DAVID CHIEN *
This paper briefly describes and evaluates some of the more important and frequently used models to estimate greenhouse gas emissions by a number of U.S. government agencies. Among the models covered are: National Energy Modeling System (NEMS); MARKAL-MACRO; MiniCAM; Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) Model; and Transitional Alternative Fuels and Vehicles (TAFV) Model. These models have been used by the U.S. Congress and federal agencies to assess U.S. strategies to meet the Kyoto Accord, which would require the United States to maintain U.S. carbon emissions at 7% below 1990 levels between 2008 and 2012. In this paper, each model is described and its capabilities and limitations highlighted. Model perspectives are provided from a user's viewpoint, so that potential users will have a full understanding of the capabilities of these models and the resources needed to build, update, and maintain them.
KEYWORDS: Greenhouse gases, transportation forecasting, modeling.
In December 1997, approximately 160 nations met in Kyoto, Japan, and developed the Kyoto Protocol, which would limit greenhouse gas (GHG) emissions worldwide. In the protocol, the United States agreed to reduce GHG emissions levels to 7% below 1990 levels from 2008 to 2012 (although the U.S. government has not formally ratified the agreement). In considering the impacts of the Kyoto Protocol and other GHG emissions reduction policies or programs, the U.S. Department of Transportation (DOT), the U.S. Department of Energy (DOE), and the U.S. Environmental Protection Agency (EPA) have employed a number of useful models. These models range from pure energy and environmental analytical tools to integrated energy-environmental-economic models that capture the interactions of GHG reduction technologies and policies within the economy. Because such modeling efforts have significant impacts on policy and program decisionmaking, a critical review of those models is important.
This paper describes a number of models used by the U.S. government to better understand the impact of future technologies and policies on U.S. GHG emissions in the transportation sector.1 In particular, the paper focuses on five models:
This paper does not thoroughly evaluate detailed inputs to the models (i.e., data and assumptions). Therefore, the author highly recommends that interested readers refer to model documentation and visit the model websites for more detailed information on inputs and assumptions. Sources for further reading for each model are provided at the end of the paper.
Although this article provides potential users with brief descriptions of the GHG models employed by the U.S. federal government, it also evaluates the models based on key operational factors often overlooked by potential users. Topics discussed here include: the size of the data inputs and sourcecode of the models, the hardware and software platform and requirements, the run time or amount of time associated with execution of the models, the resources needed to develop and maintain the models, and examples of studies that have used the models extensively. Detailed model coverage with respect to the transportation sector is also evaluated. It should be noted that the evaluations contained in this article are a snapshot in time and were up to date at the time of writing. However, the reader should keep in mind that the models receive significant improvements over time.
Most of the models reviewed in this study originated before GHG modeling became prevalent. In fact, almost none of the models were originally designed to measure greenhouse gases, but rather operational aspects, criteria pollutant (nitrogen oxidesNOx, sulfur oxidesSOx, nonmethane hydrocarbonsHC, and particulates) emissions, and energy aspects of modeling and forecasting. Energy forecasting and emissions models are a natural fit for carbon emissions estimation and forecasting. Transportation energy and emissions models usually require two components, travel and fuel efficiency or emissions factors, that are related to the technology and its age. Given these components, fuel consumption can easily be converted into carbon emissions, because the burning of carbon emission compounds is in direct proportion to its consumption.
These models have become very popular and widely used because of their importance in formulating policies designed to reduce carbon emissions. With the ratification of the Kyoto Accord (minus the United States and Australia), many countries had to devise strategies to reduce carbon emissions rapidly. The economic and operational consequences for carbon emissions reduction policies have now emerged as important factors to investigate when countries attempt to reduce carbon emissions to achieve the long-term Kyoto emissions levels.
DOT uses some of these models and integrates data from many sources. Many of the data sources used to develop these models reside at the federal agencies that created and currently maintain the models. Among those used most frequently for estimates of historical GHG emissions are the EPA 2 and EIA 3 models. Transportation and energy-related data can be found on the Transportation Energy Databook website run by ORNL and DOE,4 DOT's Bureau of Transportation Statistics (BTS) website,5 and the BTS TRANSTATS website.6 The TRANSTATS website contains National Transportation Statistics data compiled from BTS surveysOffice of Airline Information databases, the National Household Transportation Survey (NHTS), the Commodity Flow Survey (CFS), and many more.
NEMS is a computer-based energy-economy modeling system of the U.S. energy markets for the midterm period through 2025. NEMS annually projects the production, imports, conversion, consumption, and prices of energy, subject to assumptions on macroeconomic and financial factors, world energy markets, resource availability and costs, behavioral and technological choice criteria, cost and performance characteristics of energy technologies, and demographics. The purpose of NEMS is to project energy, economic, environmental, and security impacts on the United States of alternative energy policies and of different assumptions about energy markets. (USDOE EIA 2003a)
Congress and other federal agencies have used NEMS7 extensively to evaluate energy and transportation policies. The model has the advantage of extensive peer review by the U.S. transportation community including DOE, DOT, EPA, the Office of Management and Budget, the Government Accountability Office, and the National Academy of Sciences.
