolineand diesel fuel. As petroleum becomes more scarceand expensive, these and potentially other biofuelscould become important alternatives.Biodiesel. The term biodiesel applies specifically tomethyl or ethyl esters of vegetable oil or animal fat. Inprinciple, biodiesel can be blended into petroleum dieselfuel or heating oil in any fraction, so long as thefuel system that uses it is constructed of materialsthat are compatible with the blend. The actual maximumallowable fraction of biodiesel in diesel fuel variesby engine manufacturer and by specific model line.Fuel system materials are a concern, because methyland ethyl esters are strong solvents that can damagecertain plastics or rubbers.The solvent properties of biodiesel also make itunlikely that biodiesel blends could be shippedthrough petroleum product pipelines. There would bea risk of contamination when the biodiesel dissolvedany material deposited on the walls of pipes, manifolds,or storage tanks. On the positive side, the additionof biodiesel to petroleum diesel reduces engineemissions of carbon monoxide, unburned hydrocarbons,and particulates. On the negative, it tends toincrease nitrogen oxide emissions, and that may limitthe use of biodiesel in places with excess levels ofozone at ground level.The production of methyl esters is an establishedtechnology in the United States, but the product typicallyhas been too expensive to be used as fuel.Instead, methyl esters have been used in productssuch as soaps and detergents. Proctor and Gamble,Peter Cremer, Dow Haltermann, and other largefirms currently supply methyl esters to the industrialmarket. Most dedicated biodiesel producers are muchsmaller, and delivery of a consistent product is provingto be a challenge.Several other processes for making diesel fuel frombiomass are under consideration. The most mature ofthese technologies is biomass-to-liquids (BTL). Thebiomass is first reacted with steam in the presence ofa catalyst to form carbon monoxide and hydrogen, or44 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focussynthesis gas. Any other elements contained in thebiomass are removed during the gasification step.The carbon monoxide and hydrogen are then reactedto form liquid hydrocarbons and water.Although BTL products are high in quality, BTLplants face several challenges. They have high capitaland operating costs, and their feedstock handlingcosts are especially high. BTL gasifiers are significantlymore expensive than the gasifiers used in CTLor GTL facilities. Furthermore, the cost of a BTLplant per barrel of output is several times the cost ofexpanding an existing petroleum refinery or buildinga new one. As a result, while new BTL plants arebeing built in Germany, there is no commercial productionof BTL in the United States. BTL productionand its market implications are discussed under“Nonconventional Liquid Fuels,” below.In another process, vegetable oils and animal fats canbe reacted with hydrogen to yield hydrocarbons thatblend readily into diesel fuel. The oil or fat is pressurizedand combined in a reactor with hydrogen in thepresence of a catalyst similar to those used in hydrotreatersat petroleum refineries. The products of theprocess are bioparaffins. Bioparaffin diesel fuel issimilar in quality to BTL diesel, with the added benefitof being free of byproducts. The improvement inquality over methyl esters (biodiesel) is not free, however.A bioparaffin plant is less expensive than a BTLplant but more expensive than a biodiesel plant,because the bioparaffin reaction takes place underpressure, and a hydrogen plant is needed. Bioparaffinsalso share with biodiesel the problem offeedstock costs. Vegetable oils are expensive, especiallyif they are food grade. The catalyst needed alsoadds significant expense. The world’s first bioparaffinplant is being built at a petroleum refinery in Finland,but there are no plans for U.S. bioparaffin capacity atthis time.Ethanol. Ethanol can be blended into gasoline readilyat up to 10 percent by volume. All cars and lighttrucks built for the U.S. market since the late 1970scan run on gasoline containing 10 percent ethanol.Automakers also produce a limited number of vehiclesfor the U.S. market that can run on blends of upto 85 percent ethanol. Ethanol adds oxygen to the gasoline,which reduces carbon monoxide emissions fromvehicles with less sophisticated emissions controls. Italso dilutes sulfur and aromatic contents andimproves octane. Because newer vehicles with moresophisticated emissions controls show little or nochange in emissions with the addition of oxygen togasoline, ethanol blending in the future will dependlargely on octane requirements, limits on gasolinesulfur and aromatics levels, and mandates for the useof renewable motor fuels.Ethanol production from starches and sugars, such ascorn, is a well-known technology that continues toevolve. In the United States, most fuel ethanol currentlyis distilled from corn, yielding byproducts thatare used as supplements in animal feed. Three factorsmay limit ethanol production from starchy and sugarycrops: all such crops are also used for food, andonly a limited fraction of the available supply could bediverted for fuel use without driving up crop prices tothe point where ethanol production would no longerbe economical; there is a limit to the amount of suitableland available for growing the feedstock crops;and only a portion of the plant material from thefeedstock can be used to produce ethanol. For example,corn grain can be used in ethanol plants, but thestalks, husks, and leaves are waste material, onlysome of which needs to be left on cornfields to preventerosion and replenish soil nutrients.The underutilization of crop residue has drivendecades of research into ethanol production from cellulose;however, several obstacles continue to preventcommercialization of the process, including how toaccelerate the hydrolysis reaction that breaks downcellulose fibers and what to do with the lignin byproduct.Research on acid hydrolysis and enzymatichydrolysis is ongoing. The favored proposal for dealingwith the lignin is to use it as a fuel for CHP plants,which could provide both thermal energy and electricityfor cellulose ethanol plants, as well as electricityfor the grid; however, CHP plants are expensive.Currently, Canada’s Iogen Corporation is trying tocommercialize an enzymatic hydrolysis technologyfor ethanol production. The company estimates that aplant with ethanol capacity of 50 million gallons peryear and lignin-fired CHP will cost about $300 millionto build. By comparison, a corn ethanol plant with acapacity of 50 million gallons per year could be builtfor about $65 million, and the owners would not bearthe risk associated with a new technology. Co-locationof cellulose ethanol plants with existing coal-firedelectric power plants could reduce the capital cost ofthe ethanol plants but would also limit sitingpossibilities.Electricity ProductionSome of the electricity generating technologies andfuels represented in NEMS are currently uneconomical,and there are still other fossil, renewable, andnuclear options under development that are notEnergy Information Administration / Annual Energy Outlook 2006 45

Issues in Focusexplicitly represented. Those technologies are notexpected to be important throughout most of the projections,but with successful development they couldhave impacts in the market in the later years.Fossil FuelsAdvanced Coal Power. FutureGen is a demonstrationproject announced by DOE in February 2003 that willhave 275 megawatts of electricity generation capacityand will also produce hydrogen for other uses. Of theproject’s $1 billion cost, 80 percent will come fromDOE, and 20 percent is expected to be providedthrough a consortium of firms from the coal and electricpower industries. The demonstration plant,fueled by coal, will include carbon capture andsequestration equipment to limit GHG emissions. Itwill operate in an IGCC configuration and sequesterapproximately 1 million metric tons of CO 2 annually.The sequestered CO 2 will be used to enhance oilrecovery in depleted oil fields. SO 2 and mercury emissionsfrom the plant will also be captured.In 2003, it was anticipated that the FutureGen projectwould be operational within 10 years. Site selectionand environmental impact studies are expectedto be completed in 2007. The site must include geologicalformations that can be used to store at least 90percent of the plant’s CO 2 emissions, with an annualleakage rate below 0.01 percent.If the project proves to be technically and economicallysuccessful, it could offer a partial solution for thecontinued use of fossil fuels without contributing furtherto rising atmospheric concentrations of GHGs,by injecting CO 2 into depleted oil and gas wells whileadequate space is available. Coal gasification plantswith carbon capture and sequestration equipmenthave yet to be demonstrated, however, and manychallenges remain. The capital costs for IGCC plantswith carbon capture and sequestration equipment aremuch higher than those for conventional coal-firedplants, and their conversion efficiencies are lower.Moreover, the current conventional solvent-basedabsorption process for carbon capture remains energyintensive.Advanced Fuel Cells. Fuel cells operate similarly tobatteries but do not lose their charge. Instead, theyrely on a supply of hydrogen, which is broken into freeprotons and electrons within the cell. There are severaltypes of fuel cells, using different materials andoperating at different temperatures. Stationarypower fuel cells can be connected to the electricitygrid, and smaller cells are envisioned for the transportationsector. Although the costs of fuel cells havebeen reduced since their inception, they currentlyremain too high for widespread market penetration.Phosphoric acid fuel cells, which operate at relativelylow temperatures, are currently being used in severalapplications with efficiency rates of 37 to 42 percent.An advantage of this cell type is that relatively impurehydrogen is tolerated, broadening the source ofpotential fuels. The major disadvantage is the highcost of the platinum catalyst.Molten carbonate fuel cells, which use nickel in placeof more costly metals, can achieve a 50-percent efficiencyrate and are operating experimentally aspower plants. Solid oxide fuel cells, also currentlybeing developed, use ceramic materials, operate atrelatively high temperatures, and can achieve similarefficiencies of around 50 percent. They have applicationsin the electric power sector, providing exhaustto turn gas turbines, and could also have future usesin the transportation sector.The costs of fuel cells must be reduced significantlybefore they can become competitive in U.S. markets,and an inexpensive, plentiful source of hydrogen fuelmust also be found. If those hurdles can be met, fuelcells offer several advantages over current generationtechnologies: they are small, quiet, and clean, andbecause no combustion is involved, their only byproductis water.Carbon Capture with SequestrationCapturing CO 2 from the combustion of fossil fuelsmay allow for their continued use without significantadditional contributions to GHG emissions andglobal warming. Currently, however, sequestrationtechnologies are too costly for implementation on asignificant scale. One of the greatest challenges is separationof CO 2 from other emissions, given typicalCO 2 concentrations of 3 to 12 percent in the smokestackgases of coal-fired power plants.One potential solution for capturing CO 2 is the use ofamine scrubbers. Amines react with CO 2 , and theresulting product can be heated and separated in adesorber. Another option is the IGCC process to beused in FutureGen, which will produce highly concentratedCO 2 ready for storage.Carbon storage will most likely be underground. Forexample, enhanced oil recovery technologies pumpCO 2 into depleted oil and natural gas fields to extendtheir yields and lifetimes. Other options includeplacing the CO 2 in coalbeds and saline formations.Ocean storage is a possibility, although the potential46 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focusenvironmental impacts are unknown. Preliminarygeological studies have shown that underground storage,if successful, has the potential to store all the CO 2from industrial and power sector emissions for severaldecades. Major issues include the proximity ofthe geologic storage formations to potential CO 2 productionsites, the long-term permanence of the storagesites, and the development of the monitoringsystems needed to ensure that leakage is limited andcontrolled.In 2005, DOE announced the second phase of sevenpartnerships involving small, field-level demonstrationsto determine the feasibility of carbon sequestrationtechnologies. In one project, ConocoPhillips,Shell, and Scottish and Southern Energy will begindesigning the world’s first industrial-scale facility togenerate “carbon-free electricity” from hydrogen.The planned project will convert natural gas to hydrogenand CO 2 , then use the hydrogen gas as fuel for a350-megawatt power station, reducing the amount ofCO 2 emitted to the atmosphere by 90 percent. TheCO 2 will be exported to a North Sea oil reservoir forincreased oil recovery and eventual storage. Smallerdemonstration projects are already operating in Algeriaand Norway.RenewablesIn the face of international concern over GHG emissions,the eventual peaking of world oil production,and recent volatility in fossil fuel prices, many haveseen promise in exploiting an ever-increasing rangeof renewable energy resources. Renewable energyresources used to generate electricity generally reducenet GHG emissions compared to fossil generation,are accepted as being nondepletable on a timescale of interest to society, and tend to have low andstable operating costs.To date, however, market adoption of most renewabletechnologies has been limited by the significant capitalexpense of capturing and concentrating theoften diffuse energy fluxes of wind, solar, ocean,and other renewable resources. With the most successfulrenewable generation technology, hydropower,nature has largely concentrated the diffuseenergy of falling water through the geography ofwatersheds. The challenge for emerging technologies,as well as those on the horizon, will be to minimizeboth the monetary and environmental costs of collectingand converting renewable energy fuels to moreportable and useful forms.Wind. Through a combination of significant costreductions over the past 20 years and policy supportin the United States, Europe, and elsewhere, electricitygeneration from wind energy has increasedsubstantially over the past 5 to 10 years. In fact, insome areas of Western Europe, viable new sites forwind are seen as severely limited, because the bestsites already are being exploited, leaving sites withpoor resources, too close to populated areas, and/or inotherwise undesirable locations. In response, a numberof European countries have begun to build windplants offshore, where they are more remote frompopulation centers and can take advantage of betterresources. Although firm data on costs has beenscarce, it is believed that offshore wind plants costsubstantially more to construct, to transmit power,and to maintain than comparable onshore windplants.There have been a number of proposals for offshorewind plants in the United States, including at leasttwo under serious consideration for near-term development,off Cape Cod, Massachusetts, and LongIsland, New York. The United States has substantiallylarger and better wind resources than mostcountries of Europe, and thus is unlikely to see itsonshore resources exhausted in the mid-term outlook.Still, localized factors such as State renewableenergy requirements and constraints on electricitytransmission from conventional power plants intocoastal areas may make some offshore resources economicallyattractive, despite the abundance of lowercost wind resources further inland. Because NEMSmodels 13 relatively large electricity markets, it cannotfully account for localized effects at the State ormetropolitan level, and thus is likely to miss the feweconomical opportunities for offshore development ofwind-powered generators.Hydropower. In addition to ocean-based wind powertechnologies, there are a number of technologies thatcould harness energy directly from ocean waters.They include wave energy technologies (which indirectlyharness wind energy, in that ocean waves usuallyare driven by surface winds), tidal energytechnologies, “in-stream” hydropower, and oceanthermal energy technologies.Although a number of wave energy technologies areunder development, including some that may be nearpre-commercial demonstration, the publicly availabledata on resource quantity, quality, and distributionand on technology cost and performance are inadequateto describe the specifics of the technologies. Ahandful of tidal power stations around the world dooperate on a commercial basis, but prime tidalresources are limited, and the technology seemsEnergy Information Administration / Annual Energy Outlook 2006 47

Issues in Focusunlikely to achieve substantial market penetrationunless more marginal resources can be harnessedeconomically.In-stream hydropower technologies generally usefreestanding or tethered hydraulic turbines to capturethe kinetic energy of river, ocean, or tidal currentswithout dams or diversions. As with waveenergy technologies, while some of these technologiesappear to be in fairly advanced pre-commercial development,there is insufficient available information tosupport reasonable market assessment within theNEMS framework.Ocean thermal technologies harness energy fromtemperature differentials between surface waters andwaters at depth. These technologies have receivedfunding from the Federal Government in the past,and U.S. development continues today under fullyprivate funding. To date, however, there have been nonew pre-commercial demonstrations beyond thosepreviously funded by the Federal Government.Resources suitable for ocean thermal energy developmentare geographically limited to tropical or neartropicalwaters near land, with a relatively steep continentalshelf. (Although a fully offshore deepwatertechnology is plausible, it would be significantly moreexpensive than a shore-based implementation.)These requirements eliminate virtually the entirecontinental United States as a potential resourcebase, and the technology is not included in AEO2006.Geothermal. Although U.S. geothermal resourceshave been exploited for decades to produce electricity,commercial development to date has been limited tohydrothermal deposits at relatively shallow depths.In hydrothermal deposits, hot rock close to the surfaceheats naturally occurring groundwater, which isextracted at relatively low cost to drive a conventionalgenerator. Steam may be used directly from theground, or superheated water may be used to heat asecondary working fluid that drives the turbine. Suitablehydrothermal deposits, however, are limited inquantity and location, and in most cases they wouldbe too expensive for development in the mid-term.Enhanced geothermal technologies to exploit deeper,drier resources are not likely to be cost-effective forwidespread commercial deployment until well after2030.Solar. Sunlight is a renewable resource that is almostuniversally available. NEMS models several differenttechnologies for harnessing solar energy, includingPV cells deployed at end-user locations, PV deployedat central, utility-owned locations, and thermalconversion of sunlight to electricity. Each is basedon commercially available technologies, with substantialallowances made for future improvements incost and performance. In view of the significant contributionof government-funded R&D to the progressof solar energy technologies, much of the futureimprovements occur independently from actual marketgrowth (although significant market growth isprojected).Research is continuing on a number of solar technologies—bothdirect conversion and thermal conversion—thatcould substantially improve the efficiencyor reduce the cost of producing electricity fromsunlight. Examples include organic PV, highlyconcentrated PV, “solar chimneys,” and a range ofimprovements to PV efficiency and manufacturing.Given the wide variety of potential technologiesand uncertainty as to the success of any particularone, solar technology is modeled from the knowncost and performance parameters of commercialtechnologies, along with both production-based andproduction-independent improvements in cost andperformance.HydrogenWidespread use of hydrogen as an energy carrier hasbeen presented by some as a long-term solution to thelimitations of our largely fossil-energy based economy.Significant quantities of molecular hydrogen(H 2 ) are not found in nature but must be releasedfrom water, hydrocarbons, or other “chemical reservoirs”of hydrogen. Thus, hydrogen is an energy carrier,in much the same way that electricity is anenergy carrier, rather than a primary source ofenergy. Hydrogen has a wide variety of potential enduses, including the production of electricity; buthydrogen production based on fossil fuels (primarilythrough methane steam reforming or other thermochemicalprocesses), currently the least costly meansof production, would at best provide only limitedrelief from the use of fossil fuels (by increasing theefficiency of energy end uses) and potentially couldlead to more use of fossil fuels (by reducing overall“wells-to-wheels” system efficiency).Hydrogen could also be produced from non-fossilfuels, including nuclear and renewable resources,either through electrolysis of water or by directthermochemical conversion. Significant use of hydrogenwould likely evolve as a system, with developmentand deployment of technologies for production, distribution,and end use closely linked. Many technologiesfor producing hydrogen are commerciallyavailable today, but they are expensive. Without significanttechnological progress, it seems unlikely that48 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focussubstantial incremental amounts of hydrogen will beproduced before 2030.NuclearThe nuclear cost assumptions for AEO2006 are basedon the realized costs of advanced nuclear powerplants whose designs have been certified by the U.S.Nuclear Regulatory Commission (NRC) and/or havebeen built somewhere in the world—specifically, thegeneration 3 light-water reactors (LWRs). To accountfor technological improvements, it is assumed thatcosts will fall, with cost reductions reflecting incrementalimprovements in the designs of reactors asthey evolve from the generation 3 to generation 3+.Recently, some vendors have reported cost estimatesfor generation 3+ reactors that are much lower thanthose assumed in NEMS, even after allowing for costreductions; however, their estimates were based onincomplete designs, and history has shown that costestimates based on incomplete designs tend to beunreliable [44]. For AEO2006, the vendor estimatesare used in a sensitivity analysis.Although the nuclear capital cost assumptions usedin both the reference case and the sensitivity analysisare representative of the costs of building LWRswhose designs reflect incremental improvementsover those that have been built in the Far East or arebeing built in Europe, a number of small-scale andlarge-scale LWR designs that differ significantly fromgeneration 3 plants could be commercially availableby 2030 [45]. Because of technical and economicuncertainties, however, they are not included inAEO2006.A number of non-LWR designs for nuclear powerplants have also been suggested, including variantson the traditional fast breeder technology, such aslead-cooled and sodium-cooled reactors. Thesedesigns are often referred to as “generation 4”nuclear power plants. The technologies have all theadvantages and disadvantages of the traditionalbreeder reactors that have been built in Europe andthe Far East, and because of their large size theywould be more economically advantageous in regulatedelectricity markets, where financial risks arenot borne entirely by investors.Examples of the small, modular power plant designsinclude the Pebble Bed Modular Reactor (PBMR), theGas-Turbine Modular Helium (GT-MH) reactor andthe International Reactor Innovative and Secure(IRIS) reactor. In theory at least, these plants mightbe built in competitive markets where it is economicallyadvantageous to add small amounts of capacityin response to volatile and uncertain electricity prices[46].The PBMR and the GT-MH reactor are also designedto operate at much higher temperatures than theLWRs currently in operation. Thus, both of thesedesigns could potentially be used to produce both electricityand hydrogen. In fact, EPACT2005 authorizes$1.25 billion to build a prototype of such a reactorthat could be used to cogenerate electricity and hydrogen.The law specifies that a prototype reactor shouldbe completed by 2021. The economic potential of sucha reactor is considerable, in that the hydrogen couldbe used in fuel cells or in other industrial processes;however, the technological uncertainties involved aresubstantial.Advanced Technologies for Light-DutyVehiclesA fundamental concern in projecting the futureattributes of light-duty vehicles—passenger cars,sport utility vehicles, pickup trucks, and minivans—is how to represent technological change and the marketforces that drive it. There is always considerableuncertainty about the evolution of existing technologies,what new technologies might emerge, and howconsumer preferences might influence the directionof change. Most of the new and emerging technologiesexpected to affect the performance and fuel use oflight-duty vehicles over the next 25 years are representedin NEMS; however, the potential emergence ofnew, unforeseen technologies makes it impossible toaddress all the technology options that could comeinto play. The previous section of “Issues in Focus”discussed several potential technologies that currentlyare not represented in NEMS. This section discussessome of the key technologies represented inNEMS that are expected to be implemented inlight-duty vehicles over the next 25 years.The NEMS Transportation Module represents technologiesfor light-duty vehicles that allow them tocomply with current standards for safety, emissions,and fuel economy or may improve their efficiencyand/or performance, based on expected consumerdemand for those attributes. Technologies that canimprove vehicle efficiency take two forms: those thatrepresent incremental improvements to or advancementsin the various components of conventionalpower trains, and those that represent significantchanges in power train design. Advanced technologiesused in vehicles with new power train designsinclude, primarily, electric power propulsion systemsin hybrid, fuel cell, and battery-powered vehicles.Energy Information Administration / Annual Energy Outlook 2006 49

Issues in FocusHistorically, the development of new technologies forlight-duty vehicles has been driven by the challenge ofmeeting increased demand for larger, quieter, morepowerful vehicles while complying with emissions,safety, and fuel economy standards. The auto industryhas met those challenges and, through technologicalinnovation, delivered larger, more powerfulvehicles with improved fuel economy.In 1980, the average new car weighed 3,101 pounds,had 100 horsepower, and averaged 24.3 miles per gallon.In 2004, the average new car weighed 3,454pounds (an 11-percent increase), had 181 horsepower(an 81-percent increase), and averaged 29.3 miles pergallon (a 21-percent increase). Improvements in newlight trucks (including sport utility vehicles) from1980 to 2004 have been even more profound: theiraverage weight has increased by 20 percent to 4,649pounds, their horsepower has increased by 91 percentto 231, and their average fuel economy has increasedby 16 percent to 21.5 miles per gallon [47].The majority of improvements in horsepower and fueleconomy for new light-duty vehicles have resultedfrom changes in conventional vehicle components,including fuel delivery systems, valve train design,aerodynamics, and transmissions. In 1980, almost allnew light-duty vehicles employed carburetors for fueldelivery; in 2004, all new light-duty vehicles used portfuel injection systems, which improve engine efficiencythrough very precise electronic control of fueldelivery. Advances have also been made in valve traindesign, improving efficiency by reducing enginepumping losses. In 1980, all engine designs used twovalves per cylinder; in 2004, engines with four valvesper cylinder were installed in 74 percent of new carsand 43 percent of new light trucks.Increases in light-duty vehicle horsepower and fueleconomy are projected to continue in the AEO2006cases at rates similar to their historical rates, whilevehicle weight remains relatively constant. For example,between 2005 and 2030 new car horsepowerincreases by 19 percent, to 215, in the reference case,while fuel economy increases by 15 percent to 33.8miles per gallon; and the horsepower of new lighttrucks increases by 14 percent, to 264, and fuel economyincreases by 23 percent to 26.4 miles per gallon,while their weight increases by 4 percent to 4,828pounds. Most of the improvements result from innovationsin conventional vehicle components.To project potential improvement in new light-dutyvehicle fuel economy, 63 conventional technologiesare represented in the Transportation Module. Thetechnologies are grouped into six vehicle system categories:engine, transmission, accessory load, body,drive train, and independent (related to safety andemissions). Table 13 summarizes the technologiesexpected to have significant impacts over the projectionperiod, the expected range of efficiency improvements,and initial costs.Engineering relationships among the technologiesare also modeled in the Transportation Module.The engineering relationships account for: (1)co-relationships, where the existence of one technologyis required for the existence of another; (2) synergisticeffects, reflecting the combined efficiencyimpact of two or more technologies; (3) supersedingrelationships, which remove replaced technologies;and (4) mandatory technologies, needed to meetsafety and emissions regulations. In addition to theengineering relationships, reductions in technologycost are captured as unit production increases orcumulative production reaches a design cyclethreshold.Technologies expected to show the greatest increasein market penetration, and thus the greatest impacton new car and light truck efficiency, include lightweightmaterials, improved aerodynamics, enginefriction reduction, improved pumps, and low rollingresistance tires (Figures 17 and 18). These technologiesrepresent the most cost-effective options forimproving fuel economy while meeting consumerexpectations for vehicle performance and comfort.The weight of new cars remains relatively constant asa result of increased market penetration of highstrengthlow-alloy steel (63 percent by 2030), aluminumcastings (24 percent by 2030), and aluminumbodies and closures (12 percent by 2030). Variablevalve timing and lift and camless valve actuation arealso expected to have a significant impact on new carefficiency, with installations increasing to approximately30 percent and 4 percent, respectively, in2030. The use of unit body construction in new lighttrucks increases from 23 percent in 2004 to 36 percentin 2030 as more sport utility vehicles and pickuptrucks are developed from car-based platforms.The efficiency of new light-duty vehicles also improveswith increased market penetration of hybridand diesel vehicles. Depending on the make andmodel, the incremental cost of a power-assistedhybrid vehicle (a “full hybrid”), currently estimatedat $3,000 to $10,000, decreases to between $1,500 and$5,400 in 2030 [48]. As a result, the penetration ofhybrid vehicles increases from 0.5 percent of newlight-duty vehicle sales in 2004 to 9.0 percent in 2030.50 Energy Information Administration / Annual Energy Outlook 2006

Issues in FocusMarket penetration of diesel vehicles increases fromabout 2 percent in 2004 to more than 8 percent in2030. Battery and fuel cell powered vehicles also penetratethe light-duty vehicle market as a result of legislativemandates, but with very high vehicle costs,limited driving range, and the lack of a refuelinginfrastructure, they account for only 0.1 percent ofnew vehicle sales in 2030.Nonconventional Liquid FuelsHigher prices for crude oil and refined petroleumproducts are opening the door for nonconventionalliquids to displace petroleum in the traditional fuelsupply mix. Growing world demand for diesel fuel ishelping to jump-start the trend toward increasingproduction of nonconventional liquids, and technologicaladvances are making the nonconventionalTable 13. Technologies expected to have significant impacts on new light-duty vehiclesVehicle componentand technologyEngineAdvanced valve trainFriction reductionCylinder deactivationTechnology descriptionFour valves per cylinder; variable valve timing and lift; camlessvalve actuationLow-mass pistons and valves; reduced piston ring and valvespring tension; improved surface coatings and tolerancesReduced cylinder operation at light load, lowering displacementand reducing pumping lossesExpected efficiencyimprovement(percent)Initial incrementalcost (2000 dollars)2.5-8.0 45-7502.0-6.5 25-1774.5 250Lean burn Direct injection fuel system, enabling very lean air-fuel ratios 5.0 250TransmissionControl systemTransmissionElectronic controls, improving efficiency through shift logic andtorque converter lockup5-speed and 6-speed automatics; continuously variabletransmissions0.5-2.0 8-606.5-10.0 435-615Accessory loadImproved pumps Reduced engine load from oil, water, and power steering pumps 0.3-0.5 10-15Electric pumps Electrically powered pumps, replacing mechanical pumps 1.0-2.0 50-150BodyImproved materialsHigh-strength alloy steel; aluminum castings; lightweightinteriors; aluminum body and closures3.3-13.2 0.4-1.2 dollars per poundof vehicle weight reductionUnit body construction Elimination of body-on-chassis structure 4.0 100Improved aerodynamics Reduction in drag coefficient, with improvements specific to2.3-8.0 40-225body typeDrive trainAdvanced tires Reduced rolling resistance 2.0-6.0 30-135Improved 4-wheel drive Reduced weight; improved electronic controls 2.0 100IndependentSafety and emissions Improved safety and emission systems -3.0 200Figure 17. Market penetration of advancedtechnologies in new cars, 2004 and 2030(percent of total new cars sold)Advanced transmissionsLightweight materialsImproved aerodynamicsUnit body construction4-wheel drive improvementsEngine friction reductionImproved pumpsLow rolling resistance tires200420300 20 40 60 80 100Figure 18. Market penetration of advancedtechnologies in new light trucks, 2004 and 2030(percent of total new light trucks sold)Advanced transmissionsLightweight materialsImproved aerodynamicsUnit body construction4-wheel drive improvementsEngine friction reductionImproved pumpsLow rolling resistance tires200420300 20 40 60 80 100Energy Information Administration / Annual Energy Outlook 2006 51

Issues in Focusalternatives more viable commercially. Those trendsare reflected in the AEO2006 projections.In the reference case, based on projections for theUnited States and project announcements coveringother world regions through 2030, the supply ofsyncrude, synthetic fuels, and liquids produced fromrenewable fuels approaches 10 million barrels perday worldwide in 2030. In the high price case, nonconventionalliquids represent 16 percent of totalworld oil supply in 2030, at more than 16.4 millionbarrels per day. The U.S. share of world nonconventionalliquids production in 2030 is 15 percentin the reference case and nearly 20 percent in the highprice case (Table 14).The term “nonconventional liquids” applies to threedifferent product types: syncrude derived from thebitumen in oil sands, from extra-heavy oil, or from oilshales; synthetic fuels created from coal, natural gas,or biomass feedstocks; and renewable fuels—primarily,ethanol and biodiesel—produced from a variety ofrenewable feedstocks. Generally, these resources areeconomically competitive only when oil prices reachrelatively high levels.Synthetic Crude OilsAt present, two nonconventional oil resources—bitumens(oil sands) and extra-heavy crude oils—areactively being developed and produced. With technologyinnovations ongoing and production costs decliningsteadily, their production increases in theAEO2006 projections, provided that the world oilprice remains above $30 per barrel. Development of athird nonconventional resource, shale oil, is morespeculative. The greatest risks facing syncrude productionare higher production costs and lower crudeoil prices. In AEO2006, production of syncrude worldwideincreases to 5.3 million barrels per day in the referencecase and 8.5 million barrels per day in the highprice case in 2030.Oil sands. Bitumen, the “oil” in oil sands, is composedof carbon-rich, hydrogen-poor long-chain molecules.Its API gravity is less than 10, and its viscosity is sohigh that it does not flow in a reservoir. It can containundesirable quantities of nitrogen, sulfur, and heavymetals.The percentage of bitumen in oil sands depositsranges from 1 to 20 percent [49]. After the bitumen isextracted from the sand matrix, various processes,including coking, distillation, catalytic conversion,and hydrotreating, must be applied to createsyncrude. On average, about 1.16 barrels of bitumenis required to produce 1 barrel of syncrude. Canada’sresource of 2.5 trillion barrels of in-place bitumen isestimated to be 81 percent of the world total [50]. Economicallyrecoverable deposits in Canada amount toabout 315 billion barrels of bitumen under currenteconomic and technological conditions [51], and in2004 Canada shipped more than 87 million barrels oflight, sweet syncrude [52]. If fully developed, the bitumenresources in Canada could supply more than 40years of U.S. oil consumption at current demandlevels.Currently, there are two methods for extracting bitumenfrom oil sands: open-pit mining and in situ recovery.For deposits near the surface, open-pit mining isused to extract the bitumen by physically separatingit from the sand and clay matrix, at recovery ratesapproaching 95 percent. For deposits deeper than 225feet, the in situ process is used. Two wells are drilled,one of which is used to inject steam into the deposit toheat the sand and lower the viscosity of the bitumenand the other to collect the flowing bitumen and bringit to the surface. Addition of gas condensate, lightcrude, or natural gas can also reduce viscosity andallow the bitumen to flow. Much of today’s productioncomes from open-pit mining operations; however, 80percent of the Canadian oil sands reserves are toodeep for open-pit mining.Table 14. Nonconventional liquid fuels production in the AEO2006 reference and high price cases, 2030(million barrels per day)Synthetic crude oils Synthetic fuels Renewable fuelsTotal production Oil sands Extra-heavy oil Shale oil CTL GTL BTL Biodiesel Ethanol TotalReference caseUnited States — — — 0.8 — — 0.02 0.7 1.5World 2.9 2.3 0.05 1.8 1.1 — — 1.7 a 9.9High price caseUnited States — — 0.4 1.7 0.2 — 0.03 0.9 3.2World 4.9 3.1 0.5 2.3 2.6 — — 3.0 a 16.4a Includes biodiesel.52 Energy Information Administration / Annual Energy Outlook 2006

Issues in FocusAccording to most analysts, oil sands syncrude productionis economically viable, covering fixed andvariable costs, only when syncrude prices exceed $30per barrel. The variable costs of producing syncrudehave declined to around $5 per barrel today, fromestimates of $10 per barrel in the late 1990s and $22per barrel in the 1980s.Syncrude tends to yield poor quality distillate andgas-oil products owing to its low hydrogen content.Refineries processing oil sands syncrude need moresophisticated conversion capacity including catalyticcracking, hydrocracking, and coking to create higherquality fuels suitable for transportation markets.Extra-heavy oil. Extra-heavy oil is crude oil with APIgravity less than 10 and viscosity greater than 10,000centipoise. Unlike bitumen, extra-heavy oil will flowin reservoirs, albeit much more slowly than ordinarycrude oils. Extra-heavy oil deposits are located in atleast 30 countries. One singularly large deposit, representingthe majority of the known extra-heavy oilresource is located in the Orinoco oil belt of easternVenezuela. Petroleos de Venezuela SA (PDVSA) estimatesthat 1.36 trillion barrels of extra-heavy oil arein place in the Orinoco belt, with an estimated 270 billionbarrels of currently recoverable reserves.There are three main recovery methods: cyclicsteam injection/steam flood; diluents and gas lift;and steam-assisted gravity drainage (SAGD) usingstacked horizontal wells. Other methods substituteCO 2 for natural gas injection or solvents for steaminjection. The Orinoco projects currently use a twostepupgrading process, partially upgrading the bitumenin the field, followed by deep conversion refiningin the importing country.Extra-heavy oil recovery rates currently range from 5to 10 percent of oil in place, although R&D efforts aresteadily and significantly improving the performance.Lifting and processing costs range from $8 to $11 perbarrel (2004 dollars) [53]. According to the latestPDVSA filings with the U.S. Securities and ExchangeCommission, production of extra-heavy crude oilfrom the Orinoco area totaled 430,000 barrels per dayin 2003 [54].It is not clear that PDVSA can continue to provide themassive capital investment necessary to sustain thegrowth of its extra-heavy oil production in the future.Relationships with possible foreign investors havebeen strained due to actions by the Venezuelan governmentto renegotiate existing contracts and tostructure new ones so as to sharply reduce potentialreturns to investors. In addition, the recent deteriorationof political relations between Venezuela and theUnited States could limit the market for Orinocoproducedextra-heavy crude oils.Shale oil. The term “oil shale” is something of a misnomer.First, the rock involved is not a shale; it is acalcareous mudstone known as marlstone. Second,the marlstone does not contain crude oil but insteadcontains an organic material, kerogen, that is a primitiveprecursor of crude oil. When oil shale is heated atmoderate to high temperatures for a sufficient periodof time, kerogen can be cracked to smaller organicmolecules like those typically found in crude oils andthen converted to a vapor phase that can be separatedby boiling point and processed into a variety of liquidfuels in a distillation process. The synthetic liquid distilledfrom oil shale is commonly known as shale oil.Oil shale has also been burned directly as a solid fuel,like coal, for electricity generation.The global resource of oil shale base is huge—estimatedat a minimum of 2.9 trillion barrels of recoverableoil [55], including 750 billion barrels in theUnited States, mostly in Utah, Wyoming, and Colorado[56]. Deposits that yield greater than 25 gallonsper ton are the most likely to be economically viable[57]. Based on an estimated yield of 25 gallons ofsyncrude from 1 ton of oil shale, the U.S. resource, iffully developed, could supply more than 100 years ofU.S. oil consumption at current demand levels.There are two principal methods for oil shale extraction:underground mining and in situ recovery.Underground mining, followed by surface retorting,is the primary approach used by petroleum companiesin demonstration plants built in the mid to late1970s. In this approach, oil shale is mined from theground and then transferred to a processing facility,where the kerogen is heated in a retort (a large, cylindricalfurnace) to around 900 degrees Fahrenheit andenriched with hydrogen to release hydrocarbonvapors that are then condensed to a liquid. There issome risk that, despite its apparent promise, theunderground mining/surface retorting technologyultimately will not be viable, because of its potentiallyadverse environmental impacts associated with wasterock disposal and the large volumes of water requiredfor remediation of waste disposal piles.A comprehensive in situ process is currently underexperimental development by Shell Oil [58]. Shalerock is heated to 650-750 degrees Fahrenheit, causingwater in the shale to turn into steam that “microfractures”the formation. The in situ process generatesa greater yield from a smaller land surface area ata lower cost than open-pit mining. The technologyEnergy Information Administration / Annual Energy Outlook 2006 53

Issues in Focusalso avoids several adverse issues connected to miningand waste rock remediation, minimizes waterusage, and has the potential to recover at least 10times more oil per acre than the conventional surfacemining and retorting process; however, it could takeas long as 15 years to demonstrate the commercialviability of the Shell in situ process.For a conventional mining and retorting process, $55to $70 per barrel (2004 dollars) is the estimatedbreakeven price. That estimate is based in part ontechnical literature from the late 1970s and early1980s, however, and thus may no longer be relevanttoday. The older estimates are likely to understatethe cost of waste rock remediation. Advances inequipment technology over the years could increaseoperating efficiencies and reduce costs. A 1 millionbarrel per day shale oil industry based on undergroundmining/surface retorting would require miningand remediation of more than 500 million tons ofoil shale rock per year—about one-half of the annualtonnage of domestic coal production. The processwould also consume approximately 3 million barrelsof water per day [59].A 2005 industry study prepared for the NationalEnergy Technology Laboratory estimates that crudeoil prices (WTI basis) would need to be in the range of$70 to $95 per barrel for a first-of-kind shale oil operationto be profitable [60] but could drop to between$35 and $48 per barrel within a dozen years as aresult of experience-based learning (“learning-bydoing”).In the AEO2006 high price case, assumingthe use of underground mining with surface retorting,U.S. oil shale production begins in 2019 andgrows to 410,000 barrels per day in 2030.which in turn is hydrocracked to producehydrocarbons of various chain lengths. End productsare determined by catalyst selectivity and reactionconditions, and product yields are adjustable withinranges, depending on reaction severity and catalystselection. Potential products include naphtha, kerosene,diesel, methanol, dimethyl ether, alcohols, wax,and lube oil stock. A product workup section separatesthe liquids and completes the transformationinto final products. The diesel fuel produced(“Fischer-Tropsch diesel”) is limited by a lack of naturallubricity, which can be remedied by additives [61].Water and CO 2 are typically produced as byproductsof the process.Coal-to-Liquids. A CTL plant transforms coal into liquidfuels. CTL is economically competitive at an oilprice in the low to mid-$40 per barrel range and a coalcost in the range of $1 to $2 per million Btu, dependingon coal quality and location.A CTL plant requires several decades of coal reservesto justify construction. Given the economies of scalerequired, 30,000 barrels per day is regarded as a minimumplant size. Coal reserves of approximately 2 to 4billion tons are required to support a commercial CTLplant with a capacity of 70,000 to 80,000 barrels perday over its useful life [62]. Capital expenses are estimatedto be in the range of $50,000 to $70,000 (2004dollars) per barrel of daily capacity. The front-end(coal handling) portion of a CTL plant accounts forabout one-half of the capital cost [63].Figure 19. System elements for production ofsynthetic fuels from coal, natural gas, and biomassCoalSynthetic FuelsNatural gasSynfuels can be produced from coal, natural gas, orbiomass feedstocks through chemical conversion intosyncrude and/or synthetic liquid products. Hugeindustrial facilities gasify the feedstocks to producesynthesis gas (carbon monoxide and hydrogen) as aninitial step. Synfuel plants commonly employ theFischer-Tropsch process, with front-end processingfacilities that vary, depending on the feedstock. Themanufacturing process for the synthetic fuels typicallybypasses the traditional oil refining system, creatingfuels that can go directly to final markets. Asimplified flow diagram of the synthetic fuels processis shown in Figure 19.In the basic Fischer-Tropsch reaction, syngas is fed toa reactor where it is converted to a paraffin wax,Front end:feedstockhandlingand syngasproductionSteam/heatElectricityWaterHydrogen/carbon dioxideBiomassSyngasFischer-TropschsynthesisByproductsandco-productioncapabilitySyncrudeProductseparationandworkupsectionNaphthaDieselKeroseneLubricants/waxes54 Energy Information Administration / Annual Energy Outlook 2006

Issues in FocusThere are two leading technologies for convertingcoal into transportation fuels and liquids. The originalprocess, indirect coal liquefaction (ICL), gasifiescoal to produce a syngas and rebuilds small moleculesin the Fischer-Tropsch process to produce the desiredfuels. Direct coal liquefaction (DCL) breaks the coaldown to maximize the proportion of compounds withthe correct molecular size for liquid products. Theprocess reacts coal molecules with hydrogen underhigh temperatures and pressures to produce asyncrude that can be refined into products. The conversionefficiency of DCL is greater than that of ICLand requires higher quality coal; however, DCL currentlyexists only in the laboratory and at pilot plantscale. China’s first two CTL plants, which will use theDCL process, are slated to be operational after 2008[64].When combined with related processes such as CHPor IGCC, CTL can be considered a byproduct, withFischer-Tropsch added as a part of a poly-generationconfiguration (steam, electricity, chemicals, andfuels). Revenues from the sale of electricity and/orsteam can significantly offset CTL production costs[65]. Prospects for CTL production could be constrained,however, by plant siting issues that includewaste disposal, water supply, and wastewater treatmentand disposal. Water-cooling limitations can beovercome through the use of air-cooling, although itadds to the cost of production. CTL requires water forthe front-end steps of coal preparation, and processingof coal with excessive moisture content can alsoproduce contaminated water that requires disposal.These issues are similar to those associated with typicalcoal-fired power plants.AEO2006 projects 800,000 barrels per day of domesticCTL production in the reference case and 1.7 millionbarrels per day in the high price case in 2030. Most ofthis activity initially occurs in coal-producing regionsof the Midwest. Worldwide CTL production in 2030totals 1.8 million barrels per day in the reference caseand 2.3 million barrels per day in the high price case.Gas-to-Liquids. GTL is the chemical conversion ofnatural gas into a slate of petroleum fuels. The processbegins with the reaction of natural gas with air(or oxygen) in a reformer to produce syngas, which isfed into the Fischer-Tropsch reactor in the presenceof a catalyst, producing a paraffin wax that ishydrocracked to products. A product workup sectionthen separates out the individual products. Distillateis the primary product, ranging from 50 percent to 70percent of the total yield.Given the significant capital costs of a GTL plant,natural gas reserves of 4 to 5 trillion cubic feet arerequired to provide a feedstock supply of 500 to 600million cubic feet per day over 25 years to support aplant with nominal capacity of 75,000 barrels per day.GTL competes with LNG for reserves of inexpensive,stranded natural gas located in scattered worldregions. Stranded natural gas lies far from marketsand would otherwise require major pipeline investmentsto commercialize. One processing advantagefor GTL plants is that they can use natural gas withhigh CO 2 content as a feedstock and can targetsmaller fields than are required for LNG production.Competition between GTL and LNG plants for theworld’s stranded natural gas supplies is not a limitingissue, however. All the GTL and LNG plants envisionedbetween now and 2030 would tap less than 15percent of the total world supply of stranded naturalgas.Capital costs for GTL plants range from $25,000 to$45,000 (2004 dollars) per barrel of daily capacity,depending on production scale and site selection.Those costs have dropped significantly, however,from more than $100,000 per barrel of total installedcapacity for the earliest plants. Opportunities to furtherlower the capital costs include reducing the sizeof air separation units, syngas reformers, andFischer-Tropsch reactors. Another opportunity lies inreducing cobalt and precious metals content in catalysts.An industry goal is to reduce GTL capital costsbelow $20,000 per barrel, but recent increases in steelprices and process equipment are making the goalmore elusive. By comparison, the cost of a conventionalpetroleum refinery is around $15,000 per barrelper day. In terms of engineering and constructionmetrics, a GTL facility with a capacity of 34,000 barrelsper day is roughly equivalent to a grassrootsrefinery with a capacity of 100,000 barrels per day[66].GTL is profitable when crude oil prices exceed $25 perbarrel and natural gas prices are in the range of $0.50to $1.00 per million Btu. The economics of GTL areextremely sensitive to the cost of natural gas feedstocks.As in the case of LNG, the presence of naturalgas liquids (NGL) in the feedstock stream can augmenttotal producer revenues, reducing the effectivecost of the natural gas input. In addition, the GTLprocess is exothermic, generating excess heat that canbe used to produce electricity, steam, or desalinatedwater and further enhance revenue streams.The technologies used for GTL are similar to thosethat have been employed for decades in methanol andEnergy Information Administration / Annual Energy Outlook 2006 55

