The National Energy Modeling System: An Overview...

DOE/EIA-0581(2003)

The National Energy ModelingSystem:

An Overview 2003

March 2003

Energy Information AdministrationOffice of Integrated Analysis and Forecasting

U.S. Department of EnergyWashington, DC 20585

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.

This publication is on the WEB at:

www.eia.doe.gov/oiaf/aeo/overview/index.html

The National Energy Modeling System: An Overview 2003 provides a summary description of the National En-ergy Modeling System (NEMS), which was used to generate the forecasts of energy production, demand, im-ports, and prices through the year 2025 for the Annual Energy Outlook 2003 (AEO2003),(DOE/EIA-0383(2003)), released in January 2003. AEO2003 presents national forecasts of energy markets forfive primary cases—a reference case and four additional cases that assume higher and lower economic growthand higher and lower world oil prices than in the reference case. The Overview presents a brief description ofthe methodology and scope of each of the component modules of NEMS. The model documentation reports listedin the appendix of this document provide further details.

The Overview was prepared by the Energy Information Administration (EIA), Office of Integrated Analysis andForecasting under the direction of Mary J. Hutzler ([email protected], 202/586-2222), Director of the Officeof Integrated Analysis and Forecasting; Scott Sitzer ([email protected], 202/586-2308), Director of the Coaland Electric Power Division; John J. Conti ([email protected], 202-586-4430), Director, International,Economic, and Greenhouse Gases Division; Paul D. Holtberg ([email protected], 202/586-1284), Direc-tor of the Demand and Integration Division; James M. Kendell ([email protected], 202/586-9646), Directorof the Oil and Gas Division; and Andy S. Kydes ([email protected], 202/586-2222), Senior Technical Advisor.Detailed questions concerning NEMS may be addressed to the following analysts:

AEO2003 is available on the EIA Home Page on the Internet (http://www.eia.doe.gov/oiaf/aeo/index.html). As-sumptions underlying the projections are available in Assumptions to the Annual Energy Outlook 2003 athttp://www.eia.doe.gov/oiaf/aeo/assumption/index.html. Tables of regional projections and other underlyingdetails of the reference case are available at http://www.eia.doe.gov/oiaf/aeo/supplement/index.html. Modeldocumentation reports and The National Energy Modeling System: An Overview 2003 are also available on theHome Page at http://www.eia.doe.gov/bookshelf/docs.html.

For ordering information and for questions on energy statistics, please contact EIA’s National Energy Informa-tion Center.

National Energy Information Center, EI-30Energy Information Administration

Forrestal BuildingWashington, DC 20585

Telephone: 202/586-8800FAX: 202/586-0727TTY: 202/586-1181

9 a.m. to 5 p.m., eastern time, M-FE-mail: [email protected]

World Wide Web Site: http://www.eia.doe.gov

FTP Site: ftp://ftp.eia.doe.gov

ii Energy Information Administration/The National Energy Modeling System: An Overview 2003

PREFACE

AEO2003 Paul D. Holtberg ([email protected], 202/586-1284)Integrating Module/Carbon Emissions Daniel H. Skelly ([email protected], 202/586-1722)Macroeconomic Activity Module Ronald Earley ([email protected], 202/586-1398)International Energy Module G. Daniel Butler ([email protected], 202/586-9503)Residential Demand Module John Cymbalsky ([email protected], 202/586-4815)Commercial Demand Module Erin Boedecker ([email protected], 202/586-4791)Industrial Demand Module Brian Unruh ([email protected], 202/586-1344)Transportation Demand Module John Maples ([email protected], 202/586-1757)Electricity Market Module Jeffrey Jones ([email protected], 202/586-2038)Renewable Fuels Module Thomas Petersik ([email protected], 202/586-6582)Oil and Gas Supply Module Ted McCallister ([email protected], 202/586-4820)Natural Gas Transmission and Distribution Module Joseph Benneche ([email protected], 202/586-6132)Petroleum Market Module Han-Lin Lee ([email protected], 202/586-4247)Coal Market Module Diane Kearney([email protected], 202/586-2415)

Introduction · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 1

Overview of NEMS· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 7

Carbon Dioxide and Methane Emissions · · · · · · · · · · · · · · · · · · · · · · · · · 13

Macroeconomic Activity Module· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 15

International Energy Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 17

Residential Demand Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 21

Commercial Demand Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 27

Industrial Demand Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 33

Transportation Demand Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 37

Electricity Market Module· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 43

Renewable Fuels Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 49

Oil and Gas Supply Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 53

Natural Gas Transmission and Distribution Module · · · · · · · · · · · · · · · · · · · 59

Petroleum Market Module· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 63

Coal Market Module · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 69

Appendix: Bibliography · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 73

CONTENTS

Energy Information Administration/The National Energy Modeling System: An Overview 2003 iii

CONTENTS

iv Energy Information Administration/The National Energy Modeling System: An Overview 2003

Figure 1. Census Divisions · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · 9

Figure 2. National Energy Modeling System · · · · · · · · · · · · · · · · · · · · · · · 10

Figure 3. Macroeconomic Activity Module Structure · · · · · · · · · · · · · · · · · · · 16

Figure 4. International Energy Module Structure · · · · · · · · · · · · · · · · · · · · 18

Figure 5. Residential Demand Module Structure · · · · · · · · · · · · · · · · · · · · 22

Figure 6. Commercial Demand Module Structure · · · · · · · · · · · · · · · · · · · · 28

Figure 7. Industrial Demand Module Structure · · · · · · · · · · · · · · · · · · · · · 34

Figure 8. Transportation Demand Module Structure · · · · · · · · · · · · · · · · · · 38

Figure 9. Electricity Market Module Structure· · · · · · · · · · · · · · · · · · · · · · 44

Figure 10. Electricity Market Module Supply Regions · · · · · · · · · · · · · · · · · · 45

Figure 11. Renewable Fuels Module Structure · · · · · · · · · · · · · · · · · · · · · · 50

Figure 12. Oil and Gas Supply Module Regions · · · · · · · · · · · · · · · · · · · · · 54

Figure 13. Oil and Gas Supply Module Structure · · · · · · · · · · · · · · · · · · · · 55

Figure 14. Foreign Natural Gas Trade via Pipeline and Liquefied Natural Gas Terminals 57

Figure 15. Natural Gas Transmission and Distribution Module Structure · · · · · · · 60

Figure 16. Natural Gas Transmission and Distribution Module Network· · · · · · · · 61

Figure 17. Petroleum Market Module Structure· · · · · · · · · · · · · · · · · · · · · 64

Figure 18. Petroleum Administration for Defense Districts · · · · · · · · · · · · · · · 64

Figure 19. Coal Market Module Demand Regions · · · · · · · · · · · · · · · · · · · · 70

Figure 20. Coal Market Module Supply Regions · · · · · · · · · · · · · · · · · · · · · 71

Figure 21. Coal Market Module Structure · · · · · · · · · · · · · · · · · · · · · · · · 72

The National Energy Modeling System (NEMS) is acomputer-based, energy-economy modeling systemof U.S. energy markets for the midterm periodthrough 2025. NEMS projects the production, im-ports, conversion, consumption, and prices of energy,subject to assumptions on macroeconomic and finan-cial factors, world energy markets, resource avail-ability and costs, behavioral and technological choicecriteria, cost and performance characteristics of en-ergy technologies, and demographics. NEMS was de-signed and implemented by the Energy InformationAdministration (EIA) of the U.S. Department of En-ergy (DOE).

The National Energy Modeling System: An Overview2003 presents an overview of the structure andmethodology of NEMS and each of its components.This chapter provides a description of the design andobjectives of the system, followed by a chapter on theoverall modeling structure and solution algorithm.The remainder of the report summarizes the meth-odology and scope of the component modules ofNEMS. The model descriptions are intended forreaders familiar with terminology from economics,operations research, and energy modeling. More de-tailed model documentation reports for all theNEMS modules are also available from EIA (Appen-dix, “Bibliography”).

Purpose of NEMS

NEMS is used by EIA to project the energy, eco-nomic, environmental, and security impacts on theUnited States of alternative energy policies and ofdifferent assumptions about energy markets. Projec-tions are made for each year from the presentthrough 2025. The forecast horizon is periodicallyextended to approximately 20 to 25 years into the fu-ture. This time period is one in which technology, de-mographics, and economic conditions are sufficientlyunderstood in order to represent energy marketswith a reasonable degree of confidence. NEMS pro-vides a consistent framework for representing thecomplex interactions of the U.S. energy system andits response to a wide variety of alternative assump-tions and policies or policy initiatives. As an annualmodel, NEMS can also provide the impacts of transi-tions to new energy programs and policies.

Energy resources and prices, the demand for specificenergy services, and other characteristics of energy

markets vary widely across the United States. To ad-dress these differences, NEMS is a regional model.The regional disaggregation for each module reflectsthe availability of data, the regional format typicallyused to analyze trends in the specific area, geology,and other factors, as well as the regions determinedto be the most useful for policy analysis. For exam-ple, the demand modules (e.g., residential, commer-cial, industrial and transportation) use the nine Cen-sus divisions, the Electricity Market Module uses 15supply regions based on the North American ElectricReliability Council (NERC) regions, the Oil and GasSupply Module uses 7 onshore and 3 offshore supplyregions based on geologic breakdowns, and the Pe-troleum Market Module uses 3 regions based on com-binations of the five Petroleum Administration forDefense Districts.

Baseline forecasts are developed with NEMS andpublished annually in the Annual Energy Outlook.In accordance with the requirement that EIA remainpolicy-neutral, the Annual Energy Outlook projec-tions are based on Federal, State, and local laws andregulations in affect at the time of the forecast. Anal-yses are also prepared in response to requests forspecial studies by the White House, U.S. Congress,the DOE Office of Policy, other offices in DOE, andother government agencies. The first version ofNEMS, completed in December 1993, was used todevelop the forecasts presented in the Annual En-ergy Outlook 1994, This report describes the ver-sion of NEMS used for the Annual Energy Outlook2003,1 which was extended to 2025 for the first time.

The forecasts produced by NEMS are not consideredto be statements of what will happen but of whatmight happen, given the assumptions and methodol-ogies used. Assumptions include, for example, the es-timated size of the economically recoverable resourcebase of fossil fuels, changes in world energy supplyand demand, the rate at which new energy technolo-gies are developed and the rate and extent of technol-ogy adoption and penetration.

Analytical Capability

NEMS can be used to analyze the effects of existingand proposed government laws and regulations re-lated to energy production and use; the potential im-pacts of new and advanced energy production, con-version, and consumption technologies; the impacts

Energy Information Administration/The National Energy Modeling System: An Overview 2003 1

INTRODUCTION

1 Energy Information Administration, Annual Energy Outlook 2003, DOE/EIA-0383(2003) (Washington, DC, January2003).

and costs of carbon emissions reductions, the im-pacts of increased use of renewable energy sources;the potential savings from increased efficiency of en-ergy use; and the changes in emission levels that arelikely to result from such policies as the Clean AirAct Amendments of 1990, regulations on the use ofalternative or reformulated fuels, and climatechange policy. Specific energy topics that can be, orhave been, addressed by NEMS include the follow-ing:

• Impacts of existing and proposed energy taxpolicies on the U.S. economy and energysystem

• Impacts on energy prices, energyconsumption, and electricity generation inresponse to carbon mitigation policies such ascarbon fees, limits on carbon emissions, orpermit trading systems

• Responses of the energy and economic systemsto changes in world oil market conditions as aresult of changing levels of foreign productionand demand in the developing countries

• Impacts of new technologies on consumptionand production patterns and emissions

• Effects of specific policies, such as mandatoryappliance efficiency and building shellstandards or renewable tax credits, on energyconsumption

• Impacts of fuel-use restrictions, for example,required use of oxygenated and reformulatedgasoline or mandated use of alternative-fueled vehicles, on emissions and energysupply and prices

• Impacts on the production and price of crudeoil and natural gas resulting fromimprovements in exploration and productiontechnologies

• Impacts on the price of coal resulting fromimprovements in productivity.

In addition to producing the analyses in the AnnualEnergy Outlook, NEMS is used for one-time analyti-cal reports and papers, such as Measuring Changesin Energy Efficiency for the Annual Energy Outlook2002,2 which describes the construction of an aggre-gate energy efficiency index based on projections ofsectoral and subsector energy consumption andsubsector–specific energy service indicators. The re-sults are compared with the ratio energy to realgross domestic product, which typically is presentedas a measure of energy intensity. Other analyticalpapers on topics of current interest in energy mar-kets are prepared, which either underlie the as-sumptions and methodology of NEMS or are applica-tions of NEMS to current issues. In the past, some ofthese papers have been collectively published in Is-sues in Midterm Analysis and Forecasting, and in thefuture they will be available at http://www.eia.doe.gov/ oiaf/analysis.html.

NEMS has also been used for a number of specialanalyses at the request of the White House, U.S.Congress, other offices of DOE and other govern-ment agencies, who specify the scenarios and as-sumptions for the analysis. Some recent examplesinclude:

• Analysis of Corporate Average Fuel Economy(CAFE) Standards for Light Trucks andIncreased Alternative Fuel Use,3 requested bySenator Murkowski to analyze the effects ofproposed provisions in S. 1766 and H.R. 4calling for more stringent corporate averagefuel economy standards on energy supply,demand, and prices, import dependence, andemissions.

• Analysis of Efficiency Standards for AirConditioners, Heat Pumps, and OtherProducts,4 requested by Senator Murkowski toevaluate the effects of the provisions in H.R. 4and S. 1766 that pertain to efficiency in the

2 Energy Information Administration/The National Energy Modeling System: An Overview 2003

INTRODUCTION

2 Energy Information Administration, Measuring Changes in Energy Efficiency for the Annual Energy Outlook 2002,(Washington, DC, 2002).

3 Energy Information Administration, Analysis of Corporate Average Fuel Economy (CAFE) Standards for Light Trucksand Increased Alternative Fuel Use, SR/OIAF/2002-05 (Washington, DC, March 2002).

4 Energy Information Administration, Analysis of Efficiency Standards for Air Conditioners, Heat Pumps, and OtherProducts, SR/OIAF/2002-01 (Washington, DC, February 2002).

residential, commercial and industrialsectors.

• Analysis of Strategies for Reducing MultipleEmissions from Electric Power Plants WithAdvanced Technology Scenarios,

5 requestedby Senators Jeffords and Lieberman toanalyze the impacts of technology improve-ments and other market–based opportunitieson the costs of emissions reductions.

• Impact of Renewable Fuels Standard/MTBEProvisions of S. 1766,6 requested by SenatorMurkowski to evaluate the Renewable FuelsStandard and methyl teritirary butyl etherprovisions of S. 1766.

• Impact of Renewable Fuels Standard/MTBEProvisions of S. 517,7 was completed as anaddendum to the service report Impact ofRenewable Fuels Standard/MTBE Provisionsof S. 17666 for Senators Murkowski andDaschle. The addendum to the service reportprovidws additional analysis of the impact ofthe Renewable Fuels Standard (RFS) andmethyl tertiary butyl ether (MTBE) banprovisions of S. 517. The projected consumercost of the S. 517 provisions is compared with aReference Case that assumes a 2 percentoxygen requirement is maintained and thatalready–scheduled MTBE restrictuons orbans become effective in 14 States.

• Analysis of Strategies for Reducing MultipleEmissions from Power Plants: Sulfur Dioxide,Nitrogen Oxides, and Carbon Dioxide

8,requested by the U.S. House Subcommittee onNational Economic Growth, NaturalResources, and Regulatory Affairs to analyzethe potential costs of multi–pollutantstrategies to reduce the emissions fromelectric power plants.

• Impacts of a 10–Percent Renewable PortfolioStandard,9 requested by Senator Murkowskito examine the Renewable Portfolio Standard(RPS) called for in S. 1766. The analysisincludes scenarios with a 10 percent RPS, asprovided in S. 1766, and a 20 percent RPS.

• Impacts of the Kyoto Protocol on U.S. EnergyMarkets & Economic Activity,10 requested bythe U.S. House Committee on Science toanalyze the impacts of the Kyoto Protocol onU.S. energy markets and the economy in the2008 to 2012 time frame.

• Reducing Emissions of Sulfur Dioxide,Nitrogen Oxides, and Mercury from ElectricPower Plants,11 requested by Senators Smith,Voinovich, and Brownback that describes theimpacts of scenarios with alternative powersector emission caps on nitrogen oxides, sulfurdioxides, and mercury.


INTRODUCTION

5 Energy Information Administration, Analysis of Strategies for Reducing Multiple Emissions from Electric Power PlantsWith Advanced Technology Scenarios, SR/OIAF/2001-05 (Washington, DC, October 2001).

6 Energy Information Administration, Impact of Renewable Fuels Standard/MTBE Provisions of S. 1766,SR/OIAF/2002-06 (Washington, DC, March 2002).

7 Energy Information Administration, Impact of Renewable Fuels Standard/MTBE Provisions of S. 517, SR/OIAF/2002-06Addendum (Washington, DC, April 2002).

8 Energy Information Administration, Analysis of Strategies for Reducing Multiple Emissions from Power Plants: SulfurDioxide, Nitrogen Oxides, and Carbon Dioxide, SR/OIAF/2000-05 (Washington, DC, December 2000).

9 Energy Information Administration, Impacts of a 10–Percent Renewable Portfolio Standard, SR/OIAF/2002-03(Washington, DC, February 2002).

10 Energy Information Administration, Impacts of the Kyoto Protocol on U.S. Energy Markets & Economic Activity,SR/OIAF/98-03 (Washington, DC, October 1998).

11 Energy Information Administration, Reducing Emissions of Sulfur Dioxide, Nitrogen Oxides, and Mercury from ElectricPower Plants, SR/OIAF/2001-04 (Washington, DC, September 2001).

• The Effects of the Alaska Oil and Natural GasProvisions of H.R. 4 and S. 1766 on U.S.Energy Markets,

12 requested by SenatorMurkowski to evaluate the effects of theprovisions in H.R. 4 proposing crude oilproduction in the Arctic National WildlifeRefuge and provisions in S. 1766 concerningthe construction of a pipeline bringingAlaskan natural gas to the Lower–48 States.

• The Transition to Ultra–Low–Sulfur DieselFuel: Effects on Prices and Supply,13 requestedby U.S. House Committee on Science to assessthe possible impact of the new sulfurrequirements on the diesel fuel market,specifically the implications for vehicle fuelefficiency, production, distribution, and cost.

• The Comprehensive Electricity CompetitionAct: A Comparison of Model Results,

14re-

quested by Secretary of Energy BillRichardson to evaluate the effects of theClinton Administration’s restructuringproposal using the parameter settings andassumptions from the Policy Office ElectricityModeling System analysis.

• U.S. Natural Gas Markets: Mid–TermProspects for Natural Gas Supply,15 requestedby Secretary of Energy, Spencer Abrahamdescribes the recent behavior of natural gasmarkets with respect to natural gas prices,their potential future behavior, the potentialfuture supply contribution of liquefied naturalgas and increased access to Federallyrestricted resources, and the need forimproved natural gas data.

• U.S. Natural Gas Markets: Recent Trends andProspects for the Future,16 requested by

Secretary of Energy, Spencer Abrahamexamines recent trends and prospects for thefuture of the U.S. natural gas markets inresponse to the dramatic increase in naturalgas prices during 2000 and early 2001 thatraised concerns about the future of naturalgas prices and the potential for natural gas tofuel the growth of the U.S. economy.

Representations of Energy MarketInteractions

NEMS is designed to represent the important inter-actions of supply and demand in U.S. energy mar-kets. In the United States, energy markets aredriven primarily by the fundamental economic inter-actions of supply and demand. Government regula-tions and policies can exert considerable influence,but the majority of decisions affecting fuel prices andconsumption patterns, resource allocation, and en-ergy technologies are made by private individualswho value attributes other than life cycle costs orcompanies attempting to optimize their own eco-nomic interests. NEMS represents the market be-havior of the producers and consumers of energy at alevel of detail that is useful for analyzing the implica-tions of technological improvements and policy ini-tiatives.

Energy Supply/Conversion/Demand Interactions

NEMS is designed as a modular system. Fourend-use demand modules represent fuel consump-tion in the residential, commercial, transportation,and industrial sectors, subject to delivered fuelprices, macroeconomic influences, and technologycharacteristics. The primary fuel supply and conver-sion modules compute the levels of domestic produc-tion, imports, transportation costs, and fuel pricesthat are needed to meet domestic and export de-mands for energy, subject to resource base character-


INTRODUCTION

12 Energy Information Administration, The Effects of the Alaska Oil and Natural Gas Provisions of H.R. 4 and S. 1766 onU.S. Energy Markets, SR/OIAF/2002-02 (Washington, DC, February 2002).

13 Energy Information Administration, The Transition to Ultra–Low–Sulfur Diesel Fuel: Effects on Prices and Supply,SR/OIAF/2001-01 (Washington, DC, May 2001).

14 Energy Information Administration, The Comprehensive Electricity Competition Act: A Comparison of Model Results,SR/OIAF/99-04 (Washington, DC, September 1999).

15 Energy Information Administration, U.S. Natural Gas Markets: Mid–Term Prospects for Natural Gas Supply,SR/OIAF/2001-06 (Washington, DC, December 2001).

16 Energy Information Administration, U.S. Natural Gas Markets: Recent Trends and Prospects for the Future,SR/OIAF/2001-02 (Washington, DC, May 2001).

istics, industry infrastructure and technology, andworld market conditions. The modules interact tosolve for the economic supply and demand balancefor each fuel. Because of the modular design, eachsector can be represented with the methodology andthe level of detail, including regional detail, that isappropriate for that sector. The modularity also fa-cilitates the analysis, maintenance, and testing ofthe NEMS component modules in the multi-user en-vironment.

Domestic Energy System/Economy Interactions

The general level of economic activity, representedby gross domestic product, has traditionally beenused as a key explanatory variable or driver for pro-jections of energy consumption at the sectoral andregional levels. In turn, energy prices and other en-ergy system activities influence economic growthand activity. NEMS captures this feedback betweenthe domestic economy and the energy system. Thus,changes in energy prices affect the key macroeco-nomic variables—such as gross domestic product,disposable personal income, industrial output, hous-ing starts, employment, and interest rates—thatdrive energy consumption and capacity expansiondecisions.

Domestic/World Energy Market Interactions

World oil prices play a key role in domestic energysupply and demand decision–making and oil priceassumptions are a typical starting point for energysystem projections. The level of oil production andconsumption in the U.S. energy system also has asignificant influence on world oil markets and prices.In NEMS, an international energy module repre-sents world oil production and demand, as well asthe interactions between the domestic and world oilmarkets, and this module calculates the averageworld crude oil price and the supply of specific crudeoils and petroleum products. As a result, domesticand world oil market projections are internally con-sistent. Imports and exports of natural gas, electric-ity, and coal—which are less influenced by volatileworld conditions—are represented in the individualfuel supply modules.