The structure of NEMS consists of an integrated modeling system representing all demand sectors of the economy (residential, commercial, industrial, and transportation), including a macroeconomic component and all energy supply sources (i.e., crude oil supply; oil refinery; oil distribution; natural gas including exploration, drilling, and distribution; electricity including nuclear, coal, natural gas, residual fuel, and small generators like wind and solar; coal; and renewable fuels).
The transportation sector is important, because it consumes over 27% of all energy, and approximately 98% of transportation consumption comes from petroleum use (USDOE 2002a). The NEMS Transportation Demand Module (TRAN) provides wide coverage of the aggregate transportation system including the following submodules: Light-Duty Vehicles (LDV), Aviation, Freight Transport (truck, rail, waterborne), Miscellaneous (transit, recreational boats, aviation gasoline), and Emissions (USDOE EIA 2003a).
The LDV Submodule covers 6 areas: Fuel Economy6 car and 6 light-truck EPA size classes across 63 advanced subsystems and fuel savings technologies; Regional Sales9 Census Divisions; Alternative Fuel Vehicles12 types of vehicles; LDV Stockvehicle retirement curves and capital stocks by 20 vintages and vehicle types; Vehicle-Miles Traveled (VMT)by car and light truck as a function of income per capita and the cost of driving per mile; and LDV Fleet for business, government, and utility fleets (as part of the Energy Policy Act).
The Aviation Submodule includes two components. The Air Travel Demand Submodule forecasts revenue passenger-miles for international and domestic travel, revenue ton-miles for freight, and seat-miles demanded. The Aircraft Fleet Efficiency Submodule covers six fuel-saving technologies for regional, narrow, and wide-body aircraft: ultra-high bypass, propfan, improved thermodynamics, hybrid laminar flow, improved aerodynamics, and weight reduction. The submodule also contains 48 vintages of aircraft with aircraft survival curves and stock model representation.
The Freight Transport Submodule includes truck, rail, and waterborne. The Freight Truck Submodule uses macroeconomic gross outputs by North American Industrial Classification System (NAICS) code in determining VMT. The CFS and the Vehicle Inventory and Use Survey are used extensively to establish the connection between commodities and mode of travel. The truck stock model determines capital stocks by three truck size classes (Class 3, Classes 4 through 6, and Classes 7 and 8) and by vehicle age (20 vintages). Technology choice is based on future emissions standards, commercial availability, fuel prices, capital cost, efficiency improvement, and other cost-effectiveness criteria such as discount rates and payback period. There are 32 advanced subsystem and emissions control technologies (Argonne 1999). Gasoline, diesel, natural gas, and liquid petroleum gas are the fuels represented in the Truck Submodule.
Rail and Waterborne Submodules use ton-miles traveled estimated equations based on industrial output by NAICS code. Energy efficiency for old and new vehicles is estimated. A major drawback of the model is the lack of capital stocks and vintaging by age. Therefore, the growth rates of efficiency improvements must be made exogenously based on trends rather than an explicit endogenous calculation of the model. Specific technology representation and turnover cannot be endogenously determined, which limits the effect of advanced technologies over time, unless of course the modeler pre-determines this in the exogenous input file. Overall, the Rail and Waterborne Submodules have no sensitivity to fuel prices or the cost of travel in either the travel or efficiency forecasts.
The Miscellaneous Submodule includes mass transit, which covers six transit modes: three types of passenger railtransit, commuter, and intercity; and three types of passenger busestransit, intercity, and school. Travel is estimated for all six transit modes as a function of the relative historical growth rate of passenger-miles of travel relative to light-duty vehicle passenger-miles. Growth rates of efficiency improvements are calculated based on the growth rates of similar technology modes. This assumes that technology advancements will parallel those in modes using the same vehicles. For example, mass transit rail efficiencies would then be assumed to grow at the same rate as Class I freight rail. Therefore, the same caveats from the rail and waterborne models apply to the Mass Transit Module, because both lack explicit model responsiveness to fuel prices and travel costs.
TRAN also has a module that forecasts emissions of the criteria pollutants SOx, NOx, HC, carbon (CO), and carbon dioxide (CO2). Most recently, TRAN incorporated the EPA Mobile 6.0 model, which is used by EPA and several state governments to calculate regional emissions. CO2 and total CO emissions can be calculated by fuel type and by transportation mode, which allows the user to associate various policies or investments with an increase or decrease in carbon emissions.
Finally, the Macroeconomic Activity Module (MAM) currently consists of the Global Insight (formerly DRI/WEFA) Model of the U.S. economy, the Industry Model, the Employment Model, and the Regional Model. MAM uses the input-output (I-O) National Accounts data (from the Bureau of Economic Analysis of the U.S. Department of Commerce). One issue in using the I-O accounts data is that they undercount the effects of the transportation system on the economy, due to the exclusion of almost all private commercial businesses, which have their own private transportation and are currently counted under commercial operations. A potential improvement to the model would be to adjust the I-O data using the BTS Transportation Satellite Accounts (TSA). The TSA measures the private transportation associated with commercial operations to provide more detailed data for the I-O accounts.