Issues in Focusammonia plants, and most are relatively mature;however, the suite of integrated GTL technologieshas not been used on a commercial scale. One loominguncertainty with regard to GTL is whether a provenpilot plant can be scaled up to the size of a commercialplant while reducing capital and operating costs. Akey engineering goal is to improve the thermal efficiencyof the GTL process, which is more complexthan either LNG liquefaction or petroleum refining.The leading GTL processes include those developedby Shell, Sasol, Exxon, Rentech, and Syntroleum. Atthis time, there is no indication as to which technologywill prevail. Currently, the proponents of thesevarious processes have nearly 800,000 barrels per dayof first generation capacity under development inQatar.AEO2006 projects domestic GTL production originatingin Alaska, reflecting a longstanding proposal tomonetize stranded natural gas on the North Slope.GTL liquids would be transported to the lower 48refining system. In 2030, domestic GTL productiontotals 200,000 barrels per day in the high price case,even though it competes directly with the Alaska naturalgas pipeline project. In AEO2006, both investmentsare feasible simultaneously. What will actuallyoccur depends on how and where Alaska natural gasstakeholders ultimately decide to make their investments.GTL production worldwide exceeds 1.1 millionbarrels per day in the reference case and 2.6million barrels per day in the high price case in 2030.Biomass-to-Liquids. BTL encompasses the productionof fuels from waste wood and other non-foodplant sources, in contrast to conventional biodieselproduction, which is based primarily on food-relatedcrops. Because BTL does not ordinarily usefood-related crops, it does not conflict with increasingfood demands, although crops grown for BTLfeedstocks would compete with food crops for land.BTL gasification technology is based on the CTL process.The resulting syngas is similar, but the distributionof the hydrocarbon components differs. BTL useslower temperatures and pressures than CTL. LikeGTL, the BTL reaction is exothermic and requires acatalyst [67]. There are at least 13 known processescovering directly and indirectly heated gasifiers forthis step.BTL originates from renewable sources, includingwood waste, straw, grain waste, crop waste, garbage,and sewage/sludge. According to a leading processdeveloper, 5 tons of biomass yields 1 ton of BTL [68].One hectare (2.471 acres) of land generates 4 tons ofBTL. A modestly sized BTL plant under sustainedoperation would require the biomass of slightly morethan 12,000 acres [69]. Unlike biodiesel or ethanol,BTL uses the entire plant and, thereby, requires lessland use.BTL fuels are several times more expensive to producethan gasoline or diesel. Without taxes and distributionexpenses, a leading European developerestimates BTL production costs approaching $3.35per gallon by 2007 and falling to $2.43 per gallon by2020 [70]. This equates to a crude oil equivalent pricein the high $80 per barrel range at current capital costlevels.BTL technology is at the pilot-plant stage of development.The capital cost of a commercial-scale BTLplant could approach $140,000 (2004 dollars) per barrelof capacity, according to a study conducted forDOE by Bechtel in 1998 [71]. The estimated initialinvestment level is comparable with those for earlyCTL and GTL plants, which have since declined by 50percent or more. Technological innovations over timeand economies of scale could further reduce BTLcosts. The first commercial-scale BTL plant, with acapacity just over 4,000 barrels per day. is planned tobegin operation in Germany after 2008, followed byfour additional facilities. About two-thirds of a BTLplant’s capital cost is related to biomass handling andgasification. BTL front-end technology is new andevolving and has parallels with cellulose ethanoltechnology.Large BTL plants require huge catchment (staging)areas and incur high transportation costs to movefeedstocks to a central plant. From a process standpoint,the main challenge for BTL is the high cost ofremoving oxygen. It is unclear whether gasificationand other processing steps can achieve the cost reductionsnecessary to make it more competitive. Catalystcosts are high, as they are for other Fischer-Tropschprocesses. Without additional technological advancesto lower costs, BTL could be limited to the productionof fuel extenders rather than primary fuels.Renewable BiofuelsNot to be confused with BTLs are the renewablebiofuels, ethanol and biodiesel. These fuels can beblended with conventional fuels, which enhancestheir commercial attractiveness. Biofuels have highproduction costs and are about 2 to 3 times moreexpensive than conventional fuels. Renewable biofueltechnology is relatively mature for corn-based ethanolproduction, and future innovations are notexpected to bring its costs down substantially. Future56 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focuscost reductions are likely to be achieved by increasingproduction scale and implementing incremental processoptimizations. Energy is a significant componentof operating costs, followed by catalysts, chemicals,and labor. Production costs are highly localized.The greatest challenge facing biofuels production is tosecure sufficient raw material feedstock for conversioninto finished fuels. Production of biofuelsrequires significant land use dedicated to the growthof feedstock crops, and land prices could represent asignificant constraint.Ethanol. Ethanol, the most widely used renewablebiofuel, can be produced from any feedstock that containsplentiful natural sugars. Popular feedstocksinclude sugar beets (Europe), sugar cane (Brazil), andcorn (United States). Ethanol is produced by fermentingsugars with yeast enzymes that convert glucose toethanol. Crops are processed to remove sugar (bycrushing, soaking, and/or chemical treatment), thesugar is fermented to alcohol using yeasts andmicrobes, and the resulting mix is distilled to obtainanhydrous ethanol.There are two ethanol production technologies: sugarfermentation and cellulose conversion. Sugar fermentationis a mature technology, whereas celluloseconversion is new and still under development. Cellulose-to-biofuel(bioethanol) can use a variety of feedstocks,such as forest waste, grasses, and solidmunicipal waste, to produce synthetic fuel.Capital costs for a corn-based ethanol plant can rangefrom $21,000 to $33,000 (2004 dollars) per barrel ofcapacity, depending on size [72]. Manufacturing costscan be as low as $0.75 per gallon, as demonstratedby the low-cost production in Brazil, where climateconditions are favorable and labor costs are low.One industry risk is drought, which can limit theavailability of feedstocks. Another issue is competitionwith the food supply. Based on current land use,industry trade sources estimate that annual corn ethanolproduction in the United States is limited toCapital costs in transition for synthetic fuel facilitiesThe chart below shows the range of capital investmentcosts for the synthetic fuel technologies. A traditionalcrude oil refinery is shown as a point ofreference. Each of the alternative fuel technologiesis more expensive than an oil refinery, with a rangeof capital costs for each technology resulting fromindividual site location factors, facility layouts, competingvendor technologies, and production scale.Over time, investment costs for synthetic fuel facilitiesare expected to decrease as a result of “learning-by-doing.”As the installed base of synthetic fuelplants grows, cost reductions are expected to parallelthose seen in the past for LNG liquefaction facilities,which have achieved cost reductions oftwo-thirds over the past three decades.At present, observed capital costs generally areinversely proportional to installed capacity. There isabout 300,000 barrels per day of installed corn ethanolcapacity in the United States, whereas biodieselcapacity amounts to about 12,000 barrels per dayof dedicated capacity plus another 7,000 barrelsper day of swing capacity from the oleochemicalindustry.(Malaysia and South Africa) and global CTL capacitytotals 150,000 barrels per day at the original developmentplants in South Africa. There is no commercialBTL capacity in the United States or elsewherein the world, except for pilot plants.Putting the current production capacity of these variousfuels into perspective with traditional oil-basedfuels, U.S. refining capacity for all nonconventionalliquid fuels is over 17 million barrels per day, out of aworldwide total that is approaching 83 million barrelsper day.Range of capital investment costs for synthetic fuelfacilities (thousand 2004 dollars per daily barrelof capacity)Oil RefineryBiodieselEthanol (Corn)Gas-To-LiquidsThe liquefaction industry is still in its infancy. Atpresent there are no commercial GTL or CTL plantsin the United States other than pilot plants. Worldwide,GTL capacity is nearly 60,000 barrels per dayCoal-To-LiquidsBiomass-To-Liquids0 20 40 60 80 100 120 140 160Energy Information Administration / Annual Energy Outlook 2006 57

Issues in Focusapproximately 12 billion gallons to avoid disruptingfood markets.AEO2006 projects 700,000 barrels per day of ethanolproduction in 2030 in the reference case, representingabout 47 percent of world production. The high pricecase projects production of 900,000 barrels per day in2030, representing 30 percent of the world total.Worldwide, ethanol production (including biodiesel)in 2030 totals nearly 1.7 million barrels per day in thereference case and 3 million barrels per day in thehigh price case.Biodiesel. Biodiesel is produced from a variety of feedstocks,including soybean oil (United States), palm oil(Malaysia), and rapeseed and sunflower oil (Europe).The technology is mature and proven. In general, thefeedstock for biodiesel undergoes an esterificationprocess, which removes glycerin and allows the oil toperform like traditional diesel. Although biodiesel hasbeen produced and used in stationary applications(heat and power generation) for nearly a century, itsuse as a transportation fuel is recent. Today it is usedprimarily as an additive to “stretch” conventional dieselsupplies, rather than as a standalone primary fuel.One technical limitation of biodiesel is its blend instabilityand tendency to form insoluble matter. In theUnited States, those limitations are further aggravatedby the introduction of new ULSD into thenational fuel supply [73].Capital costs for biodiesel production facilities aresimilar to those for ethanol facilities, ranging from$9,800 to $29,000 (2004 dollars) per daily barrel ofcapacity, depending on size [74, 75]. Feedstocks forbiodiesel, which can be expensive, include inedibletallow ($41 per barrel), jatropha oil ($43 per barrel),palm oil ($46 per barrel), soybean oil ($73 per barrel),and rapeseed oil ($78 per barrel) [76]. On a gasoline-equivalentbasis, production costs in the UnitedStates range from 80 cents per gallon for biodieselfrom waste grease to $1.14 per gallon for biodieselfrom soybeans oil. U.S. biodiesel production totals20,000 barrels per day in 2030 in the AEO2006 referencecase and 30,000 barrels per day in the high pricecase.Mercury Emissions Control TechnologiesThe AEO2006 reference case assumes that States willcomply with the requirements of the EPA’s newCAMR regulation. CAMR is a two-phase program,with a Phase I cap of 38 tons of mercury emitted fromall U.S. power plants in 2010 and a Phase II cap of 15tons in 2018. Mercury emissions in the electricitygeneration sector in 2003 are estimated at around 50tons. Generators have a variety of options to meet themercury limits, such as: switching to coal with a lowermercury content, relying on flue gas desulfurizationor selective catalytic reduction equipment to reducemercury emissions, or installing conventional activatedcarbon injection (ACI) technology.The reference case assumes that conventional ACItechnology will be available as an option for mercurycontrol. Conventional ACI has been shown to be effectivein removing mercury from bituminous coals buthas not performed as well on subbituminous or lignitecoals. On the other hand, brominated ACI—a relativelynew technology—has shown promise in its abilityto control mercury emissions from subbituminousand lignite coals. Therefore, an alternative mercurycontrol technology case was developed to analyze thepotential impacts of brominated ACI technology.Preliminary tests sponsored by DOE indicate thatbrominated ACI can achieve high efficiencies inremoving mercury (approximately 90 percent orhigher for subbituminous coal and lignite, comparedwith about 60 percent for conventional ACI) at relativelylow carbon injection rates [77]. For the sensitivitycase, the mercury removal efficiency equationswere revised to reflect the latest brominated ACI dataavailable from DOE-sponsored tests [78]. BrominatedACI is about 33 percent more expensive than conventionalACI, and this change was also incorporated inthe alternative case. Other than the change in mercuryremoval efficiency and the higher cost ofbrominated ACI, the mercury emissions case uses thereference case assumptions.Figure 20 compares mercury emissions in the referenceand mercury control technology cases. Bothcases show substantial reductions in mercury emissions,with the greatest reductions occurring around2010 to 2012, when the CAMR Phase I cap has to bemet. The availability of brominated ACI results inslightly greater reductions in mercury emissions inthe 2010-2012 period, as generators are able to utilizethe technology to overcomply and bank allowancesfor later use. In the reference case, mercury emissionsfrom U.S. power plants total 37 tons in 2012, comparedwith 31 tons in the mercury control technologycase. In 2030, emissions are approximately the samein the two cases, at 15.3 and 15.6 tons.Figure 21 shows mercury allowance prices in the referenceand mercury control technology cases. Whenbrominated ACI is assumed to be available, it has asubstantial impact on mercury allowance prices in58 Energy Information Administration / Annual Energy Outlook 2006

Issues in Focusthe early years of the projection. In 2010, mercuryallowance prices are reduced from $23,400 per poundin the reference case to $8,700 per pound in themercury control technology case, a reduction of 63percent. The mercury control technology case incorporatesimproved ACI performance data for a limitednumber of plant configurations (those for whichdata were available from the DOE-sponsored tests),because not all plant configurations had been testedwith brominated ACI technology at the time [79]. Inthe alternative case, the difference in allowanceprices between the reference and mercury controltechnology cases narrows over the forecast horizon.Mercury allowance prices have a substantial impacton the market for pollution control equipment. Themercury control technology case shows that, asexpected, increased use of brominated ACI wouldgreatly influence the ACI equipment market. Figure22 compares the amounts of coal-fired capacityexpected to be retrofitted with ACI systems in thereference and mercury control technology cases.The impact is significant in the alternative caseFigure 20. Mercury emissions from the electricitygeneration sector, 2002-2030 (short tons per year)605040302010Mercury controltechnology caseReference case02002 2010 2015 2020 2025 2030Figure 21. Mercury allowance prices, 2010-2030(thousand 2004 dollars per pound)8060ReferenceMercury control technologythroughout the projection period. In the referencecase, about 125 gigawatts of coal-fired capacity isretrofitted with ACI by 2030. In the mercury controltechnology case, as a result of more effective mercuryremoval with brominated ACI, only about 88gigawatts of coal-fired capacity is retrofitted with ACIby 2030.The mercury control technology case assumes thatbrominated ACI will be commercially available before2010 (CAMR Phase I), and that the cost and performancelevels seen in the initial DOE-sponsored testswill be replicable in the systems being offered commercially.Under these assumptions, comparison ofthe reference and mercury control technology caseshighlights several important points. The mercuryemissions levels are similar in the two cases, butallowance prices are much lower in the alternativecase, through 2020. Corresponding to the differencein allowance prices, significantly less coal-fired capacityis retrofitted with ACI in the mercury control technologycase than in the reference case. Overall,electricity generators are able to comply with theCAMR requirements more easily when they haveaccess to the brominated ACI technology, whileachieving the same reductions in mercury emissionsas in the reference case and complying with theCAMR caps.U.S. Greenhouse Gas Intensity and theGlobal Climate Change InitiativeOn February 14, 2002, President Bush announced theAdministration’s Global Climate Change Initiative[80]. A key goal of the Climate Change Initiative is toreduce U.S. GHG intensity—defined as the ratio oftotal U.S. GHG emissions to economic output—by 18percent over the 2002 to 2012 time frame.Figure 22. Coal-fired generating capacityretrofitted with activated carbon injection systems,2010-2030 (gigawatts)150125100ReferenceMercury control technology407550202502010 2015 2020 2025 203002010 2015 2020 2025 2030Energy Information Administration / Annual Energy Outlook 2006 59

Issues in FocusAEO2006 projects energy-related CO 2 emissions,which represented approximately 83 percent of totalU.S. GHG emissions in 2002. Projections for theother GHGs are derived from an EPA “no-measures”case, a recent update to the “business-as-usual” casecited in the White House Greenhouse Gas Policy BookAddendum [81] released with the Climate ChangeInitiative. The projections from the Policy Book werebased on several EPA-sponsored studies conducted inpreparation for the U.S. Department of State’s ClimateAction Report 2002 [82]. The no-measures casewas developed by EPA in preparation for a planned2006 “National Communication” to the UnitedNations in which a “with-measures” policy case is tobe published [83]. Table 15 combines the AEO2006reference case projections for energy-related CO 2emissions with the projections for other GHGs.Although AEO2006 does not include cases that specificallyaddress alternative assumptions about GHGintensity, the integrated high technology case doesgive some indication of the feasibility of meeting the18-percent intensity reduction target. In the integratedhigh technology case, which combines the hightechnology cases for the residential, commercial,industrial, transportation, and electric power sectors,CO 2 emissions in 2012 are projected to be 166 millionmetric tons less than the reference case projection. Asa result, U.S. GHG intensity would fall by 18.6 percentfrom 2002 to 2012, more than enough to meetthe Administration’s goal of 18 percent (Figure 23).Figure 23. Projected change in U.S. greenhouse gasintensity in three cases, 2002-2020 (percent)0According to the combined emissions projections inTable 15, the GHG intensity of the U.S. economy isexpected to decline by 17 percent between 2002 and2012, and by 28 percent between 2002 and 2020 in thereference case. The Administration’s goal of reducingGHG intensity by 18 percent by 2012 would requireemissions reductions of about 116 million metric tonsCO 2 equivalent from the projected levels in the referencecase.-5-10-15-20-25-30-352012 goal2002 2007 2012 2016 20202005technologycaseReferenceHightechnologycaseTable 15. Projected changes in U.S. greenhouse gas emissions, gross domestic product, and greenhouse gasintensity, 2002-2020ProjectionPercent ChangeMeasure2002 2012 2020 2002-2012 2002-2020Greenhouse gas emissions(million metric tons carbon dioxide equivalent)Energy-related carbon dioxide 5,746 6,536 7,119 13.7 23.9Methane 626 686 739 9.5 18.0Nitrous oxide 335 351 366 4.9 9.3Gases with high global warming potential 143 245 339 71.2 136.6Other carbon dioxide and adjustmentsfor military and international bunker fuel 62 79 86 26.7 37.2Total greenhouse gases 6,913 7,897 8,649 14.2 25.1Gross domestic product (billion 2000 dollars) 10,049 13,793 17,541 37.3 74.6Greenhouse gas intensity(thousand metric tons carbon dioxideequivalent per billion 2000 dollars of grossdomestic product) 688 573 493 -16.8 -28.360 Energy Information Administration / Annual Energy Outlook 2006

Market TrendsThe projections in AEO2006 are not statements ofwhat will happen but of what might happen, giventhe assumptions and methodologies used. Theprojections are business-as-usual trend estimates,given known technology, technological and demographictrends, and current laws and regulations.Thus, they provide a policy-neutral reference casethat can be used to analyze policy initiatives. EIAdoes not propose, advocate, or speculate on futurelegislative and regulatory changes. All laws areassumed to remain as currently enacted; however,the impacts of emerging regulatory changes, whendefined, are reflected.Because energy markets are complex, models aresimplified representations of energy production andconsumption, regulations, and producer and consumerbehavior. Projections are highly dependenton the data, methodologies, model structures,and assumptions used in their development.Behavioral characteristics are indicative of realworldtendencies rather than representations ofspecific outcomes.Energy market projections are subject to muchuncertainty. Many of the events that shape energymarkets are random and cannot be anticipated,including severe weather, political disruptions,strikes, and technological breakthroughs. In addition,future developments in technologies, demographics,and resources cannot be foreseen withcertainty. Many key uncertainties in the AEO2006projections are addressed through alternative cases.EIA has endeavored to make these projections as objective,reliable, and useful as possible; however,they should serve as an adjunct to, not a substitutefor, a complete and focused analysis of public policyinitiatives.

Trends in Economic ActivityStrong Economic Growth Is ExpectedTo Continue Through 2030Figure 24. Average annual growth rates of realGDP, labor force, and productivity in three cases,2004-2030 (percent per year)4.0 ReferenceHigh growth3.0Low growth2.01.00.0Real GDP Labor force ProductivityAEO2006 presents three views of economic growthfor the forecast period from 2004 through 2030.Although probabilities are not assigned, the referencecase reflects the most likely view of how the economywill unfold over the period. In the reference case, theNation’s economic growth, measured in terms of realGDP based on 2000 chain-weighted dollars, is projectedto average 3.0 percent per year (Figure 24). Thelabor force is projected to grow by 0.8 percent per yearon average; labor productivity growth in the nonfarmbusiness sector is projected to average 2.3 percent peryear; and investment growth is projected to average4.0 percent per year. Disposable income grows by 3.1percent per year in the reference case and disposableincome per capita by 2.2 percent per year. Nonfarmemployment grows by 1.1 percent per year, whileemployment in manufacturing shrinks by 0.5 percentper year.The high and low economic growth cases show theeffects of alternative growth assumptions on theenergy market projections. The high growth caseassumes higher growth rates for population (1.2 percentper year), nonfarm employment (1.4 percent),and productivity (2.7 percent). With higher productivitygains and employment growth, projected inflationand interest rates are lower than in the referencecase. The low growth case assumes lower growthrates for population (0.5 percent per year), nonfarmemployment (0.7 percent per year), and productivity(1.8 percent per year), resulting in higher projectionsfor prices and interest rates and lower projections forindustrial output growth.Unemployment, Interest, andInflation Rates Near Historical NormsFigure 25. Average annual unemployment, interest,and inflation rates, 2004-2030 (percent per year)0.0Unemployment4.7InterestAA utility bond rate7.710-year Treasury note5.9Federal funds rate4.9InflationGDP chain-type index2.5CPI: all urban2.7CPI: energy2.6WPI: all commodities 1.5WPI: fuel and power0.02.6In the reference case, U.S. economic indicators generallyare projected to follow historical trends, on average,from 2004 through 2030. Economic factors thatare widely viewed as barometers for conditions in themarkets for labor, credit, and goods and servicesinclude: the average annual unemployment rate;yields on Federal funds, 10-year U.S. Treasury notes,and AA utility corporate bonds; and average annualinflation rates as measured by various wholesale andretail price indexes (Figure 25). For AEO2006, unemploymentand interest rates are calculated as annualaverages over the 2004-2030 period, and inflationrates are calculated as average annual percentchanges in the price indexes.From 2004 through 2010, the economy (in terms ofreal GDP) is projected to grow more rapidly than itsprojected long-term average growth rate in the referencecase. Over the same period, the unemploymentrate is projected to decline from 5.5 percent in 2004 to4.7 percent in 2010. After an initial rise through 2015,the Federal funds rate is projected to decline to its historicalnorm of 5 percent. Longer term rates areexpected to be higher than the Federal funds rate,with the 10-year Treasury note yielding 6 percent andAA utility corporate bonds yielding approximately 8percent per year, on average, for the entire forecastperiod. The reference case projects an average annualinflation rate of 2.7 percent, as measured by all urbanCPI—slightly higher than the CPI for energy commoditiesand services or the wholesale price index(WPI) for fuel and power.62 Energy Information Administration / Annual Energy Outlook 2006

Trends in Economic ActivityOutput Growth for Energy-IntensiveIndustries Is Expected To SlowEnergy Expenditures Relative to GDPAre Projected To DeclineFigure 26. Sectoral composition of output growthrates, 2004-2030 (percent per year)Figure 28. Energy expenditures as share of grossdomestic product, 1970-2030 (nominal expendituresas percent of nominal GDP)Total gross output2.9Services3.2HistoryProjections25Industrial2.1Energy intensityManufacturing2.3(thousand Btu per20real dollar of GDP)Nonmanufacturing1.417.49.3Agriculture: crops1.2155.8Agriculture: other 0.7Coal mining1.6101973 2004 2030Oil and gas extraction 0.1Other mining1.55All energyPetroleumConstruction1.7Natural gas00.0 1.0 2.0 3.0 4.0 1970 1980 1990 2000 2010 2020 2030The industrial sector (all non-service industries) hasshown slower output growth than the economy as awhole in recent decades, with imports meeting agrowing share of demand for industrial goods. Thattrend is expected to continue in the reference caseprojections.Within the industrial sector, the output of manufacturingindustries is projected to grow more rapidlythan that of nonmanufacturing industries, whichinclude agriculture, mining, and construction (Figure26). With higher energy prices and more foreign competitionexpected, however, the energy-intensivemanufacturing sectors [84] are projected to grow byonly 1.3 percent per year from 2004 through 2030,compared with a projected 2.6-percent averageannual rate of growth for non-energy-intensive manufacturingoutput (Figure 27).Figure 27. Sectoral composition of manufacturingoutput growth rates, 2004-2030 (percent per year)ManufacturingEnergy-intensiveFoodPaperInorganic chemicalsOrganic chemicalsResinsAgricultural chemicalsGlassCementIron and steelAluminumPetroleum refineriesNon-energy-intensiveMetal-based durablesOther manufacturing-1.0-0.51.11.31.92.30.60.90.12.11.10.90.81.02.63.01.90.0 1.0 2.0 3.0The ratio of total expenditures for energy relative tototal GDP (both in nominal dollars) provides an indicationof the importance of energy expenditures inthe aggregate economy. Before the oil embargo of1973-74, total energy expenditures were equal to 8percent of U.S. GDP, petroleum expenditures justunder 5 percent, and natural gas expenditures 1 percent.Following the price shocks of the 1970s andearly 1980s, those shares rose dramatically—to 14percent, 8 percent, and 2 percent, respectively, in1981. Since then they have fallen consistently, to2004 levels of about 7 percent for total energy expenditures,4 percent for petroleum expenditures, andjust over 1 percent for natural gas expenditures.Although recent developments in the world oil markethave pushed the shares upward, they are projectedto decline from current levels in the referencecase. In 2030, total nominal energy expenditures areprojected to equal 5 percent of nominal GDP, petroleumexpenditures 3 percent, and natural gas expendituresless than 1 percent (Figure 28).The overall decline in energy expenditures relative toGDP has resulted in large part from a decline in worldoil prices (in real dollar terms) from their peak in1981. And although oil prices have risen recently,their long-term trajectory in the AEO2006 referencecase is relatively flat in real terms. Another reason forthe declining share of energy expenditures has been asteady decline in the energy intensity of the U.S.economy, measured as energy consumption (thousandBtu) per dollar of real GDP. Structural shifts inthe economy and improvements in energy efficiencyhave allowed for the decline in energy intensity,which is projected to continue through 2030.Energy Information Administration / Annual Energy Outlook 2006 63

International Oil MarketsOil Price Cases Show Uncertaintyin Prospects for World Oil MarketsFigure 29. World oil prices in three cases, 1980-2030(2004 dollars per barrel)10080604020HistoryProjections01980 1995 2004 2015 2030High priceReferenceLow priceWorld oil price projections in the AEO2006 referencecase, in terms of the average price of importedlow-sulfur crude oil to U.S. refiners, are considerablyhigher than those presented in the AEO2005 referencecase. The higher price path in the reference casedoes not result from different assumptions about theultimate size of world oil resources but rather anticipatesa lower level of future investment in productioncapacity in key resource-rich regions and a reassessmentof the willingness of OPEC to produce at higherrates than projected in last year’s outlook.The historical record shows substantial variability inworld oil prices, and there is arguably even moreuncertainty about future prices in the long term.AEO2006 considers three price cases, allowing anassessment of alternative views on the course offuture oil prices (Figure 29). In the reference case,world oil prices moderate from current levels to $47per barrel in 2014, before rising to $57 per barrel in2030 (2004 dollars). The low and high price casesdefine a wide range of potential world oil price paths,which in 2030 range from $34 to $96 per barrel.This variability is meant to show the uncertaintyabout prospects for future world oil resources andeconomics.In all three price cases, non-OPEC suppliers produceto capacity. Thus, the variation in price paths has thegreatest impact on the need for OPEC supply in thelong term. In 2030, the call on OPEC is 46.8 millionbarrels per day in the reference case and 51.3 millionbarrels per day in the low price case, but only 31.7million barrels per day in the high price case—notmuch more than current OPEC production levels.Oil Imports Reach More Than18 Million Barrels per Day by 2030Figure 30. U.S. gross petroleum imports by source,2004-2030 (million barrels per day)201510502004 2005 2010 2015 2020 2025 2030OtherFar EastCaribbeanEuropeNorth AmericaOther OPECOPECPersian GulfTotal U.S. gross petroleum imports increase in thereference case, from 13.1 million barrels per day in2004 to 18.3 million in 2030 (Figure 30), deepeningU.S. reliance on imported oil in the long term. In2030, gross petroleum imports account for 64 percentof total U.S. petroleum supply.More than one-half of the increase in U.S. grossimports comes from OPEC suppliers. Crude oilimports from the North Sea decline as productionebbs, and West Coast refiners import small volumesof crude oil from the Far East to replace a decline insupplies of Alaskan crude oil. Canada and Mexico continueto be important sources of U.S. petroleum supply.Much of the future Canadian contribution comesfrom the development of its enormous oil sandsresource base; however, the availability of suchnonconventional oil supplies is linked to world oilprices. In the high price case, nonconventional suppliesare more competitive with conventional sources,rising to about 21.1 million barrels per day worldwidein 2030. In the low price case, nonconventional productiontotals only 7.1 million barrels per day in 2030.U.S. imports of refined petroleum products alsoincrease. Most of the increase comes from refiners inthe Caribbean Basin, North Africa, and the MiddleEast, where refining capacity is expected to expandsignificantly. Vigorous growth in demand for lighterpetroleum products in developing countries meansthat U.S. refiners are likely to import smaller volumesof light, low-sulfur crude oils.64 Energy Information Administration / Annual Energy Outlook 2006

Energy DemandAverage Energy Use per PersonIncreases Through 2030Figure 31. Energy use per capita and per dollar ofgross domestic product, 1980-2030 (index, 1980 = 1)1.41.21.00.80.60.40.2HistoryProjectionsEnergyuse percapitaEnergyuse perdollarof GDPCoal and Petroleum Lead Increasesin Primary Energy UseFigure 32. Primary energy use by fuel, 2004-2030(quadrillion Btu)150125100755025Other renewablesHydropowerNuclearCoalNatural gasPetroleum0.01980 1995 2004 2015 203002004 2010 2015 2020 2025 2030Population growth is a key determinant of totalenergy consumption, closely linked to rising demandfor housing, services, and travel. Energy consumptionper capita, controlling for population growth, showsthe combined effect of other factors, such as economicgrowth and technology improvement. In the AEO-2006 reference case, energy consumption per capitagrows faster than it has in recent history (Figure 31),as a result of continued growth in disposable income.In dollar terms, the economy as a whole is becomingless dependent on energy, the Nation’s growing relianceon imported fuel notwithstanding. Projectedenergy intensity, as measured by energy use per 2000dollar of GDP, declines at an average annual rate of1.8 percent in the reference case. Efficiency gains andfaster growth in less energy-intensive industriesaccount for much of the decline, more than offsettingthe expected growth in demand for energy services inbuildings, transportation, and electricity generation.Energy intensity declines more rapidly in the nearterm, as consumers and businesses react to highenergy prices. As energy prices moderate over the longerterm, energy intensity declines at a slower rate. Asimilar pattern occurred from 1986 to 1992, whenenergy prices were generally falling.AEO2006 does not assume policy-induced conservationmeasures beyond those in existing legislationand regulation, nor does it assume behavioralchanges that could result in greater energy conservation,beyond those experienced in the past.Total primary energy consumption, including energyfor electricity generation, grows by 1.1 percent peryear from 2004 to 2030 (Figure 32). Fossil fuelsaccount for 88 percent of the growth, with coal useincreasing by 53 percent, petroleum by 34 percent,and natural gas by 20 percent over that period. Theincrease in coal consumption occurs primarily in theelectric power sector, with strong growth in electricitydemand and favorable economics promptingincreases in coal-fired baseload capacity. More thanone-half of the increase in coal consumption isexpected after 2020, when higher natural gas pricesmake coal the fuel choice for most new power plants.Growth in natural gas consumption for power generationis restrained by its high price relative to coal.Industry and buildings account for 71 percent of theincrease in natural gas consumption from 2004 to2030.Transportation accounts for 87 percent of theincrease in petroleum consumption, dominated bygrowth in fuel use for light-duty vehicles. Fuel use byfreight trucks, second in energy use among travelmodes, grows by 1.9 percent per year on average, thefastest annual rate among the major forms of transport.Petroleum use in the buildings sectors, mostlyfor space heating, declines slightly in the projection.AEO2006 projects rapid growth in energy productionfrom nonhydroelectric renewable sources, partly as aresult of State mandates for renewable electricitygeneration and renewable energy production taxcredits. An increase in power generation from nuclearenergy is also projected, as tax credits spur constructionof new nuclear plants between 2014 and 2018.Energy Information Administration / Annual Energy Outlook 2006 65

Energy DemandPetroleum and Electricity LeadGrowth in Energy ConsumptionFigure 33. Delivered energy use by fuel, 1980-2030(quadrillion Btu)60504030HistoryProjectionsPetroleumU.S. Primary Energy Use Climbs to134 Quadrillion Btu in 2030Figure 34. Primary energy consumption by sector,1980-2030 (quadrillion Btu)60HistoryProjectionsResidentialCommercial50IndustrialTransportation403020Natural gasElectricity2010Coal0Renewables1980 1995 2004 2015 20301001980 1990 2000 2010 2020 2030Delivered energy use (excluding losses in electricitygeneration) grows by 1.1 percent per year on averagefrom 2004 through 2030. Petroleum use, whichmakes up more than one-half of total delivered energyuse, grows at about the same rate (Figure 33). Thefastest growth in petroleum use is projected for transportationenergy use. Although high fuel prices tendto restrain travel by passenger and commercial vehicles,economic growth and more travel per capitaincrease demand for gasoline and diesel fuel, assumingno changes in consumer behavior.Past trends in electricity consumption are expected tocontinue, with future increases resulting from stronggrowth in commercial floorspace, continued penetrationof electric appliances in the residential sector,and increases in industrial output. Natural gas usegrows more slowly than overall delivered energydemand, in contrast to its more rapid growth duringthe 1990s. As a result, natural gas meets 21 percent oftotal end-use energy requirements in 2030, comparedwith 24 percent in 2004.End-use demand for energy from marketed renewables,such as wood, grows by 1.0 percent per year.Biomass used in the industrial sector, mostly as abyproduct fuel in the pulp and paper industry, is thelargest source of renewable fuel for end use. Demandfor purchased wood for home heating remains steadyin the projections, with potential growth constrainedby its limited availability and inconvenience. Energyfrom nonmarketed renewables, such as solar and geothermalheat pumps, more than doubles over the projectionperiod but is less than 1 percent of deliveredresidential energy use throughout the period.Primary energy use (including electricity generationlosses) increases by one-third over the next 26 years(Figure 34). The projected growth rate of energy consumptionin the AEO2006 reference case approximatelymatches the average from 1981 to 2004.Demand for energy in the early 1980s fell in the faceof recession, high energy prices, and changing regulations;but beginning in the mid-1980s, declining realenergy prices and economic expansion contributed toa marked increase in energy consumption. The longtermupward trend in energy consumption is projectedto continue in the reference case, with growthslowed somewhat by rising energy prices.The most rapid growth in sectoral energy use is in thecommercial sector, where services continue to expandmore rapidly than the economy as a whole. Thegrowth rate for residential energy use is about halfthat for the commercial sector, with demographictrends being a dominant factor. Transportationenergy use grows at a slightly slower rate than it hassince 1980, despite high fuel prices. Increases intravel by personal and commercial vehicles are onlypartially offset by vehicle efficiency gains. Primaryenergy use in the industrial sector grows more slowlythan in the other sectors, with efficiency gains, higherreal energy prices, and shifts to less energy-intensiveindustries moderating the expected growth.Alternative cases have been developed to explore thekey uncertainties in the forecast, including two economicgrowth cases and two world oil price cases.Detailed projections and comparisons with the referencecase are provided in Appendixes B, C, and D.66 Energy Information Administration / Annual Energy Outlook 2006

Residential Sector Energy DemandDemographic Shifts Lead to Changesin Residential Energy Use per CapitaFigure 35. Delivered residential energyconsumption per capita by fuel, 1980-2030(index, 1980 = 1)2.0 History Projections1.5ElectricityEnergy Use per Household for Spaceand Water Heating Is Expected To FallFigure 36. Delivered residential energyconsumption by end use, 2001, 2004, 2015, and 2030(million Btu per household)50 200120042015402030301.00.5Natural gasPetroleum20100.01980 1995 2004 2015 20300SpaceheatingSpacecoolingWaterheatingRefrigerationLightingAllotherResidential electricity use per person has increasedsignificantly since 1980 (Figure 35). With the U.S.population migrating south and west, electricity usefor air conditioning has become more important thannatural gas and petroleum for space heating. TheSouth and West Census regions, which accounted for52 percent of the U.S. population in 1980, increased to59 percent in 2004 and continue increasing to nearly65 percent in 2030.The type and size of houses, household energy uses,and the fuels chosen vary by region. In the Northeast,37 percent of households (compared with 27 percentnationally) live in multifamily units, which generallyare smaller and use less energy per household thanother types of housing. Fuel use for space heating isrelatively more important in this region than in otherregions. The Northeast, which accounted for 21 percentof the U.S. population in 1980, decreased to 19percent in 2004 and falls to 16 percent in 2030. This isone of the factors that contributes to a decline in heatingoil use per capita in the U.S. residential sector.Natural gas use per capita has remained relativelyconstant in the residential sector since 1990. In 2004,56 percent of all households and 63 percent of newsingle-family households used natural gas for homeheating. Natural gas consumption per householddeclines as a growing share of the population livesin warmer climates; but per capita consumption ofnatural gas does not change significantly, because theaverage size of new houses increases.The size, type, and location of housing affect not onlythe type and amount of energy consumed per householdbut also which services are used more intensively.Larger houses require more energy to heat,cool, and illuminate, and as housing size continues togrow, energy use per household for these services canbe expected to grow, all else being equal. Energyconsumption for space heating, water heating, andrefrigeration per household decreases over time,while energy use for lighting and all other applicationsgrows, despite continuing increases in theirenergy efficiency (Figure 36).In 2004, houses required 101 million Btu of deliveredenergy on average to provide all services. Energy usefor space heating, the largest single energy-consumingservice, declines by about 9 million Btu per household(20 percent) from 2004 to 2030 as a result ofincreasing energy efficiency and a 5-percent decreasein the average number of heating degree-days peryear. Energy use for space cooling per householdincreases slightly, based on an 8-percent increase incooling degree-days and the expectation that centralair conditioning will be installed in more existinghomes over time.The “all other” category shows the fastest growth ona per household basis, as higher incomes and newuses lead to additional purchases of electronics andother miscellaneous devices. In 2004, 21 percent ofthe energy used in the average home was for smallappliances. Their share of energy use per householdgrows to 29 percent in 2030, as more large-screentelevision sets, computers, and related equipment arepurchased.Energy Information Administration / Annual Energy Outlook 2006 67

Residential Sector Energy DemandIncreases in Energy EfficiencyAre Projected To ContinueFigure 37. Efficiency indicators for selectedresidential appliances, 2004 and 2030(index, 2004 stock efficiency = 1)2.01.51.00.50.0NaturalgasfurnacesResistancewaterheatersCentralair conditionersRefrigeratorsbuildingNewshell efficiency2004 stock2030 stock2004best availableCurrentstandardThe energy efficiency of new household appliancesplays a large role in determining the type and amountof energy used in the residential sector. As a result ofstock turnover and purchases of more efficient equipment,the amount of energy used by residential consumerson a per household basis has fallen over time,and many technologies exist today that can furtherreduce residential energy consumption if they arepurchased and used by more consumers (Figure 37).The most efficient technologies can provide significantlong-run savings in energy bills, but their higherpurchase costs (and in some cases unsuitability forretrofit applications) may restrict their market penetration.For example, condensing technology for naturalgas furnaces, which reclaims heat from exhaustgases, can raise efficiency by more than 20 percentover units that just meet the current standard. Incontrast, there is little room for efficiency improvementsin electric resistance water heaters, becausethe technology is approaching its thermal limit.In 2004, 8 percent of all new single-family homes werecertified as ENERGY STAR compliant, implying atleast a 30-percent energy savings for heating andcooling relative to comparable homes built to currentcode. Four States—Texas, California, Arizona, andNevada—account for two-thirds of all ENERGYSTAR home completions, concentrating energy savingsin areas with relatively moderate climates.ENERGY STAR completions, as a percent of totalcompletions, are expected to increase over time asbuilders become more familiar with the requiredbuilding practices and the cost of the more efficientcomponents used decreases.Advanced Technologies Could ReduceResidential Energy UseFigure 38. Variation from reference case deliveredresidential energy use in three alternative cases,2004-2030 (million Btu per household)50-5-102005 technologyReferenceHigh technologyBest available-15technology2004 2010 2015 2020 2025 2030The reference case includes the effects of several policiesaimed at increasing residential end-use efficiency,including minimum efficiency standards andvoluntary energy savings programs designed to promoteenergy efficiency through innovations in manufacturing,building, and mortgage financing. In the2005 technology case, which assumes no increase inthe efficiency of equipment or building shells beyondthat available in 2005, energy use per household in2030 would be 5 percent higher than projected in thereference case (Figure 38). In the best available technologycase, which assumes that the most energyefficienttechnology is always chosen regardless ofcost, energy use per household in 2030 would be 15percent lower than in the reference case and 19 percentlower than in the 2005 technology case. In thehigh technology case, which assumes earlier availabilityof the most energy-efficient technologies, withlower costs and higher efficiencies, but does not constrainconsumer choices, energy use per householdwould be 6 percent lower than projected in the referencecase but higher than in the best available case.In the high technology case, the consumer discountrates used to determine the purchased efficiency of allresidential appliances do not vary from those used inthe reference case; that is, consumers value efficiencyequally across the two cases. Energy-efficient equipment,such as central air conditioners with efficiencyratings 50 percent higher than the current standard,can significantly reduce electricity use for space cooling.Likewise, home builders can construct homesthat use 50 percent less energy for heating and coolingrelative to current code in most regions of thecountry.68 Energy Information Administration / Annual Energy Outlook 2006

Commercial Sector Energy DemandEconomic, Population Growth ShapeTrends in Commercial Energy UseFigure 39. Delivered commercial energyconsumption per capita by fuel, 1980-2030(index, 1980 = 1)2.5 History Projections2.01.5ElectricityEfficiency Gains Moderate Increasesin Commercial Energy IntensityFigure 40. Delivered commercial energy intensityby end use, 2004, 2015, and 2030(thousand Btu per square foot)75 200420152030501.0Natural gas250.50.01980 1995 2004 2015 2030Petroleum0SpaceheatingWaterheatingLightingCoolingOffice:PCsOffice:otherMisc.gasAllotherRecent trends in commercial fuel use are expected tocontinue, with growth in overall consumption similarto its pace in recent history (Figure 39). Commercialdelivered energy use (excluding primary energy lossesin electricity generation) grows at about the samerate as commercial floorspace, by 1.6 percent per yearfrom 2004 through 2030.Commercial floorspace growth and, in turn, commercialenergy use are driven by trends in economic andpopulation growth. Growth in disposable incomeleads to increased demand for services ranging fromhotels and restaurants to stores, theaters, galleries,and arenas. These establishments continue to growmore electricity-based, as well as depending on electricity-basedsupport services such as electronic processingcenters and Internet providers to completebusiness transactions.Increases in cooling demand contribute to the growthin commercial electricity use per capita, as the commercialsector expands to serve expected populationgrowth in the South and West. Slower populationgrowth in the Northeast, where heating oil is usedmore extensively, contributes to a decline in percapita petroleum use. Population effects on projectedcommercial energy use are not limited to regionaltrends. The share of the population over age 65increases from 12 to 20 percent between 2004 and2030, increasing the need for healthcare and assistedliving facilities and for electricity to power medicaland monitoring equipment in those facilities.The determinants of commercial energy demandinclude both structural and demographic components.The Nation’s continued move to a service economyimplies growth in commercial services that useenergy intensively, such as health care; however, newconstruction must meet building codes and efficiencystandards, offsetting potential increases in energyintensity (consumption per square foot of commercialfloorspace).Energy intensity for the major commercial end usesdeclines as more energy-efficient equipment isadopted (Figure 40). Minimum efficiency standards,including those in EPACT2005, contribute to projectedefficiency improvements in commercial spaceheating, space cooling, water heating, and lighting,moderating the growth in commercial energydemand. An increase in building shell efficiency alsocontributes to the trend. In addition, the prospect ofhigh fossil fuel prices factors into expected efficiencyincreases for space and water heating equipment.Increases in energy intensity are expected only forend uses that have not yet saturated the commercialmarket, including electricity use for office equipment,as well as new telecommunications technologies andmedical imaging equipment in the “all other” end-usecategory. The “all other” category also includesventilation, refrigeration, minor fuel consumption,municipal water services, service station equipment,elevators, vending machines, and a myriad of otheruses.Energy Information Administration / Annual Energy Outlook 2006 69