Economic Decision making Over Time

The production and consumption of energy productstoday are influenced by past investment decisions todevelop energy resources and acquire energy-usingcapital stock. Similarly, the production and con-sumption of energy in a future time period will be in-fluenced by decisions made today and in the past.

Current investment decisions depend on expecta-tions about future markets. For example, expecta-tions of rising energy prices in the future increasethe likelihood of current decisions to invest in moreenergy-efficient technologies or alternative energysources. A variety of assumptions about planning ho-rizons, the formation of expectations about the fu-ture, and the role of those expectations in economicdecision making are applied within the individualNEMS modules.

Technology Representation

A key feature of NEMS is the representation of tech-nology and technology improvement over time. Fiveof the sectors—residential, commercial, transpor-tation, electricity generation, and refining–includeextensive treatment of individual technologies andtheir characteristics, such as the initial capital cost,operating cost, date of availability, efficiency, andother characteristics specific to the sector. Techno-logical progress is lighting technologies results in agradual reduction in cost and is modeled as a func-tion of time in these end–use sectors. In addition, theelectricity sector accounts for technological optimismin the capital costs of first-of-a-kind generating tech-nologies and for a decline in cost as experience withthe technologies is gained both domestically and in-ternationally. In each of these sectors, equipmentchoices are made for individual technologies as newequipment is needed to meet growing demand for en-ergy services or to replace retired equipment.

In the other sectors—industrial, oil and gas supply,and coal supply—the treatment of technologies ismore limited due to a lack of data on individual tech-nologies. In the industrial sector, only the combinedheat and power and motor technologies are explicityconsidered and characaterized. Cost reductions re-sulting from technological progress in combined heatand power technologies is represented as a functionof time as experience with the technologies grows.Technological progress is not explicity modeled forthe industrial motor technologies. Other technolo-gies in the energy-intensive industries are repre-sented by technology bundles, with technology possi-bility curves representing efficiency improvementover time. In the oil and gas supply sector, technolog-ical progress is represented by econometricallyestimated improvements in finding rates, successrates, and costs. Productivity improvements overtime represent technological progress in coal produc-tion.


INTRODUCTION

External Availability

In accordance with EIA requirements, NEMS is fullydocumented and archived. EIA has been runningNEMS on three EIA workstations under the Win-dows 2000 operating system. The archive file pro-vides the source language, input files, and outputfiles to replicate the Annual Energy Outlook refer-ence case runs on an identically equipped computer;however, it does not include the proprietary portionsof the model, such as the Global Insights, Inc. (for-merly DRI/WEFA) macroeconomic model and the op-timization modeling libraries. NEMS can be run on

a high–powered individual PC as long as the re-quired proprietary software resides on the PC. Be-cause of the complexity of NEMS, and the relativelyhigh cost of the proprietary software, NEMS is notwidely used outside of the Department of Energy.However, NEMS, or portions of it, is installed at theLawrence Berkeley National Laboratory, Oak RidgeNational Laboratory, the Electric Power ResearchInstitute, the National Energy Technology Labora-tory, the National Renewable Energy Laboratory,and several private consulting firms.


INTRODUCTION

NEMS represents domestic energy markets by ex-plicitly representing the economic decision makinginvolved in the production, conversion, and con-sumption of energy products. Where possible, NEMSincludes explicit representation of energy technolo-gies and their characteristics.

Since energy costs and availability and en-ergy-consuming characteristics vary widely acrossregions, considerable regional detail is included.Other details of production and consumption cate-gories are represented to facilitate policy analysisand ensure the validity of the results. A summaryof the detail provided in NEMS is shown below.


OVERVIEW OF NEMS

Energ Acti it Categories Regions

Residential demand Sixteen end-use servicesThree housing typesThirty–four end–use technologies

Nine Census divisions

Commercial demand Ten end–use servicesEleven building typesTen distributed generation technologiesSixty–four end-use technologies


Industrial demand Seven energy–intensive industriesEight non–energy–intensive industriesCogeneration

Four Census regions, shared to nineCensus divisions

Transportation demand Six car sizesSix light truck sizesSixty–three conventional fuel-saving technologies

for light–duty vehiclesGasoline, diesel, and thirteen alternative–fuel

vehicle technologies for light-duty vehiclesTwenty vintages for light-duty vehiclesNarrow and wide–body aircraftSix advanced aircraft technologiesMedium and heavy freight trucksThirty–seven advanced freight truck technologie


Electricity Eleven fossil generation technologiesTwo distributed generation technologiesSeven renewable generation technologiesConventional and advanced nuclearMarginal and average cost pricingGeneration capacity expansionSeven environmental control technologies

Fifteen electricity supply regions (includingAlaska and Hawaii) based on the NorthAmerican Electric Reliability Council regionsand subregions

Nine Census divisions for demand

Renewables Wind, geothermal, solar thermal, solar photovoltaic,landfill gas, biomass, conventional hydropower

Fifteen electricity supply regions

Oil supply OnshoreDeep and shallow offshore

Six lower 48 onshore regionsThree lower 48 offshore regionsThree Alaska regions

Natural gas supply Conventional lower–48 onshoreLower–48 deep and shallow offshoreCoalbed methaneGas shalesTight sandsCanadian, Mexican, and liquefied natural gasAlaskan Gas

Six lower 48 onshore regionsThree lower 48 offshore regionsThree Alaska regionsEight liquefied natural gas import regions

Natural gas transmissionand distribution

Core vs. noncorePeak vs. offpeakPipeline capacity expansion

Twelve lower 48 regionsTen pipeline border points

Refining Five crude oil categoriesFourteen product categoriesMore than 40 distinct technologiesRefinery capacity expansion

Three refinery regions aggregatedfrom Petroleum Administration for DefenseDistricts

Coal supply Three sulfur categoriesFour thermal categoriesUnderground and surface mining typesImports and Exports

Eleven supply regionsSixteen demand regionsSixteen export regionsTwenty import regions

Summary of NEMS Detail

Major Assumptions

Each module of NEMS embodies many assumptionsand data to characterize the future production, con-version, or consumption of energy in the UnitedStates. Two major assumptions concern economicgrowth in the United States and world oil prices, asdetermined by world oil supply and demand.

The Annual Energy Outloook 2003 (AEO2003) in-cludes five primary fully–integrated cases: a refer-ence case, a high and low economic growth case, anda high and low world oil price case. The referencecase uses mid–range assumptions for both the eco-nomic growth rate and the world oil price. The pri-mary determinant for differenct economic growthrates are growth in the labor force and productivity,while different assumptions on oil production in theOrganization of Petroleum Exporting Countries(OPEC) lead to different levels of world oil prices.

In addition to the five primary fully–integratedcases, AEO2003 includes 21 other cases that explorethe impacts of varying key assumptions in the indi-vidual components of NEMS. Many of these cases in-volve changes in the assumptions that impact thepenetration of new or improved technologies, whichis a major uncertainty in formulating projections offuture energy markets. Some of these cases are runas fully integrated cases (e.g., 2003 technology case,high technology case, high renewables, slow andrapid oil and gas technology cases, and low and highcoal mining cost cases). Others exploit the modularstructure of NEMS by running only a portion of theentire modeling system in order to focus on thefirst–order impacts of changes in the assumptions(e.g., advanced nuclear cost case, high electricity de-mand case, low and high fossil electric generatingtechnology cases, and low and high technology casesin the residential, commercial, industrial, and trans-portation sectors).

NEMS Modular Structure

Overall, NEMS represents the behavior of energymarkets and their interactions with the U.S. econ-omy. The model achieves a supply/demand balancein the end-use demand regions, defined as the nineCensus divisions (Figure 1), by solving for the pricesof each energy product that will balance the quanti-ties producers are willing to supply with the quanti-ties consumers wish to consume. The system reflectsmarket economics, industry structure, and existingenergy policies and regulations that influence mar-ket behavior.

NEMS consists of four supply modules (oil and gas,natural gas transmission and distribution, coal, andrenewable fuels); two conversion modules (electricityand petroleum refineries); four end-use demandmodules (residential, commercial, transportation,and industrial); one module to simulate energy/econ-omy interactions (macroeconomic activity); one mod-ule to simulate world oil markets (international en-ergy activity); and one module that provides themechanism to achieve a general market equilibriumamong all the other modules (integrating module).Figure 2 depicts the high-level structure of NEMS.

Because energy markets are heterogeneous, a singlemethodology does not adequately represent all sup-ply, conversion, and end-use demand sectors. Themodularity of the NEMS design provides the flexi-bility for each component of the U.S. energy systemto use the methodology and coverage that is most ap-propriate. Furthermore, modularity provides the ca-pability to execute the modules individually or in col-lections of modules, which facilitates the develop-ment and analysis of the separate component mod-ules. The interactions among these modules are con-trolled by the integrating module.

The NEMS global data structure is used to coordi-nate and communicate the flow of informationamong the modules. These data are passed throughcommon interfaces via the integrating module. Theglobal data structure includes energy market pricesand consumption; macroeconomic variables; energyproduction, transportation, and conversion informa-tion; and centralized model control variables, param-eters, and assumptions. The global data structureexcludes variables that are defined locally within themodules and are not communicated to other mod-ules.

A key subset of the variables in the global data struc-ture is the end-use prices and quantities of fuelswhich are used to equilibrate the NEMS energy bal-ance in the convergence algorithm. These deliveredprices of energy and the quantities demanded are de-fined by product, region, and sector. The deliveredprices of fuel encompass all the activities necessaryto produce, import, and transport fuels to the enduser. The regions for the price and quantity variablesin the global data structure are the nine Census divi-sions. The four Census regions (shown in Figure 1 bybreaks between State groups) and nine Census divi-sions are a common, mainstream level of regionalitywidely used by EIA and other organizations for datacollection and analysis.


OVERVIEW OF NEMS


OVERVIEW OF NEMS

MEVT

NH

MA

RICT

CA

WA

OR

NM

WY

MT

ID

UTNV

AZ

MIWI

IL IN OH

MN

IA

ND

SD

NE

KS MO

LATX

OKAR

KY

ALMS

TN

DE

MDVAWV

NC

SC

GA

FL

NY

PA NJ

HI

Pacific

Mountain

West North

Central East North

Central

Middle

Atlantic

New

England

West South Central

East South

Central

South Atlantic

CO

Figure 1. Census Divisions

Division 1 Division 3 Division 5 Division 7 Division 9

New England East North Central South Atlantic West South Central Pacific

Connecticut Illinois Delaware Arkansas Alaska

Maine Indiana District of Columbia Louisiana California

Massachusetts Michigan Florida Oklahoma Hawaii

New Hampshire Ohio Georgia Texas Oregon

Rhode Island Wisconsin Maryland Washington

Vermont North Carolina Division 8

Division 4 South Carolina Mountain

Division 2 West North Central Virginia Arizona

Middle Atlantic Iowa West Virginia Colorado

New Jersey Kansas Idaho

New York Minnesota Division 6 Montana

Pennsylvania Missouri East South Central Nevada

Nebraska Alabama New Mexico

North Dakota Kentucky Utah

South Dakota Mississippi Wyoming

Tennessee

Integrating Module

The NEMS integrating module controls the entireNEMS solution process as it iterates to determine ageneral market equilibrium across all the NEMSmodules. It has the following functions:

• Manages the NEMS global data structure

• Executes all or any of the user-selectedmodules in an iterative convergence algorithm

• Checks for convergence and reports variablesthat remain out of convergence

• Implements convergence relaxation onselected variables between iterations toaccelerate convergence

• Updates expected values of the key NEMSvariables.

The integrating module executes the demand, con-version, and supply modules iteratively until itachieves an economic equilibrium of supply and de-mand in all the consuming and producing sectors.Each module is called in sequence and solved, as-suming that all other variables in the energy mar-kets are fixed. The modules are called iteratively un-til the end-use prices and quantities remain constant

within a specified tolerance, a condition defined asconvergence. Equilibration is achieved annuallythroughout the midterm period, currently 2025, foreach of the nine Census divisions.

In addition, the macroeconomic and internationalmodules are executed iteratively to incorporate thefeedback on the economy and international marketsfrom changes in the domestic energy markets.Convergence tests check the stability of a set of keymacroeconomic and international trade variables inresponse to interactions with the domestic energysystem.

The NEMS algorithm executes the system of mod-ules until convergence is reached. The solution pro-cedure for one iteration involves the execution of allthe component modules, as well as the updating ofexpectation variables (related to foresight assump-tions) for use in the next iteration. The system is exe-cuted sequentially for each year in the forecast pe-riod. During each iteration within a year, each of themodules is executed in turn, with intervening con-vergence checks that isolate specific modules thatare not converging. A convergence check is made foreach price and quantity variable to see whether thepercentage change in the variable is within the as-sumed tolerance. To avoid unnecessary iterations forchanges in insignificant values, the quantity conver-


OVERVIEW OF NEMS

Oil and GasSupply Module

Natural GasTransmission

and DistributionModule

Coal MarketModule

RenewableFuels Module

Supply

ResidentialDemand Module

CommercialDemandModule

TransportationDemand Module

IndustrialDemand Module

Demand

ElectricityMarketModule

PetroleumMarketModule

Conversion

MacroeconomicActivityModule

InternationalEnergyModule

Integrating

Module

Figure 2. National Energy Modeling System

gence check is omitted for quantities less than auser-specified minimum level. The order of executionof the modules may affect the rate of convergence butwill generally not prevent convergence to an equilib-rium solution or significantly alter the results. Anoptional relaxation routine can be executed todampen swings in solution values between itera-tions. With this option, the current iteration valuesare reset partway between solution values from thecurrent and previous iterations.

Because of the modular structure of NEMS and theiterative solution algorithm, any single module orsubset of modules can be executed independently.Modules not executed are bypassed in the calling se-quence, and the values they would calculate and pro-vide to the other modules are held fixed at the valuesin the global data structure, which are the solutionvalues from a previous run of NEMS. This flexibilityis an aid to independent development, debugging,and analysis.


OVERVIEW OF NEMS

The emissions policy submodule, part of the inte-grating module, estimates the energy–related emis-sions of carbon dioxide and methane. Carbon dioxideemissions are dependent on the fossil fuel consumed,the carbon content of the fuel, and the fraction of thefuel consumed in combustion. The product of thecarbon dioxide coefficient and the combustion frac-tion yields a carbon dioxide emission factor. For fueluses of fossil energy, the combustion fractions are as-sumed to be 0.99 for liquid fuels and 0.995 for gas-eous fuels. The carbon dioxide potential of nonfueluses of energy, such as asphalt and petrochemicalfeedstocks, is assumed to be sequestered in the prod-uct and not released to the atmosphere. The coeffi-cients for carbon dioxide emissions are updated eachyear from the Energy Information Administration’sannual, Emissions of Greenhouse Gases in the UnitedStates.17

The energy-related methane emissions are esti-mated as a function of energy production and con-sumption drivers. Methane emissions occur in vari-ous phases of the production and transportation ofcoal, oil, and natural gas. Additional emissions occuras a result of incomplete combustion of fossil fuelsand wood. The methane equations in NEMS are de-rived from methodologies and data sources in En-ergy Information Administration’s annual Emis-sions of Greenhouse Gases in the United States.18

The emissions policy submodule also allows for sev-eral carbon dioxide policy evaluation options to beanalyzed within NEMS. Although these policy op-tions are not assumed in the Annual Energy Out-look, the options have been used in special analysesto simulate potential market–based approaches tomeet national carbon dioxide emission objectives.The policy options implemented are as follows:

• Carbon Dioxide Tax. A tax on carbon dioxideemissions from fossil fuels is added to raisedelivered fossil fuel prices. The resultinghigher prices then induce changes in fossil fueluse and carbon dioxide emissions, as well as

changes in some long–term decision making,such as generating capacity decisions in theelectricity market module.

• Auction of Permits. This option simulates anauction on carbon dioxide emissions permitsto meet an overall cap on emissions. A carbondioxide permit price is computed that clearsthe auction market. The permit fee is treatedas a carbon dioxide tax and used as anadjustment to the fossil fuel prices. A newprice is set each NEMS iteration until theemissions reach the goal. The revenuegenerated from the auction is calculatedassuming there is no initial allocation ofemission permits.

• Market for Permits (Cap and Trade). A marketfor tradable carbon dioxide emissions permitsis simulated assuming that an initial distribu-tion of marketable permits to emission sourcestakes place. The permits are transferable butare not banked between years. As with thecarbon dioxide tax and auction options, thefull market price of the permits is added to theenergy prices. The system of marketablepermits is implemented in the same way asthe permit auction, with the exception of thecalculation of revenues from permit sales.Similar treatment is warranted because themarginal cost of a free permit is equivalent toone purchased at auction, given theopportunity cost of holding the distributedpermit.19

The options above can be implemented for all con-suming sectors of NEMS or for the electricity gener-ating sector alone. The use of any of these emissionspolicy options in NEMS requires a macroeconomicanalysis to assess the fiscal and monetary issues, aswell as the possible international trade effects. Theanalysis depends on such factors as how revenuesgenerated from the policy would be used, how mone-tary authorities would react to the fiscal policy


CARBON DIOXIDE AND METHANE EMISSIONS

17 Energy Information Administration, Emissions of Greenhouse Gases in the United States 2001, DOE/EIA-0573(2001)(Washington, DC, December 2002).

18 Energy Information Administration, Emissions of Greenhouse Gases in the United States 2001, DOE/EIA-0573(2001)(Washington, DC, December 2002).

19 In an open, competitive permit market, the permit will tend to be priced at the marginal cost of reducing carbon dioxideemissions, regardless of the initial distribution of permits. If permits are purchased by suppliers and passed through to the fuelprice, the marginal cost of the carbon dioxide emissions by a particular sector in a region will be reflected in the individualend–use fuel cost for tht sector.

changes, and how international agreements to re-duce carbon dioxide emissions would be imple-mented.

A limitation of these policy options is that they ad-dress energy–related carbon dioxide emissions only.

Work on NEMS is underway on a capability to esti-mate reductions of other greenhouse gas emissions.This capability, drawing on marginal cost of abate-ment curves for such gases, will enahle the economicanalysis of policies targeted at capping total green-house gases in an internally consistent framework.


CARBON DIOXIDE AND METHANE EMISSIONS

The Macroeconomic Activity Module (MAM) linksNEMS to the rest of the economy by providing projec-tions of economic driver variables for use by the sup-ply, demand, and conversion modules of NEMS. Thederivation of the baseline macroeconomic forecastlays a foundation for the determination of the energydemand and supply forecast. MAM is used to presentalternative macroeconomic growth cases to provide arange of uncertainty about the growth potential forthe economy and its likely consequences for the en-ergy system. MAM is also able to address the macro-economic impacts associated with changing energymarket conditions, such as alternative world oilprice assumptions. Outside of the Annual EnergyOutlook setting, MAM represents a system of linkedmodules which can assess the potential impacts onthe economy of changes in energy events or policyproposals. These economic impacts then feed backinto NEMS for an integrated solution. MAM consistsof five modules:

• Global Insight Model of the U.S. Economy

• Global Insight Industry Model

• Global Insight Employment Model

• Global Insight Regional Model

• Energy Information Administration (EIA)Commercial Floorspace Model

The Global Insight Model of the U.S. Economy (Mac-roeconomic Model) is the same model used by GlobalInsight, Inc. (formerly DRI-WEFA) to generate theeconomic forecasts behind the company’s monthlyassessment of the U.S. economy. The Industry, Em-ployment, and Regional Models are derivatives ofGlobal Insight’s industry, employment, and regionalmodels. The models have been tailored in order toprovide the industry and regional detail required byNEMS. The Commercial Floorspace Model was de-veloped by EIA to complement the set of Global In-

sight models. This system of models provides a fullyintegrated approach to forecasting economic activityat the national, industry and regional levels. The setof models is designed to run in a recursive manner(see Figure 3).

Global Insight’s Macroeconomic Model determinesthe national economy’s growth path and final de-mand mix. The Macroeconomic Model provides fore-casts of over 1300 concepts spanning final demands,aggregate supply, prices, incomes, internationaltrade, industrial detail, interest rates and financialflows.

The Industry Model takes the final demand projec-tions from the Macroeconomic Model as inputs toprovide projections of output and other key indica-tors for 130 sectors, covering the entire economy.This is later aggregated to 45 sectors to provide infor-mation to NEMS. The Industry Model insures thatsupply by industry is consistent with the final de-mands (consumption, investment, governmentspending, exports and imports) generated in theMacroeconomic Model.

The Employment Model takes the industry outputprojections from the Industry Model and nationalwage rates, productivity trends and average work-week trends from the Macroeconomic Model to pro-ject employment for the 45 NEMS industries. Thesum of non–agricultural employment is constrainedto sum to the national total projected by the Macro-economic Model.

The Regional Model determines the level of industryoutput and employment, population, incomes, andhousing activity in each of nine Census regions. TheCommercial Floorspace Model calculates regionalfloorspace for 13 types of building use by Census Di-vision.

Integrated forecasts of NEMS center around esti-mating the state of the energy-economy system un-der a set of alternative energy conditions: Typically,


MACROECONOMIC ACTIVITY MODULE

MAM Outputs Inputs from NEMS Exogenous Inputs

Gross domestic product

Other economic activity measures, including housing

starts, commercial floorspace growth, vehicle sales,

population

Price indices and deflators

Production and employment for manufacturing

Production and employment for nonmanufacturing

Interest rates

Petroleum, natural gas, coal, and

electricity prices

Oil, natural gas, and coal production

Electric and gas industry output

Refinery output

End-use energy consumption by fuel

Macroeconomic variables defining

alternative economic growth cases

the forecasts fall into the following four types of inte-grated NEMS simulations:

• Baseline Projection

• Alternative World Oil Prices

• Proposed Energy Fees or Emissions Permits

• Proposed Changes in Combined Average FuelEconomy (CAFÉ) Standards

In these integrated NEMS simulations, forecast pe-riod baseline values for over 240 macroeconomic anddemographic variables from MAM are passed toNEMS which solves for demand, supply and prices ofenergy for the forecast period. These energy pricesand quantities are passed back to MAM and solvedin the Macroeconomic Model, the Industry Modeland the Employment Model in the EViews environ-ment.20 The Regional Model and the CommercialFloorspace Model and NEMS are run in theFORTRAN environment.