Despite these issues, the MAM is a key element for measuring the impacts of potential GHG strategies on the economy. This makes it one of the most important components of NEMS, because it is essential to the convergence process and it fully integrates the economy with the modeling process, which many of the other GHG models reviewed in this article do not. Reaching equilibrium in a model of this size is of paramount importance, especially because feedback effects of prices on transportation services have a tendency to be dampened significantly when macroeconomic feedback with the rest of the model is turned on. What does this tell us? The conclusion is that models that do not have this capability can overstate the effects of any given policy that may be implemented, because they do not account for economic changes and responses to those changes. Reaching equilibrium is critical to the accuracy in measuring costs and benefits of any policy or program.
The NEMS model operates at a Census region and Census division level. Therefore, extrapolation and interpolation are needed to subdivide the estimates down to the state level. Local- or county-level forecasts are not applicable to the model. TRAN does not explicitly account for modal switching (shifting from one mode to another). Policies designed to shift ridership from one mode to another are currently not measurable nor easy to implement. The travel equations for most modes do use many of the same economic variables, which will result in simultaneous modal switching, but each equation contains a different set of variables that affect travel. Some modes (e.g., rail, waterborne, and all modes of transit) are not technology-based nor do they contain stock models, which make technology policy options limited for those modes of travel.
One of the drawbacks to using the model is also one of NEMS' greatest strengths. The size of the entire NEMS model is very large and detailed, requiring over 10 to 15 megabytes (MB) of storage just for the "restart file," which contains the starting values for the model each year. In order to do a "standalone" run, which consists of running only one module and keeping the others at reference case levels, would require 100 MB of storage space. Although NEMS can be installed on an individual personal computer (PC), the storage requirements are substantial. Hardware should consist of 512 MB of random access memory (RAM) and a 486 or Pentium processor. The model operates in Compaq Visual FORTRAN and requires the EViews software. If the user wanted to run the supply models also, then OML, a linear programming software, is also necessary.
When running in standalone mode with only one module endogenously active, the transportation module will return a solution within a minute. However, submitting a fully integrated run with all of the modules turned on or active would take about two to four hours depending on how many changes were made to the model. The current NEMS model at EIA employs approximately 40 full-time employees and 4 full-time contractors. Therefore, enhancing, updating, and maintaining the model requires significant resources. However, several agencies and national laboratories work with versions of the NEMS models and usually employ two to four people to operate and maintain the model for their uses. These NEMS model clones require EIA updates on an annual basis, which are posted on EIA's website at http://www.eia.doe.gov.
The MARKAL-MACRO Model 8 at DOE is an integration of two models, MARKAL and MACRO. MARKAL is the "bottom-up" technological model of energy and the environment, which includes depletable and renewable natural resources, processing of energy resources, and end-use technologies to meet the projected energy service demands in all sectors. MACRO is the "top-down" macroeconomic growth model that links MARKAL to the economy and maximizes utility (discounted sum of consumption). Top-down refers to models that are usually more aggregate in nature and estimate by forecasting a particular variable as a function of other aggregate causal factors. Bottom-up modeling approaches are more detailed at the individual equipment level and then sum up to the total in order to forecast variables.
MARKAL-MACRO finds the least-cost dynamic equilibrium under specific market and policy assumptions. DOE calibrates the MARKAL-MACRO Model to the NEMS outputs annually. The MARKAL-MACRO Model, used by over 45 countries, was developed by Brookhaven National Laboratory and then further improved by 18 Organization for Economic Cooperation and Development (OECD) countries.
MARKAL-MACRO optimizes the mix of fuels and technologies based on the consumer discount rate, technology characteristics, and fuel prices. Marginal costs for technologies and applications are used to determine the most efficient level of energy inputs along with technology costs and energy efficiencies. The model forecasts emissions sources and levels for CO2, SOx, NOx, and any user-specified pollutants and wastes. The value of carbon rights (marginal cost of emissions) is one of the important outputs of the model. Outputs are solved in five-year intervals through 2050. Transportation coverage includes passenger cars, light trucks, heavy trucks, buses, airplanes, shipping, passenger rail, and freight rail.
The model can output a business-as-usual energy and carbon emissions profile. Identification of dynamic technology paths to meet emissions growth targets is one of the more common uses of the model outputs. The MARKAL-MACRO Model has facilitated the study by many countries of the costs of alternative approaches to reducing CO2 emissions. Policy options would include fuel switching, substitution of capital and/or labor for energy services, demand reduction, emissions taxes, etc. MARKAL-MACRO can also identify opportunities for reducing CO2 emissions through supply and demand technologies. Based on the technologies chosen, the model can calculate the cost of CO2 emissions reductions.
The MARKAL-MACRO Model has proved useful in a number of areas. DOE has used it to analyze the Energy Policy Act of 1992. EIA has also built an international version of the MARKAL Model called SAGE to generate the annual International Energy Outlook. EPA is developing a national MARKAL database and scoping out a regional MARKAL representation of the U.S. economy. The MARKAL family of models is used by over 45 countries to support energy and environmental planning. The International Energy Agency (IEA) has a version of the Global MARKAL Model that they use to look into future energy technology perspectives. Most recently the model has focused on externalities measurement, hydrogen economy development, cost-competitive life cycle analysis, oil market response, technology learning, and country analysis.