Commercial Sector Energy DemandCurrent Technologies ProvidePotential Energy SavingsFigure 41. Efficiency indicators for selectedcommercial equipment, 2004 and 2030(index, 2004 stock efficiency = 1)1.51.0Future standard2004 stock2030 stock2004best availableCurrent standardAdvanced Technologies Could ReduceCommercial Energy UseFigure 42. Variation from reference case deliveredcommercial energy intensity in three alternativecases, 2004-2030 (thousand Btu per square foot)1052005 technology0.50-5ReferenceHigh technology0.0GasfurnacesOil-firedboilersCommercial Electricunit air resistanceconditioners waterheaters-10Best available-15technology2004 2010 2015 2020 2025 2030The stock efficiency of energy-using equipment in thecommercial sector increases in the AEO2006 referencecase (Figure 41). As equipment is replaced andnew buildings are built, technology advances are expectedto reduce commercial energy intensity in mostend-use services, although the long equipment servicelives for many technologies moderate the pace ofefficiency improvement in the forecast. For much ofthe equipment covered by the EPACT1992, the existingFederal efficiency standards are higher than theaverage efficiency of the 2004 stock, ensuring someincrease in the stock average efficiency as new equipmentis added. EPACT2005 includes efficiency standardsfor a variety of commercial technologies, suchas air-cooled air conditioners, guaranteeing furtherincreases in stock efficiency. Future updates to theFederal standards could have significant effects oncommercial energy consumption, but they are not includedin the reference case.Currently available technologies have the potential toreduce commercial energy consumption significantly.Improved heat exchangers for oil-fired boilers canraise efficiency by 8 percent over the current standard;and the use of multiple compressors andenhanced heat exchanger surfaces can increase theefficiency of unit air conditioners by more than50 percent over the current standard and more than20 percent over the new standard. When a business isconsidering an equipment purchase, however, theadditional capital investment required for the mostefficient technologies often carries more weight thando future energy savings, limiting the adoption ofmore efficient technologies.The AEO2006 reference case incorporates efficiencyimprovements for commercial equipment and buildingshells, resulting in little change in projected commercialenergy intensity (delivered energy use persquare foot of floorspace) over the projection period,despite increased demand for services. The 2005 technologycase assumes that future equipment andbuilding shells will be no more efficient than thoseavailable in 2005. The high technology case assumesearlier availability, lower costs, and higher efficienciesfor more advanced equipment than in the referencecase and more rapid improvement in buildingshells. The best available technology case assumesthat only the most efficient technologies will be chosen,regardless of cost, and that building shells willimprove at a faster rate than assumed in the hightechnology case.In the 2005 technology case, energy use per squarefoot in 2030 is 6 percent higher than in the referencecase (Figure 42), as a result of an 0.3-percent averageannual increase in commercial delivered energyintensity. The high technology case projects a3-percent energy savings per square foot in 2030 relativeto the reference case. In the best available technologycase, commercial delivered energy intensity in2030 is 12 percent lower than in the reference case.More optimistic assumptions result in additionalprojected energy savings from both renewable andconventional fuel-using technologies. In 2030, commercialsolar PV systems generate more than threetimes as much electricity in the best technology caseas in the reference case.70 Energy Information Administration / Annual Energy Outlook 2006

Industrial Sector Energy DemandAdvanced Technologies Could SlowElectricity Sales Growth for BuildingsFigure 43. Buildings sector electricity generationfrom advanced technologies in two alternativecases, 2030 (percent change from reference case)300250High technologyBest available technologyEconomic Growth, Structural ChangeShape Industrial Energy IntensityFigure 44. Industrial energy intensity by twomeasures, 1980-2030 (thousand Btu per 2000 dollar)8 History Projections62001501005042Energy useper dollar value of shipmentsEnergy useperdollarofGDP0Photovoltaics Fuel cells Microturbines01980 1995 2004 2015 2030Alternative technology cases for the residential andcommercial sectors include varied assumptions forthe availability and market penetration of advanceddistributed generation technologies (solar PV systems,fuel cells, and microturbines). In the high technologycase, buildings generate 1.2 billionkilowatthours (10 percent) more electricity in 2030than in the reference case (Figure 43), most of whichoffsets residential and commercial electricity purchases.In the best available technology case, electricitygeneration in buildings in 2030 is 10.9 billionkilowatthours (90 percent) higher than in the referencecase, with solar systems responsible for 96 percentof the increase relative to the reference case. Theoptimistic assumptions of the best technology caseaffect solar PV systems more than fuel cells andmicroturbines, because there are no fuel expenses forsolar systems. In the 2005 technology case, assumingno further technological progress or cost reductionsafter 2005, electricity generation in buildings in 2030is 2.8 billion kilowatthours (23 percent) lower than inthe reference case.Some of the heat produced by fossil-fuel-fired generatingsystems may be used to satisfy heating requirements,increasing system efficiency and enhancingthe attractiveness of the advanced technologies. Onthe other hand, the additional natural gas use for fuelcells and microturbines in the high technology andbest technology cases offsets some of the reductionsin energy costs that result from improvements inbuilding shells and end-use equipment. In addition,the prospect of high natural gas prices may slow orlimit their adoption.For the U.S. industrial sector, delivered energyconsumption in 2004 was approximately the same asin 1980, although GDP more than doubled and thevalue of shipments in the industrial sector was 50 percenthigher. Thus, aggregate industrial energy intensity,measured as industrial delivered energy perdollar of GDP, declined by 3.0 percent per year, andindustrial delivered energy per dollar of industrialvalue of shipments declined by 1.6 percent per yearfrom 1980 to 2004 (Figure 44). Factors contributingto the decline in industrial energy intensity includeda greater focus on energy efficiency after the energyprice shocks of the 1970s and 1980s and a reduction inthe share of manufacturing activity accounted for bythe most energy-intensive industries. As the economyevolved, a larger portion of the nation’s output wasprovided by the services sector and a smaller portionby the industrial sector.In the AEO2006 reference case, these trends continueat a slower pace through 2030. Industrial energy useper dollar of GDP declines by 2.1 percent per year onaverage from 2004 through 2030, and energy use perdollar of industrial value of shipments declines by 1.2percent per year. The rates of decline in industrialenergy intensity are less rapid than those from 1980to 2004, in part because the nonmanufacturing portionof industrial value of shipments (agriculture,mining, and construction) grows more slowly thanthe manufacturing portion, which includes the moreenergy-intensive manufacturing sectors.Energy Information Administration / Annual Energy Outlook 2006 71

Industrial Sector Energy DemandMost Energy-Intensive IndustriesAre in the Manufacturing SectorFigure 45. Energy intensity in the industrial sector,2004 (thousand Btu per 2000 dollar of shipment)NonmanufacturingAgricultureMiningConstructionManufacturingEnergy-intensiveFoodPaperBulk chemicalsGlassCementIron and steelAluminumPetroleum refineriesNon-energy-intensiveMetal-based durablesOther manufacturing0 10 20 30 40 50 60In the industrial sector, the manufacturing subsectoris more energy-intensive than the nonmanufacturingsubsector (Figure 45), using about 50 percent moreenergy per dollar of output in 2004 [85]. From 1985 to2004, energy intensity declined more rapidly fornonmanufacturing industries than for manufacturing,primarily because most of the historical reductionin energy intensity for the manufacturingsubsector had already occurred by 1985 in response tothe high energy prices of the late 1970s and early1980s. In the nonmanufacturing subsector, much ofthe decline in energy intensity from 1985 to 2004resulted from a compositional shift: the relativelylow-intensity construction industry grew more rapidly,particularly in the late 1990s and early 2000s,than the relatively high-intensity mining sector.From 2004 levels, energy intensity in the manufacturingsubsector, based on value of shipments,declines in the reference case at an average annualrate of 1.2 percent through 2030, compared with anaverage decline of 1.0 percent per year in thenonmanufacturing subsector. The improvement inaggregate energy intensity for the manufacturingsubsector is accelerated by a compositional shift. In2004, the energy-intensive group of manufacturingindustries accounted for 21 percent of industrialvalue of shipments and the non-energy-intensivegroup 54 percent. With more rapid output growth inthe non-energy-intensive group from 2004 to 2030,the 2030 shares are 17 percent and 61 percent,respectively.Energy-Intensive Industries GrowLess Rapidly Than Industrial AverageFigure 46. Average growth in industrial output anddelivered energy consumption by sector, 2004-2030(percent per year)NonmanufacturingAgricultureOutputMiningConstructionManufacturingEnergy useEnergy-IntensiveFoodPaperBulk chemicalsGlassCementIron and steelAluminumPetroleum refineriesNon-energy-intensiveMetal-based durablesOther manufacturing-1.0 0.0 1.0 2.0 3.0 4.0The shift in value of shipments from the industrialsector to the service sectors seen in recent decadescontinues in the AEO2006 reference case. Totalindustrial sector value of shipments grows by 2.1 percentper year on average from 2004 through 2030,while GDP grows by 3.0 percent per year. Among theindustrial subsectors, average annual growth ratesrange from 3.0 percent for metal-based durables to0.5 percent for bulk chemicals (Figure 46).The energy-intensive manufacturing subsector accountedfor nearly two-thirds of industrial deliveredenergy consumption in 2004. From 2004 through2030, the value of shipments for the energy-intensivesubsector grows by an average of 1.3 percent per year,while the non-energy-intensive subsector grows atabout twice that rate. The energy-intensive industriesmaintain their 2004 share of industrial energyconsumption in 2030. With the growth of coal-toliquidsproduction in the refining sector, the bulkchemicals industry accounts for a smaller share of totalindustrial energy use in 2030 than it did in 2004;however, it remains the largest industrial energy consumer,accounting for nearly 25 percent of total industrialenergy consumption in 2030. Together, thepaper, bulk chemicals, and petroleum refining subsectorsaccount for more than 50 percent of all the energyconsumed in the industrial sector.Nonfuel use of energy in the industrial sector is concentratedin the bulk chemicals and constructionindustries. More than 60 percent of the energy consumedin the bulk chemicals industry is in the form offeedstock, and asphalt accounts for more than 50 percentof the energy consumed in construction.72 Energy Information Administration / Annual Energy Outlook 2006

Industrial Sector Energy DemandEnergy Intensity Declines in MostIndustrial SubsectorsFigure 47. Projected energy intensity in 2030relative to 2004, by industry (percent of 2004 value)NonmanufacturingAgricultureMiningConstructionManufacturingEnergy-IntensiveFoodPaperBulk chemicalsGlassCementIron and steelAluminumPetroleum refineriesNon-energy-intensiveMetal-based durablesOther manufacturing0.0 0.5 1.0 1.5Within the industrial sector, the energy intensity ofdifferent industries (Figure 47) can change for a varietyof reasons. For example, there may be a change inthe types of activities or the methods used in a givensubsector, or more energy-efficient new capacity maybe installed to accommodate growth or replaceworn-out machinery.For each industry with a relatively rapid projectedchange in energy intensity from 2004 to 2030, there isa unique explanation. In the steel industry, most newcapacity is expected to use electric arc furnace technology,which has a lower energy intensity than theolder blast furnace/basic oxygen furnace technology.Thus, the average energy intensity for the iron andsteel industry declines by 21 percent overall from2004 to 2030. In the U.S. aluminum industry, no newprimary smelting capacity is expected to be constructed,and secondary smelting, a less energyintensiveprocess of melting scrap, is expected tobecome the dominant technology. As a result, theenergy intensity of the aluminum industry in 2030 isnearly one-third less than in 2004. In the metal-baseddurables industry, a robust growth projection, withoutput 114 percent greater in 2030 than in 2004,requires substantial amounts of new, more energyefficientcapital stock to meet demand for theindustry’s output. In the petroleum refining industry,coal consumption increases by 1.6 quadrillion Btufrom 2004 to 2030, as CTL production grows. Consequently,its energy intensity increases over time.Alternative Technology Cases ShowRange of Industrial Efficiency GainsFigure 48. Variation from reference case deliveredindustrial energy use in two alternative cases,2004-2030 (quadrillion Btu)3.02.01.00.0-1.0-2.0-3.02004 2010 2015 2020 2025 20302005 technologyReferenceHigh technologyThe technology cases for the industrial sector representalternative views of the evolution and adoptionof energy-reducing technologies. In some sectors,energy-reducing technologies make significant contributionsto lower energy intensity. For example,energy intensity in mining would increase if therewere more widespread adoption of technologies toproduce fuels from oil shale. In other subsectors, suchas glass, technologies or techniques that tend toimprove output quality have the ancillary effectof reducing energy consumption. Generally, themanufacturing sector has more potential than thenonmanufacturing sector for the adoption of energyreducingtechnologies.In the high technology case, increased biomassrecovery and additions of CHP capacity offset process-relatedreductions in on-site energy consumption.Industrial cogeneration capacity increases morerapidly in the high technology case (3.4 percent peryear) than in the reference case (3.1 percent per year)[86]. Still, total industrial delivered energy consumptionin 2030 is 2 quadrillion Btu less than in the referencecase for the same level of output, and industrialenergy intensity improves by 1.4 percent per year onaverage, compared with 1.2 percent in the referencecase (Figure 48).In the 2005 technology case, industrial deliveredenergy use in 2030 is 2.4 quadrillion Btu higher thanin the reference case. Although the energy efficiencyof new equipment remains at its 2005 level in thiscase, average efficiency improves as old equipment isretired, and aggregate industrial energy intensityimproves by 0.9 percent per year.Energy Information Administration / Annual Energy Outlook 2006 73

Transportation Sector Energy DemandTransportation Energy Use Per Capitain 2030 Is 15 Percent Over 2004 LevelFigure 49. Transportation energy use per capita,1980-2030 (index, 2004 = 1)Travel Demand Is Projected To Growfor All Modes of TransportationFigure 50. Transportation travel demand by mode,1980-2030 (index, 2004 = 1)1.21.0OtherTotalLight-duty2.01.5Heavy trucksAircraftLight-duty0.80.61.00.40.2HistoryProjections0.01980 1995 2004 2015 2030Total delivered energy consumption for transportationin the AEO2006 reference case grows from 27.8quadrillion Btu in 2004 to 39.7 quadrillion Btu in2030, an increase of 43 percent. On a per-capita basis,however, the corresponding increase is only 15 percent(Figure 49). By mode, the most rapid increasesare in demand for freight movement and air travel.Energy use for freight trucks increases by 61 percentfrom 2004 to 2030, followed by increases of 47 percentfor aircraft and 42 percent for light-duty vehicles.The increase in diesel fuel consumption by heavyfreight trucks averages 1.9 percent annually from2004 through 2030, primarily as a result of growth inindustrial output that averages 2.1 percent per year.Economic growth is the primary reason for theincrease in demand for air travel, which results in a1.5-percent average annual increase in jet fuel use.Demand for light-duty vehicle fuels increases from16.2 quadrillion Btu in 2004 to 23.0 quadrillion Btu in2030. Although the vast majority of light-duty fueluse continues to be gasoline, diesel fuel consumptiongrows from 1.6 percent of total light-duty vehicle fuelconsumption in 2004 to 5.2 percent in 2030. Transportationdemand for alternative fuels, mostly ethanolused in gasoline blending and liquefied petroleumgas (LPG), increases from 1.9 percent of total transportationenergy use in 2004 to 5.9 percent in 2030.0.5HistoryProjections0.01980 1995 2004 2015 2030From 2004 to 2030, demand for transportation servicesincreases for all modes of travel (Figure 50).Light-duty vehicle travel grows by 1.8 percent annuallythrough 2030, significantly slower than the averageof 2.9 percent per year over the past 3 decades.Approximately 50 percent of the growth in light-dutyvehicle travel can be attributed to growth in the drivingage population, which increases by 0.9 percentannually. Higher fuel prices through 2008 slow thegrowth in demand for light-duty vehicle travel, but asfuel prices stabilize and per capita disposable incomerises, there is a more rapid increase in travel demand.Historically, freight truck travel has grown by 3.0percent annually. In the reference case, its growthaverages 2.3 percent per year from 2004 through2030. Although output grows in many manufacturingsectors, most of the future increase in demand forfreight movement is tied to increased output from theelectronics, food, plastics, furniture, and miscellaneoussectors.Demand for air travel increases by 1.8 percent annuallyfrom 2004 through 2030, down from its historicalannual growth rate of 3.3 percent. The airline industryis expected to recover from current financial conditionsand experience a strong recovery periodthrough 2011, when growth slows as congested conditionsbegin to affect the market. By 2019, severe constraintsassociated with available infrastructuremake airport capacity expansion a requirement forincreasing air travel.74 Energy Information Administration / Annual Energy Outlook 2006

Transportation Sector Energy DemandNew Technologies Promise ImprovedFuel Economy for Light-Duty VehiclesFigure 51. Average fuel economy of new light-dutyvehicles, 1980-2030 (miles per gallon)3530252015105HistoryProjections01980 1995 2004 2015 2030CarsAll light-dutyvehiclesLight trucksThe average fuel economy of new light-duty vehicles,which peaked at 26.2 miles per gallon in 1987,declined to 24.9 miles per gallon in 2004 (Figure 51).The downward trend in light-duty vehicle fuel economyresulted from a rapid increase in sales of lighttrucks (sport utility vehicles, minivans, and pickups),which were required to meet a CAFE standard of 20.7miles per gallon, compared with 27.5 miles per gallonfor cars. In April 2003, NHTSA increased the CAFEstandards for light trucks to 21 miles per gallon formodel year 2005, 21.6 miles per gallon for model year2006, and 22.2 miles per gallon for 2007, and morerecently the agency has proposed a restructuring oflight truck standards, with additional increases infuel efficiency standards for model years 2008through 2011. AEO2006 assumes no changes in currentlypromulgated fuel efficiency standards for carsand light trucks.Reversing the historic trend, the average fuel economyof new light-duty vehicles increases in the referencecase as a result of advances in fuel-savingtechnologies. Fuel economy for new light-duty vehiclesis 29.2 miles per gallon in 2030. Although higherpersonal incomes are expected to increase demand forlarger, more powerful vehicles, and the averagehorsepower for new cars is 27 percent above the 2004average in 2030, advanced technologies and materialspermit increases in performance and size of newvehicles without sacrificing improvements in fueleconomy.Advanced Technologies Are ProjectedTo Exceed 25 Percent of Sales by 2030Figure 52. Sales of advanced technology light-dutyvehicles by fuel type, 2015 and 2030(thousand vehicles sold)2,500 HybridsAlcohol2,000Turbo direct injection dieselGaseousElectric1,500Fuel cell1,00050002015 2030Sales of advanced technology vehicles, representingautomotive technologies that use alternative fuels orrequire advanced engine technology, reach 5.7 millionper year (Figure 52) and make up more than 25 percentof total light-duty vehicle sales in 2030. Hybridelectric vehicles (including those specifically designedto use electric motors and batteries in combinationwith a combustion engine to drive the vehicle andthose incorporating only an integrated starter generatorfor fuel economy) are anticipated to sell well,with 1.1 million units sold in 2015, increasing to 2.4million units in 2030. Sales of turbo direct injectiondiesel vehicles increase to 638,500 units in 2015 and1.7 million units in 2030. Sales of alcohol flexiblefueledvehicles continue to increase, with 1.3 millionsold in 2030.About 40 percent of advanced technology sales are asa result of Federal and State mandates for fuel economystandards, emissions programs, or other energyregulations. Currently, manufacturers selling alcoholflexible-fueled vehicles receive fuel economy creditsthat count toward compliance with CAFE regulations.In the AEO2006 reference case, the majority ofgasoline hybrid, electric, and fuel cell vehicle salesresult from compliance with low-emission vehicleprograms in California, Connecticut, Maine, Massachusetts,New Jersey, New York, Rhode Island,Washington, and Vermont. AEO2006 does notinclude the impacts of California Assembly Bill 1493,which effectively sets carbon emission standards forlight-duty vehicles, because of uncertainty about theState’s ability to enforce the standards.Energy Information Administration / Annual Energy Outlook 2006 75

Transportation Sector Energy DemandVehicle Technology Advances ReduceTransportation Energy DemandFigure 53. Changes in projected transportationfuel use in two alternative cases, 2010 and 2030(percent change from reference case)302010High technology2005 technology2010 2030Technology Assumptions IncludeImprovements in Vehicle EfficiencyFigure 54. Changes in projected transportationfuel efficiency in two alternative cases,2010 and 2030 (percent change from reference case)40202010 2030High technology2005 technology0-100-20-30LightdutyvehiclesFreighttrucksAircraftLightdutyvehiclesFreighttrucksAircraft-20LightdutyvehiclesFreighttrucksAircraftLightdutyvehiclesFreighttrucksAircraftIn the AEO2006 reference case, delivered energy usein the transportation sector increases from 27.8quadrillion Btu in 2004 to 39.7 quadrillion Btu in2030. In the high technology case, the projection for2030 is 2.8 quadrillion Btu (7.1 percent) lower, withabout 54 percent (1.5 quadrillion Btu) of the differenceattributed to efficiency improvements in lightdutyvehicles (Figure 53) as a result of increasedpenetration of advanced technologies, includingvariable valve lift, electrically driven power steeringpumps, and advanced electronic transmission controls.Similarly, projected fuel use by heavy freighttrucks in 2030 is 0.1 quadrillion Btu (0.9 percent)lower in the high technology case than in the referencecase, and advanced aircraft technologies reducefuel use for air travel by 1.0 quadrillion Btu (23.7 percent)in 2030.In the 2005 technology case, with new technology efficienciesfixed at 2005 levels, efficiency improvementscan result only from stock turnover. As a result, totaldelivered energy demand for transportation in 2030 is3.7 quadrillion Btu (9.2 percent) higher in 2030 in the2005 technology case than in the reference case. Fueluse for air travel in 2030 is 1.0 quadrillion Btu (23.4percent) higher in the 2005 technology case than inthe reference case, and freight trucks use 0.9 quadrillionBtu (11.9 percent) more fuel in 2030 [87].The high technology case assumes lower costs andhigher efficiencies for new transportation technologies.Advances in conventional technologies increasethe average fuel economy of new light-duty vehicles in2030 from 29.2 miles per gallon in the reference caseto 32.1 miles per gallon in the high technology case.The average efficiency of the light-duty vehicle stockis 20.6 miles per gallon in 2010 and 24.4 miles per gallonin 2030 in the high technology case, comparedwith 20.4 miles per gallon in 2010 and 22.5 miles pergallon in 2030 in the reference case (Figure 54).For freight trucks, average stock efficiency in thehigh technology case is 0.6 percent higher in 2010 and1.1 percent higher in 2030 than the reference caseprojection of 6.8 miles per gallon. Advanced aircrafttechnologies increase aircraft efficiency by 9.3 percentin 2010 and 31.0 percent in 2030 relative to thereference case projections.In the 2005 technology case, the average fuel economyof new light-duty vehicles is 26.2 miles per gallon in2030, and the average for the entire light-duty vehiclestock is 20.7 miles per gallon in 2030. For freighttrucks, the average stock efficiency in 2030 is 6.1miles per gallon. Aircraft efficiency in 2030 averages61.6 seat-miles per gallon in the 2005 technology case,compared with 76.0 seat-miles per gallon in the referencecase.76 Energy Information Administration / Annual Energy Outlook 2006

Electricity Demand and SupplyContinued Growth in Electricity UseIs Expected in All SectorsFigure 55. Annual electricity sales by sector,1980-2030 (billion kilowatthours)2,5002,0001,5001,000HistoryProjectionsCommercialResidentialIndustrialEarly Capacity Additions Use NaturalGas, Coal Plants Are Added LaterFigure 56. Electricity generation capacity additionsby fuel type, including combined heat and power,2005-2030 (gigawatts)120 Natural gasCoalRenewables80Nuclear4050001980 1995 2004 2015 203002005-20102011-20152016-20202021-20252026-2030Total electricity sales increase by 50 percent in theAEO2006 reference case, from 3,567 billion kilowatthoursin 2004 to 5,341 billion kilowatthours in 2030(Figure 55). The largest increase is in the commercialsector, as service industries continue to drive economicgrowth. By customer sector, electricity demandgrows by 75 percent from 2004 to 2030 in thecommercial sector, by 47 percent in the residentialsector, and by 24 percent in the industrial sector.Efficiency gains are expected in both the residentialand commercial sectors as a result of new standardsin EPACT2005 and higher energy prices that promptmore investment in energy-efficient equipment. Inthe residential sector, the increase in electricitydemand that results from a trend toward houses withmore floorspace, in addition to population shifts towarmer regions, is mitigated by an increase in theefficiency of air conditioners and refrigerators. In thecommercial sector, increases in demand as a consequenceof larger building sizes and more intensive useof electrical equipment is offset by increases in theefficiency of heating, cooling, lighting, refrigeration,and cooking appliances.Personal computers become more energy-efficient onaverage as residents and companies replace monitorsthat use cathode ray tubes with new models thatuse more efficient flat screens. New telecommunicationstechnologies and medical imaging equipmentincrease electricity demand in the “other” commercialend-use category, which accounts for one-half ofthe increase in commercial demand. In the industrialsector, increases in electricity sales are offset by rapidgrowth in on-site generation.With growing electricity demand and the retirementof 65 gigawatts of inefficient, older generating capacity,347 gigawatts of new capacity (including end-useCHP) will be needed by 2030. Most retirements areexpected to be oil- and natural-gas-fired steam capacity,along with smaller amounts of oil- and natural-gas-firedcombustion turbines and coal-firedcapacity, which are not cost-competitive with newerplants.Capacity decisions depend on the costs and operatingefficiencies of different options, fuel prices, and theavailability of Federal tax credits for investments insome technologies. Natural gas plants are generallythe least expensive capacity to build but are characterizedby comparatively high fuel costs. Coal,nuclear, and renewable plants are typically expensiveto build but have relatively low operating costs and, inaddition, receive tax credits under EPACT2005.Coal-fired and natural-gas-fired plants account forabout 50 percent and 40 percent, respectively, of thecapacity additions from 2004 to 2030 (Figure 56).Coal-fired capacity is generally more economical tooperate than natural-gas-fired capacity, because coalprices are considerably lower than natural gas prices.As a result, new natural-gas-fired plants are built toensure reliability and operate for comparatively fewhours when electricity demand is high.About 8 percent of the expected capacity expansionconsists of renewable generating units. New nuclearcapacity additions total 6 gigawatts, but no additionalnew nuclear plants are built after 2020, when theEPACT2005 production tax credit expires.Energy Information Administration / Annual Energy Outlook 2006 77

Electricity SupplyCapacity Additions Are ExpectedTo Be Required in All RegionsFigure 57. Electricity generation capacityadditions, including combined heat and power,by region and fuel, 2005-2030 (gigawatts)ECARERCOTMAACMAINMAPPNYNEFLSERCSPPNWPRACA0 20 40 60 80Natural gasNuclearRenewable/otherCoalMost areas of the United States currently haveexcess generation capacity, but all electricity demandregions (see Appendix F for definitions) are expectedto need additional, currently unplanned, capacity by2030 (Figure 57). The largest amounts of new capacityare expected in the Southeast (FL and SERC) andthe West (NWP, RA, and CA). In the Southeast, electricitydemand represents a relatively large share oftotal U.S. electricity sales, and its need for new capacityis greater than in other regions.With natural gas prices rising in the reference case,coal-fired plants make up most of the capacity additionsthrough 2030. The largest concentrations ofnew coal-fired plants are in the Southeast and theWest. In the Southeast, new coal-fired plants are builtin view of the size of the electricity market and thecorresponding need for additional capacity. In theWest, where the capacity requirement is muchsmaller, the choice to build mostly coal-fired plants isbased on the region’s lower-than-average coal pricesand higher-than-average natural gas prices.Nationwide, some new natural-gas-fired plants arebuilt to maintain a diverse capacity mix or to serve asreserve capacity. Most are located in the Midwest(MAPP, MAIN, and ECAR) and South (ERCOT andSPP). The Midwest has a surplus of coal-fired generatingcapacity and does not need to add many newcoal-fired plants. In the South, natural gas prices arelower than the national average, and natural-gasfiredplants are more economical than in otherregions.Least Expensive Technology OptionsAre Likely Choices for New CapacityFigure 58. Levelized electricity costs for new plants,2015 and 2030 (2004 mills per kilowatthour)755025Incremental transmission costsVariable costs, including fuelFixed costsCapital costs020152030Coal GascombinedcycleWind Nuclear Coal GascombinedcycleWind NuclearTechnology choices for new generating capacity aremade to minimize cost while meeting local and Federalemissions constraints. The choice of technologyfor capacity additions is based on the least expensiveoption available (Figure 58) [88]. The reference caseassumes a capital recovery period of 20 years. In addition,the cost of capital is based on competitive marketrates, to account for the risks of siting new units.Capital costs decline over time (Table 16), at ratesthat depend on the current stage of development foreach technology. For the newest technologies, capitalcosts are initially adjusted upward to reflect the optimisminherent in early estimates of project costs. Asproject developers gain experience, the costs areassumed to decline. The decline continues at a progressivelyslower rate as more units are built. Theefficiency of new plants is also assumed to improvethrough 2015, with heat rates for advanced combinedcycle and coal gasification units declining to 6,333 and7,200 Btu per kilowatthour, respectively.Table 16. Costs of producing electricityfrom new plants, 2015 and 2030Advancedcoal2015 2030AdvancedcombinedcycleAdvancedcoalAdvancedcombinedcycleCosts2004 mills per kilowatthourCapital 30.34 11.33 27.78 10.76Fixed 4.73 1.40 4.73 1.40Variable 14.58 36.97 15.82 40.18Incrementaltransmission 3.47 2.88 3.40 2.94Total 53.12 52.58 51.73 55.2878 Energy Information Administration / Annual Energy Outlook 2006

Electricity SupplyEPACT2005 Tax Credits Are ExpectedTo Stimulate New Nuclear BuildsFigure 59. Electricity generation from nuclearpower, 1973-2030 (billion kilowatthours)1,000State Programs Support RenewableGenerating Capacity AdditionsFigure 60. Additions of renewable generatingcapacity, 2004-2030 (gigawatts)148001210UnplannedOther planned600840064State mandates20020HistoryProjections1970 1985 1995 2004 2015 20300HydropowerBiomass GeothermalLandfillgasSolarWindIn the AEO2006 reference case, nuclear capacityincreases from 99.6 gigawatts in 2004 to 108.8 gigawattsin 2030. The increase includes 6.0 gigawatts ofcapacity at new plants stimulated by EPACT2005 taxincentives and 3.2 gigawatts of capacity expansion atexisting plants. EPACT2005 provides an 8-year productiontax credit of 1.8 cents per kilowatthour for upto 6 gigawatts of capacity built before 2021. If thecapacity limit is reached before 2020, the credit programends, and no additional units are expected. Theincrease in capacity at existing units assumes that alluprates approved, pending, or expected by the NRCwill be carried out.All existing nuclear plants are expected to continueoperating through 2030, although most will bebeyond their original license expiration dates by then.As of September 2005, the NRC had approved licenserenewals for 35 nuclear units, and 14 other applicationswere being reviewed. As many as 28 additionalapplicants have announced intentions to pursuelicense renewals over the next 7 years, indicating astrong interest in maintaining the existing stock ofnuclear plants.Because of the increase in capacity, from new capacityand uprates, and the continued strong performanceof existing units, nuclear generation grows from 789billion kilowatthours in 2004 to 871 billion kilowatthoursin 2030 (Figure 59). That increase is not sufficient,however, for nuclear power to maintain itscurrent 20-percent share of total generation. In 2030,even with a national average capacity factor of morethan 90 percent, nuclear power accounts for about 15percent of total U.S. generation.From 2004 to 2030, 26.4 gigawatts of new renewablegenerating capacity is added in the reference case,including 21.9 gigawatts in the electric power sectorand 4.5 gigawatts in the end-use sectors. Nearlyone-half of the total (11.7 gigawatts in the electricpower sector and 0.75 in the end-use sectors) is atleast partially stimulated by State programs, with theremainder resulting from commercial projects.Overall, 9.0 gigawatts of central-station capacity, primarilyin near-term projects, results from specificState standards: 3.7 gigawatts in Texas, 3.4 in California,0.9 in Nevada, and 0.5 in Minnesota. ThreeStates—Montana, New Mexico, and New York—add100 to 200 megawatts each. Ten States—Arizona,Colorado, Hawaii, Illinois, Massachusetts, Maine,New Jersey, Pennsylvania, Vermont, and Wisconsin—addless than 100 megawatts each. SeveralStates without specific requirements also add newrenewable capacity, including nearly 400 megawattsin Washington, 300 in Oklahoma, 200 in Iowa, 150 inKansas, and smaller amounts elsewhere.The combination of Federal production tax creditsand State programs results primarily in new windpower. More than 93 percent of the capacity additionsstimulated by State programs are wind plants (Figure60). State programs also spur small amounts of PVand solar thermal capacity, totaling 180 megawatts.On the other hand, with the Federal production taxcredit assumed to expire on December 31, 2007, itspotential to trigger capacity additions using technologieswith longer lead times, such as biomass, geothermal,and hydropower, is limited.Energy Information Administration / Annual Energy Outlook 2006 79

Electricity SupplyRenewables Are Expected To BecomeMore Competitive Over TimeFigure 61. Levelized and avoided costs for newrenewable plants in the Northwest, 2015 and 2030(2004 mills per kilowatthour)Natural Gas and Coal Meet MostNeeds for New Electricity SupplyFigure 62. Electricity generation by fuel,2004 and 2030 (billion kilowatthours)4,0002015GeothermalSolar thermalBiomassLevelized costWindAvoided cost2030GeothermalSolar thermalBiomassWind3,0002,0001,000200420300 25 50 75 100 125 1500Coal Nuclear NaturalgasRenewablesPetroleumThe competitiveness of both conventional and renewablegeneration resources is based on the most costeffectivemix of capacity that satisfies the demand forelectricity across all hours and seasons. Baseloadtechnologies tend to have low operating costs and setthe market price for power only during the hours ofleast demand. Dispatchable geothermal and biomassresources compete directly with new coal and nuclearplants, which to a large extent determine the avoidedcost [89] for baseload energy (Figure 61). In someregions and years, new geothermal or biomass plantsmay be competitive with new coal-fired plants, buttheir development is limited by the availability of geothermalresources or competitive biomass fuels.Wind and solar are intermittent technologies that canbe used only when resources are available. With relativelylow operating costs and limited resource availability,their avoided costs are determined largely bythe operating costs of the most expensive units inoperation when their resources are available. Solargenerators tend to operate during peak load periods,when natural-gas-fired combustion turbines andcombined-cycle units with higher fuel costs tend todetermine the avoided cost. The levelized cost of solarthermal generation is significantly higher than itsavoided cost through 2030. The availability of windresources varies among regions, but wind plants generallytend to displace intermediate load generation.Thus, the avoided costs of wind power will be determinedlargely by the low-to-moderate operating costsof combined-cycle and coal-fired plants, which setpower prices during intermediate load hours. In someregions and years, the levelized costs for wind powerare below its avoided costs.Coal-fired power plants (including utilities, independentpower producers, and end-use CHP) continue tosupply most of the Nation’s electricity through 2030(Figure 62). Coal-fired plants accounted for 50 percentof all electricity generation in 2004, and theirshare increases to 57 percent in 2030. In the nearterm, coal use increases gradually as a result ofgreater utilization of existing facilities, but its shareof total generation remains relatively constant. In thelonger term, the share of coal-fired generationincreases as new plants begin to operate.Because of comparatively high fuel prices, naturalgas-firedplants are not used as intensively as coalfiredplants. Natural-gas-fired plants provided 18 percentof total supply in 2004, and their share declinesslightly to 17 percent in 2030. Natural-gas-fired generationincreases initially as the recent wave ofnewer, more efficient plants come online, but itdeclines toward the end of the forecast period as naturalgas prices continue to rise.Both nuclear and renewable generation increase asnew plants are built, stimulated by Federal tax incentivesand rising fossil fuel prices. Modest increases innuclear generation also result from improvements inplant performance and expansion of existing facilities,but the share of generation from nuclear plantsdeclines from 20 percent in 2004 to 15 percent in 2030as total generation grows at a faster rate than nucleargeneration. The share of generation from renewablecapacity increases slightly, to account for about 9 percentof total electricity supply in 2030.80 Energy Information Administration / Annual Energy Outlook 2006

Electricity From Renewable SourcesTechnology Advances, Tax ProvisionsIncrease Renewable GenerationFigure 63. Grid-connected electricity generationfrom renewable energy sources, 1980-2030(billion kilowatthours)400HistoryProjectionsBiomass, Wind, and GeothermalLead Growth in RenewablesFigure 64. Nonhydroelectric renewable electricitygeneration by energy source, 2004-2030(billion kilowatthours)300300200ConventionalhydropowerOtherrenewables250200150GeothermalSolarWind10010050Biomass01980 1995 2004 2015 203002004 2010 2020 2030MSW/LFGDespite technology improvements, rising fossil fuelcosts, and public support, the contribution of renewablefuels to U.S. electricity supply remains relativelysmall in the AEO2006 reference case at 9.4 percent oftotal generation in 2030, up from 9.0 percent in2004 (Figure 63). Although conventional hydropowerremains the largest source of renewable generationthrough 2030, a lack of untapped large-scale sites,coupled with environmental concerns, limits itsgrowth, and its share of total generation falls from 6.8percent in 2004 to 5.1 percent in 2030. Electricitygeneration from nonhydroelectric alternative fuelsincreases, however, bolstered by technology advancesand State and Federal supports. The share of nonhydropowerrenewables increases by 95 percent, from2.2 percent of total generation in 2004 to 4.3 percentin 2030.Biomass is the largest source of renewable electricitygeneration among the nonhydropower renewablefuels. Co-firing with coal is relatively inexpensivewhen low-cost biomass resources are available; and asthe cost of biomass increases over time, new dedicatedbiomass facilities, such as IGCC plants, are built.Electricity generation from biomass increases from0.9 percent of total generation in 2004 to 1.7 percentin 2030, with 38 percent of the increase comingfrom biomass co-firing, 36 percent from dedicatedpower plants, and 26 percent from new on-site CHPcapacity.Electricity generation from wind and geothermalenergy also increases in the reference case (Figure64). There is considerable uncertainty about thegrowth potential of wind power, which depends on avariety of factors, including fossil fuel costs, Staterenewable energy programs, technology improvements,access to transmission grids, public concernsabout environmental and other impacts, and thefuture of Federal production tax credits. In the referencecase, generation from wind power increasesfrom 0.4 percent of total generation in 2004 to 1.1 percentin 2030. Generation from geothermal facilitiesincreases from 0.4 percent of total generation in 2004to 0.9 percent in 2030, despite limited opportunitiesfor the development of new sites. Most of the suitablesites, restricted mainly to Nevada and California,involve relatively high up-front costs and performancerisks; and although geothermal power plantsare eligible for the Federal production tax creditthrough 2007, the long construction lead timesrequired make it unlikely that significant new capacitycould be built in time to benefit from the currentcredit.Among the other alternative fuel technologies, generationfrom municipal solid waste (MSW) and LFGslips from 0.5 percent of total generation in 2004 to0.5 percent in 2030. Solar technologies in generalremain too costly for grid-connected applications, butdemonstration programs and State policies supportsome growth in central-station solar PV, and smallscalecustomer-sited PV applications grow rapidly[90]. Grid-connected solar generation remains at lessthan 0.1 percent of total generation through 2030.Energy Information Administration / Annual Energy Outlook 2006 81

Electricity Fuel Costs and PricesFuel Costs Drop From Recent Highs,Then Increase GraduallyFigure 65. Fuel prices to electricity generators,1995-2030 (2004 dollars per million Btu)108642HistoryProjectionsPetroleumNatural gasCoalNuclear01995 2004 2010 2015 2020 2025 2030Electricity production costs are a function of the costsfor fuel, operations and maintenance, and capital. Inthe reference case, fuel costs account for about twothirdsof the generating costs for new natural-gasfiredplants, less than one-third for new coal-firedunits, and less than one-tenth for new nuclear powerplants in 2030. For most fuels, delivered prices to electricitygenerators peak by 2006, fall in the middleyears of the projections, and then increase steadilythrough 2030. As a yearly average, natural gas pricesdrop to $5.06 in 2016 and then rise to $6.26 permillion Btu in 2030 (Figure 65). Similarly, petroleumprices decline to $6.39 in 2013 and then rise to $7.61per million Btu in 2030. Coal prices remain relativelylow, with highs of about $1.50 per million Btu at thebeginning and end of the projection period and lows ofabout $1.40 in the middle years. Nuclear fuel costs,averaging $0.45 per million Btu in 2004, rise to $0.60per million Btu in 2030.Electricity generation from natural-gas-fired powerplants, which have relatively low capital costs andemissions levels, increased in the early years of thisdecade. More recently, higher fuel prices haveincreased the cost of natural-gas-fired generation.For example, the price of natural gas to generatorsjumped by 37 percent from 2004 to 2005. With naturalgas prices rising after 2016 in the reference case,the natural gas share of total electricity generationdrops, and both coal-fired and renewable generationincrease.Electricity Prices Moderate in theNear Term, Then Rise GraduallyFigure 66. Average U.S. retail electricity prices,1970-2030 (2004 cents per kilowatthour)12 History Projections1086421.7Average price(nominal cents)14.11970 203001970 1985 1995 2004 2015 2030Electricity prices are determined primarily by thecosts of generation, which make up about two-thirdsof the total retail price. The 2004-2005 spikes in naturalgas and petroleum prices, along with elevated coalprices, led to a jump in electricity prices. Averageretail prices (in 2004 dollars) fall to 7.1 cents per kilowatthourin 2015, as new sources of natural gas andcoal are brought on line. After 2015, natural gas andpetroleum prices rise steadily, and power producersincrease their reliance on lower priced coal. As aresult, retail electricity prices rise gradually, to 7.5cents per kilowatthour in 2030 (Figure 66).Customers in States with competitive retail marketsfor electricity see the effects of natural gas prices intheir electricity bills more rapidly than those in regulatedStates, because their prices are determined to agreater extent by the marginal cost of energy—theaverage operating cost of the last, most expensiveunit run each hour—rather than the average of allplant costs. Natural gas plants, with their higheroperating costs, often set the hourly marginal price.Distribution costs, which accounted for more thanone-quarter of retail electricity prices in 2004, declineby 14 percent from 2004 to 2030, as the cost of distributioninfrastructure is spread over a growing customerbase, and technology improvements lower thecosts of billing, metering, and call-center services.Because of the additional investment needed to meetconsumers’ growing electricity use and to facilitatecompetitive wholesale energy markets, transmissioncosts rise by 27 percent, increasing their share of thetotal electricity price from 7 percent to 9 percent.82 Energy Information Administration / Annual Energy Outlook 2006

Electricity Alternative CasesFaster Economic Growth StimulatesCapacity Additions, Particularly CoalFigure 67. Cumulative new generating capacityby technology type in three economic growth cases,2004-2030 (gigawatts)300 Natural gasCoal steam250RenewablesNuclear200150100500Low growth Reference High growthThe need for new generating capacity, particularlycoal-fired capacity, is influenced by economic growth.From 2004 to 2030, average annual GDP growthranges from 2.4 percent in the low economic growthcase to 3.5 percent in the high economic growth case.The difference leads to a 21-percent variation in thelevel of electricity demand in 2030 between the lowand high economic growth cases, with a correspondingdifference of 215 gigawatts in the amount of newcapacity added from 2004 through 2030, includingCHP in the end-use sectors.Most (65 percent) of the capacity added in the higheconomic growth case, relative to the reference case,consists of new coal-fired plants. Higher demand forelectricity and lower interest rates in the high economicgrowth case make new coal plants attractive.The stronger demand growth assumed in the highgrowth case also stimulates additions of renewableplants and, to a lesser degree, new natural-gas-firedcapacity (Figure 67). In the low economic growthcase, total capacity additions are reduced by 104gigawatts, and 73 percent of that projected reductionis in coal-fired capacity additions.Average electricity prices in 2030 are 4 percent higherin the high economic growth case than in the referencecase, due to higher natural gas prices and thecosts of building additional generating capacity. Electricityprices in 2030 are 4 percent lower in the loweconomic growth case than in the reference case. Inthe high economic growth case, a 9-percent increasein consumption of fossil fuels results in a 10-percentincrease in CO 2emissions from electricity generatorsin 2030.Natural-Gas-Fired Capacity AdditionsVary With Cost and PerformanceFigure 68. Cumulative new generating capacity bytechnology type in three fossil fuel technology cases,2004-2030 (gigawatts)250 CoalAdvanced coal200Natural gasAdvanced gasRenewables150Nuclear100500Low fossil Reference High fossilThe cost and performance characteristics for variousfossil fuel generating technologies in the AEO2006reference case are determined in consultation withindustry and government specialists. To test the significanceof uncertainty in the assumptions, alternativecases vary key parameters. In the high fossil fuelcase, capital costs, heat rates, and operating costs foradvanced fossil-fired generating technologies in 2030are assumed to be 10 percent lower than in the referencecase. The low fossil fuel case assumes no changein capital costs and heat rates for advanced technologiesfrom their 2006 levels.With different cost and performance assumptions,the mix of generating technologies changes (Figure68). In the high fossil case, natural gas technologiesmake up the largest share of new capacity additions;in the reference and low fossil cases, coal technologiesaccount for most of the new capacity additions. In thehigh fossil case, advanced technologies are used for 79percent of all natural-gas-fired capacity additions and71 percent of all coal-fired capacity additions by 2030;in the low fossil case, advanced technologies are usedfor only 54 percent of natural-gas-fired capacity additionsand a negligible percentage of coal-fired capacityadditions, but almost 10 gigawatts of nuclear generatingcapacity is added by 2030. The average efficiencyof fossil-fuel-fired power plants varies only slightlyamong the three cases—from 36 percent in the lowfossil case to almost 38 percent in the high fossil casein 2030—because plant owners are not expected toupgrade the large base of older generating units.Energy Information Administration / Annual Energy Outlook 2006 83