MACROECONOMIC ACTIVITY MODULE

National EmploymentVariables

National InterindustryVariables

Oil, Natural Gas and CoalProduction; Refinery

Activity; Electric and GasIndustry Output

Energy Prices and End-useConsumption

National MacroeconomicVariables

Macroeconomic Activity Module

NationalMacroeconomic

Variables & IndustrialShipments

Macroeconomic

Submodule

IndustrySubmodule

Employment

Submodule

Regional

Submodule

MacroeconomicGrowth Cases

Exogenous

NEMS

Commercial FloorspaceSubmodule

Regional MacroeconomicIndustry Employment

Floorspace Variablesand Commercial

Figure 3. Macroeconomic Activity Module Structure

20 Eviews is a model building and operating software package maintained by QMS (Quantitative Micro Software.)

The international energy module (IEM) consists offour submodules (Figure 4) that perform thefollowing functions:

• world oil market submodule calculates theaverage annual world oil price (importedrefiner acquisition cost) that is consistent withworldwide petroleum demand and supplyavailability

• crude oil supply submodule provides im-ported crude oil supply curves for five crude oilquality classes

• petroleum products supply submodule pro-vides imported refined product supply curvesfor twelve types of refined products

• oxygenates supply submodule providesimported oxygenates supply curves for methyltertiary butyl ether (MTBE) and methanol.

The world oil price that is generated by the world oilmarket submodule is used by all the modules ofNEMS as well as the other submodules of IEM. Theimport supply curves for crude oils, refined products,and oxygenates are used by the petroleum marketmodule.

World Oil Market Submodule

In NEMS, the U.S. oil market is modeled in consider-able detail, while foreign markets use a less detailedapproach. EIAs modeling of the near–to mid–termworld oil market depends on two key assumptions:(1) oil is the marginal fuel and (2) the Organization ofPetroleum Exporting Countries (OPEC) is the mar-ginal supplier of oil. The first assumption impliesthat competition between oil and other fuels is notsignificant enough to impact the world oil price. In

addition, prices remain sufficiently low such that themarket penetration of new technologies that wouldreduce the demand for oil is inhibited. In the secondassumption, OPEC producers are assumed to ex-pand oil production capacity in order to meet thegrowth in worldwide oil demand.

The various price cases examined by EIA differ inthe magnitude to which OPEC producers expandtheir production capacity. Lower prices imply consid-erable capacity expansion activity with a probableassist from foreign investment interests. Higherprices imply an unwillingness on the part of OPECproducers to invite foreign investment participation.The world oil market submodule forecasts the worldoil price and produces a regional world oil marketsupply/demand balance that is consistent with theforecasted price. The world oil price forecast is basedupon a regression analysis of the price in the previ-ous time period and the percent utilization of OPECproduction capacity. IEM has either the capabilityto forecast world oil prices given OPEC productioncapacity estimates or the capability to forecastOPEC production given an exogenous world oil pricepath.

Crude Oil Supply Submodule

The crude oil supply submodule consists of a set ofimport supply curves to all five Petroleum Adminis-tration for Defense Districts (PADDs) for each of fivequality classes of crude oils and for each simulationyear. The petroleum market module uses the supplycurves to determine the quantities and prices of thecrude oils to be imported. Because the petroleummarket module is a linear programming formula-tion, the imported crude oil supply curves are formu-lated as 3–step, piecewise–linear functions. The fiveclasses of imported crude oils categorized by sulfurcontent and American Petroleum Institute (API)

INTERNATIONAL ENERGY MODULE


IEM Outputs Inputs from NEMS Exogenous Inputs

World oil price

Crude oil import supply curves

Refined product import supply curves

Oxygenate import supply curves

Domestic crude oil production

Domestic natural gas liquids production

Domestic gas-to-liquids production

Domestic coal-to-liquids production

Domestic other liquids production

Domestic refinery gain

Domestic product supplied

GDP price deflators

Domestic crude oil imports

Domestic refined product imports

Domestic oxygenate imports

Domestic unfinished oils imports

OPEC production capacity path

Reference non-U.S. oil supply and demand

Non-U.S. economic parameters

Base import supply curves for crude oils, refined

products, and oxygenates

gravity include: low–sulfur light, medium–sulfurheavy, high–sulfur light, high–sulfur heavy,high–sulfur very heavy.

The imported crude oil supply curves are developedexogenous to NEMS using a large–scale linear pro-gramming formulation of international refining andtransportation. This formulation, known as the

World Oil Refining, Logistics, and Demand(WORLD) model, is run repetitively, parameterizingon the import levels of the five crude oil classes intoeach PADD. From these runs, base price/quantity re-lationships for imported crude oils are established.Within NEMS, these base relationships are shiftedas a function of the world oil price and presented tothe petroleum market module as a flexible set of



World Oil Price

PetroleumMarket Module


InternationalEnergy Module

NEMS

Exogenous

OPEC ProductionCapacity Path, Non-U.S.Oil Supply and Demand,Non-U.S. EconomicParameters

Exogenous

Exogenous

World Oil Price

OxygenatesSupply

Submodule

PetroleumProductsSupply

Submodule

World OilMarket

Submodule

Crude OilSupply

Submodule

World Oil Price

ProductSupply Curves

OxygenateSupply Curves

Crude Oil, NaturalGas Liquids, Gas-to-Liquids,

Coal-to-Liquids,Other Liquids Produc-

tion, Refinery Gain,Unfinished Oils Imports

Product Supplied

Crude Oil Imports

Crude Oil ImportSupply Curves

Product Imports

Product ImportSupply Curves

GDP Deflator

Exogenous

Crude OilSupply Curves

World Oil Price

World Oil Price

OxygenateImports

OxygenateImport Supply

Curves

Figure 4. International Energy Module Structure

crude oil import alternatives. By observing whichimport supply curves are selected by the petroleummarket module, it becomes possible to map these se-lections back into the WORLD model in order to pro-vide estimates of future sources of crude oil importsto the United States.

Petroleum Products SupplySubmodule

The petroleum products supply submodule consistsof a set of import supply curves to all five PADDs foreach of twelve refined product types and for eachsimulation year. The petroleum market module usesthe supply curves to determine the quantities andprices of refined products to be imported. Becausethe petroleum market module is a linear program-ming formulation, the imported refined product sup-ply curves are formulated as 3–step, piecewise–lin-ear functions. The twelve types of imported refinedproducts include: traditional gasoline (including avi-ation), reformulated gasoline, reformulated gasolineblending stocks for oxygenated blending (RBOB),traditional distillate fuel, low-sulfur No. 2 heatingoil, low–sulfur diesel fuel, high– and low–sulfur re-sidual fuel, jet fuel (including naphtha jet), liquefiedpetroleum gases, petrochemical feedstocks, andother petroleum products.

Similar to the imported crude oil supply curves, theimported refined product supply curves are also de-veloped exogenous to NEMS using the WORLDmodel. By observing which import supply curves areselected by the petroleum market module, it becomespossible to map these selections back into theWORLD model in order to provide estimates of fu-

ture sources of refined product imports to the UnitedStates.

Oxygenates Supply Submodule

The oxygenates supply submodule consists of a set ofimport supply curves to all five PADDs for the oxy-genates MTBE and methanol and for each simula-tion year. The petroleum market module uses thesupply curves to determine the quantities and pricesof oxygenates to be imported. Because the petroleummarket module is a linear programming formula-tion, the imported oxygenate supply curves are for-mulated as 3–step, piecewise–linear functions. Simi-lar to the imported crude oil supply curves, the im-ported oxygenate supply curves are developed exoge-nous to NEMS using the WORLD model. By observ-ing which import supply curves are selected by thepetroleum market module, it becomes possible tomap these selections back into the WORLD model inorder to provide estimates of future sources of oxy-genate imports into the United States.

Because of the potential expansion of the U.S. etha-nol industry and the lack of commercial markets forother oxygenates, it is assumed that ethanol, ethyltertiary butyl ether (ETBE), tertiary amyl methylether (TAME), and tertiary butyl alcohol (TBA) areall supplied from domestic sources. Therefore, IEMdoes not provide import supply curves for these oxy-genates.

By presenting NEMS with a flexible array of importchoices, valuable insights can be gained on such is-sues as the future crude oil/refined product importcomposition, potential U.S. refinery expansion (bothdistillation capacity and downstream capacity), andfuture sources of petroleum imports (including Per-sian Gulf import dependence).



The residential demand module (RDM) forecasts en-ergy consumption by Census division for seven mar-keted energy sources plus solar and geothermal en-ergy. RDM is a structural model and its forecasts arebuilt up from projections of the residential housingstock and of the energy-consuming equipment con-tained therein. The components of RDM and its in-teractions with the NEMS system are shown in Fig-ure 5. NEMS provides forecasts of residential energyprices, population, disposable income, and housingstarts, which are used by RDM to develop forecasts ofenergy consumption by end–use service, fuel type,and Census division.

RDM incorporates the effects of four broadly-defineddeterminants of energy consumption: economic anddemographic effects, structural effects, technologyturnover and advancement effects, and energy mar-ket effects. Economic and demographic effects in-clude the number, dwelling type (single-family,multifamily or mobile homes), occupants per house-hold, disposable income, and location of housingunits. Structural effects include increasing averagedwelling size and changes in the mix of desiredend-use services provided by energy (new end usesand/or increasing penetration of current end uses,such as the increasing popularity of electronic equip-ment and computers). Technology effects includechanges in the stock of installed equipment causedby normal turnover of old, worn out equipment withnewer versions which tend to be more energy effi-cient, the integrated effects of equipment and build-ing shell (insulation level) in new construction, andin the projected availability of even more en-ergy-efficient equipment in the future. Energy mar-ket effects include the short-run effects of energyprices on energy demands, the longer-run effects ofenergy prices on the efficiency of purchased equip-ment and the efficiency of building shells, and limita-tions on minimum levels of efficiency imposed by leg-islated efficiency standards.

Housing Stock Submodule

The base housing stock by Census division anddwelling type is derived from EIA’s 1997 ResidentialEnergy Consumption Survey (RECS). Each element

of the base stock is retired on the basis of a constantrate of decay for each dwelling type. RDM receivesas an input from the macroeconomic activity moduleforecasts of housing additions by type and Census di-vision. RDM supplements the surviving stocks fromthe previous year with the forecast additions bydwelling type and Census division. The averagesquare footage of new construction is based on recentupward trends developed from the RECS and theCensus Bureau’s Characteristics of New Housing.

Appliance Stock Submodule

The installed stock of appliances is also taken fromthe 1997 RECS. The efficiency of the appliance stockis deived from historical shipments by efficiencylevel over a multi–year interval for the followingequipment: heat pumps, gas furnaces, central airconditioners, room air conditioners, water heaters,refrigerators, freezers, stoves, dishwashers, clotheswashers, and clothes dryers. A linear retirementfunction with both minimum and maximum equip-ment lives is used to retire equipment in survivinghousing units. For equipment where shipment dataare available, the efficiency of the retiring equip-ment varies over the projection. In early years, theretiring efficiency tends to be lower as the older, lessefficient equipment in the stock turns over first.Also, as housing units retire, the associated appli-ances are removed from the base appliance stock aswell. Additions to the base stock are tracked sepa-rately for housing units existing in 1997 and for cu-mulative new construction.

As appliances are removed from the stock, they arereplaced by new appliances with generally higher ef-ficiencies due to technology improvements, equip-ment standards, and market forces. Appliancesadded due to new construction are accumulated andretired parallel to appliances in the existing stock.Appliance stocks are maintained by fuel, end use,and technology as shown in the above table.

Technology Choice Submodule

Fuel-specific equipment choices are made for bothnew construction and replacement purchases. Fornew construction, initial heating system shares

RESIDENTIAL DEMAND MODULE


RDM Outputs Inputs from NEMS Exogenous Inputs

Energy demand by service and fuel type

Changes in housing and appliance stocks

Appliance stock efficiency

Energy product prices

Housing starts

Population

Current housing stocks and retirement rates

Current appliance stocks and life expectancy

New appliance types, efficiencies, and costs

Housing shell retrofit indices

Unit energy consumption

Square footage



NEMS

MacroeconomicActivity Module



Housing Starts,Population

TechnologyChoice

Submodule

FuelConsumptionSubmodule

Residential Demand Module

HousingStock

Submodule

Electricity Prices



CoalMarketModule

Electricity Demand

Natural Gas Prices

Natural Gas Demand

Petroleum ProductPrices

Petroleum Demand

Coal Prices

Coal Demand

ShellIntegrity

Submodule

Stock of Structures by Type and Vintage

ApplianceStock

Submodule

Building Shell Efficiencies

Surviving Stock of Appliances

Average Efficiency of Appliance Stock

Exogenous

Base Year Housing Stock andRetirement Rates, Appliance

Stocks and Life Expectancies,New Appliance Types, Efficiencies

and Costs, Housing ShellRetrofit Indices, Unit Energy

Consumption, Square Footage

DistributedGenerationSubmodule

Figure 5. Residential Demand Module Structure

(taken from the most recently available Census Bu-reau survey data covering new construction, cur-rently 2001) are adjusted based on relative life cyclecosts for all competing technology and fuel combina-tions. Once new home heating system shares are es-tablished, the fuel choices for other services, such aswater heating and cooking, are determined based onthe fuel chosen for space heating. For replacementpurchases, fuel switching is allowed for an assumedpercentage of all replacements but is dependent onthe estimated costs of fuel–switching (for example,switching from electric to gas heating is assumed toinvolve the costs of running a new gas line).

For both replacement equipment and new construc-tion, a “second-stage” of the equipment choice deci-sion requires selecting from several projected avail-able efficiency levels. The projected efficiency rangeof available equipment represents a “menu” of effi-ciency levels and installed cost combinations pro-jected to be available at the time the choice is beingmade. Costs and efficiencies for selected appliancesare shown in the table on page 25, derived from thereport Assumptions to the Annual Energy Outlook

2003.21 At the low end of the efficiency range are theminimum levels required by legislated standards. Inany given year, higher efficiency levels are associ-ated with higher installed costs. Thus, purchasinghigher than the minimum efficiency involves atrade-off between higher installation costs and fu-ture savings in energy expenditures. In RDM,these trade-offs are calibrated to recent shipment,cost, and efficiency data. Changes in projected pur-chases by efficiency level are based on changes in ei-ther the installed capital costs or changes in thefirst-year operating costs across the available effi-ciency levels. As energy prices increase, the incen-tive of greater energy expenditures savings will pro-mote increased purchases of higher-efficiency equip-ment. In some cases, due to government programs orgeneral projections of technology improvements, pro-jected increases in efficiency or decreases in the in-stalled costs of higher-efficiency equipment will alsopromote purchases of higher-efficiency equipment.

Shell Integrity Submodule

Shell integrity is also tracked separately for the ex-isting housing stock and new construction. Shell in-tegrity for existing construction is assumed to re-spond to increases in real energy prices by becomingmore efficient. There is no change in existing shellintegrity when real energy prices decline. New shellefficiencies are based on the cost and performance ofthe heating and cooling equipment as well as theshell dharacteristics. Several efficiency levels ofshell characteristics are available throughout theprojection period and can change over time based onchanges to building codes. All shell efficiencies aresubject to a maximum shell efficiency based on stud-ies of currently available residential constructionmethods.

Distributed Generation Submodule

Distributed generation equipment with explicit tech-nology characterizations is also modeled for residen-tial customers. Currently, two technologies arecharacterized, photovoltaics and fuel cells. Thesubmodule incorporates historical estimates ofphotovoltaics (residential-sized fuel cells are not ex-pected to be commercialized until after 2001) fromits technology characterization and exogenous pene-tration input file. Program-based photovoltaic esti-mates for the Department of Energy’s Million Solar



Space Heating Equipment: electric furnace, electric air-source

heat pump, natural gas furnace, natural gas hydronic,

kerosene furnace, liquefied petroleum gas, distillate furnace,

distillate hydronic, wood stove, ground-source heat pump,

natural gas heat pump.

Space Cooling Equipment: room air conditioner, central air

conditioner, electric air-source heat pump, ground-source heat

pump, natural gas heat pump.

Water Heaters: solar, natural gas, electric, distillate, liquefied

petroleum gas.

Refrigerators: 18 cubic foot top-mounted freezer, 25 cubic foot

side-by-side with through-the-door features.

Freezers: chest - manual defrost, upright - manual defrost

Lighting: incandescent, compact fluorescent, mercury vapor

Clothes Dryers: natural gas, electric

Cooking: natural gas, electric, liquefied petroleum gas.

Dishwashers

Clothes Washers

Fuel Cells

Solar Photovoltaic

NEMS Residential Module Equipment Summary

21 Energy Information Administration, Assumptions to the Annual Energy Outlook 2003,http://www.eia.doe.gov/oiaf/aeo/assumption/pdf/0554(2003).pdf (Washington, DC, January 2003).

Roofs program are also input to the submodule fromthe exogenous penetration portion of the input file.Endogenous, economic purchases are based on a pen-etration function driven by a cash flow model whichsimulates the costs and benefits of distributed gener-ation purchases. The cash flow calculations are de-veloped from NEMS projected energy prices coupledwith the technology characterizations provided fromthe input file.

Potential economic purchases are modeled byCensus division and technology for all years subse-quent to the base year. The cash flow model developsa 30-year cost-benefit horizon for each potential in-vestment. It includes considerations of annual costs(down payments, loan payments, maintenance costsand, for fuel cells, gas costs) and annual benefits (in-terest tax deductions, any applicable tax credits,electricity cost savings, and water heating savingsfor fuel cells) over the entire 30-year period. Pene-tration for a potential investment in eitherphotovoltaics or fuel cells is a function of whether itachieves a cumulative positive cash flow, and if so,how many years it takes to achieve it.

Once the cumulative stock of distributed equipmentis projected, reduced residential purchases of elec-tricity are provided to NEMS. For fuel cells, in-creased residential natural gas consumption is also

provided to NEMS based on the calculated energy in-put requirements of the fuel cells, partially offset bynatural gas water heating savings from the use ofwaste heat from the fuel cell.

Fuel Consumption Submodule

The fuel consumption submodule modifies base yearenergy consumption intensities in each forecastyear. Base year energy consumption for each end useis derived from energy intensity estimates from the1997 RECS. The base year energy intensities aremodified for the following effects: (1) increases in ef-ficiency, based on a comparison of the projected ap-pliance stock serving this end use relative to the baseyear stock, (2) changes in shell integrity for spaceheating and cooling end uses, (3) changes in real fuelprices—(short-run price elasticity effects), (4)changes in square footage, (5) changes in the numberof occupants per household, (6) changes in disposableincome, (7) changes in weather relative to the baseyear, (8) adjustments in utilization rates caused byefficiency increases (efficiency “rebound” effects),and (9) reductions in purchased electricity and in-creases in natural gas consumption from distributedgeneration. Once these modifications are made, totalenergy use is computed across end uses and housingtypes and then summed by fuel for each Census divi-sion.





Equipment TypeRelative

Performance1

2001Installed Cost(2001 dollars)2 Efficiency3

2015Installed Cost(2001 dollars)2 Efficiency3

ApproximateHurdle4

Rate

Electric Heat Pump MinimumBest

$2,930$5,600

10.018.0

$3,500$5,600

12.018.0

15%

Natural Gas Furnace MinimumBest

$1,300$2,700

0.800.97

$1,300$1,950

0.800.97

15%

Room Air Conditioner MinimumBest

$540$760

8.711.7

$540$760

9.712.0

140%

Central Air Conditioner MinimumBest

$2,080$3,500

10.018.0

$2,300$3,500

12.018.0

25%

Refrigerator (18 cubic ft) MinimumBest

$600$950

690515

$600$950

478400

19%

Electric Water Heater MinimumBest

$337$1,200

0.862.60

$500$1,100

0.902.6

83%

Solar Water Heater N/A $3,200 2.0 $2,533 2.0 83%

Characteristics of Selected Equipment

1Minimum performance refers to the lowest efficiency equipment available. Best refers to the highest effi-ciency equipment available.2Installed costs represents the capital cost of the equipment plus the cost to install it, excluding any financecosts.3Efficiency measurements vary by equipment type. Electric heat pumps and central air conditioners arerated for cooling performance using the Seasonal Energy Efficiency Ratio (SEER); natural gas furnaces arebased on Annual Fuel Utilization Efficiency; room air conditioners are based on Energy Efficiency Ratio(EER); refrigerators are based on kilowatt-hours per year; and water heaters are based on Energy Factor(delivered Btu divided by input Btu).4The hurdle rate represents the consumer’s “willingness” to invest in energy efficiency is by weighing thefirst cost and operating cost of competing technologies. The higher the hurdle rate, the less likely a con-sumer will invest in energy efficiency. These rates include all financial and non–financial factors (such assize, color) that influence a consumer’s purchase decision.

Source: Arthur D. Little, EIA Technology Forecast Updates, Reference Number 8675309, October 2001.

The commercial demand module (CDM) forecastsenergy consumption by Census division for eightmarketed energy sources plus solar and geothermalenergy. For the three major commercial sector fuels,electricity, natural gas and distillate oil, CDM is astructural model and its forecasts are built up fromprojections of the stock of commercial floorspace andenergy–consuming equipment. For the remainingfive marketed minor fuels, simple econometricprojections are made.

The commercial sector encompasses business estab-lishments that are not engaged in industrial ortransportation activities. Commercial sector energyis consumed mainly in buildings, except for a rela-tively small amount for services such as street lightsand water supply. CDM incorporates the effects offour broadly–defined determinants of energy con-sumption: economic and demographics, structural,technology turnover and change, and energy mar-kets. Demographic effects include total floorspace,building type and location. Structural effects includechanges in the mix of desired end–use services pro-vided by energy (such as the penetration of telecom-munications equipment, personal computers andother office equipment). Technology effects includechanges in the stock of installed equipment causedby the normal turnover of old, worn out equipment tonewer versions which tend to be more energy effi-cient, the integrated effects of equipment and build-ing shell (insulation level) in new construction, andthe projected availability of equipment with evengreater energy–efficiency. Energy market effects in-clude the short–run effects of energy prices on en-ergy demands, the longer–run effects of energyprices on the efficiency of purchased equipment, andlimitations on minimum levels of efficiency imposedby legislated efficiency standards. The model struc-ture carries out a sequence of five basic steps, asshown in Figure 6. The first step is to forecast com-mercial sector floorspace. The second step is to fore-cast the energy services (space heating, lighting,etc.) required by the projected floorspace. The thirdstep is to project the electricity generation and waterand space heating supplied by distributed genera-

tion and combined heat and power (CHP) technolo-gies. The fourth step is to select specific technologies(natural gas furnaces, fluorescent lights, etc.) tomeet the demand for energy services. The last step isto determine how much energy will be consumed bythe equipment chosen to meet the demand for energyservices.