There are some limitations of MARKAL-MACRO. While it can provide an alternative and complimentary approach in longer term analysis (e.g., projection of renewable fuel penetration and reduction of CO2 emissions), the model does not cover as much detail in all sectors as the NEMS model. The MARKAL-MACRO Model uses a simple approach to forecast energy service demands based on economic indicators such as housing stocks, commercial floor space, industrial production index, and VMT. Modeling at the individual equipment level would be difficult and require off-line analysis combined with aggregate implementation in MARKAL-MACRO. Individual sector modeling is relatively aggregate and may also require similar off-line analysis.
The data inputs to the model use about 7 to 20 MB of storage space, and the sourcecode is approximately 7 to 10 MB. The model can be run on a Pentium 4 processor with a 2 gigahertz (GHz) processor speed and 256 MB of RAM. Model execution is fairly quick at around five minutes. The model is quite complicated and requires special skills to run, similar to the NEMS model but with fewer people. Generally, two national laboratory analysts use and maintain the model for DOE. The MARKAL-MACRO Model is written in GAMS (General Algebraic Modeling System) programming language.9
The MiniCAM Model,10 maintained by the Pacific Northwest National Laboratory (PNNL) forecasts CO2 and other GHG emissions, and it estimates the impacts on GHG atmospheric concentrations, climate, and the environment. Although the model is a top-down agriculture-energy-economy model, it contains bottom-up assumptions about end-use energy efficiency. MiniCAM Model projections are made through 2100 and, therefore, the model has more futuristic technologies than NEMS. The model outputs forecasts in 15-year increments. Projections cover the entire planet in 14 global regions: the United States, Canada, Western Europe, Australia and New Zealand, Japan, the former Soviet Union, Eastern Europe, China, Southeast Asia, the Middle East, Africa, Latin America, South Korea, and India. Projections for Mexico, Argentina, and Brazil are under development.
MiniCAM is comprised of three larger models: the Edmonds-Reilly-Barns Model (ERB), the Agriculture Land Use Model (AGLU), and the Model for the Assessment of GHG Induced Climate Change (MAGICC). ERB represents the energy/economy/emissions system, including supply and demand of energy, the energy balance, GHG emissions, and long-term trends in economic output. AGLU simulates global land-use change from the production of composite crops, animal products, and forest products, and tracks GHG emissions associated with land use. MAGICC models the atmospheric/climate/sea-level system, which includes a gas cycle, climate, and sea-level model. MAGICC outputs atmospheric composition, radiative forcing, global mean temperature change, and sea-level rise.
Energy supply and demand are calculated in the model. Energy supply of renewable and nonrenewable sources is dependent on resource constraints, behavioral assumptions, and energy prices by region. Energy demand is a function of population, labor productivity, economic activity, technological change, energy prices, and energy taxes and tariffs. Transportation is one of three sectors (residential/commercial and industrial are the other two sectors) covered in the model, and passenger and freight technologies and modes are included. Model inputs consist of total service, service cost, vehicle technologies and their characteristics, price and income elasticities, technical change, percentage of population licensed to drive, and average speeds. The transportation system coverage includes automobiles, light trucks, buses, rail, air, and motorcycles for passenger modes; and trucks, rail, air, ship, pipeline, and motorcycles for freight modes. Seven major energy sources are modeled: oil, gas, coal, biomass, resource-constrained renewables, nuclear, and solar.
Similar to the MARKAL-MACRO, individual equipment policies and certain detailed technology modeling would require off-line analysis and then aggregate implementation within MiniCAM. The model capabilities lie only at the national level and do not extend to the regional or state level. Due to the combination of three models within MiniCAM, the complexity of running the model may require specialized knowledge of the operations. In addition, because the model also contains scientific climate and atmospheric conditions, interpretation of the MAGICC results may require specialized knowledge in climatology. However, the three models within the MiniCAM can be run independently to narrow the analytical focus, and the results from each model can be interpreted separately.
The current MiniCAM Model11 has an executable file size of about 1 MB and the data input files are about the same size. The sourcecode is approximately 903 kilobytes. Run time is approximately 30 seconds on a Pentium 4 (1.7 GHz), depending on the number of scenarios run at one time. MiniCAM can operate on a Pentium III or higher speed processor. FORTRAN is the modeling language using a MicroSoft Visual Studio compiler. However, there is a graphics user interface (GUI) front-end to the model if desired, which requires MicroSoft Acesss and Excel software. With the GUI, the user can run multiple scenarios at once and query, view, and chart results. Currently five or more people use MiniCAM at PNNL, and its maintenance requires two individuals.