Electricity Alternative CasesNew Nuclear Plants Are CompetitiveWhen Lower Costs Are AssumedFigure 69. Levelized electricity costs for new plantsby fuel type in two nuclear cost cases, 2015 and 2030(2004 cents per kilowatthour)108642015Incremental transmission costs2030 Variable costs, including fuelFixed costsCapital costsLower Cost Assumptions IncreaseBiomass and Geothermal CapacityFigure 70. Nonhydroelectric renewable electricitygeneration by energy source in three cases,2010 and 2030 (billion kilowatthours)400 High renewablesLow renewables300200ReferenceGeothermalSolarWind20CoalNatural gascombinedcycleNuclear:referencecaseNuclear: Nuclear:advanced vendorcost case estimate case10002010 2030BiomassMSW/LFGThe reference case assumptions for the cost andperformance characteristics of new technologies arebased on cost estimates by government and industryanalysts, allowing for uncertainties about newdesigns. Because no new nuclear plants have beenordered in this country since 1977, there is no reliableestimate of what they might cost. In recent years, variousnuclear vendors have argued that their newplants will be simpler and less costly. Two alternativenuclear cost cases address this uncertainty. Theadvanced nuclear cost case assumes capital and operatingcosts 20 percent below those in the referencecase in 2030, reflecting a 31-percent reduction inovernight capital costs from 2006 to 2030. The vendorestimate case assumes reductions relative to the referencecase of 18 percent initially and 44 percent in2030, consistent with estimates from British NuclearFuels Limited (Westinghouse) for the manufacture ofits AP1000 advanced pressurized-water reactor.Nuclear generating costs in the alternative nuclearcost cases are competitive with the generating costsfor new coal- and natural-gas-fired units toward theend of the projection period (Figure 69). (The figureshows average generating costs, assuming generationat the maximum capacity factor for each technology;the costs and relative competitiveness of the technologiescould vary by region.) In the reference case, Federaltax credits result in 6 gigawatts of new nuclearcapacity. In the advanced nuclear case 34 gigawatts ofnew nuclear capacity is added between 2004 and2030, and in the vendor estimate case 77 gigawatts ofnew nuclear capacity is added. The additional nuclearcapacity displaces primarily new coal-fired capacity.The impacts of key assumptions about the availabilityand cost of nonhydroelectric renewable energy resourcesfor electricity generation are shown in twoalternative technology cases. In the low renewablescase, the cost and performance of generators usingrenewable resources are assumed to remain unchangedthroughout the forecast. The high renewablescase assumes cost reductions of 10 percent in2030 on a site-specific basis for hydroelectric, geothermal,biomass, wind, and solar capacity.In the low renewables case, construction of newrenewable capacity is less than projected in the referencecase (Figure 70). In the high renewables case,more additions of biomass, geothermal, and windcapacity are projected through 2030 than in the referencecase, with most of the incremental capacityadded between 2020 and 2030.Biomass, available in some quantity in all U.S.regions, provides desirable baseload and intermediate-loadcapacity. In the high renewables case, thelargest increase in generation relative to the referencecase is seen for biomass, which nearly doubles in2030. Geothermal resources are limited to the West,and despite limited opportunities for expansion, generationincreases by 39 percent from 2004 to 2030.Generation from wind power, with significantexpansion in the reference case over current capacity,increases by a modest 15 percent over the referencecase in 2030, and there is little or no increase ingeneration from solar and hydropower. Even with theassumptions of reduced costs, nonhydropower renewablesaccount for less than 6 percent of total generationin 2030 in the high renewables case.84 Energy Information Administration / Annual Energy Outlook 2006

Natural Gas DemandIncreases in Natural Gas Use AreModerated by High PricesFigure 71. Natural gas consumption by sector,1990-2030 (trillion cubic feet)12HistoryProjectionsU.S. Natural Gas Consumption Growsthe Most East of the Mississippi RiverFigure 72. Increases in natural gas consumptionby Census division, 2004-2030 (percent per year)New England10864IndustrialElectricitygeneratorsResidentialCommercialMiddle AtlanticSouth AtlanticEast North CentralEast South CentralWest North CentralWest South Central201990 1995 2004 2010 2015 2020 2025 2030TransportationIn the AEO2006 reference case, total natural gas consumptionincreases from 22.4 trillion cubic feet in2004 to 26.9 trillion cubic feet in 2030. Most of theincrease is seen before 2017, when total U.S. naturalgas consumption reaches just under 26.5 trillion cubicfeet. After 2017, high natural gas prices limit consumptionto about 27 trillion cubic feet through 2030.Consequently, the natural gas share of total energyconsumption drops from 23 percent in 2004 to 21 percentin 2030.Currently, high natural gas prices discourage the constructionof new natural-gas-fired electricity generationplants. As a result, only 130 gigawatts of newnatural-gas-fired capacity is added from year-end2004 through 2030, as compared with 154 gigawattsof new coal-fired capacity. Natural gas consumptionin the electric power sector peaks at 7.5 trillion cubicfeet in 2019, then starts falling as new coal-fired electricitygeneration increasingly displaces natural-gasfiredgeneration. Natural gas use for electricity generationdeclines to 6.4 trillion cubic feet in 2030 (Figure71).With natural gas prices remaining relatively highthroughout the projection period, consumption ofnatural gas in the industrial sector gas grows slowly,from 8.5 trillion cubic feet in 2004 to 10.0 trillioncubic feet in 2030. Natural gas consumption increasesin all the major industrial sectors, with the exceptionof the refining industry. High prices also limit consumptionincreases in the buildings sector (residentialand commercial), where natural gas use growsfrom 7.9 trillion cubic feet in 2004 to 9.6 trillion cubicfeet in 2030.MountainPacific0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4From 2004 to 2030, 60 percent of the projectedgrowth in lower 48 end-use consumption of naturalgas occurs east of the Mississippi River (Figure 72).Variation in regional growth rates for natural gasconsumption results from different prospects for populationgrowth, economic activity, and natural-gas-firedelectricity generation. The most rapidincreases in natural gas consumption, averaging 1.3percent per year from 2004 through 2030, are in theSouth Atlantic and East South Central Census divisions.In the West North Central division, consumptiongrows by 1.0 percent per year, and growth ratesin the other Census divisions are less than that,including annual averages of 0.5 percent in New England,0.7 percent in the Middle Atlantic, 0.8 percent inthe East North Central, 0.5 percent in the West SouthCentral, 0.8 percent in the Mountain, and 0.3 percentin the Pacific divisions.The Rocky Mountain and Alaska regions providemost of the increase in domestic natural gas productionfrom 2004 to 2030. Because 60 percent of the projectedgrowth in natural gas consumption occurs eastof the Mississippi River, new natural gas pipelines arebuilt from supply regions in the West to meet naturalgas demand in the East, including a North SlopeAlaska pipeline. An exception is the construction ofnew pipeline capacity originating in the Rocky Mountainsto provide its increasing production to PacificCoast markets. Also, some additional pipeline constructionis expected to provide new LNG terminalsaccess to the major consumption markets and to linkdeepwater natural gas production to major onshoretransmission pipelines.Energy Information Administration / Annual Energy Outlook 2006 85

Natural Gas SupplyUnconventional Production Becomesthe Largest Source of U.S. Gas SupplyNet Imports of Natural Gas Growin the ProjectionsFigure 73. Natural gas production by source,1990-2030 (trillion cubic feet)108HistoryProjectionsLower 48 NAunconventionalFigure 74. Net U.S. imports of natural gasby source, 1990-2030 (trillion cubic feet)54HistoryProjectionsOverseasLNG642Lower 48 ADAlaska01990 1995 2004 2010 2015 2020 2025 2030Lower 48 NAconventionalonshoreLower 48 NAoffshoreA large proportion of the onshore lower 48 conventionalnatural gas resource base has been discovered.New reservoir discoveries are expected to be smallerand deeper, and thus more expensive and riskier todevelop and produce. Much of the onshore lower 48nonassociated (NA) conventional natural gas productionin the reference case comes from existing largefields, as lower 48 NA onshore conventional naturalgas production declines from 4.8 trillion cubic feet in2004 to 4.2 trillion cubic feet in 2030 (Figure 73). Productionof associated-dissolved (AD) natural gas fromlower 48 crude oil reserves also declines, from 2.4 trillioncubic feet in 2004 to 2.3 trillion cubic feet in 2030.Incremental production of lower 48 onshore naturalgas production comes primarily from unconventionalresources, including coalbed methane, tight sandstones,and gas shales. Lower 48 unconventional productionincreases from 7.5 trillion cubic feet in 2004to 9.5 trillion cubic feet in 2030.Considerable natural gas resources remain in the offshoreGulf of Mexico, especially in the deep waters.Deepwater natural gas production in the Gulf of Mexicoincreases from 1.8 trillion cubic feet in 2004 to apeak of 3.2 trillion cubic feet in 2014, then declines to2.1 trillion cubic feet in 2030. Production in the shallowerwaters of the Gulf of Mexico declines throughoutthe projection period, from 2.4 trillion cubic feetin 2004 to 1.8 trillion cubic feet in 2030.The Alaska pipeline begins transporting natural gasto the lower 48 States in 2015. In 2030, Alaska’s naturalgas production totals 2.1 trillion cubic feet.321Canada0Mexico-11990 1995 2004 2010 2015 2020 2025 2030With U.S. natural gas production declining, importsof natural gas rise to meet increasing domesticconsumption. A decline in Canada’s non-Arctic conventionalnatural gas production is only partiallyoffset by its Arctic and unconventional production.Although a MacKenzie Delta natural gas pipeline isexpected to begin transporting natural gas in 2011 inthe reference case, net imports from Canada fall from3.2 trillion cubic feet in 2004 to 1.5 trillion cubic feetin 2019. After 2019, net imports from Canada begin toincrease, as unconventional production eventuallyoffsets the decline in conventional production. Netimports of natural gas from Canada total 1.8 trillioncubic feet in 2030.Most of the projected growth in U.S. natural gasimports is in the form of LNG, some of which flowsinto the United States by pipeline from Mexico.The total capacity of U.S. LNG receiving terminalsincreases rapidly, from 1.4 trillion cubic feet in 2004to 4.9 trillion cubic feet in 2015, when net LNGimports total 3.1 trillion cubic feet. Construction ofnew LNG terminals slows after 2015. With terminalcapacity of 5.8 trillion cubic feet in 2030, U.S. netLNG imports total 4.4 trillion cubic feet (Figure 74).Net exports of U.S. natural gas to Mexico decreasefrom 2004 through 2011, as new LNG terminals arebuilt in Mexico. After 2011 U.S. net exports of naturalgas to Mexico increase, to 560 billion cubic feet in2030.86 Energy Information Administration / Annual Energy Outlook 2006

Natural Gas PricesProjected Natural Gas PricesRemain Above Historical LevelsFigure 75. Lower 48 natural gas wellhead prices,1990-2030 (2004 dollars per thousand cubic feet)10 History ProjectionsDelivered Natural Gas PricesFollow Trends in Wellhead PricesFigure 76. Natural gas prices by end-use sector,1990-2030 (2004 dollars per thousand cubic feet)14 History Projections864212108642ResidentialCNG vehiclesCommercialIndustrialElectricityWellhead01990 1995 2004 2010 2015 2020 2025 2030In the reference case, wellhead natural gas pricesdecline from current levels to an average of $4.46(2004 dollars) per thousand cubic feet in 2016, thenrise to $5.92 per thousand cubic feet in 2030 (Figure75). Current high prices for natural gas are expectedto accelerate the construction of new LNG terminalcapacity, resulting in a significant increase in totalU.S. LNG receiving capacity by 2015. High naturalgas prices are also expected to increase support forthe construction of an Alaska natural gas pipelinethat begins operations in 2015, and to stimulateproduction of unconventional natural gas. On thedemand side, high prices reduce the growth of naturalgas consumption.As a result of the development of new natural gassupplies and slower growth in consumption, wellheadnatural gas prices decline through 2016. After 2016,as the cost of developing the remaining U.S. naturalgas resource base increases, wellhead natural gasprices increase. World LNG prices also increase after2016 in the reference case, slowing the growth of U.S.LNG imports.01990 1995 2004 2010 2015 2020 2025 2030Trends in delivered natural gas prices largely reflectchanges in wellhead prices. In the reference case,prices for natural gas delivered to the end-use sectorsdecline through 2016 as wellhead prices decline, thenincrease along with wellhead prices (Figure 76).On average, end-use transmission and distributionmargins remain relatively constant, because the costof adding new facilities largely offsets the reduceddepreciation expenses of existing facilities. Transmissionand distribution margins in the end-use sectorsreflect both the volumes of natural gas delivered andthe infrastructure arrangements of the different sectors.The industrial and electricity generation sectorshave the lowest end-use prices, because they receivemost of their natural gas directly from interstatepipelines, avoiding local distribution charges. In addition,summer-peaking electricity generators reducetransmission costs by using interruptible transportationservices during the summer, when there is sparepipeline capacity. As power generators take a largershare of the natural gas market, however, they areexpected to rely more on higher cost firm transportationservice.The reference case assumes that sufficient transmissionand distribution capacity will be built to accommodatethe growth in natural gas consumption. Iffuture public opposition were to prevent the buildingof new infrastructure, delivered prices could behigher than projected in the reference case.Energy Information Administration / Annual Energy Outlook 2006 87

Natural Gas Alternative CasesTechnology Advances Could ModerateFuture Natural Gas PricesFigure 77. Lower 48 natural gas wellhead pricesin three technology cases, 1990-2030(2004 dollars per thousand cubic feet)10HistoryProjectionsNatural Gas Supply ProjectionsReflect Technological Progress RatesFigure 78. Natural gas production and net importsin three technology cases, 1990-2030(trillion cubic feet)25HistoryProjectionsRapid technology86Slow technologyReferenceRapid technology2015ProductionReferenceSlow technology41025Net importsSlow technologyReferenceRapid technology01990 1995 2004 2010 2020 2030Technological progress affects natural gas productionby reducing production costs and expanding the economicallyrecoverable gas resource base. An exampleis the relatively recent development of technologiesfor producing unconventional natural gas resources,which allow previously uneconomical deposits to beproduced profitably, whereas 50 years ago industrytechnology was capable of exploiting only conventionaldeposits.In the slow oil and gas technology case, advances inexploration and production technologies are assumedto be 50 percent slower than those assumed in thereference case, which are based on historical rates.As a result, domestic natural gas development costsare higher, production is lower, wellhead prices arehigher at $6.36 per thousand cubic feet in 2030(Figure 77), natural gas consumption is reduced, andLNG imports are higher than in the reference case.The rapid technology case assumes 50 percent fastertechnology progress than in the reference case,resulting in lower development costs, higher productionlevels, lower wellhead prices ($5.20 per thousandcubic feet in 2030), increased consumption of naturalgas, and lower LNG imports than in the referencecase. Technically recoverable natural gas resources(Table 17) are expected to be adequate to support thehigher production levels in the rapid technology case[91].Table 17. Technically recoverable U.S. natural gasresources as of January 1, 2004 (trillion cubic feet)Proved Unproved Total189.0 1,115.0 1,304.001990 1995 2004 2010 2020 2030In the rapid technology case, lower wellhead pricesfor natural gas lead to increased consumption andlower import levels. Natural gas production increasesto meet the increased demand (Figure 78). In 2030,natural gas production is 24.4 trillion cubic feet(17 percent higher than in the reference case), netnatural gas imports are 4.4 trillion cubic feet (20 percentlower), and domestic natural gas consumption is29.4 trillion cubic feet (9 percent higher).In the slow technology case, higher wellhead pricesreduce domestic consumption of natural gas, increasenatural gas imports, and reduce domestic production.In 2030, natural gas production is 18.8 trillion cubicfeet (10 percent lower than in the reference case), netnatural gas imports are 6.4 trillion cubic feet (14 percenthigher), and domestic natural gas consumptionis 25.6 trillion cubic feet (5 percent lower).Canada’s natural gas production also varies withchanges in assumptions about technological progressrates. In the rapid technology case, U.S. imports ofnatural gas from Canada in 2030 increase to 2.0 trillioncubic feet, 10 percent higher than in the referencecase. In the slow technology case, imports fromCanada in 2030 fall to 1.5 trillion cubic feet, 16 percentlower than in the reference case.Lower domestic prices for natural gas reduce netimports of LNG, and higher prices increase netimports. In the rapid and slow technology cases, netLNG imports in 2030 are 3.2 and 5.3 trillion cubicfeet, respectively, compared with 4.4 trillion cubicfeet in the reference case.88 Energy Information Administration / Annual Energy Outlook 2006

Natural Gas Alternative CasesNatural Gas Prices Vary WithAssumptions About Resource LevelsFigure 79. Lower 48 natural gas wellhead pricesin three price cases, 1990-2030 (2004 dollarsper thousand cubic feet)10 HistoryProjections864High priceReferenceLow priceLNG Imports Are the Source of SupplyMost Affected in the Price CasesFigure 80. Net imports of liquefied natural gasin three price cases, 1990-2030 (trillion cubic feet)10864HistoryProjectionsLow priceReference22High price01990 1995 2004 2010 2015 2020 2025 2030The high and low price cases assume that theunproven domestic natural gas resource base is 15percent lower and 15 percent higher, respectively,than the estimates used in the reference case. As theestimate of the domestic natural gas resource baseincreases, projected natural gas prices decline,because the more abundant resource base keepsnatural gas exploration and production costs lowerover time. With the lower resource level in the highprice case, the wellhead price for natural gas in 2030rises to $7.71 per thousand cubic feet (2004 dollars),30 percent higher than in the reference case ($5.92per thousand cubic feet). In the low price case, with ahigher resource level, the wellhead price in 2030 fallsto $4.97 per thousand cubic feet in 2030, 16 percentlower than in the reference case (Figure 79).The high and low price cases affect domestic consumption,production, and imports. In the high pricecase, domestic natural gas consumption in 2030 is2.9 trillion cubic feet (11 percent) lower than in thereference case. In the low price case, domestic naturalgas consumption in 2030 is 4.2 trillion cubic feet(16 percent) higher than in the reference case.Demand for natural gas in the electricity generationsector is more responsive to prices than demand inthe other end-use sectors and shows more variationin the two natural gas price cases. In 2030, naturalgas consumption in the electric power sector variesfrom 4.0 trillion cubic feet in the high price case to9.9 trillion cubic feet in the low price case, as comparedwith 6.4 trillion cubic feet in the reference case.01990 1995 2004 2010 2015 2020 2025 2030Among the major sources of natural gas supply, LNGimports are the most affected in the three price cases.Net imports of LNG in the reference case are 4.4 trillioncubic feet in 2030; in the low price case netimports in 2030 increase to 7.4 trillion cubic feet, andin the high price case they fall to 1.9 trillion cubic feet(Figure 80).Higher world oil prices are expected to result in a shiftaway from petroleum consumption and toward naturalgas consumption in all sectors of the internationalenergy market. In addition, some LNG contractprices are tied directly to crude oil prices, puttingfurther upward pressure on LNG prices. Finally,higher oil prices are expected to promote increases inGTL production, which in turn would lead to moreprice pressure on world natural gas supplies. In thehigh price case, the result is higher prices for naturalgas and LNG, both in the United States and internationally,reducing U.S. LNG imports, new LNGreceiving capacity, and the utilization rates for existingLNG terminals.With higher and lower wellhead prices for natural gasin the high and low price cases, domestic consumptionis reduced or increased by about the same amount asLNG imports. As a result, domestic natural gas productiondoes not vary significantly among the threecases. From 18.5 trillion cubic feet in 2004, U.S. naturalgas production increases to 20.8 trillion cubic feetin 2030 in the reference case, 21.2 trillion cubic feet inthe high price case, and 21.4 trillion cubic feet in thelow price case.Energy Information Administration / Annual Energy Outlook 2006 89

Natural Gas Alternative CasesLNG Supply Cases AddressUncertainty in Future LNG ImportsFigure 81. Net imports of liquefied natural gasin three LNG supply cases, 1990-2030(trillion cubic feet)10 HistoryProjectionsHigh LNGLNG Import Levels Have a DirectEffect on Domestic Natural Gas PricesFigure 82. Lower 48 natural gas wellhead pricesin three LNG supply cases, 1990-2030(2004 dollars per thousand cubic feet)8 HistoryProjections864Reference64Low LNGReferenceHigh LNG2Low LNG201990 1995 2004 2010 2015 2020 2025 2030The AEO2006 reference case assumes that two LNGterminals under construction as of August 1, 2005,will be completed: the Cheniere Energy terminals inFreeport, Texas (1.5 billion cubic feet per day), andCameron Parish, Louisiana (2.6 billion cubic feet perday), with assumed in-service dates of 2008 and 2009,respectively. The reference case also assumes expansionsat three of the four existing terminals, includinga proposed expansion of 0.8 billion cubic feet per dayat Cove Point, Maryland, and approved expansions of1.1 billion cubic feet per day at Lake Charles, Louisiana,and 0.54 billion cubic feet per day at Elba Island,Georgia. In addition, the reference case assumes thatnew facilities will be built to serve the Gulf Coast,Southern California, Florida, and New England.High and low LNG supply cases were developed toexamine the impacts of variations in LNG supply ondomestic natural gas supply, consumption, andprices. The low LNG supply case assumes a futurelevel of LNG imports 30 percent lower than in thehigh price case, with imports in 2030 at 1.3 trillioncubic feet, compared with 4.4 trillion cubic feet in thereference case (Figure 81). The high LNG supply caseassumes future LNG imports 30 percent higher thanin the low price case, with imports in 2030 at 9.6 trillioncubic feet.01990 1995 2004 2010 2015 2020 2025 2030In the high LNG supply case, domestic natural gasproduction and wellhead prices are lower than thosein the reference case, and natural gas consumption ishigher. In 2030, natural gas wellhead prices in thehigh LNG case are 10 percent lower than in the referencecase, at $5.35 per thousand cubic feet (Figure82), and natural gas production is 8 percent lowerthan in the reference case, at 19.1 trillion cubic feet.Domestic natural gas consumption in 2030 in thehigh LNG case is 12 percent higher than in the referencecase, at 30.1 trillion cubic feet.In the low LNG supply case the total supply of naturalgas to U.S. consumers is less than in the referencecase, leading to higher prices, lower consumption, andmore domestic natural gas production. In 2030, naturalgas wellhead prices in the low LNG case are 7 percenthigher than in the reference case, at $6.36thousand cubic feet, natural gas production is 6 percenthigher, at 22.0 trillion cubic feet, and domesticnatural gas consumption is 6 percent lower, at 25.3trillion cubic feet.Lower and higher wellhead prices for natural gas inthe high and low LNG supply cases have the greatestimpact on natural gas consumption in the electricpower sector, both because of its high projectedgrowth rate and because of the competition betweencoal and natural gas. In the high LNG case, naturalgas consumption for electricity generation in 2030 is44 percent higher than in the reference case, at 9.2trillion cubic feet, and in the low LNG case it is 21 percentlower than in the reference case, at 5.0 trillioncubic feet.90 Energy Information Administration / Annual Energy Outlook 2006

Oil Prices and ProductionOil Prices Decline in the Short Term,Then Rise Through 2030Figure 83. World oil prices in the reference case,1990-2030 (2004 dollars per barrel)70 History ProjectionsDomestic Crude Oil ProductionBegins To Decline After 2016Figure 84. Domestic crude oil production by source,1990-2030 (million barrels per day)8 HistoryProjections60504030201001990 1995 2004 2010 2015 2020 2025 2030The world oil price in AEO2006 is defined as theweighted average price of all crude oil containing lessthan 0.5 percent sulfur by weight that is imported byU.S. refiners. The price projection in the referencecase is based on the mean conventional petroleumresource estimate (Table 18) reported by the USGSand the U.S. Minerals Management Service [92]. Thereference case assumes that the Arctic National WildlifeRefuge (ANWR) will continue to be off limits topetroleum exploration and development.In the AEO2006 reference case, as new oil fields arebrought into production worldwide, world oil pricesdecline to $46.90 per barrel (2004 dollars) in 2014,then rise to $56.97 in 2030 (Figure 83). The increaseafter 2014 reflects rising costs for the developmentand production of non-OPEC oil resources. There isconsiderable uncertainty associated with the priceprojections, related to world economic growth, worldoil demand, OPEC’s long-term oil production policies,and international political stability.Table 18. Technically recoverable U.S. crude oilresources as of January 1, 2004 (billion barrels)Proved Unproved Total23.1 124.1 147.2642Alaska01990 1995 2004 2010 2015 2020 2025 2030TotalLower 48 onshoreDeepwater offshoreShallow water offshoreA large proportion of the total U.S. resource base ofonshore conventional oil has been produced, and newoil reservoir discoveries are likely to be smaller, moreremote, and increasingly costly to exploit. Whilehigher oil prices make it economical to produce highercost resources, lower 48 onshore oil productiondeclines in the reference case from 2.9 million barrelsper day in 2004 to 2.3 million barrels per day in 2030(Figure 84).The remaining onshore conventional oil resourcebase is not expected to provide significant new suppliesof oil, with the exception of production from theNational Petroleum Reserve in Alaska. Oil productionin the Reserve begins in 2007, increases to a peakof 458,000 barrels per day in 2016, and declinesthereafter. As a result, Alaska’s total oil production—which falls from 906,000 barrels per day in 2004 to828,000 barrels per day in 2007—rebounds to 902,000barrels per day in 2014 before beginning a steadydecline to 274,000 barrels per day in 2030.Considerable oil resources remain in the offshore,especially in the deep waters of the Gulf of Mexico. Oilproduction in the shallow waters of the Gulf, startingfrom 0.4 million barrels per day in 2004, slips to 0.3million barrels per day in 2030, while deepwater productionincreases from 1.0 million barrels per day in2004 to a peak of 2.2 million barrels per day in 2016and then declines to 1.7 million barrels per day in2030. As a result, total U.S. offshore oil productionincreases from 1.6 million barrels per day in 2004 to2.5 million barrels per day in 2016, then falls back to2.0 million barrels per day in 2030.Energy Information Administration / Annual Energy Outlook 2006 91

Oil Price CasesPrice Cases Assess AlternativeFutures for World Oil MarketFigure 85. World oil prices in three cases, 1990-2030(2004 dollars per barrel)10080604020HistoryProjections01990 1995 2004 2010 2015 2020 2025 2030High priceReferenceLow priceThe high and low price cases reflect different assumptionsabout the size of the conventional world oilresource, and they project different market shares forOPEC and non-OPEC oil production. The high pricecase assumes that the world conventional crude oilresource base is 15 percent smaller than the USGSmean oil resource estimate. In the high price case,world oil production reaches 102 million barrels perday in 2030, with OPEC contributing 31 percent oftotal world oil production. World oil prices increase to$76.30 per barrel (2004 dollars) in 2015 and $95.71per barrel in 2030 (Figure 85).The low price case assumes that the conventionalworldwide oil resource base is 15 percent larger thanthe USGS mean estimate. In the low price case, worldoil production reaches 128 million barrels per day in2030, with OPEC contributing 40 percent of totalworld oil production. World oil prices, in terms of theaverage price of imported low-sulfur crude oil to U.S.refiners, drop to $33.78 per barrel in 2015 and remainrelatively stable thereafter.U.S. Oil Production is MarginallySensitive to World Oil PricesFigure 86. Total U.S. petroleum production in threeprice cases, 1990-2030 (million barrels per day)12108642HistoryProjections01990 1995 2004 2010 2015 2020 2025 2030High priceReferenceLow priceThe high price case assumes that conventionaldomestic oil resources are 15 percent less than in thereference case, and the low price case assumes theyare 15 percent greater. The difference has a directeffect on the cost and availability of newly developeddomestic oil supplies. A higher (or lower) oil price alsoinduces more (or less) exploration activity and thedevelopment of more (or less) expensive oil resources.In all cases, a significant portion of total domestic oilproduction comes from large, existing oil fields, suchas the Prudhoe Bay Field.Oil prices also determine whether unconventional oilproduction (such as oil shale, CTL, and GTL) is economical,as illustrated in the alternative price cases.CTL production is projected in both the reference andhigh price cases; however, GTL production andsyncrude production from oil shale, both of whichrequire higher prices before they become economical,are projected only in the high price case.With higher oil prices, unconventional sources of oilbecome economical, and unconventional productionincreases. In the high price case, total conventionaland unconventional domestic petroleum production(including NGL and refinery processing gain) in 2030is 20 percent higher than in the reference case, at10.3 million barrels per day. In the low price case,total production is 10 percent lower than in thereference case, at 7.7 million barrels per day in 2030(Figure 86).92 Energy Information Administration / Annual Energy Outlook 2006

Oil Price and Technology CasesU.S. Syncrude Production FromOil Shale Requires Higher Oil PricesFigure 87. U.S. syncrude production from oil shalein the high price case, 2004-2030(thousand barrels per day)500More Rapid Technology AdvancesCould Raise U.S. Oil ProductionFigure 88. Total U.S. crude oil productionin three technology cases, 1990-2030(million barrels per day)10400830020064Rapid technologyReferenceSlow technology100202004 2010 2015 2020 2025 2030In the United States, the commercial viability ofsyncrude produced from oil shale largely depends onoil prices. Although the production costs for oil shalesyncrude decline through 2030 in all cases, it becomeseconomical only in the high price case, with productionstarting in 2019 and increasing to 405,000 barrelsper day in 2030, when it represents 4 percent ofU.S. petroleum production, including NGL and refineryprocessing gain (Figure 87).Production costs for oil shale syncrude are highlyuncertain. Development of this domestic resourcecame to a halt in the mid-1980s, during a period of lowoil prices. The cost assumptions used in developingthe projections represent an oil shale industrybased on underground mining and surface retorting;however, the development of a true in situ retortingtechnology could substantially reduce the cost of producingoil shale syncrude.The development of U.S. oil shale resources is alsouncertain from an environmental perspective. Oilshale costs will remain highly uncertain until thepetroleum industry builds a demonstration project.An oil shale industry based on underground miningand surface retorting could face considerable publicopposition because of its potential environmentalimpacts, involving scenic vistas, waste rock disposaland remediation, and water availability and contamination.Consequently, there is a high level ofuncertainty in the projection for oil shale syncrudeproduction in the high price case.HistoryProjections01990 1995 2004 2010 2015 2020 2025 2030The rapid and slow oil and gas technology casesassume rates of technological progress in the petroleumindustry that are 50 percent higher and 50 percentlower, respectively, than the historical rate. Therate of technological progress determines the cost ofdeveloping and producing the remaining domestic oilresource base. Higher (or lower) rates of technologicalprogress result in lower (or higher) oil developmentand production costs, which in turn allow more (orless) oil production.With domestic oil consumption determined largely byoil prices and economic growth rates, oil consumptiondoes not change significantly in the technology cases.Domestic crude oil production in 2030, which is4.6 million barrels per day in the reference case,increases to 4.9 million barrels per day in the rapidtechnology case and drops to 4.2 million barrels perday in the slow technology case (Figure 88). The projectedchanges in domestic oil production result indifferent projections for petroleum imports. In 2030,projected net crude oil and petroleum productimports range from 16.7 million barrels per day in therapid technology case to 17.7 million barrels per dayin the slow technology case, as compared with 17.2million barrels per day in the reference case. U.S.dependence on petroleum imports in 2030 rangesfrom 61 percent in the rapid technology case to 64percent in the slow technology case.Cumulatively, from 2004 through 2030, U.S. totalcrude oil production is projected to be 1.9 billion barrels(3.8 percent) higher in the rapid technology caseand 2.1 billion barrels (4.1 percent) lower in the slowtechnology case than in the reference case.Energy Information Administration / Annual Energy Outlook 2006 93

ANWR Alternative Oil CaseDrilling in ANWR Could SustainAlaska’s Oil ProductionANWR Oil Production Could LowerU.S. Net Oil Imports Through 2030Figure 89. Alaskan oil production in thereference and ANWR cases, 1990-2030(million barrels per day)2.01.5Figure 90. U.S. net imports of oil in thereference and ANWR cases, 1990-2030(million barrels per day)2015ReferenceANWR case1.0ANWR case100.5HistoryProjections0.01990 1995 2004 2010 2015 2020 2025 2030ReferenceWhether Federal oil and natural gas leasing inANWR will ever occur remains uncertain. The AEO-2006 ANWR alternative case suggests the potentialimpact of opening ANWR to leasing. The ANWR caseuses the same assumptions as the reference case,except that oil and natural gas development and productionare allowed in ANWR, starting in 2005.The opening of ANWR to development in 2005 resultsin the initiation of ANWR oil production in 2015. Oilproduction from ANWR grows to a peak of 780,000barrels per day in 2024, then declines to 650,000 barrelsper day in 2030. In the reference case, with no oilproduction from ANWR, Alaska’s total oil productiongrows to 900,000 barrels per day in 2014 and thendeclines to 270,000 barrels per day in 2030. In theANWR case, Alaskan oil production rises to 1.4 millionbarrels per day in 2021 and then falls to 930,000barrels per day in 2030 (Figure 89).World oil prices are slightly lower in the ANWR casethan in the reference case. The largest difference is79 cents per barrel in 2024 (in 2004 dollars), whenANWR oil production is at its peak. After 2024, asANWR production declines, the difference narrows to68 cents per barrel in 2030.5HistoryProjections01990 1995 2004 2010 2015 2020 2025 2030The opening of ANWR to Federal oil and natural gasleasing increases domestic oil production. In the referencecase, U.S. total crude oil production peaks in2010 at 5.9 million barrels per day, then declines to4.6 million barrels per day in 2030. In the ANWR case,total domestic oil production peaks in 2020 at 6.2 millionbarrels per day and then falls to 5.2 million barrelsper day in 2030.Every additional barrel of oil produced in ANWReffectively displaces a barrel of imported crude oil. In2024, when ANWR production peaks in the alternativecase, the import share of total domestic petroleumsupply is 57 percent (14.7 million barrels perday), compared with 60 percent (15.4 million barrelsper day) in the reference case (Figure 90). In 2030,when ANWR production is declining, the importshare of total domestic petroleum supply is 60 percentin the ANWR case and 62 percent in the referencecase.Although the opening of ANWR to Federal oil andnatural gas leasing reduces projected oil prices, theimpact on domestic oil consumption is negligible. In2024, when projected ANWR oil production is highestand the reduction in oil prices is largest, domesticconsumption of petroleum products is only about60,000 barrels per day higher in the ANWR case thanin the reference case. The difference in domestic oilconsumption is the same in 2030.94 Energy Information Administration / Annual Energy Outlook 2006

Refined Petroleum ProductsTransportation Uses Lead Growthin Petroleum ConsumptionFigure 91. Consumption of petroleum products,1990-2030 (million barrels per day)3025201510HistoryProjectionsTotalMotorgasolineExpansion at Existing RefineriesIncreases U.S. Refining CapacityFigure 92. Domestic refinery distillation capacity,1990-2030 (million barrels per day)2520151015.7History16.9Projections17.6 17.9 18.1 18.519.35OtherDistillateJet fuel0Residual1990 1995 2004 2010 2015 2020 2025 2030501990 2004 2010 2015 2020 2025 2030Between 70 and 74 percent of U.S. petroleum use isfor transportation, and much of the projected growthin domestic consumption reflects growth in the use oftransportation fuels (Figure 91). Gasoline, distillatefuel (ultra-low-sulfur diesel), and jet fuel are the maintransportation fuels. In the AEO2006 reference case,improvements in technology increase the efficiency ofmotor vehicles and aircraft, but growth in demand foreach mode of transit far outpaces increases in fuelefficiency, as transportation demand grows in proportionto increases in population and GDP.In the residential sector, the use of distillate for homeheating declines as natural gas and LPG are usedincreasingly as substitutes. Both burn more cleanlythan distillate, eliminating the annual maintenancethat is needed for an oil-fired furnace or boiler. Naturalgas, where available, is more convenient than distillateor LPG.In the industrial and commercial sectors, distillate isused as a fuel for heating and for diesel engines. In thenear term, high prices for distillate lead to fuelswitching away from heating oil; but as prices moderate,there is some switching back to distillate forheating uses.Residual fuel is blended from the heaviest crude oilcomponents. Undiluted residual fuel is used to powerships and electricity plants. Residual fuel diluted withdistillate is used to fire boilers and to power somelocomotives. Residual fuel consumption declinesin the reference case as environmental restrictionstighten, and because refiners find it more attractiveto upgrade residual fuel to lighter products.Distillation capacity at U.S. refineries expands in thereference case (Figure 92) as demand for refinedpetroleum products increases. More than 30 yearshave passed since a new U.S. refinery was built, andmost of the expansion occurs at existing sites. Althoughit is not difficult technically for refiners andrefinery process developers to expand the capacity ofexisting units, obtaining permits is difficult, and gettingpermits to build a new refinery is even harder.Nonetheless, a startup company has announced plansto open a major new refinery in Arizona in 2010.The most basic refinery operation is atmosphericdistillation of crude oil. Crude oil is heated to about750 degrees Fahrenheit and then fed into a towerwhere it separates into fractions according to the boilingpoints of the many compounds it contains. Theseparated fractions are sent on to other units inthe refinery for further processing and, ultimately,blending into finished products.Other processing units in a refinery generally expandat about the same rate as distillation capacity; however,tighter product specifications, poorer crude oilquality, and dwindling demand for residual fuelincrease the capacity needed for two processes, cokingand hydrotreating. Coking is used to break the heaviestfractions of crude oil into elemental carbon, orcoke, and lighter fractions. Material used in the cokerwould otherwise be usable only as residual fuel orasphalt. Hydrotreating capacity, which is used to takesulfur out of petroleum products, allows refiners tomeet tighter limits on sulfur content and to runhigher sulfur crude oils through their refineries.Energy Information Administration / Annual Energy Outlook 2006 95

Refined Petroleum ProductsImports of Petroleum ProductsIncrease With Rising U.S. DemandFigure 93. U.S. petroleum product demand anddomestic petroleum supply, 1990-2030(million barrels per day)30252015105HistoryProjectionsDemandSupplyU.S. Motor Gasoline Prices Rise andFall With Changes in World Oil PriceFigure 94. Components of retail gasoline prices,2004, 2015, and 2030 (2004 dollars per gallon)2.52.01.51.00.5World oil priceTotal priceWholesaleDistributionState taxesFederal taxes01990 1995 2004 2010 2015 2020 2025 20300.02004 2015 2030U.S. petroleum market regulations before the 1980sencouraged the U.S. refinery industry to overinvestin capacity. In the 1980s, deregulation encouragedthe shutdown of inefficient refineries, and strongdemand growth in response to the low oil prices ofthe late 1980s and the 1990s eliminated any excesscapacity that remained. In the AEO2006 referencecase, refinery utilization increases from 93 percent in2004 to 95 percent in 2030. In the 1980s, capacity utilizationat U.S. refineries averaged only 69 percent.The most advantageous locations for refineries arenear crude oil production sites or where demand forpetroleum products is concentrated. As both a majorproducer and consumer of petroleum products, theUnited States has a large refinery complex, but U.S.demand for petroleum exceeded domestic productionlong ago, and the Nation has been a net importer ofcrude oil for more than 50 years (Figure 93).In the reference case, demand for refined productscontinues to increase more rapidly than refiningcapacity, and petroleum product imports increase tofill the gap. Historically, the availability of productimports has been limited by a lack of foreign refineriescapable of meeting the stringent U.S. standardsfor petroleum products. More recently, petroleumdemand has grown rapidly in Eastern Europe andAsia, and those nations are moving to adopt the samequality standards as the developed world. As a result,refineries throughout the world are becoming moresophisticated, and more of them will be able to provideproducts suitable for the U.S. market in thefuture, which they may do if it is profitable.Changes in crude oil prices have a direct impact onwholesale prices for petroleum products. In the referencecase, the world price (the price of importedlow-sulfur light crude oil in 2004 dollars) reachs a lowof $47.79 per barrel in 2015, then begins a slowincrease that continues to a level of $56.97 per barrelin 2030. The U.S. average gasoline price in 2015 is$2.00 per gallon and $2.19 per gallon in 2030 (Figure94). Accordingly, the wholesale price makes up 69percent of the retail price for transportation gasolinein 2015 and 73 percent in 2030. In comparison, fortransportation diesel fuel, the wholesale price is 71percent of the retail price in 2015 and 74 percent in2030.The most recent increase in the Federal excise tax onmotor fuels was enacted in 1993. Consistent with historicaltrends, State taxes on gasoline decline slightlyin real terms in the reference case, and Federal taxesdecline substantially. As a result, Federal taxes ongasoline and highway diesel in 2030 are only 52 percentof their 2004 levels. State and Federal taxesmake up 19 percent of retail gasoline prices in 2015and 16 percent in 2030. For transportation diesel,taxes make up 20 percent of the retail price in 2015and 16 percent in 2030.96 Energy Information Administration / Annual Energy Outlook 2006

Ethanol and Synthetic FuelsU.S. Demand for Ethanol Fuel VariesWith World Oil Price ProjectionsFigure 95. U.S. ethanol fuel consumption in threeprice cases, 1995-2030 (billion gallons per year)1510HistoryProjectionsHigh priceReferenceLow priceSynthetic Fuel Production GrowsRapidly in the High Price CaseFigure 96. Coal-to-liquids and gas-to-liquidsproduction in two price cases, 2004-2030(million barrels per day)2.01.5High price51.00.5Reference01995 2000 2004 2010 2015 2020 2025 2030EPACT2005 repealed the oxygenate requirementfor Federal RFG. The only economically feasible oxygenatesare ethanol and MTBE. It is easier to meetthe other requirements for RFG, such as volatilityand aromatics emissions limits, with MTBE; however,MTBE readily contaminates groundwater whenblended gasoline is leaked or spilled. Refiners see therepeal of the oxygenate requirement as increasingtheir liability for MTBE pollution of water. They areexpected to stop making and blending MTBE by 2008,but ethanol blending into RFG is expected to continue,because ethanol is a clean, high-octane blendingcomponent that can be used to replace MTBE.Ethanol is a substitute for hydrocarbons, and whencrude oil prices increase, more ethanol is used to meetdemand for gasoline (Figure 95). In 2030, ethanolblending into gasoline ranges from about 5 percent ofthe gasoline pool in the low price case to almost 9 percentin the high price case.Virtually all the fuel ethanol produced in the UnitedStates is distilled from corn. EPACT2005 requires theuse of 250 million gallons per year of ethanol distilledfrom cellulosic materials, starting in 2012. Decliningcorn prices in real terms and improvements in grainethanol technology prevent further penetration ofcellulosic ethanol use in the reference case. Corn ethanolproduction is near practical limits in the referencecase, however, and production of ethanol fromcellulose feedstocks begins in 2010. In the high pricecase, cellulosic ethanol production exceeds the levelmandated in EPACT2005.0.02004 2010 2015 2020 2025 2030GTL and CTL processes are used to convert naturalgas and coal, respectively, into high-quality blendingcomponents for diesel fuel. Naphthas, waxes, andlubrication oil components are produced asbyproducts.Per unit of capacity, CTL and GTL plants are moreexpensive to construct than are petroleum refineries.The natural gas needed to feed a GTL plant is alsoexpensive. In the reference case, the cost of naturalgas makes GTL unattractive, and no U.S. plants arebuilt by 2030. Coal, however, is much cheaper thannatural gas, and CTL fuels enter the market in 2011(Figure 96). CTL production in 2030 totals 760,000barrels per day in the reference case and makes up 13percent of distillate fuel supply.Higher crude oil prices encourage the substitution ofnatural gas and coal for oil. In the high price case,GTL enters the market in 2020, and productiongrows to 194,000 barrels per day in 2030. CTL productiongrows to 1.69 million barrels per day in 2030in the high price case. Together, GTL and CTL provide32 percent of the Nation’s distillate fuel supply in2030 in the high price case. Neither GTL nor CTLfuels are economically feasible in the low price case.Energy Information Administration / Annual Energy Outlook 2006 97