Floorspace Submodule

The base stock of commercial floorspace by Censusdivision and building type is derived from EIA’s 1999Commercial Buildings Energy Consumption Survey(CBECS). CDM receives forecasts of total floorspaceby building type and Census division from the mac-roeconomic activity module (MAM) based on GlobalInsight, Inc. (formerly DRI–WEFA) definitions of thecommercial sector. These forecasts embody both eco-nomic and demographic effects on commercialfloorspace. Since the definition of commercialfloorspace from Global Insight, Inc. is not calibratedto CBECS, CDM estimates the surviving floorspacefrom the previous year and then calibrates its newconstruction so that growth in total floorspacematches that from MAM by building type and Cen-sus division.

CDM models commercial floorspace for the following11 building types:

• Assembly• Education• Food sales• Food service• Health care• Lodging• Office–large• Office–small• Mercantile and service• Warehouse• Other

COMMERCIAL DEMAND MODULE


CDM Outputs Inputs from NEMS Exogenous Inputs


Changes in floorspace and appliance stocks


Interest rates

Floorspace growth

Existing commercial floorspace

Floorspace survival rates

Appliance stocks and survival rates

New appliance types, efficiencies, costs

Energy use intensities



NEMS




Floorspace Additions,Interest Rates

EquipmentChoice

Submodule

Commercial Demand Module

FloorspaceSubmodule

Electricity Prices



CoalMarketModule

Electricity Demand

Natural Gas Prices

Natural Gas Demand


Petroleum Demand

Coal Prices

Coal Demand

EnergyConsumptionSubmodule

Stock of Commercial Floorspace

EnergyServiceDemand

Submodule

End-Use Service Demands

Stocks of Energy-Consuming Equipment

Exogenous

Existing Commercial Floorspaceand Survival Rates,

Appliance Stocks and Survival Rates,New Appliance Types,

Efficiencies, Costs,Energy Use Intensities

DistributedGeneration/

CHPSubmodule

Figure 6. Commercial Demand Module Structure

Energy Service Demand Submodule

Energy consumption is derived from the demand forenergy services. So the next step is to forecast energyservice demands for the projected floorspace. CDMmodels service demands for the following tenend–use services:

• Heating• Cooling• Ventilation• Water heating• Lighting• Cooking• Refrigeration• Office equipment personal computer

(PC)• Office equipment other• Other end uses.

Different building types require unique combina-tions of energy services. A hospital must have morelight than a warehouse. An office building in theNortheast requires more heating than one in theSouth. Total service demand for any service dependson the floorspace, type, and location of buildings.Base service demand by end use by building type andCensus division is derived from estimates developedfrom CBECS energy consumption. Projected servicedemands are adjusted for trends in new constructionbased on CBECS data concerning recent construc-tion.

Distributed Generation and CHPSubmodule

Commercial consumers may decide to purchaseequipment to generate electricity (and perhaps pro-vide heat as well) rather than depend on purchasedelectricity to fulfill all of their electric power require-ments. The third basic step of the commercial mod-ule structure projects electricity generation, fuel con-sumption, water heating, and space heating suppliedby ten distributed generation and CHP technologies.The characterized technologies include: photovoltaicsolar systems; natural gas fuel cells, reciprocatingengines, turbines and microturbines; diesel engines;coal–fired CHP; and municipal solid waste, wood,and hydroelectric generators.

Existing electricity generation by CHP technologiesis derived from historical data contained in the mostrecent year’s version of Form EIA–860B, AnnualElectric Generator Report–Nonutility. The esti-mated units form the installed base of CHP equip-ment that is carried forward into future years andsupplemented with any projected additions. Pro-

gram driven installations of solar photovoltaic sys-tems and fuel cells are also included based on infor-mation from the Departments of Energy and De-fense. For years following the base year, an endoge-nous forecast of distributed generation and CHP isdeveloped based on the economic returns projectedfor distributed generation technologies. A detailedcash–flow approach is used to estimate the numberof years required to achieve a positive cumulativecash flow. The calculations include the annual costs(down payments, loan payments, maintenance costs,and fuel costs) and returns (tax deductions, tax cred-its, and energy cost savings) from the investmentcovering a 30–year period from the time of the in-vestment decision. Penetration of these technologiesis a function of how quickly an investment in a tech-nology is estimated to recoup its flow of costs. Interms of NEMS projections, investments in distrib-uted generation reduce purchases of electricity. Fuelconsuming technologies also generate waste heatwhich is assumed to be partially captured and usedto offset commercial water heating and space heat-ing energy use.

Equipment Choice Submodule

Once service demands are projected, the next step isto project the type and efficiency of equipment thatwill be used to satisfy the demands. The bulk ofequipment required to meet service demand willcarry over from the equipment stock of the previousmodel year. However, equipment must always bepurchased to satisfy service demand for new con-struction. It must also be purchased for equipmentwhich has either worn out (replacement equipment)or reached the end of its economically useful life (ret-rofit equipment). For required equipment replace-ments, CDM uses a constant decay rate based onequipment life. A technology will be retrofitted onlyif the combined annual operating and maintenancecosts plus annualized capital costs of a potentialtechnology are lower than the annual operating andmaintenance costs of an existing technology.

Equipment choices are made based on a comparisonof annualized capital and operating and mainte-nance costs across all allowable equipment for a par-ticular end–use service. In order to add inertia to theequipment choices, only subsets of the total menu ofpotentially available equipment may be allowed fordefined market segments. For example,only 8 per-cent of floorspace in large office buildings may con-sider all available equipment using any fuel or tech-nology when making space heating equipment re-placement decisions. A second segment equal to 35percent of floorspace, must select from technologies



using the same fuel as already installed. A third seg-ment, the remaining 57 percent of floorspace, is con-strained to consider only different efficiency levels ofthe same fuel and technology already installed. Forlighting, all replacement choices are limited to thesame technology, where technologies are broadly de-fined to encompass principal competing technologies(outdoor lighting types do not compete for indoorlighting service demand).

When computing annualized costs for determiningequipment choices, commercial floorspace is seg-mented by what are referred to as hurdle rates or im-plicit discount rates (to distinguish them from thegenerally lower and more common notion of financialdiscount rates). Seven segments are used to simulateconsumer behavior when purchasing commercialequipment. The segments range from rates as low asthe 10–year Treasury bond rate, to rates highenough to guarantee that only equipment with thelowest capital cost (and least efficiency) is chosen. Asreal energy prices increase (decrease) there is an in-centive for all but the highest implicit discount ratesegments to purchase increased (decreased) levels ofefficiency.

The equipment choice submodule is designed tochoose among a discrete set of technologies that arecharacterized by a menu which defines availability,capital costs, maintenance costs, efficiencies, andequipment life. Technology characteristics for se-lected space heating equipment are shown in the ta-ble on page 31, derived from the report Assumptionsto the Annual Energy Outlook 2003.22 This menu ofprojected equipment models projects technological

innovation, market developments, and policy inter-ventions. For the Annual Energy Outlook 2003, thetechnology types that are included for seven of theten service demand categories are listed in the tableon page 32.

The remaining three end–use services (PC–relatedoffice equipment, other office equipment, and otherend uses) are considered minor services and are fore-cast using exogenous equipment efficiency and mar-ket penetration trends.

Energy Consumption Submodule

Once the required equipment choices have beenmade, the total stock and efficiency of equipment fora particular end use are determined. Energy con-sumption by fuel can be calculated from the amountof service demand satisfied by each technology andthe corresponding efficiency of the technology. Atthis stage, adjustments to energy consumption arealso made.

These include adjustments for changes in real en-ergy prices (short–run price elasticity effects), ad-justments in utilization rates caused by efficiency in-creases (efficiency rebound effects), and changes forweather relative to the CBECS survey year. Oncethese modifications are made, total energy use iscomputed across end uses and building types for thethree major fuels, for each Census division. Com-bining these projections with the econometric/trendprojections for the five minor fuels yields total pro-jected commercial energy consumption.



22 Energy Information Administration, Assumptions to the Annual Energy Outlook 2003,http://www.eia.doe.gov/loiaf/aeo/assumption/pdf/0554(2003).pdf (Washington, DC, January 2003).



Characteristics of Selected Space Heating Equipment in the Commercial Sector

Equipment Type1 Vintage Efficiency2

Capital Cost(2001 dollars perthousand Btu per

hour)3

Maintenance Cost(2001 dollars perthousand Btu per

hour)3

ServiceLife

(years)

Electric Heat Pump Current Standard 6.8 $81.39 $3.33 14

2000- typical 7.5 $97.92 $3.33 14

2000- high efficiency 9.8 $155.56 $3.33 14

2005- typical 7.5 $97.22 $3.33 14

2005- high efficiency 9.8 $155.56 $3.33 14

2010 - typical 7.5 $97.22 $3.33 14

2010 - high efficiency 9.8 $155.56 $3.33 14

2020 - typical 7.8 $97.22 $3.33 14

2020 - high efficiency 10.0 $150.00 $3.33 14

Ground-Source Heat Pump 2000- typical 3.4 $167.50 $1.46 20

2000- high efficiency 4.0 $229.17 $1.46 20

2005- typical 3.4 $166.67 $1.46 20

2005- high efficiency 4.3 $229.17 $1.46 20

2010 - typical 3.4 $166.67 $1.46 20

2010 - high efficiency 4.3 $208.33 $1.46 20

2020- typical 3.8 $166.67 $1.46 20

2020 -high efficiency 4.5 $197.92 $1.46 20

Electric Boiler Current Standard 0.98 $21.83 $0.14 21

Packaged Electric 1995 0.93 $19.77 $3.49 18

Natural Gas Furnace Current Standard 0.80 $9.11 $1.00 15

2000- high efficiency 0.92 $14.82 $0.88 15

2010 - typical 0.81 $8.70 $0.96 15

Natural Gas Boiler Current Standard 0.80 $16.11 $0.55 25

2000 - high efficiency 0.87 $33.82 $0.69 25

2005- typical 0.81 $17.87 $0.55 25

2005- high efficiency 0.90 $31.68 $0.67 25

Natural Gas Heat Pump 2005- absortion 1.4 $173.61 $4.17 15

Distillate Oil Furnace Current Standard 0.81 $14.25 $1.00 15

2000 0.86 $23.46 $1.00 15

2010 0.89 $22.69 $1.00 15

Distillate Oil Boiler Current Standard 0.83 $15.76 $0.13 20

2000- high efficiency 0.88 $18.83 $0.12 20

2005- typical 0.83 $15.76 $0.13 20

2005- high efficiency 0.88 $18.83 $0.12 20

1Equipment listed is for the New England Census division but is also representative of the technology data for the rest of the United States.2Efficiency measurements vary by equipment type. Electric air-source and natural gas heat pumps are rated for heating performance usingthe Heating Seasonal Performance Factor (HSPF); natural gas and distillate furnaces are based on Annual Fuel Utilization Efficiency;ground-source heat pumps are rated on coefficient of performance; and boilers are based on combustion efficiency.3Capital and maintenance costs are given in 2001 dollars.Source: Energy Information Administration, “Technology Forecast Updates - Residential and Commercial Building Technologies - ReferenceCase”, Arthur D. Little, Inc., Reference Number 8675309, October 2001.



End-Use Service by Fuel Technology Types

Electric Space Heating: air-source heat pump, ground-source heat pump, boiler, packaged space heating

Natural Gas Space Heating: boiler, furnace, engine-driven heat pump, absorption heat pump

Fuel Oil Space Heating: boiler, furnace

Electric Space Cooling:air-source heat pump, ground-source heat pump, reciprocating chiller, centrifugal chiller,rooftop air conditioner, residential style central air conditioner, window unit

Natural Gas Space Cooling:absorption chiller, engine-driven chiller, rooftop air conditioner, engine-driven heat pump,absorption heat pump

Electric Water Heating: electric resistance, heat pump water heater, tankless water heater

Natural Gas Water Heating: natural gas water heater, tankless water heater

Fuel Oil Water Heating: fuel oil water heater

Ventilation:small Constant Air Volume (CAV) system, large CAV system, small Variable Air Volume (VAV)system, large VAV system, fan coil unit, multi-zone CAV system

Electric Cooking: range, convection oven, deck oven, fryer, griddle, other electric

Natural Gas Cooking: range, range w/power burner, deck oven, fryer, infrared fryer, griddle, infrared griddle, other

Incandescent Style Lighting:incandescent, compact fluorescent, halogen, halogen-infrared, coated filament, hafniumcarbide

Four-foot Fluorescent Lighting:magnetic ballast, electronic ballast, electronic w/controls, electronic w/reflectors, scotopic,electrodeless

Eight-foot Fluorescent Lighting:magnetic ballast, electronic ballast, magnetic-high output, electronic-high output, scotopic,electrodeless

High Intensity Discharge Lighting: metal halide, mercury vapor, high pressure sodium, sulfur

Refrigeration: centralized refrigeration system, walk-in cooler, walk-in freezer, reach-in refrigerator, reach-infreezer, ice machine, refrigerated vending machine

Commercial End-Use Technology Types

The industrial demand module (IDM) forecasts en-ergy consumption for fuels and feedstocks for ninemanufacturing industries and six nonmanufactur-ing industries, subject to delivered prices of energyand macroeconomic variables representing the valueof shipments for each industry. The module includeselectricity generated through combined heat andpower (CHP) systems that is either used in the in-dustrial sector or sold to the electricity grid. The IDMstructure is shown in Figure 7.

Industrial energy demand is projected as a combina-tion of “bottom up” characterizations of the en-ergy-using technology and “top down” econometricestimates of behavior. The influence of energy priceson industrial energy consumption is modeled interms of the efficiency of use of existing capital, theefficiency of new capital acquisitions, and the mix offuels utilized, given existing capital stocks. Energyconservation from technological change is repre-sented over time by trend-based “technology possi-bility curves.” These curves represent the aggregateefficiency of all new technologies that are likely topenetrate the future markets as well as the aggre-gate improvement in efficiency of 1998 technology.

IDM incorporates three major industry categories:energy–intensive manufacturing industries, non-energy–intensive manufacturing industries, andnonmanufacturing industries. The level and type ofmodeling and the attention to detail is different foreach. Manufacturing disaggregation is at the 3-digitNorth American Indusrial Classification System(NAICS) level, with some further disaggregation ofmore energy–intensive or large energy–consumingindustries. Industries treated in more detail includefood, paper, chemicals, glass, cement, steel, and alu-minum. Energy product demands are calculated in-dependently for each industry.

Each industry is modeled (where appropriate) asthree interrelated components: buildings (BLD),boilers/steam/CHP (BSC), and process/assembly(PA) activities. Buildings are estimated to accountfor 6 percent of energy consumption in manufactur-

ing industries (in nonmanufacturing industries,building energy consumption is assumed to be negli-gible).

Consequently, IDM uses a simple modeling ap-proach for the BLD component. Energy consumption

INDUSTRIAL DEMAND MODULE


Energy-IntensiveManufacturing

NonmanufacturingIndustries

Food and Kindred Products

(NAICS 311)

Paper and Allied Products

(NAICS 322)

Bulk Chemicals

(NAICS 325)

Glass and Glass Products

(NAICS 3272)

Hydraulic Cement

(NAICS 32731)

Blast Furnaces and Basic

Steel

(NAICS 331111)

Aluminum

(NAICS 3313)

Agricultural Production - Crops

(NAICS 111)

Other Agriculture including

Livestock

(NAICS 112-115)

Coal Mining

(NAICS 2121)

Oil and Gas Extraction

(NAICS 211)

Metal and Other Nonmetallic

Mining

(NAICS 2122-2123)

Construction

(NAICS 233-235)

Nonenergy-IntensiveManufacturing

Metals-Based Durables

(NAICS 332-336)

Other Manufacturing

(all remaining manufacturing

NAICS)

NAICS = North American Industry Classification System.

Economic Subsectors Within the IDM

IDM Outputs Inputs from NEMS Exogenous Inputs


Electricity sales to grid

Cogeneration output and fuel

consumption


Economic output by industry

Refinery fuel consumption

Lease and plant fuel consumption

Cogeneration from refineries and oil and gas

production

Production stages in energy-intensive industries

Technology possibility curves

Unit energy consumption

Stock retirement rates

in industrial buildings is assumed to grow at thesame rate as the average growth rate of employmentand output in that industry. The BSC componentconsumes energy to meet the steam demands fromthe other two components and to provide internallygenerated electricity to the BLD and PA compo-nents. The boiler component consumes fossil fuels toproduce steam, which is passed to the PA compo-nent.

IDM models “traditional” combined heat and power(CHP) based on steam demand from the BLD and thePA components. The “nontraditional” CHP units arerepresented in the electricity market module sincethese units are mainly grid-serving, electric-ity-price-driven entities.

CHP capacity, generation, and fuel use are calcu-lated from exogenous data on existing and plannedcapacity additions and new additions determinedfrom an engineering and economic evaluation. Ex-isting CHP capacity and planned additions are de-rived from Form EIA-860B, “Annual Electric Gener-ator Report-Nonutility,” formerly Form EIA-867,“Annual Nonutility Power Producer Report.” Ex-isting CHP capacity is assumed to remain in servicethroughout the forecast or, equivalently, to be refur-bished or replaced with similar units of equal capac-ity.

EIA has comprehensively reviewed and revised howit collects, estimates, and reports fuel use for facili-ties producing electricity. The review addressed both



Industrial Demand Module

Unit Energy Consumptionby Industry,

Production Stage Informationfor Energy-Intensive Industries,Technology Possibility Curves,

Stock Retirement Rates

Exogenous

Boilers/Steam/CHPComponent

BuildingsComponent

Process/AssemblyComponent

Steam and CHP Electricity

N

E

M

S

Macroeconomic

Activity Module

Electricity

Market

Module

Coal Market

Module

Oil and Gas

Supply Module

Natural Gas

Transmission

and Distribution

Module

Petroleum Market

Module

Economic Activity bySector

Electricity Prices

Electricity Demand

Electricity Sales to Grid

CHP Output

and Fuel Consumption

Coal Prices

Coal Demand

CombinedHeat and Power

Natural Gas Prices

Natural Gas Demand

Lease and Plant FuelConsumption


Refinery Fuel Consumption

CHP

Petroleum Demand

Energy Byproducts

Figure 7. Industrial Demand Module Structure

inconsistent reporting of the fuels used for electricpower across historical years and changes in theelectric power marketplace that have been inconsis-tently represented in various EIA survey forms andpublications. In comparison with EIA’s past energydata publications, the impact of the definitionchanges for the industrial sector is to reduce mea-sured natural gas consumption. For example, thepreviously reported value for 2000 has been revisedfrom 9.39 trillion cubic feet to 8.25 trillion cubic feet.In comparison with past energy data publications,the impact of the definition changes and new datasources for total energy use increases measured nat-ural gas consumption. Total natural gas consump-tion in 2000 is 0.6 trillion cubic feet higher than waspreviously reported. A more detailed discussion ofthis update is available in EIA’s Annual Energy Re-view 2001, Appendix H, “Estimating and PresentingPower Sector Fuel Use in EIA Publications andAnalyses,” web site www.eia.doe.gov/emeu/aer/pdf/pages/sec_h.pdf.

Calculation of unplanned CHP capacity additionsbegins in 2001. Modeling of unplanned capacity ad-ditions is done in two parts: biomass-fueled and fos-sil-fueled. Biomass CHP capacity is assumed to beadded to the extent possible as additional biomasswaste products are produced, primarily in the pulpand paper industry. The amount of biomass CHP ca-pacity added is equal to the quantity of new biomassavailable (in Btu), divided by the total heat rate frombiomass steam turbine CHP.

Additions to fossil-fueled CHP capacity are limited togas turbine plants. It is assumed that the technicalpotential for CHP is based primarily on supplyingthermal requirements. First, the model assesses theamount of capacity that could be added to generatethe industrial steam requirements not met by exist-ing CHP. The second step is an economic evaluationof gas turbine prototypes for each steam load seg-ment. Finally, CHP additions are projected based ona range of acceptable payback periods.

The PA component accounts for the largest share ofdirect energy consumption for heat and power, 55percent. For the seven most energy-intensive indus-tries, process steps or end uses are modeled using en-gineering concepts. The production process is decom-posed into the major steps, and the energy relation-ships among the steps are specified.

The energy intensities of the process steps or enduses vary over time, both for existing technology andfor technologies expected to be adopted in the future.In IDM, this variation is based on engineering judg-ment and is reflected in the parameters of technologypossibility curves, which show the declining energyintensity of existing and new capital relative to the1998 stock.

IDM uses “technology bundles” to characterize tech-nological change in the energy-intensive industries.These bundles are defined for each production pro-cess step for five of the industries and for end use intwo of the industries. The process step industries arepulp and paper, glass, cement, steel, and aluminum.The end-use industries are food and bulk chemicals.

Machine drive electricity consumption in the food,bulk chemicals, metal-based durables, and balanceof manufacturing sectors is calculated by a motorstock model. The beginning stock of motors is modi-fied over the forecast horizon as motors are added toaccommodate growth in shipments for each sector,as motors are retired and replaced, and as failed mo-tors are rewound. When a new motor is added, ei-ther to accommodate growth or as a replacement, aneconomic choice is made between purchasing a motor



End Use Characterization

Food: direct fuel, hot water/steam, refrigeration, and

other electric.

Bulk Chemicals: direct fuel, hot water/steam,

electrolytic, and other electric.

Process Step Characterization

Pulp and Paper: wood preparation, waste pulping,

mechanical pulping, semi-chemical pulping, Kraft

pulping, bleaching, and papermaking.

Glass: batch preparation, melting/refining, and forming.

Cement: dry process clinker, wet process clinker, and

finish grinding.

Steel: coke oven, open hearth steel making, basic

oxygen furnace steel making, electric arc furnace steel

making, ingot casting, continuous casting, hot rolling,

and cold rolling.

Aluminum: only primary aluminum smelting is explicitly

included.

Fuel Consuming Activities for theEnergy-Intensive Manufacturing Subsectors

which meets the EPACT minimum for efficiency or apremium efficiency motor. There are seven motorsize groups in each of the four industries. TheEPACT efficiency standards only apply to the fivesmallest groups (up to 200 horsepower). As the mo-tor stock changes over the forecast horizon, the over-all efficiency of the motor population changes aswell.

The unit energy consumption is defined as the en-ergy use per ton of throughput at a process step or asenergy use per dollar of shipments for the end use in-dustries. The “Existing UEC” is the current averageinstalled intensity as of 1998. The “New 1998 UEC”is the intensity expected to prevail for a new installa-tion in 1999. Similarly, the “New 2025 UEC” is theintensity expected to prevail for a new installation in2025. For intervening years, the intensity is interpo-lated.