The GREET model is intended to serve as an analytical tool for use by researchers and practitioners in estimating fuel-cycle energy use and emissions associated with alternative transportation fuels and advanced vehicle technologies.12
GREET, maintained by Argonne National Laboratory (ANL), provides full fuel-cycle emissions analysis from wells to wheels, which represents emissions from all phases of production, distribution, and use of transportation fuels. Besides the fuel-cycle GREET model (GREET Series 1), there is a vehicle cycle model (GREET Series 2) that simulates emissions and energy use of direct input sources such as vehicle production, disposal, and recycling. ANL is currently finalizing a new version of GREET Series 2 for release. The strength of GREET is that it analytically compares energy use and emissions from vehicle technologies matched with many fuels, especially very advanced alternative fuels, over the entire fuel cycle. Emissions in the model include the following: GHGs (CO2, methane, and nitrous oxide), NO x, HC, CO, sulfur dioxide, and particulate matter. Other versions of the model have also included toxic pollutants, such as formaldehyde, acetaldehyde, 1,3-butadidine, and benzene (Winebrake et al. 2000).
The beauty of GREET is that it has a substantial combination of vehicle technologies and fuel types. GREET contains the following powertrains: conventional, direct injection, spark ignition, compression ignition, hybrid electric vehicles (which can be grid connected or not), electric vehicles, and fuel cell vehicles. Fuel types are also numerous: gasoline (reformulated or nonreformulated), diesel and low sulfur diesel, compressed natural gas, liquefied petroleum gas, liquefied natural gas, dimethyl ether, Fischer-Tropsch (FT) diesel, gaseous and liquid hydrogen, methanol, ethanol, biodiesel, and electricity. The GREET model contains more than 85 fuel production pathways and more than 75 vehicle/fuel combinations. These powertrains and fuel types can be produced from several feedstocks: petroleum, natural gas, flared gas, landfill gas, corn, cellulosic biomass, soybeans, and electricity. GREET is an excellent model to determine individual vehicle emissions and would be valuable in assisting evaluation of new transportation fuels and advanced vehicle technologies. EPA has incorporated GREET into their air emissions MOVES Model.
GREET applies only to light-duty vehicles; however, this does not preclude it from being used for other vehicle types in the future. GREET does not include a vehicle choice model to forecast what people might purchase based on consumer preferences, but GREET output (total fuel-cycle emissions factors) can be used with future vehicle technology projections to get a more complete picture of the environmental impacts of these vehicle populations. Additionally, the model may be used in combination with policy options to reduce emissions and set emissions standards to achieve a goal.
The hardware requirements to run and operate the model are GREETGUI (GREET with a GUI interface or front end), which operates on PCs with Microsoft Windows 2000 or later. Minimum hardware requirements are a Pentium III processor at 166 megahertz (MHz) or higher, at least 64 MB RAM; and at least 30 MB of free space on the hard drive. The recommended hardware profile is a Pentium processor at 400 MHz or higher, 128 MB or more of RAM, 100 MB of free hard disk space or more (Argonne 2001). GREET uses an Excel spreadsheet and Visual Basic. Use of GREET requires installation of MS Excel on a PC. Future plans are to convert it to C language in 2006.
GREET can also be run as a spreadsheet model that takes about 5 MB on an Excel spreadsheet. GREET recently added a Monte Carlo simulation module that stochastically generates a distribution rather than a point estimate. Running the model would normally be almost instantaneous, but for Monte Carlo simulations with Crystal Ball commercial software, run times may be approximately 3 hours. Four people developed and are currently maintaining and running GREET at ANL.
The Transitional Alternative Fuels and Vehicle Model13 represents economic decisions among auto manufacturers, vehicle purchasers, and fuel suppliers, including distribution to end users. The model simulates decisions during a transition from current fuels to alternative fuels and traditional vehicles to advanced technology vehicles. Limited availability of alternative fuels, including refueling infrastructure, and availability of alternative fuel vehicle technologies are interdependent. TAFV assumes retail alternative fuel providers will maximize profits and spread capital costs across outlets to increase availability.
TAFV contains a model for predicting the choice of alternative fuel and alternative vehicle technologies for light-duty motor vehicles. The nested multinomial logit mathematical framework is used to estimate vehicle choice among technologies and fuel type combinations based on consumer preferences and vehicle attributes. Vehicle choice is dependent on prices, fuel availability, and the diversity of vehicle offerings (all endogenous) as well as luggage space, refueling time, vehicle performance, and cargo space (all exogenous parameters). Alternative fuel vehicles have three costs to vehicle manufacturers: capital costs, variable costs, and costs associated with diverse vehicle offerings. Calibration of the model through some key parameters, such as the value of time, the value of fuel availability, and discount rates, is based on existing literature. A spreadsheet model has been developed for calibration and preliminary testing. TAFV includes a range of vehicle- and fuel-related policies, including taxes or subsidies, federal mandates for vehicle acquisition (i.e., policies such as the Low Emission Vehicle Program and the Energy Policy Act). In addition, TAFV tracks GHG emissions from fuel production and vehicles using GREET-based emissions factors.
Limitations of TAFV are that it includes only light-duty vehicles; growth rates in transportation demand and oil and gas prices are exogenous; it is national (U.S.) in scope, omitting regional detail and international trends in vehicle use or GHG emissions; and it assumes competitive behavior under complete foresight.