Coal Supply and DemandMarket Share of Western CoalContinues To IncreaseFigure 97. Coal production by region, 1970-2030(million short tons)2,0001,5001,000500HistoryProjections01970 1985 1995 2004 2015 2030TotalWestAppalachiaInteriorU.S. coal production has remained near 1,100 milliontons annually since 1996. In the AEO2006 referencecase, increasing coal use for electricity generation atexisting plants and construction of a few new coalfiredplants lead to annual production increases thataverage 1.1 percent per year from 2004 to 2015, whentotal production is 1,272 million tons. The growth incoal production is even stronger thereafter, averaging2.0 percent per year from 2015 to 2030, as substantialamounts of new coal-fired generating capacity areadded, and several CTL plants are brought on line.Western coal production, which has grown steadilysince 1970, continues to increase through 2030(Figure 97), especially in the Powder River Basin,where vast reserves are contained in thick seamsaccessible to surface mining. Easing of rail transportationbottlenecks will be needed for producers in theWest to exploit the market opportunities presentedby slow growth in Appalachian coal production and bydemand for coal at new power plants built to serveelectricity markets in the Southwest and California.Appalachian coal production remains nearly flat inthe reference case. Although producers in CentralAppalachia are well situated geographically to supplycoal to new generating capacity in the Southeast, theAppalachian basin has been mined extensively, andproduction costs have been increasing more rapidlythan in other regions. The Eastern Interior coal basin(Illinois, Indiana, and western Kentucky), with extensivereserves of mid- and high-sulfur bituminouscoals, does benefit from the new builds of coal-firedgenerating capacity in the Southeast.More Eastern Power PlantsAre Expected To Use Western CoalFigure 98. Distribution of domestic coal by demandand supply region, 2003 and 2030 (million shorttons)Supply regionAppalachiaInteriorWestTotalSupply regionAppalachiaInteriorWestTotalEastern demand regionWestern demand region0 200 400 600 800 1,00020032030In the reference case, low-cost Western coal continuesto gain market share east of the Mississippi River andremains the dominant supplier in markets west of theMississippi River (Figure 98). Use of low-sulfur Westerncoal continues to increase through 2030, eventhough 141 gigawatts of existing coal-fired capacity isretrofitted with flue gas desulfurization equipmentand another 174 gigawatts of new environmentallycompliant coal-fired capacity is built. Even in theabsence of sulfur compliance costs, Western coal isthe lowest cost fuel option for electricity generation inmany areas of the country. As a result, each year typicallysees more coal-fired plants switching to Westerncoal, particularly from the Powder River Basin. In2004, approximately 20 plants, many located east ofthe Mississippi River, used Powder River Basin coalfor the first time.Although two new pieces of environmental legislationenacted in 2005, CAIR and CAMR, will increase thecost of coal-fired generation, they have only minorimpacts on overall coal use in the electric power sectoror regional coal production patterns. As a result of thestricter caps on SO 2 emissions in CAIR, allowanceprices increase substantially, virtually eliminating by2030 the use of medium- and high-sulfur coals (containingmore than 0.6 pounds sulfur per million Btu)at power plants not equipped with scrubbers. In 2004,medium- and high-sulfur coals accounted for about 40percent of the 638 million tons of coal consumed atgenerating units without scrubbers [93]. In 2030,coal-fired power plants without scrubbers consumeonly 233 million tons.98 Energy Information Administration / Annual Energy Outlook 2006

Coal Mine Employment and Coal PricesCoal Mine Employment IncreasesAs Production ExpandsFigure 99. U.S. coal mine employment by region,1970-2030 (number of jobs)250,000200,000150,000100,000HistoryProjectionsTotal50,000AppalachiaWest0Interior1970 1985 1995 2004 2015 2030Most jobs in the U.S. coal industry remain east of theMississippi River, mainly in the Appalachian region(67 percent in 2004). Most coal production, however,occurs west of the Mississippi River (56 percent in2004), with the major share from the Powder RiverBasin. As coal demand increases, pressure to keepprices low will shift more production to mines withhigher labor productivity. Large surface mines in thePowder River Basin take advantage of economies ofscale, using large earth-moving equipment and combiningadjacent mines to increase operating flexibility.Underground mines in the Northern Appalachiaand Rocky Mountain supply regions use highly productiveand increasingly automated longwall equipmentto maximize production while reducing thenumber of miners required.In the reference case, labor productivity remains nearcurrent levels in most coal supply regions, reflectingthe trend of the past 5 years. Higher stripping ratiosand the additional labor needed to maintain moreextensive underground mines offset productivitygains achieved from improved equipment, automation,and technology. Productivity in some areas ofthe East declines as operations move to marginalreserve areas. Regulatory restrictions on surfacemines and fragmentation of underground reserveslimit the benefits that can be achieved by Appalachianproducers from economies of scale.Some 27,000 additional mining jobs are createdbetween 2004 and 2030 (Figure 99). In the East, joblosses in Central Appalachia are more than offset byadditional jobs at more productive mines in NorthernAppalachia.Average Minemouth Coal PricesIncrease SlowlyFigure 100. Average minemouth price of coal byregion, 1990-2030 (2004 dollars per short ton)7060504030201021.76Average price(nominal dollars) 40.791990 2030HistoryProjections01990 1995 2004 2010 2015 2020 2025 2030AppalachiaInteriorU.S. averageWestFrom 1990 to 1999, the average minemouth priceof coal declined by 4.9 percent per year, from $29.09per ton (2004 dollars) to $18.54 per ton (Figure 100).Increases in U.S. coal mining productivity of 6.3 percentper year during the period helped to reducemining costs and contributed to the decline in prices.Since 1999, growth in U.S. coal mining productivityhas slowed to 0.6 percent per year, and the averageminemouth coal price has increased by 1.6 percentper year, to $20.07 per ton in 2004.In the reference case, the average minemouth coalprice drops slightly from 2010 to 2020, as mine capacityutilization declines and production shifts awayfrom higher cost Central Appalachian mines. After2020, rising natural gas prices and the need forbaseload generating capacity result in the constructionof 126 gigawatts of new coal-fired generatingplants (72 percent of all coal builds from 2004 to 2030in the reference case), and production in most of themajor coal supply basins increases. The substantialinvestment in new mining capacity required to meetincreasing demand during the period, combined withlow productivity growth and rising utilization of miningcapacity, leads to an increase in the averageminemouth price, from $20.20 per ton in 2020 to$21.73 per ton in 2030. Strong growth in productionin the Interior and Western supply regions, combinedwith limited improvement in coal mining productivity,results in minemouth price increases of 1.4 and1.2 percent per year, respectively, in those regionsfrom 2004 through 2030. With little increase in production,average minemouth prices in Appalachiaincrease by only 0.1 percent per year over the sameperiod.Energy Information Administration / Annual Energy Outlook 2006 99

Coal Transportation and ImportsRising Regional Coal TransportationRates Depart from Historical TrendDemand for Imported Coal Increasesin the East and SoutheastFigure 101. Changes in regional coaltransportation rates, 2010, 2025, and 2030(percent increase from 2004 rates)1086Eastern coalWestern coalFigure 102. U.S. coal exports and imports,1970-2030 (million short tons)12510075HistoryProjectionsImports450225Exports02010 2025 203001970 1985 1995 2004 2015 2030Coal transportation rates (in constant 2004 dollars),rise in the reference case, ending the decreasing trendof the past 20 years. Historically, infrastructureinvestments and subsequent overcapacity, as well asthe efficiency gains associated with consolidation ofthe railroad industry, have steadily reduced coaltransportation rates. Productivity improvementscontinue in the forecast, but they are dampened bylarger demands on rail infrastructure and an expectationthat investments will be made incrementally, asneeded, rather than in anticipation of higher demand.Periodic bottlenecks are likely as railroads adapt toincreasing traffic flows from western mines andchanging coal distribution patterns in the East. Inconstant dollars, coal transportation costs peak in2010, then fall to 2.2 percent and 3.3 percent above2004 levels in 2030 for coal originating in the Westand East, respectively (Figure 101). In general, westernsuppliers are at a greater disadvantage than easternsuppliers when transportation rates rise, becausewestern coal typically travels over longer distances.Despite the increases in transportation rates, thenational average continues to decline, because 76 percentof the increase in demand from 2004 to 2030 isfrom CTL plants and new electric power plants, manyof which are expected to be built near sources of coalsupply. In 2030, the average coal transportation ratefor new electric power capacity is $7.14 per short ton(2004 dollars), compared with $8.63 for existingcapacity.U.S. imports of low-sulfur coal rise from 27 milliontons in 2004 to 99 million tons in 2030 (Figure 102).In addition to further displacement of more expensiveCentral and Southern Appalachian coal at existingpower plants, imports fuel some of the new coal-firedgenerating capacity expected to be built in the U.S.East and Southeast. Much of the additional importtonnage originates from mines in Colombia, Venezuela,and Indonesia.U.S. coal exports have been in steady decline fromtheir 1996 level of 90 million tons, falling to 40 milliontons in 2002, despite a substantial increase in worldcoal trade (from 503 million tons to 656 million tons).Low-cost supplies of coal from China, Colombia, Indonesia,Russia, and Australia satisfied much of thegrowth in international demand for steam coal duringthe period, and low-cost supplies of coking coal fromAustralia supplanted substantial amounts of U.S.coking coal in world markets. Since 2002, however,U.S. exports have rebounded, including increases insteam coal exports to Canada in 2003 and coking coalto overseas customers in 2004.Although U.S. exports remain near their 2004 levelfor the next several years, their share of total worldcoal trade ultimately falls from 6 percent in 2004 to1 percent in 2030, as international competition intensifiesand imports of coal to Europe and the Americas(excluding the United States) grow more slowlyor decline. With the planned decommissioning ofOntario’s five coal-fired generating plants, U.S. coalexports to Canada decline from 19 million tons in2004 to 7 million tons in 2030.100 Energy Information Administration / Annual Energy Outlook 2006

Coal ConsumptionCoal-Fired Generators Can ComplyWith CAIR and CAMRFigure 103. Electricity and other coal consumption,1970-2030 (million short tons)2,0001,5001,000HistoryProjectionsTotalElectricityEmerging Coal-to-Liquids IndustryIncreases Industrial Coal UseFigure 104. Coal consumption in the industrial andbuildings sectors and at coal-to-liquids plants,2004, 2015, and 2030 (million short tons)200150100Coke plantsOther industrialCoal-to-liquidsResidential/commercial500Coal-to-liquids0Other1970 1985 1995 2004 2015 20305002004 2015 2030EPA’s CAIR and CAMR regulations tighten restrictionson emissions of SO 2 and NO x and, for the firsttime, address mercury emissions from electric powerplants. Even with the new regulations, however, theaverage capacity utilization of coal-fired power plantsrises from 72 percent in 2004 to 80 percent in 2012.Coal-fired power plants fitted with emissions controlequipment remain competitive with natural-gas-firedgenerators because of their lower fuel costs. Coal consumptionin the electric power sector rises to 1.5 billiontons in 2030 (Figure 103). To comply with CAIRand CAMR, selective catalytic reduction equipment isadded to 118 gigawatts of coal-fired capacity between2004 and 2030 in the reference case, flue gasdesulfurization equipment is added to 141 gigawatts,and supplemental fabric filters are added to 126gigawatts between 2004 and 2030. Activated carbon,a sorbent added to post-combustion flue gases toremove mercury, is also used in some plants.In the projections, a mix of advanced IGCC and conventionalcoal-fired capacity is built in the electricpower sector; 55 percent of the 154 gigawatts of newcoal capacity is IGCC, which has low emissions ofboth SO 2 and mercury. A typical IGCC plant mayremove 99 percent of the sulfur and 95 percent of themercury present in bituminous coal. In addition, sustainedhigh world oil prices combined with competitivecoal prices stimulate investment in 19 gigawattsof CTL capacity, requiring 190 million tons of coal peryear, by 2030. SO 2 and mercury emissions from CTLplants are comparable with those from IGCC plants.Although the electric power sector accounts for thebulk of U.S. coal consumption, 89 million tons of coalcurrently is consumed in the industrial and buildings(residential and commercial) sectors (Figure 104). Inthe industrial sector, steam coal is used to manufactureor produce cement, paper, chemicals, food, primarymetals, and synthetic fuels; as a boiler fuel toproduce process steam and electricity; as a directsource of heat; and as a feedstock. Coal consumptionin the other industrial sector (excluding production ofcoal-based synthetic liquids) increases slightly in theAEO2006 reference case.Coal is also used to produce coke, which in turn isused as a source of energy and as a raw material inputat blast furnaces to produce iron. A continuing shiftfrom coke-based production at integrated steel millsto electric arc furnaces, combined with a relativelyflat outlook for U.S. steel production, leads to a slightdecline in consumption of coal at coke plants.Outside the electric power sector, most of the increasein coal demand in the reference case is for productionof coal-based synthetic liquids. High world oil pricesspur investment in the CTL industry, leading to theconstruction of new plants in the West and Midwestthat produce just under 0.8 million barrels of liquidsper day in 2030. In AEO2006, CTL technology is representedas an IGCC coal plant equipped with aFischer-Tropsch reactor to convert the synthesis gasto liquids. Of the total amount of coal consumed ateach plant, 49 percent of the energy input is retainedin the product, 20 percent is used for conversion,and 31 percent is used for grid-connected electricitygeneration.Energy Information Administration / Annual Energy Outlook 2006 101

Coal Alternative CasesHigh Economic Growth, High Oil andGas Prices Increase Coal DemandFigure 105. Projected variation from the referencecase projection of U.S. total coal demand in fourcases, 2030 (million short tons)4002000High economicgrowth caseOther demandCoal-to-liquidsElectric powerHigh price caseHigher Mining and TransportationCosts Reduce Demand for CoalFigure 106. Average delivered coal prices in threecost cases, 1990-2030 (2004 dollars per short ton)50403020HistoryProjectionsHigh costReferenceLow cost-20010-400Low economicgrowth caseLow price case01990 1995 2004 2010 2015 2020 2025 2030In comparison with the reference case, electricitydemand is higher in the high economic growth caseand lower in the low growth case. Accordingly, coalconsumption also rises and falls in the high and lowgrowth cases, respectively (Figure 105). As in the referenceand high growth cases, the first CTL plantcomes on line in 2011 in the low economic growthcase; but total CTL capacity in 2030 in the low growthcase is only 62 percent of that in the high growth case.In the high price case, higher natural gas prices discouragenatural-gas-fired generation and boost coalfiredgeneration. Delivered natural gas prices to theelectric power sector in 2030 are $1.63 per million Btuhigher in the high price case than in the referencecase, whereas coal prices are only 10 cents per millionBtu higher than in the reference case. In the referencecase, coal fuels 57 percent of total electricity generationin 2030 in the reference case, as comparedwith 64 percent in the high price case and 46 percentin the low price case.Higher world oil prices in the high price case favorincreased investment in CTL, and the demand forcoal at CTL facilities increases to 20 percent of totalcoal consumption in 2030. In the low price case, noCTL plants are operating in 2030. Because electricitygeneration at CTL plants displaces some generationin the electric power sector in the high price case, coaldemand in the electric power sector is only 75 milliontons higher than in the reference case. In the lowprice case there is more natural-gas-fired electricitygeneration, and as a result coal demand in the electricpower sector is 151 million tons lower than in the referencecase.Alternative assumptions about future coal miningand transportation costs affect coal prices and, consequently,demand. The two alternative coal cost casesdeveloped for AEO2006 examine the impacts on U.S.coal markets of alternative assumptions about miningproductivity, labor costs and mine equipmentcosts on the production side, and railroad productivityand rail equipment costs on the transportationside. Adjustments of about 2.5 percent from the referencecase assumptions are based on variations in historicalgrowth rates for the coal mining and railtransportation industries since 1980.In the high cost case, the average delivered coal pricein 2030, in constant 2004 dollars, is $45.39 per ton—50 percent higher than in the reference case (Figure106). As a result, U.S. coal consumption is 284 milliontons (16 percent) lower than in the reference case in2030, reflecting both a switch from coal to naturalgas, nuclear, and renewables in the electricity sectorand reduced production of coal-based synthetic liquids.In the electric power sector, 111 gigawatts ofnew coal-fired generating capacity is built by 2030 inthe high cost case—63 gigawatts less than in the referencecase. CTL production in 2030 in the high costcase totals only 0.2 million barrels per day, or 77 percentless than in the reference case.In the low cost case, the average delivered coal pricein 2030 is $21.42 per ton—29 percent lower than inthe reference case—and total coal consumption is 160million tons (9 percent) higher than in the referencecase.102 Energy Information Administration / Annual Energy Outlook 2006

Carbon Dioxide EmissionsHigher Energy Consumption ForecastIncreases Carbon Dioxide EmissionsFigure 107. Carbon dioxide emissions by sector andfuel, 2004 and 2030 (million metric tons)4,0008,114 Petroleum20045,900Natural gas2030Coal3,000ElectricityTotal carbon dioxideemissions2,0001,00002004 2030ResidentialCommercialIndustrialCO 2 emissions from the combustion of fossil fuels areproportional to fuel consumption. Among fossil fueltypes, coal has the highest carbon content, naturalgas the lowest, and petroleum in between. In theAEO2006 reference case, the shares of these fuelschange slightly from 2004 to 2030, with more coal andless petroleum and natural gas. The combined shareof carbon-neutral renewable and nuclear energy isstable from 2004 to 2030 at 14 percent. As a result,CO 2 emissions increase by a moderate average of 1.2percent per year over the period, slightly higher thanthe average annual increase in total energy use (Figure107). At the same time, the economy becomes lesscarbon intensive: the percentage increase in CO 2emissions is one-third the increase in GDP, andemissions per capita increase by only 11 percent overthe 26-year period.The factors that influence growth in CO 2 emissionsare the same as those that drive increases in energydemand. Among the most significant are populationgrowth; increased penetration of computers, electronics,appliances, and office equipment; increases incommercial floorspace; growth in industrial output;increases in highway, rail, and air travel; and continuedreliance on coal and natural gas for electric powergeneration. The increases in demand for energy servicesare partially offset by efficiency improvementsand shifts toward less energy-intensive industries.New CO 2 mitigation programs, more rapid improvementsin technology, or more rapid adoption of voluntaryprograms could result in lower CO 2 emissionslevels than projected here.TransportationElectricity generationEmissions Projections Change WithEconomic Growth AssumptionsFigure 108. Carbon dioxide emissions inthree economic growth cases, 1990-2030(million metric tons)10,0008,0006,0004,0002,000HistoryProjections01990 2004 2010 2020 2030High growthReferenceLow growthThe high economic growth case assumes highergrowth in population, labor force, and productivitythan in the reference case, leading to higher industrialoutput, higher disposable income, lower inflation,and lower interest rates. The low economicgrowth case assumes the reverse. The spread in GDPprojections increases over time, with GDP in thealternative cases varying by about 15 percent fromthe reference case in 2030.Alternative projections for industrial output, commercialfloorspace, housing, and transportation influencethe demand for energy and result in variations inCO 2 emissions (Figure 108). Emissions in 2030 are 10percent lower in the low growth case and 10 percenthigher in the high growth case. The strength of therelationship between economic growth and emissionsvaries by end-use sector. It is strongest for the industrialsector and, to a lesser extent, the transportationsector, where economic activity strongly influencesenergy use and emissions, and where fuel choices arelimited. It is weaker in the commercial and residentialsectors, where population and building characteristicshave large influences and vary less across thethree cases.In the electricity sector, changes in electricity salesacross the cases affect the amount of new, more efficientgenerating capacity required, reducing the sensitivityof energy use to GDP. However, the choice ofcoal for most new baseload capacity increases CO 2intensity in the high growth case while decreasing itin the low case, offsetting the effects of changes inefficiency across the cases.Energy Information Administration / Annual Energy Outlook 2006 103

Carbon Dioxide and Sulfur Dioxide EmissionsTechnology Advances Could ReduceCarbon Dioxide EmissionsFigure 109. Carbon dioxide emissions in threetechnology cases, 2004, 2020, and 2030(million metric tons)10,0008,0006,000ReferenceHigh technology2005 technologySulfur Dioxide Emissions Fall Sharplyin Response to Tighter RegulationsFigure 110. Sulfur dioxide emissions fromelectricity generation, 1990-2030(million short tons)20151015.9History12.110.9Projections4,0002,00055.94.64.0 3.83.702004 2020 203001990 1995 2004 2010 2015 2020 2025 2030Future CO 2 emissions depend, in part, on the timing,effectiveness, and costs of new energy technologies.The reference case assumes continuing improvementin energy-consuming and producing technologies.The high technology case assumes earlier introduction,lower costs, and higher efficiencies for energytechnologies in the end-use sectors, as well asimproved costs and efficiencies for advanced fossil-firedand new renewable generating technologiesin the electric power sector [94]. As in the referencecase, however, technology adoption is assumed to beconsistent with past patterns of market behavior.Energy use grows more slowly in the high technologycase, with prospects for greater energy savings constrainedby gradual turnover of energy-using equipmentand buildings. Increased use of renewables andless new coal-fired generating capacity accompanythe efficiency improvements in the high technologycase. As a result, CO 2 emissions in 2030 are 9 percentlower in the high technology than in the referencecase, while total energy consumption is only 8 percentlower (Figure 109).In contrast, the 2005 technology case assumes thatonly the equipment and vehicles available in 2005 willbe available through 2030, with no further improvementsin efficiency for new building shells and electricpower plants. Consequently, more energy is used,and CO 2 emissions in 2030 are 9 percent higher thanin the reference case, with the difference quantifyingthe effects of technology improvement assumptionson the reference case projections.EPA’s CAIR regulation, promulgated in March 2005,caps emissions of SO 2 for the District of Columbia and28 eastern and midwestern States that were determinedby the EPA to contribute to nonattainment ofthe NAAQS for PM 2.5 and ozone. CAIR is scheduled tosupersede Title IV of the Clean Air Act through theuse of a cap and trade approach. Phase I of CAIRcomes into effect in 2010 for SO 2. Phase II takes effectin 2015. States can achieve the required emissionsreductions by using one of two compliance options:meet the State’s emissions budget by requiring powerplants to participate in an EPA-administered interstatecap and trade system that caps emissions in twostages; or meet an individual State emissions budgetthrough measures of the State’s choosing.Power companies are projected to add flue gasdesulfurization equipment to 141 gigawatts of capacityin order to comply with State or Federal initiatives.As a result of those actions and the growing useof lower sulfur coal, SO 2 emissions drop from 10.9million short tons in 2004 to 3.7 million short tons in2030 (Figure 110). The SO 2 emissions allowanceprice rises to nearly $890 per ton in 2015 and remainsbetween $880 and $980 per ton from 2015 through2030. The reference case projections indicate thatthe level of SO 2 reductions called for in CAIR can beachieved without significantly raising electricityprices, which are determined by many factors,including natural gas prices, environmental compliancecosts, and the status of electricity deregulationactivities.104 Energy Information Administration / Annual Energy Outlook 2006

Nitrogen Oxide and Mercury EmissionsNitrogen Oxide Emissions FallAs New Regulations Take EffectFigure 111. Nitrogen oxide emissionsfrom electricity generation, 1990-2030(million short tons)866.7History6.4ProjectionsNew Environmental RegulationsReduce Mercury EmissionsFigure 112. Mercury emissions from electricitygeneration, 1995-2030 (short tons)60504044.5History51.453.337.7Projections423.72.3 2.1 2.1 2.22.230201024.018.716.615.301990 1995 2004 2010 2015 2020 2025 203001995 2000 2004 2010 2015 2020 2025 2030In the reference case, NO x emissions from electricitygeneration in the U.S. power sector fall as newregulations take effect. The required reductions areintended to reduce the formation of ground-levelozone, for which NO x emissions are a major precursor.Together with VOCs and hot weather, NO x emissionscontribute to unhealthy air quality in manyareas during the summer months.EPA’s CAIR will apply to NO x emissions from 28 easternand midwestern States and the District ofColumbia. Each State will be subject to two NO x limitsunder CAIR: a 5-month summer season limit andan annual limit. These caps are expected to stimulateadditions of emission control equipment to someexisting coal-fired power plants.National NO x emissions fall from 3.7 million shorttons in 2004 to 2.2 million short tons in 2030 inthe reference case (Figure 111). The largest decreaseoccurs in 2009, when Phase I of CAIR is implemented,and there is a smaller reduction in 2015with the start of Phase II caps. Between 2009 and2030, NO x allowance prices range from roughly$2,000 to $2,500 per ton, and they are expected to behighly volatile as the emission caps tighten. Theseprojections are indicative of the general range anddirection of the allowance prices. Power companiesare expected to add selective catalytic reductionequipment to 118 gigawatts of coal-fired capacity inorder to comply with both Federal and State initiatives;however, as with the requirements for SO 2 compliance,the CAIR NO x caps are not expected to lead tosignificantly higher electricity prices for consumers.EPA’s CAMR regulation, also promulgated in March2005, establishes a cap and trade program to reducemercury emissions from coal-fired power plants inthe United States. In addition to nationwide caps,each new and existing coal-fired power plant mustmeet mercury emissions standards based on coaltype. Emissions of mercury must be reduced in twophases: the national Phase I mercury cap is 38 shorttons in 2010, and the Phase II cap is 15 short tons in2018.Emissions of mercury depend on a variety of sitespecificfactors, including the amounts of mercuryand other compounds (such as chlorine) in the coal,the boiler type and configuration, and the presence ofpollution control equipment, such as fabric filters,electrostatic precipitators, flue gas desulfurization,and selective catalytic reduction equipment.The AEO2006 reference case assumes that States willcomply with CAMR regulations. As a result, mercuryemissions decline from 53.3 short tons in 2004 to 15.3short tons in 2030 (Figure 112). National emissionswill be slightly higher than the Phase II cap in 2018due to the use of banked allowances from earlieryears. Electricity generators are expected to retrofitabout 126 gigawatts of coal-fired capacity with ACItechnology in order to comply with the CAMR caps.Mercury allowance prices increase steadily from 2010on, to about $62,000 per pound in 2030. As with theCAIR requirements for SO 2 and NO x compliance, theCAMR mercury caps are not expected to lead to significantlyhigher electricity prices for consumers.Energy Information Administration / Annual Energy Outlook 2006 105

Forecast Comparisons

Forecast ComparisonsOnly GII produces a comprehensive energy projectionwith a time horizon similar to that of AEO2006.Other organizations address one or more aspects ofthe energy markets. The most recent projection fromGII, as well as others that concentrate on economicgrowth, international oil prices, energy consumption,electricity, natural gas, petroleum, and coal, are comparedhere with the AEO2006 projections.Economic GrowthIn the AEO2006 reference case, the projected growthin real GDP, based on 2000 chain-weighted dollars, is3.0 percent per year from 2004 to 2030 (Table 19). Forthe period from 2004 to 2025, real GDP growth in theAEO2006 reference case is similar to the averageannual growth projected in AEO2005. The AEO2006projections of economic growth are based on theAugust short-term forecast of GII, extended by EIAthrough 2030 and modified to reflect EIA’s view onenergy prices, demand, and production.The projected average annual GDP growth rate forthe United States from 2004 through 2010 rangesfrom 2.8 percent to 3.3 percent. The AEO2006 referencecase projects annual growth of 3.3 percent,matching the average annual real GDP growth projectedby the Office of Management and Budget(OMB), the Congressional Budget Office (CBO), andthe consensus Blue Chip forecast. GII and EnergyVentures Analysis, Inc. (EVA) project real GDPgrowth at 3.2 percent per year. Two other organizationsproject somewhat lower annual growth:Interindustry Forecasting at the University of Maryland(INFORUM) at 2.9 percent and Energy andEnvironmental Analysis, Inc. (EEA) at 2.8 percent.When the projection period is extended to 2015, theuncertainty in the projected rate of GDP growth isreflected in the wider range of the projections (2.5 to3.2 percent per year). AEO2006 remains in the upperhalf of the range, whereas the CBO projection shows aconsiderable slowing of GDP growth from 2010through 2015. There are few public or private projectionsof GDP growth rates for the United States thatextend to 2030. The AEO2006 reference case projectionreflects a slowing of the GDP growth rate after2020, consistent with an expected slowing of populationgrowth.World Oil PricesComparisons with other oil price projections areshown in Table 20. The world oil prices in EIA’sAEO2006 are generally toward the high end of the oilprice projections. Of the nine other publicly availablelong-term projections, only two—Petroleum IndustryResearch Associates, Inc. (PIRA) and Petroleum Economics,Ltd. (PEL)—have projections of world oilprices for specific years after 2010 that exceed theAEO2006 reference case projections. Four of thenine—GII, Altos Partners (Altos), Strategic Energyand Economic Research, Inc. (SEER), and the InternationalEnergy Agency (IEA) Reference Scenario—have prices lower than those in the AEO2006 lowprice case for at least some years. All the projections—exceptfor the price forecast from Altos, whichhas not been revised since July 2003—have raisedtheir long-term price expectations relative to lastyear’s releases.Table 19. Forecasts of annual average economicgrowth, 2004-2030Average annual percentage growthForecast 2004-2010 2010-2015 2015-2020 2020-2030AEO2005 3.2 3.1 3.0 NAAEO2006Reference 3.3 3.0 3.1 2.8Low growth 2.6 2.3 2.7 2.4High growth 3.9 3.5 3.4 3.7GII 3.2 3.0 3.0 2.8OMB 3.3 NA NA NACBO 3.3 2.6 NA NABlue Chip 3.3 3.2 NA NAINFORUM 2.9 2.5 2.6 NAEEA 2.8 2.8 2.8 2.8EVA 3.2 NA NA NANA = not available.Table 20. Forecasts of world oil prices, 2010-2030(2004 dollars per barrel)Forecast 2010 2015 2020 2025 2030AEO2005 (reference case) 27.18 28.97 30.88 32.95 NAAEO2006Reference 47.29 47.79 50.70 54.08 56.97High price 62.65 76.30 85.06 90.27 95.71Low price 40.29 33.78 33.99 34.44 33.73GII 37.82 34.06 31.53 33.50 34.50Altos 27.58 31.14 34.02 37.89 40.03IEA (reference) 35.00 36.00 37.00 38.00 39.00IEA (deferred investment) 41.00 43.50 46.00 49.00 52.00PEL 47.84 47.84 49.80 50.77 NAPIRA 44.10 49.95 63.35 NA NAEEA 46.74 43.85 42.79 41.76 NADB 31.75 31.75 31.75 31.75 31.75SEER 29.54 31.00 32.00 34.18 36.50Delphi NA 52.50 57.50 62.50 72.50NA = not available.108 Energy Information Administration / Annual Energy Outlook 2006

Forecast ComparisonsThe world oil price measure does vary by projection.In some cases, the measure is the WTI spot price,Brent equivalent, weighted average U.S. refineracquisition cost of imported crude oil, or a basket ofcrude oils. For AEO2006, EIA redefined its world oilprice path to represent the average U.S. refinersacquisition price of imported low-sulfur light crude oil(see “Issues in Focus” for discussion). Those pricesare considered comparable to the WTI prices mostoften cited in the trade press as a proxy for worldoil prices. The different price measures used in thevarious projections do not wholly explain the differentprice expectations among the projections. Forinstance, GII publishes a WTI spot price forecast thatis considerably lower than the AEO2006 referencecase prices, and PIRA publishes a WTI spot price forecastthat is considerably higher than the AEO2006reference case prices in most years (Table 20).Recent variability in crude oil prices demonstratesthe uncertainty inherent in projections for crude oilmarkets. The oil price paths projected by severalorganizations, including EIA, illustrate the uncertainty.For example, for 2010, the price range in theprojections is from a low of about $28 per barrel byAltos to a high of almost $48 per barrel projected byPEL. The range in the projections widens in 2020,from a low of $32 per barrel (GII and DB) to a high of$63 per barrel (PIRA). In 2030, the band of prices representedby the published projections narrows to $23per barrel, probably in part because the PIRA forecasthorizon ends in 2020.To construct the world oil price cases for AEO2006,EIA employed input from a Delphi group of energyanalysts. In August 2005, an informal, nonrandomsample of expert oil analysts from outside DOE wereinvited to participate, with the stipulation that theresponses were to reflect the analysts’ personal viewsand not necessarily the views of the organizationswith which they were affiliated. In addition, the analystswere told that their responses would be anonymous.Seventeen analysts were surveyed, and eightresponses were received. The median response fromthe Delphi group was generally higher than any of theother published projections, though still fallingwithin the range defined by the AEO2006 low andhigh price cases (Table 20). The group expected oilprices to continue rising through the 2005-2030 timeperiod, to nearly $73 per barrel in 2030—more than$20 per barrel higher than the nearest alternative,the Deferred Investment Scenario published by IEA.Total Energy ConsumptionThe AEO2006 projects higher growth in end-use sectorconsumption of petroleum, natural gas, and coalthan occurred from 1980 to 2004 but lower growth inelectricity consumption (Table 21). Much of the projectedgrowth in petroleum consumption is driven byincreased demand in the transportation sector, withcontinued growth in personal travel and freighttransport projected to result from demographictrends and economic expansion. Natural gas consumptionis expected to increase in the residential,commercial, and industrial sectors, despite relativelyhigh prices. Natural gas is cleaner than other fuels,does not require on-site storage, and has tended to bepriced competitively with oil for heating. Coal consumptionas a boiler fuel in the commercial andindustrial sectors is expected to decline slightly, withpotential use in new boilers limited by environmentalrestrictions; however, the projections for industrialcoal consumption include its use in CTL plants, atechnology that is expected to become competitive atthe high oil prices assumed in AEO2006.While strong growth in electricity use is projected tocontinue in the AEO2006 projections, the pace slowsfrom historical rates. Some rapidly growing applications,such as air conditioning and computers, slow aspenetration approaches saturation levels. Electricalefficiency also continues to improve, due in large partto efficiency standards, and the impacts tend to accumulatewith the gradual turnover of appliance stocks.The AEO2006 projections are generally consistentwith the outlook from GII; however, GII projectsslightly faster growth in petroleum and natural gasconsumption and slightly slower growth in electricityTable 21. Forecasts of average annual growth ratesfor energy consumption, 2004-2030 (percent)ProjectionsHistoryEnergy use 1980-2004 AEO2006 GIIPetroleum* 0.9 1.2 1.3Natural gas* 0.2 0.7 0.9Coal* -1.5 2.0 -0.4Electricity 2.2 1.6 1.5Delivered energy 0.7 1.1 1.1Electricity losses 1.9 1.2 0.9Primary energy 1.0 1.2 1.1*Excludes consumption by electricity generators in the electricpower sector; includes consumption for end-use combined heat andpower generation.Energy Information Administration / Annual Energy Outlook 2006 109

Forecast Comparisonsconsumption and losses. The differences can beattributed largely to the higher oil and natural gasprices assumed in AEO2006. Differences between theAEO2006 and GII projections for coal result from anincrease in coal use for CTL in AEO2006.ElectricityThe AEO2006 projections for the electricity generationsector assume that new generating capacity willbe built by independent power producers rather thanutilities. Retail price projections are based on averagecosts for electricity supply regions that are still regulated;marginal costs for regions that are competitive;and a mixture of average and marginal costs,weighted by the amounts of load, in regions with amix of regulated and competitive markets. As of 2005,only four electricity market regions had fully competitiveretail markets in operation; seven had mixedcompetitive and regulated retail markets; and twohad fully regulated markets. The AEO2006 casesassume that no additional retail markets will berestructured, but that partial restructuring (particularlyin wholesale markets) will lead to increasedcompetition in the electric power industry, loweroperating and maintenance costs, and early retirementof inefficient generating units.Comparison of the AEO2006 and GII projectionsshows some variation in electricity sales (Table 22).The projections for total electricity sales in 2030range from 4,828 billion kilowatthours (AEO2006 loweconomic growth case) to 5,854 billion kilowatthours(AEO2006 high economic growth case). The rate ofdemand growth ranges from 1.2 percent (AEO2006low economic growth) to 1.9 percent (AEO2006 higheconomic growth). All price projections reflect competitionin wholesale markets and slow growth in electricitydemand relative to GDP growth, exertingdownward pressure on real electricity prices through2030. Rising natural gas and coal prices balance someof the downward pressure and tend to push electricityprices up in the later years of the projections.The AEO2006 reference case shows a slight decline inreal electricity prices over the full period of the projection(except for the industrial sector), although averageprices increase slightly during the last severalyears as capacity margins tighten and natural gasprices climb. In contrast, GII projects a decline inelectricity prices over the second half of the projection,as lower delivered natural gas prices to generators($5.08 per million Btu in the GII projection,compared with $6.26 in the AEO2006 reference casein 2030) contribute to a small decrease in averageelectricity prices, from 7.6 cents per kilowatthour in2015 to 7.4 cents per kilowatthour in 2030. Thehigher natural gas price in the AEO2006 referencecase leads to an increase in average electricity price,from 7.1 cents per kilowatthour in 2015 to 7.5 centsper kilowatthour in 2030.Both the AEO2006 reference case and GII projectionsinclude some planned capacity additions in the nearterm, with the AEO2006 reference case expectingabout 29 gigawatts through 2006 and GII expectingabout 25 gigawatts. Virtually all the projected capacityadditions are natural gas fired. Both projectionsshow electricity prices falling in the near term as aresult of excess total capacity.Except for GII, all the projections for electricitydemand show the fastest growth in the commercialsector, and more additions of cycling and baseloadcapability than peaking units. All the projectionsshow significant net additions to coal-fired capacity,including 167 gigawatts through 2030 in the AEO-2006 reference case and 136 gigawatts through 2030in the GII projection. Both GII and the AEO2006 referencecase project no nuclear retirements; however,each of the three AEO2006 cases (reference and highand low economic growth) projects 6 gigawatts ofnuclear capacity additions by 2030 as a result of theincentives in EPACT2005.The fuel mix in the EVA projection differs from thatin the AEO2006 reference case and the other projections.Except for EVA, all the projections show coalmeeting about one-half and natural gas about onequarterof the growth in electricity generation capacityover the projection period. The EVA projectionassumes that legislation similar to the Clear SkiesAct—including further restrictions on SO 2 ,NO x , andmercury emissions—will be in effect by 2010. TheEVA projection also includes a tax of $5 per ton onCO 2 emissions, beginning in 2013. AEO2006 includesthe impact of the EPA’s new CAIR and CAMR regulations,which have environmental effects similar tothose of the Clear Skies Act; however, AEO2006 doesnot assume any tax on CO 2 emissions. In the EVAprojection, the combination of further environmentalrestrictions, a tax on CO 2 , and aggregate State-levelRPS requirements leads to greater growth in nonhydroelectricgeneration.110 Energy Information Administration / Annual Energy Outlook 2006

Forecast ComparisonsTable 22. Comparison of electricity forecasts, 2015 and 2030 (billion kilowatthours, except where noted)Projection 2004ReferenceAEO2006LoweconomicgrowthHigheconomicgrowthOther forecastsGII EVA EEA SEER PIRA2015Average end-use price(2003 cents per kilowatthour) 7.6 7.1 6.9 7.3 7.6 NA NA NA NAResidential 8.9 8.3 8.1 8.5 8.8 9.0 NA NA NACommercial 8.0 7.4 7.2 7.6 8.2 8.0 NA NA NAIndustrial 5.3 5.1 4.9 5.3 5.2 5.8 NA NA NANet energy for load, including CHP 3,614 4,813 4,642 4,984 4,663 4,966 4,970 4,875 4,658Coal 1,977 2,277 2,245 2,360 2,217 2,267 2,281 2,211 2,293Oil 136 120 116 126 56 37 96 126 90Natural gas a 326 1,018 929 1,069 1,080 1,286 1,323 1,238 1,004Nuclear 789 829 807 840 814 842 811 826 819Hydroelectric/other b 349 482 469 495 496 521 381 457 452Nonutility sales to grid c 26 62 57 70 NA NA 41 NA NANet imports 11 23 19 25 17 13 38 17 22Electricity sales 3,567 4,300 4,147 4,449 4,239 4,638 4,456 NA NAResidential 1,293 1,576 1,539 1,613 1,593 1,697 1,575 NA NACommercial/other d 1,253 1,620 1,583 1,650 1,493 1,718 1,602 NA NAIndustrial 1,021 1,103 1,024 1,185 1,153 1,225 1,278 NA NACapability, including CHP (gigawatts) e 965 1,002 977 1,026 1,008 1,055 1,046 NA NACoal 314 326 323 336 331 338 331 NA NAOil and natural gas 433 439 422 451 429 487 478 NA NANuclear 100 104 101 105 101 105 102 NA NAHydroelectric/other 118 133 131 134 147 125 136 NA NA2030Average end-use price(2002 cents per kilowatthour) 7.6 7.5 7.2 7.8 7.4 NA NA NA NAResidential 8.9 8.5 8.2 8.8 8.5 NA NA NA NACommercial 8.0 7.8 7.4 8.2 8.0 NA NA NA NAIndustrial 5.3 5.4 5.2 5.7 5.0 NA NA NA NANet energy for load, including CHP 3,614 6,119 5,496 6,748 5,828 NA NA 6,237 NACoal 1,977 3,381 2,835 3,897 3,032 NA NA 3,221 NAOil 136 131 121 138 27 NA NA 127 NANatural gas a 326 993 1,010 990 1,453 NA NA 1,407 NANuclear 789 871 856 871 774 NA NA 926 NAHydroelectric/other b 349 550 517 609 542 NA NA 528 NANonutility sales to grid c 26 179 143 229 NA NA NA NA NANet imports 11 14 13 15 12 NA NA 28 NAElectricity sales 3,567 5,341 4,828 5,854 5,289 NA NA NA NAResidential 1,293 1,897 1,759 2,036 2,001 NA NA NA NACommercial/other d 1,253 2,182 1,997 2,366 1,926 NA NA NA NAIndustrial 1,021 1,262 1,073 1,453 1,362 NA NA NA NACapability, including CHP (gigawatts) e 965 1,248 1,134 1,362 1,209 NA NA NA NACoal 314 481 405 555 449 NA NA NA NAOil and natural gas 433 513 483 545 501 NA NA NA NANuclear 100 109 107 109 101 NA NA NA NAHydroelectric/other 118 145 139 154 158 NA NA NA NAa Includes supplemental gaseous fuels. b “Other” includes conventional hydroelectric, pumped storage, geothermal, wood, wood waste,municipal waste, other biomass, solar and wind power, plus a small quantity of petroleum coke. c For AEO2006, includes only net sales fromcombined heat and power plants. d “Other” includes sales of electricity to government, railways, and street lighting authorities. e EIAcapacity is net summer capability, including combined heat and power plants. GII capacity is nameplate, excluding cogeneration plants.CHP = combined heat and power. NA = not available.Sources: 2004 and AEO2006: AEO2006 National Energy Modeling System, runs AEO2006.D111905A (reference case), LM2006.D113005A (low economic growth case), and HM2006.D112505B (high economic growth case). GII: Global Insight, Inc., Summer 2005 U.S.Energy Outlook (August 2005). EVA: Energy Ventures Analysis, Inc., FUELCAST: Long-Term Outlook (August 2005). EEA: Energy andEnvironmental Analysis, Inc., EEA’s Compass Service Base Case (October 2005). SEER: Strategic Energy and Economic Research, Inc.,2005 Energy Outlook (October 2005). PIRA: PIRA Energy Group (October 2005).Energy Information Administration / Annual Energy Outlook 2006 111