The rate at which the average intensity declines isdetermined by the rate and timing of new additionsto capacity. In IDM, the rate and timing of new addi-tions are functions of retirement rates and industrygrowth rates.

IDM uses a vintaged capital stock accounting frame-work that models energy use in new additions to thestock and in the existing stock. This capital stock isrepresented as the aggregate vintage of all plantsbuilt within an industry and does not imply the in-clusion of specific technologies or capital equipment.

The capital stock is grouped into three vintages: old,middle, and new. The old vintage consists of capitalin production prior to 1998, which is assumed to re-tire at a fixed rate each year. Middle-vintage capitalis that added after 1998, excluding the year of theforecast. New production capacity is built in the fore-

cast years when the capacity of the existing stock ofcapital in IDM cannot produce the output forecastedby the NEMS regional submodule of the macroeco-nomic activity module. Capital additions during theforecast horizon are retired in subsequent years atthe same rate as the pre-1998 capital stock.

The energy-intensive and/or large energy-consum-ing industries are modeled with a structure that ex-plicitly describes the major process flows or “stagesof production” in the industry (some industries havemajor consuming uses).

Technology penetration at the level of major pro-cesses in each industry is based on a technology pen-etration curve relationship. A second relationshipcan provide additional energy conservation resultingfrom increases in relative energy prices. Major pro-cess choices (where applicable) are determined by in-dustry production, specific process flows, and exoge-nous assumptions.

IDM achieves fuel switching by application of a logitfunction methodology for estimating fuel shares inthe boilers/steam/CHP component. A small amountof additional fuel switching capability takes placewithin the PA component.

Recycling, waste products, and byproduct consump-tion are modeled using parameters based on off-lineanalysis and assumptions about the manufacturingprocesses or technologies applied within industry.These analyses and assumptions are mainly basedupon environmental regulations such as governmentrequirements about the share of recycled paper usedin offices. IDM also accounts for trends within indus-try toward the production of more specialized prod-ucts such as specialized steel which can be producedusing scrap material versus raw iron ore.



The transportation demand module (TRAN) fore-casts the consumption of transportation sector fuelsby transportation mode, including the use ofrenewables and alternative fuels, subject to deliv-ered prices of energy fuels and macroeconomic vari-ables, including disposable personal income, grossdomestic product, level of imports and exports, in-dustrial output, new car and light truck sales, andpopulation. The structure of the module is shown inFigure 8.

NEMS projections of future fuel prices influence fuelefficiency, vehicle-miles traveled, and alterna-tive-fuel vehicle (AFV) market penetration for thecurrent fleet of vehicles. Alternative-fuel shares areprojected on the basis of a multinomial logit vehicleattribute model, subject to State and Federal govern-ment mandates.

Fuel Economy Submodule

This submodule projects new light-duty vehicle fuelefficiency by 12 U.S. Environmental ProtectionAgency (EPA) vehicle size classes and 15 enginetechnologies (gasoline, diesel, and 13 AFV technolo-gies) as a function of energy prices and in-come-related variables. There are 59 fuel-savingtechnologies which vary in cost and marginal fuelsavings by size class. Characteristics of a sample ofthese technologies are shown on page 40, a completelist is published in Assumptions to the Annual En-ergy Outlook 2003.23 Technologies penetrate themarket based on a cost-effectiveness algorithmwhich compares the technology cost to the dis-counted stream of fuel savings and the value of per-formance to the consumer. In general, higher fuelprices lead to higher fuel efficiency estimates within

each size class, a shift to a more fuel-efficient sizeclass mix, and an increase in the rate at which alter-native-fuel vehicles enter the marketplace.

Regional Sales Submodule

Vehicle sales from the macroeconomic activity mod-ule are divided into car and light truck salesbased on demographic analysis. The remainder ofthe submodule is a simple accounting mechanismthat uses endogenous estimates of new car andlight truck sales and the historical regional vehiclesales adjusted for regional population trends to pro-duce estimates of regional sales, which are subse-quently passed to the alternative-fuel vehicle andthe light-duty vehicle stock submodules.

Alternative-Fuel Vehicle Submodule

This submodule projects the sales shares of alterna-tive-fuel technologies as a function of time, technol-ogy attributes, costs, and fuel prices. The alterna-tive-fuel technologies are listed on the next page. Ve-hicle attributes are shown on page 40, derived fromAssumptions to the Annual Energy Outlook 2003.Both conventional and new technology vehicles areconsidered. The alternative–fuel vehicle submodulereceives regional new car and light truck sales bysize class from the regional sales submodule.

The forecast of vehicle sales by technology utilizes anested multinomial logit (NMNL) model that pre-dicts sales shares based on relevant vehicle and fuelattributes. The nesting structure first predicts theprobability of fuel choice for multi-fuel vechicleswithin a technology set. The second level nestingpredicts penetration among similar technologieswithin a technology set (i.e. Gasoline versus diesel

TRANSPORTATION DEMAND MODULE


TRAN Outputs Inputs from NEMS Exogenous Inputs

Fuel demand by mode

Sales, stocks and characteristics of vehicle

types by size class

Vehicle-miles traveled

Fuel efficiencies by technology type

Alternative-fuel vehicle sales by

technology type

Light-duty commercial fleet vehicle

characteristics



Disposable personal income

Industrial output

Vehicle sales

International trade

Natural gas pipeline consumption

Current and projected demographics

Existing vehicle stocks by vintage and

fuel efficiency

Vehicle survival rates

New vehicle technology characteristics

Fuel availability

Commercial availability

Vehicle safety and emissions regulations

Vehicle miles-per-gallon degradation rates

23 Energy Information Administration, Assumptions to the Annual Energy Outlook 2003,http:/www.eia.doe.gov/oiaf/aeo/assumption/pdf/0554(2003).pdf (Washngton, DC, January 2003).

hybrids). The third level choice determines marketshare among the different technology sets.24 Thetechnology sets include:

• Conventional fuel capable (gasoline, diesel,bi-fuel and flex-fuel),

• Hybrid (gasoline and diesel),

• Dedicated alternative fuel (CNG, LPG,methanol, and ethanol),

• Fuel cell (gasoline, methanol, and hydrogen),and



Methanol flex-fueled

Methanol neat (85 percent methanol)

Ethanol flex-fueled

Ethanol neat (85 percent ethanol)

Compressed natural gas (CNG)

CNG bi-fuel

Liquefied petroleum gas (LPG)

LPG bi-fuel

Electric

Diesel-electric hybrid

Fuel cell gasoline

Fuel cell hydrogen

Fuel cell methanol

Alternative Fuel Vehicles

Transportation Demand Module

FreightTransport

Submodule

MiscellaneousEnergy UseSubmodule

AircraftFleet Efficiency

Submodule

CommercialLight TruckSubmodule

Vehicle MilesTraveled

Submodule

FuelDemand

Submodule

Light-DutyVehicle

CommercialFleet

Submodule

Alternative-Fuel VehicleSubmodule

RegionalSales

Submodule

FuelEconomy

Submodule

Air TravelDemand

Submodule

NEMS

Gross DomesticProduct

Disposable PersonalIncome

Industrial Output

International Trade

New Car andLight Truck Sales

Petroleum Demand

Natural Gas Prices

Pipeline FuelConsumption

Natural Gas Demand

Electricity Prices

Electricity Demand


Natural GasTransportation

andDistribution

Module



Exogenous

Current and ProjectedDemographics, Existing Vehicle

Stocks by Vintage and Fuel Efficiency,Vehicle Survival Rates, New VehicleTechnology Characteristics, FuelAvailability, CommercialAvailability, Vehicle Safety andEmissions Regulations, VehicleMiles-per-Gallon Degradation Rates


Figure 8. Transportation Demand Module Structure

24 Greene, David L. and S.M. Chin, “Alternative Fuels and Vehicles (AFV) Model Changes,” Center for TransportationAnalysis, Oak Ridge National Laboratory, page 1, (Oak Ridge, TN, November 14, 2000).

• Electric battery powered (lead acid,nickel-metal hydride, lithium polymer)25

The vehicles attributes considered in the choice algo-rithm include: price, maintenance cost, battery re-placement cost, range, multi-fuel capability, homerefueling capability, fuel economy, acceleration andluggage space. With the exception of maintenancecost, battery replacement cost, and luggage space,

vehicle attributes are determined endogenously.26

The fuel attributes used in market share estimationinclude availability and price. Vehicle attributesvary by six EPA size classes for cars and light trucksand fuel availability varies by Census division. TheNMNL model coefficients were developed to reflectpurchase decisions for cars and light trucks sepa-rately.

Light-Duty Vehicle Stock Submodule

This submodule specifies the inventory of light-dutyvehicles from year to year. Survival rates are appliedto each vintage, and new vehicle sales are introducedinto the vehicle stock through an accounting frame-work. The fleet of vehicles and their fuel efficiency

characteristics are important to the translation oftransportation services demand into fuel demand.

TRAN maintains a level of detail that includestwenty vintage classifications and six passenger carand six light truck size classes corresponding to EPAinterior volume classifications for all vehicles lessthan 8,500 pounds, as follows:

Vehicle-Miles Traveled (VMT)Submodule

This submodule projects travel demand for automo-biles and light trucks. VMT per capita estimates arebased on the fuel cost of driving per mile, per capitadisposable personal income, and an adjustment forfemale-to-male driving ratios. Total VMT is calcu-lated by multiplying VMT per capita by the drivingage population.

Light-Duty Vehicle Commercial FleetSubmodule

This submodule generates estimates of the stock ofcars and trucks used in business, government, andutility fleets. It also estimates travel demand, fuel ef-ficiency, and energy consumption for the fleet vehi-cles prior to their transition to the private sector atpredetermined vintages.

Commercial Light Truck Submodule

The commercial light truck submodule estimatessales, stocks, fuel efficiencies, travel, and fuel de-mand for all trucks greater than 8,500 pounds andless than 10,000 pounds.

Air Travel Demand Submodule

This submodule estimates the demand for both pas-senger and freight air travel. Passenger travel isforecasted by domestic travel, which is dissaggre-regated between business and personal travel, andinternational travel. Dedicated air freight travel isdisaggregated between the total air freight demandand air freight carried in the lower hull of commer-cial passenger aircraft. In each of the market seg-ments, the demand for air travel is estimated as afunction of the cost of air travel (including fuel costs)



Cars:

Mini-compact - less than 85 cubic feet

Subcompact - between 85 and 99 cubic feet

Compact - between 100 and 109 cubic feet

Mid-size - between 110 and 119 cubic feet

Large - 120 or more cubic feet, including all station

wagons (small, mid-size, and large)

Two-seater - designed to seat two adults

Trucks:Passenger vans

Cargo vans

Small pickups - trucks with gross vehicle weight rating

(GVWR) less than 4,500 pounds

Large pickups - trucks with GVWR 4,500 to 8,500 pounds

Small utility

Large utility

Light Vehicle Size Classes

25 U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, prepared by Interlaboratory WorkingGroup, Scenarios of U.S. Carbon Reductions: Potential Impacts of Energy Technologies by 2010 and Beyond,(Washington, DC, 1998).

26 Energy and Environmental Analysis, Inc., Updates to the Fuel Economy Model (FEM) and Advanced Technology Vehicle(ATV) Module of the National Energy Modeling System (NEMS) Transportation Model, Prepared for the EnergyInformation Administration (EIA), (Arlington, VA, October 23, 2000).

and economic growth (GDP, disposable income, andmerchandise exports).

Aircraft Fleet Efficiency Submodule

This submodule forecasts the total stock and the av-erage fleet efficiency of narrow body and wide bodyaircraft required to meet the projected travel de-mand. The stock estimation is based on the growth oftravel demand and a logistic function that calculatesthe survival of the older planes. The overall fleet effi-ciency is determined by the weighted average of thesurviving aircraft efficiency (including retrofits) andthe efficiencies of the newly acquired aircraft. Theefficiency improvements of the new aircraft are de-termined by technology choice which depends on thetrigger fuel price, the time in which the technology iscommercially viable, and by the expected efficiency

gains of aircraft incorporating those technologies.Technology characteristics are shown on page 41.

Freight Transport Submodule

This submodule translates NEMS estimates of in-dustrial production into ton-miles traveled require-ments for rail and ship travel, and into vehicle-milestraveled for trucks, then into fuel demand by mode offreight travel. The freight truck stock is subdividedinto medium and heavy-duty trucks. VMT freight es-timates by truck size class and technology are basedon matching freight needs, as measured by thegrowth in industrial output by Standard IndustrialClassification (SIC) code, to VMT levels associatedwith truck stocks and new vehicles. Rail and ship-ping ton-miles traveled are also estimated as a func-tion of growth in industrial output.



Fractional Fuel

Efficiency Change First Year IntroducedFractional Horsepower

Change

Material Substitution IV 0.099 2006 0

Drag Reduction IV 0.063 2002 0

5–Speed Automatic 0.065 1995 0

CVT 0.105 1998 0

Automated Manual Trans 0.080 2006 0

VVL–6 Cylinder 0.050 2000 0.10

Camless Valve Actuation 6 Cylinder 0.110 2008 0.13

Electric Power Steering 0.020 2004 0

42V–Launch Assist and Regen 0.030 2005 -0.05

Selected Technology Characteristics for Automobiles

Year GasolineDieselFlex

EthanolFlex LPG

Electric-DieselHybrid

Fuel CellHydrogen

Vehicle Price (thousand 2001 dollars) 2001 25.0 27.2 27.3 30.7 36.1 81.0*

2025 26.5 28.4 26.8 32.2 28.9 53.8

Vehicle Miles per Gallon 2001 26.7 35.9 26.9 27.7 42.6 52.9*

2025 27.6 34.5 27.9 28.5 40.0 49.8

Vehicle Range (miles) 2001 450 610 330 390 590 450*

2025 470 630 340 400 610 470

Fuel Availability Relative to Gasoline 2001 1.00 1.00 1.00** 0.30 1.00 0.00*

2025 1.00 1.00 1.00 0.40 1.00 0.90

*Data for fuel cell hydrogen automobiles is for 2005, first year of availability.**Due to availability using gasoline.

Examples of Midsize Automobile Attributes

Freight truck fuel efficiency growth rates are tied tohistorical growth rates by size class and are also de-pendent on the maximum penetration, introductionyear, fuel trigger price (based on cost-effectiveness),and fuel economy improvement of advanced technol-ogies, which include alternative-fuel technologies. Asubset of the technology characteristics are shown onpage 42. In the rail and shipping modes, energy effi-ciency estimates are structured to evaluate the po-

tential of both technology trends and efficiency im-provements related to energy prices.

Miscellaneous Energy Use Submodule

This submodule projects the use of energy in militaryoperations, mass transit vehicles, recreational boats,and lubricants, based on endogenous variableswithin NEMS (e.g., vehicle fuel efficiencies) and ex-ogenous variables (e.g., the military budget).



TechnologyIntroduction

Year

Jet Fuel PricesNecessary for

Cost-Effectiveness(1997 dollars per gallon)

Seat-Milesper Gallon

Gain Over 1990(Percent)

Narrow Body Wide Body

Engines

Ultra-high Bypass 1995 0.69 10 10

Propfan 2000 1.36 23 0

Thermodynamics 2010 1.22 20 20

Aerodynamics

Hybrid Laminar Flow 2020 1.53 15 15

Advanced Aerodynamics 2000 1.70 18 18

Other

Weight Reducing Materials 2000 - 15 15

Aircraft Technology Characteristics



Fuel EconomyImprovement

(percent)

MaximumPenetration

(percent)Introduction

YearCapital Cost

(2001 dollars)

Medium Heavy Medium Heavy Medium Heavy Medium Heavy

Aero Dynamics: bumper,underside air baffles, wheel well covers 2.3 2.7 50 66 2005 2005 $280 $550

Low rolling resistence tires3.6 2.3 50 40 2004 2005 $800 $1,500

Transmission: lock-up, electroniccontrols, reduced friction 1.8 1.8 100 100 2005 2005 $900 $1,000

Diesel Engine:hybrid electric powertrain 36.0 N/A 15 N/A 2010 N/A $8,000 N/A

Reduce waste heat, thermal mgmtN/A 9.0 N/A 35 N/A 2010 N/A $2,000

Gasoline Engine:

Direct injection10.8 N/A 25 N/A 2008 N/A $700 N/A

Weight Reduction4.5 9.0 20 30 2007 2005 $2,000 $2,000

Diesel EmissionNOx non-thermal plasma catalyst -1.5 -1.5 25 25 2006 2007 $1,200 $1,250

PM catalytic filter-2.5 -1.5 95 95 2006 2006 $1,250 $1,500

HC/CO: oxidation catalyst -0.5 -0.5 95 95 2002 2002 $200 $250

NOX adsorbers -3.0 -3.0 90 90 2006 2007 $2,000 $2,500

Freight Truck Technology Characteristics

The electricity market module (EMM) represents thegeneration, transmission, and pricing of electricity,subject to: delivered prices for coal, petroleum prod-ucts, and natural gas; the cost of centralized genera-tion from renewable fuels; macroeconomic variablesfor costs of capital and domestic investment; andelectricity load shapes and demand. The submodulesconsist of capacity planning, fuel dispatching, fi-nance and pricing, and load and demand–side man-agement (Figure 9). In addition, nonutility supplyand electricity trade are represented in the fuel dis-patching and capacity planning submodules.Nonutility generation from combined heat andpower (CHP) and other facilities whose primarybusiness is not electricity generation is representedin the demand and fuel supply modules. All othernonutility generation is represented in the EMM.The generation of electricity is accounted for in 15supply regions (Figure 10), and fuel consumption isallocated to the 9 Census divisions.

The EMM determines airborne emissions producedby the generation of electricity. It represents limitsfor sulfur dioxide and nitrogen oxides specified in theClean Air Act Amendments of 1990. The EMM canalso examine potential legislation that requires morestringent restrictions on sulfur dioxide and nitrogenoxides as well as limits on mercury and carbon diox-ide.

Operating (dispatch) decisions are provided by thecost–minimizing mix of fuel and variable operatingand maintenance (O&M) costs, subject to environ-mental costs. Capacity expansion is determined bythe least–cost mix of all costs, including capital,O&M, and fuel. Electricity demand is represented byload curves, which vary by region, season, and timeof day. The solution to the submodules of EMM is si-multaneous in that, directly or indirectly, the solu-tion for each submodule depends on the solution to

every other submodule. A solution sequence throughthe submodules can be viewed as follows:

• The load and demand–side managementsubmodule processes electricity demand toconstruct load curves

• The electricity capacity planning submoduleprojects the construction of new utility andnonutility plants, the level of firm powertrades, and the addition of scrubbers forenvironmental compliance

• The electricity fuel dispatch submoduledispatches the available generating units,both utility and nonutility, allowing surpluscapacity in select regions to be dispatched foranother regions needs (economy trade)

• The electricity finance and pricing submodulecalculates total revenue requirements for eachoperation and computes average andmarginal–cost based electricity prices.

Electricity Capacity PlanningSubmodule

The electricity capacity planning (ECP) submoduledetermines how best to meet expected growth inelectricity demand, given available resources, ex-pected load shapes, expected demands and fuelprices, environmental constraints, and costs for util-ity and nonutility technologies. When new capacityis required to meet growth in electricity demand, thetechnology choosen is determined by the timing ofthe demand increase, the expected utilization of thenew capacity, the operating efficiencies, and the con-struction and operating costs of available technolo-gies.

The expected utilization of the capacity is importantin the decision–making process. A technology with


ELECTRICITY MARKET MODULE

EMM Outputs Inputs from NEMS Exogenous Inputs

Electricity prices and price components

Fuel demands

Capacity additions

Capital requirements

Emissions

Renewable capacity

Avoided costs

Electricity sales

Fuel prices

Cogeneration supply and fuel consumption

Electricity sales to the grid

Renewable technology characteristics, allowable capacity,

and costs

Renewable capacity factors


Interest rates

Financial data

Tax assumptions

Capital costs

Operation and maintenance costs

Operating parameters

Emissions rates

New technologies

Existing facilities

Transmission constraints

Hydropower capacity and capacity

factors

relatively high capital costs but comparatively lowoperating costs (primarily fuel costs) may be the ap-propriate choice if the capacity is expected to operatecontinuously (base load). However, a plant type withhigh operating costs but low capital costs may be themost economical selection to serve the peak load (i.e.,the highest demands on the system), which occursinfrequently. Intermediate or cycling load occupies amiddle ground between base and peak load and isbest served by plants that are cheaper to build thanbaseload plants and cheaper to operate than peakload plants.

Technologies are compared on the basis of total capi-tal and operating costs incurred over a 20–year pe-riod. As new technologies become available, they arecompeted against conventional plant types. Fos-sil–fuel, nuclear, and renewable central–station gen-erating technologies are represented, as listed onpage 46. The EMM also considers two distributedgeneration technologies–baseload and peak.

Uncertainty about investment costs for new technol-ogies is captured in ECP using technological opti-mism and learning factors. The technological opti-mism factor reflects the inherent tendency to under-estimate costs for new technologies. The degree oftechnological optimism depends on the complexity ofthe engineering design and the stage of develop-ment. As development proceeds and more data be-come available, cost estimates become more accurateand the technological optimism factor declines.

Learning factors represent reductions in capitalcosts due to learning–by–doing. For new technolo-gies, cost reductions due to learning also account forinternational experience in building generating ca-pacity.

he decrease in overnight capital costs due to learningdepends on the stage of technological development.The cost for a revolutionary technology is assumed todecrease by 10 percent for the first three doublings ofcapacity constructed, 5 percent for the next five



Electricity Market Module

Electricity Capacity

Planning

Submodule

Electricity Fuel

Dispatch

Submodule

Electricity Finance

and Pricing

Submodule

Load and

Demand-Side

Management

Submodule

Available Capacity

Fuel Demands

Average Electricity Prices

Load Curves

CapacityAdditions

Electricity Sales

CHP Supply andFuel Consumption

Sales to Grid

Electricity Prices

Natural Gas Peak andOffpeak, Core and Noncore

Demand

Petroleum Product Prices

Petroleum Demand

Electricity Price

Electricity Price

Coal Prices

Coal Demand

Renewable Capacity, AvoidedCosts, Discount Rate, Biomass

Consumption

Electricity Prices

Electric Industry Output

Macroeconomic

Activity Module

Renewable

Fuels

Module

Coal Market

Module

Oil and Gas

Supply Module

Petroleum

Market Module

Natural Gas

Transmission and

Distribution

Module

Demand

Modules

Exogenous

Financial Data, Tax Assumptions, CapitalCosts, Operating and Maintenance Costs,

Operating Parameters, Emission Rates,Existing Facilities, New Technologies,

Transmission Constraints, HydropowerCapacity and Capacity Factors

Renewable TechnologyCharacteristics, Allowable Capacity,

Capacity Factors, Costs

Expectations

Interest Rates, GrossDomestic Product

Natural Gas Prices

Figure 9. Electricity Market Module Structure

doublings, and 1 percent for every doubling thereaf-ter. The cost for an evolutionary technology is as-sumed to decrease by 5 percent for the first fivedoublings and 1 percent for every doubling thereaf-ter. The cost for a conventional technology is as-sumed to decrease by 1 percent for every doubling ofcapacity constructed.