The model is quite small at 208 kilobytes, but inputs could be a few megabytes of spreadsheet data. The main program is written in the GAMS language. TAFV uses the MINOS5 and CONOPT2 nonlinear optimization solvers. The sourcecode is about 111 kilobytes in GAMS language, but the model requires more than 100 MB to execute. A model run takes approximately 30 to 60 minutes on a Pentium III 1,000 MHz PC to solve for the dynamic market equilibria (endogenous prices and quantities). Work files generated during a run can approach 1 gigabyte. Users should have 128 MB or more memory. TAFV can be run on Windows, Linux, and Unix, depending on which platform the licensed GAMS software resides. Maintenance currently involves a team of two; however, plans in the future are for a team of five over the next two years. TAFV has formed the foundation for an extended hydrogen vehicle transition model under development, HyTrans.
Table 1 contains a short summary of the models reviewed in this study. The appendix provides documentation and more detailed information about each model, especially publications and studies that use the models. The appendix is meant to provide potential model users with a better understanding of how to apply the models to a specific problem and use.
Several very good models are available from which to choose when conducting GHG emissions studies, scenarios, or emissions estimates and forecasts for the transportation sector. Depending on the level of regionality and detail required, the model of choice will vary. Some of the models are more applicable at the aggregate level, such as MiniCAM and MARKAL-MACRO. Others such as NEMS, GREET, and TAFV are very detailed at the technology level. Maintenance, usability, resources, and analytical capabilities should be matched to the model choice.
Argonne National Laboratory. 1999. Heavy- and Medium-Duty Truck Fuel Economy and Market Penetration Analysis for the NEMS Transportation Sector Model, prepared for the U.S. Department of Energy, Energy Information Administration. Washington, DC. August.
______. 2001. Development and Use of GREET 1.6 Fuel Cycle Model for Transportation Fuels and Vehicle Technologies, ANL/ESD/TM163. Center for Transportation Research, Energy Systems Division, Argonne, Illinois. June.
Energy Technology Systems Analysis Program. 2002. ETSAP Newsletter 8(2). August. Available at http://www.etsap.org.
Pacific Northwest National Laboratory, Advanced International Studies Group. 2001. China-Korea-U.S. Economic Environmental Workshop Conference Proceedings, May 2325, 2001. Available at http://www.pnl.gov/aisu/pubs/chinmod2.pdf and http://sedac.ciesin.org/mva/minicam/MCHP.html.
U.S. Department of Energy (USDOE), Energy Information Administration (EIA). 2002a. Emissions of Greenhouse Gases in the United States 2001, DOE/EIA-0573(2002). Washington, DC. December.
______. 2002b. Transportation Energy Databook: Edition 22, ORNL 69-67. Oak Ridge, TN: Oak Ridge National Laboratory. September.
______. 2003a. The National Energy Modeling System: An Overview 2003, DOE/EIA-0581(2003). Washington, DC. March.
______. 2003b. The Transportation Sector Model of the National Energy Modeling System: Model Documentation Report, DOE/EIA-M070(2003). Washington, DC. February.
U.S. Department of Transportation, Center for Climate Change and Environmental Forecasting. 2003. Transportation Greenhouse Gas Emissions Data & Models: Review and Recommendations. Cambridge, MA: USDOT, Volpe National Transportation Systems Center. March.
U.S. Environmental Protection Agency. 2002. The U.S. Greenhouse Gas Inventory: In Brief, EPA 430-F-02-008. Washington, DC. April.
Winebrake, J.J., H. Dongquan, and M. Wang. 2000. Fuel-Cycle Emissions for Conventional and Alternative Fuel Vehicles: An Assessment of Air Toxics, ANL/ESD-44. Argonne, Illinois: Argonne National Laboratory, Center for Transportation Research, Energy Systems Division. August.
General Topics of Energy-Related
MARKAL-MACRO was used for a project on "Policies and Measures for Common Action" conducted by the Annex I Expert Group on the United Nations (UN) Framework Convention on Climate Change.15
As part of a study by the OECD Secretariat on the environmental implications of energy and transport subsidies, the Italian participant used an "elastic" version of MARKAL to evaluate the impact of removing financial subsidies from the electric sector in Italy. The many ways in which financial interventions affect the electric supply industry were searched out, and MARKAL was used to assess their effect on electric and energy system costs and CO2 emissions.
The Energy Technology Systems Analysis Programme (ETSAP) of IEA continues to provide a multinational capability to determine the most cost-effective national choices to limit future emissions of greenhouse gases by using consistent methodology that offers a basis for international agreement on abatement measures. The basic MARKAL Model continues to serve national interests, as illustrated by its use for a major national research and development (R&D) appraisal in the United Kingdom, its use to help develop the national least-cost energy strategy in the United States, and its acceptance by a wider international community. Outside ETSAP, MARKAL was used in Taiwan and (in the form of MENSA) in Australia to inform the debate on response strategies under the UN Framework Convention on Climate Change.
With the cooperation of the participants from Italy, Japan, the United Kingdom, and the United States, ETSAP contributed to the IEA study, "Electricity and the Environment." Detailed descriptions were provided of technologies available for electricity supply and demand in the short and medium term, including technical performance and engineering costs. Specific data were drawn from the MARKAL databases of the four cooperating countries.