Forecast ComparisonsNatural GasPublished projections of natural gas prices, production,consumption, and imports (Table 23) differ considerably.The differences highlight the uncertaintyof future market trends. Because the projectionsdepend heavily on the underlying assumptions thatshape them, the assumptions made in each should beconsidered when they are compared.The AEO2006 reference case in general projectslower total natural gas consumption than in the otherprojections, and it is the only one showing a period ofdecline. The exception is in the early part of the projectionperiod: in 2015, PIRA and Deutsche Bank AG(DB) project lower natural gas consumption than theAEO2006 reference case, but by 2025 the AEO2006reference case projects lower consumption than anyof the others. The primary reason is that AEO2006expects a stronger demand response to higher naturalgas prices, particularly in the electricity generationsector.The highest projected level of total natural gas consumptionis in the EVA projection, due to stronggrowth in natural gas consumption for electric powerTable 23. Comparison of natural gas forecasts, 2015, 2025, and 2030 (trillion cubic feet, except where noted)Projection 2004AEO2006referencecaseOther forecastsGII a EEA b EVA PIRA DB SEER Altos2015Dry gas production c 18.46 20.36 19.19 21.12 18.64 d 17.61 21.38 19.68 20.74Net imports 3.40 5.10 6.80 7.11 9.67 7.33 4.30 7.86 7.92Pipeline 2.81 2.05 e 2.17 2.82 4.78 3.28 1.75 3.00 1.82LNG 0.59 3.05 4.63 4.29 4.89 4.05 2.55 4.85 6.10Consumption 22.41 25.91 26.16 27.98 28.32 25.32 25.67 28.18 NAResidential 4.88 5.36 5.15 5.49 5.33 5.24 5.53 5.45 5.41Commercial 3.00 3.36 3.09 3.35 3.41 3.53 3.53 3.28 3.54Industrial f 7.41 8.08 7.57 g 6.98 h 7.99 i 6.61 j 8.17 7.83 7.53Electricity generators k 5.32 7.14 8.44 l 10.08 m 9.42 8.01 n 6.63 9.61 9.30Other o 1.80 1.97 1.92 2.08 2.17 p 1.95 1.81 2.01 NALower 48 wellhead price(2004 dollars per thousand cubic feet) 5.49 4.52 4.73 5.91 5.53 5.55 q 5.03 4.65 4.15End-use prices(2004 dollars per thousand cubic feet)Residential 10.72 10.11 9.21 9.33 NA NA NA 9.68 NACommercial 9.38 8.37 8.11 8.57 NA NA NA 7.97 NAIndustrial f 6.29 5.32 6.09 r 6.81 NA NA NA 5.75 NAElectricity generators k 6.07 5.21 5.13 6.62 NA NA NA 5.32 NA2025Dry gas production c 18.46 21.16 20.46 21.38 19.27 d NA 18.95 21.53 25.77Net imports 3.40 5.37 8.64 8.89 11.80 NA 8.19 8.47 7.69Pipeline 2.81 1.24 e 1.61 1.81 3.64 NA 4.75 1.90 0.70LNG 0.59 4.13 7.03 7.07 8.16 NA 3.44 6.57 6.99Consumption 22.41 26.99 29.28 30.33 31.08 NA 27.74 30.44 NAResidential 4.88 5.57 5.61 5.88 5.44 NA 6.11 5.89 6.09Commercial 3.00 3.77 3.34 3.56 3.76 NA 3.99 3.49 4.19Industrial f 7.41 8.51 8.14 g 7.64 h 8.95 i NA 9.03 8.37 7.73Electricity generators k 5.32 7.05 10.10 l 11.14 m 10.55 NA 6.97 10.50 11.37Other o 1.80 2.08 2.09 2.12 2.38 p NA 1.64 2.19 NALower 48 wellhead price(2003 dollars per thousand cubic feet) 5.49 5.43 4.52 6.45 6.07 NA 5.03 5.13 5.67End-use prices(2003 dollars per thousand cubic feet)Residential 10.72 11.10 8.82 9.71 NA NA NA 9.92 NACommercial 9.38 9.11 7.73 8.99 NA NA NA 8.30 NAIndustrial j 6.29 6.18 5.81 r 7.22 NA NA NA 6.07 NAElectricity generators o 6.07 6.02 4.90 6.86 NA NA NA 5.61 NANA = not available. See notes and sources at end of table.112 Energy Information Administration / Annual Energy Outlook 2006

Forecast Comparisonsgeneration. Altos projects the strongest growth in residentialand commercial sector natural gas consumptionthrough both 2025 and 2030, whereas the GIIand EVA projections have the lowest projected consumptionlevels. The AEO2006 reference case projectionfor residential natural gas consumption in 2030is lower than all but the EVA projection, but its commercialsector projection is higher than the GII, EVA,and SEER projections. Natural gas consumption inthe industrial and electric power sectors is more difficultto compare, given potential definitional differences.The combined total of industrial and electricpower sector natural gas consumption from 2004 to2030 is projected to grow the fastest in the EVA andAltos projections; the DB projection shows muchslower growth but still faster than is projected in theAEO2006 reference case. The DB combined total in2030 exceeds the AEO2006 reference case by lessthan 10 percent, whereas the GII, EVA, SEER, andAltos projections all exceed the AEO2006 by morethan 25 percent.Domestic natural gas production provides a decreasingshare and net imports an increasing share of totalnatural gas supply in all the projections. The EVAprojection shows the greatest increase in the netimport share of supply, at more than 41 percent oftotal supply in 2030. More than 34 percent of supplyis projected to come from imports in 2030 in the DBprojection, and GII and SEER both show net importsTable 23. Comparison of natural gas forecasts, 2015, 2025, and 2030 (continued)(trillion cubic feet, except where noted)Projection 2004AEO2006referencecaseOther forecastsGII a EEA b EVA PIRA DB SEER Altos2030Dry gas production c 18.46 20.83 21.40 NA 18.96 d NA 18.95 21.70 28.13Net imports 3.40 5.57 9.06 NA 13.30 NA 9.86 9.33 7.92Pipeline 2.81 1.22 e 1.37 NA 2.80 NA 1.75 1.00 0.20LNG 0.59 4.36 7.68 NA 10.50 NA 8.11 8.33 7.72Consumption 22.41 26.86 30.64 NA 32.39 NA 28.81 31.56 NAResidential 4.88 5.64 5.84 NA 5.49 NA 6.42 6.12 6.48Commercial 3.00 3.99 3.48 NA 3.96 NA 4.20 3.63 4.56Industrial f 7.41 8.81 8.48 g NA 9.45 i NA 9.49 8.73 7.85Electricity generators k 5.32 6.38 10.67 l NA 11.01 NA 7.14 10.85 12.54Other o 1.80 2.04 2.17 NA 2.48 p NA 1.56 2.24 NALower 48 wellhead price(2004 dollars per thousand cubic feet) 5.49 5.92 4.65 NA 6.52 NA 5.02 5.42 6.30End-use prices(2004 dollars per thousand cubic feet)Residential 10.72 11.67 8.86 NA NA NA NA 10.16 NACommercial 9.38 9.58 7.79 NA NA NA NA 8.60 NAIndustrial f 6.29 6.65 5.90 r NA NA NA NA 6.37 NAElectricity generators k 6.07 6.41 5.02 NA NA NA NA 5.92 NANA = not available.a February 2005 (previously DRI-WEFA). Conversion factors: 1,000 cubic feet = 1.027 million Btu for production, 1.028 million Btu forend-use consumption, 1.019 million Btu for electric power. b The EEA projection shows a cyclical price trend; forecast values for an isolatedyear may be misleading. c Does not include supplemental fuels. d Includes supplemental fuels. e Includes LNG imports into Florida via theBahamas. f Includes consumption for industrial combined heat and power (CHP) plants and a small number of electricity-only plants;excludes consumption by nonutility generators. g Excludes gas used in cogeneration or other nonutility generation. h Includes natural gasconsumed in cogeneration. i Includes transportation fuel consumed in natural gas vehicles. j Excludes gas demand for nonutility generation.k Includes consumption of energy by electricity-only and CHP plants whose primary business is to sell electricity, or electricity and heat, tothe public; includes electric utilities, small power producers, and exempt wholesale generators. l Includes gas used in cogeneration or othernonutility generation. m Includes independent power producers; excludes cogenerators. n Equals the sum of natural gas demand fornonutility generation (NUG) and for utility generation. o Includes lease, plant, and pipeline fuel and fuel consumed in natural gas vehicles.p Includes lease, plant, and pipeline fuel. q Henry Hub daily cash price for natural gas, in 2004 dollars per thousand cubic feet. r On-systemsales or system gas (i.e., does not include gas delivered for the account of others).Sources: 2004 and AEO2006: AEO2006 National Energy Modeling System, run AEO2006.D111905A (reference case). GII: GlobalInsight, Inc., Summer 2005 U.S. Energy Outlook (August 2005). EEA: Energy and Environmental Analysis, Inc., EEA’s Compass ServiceBase Case (October 2005). EVA: Energy Ventures Analysis, Inc., FUELCAST: Long-Term Outlook (August 2005). PIRA: PIRA EnergyGroup (October 2005). DB: Deutsche Bank AG, e-mail from Adam Sieminski on October 31, 2005. SEER: Strategic Energy and EconomicResearch, Inc., 2005 Energy Outlook (October 2005). Altos: Altos Partners North American Regional Gas Model (NARG) Long-Term BaseCase (October 7, 2005).Energy Information Administration / Annual Energy Outlook 2006 113

Forecast Comparisonsproviding about 30 percent of total natural gas supply.The AEO2006 reference case and Altos projectthat net imports will meet the smallest share of totalsupply—21 percent and 22 percent, respectively—in2030. Most of the projections show a notable declinein pipeline imports over the forecast period. Only DBshows an increase from 2015 to 2025. EVA’s pipelineimport projection, although significantly greater thanthe rest, also declines after 2015. Much of the variationin imports reflects different projections of netLNG imports in 2030, ranging from a low of 4.4 trillioncubic feet in the AEO2006 reference case to 10.5trillion cubic feet in the EVA projection.The AEO2006 reference case projections for wellheadnatural gas prices in 2025 and 2030 fall within therange of the other projections, with the EEA, EVA,and Altos projections higher than AEO2006 and theothers lower. In the earlier years, however, all theprojections with the exception of Altos show wellheadnatural gas prices exceeding those in the AEO2006reference case. Of the three projections that projectend-use prices for 2030 (AEO2006, GII, and SEER),the AEO2006 reference case and SEER show thehighest end-use-to-wellhead margins for the electricpower sector ($0.50 and $0.51, respectively). TheAEO2006 reference case shows the lowest end-useto-wellheadmargins for the industrial sector. WhileGII’s margins for the electric power sector are thelowest, some of the difference may be definitional. Forthe residential and commercial sectors, the projectedmargins in the AEO2006 reference case exceed theother projections by more than 15 percent.PetroleumAs discussed earlier in this report, crude oil pricesin the AEO2006 reference case are substantiallyhigher than they were in earlier AEOs. They arealso considerably higher than those in most of theother projections. The AEO2006 reference case showsthe weighted average refiners acquisition cost ofimported crude oil (the price basis used in most of theother forecasts) ranging from $43 to $50 per barrel(2004 dollars) between 2015 and 2030 and the averagerefiners acquisition cost of imported low-sulfurlight crude oil (the reference price used in AEO2006)ranging from $48 to $57 per barrel (2004 dollars) overthe same period. DB assumes that the refiners acquisitioncost of crude oil will average $31.75 per barrelfrom 2010 through 2030; GII assumes that the refinersacquisition cost of crude oil will be between $28and $31 per barrel from 2015 through 2030. PIRAgives its oil price forecast in terms of WTI, alow-sulfur, light crude oil, assuming prices of $50 perbarrel in 2015 and $63 per barrel in 2020.Despite much lower crude oil price projections, GIIand DB project gasoline consumption levels that areessentially the same as those in the AEO2006 referencecase (Table 24). The GII and DB projections forgasoline demand are within 1 percent of the AEO2006reference case from 2015 to 2030. PIRA sees slowergrowth in gasoline demand, 14 percent below theAEO2006 reference case in 2015, due to more rapidimprovement in vehicle efficiency.In comparison with the AEO2006 reference case, projecteddistillate consumption is about 2 percent lowerin the DB and PIRA projections in 2015 and 5 percentlower in 2030 in the DB projection. GII also projectslower levels of distillate consumption than theAEO2006 reference case, 6 percent less in 2015 and13 percent less in 2030. Most of the variation isaccounted for by the projected level of highway dieselconsumption.The projected pattern of growth in jet fuel consumptionvaries significantly by projection, and the basisof the variation is not always clear. Relative to theAEO2006 reference case, PIRA projects slightlyhigher jet fuel consumption in 2015, whereas GII projectshigher jet fuel consumption only toward the endof the projection (25 percent higher in 2030 but 4 percentlower in 2015). DB also projects lower jet fuelconsumption in the middle years, 9 percent below theAEO2006 reference case in 2015, but is nearly identicalwith the AEO2006 reference case in the later yearsof the projection.The projections also differ substantially on the projectedfuture use of residual fuel oil. PIRA and GIIproject a steady decline in residual fuel oil consumption,but DB sees some growth in the future. In theGII projection, residual fuel oil consumption is 3 percentbelow that in the AEO2006 reference case in2015 and 18 percent below in 2030. Both GII andPIRA project deep declines in residual fuel oil consumptionfor electricity generation. The AEO2006reference case projects more modest reductionsthrough 2015 and then slow growth for the remainderof the projection. The DB projections are 14 percentand 17 percent above the AEO2006 reference case in2015 and 2030, respectively.114 Energy Information Administration / Annual Energy Outlook 2006

Forecast ComparisonsDomestic crude oil production declines in all the projections,but at different rates. As compared with theAEO2006 reference case, domestic crude oil productiondeclines more rapidly in the earlier years andmuch more slowly in the later years of the GII projection.GII projects domestic crude oil production 14percent lower than in the AEO2006 reference case in2015 but essentially the same in 2030. DB and PIRAproject a much more rapid decline in domestic crudeoil production: both are about 15 percent below theAEO2006 reference case in 2015, and DB projects afurther decline, to 19 percent below the AEO2006 referencecase in 2030.The projections do not agree on domestic productionof NGL. The AEO2006 reference case projects NGLproduction slightly above current levels in 2015 and2030, with peak production in 2020. DB is bearishon NGL production, projecting 19 percent lowerlevels than in the AEO2006 reference case in 2015and 42 percent lower in 2030. GII, on the other hand,is bullish on NGL production, projecting domesticproduction 24 percent above the levels in theAEO2006 reference case in 2015 and 38 percent abovein 2030. EVA and DB project the lowest totals ofdomestic crude oil and NGL production in 2015 and2030.Declining domestic production of crude oil and risingpetroleum product demand imply greater dependenceon imports in all the projections. The decreases incrude oil production are offset somewhat by projectedincreases in NGL production in the AEO2006 referencecase and GII. DB projects substantial declines incrude oil and NGL production and therefore projectsthe highest levels of net imports of crude and petroleumproducts. DB projects import shares 9 percentagepoints above the AEO2006 reference case in 2015and 14 percentage points above in 2030. GII projectsimport shares that are 8 percentage points above theAEO2006 reference case in 2015 and 2030.AEO2006 also includes alternative price cases. TheAEO2006 low price case assumes that the averageTable 24. Comparison of petroleum forecasts, 2015 and 2030 (million barrels per day, except where noted)Projection 2004AEO2006Other forecastsReference Low price High price GII DB EVA PIRA2015Crude oil and NGL production 7.23 7.72 7.34 6.49 6.56 NA 7.88 7.61Crude oil 5.42 5.84 5.02 4.98 NA 4.99 5.99 5.76Natural gas liquids 1.81 1.88 2.32 1.51 NA NA 1.89 1.85Total net imports 12.11 13.23 15.08 15.31 NA 14.37 14.06 11.87Crude oil 10.06 10.47 11.28 NA NA 11.74 11.06 9.65Petroleum products 2.05 2.76 3.79 NA NA 2.63 3.00 2.22Petroleum demand 20.76 23.53 23.71 23.43 NA 23.01 24.48 22.21Motor gasoline 9.10 10.63 10.69 10.39 NA 9.14 11.07 9.85Jet fuel 1.63 2.06 1.98 1.88 NA 2.11 2.09 2.03Distillate fuel 4.06 4.91 4.60 4.81 NA 4.83 5.05 4.72Residual fuel 0.87 0.73 0.71 0.83 NA 0.72 0.83 0.66Other 5.10 5.20 5.74 5.51 NA 6.21 5.44 4.95Import share of product supplied (percent) 58 56 64 65 NA 62 57 532030Crude oil and NGL production 7.23 6.44 7.17 4.78 4.70 NA 6.41 6.85Crude oil 5.42 4.57 4.59 3.69 NA NA 4.49 4.96Natural gas liquids 1.81 1.87 2.58 1.09 NA NA 1.92 1.89Total net imports 12.11 17.24 19.69 21.13 NA NA 20.21 13.28Crude oil 10.06 13.51 13.01 NA NA NA 15.51 11.24Petroleum products 2.05 3.73 6.67 NA NA NA 4.70 2.04Petroleum demand 20.76 27.57 28.24 27.74 NA NA 29.57 25.17Motor gasoline 9.10 12.49 12.59 12.25 NA NA 13.68 10.96Jet fuel 1.63 2.31 2.89 2.29 NA NA 2.33 2.09Distillate fuel 4.06 6.09 5.31 5.81 NA NA 6.29 5.99Residual fuel 0.87 0.78 0.64 0.91 NA NA 1.01 0.70Other 5.10 5.89 6.80 6.49 NA NA 6.26 5.44Import share of product supplied (percent) 58 62 70 76 NA NA 68 53NA = Not available.Sources: 2004 and AEO2006: AEO2006 National Energy Modeling System, runs AEO2006.D111905A (reference case), LP2006.D113005A (low price case), and HP2006.D120105A (high price case). GII: Global Insight, Inc., Summer 2005 U.S. Energy Outlook (August2005). DB: Deutsche Bank AG, e-mail from Adam Sieminski on October 31, 2005. EVA: Energy Ventures Analysis, Inc., FUELCAST:Long-Term Outlook (August 2005). PIRA: PIRA Energy Group (October 2005).Energy Information Administration / Annual Energy Outlook 2006 115

Forecast Comparisonsrefiners acquisition cost of imported crude oil willremain at about $28 per barrel from 2015 to 2030(2004 dollars), and that the average refiners acquisitioncost of imported low-sulfur light crude oil willremain at about $34 per barrel over the same period.The AEO2006 low price case is somewhat lower thanGII’s crude oil price path. The AEO2006 high pricecase assumes that the average refiners acquisitioncost of imported crude oil will range between $72 and$90 per barrel from 2015 to 2030, and that the averagerefiners acquisitions cost of imported low-sulfurlight crude oil will range between $76 and $96 perbarrel over the same period. The crude oil prices inthe AEO2006 high price case are well above PIRA’sprojected levels. The AEO2006 low price case showsthe highest levels of total petroleum demand in 2015and 2030, and the AEO2006 high price case shows thelowest. The projected demand reduction in the AEO-2006 high price case also results in the least relianceon imports to meet petroleum demand in 2015 and2030. The DB projection shows the greatest relianceon petroleum imports, because it assumes the lowestlevels of domestic crude oil and NGL production.CoalThe coal projections for the AEO2006 reference caseand economic growth cases (Table 25) incorporateCAAA90, CAIR, and CAMR. EVA’s forecast assumeslegislation similar to the Clear Skies Act but alsoincludes a fee of $5 per ton on CO 2 emissions, beginningin 2013. The AEO2006, PIRA, and GII projectionsdo not include assumptions about reductions inCO 2 emissions for the United States. In addition toenvironmental assumptions, differences among theTable 25. Comparison of coal forecasts, 2015, 2025, and 2030 (million short tons, except where noted)Projection 2004ReferenceAEO2006LoweconomicgrowthHigheconomicgrowth116 Energy Information Administration / Annual Energy Outlook 2006Other forecastsPIRA EVA GII2015Production 1,125 1,272 1,251 1,318 1,250 1,234 1,149Consumption by sectorElectric power 1,015 1,161 1,145 1,199 1,171 1,140 1,071Coke plants 24 22 21 23 NA 29 19Coal-to-liquids 0 22 19 27 NA NA NAIndustrial/other 65 71 69 72 88 a 65 66Total 1,104 1,276 1,254 1,321 1,259 1,234 1,156Net coal exports 20.7 -4.8 -4.8 -4.8 -8.0 -17.3 -7.7Exports 48.0 22.0 22.0 22.0 NA 28.0 28.6Imports 27.3 26.7 26.7 26.8 NA 45.3 36.3Minemouth price(2004 dollars per short ton) 20.07 20.39 20.04 20.67 NA 19.69 b 17.82 d(2004 dollars per million Btu) 0.98 1.01 0.99 1.02 NA 0.99 c 0.86 dAverage delivered priceto electricity generators(2004 dollars per short ton) 27.43 28.12 27.74 28.50 NA 29.45 b 28.17 e(2004 dollars per million Btu) 1.36 1.40 1.39 1.42 NA 1.48 b 1.362025Production 1,125 1,530 1,394 1,710 NA 1,404 1,296Consumption by sectorElectric power 1,015 1,354 1,248 1,486 NA 1,329 1,226Coke plants 24 21 19 23 NA 26 16Coal-to-liquids 0 146 115 192 NA NA NAIndustrial/other 65 71 68 73 NA 60 67Total 1,104 1,592 1,450 1,774 NA 1,415 1,309Net coal exports 20.7 -62.8 -57.9 -65.5 NA -29.2 -15.1Exports 48.0 19.6 19.6 18.4 NA 30.1 23.4Imports 27.3 82.4 77.4 84.0 NA 59.3 38.5Minemouth price(2004 dollars per short ton) 20.07 20.63 19.40 21.73 NA 20.15 b 16.12 d(2004 dollars per million Btu) 0.98 1.03 0.98 1.09 NA 1.02 c 0.78 dAverage delivered priceto electricity generators(2004 dollars per short ton) 27.43 29.02 27.48 30.87 NA 30.12 b 25.84 e(2004 dollars per million Btu) 1.36 1.44 1.37 1.52 NA 1.53 b 1.25Btu = British thermal unit. NA = Not available. See notes and sources at end of table.

Forecast ComparisonsAEO2006, EVA, PIRA, and GII projections reflectvariation in other assumptions, including those abouteconomic growth, the natural gas outlook, and worldoil prices.While all the projections show increases in coal consumptionover their projection horizons, theAEO2006 reference case projects the highest level oftotal coal consumption. Given its more restrictiveenvironmental assumptions after 2012 and an averageeconomic growth rate of 2.5 percent per year from2004, EVA projects lower levels of coal consumption(11 percent lower in 2025) than the AEO2006 referencecase. The EVA and PIRA projections for totalcoal consumption in the 2015-2020 period mostclosely resemble those in the AEO2006 low economicgrowth case. GII’s projection, which does not includea carbon tax, has the lowest projection of total coalconsumption. Although the GII projection shows 21percent less total coal consumption than theAEO2006 reference case in 2030, GII’s outlook forcoal consumption in the electric power sector in 2030is virtually identical to that in the AEO2006 low economicgrowth case.In contrast to the AEO2006 reference case, the otherprojections show natural gas with a larger share ofelectricity generation than coal’s. GII, PIRA, andEVA expect imports of LNG to be greater than projectedin the AEO2006 reference case. Although EVAand the AEO2006 reference case project similar levelsof generation in the electric power sector, theAEO2006 reference case also projects 19 gigawatts ofgeneration capacity at CTL plants by 2030, representing11 percent of total coal consumption in 2030.For coke plants, both GII and the AEO2006 referencecase project declining consumption of coal. EVA differsfrom the other projections and projects anincrease in coal consumption at coke plants, peakingat around 30 million tons before falling to 26 milliontons in 2025—2 million tons higher than 2004Table 25. Comparison of coal forecasts, 2015, 2025, and 2030 (continued)(million short tons, except where noted)Projection 2004ReferenceAEO2006LoweconomicgrowthHigheconomicgrowthOther forecastsPIRA EVA GII2030Production 1,125 1,703 1,497 1,936 NA NA 1,395Consumption by sectorElectric power 1,015 1,502 1,331 1,680 NA NA 1,330Coke plants 24 21 19 23 NA NA 14Coal-to-liquids 0 190 153 247 NA NA NAIndustrial/other 65 72 68 75 NA NA 67Total 1,104 1,784 1,571 2,025 NA NA 1,411Net coal exports 20.7 -82.7 -69.3 -89.0 NA NA -18.7Exports 48.0 16.7 16.4 16.8 NA NA 22.3Imports 27.3 99.4 85.7 105.8 NA NA 41.0Minemouth price(2004 dollars per short ton) 20.07 21.73 19.91 23.05 NA NA 15.65 d(2004 dollars per million Btu) 0.98 1.09 1.00 1.15 NA NA 0.76 dAverage delivered priceto electricity generators(2004 dollars per short ton) 27.43 30.58 28.28 32.79 NA NA 25.23 e(2004 dollars per million Btu) 1.36 1.51 1.41 1.61 NA NA 1.22Btu = British thermal unit. NA = Not available.a Includes coal consumed at coke plants.b The average coal price is a weighted average of the projected spot market price for the electric power sector only and was converted from2005 dollars to 2004 dollars to be consistent with AEO2006.c Estimated by dividing the minemouth price in dollars per short ton by the average heat content of coal delivered to the electric powersector.d The minemouth prices are average prices for the electric power sector only and are calculated as a weighted average from Census regionprices.e Calculated by multiplying the delivered price of coal to the electric power sector in dollars per million Btu by the average heat content ofcoal delivered to the electric power sector.Sources: 2004 and AEO2006: AEO2006 National Energy Modeling System, runs AEO2006.D111905A (reference case), LM2006.D113005A (low economic growth case), and HM2006.D112505B (high economic growth case). PIRA: PIRA Energy Group (October 2005).EVA: Energy Ventures Analysis, Inc., FUELCAST: Long-Term Outlook (August 2005). GII: Global Insight, Inc., U.S. Energy Outlook(Summer 2005).Energy Information Administration / Annual Energy Outlook 2006 117

Forecast Comparisonsconsumption. In the GII projection, coke plants consumeonly 14 million tons of coal in 2030, comparedwith 21 million tons in the AEO2006 reference case.The AEO2006 reference case shows no change inindustrial/other coal consumption, whereas EVA projectsa drop in industrial/other consumption, to 60million short tons in 2025.In the GII projections, minemouth coal prices (electricpower sector only) appear to peak by 2010 andthen fall below 2004 levels by 2030 (in real dollars).The AEO2006 reference case shows a similar downwardtrend after 2010 in its average nationalminemouth price (all sectors combined) through2020; however, prices rise after 2020, in response tosubstantial growth in coal demand from the electricpower sector and CTL. GII’s average delivered priceof coal to the electric power sector in 2030 is 19 percentlower than the AEO2006 reference case (on a Btubasis). The average delivered price of coal to the electricitysector in the GII forecast is still 13 percentlower (on a Btu basis) than the AEO2006 low economicgrowth case, despite comparable levels of coalconsumption in the electricity sector.All the forecasts reviewed meet coal demand primarilythrough domestic production. AEO2006 projectsthe largest increase in production over the forecasthorizon, 51 percent higher in 2030 than in 2004. Aswith consumption, the PIRA and EVA projections forcoal production most closely resemble those in theAEO2006 low economic growth case. GII projects coalconsumption levels for 2015, 2025, and 2030 that aremore than 100 million tons less than projected in theAEO2006 reference case.In all the projections, gross exports of coal represent asmall and declining part of domestic coal production.EVA projects the most exports, 30 million tons in2025, and the other projections are around 20 milliontons. The AEO2006 reference case shows coal exportsfalling to 17 million tons in 2030, and GII projects 22million tons. In the AEO2006 reference case, theexport share of total U.S. coal production falls from 4percent in 2004 to roughly 1 percent in 2030. Currently,coal is the only domestic U.S. energy resourcefor which exports exceed imports. All the projectionsexpect the United States to become a net importer ofcoal over the projection period. GII projects the lowestlevel of coal imports, only 14 million tons in 2030. TheAEO2006 reference case projection for coal imports in2025 is 23 million tons higher than the EVA projection,which is the next highest.118 Energy Information Administration / Annual Energy Outlook 2006

List of AcronymsACI Activated carbon injectionAD Associated-dissolved (natural gas)AEO Annual Energy OutlookAEO2005 Annual Energy Outlook 2005AEO2006 Annual Energy Outlook 2006Altos Altos PartnersANWR Arctic National Wildlife RefugeAPI American Petroleum InstituteBLGCC Black liquor gasification coupled with acombined-cycle power plantBOE Barrels of oil equivalentBTL Biomass-to-liquidsBtu British thermal unitsCAAA90 Clean Air Act Amendments of 1990CAFE Corporate average fuel economyCAIR Clean Air Interstate RuleCAMR Clean Air Mercury RuleCARB California Air Resources BoardCBO Congressional Budget OfficeCHP Combined heat and powerCO 2 Carbon dioxideCPI Consumer price indexCRI Color rendering indexCTL Coal-to-liquidsDB Deutsche Bank AGDCL Direct coal liquefactionDOE U.S. Department of EnergyE85 Fuel containing a blend of 70 to 85 percent ethanolEEA Energy and Environmental Analysis, Inc.EIA Energy Information AdministrationEMF Stanford University Energy Modeling ForumEPA U.S. Environmental Protection AgencyEPACT1992 Energy Policy Act of 1992EPACT2005 Energy Policy Act of 2005ERO Electric Reliability OrganizationETBE Ethyl tertiary butyl etherEVA Energy Ventures Analysis, Inc.FERC Federal Energy Regulatory CommissionGATS Generation Attribute Tracking SystemGDP Gross domestic productGHG Greenhouse gasGII Global Insight, Inc.GT-MH Gas-Turbine Modular Helium reactorGTL Gas-to-liquidsH 2 Molecular hydrogenHCCI Homogeneous charge compression ignitionICL Indirect coal liquefactionIEA International Energy AgencyIGCC Integrated gasification combined-cycleINFORUM Interindustry Forecasting at the University ofMarylandIRAC Average imported refiner acquisition cost of crudeoil to the United StatesIRIS International Reactor Innovative and Secure reactorLED Light-emitting diodeLFG Landfill gasLNG Liquefied natural gasLPGLWRMACTMRETSMSWMTBENANAAQSNARGNASNEMSNEPOOLNGLNHTSAnmNO xNRCNYMEXOLEDOMBOPECPATHPBMRPDVSAPELPIRAPM 2.5ppmPTCPURPAPVR&DRD&DRFGRFSRPSSAFETEA-LUSAGDSCRSEERSIPSNCRSO 2SSLSyncrudeTAMETANULSDUSGSVOCWPIWREGISWTIZEHLiquefied petroleum gasLight-water reactorMaximum Achievable Control TechnologyMidwest Renewable Energy Tracking SystemMunicipal solid wasteMethyl tertiary butyl etherNonassociated (natural gas)National Ambient Air Quality StandardsAltos Partners North American Regional Gas ModelNational Academy of SciencesNational Energy Modeling SystemNew England Power PoolNatural gas liquidsNational Highway Traffic Safety AdministrationNanometer (one-billionth of a meter)Nitrogen oxidesU.S. Nuclear Regulatory CommissionNew York Mercantile ExchangeOrganic light-emitting diodeOffice of Management and BudgetOrganization of the Petroleum Exporting CountriesPartnership for Advanced Technology in HousingPebble Bed Modular ReactorPetroleos de Venezuela SAPetroleum Economics, Ltd.Petroleum Industry Research Associates, Inc.Fine particulate matterParts per millionProduction tax creditPublic Utility Regulatory Policies ActPhotovoltaicResearch and developmentResearch, development, and demonstrationReformulated gasolineRenewable fuels standardRenewable portfolio standardSafe, Accountable, Flexible, and EfficientTransportation Equity Act: A Legacy for UsersSteam-assisted gravity drainageSelective catalytic reductionStrategic Energy and Economic Research, Inc.State Implementation PlanSelective noncatalytic reductionSulfur dioxideSolid-state lightingSynthetic crudeTertiary amyl methyl etherTotal acid numberUltra-low-sulfur diesel fuelU.S. Geological SurveyVolatile organic compoundWholesale price indexWestern Renewable Energy Generation InformationTracking SystemWest Texas Intermediate (crude oil)Zero Energy HomeEnergy Information Administration / Annual Energy Outlook 2006 119

Notes and SourcesText NotesLegislation and Regulations[1] For the complete text of the Energy Policy Act of 2005,see web site http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=109_cong_public_laws&docid=f:publ058.109.pdf.[2] See, for example, web site http://energy.senate.gov/public/_files/PostConferenceBillSummary.doc.[3] Joint Committee on Taxation, Description and TechnicalExplanation of the Conference Agreement of H.R.6, TitleXIII, The “Energy Tax Incentives Act of 2005,”JCX-60-05 (Washington, DC, July 28, 2005), pp. 6-8, website www. house.gov/jct/x-60-05.pdf.[4] Other Federal credit assistance programs, such as thatcreated by the Transportation Infrastructure Financeand Innovation Act of 1998 (TIFIA), have used loan guaranteesto leverage limited Federal resources and stimulateprivate capital investment. With a budget authorizationof $130 million for fiscal year 2003, the TIFIAprogram was able to support loans valued at $2.6 billion.See web site http://tifia.fhwa.dot.gov.[5] Other States that have adopted the California emissionstandards include Connecticut, Maine, Massachusetts,New Jersey, New York, Rhode Island, Vermont, andWashington.[6] On December 7, 2004, the Alliance of Automobile Manufacturersand several California auto dealerships filedsuit in the U.S. District Court in Fresno, California,against A.B. 1493.[7] Energy Information Administration, Annual EnergyOutlook 2005, DOE/EIA-0383(2005) (Washington, DC,February 2005), pp. 27-31, web site www.eia.doe.gov/oiaf/archive/aeo05/index.html.[8] National Highway Traffic Safety Administration,Average Fuel Economy Standards for Light TrucksModel Years 2008-2011, Notice of Proposed Rulemaking,49 CFR Parts 523, 533, and 537, Docket No.2005-22223, RIN 2127-AJ61 (Washington, DC, August2005), web site www.nhtsa.dot.gov/cars/rules/rulings/LightTrucksRuling-2008-2001/ProposedRulemaking/CAFE-LigthTrucks-PR-24Aug05.pdf.[9] Energy Information Administration, “State RenewableEnergy Requirements and Goals: Status Through 2003,”Annual Energy Outlook 2005, DOE/EIA-0383(2005)(Washington, DC, February 2005), pp. 20-23, web sitewww.eia.doe.gov/oiaf/archive/aeo05/index.html.[10]Vermont Senate Bill 52, Sec. 2 (8002)(2) (June 14,2005).[11]Federal Register, Vol. 70, No. 91 (May 12, 2005), 40CFR Parts 51, 72, 73, 74, 77, 78, and 96.[12]U.S. Environmental Protection Agency, “Clean Air InterstateRule,” web site www.epa.gov/cair.[13]States are required to meet both seasonal and annualNOx caps. The SO2 caps are annual only.[14]Federal Register, Vol. 70, No. 95 (May 18, 2005), 40CFR Parts 60, 72, and 75.[15]For the complete text of SAFETEA-LU, Public Law109-59, see web site http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=109_cong_public_laws&docid=f:publ059.109.pdf.Issues in Focus[16]The USGS provides three point estimates of undiscoveredand inferred resources: the mean, a 5-percent confidenceinterval, and a 95-percent confidence interval withno price relationship. AEO2006 assumes that proven reservesare not subject to much uncertainty.[17]For readers interested in the international effects ofhigher oil prices, an International Energy Agency paper,“Impact of Higher Oil Prices on the World Economy”(2003) is available from web site www.iea.org/Textbase/publications/free_new_Desc.asp? PUBS_ID=886.[18]The more that is spent back in the U.S. economy, thelower will be the net effect. If the receivers of the extra income,domestic oil companies and oil-exporting countries,do not spend it back in the U.S. economy, aggregate demandfor goods and services will be reduced in the shortterm. Even if all additional oil revenues are spent back inthe United States, there still will be distribution effectsinvolving a move toward different categories of consumption.There will also be an indirect impact on demand forU.S. goods and services through third-country effects;when higher oil prices have negative effects on economicgrowth in other countries, their demand for imports fromthe United States will be reduced.[19]See K.A. Mork, “Oil and the Macroeconomy WhenPrices Go Up and Down: An Extension of Hamilton’s Results,”Journal of Political Economy, Vol. 97 (1989), pp.740-744.[20]There have been several recent surveys of past researchon the economic impacts of oil price shocks. See S.P.A.Brown, M.K. Yücel, and J. Thompson, “Business Cycles:The Role of Energy Prices,” in Encyclopedia of Energy,C.J. Cleveland, Ed. (New York, NY: Academic Press,2004); S.P.A. Brown and M.K. Yucel, “Energy Prices andAggregate Economic Activity: An Interpretative Survey,”Quarterly Review of Economics and Finance, Vol. 42(2002), pp. 193-208; and D.W. Jones, P.N. Leiby, and I.K.Paik, “Oil Price Shocks and the Macroeconomy: WhatHas Been Learned Since 1996,” Energy Journal, Vol. 25,No. 2 (2004).[21]H.G. Huntington, The Economic Consequences ofHigher Crude Oil Prices, EMF SR 9 (Stanford, CA, October2005), web site www.stanford.edu/group/EMF/publications/doc/EMFSR9.pdf.[22]Results of Global Insight’s Macroeconomic Model of theU.S. Economy are from N. Gault, “Impacts on the U.S.Economy: Macroeconomic Models,” Presented at the EnergyModeling Forum Workshop on Macroeconomic Impactsof Oil Shocks (Arlington, VA, February 8, 2005),web site www.stanford.edu/group/EMF/research/doc/gault.pdf. Federal Reserve Bank macroeconomic modelresults are from D. Reifschneider, R. Tetlow, and J. Williams,“Aggregate Disturbances, Monetary Policy, andthe Macroeconomy: The FRB/US Perspective,” FederalReserve Bulletin (January 1999), web site www.federalreserve.gov/pubs/bulletin/1999/0199lead.pdf. Resultsfrom the NiGEM global macroeconomic model arefrom R. Barrell and O. Pomerantz, “Oil Prices and theWorld Economy,” National Institute of Economic and SocialResearch Discussion Paper 242 (London, UK, December2004), web site www.niesr.ac.uk/pubs/dps/dp242.pdf.120 Energy Information Administration / Annual Energy Outlook 2006

Notes and Sources[23]R. Jimenez-Rodriguez and M. Sanchez. “Oil PriceShocks and Real GDP Growth: Empirical Evidence forSome OECD Countries,” Applied Economics, Vol. 37, No.2 (February 2005), pp. 201-228.[24]H.G. Huntington, “Energy Disruptions, InterfirmPrice Effects and the Aggregate Economy,” Energy Economics,Vol. 25, No. 2 (March 2003), pp. 119-136.[25]Quality and prices vary among imported light, sweetcrudes. For example, Nigerian Bonny Light usually isworth 15 cents a barrel more than WTI, and NorwegianOseberg is discounted by about 55 cents a barrel. SeeNYMEX Light Sweet Crude Oil Futures Contract Specifications,web site www.nymex.com/CL_spec.aspx.[26]Energy Information Administration, Form EIA-856,“Monthly Foreign Crude Oil Acquisition Report” (1985-2005).[27]Energy Information Administration, Form EIA-856,“Monthly Foreign Crude Oil Acquisition Report” (1985-2005).[28]See Purvin & Gertz, Inc., “Global Petroleum MarketOutlook—An Online Global Service,” Petroleum Balances(2004), web site www.purvingertz.com/studies.html.[29]U.S. Environmental Protection Agency, “Tier 2/Gasoline Sulfur Final Rule” (Washington, DC, February2000), web site www.epa.gov/tier2/finalrule.htm; “Controlof Air Pollution From New Motor Vehicles: Heavy-Duty Engine and Vehicle Standards and Highway DieselFuel Sulfur Control Requirements Final Rule” (Washington,DC, January 2001), web site www.epa.gov/fedrgstr/EPA-AIR/2001/January/Day-18/a01a.htm; and “Controlof Emissions of Air Pollution From Nonroad DieselEngines and Fuel Final Rule” (Washington, DC, June2004), web site www.epa.gov/fedrgstr/EPA-AIR/2004/June/Day-29/a11293a.htm.[30]Navigant Consulting, Inc., Energy Savings Potential ofSolid State Lighting in General Illumination Applications(Washington DC, November 2003), prepared forBuilding Technologies Program, Office of Energy Efficiencyand Renewable Energy, U.S. Department ofEnergy, web site www.netl.doe.gov/ssl/PDFs/SSL%20Energy%20Savi_ntial%20Final.pdf. General researchgoals are provided in Optoelectronics Industry DevelopmentAssociation, 2002 Update: The Promise of SolidState Lighting for General Illumination (Washington,DC, 2002), web site http://lighting.sandia.gov/lightingdocs/OIDA_SSL_Roadmap_Summary_2002.pdf.[31]The home would be both an energy consumer and, atcertain times, a net energy provider. Over the course of ayear, its net energy purchases would be zero or near zero.U.S. Department of Energy, Building America Researchis Leading the Way to Zero Energy Homes (Washington,DC, May 2005), web site www.nrel.gov/docs/fy05osti/37547.pdf.[32]Non-grid-connected housing units that derive most orall of their energy from renewable sources exist, but mostare in remote areas and not designed for year-round living.[33]Partnership for Advancing Housing Technology(PATH), web site www.pathnet.org.[34]These are the six high-priority areas for emerging technologiesidentified in American Council for an Energy-Efficient Economy, Emerging Energy-Saving Technologiesand Practices for the Buildings Sector as of 2004,Report Number A042 (Washington, DC, October 2004),web site http://aceee.org/pubs/a042full.pdf.[35]G.A. Smook, Handbook for Pulp & Paper Technologists,2nd Edition (Vancouver, BC: Angus Wilde Publications,1992), p. 140.[36]Forest Products Industry Technology Alliance, Agenda2020: 2003 Progress Report (Washington, DC, 2003),p. 5, web site www.agenda2020.org/PDF/2003_Progress_Report.pdf.[37]National Academy of Sciences, National ResearchCouncil, Energy Research at DOE: Was It Worth It? EnergyEfficiency and Fossil Energy Research 1978 to 2000(Washington, DC: National Academy Press, 2001), website www.nap.edu/catalog/10165.html.[38]A nanometer is one-billionth of a meter.[39]National Nanotechnology Initiative, web site www.nano.gov.[40]J. McConnico, “Breakthrough at WFU NanotechnologyCenter Aids Quest for Viable Alternative EnergySources” (Wake Forest University News Service, November7, 2005), web site www.wfu.edu/wfunews/2005/110705n.html.[41]Y. Chen, J. Au, P. Kazlas, A. Ritenour, H. Gates, and M.McCreary, “Electronic Paper: Flexible Active-MatrixElectronic Ink Display,” Nature, Vol. 423 (May 8, 2003),p. 136.[42]M. Brown, “Nano-Bio-Info Pathways to ExtremeEnergy Efficiency,” Presentation to the American Associationfor the Advancement of Science Annual Meeting(Washington, DC, February 21, 2005), web site www.ornl.gov/sci/eere/aaas/ExtremeEfficiency8.pdf.[43]The estimates of “in-place” resources do not take intoconsideration whether the resources are either technicallyor economically recoverable.[44]M. Kray, “Testimony on Behalf of NuStart Energy Development,LLC, United States Senate, Committee onEnvironment & Public Works, Subcommittee on CleanAir, Climate Change and Nuclear Safety” (May 20, 2004),web site www.nei.org/documents/Testimony_Kray_05-20-04.pdf; and E.W. Merrow, K.E. Phillips, and C.W.Myers, Understanding Cost Growth and PerformanceShortfalls in Pioneer Process Plants, R-2569-DOE (SantaMonica, CA: The Rand Corporation, September 1981),web site www.rand.org/pubs/reports/R2569.[45]For more details, see Energy Information Administration,“New Reactor Designs,” web site www.eia.doe.gov/cneaf/nuclear/page/analysis/nucenviss2.html.[46]See, for example, A. Dixit and R. Pyndick, InvestmentUnder Uncertainty (New York, NY: McGraw-Hill, 2001).[47]R.M. Heavenrich, Light-Duty Automotive Technologyand Fuel Economy Trends: 1975 Through 2005,EPA420-R-05-001, Appendix E, “Data Stratified by VehicleType” (Washington, DC: U.S. Environmental ProtectionAgency, July 2005), web site www.epa.gov/otaq/cert/mpg/fetrends/420r05001e.pdf.Energy Information Administration / Annual Energy Outlook 2006 121