Capital costs for all new electricity generating tech-nologies (fossil, nuclear, and renewable) decrease inresponse to foreign and domestic experience. For-eign units of new technologies are assumed to con-tribute to reductions in capital costs for units thatare installed in the United States to the extent that(1) the technology characteristics are similar to thoseused in U.S. markets, (2) the design and constructionfirms and key personnel compete in the U.S. market,(3) the owning and operating firm competes activelyin the United States, and (4) there exists relativelycomplete information about the status of the associ-ated facility. If the new foreign units do not satisfyone or more of these requirements, they are given a

reduced weight or not included in the learning ef-fects calculation. Capital costs, heat rates, and firstyear of availability from the Annual Energy Outlook2003 reference case are shown on page 46; capitalcosts represent the costs of building new plants be-ginning in 2002. For renewable technologies, thecapital costs are for California, which is representa-tive of the United States. Additional informationabout costs and performance characteristics can befound on page 73 of the Assumptions to the AnnualEnergy Outlook 2003.27

Initially, investment decisions are determined inECP using cost and performance characteristicsthat are represented as single point estimates corre-sponding to the average (expected) cost. However,these parameters are also subject to uncertainty andare better represented by distributions. If the distri-butions of two or more options overlap, the optionwith the lowest average cost is not likely to capturethe entire market. Therefore, ECP uses amarket–sharing algorithm to adjust the initial solu-



Figure 10. Electricity Market Module Supply Regions

27Energy Informtion Administration, Assumptions to the Annual Energy Outlook 2003,http://www.eia.doe.gov/oiaf/aeo/assumption/pdf/0554(2003).pdf (January 2003).

tion and reallocate some of the capacity expansiondecisions to technologies that are competitive but donot have the lowest average cost.

Fossil–fired steam and nuclear plant retirementsare calculated endogenously within the model.Plants are retired if the market price of electricity isnot sufficient to support continued operation. Theexpected revenues from these plants are comparedto the annual going–forward costs, which aremainly fuel and operations and maintenance costs.A plant is retired if these costs exceed the revenuesand the overall cost of electricity can be reduced bybuilding replacement capacity.

The ECP submodule also determines whether tocontract for unplanned firm power imports fromCanada and from neighboring electricity supply re-gions. Imports from Canada are competed usingsupply curves developed from cost estimates for po-tential hydroelectric projects in Canada. Importsfrom neighboring electricity supply regions are com-peted in ECP based on the cost of the unit in the ex-porting region plus the additional cost of transmit-ting the power. Transmission costs are computed asa fraction of revenue.

After building new capacity, the submodule passestotal available capacity to the electricity fuel dis-patch submodule and new capacity expenses to theelectricity finance and pricing submodule.



Fossil

Existing coal steam plants (with or without environmental controls)

New pulvberized coal with environmental controls

Advanced clean coal technology

Advanced clean coal technology with sequestration

Oil/Gas steam

Conventional combined cycle

Advanced combined cycle

Advanced combined cycle with sequestration

Conventional combustion turbine

Advanced combustion turbine

Fuel Cells

Nuclear

Conventional nuclear

Advanced nuclear

Renewables

Conventional hydropower

Pumped storage

Geothermal

Solar-thermal

Solar-photovoltaic

Wind

Wood

Municipal solid waste

Environmental controls include flue gas desulfurization (FGD), selectivecatalytic reduction (SCR), selective non-catalytic reduction (SNCR), fabricfilters, spray cooling, activated carbon injection (ACI), and particulateremoval equipment.

Central–Station Generating Technologies

TechnologyCapital Costs1/

(2001$/Kwhr)

2002 HeatRates

(Btu/Kwhr)OnlineYear2/

Advanced Combustion Turbine 460 8,550 2004

Conventional Combustion Turbine 409 10,450 2004

Advanced Gas/Oil Combined Cycle 608 7,000 2005

Conventional Gas/Oil Combined Cycle 536 7,500 2005

Scrubbed Coal New 1,154 9,000 2006

Integrated Gas Combined Cycle 1,367 8,000 2006

Fuel Cells 2,137 7,500 2005

Advanced Nuclear 2,117 10,400 2007

Biomass 1,763 8,911 2006

Solar Thermal 2,594 10,280 2005

Solar Photovoltaic 3,915 10,280 2004

Wind 1,003 10,280 2005

2002 Overnight Capital Costs (including Contingencies), 2002 Heat Rates, and Online Year by Technologyfor the AEO2003 Reference Case

1/Overnight capital cost include contingency factors, and exclude regional multipliers and learning effects. Interest charges are also excluded. Theserepresent costs of new projects initiated in 2002.2/Online year represents the first year that a new unit could be completed, given an order date of 2002.

Electricity Fuel Dispatch Submodule

Given available capacity, firm purchased–poweragreements, fuel prices, and load curves, the electric-ity fuel dispatch (EFD) submodule minimizes vari-able costs as it solves for generation facility utiliza-tion and economy power exchanges to satisfy de-mand in each time period and region. The submoduleuses merit order dispatching; that is, utility, inde-pendent power producer, and small power producerplants are dispatched until demand is met in a se-quence based on their operating costs, withleast–cost plants being operated first. Limits onemissions of sulfur dioxide from generating unitsand the engineering characteristics of units serve asconstraints. Coal–fired capacity can cofire with bio-mass in order to lower operating costs and/or emis-sions.

During off–peak periods, the submodule institutesload following, which is the practice of runningplants near their minimum operating levels ratherthan shutting them down and incurring shutoff andstartup costs. In addition, to account for scheduledand unscheduled maintenance, the capacity of eachplant is derated (lowered) to the expected availabil-ity level. Finally, the operation of utility andnonutility plants for each region is simulated oversix seasons to reflect the seasonal variation in elec-tricity demand.

Interregional economy trade is also represented inthe EFD submodule by allowing surplus generationin one region to satisfy electricity demand in an im-porting region, resulting in a cost savings. Economytrade with Canada is determined in a similar man-ner as interregional economy trade. Surplus Cana-dian energy is allowed to displace energy in an im-porting region if it results in a cost savings. After dis-patching, fuel use is reported back to the fuel supplymodules and operating expenses and revenues fromtrade are reported to the electricity finance and pric-ing submodule.

Electricity Finance and PricingSubmodule

The costs of building capacity, buying power, andgenerating electricity are tallied in the electricity fi-nance and pricing (EFP) submodule, which simu-lates both competitive electricity pricing and thecost–of–service method often used by State regula-tors to determine the price of electricity. While thereis considerable near–term uncertainty, twenty–twoStates and the District of Columbia either had orwere still planning to initiate competitive retail

pricing of electricity at the time the AEO2003 wasprepared. In these States, it is assumed that suchprograms are in effect as of the date established byeach State’s legislation and/or regulations. The re-maining States are assumed to continue traditionalcost-of-service regulation through 2025.

Using historical costs for existing plants(derivedfrom various sources such as Federal Energy Regula-tory Commission (FERC) Form 1, Annual Report ofMajor Electric Utilities, Licensees and Others, andForm EIA–412, Annual Report of Public ElectricUtilities), cost estimates for new plants, fuel pricesfrom the NEMS fuel supply modules, unit operatinglevels, plant decommissioning costs, plant phase–incosts, and purchased power costs, the EFPsubmodule calculates total revenue requirements foreach area of operation—generation, transmission,and distribution—for pricing of electricity in thefully regulated States. Revenue requirementsshared over sales by customer class yield the price ofelectricity for each class. Electricity prices are re-turned to the demand modules. In addition, thesubmodule generates detailed financial statements.

For those States for which it is applicable, EFP alsodetermines competitive prices for electricity genera-tion. Unlike cost–of–service prices, which are basedon average costs, competitive prices are based onmarginal costs. Marginal costs are primarily the op-erating costs of the most expensive plant required tomeet demand. The competitive price also includes areliability price adjustment, which represents thevalue consumers place on reliability of service whendemands are high and available capacity is limited.Prices for transmission and distribution are as-sumed to remain regulated, so the delivered electric-ity price under competition is the sum of the mar-ginal price of generation and the average price oftransmission and distribution.

Load and Demand-Side ManagementSubmodule

The load and demand–side management (LDSM)submodule generates load curves representing thedemand for electricity. The demand for electricityvaries over the course of a day. Many differenttechnologies and end uses, each requiring a differentlevel of capacity for different lengths of time, arepowered by electricity. For operational and planninganalysis, an annual load duration curve, which rep-resents the aggregated hourly demands, is con-structed. Because demand varies by geographic areaand time of year, the LDSM submodule generatesload curves for each region and season.



Emissions

EMM tracks emission levels for sulfur dioxide (SO2)and nitrogen oxides (NOx). Facility development,retrofitting, and dispatch are constrained to complywith the pollution constraints of the Clean Air ActAmendments of 1990 (CAAA90) and other pollutionconstraints. An innovative feature of this legislationis a system of trading emissions allowances. Thetrading system allows a utility with a relatively lowcost of compliance to sell its excess compliance (i.e.,the degree to which its emissions per unit of powergenerated are below maximum allowable levels) toutilities with a relatively high cost of compliance.The trading of emissions allowances does not changethe national aggregate emissions level set by

CAAA90, but it does tend to minimize the overallcost of compliance.

In addition to SO2 and NOX, the EMM also deter-mines mercury and carbon dioxide emissions. It rep-resents control options to reduce emissions of thesefour gases, either individually or in any combination.Fuel switching from coal to natural gas, renewables,or nuclear can reduce all of these emissions. Fluegas desulfurization equipment can decrease SO2 andmercury emissions. Selective catalytic reduction canreduce NOX and mercury emissions. Selectivenon-catalytic reduction and low–NOX burners canlower NOX emissions. Fabric filters, spray cooling,and activated carbon injection can reduce mercuryemissions. Lower emissions resulting from demandreductions are determined in the end-use demandmodules.



The renewable fuels module (RFM) represents re-newable energy resoures and large–scale technolo-gies used for grid-connected U.S. electricity supply(Figure 11). Since most renewables (biomass, con-ventional hydroelectricity, geothermal, landfill gas,solar photovoltaics, solar thermal, and wind) areused to generate electricity, the RFM primarily in-teracts with the electricity market module (EMM).

New renewable energy generating capacity is eithermodel–determined or based on surveys or other pub-lished information. A new unit is only included insurveys or acccepted from published information if itis reported to or identified by the Energy Informa-tion Administration and the unit meets EIA criteriafor inclusion (the unit exists, is under construction,under contract, is publicly declared by the vendor, oris mandated by state law, such as under a state re-newable portfolio standard). EIA may also assumeminimal builds for reasons based on historical expe-rience (floors). The penetration of grid-connected re-newable energy generating technologies, with theexception of landfill gas, is determined by the EMM.

Each renewable energy submodule of the RFM istreated independently of the others, except for theirleast-cost competition in the EMM. Because vari-able operation and maintenance costs for renewabletechnologies are lower than for any other major gen-erating technology, and because they generally pro-duce little or no air pollution, all available renewablecapacity, except biomass, is assumed to be dis-patched first by the EMM. Because of its potentiallysignificant fuel cost, biomass is dispatched accordingto its variable cost by the EMM.

The near–term costs of renewable energy technolo-gies can increase due to infrastructure constraints ifnew capacity is projected to increase by greater than50 percent a year nationally. With significantgrowth over time, costs in the longer term are alsoassumed to be higher because of growing constraintson the availability of sites, natural resource degrada-

tion, and the need to upgrade existing transmissionor distribution networks.

Geothermal-Electric Submodule

The geothermal-electric submodule provides theEMM the amounts of new geothermal capacity thatcan be built at 51 individual sites, along with relatedcost and performance data. The information is ex-pressed in the form of a three–step supply functionthat represents the aggregate amount of new capac-ity and associated costs that can be offered in eachyear in each region.

Geothermal resource data are based on Sandia Na-tional Laboratory’s 1991 geothermal resource as-sessment. Only hydrothermal (hot water and steam)resources are considered. Hot dry rock resources arenot included, because they are not expected to be eco-nomically accessible during the NEMS forecast hori-zon.

Capital and operating costs are estimated sepa-rately, and life-cycle costs are calculated by theRFM. The costing methodology incorporates the ef-fects of Federal and State energy tax constructionand production incentives (if any).

Wind-Electric Submodule

The wind-electric submodule projects the availabil-ity of wind resources as well as the cost and perfor-mance of wind turbine generators. This informationis passed to EMM so that wind turbines can be builtand dispatched in competition with other electricitygenerating technologies. The wind turbine data areexpressed in the form of energy supply curves thatprovide the maximum amount, capital cost, and ca-pacity factor of turbine generating capacity thatcould be installed in a region in a year, given theavailable land area and wind speed.

Solar-Electric Submodule

The solar-electric submodule represents both photo-voltaic and high-temperature thermal electric (con-

RENEWABLE FUELS MODULE


RFM Outputs Inputs from NEMS Exogenous Inputs

Energy production capacities

Capital costs

Operating costs (including wood supply prices for the

wood submodule)

Capacity factors

Available capacity

Biomass fuel costs

Biomass supply curves

Installed energy production capacity


Population

Interest rates

Avoided cost of electricity

Discount rate

Capacity additions

Biomass consumption

Site-specific geothermal resource quality

data

Site-specific wind resource quality data

Plant utilization (capacity factor)

Technology cost and performance

parameters

Landfill gas capacity

centrating solar power) installations. Only cen-tral-station, grid-connected applications constructedby a utility or independent power producer are con-sidered in this portion of the model.

The solar-electric submodule provides the EMMwith time-of-day and seasonal solar availability datafor each region, as well as current costs. The EMM

uses this data to evaluate the cost and performanceof solar-electric technologies in regional grid applica-tions. The commercial and residential demand mod-ules of NEMS also model photovoltaic systems in-stalled by consumers, as discussed in the demandmodule descriptions under “Distributed Genera-tion.”



Cost Characteristics,Capacity Factor,

Available Capacity

Avoided Cost,Installed Capacity,

Discount Rate

Cost Characteristics,Time Segment Capacity

Factors, Available Capacityby Wind Class

Capacity Additions


Installed Capacity


Fuel Cost,Available Capacity

Biomass Consumption


DemandModules

BiomassConsumption

Gross DomesticProduct, Population,

Interest Rates


GeothermalElectric

Submodule

Wind-ElectricSubmodule

Solar-ElectricSubmodule

Cost Characteristics,Time Segment Capacity

Factors, Available Capacity

RenewableFuels

ModuleExogenous

Technological Characteristics,Capital Costs, OperatingCosts, Resource Data,Landfill Gas Capacity

Landfill GasSubmodule

BiomassFuels

Submodule PetroleumMarketModule

Fuel Cost forBiomass from

Cellulose

-

Figure 11. Renewable Fuels Module Structure

Landfill Gas Submodule

The landfill gas submodule provides annual projec-tions of electricity generation from methane fromlandfills (landfill gas). The submodule uses thequantity of municipal solid waste (MSW) that is pro-duced, the proportion of MSW that will be recycled,and the methane emission characteristics of threetypes of landfills to produce forecasts of the futureelectric power generating capacity from landfill gas.The amount of methane available is calculated byfirst determining the amount of total waste gener-ated in the United States. The amount of total wastegenerated is derived from an econometric equationthat uses gross domestic product and population asthe forecast drivers. It is assumed that no new massburn waste–to–energy (municipal solid waste) facili-ties will be built and operated during the forecast pe-riod in the United States. It is also assumed that op-erational mass-burn facilities will continue to oper-ate and retire as planned throughout the forecast pe-riod. The landfill gas submodule passes cost and per-formance characteristics of the landfill gas–to–elec-tricity technology to the EMM for capacity planningdecisions. The amount of new landfill-gas-to-elec-

tricity capacity competes with other technologies us-ing supply curves that are based on the amount ofhigh, medium, and low methane producing landfillslocated in each EMM region.

Biomass Fuels Submodule

The biomass fuels submodule provides biomass-firedplant technology characterizations (capital costs, op-erating costs, capacity factors, etc.) and fuel informa-tion for EMM, thereby allowing biomass-fueledpower plants to compete with other electricity gen-erating technologies.

Biomass fuel prices are represented by a supplycurve constructed according to the accessibility ofresources to the electricity generation sector. Thesupply curve employs resource inventory and costdata for four categories of biomass fuel - urban woodwaste and mill residues, forest residues, energycrops, and agricultural residues.28 Fuel distributionand preparation cost data are built into thesecurves. The supply schedule of biomass fuel prices iscombined with other variable operating costs associ-ated with burning biomass. The aggregate variablecost is then passed to EMM.



28Urban Wood Waste and Mill Residues: Antares Group, Inc; Forest and Crop Residues: Oak Ridge National Laboratory;Energy Crops: Oak Ridge Energy Crop County Level Database (December 20, 1996); and Agricultural Residues: OakRidge National Laoboratory.

The oil and gas supply module (OGSM) consists of aseries of process submodules that project the avail-ability of:

• Domestic crude oil production and dry naturalgas production from onshore, offshore, andAlaskan reservoirs

• Imported pipeline–quality gas from Mexicoand Canada

• Imported liquefied natural gas.

The OGSM regions are shown in Figure 12.

The driving assumption of OGSM is that domestic oiland gas exploration and development are under-taken if the discounted present value of the recov-ered resources at least covers the present value oftaxes and the cost of capital, exploration, develop-ment, and production. Crude oil is transported to re-fineries, which are simulated in the petroleum mar-ket module, for conversion and blending into refinedpetroleum products. The individual submodules ofthe oil and gas supply module are solved independ-ently, with feedbacks achieved through NEMS solu-tion iterations (Figure 13).

Technological progress is represented in OGSMthrough annual increases in the finding rates andsuccess rates, as well as annual decreases in costs.For conventional onshore, a time trend was used ineconometrically estimated equations as a proxy fortechnology. Reserve additions per well (or findingrates) are projected through a set of equations thatdistinquish between new field discoveries and dis-coveries (extensions) and revisions in known fields.The finding rate equations capture the impacts oftechnology, prices, and declining resources. Anotherrepresentation of technology is in the success rateequations. Success rates capture the impact of tech-nology and saturation of the area through cumula-tive drilling. Technology is further represented inthe determination of drilling, lease equipment and

operating costs. Technological progress puts down-ward pressure on the drilling, lease equipment, andoperating cost projections. For unconventional gas, aseries of eleven different technology groups are rep-resented by time–dependent adjustments to factorswhich influence finding rates, success rates, andcosts.

Lower 48 Onshore and ShallowOffshore Supply Submodule

The lower 48 onshore supply submodule projectscrude oil and natural gas production from conven-tional recovery techniques. This submodule accountsfor drilling, reserve additions, total reserves, produc-tion-to-reserves ratios for each lower 48 onshore sup-ply region.

The basic procedure is as follows:

• First, the prospective costs of a representativedrilling project for a given fuel category andwell class within a given region are computed.Costs are a function of the levels of drillingactivity, average well depth, rig availabilityand the effects of technological progress.

• Second, the present value of the discountedcash flows (DCF) associated with therepresentative project is computed. Thesecash flows include both the capital andoperating costs of the project, includingroyalties and taxes, and the revenues derivedfrom a declining well production profile,computed after taking into account theprogressive effects of resource depletion andvalued at constant real prices as of the year ofinitial valuation.

• Third, drilling levels are calculated as afunction of projected profitability as measuredby the projected DCF levels for each projectand national level cashflow.

OIL AND GAS SUPPLY MODULE


OGSM Outputs Inputs from NEMS Exogenous Inputs

Crude oil production

Domestic and Canadian nonassociated natural

gas supply curves

Mexican and liquefied natural gas imports

and exports

Cogeneration from oil and gas production

Reserves and reserve additions

Drilling levels

Associated-dissolved gas production

Domestic and Canadian

natural gas production and

wellhead prices

Crude oil demand

World oil price

Electricity price

Gross Domestic Product

Inflation Rate

Resource levels

Initial finding rate parameters and costs

Production profiles

Tax parameters

Mexican natural gas consumption and

capacities

Liquefied natural gas costs and capacities

• Fourth, regional finding rate equations areused to forecast new field discoveries from newfield wildcats, new pools and extensions fromother exploratory drilling, and reserverevisions from development drilling.

• Fifth, production is determined on the basis ofreserves, including new reserve additions,previous productive capacity, flow from newwells, and, in the case of natural gas, fueldemands. This occurs within the marketequilibration of the natural gas transmissionand distribution module (NGTDM) for naturalgas and within OGSM for oil.

Unconventional Gas Recovery SupplySubmodule

Unconventional gas is defined as gas produced fromnonconventional geologic formations, as opposed toconventional (sandstones) and carbonate rock forma-tions. The three nonconventional geologic forma-tions considered are low–permeability or tight sand-stones, gas shales, and coalbed methane.

For unconventional gas, a play–level model calcu-lates the economic feasibility of individual playsbased on locally specific wellhead prices and costs,resource quantity and quality, and the various ef-fects of technology on both resources and costs. Ineach year, an initial resource characterization deter-mines the expected ultimate recovery (EUR) for thewells drilled in a particular play. Resource profilesare adjusted to reflect assumed technological im-pacts on the size, availability, and industry knowl-edge of the resources in the play. Subsequently,prices received from NGTDM and endogenously de-termined costs adjusted to reflect technologicalprogress are utilized to calculate the economic profit-ability (or lack thereof) for the play. If the play isprofitable, drilling occurs according to an assumedschedule, which is adjusted annually to account fortechnological improvements, as well as varying eco-nomic conditions. This drilling results in reserve ad-ditions, the quantities of which are directly related tothe EURs for the wells in that play. Given these re-serve additions, reserve levels and expected produc-tion–to–reserves (P/R) ratios are calculated at boththe OGSM and the NGTDM region level. The resul-tant values are aggregated with similar values from



Rocky Mountain

West Coast

Pacific

Midcontinent

Southwest

Northeast

Gulf Coast

Gulf of MexicoOffshore

North Slope

OtherAlaska

North SlopeOnshore

Atlantic

Figure 12. Oil and Gas Supply Module Regions

the conventional onshore and offshore submodules.The aggregate P/R ratios and reserve levels are thenpassed to NGTDM, which determines the prices andproduction for the following year through marketequilibration.