Although a common set of runs among the ETSAP participants was delayed, four countries participated in CHALLENGE, a cooperative international project on energy and environment systems analysis. CHALLENGE consists of a network of scientists from Eastern and Western European countries. The project is intended to facilitate international negotiations and cooperation by providing a scientific basis for decisions on response strategies to reduce environmental stresses and climate risks due to energy use.
During Annex V, some participating countries provided inputs to major international studies by IEA, OECD, and the Annex I Expert Group on the UN Framework Convention on Climate Change.
ETSAP originated as an IEA program to help establish energy technology R&D priorities on the basis of the needs of all the IEA countries. A common methodology and comparable databases have been the touchstone of the program since its very beginnings. The standard MARKAL Model has continued to be the focus of the group's analyses, and recurring efforts have been made to assure reasonable consistency in the national databases.
The major applications of the GREET Model (reports available at www.transportational.gov) consist of the following:
• Numerous energy-related studies for Congress or federal agencies:
U.S. Department of Energy, Energy Information Administration, Analysis of Corporate Average Fuel Economy (CAFÉ) Standards for Light Trucks and Increased Alternative Fuel Use, SR/OIAF/2002-05 (Washington, DC: March 2002).
______. Analysis of Efficiency Standards for Air Conditioners, Heat Pumps, and Other Products, SR/OIAF/2002-01 (Washington, DC: February 2002).
______. Analysis of Strategies for Reducing Multiple Emissions from Power Plants: Sulfur Dioxide, Nitrogen Oxides, and Carbon Dioxide, SR/OIAF2000-05 (Washington, DC: December 2002).
______. Impact of Renewable Fuel Standard/MTBE Provisions of S. 1766, SR/OIAF/2002-06 (Washington, DC: March 2002).
______. Impact of Renewable Fuel Standard/ MTBE Provisions of S. 517: Addendum, SR/OIAF/2002-06 (Washington, DC: April 2002).
______. Impacts of a 10-Percent Renewable Portfolio Standard, SR/OIAF/2002-03 (Washington, DC: February 2002).
______. Impacts of the Kyoto Protocol on U.S. Energy Markets & Economic Activity, SR/OIAF/98-03 (Washington, DC: October 2002).
______. Measuring Changes in Energy Efficiency for the Annual Energy Outlook 2002 (Washington, DC: 2002).
______. Reducing Emissions of Sulfur Dioxide, Nitrogen Oxides, and Mercury from Electric Power Plants, SR/OIAF/2001-04 (Washington, DC: September 2001).
______. Strategies for Reducing Multiple Emissions from Electric Power Plants with Advanced Technology Scenarios, SR/OIAF/2001-05 (Washington, DC; October 2001).
• Carbon and vehicle emissions modeling for Congress, EPA, and DOE:
Interlaboratory Working Group on Energy-Efficient and Low-Carbon Technologies, Scenarios for a Clean Energy Future (Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, Pacific Northwest National Laboratory, National Renewable Energy Laboratory, and Argonne National Laboratory), ORNL/CON-476 and LBNL-44029 (Oak Ridge, TN: November 2000).
______. Scenarios of U.S. Carbon Reductions: Potential Impacts of Energy-Efficient and Low-Carbon Technologies by 2010 and Beyond (Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, Pacific Northwest National Laboratory, National Renewable Energy Laboratory, and Argonne National Laboratory) (Oak Ridge, TN: September 1997).
U.S. Department of Energy, Energy Information Administration, Analysis of the Climate Change Technology Initiative: Fiscal Year 2001, prepared for the U.S. House of Representatives Committee on Science, SR/OIAF/2000-01 (Washington, DC: April 2000).
______. Analysis of the Climate Change Technology Initiative, prepared for the U.S. House of Representatives Committee on Science, SR/OIAF/99-01 (Washington, DC: April 1999).
______. Analysis of the Impacts of an Early Start for Compliance with the Kyoto Protocol, prepared for the U.S. House of Representatives Committee on Science, SR/OIAF/99-02 (Washington, DC: July 1999).
______. Impacts of the Kyoto Protocol on U.S. Energy Markets and Economic Activity, prepared for the U.S. House of Representatives Committee on Science, SR/OIAF/98-03 (Washington, DC: October 1998).
______. Service Report: Analysis of Carbon Stabilization Cases, prepared for the U.S. Department of Energy, Office of Policy and International Affairs, SR-OIAF/97-01 (Washington, DC: October 1997).
• Transportation-specific model runs for the White House and other governmental agencies:
U.S. Department of Energy, Energy Information Administration, Analysis of Corporate Average Fuel Economy (CAFÉ) Standards for Light Trucks and Increased Alternative Fuel Use, SR/OIAF/2002-05 (Washington, DC: March 2002).
______. The Transition to Ultra-Low-Sulfur Diesel Fuel: Effects on Prices and Supply, SR/OIAF/2001-01 (Washington, DC: May 2001). This study evaluates EPA's ultra-low-sulfur diesel fuel regulations for heavy-duty trucks at the request of the House Committee on Science.