Notes and Sources[48]These costs represent the range of difference in cost betweenhybrid and conventional vehicles. The hybrid vehiclesrepresented here have electric drive train componentsthat are used to move the vehicle.[49]Moody International Group, “Canada Oil—SandyDepths,” Milestones, Vol. 5, No. 1 (May 2005), p. 2,web site www.moodyint.com/news/documents/MoodyVol5Iss1FINAL.pdf.[50]Alberta Energy and Utilities Board, Alberta’s Reserves2003 and Supply/Demand Outlook 2004-2013, StatisticalSeries (ST) 2004-98 (Calgary, Alberta: June 2004).[51]National Energy Board, Canada’s Oil Sands: A Supplyand Market Outlook to 2015 (Calgary, Alberta: October2000).[52]Syncrude Canada, Ltd., 2004 Sustainability Report(Fort McMurray, Alberta: 2004), p. 14, “2004 Results,”web site http://sustainability.syncrude.ca/sustainability2004/download/SyncrudeSD2004.pdf.[53]F. Birol and W. Davie, “Oil Supply Costs and EnhancedProductivity,” Energy Prices and Taxes (4th Quarter2001), pp. xvii-xxii, web site http://data.iea.org/ieastore/assets/products/eptnotes/feature/4Q2001B.pdf.[54]Petroleos De Venezuela, S.A., U.S. Securities and ExchangeCommission Form 20-F (October 7, 2005).[55]J.R. Dyni, “Geology and Resources of Some WorldOil-Shale Deposits,” Oil Shale, Vol. 20, No. 3 (2003), pp.193-252.[56]J.R. Donnell, “Geology and Oil-Shale Resources of theGreen River Formation,” in Proceedings of the First Symposiumon Oil Shale (Colorado School of Mines, 1964),pp. 153-163.[57]T. Dammer, “Strategic Significance of America’s OilShale Resource,” Presented at the 2005 EIA MidtermEnergy Outlook Conference (Washington, DC, April12, 2005), web site www.eia.doe.gov/oiaf/aeo/conf/pdf/dammer.pdf.[58]“Shell Signs Chinese Oil-Shale Deal; Begins USGroundwater-Freezing Test,” Syngas Refiner, Vol. 1, No.9 (September 2005), p. 46.[59]J.T. Bartis, T. LaTourrette, L. Dixon, D.J. Peterson,and G. Cecchine, Oil Shale Development in the UnitedStates: Prospects and Policy Issues (Santa Monica, CA:RAND Corporation, 2005), p. xiii, web site www.rand.org/pubs/monographs/2005/RAND_MG414.pdf.[60]J.T. Bartis et al., p. 15.[61]“Fischer-Tropsch Condensate Boosts Lubricity in SyntheticDiesel: New SASOL Study,” World Fuels Today(October 6, 2005), p. 3.[62]“Coal-To-Liquids Looks Profitable If Oil Prices StayAbove $30 to $40/Barrel Range,” World Fuels Today, (December22, 2004), p. 1.[63]Rentech, Inc., “The Economic Viability of an FT FacilityUsing PRB Coals,” Presentation to the Wyoming Governor’sOffice and the Wyoming Business Council (April14, 2005), web site www.rentechinc.com/pdfs/WBC-Presentation-4-14-05-b.pdf.[64]“Shenhua Coal Liquefaction Project Planned for Executionin Inner Mongolia,” Business China (January 24,2005).[65]World Fuels Today (August 16, 2005), p. 2.[66]S. Clarke and B. Ghaemmaghami, “Taking GTLForward—Engineering a Gas-to-Liquids Project,” TCEToday (July 2003), pp. 34-37, web site www.fwc.com/publications/tech_papers/files/GTL06%2Epdf.[67]U. Zuberbuhler, M. Specht, and A. Bandi, “Gasificationof Biomass: An Overview on Available Technologies,”Centre for Solar Energy and Hydrogen Research (ZSW),Birkenfeld, Germany (August 30, 2005), p. 2.[68]CHOREN Industries GmbH, Freiberg, Germany.[69]Correspondence with Matthias Rudloff, Head of BusinessDevelopment, CHOREN Industries GmbH (November29, 2005).[70]M. Rudloff, CHOREN Industries GmbH, “Biomass toLiquids (BtL) from the Carbo-V Process: Technology andthe Latest Developments,” Presentation at the SecondWorld Conference and Technology Exhibition (Rome, Italy,May 10-14, 2004).[71]Bechtel, Aspen Process Flowsheet Simulation Model ofa Battelle Biomass-Based Gasification, Fischer-TropschLiquefaction and Combined-Cycle Power Plant: TopicalReport, May 1998,” DE-AC22-93PC91029--16 (Pittsburgh,PA: Pittsburgh Energy Technology Center, October30, 1998), Appendix B-3.[72]Business Advisory Services, “Illinois Ethanol PrefeasibilityEvaluator,” web site www.iletohprefeas.com.[73]J.A. Waynick, Sr., Characterization of Biodiesel Oxidationand Oxidation Products: CRC Project No. AVFL-2b,Task 1 Results (San Antonio, TX: Southwest ResearchInstitute, August 2005), web site www.nrel.gov/vehiclesandfuels/npbf/pdfs/39096.pdf.[74]Frazier, Barnes & Associates, LLC (Memphis, TN).[75]“Oil Refiners Might Blend or Co-Process ‘Renewables’With Crude-Based Feedstocks,” World Fuels Today (September20, 2005), p. 4.[76]“Oil Refiners Might Blend or Co-Process ‘Renewables’With Crude-Based Feedstocks,” World Fuels Today (September20, 2005), p. 4.[77]Approximately 2 to 3 pounds of brominated ACI permillion metric actual cubic feet of flue gas.[78]A.P. Jones, J.W. Hoffman, T.J. Feeley, III, and J.T.Murphy, DOE/NETL Mercury Control Technology CostEstimate Report – 2005 Update (National Energy TechnologyLaboratory, June 2005).[79]Configuration refers to the type of air pollution controldevice at a plant, such as cold-side electrostatic precipitator,fabric filter, or flue gas desulfurization unit.[80]“President Announces Clear Skies & Global ClimateChange Initiatives” (February 14, 2002), web site www.whitehouse.gov/news/releases/2002/02/20020214-5.html.[81]See “Addendum” in the “Global Change Policy Book,”web site www.whitehouse.gov/news/releases/2002/02/climatechange.html. The business-as-usual (BAU) projectionscited in the addendum are somewhat higher thana “Policies and Measures” case developed by the EPA forthe U.S. Climate Action Report 2002.122 Energy Information Administration / Annual Energy Outlook 2006

Notes and Sources[82]U.S. Department of State, U.S. Climate Action Report2002 (Washington, DC, May 2002), Chapter 5, “ProjectedGreenhouse Gas Emissions,” pp. 70-80, web site http://yosemite.epa.gov/oar/globalwarming.nsf/content/ResourceCenterPublicationsUSClimateActionReport.html.[83]Personal communication from Casey Delhotal, EPA, toDaniel Skelly, EIA, July 7, 2005. EIA adjusted the EPAno-measures case projections to extrapolate from themost recent 2002-to-2004 data on these gases as publishedby EIA, as well as to estimate the intervening yearsof the projections, because the projections were providedonly for every 5 years, beginning in 2005 and ending in2020.Market Trends[84]The energy-intensive manufacturing sectors includefood, paper, bulk chemicals, petroleum refining, glass, cement,steel, and aluminum.[85]EIA collects aggregate data on total energy consumptionin the industrial sector and detailed energy consumptionin the manufacturing sector. The difference betweenthe aggregate industrial sector energy data and thedetailed manufacturing sector data is nonmanufacturingindustrial energy consumption. Nonmanufacturing energyconsumption is allocated to the agriculture, mining,and construction sectors on the basis of information fromthe Energy Information Administration, the U.S. CensusBureau, and the U.S. Department of Agriculture.[86]The alternative technology cases change technologycharacterizations only for sectors represented in theNEMS industrial model. Consequently, refining valuesare unchanged from those in the reference case projections.The reference case projections for the refining sectorinclude considerable cogeneration capacity additionsassociated with coal-to-liquids production. Excluding therefining sector, cogeneration capacity increases by 2.4percent per year in the high technology case, as comparedwith 1.8 percent per year in the reference case.[87]U.S. Department of Energy, Office of Energy Efficiencyand Renewable Energy, Scenarios of U.S. Carbon Reductions:Potential Impacts of Energy Technologies by 2010and Beyond, ORNL/CON-444 (Washington, DC, September1997); J. DeCicco et al., Technical Options for Improvingthe Fuel Economy of U.S. Cars and Light Trucksby 2010-2015 (Washington, DC: American Council for anEnergy Efficient Economy, April 2001); M.A. Weiss et al.,On the Road in 2020: A Life-Cycle Analysis of New AutomotiveTechnologies (Cambridge, MA: Massachusetts Instituteof Technology, October 2000); A. Vyas, C. Saricks,and F. Stodolsky, Projected Effect of Future Energy Efficiencyand Emissions Improving Technologies on FuelConsumption of Heavy Trucks (Argonne, IL: Argonne NationalLaboratory, 2001); and Energy and EnvironmentalAnalysis, Inc., Documentation of Technologies includedin the NEMS Fuel Economy Model for Passenger Carsand Light Trucks (prepared for Energy Information Administration,September 30, 2002).[88]Unless otherwise noted, the term “capacity” in the discussionof electricity generation indicates utility,nonutility, and combined heat and power capacity. Costsreflect the arithmetic average of regional costs.[89]Avoided cost estimates the incremental cost of fuel andcapacity displaced by a unit of the specified resource andmore accurately reflects its as-dispatched energy valuethan comparison to the levelized cost of other individualtechnologies. It does not reflect system reliability cost,nor does it necessarily indicate the lowest cost alternativefor meeting system energy and capacity needs.[90]AEO2006 does not include off-grid photovoltaics (PV).Based on annual PV shipments from 1989 through 2003,EIA estimates that as much as 149 megawatts of remoteelectricity generation PV applications (i.e., off-grid powersystems) were in service in 2003, plus an additional 414megawatts in communications, transportation, and assortedother non-grid-connected, specialized applications.See Energy Information Administration, AnnualEnergy Review 2004, DOE/EIA-0384 (2004) (Washington,DC, August 2005), Table 10.6, “Photovoltaic Cell andModule Shipments by End Use and Market Sector,1989-2003.” The approach used to develop the estimate,based on shipment data, provides an upper estimate ofthe size of the PV stock, including both grid-based andoff-grid PV. It overestimates the size of the stock, becauseshipments include a substantial number of units that areexported, and each year some of the PV units installedearlier are retired from service or abandoned.[91]Stranded Arctic natural gas and resources in areaswhere drilling is officially prohibited are not included inTable 17.[92]Resources from the Arctic Offshore Outer Continentalshelf and resources in areas where drilling is officiallyprohibited are not included in Table 18.[93]Energy Information Administration (EIA), FormEIA-767, “Steam-Electric Plant Operation and DesignData”; and EIA, Office of Integrated Analysis and Forecasting,Plant File.[94]Assumptions for the Integrated High Technology caseare documented in Energy Information Administration(EIA), Assumptions to the Annual Energy Outlook 2006,DOE/EIA-0554(2006) (Washington, DC, February 2006).Primary sources and assumptions include the following:Buildings: EIA, Technology Forecast Updates—Residentialand Commercial Building Technologies—AdvancedAdoption Case (Navigant Consulting, Inc., September2004). Industrial: EIA, Industrial Technology and DataAnalysis Supporting the NEMS Industrial Model (FocisAssociates, October 2005). Transportation: EIA, Documentationof Technologies included in the NEMS FuelEconomy Model for Passenger Cars and Light Trucks(Energy and environmental Analysis, Inc., September2002), and A. Vyas, C. Saricks, and F. Stodolsky, ProjectedEffect of Future Energy Efficiency and EmissionsImproving Technologies on Fuel Consumption of HeavyTrucks (Argonne, IL: Argonne National Laboratory,2001). Fossil-fired generating technologies: By assumption,capital costs, heat rates, and operating costs for advancedcoal and gas technologies fall more rapidly overtime, reaching levels 10 percent lower than in the referencecase in 2030. Advanced nuclear technology: By assumption,capital and operating costs fall more rapidlyover time, reaching levels 20 percent lower than in thereference case in 2030.Energy Information Administration / Annual Energy Outlook 2006 123

Notes and SourcesTable Notes and SourcesNote: Tables indicated as sources in these notes referto the tables in Appendixes A, B, C, and D of thisreport.Table 1. Total energy supply and disposition in theAEO2006 reference case: summary, 2003-2030:AEO2006 National Energy Modeling System, run AEO-2006.D111905A. Notes: Quantities are derived from historicalvolumes and assumed thermal conversion factors.Other production includes liquid hydrogen, methanol, supplementalnatural gas, and some inputs to refineries. Netimports of petroleum include crude oil, petroleum products,unfinished oils, alcohols, ethers, and blending components.Other net imports include coal coke and electricity. Somerefinery inputs appear as petroleum product consumption.Other consumption includes net electricity imports, liquidhydrogen, and methanol.Table 2. CARB emissions standards for light-duty vehicles,model years 2009-2016: California Air ResourcesBoard, California Exhaust Emission Standards and TestProcedures for 2001 and Subsequent Model Passenger Cars,Light-Duty Trucks, and Medium-Duty Vehicles (Sacramento,CA, August 4, 2005).Table 3. Proposed light truck CAFE standards bymodel year and footprint category: National HighwayTraffic Safety Administration, “Average Fuel EconomyStandards for Light Trucks Model Years 2008-2011,” 49CFR Parts 523, 533 and 537, Docket No. 2005-22223, RIN2127-AJ61 (Washington, DC, August, 2005).Table 4. Key projections for light truck fuel economyin the alternative CAFE standards case, 2011-2030:AEO2006 National Energy Modeling System, runsAEO2006.D111905A and ALTCAFE.D121505A.Table 5. Basic features of State renewable energy requirementsand goals enacted since 2003: Energy InformationAdministration, Office of Integrated Analysisand Forecasting.Table 6. Major changes in existing State renewableenergy requirements and goals since 2003: Energy InformationAdministration, Office of Integrated Analysisand Forecasting.Table 7. New U.S. renewable energy capacity, 2004-2005: Energy Information Administration, Office of IntegratedAnalysis and Forecasting.Table 8. Estimates of national trends in annual emissionsof sulfur dioxide and nitrogen oxides, 2003-2020: EPA Data: U.S. Environmental Protection Agency,Regulatory Impact Analysis for the Final Clean Air InterstateRule, EPA-452/R-05-002 (Washington, DC, March2005), Table 7-2, p. 196. Projections: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A.Table 9. Macroeconomic model estimates of economicimpacts from oil price increases: H.G. Huntington,The Economic Consequences of Higher Crude OilPrices, EMF SR 9 (Stanford, CA, October 2005), web sitewww.stanford.edu/group/EMF/publications/doc/EMFSR9.pdf.Table 10. Time-series estimates of economic impactsfrom oil price increases: R. Jimenez-Rodriguez and M.Sanchez. “Oil Price Shocks and Real GDP Growth: EmpiricalEvidence for Some OECD Countries,” Applied Economics,Vol. 37, No. 2 (February 2005), pp. 201-228.Table 11. Summary of U.S. oil price-GDP elasticities:H.G. Huntington, The Economic Consequences of HigherCrude Oil Prices, EMF SR 9 (Stanford, CA, October 2005),web site www.stanford.edu/group/EMF/publications/doc/EMFSR9.pdf.Table 12. Economic indicators in the reference, highprice, and low price cases, 2005-2030: AEO2006 NationalEnergy Modeling System, runs AEO2006.D111905A,LP2006.D120105A, and HP2006.D113005A.Table 13. Technologies expected to have significantimpacts on new light-duty vehicles: Energy InformationAdministration, Office of Integrated Analysis andForecasting.Table 14. Nonconventional liquid fuels productionin the AEO2006 reference and high price cases, 2030:AEO2006 National Energy Modeling System, runs AEO-2006.D111905A and HP2006.D113005A.Table 15. Projected changes in U.S. greenhouse gasemissions, gross domestic product, and greenhousegas intensity, 2002-2020: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Table 16. Costs of producing electricity from newplants, 2015 and 2030: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Table 17. Technically recoverable U.S. natural gasresources as of January 1, 2004: Energy InformationAdministration, Office of Integrated Analysis and Forecasting.Table 18. Technically recoverable U.S. crude oil resourcesas of January 1, 2004: Energy Information Administration,Office of Integrated Analysis and Forecasting.Figure Notes and SourcesNote: Tables indicated as sources in these notes referto the tables in Appendixes A, B, C, and D of thisreport.Figure 1. Energy prices, 1980-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Table A1.Figure 2. Delivered energy consumption by sector,1980-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A2.Figure 3. Energy consumption by fuel, 1980-2030:History: Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC,August 2005). Projections: Tables A1 and A18.124 Energy Information Administration / Annual Energy Outlook 2006

Notes and SourcesFigure 4. Energy use per capita and per dollar ofgross domestic product, 1980-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Energy use per capita: Calculated fromdata in Table A2. Energy use per dollar of GDP: TableA19.Figure 5. Electricity generation by fuel, 1980-2030:History: Energy Information Administration (EIA), FormEIA-860B, “Annual Electric Generator Report—Nonutility”;EIA, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005); and Edison ElectricInstitute. Projections: Table A8.Figure 6. Total energy production and consumption,1980-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A1.Figure 7. Energy production by fuel, 1980-2030: History:Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC, August2005). Projections: Tables A1 and A17.Figure 8. Projected U.S. carbon dioxide emissions bysector and fuel, 1990-2030: History: Energy InformationAdministration, Emissions of Greenhouse Gases in theUnited States 2004, DOE/EIA-0573(2004) (Washington,DC, December 2005). Projections: Table A18.Figure 9. Sulfur dioxide emissions in selected States,1980-2003: U.S. Environmental Protection Agency, website http://cfpub.epa.gov/gdm.Figure 10. World oil prices in the AEO2005 andAEO2006 reference cases: AEO2005: Energy InformationAdministration, AEO2005 National Energy ModelingSystem, run AEO2005.D102004A. AEO2006: Table A1.Figure 11. World oil prices in three AEO2006 cases,1980-2030: Table C1.Figure 12. Changes in world oil price and U.S. realGDP in the AEO2006 high and low price cases, 2004-2030: AEO2006 National Energy Modeling System, runsLP2006.D120105A and HP2006.D113005A.Figure 13. GDP elasticities with respect to oil pricechanges in the high price case, 2006-2030: AEO2006National Energy Modeling System, run HP2006.D113005A. Note: The figure shows profiles of year-by-yearand period average GDP elasticities with respect to oil pricechanges in the high price case. The elasticities are computedas follows: (1) Year-by-year elasticity =(Percentagechange from baseline real GDP )/(Percentage change frombaseline oil price ). (2) Period average elasticity =(percentagechange in cumulative high price GDPs from cumulativebaseline GDPs )/(Percentage change in cumulative highprices from cumulative baseline prices ).Figure 14. Purvin & Gertz forecast for world oil productionby crude oil quality, 1990-2020: Purvin &Gertz, Inc., “Global Petroleum Market Outlook—An OnlineGlobal Service,” Petroleum Balances (2004), web sitewww.purvingertz.com/studies.html.Figure 15. Sulfur content specifications for U.S. petroleumproducts, 1990-2014: U.S. Environmental ProtectionAgency, “Tier 2/Gasoline Sulfur Final Rule” (Washington,DC, February 2000), web site www.epa.gov/tier2/finalrule.htm; “Control of Air Pollution From New MotorVehicles: Heavy-Duty Engine and Vehicle Standards andHighway Diesel Fuel Sulfur Control Requirements FinalRule” (Washington, DC, January 2001), web site www.epa.gov/fedrgstr/EPA-AIR/2001/January/Day-18/a01a.htm;and “Control of Emissions of Air Pollution From NonroadDiesel Engines and Fuel Final Rule” (Washington, DC,June 2004), web site www.epa.gov/fedrgstr/EPA-AIR/2004/June/Day-29/a11293a.htm.Figure 16. U.S. hydrotreating capacity, 1990-2030:History: Energy Information Administration, Office of Oiland Gas, derived from Form EI-810 “Monthly Refinery Report.”Projections: AEO2006 National Energy ModelingSystem, run AEO2006.D111905A.Figure 17. Market penetration of advanced technologiesin new cars, 2004 and 2030: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A.Figure 18. Market penetration of advanced technologiesin new light trucks, 2004 and 2030: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A.Figure 19. System elements for production of syntheticfuels from coal, natural gas, and biomass:Energy Information Administration, Office of IntegratedAnalysis and Forecasting.Figure 20. Mercury emissions from the electricitygeneration sector, 2002-2030: AEO2006 National EnergyModeling System, runs AEO2006.D111905A andACI2006.D112305A.Figure 21. Mercury allowance prices, 2010-2030:AEO2006 National Energy Modeling System, runs AEO-2006.D111905A and ACI2006.D112305A.Figure 22. Coal-fired generating capacity retrofittedwith activated carbon injection systems, 2010-2030:AEO2006 National Energy Modeling System, runs AEO-2006.D111905A and ACI2006.D112305A.Figure 23. Projected change in U.S. greenhouse gasintensity in three cases, 2002-2020: AEO2006 NationalEnergy Modeling System, runs AEO2006.D111905A,LTRKITEN.D121905A, and HTRKITEN.D121905A.Figure 24. Average annual growth rates of real GDP,labor force, and productivity in three cases, 2004-2030: Table B4.Figure 25. Average annual unemployment, interest,and inflation rates, 2004-2030: Table A19.Figure 26. Sectoral composition of output growthrates, 2004-2030: AEO2006 National Energy ModelingSystem, run AEO2006.D111905A.Figure 27. Sectoral composition of manufacturingoutput growth rates, 2004-2030: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Energy Information Administration / Annual Energy Outlook 2006 125

Notes and SourcesFigure 28. Energy expenditures as share of gross domesticproduct, 1970-2030: History: U.S. Departmentof Commerce, Bureau of Economic Analysis; and Energy InformationAdministration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: AEO2006 National Energy Modeling System,run AEO2006.D111905A.Figure 29. World oil prices in three cases, 1980-2030:History: Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC,August 2005). Projections: Table C1.Figure 30. U.S. gross petroleum imports by source,2004-2030: AEO2006 National Energy Modeling System,run AEO2006.D111905A.Figure 31. Energy use per capita and per dollar ofgross domestic product, 1980-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Energy use per capita: Calculated fromdata in Table A2. Energy use per dollar of GDP: TableA19.Figure 32. Primary energy use by fuel, 2004-2030:History: Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC,August 2005). Projections: Tables A1 and A17.Figure 33. Delivered energy use by fuel, 1980-2030:History: Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC,August 2005). Projections: Table A2.Figure 34. Primary energy consumption by sector,1980-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A2.Figure 35. Delivered residential energy consumptionper capita by fuel, 1980-2030: History: Energy InformationAdministration, State Energy Data Report 2001,DOE/EIA-0214(2001) (Washington, DC, November 2004),and Annual Energy Review 2004, DOE/EIA-0384(2004)(Washington, DC, August 2005). Projections: AEO2006National Energy Modeling System, run AEO2006.D111905A.Figure 36. Delivered residential energy consumptionby end use, 2001, 2004, 2015, and 2030: History:Energy Information Administration, “Residential EnergyConsumption Survey.” Projections: Table A4. Note: Although2001 is the last year of historical data for many ofthe detailed end-use consumption concepts (e.g., spaceheating, cooling), 2004 data, taken from EIA’s Annual EnergyReview 2004, is used as the base year for the more aggregatestatistics shown in AEO2006. For illustrative purposes,EIA estimates for the detailed end-use consumptionconcepts, consistent with this historical information, areused to show growth rates.Figure 37. Efficiency indicators for selected residentialappliances, 2004 and 2030: Energy Information Administration,Technology Forecast Updates—Residentialand Commercial Building Technologies—Advanced AdoptionCase (Navigant Consulting, Inc., September 2004); andAEO2006 National Energy Modeling System, run AEO-2006.D111905A.Figure 38. Variation from reference case deliveredresidential energy use in three alternative cases,2004-2030: Table D1.Figure 39. Delivered commercial energy consumptionper capita by fuel, 1980-2030: History: Energy InformationAdministration, State Energy Data Report 2001,DOE/EIA-0214(2001) (Washington, DC, November 2004),and Annual Energy Review 2004, DOE/EIA-0384(2004)(Washington, DC, August 2005). Projections: AEO2006National Energy Modeling System, run AEO2006.D111905A.Figure 40. Delivered commercial energy intensity byend use, 2004, 2015, and 2030: Table A5.Figure 41. Efficiency indicators for selected commercialequipment, 2004 and 2030: Energy InformationAdministration, Technology Forecast Updates—Residentialand Commercial Building Technologies—AdvancedAdoption Case (Navigant Consulting, Inc., September2004); and AEO2006 National Energy Modeling System,run AEO2006.D111905A.Figure 42. Variation from reference case deliveredcommercial energy intensity in three alternativecases, 2004-2030: Table D1.Figure 43. Buildings sector electricity generationfrom advanced technologies in two alternativecases, 2030: AEO2006 National Energy Modeling System,runs AEO2006.D111905A, BLDHIGH.D112205A, andBLDBEST.D112205C.Figure 44. Industrial energy intensity by two measures,1980-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004)(Washington, DC, August 2005). Projections: AEO2006National Energy Modeling System, run AEO2006.D111905A.Figure 45. Energy intensity in the industrial sector,2004: AEO2006 National Energy Modeling System, runAEO2006.D111905A.Figure 46. Average growth in industrial output anddelivered energy consumption by sector, 2004-2030:AEO2006 National Energy Modeling System, run AEO-2006.D111905A.Figure 47. Projected energy intensity in 2030 relativeto 2004, by industry: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Figure 48. Variation from reference case deliveredindustrial energy use in two alternative cases, 2004-2030: Table D2.Figure 49. Transportation energy use per capita,1980-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A.Figure 50. Transportation travel demand by mode,1980-2030: History: Federal Highway Administration,Highway Statistics 2002 (Washington, DC, November2003), Table VM-1, p. V-57, web site www.fhwa.dot.gov/policy/ohim/hs02/pdf/vm1.pdf, and previous annual issues.Projections: AEO2006 National Energy Modeling System,run AEO2006.D111905A.126 Energy Information Administration / Annual Energy Outlook 2006

Notes and SourcesFigure 51. Average fuel economy of new light-dutyvehicles, 1980-2030: History: National Highway TrafficSafety Administration, Summary of Fuel Economy Performance(Washington, DC, March 2004), web site www.nhtsa.gov/cars/rules/CAFE/docs/Summary-Fuel-Economy-Pref-2004.pdf. Projections: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Figure 52. Sales of advanced technology light-dutyvehicles by fuel type, 2015 and 2030: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A.Figure 53. Changes in projected transportation fueluse in two alternative cases, 2010 and 2030: Table D3.Figure 54. Changes in projected transportation fuelefficiency in two alternative cases, 2010 and 2030:Table D3.Figure 55. Annual electricity sales by sector, 1980-2030: History: Energy Information Administration, AnnualEnergy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A8.Figure 56. Electricity generation capacity additionsby fuel type, including combined heat and power,2005-2030: Table A9.Figure 57. Electricity generation capacity additions,including combined heat and power, by region andfuel, 2005-2030: AEO2006 National Energy Modeling System,run AEO2006.D111905A.Figure 58. Levelized electricity costs for new plants,2015 and 2030: AEO2006 National Energy Modeling System,run AEO2006.D111905A.Figure 59. Electricity generation from nuclearpower, 1973-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:Table A8.Figure 60. Additions of renewable generating capacity,2004-2030: AEO2006 National Energy Modeling System,run AEO2006.D111905A.Figure 61. Levelized and avoided costs for new renewableplants in the Northwest, 2015 and 2030:AEO2006 National Energy Modeling System, run AEO-2006.D111905A.Figure 62. Electricity generation by fuel, 2004 and2030: Table A8.Figure 63. Grid-connected electricity generationfrom renewable energy sources, 1980-2030: History:Energy Information Administration, Annual Energy Review2004, DOE/EIA-0384(2004) (Washington, DC, August2005). Projections: Table A16. Note: Data for nonutilityproducers are not available before 1989.Figure 64. Nonhydroelectric renewable electricitygeneration by energy source, 2004-2030: Table A16.Figure 65. Fuel prices to electricity generators,1995-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A.Figure 66. Average U.S. retail electricity prices,1970-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A8.Figure 67. Cumulative new generating capacity bytechnology type in three economic growth cases,2004-2030: AEO2006 National Energy Modeling System,runs AEO2006.D111905A, LM2006.D113005A, andHM2006.D112505B.Figure 68. Cumulative new generating capacity bytechnology type in three fossil fuel technologycases, 2004-2030: Table D6.Figure 69. Levelized electricity costs for new plantsby fuel type in two nuclear cost cases, 2015 and2030: AEO2006 National Energy Modeling System,runs AEO2006.D111905A, ADVNUC5A.D120105A, andADVNUC20.D120105A. Note: Includes generation and interconnectioncosts.Figure 70. Nonhydroelectric renewable electricitygeneration by energy source in three cases, 2010 and2030: Table D7.Figure 71. Natural gas consumption by sector, 1990-2030: History: Energy Information Administration, AnnualEnergy Review 2004, DOE/EIA-0384 (2004) (Washington,DC, August 2005). Projections: Table A13.Figure 72. Increases in natural gas consumption byCensus division, 2004-2030: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Figure 73. Natural gas production by source, 1990-2030: Table A14.Figure 74. Net U.S. imports of natural gas by source,1990-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: AEO2006 NationalEnergy Modeling System, run AEO2006. D111905A.Figure 75. Lower 48 natural gas wellhead prices,1990-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A14.Figure 76. Natural gas prices by end-use sector,1990-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A13.Figure 77. Lower 48 natural gas wellhead prices inthree technology cases, 1990-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Table D9.Figure 78. Natural gas production and net importsin three technology cases, 1990-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Table D9.Figure 79. Lower 48 natural gas wellhead prices inthree price cases, 1990-2030: History: Energy InformationAdministration, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:Table C1.Figure 80. Net imports of liquefied natural gas inthree price cases, 1990-2030: History: Energy InformationAdministration, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:AEO2006 National Energy Modeling System, runsAEO2006.D111905A, LP2006.D120105A, and HP2006.D113005A.Energy Information Administration / Annual Energy Outlook 2006 127

Notes and SourcesFigure 81. Net imports of liquefied natural gas inthree LNG supply cases, 1990-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: AEO2006 National Energy Modeling System,runs AEO2006.D111905A, LOLNG06.D120405A, andHILNG06.D120405.Figure 82. Lower 48 natural gas wellhead prices inthree LNG supply cases, 1990-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Table D11.Figure 83. World oil prices in the reference case,1990-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A1.Figure 84. Domestic crude oil production by source,1990-2030: History: Energy Information Administration,Petroleum Supply Annual 1990, DOE/EIA-0340(90) (Washington,DC, May 1991), and Office of Integrated Analysisand Forecasting. Projections: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Figure 85. World oil prices in three cases, 1990-2030:History: Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC,August 2005). Projections: Table C1.Figure 86. Total U.S. petroleum production in threeprice cases, 1990-2030: History: Energy InformationAdministration, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:Table C4.Figure 87. U.S. syncrude production from oil shale inthe high price case, 2004-2030: AEO2006 National EnergyModeling System, run HP2006.D113005A.Figure 88. Total U.S. crude oil production in threetechnology cases, 1990-2030: History: Energy InformationAdministration, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:Table D10.Figure 89. Alaskan oil production in the referenceand ANWR cases, 1990-2030: History: Energy InformationAdministration, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:Table D12.Figure 90. U.S. net imports of oil in the referenceand ANWR cases, 1990-2030: History: Energy InformationAdministration, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:Table D12.Figure 91. Consumption of petroleum products,1990-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A11.Figure 92. Domestic refinery distillation capacity,1990-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A11.Figure 93. U.S. petroleum product demand and domesticpetroleum supply, 1990-2030: History: EnergyInformation Administration, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Tables A11.Figure 94. Components of retail gasoline prices,2004, 2015, and 2030: AEO2006 Energy Modeling System,run AEO2006.D111905A.Figure 95. U.S. ethanol fuel consumption in threeprice cases, 1995-2030: History: Energy InformationAdministration, Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington, DC, August 2005). Projections:AEO2006 National Energy Modeling System, runsAEO2006.D111905A, LP2006.D120105A, and HP2006.D113005A.Figure 96. Coal-to-liquids and gas-to-liquids productionin two price cases, 2004-2030: Table C4.Figure 97. Coal production by region, 1970-2030: History:1970-1990: Energy Information Administration(EIA), The U.S. Coal Industry, 1970-1990: Two Decades ofChange, DOE/EIA-0559 (Washington, DC, November2002); 1991-2000: EIA, Coal Industry Annual, DOE/EIA-0584 (various years); 2001-2004: EIA, Annual CoalReport 2004, DOE/EIA-0584(2004) (Washington, DC, November2005), and previous issues; and EIA, Short-TermEnergy Outlook (Washington, DC, September 2005). Projections:Table A15.Figure 98. Distribution of domestic coal by demandand supply region, 2003 and 2030: 2003: Energy InformationAdministration (EIA), Form EIA-6, “Coal DistributionReport—Annual.” Projections: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A. Note:The Eastern Demand Region includes the New England,Middle Atlantic, South Atlantic, East North Central, andEast South Central Census Divisions. The Western DemandRegion includes the West North Central, West SouthCentral, Mountain, and Pacific Census Divisions.Figure 99. U.S. coal mine employment by region,1970-2030: History: 1970-1976: U.S. Department of theInterior, Bureau of Mines, Minerals Yearbook (variousyears); 1977-1978: Energy Information Administration(EIA), Energy Data Report, Coal—Bituminous and Lignite,DOE/EIA-0118 (Washington, DC, various years), and EnergyData Report, Coal—Pennsylvania Anthracite, DOE/EIA-0119 (Washington, DC, various years); 1979-1992:EIA, Coal Production, DOE/EIA-0118 (various years);1993-2000: EIA, Coal Industry Annual, DOE/EIA-0584(Washington, DC, various years); 2001-2004: EIA, AnnualCoal Report 2004, DOE/EIA-0584(2004) (Washington, DC,November 2005), and previous issues. Projections: AEO-2006 National Energy Modeling System, run AEO2006.D111905A.Figure 100. Average minemouth price of coal by region,1990-2030: History: 1990-2000: Energy InformationAdministration (EIA), Coal Industry Annual, DOE/EIA-0584 (various years); 2001-2004: EIA, Annual CoalReport 2004, DOE/EIA-0584(2004) (Washington, DC, November2005), and previous issues. Projections: AEO2006National Energy Modeling System, run AEO2006.D111905A.128 Energy Information Administration / Annual Energy Outlook 2006

Notes and SourcesFigure 101. Changes in regional coal transportationrates, 2010, 2025, and 2030: AEO2006 National EnergyModeling System, run AEO2006.D111905A.Figure 102. U.S. coal exports and imports, 1970-2030:History: Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC,August 2005). Projections: Table A15.Figure 103. Electricity and other coal consumption,1970-2030: History: Energy Information Administration,Annual Energy Review 2004, DOE/EIA-0384(2004) (Washington,DC, August 2005). Projections: Table A15.Figure 104. Coal consumption in the industrial andbuildings sectors and at coal-to-liquids plants, 2004,2015, and 2030: Table A15.Figure 105. Projected variation from the referencecase projection of U.S. total coal demand in fourcases, 2030: AEO2006 National Energy Modeling System,runs AEO2006.D111905A, LP2006.D120105A, HP2006.D113005A, LM2006.D113005A, and HM2006.D112505B.Figure 106. Average delivered coal prices in threecost cases, 1990-2030: History: Energy InformationAdministration (EIA), Quarterly Coal Report, October-December 2004, DOE/EIA-0121(2004/4Q) (Washington,DC, March 2005), and previous issues; EIA, Electric PowerMonthly, June 2005, DOE/EIA-0226(2005/06) (Washington,DC, June 2005); and EIA, Annual Energy Review 2004,DOE/EIA-0384(2004) (Washington, DC, August 2005).Projections: Table D13. Note: Historical prices areweighted by consumption but exclude residential/commercialprices and export free-alongside-ship (f.a.s.) prices. Projectedprices are weighted in the same way as historicalprices and also exclude import quantities and prices.Figure 107. Carbon dioxide emissions by sector andfuel, 2004 and 2030: 2004: Energy Information Administration,Emissions of Greenhouse Gases in the United States2004, DOE/EIA-0573(2004) (Washington, DC, December2005). Projections: Table A18.Figure 108. Carbon dioxide emissions in three economicgrowth cases, 1990-2030: History: Energy InformationAdministration, Emissions of Greenhouse Gases inthe United States 2004, DOE/EIA-0573(2004) (Washington,DC, December 2005). Projections: Table B2.Figure 109. Carbon dioxide emissions in three technologycases, 2004, 2020, and 2030: History: EnergyInformation Administration, Emissions of GreenhouseGases in the United States 2004, DOE/EIA-0573(2004)(Washington, DC, December 2005). Projections: TableD4.Figure 110. Sulfur dioxide emissions from electricitygeneration, 1990-2030: History: 1990 and 1995: U.S.Environmental Protection Agency, National Air PollutantEmissions Trends, 1990-1998, EPA- 454/R-00-002 (Washington,DC, March 2000). 2004: U.S. Environmental ProtectionAgency, Acid Rain Program Preliminary SummaryEmissions Report, Fourth Quarter 2004, web site www.epa.gov/airmarkets/emissions/prelimarp/index.html. Projections:AEO2006 National Energy Modeling System, runAEO2006.D111905A.Figure 111. Nitrogen oxide emissions from electricitygeneration, 1990-2030: History: 1990 and 1995:U.S. Environmental Protection Agency, National Air PollutantEmissions Trends, 1990-1998, EPA- 454/R-00-002(Washington, DC, March 2000). 2004: U.S. EnvironmentalProtection Agency, Acid Rain Program Preliminary SummaryEmissions Report, Fourth Quarter 2004, web sitewww.epa.gov/airmarkets/emissions/prelimarp/index.html.Projections: AEO2006 National Energy Modeling System,run AEO2006.D111905A.Figure 112. Mercury emissions from electricity generation,1995-2030: History: 1995, 2000, and 2004:Energy Information Administration, Office of IntegratedAnalysis and Forecasting. Projections: AEO2006 NationalEnergy Modeling System, run AEO2006.D111905A.Energy Information Administration / Annual Energy Outlook 2006 129

Appendixes

Appendix AReference Case Energy Information Administration / Annual Energy Outlook 2006 133

Reference Case 134 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 135

Reference Case 136 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 137

Reference Case 138 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 139

Reference Case 140 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 141

Reference Case 142 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 143

Reference Case 144 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 145

Reference Case 146 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 147

Reference Case 148 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 149

Reference Case 150 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 151

Reference Case 152 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 153

Reference Case 154 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 155

Reference Case 156 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 157

Reference Case 158 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 159

Reference Case 160 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 161

Reference Case 162 Energy Information Administration / Annual Energy Outlook 2006

Reference Case Energy Information Administration / Annual Energy Outlook 2006 163

Appendix BEconomic Growth Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 165

Economic Growth Case Comparisons 166 Energy Information Administration / Annual Energy Outlook 2006

Economic Growth Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 167

Economic Growth Case Comparisons 168 Energy Information Administration / Annual Energy Outlook 2006

Economic Growth Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 169

Economic Growth Case Comparisons 170 Energy Information Administration / Annual Energy Outlook 2006

Economic Growth Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 171

Appendix CPrice Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 173

Price Case Comparisons 174 Energy Information Administration / Annual Energy Outlook 2006

Price Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 175

Price Case Comparisons 176 Energy Information Administration / Annual Energy Outlook 2006

Price Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 177

Price Case Comparisons 178 Energy Information Administration / Annual Energy Outlook 2006

Price Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 179

Price Case Comparisons 180 Energy Information Administration / Annual Energy Outlook 2006

Price Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 181

Price Case Comparisons 182 Energy Information Administration / Annual Energy Outlook 2006

Price Case Comparisons Energy Information Administration / Annual Energy Outlook 2006 183

Appendix DResults from Side Cases 184 Energy Information Administration / Annual Energy Outlook 2006

Results from Side Cases Energy Information Administration / Annual Energy Outlook 2006 185

Results from Side Cases 186 Energy Information Administration / Annual Energy Outlook 2006

Results from Side Cases Energy Information Administration / Annual Energy Outlook 2006 187

Results from Side Cases 188 Energy Information Administration / Annual Energy Outlook 2006

Results from Side Cases Energy Information Administration / Annual Energy Outlook 2006 189

Results from Side Cases 190 Energy Information Administration / Annual Energy Outlook 2006

Results from Side Cases Energy Information Administration / Annual Energy Outlook 2006 191

Results from Side Cases 192 Energy Information Administration / Annual Energy Outlook 2006

Results from Side Cases Energy Information Administration / Annual Energy Outlook 2006 193

Results from Side Cases 194 Energy Information Administration / Annual Energy Outlook 2006

Results from Side Cases Energy Information Administration / Annual Energy Outlook 2006 195

Results from Side Cases 196 Energy Information Administration / Annual Energy Outlook 2006

Results from Side Cases Energy Information Administration / Annual Energy Outlook 2006 197

Results from Side Cases 198 Energy Information Administration / Annual Energy Outlook 2006

Appendix ENEMS Overview and Brief Description of CasesThe National Energy Modeling SystemThe projections in the Annual Energy Outlook 2006(AEO2006) are generated from the National EnergyModeling System (NEMS) [1], developed and maintainedby the Office of Integrated Analysis andForecasting (OIAF) of the Energy Information Administration(EIA). In addition to its use in the developmentof the AEO projections, NEMS is also used inanalytical studies for the U.S. Congress, the WhiteHouse, and other offices within the Department ofEnergy. The AEO projections are also used by analystsand planners in other government agencies andoutside organizations.The projections in NEMS are developed with the useof a market-based approach to energy analysis. Foreach fuel and consuming sector, NEMS balancesenergy supply and demand, accounting for economiccompetition among the various energy fuels andsources. The time horizon of NEMS is the long-termperiod through 2030, approximately 25 years into thefuture. In order to represent regional differences inenergy markets, the component modules of NEMSfunction at the regional level: the nine Census divisionsfor the end-use demand modules; productionregions specific to oil, gas, and coal supply and distribution;the North American Electric ReliabilityCouncil (NERC) regions and subregions for electricity;and the Petroleum Administration for DefenseDistricts (PADDs) for refineries.NEMS is organized and implemented as a modularsystem. The modules represent each of the fuel supplymarkets, conversion sectors, and end-use consumptionsectors of the energy system. NEMS alsoincludes macroeconomic and international modules.The primary flows of information between each ofthese modules are the delivered prices of energy to theend user and the quantities consumed by product,region, and sector. The delivered fuel prices encompassall the activities necessary to produce, import,and transport fuels to the end user. The informationflows also include other data on such areas as economicactivity, domestic production, and internationalpetroleum supply.The integrating module controls the execution ofeach of the component modules. To facilitate modularity,the components do not pass information toeach other directly but communicate through acentral data file. This modular design provides thecapability to execute modules individually, thusallowing decentralized development of the systemand independent analysis and testing of individualmodules, and permits the use of the methodology andlevel of detail most appropriate for each energy sector.NEMS calls each supply, conversion, and end-usedemand module in sequence until the delivered pricesof energy and the quantities demanded have convergedwithin tolerance, thus achieving an economicequilibrium of supply and demand in the consumingsectors. Solution is reached annually through thelong-term horizon. Other variables are also evaluatedfor convergence, such as petroleum product imports,crude oil imports, and several macroeconomicindicators.Each NEMS component represents the impacts andcosts of legislation and environmental regulationsthat affect that sector and reports key emissions.NEMS represents current legislation and environmentalregulations as of October 31, 2005, such as theEnergy Policy Acts of 2005 [2] and 1992 [3], the CleanAir Act Amendments (CAAA), and the costs of compliancewith regulations, such as the Clean Air InterstateRule (CAIR) and Clean Air Mercury Rule(CAMR), both of which were finalized and publishedon the U.S. Environmental Protection Agency webpage in March 2005 and in the Federal Register inMay 2005.In general, the historical data used for the AEO2006projections were based on EIA’s Annual EnergyReview 2004, published in August 2005 [4]; however,data were taken from multiple sources. In some cases,only partial or preliminary data were available for2004. Carbon dioxide emissions were calculated byusing carbon dioxide coefficients from the EIA report,Emissions of Greenhouse Gases in the United States2004, published in December 2005 [5].Historical numbers are presented for comparisononly and may be estimates. Source documents shouldbe consulted for the official data values. Some definitionaladjustments were made to EIA data for the projections.For example, the transportation demandsector in AEO2006 includes electricity used by railroads,which is included in the commercial sector inEIA’s consumption data publications. Footnotes inthe appendix tables of this report indicate the definitionsand sources of historical data.Energy Information Administration / Annual Energy Outlook 2006 199