Offshore Supply Submodule

This submodule uses a field–based engineering andeconomic analysis approach to project reserve addi-tions and production from resources in the shallowand deep water offshore Gulf of Mexico Outer Conti-nental Shelf and Pacific regions. Two structural com-ponents make up the offshore supply submodule, anexogenous price/supply data generation routine andan endogenous reserves and production timing algo-rithm.

The price/supply data generation methodology em-ploys a rigorous field–based DCF approach. Thisoffline model utilizes key field properties data, algo-rithms to determine key technology components, al-

gorithms to determine the exploration, developmentand production costs, and computes a minimum ac-ceptable supply price (MASP) at which the dis-counted net present value of an individual prospectequals zero. The MASP and the recoverable re-sources for the different fields are aggregated byplanning region and by resource type to generate re-source–specific price–supply curves. In addition tothe overall supply price and reserves, costs compo-nents for exploration, development drilling, produc-tion platform, and operating expenses, as well as ex-ploration and development well requirements, arealso carried over to the endogenous component.

After the exogenous price/supply curves have beendeveloped, they are transmitted to an endogenous al-gorithm. This algorithm makes choices for field ex-ploration and development based on relative eco-nomics of the project profitability compared with theequilibrium crude oil and natural gas prices deter-mined by the petroleum market module and natural



NEMS

Oil and Gas

Supply Module

Foreign

Natural Gas

Supply

Submodule

Resource Levels,Finding Rates,

Costs, ProductionProfiles, Taxes

Exogenous

Exogenous

Natural Gas

Transmission

and Distribution

Module

Industrial

Demand Module

Electricity Market

ModuleElectricity Price

Domestic

Oil and Gas

Supply

Submodules

Petroleum

Market Module

Natural Gas

Mexican and Liquefied

Associated-DissolvedGas Production

Supply Curves forDomestic and

Canadian Natural Gas

Crude OilProduction

Crude Oil Demand

Combined Heatand Power

Domestic and CanadianNatural Gas Production and

Expectations

International EnergyModule

World Oil Price

Wellhead Prices

Imports and Exports

Natural Gas

Mexican and Liquefied

Imports and Exports

Figure 13. Oil and Gas Supply Module Structure

gas transmission and distribution module. Develop-ment of economically recoverable resources intoproved reserves is constrained by drilling activity.Proved reserves are translated into production basedon a P/R ratio. The drilling activity and the P/R ratioare both determined by extrapolating the historicalinformation.

Alaska Oil and Gas Submodule

This submodule projects the crude oil and naturalgas produced in Alaska. The Alaskan oil submoduleis divided into three sections: new field discoveries,development projects, and producing fields. Oiltransportation costs to lower 48 facilities are used inconjunction with the relevant market price of oil tocalculate the estimated net price received at thewellhead, sometimes called the netback price. A dis-counted cash flow method is used to determine theeconomic viability of each project at the netbackprice.

Alaskan oil supplies are modeled on the basis of dis-crete projects, in contrast to the onshore lower 48conventional oil and gas supplies, which are modeledon an aggregate level. The continuation of the explo-ration and development of multiyear projects, aswell as the discovery of new fields, is dependent onprofitability. Production is determined on the basisof assumed drilling schedules and production pro-files for new fields and developmental projects, his-torical production patterns, and announced plans forcurrently producing fields.

Alaskan gas production is set separately for any gastargeted to flow through a pipeline to the lower 48States and gas produced for consumption in theState and for export to Japan. The latter is set basedon a forecast of Alaskan consumption in the NGTDMand an exogenous specification of exports. NorthSlope production for the pipeline is dependent onconstruction of the pipeline, set to commence if thelower 48 average wellhead price is maintained at alevel exceeding the established comparable cost ofdelivery to the lower 48 States.

Foreign Natural Gas SupplySubmodule

The foreign natural gas supply submodule (FNGSS)establishes production in the Western CanadianSedimentary Basin (WCSB) and Eastern Canada,natural gas trade via pipeline with Mexico, as wellas liquefied natural gas (LNG) trade. The receivingregions for foreign gas supplies correspond to thoseof the natural gas integrating framework estab-lished for NGTDM. Within NGTDM, pipeline natu-

ral gas imports flow from two sources: Canada andMexico. U.S. natural gas trade with Canada is repre-sented by seven entry/exit points, and trade withMexico is represented by three entry/exit points (Fig-ure 14).

OGSM provides NGTDM with the begin-ning–of–year natural gas proved reserves from theWCSB and an associated expected production–to–re-serve ratio. NGTDM uses this information to estab-lish a short–term supply curve for the region. Alongwith exogenously specified forecasts for exports ofgas to Canada, other Canadian supplies, and Cana-dian consumption, this supply curve is used to deter-mine the wellhead gas production and price in theWCSB and the level and price of imports from Can-ada at the seven border crossings. Based on theWCSB gas wellhead price, OGSM forecasts drillingactivity in the WCSB using an econometrically de-rived equation, along with the associated reserve ad-ditions. The finding rate is set using an assumed ex-ponential decline function which responds to thedrilling activity. The reserve additions are added tothe beginning-of-year proved reserves from the cur-rent forecast year, after the forecasted productionlevels are subtracted, to establish the begin-ning–of–year proved reserves for the next forecastyear. Construction is set to commence on a pipelineto bring natural gas from the MacKenzie Delta tomarket after the lower 48 average wellhead price ismaintained at a level exceeding the established com-parable cost.

Mexican gas trade is a highly complex issue. A rangeof noneconomic factors influences, if not determines,flows of gas between the United States and Mexico.The uncertainty is so great that not only is the mag-nitude of flow for any future year in doubt, but alsothe direction of flow. Reasonable scenarios have beendeveloped and defended in which Mexico may be ei-ther a net importer or exporter of hundreds of bil-lions of cubic feet of gas by 2025 or sooner.

Despite the uncertainty and the significant influenceof noneconomic factors that influence Mexican gastrade with the United States, a methodology to antic-ipate the path of future Mexican imports and ex-ports has been incorporated into FNGSS. This out-look is generated using assumptions regarding re-gional supply and regional/sectoral demand growthfor natural gas in Mexico that have been developedfrom an assessment of current and expected industryand market circumstances as indicated in industryannouncements, or articles or reports in relevantpublications. Excess supply is assumed to be avail-able for export to the United States, and any short-



fall is assumed to be met by imports from the UnitedStates. The importation of liquefied natural gas intoBaja, Mexico is expected to commence when the mar-ket price, established as a netback from the price inCalifornia, exceeds the assumed cost.

Liquefaction is a process whereby natural gas iscooled to minus 260 degrees Fahrenheit, causing it tobe converted from a gas to a liquid. This also reducesits volume significantly, making it possible to trans-port to distant markets. This allows stranded gas, orgas that would otherwise be inaccessible due eitherto lack of nearby markets or lack of pipeline infra-

structure to deliver it to local markets, to be mone-tized.

Costs of producing, liquefying, transporting, andre-gasifying the gas for delivery via pipeline toend-users are input to the FNGSS. The summationsof these values for each location serve as economicthresholds that must be achieved before investmentin expansion at an existing, or construction of a new,LNG project occurs. Imported LNG costs competewith the purchase price of gas prevailing in the vicin-ity of the import terminal. This is a significant ele-ment in evaluating the competitiveness of LNG sup-plies, since LNG terminals vary greatly in their



Elba Island, GA

Cove Point, MD

Everett, MA

Lake Charles, LA

Cook Inlet,

WesternCanadian

SedimentaryBasin

Pacific

CA

AZ/NM

Mountain West NorthCentral

West SouthCentral

EastSouth

Central

SouthAtlantic

East

North Central

Mid-Atlantic

NewEngland

FL

EasternFrontier

Pipeline imports/exports

Existing LNG terminals.

NorthSlope

MacKenzieDelta

Potential LNG terminals.

Figure 14. Foreign Natural Gas Trade via Pipeline and Liquefied Natural Gas Terminals

proximity to domestic producing areas. Terminalsclose to major consuming markets and far from com-peting producing areas may provide a sufficient eco-nomic advantage to make LNG a competitive gassupply source in some markets.

In addition to costs, extensive operational assump-tions are required to determine LNG imports. Domi-nant general factors affecting the outlook include ex-pected developments with respect to the use of exist-ing capacity, expansion at existing sites, and con-struction at additional locations. The LNG forecastalso requires the specification of a combination offactors: available gasification capacity, schedules forand lags between constructing and opening a facility,tanker availability, expected utilization rates, andworldwide liquefaction capacity. For inactive termi-nals, it is necessary to determine the length of timerequired to restart operations, normally between 12and 18 months. These considerations are taken intoaccount when the economic viability of LNG suppliesis determined.

The algorithm for representing LNG regasificationcapacity expansion in the United States comparesestimated costs for bringing LNG into various re-gions in the United States with the average marketprice in the region over the previous three years ofthe forecast. If the market price has been sustained

above the estimated cost, construction of additionalregasification capacity is expected to occur. The re-gions represented are: New England Census Divi-sion, Middle Atlantic Census Division, South Atlan-tic Census Division (excluding Florida), Florida,East South Central Census Division, West SouthCentral Census Division, California, and Washing-ton/Oregon. The incremental expansion volumes arespecified exogenously, along with the expected utili-zation of the capacity across time. Under special cir-cumstances (e.g., rapid consumption growth) theseutilization rates are adjusted endogenously. The as-sumed costs for bringing LNG into the United Statesreflect the least cost aggregation of cost estimates forproduction, liquefaction, transportation, and regasi-fication from potential supply sources to each of thecoastal regions of the United States. Build decisionsoccur under various restrictions, such as the limita-tion that new capacity can not be added in a regionuntil existing capacity has been expanded to a speci-fied limit and all of this capacity is fully utilizedwithin the region. In deciding upon capacity expan-sion, the model does not attempt to anticipate futuremarket situations, factor in regional demand forLNG (except through its indirect impact on prices),nor select between potential regasification sites. Themodel accounts for LNG exports to Japan fromAlaska using an exogenously–specified forecast.



The natural gas transmission and distribution mod-ule (NGTDM) of NEMS represents the natural gasmarket and determines regional market–clearingprices for natural gas supplies and for end–use con-sumption, given the information passed from otherNEMS modules (Figure 15). A transmission and dis-tribution network (Figure 16), composed of nodesand arcs, is used to simulate the interregional flowand pricing of gas in the contiguous United Statesand Canada in both the peak (December throughMarch) and offpeak (April through November) pe-riod. This network is a simplified representation ofthe physical natural gas pipeline system and estab-lishes the possible interregional flows and associatedprices as gas moves from supply sources to end users.

Flows are further represented by establishing arcsfrom transshipment nodes to each demand sectorrepresented in an NGTDM region (residential, com-mercial, industrial, electric generators, and trans-portation). Mexican exports and net storage injec-tions in the offpeak period are also represented asflow exiting a transshipment node. Similarly, arcsare also established from supply points into a trans-shipment node. Each transshipment node can haveone or more entering arcs from each supply sourcerepresented: U.S. or Canadian onshore or U.S. off-shore production, liquefied natural gas imports, sup-plemental gas production, gas produced in Alaskaand transported via pipeline, Mexican imports, ornet storage withdrawals in the region in the peak pe-riod. Most of the types of supply listed above are setindependently of current year prices and beforeNGTDM determines a market equilibrium solution.

Only the onshore and offshore lower 48 U.S. andWestern Canadian Sedimentary Basin production,along with net storage withdrawals, are representedby short–term supply curves and set dynamicallyduring the NGTDM solution process. The construc-tion of natural gas pipelines from Alaska and Can-ada’s MacKenzie Delta are triggered when marketprices exceed estimated project costs. The flow of gasduring the peak period is used to establish interre-gional pipeline and storage capacity requirementsand the associated expansion. These capacity levelsprovide an upper limit for the flow during the offpeakperiod.

Arcs between transshipment nodes, from the trans-shipment nodes to end–use sectors, and from supplysources to transshipment nodes are assigned tariffs.The tariffs along interregional arcs reflect reserva-tion (represented with volume dependent curves)and usage fees and are established in the pipelinetariff submodule. The tariffs on arcs to end–use sec-tors represent the interstate pipeline tariffs in theregion, intrastate pipeline tariffs, and distributormarkups set in the distributor tariff submodule.Tariffs on arcs from supply sources represent gather-ing charges or other differentials between the priceat the supply source and the regional market hub.The tariff associated with injecting, storing, andwithdrawing from storage is assigned to the arc rep-resenting net storage withdrawals in the peak pe-riod. During the primary solution process in the in-terstate transmission submodule, the tariffs alongan interregional arc are added to the price at thesource node to arrive at a price for the gas along thearc right before it reaches its destination node.


NATURAL GAS TRANSMISSIONAND DISTRIBUTION MODULE

NGTDM Outputs Inputs from NEMS Exogenous Inputs

Natural gas end–use and electric generator

prices

Domestic and Canadian natural gas

wellhead prices

Domestic natural gas production

Canadian natural gas imports and production

Lease and plant fuel consumption

Pipeline fuel use

Pipeline and distribution tariffs

Interregional natural gas flows

Storage and pipeline capacity expansion

Supplemental gas production

Natural gas demands

Domestic and Canadian

natural gas supply curves

Mexican and liquefied natural

gas imports and exports

Macroeconomic variables

Associated-dissolved natural gas

production

Historical consumption patterns

Historical flow patterns

Rate design specifications

Company-level financial data

Pipeline and storage capacity and utilization data

Historical end-use prices

State and Federal tax parameters

Pipeline and storage expansion cost data

Supplemental gas production


NATURAL GAS TRANSMISSION AND DISTRIBUTION MODULE

NEMS


Oil and GasSupplyModule


DemandModules


Exogenous

Historical Consumption and

Company-Level Financial Data,Pipeline and Storage Capacity,

Utilization and Cost Data,State and Federal Tax Parameters,

Supplemental Gas Production

Flow Patterns, RateDesign Specifications,

InterstateTransmissionSubmodule

DistributorTariff

Submodule

Natural Gas Transmission andDistribution Module

ExpansionCost Curves(Exogenous)

PipelineTariff

Submodule

DistributorMarkups

Lease and Plant FuelConsumption,

Pipeline Fuel Consumption

MacroeconomicVariables

Natural Gas Prices,Production, andIndustry Activity

Domestic and CanadianNatural Gas

Supply Curves

Associated-DissolvedGas Production

Mexican Gas Importsand Exports

Liquefied Natural GasImports and Exports

Domestic andCanadian Natural Gas

Wellhead Prices

Domesticand Canadian Natural Gas

Production

Natural Gas Pricesand Production

Natural Gas Core andNoncore Demand

Natural Gas End-Use Prices

Peak and Offpeak,Core and Noncore

Demand

Natural Gas Prices toElectric Generators

InterstateTariffsCurves

Interstate Flowsand Capacity

Expansion

Figure 15. Natural Gas Transmission and Distribution Module Structure

Interstate Transmission Submodule

The interstate transmission submodule (ITS) is themain integrating module of NGTDM. One of its ma-jor functions is to simulate the natural gas price de-termination process. ITS brings together the majoreconomic factors that influence regional natural gastrade on a seasonal basis in the United States, thebalancing of the demand for and the domestic supplyof natural gas, including competition from importednatural gas. These are examined in combinationwith the relative prices associated with moving thegas from the producer to the end user where andwhen (peak versus offpeak) it is needed. In the pro-cess, ITS simulates the decision–making process forexpanding pipeline and/or seasonal storage capacityin the U.S. gas market, determining the amount of

pipeline and storage capacity to be added between orwithin regions in NGTDM. Storage serves as the pri-mary link between the two seasonal periods repre-sented.

ITS employs an iterative heuristic algorithm, alongwith an acyclic hierarchical representation of theprimary arcs in the network, to establish a marketequilibrium solution. Given the consumption levelsfrom other NEMS modules, the basic process fol-lowed by ITS involves first establishing the back-ward flow of natural gas in each period from the con-sumers, through the network, to the producers,based primarily on the relative prices offered for thegas from the previous ITS iteration. This process isperformed for the peak period first since the netwithdrawals from storage during the peak period



Figure 16. Natural Gas Transmission and Distribution Module Network

will establish the net injections during the offpeakperiod. Second, using the model’s supply curves,wellhead prices are set corresponding to the desiredproduction volumes. Also, using the pipeline andstorage tariffs from the pipeline tariff submodule,pipeline and storage tariffs are set corresponding tothe associated flow of gas, as determined in the firststep. These prices are then translated from the pro-ducers, back through the network, to the city gateand the end users, by adding the appropriate tariffsalong the way. A regional storage tariff is added tothe price of gas injected into storage in the offpeak toarrive at the price of the gas when withdrawn in thepeak period. This process is then repeated until thesolution has converged. Finally, end–use prices arederived for residential, commercial, and transporta-tion customers, as well as for both core and noncoreindustrial and electric generation sectors using thedistributor tariffs provided by the distributor tariffsubmodule.

Pipeline Tariff Submodule

The pipeline tariff submodule (PTS) provides usagefees and volume dependent curves for computingunitized reservation fees (or tariffs) for interstatetransportation and storage services within ITS.These curves extend beyond current capacity levelsand relate incremental pipeline or storage capacityexpansion to corresponding estimated rates. The un-derlying basis for each tariff curve in the model is aforecast of the associated regulated revenue require-ment. Econometrically estimated forecasting equa-tions within a general accounting framework areused to track costs and compute revenue require-ments associated with both reservation and usagefees under current rate design and regulatory sce-narios. Other than an assortment of macroeconomicindicators, the primary input to PTS from other mod-ules in NEMS is the level of pipeline and storage ca-pacity expansions in the previous forecast year.

Once an expansion is forecast to occur, PTS calcu-lates the resulting impact on the revenue require-ment. PTS assumes rolled–in (or average), not in-cremental, rates for new capacity. The pipeline tariffcurves generated by PTS are used within the ITSwhen determining the relative cost of purchasingand moving gas from one source versus another inthe peak and offpeak seasons.

Distributor Tariff Submodule

The distributor tariff submodule (DTS) sets distribu-tor markups charged by local distribution companiesfor the distribution of natural gas from the city gateto the end user. End–use distribution service is dis-tinguished within DTS by sector (residential, com-mercial, industrial, electric generators, and trans-portation), season (peak and offpeak), and servicetype (core and noncore). DTS sets distribution tariffsby estimating or assuming the annual change inthese tariffs, starting from base–year values that areestablished using historical data.

The annual change in distributor tariffs for residen-tial, commercial, and industrial core customers de-pends on an assumed increase in operational effi-ciencies combined with a depreciation rate, as wellas the annual change in natural gas consumptionand in national average capital and employmentcosts. Distributor markups to the noncore industrialcustomers are assumed not to change over time. Dis-tributor markups to electric generators are allowedto change in response to annual changes in consump-tion levels within the sector. The natural gas vehiclesector markups are calculated separately for fleetand personal vehicles. Markups for fleet vehicles areset and held constant at historical levels, with taxesadded. Markups for personal vehicles are set at theindustrial sector core price, plus taxes, plus an as-sumed distribution cost. This price is capped at thegasoline equivalent price, as long as minimum costsare covered.



The petroleum market module (PMM) represents do-mestic refinery operations and the marketing of pe-troleum products to consumption regions. PMMsolves for petroleum product prices, crude oil andproduct import activity (in conjunction with the in-ternational energy module and the oil and gas sup-ply module), and domestic refinery capacity expan-sion and fuel consumption. The solution is derived,satisfying the demand for petroleum products andincorporating the prices for raw material inputs andimported petroleum products, the costs of invest-ment, and the domestic production of crude oil andnatural gas liquids. The relationship of PMM toother NEMS modules is illustrated in Figure 17.

PMM is a regional, linear–programming representa-tion of the U.S. petroleum market. Refining opera-tions are represented by a three–region linear pro-gramming formulation of the five Petroleum Admin-istration for Defense Districts (PADDs) (Figure 18).PADDs I and V are each treated as single regions,while PADDs II, III, and IV are aggregated into oneregion. Each region is considered as a single firmwhere more than 40 distinct refinery processes aremodeled. Refining capacity is allowed to expand ineach region, but the model does not distinguish be-tween additions to existing refineries or the buildingof new facilities. Investment criteria are developedexogenously, although the decision to invest is en-dogenous.

PMM assumes that the petroleum refining and mar-keting industry is competitive. The market will movetoward lower–cost refiners who have access to crudeoil and markets. The selection of crude oils, refineryprocess utilization, and logistics (transportation)will adjust to minimize the overall cost of supplyingthe market with petroleum products.

Although the petroleum market responds to pres-sure, it rarely strays from the underlying refiningcosts and economics for long periods of time. If de-mand is unusually high in one region, the price willincrease, driving down demand and providing eco-nomic incentives for bringing supplies in from otherregions, thus restoring the supply/demand balance.

Existing regulations concerning product types andspecifications, the cost of environmental compliance,and Federal and State taxes are also modeled. PMMincorporates taxes imposed by the 1993 Budget Rec-onciliation Act as well as costs resulting from theClean Air Act Amendments of 1990 (CAAA90) andother environmental legislation. The costs of produc-ing new formulations of gasoline and diesel fuel as aresult of the CAAA90 are determined within the lin-ear–programming representation by incorporatingspecifications and demands for these fuels.

An important innovation in NEMS involves the rela-tionship between the domestic and internationalmarkets. Whereas earlier models postulated entirelyexogenous prices for oil on the international market(the world oil price), NEMS includes an interna-tional energy module that estimates supply curvesfor imported crude oils and products based on,among other factors, U.S. participation in interna-tional trade.

Regions

PMM models U.S. crude oil refining capabilitiesbased on the five PADDs which were establishedduring World War II and are still used by EIA fordata collection and analysis. The use of PADD datapermits PMM to take full advantage of EIA’s histori-cal database and allows analysis and forecastingwithin the same framework used by the petroleumindustry.