______. The Impacts of Increased Diesel Penetration in the Transportation Sector, prepared by the Office of Integrated Analysis and Forecasting (Washington, DC: August 1998).
Edmonds, J. and J. Reilly, Global Energy: Assessing the Future (New York, NY: Oxford University Press, 1985).
Edmonds, J.A., J.M. Reilly, R.H. Gardner, and A. Brenkert, "Uncertainty in Future Global Energy Use and Fossil Fuel CO2 Emissions 1975 to 207," TR036, DO3/NBB-0081 Dist. Category UC-11, prepared for the U.S. Department of Commerce (Springfield, VA: National Technical Information Service, 1986).
Edmonds, J.A., M.A. Wise, and C.N. MacCracken, Advanced Energy Technologies and Climate Change: An Analysis Using the Global Change Assessment Model (GCAM), PNL-9798 (Richland, WA: Pacific Northwest Laboratory, 1994).
Richels, R. and J. Edmonds, "The Economics of Stabilizing Atmospheric CO2 Concentrations," Energy Policy 23(415):373, 1995.
Wang, M.Q., GREET 1.5: Transportation Fuel-Cycle Model, Volume 1 (Argonne, IL: Argonne National Laboratory, 1999).
Wang, M.Q. and H.S. Huang, A Full Fuel-Cycle Analysis of Energy and Emissions Impacts of Transportation Fuels Produced from Natural Gas (Argonne, IL: Argonne National Laboratory, 1999).
Wang, M., C. Saricks, and M. Wu, "Fuel Ethanol Produced from Midwest US Corn: Help or Hindrance to the Vision of Kyoto?" Journal of the Air and Waste Management Association 49(7):756772, 1999.
Winebrake, J.J., M.Q. Wang, and D. He, "Toxic Emissions from Mobile Sources: A Total Fuel Cycle Analysis of Conventional and Alternative-Fuel Vehicles," Journal of the Air and Waste Management Association 51(7):10731086, July 2001.
Leiby, P. and J. Rubin, "Transitions in Light-Duty Vehicle Transportation: Alternative Fuel and Hybrid Vehicles and Learning," Transportation Research Record 1842:127134, 2003.
______. "Flexible Greenhouse Gas Emission Banking Systems," Maine Agriculture and Forest Experiment Station Miscellaneous Report, No. 427, April 2002.
______. "Effectiveness and Efficiency of Policies to Promote Alternative Fuel Vehicles," Transportation Research Record 1750:8491, 2001.
______. "The Alternative Fuel Transition: Results from the TAFV Model of Alternative Fuel Use in Light-Duty Vehicles 19962010 Final Report, TAFV Version 1," Maine Agricultural and Forest Experiment Station Miscellaneous Report, No. 417, September 2000.
______. "Dynamic Analysis of Achievable Potential and Costs for Alternative Fuel Vehicles," invited presentation at the International Energy Agency, International Workshop on Technologies to Reduce Greenhouse Gas Emissions, Washington, DC, May 1999.
______. "Sustainable Transportation: Analyzing the Transition to Alternative Fuel Vehicles," Transportation Research Board Circular 492:5482, August 1999.
______. "The Transitional Alternative Fuels and Vehicles Model," Transportation Research Record 1587:1018, 1997.
Leiby, P.N., J. Rubin, and D. Bowman, "Efficacy of Policies to Promote New Vehicle Technologies: Alternative Fuel Vehicles and Hybrid Vehicles," Proceedings of the 25th Annual IAEE International Conference, Aberdeen, Scotland, 2629 June 2000.
Rubin, J. and P. Leiby, "An Analysis of Alternative Fuel Credit Provisions of U.S. Automotive Fuel Economy Standards," Energy Policy 28(9):589602, 2000.
1. Although this article was written by the author, information presented here relies heavily on a more detailed document prepared by Kevin Greene of the Volpe National Transportation Systems Center (USDOT 2003).
3. Available at http://www.eia.doe.gov/env/ghg.html.
4. Available at http://www-cta.ornl.gov/cta/data/Index.html.
5. Available at http://www.bts.gov/.
6. Available at http://www.transtats.bts.gov/.
10. MiniCAM Model contact: Son H. Kim, Pacific Northwest National Laboratory, email@example.com (301-314-6763) or Mariana Vertenstein, firstname.lastname@example.org (303-497-1349); also see http://www.pnl.gov/aisu/pubs/chinmod2.pdf and http://sedac.ciesin.org/mva/minicam/MCHP.html.
11. As of 2005, the version of the MiniCAM Model described here has been superseded by ObjECTS-MiniCAM, a C++ version of the model that incorporates object-oriented programming designs for increased flexibility, maintenance, and modeling detail.
13. TAFV Model contacts: Paul Leiby, email@example.com (865-574-7720) and David Greene, Oak Ridge National Laboratory, and Jonathan Rubin, University of Maine; also see http://pzl1.ed.ornl.gov/altfuels.htm.
Coresponding author: D. Chien, Office of Advanced Studies Bureau of Transportation Statistics, Research and Innovative Technology Administration, U.S. Department of Transportation, 400 Seventh St., SW, Room 3430, Washington, DC 20590. E-mail: David.Chien@dot.gov