NEMS Overview and Brief Description of CasesThe AEO2006 projections for 2005 and 2006 incorporateshort-term projections from EIA’s September2005 Short-Term Energy Outlook (STEO). For shorttermenergy projections, readers are referred tomonthly updates of the STEO [6].Component ModulesThe component modules of NEMS represent the individualsupply, demand, and conversion sectors ofdomestic energy markets and also include internationaland macroeconomic modules. In general, themodules interact through values representing theprices or expenditures of energy delivered to the consumingsectors and the quantities of end-use energyconsumption.Macroeconomic Activity ModuleThe Macroeconomic Activity Module provides a set ofessential macroeconomic drivers to the energy modulesand a macroeconomic feedback mechanismwithin NEMS. Key macroeconomic variables includegross domestic product (GDP), industrial output,interest rates, disposable income, prices, new housingstarts, new light-duty vehicle sales, and employment.The module uses the following models from GlobalInsight, Inc. (GII): Macroeconomic Model of the U.S.Economy, national Industry Model, and nationalEmployment Model. In addition, EIA has constructeda Regional Economic and Industry Model to projectregional economic drivers and a CommercialFloorspace Model to project 13 floorspace types in 9Census divisions. The accounting framework forindustrial output uses the North American IndustryClassification System (NAICS).International ModuleThe International Module represents world oil markets,calculating the average world oil price and computingsupply curves for 5 categories of importedcrude oil for the Petroleum Market Module (PMM) ofNEMS. The module allows changes in U.S. importrequirements. In addition, 17 international petroleumproduct supply curves, including supply curvesfor oxygenates and unfinished oils, are also calculatedand provided to the PMM. A world oil supply/demandbalance is created, including estimates for 16 oil consumptionregions and 19 oil production regions. Theoil production estimates include both conventionaland nonconventional supply recovery technologies.Residential and Commercial Demand ModulesThe Residential Demand Module projects energy consumptionin the residential sector by housing typeand end use, based on delivered energy prices, themenu of equipment available, the availability ofrenewable sources of energy, and housing starts. TheCommercial Demand Module projects energy consumptionin the commercial sector by building typeand nonbuilding uses of energy and by category of enduse, based on delivered prices of energy, availability ofrenewable sources of energy, and macroeconomicvariables representing interest rates and floorspaceconstruction.Both modules estimate the equipment stock for themajor end-use services, incorporating assessments ofadvanced technologies, including representations ofrenewable energy technologies and effects of bothbuilding shell and appliance standards. The CommercialDemand Module incorporates combined heat andpower (CHP) technology. The modules also includeprojections of distributed generation. Both modulesincorporate changes to “normal” heating and coolingdegree-days by Census division, based on State-levelpopulation projections. The Residential DemandModule projects that the average square footage ofboth new construction and existing structures isincreasing based on trends in the size of new constructionand the remodeling of existing homes.Industrial Demand ModuleThe Industrial Demand Module projects the consumptionof energy for heat and power and forfeedstocks and raw materials in each of 16 industrygroups, subject to the delivered prices of energy andmacroeconomic variables representing employmentand the value of shipments for each industry. Asnoted in the description of the Macroeconomic Module,the value of shipments is based on NAICS. Theindustries are classified into three groups—energyintensivemanufacturing, non-energy-intensive manufacturing,and nonmanufacturing. Of the 8 energyintensiveindustries, 7 are modeled in the IndustrialDemand Module, with components for boiler/steam/cogeneration, buildings, and process/assembly use ofenergy. Bulk chemicals are further disaggregated toorganic, inorganic, resins, and agricultural chemicals.A representation of cogeneration and a recycling componentare also included. The use of energy for petroleumrefining is modeled in the Petroleum Market200 Energy Information Administration / Annual Energy Outlook 2006

NEMS Overview and Brief Description of CasesModule, and the projected consumption is included inthe industrial totals.Transportation Demand ModuleThe Transportation Demand Module projects consumptionof fuels in the transportation sector, includingpetroleum products, electricity, methanol,ethanol, compressed natural gas, and hydrogen, bytransportation mode, vehicle vintage, and size class,subject to delivered prices of energy fuels and macroeconomicvariables representing disposable personalincome, GDP, population, interest rates, and thevalue of output for industries in the freight sector.Fleet vehicles are represented separately to allowanalysis of CAAA and other legislative proposals.The module also includes a component to assess thepenetration of alternative-fuel vehicles explicitly. Theair transportation module explicitly represents theindustry practice of parking aircraft to reduce operatingcosts and the movement of aircraft from passengerto cargo markets as aircraft age [7]. For air freightshipments, the model employs narrow-body andwide-body aircraft only. The model also uses an infrastructureconstraint that limits growth in air travel tolevels commensurate with industry-projected infrastructureexpansion and capacity growth.Electricity Market ModuleThe Electricity Market Module (EMM) representsgeneration, transmission, and pricing of electricity,subject to delivered prices for coal, petroleum products,natural gas, and biofuels; costs of generation byall generation plants, including capital costs; macroeconomicvariables for costs of capital and domesticinvestment; enforced environmental emissions lawsand regulations; and electricity load shapes anddemand. There are three primary submodules—capacity planning, fuel dispatching, and finance andpricing. Nonutility generation, distributed generation,and transmission and trade are modeled in theplanning and dispatching submodules. The levelizedcost of uranium fuel for nuclear generation is incorporateddirectly in the EMM.All specifically identified CAAA compliance optionsthat have been promulgated by the EPA are explicitlyrepresented in the capacity expansion and dispatchdecisions; those that have not been promulgated arenot incorporated (e.g., fine particulate proposal). Allspecifically identified EPACT2005 financial incentivesfor power generation expansion and dispatchhave been implemented. Several States, primarily inthe Northeast, have recently enacted air emissionregulations that affect the electricity generation sector.Where firm State compliance plans have beenannounced, regulations are represented in AEO2006.Renewable Fuels ModuleThe Renewable Fuels Module (RFM) includessubmodules representing renewable resource supplyand technology input information for central-station,grid-connected electricity generation technologies,including conventional hydroelectricity, biomass(wood, energy crops, and biomass co-firing), geothermal,landfill gas, solar thermal electricity, solarphotovoltaics, and wind energy. The RFM containsrenewable resource supply estimates representingthe regional opportunities for renewable energydevelopment. Investment tax credits for renewablefuels are incorporated, as currently legislated in theEPACT1992 and EPACT2005. EPACT1992 providesa 10-percent tax credit for business investment insolar energy (thermal non-power uses as well aspower uses) and geothermal power. EPACT2005increases the tax credit to 30 percent for solar energysystems installed before January 1, 2008. The creditshave no expiration dates.Production tax credits for wind, geothermal, landfillgas, and some types of hydroelectric and biomass-fueledplants are also represented. They providea tax credit of up to 1.9 cents per kilowatthour forelectricity produced in the first 10 years of plant operation.New plants that come on line before January 1,2008, are eligible to receive the credit. Significantchanges made for AEO2006 in the accounting of newrenewable energy capacity resulting from Staterenewable portfolio standards, mandates, and goalsare described in Assumptions to the Annual EnergyOutlook 2006 [8].Oil and Gas Supply ModuleThe Oil and Gas Supply Module represents domesticcrude oil and natural gas supply within an integratedframework that captures the interrelationshipsamong the various sources of supply: onshore, offshore,and Alaska by both conventional andnonconventional techniques, including natural gasrecovery from coalbeds and low-permeability formationsof sandstone and shale. This framework analyzescash flow and profitability to computeinvestment and drilling for each of the supplysources, based on the prices for crude oil and naturalEnergy Information Administration / Annual Energy Outlook 2006 201

NEMS Overview and Brief Description of Casesgas, the domestic recoverable resource base, and thestate of technology. Oil and gas production functionsare computed at a level of 12 supply regions, including3 offshore and 3 Alaskan regions. This module alsorepresents foreign sources of natural gas, includingpipeline imports and exports to Canada and Mexico,and liquefied natural gas (LNG) imports and exports.Crude oil production quantities are input to thePetroleum Market Module in NEMS for conversionand blending into refined petroleum products. Supplycurves for natural gas are input to the Natural GasTransmission and Distribution Module for use indetermining natural gas prices and quantities. InternationalLNG supply sources and options for regionalexpansions of domestic regasification capacity arerepresented, based on the projected regional costsassociated with international gas supply, liquefaction,transportation, and regasification and worldnatural gas market conditions.Natural Gas Transmission and DistributionModuleThe Natural Gas Transmission and DistributionModule represents the transmission, distribution,and pricing of natural gas, subject to end-use demandfor natural gas and the availability of domestic naturalgas and natural gas traded on the internationalmarket. The module tracks the flows of natural gas inan aggregate domestic pipeline network, connectingthe domestic and foreign supply regions with 12demand regions. This capability allows the analysis ofimpacts of regional capacity constraints in the interstatenatural gas pipeline network and the identificationof pipeline capacity expansion requirements. Theflow of natural gas is determined for both a peak andoff-peak period in the year. Key components of pipelineand distributor tariffs are included in separatepricing algorithms.Petroleum Market ModuleThe Petroleum Market Module (PMM) projects pricesof petroleum products, crude oil and product importactivity, and domestic refinery operations (includingfuel consumption), subject to the demand for petroleumproducts, the availability and price of importedpetroleum, and the domestic production of crude oil,natural gas liquids, and alcohol and biodiesel fuels.The module represents refining activities in the fivePetroleum Administration for Defense Districts(PADDs), using the same crude oil types representedin the International Energy Module. It explicitlymodels the requirements of CAAA and the costs ofautomotive fuels, such as conventional and reformulatedgasoline, and includes biofuels production forblending in gasoline and diesel.AEO2006 reflects State legislation that bans or limitsthe use of the gasoline blending component methyltertiary butyl ether (MTBE) in Arizona, California,Colorado, Connecticut, Illinois, Indiana, Iowa, Kansas,Kentucky, Maine, Michigan, Minnesota, Missouri,Montana, Nebraska, New Hampshire, NewJersey, New York, North Carolina, Ohio, RhodeIsland, South Dakota, Vermont, Washington, andWisconsin. Furthermore, MTBE is assumed to bephased out by the end of 2008 as a result ofEPACT2005, which allows refiners to discontinue useof oxygenates in reformulated gasoline, and becauseof concern about MTBE contamination of surfacewater and groundwater resources.The nationwide phase-in of gasoline with an annualaverage sulfur content of 30 ppm between 2005 and2007, regulations that limit the sulfur content ofhighway diesel fuel to 15 ppm starting in mid-2006and of all nonroad and locomotive/marine diesel to 15ppm by mid-2012, and the renewable fuels standardof 7.5 billion gallons by 2012 are represented inAEO2006. Growth in demand and the costs of the regulationslead to capacity expansion for refinery-processingunits, assuming a financing ratio of 60percent equity and 40 percent debt, with a hurdle rateand an after-tax return on investment of about 9 percent[9]. End-use prices are based on the marginalcosts of production, plus markups representing productand distribution costs and State and Federal taxes[10]. Expansion of refinery capacity at existing sites ispermitted in all of the five refining regions modeled.Fuel ethanol and biodiesel are included in the PMM,because they are commonly blended into petroleumproducts. The module allows ethanol blending intogasoline at 10 percent by volume or less, as well aslimited quantities of E85, a blend of up to 85 percentethanol by volume. Ethanol is produced primarily inthe Midwest from corn or other starchy crops, and itis expected to be produced from cellulosic material inother regions in the future. Biodiesel is producedfrom soybean oil or yellow grease (primarily, recycledcooking oil). Both soybean oil biodiesel and yellowgrease biodiesel are assumed to be blended into highwaydiesel.Alternative fuels such as coal-to-liquids (CTL) andgas-to-liquids (GTL) are modeled in the PMM, based202 Energy Information Administration / Annual Energy Outlook 2006

NEMS Overview and Brief Description of Caseson their economics relative to competing feedstocksand products. CTL facilities are likely to be built atlocations close to coal supply sources, where liquidproducts and electricity could also be distributed tonearby demand regions. GTL facilities may be builton the North Slope of Alaska but would compete withthe Alaska Natural Gas Transportation System(ANGTS) for available natural gas resources. BothCTL and GTL are discussed in more detail in “Issuesin Focus.”Coal Market ModuleThe Coal Market Module (CMM) simulates mining,transportation, and pricing of coal, subject to theend-use demand for coal differentiated by heat andsulfur content. U.S. coal production is represented inthe CMM using 40 separate supply curves—differentiatedby region, mine type, coal rank, and sulfur content.The coal supply curves include a response tocapacity utilization of mines, mining capacity, laborproductivity, and factor input costs (mining equipment,mining labor, and fuel requirements). Projectionsof U.S. coal distribution are determined in theCMM through the use of a linear programming algorithmthat determines the least-cost supplies of coalfor a given set of coal demands by demand region andsector, accounting for minemouth prices, transportationcosts, existing coal supply contracts, and sulfurand mercury allowance costs. Over the projectionhorizon, coal transportation costs in the CMM areprojected to vary in response to changes in railroadproductivity and the user cost of rail transportationequipment.The CMM produces projections of U.S. steam andmetallurgical coal exports and imports, in the contextof world coal trade. The CMM’s linear programmingalgorithm determines the pattern of world coal tradeflows that minimizes the production and transportationcosts of meeting a pre-specified set of regionalworld coal import demands, subject to constraints onexport capacities and trade flows. The internationalcoal market component of the module computes tradein 3 types of coal for 16 export and 20 import regions.U.S. coal production and distribution are computedfor 14 supply and 14 demand regions.Annual Energy Outlook 2006 CasesTable E1 provides a summary of the cases used toderive the AEO2006 projections. For each case, thetable gives the name used in this report, a briefdescription of the major assumptions underlying theprojections, a designation of the mode in which thecase was run in NEMS (either fully integrated, partiallyintegrated, or standalone), and a reference tothe pages in the body of the report and in this appendixwhere the case is discussed. The following sectionsdescribe the cases listed in Table E1. The referencecase assumptions for each sector are described at website www.eia.doe.gov/oiaf/aeo/assumption/. Regionalresults and other details of the projections areavailable at web site www.eia.doe.gov/oiaf/aeo/supplement/.Macroeconomic Growth CasesIn addition to the AEO2006 reference case, the loweconomic growth and high economic growth caseswere developed to reflect the uncertainty in projectionsof economic growth. The alternative cases areintended to show the effects of alternative growthassumptions on energy market projections. The casesare described as follows:• The low economic growth case assumes lowergrowth rates for population (0.5 percent per year),nonfarm employment (0.7 percent per year), andproductivity (1.8 percent per year), resulting inhigher prices and interest rates and lower growthin industrial output. In the low economic growthcase, economic output increases by 2.4 percent peryear from 2004 through 2030, and growth in GDPper capita averages 1.9 percent per year.• The high economic growth case assumes highergrowth rates for population (1.1 percent per year),nonfarm employment (1.4 percent per year), andproductivity (2.7 percent per year). With higherproductivity gains and employment growth, inflationand interest rates are lower than in the referencecase, and consequently economic outputgrows at a higher rate (3.5 percent per year) thanin the reference case (3.0 percent). GDP per capitagrows by 2.4 percent per year, compared with 2.2percent in the reference case.Price CasesThe world oil price in AEO2006 is represented by theaverage U.S. refiners acquisition costs of importedlow-sulfur light crude oil, in order to be more consistentwith prices typically reported in the media. Thelow-sulfur light crude oil price is similar to the WestTexas Intermediate (WTI) crude oil price. AEO2006also includes a projection of the annual average U.S.Energy Information Administration / Annual Energy Outlook 2006 203

NEMS Overview and Brief Description of CasesTable E1. Summary of the AEO2006 casesCase nameReferenceLow Economic GrowthHigh Economic GrowthLow PriceHigh PriceResidential:2005 TechnologyResidential:High TechnologyResidential: BestAvailable TechnologyCommercial:2005 TechnologyCommercial:High TechnologyCommercial: BestAvailable TechnologyIndustrial:2005 TechnologyIndustrial:High TechnologyTransportation:2005 TechnologyDescriptionBaseline economic growth (3.0 percent per year), worldoil price, and technology assumptions. Completeprojection tables in Appendix A.Gross domestic product grows at an average annual rateof 2.4 percent from 2004 through 2030. Subset ofprojection tables in Appendix B.Gross domestic product grows at an average annual rateof 3.5 percent from 2004 through 2030. Subset ofprojection tables in Appendix B.More optimistic assumptions for worldwide crude oil andnatural gas resources than in the reference case. Worldoil prices are $28 per barrel in 2030, compared with $50per barrel in the reference case, and lower 48 wellheadnatural gas prices $4.96 per thousand cubic feet in 2030,compared with $5.92 in the reference case. Subset ofprojection tables in Appendix C.More pessimistic assumptions for worldwide crude oil andnatural gas resources than in the reference case. Worldoil prices are about $90 per barrel in 2030 and lower 48wellhead natural gas prices $7.72 per thousand cubic feetin 2030. Subset of projection tables in Appendix C.Future equipment purchases based on equipmentavailable in 2005. Existing building shell efficiencies fixedat 2005 levels. Partial projection tables in Appendix D.Earlier availability, lower costs, and higher efficienciesassumed for more advanced equipment. Building shellefficiencies increase by 22 percent from 2003 values by2030. Partial projection tables in Appendix D.Future equipment purchases and new building shellsbased on most efficient technologies available. Buildingshell efficiencies increase by 26 percent from 2003 valuesby 2030. Partial projection tables in Appendix D.Future equipment purchases based on equipmentavailable in 2005. Building shell efficiencies fixed at 2005levels. Partial projection tables in Appendix D.Earlier availability, lower costs, and higher efficienciesassumed for more advanced equipment. Building shellefficiencies for new and existing buildings increase by10.4 and 7.4 percent, respectively, from 1999 values by2030. Partial projection tables in Appendix D.Future equipment purchases based on most efficienttechnologies available. Building shell efficiencies for newand existing buildings increase by 12.4 and 8.9 percent,respectively, from 1999 values by 2030. Partial projectiontables in Appendix D.Efficiency of plant and equipment fixed at 2005 levels.Partial projection tables in Appendix D.Earlier availability, lower costs, and higher efficienciesassumed for more advanced equipment. Partial projectiontables in Appendix D.Efficiencies for new equipment in all modes of travel fixedat 2005 levels. Partial projection tables in Appendix D.IntegrationmodeFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedWithcommercialWithcommercialWithcommercialWithresidentialWithresidentialWithresidentialReferencein textReference inAppendix E— —p. 62 p. 203p. 62 p. 203p. 64 p. 206p. 64 p. 206p. 68 p. 206p. 68 p. 207p. 68 p. 207p. 70 p. 207p. 70 p. 207p. 70 p. 207Standalone p. 73 p. 207Standalone p. 73 p. 207Standalone p. 76 p. 208204 Energy Information Administration / Annual Energy Outlook 2006

Table E1. Summary of the AEO2006 cases (continued)NEMS Overview and Brief Description of CasesCase nameTransportation:High TechnologyTransportation:Alternative CAFEIntegrated2005 TechnologyIntegratedHigh TechnologyElectricity: AdvancedNuclear CostElectricity: NuclearVendor EstimateElectricity: Low FossilTechnologyElectricity: High FossilTechnologyElectricity: MercuryControl TechnologiesRenewables:Low RenewablesRenewables:High RenewablesOil and Gas:Slow TechnologyOil and Gas:Rapid TechnologyOil and Gas: Low LNGDescriptionReduced costs and improved efficiencies assumed foradvanced technologies. Partial projection tables inAppendix D.Assumes that manufacturers adhere to the proposedfleetwide increases in light truck CAFE standards to 24miler per gallon for model year 2011.Combination of the residential, commercial, industrial,and transportation 2005 technology cases, electricity lowfossil technology case, and assumption of renewabletechnologies fixed at 2005 levels. Partial projection tablesin Appendix D.Combination of the residential, commercial, industrial,and transportation high technology cases, electricity highfossil technology case, high renewables case, andadvanced nuclear cost case. Partial projection tables inAppendix D.New nuclear capacity assumed to have 20 percent lowercapital and operating costs in 2030 than in the referencecase. Partial projection tables in Appendix D.New nuclear capacity assumed to have lower capitalcosts based on vendor goals. Partial projection tables inAppendix D.New advanced fossil generating technologies assumednot to improve over time from 2006. Partial projectiontables in Appendix D.Costs and efficiencies for advanced fossil-fired generatingtechnologies improve by 10 percent in 2030 fromreference case values. Partial projection tables inAppendix D.Cost and performance for halogenated activated carboninjection technology used to determine its impact onmercury removal requirements from coal-fired powerplants.New renewable generating technologies assumed not toimprove over time from 2006. Partial projection tables inAppendix D.Levelized cost of energy for nonhydropower renewablegenerating technologies declines by 10 percent in 2030from reference case values. Lower capital cost forcellulose ethanol plants. Partial projection tables inAppendix D.Cost, finding rate, and success rate parameters adjustedfor 50-percent slower improvement than in the referencecase. Partial projection tables in Appendix D.Cost, finding rate, and success rate parameters adjustedfor 50-percent more rapid improvement than in thereference case. Partial projection tables in Appendix D.LNG imports exogenously set to 30 percent less thanthe results from the high price case, with remainingassumptions from the reference case. Partial projectiontables in Appendix D.IntegrationmodeReferencein textReference inAppendix EStandalone p. 76 p. 208Standalone p. 24 p. 208FullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedFullyintegratedp. 60 —p. 60 —p. 84 p. 208p. 84 p. 208p. 83 p. 209p. 83 p. 208p. 59 p. 209p. 84 p. 209p. 84 p. 209p. 88 p. 210p. 88 p. 209p. 90 p. 210Energy Information Administration / Annual Energy Outlook 2006 205

NEMS Overview and Brief Description of CasesTable E1. Summary of the AEO2006 cases (continued)Case nameOil and Gas: High LNGDescriptionLNG imports exogenously set to 30 percent more thanthe results from the low price case, with remainingassumptions from the reference case. Partial projectiontables in Appendix D.IntegrationmodeReferencein textReference inAppendix EFullyintegratedp. 90 p. 210Oil and Gas: ANWRFederal oil and gas leasing permitted in theArctic National Wildlife Refuge starting in 2005. Partialprojection tables in Appendix D.Fullyintegratedp. 94 p. 210Coal: Low CostProductivity for coal mining and coal transportationassumed to increase more rapidly than in the referencecase. Coal mining wages, mine equipment and coaltransportation equipment costs assumed to be lower thanin the reference case. Partial projection tables inAppendix D.Fullyintegratedp. 102 p. 210Coal: High CostProductivity for coal mining and coal transportationassumed to increase more slowly than in the referencecase. Coal mining wages, mine equipment and coaltransportation equipment costs assumed to be higherthan in the reference case. Partial projection tables inAppendix D.Fullyintegratedp. 102 p. 210refiners acquisition cost of imported crude oil (IRAC),which is more representative of the average cost of allcrude oil used by refiners.The historical record shows substantial variability inworld oil prices, and there is arguably even moreuncertainty about future prices in the long term.AEO2006 considers three price cases (reference case,low price case, and high price case) to allow an assessmentof alternative views on the course of future oiland natural gas prices. In the reference case, world oilprices moderate from current levels through 2015before beginning to rise to $57 per barrel in 2030(2004 dollars). The low and high price cases define awide range of potential price paths (from $34 to $96per barrel in 2030). The two cases reflect differentassumptions about the availability of world oil andnatural gas resources and production costs; they donot assume changes in OPEC behavior. Because thelow and high price cases are not directly integratedwith a world economic model, the impact of world oilprices on international economies is not directlyaccounted for in this analysis.• The reference case represents EIA’s current judgmentregarding the expected behavior of OPECproducers in the long term, adjusting productionto keep world oil prices in a range of $40 to $50 perbarrel, in keeping with OPEC’s stated goal ofkeeping potential competitors from eroding itsmarket share. Because OPEC (and particularlythe Persian Gulf nations) is expected to be thedominant supplier of oil in the international marketover the long term, its production choices willsignificantly affect world oil prices.• The low price case assumes greater world crudeoil and natural gas resources which are less expensiveto produce and a future market where all oiland natural gas production becomes more competitiveand plentiful than the reference case.• The high price case assumes that world crude oiland natural gas resources, including OPEC’s, arelower and require greater cost to produce thanassumed in the reference case.Buildings Sector CasesIn addition to the AEO2006 reference case, threestandalone technology-focused cases using the Residentialand Commercial Demand Modules of NEMSwere developed to examine the effects of changes toequipment and building shell efficiencies.For the residential sector, the three technologyfocusedcases are as follows:• The 2005 technology case assumes that all futureequipment purchases are based only on the rangeof equipment available in 2005. Existing buildingshell efficiencies are assumed to be fixed at 2005levels.206 Energy Information Administration / Annual Energy Outlook 2006

NEMS Overview and Brief Description of Cases• The high technology case assumes earlier availability,lower costs, and higher efficiencies formore advanced equipment [11]. Building shellefficiency in 2030 is assumed to be 22 percenthigher than the 2003 level.• The best available technology case assumes thatall future equipment purchases are made from amenu of technologies that includes only the mostefficient models available in a particular year,regardless of cost. Building shell efficiency in 2030is assumed to be 26 percent higher than the 2003level.For the commercial sector, the three technologyfocusedcases are as follows:• The 2005 technology case assumes that all futureequipment purchases are based only on the rangeof equipment available in 2005. Building shell efficienciesare assumed to be fixed at 2005 levels.• The high technology case assumes earlier availability,lower costs, and/or higher efficiencies formore advanced equipment than in the referencecase [12]. Building shell efficiencies for new andexisting buildings in 2030 are assumed to be 10.4percent and 7.4 percent higher, respectively, thantheir 1999 levels—a 25-percent improvement relativeto the reference case.• The best available technology case assumes thatall future equipment purchases are made from amenu of technologies that includes only the mostefficient models available in a particular year,regardless of cost. Building shell efficiencies fornew and existing buildings in 2030 are assumed tobe 12.4 percent and 8.9 percent higher, respectively,than their 1999 values—a 50-percentimprovement relative to the reference case.Two additional integrated cases were developed, incombination with assumptions for electricity generationfrom renewable fuels, to analyze the sensitivityof the projections to changes in generating technologiesthat use renewable fuels and in the availability ofrenewable energy sources. For the Residential andCommercial Demand Modules:• The high renewables case assumes greater improvementsin residential and commercial photovoltaicsystems than in the reference case. Thehigh renewables assumptions result in capitalcost estimates for 2030 that are approximately 10percent lower than reference case costs for distributedphotovoltaic technologies.• The low renewables case assumes that costs andperformance levels for residential and commercialphotovoltaic systems remain constant at 2005 levelsthrough 2030.Industrial Sector CasesIn addition to the AEO2006 reference case, twostandalone cases using the Industrial Demand Moduleof NEMS were developed to examine the effects ofless rapid and more rapid technology change andadoption. The Industrial Demand Module was alsoused as part of an integrated high renewables case.For the industrial sector:• The 2005 technology case holds the energy efficiencyof plant and equipment constant at the2005 level over the projection period. In this case,delivered energy intensity falls by 0.9 percentannually. Because the level and composition ofindustrial output are the same in the reference,2005 technology, and high technology cases, anychange in primary energy intensity in the twotechnology cases is attributable to efficiencychanges. The 2005 technology case was run withonly the Industrial Demand Module, rather thanin fully integrated NEMS runs. Consequently, nopotential feedback effects from energy marketinteractions were captured.• The high technology case assumes earlier availability,lower costs, and higher efficiency for moreadvanced equipment [13] and a more rapid rate ofimprovement in the recovery of biomass byproductsfrom industrial processes (0.7 percent peryear, as compared with 0.4 percent per year in thereference case). The same assumption is alsoincorporated in the integrated high renewablescase, which focuses on electricity generation.While the choice of 0.7 percent recovery is anassumption of the high technology case, it is basedon the expectation that there would be higherrecovery rates and substantially increased use ofCHP in that case. Changes in aggregate energyintensity result both from changing equipmentand production efficiency and from changing compositionof industrial output. Because the compositionof industrial output remains the same as inthe reference case, delivered energy intensity fallsby 1.4 percent annually in the high technologycase. In the reference case, delivered energy intensityfalls by 1.2 percent annually between 2004and 2030.Energy Information Administration / Annual Energy Outlook 2006 207

NEMS Overview and Brief Description of CasesTransportation Sector CasesIn addition to the AEO2006 reference case, twostandalone cases using the Transportation DemandModule of NEMS were developed to examine theeffects of less rapid technology change and adoptionand more rapid technology change and adoption. Forthe transportation sector:• The 2005 technology case assumes that new vehiclefuel efficiencies remain constant at 2005 levelsthrough the projection horizon, unless emissionsand/or efficiency regulations require the implementationof technology that affects vehicle efficiency.For example, the new light truck corporateaverage fuel economy (CAFE) standards requirean increase in fuel economy through 2007, andincreases in heavy truck emissions standards arerequired through 2010. As a result, the technologyavailable for light truck efficiency improvement isfrozen at 2007 levels, and the technology availableto heavy trucks is frozen at 2010 levels.• In the high technology case, the characteristics oflight-duty conventional and alternative-fuel vehiclesreflect more optimistic assumptions aboutincremental improvements in fuel economy andcosts [14]. In the air travel sector, the high technologycase reflects lower costs for improved thermodynamics,advanced aerodynamics, andweight-reducing materials, providing a 25-percentimprovement in new aircraft efficiency relativeto the reference case in 2025. In the freighttruck sector, the high technology case assumesmore incremental improvement in fuel efficiencyfor engine and emissions control technologies[15]. More optimistic assumptions for fuel efficiencyimprovements are also made for the railand shipping sectors.Both cases were run with only the TransportationDemand Module rather than as fully integratedNEMS runs. Consequently, no potential macroeconomicfeedback on travel demand was captured, norwere changes in fuel prices incorporated.In addition to these standalone cases, EIA also developedan alternative CAFE case designed to examinethe potential energy impacts of proposed reforms tothe structure of CAFE standards for light trucks andincreases in light truck CAFE standards for modelyears 2008 through 2011 [16]. The alternative CAFEcase assumes that manufacturers will adhere to theproposed fleet-wide increases in light truck CAFEstandards, to 24 miles per gallon for model year 2011.Electricity Sector CasesIn addition to the reference case, four integratedcases with alternative electric power assumptionswere developed to analyze uncertainties about thefuture costs and performance of new generating technologies.Two of the cases examine alternativeassumptions for nuclear power technologies, and twoexamine alternative assumptions for fossil fuel technologies.Reference case values for technology characteristicsare determined in consultation with industryand government specialists; however, there is alwaysuncertainty surrounding newer, untested designs.The electricity cases analyze what could happen ifcosts of advanced designs are either higher or lowerthan assumed in the reference case. The cases arefully integrated to allow feedback between the potentialshifts in fuel consumption and fuel prices.Nuclear Technology Cases• The cost assumptions for the advanced nuclearcost case reflect a 20-percent reduction in the capitaland operating costs for advanced nuclear technologyin 2030, relative to the reference case. Thereference case, which assumes that some learningoccurs regardless of new orders and construction,projects a 14-percent reduction in the capital costsof nuclear power plants between 2006 and 2030.The advanced nuclear cost case assumes a 31-percent reduction between 2006 and 2030.• The nuclear vendor estimate case uses assumptionsthat are consistent with estimates from BritishNuclear Fuels Limited (Westinghouse) for themanufacture of its AP1000 advanced pressurized-waterreactor. In this case, the overnight capitalcost of a new advanced nuclear unit isassumed to be 18 percent lower initially thanassumed in the reference case and 44 percentlower in 2030. In both of the alternative nuclearcases, cost and performance characteristics for allother technologies are as assumed in the referencecase.Fossil Technology Cases• In the high fossil technology case, capital costs,heat rates, and operating costs for advanced coaland natural gas generating technologies areassumed to be 10 percent lower than referencecase levels in 2030. Because learning is assumedto occur in the reference case, costs and performancein the high case are reduced from initiallevels by more than 10 percent. Heat rates in the208 Energy Information Administration / Annual Energy Outlook 2006

NEMS Overview and Brief Description of Caseshigh fossil technology case fall to between 16 and22 percent below initial levels, and capital costsare reduced by 22 to 26 percent between 2006 and2030, depending on the technology.• In the low fossil technology case, capital costs andheat rates for coal gasification combined-cycleunits and advanced combustion turbine and combined-cycleunits do not decline during the projectionperiod but remain fixed at the 2006 valuesassumed in the reference case.Details about annual capital costs, operating andmaintenance costs, plant efficiencies, and other factorsused in the high and low fossil technology casesare described in the detailed assumptions, whichare available at web site www.eia.doe.gov/oiaf/aeo/assumption/.An additional integrated case was also run to analyzethe potential impacts of improved mercury controltechnologies to comply with CAMR. A detaileddescription of the rule is included in “Legislation andRegulations.”• In the mercury control technology case, the costand performance for halogenated activated carboninjection technology are used to determine itsimpact on mercury removal requirements fromcoal-fired power plants. Conventional activatedcarbon injection has not been effective in achievinghigh mercury removal rates from subbituminousand lignite coals, but preliminary tests showthat high levels of mercury removal can beachieved with relatively low rates of brominatedactivated carbon injection. If brominated activatedcarbon becomes commercially available by2018, it could have significant impacts on the costof achieving mercury removal targets.Renewable Fuels CasesIn addition to the AEO2006 reference case, two integratedcases with alternative assumptions aboutrenewable fuels were developed to examine the effectsof less aggressive and more aggressive improvementin renewable technologies. The cases are as follows:• In the low renewables case, capital costs, operationsand maintenance costs, and performancelevels for wind, solar, biomass, and geothermalresources are assumed to remain constant at 2006levels through 2030.• In the high renewables case, the levelized costs ofenergy for nonhydroelectric generating technologiesusing renewable resources are assumed todecline to 10 percent below the reference casecosts for the same resources in 2030. For mostrenewable resources, lower costs are accomplishedby reducing the capital costs of new plantconstruction. To reflect recent trends in windenergy cost reductions, however, it is assumedthat wind plants ultimately achieve the 10-percentcost reduction through a combination ofperformance improvement (increased capacityfactor) and capital cost reductions. Biomass suppliesare also assumed to be 10 percent greater foreach supply step. Annual limits are placed on thedevelopment of geothermal sites, because theyrequire incremental development to assure thatthe resource is viable. In the high renewables case,the annual limits on capacity additions at geothermalsites are raised from 25 megawatts per yearthrough 2015 to 50 megawatts per year for all projectionyears. All other cases are assumed to retainthe 25-megawatt limit through 2015. Other generatingtechnologies and projection assumptionsremain unchanged from those in the referencecase. In the high renewables case, the rate ofimprovement in recovery of biomass byproductsfrom industrial processes is also increased. Morerapid improvement in cellulosic ethanol productiontechnology is also assumed, resulting in lowercost for cellulose ethanol at any level of outputthan in the reference case.Oil and Gas Supply CasesTwo alternative technology cases were created toassess the sensitivity of the projections to changes inthe assumed rates of progress in oil and natural gassupply technologies. In addition, high and low LNGsupply cases were developed to examine the impactsof variations in LNG supply on the domestic naturalgas market.• In the rapid technology case, the parameters representingthe effects of technological progresson finding rates, drilling, lease equipment andoperating costs, and success rates for conventionaloil and natural gas drilling in the referencecase were increased by 50 percent. A number ofkey exploration and production technologies forunconventional natural gas were also increasedby 50 percent in the rapid technology case. KeyCanadian supply parameters were also modifiedto simulate the assumed impacts of more rapid oiland natural gas technology penetration on theCanadian supply potential. All other parametersEnergy Information Administration / Annual Energy Outlook 2006 209

NEMS Overview and Brief Description of Casesin the model were kept at the reference case values,including technology parameters for othermodules, parameters affecting foreign oil supply,and assumptions about imports and exports ofLNG and natural gas trade between the UnitedStates and Mexico. Specific detail by region andfuel category is presented in Assumptions to theAnnual Energy Outlook 2006, available at website www. eia.doe.gov/oiaf/aeo/assumption/.• In the slow technology case, the parameters representingthe effects of technological progress onfinding rates, drilling, lease equipment and operatingcosts, and success rates for conventional oiland natural gas drilling in the AEO2006 referencecase were reduced by 50 percent. A number of keyexploration and production technologies forunconventional natural gas were also reduced by50 percent in the slow technology case. Key Canadiansupply parameters were also modified to simulatethe assumed impacts of slow oil and naturalgas technology penetration on Canadian supplypotential. All other parameters in the model werekept at the reference case values.• The high LNG case exogenously specifies LNGimports at levels 30 percent higher than projectedin the low price case. The intent is to project thepotential impact on domestic markets if LNGimports turn out to be higher than projected inthe reference case.• The low LNG case exogenously specifies LNGimports at levels 30 percent lower than projectedin the high price case. The intent is to project thepotential impact on domestic markets if LNGimports turn out to be lower than projected in thereference case.• The ANWR case assumes that the U.S. Congresswill approve leasing in the 1002 Area Federallands in the Arctic National Wildlife Refuge for oiland natural gas exploration and production.Petroleum Market CasesIn addition to the AEO2006 reference case, a case thatis part of the integrated high renewable case evaluatesthe impact of more optimistic assumptions aboutbiomass supplies on the production and use of cellulosicethanol.• The high renewables case uses more optimisticassumptions about the availability of renewableenergy sources. The supply curve for cellulosicethanol is shifted in each projection year relativeto the reference case, making larger quantitiesavailable at any given price earlier than in the referencecase. More rapid improvement in cellulosicethanol production technology is also assumed,resulting in lower cost for cellulose ethanol at anylevel of output than in the reference case.Coal Market CasesTwo alternative coal cost cases examine the impactson U.S. coal supply, demand, distribution, and pricesthat result from alternative assumptions about miningproductivity, labor costs, and mine equipmentcosts on the production side, and railroad productivityand rail equipment costs on the transportationside. For the coal cost cases, adjustments to the referencecase assumptions for coal mining and railroadproductivity were based on variations in growth ratesobserved in the data for these industries since 1980.The low and high coal cost cases represent fully integratedNEMS runs, with feedback from the macroeconomicactivity, international, supply, conversion, andend-use demand modules.• In the low coal cost case, average annual productivitygrowth rates for coal mining and railroadproductivity are 2.5 percent and 2.6 percenthigher, respectively, than in the AEO2006 referencecase. On the mining side, adjustments to referencecase productivity are applied at the supplycurve level, while adjustments to railroad productivityare made at the regional level. Coal miningwages and mine equipment costs, which remainconstant in real dollars in the reference case, areassumed to decline by 1.0 percent per year in realterms in the low coal cost case. Railroad equipmentcosts, which are projected to increase by 2.1percent per year in constant dollars in the referencecase, are assumed to increase at a slower rateof 1.1 percent per year.• In the high coal cost case, average annual productivitygrowth rates for coal mining and railroadproductivity are 2.5 percent and 2.6 percentlower, respectively, than in the AEO2006 referencecase. Coal mining wages and mine equipmentcosts are assumed to increase by 1.0 percentper year in real terms. Railroad equipment costsare assumed to increase by 3.1 percent per year.Additional details about the productivity, wage, andequipment cost assumptions for the reference andalternative coal cost cases are provided in AppendixD.210 Energy Information Administration / Annual Energy Outlook 2006

NEMS Overview and Brief Description of CasesNotes[1]Energy Information Administration, The NationalEnergy Modeling System: An Overview 2003, DOE/EIA-0581(2003) (Washington, DC, March 2003).[2]Energy Policy Act of 2005, P.L. 109-58.[3]Energy Policy Act of 1992, P.L. 102-486, Title III, Section303, and Title V, Sections 501 and 507.[4]Energy Information Administration, Annual EnergyReview 2004, DOE/EIA-0384(2004) (Washington, DC,August 2005).[5]Energy Information Administration, Emissions ofGreenhouse Gases in the United States 2004, DOE/EIA-0573(2004) (Washington, DC, December 2005).[6]Energy Information Administration, Short-TermEnergy Outlook, web site www.eia.doe.gov/emeu/steo/pub/contents.html. Portions of the preliminary informationwere also used to initialize the NEMS PetroleumMarket Module projection.[7]Jet Information Services, Inc., World Jet InventoryYear-End 2003 (Woodinville, WA, March 2004), andpersonal communication from Bill Francoins (JetInformation Services) and Thomas C. Hoang (Boeing).[8]Energy Information Administration, Assumptions tothe Annual Energy Outlook 2006, DOE/EIA-0554(2006) (Washington, DC, to be published).[9]The hurdle rate for a coal-to-liquids (CTL) plant isassumed to be 12.3 percent because of the higher economicrisk involved in this technology.[10]For gasoline blended with ethanol, the tax credit of 51cents (nominal) per gallon of ethanol is assumed to beextended through 2030, based on the fact that the ethanoltax credit has been continuously in force for thepast 25 years and was recently extended from 2007 to2010 by the American Jobs Creation Act of 2004.[11]High technology assumptions are based on EnergyInformation Administration, EIA—Technology ForecastUpdates—Residential and Commercial BuildingTechnologies—Advanced Adoption Case (NavigantConsulting, Inc., September 2004).[12]High technology assumptions are based on EnergyInformation Administration, EIA—Technology ForecastUpdates—Residential and Commercial BuildingTechnologies—Advanced Adoption Case (NavigantConsulting, Inc., September 2004).[13]These assumptions are based in part on Energy InformationAdministration, Industrial Technology andData Analysis Supporting the NEMS Industrial Model(FOCIS Associates, October 2005).[14]Energy Information Administration, Documentationof Technologies Included in the NEMS Fuel EconomyModel for Passenger Cars and Light Trucks (Energyand Environmental Analysis, September 2003).[15]A. Vyas, C. Saricks, and F. Stodolsky, Projected Effectof Future Energy Efficiency and Emissions ImprovingTechnologies on Fuel Consumption of Heavy Trucks(Argonne, IL: Argonne National Laboratory, 2001).[16]National Highway Traffic Safety Administration,Average Fuel Economy Standards for Light TrucksModel Years 2008-2011, 49 CFR Parts 523, 533, and537, Docket No. 2005-22223, RIN 2127-AJ61 (Washington,DC, August 2005).Energy Information Administration / Annual Energy Outlook 2006 211

Appendix FRegional MapsEnergy Information Administration / Annual Energy Outlook 2006 213

Regional Maps214 Energy Information Administration / Annual Energy Outlook 2006

Regional Maps Energy Information Administration / Annual Energy Outlook 2006 215

Regional MapsF4. Natural Gas Transmission and Distribution Model RegionsAlaskaMacKenzieW. CanadaPrimary FlowsSecondary FlowsPipeline Border CrossingSpecific LNG TerminalsGeneric LNG TerminalsE. CanadaCanadaOffshoreand LNGPacific(9)Pacific(9)CA(12)Mountain(8)W. North Central(4)E. NorthCentral(3)Mid.Atlantic(2)NewEngl.(1)AZ/NM(11)W. South Central(7)E. SouthCentral(6)S. Atlantic(5)MexicoFL(10)BahamasSource: Energy Information Administration. Office of Integrated Analysis and Forecasting.216 Energy Information Administration / Annual Energy Outlook 2006

Regional Maps Energy Information Administration / Annual Energy Outlook 2006 217

Regional Maps500 0SCALE IN MILES1000 0SCALE IN MILES218 Energy Information Administration / Annual Energy Outlook 2006

Regional MapsEnergy Information Administration / Annual Energy Outlook 2006 219

Appendix GConversion Factors Energy Information Administration / Annual Energy Outlook 2006 221

The Energy Information Administration2006 EIA Energy Outlook and Modeling ConferenceRenaissance Hotel, Washington, DC March 27, 20068:30 a.m. - 8:45 Opening Remarks - Guy F. Caruso, Administrator, Energy Information Administration8:50 a.m. - 9:20 Overview of the Annual Energy Outlook 2006 - John Conti, Director,Office of Integrated Analysis and Forecasting, Energy Information Administration9:25 a.m. - 10:25 Keynote Address: International Oil Markets - Speaker to be announced10:40 a.m. - 12:10 Concurrent Sessions A1. Global Oil Market Outlook: Short-Term Issues2. The Future Relationship of Oil and Natural Gas Prices in the U.S.3. Nuclear—Is EPACT Enough?1:40 p.m. - 3:10 Concurrent Sessions B1. World Outlook for Unconventional Oil Production2. Globalization of the Natural Gas Market3. Return to Coal—From Where and at What Price?3:25 p.m. - 4:55 Concurrent Sessions C1. How Far Ethanol?2. Unconventional Natural Gas: Industry Savior or Bridge?3. If Not Natural Gas Then ...?HotelThe conference will be held at the Renaissance Hotel, (202) 898-9000. The Renaissance Hotel is located at 999 NinthStreet, NW, Washington, DC 20001, near the Gallery Place Metro station.InformationFor information, contact Peggy Wells, Energy Information Administration, at (202) 586-8845, peggy.wells@eia.doe.gov.Conference HandoutsHandouts provided in advance by the conference speakers will be posted online as available, at www.eia.doe.gov/oiaf/aeo/conf/handouts.html. They will not be provided at the conference.Register online at www.eia.doe.gov/oiaf/aeo/conf/Or mail or fax this form to:Peggy WellsEnergy Information Administration, EI-841000 Independence Avenue, SWWashington, DC 20585Phone: (202) 586-8845Fax: (202) 586-3045Conference RegistrationConference registration is free, but space is limited.Please register by March 20, 2006.Or register by e-mail to peggy.wells@eia.doe.gov.Please provide the information requested below:Name: _____________________________________________Title: _____________________________________________Organization: _______________________________________Address: ______________________________________________________________________________________Phone: ___________________________________________Fax: ___________________________________________E-mail: ___________________________________________Please indicate which sessions you will be attending: Opening Remarks/Overview/Keynote AddressConcurrent Sessions A Global Oil Market Outlook: Short-Term Issues The Future Relationship of Oil and Natural GasPrices in the U.S. Nuclear—Is EPACT Enough?Concurrent Sessions B World Outlook for Unconventional Oil Production Globalization of the Natural Gas Market Return to Coal—From Where and at What Price?Concurrent Sessions C How Far Ethanol? Unconventional Natural Gas: Industry Savioror Bridge? If Not Natural Gas Then ...?

AnnouncingThe Fourteenth AnnualEIA Energy Outlookand Modeling ConferenceMarch 27, 2006Renaissance HotelWashington, DCSee Previous Page for Information andRegistration Form