PETROLEUM MARKET MODULE

PMM Outputs Inputs from NEMS Exogenous Inputs

Petroleum product prices

Crude oil imports and exports

Crude oil demand

Petroleum product imports and exports

Refinery activity and fuel use

Ethanol demand and price

Combined heat and power (CHP)

Natural gas plant liquids production

Processing gain

Capacity additions

Capital expenditures

Revenues

Petroleum product demand by sector

Domestic crude oil production

World oil price

International crude oil supply curves

International product supply curves

International oxygenates supply curves

Natural gas prices

Electricity prices

Natural gas production

Macroeconomic variables

Biomass supply curves

Coal prices

Processing unit operating parameters

Processing unit capacities

Product specifications

Operating costs

Capital costs

Transmission and distribution costs

Federal and State taxes

Agricultural feedstock quantities and costs

CHP unit operating parameters

CHP unit capacities



N

E

M

S

Petroleum

Market

ModuleOil and Gas

Supply

Module

Natural Gas

Transmission

and

Distribution

Module

N

E

M

S

Exogenous

Expectations

Crude Oil Production

Crude Oil Demand

Natural Gas Prices

and Production

Demand

Modules

Electricity

Market

Module

International

Energy

Module

Macroeconomic

Activity

Module

PetroleumDemand

Petroleum Prices

Electricity Prices

PetroleumDemands

World Oil Price

Domestic OilProduction and Demand

Product Imports,Exports

Crude Oil andProduct Supply Curves

Oxygenate SupplyCurves and Prices

Refinery Activity

Interest Rates, GrossDomestic

Product, Deflators

Refinery FuelUse, CHP

Petroleum Prices

Processing Unit Parametersand Capacities, Product Specifications,

Operating and Capital Costs,Transmission and Distribution Costs,Federal and State Taxes,Agricultural Feedstock Quantities and

Capacities

Oil Production

Petroleum Prices

Coal Market

ModuleDiesel Fuel

Prices

RenewableFuels Module Biomass Supply

Curves

Costs, CHP Unit Parameters and

Figure 17. Petroleum Market Module Structure

IVII

III

IV

Figure 18. Petroleum Administration for Defense Districts

Product Categories

Product categories, specifications, and recipe blendsmodeled in PMM include the following:

Fuel Use

PMM determines refinery fuel use by refining regionfor purchased electricity, natural gas, distillate fuel,residual fuel, liquefied petroleum gas, and other pe-troleum. The fuels (natural gas, petroleum, othergaseous fuels, and other) consumed within the refin-ery to generate electricity from CHP facilities arealso measured.

Crude Oil Categories

Both domestic and imported crude oil are aggregatedinto five categories, as defined by the followingranges of gravity and sulfur:

This aggregation of crude oil types allows PMM to ac-count for changes in crude oil composition over time.A composite crude with the appropriate yields andqualities is developed for each category by averagingcharacteristics of specific crude oil streams that fallinto each category. While the domestic and foreigncategories are the same, the composites for each type

may differ, because different crude oil streams makeup the composites.

Refinery Processes

Not every refinery processing unit is represented inPMM. The refinery processes represented were cho-sen because they have the most significant impact onproduction. The following distinct processes are rep-resented:



Motor gasoline: conventional, oxygenated (2.7% oxygen),

Federal reformulated (2.0% oxygen), California reformulated

(no oxygen).

Jet fuels: kerosene-based.

Distillates: kerosene, heating oil, low-sulfur (500 ppm) and

ultra–low–sulfur(15 ppm) highway diesel.

Residual fuels: low-sulfur, high-sulfur.

Liquefied petroleum gas (LPG): propane, LPG mixes.

Petrochemical feedstocks: petrochemical naphtha,

petrochemical gas oil, propylene, aromatics.

Other: asphalt and road oil, still gas, petroleum coke,

lubricating products and waxes, special naphthas.

Petroleum Products Modeled in PMM

Category Sulfur Gravity

Low-sulfur light 0.5% >24

Medium-sulfur heavy 0.35-1.1% >24

High-sulfur light >1.1% >32

High-sulfur heavy >1.1% 24-33

High-sulfur very heavy >0.7% <23

Crude Oil Categories in PMM

Atmospheric distillation, vacuum distillationDelayed cokerFluid coker-includes flexicoking modeFluid catalytic cracker (FCC)-includes distillate, vacuum gas oil, cokergas oil, atmospheric residual, unallowable feeds; conversion ranges 65to 85 percent; ZSM catalyst mode, low severity mode

Gas oil hydrocracker (including advanced technologies)Residuum hydrocrackerNaphtha hydrotreaterDistillate desulfurizationFCC feed hydrofinerResiduum desulfurizer (including atmospheric)Lube and wax unitsModule distillate deep hydrotreaterSolvent deasphaltingCatalytic reforming-separate units for semi-regenerative high-pressure,low-pressure cyclic; severity range as appropriate to the reactor; allowlight straight run through heavy naphtha virgin streams, heavynaphthas from FCC, coker, and hydrocracker operations; allow newhighly active catalyst operation

Naphtha splitterFCC gasoline fractionationFCC naphtha desulfurizationC2-C5 dehydrogenationButane splitterButane isomerizationIsomerization for pentanes, hexanesIsomarization - butane, butylene, buteneAlkylationAromatics recoveryDiesel hydro desulfurization (S-Zorb and others)Coal–to–liquids facility, including cogenerationGas–to–liquids process, including Syntroleum, Shell middle distillateIso-OctaneOlifin hydrogenationFCC olefins to gasoline and distillateThermal Cracker, including C2-C4 feed, Naphtha feed, Gas Oil feed,

VisbreakerHydrocracker— naphtha, partial, low conversionHigh density jet fuel pre/processing— prefractionation, hydrotreatingPolymerization of prolylene, butyleneMTBE, ETBE, and TAME production (captive and merchant)Gas processingHydrogen generation, purificationSteam generationPower generationCogenerationSulfur plantFuel mixerMethanol production

ETBE=Ethyl tertiary butyl ether.MTBE=Methyl tertiary butyl ether.TAME= Tertiary amyl methyl ether.

Refinery Processing Units Modeled in PMM



Natural Gas Plants

The outputs of natural gas processing plants, includ-ing ethane, propane, butane, isobutane, and naturalgasoline are modeled in PMM. These products movedirectly into the market to meet demand or are in-puts to the refinery.

Ethanol

PMM contains an ethanol submodule which providesthe PMM linear program with regional ethanol sup-plies and prices. Ethanol quantity/price curves arecalculated in each Census Division for bothcorn–derived and cellulose–derived ethanol, allow-ing PMM to forecast transportation ethanol demand.The supply curves take into account feedstock costs,feedstock conversion costs, and energy prices. Cornand corn co–product quantities and costs are pro-vided exogenously from the USDA AgriculturalBaseline Projections to 2011.29 Cellulose feedstocksupply/price curves are provided by the renewablefuels module of NEMS.

End-Use Markups

The linear–programming portion of the model pro-vides unit prices of products sold in the refinery re-gions (refinery gate) and in the demand regions(wholesale). End–use markups are added to producea retail price for each of the Census Divisions. Themarkups are based on an average of historical mark-ups, defined as the difference between the end–useprices by sector and the corresponding wholesaleprice for that product. The average is calculated us-ing data from 1989 to the present. Because of thelack of any consistent trend in the historical end–usemarkups, the markups remain at the historical aver-age level over the forecast period.

State and Federal taxes are also added to transpor-tation fuel prices to determine final end–use prices.Recent tax trend analysis indicates that State taxes

increase at the rate of inflation, while Federal taxesdo not. In PMM, therefore, State taxes are held con-stant in real terms throughout the forecast whileFederal taxes are deflated at the rate of inflation.

Gasoline Types

Federal and State legislations have resulted in theproduction of several blends of gasoline. PMM cate-gorizes these blends into four gasoline types:conventional gasoline, oxygenated gasoline, Federalreformulated gasoline, and California reformulatedgasoline. The conventional category includes gaso-line blended with 10–percent ethanol, also known asgasohol. Oxygenated gasoline is conventional gaso-line containing a minimum of 2.7–percent oxygen byweight for use in specific regions of the United Statesduring the winter months to reduce carbon monox-ide.

Federal reformulated gasoline is blended accordingto U.S. Environmental Protection Agency (EPA)Complex Model II specifications with a minimum ox-ygen content of 2.0 percent by weight for use in ozonenon–attainment areas. PMM uses either ethanol orethers (MTBE, ETBE, and other ethers) to obtain the2.0 percent oxygen requirement. However, after2004, the model uses only ethanol in Census Division9 to make Federal reformulated gasoline, because ofCalifornia legislation which bans the use of MTBE ingasoline by the end of 2003. California reformulatedgasoline is blended to the California Air ResourcesBoard (CARB) specifications, which are more severethan the Federal reformulated standards, but haveno minimum oxygen requirement. Because abouttwo–thirds of California’s gasoline consumption oc-curs within Federal ozone nonattainment areas, gas-oline in these areas is also assumed to meet the Fed-eral oxygen requirement of 2.0 percent. Althoughthe reference case assumes current laws and regula-tions, additional product specifications can be mod-eled for policy analysis.

29 U.S. Department of Agriculture, USDA Agricultural Baseline Projections to 2011, Staff Report WAOB-2002-01 (Washington,DC, February 2002).

Ultra–Low–Sulfur Diesel

In December 2000, EPA promulgated the “ul-tra-low-sulfur diesel” (ULSD) regulation for high-way diesel. By definition, ULSD is highway dieselthat contains no more than 15–ppm sulfur at thepump. The new regulation contains the “80/20” rule,which requires the production of 80 percent ULSDand 20 percent 500 ppm highway diesel betweenJune 2006 and June 2010, and a 100–percent re-quirement for ULSD thereafter. Because NEMS isan annual average model, the full impact of the 80/20rule cannot be seen until 2007, and the impact of the100– percent requirement cannot be seen until 2011.Major assumptions related to the implementation ofthe ULSD rule are as follows:

• Highway diesel at the refinery gate willcontain a maximum of 7–ppm sulfur.Although sulfur content is limited to 15 ppmat the pump, there is a general consensus thatrefineries will need to produce diesel below 10ppm sulfur in order to allow for contamini-nation during the distribution process.

• The amount of ULSD downgraded to a lowervalue product because of sulfur contaminationin the distribution system is assumed to be 10percent at the onset of the program, decliningto 4.4 percent at full implementation. Thedecline reflects an expectation that thedistribution system will become more efficientat handling ULSD with experience.

• Demand for highway-grade diesel, both 500and 15 ppm combined, is assumed to beequivalent to the total transportationdistillate demand. Historically, highway-grade diesel supplied has nearly matched totaltransportation distillate sales, although somehighway-grade diesel has gone to nontrans-portation uses such as construction andagriculture.

• ULSD production is modeled throughimproved distillate hydrotreating units aswell as the ConocoPhillips S–Zorb process.Revamping (retrofitting) existing units toproduce ULSD will be undertaken byrefineries representing two–thirds of highwaydiesel production; the remaining refinerieswill build new units. The capital cost of a

revamp is assumed to be 50 percent of the costof adding a new unit.

• No change in the sulfur level of non-roaddiesel is assumed, because the EPA has notyet promulgated non-road diesel standards.

Gas–To–Liquids and Coal–To–Liquids

If prices for lower sulfur distillates reach a highlevel, it is assumed that gas-to-liquids (GTL) facili-ties will be built on the North Slope of Alaska to con-vert stranded natural gas into distillates, to betransported on the Trans-Alaskan Pipeline System(TAPS) to Valdez and shipped to markets in thelower 48 States. The facilities are assumed to bebuilt incrementally, no earlier than 2005, with out-put volumes of 50,000 barrels per day, at a cost of$21,500 per barrel of daily capapcity (2001 dollars).Operating costs are assumed to be $3.99 per barrel.Transportation costs to ship the GTL product fromthe North Slope to Valdez along the TAPS rangefrom $2.75 to $4.45 per barrel, depending on total oilflow on the pipeline and the potential need for GTLto maintain the viability of the TAPS line if Alaskanoil production declines. Initially, the natural gas feedis assumed to cost $0.82 per million cubic feet (2001dollars).

It is also assumed that coal-to-liquids (CTL) facilitieswill be built when low sulfur distillate prices arehigh. One CTL facility is capable of processing16,400 tons per day bituminous coal and will have aproduction capacity of 33,200 barrels per day of syn-thetic fuels and 696 MW of CHP electric generatingcapacity. CTL facilities are assumed to be built nearexisting refineries. For the East Coast, potentialCTL facilities are assumed to be built near the Dela-ware River basin; for the Central region, near the Il-linois River basin or near Billings, Montana; and forthe West Coast, in the vicinity of Puget Sound inWashington State. The CTL yields are assumed tobe similar to those from a GTL facility because bothinvolve the Fischer-Tropsch process to convertsyngas (CO + H2) to liquid hydrocarbons. The pri-mary yields would be distillate and kerosene, withadditional yields of naphthas and LPG. Petroleumproduct production from CTL’s is assumed to be com-petitive when distillate prices rise above the cost ofCTL production – adjusted for credits from the saleof electricity co-products.



The coal market module (CMM) represents the min-ing, transportation, and pricing of coal, subject toend–use demand. Coal supplies are differentiated byheat and sulfur content. CMM also determines theminimum cost pattern of coal supply to meet exoge-nously defined U.S. coal export demands as a part ofthe world coal market. Coal supply is projected on acost–minimizing basis, constrained by existing con-tracts. Twelve different coal types are differentiatedwith respect to thermal grade, sulfur content, andunderground or surface mining. The domestic pro-duction and distribution of coal is forecast for 13 de-mand regions and 11 supply regions (Figures 19 and20).

The CMM components are solved simultaneously.The sequence of solution among components can besummarized as follows. Coal supply curves are pro-duced by the coal production submodule and input tothe coal distribution submodule. Given the coal sup-ply curves, distribution costs, and coal demands, thecoal distribution submodule projects delivered coalprices. The module is iterated to convergence withrespect to equilibrium prices to all demand sectors.The structure of the CMM is shown in Figure 21.

Coal Production Submodule

This submodule produces annual coal supply curves,relating annual production to minemouth prices.The supply curves are constructed from an econo-metric analysis of prices as a function of productivecapacity, capacity utilization, and costs. A separatesupply curve is provided for surface and under-ground mining for all significant production by coalrank (bituminous, subbituminous and lignite), coalgrade (steam or metallurgical), and sulfur level ineach supply region. Constructing curves for the coaltypes available in each region yields a total of 36curves that are used as inputs to the coal distribu-

tion submodule. Supply curves are updated for eachyear in the forecast period.

The factors accounted for in constructing the supplycurves are labor productivity and the costs of factorinputs (mining equipment, mine labor, and fuel). La-bor productivity projections are developed and ap-plied to each supply curve, based on historical data.The projections incorporate an assumption that therate of improvement will decline as the rate of tech-nology penetration slows. Labor costs are tied to la-bor productivity and wage rates. It is assumed, in thereference case, that wage rates keep pace with infla-tion.

Coal Distribution Submodule

The coal distribution submodule is a linear programthat determines the least–cost supplies of coal for agiven set of coal demands by demand region and sec-tor, accounting for transportation costs from the dif-ferent supply curves, heat and sulfur content, exist-ing coal supply contracts, technical limitations ofolder boiler types, and sulfur allowance costs underthe Clean Air Act Amendments of 1990. Existingsupply contracts between coal producers and utili-ties are incorporated in the model as minimum flowsbetween specific supply curves and region–sulfurlevel combinations. The minimum flows are gener-ally assumed to remain in effect for the duration ofthe contract and then be replaced by market–deter-mined flows.

Coal transportation costs are simulated using inter-regional coal transportation costs derived by sub-tracting reported minemouth costs for each supplycurve from reported delivered costs for each demandtype in each demand region. Transportation ratesare assumed to change in response to railroad laborproductivity, labor costs, diesel fuel costs, and equip-ment costs. In recent years, railroad rates have de-


COAL MARKET MODULE

CMM Outputs Inputs from NEMS Exogenous Inputs

Coal production and distribution

Minemouth coal prices

End-use coal prices

Coal exports

Transportation rates

Coal quality by source, destination, and end-use sector

World coal flows

Coal demand

Interest rates

Price indices and deflators

Diesel fuel prices

Electricity prices

Base year productive capacity, capacity

utilization, prices, and coal quality parameters

Contract quantities

Labor productivity

Labor costs

Labor cost escalators

Domestic transportation costs

International transportation costs

International supply curves

International coal import demands

Demand for U.S. coal imports

clined because of operating efficiencies from suchmeasures as improved scheduling and lower fuel costper ton–mile that have resulted from low crude oilprices, more efficient diesel engines, and larger andlighter aluminum cars.

Coal Export Component

The coal export component of the coal distributionsubmodule projects quantities of coal imported andexported from the United States. The quantities aredetermined within a world trade context, based onassumed characteristics of foreign coal supply anddemand. The component disaggregates coal into 16export regions and 20 import regions, as shown onthe following page.

The export component is a part of the linear programthat optimizes domestic coal supply. It determinesworld coal trade distribution by minimizing overallcosts for coal, subject to U.S. coal supply prices and anumber of constraints. Supply costs (mining andpreparation plus transportation) for each coal exportregion, coal type, and end use compete in two de-mand sectors (coking and steam). The componentalso incorporates within the model structure supplydiversity constraints that reflect the observed ten-dency of coal–importing countries to avoid excessivedependence upon one source of supply, even at asomewhat higher cost.


COAL MARKET MODULE

1. NE

2. YP

6. EN

7. KT

3. SA

5. OH9. CW

4. GF8. AM10. WS

12. ZN13. PC

11. MT

RegionCode Region Content

1

2

3

4

5

6

7

New England

Middle Atlantic

South Atlantic

Georgia and Florida

Ohio

East North Central

Kentucky and Tennessee

RegionCode Region Content

8

9

10

11

12

13

Alabama and Mississippi

West North Central

West South Central

Montana, Wyoming, and Idaho

Mountain

Pacific

Figure 19. Coal Market Module Demand Regions


COAL MARKET MODULE

RIWI

L

W A

ID

MT

NE

OR

UT

CANV

AZ NM

CO

WY

TXOK

KS

AR

LA

MO

IA

SD

ND

MN

IL MI

INOH

KY

TN

MS AL GASC

NC

VA

PA

NYVT

NH

MA

NJ

DE

ME

AK

FL

B

B

2001000100

SCALE IN MILES

MI

MD

CT

4002000200

SCALE IN MILES

WV

1. Northern Appalachia 4. Eastern Interior 7. Dakota Lignite

2. Central Appalachia 5. Western Interior 8. Powder andGreen RiverBasins

3. Southern Appalachia 6. Gulf Lignite

9. Rocky Mountain

10. Southwest

11. Northwest

APPALACHIA INTERIOR NORTHERN GREAT PLAINS OTHER WEST

Figure 20. Coal Market Module Supply Regions

Coal Export Regions Coal Import Regions

U.S. East CoastU.S. Gulf CoastU.S. Southwest and WestU.S. Northern InteriorU.S. NoncontiguousAustraliaWestern CanadaInterior CanadaSouth AfricaPolandCIS (Europe)CIS (Asia)

China

Colombia

Indonesia

Venezuela

U.S. East CoastU.S. Gulf CoastU.S. Northern InteriorU.S. NoncontiguousEastern CanadaInterior CanadaScandinaviaUnited Kingdom and IrelandGermany and AustriaOther Northwestern EuropeIberiaItalyMediterranean and Eastern EuropeMexicoSouth AmericaJapan

East Asia

China and Hong Kong

ASEAN (Association of Southeast Asian Nations)

India and South Asia


COAL MARKET MODULE



DemandModules


CoalDistributionSubmodule

CoalProductionSubmodule

Coal Market Module

CoalExport

Component

CoalSupplyCurves

CoalProduction and

Distribution

Coal Prices

Coal Demand

Coal Prices

Coal Demand

Interest Rates,Price Indices, and

Deflators

Coal Productionand Prices

Diesel FuelPrices

ExportDemand

MinemouthCoal Prices

NEMS

Exogenous

Coal Import and Export DemandsInternational Transportation CostsInternational Supply Curves

Exogenous

Domestic Transportation CostsContract Quantities

Exogenous

Base Year Productive CapacityBase Year Capacity UtilizationLabor ProductivityLabor CostsLabor Cost Escalators

Electricity Prices

Figure 21. Coal Market Module Structure

The National Energy Modeling System is docu-mented in a series of model documentation reports,available on the EIA Web site at http://www.eia.doe.gov/bookshelf/docs.html or by contacting the Na-tional Energy Information Center (202/586-8800).

Energy Information Administration, IntegratingModule of the National Energy Modeling System:Model Documentation DOE/EIA-M057(2002) (Wash-ington, DC, December 2001).

Energy Information Administration, Model Docu-mentation Report: Macroeconomic Activity Module(MAM) of the National Energy Modeling System,DOE/EIA-M065(2003) (Washington, DC, February2003).

Energy Information Administration, NEMS Interna-tional Energy Module: Model Documentation Report,DOE/EIA-M071(03) (Washington, DC, April 2003).

Energy Information Administration, EIA Model Doc-umentation: World Oil Refining Logistics DemandModel, “World” Reference Manual, DOE/EIA-M058(Washington, DC, March 1994).

Energy Information Administration, Model Docu-mentation Report: Residential Sector Demand Mod-ule of the National Energy Modeling System,DOE/EIA-M067(2003) (Washington, DC, January2003).

Energy Information Administration, Model Docu-mentation Report: Commercial Sector Demand Mod-ule of the National Energy Modeling System,DOE/EIA-M066(2003) (Washington, DC, March2003).

Energy Information Administration, Model Docu-mentation Report: Industrial Sector Demand Moduleof the National Energy Modeling System,

DOE/EIA-M064(2003) (Washington, DC, February2003).

Energy Information Administration, TransportationSector Model of the National Energy Modeling Sys-tem: Model Documentation Report, DOE/EIA-M070(2003) (Washington, DC, February 2003).

Energy Information Administration, The ElectricityMarket Module of the National Energy Modeling Sys-tem: Model Documentation Report, DOE/EIA-M068(2003) (Washington, DC, March 2003).

Energy Information Administration, Documentationof the Oil and Gas Supply Module (OGSM), DOE/EIA-M063(2003) (Washington, DC, March 2003).

Energy Information Administration, Model Docu-mentation: Natural Gas Transmission and Distribu-tion Model of the National Energy Modeling System,DOE/EIA-M062(2003) (Washington, DC, March2003).

Energy Information Administration, EIA Model Doc-umentation: Petroleum Market Model of the NationalEnergy Modeling System, DOE/EIA-M059(2003)(Washington, DC, February 2003).

Energy Information Administration, Model Docu-mentation: Coal Market Module of the National En-ergy Modeling System, DOE/EIA-M060(2003)(Washington, DC, February 2003).

Energy Information Administration, Model Docu-mentation: Renewable Fuels Module of the NationalEnergy Modeling System, DOE/EIA-M069(2003)(Washington, DC, March 2003).

Energy Information Administration, Annuel EnergyReview 2001, DOE/EIA-0384(2001) (Washington,DC, November 2002).


APPENDIX

BIBLIOGRAPHY

The National Energy Modeling System: An Overview...

Documents

Transcript of The National Energy Modeling System: An Overview...